<%BANNER%>

Processing Hibiscus Beverage Using Dense Phase Carbon Dioxide

Permanent Link: http://ufdc.ufl.edu/UFE0041848/00001

Material Information

Title: Processing Hibiscus Beverage Using Dense Phase Carbon Dioxide
Physical Description: 1 online resource (151 p.)
Language: english
Creator: Ramirez Rodrigues, Milena
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: dense, hibiscus
Food Science and Human Nutrition -- Dissertations, Academic -- UF
Genre: Food Science and Human Nutrition thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROCESSING OF A HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE By Milena Maria Ramirez Rodrigues August 2010 Chair: Maurice R. Marshall Cochair: Murat O. Balaban Major: Food Science and Human Nutrition Consumer demand for natural beverages with health promoting properties that offer fresh-like sensory attributes and changes in U.S. demographics have created the opportunity for the development of new products that would target new market segments. Hibiscus sabdariffa (family Malvaceae) red calyces are rich in anthocyanins and other phenolic compounds. Fresh and dried hibiscus is used to prepare cold and hot beverages and their preparation includes an extraction step followed by a pasteurization method. Although thermal preservation of foods is effective in reducing microbial loads it can also lead to organoleptic and nutritional changes. Nonthermal processes like dense phase carbon dioxide (DPCD) are an alternative which may help preserve the color, flavor, and nutrients of food. Equivalent cold and hot water conditions were found for anthocyanins extraction of dried hibiscus in this research. Likewise, similar polyphenolic profiles and chemical composition of aroma compounds were observed between fresh and dried hibiscus. Solubility of CO2 in a hibiscus beverage (5.06 g CO2/mL at 31.0 MPa) and optimal processing conditions to inactivate yeasts and molds (Y & M) were 34.5 MPa and 6.5 min. DPCD was a viable technology for processing hibiscus beverage since it extended its shelf life for 14 weeks of refrigerated storage. Quality attributes were maintained during storage. Lower losses of anthocyanins were observed in the DPCD (9%) hibiscus beverage as compared to thermally treatment process (14%) and no major changes in total phenolics content and antioxidant capacity occurred during storage. Changes in hibiscus aroma volatiles during storage did not affect untrained panelists overall likeability of the product. Findings in this research can help in the development and marketing of hibiscus beverage.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Milena Ramirez Rodrigues.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Marshall, Maurice R.
Local: Co-adviser: Balaban, Murat O.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041848:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041848/00001

Material Information

Title: Processing Hibiscus Beverage Using Dense Phase Carbon Dioxide
Physical Description: 1 online resource (151 p.)
Language: english
Creator: Ramirez Rodrigues, Milena
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: dense, hibiscus
Food Science and Human Nutrition -- Dissertations, Academic -- UF
Genre: Food Science and Human Nutrition thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROCESSING OF A HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE By Milena Maria Ramirez Rodrigues August 2010 Chair: Maurice R. Marshall Cochair: Murat O. Balaban Major: Food Science and Human Nutrition Consumer demand for natural beverages with health promoting properties that offer fresh-like sensory attributes and changes in U.S. demographics have created the opportunity for the development of new products that would target new market segments. Hibiscus sabdariffa (family Malvaceae) red calyces are rich in anthocyanins and other phenolic compounds. Fresh and dried hibiscus is used to prepare cold and hot beverages and their preparation includes an extraction step followed by a pasteurization method. Although thermal preservation of foods is effective in reducing microbial loads it can also lead to organoleptic and nutritional changes. Nonthermal processes like dense phase carbon dioxide (DPCD) are an alternative which may help preserve the color, flavor, and nutrients of food. Equivalent cold and hot water conditions were found for anthocyanins extraction of dried hibiscus in this research. Likewise, similar polyphenolic profiles and chemical composition of aroma compounds were observed between fresh and dried hibiscus. Solubility of CO2 in a hibiscus beverage (5.06 g CO2/mL at 31.0 MPa) and optimal processing conditions to inactivate yeasts and molds (Y & M) were 34.5 MPa and 6.5 min. DPCD was a viable technology for processing hibiscus beverage since it extended its shelf life for 14 weeks of refrigerated storage. Quality attributes were maintained during storage. Lower losses of anthocyanins were observed in the DPCD (9%) hibiscus beverage as compared to thermally treatment process (14%) and no major changes in total phenolics content and antioxidant capacity occurred during storage. Changes in hibiscus aroma volatiles during storage did not affect untrained panelists overall likeability of the product. Findings in this research can help in the development and marketing of hibiscus beverage.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Milena Ramirez Rodrigues.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Marshall, Maurice R.
Local: Co-adviser: Balaban, Murat O.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041848:00001


This item has the following downloads:


Full Text





PROCESSING HIBISCUS BEVERAGE
USING DENSE PHASE CARBON DIOXIDE















By

MILENA MARIA RAMIREZ RODRIGUES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010




























2010 Milena Maria Ramirez Rodrigues




























To my parents and sister









ACKNOWLEDGMENTS

First thanks to God. Second I want to thank my parents Aidil and Agustin and my

sister Melissa for their love, dedication, and support and for helping me fulfill my dreams

and achieve what I have in life.

I would like to thank my advisor Dr. Murat Balaban for his guidance, for sharing his

knowledge and experience and for his care. I extend a special acknowledgment to Dr.

Marty Marshall who guided me through the second part of this research for his great

support, advice and care. I also want to thank my other advising committee members,

Dr. Allen Wysocki and Dr. Jose Reyes for their help in my research and Dr. Russell

Rouseff and Charles Sims for all their help and advice in the flavor chemistry and

sensory evaluation parts of my research.

Finally, I want to thank my close friends for helping me out in every way possible

and for the good times shared and to the countless persons who in some way or

another have contributed to my success in the UF Food Science Graduate Program.









TABLE OF CONTENTS

page

A C KNO W LEDG M ENTS ............................... .......................................... ...............

L IS T O F T A B L E S ............ .......... .. ................................................................................ 9

LIST O F FIG U R ES.......................................... .......... 11

A BST RA C T ............................................... ... ............ ............ 14

CHAPTER

1 INTRO DUCTIO N ............... .. ............... ....................... .... .... ........... 16

Ju stifica tio n ........................ ........................................................... .... ........ 16
H ypothe sis ........................ ................. .. .................................. 17
Specific Objectives .............................. ............. ................. 17

2 LIT E R A T U R E R EV IE W ............ .......... ...... .......... .......................................... 19

The Beverage Market ............... .... ....... .. .. ........... .......... ..... 19
Market Performance and Competitive Context of U.S. Ready to Drink Non-
Carbonated Beverages .................. ........ ...... ................................... 20
Consumption and Demographic Trends ..................... ............................... 23
Hibiscus (Hibiscus sabdariffa)......................... ...... ........................... 24
Characteristics and Economic Importance .................................. ............... 24
C om m ercial H ibiscus P roducts..................................................... ............... 27
H hibiscus Lem on B issap.............................................................. 27
Caiita Aguas Frescas (jamaica (hibiscus) flavor).............. ............ 28
Squish Hibiscus Presse ...................... ...... ......... ... .............. 28
Sim ply Hibi ...... ... ..... .... .............. .. .. ...... ............... 28
Composition and Associated Health Benefits........................................ 28
Hibiscus Extraction Process ..................... ........... ....... ........ 34
H ibiscus F lavor............................... ............... 34
Phenolic com pounds ........................ .................. .. .. ............................... 37
Classification ........................................................................ ..... .. ....... .......... 37
Phenolic Compounds Attributes .......... .... ......... ......... .. .............. 38
Contribution to flavor..................... ........ .................... 38
Antioxidant potential............................... .................... 41
Anticarcinogenic action ................ .. ................ ....... ........ 42
Anthocyanins ....................... ..................................... 42
Classification ............... .............................. ............ .. ........ 42
S ta b ility .............................. .............. ...... 4 4
Health Benefits ...... .. ................................................... ........................... 46
Beverage Processing....................................... ............... 46
T herm al P processing ...................... .......................... .... ............................ 47









Flash pasteurization ........ ................... ........ ... ...... 47
In-pack pasteurization ............. .... ............................. 48
Aseptic filling ...... ...................... .................. 48
Chilled distribution........................ ................ ......... 48
Dense Phase CO2 Processing (DPCD)............. ........................................ 48
Mechanisms of microorganisms' inactivation by DPCD............................ 49
Factors affecting microbial inactivation ...... ................ ......................... 50
Solubility of C O 2.................................................................................... 52
DPCD treatment systems ......... ......... .......... ......... ..................... 53
DPCD food applications .............. ........ ... ............................. 54
Coconut water.................................. ............... 60
Sensory Evaluation .............. ...... ..................................... ......... 60
Difference-from-Control-Test......... ........... ......... ........................ 61
Flavor Analysis ................ .. ........ ... ..................... 61
Solid Phase Micro Extraction (SPM E) ............................ ......... ............... 62
Gas chromatography-Olfactometry (GC-O)....... ....... ...... .................. 62

3 EFFECT OF COLD AND HOT WATER EXTRACTION ON THE
PHYSICOCHEMICAL AND PHYTOCHEMICAL PROPERTIES OF HIBISCUS
SABDARIFFA EXTRACTS .............................................. 64

In tro d u c tio n ................... .... .................................................. 6 4
Materials and Methods................................ ............... 65
Extracts Preparation............. .. ................................................ ... ...... 65
pH, Total Solids, and Titratable Acidity.................................. ........ .... 66
Anthocyanin Content, Total Phenolics and Antioxidant Capacity ................. 67
Characterization of Major Polyphenolics .......................... ............... 67
LC-M S identification ............... .........................._.. ..... ......... .. 68
HPLC quantification ...... ........._.......................... 69
S tatistica l A na lysis ........................................................................ ......... 6 9
Results and Discussion.......................... ....... .......... ............... 69
Effect of Extraction Conditions ........... .......... .................... ............... 69
Parameters Correlations........................................... ............... 71
Polyphenolics Identification ...................... ...... .......... ............... 72
Polyphenolics Quantification ...... .... ............ ...................... ............... 75
C conclusions .............. ................................... ....... ........ ......... 76

4 AROMA PROFILES OF BEVERAGES OBTAINED FROM FRESH AND DRIED
H IB IS C U S .................................. ......... .......... 82

Introduction .............. .. ................................. ........ .. ........... 82
Materials and Methods................................................... 83
Sample Preparation........................... ............... 83
GC-O Analysis......................................... ............... 84
Identification Procedures .............. .... .................... ..... 86
S tatistica l A na lysis ........................................................................ ......... 86
Results and Discussion.......................................... ............... 86


6









H ibiscus V olatiles C om position .............. ................ .................................... 86
GC-M S Identifications ........................ ........................................ 87
G C -O A rom a P profiles .......................................... .............. .............. 89
C o n c lu s io n ............................................................................................... 9 0

5 PROCESSING HIBSCUS BEVERAGE USING DENSE PHASE CARBON
DIOXIDE: MICROBIAL AND PHYTOHCEMICAL STABILITY.............................. 95

Introduction ......... ..... .. ............................................................... ...... 95
Materials and Methods................................................... 96
Beverage Preparation .................. ...... ................................. 96
S o lubility E xperim ent ............................................................ 97
Dense Phase CO2 Equipment ......................................................... 97
DPCD Process Optimization ...................... .................... .................... 98
Storage Experiment....................................... .......... 99
M icrobial Analysis....................................... .......... 100
pH, oBrix, and Titratable Acidity.................................................................. 100
Anthcyanin Content, Total Phenolics and Antioxidant capacity................... 101
HPLC Quantification of Polyphenolics ........... .................. ............... 101
Statistical Analysis ....................................... 102
Results and Discussion....................................... 102
S olubility M easurem ents ...................... ................ .......... ....... ........ 102
M icrobial Inactivation Study.............................. .................................... 103
Microbial Stability during Storage .......................... .............. 105
Physicochemical Stability during Storage............................... 105
Phytochemical Stability during Storage ........ ............ ...................... 106
C o n c lu s io n s .............. ..... ............ ............................... ........................................ 1 0 8

6 PROCESSING HIBISCUS BEVERAGE USING DENSE PHASE CARBON
DIOXIDE: SENSORY ATTRIBUTES AND AROMA COMPOUNDS STABILITY .... 115

Introduction ............... .................... ........ ..... ................ ......... 115
M materials and M ethods............................................ .................. 116
Beverage Preparation ....................................... ..... ................. 116
Dense Phase CO2 Equipment ...... ......... ....... .............. ............... 117
Physicochem ical Analysis ...... .... ............ ..................... ......... ...... 118
S ensory E va luation ................................ ......... .. .............. ........... 118
Identification Procedures ........... ...... .... ............... ............. 120
Color Analysis ............. .... ............................. ... ...... ......... 120
Statistical Analysis ....................................... 121
Results and Discussion....................................... 121
Physicochem ical Analysis ...... .... ............ ..................... ......... ...... 121
Sensory Evaluation .............................................. 122
Aroma Compounds .............................................. 123
C olor A analysis .. ................................. .............................................. 125
Conclusions ...................... ........... .......... ............... 126









7 SUMMARY AND CONCLUSIONS...................................... 133

APPENDIX

A EXTRACTION EXPERIMENT STATISTICAL ANALYSIS ................................ 134

B STORAGE EXPERIMENT STATISTICAL ANALYSIS...................... ......... 137

C HIBISCUS SABDARIFFA PICTURES ................ ........ .......... .................. 140

LIST O F R EFER EN C ES ................................................. ......... ............... 141

BIOGRAPHICAL SKETCH ......... ......... .. ................... .................. ............... 151









LIST OF TABLES


Table page

2-1 U.S. population projections........................ .... ............................. 24

2-2 Ingredients and nutritional facts of four hibiscus commercial products .............. 30

2-3 Nutritional composition of fresh hibiscus calyces..................... ........... .... 31

2-4 Hibiscus extraction conditions found in the literature................... ........... 35

2-5 Main classes of polyphenolic compounds ............... ................ ................. 39

2-6 Classification of food flavonoids ............... .... ............................ ............... 40

2-7 CO2 solubility of liquid foods measured at 40 oC ..................... .. ............. 53

3-1 Measured pH, total solids (TS) (g of solids/100 mL of extract), titra acidity
(TA) (g of malic acid/100 mL of extract), and color (L*, a*, b* values, color
density (CD) and hue tint (HT)) for the extracts ............................................... 77

3-2 Linear regression and correlation coefficients between measured parameters
for cold and hot water extraction processes. ..................................................... 79

3-3 Identification of anthcocyanins present in hibiscus using their spectral
characteristics with HPLC-DAD and positive ions in LC-MS and MS2............... 79

3-4 Identification of polyphenolics present in hibiscus using their spectral
characteristics with HPLC-DAD and negative ions in LC-MS and MS2, and
respective standards. ............. ..... ........................ ............ ............... 79

3-5 Polyphenolics content (mg/L) of hibiscus samples analyzed in this studyd....... 81

4-1 Extraction conditions and measured pH and oBrix values for hibiscus
sam ples included in this study ............................. .................. .. ....... ...... 92

4-2 MS identification of hibiscus volatiles. Peak areas were normalized (100) to
the largest peak in all four sam ples. .......................... ......... ..... ......... 93

4-3 Hibiscus aroma active compounds. Peak heights were normalized (100) to
the m ost intense peak in all four sam ples................................ ..................... 94

5-1 Response surface design used to test the effect of pressure and residence
time on microbial reduction logoi) at 40 oC and 8% CO2 ............................. 110

5-2 Physicochemical and phytochemical changes of unprocessed (CONTROL),
dense phase-CO2 processed (DPCD), and thermally treated (HTST) hibiscus
beverages during refrigerated storage at 4 C............................. ............... 112









5-3 Polyphenolics content (mg/L) of unprocessed (CONTROL), dense phase-
CO2 processed (DPCD), and thermally treated (HTST) hibiscus beverages
during refrigerated storage at 4 oC. ............... .... ............. ........ .......... 114

6-1 Measured pH, Brix, and titra acidity (TA) (g of malic acid/100 mL of
beverage) at weeks 0 and 5 of refrigerated storage (4 oC)............................. 127

6-2 Difference in flavor and overall likeability between fresh (reference and
hidden reference), dense phase-CO2 processed (DPCD; 34.5 MPa, 8% CO2,
6.5 min, 40 oC) and thermally treated (HTST; 75 oC, 15 s) hibiscus
beverages detected by untrained panelists (n = 75) at weeks 0 and 5 of
refrigerated storage (4 oC) ........... ..... .... ..... ....... ...................... 127

6-3 MS identification of hibiscus beverage volatiles during storage. Peak areas
were normalized (100) to the largest peak (1-Octen-3-ol) in the CONTROL
(C) w eek 0 sam ple ............................................ .......... ........... 129

A-1 SAS software output of statistical analysis for the anthocyanins concentration
data (AC) perfumed in the hibiscus extraction experiment (Chapter 3)............ 134

B-1 SAS software output of statistical analysis for the anthocyanins concentration
data (AC) perfumed in the hibiscus storage experiment (Chapter 5).............. 137









LIST OF FIGURES


Figure page

2-1 Beverage sectors and segments. (Source: Roethenbaugh 2005) ................. 19

2-2 Total U.S. sales and forecast (f) of RTD non-carbonated beverages at
inflation adjusted prices, 2002-12. (Source: Mintel 2008) ............. ............... 21

2-3 U.S. sales and forecast (f) of RTD non-carbonated beverages at current
prices, by segment, 2002-12. (Source: Mintel 2008). ................................. 21

2-4 Market share according to FDMx (Food, drug and mass merchandisers
excluding Wal-Mart) sales of leading RTD non-carbonated beverage
companies, February, 2008. (Source: Mintel 2008) ................ ........... ........ 23

2-5 Hibiscus pictures. A: hibiscus plant, B: hibiscus flower, C: hibiscus calyxes,
and D: opened hibiscus calyx with velvety capsule in the center.................... 25

2-6 Compounds found in some hibiscus extracts: 1 = protocatechuic acid, 2 =
chlorogenic acid and 3 = hibiscus or hibiscic acid .................... ............. 31

2-7 Chemical structure of anthocyanins present in hibiscus............................. 32

2-8 Structural and spectral characteristics of the major naturally occurring
aglycons. (Source: Rodriguez-Saona and Wrolstad 2005).............................. 43

2-9 Predominant structural forms of anthcoaynins present at different pH levels.
(Source: G iusti and W rolstad 2005)............................................. .... .. ............... 45

2-10 Phase diagram of carbon dioxide ...................................... ....... ................. 51

2-11 Schematic diagram of the continuous flow dense phase CO2 system................ 54

3-1 Total anthocyanins content expressed as delphinidin-3-glucoside (mg/L) for
the extracts. The upper time scale belongs to the 90 oC curve and the lower
time scale belongs to the 25 oC curve. Data represents the mean of n=9.
Values with similar letters within the are not significantly different (Tukey's
HSD, p > 0.05).... ....................................... .......... 77

3-2 Total phenolics content expressed as gallic acid equivalents (mg/L) for the
extracts. The upper time scale belongs to the 90 oC curve and the lower time
scale belongs to the 25 oC curve. Data represents the mean of n=9. Values
with similar letters within the are not significantly different (Tukey's HSD, p >
0.05). ............... .. ...................... ................. ........ 78

3-3 Antioxidant capacity (pmol of TE/mL) L) for the extracts. The upper time
scale belongs to the 90 oC curve and the lower time scale belongs to the 25









oC curve. Data represents the mean of n=9. Values with similar letters within
the are not significantly different (Tukey's HSD, p > 0.05)...... ........................ 78

3-4 HPLC chromatograms of dried hibiscus (DHE) and fresh hibiscus (FHE) hot
water extracts: (A) 520 nm, (B) 360 nm, (C) 320 nm, (D) 280 nm, and (E) 260
nm. For peak identification see Tables 3-3 and 3-4 ................ ..... ......... 80

4-1 Chemical composition of hibiscus headspace volatiles. Total number of
compounds for each class is put in parentheses. All four samples were
normalized to the total peak area of DHE (dried hibiscus hot extraction). DCE
= dried hibiscus cold extraction, FHE = fresh hibiscus hot extraction, FCE =
fresh hibiscus cold extraction......................... ............................. 92

5-1 Schematic diagram of the setup used for the hibiscus beverage thermal
treatm ent (75 oC for 15 s) .................................... ...... ..... .. ........... 109

5-2 CO2 solubility in water and a hibiscus beverage as a function of pressure
measured at 40 oC. Data represents the mean of n=3. Values with similar
letters within the are not significantly different (Tukey's HSD, p > 0.05).......... 109

5-3 Aerobic plate counts of unprocessed (CONTROL), dense phase-CO2
processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally treated
(HTST; 75 oC, 15 s) hibiscus beverage during refrigerated storage (4 oC)....... 110

5-4 Yeast/mold counts of unprocessed (CONTROL), dense phase-CO2
processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally treated
(HTST; 75 oC, 15 s) hibiscus beverage during refrigerated storage (4 oC)....... 111

5-5 Hue tint values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally treated (HTST; 75
oC, 15 s) hibiscus beverage during refrigerated storage (4 oC) ...................... 111

5-6 Concentration of anthocyanins of unprocessed (CONTROL), dense phase-
CO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally
treated (HTST; 75 oC, 15 s) hibiscus beverage during refrigerated storage (4
C ) .............................................................. ........... ...... 1 1 3

6-1 Chemical composition of hibiscus beverage headspace volatiles during
storage. Total number of compounds for each class is put in parenthesis. All
six samples were normalized to total peak area of the sample CWO
(CONTROL week 0). C = CONTROL, D = DPCD, H = HTST, W = week......... 128

6-2 L values of unprocessed (CONTROL), dense phase-CO2 processed (DPCD;
34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally treated (HTST; 75 oC, 15
s) hibiscus beverage during refrigerated storage (4 oC). ................................ 130









6-3 a* values of unprocessed (CONTROL), dense phase-CO2 processed (DPCD;
34.5 MPa, 8% C02, 6.5 min, 40 oC) and thermally treated (HTST; 75 oC, 15
s) hibiscus beverage during refrigerated storage (4 oC). .............. ............... 130

6-4 b values of unprocessed (CONTROL), dense phase-CO2 processed (DPCD;
34.5 MPa, 8% C02, 6.5 min, 40 oC) and thermally treated (HTST; 75 oC, 15
s) hibiscus beverage during refrigerated storage (4 oC). .............. ............... 131

6-5 Hue angle values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 oC) and thermally treated (HTST; 75
oC, 15 s) hibiscus beverage during refrigerated storage (4 C) ..................... 131

6-6 Chroma values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 oC) and thermally treated (HTST; 75
oC, 15 s) hibiscus beverage during refrigerated storage (4 oC) ..................... 132

6-7 AE values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 oC) and thermally treated (HTST; 75
oC, 15 s) hibiscus beverage during refrigerated storage (4 oC) ..................... 132

B-1 Pictures of dried hibiscus (A), dried hibiscus extraction process (B), hibiscus
beverage (C), hibiscus beverage in the dense phase carbon dioxide (DPCD)
feed tank (D), DPCD processing equipment (E), DPCD processed hibiscus
beverage (F), hibiscus beverage samples for analysis (G), hibiscus beverage
under refrigerated storage (H), and DPCD processed hibiscus beverage after
14 weeks of storage at 4 oC (I). Photos by Milena Ramirez.. .......................... 140









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PROCESSING OF A HIBISCUS BEVERAGE
USING DENSE PHASE CARBON DIOXIDE

By

Milena Maria Ramirez Rodrigues

August 2010

Chair: Maurice R. Marshall
Cochair: Murat O. Balaban
Major: Food Science and Human Nutrition

Consumer demand for natural beverages with health promoting properties that

offer fresh-like sensory attributes and changes in U.S. demographics have created the

opportunity for the development of new products that would target new market

segments.

Hibiscus sabdariffa (family Malvaceae) red calyces are rich in anthocyanins and

other phenolic compounds. Fresh and dried hibiscus is used to prepare cold and hot

beverages and their preparation includes an extraction step followed by a pasteurization

method. Although thermal preservation of foods is effective in reducing microbial loads it

can also lead to organoleptic and nutritional changes. Nonthermal processes like dense

phase carbon dioxide (DPCD) are an alternative which may help preserve the color,

flavor, and nutrients of food.

Equivalent cold and hot water conditions were found for anthocyanins extraction

of dried hibiscus in this research. Likewise, similar polyphenolic profiles and chemical

composition of aroma compounds were observed between fresh and dried hibiscus.









Solubility of C02 in a hibiscus beverage (5.06 g C02/mL at 31.0 MPa) and

optimal processing conditions to inactivate yeasts and molds (Y&M) were 34.5 MPa and

6.5 min. DPCD was a viable technology for processing hibiscus beverage since it

extended its shelf life for 14 weeks of refrigerated storage. Quality attributes were

maintained during storage. Lower losses of anthocyanins were observed in the DPCD

(9%) hibiscus beverage as compared to thermally treatment process (14%) and no

major changes in total phenolics content and antioxidant capacity occurred during

storage. Changes in hibiscus aroma volatiles during storage did not affect untrained

panelists overall likeability of the product.

Findings in this research can help in the development and marketing of hibiscus

beverage.









CHAPTER 1
INTRODUCTION

Justification

Anthocyanins are water-soluble pigments responsible for the red to purple to blue

colors in many fruit, vegetables, flowers, and cereal grains. The interest in anthocyanin

pigments has intensified in recent years because of their possible health benefits. Thus

in addition to their functional role as colorants, anthocyanin extracts may improve the

nutritional quality of foods and beverages (Wrolstad 2004).

Consumer demands for natural beverages with health promoting properties that

offer fresh-like sensory attributes and changes in U.S. demographics with Hispanics and

Blacks as important growth-driving demographics (Mintel 2008) have created the

opportunity for the development of new products that would target these market

segments.

Hibiscus sabdariffa (family Malvaceae) is a short-day annual shrub that grows in

many tropical and subtropical countries and is one of the highest volume specialty

botanical products in international commerce (Plotto 1999). The red calyces are the part

of the plant with commercial interest and are rich in organic acids, minerals,

anthocyanins, and other phenolic compounds.

Fresh and dried hibiscus calyces are used to prepare cold and hot beverages

which are commonly mixed with a sweetener and are characterized by an intense red

color and acidic flavor which provides a sensation of freshness. The preparation of a

hibiscus beverage includes an extraction step followed by a pasteurization method.

Although thermal preservation of foods is effective in reducing microbial loads it can

also lead to organoleptic and nutritional changes. Nonthermal processes are an









alternative which may help preserve the color, flavor, and nutrients of food. Dense

phase carbon dioxide (DPCD) is a cold pasteurization method that uses pressures

below 90 MPa in combination with carbon dioxide (C02) to inactivate microorganisms.

This non-thermal technology is mainly used in liquid foods and since the food is not

exposed to the adverse effect of heat, its fresh-like physical, nutritional, and sensory

qualities are maintained.

Hypothesis

The combination of a cold extraction process with dense phase carbon dioxide

(DPCD) processing will help prevent the degradation of anthocyanins present in a

hibiscus beverage, and thus provide a product with enhanced quality and phytochemical

activity.

Specific Objectives

1. To compare the effects of cold and hot water extraction on the physicochemical

and phytochemical properties of hibiscus extracts and to identify and quantify the

anthocyanins and major polyphenolics present in extracts obtained from fresh

and dried hibiscus by equivalent cold and hot water extraction conditions.

2. To determine the aroma profile differences between four extracts obtained from

fresh and dried hibiscus extracted at two different conditions (22 C for 4 h and

98 C for 16 min), by GC-MS and GC-olfactometry.

3. To determine the solubility of C02 in a hibiscus beverage, to optimize DPCD

processing parameters based on microbial reduction, and to monitor during

refrigerated storage the microbial, physicochemical, and phytochemcial changes

of DPCD processed hibiscus beverage compared to thermally treated and control

(untreated) beverages.









4. To determine the effect of DPCD processing on the sensory attributes and aroma

compounds of hibiscus beverage when compared to a thermally treated and a

control (untreated) and to monitor the changes in these attributes during

refrigerated storage.









CHAPTER 2
LITERATURE REVIEW

The Beverage Market

The global beverage market is comprised of four sectors: 1) hot drinks, 2) milk

drinks, 3) soft drinks, and 4) alcoholic drinks. Hot drinks include tea, coffee, and hot-

malt based products; milk drinks include white drinking milk and flavored milk products;

soft drinks are divided into five main subcategories: (bottled water; carbonated soft

drinks; dilutables including powder and liquid concentrates; 100% fruit juice and nectars

with 25-99% juice content; still drinks including ready to drink (RTD) teas, sports drinks,

and other non carbonated products with less than 25% fruit juice, and alcoholic drinks

which include beer, wine, sprits, cider, sake and flavored alcoholic beverages

(Roethenbaugh 2005). A diagram of these sectors is presented in Figure 2-1.


Hot drinks Soft drinks Milk drinks Alcoholic drinks



Tea titled White milk Beer
water


SCoffee fCarbonated Flavored Wine
drinks milk


Other- Dilutables Spirits
drinks


Fruit juice/ Other
alcoholic
nectars drinks


Still drinks


Figure 2-1. Beverage sectors and segments. (Source: Roethenbaugh 2005).









Soft drinks are normally defined as sweetened water-based beverages, usually

having a balanced acidity. Flavor, color, fruit juice or fruit pulp are often added in their

formulation. The main ingredient in soft drinks is water, and thus their primary function is

hydration. There are two basic types of soft drinks: ready-to-drink (RTD) products and

concentrates or dilute-to-taste products. The RTD sector is divided into products that

are carbonated and those that are non-carbonated (Ashurst 2005).

The market of ready to drink (RTD) non-carbonated beverages can be divided in

four segments: 1) bottled water, 2) sports/energy drinks, 3) fruit juice/juice drinks, and 4)

RTD teas and coffees.

Market Performance and Competitive Context of U.S. Ready to Drink Non-
Carbonated Beverages

Sales for ready to drink non-carbonated beverages reached $38.6 billion in 2007,

exhibiting a 35% growth, measured in current prices, during the period 2002-2007. The

market is projected to grow 33% in current prices from 2008-12, or the equivalent of

16% when considering the impact of inflation (Figure 2-2). Enhanced bottled waters,

energy drinks, and RTD teas are the categories that have driven this growth (Mintel

2008).

Fruit juice and juice drinks, bottled water, sports and energy drinks, and RTD tea

and coffee accounted for 37.4, 28.6, 26.0, and 8.1% or the RTD non-carbonated

beverage market in 2007. While the sales for fruit juice and juice drinks are forecast to

decline in the period 2007-2012, the other three categories will have increasing sales

over this period (Figure 2-3).








50.00
45.00
40.00
35.00
. 30.00
o
S25.00
* 20.00
15.00
10.00
5.00
0.00


2002 2003 2004 2005 2006 2007 2008f 2009f 2010f 2011f 2012f

Figure 2-2. Total U.S. sales and forecast (f) of RTD non-carbonated beverages at
inflation adjusted prices, 2002-12. (Source: Mintel 2008).


20,000
18,000
16,000- -
14,000
12,000
= 10,000 -
E
6 8,000
6,000
4,000 -
2,000 -

2002 2003 2004 2005 2006 2007 2008f 2009f 2010f 2011f 2012f
Fruit juice and juice drinks 0 Bottled water
SSports and energy drinks 0 RTD tea and coffee

Figure 2-3. U.S. sales and forecast (f) of RTD non-carbonated beverages at current
prices, by segment, 2002-12. (Source: Mintel 2008).


11111111111









PepsiCo dominates brand sales in this category, its RTD non-carbonated

beverage portfolio includes the top juice (Tropicana), bottled water (Aquafina), sports

drink (Gatorade), and tea (Lipton) brands. Because PepsiCo is such a strong player in

every segment, its total market share as of February, 2008 was 25.3% according to

FDMx (Food, drug and mass merchandisers excluding Wal-Mart) sales (Figure 2-4).

However, sales growth has been slow, allowing other beverage suppliers to gain a

market share (Mintel 2008).

Coca-Cola is well-positioned to gain a market share with its Glaceau, Fuze, and

Powerade brands. Its sales increased nearly 14% during 2007-08, adding 1.1% to

Coca-Cola's share in the market. The company experienced solid sales growth in every

segment (Mintel 2008).

While major beverage companies dominate the RTD non-carbonated beverage

category, mid-level entrants are defined by their strong representation in a single

segment, such as Ocean Spray in the juice segment, or Nestle in water sales. Smaller

(under the category "Other" in Figure 2-4) participants in the market have found their

niche, and are defined mostly by a single brand targeting a specific demographic, like

those in the energy/sports drink segment like Red Bull and Rockstar (Mintel 2008).

Many small, high-growth companies present alliance opportunities for big

players. These include, Hansen's Natural (natural soda and fruit juice, and energy

drinks), Jumex (Hispanic-targeted beverages), Jones Soda (unique soda flavors),

Tampico (Hispanic-targeted beverages) and AriZona Tea Co. (RTD tea) (Mintel 2008).

Private label is becoming a more important player in the RTD non-carbonated

beverage category. While its presence is strongest in commodity segments such as









juice and water, it is growing fastest in trendy beverages (sports/energy and RTD teas

and coffees) (Mintel 2008).




PepsiCo
c Coca-Cola Company
24.1%
Nestl6
C Kraft
a Cadbury Schweppes
Ocean Spray
Campbell
Private label
Other
3.2%
3.2% 3.2%










are demanding more from their beverages. Drinks should not only be thirst-quenchers

but also provide added benefits. Health and wellness increasingly plays an influential

role in consumer choices on the beverage aisle. Consumers are seeking products that

add value to their diet; however, not only must products deliver nutrition conveniently,

but the packaging must carry a convenient format (Mintel 2008).

Hispanics and blacks are important growth-driving demographics, not only

because these groups are projected to exhibit an above-average population growth, but

also because they display an above-average incidence of juice consumption.

Additionally, both groups are the key consumer in high-growth sports and energy drinks

markets (Mintel 2008).









According of the U.S. Census Bureau projection for 2050, non-Hispanic whites

will no longer make up the majority of the population. Today non-Hispanic whites make

up about 68% of the population. This is expected to fall to 46% in 2050 as a result of a

much older white population relative to minorities. Hispanic population is projected to

change from 15% to 30% of the total U.S. population while African American and Asian

Americas will reach 15 and 9% of the population by 2050 (Table 2-1). The U.S. has

nearly 305 million people today. The population is projected to reach 400 million by

2039 and 439 million in 2050 (U.S. Census Bureau 2009).

Table 2-1. U.S. population projections

2008 2050
Non-Hispanic whites 68% 46%
Hispanic 15% 30%
African Americans 12% 15%
Asian American 5% 9%
(Source: U. S. Census Bureau 2009).

Hibiscus (Hibiscus sabdariffa)

Characteristics and Economic Importance

There are more than 300 species of hibiscus around the world. One of them is

Hibiscus sabdariffa, Linn, which is a member of the Malvaceae family. The origin of H.

sabdariffa is not fully known but it is believed to be native to India and Malaysia and to

have been carried at an early date to Africa. It is widely grown in tropical and subtropical

regions including Africa, South East Asia and some countries of America. Seeds are

said to have been brought to the New World by African slaves. It is know by different

synonyms and vernacular names such as "roselle" in the U.S and England, "l'oiselle" in

France, "jamaica" or "flor de jamaica" in Mexico and Spain, "karkade" in Sudan and









Arabia, "sorrel" in the Caribbean and "byssap" in Senegal (Morton 1987; Stephens

2003). In this study the word "hibiscus" will be used to refer to Hibiscus sabdariffa.

Hibiscus (Figure 2-5) is a short-day annual shrub and can grow to a height of 1-3

m, depending on variety. The green leaves are about 8-12 cm long and the stems,

branches, leaf veins and petioles are reddish purple. Flowers are up to 12.5 cm wide,

they are yellow with a rose or maroon eye, and are made up of five petals. After the

flowers fall apart, the calyx which is a red cup-like structure consisting of 5 large sepals

with a collar epicalyxx) of 8 to 12 slim pointed bracts around the base, begins to enlarge,

becomes fleshy, crisp but juicy (3.2-5.7 cm long), and fully encloses the velvety capsule,

(1.25-2 cm long), which is green when immature, 5-valved, with each valve containing 3

to 4 kidney-shaped light-brown-seeds, (3-5 mm long). The capsule turns brown and

splits open when mature and dry (Morton 1987; De Castro and others 2004).











A B C D

Figure 2-5. Hibiscus pictures. A: hibiscus plant, B: hibiscus flower, C: hibiscus calyxes,
and D: opened hibiscus calyx with velvety capsule in the center.

Usually, hibiscus is propagated by seeds or cuttings and grows on sandy soil.

The ideal planting time in North America is from April to May, blooming occurs in

September and October, and calyces are ready for harvest in November and









December. The "fruits" should be gathered before any woody tissue develops in the

calyx. They should be tender, crisp, and plump (Stephens 2003).

Hibiscus has several uses. Its calyces, which is the part of the plant of

commercial interest, are used either fresh or dehydrated in the processing of preserves,

jellies, jams and sauces for their rich pectin content, to prepare hot and cold beverages

which are commonly mixed with a sweetener and are characterized by an intense red

color and acidic flavor which provides a sensation of freshness, in the production of

wine, and color and flavor extracts. They are also a source of soluble and insoluble

fiber. The leaves are used extensively for animal fodder and fiber and are also used in

salads, and the seeds are a source of protein and lipids and constitute a byproduct in

hibiscus production (AI-Wandawi and others 1984; EI-Adawy and Khalil 1994; Mounigan

and Badrie 2007; Sayago-Ayerdi and others 2007; Hainida and others 20008).

Traditionally fresh hibiscus calyces are harvested by hand and are either frozen,

dried in the sun or artificially preserved and are either sold into the herbal tea and

beverage industry, or local and regional markets. Five kilograms of fresh calyces

dehydrate to 0.45 kg of dried hibiscus. Industrial scale operations that use hibiscus

include production of vacuum concentrated extract, spray drying of extracts, beverages,

natural food colorant and natural food flavor (AI-Kahtani and Hassan 1990).

Hibiscus is one of the highest volume specialty botanical products in the

international commerce and demand has steadily increased over the past decades.

Approximately 15,000 metric tons of dried hibiscus enter international trade each year.

Many countries produce hibiscus but the quality markedly differs. China and Thailand

are the largest producers and control much of the world supply. Mexico, Egypt,









Senegal, Tanzania, Mali, Sudan, and Jamaica are also important suppliers but

production is mostly used domestically (Plotto 1999).

Germany and the U.S. are the main countries importing hibiscus. The biggest

German buyer is Martin Bauer, one of the oldest and largest companies in the herb

industry. They use hibiscus in numerous products including herbal teas, herbal

medicines, syrups and food coloring. Main importers in the U.S are Celestial

Seasonings and Lipton, both tea companies. Hibiscus is also used in ready to serve

beverages made by Knudson, Whole Foods and other food and beverage

manufacturers (Plotto 1999).

Commercial Hibiscus Products

Hibiscus' striking red color, refreshing properties and associated health benefits

has attracted the interest of several entrepreneurs to start a business around the idea of

manufacturing hibiscus based beverages. There are four ready to drink commercial

products that use hibiscus as the main ingredient. Following is a brief description of

these products as well as their marketing approach.

Hibiscus Lemon Bissap

Produced by the company Adina for Life Inc. (located in California) and

established in 2005, this product is marketed as a New Age beverage, with all-natural

ingredients, refreshing, and with good-for-you appeal. The label is bright, colorful, and

folksy. The company sources its hibiscus using fair trade arrangements from

independent farmers in Senegal. Adina's president is also from Senegal and considers

this product to help rescue traditional beverage mixes from his country (Anon 2006).









Cafita Aguas Frescas (jamaica (hibiscus) flavor)

Produced by the company Eat Inc. (located in North Carolina) and established in

2003, this product is marketed as 100% natural that provides health benefits and targets

mainly Hispanic consumers. It has the intent of bringing a traditional Mexican beverage

known as "Agua de jamaica" to the Hispanic population in the U.S. (Anon 2006).

Squish Hibiscus Presse

Produced by the company Squish Hibiscus Presse located in New Zealand, this

product is marketed as a beverage with unique exotic floral fruity flavor that has

beneficial properties. This is a new product in the New Zealand market that consumers

are not familiar with. The market segments to which this product is targeted are women

between 18 and 35 years old and kids (Anon 2006).

Simply Hibi

Produced by Ibis Organica, a UK based company; this company sources its raw

material from Uganda and has established a program to help improve living conditions

in that country. The product contains 87% hibiscus extract and 13% grape juice

concentrate and it is marketed as 100% natural and high in antioxidants.

The ingredients and nutritional facts for these four products are presented in

Table 2-2.

Composition and Associated Health Benefits

Hibiscus calyces are rich in organic acids including succinic, oxalic, tartaric, and

malic acids (Wong and others 2002), hibiscus acid which is a lactone form of (2S,3R)-

(+)-2-hydroxycitric acid and its 6-methyl ester (Hansawasdi and others 2000), ascrobic

acid, 3-carotene, and lycopene (Wong and others 2002). It is also high in phenolic

compounds such as protocatechuic acid (3,4-dihydroxybenzoic acid) (Tseng and others









1998; Liu and others 2002; Lin and others 2003) and chlorogenic acid (Segura-

Carretero and others 2008) (Figure 2-6), minerals (aluminum, chromium, copper and

iron) (Wrobel and others 2000), sugars (glucose, fructose, sucrose and xylose) (Pouget

and others 1990a; Wong and others 2002), water-soluble polysaccharides (Muller and

Franz 1992) and anthocyanins (Du and Francis 1973; Wong and others 2002). There

can be composition variations depending on variety, soil, climate and growing

conditions, and post harvest handling and processing. The nutritional composition of

fresh hibiscus calyces is presented in Table 2-3.

For many years, hibiscus has been used in different countries as a medicinal

herb for therapeutic purposes. According to different ethnobotanical studies, some

traditional medicines use the aqueous extract of the plant as a diuretic, for treating

gastrointestinal disorders and hypercholesterolemia, and as a diaphoretic and

antihypertensive drug (Herrera-Arellano and others 2004).

Many biological activities have been reported in aqueous extracts of Hibiscus

sabdariffa. Animal experiments have shown that the consumption of this extract has

antihypertensive (Odigie and others 2003), antiatherosclerotic (Chen and others 2003),

lipid profile reduction (Carvajal-Zarrabal and others 2005), and antioxidant properties

(Suboh and others 2004; Hirunpanich and others 2006; Liu and others 2006). Studies

with human patients have also shown that the regular consumption of hibiscus extract

has an antihypertensive effect (Haji Faraji and Haji Tarkhani 1999; Herrera-Arellano and

others 2004) and reduces serum cholesterol in men and women (Lin and others 2007).

Several compounds isolated from hibiscus extracts also possess

pharmacological activities.









Table 2-2. Ingredients and nutritional facts of four hibiscus commercial products


Cahita Aguas
Hibiscus Frescas Squish
Lemon (jamaica Hibiscus Simply
Bissap flavor) Presse Hibi*
Ingredients
Water
Hibiscus
Sugar
Fructose
Organic evaporated cane juice
Concentrated pineapple juice
Lemon juice
Camu-camu
Organic rosehips
Acerola
Ascorbic acid
Nutritional Facts
Serving size 236 mL 236 mL 375 mL
Servings per container 1.75 2 1
Calories 80 135 124
Sodium 15 mg 11 mg
Total carbohydrates 28 g 34 g 28.9 g
Sugars 19 g 34 g 28.9 g
Vitamin C 4%
Magnesium 2%
Potassium 2%
Calcium 4% 4%
Camu-camu 50 mg
Rosehips (organic) 50 mg
Acerola 167 mg
Lemon bioflavonoids 83 mg
* Nutritional data for this product is not available. Pictures were taken from the actual products by Milena
Ramirez. Nutritional data was retrieved from the bottle labels or from New nutrition business 12(1):19-22).









Table 2-3. Nutritional composition of fresh hibiscus calyces

g/100 g mg/100 g mg/100g
Water 86.58 Calcium 215 Vitamin C 12
Protein 0.96 Phosphorus 37 Riboflavin 0.028
Lipids 0.64 Iron 1.48 Niacin 0.31
Carbohydrates 11.13 Sodium 6 Thiamin 0.011
Ash 0.51 Potassium 208 Vitamin A 287 UI
Magnesium 51 Energy 49 kcal
(Source: USDA 2009)



0 HO HO OH

HO I 1/OH HO /-0 '-COOH

OH O HO COOH OH
1 2 3
Figure 2-6. Compounds found in some hibiscus extracts: 1 = protocatechuic acid, 2 =
chlorogenic acid and 3 = hibiscus or hibiscic acid

Several compounds isolated from hibiscus extracts also possess

pharmacological activities. Protocatechuic acid has antiatherosclerosis (Lee and others

2002), antitumor promotion (Tseng and others 1998; Lin and others 2003; Olvera-

Garcia and others 2008), antioxidant (Lin and others 2003), and anti-inflammatory (Liu

and others 2002) activities. Anthocyanins isolated from hibiscus exhibited antioxidant

(Wang and others 2000) and anticancer (Chang and others 2005; Hou and others 2005)

activities while hibiscus acid and its 6-methyl ester have shown to be a-amylase

inhibitors (Hansawasdi and others 2000).

Hibiscus Anthocyanins

Recently there has been a market interest in hibiscus anthocyanins due to their

beneficial health effects and high antioxidant properties which have been extensively

evaluated (Tee and others 2002; Tsai and others 2002; Tsai and Huang 2004; Prenesti









and others 2007: Sayago-Ayerdi and others 2007) and as a potential source of natural

food colorant. The two major anthocyanins present in hibiscus are: delphinidin-3-

sambubioside also known as delphinidin-3-xylosylglucoside or hibiscin and cyanidin-3-

sambubioside also known as cyaniding-3-xylosylglucoside or gossypicyanin. They

account for approximately 70 and 30% of total anthocyanins, respectively. Other

anthocyanins like delphinidin-3-glucoside, delphinidin-3-(feruloyl)rhamnoside, cyanidin-

3-glucoside, cyaniding-3-O-rutinoside, and cyaniding-3,5-diglucoside have been found

in minor concentrations in some varieties (Figure 2-7) (Du and Francis 1973; Pouget

and others 1990b; Tsai and others 2002; Wong and others 2002; Mourtzinos and

others 2008; Segura-Carretero and others 2008).

3'
4'

+ "B
HO 0O
A C5

3
5

Anthocyanin 3' 4' 5' 3 5
Cyanidin-3-sambubiioside OH OH H 2-O-,/-xylosyl-D-glucose OH
Cyanidin-3-glucoside OH OH H Glucosyl OH
Cyanidin-3,5-diglucoside OH OH H 3,5-diglucosyl OH
Cyanidin-3-rutinoside OH OH H O-rutinosyl OH
Delphinidin-3-sambubioside OH OH OH 2-O-,/-xylosyl-D-glucose OH
Delphinidin-3-glucoside OH OH OH Glucosyl OH
Delphinidin-3-(feruloyl)rhamnoside OH OH OH (feruloyl)rhamnoside OH


Figure 2-7. Chemical structure of anthocyanins present in hibiscus









The stability of hibiscus anthocyanins has been studied in model systems testing

the effect of different chemical compounds (ascorbic acid, BHA, propyl gallate,

disodium EDTA, sodium sulfite) (Pouget and others 1990a), temperature (Gradinaru

and others 2003; Dominguez-L6pez and others 2008; Cisse and others 2009), sugar

type and concentration (Tsai and others 2004), copigmentation and polymerization (Tsai

and Huang 2004) as well as their stability in various foods including jellies, beverages,

gelatin desserts, and freeze dried products. Color stability during storage has also been

tested (Esselen and Sammy 1975; Clydesdale and others 1979). Heat, light, and

humidity were all found to be detrimental to anthocyanin stability.

Some studies have shown that thermal degradation of hibiscus anthocyanins

follow first-order reaction kinetics (Gradinaru and others 2003; Dominguez-L6pez and

others 2008; Mourtzinos and others 2008). Thermal stability of hibiscus anthocyanins in

the temperature range of 60-90 C in the presence or absence of 8-cyclodextrin was

studied. The temperature-dependent degradation was modeled by the Arrhenius

equation and the activation energy for the degradation of hibiscus anthocyanins was

~54 kJ/mol. The presence of,8-cyclodextrin improved thermal stability of nutraceutical

antioxidants present in hibiscus extracts both in solution and solid state (Mourtzinos and

others 2008). Another study showed that the activation energy for the degradation of

hibiscus anthcoyanins was 66.22 kJ/mol (Duangmal and others 2008) while a third

study found that copigmentation with chlorogenic acid didn't improve their stability in

solution and activation energies for their degradation were between 55.68 and 63.22

kJ/mol (Gradinaru and others 2003).









Hibiscus Extraction Process

Some researchers have focused on hibiscus water extracts while others have

employed organic solvents to extract possible bioactive compounds. Indeed the

different extraction techniques (extraction time and temperature) make comparison

among studies difficult. Moreover different varieties have been used. Table 2-4

summarizes some of the conditions used for hibiscus extraction found in the literature.

Some research has been done regarding the optimization of hibiscus extraction

process. One study tested three different hibiscus to water ratios (1:52, 1:67, 1:62 w/v)

at three extraction times (20, 25, 30 min) in a hot extraction at 100 C. They found that

optimum conditions based on color and taste were 1:62 w/v for 30 min (Bolade and

others 2009). Wong and others (2003) found that optimum condition for hibiscus

extraction was 3.5 h at 60 C based on anthocyanins content and color.

Hibiscus Flavor

Hibiscus flavor is a combination of sweet and tart, similar to cranberry. Few

studies have been done related to hibiscus flavor. Gonzalez-Palomares and others

(2009) identified 20 volatile compounds in a hibiscus extract using SPME and GC-MS,

including terpenoids, esters, hydrocarbons, and aldehydes. They also found 14

compounds in reconstituted spray dried extracts from which only 10 were present in the

original extract and the other four were products of degradation. Thermally generated

volatiles from untreated, frozen, hot-air-dried at 50 C, and hot-air-dried at 75 C

hibiscus by steam distillation were analyzed by GC and GC-MS (Chen and others

1998). They characterized more than 37 compounds including fatty acid derivatives,

sugar derivatives, phenol derivatives, and terpenes.










Table 2-4. Hibiscus extraction conditions found in the literature


Country Hibiscus:
of solvent Extraction Extraction
origin Fa Db Solvent ratio time temperature Ac Reference
MeOH with
S0.125% citric acid 1:1.65 w/v 48 h Pouget and others 1990a
AI-Kahtani and Hassan 1990;
Sudan Water 1:10 w/w 40 min 60 C Hassan and Hobani 1998
Mexico Water 1:8 w/v 15 min 60 C Beristain and others 1994
Taiwan Water 1:30 w/v 10 min Boiling Duh and Yen 1997
Malasya MeOH 1:10 w/v 24 h 25 C Tee and others 2002
Taiwan Water 1:100 w/v 3 min Boiling Tsai and others 2002
Malaysia *Water 1:5 w/v 1 h Boiling Wong and others 2002
3% Formic acid
Egypt in MeOH 24 h 4 C Gradinaru and others 2003
30-300
Malaysia Water 1:40 w/w min 30-90 C Wong and others 2003
Mexico Water 1:8 w/v 128 min Ambient Andrade and Flores 2004
Mexico Water 1:50 w/v 10 min Boiling Herrera-Arellano and others 2004
Nigeria *Water 1:30 w/v 30 min Boiling Oboh and Elusiyan 2004
Acidified EtOH
Taiwan (1.5 mol/L HCI) 1:50 w/v Tsai and Huang 2004
Mexico Water 1:10 w/v 5 min Boiling Dominguez-Lopez and others 2008
Singapore Water 1:50 w/v 1 h Ambient d Wong and others 2006
Egypt Water 1:50 w/v 5-930 min Ambient Prenesti and others 2007
Egypt Water 1:50 w/v 3 min 100 C Prenesti and others 2007
12% v/v EtOH
Egypt in water 1:50 w/v 30 min Ambient Prenesti and others 2007
Mexico Water 1:20 w/v 5 min Boiling Sayago-Ayerdi and others 2007
Mexico Water 1:50 w/v 10 min Boiling Olvera-Garcia and others 2008


a = fresh hibiscus, b = dried hibiscus, c = agitation, d = occasional, e = sonication.










Table 2-4. Continued


Country Hibiscus:
of solvent Extraction Extraction
origin Fa Db Solvent ratio time temperature Ac Reference
Mexico Water 1:50 w/v Overnight Ambient Olvera-Garcia and others 2008
Acidified MeOH
(MeOH/HCI
S (99:1 v/v)) 1:10 w/v 4 h Ambient Segura-Carretero and others 2008
Acidified MeOH
(MeOH/HCI
S (99:1 v/v)) 1:10 w/v 30 min Ambient *e Segura-Carretero and others 2008
Acetic acid
S (15% v/v) 1:40 w/v 48 h Ambient Segura-Carretero and others 2008
Water/MeOH/HCI,
Senegal 50:50:2 1:125 w/v 30 min e Juliani and others 2009
Senegal Water 1:62.5 w/v 15 min *e Juliani and others 2009
30% v/v EtOH Gonzalez-Palomares and others
Mexico *in water 1:12.5 w/v 168 h Ambient d 2009
Nigeria Water 1:52-1:62 w/v 20-30 min 100 C Bolade and others 2009
Taiwan Water 1:40 w/v 2 h 95 C Lin and others 2007
Guatemala
and
Senegal Water 1:10 w/v 10 h 25 C Cisse and others 2009


a = fresh hibiscus, b = dried hibiscus, c = agitation, d = occasional, e = sonication.









They concluded that hibiscus aroma was a combination of terpene derivatives with

fragrance notes and sugar derivatives with a caramel like odor.

Phenolic compounds

Phenolic compounds are products of the secondary metabolism of plants.

Biogenetically they originate from two main synthetic pathways: the shikimate pathway

and the acetate pathway. Chemically, phenolics can be defined as substances that

have an aromatic ring bearing one or more hydroxyl groups, including their functional

derivatives (Bravo 1998).

Many properties of plant products are associated with the presence, type, and

content of their phenolic compounds. Of significance to producers and consumers of

foods are the astringency of foods, the beneficial health effects of certain phenolics or

their potential antinutritional properties when present in large quantities (Shahidi and

Naczk 2004).

Classification

Natural polyphenols can range from simple molecules, such as phenolic acids, to

highly polymerized compounds, such as tannins. They occur mainly in conjugated

forms, with one or more sugar residues linked to hydroxyl groups, although direct

linkages of the sugar unit to an aromatic carbon atom also exist. The associated sugars

can be present as monosaccharides, disaccharides, or even oligosaccharides. The

most common sugar residue is glucose, but galactose, rhamnose, xylose, and

arabinose can also be found, as well as glucuronic and galacturonic acids among

others. They can also be associated with carboxylic and organic acids, amines, lipids,

and other phenols (Bravo 1998).Polyphenols can be divided into at least 10 different









classes depending on their basic chemical structure (Table 2-5). Flavonoids, which are

the most important single group, can be further subdivided into 13 classes (Table 2-6).

Phenolic Compounds Attributes

Positive attributes of phenolic compounds include: contribution to flavor and

astringency, natural pigments, antimicrobial and antiviral properties, anti-inflammatory

activity, antitumor and anticancer activity, antimutagenicity, antioxidant potential, and

reduction of coronary heart disease risk (Lule and Xia 2005).

There are also some negative attributes of phenolic compounds that include: off-

flavor and taste contribution, discoloration due to enzymatic and nonenzymatic

reactions, and antinutritional activity because of interactions with proteins,

carbohydrates, minerals, and vitamins (Lule and Xia 2005).

Contribution to flavor

Phenolic compounds may contribute to the aroma and taste of numerous food

products of animal and plant origin. The presence of chlorogenic acid can be related to

the bitterness of wine, cider, and beer while hydroxycinnamates and their derivatives

are responsible for the sour-bitter taste of cranberries. Phenolic substances also

contribute to the flavor of vanilla pod and vanilla extracts. Vanillin, p-

hydroxybenzaldehyde, and p-hydroxybenzyl methyl ester have been found to be the

most abundant volatiles but simple phenolics such as p-cresol, eugenol, p-vinylguaiacol,

and p-vinylphenol as well as aromatic acids such as vanillic and salicylic acids are also

present. Ripe bananas contain volatile phenolics such as eugenol, methyleugenol,

elimicin, and vanillin. Strawberry volatiles contain esters of some phenolic acids such as

ethyl salicylic, methyl cinnamic, and ethyl benzoic acids.










Table 2-5. Main classes of polyphenolic compounds

Class Basic Skeleton Basic Structure


Simple phenols

Benzoquinones


Phenolic acids

Acetophenones

Phenylacetic acids

Hydroxycinnamic acids

Phenylpropenes

Coumarins, isocoumarins


Chromones



Naftoquinones




Xanthones



Stilbenes




Anthraquinones



Flavonoids
Lignans, neolignans
Lignins
(Source: (Bravo, 1998).


C6

C6


C6-C1

C6-C2

C6-C2

C6-C3

C6-C3

C6-C3


C6-C3


C6-C4




C6-C1-C6



C6-C2-C6




C6-C2-C6



C6-C3-C6
(C6-C3)2
(C6-C3)n


/-OH

0O 0


-QCOOH

&/COCH,

- /-CH2-COOH

-/CH=CH-COOH

\ -CH2-CH=CH2






o
1 O 0 O0


0

0

0

0
a 0


0


see Table 2-6











Classification of food flavonoids


Flavonoid
Chalcones



Dihydrochalcones






Aurones



Flavones




Flavonols





Dihydroflavonol





Flavanones





Flavanol




Flavandiol or
leucoanthocyanidin




Anthocyanidin


Basic Structure


oy-----c
0
0






O


o0
o




0





o




OH
0





O
OH
0
















OH
OH
OH
0


Table 2-6.


OM
OMe


I


I









Table 2-6. Continued

Flavonoid Basic Structure
Isoflavonoids o

O O
Bioflavonoids

0


Proanthocyanidins or
condensed tannins or s
NI






(Source: (Bravo 1998).


Thymol also is a major contributor to the flavor of essential oils from tangerine and

mandarin. Phenolic substances may be responsible for the flavor of a number of spices

and herbs (Shahidi and Naczk 2004; Lule and Xia 2005).

Antioxidant potential

One of the principal roles that have been proposed as part of the actions of

phenolics is that of an antioxidant. Their antioxidant action can arise from a combination

of several chemical events, which include enzyme inhibition, metal chelation, hydrogen

donation from suitable groups and oxidation to a nonpropagating radical. The health

implications of an antioxidant depend on how well it is absorbed by the body and how it

is metabolized, in addition to partition effects (Parr and Bolwell 2000).









Anticarcinogenic action

The possible mechanism of action of anticarcinogens can be classified into two

groups, blocking and suppressing, depending on the point of action. Some compounds

can both block and suppress. The main action of blocking agents is to stimulate the

carcinogen-detoxifying enzymes and to inhibit enzymes which have the potential to

activate precarcinogens into carcinogens (Parr and Bolwell 2000).

Anthocyanins

Anthocyanins are water-soluble pigments responsible for the red to purple to blue

colors in many fruit, vegetables, flowers, and cereal grains. In general, its concentration

in most fruits and vegetables goes from 0.1 to 1% dw. The total content of anthocyanins

varies among fruits and vegetables, their different cultivars and is also affected by

genetic make-up, light, temperature and agronomic factors (Shahidi and Naczk 2004;

Wrolstad 2004).

Since color is one of the most important quality attributes in food, anthocyanin-

rich plant extracts might have a potential use as a natural alternative to food colorants.

Anthocyanins-based colorants are manufactured for food use from horticultural crops

grown for that specific purpose as well as from processing wastes. The interest in

anthocyanins pigments has intensified in recent years because of their possible health

benefits. Thus in addition to their functional role as colorants, anthocyanins extracts

may improve the nutritional quality of foods and beverages (Wrolstad 2004).

Classification

Chemically, anthocyanins are flavonoids and are based on a C15 skeleton. The

anthcyanidins (aglycones) are the basic structure of anthocyanins. They consist of an

aromatic ring A bonded to an heterocyclic ring C that contains oxygen which is also









bonded by a carbon-carbon bond to a third aromatic ring B (C6-C3-C6) (i.e.

anthocyanins are substituted glycosides of salts of phenyl-2-benzopyrilium

(anthocyanidins)) (Figure 2-8) (Delgado-Vargas and others 2000; Gradinaru and others

2003; Castareda-Ovando and others 2009).

The differences in color and stability between anthocyanins are related to the

number of hydroxyl and methoxyl groups, the nature, position, and number of sugars

attached to the molecule, and the nature and number of aliphatic or aromatic acids

attached to sugars in the molecule (Kong and others 2003).


R1
2' 3' OH
8 1 + 1' B
HO, A_ ,0, 2 J


5 4


Aglycon Substitution pattern Amax (nm)
R1 R2 Visible spectra
Pelargonidin H H 494 (orange)
Cyanidin OH H 506 (orange-red)
Delphinidin OH OH 508 (blue-red)
Peonidin OCH3 H 506 (orange-red)
Petunidin OCH3 OH 508 (blue-red)
Malvidin OCH3 OCH3 510 (blue-red)


Figure 2-8. Structural and spectral characteristics of the major naturally occurring
aglycons. (Source: Rodriguez-Saona and Wrolstad 2005).

An increase in the number of hydroxyl groups tends to deepen the color to a

more bluish shade while an increase in methoxyl groups increase redness. Glucose,

galactose, rhamnose, and arabinose are the sugars most commonly found in









anthocyanins, usually as 3-glycosides or 3,5-diglycosides. Rutinosides (2-O-a-L-

rhamnosyl-D-glucosides), sophorosides (6-O-,/-D-glucosyl-D-glucosides), and

sambubiosides (2-O-,/-D-xylosyl-D-glucosides) also occur as well as some 3,7-

diglycosides, and 3-triosides. The most common acylating agents include cinnamic

acids (caffeic, p-coumaric, ferulic, and synaptic) and aliphatic acids (acetic, malic,

malonic, oxalic, and succinic) (Clifford 2000; Delgado-Vargas and others 2000).

There are 17 known naturally occurring anthocyanidins but only six are common

in higher plants: pelargonidin (Pg), peonidin (Pn), cyaniding (Cy), malvidin (Mv),

petunidin (Pt), and delphinidin (Dp). The glycosides of the three non-methylated

anthocyanidins (Cy, Dp and Pg) are the most widespread in nature, being present in

80% of pigmented leaves, 69% of fruits and 50% of flowers. The distribution of the six

most common anthocyanidins in the edible parts of plants is cyanidin (50%),

pelargonidin (12%), peonidin (12%), delphinidin (12%), petunidin (7%), and malvidin

(7%) (Kong and others 2003).

Stability

Anthocyanin pigments can be destroyed easily during processing and storage. A

number of factors influence their stability including pH, temperature, humidity, light,

oxygen, enzymes, as well as the presence of ascorbic acid, sugars, sulfur dioxide or

sulfite salts, metal ions and copigments. The study of anthcoyanins characteristics can

help develop products and processing conditions that will yield high-quality products

(Jackman and others 1987; Clifford 2000; Delgado-Vargas and others 2000; Gradinaru

and others 2003; Mazza and others 2004).

The effective pH range for most anthocyanins colorants is limited to acidic foods

because of the color changes and instability that occur above pH 4 ((Wrolstad, 2004). At









a given pH, equilibrium exists between four different anthocyanin structures: a quinoidal

(anhydro) base, a flavylium cation, and the colorless carbinol pseudo-base and

chalcone (Figure 2-9).

Copigments are substances that contribute to anthocyanins coloration by

protecting the anthocyanin molecule; this mechanism is unique to the anthcyanin family.

Usually copigments have no color by themselves but when added to an anthocyanin

solution they greatly enhance its color (Mazza and Brouillard 1990).


-H+


HO.


"O-gly


quinoidal base: blue
pH=7


y O--gly
O-gly

flavylium cation (oxonium form): orange to purple
pH= 1


+H2 O -H+


HO,


HO.


O-gly


'O-gly


chalcone: colorless carbinol pseudo-base (hemiketal form): colorless
pH = 4.5 pH = 4.5

Figure 2-9. Predominant structural forms of anthcoaynins present at different pH levels.
(Source: Giusti and Wrolstad 2005).









Anthcoyanins react with flavones, alkaloids, amino acids, benzoic acids,

coumarin, cinnamic acids, and a wide variety of other flavilyum compounds. This weak

hydrophobically-driven interaction (van der Waals interactions) is known as

intermolecular copigmentation. Intramolecular copigmentation occurs with the acylation

of the molecule and is more effective than the intermolecular one. The basic role of

copigments is to protect the colored flavylium cation from the nucleophilic attack of the

water molecule (Delgado-Vargas and others 2000).

The association between anthcoyanins and copigments leads to an absorbance

increase in the visible range (hyperchromic effect) and a shift of the Amax toward higher

wavelengths (bathochromic effect) (Gradinaru and others 2003).

Health Benefits

In plants, anthocyanins serve as attractants for pollination and seed dispersal,

give protection against the harmful effects of UV irradiation, and provide anti-viral and

anti-microbial activities (Wrolstad 2004).

Anthocyanins could exhibit multiple biological effects, e.g. reduced risk of

coronary heart disease and stroke, anticarcinogen activity, antioxidant/antiradical

activity, anti-inflammatory action, inhibition of blood platelet aggregation and

antimicrobial activity, treatment of diabetic retinopathy and prevention of cholesterol

induced atherosclerosis (Wang and others 1997; Espin and others 2000; Wrolstad

2004).

Beverage Processing

Beverage processing typically involves an extraction step (juices and teas

extraction), followed by blending were they can be mixed with other ingredients like

water, sweeteners, acidulants, flavorings, colors, and preservatives among others.









Beverages then go into processing, filling, and packaging. The purpose of the

processing and packaging steps is to produce a product that is wholesome and safe for

the consumer (Ashurst 2005).

Thermal Processing

Traditionally beverages are thermally processed. There are five main processes

that use heat as a way to assure microbial safety in juices and soft drinks:

1. Flash pasteurization
2. Hot filling
3. In-pack pasteurization
4. Aseptic filling
5. Chilled distribution

Process selection depends on the level of microbial contamination of the raw

materials and packaging, whether the product composition will favor the growth of

microorganisms, the ability of the product to resist heat, and the desired shelf life

(Ashurst 2005).

Flash pasteurization

Normally the juice is passed through a balance tank or feed tank before being fed

to the pasteurizer. The liquid is generally heated by hot water in a plate or tubular

(spiral) heat exchanger to the desired pasteurization temperature and held at that

temperature for a specified time in a holding tube before being cooled to the filling

temperature (usually ambient) using chilled water. Flash pasteurizers usually have a

regeneration section. The pasteurized product is then sent to a filling machine.

Hot filling

In hot filling, the product is heated (in a heat exchanger), sent to the filler hot and

filled into containers. The containers are closed and are held at or above the desired

temperature for a specified time prior before being cooled. This is usually done in a









tunnel with water sprays. In this system, not only the product but also the container is

heat treated.

In-pack pasteurization

This type of process is generally the most severe and microbiologically most

secure form of heat treatment. The filled closed pack is put into a tunnel pasteurizer

were the treatment is given by means of water sprays at various controlled

temperatures. The pasteurizer is dived into zones. First there is a heating zone where

the temperature of the container and the product is raised, next there is the pasteurizing

zone where the product is held to the pasteurizing temperature for a specified time and

finally there is a cooling zone where the product is cooled below 30 C.

Aseptic filling

Aseptic filling is a special case of flash pasteurization that often uses a higher

temperature profile. For successful aseptic filling, clean containers, clean product,

clean headspace, and clean closures should be brought together in an environment that

prevents recontamination. This operation normally takes place in a closed space over

pressure with sterile air.

Chilled distribution

Flash pasteurized product is filled cold into clean bottles on clear fillers and then

is stored in refrigerated warehouses and is sold to costumers from chill cabinets

(Ashurst 2005).

Dense Phase CO2 Processing (DPCD)

Dense phase carbon dioxide (DPCD) is a cold pasteurization method that uses

C02 under pressures below 50 MPa. This non-thermal technology is mainly used in









liquid foods. Since the food is not exposed to the adverse effect of heat, its fresh-like

physical, nutritional, and sensory qualities are maintained (Damar and Balaban 2006).

Mechanisms of microorganisms' inactivation by DPCD

The exact means of microbial inactivation are not clear but several mechanisms

may be involved:

1. pH lowering effect

C02 can lower the pH of the aqueous parts of the food by dissolving and forming

carbonic acid (H2CO3), which further dissociates into bicarbonate (HCO3-), carbonate

(C03-2) and hydrogen (H+) ions lowering extracellular pH, by the following equilibrium:

CO2 + H20 < H2CO3 H+ + HCO3- H+ + CO32 (1)

C02 penetrates the microbial cell membrane and lowers its internal pH by exceeding

the cell's buffering capacity. This change in internal pH may inactivate microorganisms

by inhibiting essential metabolic systems such as enzymes.

2. Inhibitory effect of molecular CO2 and bicarbonate ion

Bacterial enzymes may be inhibited by C02 by formation of a bicarbonate

complex, excess C02, pH lowering by dissolved C02, sorption/interaction of C02 into

the enzyme molecules, and precipitation of intracellular Ca+2 and Mg+2 carbonates.

3. Physical disruption of cells

The disruption of physical cells was the first mechanisms proposed for

microorganisms' inactivation and suggests that microbial cells' bursting' is due to the

rapid pressure release and the expansion of C02 within the cell.

4. Modification of cell membrane and extraction of cellular components

This mechanism is based on the lipo- and hydrophilicity and solvent

characteristics of C02. Extraction of intracellular substances and their transfer out of the









cell during pressure release may lead to microbial inactivation (Damar and Balaban

2006).

Factors affecting microbial inactivation

Several factors may influence microbial inactivation including:

1. Water activity and water content

DPCD is more effective as aw increases. A higher water content of the treated

product increase C02 solubility and enhances microbial inactivation.

2. Pressure

Since C02 solubility increases with increasing pressure this can help in

microorganisms' inactivation.

3. Temperature (T)

Although solubility of C02 decreases with increasing temperature, higher T can

increase the diffusivity of C02 and the fluidity of the cell membrane which can facilitate

the penetration of C02 into cells. T can also affect the change of C02 from subcritical to

supercritical phase (Tc = 31.1 C) (Figure 2-10). Under supercritical conditions the

penetration power of C02 is higher and at the near-critical region there is a rapid

change in solubility and density of C02 (Damar and Balaban 2006).

4. Initial pH

Low pH facilitates penetration of carbonic acid through the cell membrane

leading to more inactivation.

5. Cell growth phase and age

Young cells are more sensitive and are easier to inactivate than mature cells.














W Supercritical
Solid Liquid fluid phase
Pc = 7.39
MPa
M a Critical point

Gas


Temperature
II



Figure 2-10. Phase diagram of carbon dioxide

6. Type of microorganism

Different bacteria have different susceptibilities to DPCD. It has been suggested

that the nature of the cell wall could be important in the difference in sensitivity between

G(+) and G(-) bacteria. Since G(-) bacteria have thin cell walls, they are expected to be

more sensitive and their cells walls could be more easily ruptured as compared to G(+)

bacteria.

7. Type of treatment system

The system used for DPCD treatment can affect the microbial inactivation rate.

Systems that allow better contact of C02 with the food are more effective in microbial

reduction because of the more rapid saturation of the solution with C02. Usually batch

systems require longer treatment times for microbial inactivation as compared to

continuous systems. Inactivation rate can be increased in batch systems by using

agitation (Damar and Balaban 2006).









Solubility of CO2

In a continuous flow DPCD system, several variables are controlled during

processing: pressure, temperature, residence time, and %C02. The amount of CO2

used should assure a complete saturation of the liquid but since its solubility at

processing conditions is not known this can lead to the use of excess CO2 elevating

production costs.

CO2 solubility in liquid foods can be affected by pressure, temperature, and food

composition. Pressure has a direct effect on CO2 solubility meaning that as pressure

increases, CO2 solubility increases while as temperature increases, solubility of CO2

decreases. Other substances present in the food (composition) either increase or

decrease the solubility of CO2 (Calix and others 2008).

Recent studies have focused on the measurement of CO2 solubility in liquid

model systems and fruit juices. These experiments were done using an experimental

apparatus designed and built at the University of Florida (Calix and others 2008). This

system is designed to saturate a known amount of liquid by bubbling CO2 from the

bottom of a vessel under controlled pressure and temperature and afterwards the CO2

gas is expanded and measured at ambient pressure.

Table 2-7 presents solubility of CO2 for water, orange, apple, and grapefruit juice

measured at different pressures in these studies.

Solubility of CO2 in fruit juices is lower than that of pure water because of the

presence of solutes such as sugars and acids which lowers the amount of CO2 that can

dissolve (Calix and others 2008; Ferrentino and others 2009).









Table 2-7. C02 solubility of liquid foods measured at 40 C

Liquid Pressure (MPa) Solubility Reference
Water 7.58 31.0 4.71 6.32 Calix and others 2008
Orange juice 7.58 15.86 4.10 -4.98 Calix and others 2008
Apple juice 7.58 15.86 3.95 5.01 Calix and others 2008
Grapefruit juice 7.58-31.0 3.97-4.70 Ferrentino and others 2009
g of C02/100 g of liquid.

DPCD treatment systems

Batch, semi-continuous, and continuous systems have been developed for

DPCD applications. In a batch system, C02 and the food to be treated are stationary in

a container during treatment. A semi-continuous system allows a continuous flow of

C02 through the chamber while a continuous system allows flow of both C02 and the

liquid food through the system.

A typical batch system has a C02 gas cylinder, a pressure regulator, a pressure

vessel, a water bath or heater, and a C02 release valve. The sample is placed into the

pressure vessel and temperature is set to the desired value. C02 is then introduced into

the vessel until the sample is saturated at the desired pressure and temperature. The

sample is left in the vessel for a period of time and then the C02 outlet valve is opened

to release the gas. Some systems contain an agitator to decrease the time to saturate

the sample with C02 (Damar and Balaban 2006).

A continuous flow DPCD system was developed in 1999 by Praxair (Chicago, III.,

U.S.A.) (Figure 2-11).

C02 and the product are pumped through the system and are mixed before

entering the high pressure pump, which increases the pressure to the processing levels.









Holding
tube




Ci)
CO, Chiller
tank


Pump
Beverage
stream
Intensifier
S pump Expansion
valve
Processed
Pump beverage



Figure 2-11. Schematic diagram of the continuous flow dense phase C02 system

Product temperature is controlled in holding coils. Residence time is adjusted by setting

the flow rate of the product. At the end of the process, C02 is released by means of an

expansion valve (Damar and Balaban 2006).

DPCD food applications

DPCD has been applied mainly to liquid foods. This technology has been tested

in several products at the University of Florida, Food Science and Human Nutrition

Department using the continuous flow system presented in Figure 2-11, and includes:

orange, mandarin, and grapefruit juice, beer, grape juice, and coconut water among

others.

Orange juice

Several studies with orange juice (OJ) showed that DPCD treatment can improve

some physical, nutritional and quality attributes such as cloud formation and stability,









color, and ascorbic acid retention (Damar and Balaban 2006).

Kincal and others (2005) tested the capacity of the DPCD system to reduce both

natural and inoculated microbial loads of pulp-free Valencia OJ at different pressures

(38, 72, and 107 MPa), temperatures (25 and 34.5 C), residence times (3-10 min), and

C02/juice ratios (0.1-1.0). A storage study was conducted at 1.7 C with juice processed

at 107 MPa, C02/juice ratio of 1.0 and residence time of 10 minutes. Residence time

had the greater influence on microbial reduction, followed by pressure. The C02/juice

ratio and temperature showed not to be the driving forces on microbial load reduction in

this system. They proved that the system was able to achieve a 5-log reduction of the

natural flora in spoiled juice (38, 72, and 107 MPa at 25 and 34.5C, C02/juice ratio of

1.0 and residence time of 10 min), and 5-log decrease of pathogenic Escherichia coli

0157:H7 (107 MPa and residence time of 10 min), Salmonella typhimurium (38, 72, and

107 MPa and residence time of 10 min), and Listeria monocytogenes (38, 72, and 107

MPa and residence time of 10 min). During the refrigerated storage study, they

observed an increase in the bacterial number possibly because of an injury/repair

mechanism of some of the microorganisms or due to post-contamination.

A study was performed by Kincal and others (2006) to treat pulp-free Valencia

OJ at pressures of 38, 72, and 107 MPa, and C02/juice (w/w) ratios from 0.40 to 1.18

with a constant residence time of 10 min.

The highest PE inactivation (46.3%) was obtained when the pressure was 107

MPa and no heat was applied. PE activity decreased with storage time. Cloud increased

between 446 and 846% after treatments and remained 4 times higher than the control

during storage. They found no significant changes in pH and oBrix of treated samples.









TA increased slightly after treatment and remained constant throughout storage. Small

but insignificant increase in L and a values occurred after treatment. Juice color did not

change drastically during storage. Sensory evaluations of DPCD-treated and untreated

OJ were not significantly different after 2 weeks of refrigerated storage at 1.70C

(Balaban and others 2008).

Mandarin juice

Lim and others (2006) processed mandarin juice. The process variables were

temperature (25-45C), pressure (13.8-41.4 MPa), residence time (5-9 minutes) and

%C02 (2-12). They found that temperature and %C02 had a significant effect in log

reduction of total aerobic plate count while pressure and residence time were not

significant. The maximum log reduction (3.47) was observed at 35 C, 41.4 MPa, 9 min

and 7 %C02. PE inactivation ranged from 6.1 to 50.7% and maximum inactivation was

achieved at 45C, 41.4 MPa, 7 min and 7% C02. Cloud was not only retained but

enhanced. The highest cloud increase was 38.4% at 45C, 27.6 MPa, 7 min, and 2%

C02. Lightness and yellowness increased and redness decreased after treatment. pH

and oBrix didn't change after treatment while titratable acidity of treated samples was

higher than the untreated juice (Balaban and others 2008).

Grapefruit juice

Red grapefruit juice was processed using DPCD at pressures of 13.8, 24.1, and

34.5 MPa and residence time of 5, 7, and 9 min at a constant temperature of 40 C and

C02 level of 5.7% to evaluate the effect of treatments on yeasts and molds and total

aerobic bacteria inactivation. A five log reduction for yeasts and molds and total aerobic









bacteria was achieved at 34.5 MPa and 7 min of treatment (Ferrentino and others

2009).

Ferrentino and others (2009) also conducted a storage study with DPCD

processed red grapefruit juice for 6 wk at 4 C. No growth of total aerobic bacteria and

yeasts and molds was observed in the DPCD treated juice. Cloud in the juice increased

91% while PE inactivation was partial (69.17%). No significant differences in oBrix, pH,

and TA were detected between treated and untreated samples while the treated juice

had a higher lightness and redness and lower yellowness. Total phenolics content was

not affected by treatment and storage and slight differences were detected for ascorbic

acid content and antioxidant capacity.

Beer

Dagan and Balaban (2006) studied the effect of DPCD on beer quality. They

predicted a maximum log reduction in yeast population of 7.4 logs at processing

conditions of 26.5 MPa, 21 C, 9.6% C02, and residence time of 4.77 min. DPCD

pasteurization reduced haze from 146 nephelometric turbidity units (NTU) to 95 NTU.

Aroma and flavor of beer at these same conditions was not significantly different when

compared to a fresh beer sample in a difference from control test. Foam capacity and

stability of beer were minimally affected by the process; however these changes were

unnoticed by consumers.

Grape juice

Several studies were conducted with muscadine grape juice testing the effect of

DPCD on microbial reduction, physicochemical, phytochemical and quality changes

after treatment and during storage. Different processing conditions of pressure (1.2-40.2









MPa), C02 levels (0-15.7%), and constant residence time (6.25 min) and temperature

30 C) were evaluated by Del Pozo-lnsfran and others (2006a). Results showed that

processing pressure was a significant factor affecting microbial inactivation but that CO2

content was the processing parameter that had the major influence in microbial log

reduction.

Subsequent storage stability for 10 wks at 4 C with two treatments that achieved

>5 log reduction (34.5 MPa at 8 and 16% C02) were evaluated and compared to a heat

pasteurized juice (75 C, 15 s). Results showed that thermal pasteurization decreased

anthocyanins (16%), soluble phenolics (26%), and antioxidant capacity (10%) whereas

no changes were observed for both DPCD treated juices. DPCD juices also retained

higher anthocyanins (335 mg/L), polyphenolics (473 mg/L), and antioxidant capacity

(10.9 pmol of Trolox equivalents/mL) than thermally pasteurized juices at the end of

storage (Del-Pozo-lnsfran and others 2006a).

Insignificant differences in sensory attributes (color, flavor, aroma, and overall

likeability) were observed between unprocessed and DPCD juices, while significant

differences were observed between unprocessed and heat-pasteurized juices. Panelists

preferred DPCD over heat-pasteurized juices throughout the first 6 weeks of storage but

afterwards the growth of yeast and mold adversely affected juice aroma. Comparable

microbial counts were observed between DPCD and thermally pasteurized juices during

the first 5 weeks of storage (Del-Pozo-lnsfran and others 2006a).

Another study evaluated the phytochemical stability and organoleptic attributes of

an ascorbic acid fortified muscadine grape juice as affected by DPCD processing and









addition of thyme polyphenolic cofactors (Thymus vulgaris; 1:100 anthocyanin-to-

cofactor molar ratio) (Del-Pozo-lnsfran and others 2006b).

DPCD processing insignificantly altered initial juice phytochemical and

antioxidant content, whereas thermal pasteurization reduced anthocyanins (263 mg/L),

ascorbic acid (42 mg/L), soluble phenolics (266 mg/L), and antioxidant capacity (6 pmol

of Trolox equivalents/mL). Similar trends were observed during storage, and data

showed that increasing the C02 level from 8 to 16% during DPCD processing

contributed to the reduction of juice phytochemical and antioxidant degradation.

Copigmentation helped retain higher anthocyanins, soluble phenolics, and antioxidant

capacity during storage without affecting initial juice aroma and flavor characteristics

(Del-Pozo-lnsfran and others 2006b).

A third study by Del Pozo-lnsfran and others (2007) assessed the effect of DPCD

processing on polyphenol oxidase (PPO) activity, polyphenolic and antioxidant changes

in muscadine grape juice under different processing pressures (27.6, 38.3, and 48.3

MPa), C02 levels (0, 7.5, and 15%), and constant residence time (6.25 min) and

temperature (30 C). Pressure alone was responsible for a 40% decrease in PPO

activity that resulted in 16-40% polyphenolic and antioxidant losses. Increasing CO2

from 0 to 7.5% was responsible for an additional 35% decrease in enzyme activity and a

2-fold greater polyphenolic retention. However, insignificant changes in PPO activity or

polyphenolic retention were observed when C02 was increased to 15%.

Subsequently two DPCD conditions (48.3MPa at 0 and 15% C02) were

evaluated for polyphenolic and antioxidant changes during storage (4 oC, 4 wks). Juices

with residual PPO activity following processing resulted in greater polyphenolic (8-10-









fold) and antioxidant capacity (4-fold) degradation compared to control juices with no

PPO activity.

Coconut water

The effects of DPCD on microbial, physical, chemical and sensorial quality of a

coconut water beverage were evaluated by Damar and others (2009). Different

processing conditions of pressure (13.8, 24.1, and 34.5 MPa), temperature (20, 30, and

40 C), %C02 level (7, 10, 13 gC02/100g of beverage), and constant residence time of

6 min were tested. Pressure was not significant in microbial reduction whereas

temperature and %C02 levels were significant.

In the same study, DPCD-treated (34.5 MPa, 25 C, 13% C02, 6 min), heat-

pasteurized (74 C, 15 s) and untreated coconut beverages were evaluated during 9

wks of refrigerated storage (4 C). Total aerobic bacteria of DPCD and heat-treated

samples decreased while that of untreated samples increased to >105 CFU/mL after 9

wks. DPCD increased titratable acidity but did not change pH (4.20) and Brix (6.0).

Likeability of DPCD-treated coconut water was similar to untreated. Heat treated

samples were less liked at the beginning of storage. Off flavor and taste-differences-

from control scores of heated samples were higher than DPCD during the first two

weeks (Damar and others 2009).

Sensory Evaluation

The attributes of a food item are typically perceived in the following order:

appearance (color, size and shape, surface texture), odor/aroma/fragrance, consistency

and texture, and flavor aromaticss, chemical feelings, taste). However in the process of

perception most or all of the attributes overlap (Meilgaard and others 2007).









Sensory tests provide useful information about the human perception of product

changes due to ingredients, processing, packaging, or shelf-life. Sensory evaluation

includes a set of test methods with guidelines and established techniques for product

presentation and well-defined responses, statistical methods, and guidelines for

interpretation of results. There are three primary sensory tests: discrimination tests

(focus on the existence of overall differences among products), descriptive analysis

(specification of attributes), and affective or hedonic testing (measuring consumers likes

and dislikes) (Lawless and Heymann 1998).

Difference-from-Control-Test

A difference-from-control-test is used to determine whether a difference exists

between one or more samples and a control, and to estimate the size of any such

difference. One sample is designated the "control", "reference", or "standard", and all

other samples are evaluated with respect of how different each is from the control

(Meilgaard and others 2007).

Flavor Analysis

Flavor perception depends of the combined responses or our senses and the

cognitive processing of these inputs (Reineccius 2006).

Taste is the combined sensation that derives from specialized taste receptor cells

located in the mouth. It is primarily limited to the tongue and is divided into the

sensations of sweet, sour, salty, bitter, and umami while olfaction is the sensory

component that results from the interaction of volatile food components with olfactory

receptors in the nasal cavity. The stimulus for the aroma of a food can be orthonasal

(the odor stimulus enters the olfactory region directly from the nose as we sniff the food)









or retronasal (the stimulus enters from the oral cavity as we eat a food) (Reineccius

2006).

Most of the techniques used in aroma isolation take advantage of either solubility

or volatility of the aroma compounds (Reineccius 2006).

Solid Phase Micro Extraction (SPME)

SPME is a relatively new technique for the isolation of food aromas. An inert fiber

is coated with an adsorbent. The adsorbent coated fiber is placed in the headspace of a

sample, or the sample itself if liquid, and allowed to adsorb volatiles. The loaded fiber is

then thermally desorbed into a GC carrier gas flow, and the released volatiles are

analyzed. Since SPME is an equilibrium technique, the volatile profile obtained is

strongly dependant of the sample composition and careful control of all sampling

parameters is required (Reineccius 2006).

The effectiveness of SPME techniques depends on many parameters such as :

type of fiber, sample volume, temperature and extraction time, and desorption of

analytes from the fiber (Waldemar and others 2004).

Gas chromatography-Olfactometry (GC-O)

GC-O is a technique only applied to aroma studies. In olfactometric techniques,

the nose is used as a GC detector. The GC system may be set up in such a way that

the column effluent is split so that a portion of the effluent goes to a sniffing port and the

reminder goes to a GC detector (flame ionization (FID) or an MS detector). The GC-O

produces what is called an aromagram, which is a listing of the odor character of each

peak in a GC run. Mass spectroscopy is generally used in the flavor area to either

determine the identity of an unknown or to act as a mass-selective GC detector. The









GC-MS as an identification tool is unique because of its high sensitivity (10-100pg)

(Reineccius 2006).

With consumer demands for natural beverages with health promoting properties

that offer fresh-like sensory attributes, Hibiscus sabdariffa may offer an additional

market in this arena. Previous research has focused on a hot beverage or extraction

process while new technologies focus on minimal to nonthermal processing. Dense

phase carbon dioxide processing may offer an alternative to the traditional hibiscus

processing and provide consumers a product with fresh-like quality and health benefits.

This research focused on three main areas: 1) finding alternatives to the water

extraction conditions that do not involve heat but suitable for nonthermal processing, 2)

comparing DPCD to heat pasteurization by evaluating the physicochemical,

phytochemical, and sensory properties during processing and storage, and 3)

evaluating aroma and phytochemical profiles of water hibiscus extracts obtained from

fresh and dried hibiscus.









CHAPTER 3
EFFECT OF COLD AND HOT WATER EXTRACTION ON THE PHYSICOCHEMICAL
AND PHYTOCHEMICAL PROPERTIES OF HIBISCUS SABDARIFFA EXTRACTS

Introduction

Hibiscus sabdariffa L (family Malvaceae) is a tropical annual shrub. China,

Thailand, Mexico, Egypt, Senegal and Tanzania are among the main producing

countries. In Mexico, this plant is known as "flor de jamaica" or simply "jamaica". The

red calyces are the part of the plant with commercial interest and are rich in organic

acids, minerals, anthocyanins, and other phenolic compounds.

Hibiscus extracts contain two major anthocyanins, delphinidin-3-sambubioside

and cyanidin-3-sambubioside. Their spectral characteristics (Degenhardt and others

2000), MS fragmentation patterns (Giusti and others 1999), and potential antioxidant

(Wang and others 2000) and anticancer (Chang and others 2005; Hou and others 2005)

activities have been previously studied. Similarly, other polyphenolic compounds

including protocatechuic acid (Lee and others 2002; Olvera-Garcia and others 2008),

hibiscus acid and its 6-methyl ester (Hansawasdi and others 2000) have also been

found to be present in hibiscus extracts and have been associated with pharmacological

activities. Differences in variety and extraction conditions (type of solvent, concentration,

time and temperature) potentially affect the polyphenolic profile of the extracts and

makes comparison between studies difficult.

Traditionally, fresh hibiscus is either frozen or dried in the sun for preservation

and used in the production of natural color, flavor extracts and/or beverages.

Preparation of a hibiscus beverage includes an extraction step followed by a

pasteurization method. The use of non-thermal technologies such as dense phase

carbon dioxide, pulsed UV light, high hydrostatic pressure, and pulsed electric fields as









a preservation method does not justify an extraction step that involves high

temperature, and an alternative extraction at a lower temperature should be considered.

The objectives of this study were: (1) to compare the effects of cold (25C) and

hot (90C) water extraction on the physicochemical and phytochemical properties of

hibiscus extracts and (2) to identify and quantify the anthocyanins and major

polyphenolics present in extracts obtained from fresh and dried hibiscus by equivalent

cold and hot water extraction conditions.

Materials and Methods

Chemicals and Standards

Commercial standards of chlorogenic acid, gallic acid, protocatechuic acid, and

quercetin were purchased from Sigma-Aldrich (St. Louis, Mo., U.S.A.). Delphinidin-3-

glucoside and cyanidin-3-glucoside were purchased from Polyphenols Laboratories AS

(Sandnes, Norway). AAPH (2,2'-azobis(2-methylpropionamidine) dihydrochloride),

fluorescein (free acid), Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid)

and Folin-Ciocalteu's reagent were purchased from Sigma-Aldrich (St. Louis, Mo.,

U.S.A.).

Extracts Preparation

Fresh and sun dried Hibiscus sabdariffa (cv. "Criollo") were obtained from

Puebla, Mexico. Hibiscus samples were stored in glass jars, flushed with nitrogen and

kept frozen at -22 C until used. For the first part of the experiment, dried hibiscus was

mixed with distilled water at a ratio of 1:40 (w/v) and maintained at 25 C (cold

extraction, CE) or 90 C (hot extraction, HE) for four different times (30, 60, 120, and

240 min for CE and 2, 4, 8, and 16 min for HE). Eight treatments (TRT's) were tested.

For the second part of the experiment, four extracts were prepared using equivalent









cold and hot water extraction conditions. Fresh (F) and dried (D) hibiscus were mixed

with distilled water at a ratio of 1:4 and 1:40 (w/v) respectively, and extracted at both

25C for 240 min (CE) and 90C for 16 min (HE). For CE, temperature was controlled

using a Constant Temperature Circulator Bath, Model 900 (Fisher Scientific; Pittsburg,

Pa., U.S.A.) and stirring was applied using a stirrer plate Model PC-353 (Corning,

Lowell, Mass., U.S.A.) at speed #4. For HE, a Microprocessor Controlled Water Bath,

Series 280 (Precision Scientific, Winchester, Va., U.S.A.) was used to control

temperature. All the obtained extracts were filtered under vacuum (Whatman filter paper

#4) and their physicochemical and phytochemical properties were measured using the

methods described below.

pH, Total Solids, and Titratable Acidity

pH was measured using a pH meter EA920 (Orion Research, Boston, Mass.,

U.S.A.) and total solids (TS) were determined by difference in weight after drying the

samples at 105 oC for 24 h in an oven (Precision Scientific, Winchester Va., U.S.A.). A

Brinkmann Instrument (Brinkmann Instruments Co., Westbury, N.Y., U.S.A.) consisting

of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was

used to measure titratable acidity (TA). Samples of 10 mL were used and TA was

determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid

(g/100 mL).

Color, Color Density and Hue Tint

Color was measured using a ColorQuest XE colorimeter (HunterLab, Reston,

Va., U.S.A.). Samples (40 mL) were placed in a 20 mm cell and L*, a*, and b

parameters were recorded in total transmittance mode, illuminant D65, 100 observer

angle.. Color density and hue tint were determined by measuring the absorbance (A) at









420, 520, and 700 nm for samples (200 pL) using a spectrophotometer (SpectraMax

190, Molecular Devices, Sunnyvale Calif., U.S.A.) and calculated as:

Color density = (A420 nm A700 nm) + (A520 nm A700 nm) (1)

Hue tint = (A420 nm A700 nm)/(A520 nm A700 nm) (2)

as described by Giusti and Wrolstad (2005).

Anthocyanin Content, Total Phenolics and Antioxidant Capacity

Anthocyanin content was determined by the pH differential method (A510 nm and

A700 nm at pH 1.0 and 4.5, dilution factor (DF) of 4) and expressed in mg/L of delphinidin-

3-glucoside (MW = 465.2, s = 23700) as described by Giusti and Wrolstad (2005). Total

phenolics were measured using the Folin-Ciocalteu assay (A765 nm, DF of 4) and

quantified as gallic acid equivalents (mg/L) (Waterhouse 2005). Absorbance

measurements for anthocyanin content and total phenolics were made using a

SpectraMax 190 spectrophotometer (Molecular Devices, Sunnyvale Calif., U.S.A.).

Antioxidant capacity was evaluated using the oxygen radical absorbance

capacity (ORAC) assay and results were expressed as Trolox equivalents (TE) per

milliliter (pmol of TE/mL) as described by Huang and others (2002) using a SpectraMax

Gemini XPS microplate sprectrofluorometer (Molecular Devices, Sunnyvale, Calif.,

U.S.A.). Data was acquired and analyzed using SoftMax Pro 5.2 software (Molecular

Devices, Sunnyvale, Calif., U.S.A.).

Characterization of Major Polyphenolics

Equivalent cold and hot water extraction conditions (25 C for 240 min (CE) and

90 C for 16 min (HE)) were selected based on the first part of this study. Four hibiscus

extracts were prepared: dried hibiscus cold water extract (DCE), dried hibiscus hot









water extract (DHE), fresh hibiscus cold water extract (FCE), and fresh hibiscus hot

water extract (FHE). LC-MS and HPLC analysis were performed in order to identify the

major polyphenolic compounds including anthocyanins present in these extracts.

LC-MS identification

Chromatographic analyses were performed on an Agilent 1200 series HPLC

(Agilent, Palo Alto, Calif., U.S.A.) equipped with an autosampler/injector and diode array

detector. A Dionex C18 5 pm 120A column (250 x 4.6 mm) was used for compound

separation (Dionex, Sunnyvale, Calif., U.S.A.). Mobile phases consisted of water (phase

A) and 60% methanol in water (phase B), both adjusted to pH 2.4 with formic acid. A

gradient solvent program ran phase B from 0% to 60% in 20 min; 60% to 100% in 20

min; 100% for 7 min; 100% to 0% in 3 min and final conditions were held for 2 min. The

flow rate was 0.8 mL/min, and detection was done at 260, 280, 320, 360 and 520 nm.

Electrospray ionization mass spectrometry (ESI-MS) was performed with a HCT

series ion trap mass spectrometer (Bruker Daltonics, Billerica, Mass., U.S.A.). Column

effluent was monitored in positive and negative ion mode of the MS in an alternative

manner during the same run. Other experimental conditions on the mass spectrometer

were as follows: nebulizer, 45 psi; dry gas (nitrogen), 11.0 L/min; dry temperature, 350

oC; ion trap, scan from m/z 90 to 1000; smart parameter setting (SPS), compound

stability, 50%; trap drive level, 60%. The mass spectrometer was operated in Auto MS2

mode. MS2 was used to capture and fragment the most abundant ion in full scan mass

spectra.

Polyphenolics were identified by comparison of UV/vis (190-660 nm) spectral

interpretation, retention time, comparison to standards, and MS fragmentation patterns.









HPLC quantification

Anthocaynins and polyphenolics were quantified using a Dionex HPLC system

equipped with an autosampler/injector and diode array (PDA 100) detector (Dionex,

Sunnyvale, Calif., U.S.A.). Compounds were separated on a 250 x 4.6 mm Dionex C18

5 |tm 120A column (Dionex, Sunnyvale, Calif., U.S.A.). Mobile phases consisted of

water (phase A) and 60% methanol in water (phase B), both adjusted to pH 2.4 with o-

phosphoric acid. A gradient solvent program ran phase B from 0% to 60% in 20 min;

60% to 100% in 20 min; 100% for 7 min; 100% to 0% in 3 min and final conditions were

held for 2 min. The flow rate was 0.8 mL/min, and detection was done at 260, 280, 320,

360 and 520 nm.

Statistical Analysis

Each treatment condition was repeated in triplicate. Analysis of variance

(ANOVA) and mean separation using Tukey's test (a=0.05) were performed to evaluate

the differences between extraction times, temperatures, and treatments using SAS 9.0

Statistical software (SAS Institute Inc., Cary, N.C., U.S.A.).

Results and Discussion

Effect of Extraction Conditions

The physiochemical and phytochemcial parameters measured for the eight

hibiscus treatments analyzed are presented in Table 3-1 and Figures 3-1, 3-2, and 3-3.

Color density (CD), anthocyanins content (AC), total phenolics (TP), and

antioxidant capacity (AOX) increased with increasing time for both extraction

temperatures (25 and 90 C) while L values decreased with time in both cases. There

were no significant differences between the first two (30 and 60 min, and 2 and 4 min)









and last two (120 and 240 min, and 8 and 16 min) extraction times at both temperatures

for hue tint (HT) values; although the latter times were higher.

For cold water extraction (25 C), time had a significant effect (p < 0.0001) in all

the parameters measured but pH. Total solids (TS) and b increased from 30 to 60 min

and afterwards remained constant (measurements at times 60, 120, and 240 min were

not significantly different). Titratable acidity (TA) increased from 30 to 120 min and

remained constant at 240 min (measurements at 120 and 240 min were not significantly

different) while a values increased from 30 to 60 min, remained constant from 60 to 120

min and decreased at 240 min. For hot water extraction (90 C), time had a significant

effect (p < 0.0335) in all the measured parameters. pH increased from 2 to for 4 min and

remained constant until 16 min. TS and TA increased until 8 min and remained constant

until 16 min. a values were constant for times 2 and 4 min and decreased at 8 and 16

min while b values were constant at 2, 4, and 8 min, and decreased at 16 min.

There was a significant effect (p < 0.0001) of treatment conditions (temperature +

time) in all the measured parameters but pH. Treatments CE1 and HE1 were equivalent

in TS and TA (Table 3-1), treatments CE2 and HE2 were equivalent in TS, TA, b*, CD,

and AC (Figure 3-1). Treatments CE3 and HE3 were equivalent in TS, TA, and AC

while treatments CE4 and HE4 were equivalent in TS, TA, CD, and AC.

L values were significantly lower (darker color) in hot water extracts as

compared to cold water ones while a values were slightly higher in the cold water

extracts. Hue tint (HT) is a measurement of color degradation in anthocyanin containing

products. From Table 3-1, it can be observed that the extracts obtained with cold water

have lower values than the ones obtained with hot water. This indicates that









temperature had an effect on hibiscus extract's color, and thus anthocyanins. A higher

hue tint value is associated with an increase in absorbance at 420 nm (yellow tones) in

relation to that at 520 nm (red tones); this is undesirable because it is an indication of

anthocyanins degradation

Anthocyanin content was not significantly different between treatments CE2 and

HE2, CE3 and HE3, and CE4 and HE4 (Figure 3-1) so equivalent cold and hot water

extraction conditions for anthocyanins were found. As can be seen from Figure 3-2, TP

extraction was better with hot water (90 C) than with cold water (25 C). Prenesti and

others (2007) also found that hot water (100 C for 3 min) extracted a higher phenolic

content compared to cold water hibiscus extracts. The higher concentration of

polyphenolic compounds other than anthocyanins in hot water extracts may have

contributed to a higher antioxidant activity in these extracts as compared to cold water

extracts (Figure 3-3). Tsai and others (2002) found that hibiscus anthocyanins

contributed to 51% of total antioxidant capacity and that other phenolic compounds

were responsible for the remainder of activity.

Qualitative differences were observed between the cold and hot water hibiscus

extracts. Cold extracts had a clear appearance and bright red color whereas the hot

extracts presented a more opaque red color and some haze possibly associated to a

higher concentration of phenolic compounds other than anthocyanins.

Parameters Correlations

Several correlations between the parameters measured were found in both

extraction processes with r2 > 0.9. Linear regression parameters and correlation

coefficients are presented in Table 3-2. This linear behavior will only be valid before

reaching equilibrium during the extraction process, after which, the values of the









measured parameters will remain almost constant. Similarly the equilibrium for the

extraction of anthocyanins and other polypheolic compounds could be reached at

different times. Since some of the methods used in this study can be easier or faster

than others, these equations could be used to predict parameters measured with more

time consuming methods within the range of studied extraction conditions.

L and CD values were inversely correlated. A high L value (lighter color) will be

associated with a low CD value. Color density, anthocyanins content, total phenolics

and antioxidant capacity were all directly correlated.

Polyphenolics Identification

Anthocyanins and other polyphenolics identified in hibiscus extracts are

presented in Figure 3-4 and Tables 3-3 and 3-4. Compounds were identified on the

basis of their retention time, absorption spectrum, MS fragmentation pattern, and where

possible by comparison to an authentic standard.

Peak 1 (tR, 13.4 min; Amax, 271 nm) was identified as gallic acid by comparison of

the absorption spectrum with a standard. This was confirmed by MS-MS analysis that

showed the presence of a negatively charged molecule ion ([M-H]-) at m/z 169 which

fragmented to produce a secondary fragment ion (MS2) at m/z 125 (see Table 3-4). The

presence of gallic acid in hibiscus extract was measured previously by GC-MS

(Mourtzinos and others (2008).

Peak 2 (tR, 17.1 min; Amax, 259 nm) was identified as protocatechuic acid

glucoside. The absorption spectrum was compared to a protocatechuic acid standard;

the presence of the glucose molecule slightly shifted the retention time. MS analysis of

the peak reveled a [M-H]- at m/z 315 that fragmented to yield the ion m/z 153 which









corresponds to protocatechuic acid. The difference between ions 315 and 153 gave an

ion with m/z 162 which corresponds to glucose. The same MS fragmentation patterns

for protocatechuic acid glucoside were reported in dried plum (Fang and others 2002).

Protocatechuic acid isolated from hibiscus extracts was demonstrated to have anti-

atherosclerosis (Lee and others 2002), antitumor (Olvera-Garcia and others 2008;

Tseng and others 1998), antioxidant (Lin and others 2003), and anti-inflammatory (Liu

and others 2002) activities.

Peaks 3 (tR, 18.7 min; Amax, 326 nm), 4 (tR, 23.0 min; Amax, 327 nm), 5 (tR, 23.6

min; Amax, 327 nm), and 7 (tR, 24.3 min; Amax, 331 nm) were identified as caffeoylquinic

acids which are esters formed between caffeic and quinic acid. Their identification was

based on previously developed structure-diagnostic hierarchical keys (Clifford and

others 2003), UV-vis spectrum and retention time was compared relative to a

commercial 5-caffeoylquinic acid (chlorogenic acid) standard. Peaks 3, 4, 5, and 7

produced a [M-H]- at m/z 353 and MS2 ions at m/z 191 (corresponds to quinic acid), 179

(corresponds to caffeic acid), 173, and 135, (peak 7 only had MS2 ions at m/z 191 and

173). Peak 4 was identified as 5-caffeoylquinic acid (5-CQA) by comparison with an

authentic standard. According to Clifford and others (2003) 5-CQA is characterized by

an intense base peak at m/z 191 and a weak secondary ion at m/z 179. Peak 3 was

identified as 3-caffeoylquinic (3-CQA) acid since it is characterized by a base peak at

m/z 191 and a relatively intense secondary ion at m/z 179 while peak 5 was identified

as 4-caffeoylquinic (4-CQA) acid with a characteristic base peak at m/z 173 (Clifford

and others 2003). Peak 7 was tentatively identified as a caffeoylquinic acid isomer from

its absorption spectrum and fragmentation patterns (see Table 3-4). The presence of 5-









CQA has been previously reported in hibiscus extracts (Mourtzinos and others 2008;

Segura-Carretero and others 2008).

Peaks 6 (tR, 24.2 min; Amax, 529 nm) and 8 (tR, 26.2 min; Amax, 521 nm) were

identified as delphynidin-3-sambubioside (D3S) and cyanidin-3-sambubioside (C3S)

which are the two major anthocyanins present in hibiscus (Table 3-3). Identification was

based on their absorption spectrum and MS fragmentation patterns which have been

previously reported (Juliani and others 2009; Degenhardt and others 2000; Giusti and

others 1999). The difference between the MS of the molecule (597) and the aglycone

(303) for D3S gave a m/z of 294 which corresponds to xylose-glucose (132+162) known

as sambubiose. Similarly, the MS for the C3S molecule (581) and the aglycone (287)

yields the sambubiose disaccharide.

Peaks 9 (tR, 29.0 min; Amax, 359 nm), 10 (tR, 30.9 min; Amax, 348 nm), and 11 (tR,

32.0 min; Amax, 356 nm) were tentatively identified as flavonols for their characteristic

absorption spectrum with Amax ~360 nm. Peak 12 (tR, 35.7 min; Amax, 355 nm) was also

tentatively identified as quercetin-3-rutinoside by its absorption spectrum and MS

fragmentation patterns which revealed a base peak at m/z 609 and MS2 at m/z 301. The

difference between m/z 609 and 301 gave a m/z of 308 that corresponds to the

disaccharide rutinose formed between rhamnose (m/z 146) and glucose.(m/z 162).The

presence of rutinose has been previously reported in hibiscus extract as part of an

anthocaynin (cyanidin-3-rutinoside) by Segura-Carretero and others (2008). Quercetin-

3-rutinose with the same MS fragmentation patterns was found in black and green tea

(Del Rio and others 2004) and in pear skins (Lin and Harnly 2008).









Polyphenolics Quantification

Polyphenolics were quantified in the four hibiscus extracts studied (dried hibiscus

cold water extract (DCE), dried hibiscus hot water extract (DHE), fresh hibiscus cold

water extract (FCE), and fresh hibiscus hot water extract (FHE) (Table 3-5). Results

were expressed in milligrams per L of extract. Hydroxybenzoic acids accounted for ~2%

of the total polyphenolics quantified in the dried hibiscus extracts and ~0.5% in the fresh

hibiscus extracts. Caffeoylquinic acids accounted for ~45% and ~38% of total in dried

and fresh extracts, respectively while flavonols accounted for ~10% of the total in all

four extracts. Anthocyanins accounted for ~45% and ~50% of the total in the dried and

fresh hibiscus extracts, respectively.

As seen in Table 3-5, the DHE sample had the highest concentration of total

polyphenols followed by DCE, FCE, and FHE. Gallic acid was not detected in the fresh

extracts and its presence in the dried hibiscus extracts could be attributed to a

breakdown of another phenolic compound during the drying process. The concentration

of protocatechuic acid glucoside was higher in fresh hibiscus extracts and a significantly

lower concentration of caffeoylquinic acids was also observed compared with the dried

extracts.

Hibiscus anthocyanins distribution was ~68% and 64% of the total for D3S and

32 and 36% for C3S in dried and fresh extracts, respectively. This indicated that a

significantly higher concentration of C3S was found in the fresh hibiscus extracts as

compared to dried extracts. Delphinidin-3-sambubioside was present in a significantly

higher concentration in the hot water extracts as compared to the cold water ones but

no significant differences were found in the concentration of cyanidin-3-sambubioside in

the cold and hot water extracts for both fresh and dried hibiscus.









Conclusions

Equivalent cold and hot water conditions were found for anthocyanins extraction

of dried hibiscus. Similar polyphenolic profiles were observed between fresh and dried

hibiscus extracts although differences were found in the concentration of compounds.

Hydroxybenxoic acids, caffeoylquinic acids, flavonols and anthocyanins constituted the

polyphenolic compounds identified in hibiscus extracts. Findings of this research can

provide more flexibility to hibiscus processing. Extraction process selection for industrial

applications should consider availability of raw material (fresh or dried hibiscus),

processing technology, time, and economic considerations.









Table 3-1. Measured pH, total solids (TS) (g of solids/100 mL of extract), titratable
acidity (TA) (g of malic acid/100 mL of extract), and color (L*, a*, b* values,
color density (CD) and hue tint (HT)) for the extracts.

T time
TRT (C) (min) pH TS TA L* a* b* CD HT

CE1 25 30 2.37a* 0.68d 0.28d 54.18a 65.85d 45.39e 1.04f 0.35cd

CE2 25 60 2.32a 0.92bC 0.38abc 46.29b 67.65a 66.92abc 1.82de 0.35d

CE3 25 120 2.32a 0.97ab 0.40ab 43.78cd 67.50a 68.76a 2.05c 0.36c

CE4 25 240 2.31a 1.00ab 0.44a 40.79ef 67.16ab 68.22ab 2.55ab 0.36c

HE1 90 2 2.37a 0.79cd 0.33cd 44.82bc 66.73bc 65.19C 1.73e 0.38b

HE2 90 4 2.37a 0.90bc 0.37bc 42.13de 66.39C 67.03abc 2.00cd 0.38b

HE3 90 8 2.36a 0.95ab 0.39ab 39.34f 65.65d 65.70bc 2.34b 0.39a

HE4 90 16 2.33a 1.08a 0.43a 35.269 63.93e 60.29d 2.70a 0.39a
CE = Cold extraction, HE = Hot extraction. Data represents the mean of n=9. Values with similar letters
within columns are not significantly different (Tukey's HSD, p > 0.05).


time (min)
0 2 4 6 8 10 12 14 16 18



b

d






0 30 60 90 120 150 180 210 240 270
time (min)


---25 C
--90 C


Figure 3-1. Total anthocyanins content expressed as delphinidin-3-glucoside (mg/L) for
the extracts. The upper time scale belongs to the 90 C curve and the lower
time scale belongs to the 25 C curve. Data represents the mean of n=9.
Values with similar letters within the figure are not significantly different
(Tukey's HSD, p > 0.05).











time (min)
0 2 4 6 8 10 12 14 16 11

a



or

3 6 9


0 30 60 90 120 150 180 210 240 27


---25 C
---90 C


time (min)


Figure 3-2. Total phenolics content expressed as gallic acid equivalents (mg/L) for the
extracts. The upper time scale belongs to the 90 C curve and the lower time
scale belongs to the 25 C curve. Data represents the mean of n=9. Values
with similar letters within the figure are not significantly different (Tukey's
HSD, p > 0.05).





time (min)


E
LU
4-
0
0x
Q


0 2 4 6 8 10 12 14 16 11


a
bc








0 30 60 90 120 150 180 210 240 27


---25 C
---90 C


time (min)

Figure 3-3. Antioxidant capacity (pmol of TE/mL) L) for the extracts. The upper time
scale belongs to the 90 C curve and the lower time scale belongs to the 25
C curve. Data represents the mean of n=9. Values with similar letters within
the figure are not significantly different (Tukey's HSD, p > 0.05).









Table 3-2. Linear regression and correlation coefficients between measured parameters
for cold and hot water extraction processes.

Cold Extraction Hot Extraction
m b r2 m b r2
L* vs color density -8.77 62.62 -0.99 -9.61 61.48 -0.99
Color density vs anthocyanins content 0.04 -0.04 0.96 0.03 0.34 0.96
Anthocyanins content vs total phenolics 0.18 2.56 0.95 0.17 -3.65 0.95
Total phenolics vs antioxidant capacity 38.85 13.66 0.92 44.96 13.66 0.93
m = equation slope, b = equation intercept.


Table 3-3. Identification of anthcocyanins present in hibiscus using their spectral
characteristics with HPLC-DAD and positive ions in LC-MS and MS2.

HPLC-DAD Data LC-MS Data (m/z)
Peak Compounda tR (min) Amax (nm) MS (molecule) MS2 (aglycone)
6 Dpd-3-sambubioside 24.2 529 597 303
8 Cyd-3-sambubioside 26.2 521 581 287
a Abbrevaitions: Dpd, delphinidin; Cyd, cyanidin.



Table 3-4. Identification of polyphenolics present in hibiscus using their spectral
characteristics with HPLC-DAD and negative ions in LC-MS and MS2, and
respective standards.

HPLC-DAD
Data LC-MS Data (m/z)
MS2
tR Amax MS2
Peak Compound (min) (nm) MS base peak other peaks
1 Gallic acida 13.4 271 169 125 (100)d
2 Protocatechuic acid glucosideb 18.2 260 315 153 (100)
3 3-caffeoylquinic acid 18.7 326 353 191 (100) 179 (58), 173 (7), 135 (14)
4 5-caffeoylquinic acida 23.0 327 353 191 (100) 179 (2), 173 (0.4), 135 (0.70)
5 4-caffeoylquinic acid 23.6 327 353 173 (100) 191 (20), 179 (39), 135 (14)
7 Caffeoylquinic acid isomer0 24.3 331 353 191 (100) 173 (3)
12 Quercetin-3-rutinosidec 35.7 355 609 301 (100)
a w c


Confirmed with authent]ic stalndard.ofrec E R ~a~aGo
d Values in parenthesis indicate the intensity of the ion.


acid. Tentatively id d.















00 mAU WVL 520 nm

50-
8 A


FHE

min
00 50 100 150 200 250 300 350 400 450 52
00 50 100 150 200 20 30 50 40 505


WVL 360 nm


3

4

7 11

ilLJ 1LIn
A- l^ U v ^ .JJ


FHE
DHE


0 15 0 00 0 200 250 300 350 400 450


3

45

5 7


WVL 320 nm


C



mmin


50 100 150 2 0 250 3 0 350 4 0 45 0


WVL 280 nm


1

-| |__^j


4 6

I5 7


D



mim


50 100 150 200 250 300 350 400 450


6
4
8 11
5 7f 12
9 10
'J7 UA


00 50 100 150 200 250 300 350 400 450 520


Figure 3-4. HPLC chromatograms of dried hibiscus (DHE) and fresh hibiscus (FHE) hot
water extracts: (A) 520 nm, (B) 360 nm, (C) 320 nm, (D) 280 nm, and (E) 260
nm. For peak identification see Tables 3-3 and 3-4.





80


450C mAU


2.200 mAU


FHE
DHE


mAU


FHE
DHE


mAU





FHE
DHE


3



2

.-lA^A~L~_-J--V^-'
.JLn ^xULLA-- u1


WVL 260 nm


E



mlm









Table 3-5. Polyphenolics content (mg/L) of hibiscus samples analyzed in this study.

Compound (peak)e
DCE DHE FCE FHE
Hydroxybenzoic acids
Gallic acid (1)f 0.65a 0.58a nd' nd
Protocatechuic acid glucoside (2)9 0.06b 0.05b 0.19a 0.13ab
Total 0.71 0.63 0.19 0.13
Caffeoylquinic acidsh
3-caffeoylquinic acid (3) 67.53b 73.01a 51.97c 49.700
5-caffeoylquinic acid (4) 43.64b 46.23a 39.050 38.15c
4-caffeoylquinic acid 5) 17.52b 18.81a 14.12c 13.510
Caffeoylquinic acid isomer (7) 3.93b 4.13b 10.86a 12.09a
Total 132.62 142.18 116.00 99.94
Flavonols'
Unidentified (9) 5.56b 5.70ab 5.55b 5.84a
Unidentified (10) 5.52a 5.58a 5.11b 4.86c
Unidentified (11) 10.21b 9.99b 12.29a 12.10a
Quercetin-3-rutinoside (12) 8.45a 9.20a 9.38a 9.21a
Total 29.74 30.47 32.33 32.01
Anthcoyanins
Delphinidin-3-sambubioside (6) 87.32c 100.90a 87.76c 96.16b
Cyanidin-3-sambubioside (8)k 41.62b 44.88b 50.30a 50.89a
Total 128.94 145.78 138.06 147.05
Total phenolic compounds 292.01 319.06 286.58 279.13
Data represents the mean of n=6. d Values with similar letters within rows are not significantly different
(Tukey's HSD, p > 0.05). e Peak numbers refer to the compounds identified in Tables 3-3 and 3-4. fg, hj k
Quantified with gallic acid, protocatechuic acid, chlorogenic acid, quercetin, delphinidin-3-glucoside, and
cyanidin-3-glucoside standards, respectively. Abbreviations: nd, not detected.









CHAPTER 4
AROMA PROFILES OF BEVERAGES OBTAINED FROM FRESH AND DRIED
HIBISCUS

Introduction

Hibiscus sabdariffa commonly known as hibiscus or roselle, grows in many

tropical and subtropical countries and is one of the highest volume specialty botanical

products in international commerce (Plotto 1999). Hibiscus is an annual herbaceous

shrub and is a member of the Malvaceae family. The swollen calyces, which are red

and cup-like, are the part of the plant of commercial interest (Morton 1987; De Castro

and others 2004). Fresh and dried hibiscus calyxes are used to prepare cold and hot

beverages. Sweeteners and spices can be added depending on the region where it is

consumed.

Extensive work has been done in the area of hibiscus anthocyanins due to their

beneficial health effects, high antioxidant properties, and potential source as a food

colorant (Tee and others 2002; Tsai and others 2002; Tsai and Huang 2004; Prenesti

and others 2007: Sayago-Ayerdi and others 2007). Studies with human patients have

also shown that the regular consumption of hibiscus extract has an antihypertensive

effect (Haji Faraji and Haji Tarkhani 1999; Herrera-Arellano and others 2004) and

reduces serum cholesterol in men and women (Lin and others 2007).

Hibiscus flavor is a combination of sweet and tart. Few studies have been done

related to hibiscus flavor. Gonzalez-Palomares and others (2009) identified 20 volatile

compounds in hibiscus extract using SPME and GC-MS, including terpenoids, esters,

hydrocarbons, and aldehydes. They also found 14 compounds in reconstituted spray

dried extracts from which only 10 were present in the original extract and the other 4

were products of degradation. Thermally generated volatiles from untreated, frozen, hot-









air-dried at 50 C, and hot-air-dried at 75 C hibiscus by steam distillation were analyzed

by GC and GC-MS (Chen and others 1998). They characterized more than 37

compounds including fatty acid derivatives, sugar derivatives, phenol derivatives, and

terpenes.

The objective of this study was to determine the aroma profile differences

between four extracts obtained from fresh and dried hibiscus extracted at two different

conditions, by GC-MS and GC-olfactometry.

Materials and Methods

Sample Preparation

Fresh and sun dried hibiscus (Hibiscus sabdariffa cv. "Criollo") were obtained

from the same harvest (November 2006 January 2007) from Puebla, Mexico. Hibiscus

samples were stored in glass jars, flushed with nitrogen and kept frozen at -20 C until

used. Four different extracts were prepared; fresh (F) and dried (D) hibiscus were mixed

with distilled water in a ratio of 1:4 and 1:40 (w/v) respectively and extracted at 22 C for

240 min (cold extraction (CE)) and 98 C for 16 min (hot extraction (HE)). Extraction

ratios (hibiscus: water) were determined based on moisture content of fresh (90%) and

dried (9%) hibiscus (measured at 105C for 24 h in an oven (Precision Scientific,

Winchester, Va., U.S.A.)). Stirring at low speed was applied for cold extraction and no

stirring was applied for hot extraction. After extraction, samples were filtered under

vacuum using Whatman filter paper #4.

The pH of the samples was measured using a pH meter EA920 (Orion Research;

Boston, Mass., U.S.A.) and oBrix was determined with an ABBE Mark II refractometer

(Leica Inc.; Buffalo, N.Y., U.S.A.).









Headspace Volatiles Sampling

Headspace volatiles were extracted and concentrated using SPME. Ten milliliters

of hibiscus extracts were added to a 22 mL screw cap amber glass vial PTFE/silicone

septa containing a small stir bar. Samples were equilibrated for 20 min in a water bath

at 40 C, and hibiscus headspace volatiles were extracted for 30 min using a 1 cm

50/30 mm DVB/Carboxen/PDMS SPME fiber (Supelco, Bellefonte, Pa., U.S.A.). Before

each exposure the fiber was cleaned for 5 min in the injection port (200 C) of the GC-O

or GC-MS instruments.

GC-O Analysis

GC-O analysis was carried out using a HP 5890 Series II Plus GC (Palo Alto,

Calif., U.S.A) with a sniffing port and a flame ionization detector (FID). Hibiscus volatiles

from the SPME fiber were desorbed in the GC injection port. A SPME injector liner

(SPME injection sleeve, 0.75 mm i.d., Supleco; Bellefonte, Pa., U.S.A.) was used. The

GC column effluent was split between the FID and the olfactometer. The injector

temperature was 200 C, and the detector temperature was 250 C. Helium was used

as the carrier gas at 1.67 mL/min. The oven was programmed from 35 C (held for 5

min) to 250 C at 6 C/min with a final hold of 10 min. Volatiles were separated using a

DB-5 (30 m x 0.32 mm. i.d. x 0.5 |tm, J&W Scientific; Folsom, Calif., U.S.A.) or a DB-

Wax (30 m x 0.32 mm. i.d. x 0.5 |tm, Restek; Bellefonte, Pa., U.S.A.) column. Two

olfactory assessors were employed. Samples were sniffed two times by each assessor

in each column.

Aroma descriptions and approximate times were recorded for every sample.

Assessors indicated the intensity of each aroma peak using a linear potentiometer with









a 0-1 V signal. Aromagrams and FID chromatograms were recorded and integrated

using Chrom Perfect 4.4.23 software (Justin Innovations, Inc.; Palo Alto, Calif., U.S.A.).

A peak was considered aroma active only if at least half the panel found it at the same

time with a similar description. Linear retention index values were determined for both

columns using a series of alkanes (C5-C25) run under identical conditions.

GC-MS Analysis

Mass Spectrometry (GC-MS) was used to identify the odor-active volatiles

detected in the GC-O experiment. GC-MS analysis was conducted using a HP 6890 GC

coupled with a MSD 5973 (Agilent Technologies; Palo Alto, Calif., U.S.A.). Hibiscus

volatiles from the SPME fiber were desorbed in the GC injection port (splitless mode) at

200 C. The fiber was removed after 5 min exposure in the injection port. Volatiles were

separated using a DB-5 (30 m x 0.32 mm. i.d. x 0.5 pm, J&W Scientific; Folsom, Calif.,

U.S.A.) or a DB-Wax (30 m x 0.32 mm. i.d. x 0.5 |tm, Restek; Bellefonte, Pa., U.S.A.)

column. The oven was programmed from 35 C (held for 5 min) to 250 oC at 6 oC/min

with a final hold of 10 min. Helium was used as the carrier gas at 1.67 mL/min. The

mass spectrometer was operated in the total ion chromatogram (TIC) at 70 eV. Data

were collected from 35 m/z to 400 m/z. All samples were run in duplicate in each

column. Chromatograms were recorded and integrated using Enhanced Chemstation

(version 01.00) software (Agilent Technologies; Palo Alto, Calif., U.S.A.). Mass spectral

matches were made by comparison with NIST 98.1 (NIST; Gaithersburg, Md., U.S.A.)

and WILEY 8.1 (Wiley; New York, N.Y., U.S.A.) mass spectral libraries. Only those

compounds with spectral fit values equal to or greater than 850 were considered









positive identifications. Linear retention index values were determined for both columns

using a series of alkanes (C5-C25) run under identical conditions.

Identification Procedures

Identifications were based on the combined matching of retention indices (LRI

values) from DB-5 and DB-Wax columns, matches made from spectra in the NIST and

WILEY libraries, aroma descriptors, and linear retention index matches from literature.

Statistical Analysis

Analysis of variance (ANOVA) and mean separation using Tukey's test (a=0.05)

were performed to evaluate differences in pH and oBrix between the analyzed samples

using SAS 9.0 Statistical software (SAS Institute Inc., Cary, N.C., U.S.A.).

Results and Discussion

Four samples using fresh frozen and sun dried hibiscus were prepared: DHE

(dried hibiscus hot water extraction), DCE (dried hibiscus cold water extraction, FHE

(fresh hibiscus hot water extraction), and FCE (fresh hibiscus cold water extraction).

Extraction conditions and measured pH and oBrix values are presented in Table 4-1.

There was not a significant difference (p = 0.0581) in oBrix between the four samples.

Samples prepared with dried hibiscus had a significantly (p = 0.0003) lower pH as

compared to those prepared with fresh hibiscus, reason unknown.

Hibiscus Volatiles Composition

Hibiscus volatiles were divided into five chemical groups. A total of 14 aldehydes,

10 alcohols, 5 ketones, 2 terpenes, and 1 acid were identified. The relative difference in

total volatiles in terms of peak area was normalized to total peak area of DHE (dried

hibiscus hot water extraction) which was 590. Results are shown in Figure 4-1. In

relation to DHE, total peak area was reduced by 40, 59, and 98% for DCE, FHE, and









FCE samples, respectively. In both fresh and dried hibiscus, hot water extraction gave a

higher concentration of volatiles as compared with cold extraction which indicates that

temperature facilitates the extraction process but it can also lead to undesirable

degradation reactions of hibiscus aromas. In the same way, dried hibiscus extracts had

a higher concentration of volatiles as compared to fresh hibiscus. A bigger gradient in

moisture content between the dried hibiscus (9% moisture) and the extraction solvent

(water (100% moisture)) as compared to the gradient between the fresh hibiscus (90%

moisture) and water (100% moisture) may have helped make the extraction process

faster, and thus increased the concentration of volatiles extracted from the dried

hibiscus. Composition of the four samples was similar but there were major quantitative

differences. Aldehydes comprised the largest group of volatiles contributing between 57

and 63% of the total in the hot water extracts and from 37 to 45% in the cold water

extracts, followed by alcohols (23 to 24% in hot water extracts and 28 to 36% in the cold

water extracts), ketones (7-12%), acids (4-8%), and terpenes (2-3%). In the case of

FCE, acids (19%) were higher than ketones. Aldehydes were present in a higher

proportion in hot water extracts while alcohols were present in a higher proportion in

cold water extracts. This may indicate that extraction temperature could influence the

aroma profiles of the obtained extracts by accelerating the degradation or formation of

compounds.

GC-MS Identifications

A total of 32 volatiles were identified using GC-MS in hibiscus samples, 15 of

which were not identified before in hibiscus. Limonene, linalool, a-terpineol, eugenol,

and furfural were previously identified in two studies (Gonzalez-Palomares and others

2009; Chen and others 1998) while nine other compounds (hexanal, heptanal, octanal,









nonanal, 2-heptenal (E), 5-methyl-furfural, 1-hexanol, dehydroxylinalool oxide b, acetic

acid) and two other compounds (decanal and benzaldehyde) were also found in

hibiscus extracts by Chen and others (1998).and Gonzalez-Palomares and others

(2009).respectively. Differences in volatile profiles among studies can be attributed to

the hibiscus variety used and the extraction methods. The extraction solvent polarity as

well as the extraction conditions (time, solute concentration, and temperature) may

impact the aroma profile of the final product.

Table 4-2 lists the 32 volatiles detected in this study. To compare the volatiles in

the four extracts, peak areas were normalized on the single largest peak found in all

samples. This peak was the nonanal peak in the DHE sample. It was assigned a value

of 100 and the remaining peaks in all four samples were normalized to it. Twenty eight,

25, 17, and 16 volatiles were found in the DHE, DCE, FHE, and FCE respectively.

Thirteen compounds (hexanal, heptanal, limonene, octanal, 6-mehtyl-5-hepten-2-one,

nonanal, 1-octen-3-ol, acetic acid, decanal, bornylene, 2-nonenal (E), 1-octanol, and

geranylacetone) were present in all four samples and their concentration was lower in

the fresh and the cold water extracted samples. Nonanal (100) and decanal (99) were

the volatiles present in highest concentration in DHE and were also among the three

compounds present in highest concentration for the other three samples. Nonanal was

36 and decanal was 39 for DCE, 43 and 75 for FHE and 3 and 2 for FCE. Nonanal and

decanal are aldehydes that may form as a product of lipid oxidation.

Dehydroxylinalool oxide a and b were only present in extracts from dried

hibiscus. Since there are similar amounts of both compounds in the extracts obtained

with cold and hot water, these could be degradation products of linalool formed during









the drying process. Furfural and 5-methyl furfural were also only detected in the extracts

from dried hibiscus. These compounds are sugar degradation products and may also

have developed during hibiscus drying.

On the other hand, linalool and 1-hexanol-2-ethyl were only detected in the

extracts from fresh hibiscus. Their absence in the dried samples may be attributed to

the fact that these compounds may have degraded during drying and led to the

formation of other compounds.

GC-O Aroma Profiles

A total of 22 aroma compounds were found in hibiscus extracts and are listed in

Table 4-3. Peak heights were normalized to the most intense peak in all four samples.

This peak was 6-methyl-5-hepten-2-one in the DHE sample. It was assigned a value of

100 and the remaining peaks in all four samples were normalized to it. Seventeen, 16,

13, and 10 aroma active compounds were detected for DHE, DCE, FHE, and FCE

samples respectively. Seven compounds were detected in all four samples and were

confirmed with GC-MS (hexanal, 3-octanone, octanal, 6-methyl-5-hepten-2-one,

nonanal, 2,4-nonadienal (E,E), and geranylacetone).

The most intense odorants were 6-mehtyl-5-hepten-2-one and nonanal in all four

extracts followed by geranylacetone, eugenol, and 2-Nonenal (E) in the DHE sample,

geranylacetone, 2-Nonenal (E), and an unidentified compound for the DCE sample, and

linalool, geranylacetone, and octanal for the fresh hibiscus extracts,

The compound 6-methyl-5-hepten-2-one was present in all four samples and had

the highest intensity in all of them. This compound was described to have a mushroom,

dirt, green aroma and has previously been reported in tomatoes (Buttery and others

1987) and Rooibos tea (Kawakami and others 2003). Nonanal was the second highest









in aroma intensity in all four samples with a descriptor of fruity, green. Geranylacetone

was present in all samples, was among the five highest intensity compounds, and was

described as fruit-like, apple sauce smell. Geranylacetone has been found previously in

Merlot and Cabernet wines (Gurbuz and others 2006) and is one of the major

components of Rooibos tea.

In the fresh hibiscus extracts, the compounds linalool (floral, woody, citrus) and

octanal (lemon, citrus) were among the highest intensity aroma compounds. As

mentioned before, linalool was not detected in dried hibiscus extracts while octanal is

present in dried hibiscus samples and is the sixth highest intensity peak in both cold and

hot water extracts. Linalool is a compound associated with floral notes and has

previously been reported to be present in jasmine green tea (Ito and others 2002) and

citrus blossom (Jabalpurwala and others 2009) among others. Octanal has been

described to have a fruity, citrus aroma in lychee (Mahattanatawee and others 2007).

The compound 2-Nonenal (E) (cucumber, green, floral) was present in dried

hibiscus samples as one of the highest intensity peaks. Eugenol (sweet spices) was

only detected in the DHE sample while an unidentified (anise) compound was present in

the DCE sample with a high intensity.

The five highest intensity peaks for all four samples were: 2 ketones, 2 aldehydes

and 1 alcohol. The compounds 6-mehtyl-5-hepten-2-one, geranylacetone, and 2-

Nonenal (E) which are important aroma impact compounds present in hibiscus extracts

were identified for the first time in hibiscus.

Conclusion

The four hibiscus extracts studied had a similar chemical composition of aroma

compounds with hot extracted hibiscus samples having a slightly higher aldehyde









concentration and cold extracted samples a slightly higher alcohol concentration. Total

peaks concentration was the highest for the dried hibiscus hot water extract, and

decreased in both cold water extracts and fresh hibiscus extracts. There were some

differences in aroma peak intensities in the four hibiscus samples with the dried hibiscus

hot water extraction having the highest intensity. In general, hibiscus aroma is a

combination of earthy, green, floral, and fruity notes but the final flavor profile is affected

by the preservation and extraction process.









Table 4-1. Extraction conditions and measured pH and oBrix values for hibiscus
samples included in this study.

Extraction Extraction Hibiscus:
Sample temperature time water
(C) (min) ratio pH OBrix
DHE 98 16 1:40 w/v 2.48 0.01b 1.25 0.07a
DCE 22 240 1:40 w/v 2.49 0.00b 1.25 0.07a
FHE 98 16 1:4 w/v 2.55 0.01a 1.10 0.00a
FCE 22 240 1:4 w/v 2.57 0.01a 1.10 0.00a
DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FHE = fresh
hibiscus hot water extraction. FCE = fresh hibiscus cold water extraction.


100
90
80
70 -
60
50 -
40
30
20
10
0


'a


DHE


DCE FHE


FCE


Hibiscus Samples

SAldehydes (14) Alcohols (10) Ketones (5) Terpenes (2) Acids (1)


Figure 4-1. Chemical composition of hibiscus headspace volatiles. Total number of
compounds for each class is put in parentheses. All four samples were
normalized to the total peak area of DHE (dried hibiscus hot extraction). DCE
= dried hibiscus cold extraction, FHE = fresh hibiscus hot extraction, FCE =
fresh hibiscus cold extraction.


r-










Table 4-2. MS identification of hibiscus volatiles. Peak areas were normalized (100) to
the largest peak in all four samples.

LRI Normalized peak area (%)
Name CAS # CW DB5 DHE DCE FHE FCE


Hexanala
Heptanala
Limoneneabc
Dehydroxylinalool oxide a
Dehydroxylinalool oxide ba
3-Octanone
Octanala
2,2,6-Trimethylcyclohexanone
2-Heptenal, (E)a
6-methyl-5-Hepten-2-one
1-Hexanola
Nonanala
Octenal
1-Octen-3-ol
Acetic acida
Furfurala,b
1-Hexanol-2-ethyl
Decanalb
Bornylene
Benzaldehydeb,c
2-Nonenal (E)
Linaloola,b
1-Octanol
5-Methyl furfurala
2-Nonanone0
1-Nonanol
a-terpineola,b
(E,E)-2,4-Nonadienal
2-Undecenal
(E,E)-2,4-Decadienal
Geranylacetone
Eugenola,b,c
Total normalized peak area


66-25-1
111-71-7
138-86-3
13679-86-2
13679-86-2
106-68-3
124-13-0
2408-37-9
18829-55-5
110-93-0
111-27-3
124-19-6
2548-87-0
3391-86-4
64-19-7
98-01-1
104-76-7
112-31-2
464-17-5
100-52-7
18829-56-6
78-70-6
111-87-5
620-02-2
821-55-6
143-08-8
98-55-5
5910-87-2
53448-07-0
25152-84-5
3796-70-1
97-53-0


1100
1195


1210
1246
1264
1299
1329
1342
1355
1373
1405
1448
1468
1485
1496
1508
1513
1541


1555
1568
1577
1608


1678
1725
1728
1772
1836
1876


793
903
1026
993
1007


1002
1031
958
989


1100


983


832
1030
1204
1227
961
1159
1098
1071


1089
1172


1213
1362
1315
1440
1356


55.42
4.99
0.22
37.86
26.35
2.90
36.27
7.14
4.73
21.70


100.00
5.81
46.08
23.21
24.33


98.71
9.47
0.07
11.53


17.93
4.14
0.13
15.78


13.69
8.58
5.67
7.36
0.07
590.1


25.93
0.29
0.07
35.04
21.26
5.60
14.00
4.54


16.73
7.66
35.63
1.05
39.75
22.40
19.16


39.45
8.63


3.67


9.94
3.85
0.13
13.78


8.87


4.87
9.03


351.3


3.77
0.21
0.18


12.60



19.33
4.80
42.97


7.40
17.94


6.66
74.76
5.24


1.35
23.13
10.64




2.63


5.47


239.1


0.33
0.09
0.15


0.31



1.09
0.48
2.56


0.34
2.85


1.89
1.96
0.34


0.13
1.04
0.42


0.72


14.7


DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FHE = fresh
hibiscus hot water extraction. FCE = fresh hibiscus cold water extraction. a Compounds previously
reported in H. sabdariffa by Chen and others (1998). b Compounds previously reported in H. sabdariffa by
Gonzalez-Palomares and others (2009). C LRI values for this compounds were calculated using peak
areas from DB-5 column.











Table 4-3. Hibiscus aroma active compounds. Peak heights
the most intense peak in all four samples.


were normalized (100) to


# LRI Normalized peak height (%)
Name CW DB5 Aroma descriptor DHE DCE FHE FCE


Unknown
Hexanala
Unknown
3-Octanonea

Octanala
6-methyl-5-Hepten-2-onea
Nonanala
Octenala
1-Octen-3-ola

Furfurala
Decanala

2-Nonenal (E)a
Linaloola

1-Octanola
1-Nonanola
2,4-Nonadienal, (E,E)a
Unknown
2-Undecenala
Unknown

Geranylacetonea
Unknown

Eugenola
Total intensity


1054
1102
1200
1270

1304
1349
1410
1452
1477

1497
1520
1562
1570

1579
1674
1735
1754
1780
1850
1870
1940

2100


1003
980
1103


975




1154






1215
1244




1430


1350


sweet, fruity
green, grass, nutty
sweet, fruity
butter, cookie, baked

lemon, citrus
mushroom, dirt, green

fruity, green
rancid nuts

mushroom, dirt, metallic

sweet, baked bread
sweet, nutty

cucumber, green, floral
floral, woody, citrus

fresh leather, chemical
chemical, painty
rancid nuts, citrus, green
anise

green, grass
rancid nuts

fruit-like, apple sauce
sweet spices, floral

sweet spices


47.50
44.17
33.33

58.33
100.00
95.83
25.00
56.67

52.92
33.33
61.67


20.83
25.00
49.17


35.83


66.67


66.67
872.9


43.33
41.67
30.00

57.50
83.33
82.50
16.67
50.00


25.00
66.67


16.67


44.17
63.33
31.67
33.33
75.00



760.8


33.33
30.00


26.67

50.00
91.67
83.33







45.00
58.33




25.00
41.67


26.67
58.33
41.67


611.7


32.17
16.67


16.67

33.33
58.33
54.17









52.50




20.00
33.33




45.83



363.0


DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FHE = fresh
hibiscus hot water extraction. FCE = fresh hibiscus cold water extraction.
a Compounds confirmed with GC-MS.









CHAPTER 5
PROCESSING HIBSCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE:
MICROBIAL AND PHYTOHCEMICAL STABILITY

Introduction

Juices and beverages are traditionally preserved by thermal methods which are

effective in reducing microbial loads but can also lead to organoleptic and nutritional

changes. Nonthermal processes are an alternative that may help preserve the color,

flavor, and nutrients of food, and thus address consumers' demands for high quality,

fresh-like products with extended shelf life.

Dense phase carbon dioxide (DPCD) is a continuous nonthermal processing

system for liquid foods that uses pressure (590 MPa) in combination with carbon dioxide

(CO2) to inactivate microorganisms. In a continuous flow DPCD system, several

variables are controlled during processing: pressure, temperature, residence time, and

%CO2 in the liquid food. The amount of CO2 used should assure a complete saturation

of the liquid but since its solubility at processing conditions is not known this can lead to

the use of excess CO2 elevating production costs. Previous studies with muscadine

grape juice showed that DPCD was more effective than pasteurization in retaining

anthocyanins and other phenolic compounds during processing and storage (Del Pozo-

Insfran and others 2006a; 2006b). Furthermore, DPCD was effective in extending the

shelf life of coconut water (Damar and others 2009) and red grapefruit juice (Ferrentino

and others 2009) for up to 9 and 6 weeks of refrigerated storage, respectively.

Hibiscus sabdariffa, a member of the Malvaceae family, is an annual shrub

widely grown in tropical and subtropical regions including Africa, South East Asia and

some countries of America. The calyces contain anthocyanins and other phenolics and

are of commercial interest. They are used either fresh or dehydrated to prepare hot and









cold beverages which are commonly mixed with a sweetener, and are characterized by

an intense red color, acidic flavor, and a sensation of freshness. Recently there has

been increasing interest in hibiscus anthocyanins due to their beneficial health effects

and high antioxidant properties which have been extensively evaluated (Tee and others

2002; Tsai and others 2002; Tsai and Huang 2004; Prenesti and others 2007: Sayago-

Ayerdi and others 2007) and as a potential source of natural food colorant.

The objectives of this study were (1) to determine the solubility of CO2 in a

hibiscus beverage, (2) to optimize DPCD processing parameters (pressure and

residence time) based on microbial reduction, and (3) to monitor during 14 weeks of

refrigerated storage the microbial, physicochemical, and phytochemcial changes of

DPCD processed hibiscus beverage compared to thermally treated and control

(untreated) beverages.

Materials and Methods

Chemicals and Standards

Commercial standards of gallic acid, chlorogenic acid, and quercetin were

purchased from Sigma-Aldrich (St. Lous, Mo., U.S.A.). Caffeic acid was purchased from

ACROS Organics (Geel, Belgium). Delphinidin-3-glucoside and cyanidin-3-glucoside

were purchased from Polyphenols Laboratories AS (Sandnes, Norway). AAPH (2,2'-

azobis(2-methylpropionamidine) dihydrochloride), fluorescein (free acid), Trolox (6-

hydroxy-2,5,7,8-tetramethylchroman-2carboxilic acid) and Folin-Ciocalteu's reagent

were purchased from Sigma-Aldrich (St. Lous, Mo., U.S.A.).

Beverage Preparation

Dried Hibiscus sabdariffa (cv. "Criollo") (9% moisture content) obtained from

Puebla, Mexico was mixed with water (1:40 w/v) using a 200 L stainless steel mixing









tank Model UAMS (Cherry-Burrell, Iowa, U.S.A.) and maintained at 250C for 1 h. Mixing

was applied intermittently by alternating intervals of 10 min mixing and 10 min rest. The

extract was then filtered using four layers of cheesecloth. A beverage was prepared by

adding sucrose to a concentration of 100 g sucrose/L of extract and then was placed in

3 gallon sealable buckets. For the DPCD process optimization, the beverage was

incubated at 25 oC for 4 days to obtain a high initial microbial load. The spoiled

beverage was placed in the refrigerator at 4 oC for 24 h before processing with DPCD.

For the solubility and storage experiments, the beverage was prepared as mentioned

above, without incubation, one day before processing and refrigerated.

Solubility Experiment

C02 solubility in the hibiscus beverage was measured between 6.9 and 31.0 MPa

at 40 oC using an apparatus designed and built at the University of Florida, Food

Science and Human Nutrition department (Gainesville, Fla., U.S.A.) as previously

described by Ferrentino and others (2009). In this batch system, a known volume of

sample was saturated by bubbling CO2 at the desired experimental conditions and then

dissolved CO2 was measured at atmospheric pressure. Solubility of CO2 in water at the

same experimental conditions was also measured for comparison. Experiments were

done in duplicate and results were expressed as g of C02/100 g of liquid sample.

Dense Phase CO2 Equipment

The DPCD equipment located at the University of Florida (Gainesville, Fla., U.S.A)

was constructed by APV (Chicago, Ill., U.S.A.) for Praxair (Chicago, Ill., U.S.A.). A

schematic diagram of the equipment is presented in Figure 2-11. In this continuous flow

equipment, CO2 and the hibiscus beverage were pumped through the system and

mixed before entering the high-pressure pump (intensifier pump). Processing pressure









levels were controlled by this pump while the desired temperature was maintained in the

holding coil (79.2 m, 0.635 cm i.d.). Turbulent flow and mixing were reached at the

entrance of the coil by passing the mixture through a static mixer and a small diameter

tube (length of about 180 cm). Residence time was adjusted by setting the flow rate of

the mixture. An expansion valve was used at the end of the process to release the CO2

from the mixture and the beverage was collected into 1 L sterile bottles as previously

described by Damar and others (2009).

DPCD Process Optimization

Optimal processing parameters to achieve a microbial log reduction of 5 were

determined by using response surface methodology. A central composite design (Table

5-1) consisting of 11 experiments with 4 factorial points, 4 star points, and 3 central

points in which the independent variables were pressure (P) (13.8-34.5 MPa) and

residence time (RT) (5-8 min), and the dependent variables were yeasts and molds

(Y&M) and aerobic plate counts (APC) was used. With this response surface design we

were able to reduce the volume of beverage required and were able to prepare it in one

batch. The total volume needed for the 11 experiments was ~ 160L. Processing

parameters were selected based on previous research results and equipment

specifications.

Hibiscus beverage with an initial microbial load of 3.0 x 107 CFU/mL for Y&M and

4.9 x 103 CFU/mlL for APC was processed at the different experimental conditions at a

constant temperature (40 C) and constant CO2 level (8%) which was selected based

on the minimum flow that could be handled by the CO2 pump. Microbial counts from

each experimental condition were made in duplicate by serially diluting (1x10-1 to 1x10-

5) 1 mL of beverage in 9 mL sterile Butterfield's phosphate buffer (Weber Scientific,









Hamilton, N.J., U.S.A.). Microbial counts were determined by plating 1 mL of each

dilution in duplicate for yeasts and molds and aerobic count plates (3M Petrifilm

Microbiology, St. Paul, Minn., U.S.A.) and enumerating after 48 h at 35 oC and 5 d at 24

oC respectively according to the manufacturer's guidelines.

Thermal Processing Conditions

For thermal processing, hibiscus beverage was pumped by a peristaltic pump

(Cole Parmer, Chicago Ill., U.S.A.) through two stainless steel tube sections (3.2 m,

0.457 cm i.d. ea.) placed inside a temperature controlled water bath (Precision

Scientific, Chicago Ill., U.S.A.). In the first section, the beverage was heated to 75 C

(temperature was measured using a thermocouple) and then entered the second

section where it was held at 75 oC for 15 s. The beverage was then passed through a

cooling stainless steel tube (5.2 m, 0.457 cm i.d.) in a water/ice bath and chilled to ~15

oC before it was collected into 1 L sterile glass jars. Platinum-cured silicone tubing

(0.635 cm i.d.; Nalgene, Rochester, N.Y., U.S.A.) was used to connect the pump to the

stainless steel heating, holding, and cooling sections. A schematic diagram of the setup

used for the hibiscus beverage thermal processing is presented in Figure 5-2.

Storage Experiment

Fresh prepared hibiscus beverage was divided into three parts. One part was

kept as control and did not receive any treatment; the second part was processed using

DPCD at 34.5 MPa, 8% CO2, 6.5 min, and 40 oC while the third part was thermally

treated at 75 oC for 15 s. Each processing condition was repeated in triplicate. DPCD

processing parameters were determined based on the solubility and optimization

studies described above. Both the control and treated samples were stored in 1 L glass

jars. Microbial, physicochemical and phytochemical analyses were performed at weeks









0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 of refrigerated storage at 4 oC. To quantify individual

anthocyanins and other phenolic compounds present in hibiscus, HPLC analysis was

done at weeks 0, 2, 5, 8, and 14 of storage. The methods used in all the analyses are

described as follows.

Microbial Analysis

Total aerobic plate count (APC) and yeasts and molds (Y&M) were measured as

described in the optimization section.

pH, Brix, and Titratable Acidity

pH and Brix were measured using a pH meter EA920 (Orion Research; Boston,

Mass., U.S.A.) and a ABBE Mark II refractometer (Leica Inc.; Buffalo, N.Y., U.S.A.). A

Brinkmann Instrument (Brinkmann Instruments Co., Westbury, N.Y., U.S.A.) consisting

of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was

used to measure titratable acidity (TA). Samples of 10 mL were used and TA was

determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid

(g/100 mL).

Color Density and Hue Tint

Color density and hue tint were determined by measuring the absorbance (A) at

420, 520, and 700 nm of 200 pL samples using a spectrophotometer (SpectraMax 190,

Molecular Devices, Sunnyvale Calif., U.S.A.) and calculated as:

Color density = [(A420 nm A700 nm) + (A520 nm A700 nm)]

Hue tint = (A420 nm A700 nm)/(A520 nm A700 nm)

as described by Giusti and Wrolstad (2005).


100









Anthcyanin Content, Total Phenolics and Antioxidant capacity

Anthocyanin content was determined by pH differential method (A51o nm and A700

nm at pH 1.0 and 4.5, dilution factor (DF) of 4) and expressed in mg/L of delphinidin-3-

glucoside (MW = 465.2, s = 23700) (Giusti and Wrolstad 2005). Total phenolics were

measured using the Folin-Ciocalteu assay (A765 nm, DF of 4) and quantified as gallic acid

equivalents (mg/L) (Waterhouse 2005). Absorbance measurements for anthocyanin

content and total phenolics were made using a SpectraMax 190 spectrophotometer

(Molecular Devices, Sunnyvale Calif., U.S.A.).

Antioxidant capacity was evaluated using the oxygen radical absorbance

capacity (ORAC) assay and results were expressed as Trolox equivalents (TE) per

milliliter (pmol of TE/mL) as described by Huang and others (2002) using a SpectraMax

Gemini XPS microplate sprectrofluorometer (Molecular Devices, Sunnyvale, Ca.,

U.S.A.). Data was acquired and analyzed using SoftMax Pro 5.2 software (Molecular

Devices, Sunnyvale, Calif., U.S.A.).

HPLC Quantification of Polyphenolics

Polyphenolics were identified by comparison of UV/vis (190-660 nm) spectral

interpretation, retention time, and comparison to standards. Anthocyanins and

polyphenolics were quantified using a Dionex HPLC system equipped with an

autosampler/injector and diode array (PDA 100) detector (Dionex, Sunnyvale, Calif.,

U.S.A.). Compounds were separated on a 250 x 4.6 mm Dionex C18 5 pm 120A

column (Dionex, Sunnyvale, Calif., U.S.A.). Mobile phases consisted of water (phase A)

and 60% methanol in water (phase B), both adjusted to pH 2.4 with o-phosphoric acid.

A gradient solvent program ran phase B from 0% to 60% in 20 min; 60% to 100% in 20


101









min; 100% for 7 min; 100% to 0% in 3 min and final conditions were held for 2 min. The

flow rate was 0.8 mL/min, and detection was done at 260, 280, 320, 360 and 520 nm.

Statistical Analysis

Analysis of variance (ANOVA) and mean separation using Tukey's test (a=0.05)

were performed in the solubility study to evaluate the effect of pressure on CO2

solubility. Response surface methodology was used in the DPCD optimization study to

determine optimal processing conditions. Repeated measures ANOVA and mean

separation using Tukey's test (a=0.05) was performed to evaluate the effect of

treatment (fresh (CONTROL), thermal (HTST), and DPCD processed) and storage time

(0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 weeks) on the dependent variables measured.

All statistical analyses were conducted using SAS statistical software (SAS Institute

Inc., Cary, N.C., U.S.A.).

Results and Discussion

Solubility Measurements

C02 solubility in a hibiscus beverage and water was measured between 6.9 and

31.0 MPa at 40 C. Pressure had a significant effect on solubility of C02 in both the

hibiscus beverage (4.16 to 5.06 g C02/100 mL from 6.9 to 31.0 MPa) and water (4.50 to

6.32 g C02/100 mL from 6.9 to 31.0 MPa). After 17.2 MPa, CO2 solubility remained

almost constant in both the hibiscus beverage and water (Figure 5-1). CO2 solubility in

water was significantly higher than the hibiscus beverages at all pressures tested but

6.9 MPa. The presence of solutes such as sugars and acids in the hibiscus beverage

lowered the amount of CO2 that could be dissolved. Previous studies have shown that

solubility of CO2 in fruit juices is lower than that of pure water because of the presence

of solutes. CO2 solubility in orange and apple juice measured at 15.9 MPa was around


102









5% (Calix and others 2008) while that of grapefruit juice at 31.0 MPa was 4.97%

(Ferrentino and others 2009). These values are comparable to those obtained for

hibiscus beverage (5.06%).

Based on the CO2 solubility results, a first attempt to use 6% CO2 for the DPCD

experiments was tried. This concentration of CO2 (1% higher) would assure a complete

saturation of CO2 in the beverage. This decision was made to account for the fact that

the solubility test was performed using a batch system with long contact time between

the CO2 and the beverage whereas the DPCD processing equipment is a continuous

system in which lower contact times are used. After processing the hibiscus beverage

using 6% CO2 and acquiring data, the DPCD system showed that there were

fluctuations in the CO2 flow during the process because the CO2 pump was not

designed to handle such a low flow. Our second attempt was to find the minimum flow

that would assure a constant reading throughout the process. After several attempts, it

was found that the CO2 pump could maintain a steady flow of 8% CO2 and we used this

value for our DPCD processing experiments.

Microbial Inactivation Study

Initial microbial loads in the beverage obtained after incubation for 4 days were 3 x

107 CFU/mL for Y&M and 4.9 x 103 CFU/mL for APC. The APC population reached is

not very high which can be a result of microflora competition in which the low pH (2.43)

of the beverage and high sugar concentration (9.7 oBrix) favored the growth of Y&M.

The response surface experimental design and achieved log reductions for each of the

treatments tested is presented in Table 5-1. A minimum of 5 log reduction for Y&M and

0.85 log reduction for APC was achieved for all DPCD treatments.


103









Several mechanisms have been proposed for DPCD inactivation of

microorganisms (Damar and Balaban 2006). One of the main factors that lead to

microbial inactivation is the pH lowering effect when CO2 is dissolved. Since the

hibiscus beverage has an initial low pH, this reduces the lowering pH effect of CO2 and

this may be the reason for low bacteria inactivation in the hibiscus beverage. Another

mechanism for microbial inactivation is the effect caused in the microorganisms' cells

during the decompression process. This can be the mechanisms by which Y&M were

inactivated.

Two quadratic equations, (1) and (2), were obtained from the central composite

design solution to describe Y&M (r2 = 0.81) and APC (r2 = 0.55) log reduction (LR) as a

function of pressure (P) and residence time (RT). Both quadratic models were not

statistically significant and were not suitable to predict the inactivation of

microorganisms present in the hibiscus beverage within pressures and residence times

ranges studied.

Log reduction (Y&M) = 3.8745 + 0.0155-P + 0.2200-t + 0.0007-P2- 0.0037-P-t + 0.0042-t2 (1)

Log reduction (APC) = 0.6804 + 0.0187-P + 0.4493-t 0.0001-P2- 0.0019-P-t- 0.0311-t2 (2)

As can be seen from Table 5-1, treatment 8 (24.1 MPa, 8 min) showed the

highest LR for Y&M followed by treatments 10 (34.5 MPa, 6.5 min) and 11 (34.5 MPa, 8

min). On the other hand, treatment 5 (24.1 MPa, 6.5 min) had the highest log reduction

for APC and treatment 10 was among the second highest APC log reduction treatments

while treatment 11 was among the lowest APC log reduction treatments. Based on

these results, our approach was to select treatment 10 for further DPCD processing

experiments. This treatment conditions consists of the upper level pressure within our

experimental range studied (34.5 MPa) which will assure a complete solubility of CO2


104









during processing and the middle level residence time of 6.5 min which is more feasible

for industrial applications than longer times.

Microbial Stability during Storage

Microbial stability of unprocessed (CONTROL), dense phase-CO2 processed

(DPCD), and thermally treated (HTST) hibiscus beverages during storage is presented

in Figures 5-2 and 5-3. Aerobic plate counts in all three beverages (Figure 5-2)

remained constant between 2 and 3 logs during the 14 weeks of storage. The HTST

beverage showed slightly lower counts when compared to the other two beverages.

Neither the DPCD nor the HTST treatments reduced the initial bacteria population

possibly because it was difficult to observe microbial reductions when starting with a low

population. In the case of yeast and molds (Figure 5-3), the DPCD and HTST

treatments reduced the initial population by around 3 logs and both beverages where

very stable since both treatments were effective in inactivating the initial Y&M

population and there was no growth during storage. For the CONTROL, a maximum of

5 logs at week 6 was reached and declined afterwards possibly associated with the

death stage of the microorganisms. The sensory characteristics of the CONTROL

beverage indicated that fermentation was taking place. Overall the DPCD and HTST

beverages were microbiologically stable during the 14 weeks of storage favored by the

beverage low pH and storage temperature (4 oC).

Physicochemical Stability during Storage

Physicochemical changes in the studied hibiscus beverages during storage are

shown in Table 5-2. There were no significant differences between treatments

(CONTROL, DPCD, HTST) over time for pH and Brix. There was a significant effect of

treatment over time for all other parameters measured. Titratable acidity in the DPCD


105









treated beverage was significantly higher when compared with the CONTROL and

HTST beverages which can be due to the presence of residual CO2 remaining in

solution in the beverage after depressurization. A similar behavior was observed by

Calix and others (2008) in orange and apple juices.

Color density significantly decreased over time for all three treatments

(CONTROL, DPCD, and HTST). This indicates that there is a decline in the absorbance

at 520 nm which can be associated with degradation of anthocyanins. At time 14 weeks

of storage, the HTST beverage showed a significantly lower value of color density as

compared to the CONTROL and DCPD beverages. Moreover, the hue tint values

(Figure 5-4) significantly increased for all three treatments during storage which also

indicates some loss of red color in the samples.

Phytochemical Stability during Storage

Phytochemical changes during storage for the three hibiscus beverages studied

are presented in Tables 5-2 and Figure 5-5. Several polyphenolic changes during

storage were measured using authentic standards and their concentration was

expressed in mg/L of beverage (Table 5-3). This included gallic and caffeic acid,

caffeoylquinic acids which were quantified using a chlorogenic acid standard and were

identified based on their characteristic absorption spectrum at Amax 320 nm, delphinidin-

3-sambubioside and cyanidin-3-sambubioside that are the main anthocyanins present in

hibiscus extracts, and flavonols which were quantified using quercetin and identified by

their characteristic absorption spectrum at Amax 360 nm.

There was a significant effect of treatment (CONTROL, DCPD, HTST) in

anthocyanins content, total phenolics, antioxidant capacity, gallic acid, caffeic acid, and


106









flavonols content. Anthocaynins content (Figure 5-6) significantly decreased during

storage for all three treatments. A loss of 11, 9, and 14% in anthocyanins was observed

for CONTROL, DPCD, and HTST beverages respectively. At time 14 weeks, the

concentration of anthocyanins in all three treatments was significantly different, with the

CONTROL having the highest and HTST beverage the lowest concentration. There

were no major changes in total phenolics and antioxidant capacity during storage for all

three treatments. There were some slight differences between storage times possibly

related to the breakdown and formation of polyphenolic compounds. A previous study

with muscadine grape juice (Del Pozo and others 2006a) showed that losses in

anthocyanins during processing and storage were around 78% for a pasteurized juice

and only 35% for a DPCD processed juice. A similar behavior in total phenolic and

antioxidant capacity was also found. The greater losses and differences between

treatments in the grape juice as compared to the hibiscus beverage can be attributed to

a higher initial concentration of polyphenolics and higher pH of the grape juice.

As shown in Table 5-3, the concentration of gallic acid increased with increasing

storage time for the DPCD beverages and to a greater extent for the CONTROL.

Similarly, the presence of caffeic acid in the CONTROL and DPCD beverages at time

14 weeks was detected and can be a breakdown product of the caffeoylquinic acids

present in the beverage. Both phenomena could be related to polyphenolic compounds

breaking down due to microbial activity. There were no major changes in the

caffeoylquinic acids and flavonols content during storage, although at time 14 weeks

there was a significantly lower concentration of both polyphenolics in the CONTROL

beverage as compared to the other two treatments. There was a significant but small


107









decrease in the concentration of delphinidin-3-sambubioside and cyanidin-3-

sambubioside for all the three beverages during storage. Overall there were no big

phytochemical losses during storage for any of the three treatments. This can be

attributed to the low pH of the beverage, the low storage temperature and the presence

of sucrose in the beverage. A previous study (Tsai and others 2004) has shown that

sucrose solutions favored the stability of hibiscus anthocyanins by decreasing the

availability of water that is needed for the anthocyanins degradation process.

Conclusions

002 solubility in a hibiscus beverage and optimal processing conditions to

inactivate microorganisms (Y&M and APC) were determined. DPCD was found to be a

viable technology for extending the hibiscus beverage shelf life since it showed to be

microbiologically stable during the 14 weeks of refrigerated storage. Quality attributes

such as pH and Brix were not affected by DPCD whereas TA increased. A loss of only

9% anthocyanins during storage was observed for the DPCD processed hibiscus

beverage which was lower when compared to a heat pasteurization process and no

major changes in total phenolics content and antioxidant capacity occurred during

storage


108










Hibiscus
beverage


Pasteurized
beverage


Water bath Ice slush.................................
Water bath ice slush


Figure 5-1. Schematic diagram of the setup used for the hibiscus beverage thermal
treatment (75 C for 15 s).


-Hibiscus
-a-Water


5 10 15 20 25 30 35
Pressure (MPa)

Figure 5-2. C02 solubility in water and a hibiscus beverage as a function of pressure
measured at 40 C. Data represents the mean of n=3. Values with similar
letters within the figure are not significantly different (Tukey's HSD, p > 0.05).


109


d c ca
ea




e

e









Table 5-1. Response surface design used to test the effect of pressure and residence
time on microbial reduction logol) at 40 C and 8% C02.

Run Pressure Residence Beverage flow CO2 flow LR LR
(Mpa) time (min) rate (g/min) rate (g/min) (APC) (Y&M)
1 13.8 5.0 500.0 40.0 0.93abc + 0.03 5.20e 0.08
2 13.8 6.5 384.6 30.8 0.99abc + 0.03 5.63d + 0.06
3 13.8 8.0 312.5 25.0 0.92bc 0.03 5.67d 0.04
4 24.1 5.0 500.0 40.0 0.88c 0.05 5.30e 0.05
5 24.1 6.5 384.6 30.8 1.05a 0.07 5.56d 0.08
6 24.1 6.5 384.6 30.8 0.99abc + 0.09 5.62d + 0.09
7 24.1 6.5 384.6 30.8 1.04ab + 0.06 5.65d 0.06
8 24.1 8.0 312.5 25.0 1.02ab + 0.04 6.26a 0.11
9 34.5 5.0 500.0 40.0 1.04ab + 0.06 5.85c 0.06
10 34.5 6.5 384.6 30.8 1.03ab + 0.02 6.07b + 0.01
11 34.5 8.0 312.5 25.0 0.91 bc 0.08 6.09b + 0.06
Y&M = yeasts and molds, APC = aerobic plate count. LR = loglo reduction.


4--CONTROL
--DPCD
-* HTST


0 2 4 6 8 10 12 14


Storage time (weeks)

Figure 5-3. Aerobic plate counts of unprocessed (CONTROL), dense phase-CO2
processed (DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated
(HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C).


110

















-4-CONTROL
--DPCD
-*-HTST


0 2 4 6 8 10 12 14
Storage time (weeks)

Figure 5-4. Yeast/mold counts of unprocessed (CONTROL), dense phase-CO2
processed (DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated
(HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C).


0.4

0.39

0.38


0.37

0.36


---CONTROL
----DPCD
HTST


0.35

0.34

0.33


0 2 4 6 8 10 12 14


Storage time (weeks)

Figure 5-5. Hue tint values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).


111


rg~J*a~.


T

T--









Table 5-2. Physicochemical and phytochemical changes of unprocessed (CONTROL), dense phase-C02 processed
(DPCD), and thermally treated (HTST) hibiscus beverages during refrigerated storage at 4 oC.

Storage time
week 0 week 1 week 2 week 3 week 4 week 5 week 6 week week 10 week 12 week 14
pH
CONTROL 2.43b* 2.45a 2.47a 2.45b 2.45c 2.47b 2.47b 2.478 2 2. 2.48 2.488
DPCD 2.45a 2.44b 2.46b 2.46a 2.47a 2.48a 2.48a 2.48a 2.48a 2.49a 2.49a
HTST 2.45a 2.45a 2.45b 2.45b 2.47b 2.48a 2.48a 2.48a 2.48a 2.49a 2.49a
0Brix
CONTROL 9.70ab 9.70a 9.70b 9.87b 9.70b 9.67a 9.90a 10.20a 10.07a 9.93a 9.70b
DPCD 9.77a 9.70a 9.90a 9.93a 9.87a 9.70a 9.63b 10.10b 9.83b 9.57b 9.77ab
HTST 9.63b 9.67a 9.90a 9.73C 9.908 9.60b 9.97 10.00 9.82b 9.63b 9.90a
Titratable acidity (g of malic acid/100 mL)
CONTROL 0.37b 0.37b 0.38b 0.38b 0.37b 0.38b 0.38b 0.38b 0.37ab 0.38b 0.38b
DPCD 0.40a 0.41a 0.40a 0.41a 0.41a 0.40a 0.41a 0.41a 0.40a 0.41a 0.40a
HTST 0.37b 0.36b 0.38b 0.37b 0.38b 0.38b 0.38b 0.38b 0.38b 0.37b 0.38b
Color density
CONTROL 1.78ab 1.78b 1.75ab 1.74a 1.71b 1.72a 1.73a 1.74a 1.70a 1.66a 1.67a
DPCD 1.80a 1.80a 1.76a 1.75a 1.76a 1.73a 1.74a 1.74a 1.69a 1.64a 1.66a
HTST 1.76b 1.79ab 1.73b 1.75a 1.72ab 1.69a 1.72a 1.68b 1.63b 1.58b 1.58b
Total phenolics (mg/L)
CONTROL 263.86a 259.01a 247.97b 248.38a 251.13a 246.43ab 249.13a 250.54a 251.99a 253.44a 264.05a
DPCD 259.62ab 256.86a 252.26a 253.22a 248.52a 245.03b 253.91a 252.32ab 252.25a 252.17a 255.65b
HTST 254.73b 259.39a 245.04b 251.09a 241.64b 248.36a 241.61b 248.36b 247.15b 245.94b 250.06c
Antioxidant capacity (|pmol of TE/mL)
CONTROL 5.93a 6.11b 6.62a 5.77a 6.04a 5.96a 6.64a 6.53a 6.27a 6.01a 5.85b
DPCD 5.66a 6.10b 5.43b 6.19a 5.90a 6.08a 6.31a 5.59b 5.76a 5.92a 6.80a
HTST 6.38a 7.00a 5.84b 6.34a 6.42a 6.53a 6.11 5.27b 5.81a 6.35a 6.92a
Data represents the mean of n=9. Values with similar letters within columns are not significantly different (Tukey's HSD, p > 0.05).


112

















---CONTROL
---DPCD
~ HTST


0 2 4 6 8 10 12 14


Storage time (weeks)


Figure 5-6. Concentration of anthocyanins of unprocessed (CONTROL), dense phase-
CO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 oC) and thermally
treated (HTST; 75 oC, 15 s) hibiscus beverage during refrigerated storage (4
oC).


113









Table 5-3. Polyphenolics content (mg/L) of unprocessed (CONTROL), dense phase-
C02 processed (DPCD), and thermally treated (HTST) hibiscus beverages
during refrigerated storage at 4 oC.

Storage time
weekO week 2 week 5 week 8 week 14
Galllic acidd
CONTROL 0.79a* 2.53a 3.25a 3.89a 4.03a
DPCD 0.50b 0.77b 1.03b 1.27b 1.57b
HTST 0.57b 0.68b 0.63c 0.71c 0.71c
Caffeoylquinic acids
CONTROL 93.59a 98.13a 94.05a 94.11a 89.88b
DPCD 91.90a 96.78a 95.74a 95.28a 94.50a
HTST 92.51a 98.36a 93.31a 95.10a 94.73a
Caffeic acid'
CONTROL ndj nd nd nd 5.73a
DPCD nd nd nd nd 4.26b
HTST nd nd nd nd nd
Delphinidin-3-sambubiosideg
CONTROL 65.26a 64.64a 61.96ab 62.14a 58.07ab
DPCD 63.12b 61.75b 62.53a 61.15ab 59.23a
HTST 64.79ab 62.20ab 59.12b 58.72b 56.78b
Cyanidin-3-sambubiosideh
CONTROL 29.98a 29.60a 28.23a 28.10a 26.61a
DPCD 28.78a 28.46a 28.49a 27.95a 27.43a
HTST 29.99a 29.03a 27.76a 27.74a 27.40a
Flavonols'
CONTROL 23.05b 23.55a 22.44b 22.79b 22.65b
DPCD 25.22a 24.96a 25.02a 25.18a 24.23ab
HTST 25.08a 25.15a 24.64a 25.45a 24.78a
Data represents the mean of n=6. Values with similar letters within columns of each polyphneolic
category are not significantly different (Tukey's HSD, p > 0.05). de,f,g,hI Quantified with gallic acid,
chlorogenic acid, delphinidin-3-glucoside, cyanidin-3-glucoside, and quercetin standards respectively.J
Abbreviations: nd, not detected.


114









CHAPTER 6
PROCESSING HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE:
SENSORY ATTRIBUTES AND AROMA COMPOUNDS STABILITY

Introduction

Hibiscus sabdariffa (family Malvaceae) is a short-day annual shrub that grows in

many tropical and subtropical countries and is known by different synonyms and

vernacular names such as "roselle" in the U.S. and England, "l'oiselle" in France,

"jamaica" or "flor de jamaica" in Mexico and Spain, "karkade" in Sudan and Arabia,

"sorrel" in the Caribbean and "byssap" in Senegal (Morton 1987; Stephens 2003).

Traditionally fresh hibiscus calyces are harvested by hand and are either frozen

or dried, in the sun or artificially, for preservation. They are typically sold into the herbal

tea and beverage industry or in local and regional markets where they are used in the

preparation of beverages, and color and flavor extracts (Plotto 1990). Studies with

human patients have shown that the regular consumption of hibiscus extract has an

antihypertensive effect (Haji Faraji and Haji Tarkhani 1999; Herrera-Arellano and others

2004) and reduces serum cholesterol in men and women (Lin and others 2007).

The preparation of a hibiscus beverage includes an extraction step followed by a

pasteurization method. Although thermal preservation of foods is effective in reducing

microbial loads it can also lead to organoleptic and nutritional changes. Nonthermal

processes are an alternative which may help preserve the color, flavor, and nutrients of

food. Dense phase carbon dioxide (DPCD) is a cold pasteurization method that uses

pressures below 90 MPa in combination with carbon dioxide (CO2) to inactivate

microorganisms. This non-thermal technology is mainly used in liquid foods and since

the food is not exposed to the adverse effect of heat, its fresh-like physical, nutritional,

and sensory qualities are maintained.


115









Previous studies have shown that DPCD processed beverages keep their fresh-

like characteristics after processing and storage. Likeability of DPCD-treated coconut

water was similar to untreated samples while heat treated samples were less appealing

(Damar and others 2009). Similarly, no differences in sensory attributes (color, flavor,

aroma, and overall likeability) were observed between unprocessed and DPCD

muscadine grape juices but there were differences when compared to a heat-

pasteurized juice (Del-Pozo-lnsfran and others 2006a).

The objectives of this study were (1) to determine the effect of DPCD processing

on the sensory attributes and aroma compounds of hibiscus beverage when compared

to a thermally treated and a control (untreated), and (2) to monitor the changes in these

attributes during refrigerated storage.

Materials and Methods

Beverage Preparation

Dried Hibiscus sabdariffa (cv. "Criollo") (moisture content of 9%) obtained from

Puebla, Mexico was mixed with water (1:40 w/v) using a 200 L stainless steel mixing

tank Model UAMS (Cherry-Burrell, Iowa, U.S.A.) and maintained at 25C for 1 h. Mixing

was applied intermittently by alternating intervals of 10 min mixing and 10 min rest. The

extract was then filtered using four layers of cheesecloth. A beverage was prepared by

adding sucrose to a concentration of 100 g sucrose/L of extract and then was placed in

3 gallon sealable buckets and refrigerated before processing.

Processing and Storage Conditions

Fresh prepared hibiscus beverage was divided into three parts. One part was kept

as CONTROL and didn't receive any treatment; the second part was processed using

DPCD at 34.5 MPa, 8% CO2, 6.5 min, and 40 C while the third part was pasteurized at


116









75 C for 15 s (HTST). The DPCD processing conditions were confirmed to achieve >5

log reduction of yeasts/molds according to previous experiments. Both the control and

treated samples were stored in 1 L glass jars. Physicochemical, sensory and aroma

compound analysis were done at weeks 0 and 5 of refrigerated storage at 4 C .Color

analysis was performed at weeks 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 of storage.

Dense Phase CO2 Equipment

The DPCD equipment located at the University of Florida (Gainesville, Fla.,

U.S.A.) was constructed by APV (Chicago, Ill., U.S.A.) for Praxair (Chicago, Ill., U.S.A.).

It is a continuous flow equipment in which CO2 and the hibiscus beverage were pumped

through the system and mixed before entering a high-pressure pump. Processing

pressure was controlled by this pump while the desired temperature was maintained in

the holding coil (79.2 m, 0.635 cm i.d.). Turbulent flow and mixing were reached at the

entrance of the coil by passing the mixture through a static mixer and a small diameter

tube (length of about 180 cm). Residence time was adjusted by setting the flow rate of

the mixture. An expansion valve was used at the end of the process to release the CO2

from the mixture and the beverage was collected into 1 L sterile bottles as previously

described by Damar and others (2009).

Thermal Processing Conditions

For thermal processing, the hibiscus beverage was pumped by a peristaltic pump

(Cole Parmer, Chicago Ill., U.S.A.) through two stainless steel tube sections (3.2 m,

0.457 cm i.d. ea.) placed inside a temperature controlled water bath (Precision

Scientific, Chicago Ill., U.S.A.). In the first section the beverage was heated to 75 C

(temperature was measured using a thermocouple) and then entered the second

section where it was held at 75 oC for 15 s. The beverage was then passed through a


117









cooling stainless steel tube (5.2 m, 0.457 cm i.d.) in a water/ice bath and chilled to ~15

oC before it was collected into 1 L sterile glass jars. Platinum-cured silicone tubing

(0.635 cm i.d.; Nalgene, Rochester, N.Y., U.S.A.) was used to connect the pump to the

stainless steel heating, holding, and cooling sections. A schematic diagram of the setup

used for the hibiscus beverage pasteurization is presented in Figure 5-2.

Physicochemical Analysis

pH and oBrix were measured using a pH meter EA920 (Orion Research; Boston,

Mass., U.S.A.) and a ABBE Mark II refractometer (Leica Inc.; Buffalo, N.Y., U.S.A.). A

Brinkmann Instrument (Brinkmann Instruments Co., Westbury, N.Y., U.S.A.) consisting

of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was

used to measure titratable acidity (TA Samples of 10 mL were used and TA was

determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid

(g/100 mL).

Sensory Evaluation

Flavor and overall likeability of fresh and processed hibiscus beverages were

compared using a difference from control test at weeks 0 and 5 of storage. A

randomized complete block design was used, and differences from control values were

recorded on a line scale with anchors at 0 and 10 that represented "no difference" to

"extremely different" in flavor. Panelists compared the flavor of the reference

(fresh/unprocessed beverage (CONTROL)) with that of a hidden reference (fresh

beverage (CONTROL)), the thermally (HTST), and the DPCD processed beverages. A

9-point hedonic scale was also conducted in order to compare the overall likeability of

fresh (hidden reference) and processed hibiscus beverages. For the taste panel at week

5, the reference was fresh hibiscus beverage that was kept frozen at -20 oC.


118









Before sensory analysis all beverages (fresh, thermally, and DPCD processed)

were degassed in order to have equal carbonation levels by placing them in 2 L sterile

glass bottles on a stir plate with continuous stirring and vacuum (15" Hg) was pulled for

20 min using a Gast vacuum pump (Model DOA-P104-AA; Beonton Harbor, Mich.,

U.S.A.). All samples were chilled and kept in ice at a temperature of ~4C before

serving. They were then served on a tray in numbered plastic cups containing ~30 mL

of sample. A cup of deionized water and non salted crackers were also provided to the

panelists to cleanse their palate between evaluations. Both sensory tests were

performed at the University of Florida taste panel facility using 75 untrained panelists in

each test.

Headspace Volatiles Sampling

Headspace volatiles were extracted and concentrated using Solid Phase Micro

Extraction (SPME) technique. Ten milliliters of hibiscus beverage were added to a 22

mL screw cap amber glass vial PTFE/silicone septa containing a small stir bar. Samples

were equilibrated for 20 min in a water bath at 40 C and afterwards hibiscus

headspace volatiles were extracted for 30 min using a 1 cm 50/30 mm

DVB/Carboxen/PDMS SPME fiber (Supelco, Bellefonte, Pa., U.S.A.). Before each

exposure the fiber was cleaned for 5 min at 200 C in the GC-MS injection port.

GC-MS Analysis

GC-MS analysis was conducted using a HP 6890 GC coupled with a MSD 5973

(Agilent Technologies; Palo Alto, Calif., U.S.A.). Hibiscus volatiles from the SPME fiber

were desorbed into the GC injection port (splitless mode) at 200 C. A SPME injector

liner (SPME injection sleeve, 0.75 mm i.d., Supleco; Bellefonte, Pa., U.S.A.) was used.


119









The fiber was removed after 5 min exposure in the injection port. Volatiles were

separated on both a DB-5 (30 m x 0.32 mm. i.d. x 0.5 pm, J&W Scientific; Folsom,

Calif., U.S.A.) and a DB-Wax (30 m x 0.32 mm. i.d. x 0.5 |tm, Restek; Bellefonte, Pa.,

U.S.A.) column. The oven was programmed from 35 C (held for 5 min) to 250 oC at 6

oC/min with a final hold of 10 min. Helium was used as the carrier gas at 1.67 mL/min.

The mass spectrometer was operated in the total ion chromatogram (TIC) at 70 eV.

Data were collected from 35 m/z to 400 m/z. All samples were run in duplicate in each

column. Chromatograms were recorded and integrated using Enhanced Chemstation

(version 01.00) software (Agilent Technologies; Palo Alto, Calif., U.S.A.). Mass spectral

matches were made by comparison with NIST 98.1 (NIST; Gaithersburg, Md., U.S.A.)

and WILEY 8.1 (Wiley; New York, N.Y., U.S.A.) mass spectral libraries. Only those

compounds with spectral fit values equal to or greater than 850 were considered

positive identifications. Linear retention index values were determined for both columns

using a series of alkanes (C5-C25) run under identical conditions.

Identification Procedures

Identifications were based on the combined matching of retention indices (LRI

values) from DB-5 and DB-Wax columns, matches made from spectra in the NIST and

WILEY libraries and linear retention index matches from literature.

Color Analysis

Color was measured using a ColorQuest XE colorimeter (HunterLab, Reston,

Va., U.S.A.). Samples of 40 mL where placed in a 20 mm cell and L a and b

parameters were recorded in total transmittance mode, illuminant D65, 100 observer

angle. Chroma (a*2 + b*2)1/2 and hue angle (arctan b*/a*) were calculated from the


120









measured a and b values. Chroma provides a measure of color intensity, while hue

angle (0 = red-purple, 90 = yellow, 1800 = bluish-green, 2700 = blue) indicates the

sample color itself (McGuire 1992).

Color difference (AE) values were also calculated using the following formula:

AE = (L -Lo)2 (a ao)2 +(b* -bo)2 (1)

Where Lo, ao, and bo are the reference values at storage time 0 week for each of the

treatments (CONTROL, DPCD, and HTST) and L, a, and b are the values at time t =

1,2,3,... weeks of storage.

Statistical Analysis

Sensory data was recorded and analyzed using Compusense five (Compusense,

Guelph, Ontario, Canada). Analysis of variance (ANOVA) and mean comparisons using

t-test and Tukey's test were conducted at the 5% significance level. Repeated

measures ANOVA and mean separation using Tukey's test (a=0.05) was performed to

evaluate the effect of treatment (fresh (CONTROL), thermal (HTST), and DPCD

processed) and storage time.(0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 weeks) on color

parameters using SAS statistical software (SAS Institute Inc., Cary, N.C., U.S.A.).

Results and Discussion

Physicochemical Analysis

The measured pH, oBrix, and TA for the CONTROL, DPCD, and HTST

beverages at weeks 0 and 5 of refrigerated storage are presented in Table 6-1. It is

important to notice that the CONTROL at week 5 was kept frozen in order to be used as

reference in the second taste panel. No significant differences in pH were observed

between the DPCD and HTST beverages at both week 0 and 5. Significant differences


121









in TA were found between DPCD and the CONTROL, and HTST beverages at both

storage times. A higher TA in the DPCD beverage can be a result of residual CO2 in the

beverage. For the same reason, a significantly higher oBrix value was found in the

DPCD beverage when compared with HTST. This higher value can also be a result of

residual carbonic acid. An increase in TA of DPCD treated coconut water (Damar and

others 2009) and orange juice (Kincal and others 2006) was also observed in previous

DPCD studies.

Sensory Evaluation

Two taste panels were conducted during storage. In the first test (week 0), 55%

of panelists were females, 91% of males and 83% of females were in the 18-30 age

range, while for the second panel (week 5), 48% of panelist were females, 74% of

males and 94% of females were in the 18-30 age range.

For the taste panel at week 0, there were no significant differences between the

CONTROL (hidden reference) and HTST beverages. However, significant differences

were detected by panelists between the CONTROL and DPCD beverages. The ranking

for overall likeability for the three tested beverages were not significantly different (Table

6-2) which indicates that regardless of treatment panelists preference remained the

same. For the taste panel at week 5 there were not significant differences between the

CONTROL (hidden reference) and HTST beverages. However, the DPCD was rated as

significantly different from the other two by panelists. Similarly, the ranking for overall

likeability showed no significant differences between the hidden reference and the

HTST beverage but the DPCD beverage was ranked significantly lower than the other

two beverages as shown in Table 6-2. Previous studies on muscadine grape juice and

coconut water showed greater differences in flavor and overall likeability between the


122









DPCD and thermally treated samples. They also found that the DPCD sample was very

similar to the hidden reference (Damar and others 2009; Del Pozo and others 2006a).

This may indicate that thermal processing affected more the organoleptic characteristics

of the grape juice and coconut water than in the hibiscus beverage.

In both taste panels, the DPCD beverage was different from the hidden reference

and at week 5, it was also significantly different from the HTST beverage. However,

despite these differences panelist ranking for overall likeability showed no significant

differences in the taste panel at week 0 and overall likeability values where even higher

in the taste panel at week 5. DPCD differences from the other two beverages can be

attributed to two possible factors. First, even when the DPCD beverage was degasified

before sensory analysis there could still be residual C02 remaining which would result

in a carbonated beverage mouth feel which may have also affected the

acidity/sweetness balance causing the panelists to perceive the beverage as less

sweet. This can be confirmed by a higher TA in the DPCD beverage as described

previously. Second, flavor compounds in the beverage will have dissolved in the

residual C02 and lost during degasification.

Aroma Compounds

Stability of aroma compounds were monitored during weeks 0 and 5 of storage.

The chemical composition of hibiscus beverages headspace volatiles are presented in

Figure 6-1. Total peak areas for all analyzed beverages were normalized to the total

peak area of CWO (CONTROL week 0). Alcohols and aldehydes constituted the major

fractions of hibiscus beverages aroma volatiles. A total of 4 aldehydes, 6 alcohols, 2

ketones, and 1 acid were considered for analysis. A slight decrease (21%) in alcohols

and ketones was observed in the DPCD beverage after processing (week 0) while there


123









was considerable loss (88%) in all volatiles in the thermally treated (HTST) beverage

which may indicate that there was decomposition of compounds due to heating and

possible formation of other compounds that were not studied in this experiment or that

volatiles were lost by evaporation during pasteurization. At week 5 of storage, there was

a considerable loss of all volatiles in the DPCD beverage (70%), a slight decrease in

alcohols and ketones (17%) for the CONTROL, and no changes in the volatiles for the

HTST beverage. The loss of DPCD volatiles at week 5 can be a result of decomposition

or formation of compounds or possibly they could be lost during degasification as

discussed previously. A previous study with melon juice also showed that DPCD

retained more volatile compounds as compared to the pasteurized juice (Chen and

others 2009).

The volatile compounds used in this analysis were identified by GC-MS and are

presented in Table 6-3. To compare the volatiles in the hibiscus beverages at weeks 0

and 5 of storage, peak areas were normalized (100) to the largest peak (1-Octen-3-ol)

in the CONTROL (C) week 0 sample. A total of 13 compounds were considered for

analysis, 6 of which (dehydroxylinalool oxide b, octanal, 1-hexanol, acetic acid, furfural,

and decanal) have been previously identified in hibiscus extracts (Gonzalez-Palomares

and others 2009; Chen and others 1998). The compounds present in highest

concentration in all six samples were 1-octen-3-ol, decanal, octanal, 1-hexanol, and

nonanal. A mushroom-like, fruity, citrus, and fruity aroma were associated with 1-octen-

3-ol, octanal, and nonanal respectively in lychee (Mahattanatawee and others 2007),

decanal was described as sweet, waxy, orange in merlot and cabernet wines (Gurbuz


124









and others 2006) while 1-hexanol was related to green, sweet notes in both lychee and

red wines.

Color Analysis

Color stability was assessed over 14 weeks of refrigerated storage. Treatment

(CONTROL, DPCD, HTST) had a significant effect (p < 0.0001) over storage time for all

the color parameters measured and calculated (L*, a*, b*, and hue angle, chroma, and

AE). There were slight changes for the L values in all three treatments over storage but

the trend was almost constant (Figure 6-2). At time 14 weeks of storage, the DPCD

beverage had a significantly lower L value as compared with HTST and CONTROL. As

can be seen in Figure 6-3, the a values decreased slightly over time for all the

treatments and after 14 weeks of storage there were no significant differences between

the DPCD and HTST beverages. The parameter that showed the most change was b

(Figure 6-4). In the three treatments, there was a slight but significant decrease of the b

values with time. There were significant differences between treatments at time 14 with

the CONTROL having the highest and the HTST beverage the lowest b value. Hue

angle slightly decreased over storage time for the CONTROL, DPCD, and HTST

beverages (Figure 6-5). These changes indicate that the beverages will follow a color

degradation pattern from a bright red color to a red-purple color. At time 14 weeks, there

were no significant differences in hue angle between the CONTROL and DPCD

beverages, the HTST sample was significantly lower. In the same way, chroma

decreased (the beverages became less intense in color) for all three samples (Figure 6-

6) and at time 14 weeks they were significantly different with the CONTROL having the

highest and HTST the lowest value. A decrease in hue angle and chroma during


125









storage (25 oC) was previously reported in a radish anthocyanin extract (Giusti and

Wrolstad 1996). The calculated AE value showed a significant increase over storage

time. Since the b value was the parameter that contributed the most to this difference in

color, the HTST beverage showed a significantly higher value after 14 weeks of storage

followed by the DPCD and CONTROL beverages as sown in Figure 6-7. Changes in

color during storage can be attributed to the degradation of anthocyanins which are the

pigments responsible for the red color in the hibiscus beverages with the HTST

beverage showing a higher change in color when compared to the DPCD beverage.

Conclusions

Changes in hibiscus aroma volatiles during storage did not affect panelists

overall likeability of the product. DPCD was found to be a viable technology for

processing hibiscus beverages since it maintained its characteristic red color over 14

weeks of storage and retained more aroma volatiles than the heat pasteurized

beverage. Possible losses of aroma volatiles during the degasification process can be

prevented by recovering them and adding them back to the beverage. Further studies

are needed to better understand the chemistry of hibiscus aroma compounds and to

reduce their loss during storage.


126









Table 6-1. Measured pH, oBrix, and titratable acidity (TA) (g of malic acid/100 mL of
beverage) at weeks 0 and 5 of refrigerated storage (4 C).


week 0
o Brix


week 5
o Brix


CONTROL 2.43b 9.70ab 0.37b 2.47b 9.67a 0.38b
DPCD 2.45a 9.77a 0.40a 2.48a 9.70a 0.41a
HTST 2.45a 9.63b 0.37b 2.48a 9.60b 0.38b


Data represents the mean of n=9.
(Tukey's HSD, p > 0.05).


Values with similar letters within columns are not significantly different


Table 6-2. Difference in flavor and overall likeability between fresh (reference and
hidden reference), dense phase-CO2 processed (DPCD; 34.5 MPa, 8% CO2,
6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverages
detected by untrained panelists (n = 75) at weeks 0 and 5 of refrigerated
storage (4 oC)


Difference observed when compared to given reference (difference from control test). Values with
similar letters within columns are not significantly different (Tukey's HSD, p > 0.05).


127












100

80
a-
60

40

0. 20


CWO DWO HWO CWOF DW5 HW5
Hibiscus Beverage Samples

NAldehydes (4) mAlcohols (6) Ketones (2) *Acids (1)


Figure 6-1. Chemical composition of hibiscus beverage headspace volatiles during
storage. Total number of compounds for each class is put in parenthesis. All
six samples were normalized to total peak area of the sample CWO
(CONTROL week 0). C = CONTROL, D = DPCD, H = HTST, W = week.


128









Table 6-3. MS identification of hibiscus beverage volatiles during storage. Peak areas were normalized (100) to the
largest peak (1-Octen-3-ol) in the CONTROL (C) week 0 sample.

Normalized peak areas (%)
LRI week 0 week 5
# Name CAS # DB-Wax DB5 C D H C D H
1 Dehydroxylinalool oxide a 13679-86-2 1210 993 48.9 14.4 5.2 29.7 5.96 5.1
2 Dehydroxylinalool oxide ba 13679-86-2 1246 1007 43.9 11.6 4.9 19.8 5.79 4.3
3 3-Octanone 106-68-3 1264 16.3 9.5 2.1 11.9 1.24 2.1
4 Octanala 124-13-0 1299 1002 83.5 49.6 8.5 70.9 8.52 8.8
5 6-methyl-5-Hepten-2-one 110-93-0 1355 989 45.5 36.7 5.5 38.7 8.13 6.3
6 1-Hexanola 111-27-3 1373 873 80.6 62.5 9.0 72.9 11.97 9.5
7 Nonanal 124-19-6 1405 1100 67.2 66.6 6.4 54.9 9.69 7.3
8 1-Octen-3-ol 3391-86-4 1468 983 100.0 78.8 11.7 81.6 44.10 11.9
9 Acetic acida 64-19-7 1485 35.4 29.6 3.6 24.5 7.66 3.2
10 Furfuralab 98-01-1 1496 832 15.9 26.0 3.4 7.5 7.50 2.9
11 Decanalb 112-31-2 1513 1204 99.5 123.5 15.9 111.0 43.80 14.0
12 1-Octanol 111-87-5 1577 1071 30.9 20.6 3.6 28.8 8.19 3.7
13 1-Nonanol 143-08-8 1678 1172 45.2 30.3 4.9 39.3 7.88 4.9
Total normalized peak area 712.7 559.7 84.6 591.3 170.4 84.0


C = CONTROL, D = DPCD, H = HTST. a Compounds previously reported
reported in H. sabdariffa by Gonzalez-Palomares and others (2009).


in H. sabdariffa by Chen and others (1998). Compounds previously


129















S- T I T
| J I


---CONTROL
--TDPCD
- HTST


0 2 4 6 8 10 12 14
Storage time (weeks)

Figure 6-2. L values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).




70

69

68
67 ----CONTROL
67 ---DPCD
HTST


0 2 4 6 8 10 12 14


Storage time (weeks)


Figure 6-3. a* values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).


130


47

46

.j 45

44

43

42










68
67
66
65
64
63
62
61
60
59


T
i1


0 2 4 6 8 10 12 14
Storage time (weeks)

Figure 6-4. b values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).


44.5

44

43.5

43

42.5


' T
-I
/


-CONTROL
---DPCD
-HTST


0 2 4 6 8 10 12 14


Storage time (weeks)

Figure 6-5. Hue angle values of unprocessed (CONTROL), dense phase-CO2
processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated
(HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C).


131


--CONTROL
--DPCD
SHTST

















92 -- ----CONTROL
92
90 ------DPCD
91 T --HTST
HTST
90

89
0 2 4 6 8 10 12 14
Storage time (weeks)

Figure 6-6. Chroma values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).


II


SI
..^~q^ ^^


---CONTROL
-- DPCD
-HTST


0 2 4 6 8 10 12 14
Storage time (weeks)

Figure 6-7. AE values of unprocessed (CONTROL), dense phase-CO2 processed
(DPCD; 34.5 MPa, 8% C02, 6.5 min, 40 C) and thermally treated (HTST; 75
C, 15 s) hibiscus beverage during refrigerated storage (4 C).


132









CHAPTER 7
SUMMARY AND CONCLUSIONS

Findings of this research can provide more flexibility to hibiscus processing.

Extraction and process selection for industrial applications should consider availability of

raw material (fresh or dried hibiscus), final product quality and phytochemical

characteristics, and economic considerations

Equivalent cold and hot water conditions (240 min at 25 oC and 16 min at 90 oC)

were found for anthocyanins extraction of dried hibiscus. Similar polyphenolic profiles

and chemical composition of aroma compounds were observed between fresh and

dried hibiscus extracts although differences in concentration were found. Fifteen aroma

compounds were identified for the first time. In general, hibiscus aroma is a combination

of earthy, green, floral, and fruity notes but the final flavor profile is affected by the

preservation and extraction process.

Solubility of CO2 in a hibiscus beverage (5.06 g CO21mL at 31.0 MPa) and

optimal processing conditions to inactivate microorganisms (34.5 MPa and 6.5 min for a

Y&M log reduction of 6.1) were determined. DPCD was found to be a viable technology

for processing hibiscus beverages since it extended its shelf life and maintained the

characteristic red color for 14 weeks of refrigerated storage. Quality attributes such as

pH an Brix were not affected by DPCD whereas TA increased. A loss of only 9% of

anthocyanins during storage was observed in the DPCD processed hibiscus beverage

which was lower as compared to a heat pasteurization process and no major changes

in total phenolics content and antioxidant capacity occurred during storage. Changes in

hibiscus aroma volatiles during storage did not affect panelists overall likeability of the

product.


133












APPENDIX A
EXTRACTION EXPERIMENT STATISTICAL ANALYSIS


Table A-1. SAS software output of statistical analysis for the anthocyanins concentration
data (AC) perfumed in the hibiscus extraction experiment (Chapter 3).


The GLM Procedure


Class
Treatment


Class Level Information
Levels Values
8 25-120 25-240 25-30 25-60 90-16 90-2 90-4 90-8


Number of observations


Dependent Variable: AC


Source
Model
Error
Corrected Total


R-Square
0.896002


Source
Treatment
Source
Treatment


Sum of
Squares
13655.68417
1585.00755
15240.69172


Coeff Var
8.748260


Mean Square
1950.81202
24.76574


Root MSE
4.976519


Type I SS
13655.68417
Type III SS
13655.68417


F Value Pr > F
78.77 <.0001


AC Mean
56.88582


Mean Square
1950.81202
Mean Square
1950.81202


F Value
78.77
F Value
78.77


Pr > F
<.0001
Pr > F
<.0001


Tukey's Studentized Range (HSD) Test for AC
NOTE: This test controls the Type I experimentwise error rate, but it generally has a
II error rate than REGWQ.


Alpha
Error Degrees of Freedom
Error Mean Square
Critical Value of Studentized Range
Minimum Significant Difference


higher Type


0.05
64
24.76574
4.43126
7.3507


Means with the same letter are not significantly different.


Tukey Grouping Mean
A 77.464
A 70.885
B 63.211
C B 58.159
C 55.436
C D 53.197
D 47.160
E 29.574


Treatment
90-16
25-240
90-8
25-120
90-4
25-60
90-2
25-30


-- -------------------------- Temperature=25 --
The GLM Procedure


Class
time


Class Level Information
Levels Values
4 30 60 120 240


134












Number of observations 36


Dependent Variable: AC


Source
Model
Error
Corrected Total


R-Square
0.913129


Source
time
Source
time


Sum of
Squares
8057.681364
766.575146
8824.256510


Coeff Var
9.242845


Mean Square
2685.893788
23.955473


Root MSE
4.894433


Type I SS
8057.681364
Type III SS
8057.681364


F Value Pr > F
112.12 <.0001


AC Mean
52.95375


Mean Square
2685.893788
Mean Square
2685.893788


F Value
112.12
F Value
112.12


Pr > F
<.0001
Pr > F
<.0001


Tukey's Studentized Range (HSD) Test for AC
NOTE: This test controls the Type I experimentwise error rate, but it generally has a
II error rate than REGWQ.
Alpha 0.05
Error Degrees of Freedom 32
Error Mean Square 23.95547
Critical Value of Studentized Range 3.83162
Minimum Significant Difference 6.2512

Means with the same letter are not significantly different.


Tukey Grouping


Mean
70.885
58.159
53.197
29.574


time
240
120
60
30


------------------------- Temperature=90
The GLM Procedure
Dependent Variable: AC


Source
Model
Error
Corrected Total


R-Square
0.845673


Source
time
Source
time


Sum of
Squares
4484.797153
818.432403
5303.229555


Coeff Var
8.315437


Mean Square
1494.932384
25.576013


Root MSE
5.057273


Type I SS
4484.797153
Type III SS
4484.797153


F Value Pr > F
58.45 <.0001


AC Mean
60.81789


Mean Square
1494.932384
Mean Square
1494.932384


F Value
58.45
F Value
58.45


Pr > F
<.0001
Pr > F
<.0001


135


higher Type












Tukey's Studentized Range (HSD) Test for AC
NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type
II error rate than REGWQ.
Alpha 0.05
Error Degrees of Freedom 32
Error Mean Square 25.57601
Critical Value of Studentized Range 3.83162
Minimum Significant Difference 6.4592


Means with the same
Tukey Grouping
A
B
C
D


letter are
Mean
77.464
63.211
55.436
47.160


not significantly different.
N time
9 16
9 8
9 4
9 2


136












APPENDIX B
STORAGE EXPERIMENT STATISTICAL ANALYSIS


Table B-1. SAS software output of statistical analysis for the anthocyanins concentration
data (AC) perfumed in the hibiscus storage experiment (Chapter 5).


The GLM Procedure


Class
time
Treatment


Class Level
Levels
11
3


Information
Values
0 1 2 3 4 5 6 8 10 12 14
CDT


Number of observations


Dependent Variable: AC


Source
Model
Error
Corrected Total


Sum of
Squares
56.9747574
888.2431136
945.2178710


Mean Square
28.4873787
3.0212351


F Value Pr > F
9.43 0.0001


R-Square Coeff Var
0.060277 4.330435


Root MSE
1.738170


Type I SS
56.97475742
Type III SS
56.97475742


AC Mean
40.13847


Mean Square
28.48737871
Mean Square
28.48737871


F Value
9.43
F Value
9.43


Pr > F
0.0001
Pr > F
0.0001


Repeated Measures Analysis of Variance

Repeated Measures Level Information
Level of time 1 2 3 4 5 6 7 8 9 10


MANOVA Test Criteria and Exact F Statistics for the Hypothesis
H = Type III SSCP Matrix for time
E = Error SSCP Matrix

S=1 M=3.5 N=142


of no time Effect


Statistic
Wilks' Lambda
Pillai's Trace
Hotelling-Lawley Trace
Roy's Greatest Root


Value
0.00000
1.00000
215234.11193
215234.11193


MANOVA Test Criteria and F Approximations for the Hypothesis of no
H = Type III SSCP Matrix for time*Treatment
E = Error SSCP Matrix

S=2 M=3 N=142


Statistic


time*Treatment Effect


Value F Value Num DF Den DF Pr > F


137


Source
Treatment
Source
Treatment


F Value
6839662
6839662
6839662
6839662


Num DF
9
9
9
9


Den DF
286
286
286
286


Pr > F
<.0001
<.0001
<.0001
<.0001












Wilks' Lambda
Pillai's Trace
Hotelling-Lawley Trace
Roy's Greatest Root


0.37623085
0.74964003
1.32338497
0.98306355


20.03
19.12
20.97
31.35


NOTE: F Statistic for Roy's Greatest Root is an upper bound.
NOTE: F Statistic for Wilks' Lambda is exact.

Tests of Hypotheses for Between Subjects Effects


DF Type III SS
2 237.565234
294 1481.873250


Mean Square
118.782617
5.040385


F Value Pr > F
23.57 <.0001


Univariate Tests of Hypotheses for Within Subject Effects


Source
time
time*Treatment
Error(time)


DF
9
18
2646


Type III SS
15393156.09
1280.52
11765.26


Mean Square
1710350.68
71.14
4.45


Greenhouse-Geisser Epsilon
Huynh-Feldt Epsilon


F Value
384657
16.00



0.1367
0.1379


Pr > F
<.0001
<.0001


Adj Pr > F
G G H F
<.0001 <.0001
<.0001 <.0001


----------- ---------------------------------- time=0 ------------------------------------------


The GLM Procedure


Class
Treat


Class Level Information
ss Levels Val
itment 3 C


Number of observations


Lues
ST

27


Dependent Variable: AC


Source
Model
Error
Corrected Total


Sum of
Squares
12.57264263
18.44201807
31.01466070


Mean Square
6.28632131
0.76841742


F Value Pr > F
8.18 0.0020


R-Square Coeff Var
0.405377 2.048539


Type I SS
12.57264263
Type III SS
12.57264263


Tukey's Studentized Range (HSD) Test for AC


NOTE: This test controls the Type I experimentwise error rate, but it generally has a
II error rate than REGWQ.


Alpha
Error Degrees of Freedom
Error Mean Square
Critical Value of Studentized Range
Minimum Significant Difference


higher Type


0.05
24
0.768417
3.53170
1.032


138


572
574
472.9
287


<.0001
<.0001
<.0001
<.0001


Source
Treatment
Error


Source
Treatment
Source
Treatment


Root MSE
0.876594


AC Mean
42.79120


Mean Square
6.28632131
Mean Square
6.28632131


F Value
8.18
F Value
8.18


Pr > F
0.0020
Pr > F
0.0020













Means with the same letter are not significantly different.
Tukey Grouping Mean N Treatment
A 43.7506 9 C
B 42.4015 9 D
B 42.2215 9 T


---------------------------- time=14


The GLM Procedure
Dependent Variable: AC


Source
Model
Error
Corrected Total


R-Square
0.830003


Source
Treatment
Source
Treatment


Sum of
Squares
56.96992981
11.66832881
68.63825862


Coeff Var
1.824962


Mean Square
28.48496491
0.48618037


Root MSE
0.697266


Type I SS
56.96992981
Type III SS
56.96992981


F Value Pr > F
58.59 <.0001


AC Mean
38.20717


Mean Square
28.48496491
Mean Square
28.48496491


F Value
58.59
F Value
58.59


Pr > F
<.0001
Pr > F
<.0001


Tukey's Studentized Range (HSD) Test for AC

NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type
II error rate than REGWQ.
Alpha 0.05
Error Degrees of Freedom 24
Error Mean Square 0.48618
Critical Value of Studentized Range 3.53170
Minimum Significant Difference 0.8208


Means with the
Tukey Grouping
A
B
C


same letter are
Mean
39.7695
38.5811
36.2709


significantly different.
Treatment
C
D
T


139









APPENDIX C
HIBISCUS SABDARIFFA PICTURES


I-


G H I

Figure B-1. Pictures of dried hibiscus (A), dried hibiscus extraction process (B), hibiscus
beverage (C), hibiscus beverage in the dense phase carbon dioxide (DPCD)
feed tank (D), DPCD processing equipment (E), DPCD processed hibiscus
beverage (F), hibiscus beverage samples for analysis (G), hibiscus beverage
under refrigerated storage (H), and DPCD processed hibiscus beverage after
14 weeks of storage at 4 C (I). Photos by Milena Ramirez.


140









LIST OF REFERENCES


AI-Kahtani HA & Hassan BH. 1990. Spray drying of roselle (Hibiscus sabdariffa L.)
extract. J Food Sci 55(4):1073-1076.

AI-Wandawi H, AI-Shaikhly K & Abdul-Rahman M. 1984. Roselle seeds: a new protein
source. J Agric Food Chem 32(3):510-512.

Andrade I & Flores H. 2004. Optimization of spray drying of roselle extract (Hibiscus
sabdariffa L.). Proceedings of the 14th International Drying Symposium (IDS
2004). Sao Paulo, Brazil, 22-25 August 2004. p. 597-604.

Anon. 2006. Is their something brewing in the world of hibiscus?. New nutrition business
12(1):19-22.

Ashurst PR. 2005. Chemistry and technology of soft drinks and fruit juices. 2nd ed.
Oxford UK: Blackwell Publishing. 374 p.

Balaban MO, Ferrentino G, Ramirez M & Plaza ML. 2008. Review of dense phase
carbon dioxide applications to citrus juices and new development. Proceding of
the 54th Citrus Engineering Conference. Lake Alfred, Fla.

Beristain CI, Mendoza RE, Garcia HS & Vazquez A. 1994. Cocrystallization of jamaica
(Hibiscus sabdarifa L.) granules. Lebensm Wiss u Technol 27(4):347-349.

Bolade MK, Oluwalana IB & Ojo O. 2009. Commercial practice of roselle (Hibiscus
sabdariffa L.) beverage production: optimization of hot water extraction and
sweetness level. World Journal of Agricultural Sciences 5(1): 126-131.

Bravo L. 1998. Polyphenols: chemistry, dietary sources, metabolism, and nutritional
significance. Nutr Rev 56(11):317-333.

Buttery RG, Teranishi R & Ling LC. 1987. Fresh tomato aroma volatiles: a quantitative
study. J Agric Food Chem 35(4):540-544.

Calix TF, Ferrentino G & Balaban MO. 2008. Measurement of high-pressure carbon
dioxide solubility in orange juice, apple juice, and model liquid foods. J Food Sci
73(9):E439-E445.

Carvajal-Zarrabal O, Waliszewski S, Barradas-Dermitz D, Orta-Flores Z, Hayward-
Jones P, Nolasco-Hipolito C, Angulo-Guerrero O, Sanchez-Ricaho R, Infanz6n R
& Trujillo P. 2005. The consumption of Hibiscus sabdariffa dried calyx ethanolic
extract reduced lipid profile in rats. Plant Food Hum Nutr (Formerly 60(4):153-
159.


141









Castaieda-Ovando A, Pacheco-Hernandez MdL, Paez-Hernandez ME, Rodriguez JA &
Galan-Vidal CA. 2009. Chemical studies of anthocyanins: A review. Food Chem
113(4):859-871.

Cisse M, Vaillant F, Acosta O, Dhuique-Mayer C & Dornier M. 2009. Thermal
degradation kinetics of anthocyanins from blood orange, blackberry, and roselle
using the Arrhenius, Eyring, and Ball models. J Agric Food Chem 57(14):6285-
6291.

Chang Y-C, Huang H-P, Hsu J-D, Yang S-F & Wang C-J. 2005. Hibiscus anthocyanins
rich extract-induced apoptotic cell death in human promyelocytic leukemia cells.
Toxicol Appl Pharm 205(3):201-212.

Chen CC, Hsu JD, Wang SF, Chiang HC, Yang MY, Kao ES, Ho YC & Wang CJ. 2003.
Hibiscus sabdariffa extract inhibits the development of atherosclerosis in
cholesterol-fed rabbits. J Agric Food Chem 51(18):5472-5477.

Chen J, Zhang J, Feng Z, Song L, Wu J & Hu X. 2009. Influence of thermal and dense-
phase carbon dioxide pasteurization on physicochemical properties and flavor
compounds in Hami melon juice. J Agric Food Chem 57(13):5805-5808.

Chen SH, Huang TC, Ho CT & Tsai PJ. 1998. Extraction, analysis, and study on the
volatiles in roselle tea. J Agric Food Chem 46(3):1101-1105.

Clifford MN. 2000. Anthocyanins nature, occurrence and dietary burden. J Sci Food
Agr 80(7):1063-1072.

Clifford MN, Johnston KL, Knight S & Kuhnert N. 2003. Hierarchical scheme for LC-MSn
identification of chlorogenic acids. J Agric Food Chem 51(10):2900-2911.

Clydesdale FM, Main JH & Francis FJ. 1979. Roselle (Hibiscus sabdariffa L.)
anthocyanins as colorants for beverages and gelatin desserts. J Food Protect
42(3):204-207.

Dagan GF & Balaban MO. 2006. Pasteurization of beer by a continuous dense-phase
CO2 system. J Food Sci 71(3):E164-E169.

Damar S & Balaban MO. 2006. Review of dense phase CO2 technology: Microbial and
enzyme inactivation, and effects on food quality. J Food Sci 71(1):R1-R11.

Damar S, Balaban MO & Sims CA. 2009. Continuous dense-phase CO2 processing of a
coconut water beverage. Int J Food Sci Tech 44(4):666-673.

De Castro NE, Pinto JE, Cardoso MG, Morais de A, Bertolucci SK, Silva da F & Delu N.
2004. Planting time for maximization of yield of vinegar plant calyx (Hibiscus
sabdariffa L). Cienc Agrotec Lavras 28(3):542-551.


142









Degenhardt A, Knapp H & Winterhalter P. 2000. Separation and purification of
anthocyanins by high-speed countercurrent chromatography and screening for
antioxidant activity. J Agric Food Chem 48(2):338-343.

Del Pozo-lnsfran D, Balaban MO & Talcott ST. 2006a. Microbial stability, phytochemical
retention, and organoleptic attributes of dense phase CO2 processed muscadine
grape juice. J Agric Food Chem 54(15):5468-5473.

Del Pozo-lnsfran D, Balaban MO & Talcott ST. 2006b. Enhancing the retention of
phytochemicals and organoleptic attributes in muscadine grape juice through a
combined approach between dense phase CO2 processing and copigmentation.
J Agric Food Chem 54(18):6705-6712.

Del Pozo-lnsfran D, Balaban MO & Talcott ST. 2007. Inactivation of polyphenol oxidase
in muscadine grape juice by dense phase-CO2 processing. Food Res Int
40(7):894-899.

Delgado-Vargas F, Jimenez A & Paredes-Lopez O. 2000. Natural pigments:
carotenoids, anthocyanins, and betalains characteristics, biosynthesis,
processing, and stability. Crit Rev Food Sci 40(3): 173 289.

Del Rio D, Stewart AJ, Mullen W, Burns J, Lean MEJ, Brighenti F & Crozier A. 2004.
HPLC-MSn analysis of phenolic compounds and purine alkaloids in green and
black tea. J Agric Food Chem 52(10):2807-2815.

Dominguez-Lopez A, Remondetto GE & Navarro-Galindo S. 2008. Thermal kinetic
degradation of anthocyanins in a roselle (Hibiscus sabdariffa L. cv. "Criollo")
infusion. International J Food Sci Tech 43(2):322-325.

Du CT & Francis FJ. 1973. Anthocyanins of roselle (Hibiscus sabdariffa, L.). J Food Sci
38(5):810-812.

Duangmal K, Saicheua B & Sueeprasan S. 2008. Colour evaluation of freeze-dried
roselle extract as a natural food colorant in a model system of a drink. LWT-Food
Sci Technol 41(8):1437-1445.

Duh P-D & Yen G-C. 1997. Antioxidative activity of three herbal water extracts. Food
Chem 60(4):639-645.

EI-Adawy TA & Khalil AH. 1994. Characteristics of Roselle seeds as a new source of
protein and lipid. J Agric Food Chem 42(9):1896-1900.

Espin JC, Soler-Rivas C, Wichers HJ & Garcia-Viguera C. 2000. Anthocyanin-based
natural colorants: A new source of antiradical activity for foodstuff. J Agric Food
Chem 48(5):1588-1592.


143









Esselen WB & Sammy GM. 1975. Applications for roselle as a red food colorant
Food Prod Dev 9(8):37-38, 40.

Fang N, Yu S & Prior RL. 2002. LC/MS/MS characterization of phenolic constituents in
dried plums. J Agric Food Chem 50(12):3579-3585.

Ferrentino G, Plaza ML, Ramirez-Rodrigues M, Ferrari G & Balaban MO. 2009. Effects
of dense phase carbon dioxide pasteurization on the physical and quality
attributes of a red grapefruit juice. J Food Sci 74(6):E333-E341.

Giusti MM & Wrolstad RE. 1996. Radish Anthocyanin Extract as a Natural Red Colorant
for Maraschino Cherries. J Food Sci 61(4):688-694.

Giusti MM, Rodriguez-Saona LE, Griffin D & Wrolstad RE. 1999. Electrospray and
tandem mass spectroscopy as tools for anthocyanin characterization. J Agric
Food Chem 47(11):4657-4664.

Giusti MM & Wrolstad RE. 2005. Characterization and measurement of anthocyanins by
UV-visible spectroscopy. In: Wrolstad RE, Acree TE, Decker EA, Penner MH,
Reid DS, Schwartz SJ, Shoemaker CF, Smith D & Sporns P, editors. Handbook
of food analytical chemistry. Hoboken, NJ: John Wiley & Sons Inc.

Gonzalez-Palomares S, Estarron-Espinosa M, Gomez-Leyva JF & Andrade-Gonzalez I.
2009. Effect of the temperature on the spray drying of Roselle extracts (Hibiscus
sabdariffa L.). Plant Food Hum Nutr 64(1):62-67.

Gradinaru G, Biliaderis CG, Kallithraka S, Kefalas P & Garcia-Viguera C. 2003. Thermal
stability of Hibiscus sabdariffa L. anthocyanins in solution and in solid state:
effects of copigmentation and glass transition. Food Chem 83(3):423-436.

Gurbuz O, Rouseff JM & Rouseff RL. 2006. Comparison of aroma volatiles in
commercial merlot and cabernet sauvignon wines using gas chromatography-
olfactometry and gas chromatography-mass spectrometry. J Agric Food Chem
54(11):3990-3996.

Hainida KIE, Amin I, Normah H & Mohd.-Esa N. 2008. Nutritional and amino acid
contents of differently treated Roselle (Hibiscus sabdariffa L.) seeds. Food Chem
111(4):906-911.

Haji Faraji M & Haji Tarkhani AH. 1999. The effect of sour tea (Hibiscus sabdariffa) on
essential hypertension. J Ethnopharmacol 65(3):231-236.

Hansawasdi C, Kawabata J & Kasai T. 2000. a-Amylase inhibitors from roselle
(Hibiscus sabdariffa Linn.) tea. Biosci Biotechnol Biochem 64(5):1041-1043.


144









Hassan BH & Hobani Al. 1998. Flow properties of roselle (Hibiscus sabdariffa L.)
extract. J Food Eng 35(4):459-470.

Herrera-Arellano A, Flores-Romero S, Chavez-Soto M & Tortoriello J. 2004.
Effectiveness and tolerability of a standardized extract from Hibiscus sabdariffa in
patients with mild to moderate hypertension: a controlled and randomized clinical
trial. Phytomedicine 11(375-382).

Hirunpanich V, Utaipat A, Morales NP, Bunyapraphatsara N, Sato H, Herunsale A &
Suthisisang C. 2006. Hypocholesterolemic and antioxidant effects of aqueous
extracts from the dried calyx of Hibiscus sabdariffa L. in hypercholesterolemic
rats. J Ethnopharmacol 103(2):252-260.

Huang D, Ou B, Hampsch-Woodill M, Flanagan JA & Prior RL. 2002. High-throughput
assay of oxygen radical absorbance capacity (ORAC) using a multichannel liquid
handling system coupled with a microplate fluorescence reader in 96-well format.
J Agric Food Chem 50(16):4437-4444.

Hou D-X, Tong X, Terahara N, Luo D & Fujii M. 2005. Delphinidin 3-sambubioside, a
Hibiscus anthocyanin, induces apoptosis in human leukemia cells through
reactive oxygen species-mediated mitochondrial pathway. Arch Biochem Biophys
440(1):101-109.

Ito Y, Sugimoto A, Kakuda T & Kubota K. 2002. Identification of potent odorants in
Chinese jasmine green tea scented with flowers of Jasminum sambac. J Agric
Food Chem 50(17):4878-4884.

Jackman RL, Yada RY, Tung MA & Speers RA. 1987. Anthcoyanins as food colorants-a
review. J Food Biochem 11:201-247.

Jabalpurwala FA, Smoot JM & Rouseff RL. 2009. A comparison of citrus blossom
volatiles. Phytochemistry 70(11-12): 1428-1434.

Juliani HR, Welch CR, Wu Q, Diouf B, Malainy D & Simon JE. 2009. Chemistry and
quality of hibiscus (Hibiscus sabdariffa) for developing the natural-product
industry in Senegal. J Food Sci 74(2):S113-S121.

Kawakami M, Kobayashi A & Kator K. 1993. Volatile constituents of Rooibos tea
(Aspalathus linearis) as affected by extraction process. J Agric Food Chem
41(4):633-636.

Kincal D, Hill WS, Balaban MO, Portier KM, Wei CI & Marshall MR. 2005. A continuous
high pressure carbon dioxide system for microbial reduction in orange juice. J
Food Sci 70(5):M249-M254.


145









Kincal D, Hill WS, Balaban M, Portier KM, Sims CA, Wei CI & Marshall MR. 2006. A
continuous high-pressure carbon dioxide system for cloud and quality retention in
orange juice. J Food Sci 71(6):C338-C344.

Kong J-M, Chia L-S, Goh N-K, Chia T-F & Brouillard R. 2003. Analysis and biological
activities of anthocyanins. Phytochemistry 64(5):923-933.

Lawless H & Heymann H. 1998. Sensory evaluation of food: principles and practices.
NY: Chapman & Hall. 819 p.

Lee MJ, Chou FP, Tseng TH, Hsieh MH, Lin MC & Wang CJ. 2002. Hibiscus
protocatechuic acid or esculetin can inhibit oxidative LDL induced by either
copper ion or nitric oxide donor. J Agric Food Chem 50(7):2130-2136.

Lim S, Yagiz Y & Balaban MO. 2006. Continuous high pressure carbon dioxide
processing of mandarin juice. Food Sci Biotechnol 15(1): 13-18.

Lin LZ & Harnly JM. 2008. Phenolic compounds and chromatographic profiles of pear
skins (Pyrus spp.). J Agric Food Chem 56(19):9094-9101.

Lin WL, Hsieh YJ, Chou FP, Wang CJ, Cheng MT & Tseng TH. 2003. Hibiscus
protocatechuic acid inhibits lipopolysaccharide-induced rat hepatic damage. Arch
Toxicol 77(1):42-47.

Lin TL, Lin HH, Chen C-C, Lin M-C, Chou M-C & Wang C-J. 2007. Hibiscus sabdariffa
extract reduces serum cholesterol in men and women. Nutr Res 27(3):140-145.

Liu CL, Wang JM, Chu CY, Cheng M-T & Tseng TH. 2002. In vivo protective effect of
protocatechuic acid on tert-butyl hydroperoxide-induced rat hepatotoxicity. Food
Chem Toxicol 40:635-641.

Liu JY, Chen CC, Wang WH, Hsu J-D, Yang MY & Wang CJ. 2006. The protective
effects of Hibiscus sabdariffa extract on CCI4-induced liver fibrosis in rats. Food
Chem Toxicol 44(3):336-343.

Lule SU & Xia W. 2005. Food phenolics, pros and cons: a review. Food Reviews
International 21(4):367 388.

Mahattanatawee K, Perez-Cacho PR, Davenport T & Rouseff R. 2007. Comparison of
three lychee cultivar odor profiles using gas chromatography-olfactometry and
gas chromatography-sulfur detection. J Agric Food Chem 55(5): 1939-1944.

Mazza G & Brouillard R. 1990. The mechanism of co-pigmentation of anthocyanins in
aqueous solutions. Phytochemistry 29(4): 1097-1102.


146









Mazza G, Cacace JE & Colin DK. 2004. Methods of analysis for anthocyanins in plants
and biological fluids. J AOAC Int 87(1):129-145.

McGuire RG. 1992. Reporting of Objective Color Measurements. HortScience
27(12): 1254-1255.

Meilgaard MC, Civille GV & Carr BT. 2007. Sensory evaluation techniques,4th ed. Boca
Raton, FL: CRC Press.

Mintel. 2008. RTD Non-carbonated beverages U.S. June 2008. Available from:
http://www.marketresearch.com. Accessed: Feb 20, 2009.

Morton J. 1987. Roselle. In: Morton J, editor. Fruits of warm climates. Miami: Creative
Resources systems Inc. p 281-286.

Mounigan P & Badrie N. 2007. Physicochemical and sensory quality of wines from red
sorrel/roselle (Hibiscus sabdariffa L.) calyces: effects of pretreatments of
pectolase and temperature/time. Int J Food Sci Technol 42(4):469-475.

Mourtzinos I, Makris DP, Yannakopoulou K, Kalogeropoulos N, Michali I & Karathanos
VT. 2008. Thermal stability of anthocyanin extract of Hibiscus sabdariffa L. in the
presence of ,-Cyclodextrin. J Agric Food Chem 56(21):10303-10310.

Mueller BM & Franz G. 1992. Chemical structure and biological activity of
polysaccharides from Hibiscus sabdariffa. Planta Med. 58(1):60-67.

Oboh G & Elusiyan CA. 2004. Nutrient composition and antimicroibal activity of sorrel
drinks (soborodo). J Med Food 7(3):340-342.

Odigie IP, Ettarh RR & Adigun SA. 2003. Chronic administration of aqueous extract of
Hibiscus sabdariffa attenuates hypertension and reverses cardiac hypertrophy in
2K-1C hypertensive rats. J Ethnopharmacol 86(2-3):181-185.

Olvera-Garcia V, Castaio-Tostado E, Rezendiz-Lopez RI, Reynoso-Camacho R,
Gonzalez de Mejia E, Elizondo G & Loarca-Piia G. 2008. Hibiscus sabdariffa L.
extracts inhibit the mutagenicity in microsuspension assay and the proliferation of
HeLa cells. J Food Sci 73(5):T75-T81.

Parr AJ & Bolwell PG. 2000. Phenols in the plant and in man. The potential for possible
nutritional enhancement of the diet by modifying the phenols content or profile. J
Sci Food Agric 80(7):985-1012.

Plotto A. 1999. Hibiscus: post-production management for improved market access for
herbs and spices. Compendium on post-harvest operations. Available from:
http://www.fao.org/inpho/content/compend/text/ch28/ch28.htm. Accessed Apr 13,
2005.


147









Pouget M, Lejeune B, Vennat B & Pourrat A. 1990a. Extraction, Analysis and Study of
the Stability of Hibiscus Anthocyanins. Lebensm Wiss u Technol 23:103-105.

Pouget M, Vennat B, Lejeune B & Pourrat A. 1990b. Identification of Anthocyanins of
Hibiscus sabdariffa L. Lebensm. Wiss u Technol 23:101-102.

Prenesti E, Berto S, Daniele PG & Toso S. 2007. Antioxidant power quantification of
decoction and cold infusions of Hibiscus sabdariffa flowers. Food Chem
100(2):433-438.

Reineccius G. 2006. Flavor chemistry and technology. Boca Raton FL: CRC Press
Taylor & Francis Group.

Rodriguez-Saona LE & Wrolstad RE. 2005. Extractionn, isolation, and purification of
anthcoyanins. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS,
Schwartz SJ, Shoemaker CF, Smith D & Sporns P, editors. Handbook of food
analytical chemistry. Hoboken, NJ: John Wiley & Sons Inc.

Roethenbaugh G. 2005. Trends in beverage markets. In: Ashurst P, editor. Chemistry
and technology of soft drinks and fruit juices. 2nd ed. Oxford, UK: Blackwell
Publishing. p 15-34.

Sayago-Ayerdi SG, Arranz S, Serrano J & Gohi I. 2007. Dietary fiber content and
associated antioxidant compounds in roselle flower (Hibiscus sabdariffa L.)
beverage. J Agric Food Chem 55(19):7886-7890.

Segura-Carretero A, Puertas-Mejia MA, Cortacero-Ramirez S, Beltran R, Alonso-
Villaverde C, Joven J, Dinelli G & Fernandez-Gutierrez A. 2008. Selective
extraction, separation, and identification of anthocyanins from Hibiscus sabdariffa
L. using solid phase extraction-capillary electrophoresis-mass spectrometry
(time-of-flight /ion trap). Electrophoresis 29(13):2852-2861.

Shahidi F & Naczk M. 2004. Phenolics in food and nutraceuticals. Boca Raton FL: CRC
Press. 558 p.

Stephens JM. 2003. Roselle Hibiscus sabdariffa L. University of Florida IFAS
Extension #HS659.

Suboh SM, Bilto YY & Aburjai TA. 2004. Protective effects of selected medicinal plants
against protein degradation, lipid peroxidation and deformability loss of
oxidatively stressed human erythrocytes. Phytother Res 18(4):280-284.

Tee PL, Yusof S & Mohamed S. 2002. Antioxidative properties of Roselle (Hibiscus
sabdariffa L.) in linoleic acid model system. Nutr Food Sci 32(1):17-20.


148









Tsai PJ, Mclntosh J, Pearce P, Camden B & Jordan BR. 2002. Anthocyanin and
antioxidant capacity in Roselle (Hibiscus sabdariffa L.) extract. Food Res Int
35(4):351-356.

Tsai PJ, Hsieh YY & Huang TC. 2004. Effect of sugar on anthocyanin degradation and
water mobility in a roselle anthocyanin model system using 170 NMR. J Agric
Food Chem 52(10):3097-3099.

Tsai PJ & Huang HP. 2004. Effect of polymerization on the antioxidant capacity of
anthocyanins in Roselle. Food Res Int 37(4):313-318.

Tseng TH, Hsu JD, Lo MH, Chu CY, Chou FP, Huang CL & Wang CJ. 1998. Inhibitory
effect of Hibiscus protocatechuic acid on tumor promotion in mouse skin. Cancer
Lett 126(2): 199-207.

U.S. Census Bureau. 2009. U.S. Population Projections. Available from:
http://www.census.gov/population/www/projections/downloadablefiles.html.
Accessed: Mar 7, 2009.

USDA (U.S. Department of Agriculture). 2009. National nutrient database for standard
reference. Available from: http://www.nal.usda.gov/fnic/foodcomp/cgi-
bin/list_nut_edit.pl. Accessed Sep 15, 2009.

Waldemar W, Magdalena M & Janusz Co. 2004. A review of theoretical and practical
aspects of solid-phase microextraction in food analysis. Int J Food Sci Tech
39(7):703-717.

Wang H, Cao G & Prior RL. 1997. Oxygen radical absorbing capacity of anthocyanins. J
Agric Food Chem 45(2):304-309.

Wang CJ, Wang JM, Lin WL, Chu CY, Chou FP & Tseng TH. 2000. Protective effect of
Hibiscus anthocyanins against tert-butyl hydroperoxide-induced hepatic toxicity in
rats. Food Chem Toxicol 38(5):411-416.

Waterhouse AL. 2005. Determination of total phenolics. In: Wrolstad, R. E., Acree, T.
E., Decker, E. A., Penner, M. H., Reid, D. S., Schwartz, S. J., Shoemaker, C. F.,
Smith, D. & Sporns, P., editors. Handbook of food analytical chemistry. Hoboken,
NJ: John Wiley & Sons Inc.

Wong PK, Yusof S, Ghazali HM & Che Man Y. 2002. Physicochemical characteristics of
Roselle (Hibiscus sabdariffa L.). Nutr Food Sci 32(2):68-73.

Wong PK, Yusof S, Ghazali HM & Che Man YB. 2003. Optimization of hot water
extraction of roselle juice using response surface methodology: a comparative
study with other extraction methods. J Sci Food Agric 83:1273-1278.


149









Wong SP, Leong LP & William Koh JH. 2006. Antioxidant activities of aqueous extracts
of selected plants. Food Chem 99(4):775-783.

Wrobel K, Wrobel K & Urbina EMC. 2000. Determination of total aluminum, chromium,
copper, iron, manganese, and nickel and their fractions leached to the infusions
of black tea, green tea, Hibiscus sabdariffa, and Ilex paraguariensis (mate) by
ETA-AAS. Biol. Trace Elem. Res. 78(1-3):271-280.

Wrolstad RE. 2004. Anthocyanin pigments-bioactivity and coloring properties. J Food
Sci 69(5):C419-C425.


150









BIOGRAPHICAL SKETCH

Milena M. Ramirez Rodrigues was born in Puebla, Mexico. After graduating from

high school (American School of Puebla) in July 1998, she enrolled in the Food

Engineering program of the Universidad de las Americas-Puebla (UDLA). Before

finishing her bachelor's she was offered an assistantship to pursue a master's in food

science. In July 2005 she was awarded a scholarship from CONACyT (National

Mexican Council of Science and Technology) and had the opportunity to pursue her

Ph.D. in Food Science at the University of Florida. While at UF she decided to enroll in

the agribusiness master's program, from which she graduated in May 2009. After

finishing her Ph.D., Milena hopes to continue exploring her interests in new product

development and marketing.


151





PAGE 1

1 PROCESSING HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE By MILENA MARIA RAMIREZ RODRIGUES A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Milena Maria Ramirez Rodrigues

PAGE 3

3 To my parents and sister

PAGE 4

4 ACKNOWLEDGMENTS First thanks to God. Second I want to t hank my parents Aidil and Agust n and my sister Melissa for their love, dedication, and support and for helping me fulfill my dreams and achieve what I hav e in life. I would like to thank my advisor Dr. Mur at Balaban for his guidance, for sharing his knowledge and experience and for his care. I extend a special acknowledgment to Dr. Marty Marshall who guided me through t he second part of this research for his great support, advice and care. I also want to thank my other advising committee members, Dr. Allen Wys ocki a nd Dr. Jos Reye s for their help in my research and Dr. Russell Rouseff and Charles Sims for all their help and advice in the flavor chemistry and sensory evaluation part s of my research. Finally, I want to thank my close friends for helping me out in every way possible and for the good times shared and to the countless persons who in some way or another have contributed to my success in the UF Food Science Graduate Program.

PAGE 5

5 TABLE O F CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 11 ABSTRACT ................................................................................................................... 14 CHAPTER 1 INTRODUCTION .................................................................................................... 16 Justification ............................................................................................................. 16 Hypothesis .............................................................................................................. 17 Specific Objectives ................................................................................................. 17 2 LITERATURE REVIEW .......................................................................................... 19 The Beverage Market ............................................................................................. 19 Market Performance and Competitive Context of U.S. Ready to Drink NonCarbonated Beverages ................................................................................. 20 Consumption and Demographic Trends ........................................................... 23 Hibiscus ( Hibiscus sabdariffa) ................................................................................. 24 Characteristics and Economic Importance ....................................................... 24 Commercial Hibiscus Products ......................................................................... 27 Hibiscus Lemon Bissap .............................................................................. 27 Caita Aguas Frescas (jamaica (hibiscus) flavor) ...................................... 28 Squish Hibiscus Press ............................................................................. 28 Simply Hibi ................................................................................................. 28 Composition and Associated Health Benefits ................................................... 28 Hibiscus Extraction Process ............................................................................. 34 Hibiscus Flavor ................................................................................................. 34 Phenolic compounds ............................................................................................... 37 Classification .................................................................................................... 37 Phenolic Compounds Attributes ....................................................................... 38 Contribution to flavor .................................................................................. 38 Antioxidant potential ................................................................................... 41 Anticarcinogenic action .............................................................................. 42 Anthocyanins .......................................................................................................... 42 Classification .................................................................................................... 42 Stability ............................................................................................................. 44 Health Benefits ................................................................................................. 46 Beverage Processing .............................................................................................. 46 Thermal Processing ......................................................................................... 47

PAGE 6

6 Flash pasteuri zation ................................................................................... 47 In pack pasteurization ................................................................................ 48 Aseptic filling .............................................................................................. 48 Chilled distribution ...................................................................................... 48 Dense Phase CO2 Processing (DPCD) ............................................................ 48 Mechanisms of microorganisms inactivation by DPCD ............................. 49 Factors affecting microbial inactivation ...................................................... 50 Solubility of CO2 ......................................................................................... 52 DPCD treatment systems ........................................................................... 53 DPCD food applications ............................................................................. 54 Coconut water ............................................................................................ 60 Sensory Evaluation ................................................................................................. 60 Difference from Control Test ............................................................................ 61 Flavor Analysis ....................................................................................................... 61 Solid Phase Micro Extraction (SPME) .............................................................. 62 Gas chromatography Olfactometry (GC O) ...................................................... 62 3 EFFECT OF COLD AND HOT WATER EXTRACTION ON THE PHYSICOCHEMICAL AND PHYTOCHEMICAL PROPERTIES OF HIBISCUS SABDARIFFA EXTRACTS ..................................................................................... 64 Introduction ............................................................................................................. 64 Materials and Methods ............................................................................................ 65 Extracts Preparation ......................................................................................... 65 pH, Total Solids, and Titratable Acidity ............................................................. 66 Anthocyanin Content, Total Phenolics and Antioxidant Capacity ..................... 67 Characterization of Major Polyphenolics .......................................................... 67 LC MS identification ................................................................................... 68 HPLC quantification ................................................................................... 69 Statistical Analysis ............................................................................................ 69 Results and Discussion ........................................................................................... 69 Effect of Extraction Conditions ......................................................................... 69 Parameters Correlations ................................................................................... 71 Polyphenolics Identification .............................................................................. 72 Polyphenolics Quantification ............................................................................ 75 Conc lusions ............................................................................................................ 76 4 AROMA PROFILES OF BEVERAGES OBTAINED FRO M FRESH AND DRIED HIBISCUS ............................................................................................................... 82 Introduction ............................................................................................................. 82 Materials and Methods ............................................................................................ 83 Sample Preparation .......................................................................................... 83 GC O Analysis .................................................................................................. 84 Identification Procedures .................................................................................. 86 Statis tical Analysis ............................................................................................ 86 Results and Discussion ........................................................................................... 86

PAGE 7

7 Hibiscus Volatiles Composition ........................................................................ 86 GC MS Identifications ...................................................................................... 87 GC O Aroma Profiles ....................................................................................... 89 Conclusion .............................................................................................................. 90 5 PROCESSING HIBSCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE: MICROBIAL A ND PHYTOHCEMICAL STA BILITY ................................ 95 Introduction ............................................................................................................. 95 Materials and Methods ............................................................................................ 96 Beverage Preparation ...................................................................................... 96 Solubility Experiment ........................................................................................ 97 Dense Phase CO2 Equipment .......................................................................... 97 DPCD P rocess Optimization ............................................................................ 98 Storage Experiment .......................................................................................... 99 Microbial Analysis ........................................................................................... 100 pH, Brix, a nd Titratable Acidity ...................................................................... 100 Anthcyanin Content, Total Phenolics and Antioxidant capacity ...................... 101 HPLC Quantification of Polyphenolics ............................................................ 101 Statistical Analysis .......................................................................................... 102 Results and Discussion ......................................................................................... 102 Solubility Measurements ................................................................................ 102 Microbial Inactivation Study ............................................................................ 103 Microbial Stability during Storage ................................................................... 105 Physicochemical Stability during Storage ....................................................... 105 Phytochemical Stability during Storage .......................................................... 106 Conclusions .......................................................................................................... 108 6 PROCESSING HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE: SENSORY ATR IBUTES AND AROMA COM POUNDS STABILITY .... 115 Introduction ........................................................................................................... 115 Materials and Methods .......................................................................................... 116 Beverage Preparation .................................................................................... 116 Dense Phase CO2 Equipment ........................................................................ 117 Physic ochemical Analysis .............................................................................. 118 Sensory Evaluation ........................................................................................ 118 Identification Procedures ................................................................................ 120 Color Analysis ................................................................................................ 120 Statistical Anal ysis .......................................................................................... 121 Results and Discussion ......................................................................................... 121 Physicochemical Analysis .............................................................................. 121 Sensory Evaluation ........................................................................................ 122 Aroma Compounds ........................................................................................ 123 Color Analysis ................................................................................................ 125 Conclusions .......................................................................................................... 126

PAGE 8

8 7 SUMMARY AND CONCLUSI ONS ........................................................................ 133 APPENDIX A EXTRACTION EXPERIMEN T STATISTICAL ANALYS IS .................................... 134 B STORAGE EXPERIMENT S TATISTICAL ANALYSIS .......................................... 137 C HIBISCUS SABDARIFFA PICTURES .................................................................. 140 LIST OF REFERENCES ............................................................................................. 141 BIOGRAPHICAL SKETCH .......................................................................................... 151

PAGE 9

9 LIST OF TABLES Table page 2 1 U.S. population projections ................................................................................. 24 2 2 Ingredients and nutritional facts of four hibiscus commercial products ............... 30 2 3 Nutritional composition of fresh hibiscus calyces ................................................ 31 2 4 Hibiscus extraction condi tions found in the literature .......................................... 35 2 5 Main classes of polyphenolic compounds .......................................................... 39 2 6 Classification of food flavonoids ......................................................................... 40 2 7 CO2 solubility of liquid foods measured at 40 C ................................................ 53 3 1 Measured pH, total solids (TS) (g of solids/100 mL of extract), titra acidity (TA) (g of malic acid/100 mL of extract), and color (L*, a*, b* values, color density (CD) and hue tint (HT)) for the extracts. ................................................. 77 3 2 Linear regression and correlation coefficients between measured parameters for cold and hot water extraction processes. ...................................................... 79 3 3 Identification of anthcocyanins present in hibiscus using their spectral characteristics with HPLC DAD and positive ions in LC MS and MS2. ............... 79 3 4 Identification of polyphenolics present in hibiscus using their spectral characteristics with HPLC DAD and negative ions in LC MS and MS2, and respective standards. ......................................................................................... 79 3 5 Polyphenolics content (mg/L) of hibiscus samples analyzed in this studyd. ........ 81 4 1 Extraction conditions and measured pH and Brix values for hibiscus samples included in this study. ........................................................................... 92 4 2 MS identification of hibiscus volatiles. Peak areas were normalized (100) to the largest peak in all four samples. ................................................................... 93 4 3 Hibiscus aroma active compounds. Peak heights were normalized (100) to the most intense peak in all four samples. .......................................................... 94 5 1 Response surface design used to test the effect of pressure and residence time on microbial reduction (log10) at 40 C and 8% CO2. ................................ 110 5 2 Physicochemical and phytochemical changes of unprocessed (CONTROL), dense phaseCO2 processed (DPCD), and thermally treated (HTST) hibiscus beverages during refrigerated storage at 4 C. ................................................. 112

PAGE 10

10 5 3 Polyphenolics conte nt (mg/L) of unprocessed (CONTROL), dense phaseCO2 processed (DPCD), and thermally treated (HTST) hibiscus beverages during refrigerated storage at 4 C. .................................................................. 114 6 1 Measured pH, Brix, and titra acidity (TA) (g of malic acid/100 mL of beverage) at weeks 0 and 5 of refrigerated storage (4 C). .............................. 127 6 2 Difference in flavor and overall likeability between fresh (reference and hidden reference), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 m in, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverages detected by untrained panelists (n = 75) at weeks 0 and 5 of refrigerated storage (4 C) ................................................................................ 127 6 3 MS identification of hibiscus beverage volatiles during storage. Peak areas were normalized (100) to the largest peak (1Octen 3 ol) in the CONTROL (C) week 0 sample. ........................................................................................... 129 A 1 SAS software output of statistical analysis for the anthocyanins concentration data (AC) perfumed in the hibiscus extraction experiment (Chapter 3). ........... 134 B 1 SAS software output of statistical analysis for the anthocyanins concentration data (AC) perfumed in the hibiscus storage experiment (Chapter 5). ............... 137

PAGE 11

11 LIST OF FIGURES Figure page 2 1 Beverage sectors and segments. (Source: Roethenbaugh 2005). ..................... 19 2 2 Total U.S. sales and forecast (f) of RTD noncarbonated beverages at inflation adjust ed prices, 200212. (Source: Mintel 2008). .................................. 21 2 3 U.S. sales and forecast (f) of RTD noncarbonated beverages at current prices, by segment, 2002 12. (Source: Mintel 2008). ......................................... 21 2 4 Market share according to FDMx (Food, drug and mass merchandisers excluding Wal Mart) sales of leading RTD non carbonated beverage companies, February, 200 8. (Source: Mintel 2008). ........................................... 23 2 5 Hibiscus pictures. A: hibiscus plant, B: hibiscus flower, C: hibiscus calyxes, and D: opened hibi scus calyx with velvety capsule in the center. ....................... 25 2 6 Compounds found in some hibiscus extracts: 1 = protocatechuic acid, 2 = chlor ogenic acid and 3 = hibiscus or hibiscic acid .............................................. 31 2 7 Chemical structure of anthocyanins present in hibiscus ..................................... 32 2 8 Structural and spectral characteristics of the major naturally occurring aglycons. (Source: Rodriguez Saona and Wrolstad 2005). ................................ 43 2 9 Predominant structural forms of anthcoaynins present at different pH levels. (Source: Giusti and Wrolstad 2005). ................................................................... 45 2 10 Phase diagram of carbon dioxide ....................................................................... 51 2 11 Schematic diagram of the continuous flow dense phase CO2 system ................ 54 3 1 Total anthocyanins content expressed as delphinidin3 glucoside (mg/L) for the extracts. The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25 C curve. Data represents the mean of n=9. Values with similar letters within the are not significantly different (Tukeys HSD, p > 0.05). ................................................................................................... 77 3 2 Total phenolics content expressed as gallic acid equivalents (mg/L) for the extracts. The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25 C curve. Data represents the mean of n=9. Values with similar letters within the are not significantly different (Tukeys HSD, p > 0.05). .................................................................................................................. 78 3 3 Antioxidant capacity ( mol of TE/mL) L) for the extracts. The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25

PAGE 12

12 C curve. Data represents the mean of n=9. Values with similar letters within the are not significantly different (Tukeys HSD, p > 0.05). ................................ 78 3 4 HPLC chromatograms of dried hibiscus (DHE) and fresh hibiscus (FHE) hot water extracts: (A) 520 nm, (B) 360 nm, (C) 320 nm, (D) 280 nm, and (E) 260 nm. For peak identification see Tables 33 and 3 4. ........................................... 80 4 1 Chemical composition of hibiscus he adspace volatiles. Total number of compounds for each class is put in parentheses. All four samples were normalized to the total peak area of DHE (dried hibiscus hot extraction). DCE = dried hibiscus cold extraction, FHE = fresh hibiscus hot extraction, FC E = fresh hibiscus cold extraction. ............................................................................. 92 5 1 Schematic diagram of the setup used for the hibiscus beverage thermal treatment (75 C for 15 s). ................................................................................ 109 5 2 CO2 solubility in water and a hibiscus beverage as a function of pressure measured at 40 C. Data represents the mean of n=3. Values with similar letters within the are not significantly different (Tukeys HSD, p > 0.05). ......... 109 5 3 Aerobic plate counts of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ...... 110 5 4 Yeast/mold counts of unprocessed (CONTROL), dense phaseCO2 processed ( DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ...... 111 5 5 Hue tint values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ....................... 111 5 6 Concentration of anthocyanins of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally tr eated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ................................................................................................................... 113 6 1 Chemical composition of hibiscus beverage h eadspace volatiles during storage. Total number of compounds for each class is put in parenthesis. All six samples were normalized to total peak area of the sample CW0 (CONTROL week 0). C = CONTROL, D = DPCD, H = HTST, W = week. ........ 128 6 2 L* values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). .................................. 130

PAGE 13

13 6 3 a* values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). .................................. 130 6 4 b* values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). .................................. 131 6 5 Hue angle values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ....................... 131 6 6 Chroma values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ....................... 1 32 6 7 E values of unprocessed (CONTROL), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2, 6.5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). ....................... 132 B 1 Pictures of dried hibiscus (A), dried hibiscus extraction process (B), hibiscus beverage (C), hibiscus beverage in the dense phase carbon dioxide (DPCD) feed tank (D), DPCD processing equipment (E), DPCD processed hibiscus beverage (F), hibiscus beverage samples for analysis (G), hibiscus beverage under refrigerated storage (H), and DPCD processed hibiscus beverage after 14 weeks of storage at 4 C (I). Photos by Milena Ramirez. ........................... 140

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROCESSING OF A HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE By Milena Mar ia Ramirez Rodrigues August 2010 Chair: Maurice R. Marshall Coc hair : Murat O. Balaban Major: Food Science and Human Nutrition C onsumer demand for natural beverages with health promoting properties that offer fresh like sensory att ributes and changes in U.S. demographics have created the opportunity for the development of new products that would target new m arket segments. Hibiscus sabdariffa (family Malvaceae) red calyces are rich in anthocyanins and other phenolic compounds. Fresh and dried hibiscus is used to prepare cold and hot beverages and their preparation includes an extraction step followed by a pasteurization method. Although thermal preservation of foods is effective in reducing microbial loads it can also lead to organol eptic and nutritional changes. Nonthermal processes like dense phase carbon dioxide (DPCD) are an alternative which may help preserve the color, flavor, and nutrients of food. Equivalent cold and hot water conditions were found for anthocyanins extraction of dried hibiscus in this research. Likewise, s imilar polyphenolic profiles and chemical composition of aroma compounds were observed between fresh and dried hibiscus

PAGE 15

15 S olubility of CO2 in a hibiscus beverage ( 5.06 g CO2Finding s in this research can help in the development and marketing of hibiscus beverage. /mL at 31.0 MPa) and optimal processing conditions to inactivat e yeasts and molds (Y&M) were 34.5 MPa and 6.5 min. DPCD was a viable technology for processing hibisc us beverage since it extended its shelf life for 14 weeks of refrigerated storage. Quality attributes were maintained during storage. Lower losses of anthocyanins were observed in the DPCD (9%) hibiscus beverage as c ompared to thermally treatment process (14%) and no major changes in total phenolics content and antioxidant c apacity occurred during storage. Changes in hibiscus ar oma volatiles during storage did not affect untrained panelists over all likeability of the product.

PAGE 16

16 CHAPTER 1 I NTRODUCTION Justification Anthocyanins are water soluble pigments responsible for the red to purple to blue colors in many fruit, vegetables, flowers, and cereal grains. The interest in anthocyanin pigments has intensified in recent years because of their possible health benefits. Thus in a ddition to their functional role as colorants, anthocyanin extracts may improve the nutritional quality of foods and beverages (Wrolstad 2004). C onsumer demands for natural beverages with health promoting properties that offer fresh like sensory att ributes and changes in U.S. demographics with Hispanics and Blacks as important growthdriving demographics (Mintel 2008) have created the opportunity for the development of new products that would target these market segments. Hibiscus sabdariffa (family Malvac eae) is a short day annual shrub that grows in many tropical and subtropical countries and is one of the highest volume specialty botanical products in international commerce (Plotto 1999). The red calyces are the part of the plant with commercial interest and are rich in organic acids, minerals, anthocyanins, and other phenolic compounds. Fresh and dried hibiscus calyc es are used to prepare cold and hot beverages which are commonly mixed with a sweetener and are characterized by an intense red color and acidic flavor which provides a sensation of freshness. The preparation of a hibiscus beverage includes an extraction step followed by a pasteurization method. Although thermal preservation of foods is effective in reducing microbial loads it can also lead to organoleptic and nutritional changes. Nonthermal processes are an

PAGE 17

17 alternative which may help preserve the color, flavor, and nutrients of food. Dense phase carbon dioxide (DPCD) is a cold pasteurization method that uses pressures below 90 MPa in combination with carbon dioxide (CO2Hypothesis ) to inactivate microorganisms. This nonthermal technology is mainly used in liquid foods and since the food is not exposed to the adverse effect of heat, its freshlike physical, nutritional, and sensory qualities are maintai ned. The combination of a cold extraction process with dense phase carb on dioxide (DPCD) processing will help prevent the degradation of anthocyanins present in a hibiscus beverage, and thus provide a product with enhanced quality and phytochemi cal activity Specific Objectives 1. T o compare the effects of cold and hot water extraction on the physicochemical and phytochemical properties of hibiscus extracts and to identify and quantify the anthocyanins and major polyphenolics present in extracts obt ained from fresh and dried hibiscus by equivalent cold and hot water extraction conditions. 2. T o determine the aroma profile differences between four extracts obtained from fresh and dried hibiscus extracted at two different conditions (22 C for 4 h and 98 C for 16 min) by GC MS and GC olfactometry. 3. To determine the solubility of CO2 in a hibiscus beverage, to optimize DPCD processing parameters based on microbial reduction, and to monitor during refrigerated storage the microbial, physicochemical, and phytochemcial changes of DPCD processed hibiscus beverage compared to thermally treated and control (untreated) beverages.

PAGE 18

18 4. To determine the effect of DPCD processing on the sensory att ributes and aroma compounds of hibiscus bever age when compared to a ther mally treated and a control (untreated) and to monitor the changes in these attributes during refrigerated storage.

PAGE 19

19 CHAPTER 2 LITERATURE REVIEW The Beverage Market The global beverage market is comprised of four sectors: 1) hot drinks, 2) milk drinks, 3) soft drinks, and 4) alcoholic drinks. Hot drinks include tea, coffee, and hot malt based products; milk drinks include white drinking milk and flavored milk products; soft drinks are divided into five main subcategories: (bottled water; carbonated soft dri nks; dilutables including powder and liquid concentrates; 100% fruit juice and nectars with 2599% juice content; still drinks including ready to drink (RTD) teas, sports drinks, and other non carbonated products with less than 25% fruit juice and alcohol ic drinks which include beer, wine, sprits, cider, sake and flavored alcoholic beverages ( Roethenbaugh 2005). A diagram of these sectors is presented in F igure 2 1. Figure 2 1. Beverage sectors and segments. ( Source: Roethenbaugh 2005) Hot drinks Tea Coffee Other hot drinks Soft drinks Bottled water Carbonated drinks Dilutables Fruit juice/ nectars Still drinks Milk drinks White milk Flavored milk Alcoholic drinks Beer Wine Spirits Other alcoholic drinks

PAGE 20

20 Soft drinks are normally defined as sweetened water based bever ages, usually having a balanced acidity. Flavor, color, fruit juice or fruit pulp are often added in their formulation. The main ingredient in soft drinks is water and thus their primary function is hydration. There are two basic types of soft drinks: ready to drink (RTD) products and concentrates or diluteto taste products. The RTD sector is divided into products that are carbonated and those that are noncarbonated (Ashurst 2005). The market of ready t o drink (RTD) non carbonated beverages can be divided in four segments: 1) bottled water, 2) sports/energy drinks, 3) fruit juice/juice drinks, and 4) RTD teas and coffees. Market Performance and Competitive C ontext of U.S. Ready to Drink NonC arbonated Beverages Sales for ready to drink noncarbonated beverages reached $38.6 billion in 2007, exhibiting a 35% growth, measured in current prices, during the period 20022007. The market is projected to grow 33% in current prices from 200812, or the equival ent of 16% when cons idering the imp act of inflation (F igure 22 ). E nhanced bottled waters, energy drinks, and RTD teas are the categories that have driven this grow th (Mintel 2008). Fruit juice and juice drinks, bottled water, sports and energy drinks, and RTD tea and coffee accounted for 37.4, 28.6, 26.0, and 8.1% or the RTD non carbonated beverage market in 2007. While the sales for fruit juice and juice drinks are forecast t o decline in the period 200720 12, the other three categories will have increasi ng sales over this period (F igure 23).

PAGE 21

21 Figure 22 Total U.S. sales and forecast (f) of RTD non ca rbonated beverages at inflation adjusted prices, 200212 ( Source : Mintel 2008) Figure 2 3 U.S. sales and forecast (f) of RTD non carbonated beverages at current prices, by segment, 200212. ( Source: Mintel 2008) 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 2002 2003 2004 2005 2006 2007 2008f 2009f 2010f 2011f 2012f $ billion 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 2002 2003 2004 2005 2006 2007 2008f 2009f 2010f 2011f 2012f $ million Fruit juice and juice drinks Bottled water Sports and energy drinks RTD tea and coffee

PAGE 22

22 PepsiCo dominates brand sales in this category, its RTD non carbonated beverage portfolio includes the top juice (Tropicana), bottled water (Aquafina), sports dr ink (Gatorade), and tea (Lipton) brands. Because PepsiCo is such a strong player in every segment, its total market share as of February, 2008 was 25.3% according to FDMx (Food, drug and mass merc h a ndisers excluding Wal Mart) sales (F igure 2 4) However, s ales growth has been slow, allowing other beverage suppliers to gain a market share (Mintel 2008) Coca Cola is wellpositioned to gain a market share with its Glacau, Fuze, and Powerade brands. Its sales increased nearly 14% during 200708, adding 1.1% to Coca Colas share in t he market. The company experienced solid sales growth in every segment (Mintel 2008) While major beverage companies dominate the RTD noncarbonated beverage category mid level entrants are defined by their strong representation i n a single segment, such as Ocean Spray in the juice segment, or Nestl in water sales. Smaller (under the category Other in F igure 2 4) participants in the market have found their niche, and are defined mostly by a single brand targeting a specific demo graphic, like thos e in the energy/sports drink segment like Red Bull and Rockstar (Mintel 2008) Many small, highgrowth companies present alliance opportunities for big players. These include Hansens Natural (natural soda and fruit juice, and energy dri nks), Jumex (Hispanic targeted beverages), Jones Soda (unique soda flavors), Tampico (Hispanic targeted beverages) and AriZona Tea Co. (RTD tea) (Mintel 2008) Private label is becoming a more important player in the RTD noncarbonated beverage category W hile its presence is strongest in commodity segments such as

PAGE 23

23 juice and water it is growing fastest in trendy beverages (sports/energy and RTD teas and coffees) (Mintel 2008) Figure 2 4. Market share according to FDMx (Food, drug and mass merchandisers excluding Wal Mart) sales of leading RTD non carbonated beverage companies, February, 2008. ( Source: Mintel 2008) Consumption and Demographic Trends The two biggest market t r ends are health/wellness and convenience. Consumers are demanding more from their beverages. Drinks should not only be thirst quenchers but also provide added benefits. Health and wellness increasingly plays an influential role in consumer choices on the beverage aisle. Consumers are seeking products that add value to their diet; howev er, not only must products deliver nutrition conveniently, but the packaging must carry a convenient format (Mintel 2008) Hispanics and blacks are important growthdriving demographics, not only because these groups are projected to exhibit an aboveaverage population growth, but also because they display an aboveaverage incidence of juice consumption. Additionally, both groups are the key consumer in highgrowth s ports and energy drinks markets (Mintel 2008). 25.3% 16.3% 9.4% 3.2% 3.2% 3.2% 2.9% 12.4% 24.1% PepsiCo Coca Cola Company Nestl Kraft Cadbury Schweppes Ocean Spray Campbell Private label Other

PAGE 24

24 According of the U.S. Census Bureau projection for 2050, nonHispanic whites will no longer make up the majority of the population. Today nonHispanic whites make up about 68% of the population. This is expected to fall to 46% in 2050 as a result of a much older white population relative to minorities. Hispanic population is projected to change from 15% to 30% of the total U.S. population while African American and Asian Americas will reach 15 and 9% of the population by 2050 (T able 2 1). The U.S. has nearly 305 million people today. The population is projected to reach 400 million by 2039 and 439 million in 2050 (U.S. Census Bureau 2009). Table 2 1. U.S. population projections 2008 2050 Non Hispanic whites 68% 46% Hispanic 15 % 30% African Americans 12% 15% Asian American 5% 9% ( Source: U. S. Census Bureau 2009) Hibiscus ( Hibiscus sabdariffa) Characteristics and Economic Importance There are more than 300 species of hibiscus around the world. One of them is Hibiscus sabdariffa, Linn, which is a member of the Malvaceae family. The origin of H. sabdariffa is not fully known but it is believed t o be native to India and Malaysia and to have been carried at an early date to Africa. It is widely grown in tropical and subtropical regi ons including Africa, South East Asia and some countries of America. Seeds are said to have been brought to the New World by African slaves. It is know by different synonyms and vernacular names such as roselle in the U.S and England, loiselle in Fra nce, jamaica or flor de j amaica in Mexico and Spain, karkade in Sudan and

PAGE 25

25 Arabia, sorrel in the Caribbean and byssap in Senegal (Morton 1987; Stephens 2003) In this study the word hibiscus will be used to refer to Hibiscus sabdariffa Hibiscu s (Figure 2 5) is a short day annual shrub and can grow to a height of 1 3 m, depending on variety. The green leaves are about 8 12 cm long and the stems, branches, leaf veins and petioles are reddish purple. Flowers are up to 12.5 cm wide, they are yellow with a rose or maroon eye, and are made up of five petals. After the flowers fall apart, the calyx which is a red cup like structure consisting of 5 large sepals with a collar (epicalyx) of 8 to 12 slim pointed bracts around the base, begins to enlarge, b ecomes fleshy, crisp but juicy (3.2 5.7 cm long) and fully encloses the velvety capsule, (1.252 cm long) which is green when immature, 5valved, with each valve containing 3 to 4 kidney shaped light brow n seeds, (35 mm long) The capsule turns brown and splits open when mature and dry (Morton 1987; De Castro and others 2004). Figure 2 5. Hibiscus pictur es. A: hibiscus plant, B: hibiscus flow er, C: hibiscus calyxes, and D: opened hibiscus calyx with velvety capsule in the center. Usually, hibiscus is propagated by seeds or cuttings and grows on sandy soil. The ideal planting time in North America is from April to May, blooming occurs in September and October, and calyces are ready for harvest in November and D C B A

PAGE 26

26 December. The "fruits" should be gathered before any woody tissue develops in the calyx. They should be tender, crisp, and plump (Stephens 2003). Hibiscus has several uses. Its calyces which is the part of the plant of commercial interest, are used either fresh or dehydrated in the processing of preserves, jellies, jams and sauces for their rich pectin content, to prepare hot and cold beverages which are commonly mixed with a sweetener and are characterized by an intense red color and acidic flavor which provides a sensation of freshness, in the production of wine, and color and flavor extracts. They are also a source of soluble and insoluble fiber. The leaves are used extensively for animal fodder and fiber and are also used in salads and the seeds are a source of protein and lipids and constitute a byproduct in h ibiscus production (Al Wandawi and others 1984; El Adawy and Khalil 1994; Mounigan and Badrie 2007; SyagoAyerdi and others 2007; Hainida and others 20008). Traditionally fresh hibiscus calyces are harves ted by hand and are either frozen, dried in the sun or artificially preserved and are either sold into the herbal tea and beverage industry or local and regional markets. Five kilograms of fresh calyces dehydrate to 0.45 kg of dried hibiscus Industrial scale operation s that use hibiscus include production of vacuum concentrated extract, spray drying of extracts beverages, natural food colorant and natural food flavor (Al Kahtani and Hassan 1990). Hibiscus is one of the highest volume specialty botanical products in the international comm erce and d emand has steadily increased over the past decades. A pproximately 15,000 metric tons of dried hibiscus enter international trade each year Many countries produce hibiscus but the quality markedly differs. China and Thailand are the largest producers and control much of the world supply. Mexico, Egypt,

PAGE 27

27 Senegal, Tanzania, Mali Sudan, and Jamaica are also important suppliers but production is mostly used domestically (Plotto 1999). Germany and the U.S. are the main countries importing hibiscus The biggest German buyer is Martin Bauer, one of the oldest and largest companies in the herb industry They use hibiscus in numerous products including herbal teas, herbal medicines, syrups and food coloring. Mai n import ers in the U.S are Celestial Seasonings and Lipton, both tea companies. Hibiscus is also used in ready to serve beverages made by Knudson, Whole Foods and other food and beverage manufacturers (Plotto 1999). Commercial Hibiscus Products Hibiscus striking red color, refreshing properties and associated health benefits has attracted the interest of several entrepreneurs to start a business around the idea of manufacturing hibiscus based beverages. There are four ready to drink commercial products that use hibiscus as the main ingredient. Following is a brief description of these products as well as their marketing approach. Hibiscus Lemon Bissap Produced by the company Adina for Life Inc. (located in California) and established i n 2005, this product is mark eted as a New Age beverage, with all natural ingredients, refreshing, and with goodfor you appeal. The label is bright colorful and f olksy. The company sources its hibiscus using fair trade arrangements from independent farmers in Senegal. Adinas presi dent is also from Senegal and considers t his product to help rescue traditional beverage mixes from his country ( Anon 2006)

PAGE 28

28 Caita Aguas Frescas (jamaica (hibiscus) flavor) Produced by the company Eat Inc. (located in North Carolina) and estab lished in 2003, this prod uct is marketed as 100% natural that provides health benefits and targets mainly Hispanic consumers It has the intent of bringing a traditional Mexican beverage known as Agua de jamaica to the Hispanic population in the U.S. ( Anon 2006) Squish Hibiscus Press Produced by the company Squish Hibiscus Press located in New Zealand, this product is marketed as a beverage with unique exotic floral fruity flavor that has beneficial properties. This is a new product in the New Zeala nd market that consumers are not familiar with The market segments to which this product is targeted are women between 18 and 35 years old and kids ( Anon 2006) Simply Hibi Produced by Ibis Organica, a UK based company; this company sources its raw materi al from Uganda and has established a program to help improve living conditions in that country. The product contains 87% hibiscus extract and 13% grape juice concentrate and it is marketed as 100% natural and high in antioxidants. T he ingre dients and nutri tional facts for these four products are presented in T able 22 Composition and Associated Health B enefits Hibiscus calyces are rich in organic acids including succinic, oxalic, tartaric, and malic acids (Wong and others 2002) hibiscus acid which is a lactone form of (2S,3R) (+) 2 hydroxycitric acid and its 6methyl ester (Hansawasdi and others 2000) a scrobic acid, carotene, and lycopene (Wong and others 2002) It is also high in phenolic compounds such as protocatechuic acid (3,4 dihydroxybenzoic acid) (Tseng and others

PAGE 29

29 1998; Liu and others 2002; Lin and others 2003) and chlorogenic acid (Segura Carretero and others 2008) (F igure 26 ) minerals (aluminum, chromium, copper and iron) (Wrobel and others 2000), sugars (g lucose, fructose, sucrose and x ylose) (Pouget and others 1990a; Wong and others 2002) water soluble polysaccharides (Muller and Franz 1992) and anthocyanins (Du and Francis 1973; Wong and others 2002). There can be composition variations depending on variety, soil, climate and growing conditions, and post harvest handling and processing. The nutritional composition of f resh hibiscus calyces is presented in Table 2 3 F or many years hibiscus has been used in different countries as a medicinal herb for therapeutic purposes. According to different ethnobotanical studies some traditional medicines use the aqueous extract of the plant as a diuretic, for treating gastrointestinal disorders and hy percholesterolemia, and as a diaphoretic and antihypertensive drug (Herrera Arellano and others 2004) Many biological activities have been reported in aqueous extracts of Hibiscus sabdariffa. Animal experiments have shown that the consumption of this extract has antihypertensive (Odigie and others 2003) antiatherosclerotic (Chen and others 2003) lipid profile reduction (Carvajal Zarrabal and others 2005) and antioxidant properties (Su boh and others 2004; Hirunpanich and others 2006; Liu and others 2006). Studies with human patients have also shown that the regular consumption of hibiscus extract has an antihypertensive effect (Haji Faraji and Haji Tarkhani 1999; HerreraArellano and ot hers 2004) and reduces serum cholesterol in men and women (Lin and others 2007) Several compounds isolated from hibiscus extracts also possess pharmacological activities.

PAGE 30

30 Table 22 Ingredients and nutritional facts of four hibiscus commercial products Hibiscus Lemon Bissap Caita Aguas Frescas (jamaica flavor) Squish Hibiscus Press Simply Hibi* Ingredients Water Hibiscus Sugar Fructose Organic evaporated cane juice Concentrated pineapple juice Lemon juice Camu camu Organic rosehips Acerola Ascorbic acid Nutritional Facts Serving size 236 mL 236 mL 375 mL Servings per container 1.75 2 1 Calories 80 135 124 Sodium 15 mg 11 mg Total carbohydrates 28 g 34 g 28.9 g Sugars 19 g 34 g 28.9 g Vitamin C 4% Magnesium 2% Potassium 2% Calcium 4% 4% Camu camu 50 mg Rosehips (organic) 50 mg Acerola 167 mg Lemon bioflavonoids 83 mg Nutritional data for this product is not available. Pictures were taken from the actual products by Milena Ramirez. Nutritional data was retrieved from the bottle labels or from New nutrition business 12(1):1922).

PAGE 31

31 Table 2 3 Nutritio nal composition of fresh hibiscus calyces g/100g mg/100g mg/100g Water 86.58 Calcium 215 Vitamin C 12 Protein 0.96 Phosphorus 37 Riboflavin 0.028 Lipids 0.64 Iron 1.48 Niacin 0.31 Carbohydrates 11.13 Sodium 6 Thiamin 0.011 Ash 0.51 Potassium 208 Vitamin A 287 UI Magnesium 51 Energy 49 kcal (Source: USDA 2009) OH OH O H O O H O H O O O H OH O H CO OH O O H CO OH CO OH OH 1 2 3 Figure 26 Compounds found in some hibiscus extracts: 1 = protocatec huic acid, 2 = chlorogenic acid and 3 = hibiscus or hibiscic acid Several compounds isolated from hibiscus extracts also possess pharmacological activ ities. Protocatechuic acid has antiatherosclerosis (Lee and others 2002) antitumor promotion (Tseng and others 1998; Lin and others 2003; OlveraGarca and others 2008) antioxidant (Lin and others 2003), and anti inflammatory (Liu and others 2002) activities An thocyanins isolated from hibiscus exhibited antioxidant (Wang and others 2000) and anticancer (Chang and others 2005; Hou and others 2005) activities while hibiscus acid and its 6methyl ester have shown to be amylase inhibitors (Hansawasdi and others 2000). Hibiscus Anthocyanins Recently there has been a market interest in hibiscus anthocyanins due to their beneficial health effects and high antioxidant properties which have been extensively evaluated (Tee and others 2002; Tsai and others 2002; Tsai and Huang 2004; Prenesti

PAGE 32

32 and others 2007: S yago Ayerdi and others 2007) and as a potential source of natural food colorant. The two major anthocyanins present in hibiscus are: delphinidin3 sambubioside also known as delphinidin3 xylosylglucoside or hibiscin and cyanidin3 sambubioside also known as cyaniding 3 xylosylglucoside or gossypicyanin. They account for approximately 70 and 30% of total anthocyanins, respectively. Other anthocyanins like delphinidin3 glucoside, delphinidin3 (feruloyl)rhamnoside, cyanidin3 glucoside, cyaniding 3 O rutin oside and cyaniding 3,5 diglucoside have been found in minor concentrations in some varieties (F igure 27 ) (Du and Francis 1973; Pouget and others 1990b; Tsai and others 2002; Wong and others 2002; Mourtzinos and others 2008; SeguraCarretero and ot hers 2008). O+ O H C B A 3' 4' 5' 3 5 Anthocyanin 3 4 5 3 5 Cyanidin 3 sambubiioside OH OH H 2 O xylosyl D glucose OH Cyanidin 3 glucoside OH OH H Glucosyl OH Cyanidin 3,5 diglucoside OH OH H 3,5 diglucosyl OH Cyanidin 3 rutinoside OH OH H O rutinosyl OH Delphinidin 3 sambubioside OH OH OH 2 O xylosyl D glucose OH Delphinidin 3 glucoside OH OH OH Glucosyl OH Delphinidin 3 (feruloyl)rhamnoside OH OH OH (feruloyl)rhamnoside OH Figure 27 Chemical structure of anthoc yanins present in hibiscus

PAGE 33

33 The stability of hibiscus anthocyanins has been studied in model systems testing the effect of different chemical compounds (ascorbic acid, BHA, propyl gallate, disodium EDTA, sodium sulfite) (Pouget and others 1990a) temperature ( Gradinar u and others 2003; Dominguez Lpez and others 2008; Cisse and others 2009) sugar type and concentration (Tsai and others 2004) copigmentation and polymerization (Tsai and Huang 2004) as well as their stability in v arious foods including jellies, beverages, gelatin desserts, and freeze dried products C olor stability during storage has also been tested (Esselen and Sammy 1975; Clydesdale and others 1979) Heat, light, and humidity were all found to be detrimental to anthocyanin stability. Some studies have shown that thermal degradation of hibiscus anthocyanins follow first order reaction kinetics (Gradinar u and others 2003; Domin guez Lpez and others 2008; Mourtzinos and others 2008). Thermal stability of hibiscus anthocyanins in the temperature range of 6090 C in the presence or absence of cyclodextrin was studied. The temperaturedependent degradation was modeled by the Arrhenius equati on and the activation energy for the degradation of hibiscus anthoc y a nins was ~54 kJ/mol. The presence of cyclodextrin improved thermal stability of nutraceutical antioxidants present in hibiscus extract s both in solution and solid state (Mourtzinos and others 2008). Another study showed that the activation energy for the degradation of hibiscus anthcoyanins was 66.22 kJ/mol (Duangmal and others 2008) while a third study found that copigmentation with chlorogenic acid didnt improve their stability in solution and activation energies for their degradation were between 55.68 and 63.22 kJ/mol (Gradinar u and others 2003).

PAGE 34

34 Hibiscus Extraction Process Some researchers have focused on hibiscus water extracts while others have employed organic solvents to extract pos sible bioactive compounds. Indeed the different extraction techniques (extraction time and temperature) make comparison among studies difficult. Moreover different varieties have bee n used. T able 2 4 summarizes some of the conditions used for hibiscus extr action found in the literature Some research has been done regarding the optimization of hibiscus extraction process. One study tested three different hibiscus to water r atios (1:52, 1:67, 1:62 w/v) at three extraction times (20, 25, 30 min) in a hot extr action at 100 C. They found that optimum conditions based on color and taste were 1:62 w/v for 30 min (Bolade and others 2009) Wong and others (2003) found that optimum condition for hibiscus extraction was 3.5 h at 60 C based on anthocyanins content and color. Hibiscus Flavor Hibiscus flavor is a combination of sweet and tart, similar to cranberry Few studies have been done related to hibiscus flavor. Gonz alez Palomares and others (2009) identified 20 volatile compounds in a hibiscus extract using SPME and GC MS, including terpenoids, esters, hydrocarbons, and aldehydes. They also found 14 compounds in reconstituted spray dried extracts from which only 10 w ere present in the original extract and the other four were products of degradation. Thermally generated volatiles from untreated, frozen, hot air dried at 50 C, and hot air dried at 75 C hibiscus by steam distillation w ere analyzed by GC and GC MS (Chen a nd others 1998). They characterized more than 37 compounds including fatty acid derivatives, sugar derivatives, phenol derivatives, and terpenes.

PAGE 35

35 Table 24. Hibiscus extraction conditions found in the literature Country of origin F a D b Solvent Hibiscus: solvent ratio Extraction time Extraction temperature A c Reference MeOH with 0.125% citric acid 1:1.65 w/v 48 h Pouget and others 1990 a Sudan Water 1:10 w/w 40 min 60 C Al Kahtani and Hassan 1990; Hassan and Hobani 1998 Mexico Water 1:8 w/v 15 min 60 C Beristain and others 1994 Taiwan Water 1:30 w/v 10 min Boiling Duh and Yen 1997 Malasya MeOH 1:10 w/v 24 h 25 C Tee and others 2002 Taiwan Water 1:100 w/v 3 min Boiling Tsai and others 2002 Malaysia Water 1:5 w/v 1 h Boiling Wong and others 2002 Egypt 3% Formic acid in MeOH 24 h 4 C Gradinaru and others 2003 Malaysia Water 1:40 w/w 30 300 min 30 90 C Wong and others 2003 Mexico Water 1:8 w/v 128 min Ambient Andrade and Flores 2004 Mexico Water 1:50 w/v 10 min Boiling Herrera Arellano and others 2004 Nigeria Water 1:30 w/v 30 min Boiling Oboh and Elusiyan 2004 Taiwan Acidified EtOH (1.5 mol/L HCl) 1:50 w/v Tsai and Huang 2004 Mexico Water 1:10 w/v 5 min Boiling Dominguez Lopez and others 200 8 Singapore Water 1:50 w/v 1 h Ambient d Wong and others 2006 Egypt Water 1:50 w/v 5 930 min Ambient Prenesti and others 2007 Egypt Water 1:50 w/v 3 min 100 C Prenesti and others 2007 Egypt 12% v/v EtOH in water 1:50 w/v 30 min Ambient Prenesti and others 2007 Mexico Water 1:20 w/v 5 min Boiling Syago Ayerdi and others 2007 Mexico Water 1:50 w/v 10 min Boiling Olvera Garca and others 2008 a = fresh hibiscus, b = dried hibiscus c = agitation, d = occasional, e = sonication.

PAGE 36

36 Table 24. Continued Country of origin F a D b Solvent Hibiscus: solvent ratio Extraction time Extraction temperature A c Reference Mexico Water 1:50 w/v Overnight Ambient Olvera Garca and others 2008 Acidified MeOH (MeOH/HCl (99:1 v/v)) 1:10 w/v 4 h Ambient Segura Carretero and others 2008 Acidified MeOH (MeOH/HCl (99:1 v/v)) 1:10 w/v 30 min Ambient e Segura Carretero and others 2008 Acetic acid (15% v/v) 1:40 w/v 48 h Ambient Segura Carretero and others 2008 Senegal Water/MeOH/HCl, 50:50:2 1:125 w/v 30 min e Juliani and others 2009 Senegal Water 1:62.5 w/v 15 min e Juliani and others 2009 Mexico 30% v/v EtOH in water 1:12.5 w/v 168 h Ambient d Gonzalez Palomares and others 2009 Nigeria Water 1:52 1:62 w/v 20 30 min 100 C Bolade and others 2009 Taiwan Water 1:40 w/v 2 h 95 C Lin and others 2007 Guatemala and Senegal Water 1:10 w/v 10 h 25 C Cisse and others 2009 a = fresh hibiscus, b = dried hibiscus c = agitation, d = occasional, e = sonication.

PAGE 37

37 They concluded that hibiscus aroma was a combination of terpene derivatives with fragrance notes and sugar derivatives with a caramel like odor. Phenolic compounds Phenolic compounds are products of the secondary metabolism of plants. B iogenetically they originate from two main synthetic pathways: the shikimate pathway and the acetate pathway. Chemically, phenolics can be defined as substances that have an aromatic ring bearing one or more hydroxyl groups, including their functional derivatives (Bravo 1998). Many properties of plant products are associated with the presence, type, and content of their phenolic compounds. Of significance to producers and consumers of foods are the astringency of foods, the beneficial health effects of certain phenolics or their potential antinutritional properties when present in large quantities (Shahidi and Naczk 2004). Classif ication Natural polyphenols can range from simple molecules, such as phenolic acids, to highly polymerized compounds, such as tannins. They occur mainly in conjugated form s, with one or more sugar residues linked to hydroxyl groups, although direct linkage s of the sugar unit to an aromatic carbon atom also exist. The associated sugars can be present as monosaccharides, disaccharides, or even oligosaccharides. The most common sugar residue is glucose, but galactose, rhamnose, xylose, and arabinose can also be found, as well as glucuronic and galacturonic acids among others. They can also be associated with carboxylic and organic acids, amines, lipids, and other phenols (Bravo 1998). Polyphenols can be divided into at least 10 different

PAGE 38

38 classes depending on the ir basic chemical structure (T able 25) Flavonoids, which are the most important single group, can be furth er subdivided into 13 classes (T able 26). Phenolic Compounds Attributes Positive attributes of phenolic compounds include: contribution to flavor and astringency, natural pigments, antimicrobial and antiviral properties, anti inflammatory activity antitumor and anticancer activity, antimutagenicity, antioxidant potential, and reduction of coronary heart disease risk (Lule and Xia 2005). There are al so some negative attributes of phenolic compounds that include: off flavor and taste contribution, discoloration due to enzymatic and nonenzymatic reactions, and antinutritional activity because of interactions with proteins, carbohydrates, minerals, and v itamins (Lule and Xia 2005). Contribution to flavor Phenolic compounds may contribute to the aroma and taste of numerous food products of animal and plant origin. The presence of chlorogenic acid can be related to the bitterness of wine, cider, and beer while hydroxycinnamates and their derivatives are responsible for the sour bitter taste of cranberries. Phenolic substances also contribute to the flavor of vanilla pod and vanilla extracts. Vanillin, p hydroxybenzaldehyde, and p hydroxybenzyl methyl ester have been found to be the most abundant volatiles but simple phenolics such as p cresol, eugenol, p vinylguaiacol, and p vinylphenol as well as aromatic acids such as vanillic and salicylic acids are also present. Ripe bananas contain volatile phenolics suc h as eugenol, methyleugenol, elimicin, and vanillin. Strawberry volatiles contain esters of some phenolic acids such as ethyl salicylic, methyl cinnamic, and ethyl benzoic acids.

PAGE 39

39 Table 2 5 Main classes of polyphenolic compounds Class Basic Skeleton Basic Structure Simple phenols C OH 6 Benzoquinones C O O 6 Phenolic acids C6C CO OH 1 Acetophenones C 6 C COC H3 2 Phenylacetic acids C 6 C C H2CO OH 2 Hydroxycinnamic acids C 6 C C H = C H CO OH 3 Phenylpropenes C 6 C C H2C H = C H2 3 Coumarins, isocoumarins C6C O O 3 O O Chromones C6C O O 3 Naftoquinones C 6 C O O 4 Xanthones C6C1C O O 6 Stilbenes C6C2C O O 6 Anthraquinones C 6 C 2 C 6 Flavonoids C 6 C 3 C see Table 2 6 6 Lignans, neolignans (C 6 C 3 ) 2 Lignins (C 6 C 3 ) n (Source: (Bravo, 1998)

PAGE 40

40 Table 2 6 Classification of food flavonoids Flavonoid Basic Structure Chalcones O Dihydrochalcones O Aurones O O C H O O M e C O Flavones O O Flavonols O O OH Dihydroflavonol O O OH Flavanones O O Flavanol O OH Flavandiol or leucoanthocyanidin O OH OH Anthocyanidin O+ OH

PAGE 41

41 Table 26. Continued Flavonoid Basic Structure Isoflavonoids O O O O Bioflavonoids O O O O Proanthocyanidins or condensed tannins O O O (Source: (Bravo 1998). Thymol also is a major contributor to the flavor of essential oils from tangerine and mandarin. Phenolic substances may be responsible for the flavor of a number of spices and herbs (Shahidi and Naczk 2004; Lule and Xia 2005) Antioxidant potential One of the principal roles that have been proposed as part of the actions of phenolics is that of an antioxidant. Their antioxidant action can arise from a combination of several chemical events, which i nclude enzyme inhibition, metal chelation, hydrogen donation from suitable groups and oxidation to a nonpropagating radical. The health implications of an antioxidant depend on how well it is absorbed by the body and how it is metabolized, in addition to partition effects (Parr and B olwell 2000).

PAGE 42

42 Anticarcinogenic action The possible mechanism of action o f anticarcinogens can be classified into two groups, blocking and suppressing, depending on the point of action. Some compounds can both block and suppress. The main action of blocking agents is to stimulate the carcinogendetoxifying enzymes and to inhibit enzymes which have the potential to activate precarcinogens into carcinogens (Parr and Bolwell 2000). Anthocyanins Anthocyanins are water soluble pigments responsible for the red to purple to blue colors in many fruit, vegetables flowers, and cereal grains In general, its concen tration in most fruits and v egetables goes from 0.1 to 1% dw. The total content of anthocyanins varies among fruits and vegetables, t heir different cultivars and is also affected by genetic makeup, light, temperature and agronomic factors (Shahidi and Naczk 2004; Wrolstad 2004). Since color is one of the most impor tant quality attributes in food, a nthocyaninrich plant extracts might have a potential use as a natural alternative to food colorants. Anth oc y a nins based colorants are manufactured for food use from horticultural crops grown for that specific purpose as well as from processing wastes The interest in anth oc y a nins pigments has intensified in recent years because of their possible health benefits. Thus in addition to their functional role as colorants, antho c y a nins extracts may improve the nutritional quality of foods and beverages (Wrolstad 2004). Classification Chemically, ant h ocyanins are flavonoids and are based on a C15 skeleton. The anthcyanidins (aglycones) are the basic structure of anthoc yanins. They consist of an aromatic ring A bonded to an heterocyclic ring C that contains oxygen which is also

PAGE 43

43 bonded by a carboncarbon bond to a third aromatic ring B (C6 C3 C6) (i.e. anthocyanins are substituted glycosides of salts of phenyl 2 benzopyrilium ( anthocyanidins) ) (F igure 2 8 ) (DelgadoVargas and others 2000; Gradinaru and others 2003; CastaedaOvando and others 2009). The differences in color and stability between anthocyanins are related to the number of hydroxyl and methoxyl groups, the nature, position, and number of sugars attached to the molecule, and the nature and number of aliphatic or aromatic acids attached to sugars in the molecule (Kong and others 2003) R O+ O H OH OH OH R 7 6 5 8 3 2 1 4 2' 3' 4' 5' 1' 6' 1 2 C B A Aglycon Substitution pattern R1 R 2 maxVisible spectra (nm) Pelargonidin H H 494 (orange) Cyanidin OH H 506 (orange red) Delphinidin OH OH 508 (blue red) Peonidin OCH H 3 506 (orange red) Petunidin OCH OH 3 508 (blue red) Malvidin OCH OCH 3 510 (blue red) 3 Figure 28 Structura l and spectral characteristics of the major naturally occurring aglycons. (Source: Rodriguez Saona and Wrolstad 2005). An increase in the number of hydroxyl groups tends to deepen the color to a more bluish shade while an increase in methoxyl groups increase redness. Glucose, galactose, rhamnose, and arabinose are the sugars most commonly found in

PAGE 44

44 anth oc yanins, usually as 3gly cosides or 3,5diglycosides. Rutinosides (2 O L rhamnosyl D glucosides) sophorosides (6 O D glucosyl D glucosides) and sambubiosides (2 O D xylosyl D glucosides) also occur as well as some 3,7diglycosides, and 3triosides The most common acylating agents include cinnamic acids (caffeic, p coumaric, ferulic, and synaptic) and aliphatic acids (acetic, malic, malonic, oxalic, and succinic) (Clifford 2000; DelgadoVargas and others 2000). T here are 17 known naturally occurring anthocyanidins but only s ix are common in higher plants: pelarg onidin (Pg), peonidin (Pn), cyaniding (Cy), malvidin (Mv) petunidin (Pt), and delphinidin (Dp). The glycosides of the three non methylated anthocyanidins (Cy, Dp and Pg) are the most widespread in nature, being present in 80% of pigmented leaves, 69% of fruits and 50% of flowers. The distribution of the six mos t common anthocyanidins in the edible parts of plants is cyanidin (50%), pelargonidin (12%), peonidin (12%), delphinidin (12%), petunidin (7%), and malvidin (7%) (Kong and others 2003). Stability Anthocyanin pigments can be destroyed easily during process ing and storage. A num ber of factors influence their stability including pH, temperature, humidity, light oxygen, enzymes, as well as the presence of ascorbic acid, sugars, sulfur dioxide or sulfite salts, me tal ions and copigments The study of anthcoyanins characteristics can help develop products and processing conditions that will yield highquality products (Jackman and others 1987; Clifford 2000; DelgadoVargas and others 2000; Gradinaru and others 2003; Mazza and others 2004). The eff ective pH range for most anthoc y a nins colorants is limited to acidic foods because of the color changes and instability that occur above pH 4 ( (Wrolstad, 2004) At

PAGE 45

45 a given pH, equilibrium exists between four different anthocyanin structures: a quin oidal (anhydro) base, a flavylium cation, and the colorless carbinol pseudo base and chalcone (F igure 2 9 ) Copigments are substances that contribute to anthocyanins coloration by protecting the anthocyanin molecule; this mechanism is unique to the anthcyanin family. U sually copigments have no color by themselves but when added to an anthocyanin solution they gr e atly enhance its color (Mazza and Brouillard 1990). R O O H O g l y O g l y R 1 2 OH R O+ O H OH O g l y O g l y R 1 2 H + q uinoidal base: blue flavylium cation (oxonium form): orange to purple pH = 7 pH = 1 R OH R 1 2 O g l y OH O H O g l y R O O H OH O g l y O g l y R 1 2OH O H + +H O 2 chalcone: colorless carbinol pseudobase (hemiketal form): colorless pH = 4.5 pH = 4.5 Figure 2 9 Predominant structural forms of anthcoaynins present at different pH levels. ( Source: Giusti and Wrolstad 2005)

PAGE 46

46 Anthcoyanins react with flavones, alkaloids, amino acids, benzoic acids, coumarin, cinnamic acids, and a wide variety of other fl avil y u m compounds. This w eak hydrophobically driven interaction (van der Waals interactions) is known as intermolecular copigmentation. Intramolecular copigmentation occurs with the acylation of the molecule and is more effective than the intermolecular one. The basic role of copi gments is to protect the colored flavylium cation from the nucleophilic attack of the water molecule (DelgadoVargas and others 2000). The association between anthcoyanins and copigments lead s to an absorbance maxHealth B enefits toward higher wavelengths (bathoc h romic effect) (Gradinaru and others 2003). In plants anthocyani ns serve as attractants for pollination and seed dispersal give protection against the harmful effects of UV irradiation, and provide anti viral and anti microbial activities (Wrolstad 2004). A nth oc y a nins could exhibit multiple biological effects, e.g. reduced risk of coronary heart disease and strok e, anticarcinogen activity, antioxidant/antiradical activity, anti inflammatory action, inhibition of blood platelet aggregation and antimicrobial activity, treatment of diabetic retinopathy and prevention of cholesterol induced atherosclerosis (Wang and o thers 1997; Espin and others 2000; Wrolstad 2004). Beverage Processing Beverage processing typically involves an extraction step (juices and teas extraction), followed by blending were they can be mixed with other ingredients like water, sweeteners, acidulants, flavorings, colors, and preservatives among others.

PAGE 47

47 Beverages then go into processing, filling, and packaging. The purpose of the processing and packaging steps is to produce a product that is wholesome and safe for the consumer (Ashurst 2005). Thermal P rocessing Traditionally beverages are thermally processed. There are five main processes that use heat as a way to assure microbial safety in juices and soft drinks : 1. Flash pasteurization 2. Hot filling 3. In pack pasteurization 4. Aseptic filling 5. Chilled distribution Process selection depends on the level of microbial contamination of the raw materials and packaging, whether the product composition will favor the growth of microorganisms the ability of the product to resist heat and the desired shelf li fe (Ashurst 2005). Flash pasteurization Normally the juice is passed through a balance tank or feed tank before being f ed to the pasteurizer. The liquid is generally heated by hot water in a plate or tubular (spiral) heat exchanger to the desired pasteuri zation temperature and held at that temperature for a specified time in a holding tube before being cooled to the filling temperature (usually ambient) using chilled water. Flash pasteurizers usually have a regeneration section. The pasteurized product is then sent to a filling machine. Hot filling In hot filling, the product is heated (in a heat exchanger), sent to the filler hot and filled into containers. The containers are closed and are held at or above the desired temperature for a s pecified time pr ior before being cooled. This is usually done in a

PAGE 48

48 tunnel with water sprays. In this system not only the product but also the container is heat treated. In pack pasteurization This type of process is generally the most severe and microbiologically most se cure form of heat treatment. The filled closed pack is put into a tunnel pasteurizer were the treatment is given by means of water sprays at various controlled temperatures. The pasteurizer is dived into zones. First there is a heating zone where the temperature of the container and the product is raised, next there is the pasteurizing zone where the product is held to the pasteurizing temperature for a specified time and finally there is a cooling zone where the product is cooled below 30 C. Aseptic filling Aseptic filling is a special case of flash pasteurization that often uses a higher temperature profile. For successful aseptic filling, clean containers, clean product, clean headspace, and clean closures should be brought together in an environment t hat prevents recontamination. This operation normally takes place i n a closed space over pressure with sterile air. Chilled distribution Flash pasteurized product is filled cold into clean bottles on clear fillers and then is stored in refrigerated warehouses and is sold to costumers from chill cabinets (Ashurst 2005). Dense P hase CO2Dense phase carbon dioxide (DPCD) is a cold pasteurization method that uses CO P rocessing (DPCD) 2 under pressures below 50 MPa. This non thermal technology is mainly used in

PAGE 49

49 liquid foods Since the food is not exposed to t he adverse effect of heat, its fresh like physical, nutritional, and sensory qualities are maintained (Damar and Balaban 2006). Mechanisms of microorganisms inactivation by DPCD The exact means of microbial inactivation are not clear but sev eral mechanisms may be involved: 1. pH lowering effect CO2 can lower the pH of the aqueous part s of the food by dissolving and forming carbonic acid (H2CO3), which further dissociates into bicarbonate (HCO3 -) carbonate (CO3 2) and hydrogen ( H+CO) ions lowering extracellular pH by the following equilibrium : 2 + H2O 2CO3 + + HCO3 + + CO3 2 CO(1) 22. Inhibitory effect of molecular CO penetrates the microbial cell membrane and lowers its internal pH by exceeding the cells buffering capacity. This change in internal pH may inactivate microorganisms by inhibiting essential metabolic systems such as enzymes. 2 Bacterial enzymes may be inhibited by CO and bicarbonate ion 2 by formation of a bicarbonate complex, excess CO2, pH lowering by dissolved CO2, sorption /interaction of CO2 into the enzyme molecules, and precipitation of intracellular Ca+2 and Mg+23. Physical disruption of cells carbonates. The disruption of physical cells was the fir st mechanisms proposed f or microorganisms inactivation and suggests that microbial cells bursting is due to the rapid pressure release and the expansion of CO24. Modification of cell membrane and extraction of cellular components within the cell. This mechanism is based on the lipoand hydrophilicity and solvent characteristics of CO2. Extraction of intracellular substances and their transfer out of the

PAGE 50

50 cell during pressure release may lead to microbial inactivation (Damar and Balaban 2006). Factors affecting microbial inactivation Several factors may influence microbial inactivation including: 1. Water activity and water content DPCD is more effective as aw increases A higher water content of the treated product increase CO22. Pressure solubility and enhances micr obial inactivation. Since CO23. Temperature (T) solubility increases with increasing pressure this can help in microorganisms inactivation. Although solubility of CO2 decreases with increasing temperature, higher T can increase the diffusivity of CO2 and the fluidity of the cell membrane which can facilitate the penetration of CO2 into cells. T can also affect the change of CO2 from subcritical to supercritical phase (Tc = 31.1 C ) (Figure 2 10). Under supercritical conditions the penetration power of CO2 is higher and at the near critical region there is a rapid change in solubility and density of CO24. Initial pH (Damar and Balaban 2006). Low pH facilitates penetration of carbonic acid through th e cell membrane leading to more inactivation. 5. Cell growth phase and age Young cells are more sensitive and are easier to inactivate than mature cells.

PAGE 51

51 Figure 210. Phase diagram of carbon dioxide 6. Type of microorganism Different bacteria have different susceptibilities to DPCD. It has been suggested that the nature of the cell wall could be important in the difference in sensitivity between G( +) and G( ) bacteria. Since G( ) bacteria have thin cell walls they are expect ed to be more sensitive and their cells walls could be more easily ruptured as compared to G(+) bacteria. 7. Type of treatment system The system used for DPCD treatment can affect the microbial inactivation rate. Systems that allow better contact of CO2 with the food are more effective in microbial reduction because of the more rapid saturation of the solution with CO2 Usually batch systems require longer treatment times for microbial inactivation as compared to continuous systems. Inactivation rate can be i ncreased in batch systems by using agitation (Damar and Balaban 2006).

PAGE 52

52 Solubility of CO In a continuous flow DPCD system several variables are controlled during processing: pressure, temperature, residence time, and %CO2 2. The amount of CO2 used should as sure a complete saturation of the liquid but since its solubility at processing conditions is not known this can lead to the use of excess CO2CO elevating production costs. 2 solubility in liquid foods can be affected by pressure, temperature, and food composition. Pressure has a direct effect on CO2 solubility meaning that as pressure increases, CO2 solubility increases while as temperature increases, solubility of CO2 decreases. Other substances present in the food (composition) either increase or decrease the solubility of CO2Recent studies have focused on the measurement of CO (Calix and others 2008) 2 solubility in liquid model systems and fruit juices These experiments were done using an experimental apparatus designed and built at the University of Florida (Calix and others 2008). This system is designed to saturate a known amount of liquid by bubbling CO2 from the bottom of a vessel under controlled pressure and temperature and afterwards the CO2Table 27 presents solubility of CO gas is expanded and measured at ambient pressure. 2Solubility of CO for water, orange, apple, and grapefruit juice measured at different pressures in these studies. 2 in fruit juices is lower than that of pure water because of the presence of solutes such as sugars and acids which lowers the amount of CO2 that can dissolve (Calix and others 2008; Ferrentino and others 2009).

PAGE 53

53 Table 27 CO2 solubility of liquid foods measured at 40 C Liquid Pressure (MPa) Solubility Reference Water 7.58 31.0 4.71 6.32 Calix and others 2008 Orange juice 7.58 15.86 4.10 4.98 Calix and others 2008 Apple juice 7.58 15.86 3.95 5.01 Calix and others 2008 Grapefruit juice 7.58 31.0 3.97 4.70 Ferrentino and others 2009 g of CO2DPCD treatment systems /100 g of liquid. Batch, semi continuous, and continuous systems have been developed for DPCD applications. In a batch system, CO2 and the food to be treated are stationary in a container during treatment. A semi continuous system allows a continuous flow of C O2 through the chamber while a continuous system allows flow of both CO2A typical batch system has a CO and the liquid food through the system. 2 gas cylinder, a pressure regulator, a pressure vessel, a water bath or heater, and a CO2 release valve. The sample is placed into the pressure vessel and temperature is set to the desired value. CO2 is then introduced into the vessel until t he sample i s saturated at the desired pressure and temperature. The sample is left in the vessel for a period of time and then the CO2 outlet valve is opened to release the gas. Some systems contain an agitator to decrease the time to saturate the sample with CO2A continuous flow DPCD system was developed in 1999 by Praxair (Chicago, III., U.S.A.) (F igure 211) (Da mar and Balaban 2006). CO2 and the product are pumped through the system and are mixed before entering the high pressure pump, which increases the pressure to the processing levels.

PAGE 54

54 Figure 211. Schematic diagram of the continuous flow dense phase CO2Product temperature is controlled in holding coils. Residence time is adjusted by setting the flow rate of the product. At the end of the process, CO system 2DPCD food applications is released by means of an expansion valve (Damar and Balaban 2006). DPCD has been applied mainly to liquid f oods. This technology has been tested in several products at the University of Florida Fo od Science and Human Nutrition D epartment using the conti nuous flow system presented in F igure 211, and include s: orange, mandarin, and grapefruit juice, beer, grape juice, and coconut water among others. Orange juice Several studies with orange juice (OJ) show ed that DPCD tr eatment can improve some physical, nutritional and quality attributes such as cloud formation and stability,

PAGE 55

55 color, and ascorbic acid retention (Damar and Balaban 2006). Kincal and others (2005) tested the capacity of the D PCD system to reduce both natural and inoculated microbial loads of pulpfree Valencia OJ at different pressures (38, 72, and 107 MPa), temperatures (25 and 34.5 C), residence times (3 10 min), and CO2/juice ratio s (0.1 1.0). A sto rage study was conducted at 1.7 C with juice processed at 107 MPa, CO2/juice ratio of 1.0 and residence time of 10 minutes R esidence time had the greater influence on microbial reduction, followed by pressure. The CO2/juice ratio and temperature showed not to be the driving forces on microbial load reduction in this system. They proved t hat the system was able to achieve a 5 log reduction of the natural flora in spoiled juice (38, 72, and 107 MPa at 25 and 34.5 C, CO2/juice ratio of 1.0 and residence time of 10 min), and 5 log decrease of pathogenic Escherichia coli O157:H7 (107 MPa and residence time of 10 min), Salmonella typhimurium ( 3 8, 72, and 107 MPa and residence time of 10 min), and Listeria monocytogenes ( 38, 72, and 107 MPa and residence time of 10 min). Durin g the refrigerated storage study they observed an increase in the bacterial number possibly because of an injury/repair mechanism of some of the microorganisms or due to post contamination. A study was perfor med by Kincal and others (2006) to treat pulp free Valencia OJ at pressures of 38, 72, and 107 MPa, and CO2/juice (w/w) ratios from 0.40 to 1.18 with a constant residence time of 10 min. The highest PE inactivation (46.3%) was obtained when the pressure was 107 MPa and no heat was applied. PE activity decreased with storage time. Cloud increased between 446 and 846% after treatments and remained 4 times higher than the control during storage. They found no significant changes in pH and Brix of treated samples.

PAGE 56

56 TA increased slightly after treatment and remained constant throughout st orage. Small but insignificant increase in L* and a* values occurred after treatment. Juice color did not change drastically during storage. Sensory evaluations of DPCD treated and untreated OJ were not significantly different after 2 w eeks of refrigerated storage at 1. 7oC (Balaban and others 2008) Mandarin juice Lim and others ( 2006) processed mandarin juic e. The process variables were temperature (2545C ), pressure (13.841.4 MPa), residence time (59 minutes) and % CO2 (2 12). They fou nd that temperature and % CO2 had a significant effect in log reduction of total aerobic plate count while pressure and residence time were not significant. The maximum log reduction (3.47) was observed at 35 C, 41.4 MPa, 9 min and 7 % CO2. PE inactivation ranged from 6.1 to 50.7% and maximum inactivation was achieved at 45C, 41.4 MPa, 7 min and 7% CO2. Cloud was not only retained but enhanced. The highest cloud increase was 38.4% at 45C 27.6 MPa, 7 min, and 2% CO2. L ightness and yellowness increased and redness decreased after treatment. pH and B rix didnt change after treatment while t itratable acidity of treated samples was higher than the untreated juice (Balaban and others 2008). Grapefruit juice Red grapefruit juice was processed using DPCD at pressures of 13.8, 24.1, and 34.5 MPa and residence time of 5, 7, and 9 min at a constant temperature of 40 C and CO2 level of 5.7% to evaluate the effect of treatments on yeasts and molds and total aerobic bacteria inactivation. A five log reduction for yeasts and molds and total aerobic

PAGE 57

57 bacteria was achieved at 34.5 MPa and 7 min of treatment (Ferrentino and others 2009) Ferrentino and others (2009) also conducted a storage study with DPCD processed red gr apefruit juice for 6 wk at 4 C. No growth of total aerobic bacteria and yeasts and molds was observed in the DPCD treated juice. Cloud in the juice increased 91% while PE inactivation was partial (69.17%). No significant differences in Brix, pH, and TA w ere detected between treated and untreated samples while the treated juice had a higher lightness and redness and lower yellowness. Total phenolics content was not affected by treatment and storage and slight differences were detected for ascorbic acid con tent and antioxidant capacity. Beer Dagan and Balaban (2006) studied the effect of DPCD on beer quality. They predicted a maximum log reduc tion in yeast population of 7.4 logs at processing conditions of 26.5 MPa, 21 C, 9.6% CO2Grape juice and residence time of 4.77 min. DPCD pasteurization reduced haze from 146 nephelometric turbidity units (NTU) to 95 NTU. Aroma and flavor of beer at these same conditions was not significantly different when compared to a fresh beer sample in a differenc e from control test. Foam capacity and stability of beer were minimally affected by the process ; however these changes were unnoticed by consumers. Several studies were conducted with muscadine grape juice testing the effect of DPCD on microbia l reduction, physicochemical, phytochemical and quality changes after treatment and during storage. Different processing conditions of pressure (1.240.2

PAGE 58

58 MPa), CO2 levels (0 15.7%), and constant residence time (6.25 min) and temperature 30 C) were evaluat ed by Del Pozo Insfran and others (2006a) Results showed that processing pressure was a significant factor affecting microbial inactivation but that CO2Subsequent storage stability for 10 wks at 4 C with t w o treatments that achieved >5 log reduction (34.5 MPa at 8 and 16% CO2) were evaluated and compared to a heat pasteurized juice (75 C, 15 s). Results showed that thermal pasteurization decreased anthocyanins (16%), soluble phenolics (26%), and antioxidant capacity (10%) whereas no changes were observed for both DPCD treated juices. DPCD juices also retained higher anthocyanins (335 mg/L), polyphenolics (473 mg/L), and antioxidant capacity (10.9 mol of Trolox equivalents/mL) than thermally pasteurized juices at the end of storage (D el Pozo Insfran and others 2006a). content was the processing parameter that had the major influence in microbial log reduction. Insignificant differences in sensory attributes (color, flavor, aroma, and overall likeability) were observ ed between unprocessed and DPCD juices, while significant differences were observed between unprocessed and heat pasteurized juices. Panelists preferred DPCD over heat pasteurized juices throughout the first 6 weeks of storage but afterwards the growth of yeast and mold adversely affect ed juice aroma. Comparable microbial counts were observed between DPCD and thermally pasteurized juices during the first 5 weeks of storage (D el Pozo Insfran and others 2006a). Another study evaluated the phytochemical stability and organoleptic attributes of an ascorbic acid fortified muscadine grape jui ce as affected by DPCD processing and

PAGE 59

59 addition of thyme polyphenolic cofactors ( Thymus vulgaris ; 1:100 anthocyaninto cofactor molar ratio) (Del Pozo Insfran and others 2006b). DPCD processing insignificantly altered initial juice phytochemical and antioxidant content, whereas thermal pasteurization reduced anthocyanins (263 mg/L), ascorbic acid (42 mg/L), soluble phenolics (266 mg/L), and antioxidant capacity (6 mol of Trolox equivalents/mL). Si milar trends were observed during storage, and data showed that increasing the CO2 level from 8 to 16% during DPCD processing contributed to the reduction of juice phytochemical and antioxidant degradation. Copigmentation helped retain higher anthocyanins, soluble phenolics, and antioxidant capacity during storage without affecting initial juice aroma and flavor characteristics (Del Pozo Insfran and others 2006b). A third study by Del Pozo Insfran and others (2007) assessed t he e ffect of DPCD processing on polyphenol oxidase (PPO) activity, polyphenolic and antioxidant changes in muscadine grape juice under di fferent processing pressures (27.6, 38.3, and 48.3 MPa), CO2 levels (0, 7.5, and 15%) and constant residence time (6.25 min) and temperature (30 C). Pressure alone was responsible for a 40% decrease in PPO activity that resulted in 16 40% polyphenolic and antioxidant losses. Increasing CO2 from 0 to 7.5% was responsible for an additional 35% decrease in enzyme activity and a 2 fold greater polyphenolic retention. However, insigni fi cant changes in PPO activity or polyphenolic retention were observed when CO2 was increased to 15%. Subsequently two DPCD condi tions (48.3MPa at 0 and 15% CO2) were evaluated for polyphenolic and antioxidant changes during storage (4 C, 4 wks). J uices with residual PPO activity following processing resulted in greater polyphenolic (810-

PAGE 60

60 fold) and antioxidant capacity (4fold) degradation compared to control juices with no PPO activity. Coconut water The effects of DPCD on microbial, physical, chemical and sensorial quality of a coconut water beverage were evaluated by Damar and others (2009) Different processing conditions of pressure (13.8, 24.1, and 34.5 MPa), temperature (20, 30, and 40 C), %CO2 level ( 7, 10, 13 gCO2/1 00g of beverage) and constant residence time of 6 min were tested. Pressure was not significant in microbial reduction whereas temperature and %CO2 levels were significant. In the same study DPCDtreated (34.5 MPa, 25 C, 13% CO2, 6 min), heat pasteurized (74 C, 15 s) and untreated coconut bever ages were evaluated during 9 w ks of refrigerated storage (4 C). Total aerobic bacteria of DPCD and heat treated samples decreas ed while that of untreated samples increased to >105 CFU/ mL after 9 w ks. DP CD increased titratable acidity but did not change pH (4.20) and Brix (6.0). Likeability of DPCDtreated coconut water was similar to untreated. Heat treated samples were less liked at the beginning of storage. Off flavor and taste differencesfrom contr ol scores of heated samples were higher than DPCD during the first two weeks (Damar and others 2009). Sensory Evaluation The attributes of a food item are typically perceived in the following order: appearance (color, size and shape, surface texture), odor/aroma/fragrance, consistency and texture, and flavor (aromatics, chemical feelings, taste). However in the process of perception most or all of the attributes overlap (Meilgaard and others 2007)

PAGE 61

61 Sensory tests provide useful information about the human perception of product changes due to ingredients, processing, packaging, or shelf life. Sensory evaluation includes a set of test methods with guidelines and established techniques for product presentation and well defined responses, statistical methods, and guidelines for interpretation of results. There are three primary sensory tests: discrimination tests (focus on the existence of overal l differences among products), descriptive analysis (specification of attributes), and affective or hedonic testing (measuring consumers likes and dislikes) (Lawless and Heymann 1998). Differencefrom Control Test A differencefrom control test is used to determine whether a difference exists between one or more samples and a control and to estimate the size of any such difference. One sample is designated the control, reference, or standard and all other samples are evaluated with respect of how different each is from the control (Meilgaard and others 2007). Flavor A nalysis Flavor perception depends of the combined responses or our senses and the cognitive processing of these inputs (Reineccius 2006). Taste is the combined sensation that derives from specialized taste receptor cells located in the mouth. It is primarily limited to the tongue and is divided into the sensations of sweet, sour, salty, bitter, and umami while olfaction is the sensory component that results from the interaction of volatile food components with olfactory receptor s in the nasal cavity. The stimulus for the aroma of a food can be orthonasal (the odor stimulus enters the olfactory region directly from the nose as we sniff the food)

PAGE 62

62 or retronasal (the stimulus enters from the oral cavity as we eat a food) (Reineccius 2006). Most of the techniques used in aroma isolation take advantage of either solubility or volatility of the aroma compounds (Reineccius 2006). Solid Phase Micro Extraction (SPME) SPME is a relatively new technique for the isolation of food aromas. An inert fiber is coated with an adsorbent. The adsorbent coated fiber is placed in the headspace of a sample, or the sample itself if liquid, and allowed to adsorb volatiles. The loaded fiber is then thermally desorbed into a GC carrier gas flow, and the released volatiles are analyzed. Since SPME is an equilibrium technique, the volatile profile obtained is strongly dependant of the sample composition and careful control of all sampling parameter s is required (R eineccius 2006). The effectiveness of SPME techniques depends on many parameters such as : type of fiber, sample volume, temperature and extraction time, and desorption of analytes from the fiber (Waldemar and others 2004) Gas chromatography Olfactometry (GC O) GC O is a technique only applied to aroma studies In olfactometric techniques, the nose is used as a GC detector The GC system may be set up in such a way that the column effluent is split so that a portion of the effluent goes to a sniffing port and the reminder goes to a GC detector (flame ionization (FID) or an MS det ector) The GC O produces what is called an aromagram, which is a listing of the odor character of each peak in a GC run. Mass spectroscopy is generally used in the flavor area to either determine the identity of an unknown or to act as a mass selective GC detector. The

PAGE 63

63 GC MS as an identification tool is unique because of its high sensitivity (10100pg) (Reineccius 2006). With consumer demands for natural beverages with health promoting properties that offer fresh like sensory attributes, Hibiscus sabdariff a may offer an additional market in this arena. Previous research has focused on a hot beverage or extraction process while new technologies focus on minimal to nonthermal processing. Dense phase carbon dioxide processing may offer an alternative to the traditional hibiscus processing and provide consumers a product with freshlike quality and healt h benefits. This research focus ed on three main areas: 1) finding alternatives to the water extraction conditions that do not involve heat but suitable for nonthermal processing, 2) comparing DPCD to heat pasteurization by evaluating the physicochemical, phytochemical, and sensory properties during processing and storage, and 3) evaluating aroma and phytochemical profiles of water hibiscus extracts obtained fro m fresh and dried hibiscus.

PAGE 64

64 CHAPTER 3 EFFECT OF COLD AND HOT WATER EXTRACTION ON THE PH YSICOCHEMICAL AND PHYTOCHEMICAL PROPERTIES OF HIBISCUS SABDARIFFA EXTRACTS Introduction Hibiscus sabdariffa L (family Malvaceae) is a tropical annual shrub. China, Thailand, Mexico, Egypt, Senegal and Tanzania are among the main producing countries. In Mexico, this plant is known as flor de jamaica or simply jamaica. The red calyces are the part of t he plan t with commercial interest and are rich in organic acids, minerals, anthocyanins, and other phenolic compounds. Hibiscus extracts contain two major anthocyanins, delphinidin3 sambubioside and cyanidin3 sambubioside. Their spectral characteristics (Degenhardt and others 2000), MS fragmentation patterns (Giusti and others 1999), and potential antioxidant (Wang and others 2000) and anticancer (Chang and others 2005; Hou and others 2005) activities have been previously studied. Similarly, other polyphenolic c ompounds in cluding protocatechuic acid ( Lee and others 2002; Ol vera Garcia and others 2008), hibiscus acid and its 6methyl ester (Hansawasdi and others 2000) have also been found to be present in hibiscus extracts and have been associated with pharmacolog ical activities. Differences in variety and extraction conditions (type of solvent, concentration, time and temperature) potentially affect the polyphenolic profile of the extracts and makes comparison between studies difficult. Traditionall y fresh hibis cus is either frozen or dr ied in the sun for preservation and used in the production of natural color, flavor extracts and/or beverages. Preparation of a hibiscus beverage includes an extraction step followed by a pasteurization method. The use of nonther mal technologies such as dense phase carbon dioxide, pulsed UV light, high hydrostatic pressure, and pulsed electric fields as

PAGE 65

65 a preservation method does not justify an extraction step that involves high temperature, and an alternative extraction at a lower temperature should be considered. The objectives of this study were: (1) to compare the effects of cold (25C) and hot (90C) water extraction on the physicochemical and phytochemical properties of hibiscus extracts and (2) to identify and quantify the anthocyanins and major polyphenolics present in extracts obtained from fresh and dried hibiscus by equivalent cold and hot water extraction conditions Materials and Methods Chemicals and Standards Commercial standards of chlorogenic acid, gallic acid, pro tocatechuic acid, and quercetin were purchased fr om Sigma Aldrich (St. Louis, Mo., U.S.A. ). Delph inidin 3 glucoside and cyanidin 3 glucoside were purchased from Polyphenols Laboratories AS (Sandnes, Norway). AAPH (2,2 azobis(2methylpropionamidine) dihydrochloride), fluorescein (free acid), Trolox (6hydroxy 2,5,7,8 tetramethylchroman2 carboxylic acid) a nd FolinCiocalteus reagent were purchased fr om Sigma Aldrich (St. Louis, Mo., U.S.A.). Extracts Preparation Fresh and sun dried Hibiscus sabdariffa (cv. Criollo) were obtained from Puebla, Mexico Hibiscus samples were stored in glass jars, flushed with nitrogen and kept frozen at 22 C until used. For the first part of the experiment, dried hibiscus was mixed with distilled water at a r atio of 1:40 (w/v) and maintained at 25 C (cold extraction, CE) or 90 C (hot extraction HE) for four different times (30, 60, 120, and 240 min for CE and 2, 4, 8, and 16 min for HE) Eight treatments (TRTs) were tested. For the second part of the experiment, four extracts were prepared using equivalent

PAGE 66

66 cold and hot water extraction conditions. Fresh (F) and dried (D) hibiscus were mixed with distilled water at a ratio of 1:4 and 1:40 (w/v) respectively and extracted at both 25C for 240 min (CE) and 90C for 16 min (HE). For CE, temperature was controlled using a Constant Temperature Circulator Bath, Model 900 ( Fisher Scientific; Pittsburg, Pa., U.S.A.) and stirring was applied using a stirrer plate M odel PC 353 (Corning, Lowell, Mass., U.S.A.) at speed #4. For HE, a Microprocessor Controlled Water Bath, Series 280 (Prec ision Scientific, Winchester, Va., U.S.A.) was used to control temperature. All t he obtained extracts were filtered under vacuum (Whatman filter paper #4) and their physicochemical and phytochemical properties were measured using the methods described below. pH, Total Solids, and Titratable Acidity pH was measured using a pH meter E A920 (Orion Research, Boston, Mass., U.S.A.) and total solids (TS) were deter mined by difference in weight after drying the samples at 105 C for 24 h in an oven (Precision Scientific, Winchester Va., U.S.A. ). A Brinkmann Instrument (Brinkmann Instruments Co., Westbury N.Y ., U.S.A.) consisting of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was used to measure titratable acidity (TA). Samples of 10 mL were used and TA was determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid (g/100 mL). Color Color Density and Hue Tint Color was measured using a ColorQuest XE colorimeter (HunterLab, Reston, Va., U.S.A.). Samples ( 40 mL) were placed in a 2 0 mm cell and L*, a*, and b* parameters were recorded in total transmittance mode illuminant D65, 10 observer angle. Color density a nd hue tint were determined by measuring the absorbance (A) at

PAGE 67

67 420, 520, and 700 nm for samples (200 L) using a spectrophotometer (SpectraMax 190, Molecular Devices, Sunnyvale Calif ., U.S.A.) and calculated as: Color density = (A420 nm A700 nm) + (A520 nm A700 nmHue tint = (A ) (1) 420 nm A700 nm)/(A520 nm A700 nmas described by Giusti and Wrolstad (2005). ) (2) Anthocyanin Content, Total Phenolics and Antioxidant Capacity Anthocy anin content was determined by the pH differential method ( A510 nm and A700 nm at pH 1.0 and 4.5, dilution factor (DF) of 4) and expressed in mg/L of delphinidin3 glucoside (MW = 465.2, = 23700) as described by Giusti and Wrolstad (2005). Total phenolics were measured using the FolinCiocalteu assay (A765 nmAntioxidant capacity was evaluated using the oxygen radical absorbance capacity (ORAC) assay and results were expressed as Trolox equivalents (TE) per milliliter ( mol of TE/mL) as described by Huang and others (2002) using a SpectraMax Gemini XPS microplate sprectrofluorometer ( Molecular Devices, Sunnyvale, Calif ., U.S.A.). Data was acquired and analyzed using SoftMax Pro 5.2 software ( Molecular Devices, Sunnyvale, Calif ., U.S.A. ). DF of 4) and quantified as gallic acid equivalents (mg/L) (Waterhouse 2005). Absorbance measurements for anthocyanin content and total phenolics were made using a SpectraMax 190 spectrophotometer (Molecular Devices, Sunnyvale Calif ., U.S.A.). Characterization of Major Polyphenolics E quivalent cold and hot water extraction conditi ons (25 C for 240 min (CE) and 90 C for 16 min (HE)) were selected based on the first part of this study. Four hibiscus extracts were prepared: dried hibiscus cold water extract (DCE), dried hibiscus hot

PAGE 68

68 water extract (DHE), fresh hibiscus cold water extract (FCE), and fresh hibiscus hot water extract (FHE). LC MS and HP L C analysis were performed in order to identify the major polyphenolic compounds including anthocyanins present in these extracts. LC MS i dentification Chromatographic analyses were performed on an Agilent 1200 ser ies HPLC (Agilent, Palo Alto, Calif ., U.S.A. ) equipped with an autosampler/inj ector and diode array detector. A Dionex C18 5 m 120A column ( 250 x 4.6 mm ) was used for compound sep aration (Dionex, Sunnyvale, Calif ., U.S.A.) Mobile phases consisted of water (phase A) and 60% methanol in water (phase B), both adjusted to pH 2.4 with formic acid. A gradient solvent program ran phase B from 0% to 60% in 20 min; 60% to 100% in 20 min; 1 00% for 7 min; 100% to 0% in 3 min and final conditions were held for 2 min. The flow rate was 0.8 mL/min, and detection was done at 260, 280, 320, 360 and 520 nm. Electrospray ionization mass spectrometry (ESI MS) was performed with a HCT series ion trap mass spectrometer (Bruker Daltonics, Billerica, Mass ., U.S.A.). Column effluent was monitored in positive and negative ion mode of the MS in an alternative ma nner during the same run. Other experimental conditions on the mass spectrometer were as follows: nebulizer, 45 psi; dry g as (nitrogen) 11.0 L /min; dry temperature, 350 C; ion trap, scan from m/z 90 to 1000; smart parameter setting (SPS), compound stability, 50%; trap drive level, 60%. The mass spectrometer was operated in Auto MS2 mode. MS2 was used to capture and fragment the most abundant ion in full scan mass spectra Polyphenolics were identified by comparison of UV/vis (190660 nm) spectral interpretation, retention time, comparison to standards, and MS fragmentation patterns.

PAGE 69

69 HPLC q u antification Anthocaynins and polyphenolics were quantified using a Dionex HPLC system equipped with an autosampler/injector and diode array (PDA 100) detector (Dionex, Sunnyvale, Calif ., U.S.A.). Compounds were separated on a 250 x 4.6 mm Dionex C18 5 m 120A column (Dionex, Sunnyvale, Calif ., U.S.A.). Mobile phases consisted of water (phase A) and 60% methanol in water (phase B), both adjusted to pH 2.4 with o phosphoric acid. A gradient solvent program ran phase B from 0% to 60% in 20 min; 60% to 100% in 20 min; 100% for 7 min; 100% to 0% in 3 min and final conditions were held for 2 min. The flow rate was 0.8 mL/min, and detection was done at 260, 280, 320, 360 and 520 nm. Statistical Analysis Each treatment condition was repeated in triplicate. Analysi s of variance (ANOVA) and mean separation using Tukeys test the differences between extraction times, temperatures, and treatments using SAS 9.0 Statistical software (SAS Ins titute Inc., Cary, N C U.S.A.). Results and Discussion Effect of Extraction Conditions The physiochemical and phytochemcial parameters measured for the eight hibiscus treatments analyzed are presented in Table 31 and Figures 31, 3 2, and 33. C olor density (CD), antho cy a nins content (AC), total p henolics (TP) and antioxidant capacity (AOX) increased with increasing time for both extraction temperatures (25 and 90 C) while L* values decreased with time in both cases. There were no significant differences between the first two (30 and 60 min, and 2 and 4 min)

PAGE 70

70 and last two (120 and 240 min, and 8 and 16 min) extraction times at both temperatures for hue tint (HT) values ; although the latter times were higher For cold water extraction (25 C) time had a significant effect ( p < 0.0001 ) in all the parameter s measured but pH. Total solids (TS) and b* increased from 30 to 60 mi n and afterwards remained constant (measurements at times 60, 120, and 240 min were not significantly different). Titratable acidity (TA) increased from 30 to 120 min and remained constant at 240 min (measurements at 120 and 240 min were not significantly different) while a* values increased from 30 to 60 min, remained constant from 60 to 120 min and decreased at 240 min. For hot water extraction (90 C) time had a significant effect ( p < 0.0335) in all the measured parameter s. pH increased from 2 to for 4 min and remained constant until 16 min. TS and TA increased until 8 min and remained constant until 16 min. a* values were constant for times 2 and 4 min and dec reased at 8 and 16 min while b*There was a significant ef fect ( p < 0.0001 ) of treatment conditions (temperature + time) in all the measured parameters but pH. T reatments CE1 and HE1 were equivalent in TS and TA (Table 31), t reatments CE2 and HE2 were equivalent in TS, TA, b*, CD, and AC (Figure 31) Treatments CE3 and H E3 were equivalent in TS, TA and AC while treatments CE4 and HE4 were equivalent in TS, TA, CD, and AC values were constant at 2, 4, and 8 min, and decreased at 16 min. L* values were significantly lower (darker color) in hot water extracts as compared to cold water ones while a* values were slightly higher in the cold water extracts. Hue tint (HT) is a measurement of color degradation in anthocyanin co ntaining products. From Table 31 it can be observed that the extracts obtained with cold water have l ower values than the ones obtained with hot water. This indicates that

PAGE 71

71 te mperature had an effect on hibiscus extract s color and thus anthocyanins. A higher hue tint value is as sociated with an increase in absorbance at 420 nm (yellow tones) in relation to that at 520 nm (red tones); this is undesirable because it is an indication of anth o cy a nins degradation Anthocyanin content was not significantly different between treat ments CE2 and HE2, CE3 and HE3, and CE4 and HE4 (Figure 31) so equivalent cold and hot water extraction conditions for anth o cyanins were found. As can be seen from F igure 32 TP extr action was better with hot water (90 C) than with cold water (25 C). P renesti and others (2007) also found that hot water (100 C for 3 min) extracted a higher phenol ic content compared to cold water hibiscus extract s. The higher concentration of polyphenolic compounds other than anthocyanins in hot water extracts may have contributed to a higher antioxidant activity in these extracts as compared to cold water extracts (Figure 3 3). Tsai and others (2002) found that hibi scus anthocyanins contributed to 51% of total antioxidant capacity and t hat other phenolic compounds were responsible for the remainder of activity Q ualitative differences were observed between the cold and hot water hibiscus extracts. Cold extracts had a clear appearance and bright red color whereas the hot extracts presented a more opaque red color and some haze possibly associated to a higher concentration of phenolic compounds other than anthocyanins. Parameters Correlations Several correlations between the parameters measured were found in both extraction processes with r2 > 0.9. Linear regression parameters and correlation coefficients are presented in Table 32. This linear behavior will only be valid before reaching equilibrium during the extract ion process, after which, the values of the

PAGE 72

72 measured parameters will remain almost constant. Similarly the equilibrium for the extraction of anthocyanins and other polypheolic compounds could be reached at different times. Since some of the methods used in this study can be easier or faster than others, these equations could be used to predict parameters measured with more time consuming methods within the range of studied extraction conditions. L* and CD values were inversely correlated. A high L*Polyphenolics Identification val ue (lighter color) will be associated with a low CD value. Color density, anthocyanins content, total phenolics and antioxidant capacity were all directly correlated. A nthocyanins and other polyphenolics identified in hibiscus extracts are presented in Figure 34 and Tables 33 and 3 4. Compounds were identified on the basis of their retention time, absorption spectrum, MS fragmentation pattern, and where possible by comparison to an authentic standard. Peak 1 (tR, 13.4 min; m ax, 271 nm) was identified as gallic acid by comparison of the absorption spectrum with a standard. This was confirmed by MS MS analysis that showed the presence of a negatively charged molecule ion ([M H]-) at m/z 169 which fragmented to produce a secondary fragment ion (MS2Peak 2 (t ) at m/z 125 (see Table 34). T he presence of gallic acid in hibiscus extract was measured previously by GC MS ( Mourtzinos and others (2008). R, 17.1 min; max, 259 nm) was identified as protocatechuic acid glucoside. The absorption spectrum was compared to a protocatechuic acid standard; the presence of the glucose molecule slightly shifted the retention time. MS analysis of the peak reveled a [M H]at m/z 315 that fragmented to yield the ion m/z 153 which

PAGE 73

7 3 corresponds t o protocatechuic acid. The difference between ions 315 and 153 gave an ion with m/z 162 which corresponds to glucose. The same MS fragmentation patterns for protocatechuic acid glucoside were reported in dried plum (Fang and others 2002). Protocatechuic ac id isolated from hibiscus extracts was demonstrated to have anti atherosclerosis (Lee and o thers 2002), antitumor (Olvera Garcia and others 2008; Tseng and others 1998) antioxidant (Lin and others 2003), and anti inflammatory (Liu and others 2002) activit ies. Peaks 3 (tR, 18.7 min; max, 326 nm), 4 (tR, 23.0 min; max, 327 nm), 5 (tR, 23.6 min; max, 327 nm), and 7 (tR, 24.3 min; max, 331 nm) were identified as caffeoylquinic acids which are esters formed between caffeic and quinic acid. Their identification was based on previously developed structurediagnostic hierarchical keys (Clifford and others 2003), UV vis spectrum and retention time was compared relative to a commercial 5 caffeoylquinic acid (chlorogenic acid) standard Peaks 3, 4, 5, and 7 produced a [M H]at m/z 353 and MS2 ions at m/z 191 (corresponds to quinic acid), 179 (corresponds to caffeic acid), 173, and 135, (peak 7 only had MS2 ions at m/z 191 and 173). Peak 4 was identified as 5caffeoylquinic acid (5CQA) by comparison with an authentic standard. According to Clifford and others (2003) 5CQA is characterized by an intense base peak at m/z 191 and a weak secondary ion at m/z 179. Peak 3 was identified as 3caf feoylquinic (3 CQA) acid since it is characterized by a base peak at m/z 191 and a relatively intense secondary ion at m/z 179 while peak 5 was identified as 4 caffeoylquinic (4 CQA) acid with a characteristic base peak at m/z 173 (Clifford and others 2003). Peak 7 was tentatively identified as a caffeoylquinic acid isomer from its absorption spectrum and fragmentation patterns (see Table 34). The presence of 5-

PAGE 74

74 CQA has been previously reported in hibiscus extracts (Mourtzinos and others 2008; SeguraCarret ero and others 2008). Peaks 6 (tR, 24.2 min; max, 529 nm) and 8 (tR, 26.2 min; maxPeaks 9 (t 521 nm) were identified as delphynidin3 sambubioside (D3S) and cyanidin3 sambubioside (C3S) which are the two major anthocyanins present in hibiscus (Table 33). Identi fication was based on their absorption spectrum and MS fragmentation patterns which have been previously reported (Juliani and others 2009; Degenhardt and others 2000; Giusti and others 1999). The difference between the MS of the molecule (597) and the agl ycone (303) for D3S gave a m/z of 294 which corresponds to xyloseglucose (132+162) known as sambubiose. Similarly, the MS for the C3S molecule (581) and the aglycone (287) yields the sambubiose disaccharide. R, 29.0 min; max, 359 nm), 10 (tR, 3 0.9 min; max, 348 nm), and 11 (tR, 32.0 min; max, 356 nm) were tentatively identified as flavonols for their characteristic absorption spectrum with max ~360 nm. Peak 12 (tR, 35.7 min; max, 355 nm) was also tentatively identified as quercetin3 rutinos ide by its absorption spectrum and MS fragmentation patterns which revealed a base peak at m/z 609 and MS2 at m/z 301. The difference between m/z 609 and 301 gave a m/z of 308 that corresponds to the disaccharide rutinose formed between rhamnose (m/z 146) and glucose.(m/z 162).The presence of rutinose h as been previously reported in hibiscus extract as part of an anthocaynin (cyanidin3 rutinoside) by SeguraCarretero and others (2008). Quercetin3 rutinose with the same MS fragmentation patterns was found in black and green tea (Del Rio and others 2004) and in pear skins (Lin and Harnly 2008).

PAGE 75

75 Polyphenolics Quantification Polyphenolics were quantified in the four hibiscus extracts studied (dried hibiscus cold water extract (DCE ), dried hibiscus hot water extract (DHE ), fresh hibiscus cold water extract (FCE), and fresh hibiscus hot water extract (FHE) (Table 35) Results were expressed in milligrams per L of extract. Hydroxybenzoic acids accounted for ~2% of the total polyphenolics quantified in the dried hibiscus extracts and ~0.5% in the fresh hibiscus extracts. Caffeoylquinic acids accounted for ~45% and ~38% of total in dried and fresh extracts, respectively while flavonols accounted for ~10% of the total in all four extracts. Anthocyanins accounted for ~45% and ~50% of the total in the dried and fresh hibiscus extracts, respectively. As seen in Table 35, the DHE sample had the highest concentration of t otal polyphenols followed by DCE FCE, and FHE. Gallic acid was not detected in the fresh extracts a nd its presence in the dried hibiscus extracts could be attributed to a breakdown of another phenolic compound during the drying process. The concentration of protocatechuic acid glucoside was higher in fresh hibiscus extracts and a significantly lower concentration of caffeoylqu inic acids was also observed compared with the dried extracts. Hibiscus anthocyanins distribution was ~68% and 64% of the total for D3S and 32 and 36% for C3S in dried and fresh extracts, respectively. This indicated that a significantly higher concentration of C3S was found in the fresh hibi scus extracts as compared to dried extracts. Delphinidin3 sambubioside was present in a significantly higher concentration in the hot water extracts as compared to the cold water ones but no significant differences were found in the concentration of cyanidin3 sambubioside in the cold and hot water extract s for both fresh and dried hibiscus.

PAGE 76

76 Conclusions Equivalent cold and hot water conditions were found for anthoc yanins extraction of dried hibiscus Similar polyphenolic profiles were observed between fresh and dried hibiscus extracts although differences were found in the concentration of compounds Hydroxybenxoic acids, caffe oylquinic acids, flavonols and anthocyanins constituted the polyphenolic compounds identified in hibiscus extracts. Findings of this research can provide more flexibility to hibiscus processing. Extraction process selection for industrial applications should consider availability of raw material (fresh or dried hibiscus), processing technology, time, and economic considerations

PAGE 77

77 Table 31. Measured pH, total solids (TS) (g of solids/100 mL of extract), titratable acidity (TA) (g of malic acid/ 100 mL of extract), and color (L*, a*, b* values, color density (CD) and hue tint (HT)) for the extracts TRT T (C) time (min) pH TS TA L* a* b* CD HT CE1 25 30 2.37 a 0.68 d 0.28 d 54. 18 a 65. 85 d 45.39 e 1.04 f 0.35 cd CE2 25 60 2.32 a 0.92 bc 0.38 abc 46.2 9 b 67.65 a 66. 92 abc 1.82 de 0.35 d CE3 25 120 2.32 a 0.97 ab 0.40 ab 43.7 8 cd 67. 50 a 68.76 a 2.05 c 0.36 c CE4 25 240 2.31 a 1.00 ab 0.44 a 40. 79 ef 67.1 6 ab 68.2 2 a b 2.5 5 ab 0.36 c HE1 90 2 2.37 a 0.79 cd 0.33 cd 44. 82 bc 66.73 bc 65. 19 c 1.73 e 0.38 b HE2 90 4 2.37 a 0.90 bc 0.37 bc 42 13 de 66.3 9 c 67.0 3 abc 2.00 cd 0.38 b HE3 90 8 2.36 a 0.95 ab 0.39 ab 39. 34 f 65.65 d 65.7 0 bc 2.34 b 0.39 a HE4 90 16 2.33 a 1.08 a 0.43 a 35.2 6 g 63.9 3 e 60. 29 d 2.70 a 0.39 a CE = Cold extraction, HE = Hot extraction. Data represents the mean of n=9. Values with similar letters within columns are not significantly different (Tukeys HSD, p > 0.05). Figure 31. Total anthocyanins content expressed as del phinidin 3 glucoside (mg/L) for the extracts The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25 C curve. Data represents the mean of n=9. Values with similar letters within the figure are not significantly different (Tukeys HSD, p > 0.05). e cd bc a d c b a 0 2 4 6 8 10 12 14 16 18 0 10 20 30 40 50 60 70 80 90 100 0 30 60 90 120 150 180 210 240 270 time (min)Anthocyanins content (mg/L)time (min) 25 C 90 C

PAGE 78

78 Figure 32. Total phenolics content expressed as gal lic acid equivalents (mg/L) for the extracts The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25 C curve. Data represents the mean of n=9. Values with similar letters within the figure are not significantly different (Tukeys HSD, p > 0.05). Figure 33. Antioxidant capacity ( m ol of TE/mL) L) for the extracts The upper time scale belongs to the 90 C curve and the lower time scale belongs to the 25 C curve. Data represents the mean of n=9. Values with similar letters within the figure are not significantly different (Tukeys HSD, p > 0.05). f e d bc de c b a 0 2 4 6 8 10 12 14 16 18 50 100 150 200 250 300 350 400 450 500 550 600 0 30 60 90 120 150 180 210 240 270 time (min)Total phenolics (mg/L)time (min) 25 C 90 C f e de c d bc b a 0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 0 30 60 90 120 150 180 210 240 270 time (min)Antioxidant capacity (mmol of TE/mL)time (min) 25 C 90 C

PAGE 79

79 Table 32. Linear regression and correlation coefficients between measured parameter s for cold and hot water extraction processes Cold Extraction Hot Extraction m b r 2 m b r 2 L vs c olor density 8.77 62.62 0.99 9.61 61.48 0.99 Color density vs anthocyanins content 0.04 0.04 0.96 0.03 0.34 0.96 Anthocyanins content vs total p henolics 0.18 2.56 0.95 0.17 3.65 0.95 Total phenolics vs a ntioxidant capacity 38.85 13.66 0.92 44.96 13.66 0.93 m = equation slope, b = equation intercept Table 33. Identification of anthcocyanins present in hibiscus using t heir spectral characteristics with HPLC DAD and positive ions in LC MS and MS2 HPLC DAD Data LC MS Data (m/z) Peak Compound a t R (min) max (nm) MS (molecule) MS 2 (aglycone) 6 Dpd 3 sambubioside 24.2 529 597 303 8 Cyd 3 sambubioside 26.2 521 581 287 a Abbrevaitions: D pd, delphi nidin; Cyd, cyanidin. Table 34. Identification of polyphenolics present in hibiscus using t heir spectral characteristics with HPLC DAD and negative ions in LC MS and MS2 and respective standards. HPLC DAD Data LC MS Data (m/z) Peak Compound tR (min) max (nm) MS MS 2 base peak other peaks 1 Gallic acid a 13.4 271 169 125 (100) d 2 Protocatechuic acid glucoside b 18.2 2 60 315 153 (100) 3 3 caffeoylquinic acid 18.7 326 353 191 (100) 179 (58), 173 (7), 135 (14) 4 5 caffeoylquinic acid a 23 .0 327 353 191 (100) 179 (2), 173 (0.4), 135 (0.70) 5 4 caffeoylquinic acid 23.6 327 353 173 (100) 191 (20), 179 (39), 135 (14) 7 Caffeoylquinic acid isomer c 24.3 331 353 191 (100) 173 (3) 12 Quercetin 3 rutinoside c 35.7 355 609 301 (100) a Confirmed with authentic standards. b Confirmed with the standard of the acid. c Tentatively identified. d Values in parenthesis indicate the intensity of the ion.

PAGE 80

80 Figure 34 HPLC chro matograms of dried hibiscus (DHE ) and fresh hibiscus (FHE) hot water extracts: (A) 520 nm, (B) 360 nm, (C) 320 nm, (D) 280 nm, and (E) 260 nm. For peak identification see Tables 33 and 3 4. 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 52.0 100 200 400 600 mAU min FHE DHE WVL:260 nm 3 4 5 8 7 6 9 10 11 12 2 1 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 52.0 200 500 1,200 mAU min FHE DHE WVL:280 nm 3 4 5 6 7 8 1 0.0 5.0 10 .0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 52.0 200 1,000 2,200 mAU min FHE DHE WVL:320 nm 3 4 5 7 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 52.0 50 200 450 mAU min 3 4 WVL:360 nm FHE DHE 5 7 10 11 1 2 9 6 8 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 52.0 100 250 500 750 1,000 mAU min FHE DHE WVL:520 nm 6 8 A B C D E

PAGE 81

81 Table 35. Pol yphenolics content (mg/L) of hibiscus samples analyzed in this studyd Compound (peak) e DCE DHE FCE FHE Hydroxybenzoic acids Gallic acid (1) f 0.65 a 0.58 a nd l nd Protocatechuic acid glucoside (2) g 0.06 b 0.05 b 0.19 a 0.13 ab Total 0.71 0.63 0.19 0.13 Caffeoylquinic acids h 3 caffeoylquinic acid (3) 67.53 b 73.01 a 51.97 c 49.70 c 5 caffeoylquinic acid (4) 43.64 b 46.23 a 39.05 c 38.15 c 4 caffeoylquinic acid 5) 17.52 b 18.81 a 14.12 c 13.51 c Caffeoylquinic acid isomer (7) 3.93 b 4.13 b 10.86 a 12.09 a Total 132.62 142.18 116.00 99.94 Flavonols i Unidentified (9) 5.56 b 5.70 ab 5.55 b 5.84 a Unidentified (10) 5.52 a 5.58 a 5.11 b 4.86 c Unidentified (11) 10.21 b 9.99 b 12.29 a 12.10 a Quercetin 3 rutinoside (12) 8.45 a 9.20 a 9.38 a 9.21 a Total 29.74 30.47 32.33 32.01 Anthcoyanins Delphinidin 3 sambubioside (6) j 87.32 c 100.90 a 87.76 c 96.16 b Cyanidin 3 sambubioside (8) k 41.62 b 44.88 b 50.30 a 50.89 a Total 128.94 145.78 138.06 147.05 Total phenolic compounds 292.01 319.06 286.58 279.13 Data represents the mean of n=6. d(Tukeys HSD, p > 0.05). e Peak numbers refer to the compounds identified in Tables 33 and 34. Values with similar letters within rows are not significantly different f,g,h,I,j,kQuantified with gallic acid, protocatechuic acid, chlorogenic acid, quercetin, delphinidin3 glucoside, and cyanidin3 glucoside standards respectively l Abbreviations: nd, not detected.

PAGE 82

82 CHAPTER 4 AROMA PROFILES OF BEVERAGES OBTAINED FRO M FR ESH AND DRIED HIBISCUS Introduction Hibiscus sabdariffa commonly known as hibiscus or roselle, grows in many tropical and subtropical countries and is one of the highest volume specialty botanical products in international commerce (Plotto 1999). Hibiscus is an annual herbaceous shrub and is a member of the Malvaceae family. The swollen calyces, which are red and cup like are the part of the plant of commercial interest (Morton 1987; De Castro and others 2004). Fresh and dried hibiscus calyxes are used to prepare cold and hot beverages. Sweeteners and spices can be added depending on the region where it is consumed. Extensive work has been done in the area of hibiscus anthocyanins due to their beneficial heal th effects high antioxidant properties, and potential source as a food colorant (Tee and others 2002; Tsai and others 2002; Tsai and Huang 2004; Prenesti and others 2007: S yago Ayerdi and others 2007). Studies with human patients have also shown that the regular consumption of hibiscus extract has an antihypertensive effect (Haji Faraji and Haji Tarkhani 1999; HerreraArellano and others 2004) and reduces serum cholesterol in men and women (Lin and others 2007). Hibiscus flavor is a combination of sweet a nd tart. Few studies have been done related to hibiscus flavor. Gonzalez Palomares and others (2009) ident ified 20 volatile compounds in hibiscus extract using SPME and GC MS, including terpenoids, esters, hydrocarbons, and aldehydes. They also found 14 co mpounds in reconstituted spray dried extracts from which only 10 were present in the original extract and the other 4 were products of degradation. Thermally generated volatiles from untreated, frozen, hot -

PAGE 83

83 air dried at 50 C, and hot air dried at 75 C hib iscus by steam distillation w ere analyzed by GC and GC MS (Chen and others 1998). They characterized more than 37 compounds including fatty acid derivatives, sugar derivatives, phenol derivatives, and terpenes. The objective of this study was to determine the aroma profile di fferences between four extracts obtained from fresh and dried hibiscus extracted at two different conditions, by GC MS and GC olfactometry. Materials and Methods Sample Preparation Fresh and sun dried hibiscus ( Hibiscus sabdariffa cv. Criollo) were obtained from the same harvest (November 2006 January 2007) from Puebla, Mexico. Hibiscus samples were stored in glass jars, flushed with nitrogen and kept frozen at 20 C until used. Four different extracts were prepared; fresh (F) and dried (D) hibiscus were mixed with distilled water in a ratio of 1:4 and 1:40 (w/v) respectively and extracted at 22 C for 240 min (cold extraction (CE)) and 98 C for 16 min (hot extraction (HE)). Extraction ratios (hibiscus: water) were determined based on moisture content of fresh (90%) and dried (9%) hibiscus (measured at 105C for 24 h in an oven (Prec ision Scientific, Winchester, Va., U.S.A.)). Stirring at low speed was applied for cold extraction and no stirring was applied for hot extraction. After extraction, samples were filtered under vacuum using Whatman filter paper #4. The pH of the samples was measured using a pH meter E A920 (Orion Research; Boston, Ma ss., U.S.A. ) and Br ix was determined with an ABBE Mark II refractom eter (Leica Inc.; Buffalo, N Y U.S.A.).

PAGE 84

84 Headspace Volatiles Sampling Headspace volatiles were extracted and concentrated using SPME. Ten milliliters of hibiscus extracts were added to a 22 mL screw cap amber glass vial PTFE/silicone septa containing a small stir bar. Samples were equilibrated for 20 min in a water bath at 40 C and hibiscus headspace volatiles were extracted for 30 min using a 1 cm 50/30 mm DVB/Carboxen/PDMS SPM E fiber (Supelco, Bellefonte, Pa., U.S.A. ). Before each exposure the fiber was cleaned for 5 min in the injection port (200 C) of the GC O or GC MS instruments. GC O Analysis GC O analysis was carried out using a HP 5890 Series II Plus GC (Palo Alto, Ca lif ., U.S.A ) with a sniffing port and a flame ionization detector (FID). Hibi scus volatiles from the SPME fiber were desorbed in the GC injection port. A SPME injector liner ( SPME injection sleeve, 0.75 mm i.d. Supleco; Bellefonte, Pa., U.S.A. ) was used. The GC column effluent was split between the FID and the olfactometer. The injector temperature was 200 C, and the detector temperature was 250 C. Helium was used as the carrier gas at 1.67 mL/min. The oven was programmed from 35 C (held for 5 min) to 2 50 C at 6 C/min with a final hold of 10 min. Volatiles were separated using a DB 5 (30 m x 0.32 mm. i.d. x 0.5 m, J&W Scientific; Folsom, Ca lif ., U.S.A.) or a DB Wax (30 m x 0.32 mm. i.d. x 0. 5 m, Restek; Bellefonte, Pa., U.S.A. ) column. Two olfactory assessors were employed. Samples were sniffed two times by each assessor in each column. Aroma descriptions and approximate times were recorded for every sample. Assessors indicated the intensity of each aroma peak using a linear potentiometer with

PAGE 85

85 a 01 V signal. Aromagrams and FID chromatograms were recorded and integrated using Chrom Perfect 4.4.23 software (Justin Innovations, Inc.; Palo Alto, Calif ., U.S.A.). A peak was considered aroma active only if at least half the panel found it at the same time with a similar description. Linear retention index values were determined for both columns using a series of alkanes (C5C25) run under identical conditions. GC MS Analysis Mass Spectrometry ( GC MS ) was used to identify the odor active volatile s detected i n the GC O experiment GC MS analysis was conducted using a HP 6890 GC coupled with a MSD 5973 (Agi lent Technologies; Palo Alto, Calif ., U.S.A.). Hibiscus volatiles from the SPME fiber were desorbed in the GC injection port (splitless mode) at 200 C. The fiber was removed after 5 min exposure in the injection port. Volatiles were separated using a DB 5 (30 m x 0.32 mm. i.d. x 0.5 m J&W Scientific; Folsom, Ca lif ., U.S.A.) or a DB Wax (30 m x 0.32 mm. i.d. x 0.5 m, Restek; Bellefonte, Pa., U.S.A.) column. The oven was programmed from 35 C (held for 5 min) to 250 C at 6 C /min with a final hold of 10 min. Helium was used as the carrier gas at 1.67 mL/ min The mass spectrometer was operated in the total ion chromatogram (TIC) at 70 eV. Data were collected from 35 m/z to 400 m/z All samples were run in duplicate in each column. Chromatograms were recorded and integrated using Enhanced Chemstation (version 01.00) software (Agi lent Technologies; Palo Alto, Calif ., U.S.A.). Mass spectral matches were made by comparison with NIST 98.1 (NIST; Gaithersburg, Md., U.S.A. ) and WILEY 8.1 (Wiley; New York, N Y U.S.A.) mass spectral libraries. Only those compounds with spectral fit values equal to or greater than 850 were considered

PAGE 86

86 positive identifications. L inear retention index values were determined for both columns using a series of alkanes (C5C25) run under identical conditions. Identification Procedures Identifications were based on the combined matching of retention indices (LRI values) from DB 5 and DB Wax columns, matches made from spectra in the NIST and WILEY libraries aroma descriptors, and linear retention index matches from literature. Statistical Analysis Analysis of variance (ANOVA) and mean separation using Tukeys test were performed to evaluate differences in pH and Brix between the analyzed samples using SAS 9.0 Statistical software (SAS Ins titute Inc., Cary, N C U.S.A.). Results and Discussion Four samples using fresh frozen and s un dried hibiscus were prepared: DH E (dri ed hibiscus hot water extraction) DCE (dried hibiscus cold water extraction, FH E (fresh hibiscus hot water extraction) and FC E (fresh hibiscus cold water extraction). Extraction conditions and measured pH and Brix values are presented in Table 41. There was not a significant difference ( p = 0.0581 ) in Brix between the four samples. Samples prepared with dried hibiscus had a significantly ( p = 0.0003) lower pH as compared to those prepared with fresh hibi scus, reason unknown. Hibiscus Volatiles Composition Hibiscus volatiles were divided into five chemical groups. A total of 14 aldehydes, 10 alcohols, 5 ketones, 2 terpenes, and 1 acid were identified. The relative difference in total volatiles in terms of peak area was normalized to total peak area of DHE (dried hib iscus hot water extraction) which was 590. Results are shown in Figure 41. In relation to DHE, total peak area was reduced by 40, 59, and 98% for DCE, FHE, and

PAGE 87

87 FCE samples respectively. In both fresh and dried hibiscus, hot water extraction gave a higher concentration of volatiles as compared with cold extraction which indicates that temperature facilitates the extraction process but it can also lead to undesirable degradation reactions of hibiscus aromas. In the same way, dried hibiscus extracts had a hi gher concentration of volatiles as compared to fresh hibiscus. A bigger gradient in moisture content between the dried hibiscus (9% moisture) and the extraction solvent (water (100% moisture)) as compared to the gradient between the fresh hibiscus (90% moi sture) and water (100% moisture) may have helped make the extraction process faster and thus increased the concentration of volatiles extracted from the dried hibiscus Composition of the four samples was similar but there were major quantitative differences. Aldehydes comprised the largest group of volatiles contributing between 57 and 63% of the total in the hot water extracts and from 37 to 45% in the cold water extracts, followed by alcohols (23 to 24% in hot water extracts and 28 to 36% in the cold wa ter extracts), ketones (712%), acids (48%), and terpenes (23%). In the case of FCE, acids (19%) were higher than ketones. Aldehydes were present in a higher proportion in hot water extracts while alcohols were present in a higher proportion in cold water extracts. This may indicate that extraction temperature could influence the aroma profiles of the obtained extracts by accelerating the degradation or formation of compounds. GC MS Identifications A total of 32 volatiles w ere identified using GC MS in hi biscus samples, 15 of which were not identified before in hibiscus. Limonene, linalool, terpineol, eugenol, and furfural were previously identified in two studies (Gonzalez Palomares and others 2009; Chen and others 1998) while nine other compounds (hexanal, heptanal, octanal,

PAGE 88

88 nonanal, 2heptenal (E), 5methyl furfural, 1hexanol, dehydroxylinalool oxide b, acetic acid) and two other compounds (decanal and benzaldehyde) were also found in hibiscus extracts by Chen and others (1998).and Gonzalez Palomares and others (2009).respectively. Differences in volatile profiles among studies can be attributed to the hibiscus variety used and the extraction methods. The extraction solvent polarity as well as the extraction conditions (time, solute concentration, and temperature) may impact the aroma profile of the final product. Table 42 lists the 32 volatiles detected in this study. To compare the volatiles in the four extracts, peak areas were normalized on the single largest peak found in all samples. This peak was the nonanal peak in the DHE sample. It was assigned a value of 100 and the remaining peaks in all four samples were normalized to it. Twenty eight, 25, 17, and 16 volatiles were found in the DHE, DCE, FHE, and FCE respectively Thirteen compounds (hexanal, heptanal, limonene, octanal, 6mehtyl 5 hepten2 one, nonanal, 1 octen3 ol, acetic acid, decanal, bornylene, 2nonenal (E), 1octanol, and geranylacetone) were present in all four samples and their concentration was lower in the fresh and the cold water extracted samples. Nonanal (100) and decanal (99) were the volatiles present in highest concentration in DHE and were also among the three compounds present in highest concentration for the other three samples. Nonanal was 36 and decanal was 39 for DCE, 43 and 75 for FHE and 3 and 2 for FCE. Nonanal and decanal are aldehydes that may form as a product of lipid oxidation. Dehydroxylinalool oxide a and b were only present in extracts from dried hibiscus. Since there are similar amounts of both compounds in the extracts obtai ned with cold and hot water, thes e could be degradation products of linalool formed during

PAGE 89

89 the drying process. Furfural and 5methyl furfural were also only detected in the extracts from dried hibiscus. These compounds are sugar degradation products and may also have developed during hibiscus drying. On the other hand, linalool and 1 hexanol 2 eth yl were only detected in the extracts from fresh hibiscus. Their absence in the dried samples may be attributed to the fact that these compounds may have degraded during drying and led to the formation of other compounds. GC O Aroma Profiles A total of 22 aroma compounds were found in hibiscus extracts and are list ed in Table 43. Peak heights were normalized to the most intense peak in all four samples. This peak was 6methyl 5 hepten2 one in the DHE sample. It was assigned a value of 100 and the remaining peaks in all four samples were normalized to it. Seventeen, 16, 13, and 10 aroma active compounds were detected for DHE, DCE, FHE and FCE samples respectively. Seven compounds were detected in all four samples and were confirmed with GC MS (hexanal, 3octanone, octanal, 6methyl 5 hepten2 one, nonanal, 2,4nonadienal (E,E), and geranylacetone). The most intense odorants were 6mehtyl 5 hepten2 one and nonanal in all four extracts followed by geranylacetone, eugenol, and 2 Nonenal (E) in the DHE sample, geranylacetone, 2Nonenal (E), and an unidentified compound for the DCE sample, and linalool, geranylacetone, and octanal for the fresh hibiscus extracts, The compound 6 methyl 5 hepten2 one was present in all four samples and had the highest intensity in all of them. This compound was described to have a mushroom, dirt, green aroma and has previously been reported in tomatoes (Buttery and others 1987) and R ooibos tea (Kawakami and others 2003). Nonanal was the second highest

PAGE 90

90 in aroma intensity in all four samples with a descriptor of fruity, green. Geranylaceton e was present in all samples was among the five highest intensity compounds and was described as fruit like, apple sauce smell Geranylacetone has been found previously in Merlot and Cabernet wines (Gurbuz and others 2006) and is one of the major compone nts of Rooibos tea. In the fresh hibiscus extracts the compounds linalool (floral, woody, citrus) and octanal (lemon, citrus) were among the highest intensity aroma compounds. As mentioned before, linalool was not detected in dried hibiscus extracts while octanal is present in dried hibiscus samples and is the sixth highest intensity peak in both cold and hot water extracts. Linalool is a compound associated with floral notes and has previously been reported to be present in j asmine green tea (Ito and others 2002) and citrus blossom (Jabalpurwala and others 2009) among others. Octanal has been described to have a fruity, citrus aroma in lychee (Mahattanatawee and others 2007). The compound 2 Nonenal (E) (cucumber, g reen, floral ) was present in dried hibiscus samples as one of the highest intensity peaks. Eugenol (sweet spices) was only detected in the DHE sample while an unidentified (anise) compound was present in the DCE sample with a high intensity. The five highest intensit y peaks for all four samples were: 2 ketones, 2 aldehydes and 1 alcohol. The compounds 6mehtyl 5 hepten2 one, geranylacetone, and 2Nonenal (E) which are important aroma impact compounds present in hibiscus extracts were identified for the first time in hibiscus. Conclusion The four hibiscus extracts studied had a similar chemical composition of aroma compounds with hot extracted hibiscus samples having a slightly higher aldehyde

PAGE 91

91 concentration and cold extracted samples a slightly higher alcohol concentr ation. Total peaks concentration was the highest for the dried hibiscus hot water extract, and decreased in both cold water extracts and fresh hibiscus extracts. There were some differences in aroma peak intensities in the four hibiscus samples with the dr ied hibiscus hot water extraction having the highest intensity. In general hibiscus aroma is a combination of earthy, green, floral, and fruity notes but the final flavor profile is affected by the preservation and extraction process.

PAGE 92

92 Table 4 1. Extrac tion conditions and measured pH and Brix values for hibiscus samples included in this study Sample Extraction temperature (C) Extraction time (min) Hibiscus: water ratio pH Brix DH E 98 16 1:40 w/v 2.48 0.01 b 1.25 0.07 a DC E 22 240 1:40 w/v 2.49 0.00 b 1.25 0.07 a FH E 98 16 1:4 w/v 2.55 0.01 a 1.10 0.00 a FC E 22 240 1:4 w/v 2.57 0.01 a 1.10 0.00 a DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FH E = fresh hibiscus hot water extraction. FC E = fresh hibiscus cold water extraction. Figure 41. Chemical composition of hibiscus headspace volatiles. Total number of compounds for each class is put in parentheses. All four samples were normalized to the total peak area of DH E (dried hibiscus hot extraction) DCE = dried hibiscus cold extraction, FHE = fresh hibiscus hot extraction, FCE = fresh hibiscus cold extraction. 0 10 20 30 40 50 60 70 80 90 100 DHE DCE FHE FCEPeak Area PercentHibiscus Samples Aldehydes (14) Alcohols (10) Ketones (5) Terpenes (2) Acids (1)

PAGE 93

93 Table 42. MS i dentification of hibiscus volatiles Peak areas were normalized (10 0) to the largest peak i n all four samples. # Name CAS # LRI Normalized peak are a (%) CW DB5 DH E DC E FH E FC E 1 Hexanal a 66 25 1 1100 793 55.42 25.93 3.77 0.33 2 Heptanal a 111 71 7 1195 903 4.99 0.29 0.21 0.09 3 Limonene a,b,c 138 86 3 1026 0.22 0.07 0.18 0.15 4 Dehydroxylinalool oxide a 13679 86 2 1210 993 37.86 35.04 5 Dehydroxylinalool oxide b a 13679 86 2 1246 1007 26.35 21.26 6 3 Octanone 106 68 3 1264 2.90 5.60 7 Octanal a 124 13 0 1299 1002 36.27 14.00 12.60 0.31 8 2,2,6 Trimethylcyclohexanone 2408 37 9 1329 1031 7.14 4.54 9 2 Heptenal, (E) a 18829 55 5 1342 958 4.73 10 6 methyl 5 Hepten 2 one 110 93 0 1355 989 21.70 16.73 19.33 1.09 11 1 Hexanol a 111 27 3 1373 7.66 4.80 0.48 12 Nonanal a 124 19 6 1405 1100 100.00 35.63 42.97 2.56 13 Octenal 2548 87 0 1448 5.81 1.05 14 1 Octen 3 ol 3391 86 4 1468 983 46.08 39.75 7.40 0.34 15 Acetic acid a 64 19 7 1485 23.21 22.40 17.94 2.85 16 Furfural a,b 98 01 1 1496 832 24.33 19.16 17 1 Hexanol 2 ethyl 104 76 7 1508 1030 6.66 1.89 18 Decanal b 112 31 2 1513 1204 98.71 39.45 74.76 1.96 19 Bornylene 464 17 5 1541 1227 9.47 8.63 5.24 0.34 20 Benzaldehyde b,c 100 52 7 961 0.07 21 2 Nonenal (E) 18829 56 6 1555 1159 11.53 3.67 1.35 0.13 22 Linalool a,b 78 70 6 1568 1098 23.13 1.04 23 1 Octanol 111 87 5 1577 1071 17.93 9.94 10.64 0.42 24 5 Methyl furfural a 620 02 2 1608 4.14 3.85 25 2 Nonanone c 821 55 6 1089 0.13 0.13 26 1 Nonanol 143 08 8 1678 1172 15.78 13.78 27 terpineol a,b 98 55 5 1725 2.63 28 (E,E) 2,4 Nonadienal 5910 87 2 1728 1213 13.69 8.87 29 2 Undecenal 53448 07 0 1772 1362 8.58 30 (E,E) 2,4 Decadienal 25152 84 5 1836 1315 5.67 4.87 31 Geranyl acetone 3796 70 1 1876 1440 7.36 9.03 5.47 0.72 32 Eugenol a,b,c 97 53 0 1356 0.07 Total normalized peak area 590.1 351.3 239.1 14.7 DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FH E = fresh hibiscus hot water extraction. FC E = fresh hibiscus cold water extraction. a Compounds previously reported in H. sabdariffa by Chen and others (1998). b Compounds previously reported in H. sabdariffa by Gonzalez Palomares and others (2009). c LRI values for this compounds were calculated using peak areas from DB 5 column.

PAGE 94

94 Table 43. Hibiscus aroma active compounds. Peak heights were normalized (100) to the most intense peak in all four samples. # Name LRI Aroma descriptor Normalized peak height (%) CW DB5 DHE DCE FH E FC E 1 Unknown 1054 sweet, fruity 33.33 32.17 2 Hexanal a 1102 green, grass, nutty 47.50 43.33 30.00 16.67 3 Unknown 1200 sweet, fruity 44.17 41.67 4 3 Octanone a 1270 butter, cookie, baked 33.33 30.00 26.67 16.67 5 Octanal a 1304 1003 lemon, citrus 58.33 57.50 50.00 33.33 6 6 methyl 5 Hepten 2 one a 1349 980 mushroom, dirt, green 100.00 83.33 91.67 58.33 7 Nonanal a 1410 1103 fruity, green 95.83 82.50 83.33 54.17 8 Octenal a 1452 rancid nuts 25.00 16.67 9 1 Octen 3 ol a 1477 975 mushroom, dirt, metallic 56.67 50.00 10 Furfural a 1497 sweet, baked bread 52.92 11 Decanal a 1520 sweet, nutty 33.33 25.00 12 2 Nonenal (E) a 1562 1154 cucumber, green, floral 61.67 66.67 45.00 13 Linalool a 1570 floral, woody, citrus 58.33 52.50 14 1 Octanol a 1579 fresh leather, chemical 20.83 16.67 15 1 -Nonanola 1674 chemical, painty 25.00 16 2,4 Nonadienal, (E,E) a 1735 1215 rancid nuts, citrus, green 49.17 44.17 25.00 20.00 17 Unknown 1754 1244 anise 63.33 41.67 33.33 18 2 Undecenal a 1780 green, grass 35.83 31.67 19 Unknown 1850 rancid nuts 33.33 26.67 20 Geranyl acetone a 1870 1430 fruit like, apple sauce 66.67 75.00 58.33 45.83 21 Unknown 1940 sweet spices, floral 41.67 22 Eugenol a 2100 1350 sweet spices 66.67 Total intensity 872.9 760.8 611.7 363.0 DHE = dried hibiscus hot water extraction. DCE = dried hibiscus cold water extraction. FH E = fresh hibiscus hot water extraction. FC E = fresh hibiscus cold water extraction. a Compounds confirmed with GC MS.

PAGE 95

95 CHAPTER 5 PROCESSING HIBSCUS BEVERAGE USI NG DENSE PHASE CARBO N DIOXIDE : MICROBIAL AND PHYTOHCE MICAL STABILITY Introduction Juices and beverages are traditionally preserved by thermal methods which are effective in reducing microbial loads but can also lead to organoleptic and nutritional changes. Nonthermal processes are an alternative that may help preserve the color, flavor, and nutrients of food and thus address consumers demands for high quality, fresh like products with extended shelf life. Dense phase carbon dioxide (DPCD) is a continuous nonthermal processing system for liquid foods that uses pressure ( (CO2) to inactivate micr oorganisms. In a continuous flow DPCD system several variables are controlled during processing: pressure, temperature, residence time, and %CO2 in the liquid food The amount of CO2 used should assure a complete saturation of the liquid but since its sol ubility at processing conditions is not known this can lead to the use of excess CO2Hibiscus sabdariffa, a member of the Malvaceae family i s an annual shrub widely grown in tropical and subtropical regions including Africa, South East Asia and some countries of America. The calyces contain anthocyanins and other phenolics and are of commercial interest. They are used either fresh or dehydrated to prepare hot and elevating production costs. Previous studies with muscadine grape juice showed that DPCD was more effective than pas teurization in retaining anthoc yanins and other phenoli c compounds during processing and storage (Del Pozo Insfran and others 2006a; 2006b). Furthermore, DPCD was effective in extending the shelf life of coconut water (Damar and others 2009) and red grapefruit juice (Ferrentin o and others 2009) for up to 9 and 6 weeks of refrigerated storage, respectively.

PAGE 96

96 cold beverages which are commonly mixed with a sweetener and are character ized by an intense red color, acidic flavor, and a sensation of freshness. Recently there has b een increasing interest in hibiscus anthocyanins due to their beneficial health effects and high antioxidant properties which have been extensively evaluated (Tee and others 2002; Tsai and others 2002; Tsai and Huan g 2004; Prenesti and others 2007: S yagoAyerdi and others 2007) and as a potential source of natural food colorant. The objectives of this study were (1) to determine the solubility of CO2Materials and Methods in a hibiscus beverage, (2) to optimize DPCD processing param eters (pressure and residence time) based on microbial reduction, and (3) to monitor during 14 weeks of refrigerated storage the microbial, physicochemical, and phytochemcial changes of DPCD processed hibiscus beverage compared to thermally treated and control (untreated) beverages Chemicals and Standards Commercial standards of gallic acid, chlorogenic a cid, and quercetin were purchased fr om SigmaAldrich (St. Lous, Mo., U.S.A.). Caffeic acid was purchased from ACROS Organics ( Geel, Belgium ). Delphinidin3 glucoside and cyanidin3 glucoside were purchased from Polyphenols Laboratories AS (Sandnes, Norway). AAPH (2,2 azobis(2methylpropionamidine) dihydrochloride), fluorescein (free acid), Trolox (6hydroxy 2,5,7,8 tetramethylchroman2carboxilic acid) and Folin Ciocalteus reagent were purchased fr om Sigma Aldrich (St. Lous, Mo., U.S.A.). Beverage Preparation Dried Hibiscus sabdariffa (cv. C riollo) (9% moisture content ) obtained from Puebla, Mexico was mixed with water (1:40 w/v) using a 200 L stainless steel mixing

PAGE 97

97 tan k Model UAMS (Cherry Burrell, I owa U.S.A.) and maintained at 25C for 1 h. Mixing was applied intermittently by alternating intervals of 10 min mixing and 10 min rest. The extract was then filtered using four layers of cheesecloth. A beverage was prepared by adding sucrose to a concentration of 100 g sucrose/L of extract and then was placed in 3 gallon sealable buckets. For the DPCD process optimization, the beverage was incubated at 25 C for 4 days to obtain a high initial microbial load. The spoiled beverage was placed in the refrigerator at 4 C for 24 h before processing with DPCD. For the solubility and storage experiments the beverage was prepared as mentioned above, without incubation, one day before processing and refrigerated. Solubility Experiment CO2 solubility in the hibiscus beverage was measured between 6.9 and 31.0 MPa at 40 C using an apparatus designed and built at the University of Flor ida Food Science and Human Nutri tion department (Gainesville, Fla., U.S.A. ) as previously described by Ferrentino and others (2009). In this batch system a known volume of sample was saturated by bubbling CO2 at the desired exper imental conditions and then dissolved CO2 was measured at atmospheric pressure. Solubility of CO2 in water at the same experimental conditions was also measured for comparison. Exp eriments were done in duplicate and results were expressed as g of CO2Dense Phase CO /100 g of liquid sample. 2The DPCD equipment located at the Univer sity of Florida (Gainesville, Fla U.S.A ) was constructed by APV (Chicago, Il l ., U.S.A.) for Praxair (Chicago, Il l ., U.S.A.). A schematic diagram of the equipment is presented in Figure 211. In this continuous flow equipment CO Equipment 2 and the hibiscus beverage were pumped through the system and mixed before entering the highpressure pump (intensifier pump). Processing pressure

PAGE 98

98 levels were controlled by this pump while the desired temperature was maintained in the holding coil (79.2 m, 0.635 cm i.d.). Turbulent flow and mixing were reached at the entrance of the coil by passing the mixture through a static mixer and a small diameter tube (length of about 180 cm). Residence time was adjusted by setting the flow rate of the mixture. An expansion valve was used at the end of the process to release the CO2DPCD Process Optimization from the mixture and the beverage was collected into 1 L sterile bottles as previously described by Damar and others (2009). Op timal processing parameters to achieve a microbial log reduction of 5 were determined by using response surface methodology. A central composite design (Table 5 1) consisting of 11 experiments with 4 factorial points, 4 star points, and 3 central points in which the independent variables were pressure (P) (13.834.5 MPa) and residence time (RT) (5 8 min), and the dependent variables were yeasts and molds (Y&M) and aerobic plate counts (APC) was used. With this response surface design we were able to reduce the volume of beverage required and were able to prepare it in one batch. The total volume needed for the 11 experiments was ~ 160L. Processing parameters were selected based on previous research results and equipment specifications. Hibiscus beverage with an initial microbial load of 3.0 x 107 CFU/mL for Y&M and 4.9 x 103 CFU/mlL for APC was processed at the different experimental condit ions at a constant temperature (40 C) and constant CO2 level (8%) which was selected based on the minimum flow that could be handled by the CO2 pump. Microbial counts from each experimental condition were made in duplicate by serially diluting (1x101 to 1x105) 1 mL of beverage in 9 mL sterile Butterfields phosphate buffer (Weber Scientific,

PAGE 99

99 Hamilton, N. J. U.S.A.). Microbial counts were determined by plating 1 mL of each dilution in dupli cate for yeasts and molds and aerobic count plates (3M Petr ifilm Microbiology, St. Paul, M i n n ., U.S.A.) and enumerating after 48 h at 35 C and 5 d at 24 C respectively according to the manufacturers guidelines. Thermal Processing Conditions For thermal processing, hibiscus beverage was pumped by a peristaltic pump (Col e Parmer, Chicago Il l ., U.S.A.) throu gh two stainless steel tube sections ( 3.2 m, 0. 457 cm i.d. ea ) placed inside a tempe rature controlled water bath ( Precision Scientific, Chicago Il l ., U.S.A.) In the first section, the beverage was heated to 75 C (temperature was measured using a thermoc ouple) and then entered the second section w here it was he ld at 75 C for 15 s The beverage was then passed through a cooling stainless steel tube ( 5.2 m, 0.457 cm i.d.) in a water/ice bath and chilled to ~15 C before it was collected into 1 L sterile glass jars. Platinum cured s ilicone tubing (0.635 cm i.d.; Nalgene, Rochester N Y U.S.A.) was used to connect the pump to the stainless steel heating, holding and cooling sections. A schematic diagram of the setup used for the hibiscus beverage thermal processing is presented in Figure 52. Storage Experiment Fresh prepared hibiscus beverage was divided into three parts. One par t was kept as control and did not receive any treatment; the second par t was processed using DPCD at 34.5 MPa 8% CO2, 6.5 min, and 40 C whil e the third part was thermally treated at 75 C for 15 s. Each processing condition was repeated in triplicate. DPCD processing parameters were determined based on the solubility and optimization studies described above. Both the control and treated samples were stored in 1 L glass jars. Microbial, physicochemical and phytochemical analyses were performed at weeks

PAGE 100

100 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 of refrigerated storage at 4 C To quantify individual anthocyanins and other phenolic com pounds present in hibiscus, HPLC analysis was done at weeks 0, 2, 5, 8, and 14 of storage. The methods used in all the analyses are described as follows. Microbial Analysis Total aerobic plate count (APC) and yeasts and molds (Y&M) were measured as describ ed in the optimization section. pH, Brix, and Titratable Acidity pH and Brix were measured using a pH meter E A920 (Orion Research; Boston, Mass., U.S.A.) and a ABBE Mark II refractometer (Leica Inc.; Buffalo, N Y U.S.A.). A Brinkmann Instrument (Brinkmann Instruments Co., Westbury, N Y U.S.A.) consisting of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was used to measure titratable acidity (TA). Samples of 10 mL were used and TA was determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid (g/100 mL). Color Density and Hue Tint Color density and hue tint were determined by measuring the absorbance (A) at 420, 520, and 700 nm of 200 L samples using a spectrophotometer (SpectraMax 190, Molecular Devices, Sunnyvale Calif., U.S.A. ) and calculated as: Color density = [(A420 nm A700 nm) + (A520 nm A700 nmHue tint = (A )] 420 nm A700 nm)/(A520 nm A700 nmas described by Giusti and Wrolstad (2005). )

PAGE 101

101 Anthcyanin Content, Total Phenolics and Antiox idant capacity Anthocyanin content was determined by pH differential method ( A510 nm and A700 nm at pH 1.0 and 4.5, dilution factor (DF) of 4) and expressed in mg/L of delphinidin3 glucoside (MW = 465.2, = 23700) ( Giusti and Wrolst ad 2005). Total phenolics were measured using the FolinCiocalteu assay (A765 nmAntioxidant capacity was evaluated using the oxygen radical absorbance capacity (ORAC) assay and results were expressed as Trolox equivalents (TE) per milliliter ( mol of TE/mL) as described by Huang and others (2002) using a SpectraMax Gemini XPS microplate sprectrofluorometer ( Molecular Devices, Sunnyvale, Ca., U.S.A.). Data was acquired and analyzed using SoftMax Pro 5.2 software ( Molecular Devices, Sunnyvale, Calif ., U.S.A. ). DF of 4) and quantified as gallic acid equivalents (mg/L) (Waterhouse 2005). Absorbance measurements for anthocyanin content and total phenolics were made using a SpectraMax 190 spectrophoto meter (Molecular Devices, Sunnyvale Calif ., U.S.A. ). HPLC Quantification of Polyphenolics Polyphenolics were identified by comparison of UV/vis (190660 nm) spectral interpretation, retention time, and comparison to standards. Anthocyanins and polyphenolics were quantified using a Dionex HPLC system equipped with an autosampler/injector and diode array (PDA 100) detector (Dionex, Sunnyvale, Calif ., U.S.A.). Compounds were separated on a 250 x 4.6 mm Dionex C18 5 m 120A column (Dionex, Sunnyvale, Calif ., U.S.A.). Mobile phases consisted of water (phase A) and 60% methanol in water (phase B), both adjusted to pH 2.4 with o phosphoric acid. A gradient solvent program ran phase B from 0% to 60% in 20 min; 60% to 100% in 20

PAGE 102

102 min; 100% for 7 min; 100% to 0% in 3 min and final conditions were held for 2 min. The flow rate was 0.8 mL/min, a nd detection was done at 260, 280, 320, 360 and 520 nm. Statistical Analysis were performed in the solubility study to evaluate the effect of pressure on CO2All statistical analyses were conducted using SAS statistical software (SAS Institute Inc., Cary, N C U.S.A.). solubility. Response surface methodology was used in the DPCD optimization study to determine optimal processing conditions. Repeated measures ANOVA and mean treatment (fresh (CONTROL), thermal (HT ST), and D PCD processed) and storage time (0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 weeks) on the dependent variables measured. Results and Discussion Solubility Measurements CO2 solubility in a hibiscus beverage and water was measured between 6.9 and 31.0 MPa at 40 C. Pressure had a significant effect on solubility of CO2 in both the hibiscus beverage (4.16 to 5.06 g CO2/100 mL from 6.9 to 3 1.0 MPa) and water (4.50 to 6.32 g CO2/100 mL from 6.9 to 31.0 MPa). After 17.2 MPa, CO2 solubility remained almost constant in both the hibiscus beverage and water (Figure 51). CO2 solubility in water was significantly higher than the hibiscus beverages at all pressures tested but 6.9 MPa. The presence of solutes such as sugars and acids in the hibiscus beverage l owered the amount of CO2 that could be dissolve d. Previous studies have shown that solubility of CO2 in fruit juices is lower than that of pure water because of the presence of solutes CO2 solubility in orange and apple juice measured at 15.9 MPa was around

PAGE 103

103 5% (Calix and others 2008) while that of grapefruit juice at 31.0 MPa was 4.97% ( Ferrentino and others 2009). These v alues are comparable to those obtained for hibiscus beverage (5.06%). Based on the CO2 solubility results, a first attempt to use 6% CO2 for the DPCD experiments was tried. This concentration of CO2 (1% higher) would assure a complete saturation of CO2 in the beverage. This decision was made to account for the fact that the solubility test was performed using a batch system with long contact time between the CO2 and the beverage whereas the DPCD processing equipment is a continuous system in which lower contact times are used. After processing the hibiscus beverage using 6% CO2 and acquiring data, the DPCD system showed that there were fluctuations in the CO2 flow during the process because the CO2 pump was not designed to handle such a low flow. Our sec ond attempt was to find the minimum flow that would assure a constant reading throughout the process. After several attempts, it was found that the CO2 pump could maintain a steady flow of 8% CO2Microbial Inactivation Study and we used this value for our DPCD processing experiments. Initial microbial loads in the beverage obtained after incubation for 4 days were 3 x 107 CFU/mL for Y&M and 4.9 x 103 CFU/mL for APC. The APC population reached is not very high which can be a result of microflora competit ion in which the low pH (2.43) of the beverage and high sugar concentration (9.7 Brix) favored the growth of Y&M. The response surface experimental design and achieved log reductions for each of the treatments tested is presented in Table 51. A minimum o f 5 log reduction for Y&M and 0.85 log reduction for APC was achieved for all DPCD treatments.

PAGE 104

104 Several mechanisms have been proposed for DPCD inactivation of microorganisms (Damar and Balaban 2006). One of the main factors that lead to microbial inactivat ion is the pH lowering effect when CO2 is dissolved. Since the hibiscus beverage has an initial low pH, this reduces the lowering pH effect of CO2Two quadratic equations, (1) and (2), were obtained from the central composite design solution to describe Y&M (r and this may be the reason for low bacteria inactivation in the hibiscus beverage. Another mechanism for micr obial inactivation is the effect caused in the microorganisms cells during the decompression process. This can be the mechanisms by which Y&M were inactivated. 2 = 0.81) and APC (r2Log reduction (Y&M) = 3.8745 + 0. 0155 P + 0.2200 t + 0.0007 P = 0.55) log reduction (LR) as a function of pressure (P) and residence time (RT). Both quadratic models were not statistically significant and were not suitable to predict the inactivation of microorganisms pres ent in the hibiscus beverage within pressures and residence times ranges studied. 2 0.0037 P t + 0.0 042 t2Log reduction (APC) = 0.6804 + 0. 0187 P + 0.4493 t 0. 0 001 P (1) 2 0.0019 P t 0.0 311 t2As can be seen from Table 51, treatment 8 (24.1 MPa, 8 min) showed the highest LR for Y&M followed by treatments 10 (34.5 MPa, 6.5 min) and 11 (34.5 MPa, 8 min). On the other hand, treatment 5 (24.1 MPa, 6.5 min) had the highest log reduction for APC and treatment 10 was among the second highest APC log reduction treatments while treatment 11 was among the lowest APC log reduction treatments. Based on these results, our approach was to select treatment 10 for further DPC D processing experiments. This treatment conditions consists of the upper level pressure within our experimental range studied (34.5 MPa) which will assure a complete solubility of CO (2) 2

PAGE 105

105 during processing and the middle level residence time of 6.5 min which is more feasible for industrial applications than longer times. Microbial Stability during Storage Microbial stability of unprocessed (CONTROL), dense phaseCO2Physicochemical Stability during Storage processed (DPCD), and thermally treated (HTST) hibiscus beverages during storage is presented in Figures 52 and 53. Aerobic plate counts in all three beverages (Figure 52) remained constant between 2 and 3 logs during the 14 weeks of storage. The HTST beverage showed slightly lower counts when compared to the other two beverages. Neither the DPC D nor the HTST treatments reduced the initial bacteria population possibly because it was difficult to observe microbial reductions when starting with a low population In the case of yeast and molds (Figure 53), the DPCD and HTST treatments reduced the i nitial population by around 3 logs and both beverages where very stable since both treatments were effective in inactivating the initial Y&M population and there was no growth during storage. For the CONTROL, a maximum of 5 logs at week 6 was reached and declined afterwards possibly associated with the death stage of the microorganisms. The sensory characteristics of the CONTROL beverage indicated that fermentation was taking place. Overall the DPCD and HTST beverages were microbiologically stable during the 14 weeks of storage favored by the beverage low pH and storage temperature (4 C). Physicochemical changes in the studied hibiscus beverages during storage are shown in Table 52. There were no significant differ ences between treatments (CONTROL, DPCD, HTST) over time for pH and Brix. There was a significant effect of treatment over time for all other parameters measured. Titratable acidity in the DPCD

PAGE 106

106 treated beverage was significantly higher when compared with the CONTROL and HTST beverages which can be due to the presence of residual CO2Color density significantly decreased over time for all three treatments (CONTROL, DPCD, and HTST). This indicates that there is a decline in the absorbance at 520 nm which can be associated with degradation of anthocyanins. At time 14 weeks of storage the HTST beverage showed a si gnificantly lower value o f color density as compared to the CONTROL and DCPD beverages. Moreover the hue tint values (Figure 5 4) significantly increased for all three treatments during storage which also indicates some loss of red color in the samples. remaining in solution in the beverage after depressurization. A similar behavior was observed by Calix and others (2008) in orange and apple juices. P hytochemical Stability during Storage Phytochemical changes during storage for the three hibiscus beverages studied are presented in Tables 52 and Figure 55. Several polyphenolic changes during storage were measured using authentic standards and their co ncentration was expressed in mg/L of beverage (Table 53). This included gallic and caffeic acid, caffeoylquinic acids which were quantified using a chlorogenic acid standard and were identified based on their charact eristic absorption spectrum at max 320 nm, delphini din 3 sambubioside and cyanidin3 sambubioside that are the main anthocyanins present in hibiscus extracts, and flavonols which were quantified using quercetin and identified by their charact eristic absorption spectrum at max There was a significant effect of treatment (CONTROL, DCPD, HTS T) in anthocyanins content total phenolics, antioxidant capacity, gallic acid, caffeic acid, and 360 nm.

PAGE 107

107 flavonols content. Anthocaynins content (Figure 5 6 ) significantly decreased during storage for all t hree treatments. A loss of 11, 9, and 14% in anthocyanins was observed for CONTROL, DPCD, and HTST beverages respectively. At time 14 weeks the concentration of anthocyanins in all three treatments was significantly different, with the CONTROL having the highest and HTST beverage the lowest concentration. There were no major changes in total phenolics and antioxidant capacity during storage for all three treatments. There were some slight differences between storage times possibly related to the breakdown and formation of polyphenolic compounds. A previous study with muscadine grape juice (Del Pozo and others 2006a) showed that losses in anthocyanins during processing and storage were around 78% for a pasteurized juice and only 35% for a DPCD processed juic e. A similar behavior in total phenolic and antioxidant cap acity was also found. The greater losses and differences between treatments in the grape juice as compared to the hibiscus beverage can be attributed to a higher initial concentration of polyphenol ics and higher pH of the grape juice. As shown in Table 53, the concentration of gallic acid increased with increasing storage time for the DPCD beverages and to a greater extent for the CONTROL. Similarly, the presence of caffeic acid in the CONTROL and DPCD beverages at time 14 weeks was detected and can be a breakdown product of the caffeoylquinic acids present in the beverage. Both phenomena could be related to polyphenolic compounds break ing down due to microbial activity. There were no major changes in the caffeoylquinic acids and flavonols content during storage, although at time 14 weeks there was a significantly lower concentration of both polyphenolics in the CONTR OL beverage as compared to the other two treatments. There was a significant but small

PAGE 108

108 decrease in the concentration of delphinidin3 sambubioside and cyanidin3 sambubioside for al l the three beverages during storage. Overall there were no big phytochemical losses during storage for any of the three treatments. This can be attributed to the low pH of the beverage, the low storage temperature and the presence of sucrose in the beverage. A previous study (Tsai and others 2004) has shown that sucrose solutions favored the stability of hibiscus anthocyanins by decreasing the availability of w ater that is needed for the anthocyanins degradation process. Conclusions CO2 solubility in a hibiscus beverage and optimal processing conditions to inactivate microorganisms (Y&M and APC) were determined. DPCD was found to be a viable technology for extending the hibiscus beverage shelf life since it showed to be microbiologically stable during the 14 weeks of refrigerated storage. Quality attributes such as pH and Brix were not affected by DPCD whereas TA increas ed. A loss of only 9% anthoc y a nin s during storage was observed for the DPCD processed hibi scus beverage which was lower when compared to a heat pasteurization process and no major changes in total phenolics content and antioxidant capacity occurred during storage

PAGE 109

109 Figure 51. Schematic di agram of the setup used for the hibiscus beverage thermal treatment (75 C for 15 s). Figure 52 CO2 solubility in water and a hibiscus beverage as a function of pressure measured at 40 C. Data represents the mean of n=3. Values with similar letters within the figure are not significantly different (Tukeys HSD, p > 0.05). e de dc c c de b ab a a 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 5 10 15 20 25 30 35CO2Solubility (g CO2/mL)Pressure (MPa) Hibiscus Water

PAGE 110

110 Table 51. Response surface design used to test the effect of pr essure and residence time on microbial reduction (log10) at 40 C and 8% CO2Run Pressure (Mpa) Residence time (min) Beverage flow rate (g/min) CO2 flow rate (g/min) LR* (APC) LR (Y&M) 1 13.8 5.0 500.0 40.0 0.93 a bc 0.03 5.20 e 0.08 2 13.8 6.5 384.6 30.8 0.99 a bc 0.03 5.63 d 0.06 3 13.8 8.0 312.5 25.0 0.9 2 bc 0.03 5.67 d 0.04 4 24.1 5.0 500.0 40.0 0.88 c 0.05 5.3 0 e 0.05 5 24.1 6.5 384.6 30.8 1.05 a 0.07 5.56 d 0.0 8 6 24.1 6.5 384.6 30.8 0.99 abc 0.09 5.6 2 d 0.09 7 24.1 6.5 384.6 30.8 1.0 4 ab 0.06 5.65 d 0.06 8 24.1 8.0 312.5 25.0 1.02 ab 0.0 4 6.26 a 0.1 1 9 34.5 5.0 500.0 40.0 1.0 4 ab 0.06 5.85 c 0.06 10 34.5 6.5 384.6 30.8 1.03 ab 0.0 2 6.07 b 0.0 1 11 34.5 8.0 312.5 25.0 0.91 bc 0. 08 6.09 b 0.0 6 Y&M = yeasts and molds, APC = aerobic plate count. LR = log10 reduction. Figure 53 Aerobi c plate counts of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 0 1 2 3 4 0 2 4 6 8 10 12 14Log (CFU/mL)Storage time (weeks) CONTROL DPCD HTST

PAGE 111

111 Figure 5 4 Yeast/mold counts of unprocessed (CONTROL) dense phaseCO2 proc essed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). Figure 55 Hue tint values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 0 1 2 3 4 5 6 0 2 4 6 8 10 12 14Log CFU/mLStorage time (weeks) CONTROL DPCD HTST 0.33 0.34 0.35 0.36 0.37 0.38 0.39 0.4 0 2 4 6 8 10 12 14Hue tintStorage time (weeks) CONTROL DPCD HTST

PAGE 112

112 Table 52. Physicochemical and phytochemical changes of unprocessed (CONTROL), dense phaseCO2 processed (DPCD), and thermally treated (HT ST) hibiscus beverages during refrigerated storage at 4 C. Storage time week 0 week 1 week 2 week 3 week 4 week 5 week 6 week 8 week 10 week 12 week 14 pH CONTROL 2.43 b 2.45 a 2.47 a 2.45 b 2.45 c 2.47 b 2.47 b 2.47 a 2.48 a 2.48 b 2.48 a DPCD 2.45 a 2.44 b 2.46 b 2.46 a 2.47 a 2.48 a 2.48 a 2.48 a 2.48 a 2.49 a 2.49 a HTST 2.45 a 2.45 a 2.45 b 2.45 b 2.47 b 2.48 a 2.48 a 2.48 a 2.48 a 2.49 a 2.49 a Brix CONTROL 9.70 ab 9.70 a 9.70 b 9.87 b 9.70 b 9.67 a 9.90 a 10.20 a 10.07 a 9.93 a 9.70 b DPCD 9.77 a 9.70 a 9.90 a 9.93 a 9.87 a 9.70 a 9.63 b 10.10 b 9.83 b 9.57 b 9.77 ab HTST 9.63 b 9.67 a 9.90 a 9.73 c 9.90 a 9.60 b 9.97 a 10.00 c 9.82 b 9.63 b 9.90 a T itratable acidity ( g of malic acid /100 mL ) CONTROL 0.37 b 0.37 b 0.38 b 0.38 b 0.37 b 0.38 b 0.38 b 0.38 b 0.37 ab 0.38 b 0.38 b DPCD 0.40 a 0.41 a 0.40 a 0.41 a 0.41 a 0.40 a 0.41 a 0.41 a 0.40 a 0.41 a 0.40 a HTST 0.37 b 0.36 b 0.38 b 0.37 b 0.38 b 0.38 b 0.38 b 0.38 b 0.38 b 0.37 b 0.38 b Color density CONTROL 1.78 ab 1.78 b 1.75 ab 1.74 a 1.71 b 1.72 a 1.73 a 1.74 a 1.70 a 1.66 a 1.67 a DPCD 1.80 a 1.80 a 1.76 a 1.75 a 1.76 a 1.73 a 1.74 a 1.74 a 1.69 a 1.64 a 1.66 a HTST 1.76 b 1.79 ab 1.73 b 1.75 a 1.72 ab 1.69 a 1.72 a 1.68 b 1.63 b 1.58 b 1.58 b Total phenolics (mg/L) CONTROL 263.86 a 259.01 a 247.97 b 248.38 a 251.13 a 246.43 ab 249.13 a 250.54 a 251.99 a 253.44 a 264.05 a DPCD 259.62 ab 256.86 a 252.26 a 253.22 a 248.52 a 245.03 b 253.91 a 252.32 ab 252.25 a 252.17 a 255.65 b HTST 254.73 b 259.39 a 245.04 b 251.09 a 241.64 b 248.36 a 241.61 b 248.36 b 247.15 b 245.94 b 250.06 c Antioxidant capacity ( mol of TE/mL) CONTROL 5.93 a 6.11 b 6.62 a 5.77 a 6.04 a 5.96 a 6.64 a 6.53 a 6.27 a 6.01 a 5.85 b DPCD 5.66 a 6.10 b 5.43 b 6.19 a 5.90 a 6.08 a 6.31 a 5.59 b 5.76 a 5.92 a 6.80 a HTST 6.38 a 7.00 a 5.84 b 6.34 a 6.42 a 6.53 a 6.11 a 5.27 b 5.81 a 6.35 a 6.92 a Data represents the mean of n=9. Values with similar letters within columns are not significantly different (Tukeys HSD, p > 0.05).

PAGE 113

113 Figure 56 Concentration of anthocyanins of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 40 42 44 46 48 50 52 54 56 0 2 4 6 8 10 12 14Concentration of anthocyanins (mg/L)Storage time (weeks) CONTROL DPCD HTST

PAGE 114

114 Table 53. Po lyphenolics content (mg/L) of unprocessed (CONTROL), dense phaseCO2 processed (DPCD), and thermally treated (HTST) hibiscus beverages during refrigerated storage at 4 C. Storage time week 0 week 2 week 5 w eek 8 week 14 Galllic acid d CONTROL 0.79 a 2.53 a 3.25 a 3.89 a 4.03 a DPCD 0.50 b 0.77 b 1.03 b 1.27 b 1.57 b HTST 0.57 b 0.68 b 0.63 c 0.71 c 0.71 c Caffeoylquinic acids e CONTROL 93.59 a 98.13 a 94.05 a 94.11 a 89.88 b DPCD 91.90 a 96.78 a 95.74 a 95.28 a 94.50 a HTST 92.51 a 98.36 a 93.31 a 95.10 a 94.73 a Caffeic acid f CONTROL nd j nd nd nd 5.73 a DPCD nd nd nd nd 4.26 b HTST nd nd nd nd nd Delphinidin 3 sambubioside g CONTROL 65.26 a 64.64 a 61.96 ab 62.14 a 58 .07 a b DPCD 63.12 b 61.75 b 62.53 a 61.15 ab 59.23 a HTST 64.79 ab 62.20 ab 59.12 b 58.72 b 56.7 8 b Cyanidin 3 sambubioside h CONTROL 29.98 a 29.60 a 28.23 a 28.10 a 26.61 a DPCD 28.78 a 28.46 a 28.49 a 27.95 a 27.43 a HTST 29.99 a 29.03 a 27.76 a 27.74 a 27.40 a Flavonols i CONTROL 23.05 b 23.55 a 22.44 b 22.79 b 22.65 b DPCD 25.22 a 24.96 a 25.02 a 25.18 a 24.23 ab HTST 25.08 a 25.15 a 24.64 a 25.45 a 24.78 a Data represents the mean of n=6. Values with similar letters within columns of each polyphneolic category are not significantly different (Tukeys HSD, p > 0.05). d,e,f,g,h,i Quantified with gallic acid, chlorogenic acid, de lphinidin 3 glucoside, cyanidin3 glucoside, and quercetin standards respectively. j Abbreviations: nd, not detected.

PAGE 115

115 CHAPTER 6 PROCESSING HIBISCUS BEVERAGE USING DENSE PHASE CARBON DIOXIDE: SENSORY ATRIBUTES AN D AROMA COMPOUNDS STABILITY Introduction Hibiscus sabdariffa (family Malvaceae) is a short day annual shrub that grows in many tropical and subtropical countries and is kno w n by different synonyms and vernacular names such as roselle in the U.S. and England, loiselle in France, jamaica or flor de j amaica in Mexico and Spain, karkade in Sudan and Arabia, sorrel in the Caribbean and byssap in Senegal (Morton 1987; Stephens 2003). Traditionally fresh hibiscus calyces are harvested by hand and are either frozen or dried, in the sun or artificially for preservation. They are typically sold into the herbal tea and beverage industry or in local and regional markets where they are used in the preparation of beverages and color and flavor extracts (Plotto 1990). Studies with human patients have shown that the regular consumption of hibiscus extract has an antihypertensive effect (Haji Faraji and Haji Tarkhani 1999; HerreraArellano and others 2004) and reduces serum cholesterol in men and women (Lin and others 2007). The preparation of a hibiscus beverage includes an extraction step followed by a pasteurization method. Although thermal preservation of foods is effective in reducing microbial loads it can also lead to organoleptic and nutritional changes. Nonthermal processes are an alternative which may help preserve the color, flavor, and nutrients of food. Dense phase carbon dioxide (DPCD) is a cold pasteurization method that uses pressures below 90 MPa in combination with carbon dioxide (CO2) to inactivate microorganisms. This nonthermal technology is mainly used in liquid foods and since the food is not exposed to the adverse effect of heat, its freshl ike physical, nutritional, and sensory qualities are maintained.

PAGE 116

116 Previous studies have shown that DPCD processed beverages keep their freshlike characteristics after processing and storage. Likeability of DPCDtreated coconut water was similar to untreate d samples while h eat treated samples were less appealing (Damar and others 2009). Similarly, no differences in sensory attributes (color, flavor, aroma, and overall likeability) were observ ed between unprocessed and DPCD muscadine grape juices but there were differences when compared to a heat pasteurized juice (Del Pozo Insfran and others 2006a). The objectives of this study were (1) to determine the effect of DPCD processing on the sensory att ributes and aroma compounds of hibiscus beverage when compared to a thermally treated and a control (untreated) and (2) to monitor the changes in these attributes during refrigerated storage. Materials and Methods Beverage P reparation Dried Hibiscus sabdariffa (cv. C riollo) (moisture content of 9%) obtained from Puebla, Mexico was mixed with water (1:40 w/v) using a 200 L stainless steel mixing tan k Model UAMS (Cherry Burrell, I owa U.S.A.) and maintained at 25C for 1 h. Mixing was applied intermittently by alternating intervals of 10 min mixing and 10 mi n rest. The extract was then filtered using four layers of cheesecloth. A beverage was prepared by adding sucrose to a concentration of 100 g sucrose/L of extract and then was placed in 3 gallon sealable buckets and refrigerated before processing. P rocessi ng and Storage C onditions Fresh prepared hibiscus beverage was divided into three parts. One part was kept as CONTROL and didnt receive any treatment; the second part was proc essed using DPCD at 34.5 MPa 8% CO2, 6.5 min, and 40 C while the third part was pasteurized at

PAGE 117

117 75 C for 15 s (HTST). The DPCD processing conditions were confirmed to achieve >5 log reduction of yeast s/molds according to previous experiments. Both the control and treated samples were stored in 1 L glass jars. Physicochemical, sensor y and aroma compound analysis were done at weeks 0 and 5 of refrigerated storage at 4 C .Color analysis was performed at weeks 0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 of storage. Dense Phase CO2The DPCD equipment located at the Univer sity of Fl orida (Gainesville, Fl a ., U.S.A.) was constructed by APV (Chicago, Il l ., U.S.A.) for Praxair (Chicago, Il l ., U.S.A.). It is a continuous flow equipment in which CO Equipment 2 and the hibiscus beverage were pumped through the syst em and mixed before entering a high pressure pump. Processing pressure was controlled by this pump while the desired temperature was maintained in the holding coil (79.2 m, 0.635 cm i.d.). Turbulent flow and mixing were reached at the entrance of the coil by passing the mixture through a static mixer and a small diameter tube (length of about 180 cm). Residence time was adjusted by setting the flow rate of the mixture. An expansion valve was used at the end of the process to release the CO2Thermal Processing Conditions from the mixture and the beverage was collected into 1 L sterile bottles as previously described by Damar and others (2009). For thermal processing, the hibiscus beverage was pumped by a peristaltic pump (Col e Parmer, Chicago Il l ., U.S.A.) through two stainless steel tube secti ons (3.2 m, 0.457 cm i.d. ea.) placed inside a temperature controlled water bath ( Precision Scientific, Chicago Il l ., U.S.A.). In the first section the beverage was heated to 75 C (temperature was measured using a thermocouple) and then entered the second section where it was held at 75 C for 15 s. The beverage was then passed through a

PAGE 118

118 cooling stainless steel tube ( 5.2 m, 0.457 cm i.d.) in a water/ice bath and chilled to ~15 C before it was collected into 1 L sterile glass jars. Platinum cured silicone tubing (0.635 cm i.d.; Nalgene, Rochester N Y U.S.A.) was used to connect the pump to the stainless steel heating, holding, and cooling sections. A schematic diagram of the setup used for the hibiscus beverage pasteurization is presented in Figure 52. Physicochemical Analysis pH and Brix were measured using a pH meter E A920 (Orion Research; Boston, Ma ss., U.S.A.) and a ABBE Mark II refractometer (Leica Inc.; Buffalo, N Y U.S.A.). A Brinkmann Instrument (Brinkmann Instruments Co., Westbury, N Y U.S .A.) consisting of a Metrohm 655 Disomat, Metrohm 614 Impulsomat, and Metrohm 632 pH meter was used to measure titratable acidity (TA Samples of 10 mL were used and TA was determined by titration with 0.1 N NaOH until pH 8.1 and expressed as % malic acid ( g/100 mL). Sensory Evaluation Flavor and overall likeability of fresh and processed hibiscus beverages were compared using a difference from cont rol test at weeks 0 and 5 of storage. A randomized complete block design was used, and differences from control values were recorded on a line scale with anchors at 0 and 10 that represented no difference to extremely different in flavor. Panelists compared the flavor of the reference (fresh/unprocessed beverage (CONTROL)) with that of a hidden reference (fresh beverage (CONTROL)), the thermally (HTST), and the DPCD processed beverages. A 9 point hedonic scale was also conducted in order to compare the overall likeability of fresh (hidden reference) and processed hibiscus beverages. For the taste panel at week 5 the reference was fresh hibiscus beverage that was kept frozen at 20 C.

PAGE 119

119 Before sensory analysis all beverages (fresh, thermally, and DPCD processed) were degassed in order to have equal carbonation levels by placing them in 2 L sterile glass bottles on a stir plate with continuous stirring and vacuum ( 15 Hg) was pulled for 20 min using a Gast vacuum pump ( Model DOA P104 AA; Beonton Harbor, Mi ch ., U.S.A.) All samples were chilled and kept in ice at a temperature of ~4C before serving. They were then served on a tray in numbered plastic cups containing ~30 mL of sample. A cup of deionized water and non salted crackers were also provided to the panelists to cleanse their palate between evaluations. Both sensory tests were performed at the University of Florida taste panel facility using 75 untrained panelists in each test. Headspace Volatiles Sampling Headspace volatiles were extracted and concentrated using Solid Phase Micro Extraction (SPME) technique. Ten milliliters of hibiscus beverage were added to a 22 mL screw cap amber glass vial PTFE/silicone septa containing a small stir bar. Samples were equilibrated for 20 min in a water bath at 40 C and afterwards hibiscus headspace volatiles were extracted for 30 min using a 1 cm 50/30 mm DVB /Carboxen/PDMS SPM E fiber (Supelco, Bellefonte, Pa., U.S.A. ). Before each exposure th e fiber was cleaned for 5 min a t 200 C in the GC MS injection port. GC MS Analysis GC MS analysis was conducted using a HP 6890 GC coupled with a MSD 5973 (Agi lent Technologies; Palo Alto, Ca lif ., U.S.A. ) Hibiscus volatiles from the SPME fiber were desorbed into the GC injection port (splitless mode) at 200 C. A SPME injector liner (SPME injection sleeve, 0.75 mm i.d., Supleco; Bellefonte, Pa., U.S.A.) was used.

PAGE 120

120 The fibe r was removed after 5 min exposure in the injection port. Volatiles were separated on both a DB 5 (30 m x 0.32 mm. i.d. x 0.5 m J&W Scientific; Folsom, Ca lif ., U.S.A.) and a DB Wax (30 m x 0.32 mm. i.d. x 0.5 m, Restek; Bellefonte, Pa., U.S.A.) column The oven was programmed from 35 C (held for 5 min) to 250 C at 6 C /min with a final hold of 10 min. Helium was used as the carrier gas at 1.67 mL/ min The mass spectrometer was operated in the total ion chromatogram (TIC) at 70 eV. Data were collected from 35 m/z to 400 m/z All samples were run in duplicate in each column. Chromatograms were recorded and integrated using Enhanced Chemstation (version 01.00) software (Agi lent Technologies; Palo Alto, Calif ., U.S.A.). Mass spectral matches were made by comparison with NIST 98.1 (NIST; Gaithersburg, Md., U.S.A. ) and WILEY 8.1 (Wiley; New York, N Y U.S.A.) mass spectral libraries. Only those compounds with spectral fit values equal to or greater than 850 were considered positive identifications. L inear re tention index values were determined for both columns using a series of alkanes (C5C25) run under identical conditions. Identification Procedures Identifications were based on the combined matching of retention indices (LRI values) from DB 5 and DB Wax columns, matches made from spectra in the NIST and WILEY libraries and linear retention index matches from literature. Color Analysis Color was measured using a ColorQuest XE colo rimeter (HunterLab, Reston, Va., U.S.A.). Sampl es of 40 mL where placed in a 20 mm cell and L*, a*, and b* parameters were recor ded in total transmittance mode, illuminant D65, 10 observer angle. Chroma (a*2 + b*2)1/2 and hue angle (arctan b*/a*) were calculated from the

PAGE 121

121 measured a* and b* Color difference ( E ) values were also calculated using the following formula: values. C hroma provides a measure of color intensity, while hue angle (0 = redpurple, 90 = yellow, 180 = bluishgreen, 270 = blue) indicates the sample color itself (McGuire 1992). 2 0 2 0 2 0 *) ( ) ( ) ( b b a a L L E (1) Where L0, a0, and b0 are the reference values at storage time 0 week for each of the treatments (CONTROL, DPCD, and HTST) and L*, a*, and b*Statistical Analysis are the values at time t = 1,2,3, weeks of storage. Sensory data was recorded and analyzed using Compusense five (Compusense, Guelph, Ontario, Canada). Analysis of variance (ANOVA) and mean comparisons using t test and Tukeys test were conducted at the 5% significance level. Repeated measures ANOVA and mea evaluate the effect of treatment (fresh (CONTROL), thermal (HTST), and DPCD processed) and storage time.(0, 1, 2, 3, 4, 5, 6, 8, 10, 12, and 14 weeks) on color parameters using SAS statistical softw are (SAS Institute Inc., Cary, N C U.S.A.). Results and Discussion Physicochemical Analysis The measured pH, Brix, and TA for the CONTROL, DPCD, and HTST beverages at weeks 0 and 5 of refrigerated storage are presented in Table 61. It is important t o notice that the CONTROL at week 5 was kept frozen in order to be used as reference in the second taste panel. No significant differences in pH were observed between the DPCD and HTST beverages at both week 0 and 5. Significant differenc es

PAGE 122

122 in TA were found between DPCD and the CONTROL and HTST beverages at both storage times. A higher TA in the DPCD beverage can be a result of residual CO2Sensory Evaluation in the beverage. For the same reason, a significantly higher Brix value was found in the DPCD beverage when compared with HTST. This higher value can also be a result of residual carbonic acid. An increase in TA of DPCD treated coconut water (Damar and others 2009) and orange juice (Kincal and others 2006) was also observed in previous DPCD studies. Two taste panels were conducted during storage. In the first test (week 0), 55% of panelists were females, 91% of males and 83% of females were in the 1830 age range, while for the second panel (week 5), 48% of panelist were females, 74% of males and 94% of females were in the 1830 age range. For the taste panel at week 0, there were no significant differences between the CO NTROL (hidden reference) and HTST beverages. However, significant differences were detected by paneli st s between the CONTROL and DPCD beverage s. The ranking for overall likeability for the three tested beverages were not si gnificantly different (Table 6 2 ) which indicates that regardless of treatment panelists preference remained the same. For the taste panel at week 5 there were not significant differences between the CO NTROL (hidden reference) and HTST beverages. However, the DPCD was rated as significantly different from the other two by panelists. Similarly, the ranking for overall likeability showed no significant differences b etween the hidden reference and the HTST beverage but the DPCD beverage was ranked significantly lower than the other two beverages as shown i n Table 62. Previous studies on muscadine grape juice and coconut water showed greater differences in flavor and overall likea bility between the

PAGE 123

123 DPCD and thermally treated samples. They also found that the DPCD sample was very similar to the hidden r eference (Damar and others 2009; Del Pozo and others 2006a). This may indicate that thermal processing affected more the organoleptic characteristics of the grape juice and coconut water than in the hibiscus beverage. In both taste panels the DPCD beverage was different from the hidden reference and at week 5 it was also significantly different from the HTST beverage. However, despite these differences panelist ranking for overall likeability showed no significant differences in the taste panel at week 0 and overall likeability values where even higher in the taste panel at week 5. DPCD differences from the other two beverages can be attributed to two possible factors. First, even when the DPCD beverage was degasified before sensory analysis there could still be residual CO2 remaining which would result in a carbonated beverage mouth feel which may have also affected the a cidity/sweetness balance causing the panelists to perceive the beverage as less sweet This can be confirmed by a higher TA in the DPCD beverage as described previously. Second, flavor compounds in the beverage will have dissolved in the residual CO2Aroma Compounds and l ost during degasification. Stability of aroma compounds were monitored during weeks 0 and 5 of storage. The chemical composition of hibiscus beverages headspace volatiles are presented in Figure 61. Total peak areas for all analyzed beverages were normalized to the total peak area of CW0 (CONTROL week 0). Alcohols and aldehydes constituted the major fractions of hibiscus beverages aroma volatiles. A total of 4 aldehydes, 6 alcohols, 2 ketones, and 1 acid were considered for analysis. A sl ight decrease (21%) in alcohols and ketones was observed in the DPCD beverage after processing (week 0) while there

PAGE 124

124 was considerable loss (88%) in all volatiles in the thermally treated (HTST) beverage which may indicate that there was decomposition of com pounds due to heating and possible formation of other compounds that were not studied in this experiment or that volatiles were lost by evaporation during pasteurization. At week 5 of storage, there was a considerable loss of all volatiles in the DPCD beverage (70%), a slight decrease in alcohols and ketones (17%) for the CONTROL, and no changes in the volatiles for the HTST beverage. The loss of DPCD volatiles at week 5 can be a result of decomposition or formation of compounds or possibly they could be lost during degasification as discussed previously. A previous study with melon juice also showed t hat DPCD retained more volatile compounds as compared to the pasteurized juice (Chen and others 2009). The volatile compounds used in this analysis were identi fied by GC MS and are presented in Table 63. To compare the volatiles in the hibiscus beverages at weeks 0 and 5 of storage, peak areas were normalized (100) to the largest peak (1Octen 3 ol) in the CONTROL (C) week 0 sample. A total of 13 compounds were considered for analysis, 6 of which (dehydroxylinalool oxide b, octanal, 1hexanol, acetic acid, furfural, and decanal) have been previously identified in hibiscus extracts (Gonzalez Palomares and others 2009; Chen and others 1998). The compounds present in highest concentration in all six samples were 1octen 3 ol, decanal, octanal, 1hexanol, and nonanal. A mushroom like, fruity, citrus, and fruity aroma were associated with 1octen3 ol, octanal, and nonanal respectively in lychee (Mahattanatawee and ot hers 2007), decanal was described as sweet, waxy, orange in merlot and cabernet wines (Gurbuz

PAGE 125

125 and others 2006) while 1hexanol was related to green, sweet notes in both lychee and red wines. Color Analysis Color stability was assessed over 14 weeks of refr igerated storage. Treatment (CONTROL, DPCD, HTST) had a significant effect ( p < 0.0001) over storage time for all the color parameters measured and calculated (L*, a*, b*, and hue angle, chroma, and E). There were slight changes for the L* values in al l three treatments over storage but the trend was almost constant (Figure 62). At time 14 weeks of storage, the DPCD beverage had a significantly lower L* value as compared with HTST and CONTROL. As can be seen in Figure 63, the a* values decreased slight ly over time for all the treatments and after 14 weeks of storage there were no significant differences between the DPCD and HTST beverages. The parame ter that showed the most change was b* (Figure 6 4). In the three treatments there was a slight but significant decrease of the b* values with time. There were significant differences between treatments at time 14 with the CONTROL having the highest and the HTST beverage the lowest b* value. Hue angle slightly decreased over storage time for the CONTROL, DPC D, and HTST beverages (Figure 65). These changes indicate that the beverages will follow a colo r degradation pattern from a bright red color to a redpurple color. At time 14 weeks there were no significant differences in hue angle between the CONTROL and DPCD beverages, the HTST sample was significantly lower. In the same way, chroma decreased (the beverages became less intense in color) for all three samples (Figure 66) and at time 14 weeks they were significantly different with the CONTROL having the hi ghest and HTST the lowest value. A decrease in hue angle and chroma during

PAGE 126

126 storage (25 C) was previously reported in a radish anth oc yanin extract (Giusti and Wrolstad 1996). The calculated E value showed a significant increase over storage time. Since the b*Conclusions value was the parameter that contributed the most to this difference in color, the HTST beverage showed a significantly higher value after 14 weeks of storage followed by the DPCD and CONTROL beverages as sown in Figure 6 7 Changes in color during storage can be attributed to the degradation of anthocyanins which are the pigments responsible for the red color in the hibiscus beverages with the HTST beverage showing a higher change in color when compared to the DPCD beverage. Changes in hibiscus aroma volatiles during storage did not affect panelists overall likeability of the product. DPCD was found to be a viable technology for processing hibiscus beverages since it maintained its characteristic red color over 14 weeks of storage and retained more aroma volatiles than the heat pasteurized beverage. Possible losses of aroma volatiles during the degasification process can be prevented by recovering them and adding them back to the beverage. Further studies are needed to better understand t he chemistry of hibiscus aroma compounds and to reduce their loss during storage.

PAGE 127

127 Table 61. Measured pH, Brix and titratable acidity (TA) (g of malic acid/100 mL of beverage) at weeks 0 and 5 of refrigerated storage ( 4 C ) week 0 week 5 pH Brix TA pH Brix TA CONTROL 2.43 b 9.70 ab 0.37 b 2.47 b 9.67 a 0.38 b DPCD 2.45 a 9.77 a 0.40 a 2.48 a 9.70 a 0.41 a HTST 2.45 a 9.63 b 0.37 b 2.48 a 9.60 b 0.38 b Data represents the mean of n=9. Values with similar letters within columns are not significantly different (Tukeys HSD, p > 0.05). Table 62 Difference in flavor and overall likeability between fresh (reference and hidden reference), dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverages detected by untrained panelists (n = 75) at weeks 0 and 5 of refrigerated storage (4 C) Week 0 Week 5 Difference in flavor Overall likeability Difference in flavor Overall likeability Hidden reference 2.78 b ** 5.23 a 3.06 b 5.89 a D PCD 3.75 a 5.01 a 5.16 a 5.27 b HTST 3.34 ab 5.23 a 3.63 b 5.93 a Difference observed when compared to given reference (difference from control test). ** Values with similar letters within columns are not significantly different (Tukeys HSD, p > 0.05).

PAGE 128

128 Figure 61 Chemical c omposition of hibiscus beverage headspace volatiles during storage. Total number of compounds for each class is put in parenthesis. All six samples were normalized to total peak area of the sample CW0 (CONTROL week 0 ). C = CONTROL, D = DPCD, H = HTST, W = week. 0 20 40 60 80 100 120 CW0 DW0 HW0 CW0F DW5 HW5Peak Area PercentHibiscus Beverage Samples Aldehydes (4) Alcohols (6) Ketones (2) Acids (1)

PAGE 129

129 Table 63 MS iden tification of hibiscus beverage volatiles during storage. Peak areas were normalized (100) to the largest peak (1 Octen 3 ol) in the CONTROL ( C ) week 0 sample Normalized peak are as (%) # Name CAS # LRI week 0 week 5 DB Wax DB5 C D H C D H 1 Dehydroxylinalool oxide a 13679 86 2 1210 993 48.9 14.4 5.2 29.7 5.96 5.1 2 Dehydroxylinalool oxide b a 13679 86 2 1246 1007 43.9 11.6 4.9 19.8 5.79 4.3 3 3 Octanone 106 68 3 1264 16.3 9.5 2.1 11.9 1.24 2.1 4 Octanal a 124 13 0 1299 1002 83.5 49.6 8.5 70.9 8.52 8.8 5 6 methyl 5 Hepten 2 one 110 93 0 1355 989 45.5 36.7 5.5 38.7 8.13 6.3 6 1 Hexanol a 111 27 3 1373 873 80.6 62.5 9.0 72.9 11.97 9.5 7 Nonanal 124 19 6 1405 1100 67.2 66.6 6.4 54.9 9.69 7.3 8 1 Octen 3 ol 3391 86 4 1468 983 100.0 78.8 11.7 81.6 44.10 11.9 9 Acetic acid a 64 19 7 1485 35.4 29.6 3.6 24.5 7.66 3.2 10 Furfural a,b 98 01 1 1496 832 15.9 26.0 3.4 7.5 7.50 2.9 11 Decanal b 112 31 2 1513 1204 99.5 123.5 15.9 111.0 43.80 14.0 12 1 Octanol 111 87 5 1577 1071 30.9 20.6 3.6 28.8 8.19 3.7 13 1 Nonanol 143 08 8 1678 1172 45.2 30.3 4.9 39.3 7.88 4.9 Total normalized peak area 712.7 559.7 84.6 591.3 170.4 84.0 C = CONTROL, D = DPCD, H = HTST a Compounds previously reported in H. sabdariffa by Chen and others (1998). b Compounds previously reported in H. sabdariffa by Gonzalez Palomares and others (2009).

PAGE 130

130 Figure 62 L* values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). Figure 63 a values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 42 43 44 45 46 47 48 0 2 4 6 8 10 12 14L*Storage time (weeks) CONTROL DPCD HTST 65 66 67 68 69 70 0 2 4 6 8 10 12 14a*Storage time (weeks) CONTROL DPCD HTST

PAGE 131

131 Figure 64 b* values of unprocessed (CONTROL) dense phaseCO2 pro cessed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). Figure 65 Hue angle values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 59 60 61 62 63 64 65 66 67 68 0 2 4 6 8 10 12 14b*Storage time (weeks) CONTROL DPCD HTST 42 42.5 43 43.5 44 44.5 45 0 2 4 6 8 10 12 14Hue angle (degrees)Storage time (weeks) CONTROL DPCD HTST

PAGE 132

132 Figure 66 Chroma values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated ( HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). Figure 67 E values of unprocessed (CONTROL) dense phaseCO2 processed (DPCD; 34.5 MPa, 8% CO2 6 .5 min, 40 C) and thermally treated (HTST; 75 C, 15 s) hibiscus beverage during refrigerated storage (4 C). 89 90 91 92 93 94 95 96 0 2 4 6 8 10 12 14ChromaStorage time (weeks) CONTROL DPCD HTST 0 1 2 3 4 5 6 7 0 2 4 6 8 10 12 14 E Storage time (weeks) CONTROL DPCD HTST

PAGE 133

133 CHAPTER 7 SUMMARY AND CONCLUSI ONS Findings of this research can provide more flexibility to hibiscus processing. Extraction and process selection for industrial applications should consider availability of raw material (fresh or dried hibiscus), final product quality and phytochemical characteristics, and economic considerations Equivalent cold and hot water conditions (240 min at 25 C and 16 min at 90 C) were found for anthocyanins extraction of dried hibiscus S imilar polyphenolic profiles and chemical composition of aroma compounds were observed between fresh and dried hibiscus extracts although differences in concentration were found. Fifteen aroma compounds were identified for the first time. In general, hibiscus aroma is a combination of earthy, green, floral, and fruity notes but the final flavor profile is affected by the preservation and extraction process. S olubility of CO2 in a hibiscus beverage (5.06 g CO2lmL at 31.0 MPa) and optimal processing conditions to inactivate microorganisms (34.5 MPa and 6.5 mi n for a Y&M log reduction of 6.1) were determined. DPCD was found to be a viable technology for processing hibisc us beverages since it extended its shelf life and maintained the characteristic red color for 14 weeks of refrigerated storage. Quality attr ibutes such as pH an Brix were not affected by DPCD whereas TA increased. A loss of only 9% of ant hocyanins during storage was observed in the DPCD processed hibiscus beverage which was lower as compared to a heat pasteurization process and no major changes in total phenolics content and antioxidant c apacity occurred during storage. Changes in hibiscus aroma volatiles during storage did not affect panelists over all likeability of the product.

PAGE 134

134 APPENDIX A EXTRACTION EXPERIMEN T STATISTICAL ANALYSIS Table A 1. SAS software output of statistical analysis for the anthocyanins concentration data (AC) perfum ed in the hibiscus extraction experiment (Chapter 3). _____________________________________________________________________ EXTRACTION The GLM Procedure Class Level Information Class Levels Values Treatment 8 25 120 25240 2530 2560 9016 902 904 908 Number of observations 72 Dependent Variable: A C Sum of Source DF Squares Mean Square F Value Pr > F Model 7 13655.68417 1950.81202 78.77 <.0001 Error 64 1585.00755 24.76574 Corrected Total 71 15240.69172 R Square Coeff Var Root MSE A C Mean 0.896002 8.748260 4.976519 56.88582 Source DF Type I SS Mean Square F Value Pr > F Treatment 7 13655.68417 1950.81202 78.77 <.0001 Source DF Type III SS Mean Square F Value Pr > F Treatment 7 13655.68417 1950.81202 78.77 <.0001 Tukey's Studentized Range (HSD) Test for A C NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0.05 Error Degrees of Freedom 64 Error Mean Square 24.76574 Critical Value of Studentized Range 4.43126 Minimum Signi ficant Difference 7.3507 Means with the same letter are not significantly different. Tukey Grouping Mean N Treatment A 77.464 9 9016 A 70.885 9 25240 B 63.211 9 908 C B 58.159 9 25120 C 55.436 9 904 C D 53.197 9 2560 D 47.160 9 902 E 29.574 9 2530 ------------------------------------------Temperature=25 -------------------------------------The GLM Procedure Class Level Information Class Levels Values time 4 30 60 120 240

PAGE 135

135 Number of observations 36 Dependent Variable: AC Sum of Source DF Squares Mean Square F Value Pr > F Model 3 8057.681364 2685.893788 112.12 <.0001 Error 32 766.575146 23.955473 Corrected Total 35 8824.256510 R Square Coeff Var Root MSE A C Mean 0.913129 9.242845 4.894433 52.95375 Source DF Type I SS Mean Square F Value Pr > F time 3 8057.681364 2685.893788 112.12 <.0001 Source DF Type III SS Mean Square F Value Pr > F time 3 8057.681364 2685.893788 112.12 <.0001 Tukey's Studentized Range (HSD) Test for AC NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0.05 Error Degrees of Freedom 32 Error Mean Square 23.95547 Critical Value of Studentized Range 3.83162 Minimum Significant Difference 6.2512 Means with the same letter are not significantly different. Tukey Groupi ng Mean N time A 70.885 9 240 B 58.159 9 120 B 53.197 9 60 C 29.574 9 30 ------------------------------------------Temperature=90 -------------------------------------The GLM Procedure Dependent Variable: AC Sum of Source DF Squares Mean Square F Value Pr > F Model 3 4484.797153 1494.932384 58.45 <.0001 Error 32 818.432403 25.576013 Corrected Total 35 5303.229555 R Square Coeff Var Root MSE A C Mean 0.845673 8.315437 5.057273 60.81789 Source DF Type I SS Mean Square F Value Pr > F time 3 4484.797153 1494.932384 58.45 <.0001 Source DF Type III SS Mean Square F Value Pr > F time 3 4484.797153 1494.932384 58.45 <.0001

PAGE 136

136 Tukey's Studentized Range (HSD) Test for A C NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0.05 Error Degrees of Freedom 32 Error Mean Square 25.57601 Critical Value of Studentized Range 3.83162 Minimum Significant Difference 6.4592 Means with the same letter are not significantly different. Tukey Groupi ng Mean N time A 77.464 9 16 B 63.211 9 8 C 55.436 9 4 D 47.160 9 2 _________________________________________________________________________________________________

PAGE 137

137 APPENDIX B STORAGE EXPERIMENT S TATISTICAL ANALYSIS Table B 1 SAS software output of statistical analysis for the anthocyanins concentration data (AC) perfumed in the hibiscus storage experiment (Chapter 5). _____________________________________________________________________ STORAGE The GLM Procedure Class Level Information Class Levels Values time 11 0 1 2 3 4 5 6 8 10 12 14 Treatment 3 C D T Number of observations 297 Dependent Variable: AC Sum of Source DF Squares Mean Square F Value Pr > F Model 2 56.9747574 28.4873787 9.43 0.0001 Error 294 888.2431136 3.0212351 Corrected Total 296 945.2178710 R Square Coeff Var Root MSE AC Mean 0.060277 4.330435 1.738170 40.13847 Source DF Type I SS Mean Square F Value Pr > F Treatment 2 56.97475742 28.48737871 9.43 0.0001 Source DF Type III SS Mean Square F Value Pr > F Treatment 2 56.97475742 28.48737871 9.43 0.0001 Repeated Measures Analysis of Variance Repeated Measures Level Information Level of time 1 2 3 4 5 6 7 8 9 10 MANOVA Test Criteria and Exact F Statistics for the Hypothesis of no time Effect H = Type III SSCP Matrix for time E = Error SSCP Matrix S=1 M=3.5 N=142 Statistic Value F Value Num DF Den DF Pr > F Wilks' Lambda 0.00000 6839662 9 286 <.0001 Pillai's Trace 1.00000 6839662 9 286 <.0001 HotellingLawley Trace 215234.11193 6839662 9 286 <.0001 Roy's Greatest Root 215234.11193 6839662 9 286 <.0001 MANOVA Test Criteria and F Approximations for the Hypothesis of no time*Treatment Effect H = Type III SSCP Matrix for time*Treatment E = Error SSCP Matrix S=2 M=3 N=142 Statistic Value F Value Num DF Den DF Pr > F

PAGE 138

138 Wilks' Lambda 0.37623085 20.03 18 572 <.0001 Pillai's Trace 0.74964003 19.12 18 574 <.0001 HotellingLawley Trace 1.32338497 20.97 18 472.9 <.0001 Roy's Greatest Root 0.98306355 31.35 9 287 <.0001 NOTE: F Statistic for Roy's Greatest Root is an upper bound. NOTE: F Statistic for Wilks' Lambda is exact. Tests of Hypotheses for Between Subjects Effects Source DF Type III SS Mean Square F Value Pr > F Treatment 2 237.565234 118.782617 23.57 <.0001 Error 294 1481.873250 5.040385 Univariate Tests of Hypotheses for Within Subject Effects Adj Pr > F Source DF Type III SS Mean Square F Value Pr > F G G H F time 9 15393156.09 1710350.68 384657 <.0001 <.0001 <.0001 time*Treatment 18 1280.52 71.14 16.00 <.0001 <.0001 <.0001 Error(time) 2646 11765.26 4.45 GreenhouseGeisser Epsilon 0.1367 HuynhFeldt Epsilon 0.1379 ----------------------------------------------time=0 -----------------------------------------The GLM Procedure Class Level Information Class Levels Values Treatment 3 C D T Number of observations 27 Dependent Variable: AC Sum of Source DF Squares Mean Square F Value Pr > F Model 2 12.57264263 6.28632131 8.18 0.0020 Error 24 18.44201807 0.76841742 Corrected Total 26 31.01466070 R Square Coeff Var Root MSE A C Mean 0.405377 2.048539 0.876594 42.79120 Source DF Type I SS Mean Square F Value Pr > F Treatment 2 12.57264263 6.28632131 8.18 0.0020 Source DF Type III SS Mean Square F Value Pr > F Treatment 2 12.57264263 6.28632131 8.18 0.0020 Tukey's Studentized Range (HSD) Test for AC NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0.05 Error Degrees of Freedom 24 Error Mean Square 0.768417 Critical Value of Studentized Range 3.53170 Minimum Significant Difference 1.032

PAGE 139

139 Means with the same letter are not significantly different. Tukey Grouping Mean N Treatment A 43.7506 9 C B 42.4015 9 D B 42.2215 9 T ---------------------------------------------time=14 -----------------------------------------The GLM Procedure Dependent Variable: AC Sum of Source DF Squares Mean Square F Value Pr > F Model 2 56.96992981 28.48496491 58.59 <.0001 Error 24 11.66832881 0.48618037 Corrected Total 26 68.63825862 R Square Coeff Var Root MSE AC Mean 0.830003 1.824962 0.697266 38.20717 Source DF Type I SS Mean Square F Value Pr > F Treatment 2 56.96992981 28.48496491 58.59 <.0001 Source DF Type III SS Mean Square F Value Pr > F Treatment 2 56.96992981 28.48496491 58.59 <.0001 Tukey's Studentized Range (HSD) Test for AC NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. A lpha 0.05 Error Degrees of Freedom 24 Error Mean Square 0.48618 Critical Value of Studentized Range 3.53170 Minimum Significant Difference 0.8208 Means with the same letter are not significantly different. Tukey Grouping Mean N Treatment A 39.7695 9 C B 38.5811 9 D C 36.2709 9 T _________________________________________________________________________________________________

PAGE 140

140 APPENDIX C HIBISCUS SABDARIFFA PICTURES A B C D E F G H I Figure B 1. Pictures of dried hibiscus (A), dried hibiscus extraction process (B), hibiscus beverage (C), hibiscus beverage in the dense phase carbon dioxide (DPCD) feed tank (D), DPCD processing equipment (E), DPCD processed hibiscus bev erage (F), hibiscus beverage samples for analysis (G), hibiscus beverage under refrigerated storage (H) and DPCD processed hibiscus beverage after 14 weeks of storage at 4 C (I). Photos by Milena Ramirez.

PAGE 141

141 LIST OF REFERENCES Al Kahtani HA & Hassan BH. 1990. Spray drying of roselle ( Hibiscus sabdariffa L.) extract. J Food Sci 55(4):10731076. Al Wandawi H, Al Shaikhly K & Abdul Rahman M. 1984. Roselle seeds: a new protein source. J Agric Food Chem 32(3):510512. Andrade I & Flores H. 2004. Optimization of spray drying of roselle extract ( Hibiscus sabdariffa L.). Proceedings of the 14th International Drying Symposium (IDS 2004) So Paulo, Brazil, 2225 August 2004. p. 597604. Anon 2006. Is ther something brewing in the world of hibiscus?. New nutrition business 12(1):1922. Ashurst PR. 2005. Chemistry and technology of soft drinks and fruit juices. 2nd ed. Oxford UK: Blackwell Publishing. 374 p. Balaban MO, Ferrentino G, Ramirez M & Plaza ML. 2008. Review of dense phase carbon dioxide applications to citrus juices and new development. Proceding of the 54th Citrus Engineering Conference. Lake Alfred, Fla. Beristain CI, Mendoza RE, Garcia HS & Vazquez A. 1994. Cocrystallization of jamaica ( Hibiscus sabdarifa L.) granules. Lebensm Wiss u Technol 27(4):347 349. Bolade MK, Oluwalana IB & Ojo O. 2009. Commercial practice of roselle ( Hibiscus sabdariffa L.) beverage production: optimization of hot water extraction and sweetness level. World Journal of Agricultural Sc iences 5(1):126131. Bravo L. 1998. Polyphenols: chemistry, dietary sources, metabolism, and nutritional significance. Nutr Rev 56(11):317333. Buttery RG, Teranishi R & Ling LC. 1987. Fresh tomato aroma volatiles: a quantitative study. J Agric Food Chem 35(4):540544. Calix TF, Ferrentino G & Balaban MO. 2008. Measurement of highpressure carbon dioxide solubility in orange juice, apple juice, and model liquid foods. J Food Sci 73(9):E439E445. Carvajal Zarrabal O, Waliszewski S, Barradas Dermitz D, Or ta Flores Z, HaywardJones P, NolascoHiplito C, Angulo Guerrero O, Snchez Ricao R, Infanzn R & Trujillo P. 2005. The consumption of Hibiscus sabdariffa dried calyx ethanolic extract reduced lipid profile in rats. Plant Food Hum Nutr (Formerly 60(4):15 3 159.

PAGE 142

142 CastaedaOvando A, PachecoHernndez MdL, Pez Hernndez ME, Rodrguez JA & GalnVidal CA. 2009. Chemical studies of anthoc yanins: A review. Food Chem 113(4):859871. Cisse M, Vaillant F, Acosta O, DhuiqueMayer C & Dornier M. 2009. Thermal degr adation kinetics of anthocyanins from blood orange, blackberry, and roselle using the Arrhenius, Eyring, and Ball models. J Agric Food Chem 57(14):62856291. Chang Y C, Huang H P, Hsu J D, Yang S F & Wang C J. 2005. Hibiscus anthocyanins rich extract induced apoptotic cell death in human promyelocytic leukemia cells. Toxicol Appl Pharm 205(3):201212. Chen CC, Hsu JD, Wang SF, Chiang HC, Yang MY, Kao ES, Ho YC & Wang CJ. 2003. Hibiscus sabdariffa extract inhibits the development of atherosclerosis in cholesterol fed rabbits. J Agric Food Chem 51(18):54725477. Chen J, Zhang J, Feng Z, Song L, Wu J & Hu X. 2009. Influence of thermal and densephase carbon dioxide pasteurization on physicochemical properties and flavor compounds in Hami melon juice. J Agric Food Chem 57(13):58055808. Chen SH, Huang TC, Ho CT & Tsai PJ. 1998. Extraction, analysis, and study on the volatiles in roselle tea. J Agric Food Chem 46(3):11011105. Clifford MN. 2000. Anthocyanins nature, occurrence and dietary burden. J Sci Food Agr 80(7):10631072. Clifford MN, Johnston KL, Knight S & Kuhnert N. 2003. Hierarchical scheme for LC MSn identification of chlorogenic acids. J Agric Food Chem 51(10):29002911. Clydesdale FM, Main JH & Francis FJ. 1979. Roselle ( Hibiscus sabdariffa L.) anthocyanins as colorants for beverages and gelatin desserts. J Food Protect 42(3):204207. Dagan GF & Balaban MO. 2006. Pasteurization of beer by a continuous densephase CO2 system. J Food Sci 71(3):E164E169. Damar S & Balaban MO. 2006. Review of dense phase CO2 technology: Microbial and enzyme inactivation, and effects on food quality. J Food Sci 71(1):R1R11. Damar S, Balaban MO & Sims CA. 2009. Continuous densephase CO2 processing of a cocon ut water beverage. Int J Food Sci Tech 44(4):666673. De Castro NE, Pinto JE, Cardoso MG, Morais de A, Bertolucci SK, Silva da F & Del N. 2004. Planting time for maximization of yield of vinegar plant calyx ( Hibiscus sabdariffa L). Cinc Agrotec Lav ras 28(3):542 551.

PAGE 143

143 Degenhardt A, Knapp H & Winterhalter P. 2000. Separation and purification of anthocyanins by highspeed countercurrent chromatography and screening for antioxidant activity. J Agric Food Chem 48(2):338343. Del Pozo Insfran D, Balaban MO & Talcott ST. 2006a. Microbial stability, phytochemical retention, and organoleptic attributes of dense phase CO2 processed muscadine grape juice. J Agric Food Chem 54(15):54685473. Del Pozo Insfran D, Balaban MO & Talcott ST. 2006b. Enhancing the ret ention of phytochemicals and organoleptic attributes in muscadine grape juice through a combined approach between dense phase CO2 processing and copigmentation. J Agric Food Chem 54(18):67056712. Del Pozo Insfran D, Balaban MO & Talcott ST. 2007. Inactiv ation of polyphenol oxidase in muscadine grape juice by dense phaseCO2 processing. Food Res Int 40(7):894899. DelgadoVargas F, Jimnez A & Paredes Lpez O. 2000. Natural pigments: carotenoids, anthocyanins, and betalains characteristics, biosynthesis, processing, and stability. Crit Rev Food Sci 40(3):173 289. Del Rio D, Stewart AJ, Mullen W, Burns J, Lean MEJ, Brighenti F & Crozier A. 2004. HPLCMSn analysis of phenolic compounds and purine alkaloids in green and black tea. J Agric Foo d Chem 52(10):28072815. Domnguez Lpez A, Remondetto GE & NavarroGalindo S. 2008. Thermal kinetic degradation of anthocyanins in a roselle ( Hibiscus sabdariffa L. cv. "Criollo") infusion. International J Food Sci Tech 43(2):322325. Du CT & Francis FJ 1973. Anthocyanins of roselle ( Hibiscus sabdariffa L.). J Food Sci 38(5):810812. Duangmal K, Saicheua B & Sueeprasan S. 2008. Colour evaluation of freezedried roselle extract as a natural food colorant in a model system of a drink. L WT Food Sci Techn ol 41(8):14371445. Duh P D & Yen G C. 1997. Antioxidative activity of three herbal water extracts. Food Chem 60(4):639645. El Adawy TA & Khalil AH. 1994. Characteristics of Roselle seeds as a new source of protein and lipid. J Agr ic Food Chem 42(9):1896 1900. Espin JC, Soler Rivas C, Wichers HJ & Garcia Viguera C. 2000. Anthocyaninbased natural colorants: A new source of antiradical activity for foodstuff. J Agric Food Chem 48(5):15881592.

PAGE 144

144 Esselen WB & Sammy GM. 1975. Applications for roselle as a red food colorant Food Prod Dev 9(8):37 38, 40. Fang N, Yu S & Prior RL. 2002. LC/MS/MS characterization of phenolic constituents in dried plums. J Agric Food Chem 50(12):35793585. Ferrentino G, Plaza ML, Ramirez Rodrigues M, Ferrari G & Balaban MO. 2009. Effects of dense phase carbon dioxide pasteurization on the physical and quality attributes of a red grapefrui t juice. J Food Sci 74(6):E333E341. Giusti MM & Wrolstad RE. 1996. Radish Anthocyanin Extract as a Natural Red Colorant for Maraschino Cherri es. J Food Sci 61(4):688694. Giusti MM, Rodriguez Saona LE, Griffin D & Wrolstad RE. 1999. Electrospray and tandem mass spectroscopy as tools for anthocyanin characterization. J Agric Food Chem 47(11):46574664. Giusti MM & Wrolstad RE. 2005. Characteri zation and measurement of anthoc y a nins by UV vis ible spectroscopy. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS Sc hwartz SJ, Shoemaker CF, Smith D & Sporns P editors. Handbook of food analytical chemistry Hoboken, NJ: John Wiley & Sons Inc. Gonzalez Palomares S, EstarronEspinosa M, Gomez Leyva JF & AndradeGonzalez I. 2009. Effect of the temperature on the spray drying of Roselle extracts (Hibi scus sabdariffa L.). Plant Food Hum Nutr 64(1):6267. Gradinaru G, Biliaderis CG, Kallithraka S, Kefalas P & Garcia Viguera C. 2003. Thermal stability of Hibiscus sabdariffa L. anthocyanins in solution and in solid state: effects of copigmentation and glass transition. Food Chem 83(3):423436. Gurbuz O, Rouseff JM & Rouseff RL. 2006. Comparison of aroma volatiles in commercial merlot and cabernet sauvignon wines using gas chromatography olfactometry and gas chromatography mass spectrometry. J Agric Food Chem 54(11):39903996. Hainida KIE, Amin I, Normah H & Mohd. Esa N. 2008. Nutritional and amino acid contents of differently treated Roselle ( Hibiscus sabdariffa L.) seeds. Food Chem 111(4):906911. Haji Faraji M & Haji Tarkhani AH. 1999. The effect of sour tea ( Hibiscus sabdariffa) on essential hypertension. J Ethnopharmacol 65(3):231236. Hansawasdi C, Kawabata J & Kasai T. 2000. Amylase inhibitors from roselle (Hibiscus sabdariffa Linn.) tea. Biosci Biotechnol Biochem 64(5):10411043.

PAGE 145

145 Hassan BH & Hobani AI. 1998. Flow properties of roselle ( Hibiscus sabdariffa L.) extract. J Food Eng 35(4):459470. Herrera Arellano A, Flores Romero S, Chvez Soto M & Tortoriello J. 2004. Effectiveness and tolerability of a standardized extract from Hibiscus sabdariffa in patients with mild to moderate hypertension: a controlled and randomized clinical trial. P hytomedicine 11(375382). Hirunpanich V, Utaipat A, Morales NP, Bunyapraphatsara N, Sato H, Herunsale A & Suthisisang C. 2006. Hypocholesterolemic and antioxidant effects of aqueous extracts from the dried calyx of Hibiscus sabdariffa L. in hypercholester olemic rats. J Ethnopharmacol 103(2):252260. Huang D, Ou B, Hampsch Woodill M, Flanagan JA & Prior RL. 2002. Highthroughput assay of oxygen radical absorbance capacity (ORAC) using a multichannel liquid handling system coupled with a microplate fluoresc ence reader in 96 well format. J Agric Food Chem 50(16):44374444. Hou D X, Tong X, Terahara N, Luo D & Fujii M. 2005. Delphinidin 3 sambubioside, a Hibiscus anthocyanin, induces apoptosis in human leukemia cells through reactive oxygen species mediated m itoch ondrial pathway. Arch Biochem Biophys 440(1):101109. Ito Y, Sugimoto A, Kakuda T & Kubota K. 2002. Identification of potent odorants in Chinese jasmine green tea scented with flowers of Jasminum sambac J Agric Food Chem 50(17):48784884. Jackman R L, Yada RY, Tung MA & Speers RA. 1987. Anthcoyanins as fo od colorants a review. J Food Biochem 11:201 247. Jabalpurwala FA, Smoot JM & Rouseff RL. 2009. A comparison of citrus blossom volatiles. Phytochemistry 70(1112):14281434. Juliani HR, Welch CR, W u Q, Diouf B, Malainy D & Simon JE. 2009. Chemistry and quality of hibiscus ( Hibiscus sabdariffa) for developing the natural product industry in Senegal. J Food Sci 74(2):S113S121. Kawakami M, Kobayashi A & Kator K. 1993. Volatile constituents of Rooibos tea ( Aspalathus linearis ) as affected by extraction process. J Agric Food Chem 41(4):633636. Kincal D, Hill WS, Balaban MO, Portier KM, Wei CI & Marshall MR. 2005. A continuous high pressure carbon dioxide system for microbial reduction in orange juice. J Food Sci 70(5):M249M254.

PAGE 146

146 Kincal D, Hill WS, Balaban M, Portier KM, Sims CA, Wei CI & Marshall MR. 2006. A continuous highpressure carbon dioxide system for cloud and quality retention in orange juice. J Food Sci 71(6):C338C344. Kong J M, Chia L S, Goh N K, Chia T F & Brouillard R. 2003. Analysis and biological activities of anthocyanins. Phytochemistry 64(5):923933. Lawless H & Heymann H. 1998. Sensory evaluation of food: principles and practices. NY: Chapman & Hall. 819 p. Lee MJ, Chou FP, Tseng TH, Hsieh MH, Lin MC & Wang CJ. 2002. Hibiscus protocatechuic acid or esculetin can inhibit oxidative LDL induced by either copper ion or nitric oxide donor. J Agric Food Chem 50(7):21302136. Lim S, Yagiz Y & Balaban MO. 2006. Continuous high pressure carbon dioxide processing of mandarin juice. Food Sci Biotechnol 15(1):1318. L in L Z & Harnly JM. 2008. Phenolic compounds and chromatographic profiles of pear skins ( Pyrus spp.). J Agric Food Chem 56(19):90949101. Lin WL, Hsieh YJ, Chou FP, Wa ng CJ, Cheng MT & Tseng TH. 2003. Hibiscus protocatechuic acid inhibits lipopolysaccharideinduced rat hepatic damage. Arch Toxicol 77(1):4247. Lin TL, Lin H H, Chen CC, Lin M C, Chou M C & Wang C J. 2007. Hibiscus sabdariffa extract reduces serum choles t erol in men and women. Nutr Res 27(3):140145. Liu CL, Wang JM, Chu C Y, Cheng M T & Tseng T H. 2002. In vivo protective effect of protocatechuic acid on tert butyl hydroperoxideinduc ed rat hepatotoxicity. Food Chem Toxicol 40:635641. Liu JY, Chen CC, Wang WH, Hsu J D, Yang MY & Wang C J. 2006. The protective effects of Hibiscus sabdariffa extract on CCl4 induced l iver fibrosis in rats. Food Chem Toxicol 44(3):336343. Lule SU & Xia W. 2005. Food phenolics, pros and cons: a review. Food Reviews International 21(4):367 388. Mahattanatawee K, Perez Cacho PR, Davenport T & Rouseff R. 2007. Comparison of three lychee cultivar odor profiles using gas chromatography olfactometry and gas chromatography sulfur detection. J Agric Food Chem 55(5):19391 944. Mazza G & Brouillard R. 1990. The mechanism of copigmentation of anthocyanins in aqueous solutions. Phytochemistry 29(4):10971102.

PAGE 147

147 Mazza G, Cacace JE & Colin DK. 2004. Methods of analysis for anthocyanins in plants and biological fluids. J AOAC In t 87(1):129145. McGuire RG. 1992. Reporting of Objective Color Measurements. HortScience 27(12):12541255. Meilgaard MC, Civille GV & Carr BT. 2007. Sensory evaluation techniques 4th ed. Boca Raton, FL: CRC Press. Mintel. 2008. RTD Non carbonated bev er a ges U.S. June 2008. Availlable from : http://www.marketr esearch.com. Accessed: Feb 20 2009. Mort on J. 1987. Roselle. In: Morton J, editor Fruits of warm climates Miami: Cr eative Resources systems Inc. p 281 286. Mounigan P & Badrie N. 2007. Physicoc hemical and sensory quality of wines from red sorrel/roselle ( Hibiscus sabdariffa L.) calyces: effects of pretreatments of pectolase and temperature/time. Int J Food Sci Technol 42(4):469475. Mourtzinos I, Makris DP, Yannakopoulou K, Kalogeropoulos N, Mi chali I & Karathanos VT. 2008. Thermal stability of anthocyanin extract of Hibiscus sabdariffa L. in the presence of Cyclodextrin. J Agric Food Chem 56(21):1030310310. Mueller BM & Franz G. 1992. Chemical structure and biological activity of polysaccharides from Hibiscus sabdariffa. Planta Med. 58(1):6067. Oboh G & Elusiyan CA. 2004. Nutrient composition and antimicroibal activity of sorrel drinks (soborodo). J Med Food 7(3):340342. Odigie IP, Ettarh RR & Adigun SA. 2003. Chronic administration of aqueous extract of Hibiscus sabdariffa attenuates hypertension and reverses cardiac hypertrophy in 2K 1 C hypertensive rats. J Ethnopharmacol 86(2 3):181185. Olvera Garca V, CastaoTostado E, Rezendiz Lopez RI, ReynosoCamacho R, Gonzlez de Meja E, Elizondo G & LoarcaPia G. 2008. Hibiscus sabdariffa L. extracts inhibit the mutagenicity in microsuspension assay and the proliferation of HeLa cells. J Food S ci 73(5):T75T81. Parr AJ & Bolwell PG. 2000. Phenols in the plant and in man. The potential for possible nutritional enhancement of the diet by modifying the phenols content or profile. J Sci Food Agric 80(7):9851012. Plotto A. 1999. Hibiscus: post pro duction management for improved market access for herbs and spices. Compendium on post harvest operations. Available from: http://www.fao.org/inpho/content/compend/text/ch28/ch28.htm. Accessed Apr 13, 2005.

PAGE 148

148 Pouget M, Lejeune B, Vennat B & Pourrat A. 1990a. Extraction, Analysis and Study of the Stability of Hibiscus Anthocyanins. Lebensm Wiss u Technol 23:103105. Pouget M, Vennat B, Lejeune B & Pourrat A. 1990b. Identification of Anthocyanins of Hibiscus sabdariffa L. Lebensm. Wiss u Technol 23:101 102. Prenesti E, Berto S, Daniele PG & Toso S. 2007. Antioxidant power quantification of decoction and cold infusions of Hibiscus sabdariffa flowers. Food Chem 100(2):433438. Reineccius G. 2006. Flavor chemistry and technology. Boca Raton FL: CRC Press Taylor & Francis Group. Rodrigu ez Saona LE & Wrolstad RE. 2005. Extractionn, isolation, and purification of anthcoyanins. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS, Schwartz SJ, Shoemaker CF, Smith D & Spor ns P editors. Handbook of food analytic al chemistry Hoboken, NJ: John Wiley & Sons Inc. Roethenbaugh G. 2005. Trends in beverage markets. In: Ashurst P, editor. Chemistry and technology of soft drinks and fruit juices 2 nd ed. Oxford, UK: Blackwell Publishing. p 1534. SyagoAyerdi SG, Arra nz S, Serrano J & Goi I. 2007. Dietary fiber content and associated antioxidant compounds in roselle flower ( Hibiscus sabdariffa L.) beverage. J Agric Food Chem 55(19):78867890. SeguraCarretero A, Puertas Meja MA, CortaceroRamrez S, Beltrn R, Alons o Villaverde C, Joven J, Dinelli G & Fernndez Gutirrez A. 2008. Selective extraction, separation, and identification of anthocyanins from Hibiscus sabdariffa L. using solid phase extractioncapillary electrophoresis mass spectrometry (time of flight /ion trap). Electrophoresis 29(13):28522861. Shahidi F & Naczk M. 2004. Phenolics in food and nutraceuticals. Boca Raton FL: CRC Press. 558 p. Stephens JM. 2003. Roselle Hibiscus sabdariffa L. University of Florida IFAS Extension #HS659. Suboh SM, Bilto YY & Aburjai TA. 2004. Protective effects of selected medicinal plants against protein degradation, lipid peroxidation and deformability loss of oxidatively stressed human erythrocytes. Phytother Res 18(4):280284. Tee PL, Yusof S & Mohamed S. 2002. Antioxidative properties of Roselle ( Hibiscus sabdariffa L.) in linoleic acid model system. Nutr Food Sci 32(1):17 20.

PAGE 149

149 Tsai P J, McIntosh J, Pearce P, Camden B & Jordan BR. 2002. Anthocyanin and antioxidant capacity in Roselle ( Hibiscus s ab dariffa L.) extract. Food Res Int 35(4):351356. Tsai PJ, Hsieh YY & Huang TC. 2004. Effect of sugar on anthocyanin degradation and water mobility in a roselle anthocyanin model system using 17O NMR. J Agric Food Chem 52(10):30973099. Tsai PJ & Huang H P 2004. Effect of polymerization on the antioxidant capacity of anthoc yanins in Roselle. Food Res Int 37(4):313318. Tseng TH, Hsu JD, Lo MH, Chu CY, Chou FP, Huang CL & Wang C J. 1998. Inhibitory effect of Hibiscus protocatechuic acid on tumor promotion i n mouse skin. Cancer L ett 126(2):199207. U.S. Census Bureau. 2009. U.S. Popul ation Projections. Availlable from : http://www.census.gov/population/www/projections/downloadablefiles.html Accessed: Mar 7, 2009. USDA (U.S. Department of Agriculture). 2009. National nutrient database for standar reference. Availlable from: http://www.nal.usda.gov/fnic/foodcomp/cgi bin/list_nut_edit.pl Accessed Sep 15, 2009. Waldemar W, Magdalena M & Janusz Co. 2004. A review of theoretical and practical a spects of solid phase microextraction in food analysis. Int J Food Sci Tech 39(7):703717. Wang H, Cao G & Prior RL. 1997. Oxygen radical absorbing capacity of anthocyanins. J Agric Food Chem 45(2):304309. Wang CJ, Wang JM, Lin WL, Chu CY, Chou FP & Tseng TH. 2000. Protective effect of Hibiscus anthocyanins against tert butyl hydroperoxideinduced hepatic toxicity in rats. Food Chem Toxicol 38(5):411416. Waterhouse AL. 2005. Determination of total phenolics. In: Wrolstad, R. E., Acree, T. E., Decker, E. A., P enner, M. H., Reid, D. S., Schwartz, S. J., Shoemaker, C. F., Smith, D. & Sporns, P., editors. Handbook of food analytical chemistry Hoboken, NJ: John Wiley & Sons Inc. Wong P K, Yusof S, Ghazali HM & Che Man Y. 2002. Physicochemical characteristics of Ro selle ( Hibiscus sabdariffa L.). Nutr Food Sci 32(2):6873. Wong P K, Yusof S, Ghazali HM & Che Man YB. 2003. Optimization of hot water extraction of roselle juice using response surface methodology: a comparative st udy with other extraction methods. J Sci Food Agric 83:12731278.

PAGE 150

150 Wong SP, Leong LP & William Koh JH. 2006. Antioxidant activities of aqueous extracts of selected plants. Food Chem 99(4):775783. Wrobel K, Wrobel K & Urbina EMC. 2000. Determination of total aluminum, chromium, copper, iron, manganese, and nickel and their fractions leached to the infusions of black tea, green tea, Hibiscus sabdariffa and Ilex paraguariensis (mate) by ETA AAS. Biol. Trace Elem. Res. 78(13):271 280. Wrolstad RE. 2004. Anthocyanin pigments bioactivity and colori ng properties. J Food Sci 69(5):C419C425.

PAGE 151

151 BIOGRAPHICAL SKET CH Milena M. Ramirez Rodrigues was born in Puebla, Mexico. After graduating from high school (American School of Puebla) in July 1998, she enrolled in the Food Engineering program of the Universidad de las Americas Puebla (UDLA). Before finishing her bachelors she was offered an assistantship to pursue a masters in f ood science. In July 2005 she was awarded a scholarship from CONACyT (National Mexican Council of Science and Technology) and had the opportunity to pursue her Ph.D. in Food Science at the University of Florida. While at UF she decided to enroll in the agr ibusiness masters program, from which she graduated in May 2009. After finishing her Ph.D., Milena hopes to continue exploring her interests in new product development and marketing.