|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
EXAMINATION OF AROMA VOLATILES FORMED FROM THERMAL
PROCESSING OF FLORIDA RECONSTITUTED GRAPEFRUIT JUICE
WENDY ANN-MARIE BELL
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Wendy A. Bell
To my Grandparents, with love.
I would like to express my sincere gratitude to my committee chair, Dr. Rouseff,
for his guidance, encouragement and support. It was a great learning experience. To my
committee members, Dr. Charlie Sims, Bruce Welt and Kevin Goodner, I would like to
extent my gratitude for their advice and assistance. I am also grateful to the Florida
Department of Citrus for providing financial support of my graduate studies.
I thank all my lab mates and Jack Smoot, for all their help. I would also like to
thank April Elston and Kristin Nelson for sniffing my samples. Thanks go to my family
for their support. To my grandmother, I would like to extend a special gratitude for her
kindness and constant support. She has always encouraged me to work hard and to put
God first in my endeavors.
Finally I would like to acknowledge my best friend, James. He was always there,
encouraging me with his kind words. I really appreciate all that he has done.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S .................................................................. ......... ................ iv
LIST OF TABLES ............................... ...... ...................... .. vii
LIST OF FIGURE S .................................. .. .. ...................... ............. viii
A B ST R A C T ................. .......................................................................................... x
1 IN TRODU CTION ................................................. ...... .................
2 LITER A TU R E R EV IEW ............................................................. ....................... 3
G ra p e fru it .................................................................................3
T herm al Processing .......................................................... .. ... ........ .. ........ ... 3
Possible Interactions with Stainless Steel Pasteurization Tubes .................................4
O ff-flav or ............... .................... .......................................... 5
Maillard Reaction Products .......................................................5
F u ra n s ......................................................................................... ...... . 5
Furanones .................................................. ........7
2,5 dimethyl 4-hydroxy- 3(2H) furanone (Furaneol).............. ................7
4,5 Dimethyl 3-Hydroxy- 2(5H) Furanone (Sotolone) ................................8
O their D egradation Products ............................................ .......................... 9
Instrum mental A analysis ................. .. ...... ....................9.. 9
G C T echniqu es ....................................................... 9
H PLC M methods .......... .. ......................................... .. ........ .... 10
3 MATERIALS AND METHOD ............. .... ................................. 12
R agents and Standards ............................. ..................................................... 12
Apparatus Setup ................ ......... ..... ...............12
Sam ple Preparation ......................................... .. .. .... ...... ................. 13
Sensory Analysis-Difference from Control Test ....................................... .......... 15
Instrum ental M ethods ............. .. ..... .......................... ... ............ ................17
Gas Chromatography-Flame Ionization Detector (GC-FID) ............................17
Gas Chromatography-Olfactometry (GC-O) .................... ......... ......... ...... 17
Gas Chromatography-Mass Spectrometry (GC-MS) ............... ..................18
HPLC Analysis of the Furans ........................... ............... 18
Sample Preparation ......... ........... ................. .............. ........ ..... 18
H PL C Instrum entation................................................ ............................. 18
Identification and Quantification................................................... 20
4 SENSORY ANALYSIS RESULTS AND DISCUSSION........................................22
Statistical Tests for Significant D differences ........................................ ....................23
Producing Heated/Cooked Flavor.. .. ........................ ...................28
5 GAS CHROMATOGRAPHIC RESULTS AND DISCUSSION............................31
G C -O .................................... ....... ............................................ 3 3
Aroma Active Compounds in Grapefruit Juice .................................. ............... 35
Aroma Active Compounds Lost during Heating..............................................39
Aroma Active Compounds That Intensified after Heating ..............................41
Caramel Aroma Compounds Formed from Sugar Degradation Reactions ...............42
Caramel Compounds Found in Grapefruit Juice...............................................45
3-M ethyl 2(5H) Furanone............... ............................ 46
Identification of Volatiles Using GC-M S .............. ........................................... 47
6 HPLC RESULTS AND DISCUSSION...... ..................... ...............53
7 C O N C L U SIO N ......... ......................................................................... ........ .. ..... .. 64
LIST OF REFEREN CE S ........................................ ........................... ............... 65
B IO G R A PH IC A L SK E TCH ..................................................................... ..................70
LIST OF TABLES
3-1 HPLC gradient conditions with flow rate at ImL/min ................ ...............20
4-1 Difference from control sensory data from 22 panelists.......................................24
4-2 Difference from control sensory data for the second (rep 2) and third replication..25
4-3 Results of the analysis of variance for each of the three replications ...................26
4-4 LSD comparisons of the means for the three samples .................. .... ........... 28
4-5 Combination used in off-flavor duplication along with their descriptors ...............30
5-1 Preliminary identification of major grapefruit juice components based on.............33
5-2 Aroma active peaks with tentative identification ................................................36
5-3 Aroma intensities based on normalized peak heights of samples..........................38
5-4 LRI for ten sugar degradation products possessing caramel aroma.......................43
5-5 Components identified in grapefruit juice by GC-MS.........................................49
6-1 Retention time and maximum absorbance of furans standards.............................55
6-2 Concentrations of 5-HMF, furfural and Furaneol ......................................... 60
LIST OF FIGURES
2-1 Formation of furans by either the acidic pathway (1,2- enaminol) or by the basic
pathway (2,3-enediol)................ ......... ...... ... ...... ............
3-1 H eating apparatus s setup ............................................ ......................................... 14
3-2 Sample ballot for the grapefruit juice difference from control test using ................16
3-3 Graphical representation of the HPLC gradient conditions .................................20
5-1 Overlay of partial grapefruit juice FID chromatogram on a ZB-5 column..............32
5-2 Partial FID (top) and aromagram (bottom) overlay of grapefruit juice. .................34
5-3 Aroma intensities for the unheated reconstituted juice and glass-heated samples...37
5-4 Aroma active volatiles whose intensity diminished due to heating .......................40
5-5 Aroma active volatiles (peak numbers 8, 11, 16, 19, 29, 26, 32, & 35)...................42
5-6 Structure of 3(2H) furanones .............................................................................44
5-7 Structure of sotolone, a 5(2H) furanone............................................................. 45
5-8 Structure of the 3-methyl-2(5H) furanone ............. .................................................46
5-9 Segmented MS total ion chromatogram using ZB-5 column..............................48
5-10 Mass spectra of myrcene (top) and ao-pinene (bottom) .........................................50
5-11 Alpha-terpineol levels in the three sam ples .................................. ............... 51
5-12 Formation pathway for a-terpineol in the presence of limonene and linalool .........52
6-1 HPLC spectra of the seven standards ................................................................ 55
6-2 H om ofuraneol tautom ers........................... .... ...... ..... ................. ............... 56
6-3 Segment of HPLC chromatogram showing overlay .............................................57
6-4 Spectra of 5-HMF in standard (top) and in the grapefruit juice sample .................58
6-5 Spectra of furfural standard (top) and sample (bottom).........................................59
6-6 Linear calibration curve for 5-HMF analyzed in triplicate. ....................................60
6-7 Linear calibration curve for Furaneol at 290nm..................................................61
6-8 Linear calibration curve for furfural standard at 290nm ........................................62
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
EXAMINATION OF AROMA VOLATILES FORMED FROM THERMAL
PROCESSING OF FLORIDA RECONSTITUTED GRAPEFRUIT JUICE
Wendy Ann-Marie Bell
Chair: Russell L. Rouseff
Major Department: Food Science and Human Nutrition
Over fifty percent of Florida's grapefruit supply goes into juice production.
Grapefruit juice is heated after reconstitution with water to minimize spoilage
microorganisms. However, heating induces chemical changes that degrade juice flavor.
Fruit juices are traditionally heated using stainless steel surface plates or tubes. Stainless
steel contains appreciable amounts of transition metals such as nickel and chromium that
are known to catalyze chemical reactions. Acidic (pH 2.9-3.5) juice may also promote
leaching of these metals. Glass should be more inert than stainless steel and may offer a
means to produce higher quality juice. This study was undertaken to examine changes in
aroma volatiles in reconstituted grapefruit juice when heated on stainless steel and glass
Sensory analysis of grapefruit juices heated at 1000C for 10 minutes indicated that
both heated samples exhibited a heated, pineapple, metallic, and cooked off-flavor.
Unheated reconstituted juice had a fresh grapefruit juice character. Using data from a ten
point rating system, the "Difference from Control" test indicated that significant
differences (p<0.05) existed between the unheated juice and the two heat treatments.
Analysis of heated juices with GC-O showed appreciable reductions in aroma
intensity of compounds that are responsible for fresh grapefruit juice character.
Intensities of ca-pinene, myrcene, and P-sinensal were at least 45% lower in the heated
juices compared to the unheated juice. There was also a corresponding increase in aroma
intensity of compounds associated with flavor degradation such as 2,5-dimethyl-4-
hydroxy-3(2H) furanone (Furaneol) and methional in heated samples. Two other
Maillard reaction products, sotolone and 3-methyl-2(5H)-furanone, were detected by GC-
O. Both high processing temperatures and extended times are required to produce
cooked or heated off-flavors in grapefruit juice. Increases in 5-HMF concentration were
observed with heating, suggesting that Maillard reactions were involved.
Sensory experiments designed to induce cooked, heated off-flavor in unheated
reconstituted juice indicated that a combination of Furaneol, homofuraneol and bis (2-
methyl-3-furyl) disulfide, a thermal degradation product of thiamine (vitamin B ), could
produce a cooked off-flavor similar to that observed in the excessively heated juices.
Grapefruit is the second largest citrus commodity in the State of Florida. Florida
supplies over 30% of the world's grapefruit production. Over 55% of Florida grapefruit
goes into juice production while over 35% is used for fresh fruit. Grapefruit sales have
been declining in recent years due to several factors ranging from uneven flavor quality
to interactions with certain drugs. Grapefruit juice has a unique citrus flavor, but low
sugar to acid ratios and high bitterness can overwhelm the pleasant flavor aspects. Red
grapefruit is used mainly for fresh fruit or for not-from concentrate juice (NFC), while
white grapefruit juice is generally converted to "from concentrate" (FCGJ) juice.
Thermal processing alters the overall flavor of grapefruit juice but is necessary to
reduce populations of viable spoilage microorganisms, inactivate pectinesterase and,
ultimately to increase shelf life. Formation of off-flavors due to processing is highly
undesirable to consumers; therefore research is necessary to identify critical factors that
cause significant flavor alterations. Cooked and heated off-flavor was observed in both
NFC and FCOJ and had profoundly negative impacts on perceived flavor quality.
Traditional processing of grapefruit juice is achieved through the use of stainless
steel tubes or plates. Transition metals can catalyze chemical reactions known to cause
off-flavors. Glass is an inert material and should not catalyze such reactions.
Previous studies have identified several compounds that contribute negatively to
citrus juice flavor. One such class of compounds is the furans that are formed as a result
of sugar degradation. These furans impart a caramel aroma depending on the
concentration. Use of 5-HMF and furfural as markers of thermal abuse, have been
suggested by several authors. Levels of these compounds in grapefruit juice typically do
not approach their aroma threshold but as the level of these compounds increase, juice
sensory quality declines. Furaneol, 2,5-dimethyl 4 hydroxy-3(2H) furanone, imparts an
aged, pineapple like aroma to citrus juice at levels exceeding 50 [g/L. In citrus based
soft drinks, sotolone imparts a burnt spicy note at levels ranging from 1.38ng/mL to
The objectives of this study were to determine if heating grapefruit juice in contact
with glass or stainless steel resulted in distinguishable sensory flavor differences as well
determining there were differences in individual aroma volatiles. A secondary objective
was to determine what compounds were responsible for heated or cooked off flavor in
Grapefruit is the second largest citrus commodity in the State of Florida,
accounting for over 30% of the world's grapefruit production. The entire United States
of America produces 40% of the world grapefruit production. For the 2002-2003 season,
36% of the grapefruit was used for fresh fruit, 16% for the chilled juice sector and over
40% to frozen concentrate products. With the exception of 1999-2000 season, there has
been a continuous decline in grapefruit production. Sales have been declining due to
factors ranging from poor flavor quality to medical precautions. Significant changes in
the overall flavor of processed grapefruit are of concern to the grapefruit processors as
well as the consumers.
Citrus juice generally undergoes heat treatment in order to achieve reductions in
viable spoilage organisms, to inactivate pectic enzymes, and ultimately to extend the
Inactivation of pectinesterase is important since it prevents the cloud separation and
associated reactions that may alter the flavor profile. Temperatures of at least 700C to
900C with hold times varying from a few seconds to 1 minute are necessary for
inactivation of most pectic enzymes. Pectin methyl esterase is a heat resistant enzyme
found in grapefruit juice and, is inactivated at temperatures exceeding 900C (1). Most
processors pasteurize not from concentrate (NFC) or chilled juice at temperatures
exceeding 900C and majority of the time juice is being heated twice. Grapefruit juice
from concentrate (FCGJ) is subjected to at least two heat treatments. The first occurs
during the evaporator concentration process where water is removed with a multi effect
evaporator. The second heat treatment occurs after the reconstitution of the juice and the
add-back of important flavor volatiles.
Possible Interactions with Stainless Steel Pasteurization Tubes
Pasteurization and concentration processes involve heating the juice in stainless
steel tubes or plates. Transition metals in food grade stainless steel tubes may catalyze
reactions throughout the heating processes. Food grade stainless steel, 316, typically has
16-18% Cr, 10-14% Ni and 2-3% Mo. Since juice contacts steel directly, leaching of
metals may be enhanced by factors such as low pH, high temperatures, and contact time.
Studies on effect of pH and temperature on chromium release from stainless steel showed
that after heating fruit juices with pH ranging from 2.5-3.0 at 950C for lhr, 31-50g/L
chromium was released (2). Residence times in tubes for NFC pasteurized juices are
typically in the range of 6-30s. The thermally accelerated short time evaporator (TASTE)
developed by Cook Machinery (Dunedin FL) is generally used to concentrate citrus
juices. This evaporator utilizes high temperature with residence times of less than one
minute (3). Although residence time is not long, high processing temperatures, combined
with the low pH may be more damaging than either factor alone.
High temperature pasteurization induces chemical changes that may affect the
quality of the juice immediately or will initiate a chain of chemical reactions that will
ultimately produce flavor degrading compounds.
High heating temperatures along with prolonged storage have been shown to be
contributing factors to off-flavor development (4;5;6). Studies into heated/cooked off
flavor found in orange juice was carried out by heating orange juice from concentrate at
various time-temperature combinations and then subjecting the samples to varying
storage times. Sensory evaluation revealed that orange juice heated at 1000C for 10 min
produced a heated, cooked and over processed flavor but was not predominant. Samples
were rated either "dislike slightly" or lower by majority of the panelists (7).
Consumer testing of commercial orange juices revealed that that several had a
cooked flavor. Juices with this cooked flavor were given the lowest quality ratings. Both
chilled (NFC) and from concentrate products were evaluated (8).
Maillard Reaction Products
Off flavor formation in citrus due to thermal degradation of sugars is of high
importance. Non-enzymatic browning products may be directly or indirectly
contributory to the flavor.
In citrus juices, furans are important indicators of thermal abuse. Furfural and 5-
hydroxy methyl furfural (5-HMF) have been reported in citrus juices that have undergone
heat treatment and high storage temperatures (4;6).
Commercial canned and glass packed orange juice samples were stored at
temperatures ranging from 50C to 300C over a period of 16 weeks. Results showed that
storage temperatures above 16C increased the rate of furfural formation more rapidly
than at lower temperatures. Concentrations of furfural observed in canned orange juices
stored at 300C increased 28 times compared to juices stored at 50C. For every 5C
increase in storage temperature, the level of furfural doubled, thus demonstrating a hight
correlation between storage temperatures and furfural accumulation. Glass packed
orange juice samples showed a more rapid increase in furfural formation. In fact, a 72-
fold increase in furfural levels were noted in samples ranging from 50C 300C storage
A similar study utilizing commercially canned grapefruit juice revealed that storage
temperatures from 100C to 400C were enough to induce a 390-fold increase in furfural
levels over a 15 week storage time. Panelists were able to detect significant flavor
changes in juices in which furfural levels exceeded 0.175 ug/ml. Since its taste threshold
in orange juice was 80 [tg/mL and thus not directly flavor active, furfural can only be
used as an indicator for off flavor (9).
Amino acids can increase the rate of ascorbic acid degradation. Furfural is formed
through acid catalysis of ascorbic acid (6). Ascorbic acid is degraded more rapidly in
single strength juice than in concentrated juices (10).
Levels of 5-hydroxy methyl furfural (5-HMF) in grapefruit juice samples increased
up to 3000 times with a 400C rise in temperature over 15 week storage study (9).
Formation of 5-HMF increased slightly between 100C and 200C but rose exponentially
above 300C. Under acidic conditions, 5-HMF is formed by the 1,2 enolization and
dehydration of hexose sugars (fig.2-1). The R group in figure 2-1 is replaced by a methyl
group. It has a high threshold of 100 [tg/mL (11) and therefore is not flavor active (12).
In fact, sensory results did not correlate with 5-HMF concentration (9). Thus other
compounds must be responsible for the observed off-flavor.
I undergoes series of
HPHR / CHOH Dehydration &deamination
C -o with pentose R=H
I with hexose R= CH2 OH
CHOH CHPHR H
Amdr Cundergoes series of /
Amadori C /H
Compound 11 --- --.
S dehydration & deamination from hexose
with pentose R=H
with hexose R= CH3
Figure 2-1 Formation of furans by either the acidic pathway (1,2- enaminol) or by the
basic pathway (2,3-enediol)
Furanones are important flavor compounds that are found both naturally and as a
result of thermal processing in foods and beverages. Two important furanones are
Furaneol and sotolone.
2,5 dimethyl 4-hydroxy- 3(2H) furanone (Furaneol)
Furaneol gives a caramel-like, burnt pineapple aroma in high concentrations and a
sweet, strawberry like aroma at low concentrations. It is one of the major "character
impact" compounds in pineapple (13) and muscadine grapes (14). Furaneol can exist
either free or glycosidically bound in other fruits such as mango, grapes, tomato (15), and
strawberry (16;17). It is found in thermally processed foods such as beef broth (18), and
fruits such as pineapple. This furanone has been shown to mask the flavor of orange
juice, giving an aged, cooked aroma when present at 0.05 [g/mL or greater (4). In
grapefruit juice, a pineapple-like aged flavor develops with high levels of Furaneol (19).
It has a threshold of 0.03 [g/L in water (20). Orange juice samples stored at 400C had 5
times more Furaneol compared to samples stored at 40C (21). Grapefruit juice samples
stored at 500C for 15 weeks showed an increase up to 28 times that of samples stored at
200C (19). Furaneol formation in citrus juices occurs as a result of a Maillard sugar
Studies using model solutions of citrus juice showed that Furaneol is formed over a
range of pH values in the presence of rhamnose. At higher pH values Furaneol is formed
through the 2,3-enolization pathway of the Amadori compound (fig. 2-1) followed by
dehydration and molecular rearrangement (22). The amino acid arginine is also a
necessary substrate for its formation under acidic conditions. Rhamnose is a 6-deoxy
hexose sugar formed from enzymatic degradation of pectin during processing and
storage. Flavanone glycosides such as hesperidin and naringin can also degrade to form
free rhamnose (23). Acid-catalyzed fructose degradation is also a possible formation
pathway for Furaneol. In heated solutions it can decompose to smaller molecules (24) or
it can react with other groups such as thiols to produce flavorful compounds.
Like most Maillard reactions, this reaction requires only a small amount of
substrate produce potent flavors; thus; it is hard to measure the depletion of these
4,5 Dimethyl 3-Hydroxy- 2(5H) Furanone (Sotolone)
Sotolone is a powerful aromatic compound that produces a curry-like aroma in high
concentrations, however, at low levels imparts a burnt, spicy aroma (25). Sotolone has an
aroma threshold of 0.02 [g/L in air (26) and is found in wines (25), sherry and roasted
coffee (26). Recently, it has been reported in orange essence oil (27) and was found to be
a source of off-flavor in citrus soft drinks (28) producing a spicy, burnt aroma. Its
formation was monitored in model soft drinks both with and without ascorbic acid. It
was postulated that sotolone was formed from ascorbic acid or dehydroascorbic acid in
the presence of ethanol. Oxygen and metal ions are also necessary for its formation.
Small amounts of ethanol are naturally found in citrus juices.
Other Degradation Products
Methional and 2-methyl-3-furanthiol are possible off-flavor contributors formed
from strecker and thiamine thermal degradation, respectively. Methional has a cooked
potato aroma while 2-methyl-3-furanthiol has a cooked, meaty aroma. Another thiamine
degradation product, bis (2-methyl3-furyl) disulfide, a dimer of 2-methyl-3-furanthiol
(MFT), which also produces a meaty aroma. Both MFT and its dimer have been reported
in grapefruit juice from concentrate (29). Thermal processing can also degrade 3-
carotene; producing norisoprenoids such as P-damascenone, a-ionone and P-ionone.
Gas chromatography can separate complex mixtures of volatiles using high
resolution capillary columns. GC works on the principle that volatiles are sequentially
eluted in the general order of boiling point. Thus they are thermally desorbed from the
capillary column as the column (oven) temperature increases. Separation efficiency also
depends on polarity of both the compound and the stationary phase, which lines the
column. On a polar column such as DB-wax, polar compounds will be retained by the
stationary phase more than they would on a non-polar DB-5 column.
Gas chromatography-olfactometry (GC-O) is a powerful analytical technique that
combines an instrumental detector as well as human response. It offers the ability to
define aroma active volatiles in terms of odor and intensity (30). The human nose has an
odor detection threshold of 10-19 moles (31) compared to 10-12grams for a mass
spectrometer (32). Isolation of the compounds of interest from the food matrix is very
important (31). GCO analysis can involve flavor dilution techniques such as Charm or
aroma extraction dilution analysis. These techniques are useful in determining flavor
threshold of important aroma active compounds. Another technique, OSME or time-
intensity, is based on evaluating intensities of aroma over the course of the GC analysis
and in most cases is not as time consuming as the flavor dilution techniques. Drawbacks
to sniffing include fatigue, saturation and adaptation (31) therefore assessors should limit
repetitions that are done in a single day and also lengthen the time between sniffs.
HPLC is an excellent alternative method of analysis. It is a non-thermal
chromatographic technique, which allows the separation of both volatiles and non-
volatiles. Reversed phase chromatography has been one of the choice methods for furan
analysis in citrus products (19;21;28;33). Grapefruit juice is a complex mixture and
efficient separation of its components lies in sample preparation and chromatographic
conditions. There have been several sample preparation procedures for furan analysis in
citrus juices. Use of the Carrez clarifying reagent was found to be effective in removing
pulp, fat, protein and carotenoids that may co-elute with the furans (34). The same author
(9) employed a simplified method of centrifuging the juice to remove the pulp layer. In
both cases passing the remaining solution through a C-18 cartridge to selectively remove
the desired polar compounds (34). Previous studies used mobile phases of consisting of
acetonitrile-water (85:15), (34), acetonitrile-glacial acetic acid-water (9) to effectively
isolate 5-HMF and furfural. Methanol and sodium acetate buffer were found to
satisfactorily isolate Furaneol. Initial efforts to quantify Furaneol in grapefruit juice
using aqueous methanol (30%) or phosphate buffer with 30% methanol were
unsuccessful due to co-elution with another compound (19). A linear gradient and
mobile phase consisting of methanol, acetonitrile and pH4 acetate buffer provided good
separation of Furaneol (21). However, in citrus juices another compound will co-elute
with Furaneol. The technique of creating difference chromatograms by subtraction of
two detection wavelengths (35) has been employed to isolate Furaneol. Furaneol has
minimal absorbance at 335nm and appreciable absorbance at 292nm whereas the
interfering compound had an absorbance maximum at 335 nm. Thus, the difference
chromatogram should provide a peak which is due almost entirely to Furaneol.
Sotolone has previously been isolated using both reversed and normal phase
chromatography. Reversed phase chromatography using a mobile phase consisting of
acetonitrile and water adjusted to pH 2.5 with sulfuric acid was used to quantify sotolone
in citrus soft drinks (28). This compound was monitored at 235nm, a wavelength close to
its maximum absorbance. Normal phase chromatography has been the predominant
method of isolating or analyzing sotolone. The sotolone content in wines was determined
with a detection limit of 10g/L using a diol column and hexane/ dichloromethane as the
mobile phase (25). HPLC can also be used as a preparative step by collecting fractions of
sotolone and then using GC-MS to identify and quantify sotolone (36).
MATERIALS AND METHOD
Reagents and Standards
Pentane, an organic solvent used in the gas chromatographic analysis was obtained
from Fisher Scientific (Pittsburgh, PA). Acros Organics (New Jersey) supplied diethyl
ether used for solvent extraction, and the methanol used in HPLC solvents and for
preparation of standards. Standards of 3-methyl-2(5H) furanone, 4-hydroxy-5-methyl-
3(2H) furanone, 2,5-dimethyl-4-hydroxy-3(2H) furanone, 2,5-dimethyl-4-methoxy-3(2H)
furanone, 2-ethyl-4-hydroxy-5-methyl-3(2H) furanone, 5-hydroxy methyl furfural, 4,5-
dimethyl-3-hydroxy-5(2H) furanone, E-2-decenal, E,E-2,4-decadienal, methional, E,Z-
2,4-decadienal, E,E-2,4-nonadienal, E,Z-2,6-nonadienal, 1-octen-3-one, maltol and 2-
methyl-3-furanthiol were obtained from Aldrich Chemical Co. (Milwaukee, WI).
Vanillin was obtained from Sigma Chemical Co (St. Louis, MO). Linalool, octanal,
decanal, nootkatone, P-sinensal, a-terpineol, myrcene, and a-pinene Sodium phosphate
was obtained from Sigma while the phosphoric acid from Fisher Scientific.
Figure 3-1 illustrates the heating apparatus setup. An oil bath consisting of light
mineral oil (Fisher Scientific) was heated using a hot plate. A condenser with chilled (-
5C) 50:50, ethylene glycol: water, was attached to the round bottom flask to allow for
refluxing of volatiles through condensation. The mouth of the condenser was also
covered with aluminum foil. For each heating in the stainless steel and the glass
containers, temperatures of both the oil bath and the juice samples were monitored.
Mid Season White Grapefruit Concentrate was obtained from a local processor.
The concentrate was reconstituted to 10Brix using deionized water. 300ml of the
reconstituted grapefruit juice was placed in a 1000ml stainless steel round bottom flask.
This was then submerged into a pre-heated oil bath. A hold time of 10 min was
maintained once the sample reached a temperature of 100C. After heating, samples were
placed in a pre-sterilized 10oz glass flint bottle (All American Container, Tampa, FL),
and immediately cooled on ice. The heating was repeated using a glass round bottom
Sample preparation for GC analysis was carried out using liquid-liquid extraction.
20 ml of grapefruit juice were added to 10 ml of 1:1 pentane and diethyl ether for
extraction. A Mixxor-like apparatus consisting of two 50 ml syringes connected via a
stainless steel luer-lock connector was used to extract the volatiles. Extraction was
facilitated by moving syringes in a back and forth motion twenty times. The mixture was
then centrifuged at 4000 rpm for 10 minutes to break the emulsion. The solvent layer
was removed and retained, while the aqueous portion was extracted a second time. After
the second extraction, both solvent fractions were combined. The extract was dried with
Figure 3-1 Heating apparatus setup
A 50 [tl volume of 2000 [g/mL ethyl valerate and 2-heptadecanone were added as
internal standards. The solvent layer was concentrated to approximately 100[l using a
stream of high purity nitrogen and placed in a 100l1 limited volume insert, placed in a
1.8 mL sampling vial and sealed with a screw top lid. Each sample was refrigerated until
GC-O/GC-FID (GC) and GC-MS analysis. Each sample was analyzed four times for
GC-O analysis and twice for GC-MS.
Sensory Analysis-Difference from Control Test
Six batches of grapefruit juice were heated in stainless steel and glass containers.
The batches were then combined so as to present a uniform mixture to panelists.
Twenty-one to twenty-five untrained panelists from the Citrus Research and
Education Center were recruited for the study. A control and three samples, each coded
with a randomly selected 3-digit number were presented to the panelists. Order of
presentation randomized for each panelists based on six orders of presentation (ABC,
ACB, BCA, BAC, CBA, BAC). Panelists were then asked to taste the control then to
taste the sample and to rate the degree of difference between them. They were also told
that a hidden control was present among the samples. A sample ballot is shown in figure
3-2. The control was unheated reconstituted grapefruit juice.
Panelists were also asked to cleanse their palettes by eating a piece of unsalted
cracker and drinking a sip of water between samples to prevent any carryover from the
previous sample. A numerical ten-point scale similar to the hedonic scale was used, with
scale ranging from no difference to extreme difference. Color changes during the heating
process prompted the use of red lighting so as to mask the color difference. The above
procedure was repeated two additional times yielding a total of the three sensory
DIFFERENCE FROM CONTROL TEST
Today you will be tasting grapefruit juice.
Please eat a piece of cracker and drink a sip of water in between each sample to
cleanse your palette.
Three samples and one control are presented to you. Please circle the degree of
difference against the control for each sample.
It is possible that there may be no difference between the control and one of the
Control vs. # Control vs. # Control vs. #
0 =No difference 0 =No difference 0 =No difference
1= slight difference 1= slight difference 1= slight difference
2 2 2
3 = mod difference 3 = mod difference 3 = mod difference
4 4 4
5 = large difference 5 = large difference 5 = large difference
6 6 6
7 = very large difference 7 = very large difference 7 = very large difference
8 8 8
9 = extreme difference 9 = extreme difference 9 = extreme difference
Figure 3-2 Sample ballot for the grapefruit juice difference from control test using a ten-
point rating scale
Gas Chromatography-Flame Ionization Detector (GC-FID)
Volatiles from grapefruit juice extracts were separated using an HP 5890 gas
chromatograph (Hewlett Packard, Palo Alto, Ca, USA) equipped with a flame ionization
detector. One-half micro liter sample was injected onto a ZB-5 or DB-Wax capillary
column (30m x 0.32mm I.D., 0.5Oim film thickness) (J&W Scientific, Folsom, CA).
Oven temperature was programmed with an initial temperature of 400C and ramped up to
2400C for wax column and 2650C for ZB-5 column, at 70/min with a final hold time of 5
minutes. Samples were injected using splitless mode with an injector port temperature of
2200C and the detector at 2500C. The effluent was split 1:2 between a flame ionization
detector and a sniff port (Datu, Geneva, NY). Data were collected and integrated using
the ChromPerfect v. 5.0software (Justice Innovations, Denville, NJ).
Gas Chromatography-Olfactometry (GC-O)
A sniff port was attached to the GC in order to facilitate the identification of aroma
active volatiles. The GC temperature program was the same as that for the GC-FID.
Two assessors were used to evaluate the aroma quality of the samples. Each
assessor analyzed the samples in duplicate accounting for four sniffs per sample. Only
those odors that were detected in 50% of the sniffs were considered to be aroma active.
Sniffing began after the elution of the solvent peak, approximately three minutes into
each run. A sliding scale ranging from none to extreme was used to indicate the aroma
intensity. Assessor's responses were captured using a variable potentiometer, which was
collected on the ChromPerfect software. Sniffing time for each run was 25 and 29
minutes for the ZB-5 and DB-wax column respectively. Aroma quality of each sample
Gas Chromatography-Mass Spectrometry (GC-MS)
Data were collected using a Finnigan GCQ Plus GC-MS system (Thermo Electron,
San Jose, CA) using a 99.999% pure helium as the carrier gas. A DB-5 capillary column
(60m x 0.25 mm id, 0.25[im film thickness column (J&W Scientific) was used. Oven
temperature was programmed from 400C to 2500C at a rate of 70C/min. Samples were
injected using split less mode with an injector port temperature of 2200C. Electron
impact (EI) mode was used with ionization energy of 70eV. Scans were obtained after 7
minutes to avoid the solvent peak and were measured from 40-300 m/z. Data were
analyzed with XCalibur v. 1.3 software (Thermo Electron, San Jose, CA).
HPLC Analysis of the Furans
Solid Phase Extraction using a method modified by Walsh was employed (21).
10ml of grapefruit juice were centrifuged for 5 min at 3000 rpm. Three milliliters
supernatant were passed thru a C-18 cartridge (J & W Scientific, SPE Accubond
500mg/6ml, Folsom, CA) that had been pre-conditioned with 2.5 ml methanol and
washed with 6 ml water. The cartridge was then washed with 2 ml water to remove the
sugars and eluted with 2 ml methanol at a rate of 1 drop per 10s. Samples were filtered
thru a 0.45[i nylon filter (Fisher Scientific) and stored in a 1.8 ml amber vial until
injected into the HPLC.
Separation was achieved by injecting 20 [tl of the juice extract onto a Phenomenex
5[ C-18 column (4.6mm id x 250mm) (Phenomenex, Torrance, CA). A 10ul injection
volume was used for standards. Column temperature was kept at 250C.
Chromatographic analysis was carried out using Surveyor HPLC equipped with a PDA
detector. Versatility of the PDA detector enabled the monitoring of three individual
wavelengths: 235nm, 290nm and 335nm, while obtaining a spectral scan 220-380nm.
Data was collected and analyzed using Atlas software v.1.0 at an acquisition rate of 10Hz
(Thermo Electron) over a 40 minute run time.
The mobile phase consisted of methanol and phosphate buffer. Phosphate buffer
(pH4.1 +1) was made using a 0.05M sodium phosphate solution and phosphoric acid.
Linear gradient conditions are summarized in table 3-1 while figure 3-3 shows the
graphical gradient conditions. A flow rate of Iml/min was maintained and a re-
equilibration time of 10 min was applied.
Table 3-1 HPLC gradient conditions with flow rate at ImL/min
Time % Methanol % phosphate
0 5 95
10 15 85
27 30 70
33 40 60
40 5 95
50 5 95
Figure 3-3 Graphical representation of the HPLC gradient conditions
Identification and Quantification
Identification of aroma active components was achieved by comparing linear
retention indices and aroma descriptors with those from a database or published values.
Confirmation was achieved by matching fragmentation pattern of the sample with that of
standards or library. A series of alkanes (C5-C25) analyzed under the same GC conditions
- -%methanol buffer
were used to calculate linear retention indices (37). A scatter plot of retention time along
with retention indices of alkane standards were graphed on Excel. Retention indices for
the standards were obtained by multiplying the carbon number by a factor 100. A
polynomial was obtained from a regression. This equation was then used to calculate
linear retention indices for individual components in grapefruit juice. Alkane standards
were analyzed for both DB-wax and ZB-5 columns, as well as for GC-MS.
Components from HPLC analysis were identified using a combination of
comparing retention times and spectra of standards with that of sample. Quantification
was done using external standard method. Sets of standards with known concentrations
were analyzed in triplicate. Areas of those standards were plotted against the amounts
and a linear regression was obtained along with a linear equation. Responses from
unknown components were then substituted in the linear equation in order to find the
amounts and ultimately the concentration of components in juice samples.
SENSORY ANALYSIS RESULTS AND DISCUSSION
There are several sensory analysis tools that are used in difference testing of foods.
"Difference from Control" test was used in this study as it enabled the evaluation of three
treatments in one sitting. There are two objectives of the difference test. First, it
determines if a difference exists between samples and the control. Secondly, it estimates
the magnitude of the difference (38).
The three samples that were evaluated were
Unheated unheated reconstituted grapefruit juice (as well as hidden control)
Glass FCGJ heated 10 min. on glass surface
Metal FCGJ heated 10 min. on stainless steel surface
The sensory evaluation data for each replicate are shown in tables 4-1 and 4-2. Except
for one (panelist #12, table 4-2 rep 3), all panelists were able to identify the hidden
control, as having little or no difference from the control. Although overall scores (table
4-2 rep3) ranged from no difference (0) to very large or extreme (8) difference, only 68%
of the panelists were able to distinguish differences between the control and the juice
heated in metal. A difference score of three and greater was considered to show
difference, difference scores of less than three was not considered to be different from the
control. Scores for juice heated in glass ranged from moderate difference to very large
difference (7) for over 86% of the panelists. Seven of the twenty-two panelists rated the
juice heated in metal with a score of two or less compared to only three for the juice
heated in glass. Perhaps, these panelists could not distinguish the differences due to their
low bitterness threshold and hence high sensitivity. This sensitivity incapacitates
panelists from being able to distinguish flavor differences among samples. An example
of this is panelist number 15 (table 4-1 rep 1) and panelist # 4 9table 4-2 rep 2) who was
unable to detect much difference and commented that all three samples had high
bitterness levels. Also they could be anosmic or unable to detect heated off-flavors.
Although panelists' scores improved for the second and third evaluations, some panelists
were still unable to detect flavor differences in the juices. Average difference score for
the unheated reconstituted juice ranged from 1.96 in rep 1 to 0.45 in rep 3 suggesting an
improvement in identifying the hidden control. Most panelists had difficulty detecting
differences between the two heated samples e.g. panelists #1, 2, 5, 6, 8 & 10 from table
4-1 reply indicated at most, a one-point difference between the glass and metal samples.
Although there may have been a difference in overall flavor between the two heated
sample types, the severe heating produced such a strong off flavor that it was difficult for
panelists judge the degree of difference. Both samples had very strong heated/ cooked
Statistical Tests for Significant Differences
Statistical differences between samples were determined by using a two-way analysis of
variance (ANOVA) test, at the 95% confidence level, as it allows for the comparisons of
more than two samples (table 4-3). The advantage of using a two-way ANOVA is that
any error due to the panelists can be distinguished.
Table 4-1 Difference from control sensory data from 22 panelists between a control and
three samples, a hidden control (unheated reconstituted), grapefruit juice
heated in glass, and juice heated in metal using a 10-point scale.
Panelists reconstituted Glass Metal
1 1 5 5
2 0 6 5
3 4 5 1
4 0 5 3
5 5 8 7
6 2 6 5
7 2 3 5
8 1 6 7
9 0 3 3
10 0 5 4
11 0 3 5
12 1 0 3
13 4 6 4
14 1 2 5
15 0 0 0
16 4 3 2
17 3 5 3
18 4 7 2
19 3 5 1
20 2 2 4
21 2 5 7
22 1 7 7
23 5 7 7
24 3 9 1
25 1 1 3
Average 1.96 4.56 3.96
Median 2 5 4
St Dev 1.62 2.46 2.09
Min 0 0 0
Max 5 9 7
Table 4-2 Difference from control sensory data for the second (rep 2) and third
replication (rep 3).
Rep 2 Rep 3
Panelist No. reconstituted Glass Metal reconstituted Glass Metal
1 0 5 3 0 5 5
2 4 9 6 0 7 3
3 1 3 1 0 3 1
4 1 0 0 1 1 3
5 1 4 6 0 3 3
6 0 3 5 1 6 0
7 1 3 3 1 3 1
8 0 5 8 0 4 5
9 0 7 5 1 4 7
10 1 3 5 0 1 2
11 3 5 7 1 7 7
12 2 1 3 2 7 8
13 1 3 3 1 5 8
14 1 6 5 0 3 4
15 0 5 5 1 5 5
16 0 5 3 0 5 1
17 1 3 0 0 5 6
18 1 3 1 0 5 0
19 1 0 5 0 7 3
20 3 5 1 0 5 5
21 3 3 3 0 3 1
22 1 1 3
Average 1.19 3.86 3.71 0.45 4.32 3.68
Median 1 3 3 0 5 3
St Dev 1.14 2.10 2.21 0.58 1.87 2.46
Min 0 0 0 0 1 0
Max 4 9 8 2 7 8
Table 4-3 Results of the analysis of variance for each of the three replications. Repl-
table 4-1, reps 2 and 3 data are shown in table 4-2.
The F test is a statistical technique that was utilized to compare multiple means and
to test for variability among the means for the three samples (treatments) against the
variability of the observations (error). In this study the null hypothesis assumed that the
treatment means were the same, however, if the mean for the treatments were different
then the null hypothesis would be rejected. The maximum ratio between variations
among the treatments was measured using the Fcrit value. If the calculated F value were
greater than the Fcrit then there would be significant variations, which would lead to
differences among the sample means.
Source of Variation SS df MS F P-value F cnt
Panelists 158.08 24 6.59 2.08 0.015 1.75
Treatments 92.67 2 46.33 14.63 1.09E-05 3.19
Error 152 48 3.17
Total 402.747 74
Source of Variation SS df MS F P-value F crit
Panelists 107.94 20 5.40 1.89 0.043 1.84
Treatments 94.51 2 47.25 16.56 5.77E-06 3.23
Error 114.16 40 2.85
Total 316.603 62
Source of Variation SS df MS F P-value F crit
Panelists 105.15 21 5.01 1.88 0.040 1.81
Treatments 188.82 2 94.41 35.45 9.58E-10 3.22
Error 111.85 42 2.66
Total 405.82 65
From table 4-3 it can be seen that the Fcacl value is greater than the Fcrit, in all three
replicates, therefore we reject the null hypothesis and conclude that at least one of the
treatments means is different. Critical F values were approximately 3 compared to
calculated F values of 15, 17 and 35 respectively for reps 1, 2, & 3.
Since the analysis of variance test only suggests that a difference exist among the
populations, a multiple comparison test was used to rank the means and identify the
means that were different. Fisher LSD multiple comparison test was applied to the
means because it is the least conservative test in comparison to the Tukey and Duncans
test and should produce the most difference. The LSD test is defined as any observed
difference necessary to declare the corresponding population means different"(39). The
following equation was used to calculate the LSD:
LSD = t,/2~(2*EMS)/n
where t /2 was found at the degree of freedom of the error and an alpha level of 0.025
from a t-table, EMS, error mean square was taken from the ANOVA results and n was
the number of panelists.
Means for each population were subtracted from each other. Those with a difference
greater than the LSD value, were considered to be significantly different, while those that
were similar or less than the LSD value were not significantly different. From the data in
table 4-4 rep3, the difference between the glass and the metal was only 0.64 while the
difference between the unheated reconstituted juice and glass was 3.87. The calculated
LSD value was 0.99.
Table 4-4 LSD comparisons of the means for the three samples
Unheated Glass Metal LSD
Means Rep 1 1.96a 4.56b 3.96b 1.02
Means Rep 2 1.19a 3.86b 3.71b 0.50
Means Rep 3 0.45a 4.32b 3.68b 0.99
Since the difference between the unheated reconstituted juice and two heat treatments in
all three replications were greater than LSD values, it can be concluded that a significant
difference existed between unheated reconstituted and heated samples. However the
difference of means for two heated samples were less than the LSD value, therefore they
were not statistically different.
Cooked, heated, processed, pineapple flavor were some comments recorded by
panelists for heat-treated samples while the unheated reconstituted sample was identified
as having good grapefruit flavor. Panelists who had difficulty recognizing differences
between the control and two heat-treated samples recorded comments such a high
Producing Heated/Cooked Flavor
Attempts to reproduce the heated cooked off-flavor were carried out using three
furans known to produce a caramel aroma, Furaneol, sotolone, homofuraneol. In
addition, bis (2-methyl-3-furyl) disulfide, a thiamine degradation product sometimes
found in thermally abused citrus juices, was also added. Table 4-5 shows the
combination of the compounds used as well as their sensorial characteristics when added
to grapefruit juice based on a 6-member panel. It has been previously reported that
concentration greater than 0.05 [g/mL Furaneol masked the flavor of grapefruit juice.
This was seen in the combinations that had 1 lg/mL Furaneol. Little or no grapefruit
juice character detected by the panelists. However, in the combination containing 0.5
[g/mL, a slight grapefruit juice character was noticed, but caramel and sweet flavor
impression were more prominent. The combination of homofuraneol (1 [g/mL ),
Furaneol (1 lg/mL) and bis (2-methyl3-furyl) disulfide (2[g/mL) produced a cooked
Extremely low concentrations of sotolone (0.01 g/mL ) also contributed to over-
processed flavor. Sotolone is a very potent aroma active compound and would not
require a high concentration to produce a pronounced off-flavor.
Table 4-5 Combination used in off-flavor duplication along with their descriptors
No. Concentration Flavor Descriptor
1 Cg/mL homofuraneol + Sl cooked fl, caramel
1 1 Cg/mL Furaneol + 1 Cg/L aroma, no gfj character,
MFT-MFT sweet, less acid
1 Lg/mL homofuraneol More cooked/caramel
2 +1 g/mL Furaneol + 2Lg/mL aroma, less caramel aroma,
MFT-MFT less tart, sl gfj character
0.5Lg/mL Furaneol + 1 tg/mL Less cooked aroma, more
Less cooked aroma, more
3 homofuraneol +2pg/L MFT-
Sh r l gfj character, more sweet
+ Spicy, strong caramel,
4 0.5Lg/mL sotolone +
combination #1 spoiled juice, no gfj
combination #1 character
5 0.01 Lg/mL sotolone Processed, spicy aroma
GAS CHROMATOGRAPHIC RESULTS AND DISCUSSION
Gas Chromatography is a useful method of separating volatiles based on several
factors including polarity and boiling point. In this study, juice volatiles were separated
using high-resolution capillary gas chromatography. The use of at least two dissimilar
chromatographic phases is advantageous in that compounds that are not well separated
on a polar phase column can often be better separated on a non-polar column and the
reverse. Polar compounds tend to remain on a polar column longer and thus would be
better separated. Since GC is a separation technique, these components cannot be
identified without additional information. From the overlay of the FID chromatograms
(fig. 5-1) obtained from the three sample types, it was seen that some peaks were
diminished due to heating while others increased.
Peaks I, II and III showed a higher FID response for the unheated reconstituted
juice compared to that of the two heated treatments. In contrast, peak VII had a higher
response in the two unheated reconstituted samples. Table 5-1 lists some of the
volatiles that were tentatively identified using the linear retention indices on two
6 8 10 12 14 16 18
Figure 5-1 Overlay of partial grapefruit juice FID chromatogram
glass, C-juice heated in metal.
20 22 24 26 28
on a ZB-5 column. A-unheated reconstituted juice, B-juice heated in
Table 5-1 Preliminary identification of major grapefruit juice components based on
standardized retention times using FID responses from both polar (wax) and
non-polar (ZB-5) columns.
Peak RT LRI Lit LRI Preliminary LRI (wax) Lit LRI
(min) (ZB-5) (ZB-5) Identification (wax)
i 4.61 847 846 Ethyl 2-methyl 1022 1041
ii 6.39 939 939 a-pinene 978 1011
iii 7.49 989 989 myrcene 1172 1168
iv 8.5 1032 1025 limonene 1214 1205
v 9.82 1087 1074 cis-linalool oxide 1480 1448
vi 10.16 1100 1100 linalool 1552 1552
vii 12.37 1192 1207 a-terpineol 1713 1713
viii 14.75 1315 1376 a-copaene 1480 1472
ix 16.97 1432 1418 P-caryophyllene 1644 1624
x 21.38 1677 1706 p-sinensal 2254 2251
xi 23.81 1834 1814 nootkatone 2587 2595
Since the invention of GC-O as a tool to evaluate flavor volatiles, it has been
adapted by the food, environmental, and flavor industry as a major tool used to identify
those compounds that are aroma active. The selectivity and sensitivity of the human
nose combined with the resolving power of the GC provides useful information about
those volatiles that produce any aroma sensation. This is a particularly useful tool for
those potent aroma compounds that have extremely low thresholds. It can be seen that
little or no FID peaks were detected for the two peaks, 1-octen-3-one and 1-octen-3-ol,
in the region of 7-8 minutes (fig. 5-2). Their analytical concentrations were below
detection levels of the FID detector, yet their concentrations exceeded their aroma
thresholds so they were aroma active.
S5 linalool oxide
Figure 5-2 Partial FID (top) and aromagram (bottom) overlay of grapefruit juice analyzed on a ZB-5 column.
Similarly, not all volatiles found in relatively high concentrations or large FID
peaks are aroma active. Large FID peaks such as limonene and P-caryophyllene shown
in fig. 5-2 had little to no aroma activity. In the region between 24-28 min there were
several prominent FID peaks that did not produce detectable aroma activity.
Aroma Active Compounds in Grapefruit Juice
GC-O is particularly useful in explaining sensory data. However, not everyone
has the same sensitivity to all aroma compounds and there will be variability between
individuals in assessing these aromas. To compensate for individual sensory variability,
each sample was sniffed at least four times, twice per assessor. A compound was
considered aroma active only if it was detected in 50% of the sniff runs. By using two
panelists, effects of individual hypersensitivities and specific anosmias are minimized.
If one assessor is anosmic to a certain class of compounds and the other isn't, then there
is a better chance of detecting these compounds in the averaged assessor results.
GC-FID and GC-O were carried out on two dissimilar columns. The advantage to
this method is that volatiles are eluted in a different order with each column type.
Tentative identification is based upon cross-referencing calculated retention indices
with those from a library, published results or with standards.
Over forty aroma active grapefruit juice volatiles were detected using both wax
and ZB-5 columns. Shown in table 5-2 are twenty-six aroma active compounds that
were identified by matching retention indices along with aroma descriptors of
components on both columns. Standards were run under similar conditions to confirm
retention values and odorant descriptors.
Aroma intensities were normalized for each assessor to minimize differences
between assessor uses of the intensity scale. Aroma intensities detected on wax column
were normalized based on vanillin, which had the highest peak height among all the
volatiles (see table 5-3 and fig. 5-3).
Table 5-2 Aroma active peaks with tentative identification based on aroma attributes
and retention characteristics for grapefruit juice samples as detected by GC-
O on polar and non-polar columns.
ZB-5 Component Name
800 ethyl butyrate
847 ethyl-2-methyl butyrate
1032 limonene impurity
820 butanoic acid
1325 4-vinyl guaiacol
1677 3 -sinensal
1470 wine lactone
green, earthy, metallic
green, fresh, lemony
green, fatty, lemony
fresh, floral, fatty
floral, orange, greenish
faint cucumber, green
skunky, rotten, fresh
fatty, sl orange
rose, fresh, lemon pledge
metallic, cooked, burnt
7 8 28 30
7 |18 22 31 37
S1 16 41
I || 4
14 17 19
Retention Time (min)
Figure 5-3 Aroma intensities for the unheated reconstituted juice and glass-heated samples analyzed on wax column. Intensities were
normalized peak height based on vanillin, which had the highest peak height. Peak numbers correspond to those listed in
Table 5-3 Aroma intensities based on normalized peak heights of samples analyzed
using wax column. Descriptors are listed in ascending order of LRI.
Normalized Aroma Intensities
Component Calculated Unheated
No LRI Descriptors reconstituted Glass Metal
1 978 piney 6.06 2.91 2.36
2 1001 sweet, fruity 2.24
3 1022 fruity 4.94 5.22 6.88
4 1098 sulfur, stale 6.26 5.81
5 1172 green, earthy, metallic 6.09 1.88 0.83
6 1214 minty, licorice 5.45 6.52 3.83
7 1306 green, fresh, lemony 3.91 5.84 5.38
8 1315 mushroom 2.70 5.74 5.65
9 1381 green, grassy, piney 6.85 8.66 8.40
10 1390 cat litter 6.11
11 1470 cooked potato 0.65 9.01 8.12
12 1509 green, fatty, lemony 6.98 5.58 7.17
13 1514 metallic, green 2.53 6.91 2.81
14 1544 fresh, floral, fatty 4.24 3.41 4.06
15 1552 floral, orange, greenish 8.37 9.32 4.09
16 1596 faint cucumber, green, sweet 2.06 4.30 3.59
17 1633 skunky, rotten, fresh 3.72 3.31 3.06
18 1773 sweet, roasted 3.43
19 1828 fatty, sl orange 3.27 7.95 6.55
20 1840 tobacco(faint), piney 4.37 2.87 1.53
21 1861 rose, fresh, lemon pledge 2.88 7.52 6.30
22 1868 bug spray,spice 3.64 4.36 5.92
23 1875 caramel, sweet, fresh 3.34 3.57 3.90
24 1891 fresh cucumber,floral, fresh 3.18 2.36
25 1940 metallic 3.16
26 2000 burnt, spicy 5.85 2.72
27 2017 metallic,cooked, burnt 3.95 3.16 3.30
28 2038 spice 3.92
29 2043 caramel 4.35 8.00 8.61
30 2059 spicy(strong), curry 3.93 5.48
31 2154 sweet caramel 3.90 1.48
32 2163 sweet, caramel 7.35 7.84
33 2179 charred, sweet 5.07
34 2188 grapefruit, spice(lingering) 3.14 7.94 5.82
35 2201 spice 4.68 5.11 8.78
36 2207 sweet, burnt wood 2.85
37 2243 rancid fruit 3.27
38 2254 grapefruit, citrus 7.55 4.37 4.87
39 2259 nutty, musty, crayons 1.05
40 2298 grapefruit 5.86 2.37
41 2587 tropical, sour 5.20 2.89
42 2608 vanillin 9.73 4.26 10.00
Aroma Active Compounds Lost during Heating
Many of the components listed in table 2 have been previously reported in fresh
grapefruit juice (40). However, their concentrations have probably been altered due to
the heating process. During the evaporation process when making concentrate, most
volatiles are lost along with water. Although it has been reported that over 95% of the
volatiles in the fresh juice are lost in the process of concentrating grapefruit juice, over
57% of the total aroma activity remains (29). Typically, fresh juice flavor
characteristics are restored to the concentrated juice by the addition of juice essence oil,
aqueous essence and peel oil. The juice concentrate used in this study had 0.012% cold
pressed grapefruit oil added to it. This oil was the source for most of the volatiles and
many aroma active volatiles. Since essence oil and aqueous essence are obtained from
the condensate of the evaporation process, they will contain volatile heat reaction
products and thus were not used for this study. Therefore, any heated, cooked, caramel
aromas were produced during the specific heating conditions in this study. Normalized
intensities of aroma active volatiles (table 5-3) present in both the unheated
reconstituted juice as well as the two heat-treated juice samples were compared in order
to identify aroma differences due to heat treatment and contact surface.
Three aroma active compounds, c-pinene (1), myrcene (5) and P-sinensal (38),
exhibited greater aroma intensities in the unheated reconstituted juice compared to that
of the two heated juice samples (fig. 5-4). Another aroma active compound, ethyl
butyrate (2), was detected only in the unheated reconstituted juice with a normalized
intensity of 2.24 on a 10-pt scale. Ethyl butyrate imparts a sweet, fruity aroma and is
considered to be a positive contributor to grapefruit juice aroma (40). It is a low
molecular weight compound and is highly volatile, thus it likely to be degraded during
extended heating times.
Retention Time (min)
Figure 5-4 Aroma active volatiles whose intensity diminished due to heating
As seen in table 5-3, alpha-pinene (1) and myrcene (5) are two early eluting, low
boiling point compounds, which had higher aroma intensities in the unheated
reconstituted juice compared to the two heated juice treatments. In both heated
samples, the response for a-pinene was less than 45% that of the unheated reconstituted
juice while that for myrcene was less than one-third of the intensity in the unheated
reconstituted juice. A piney aroma was elicited by a-pinene while myrcene imparted a
green, metallic, earthy aroma. Both these compounds are positively correlated with
citrus juice flavor (41). They are responsible for the green notes in citrus juices.
Beta sinensal (38) is a rather high boiling point volatile, which was reduced
through heating. The aroma intensity of this aliphatic aldehyde was 42% higher in the
unheated reconstituted juice compared to that of the juice heated in glass and 36%
higher than the juice heated in metal. Beta sinensal possesses a grapefruity, citrus
(0 } iiiiiiiiiii,
0- --- In __
higher than the juice heated in metal. Beta sinensal possesses a grapefruity, citrusy
aroma and is one of the important compounds responsible for the fresh grapefruit juice
Aroma Active Compounds That Intensified after Heating
Eight aroma active compounds (1-octen-3-one (8), methional (11), (E,Z)-2,4-
nonadienal (16), (E,E)-2,4-decadienal (19), Furaneol (29), and 4-vinyl guiaicol (35) and
two unknowns (26 & 32) were present in intensities greater than 50% in comparison to
the unheated reconstituted juice (fig 5-5). Both Furaneol (29)(discussed later) and
methional (11) are thermally generated. Furaneol is formed from the reaction of
rhamnose and an amino acid in the presence of heat. Methional is believed to be a
Strecker degradation product of methionine in the presence of ascorbic acid. Its
normalized intensity in the two heated samples was 8 and 10 compared to 0.65 in the
unheated reconstituted juice. (E,Z)-2, 6-nonadienal (16) and (E,E)-2,4-decadienal (19)
are unsaturated aldehydes and are likely to be formed from fatty acid degradation (42).
Both of these aldehydes are essential flavor components of citrus oils. A faint
cucumber, green aroma was detected from (E,Z) 2,6 nonadienal while the aroma of
(E,E)-2,4-decadienal was described as fatty with a slight orange character. The two
unknown compounds (26 & 32) have a spicy and sweet, caramel aroma respectively.
Neither of the two unknowns was detected in the unheated reconstituted juice.
However the intensity of unknown # 26 in the juice heated in glass was 54% higher than
the juice heated in the metal. Unknown # 32 had similar aroma intensity in both the
heated samples. Their linear retention index values on wax were 2000 and 2163
respectively. Maltol, a Maillard reaction product elutes at an LRI close to 2000
possesses a caramel aroma and could be a possible identity. Identification of this
compound would be satisfied through the use of GC-MS and by matching the spectra of
the compound alongside that of the pure standard. However, maltol does not have a
unique mass spectrum and may be present in concentrations lower than the detection
threshold of the GC-FID or GC-MS, but not of the human nose.
Retention Time (min)
Figure 5-5 Aroma active volatiles (peak numbers 8, 11, 16, 19, 29, 26, 32, & 35) whose
intensity increased after heat treatment was applied.
Caramel Aroma Compounds Formed from Sugar Degradation Reactions
From table 5-3, it can be noted that caramel burnt aroma was detected several
times throughout the sniff run. In some cases the volatiles lingered for up to 20 sec. In
order to identify these compounds, several standards that are known to impart caramel
aromas were analyzed under similar conditions on both the wax and ZB-5 columns.
Table 5-4 lists the furans and a pyranone standard that were examined to identify the
unknown compounds responsible for the caramel aroma.
Table 5-4 LRI for ten sugar degradation products possessing caramel aroma analyzed
by GC on both polar (wax) and non-polar (ZB-5) columns.
Compounds LRI (ZB-5) Present (ZB-5) Present (Wax) LRI (DB-Wax)
Furfural 850 + 1215
5-methyl furfural 967 1560
3-methyl- 2(5H) furanone1 989 + + 2153
Furaneol1 1064 + + 2096
Norfuraneol 1054 2124
Mesifurane 1064 1602
Maltol 1114 -- 2002
Sotolone' 1118 + 2203
Homofuraneol 1145 2045
5-HMF' 1236 + 25261
1 Identified in all three samples
2 found in unheated reconstituted only
All the furans listed are oxygenated. Furaneol, mesifurane and norfuraneol elute in
close proximity on the ZB-5 column. Both Furaneol and mesifurane had identical
retention indexes values (1064) while norfuraneol had a linear retention index value of
1054. The ease of volatilization of a compound is one of several factors influencing GC
separation. Most of the furans are structurally similar and, would be expected to
volatilize easily. Also, column chemistry plays a key role in separation. Polar
compounds elute more quickly on a non-polar column but will be better retained on a
polar column. The rule "like attracts like" applies. On the ZB-5 column it was seen
that the 3(2H) furanones elute in close proximity to each other e.g. Furaneol and
mesifurane eluted at 1064, and norfuraneol eluted at 1054. The structures of these
compounds are shown below (fig. 5-6).
R3 0 R1
H3 O CH3
2,5 dimethyl-4-hydroxy-3(2H) furanone
H3C O CH3
4-methoxy-2,5- dimethyl-3(2H) furanone
H3C O CH3CH2
2 ethyl 4-hydroxy 5 methyl-3(2H) furanone
4 hydroxy-5-methyl-3(2H) furanone
Figure 5-6 Structure of 3(2H) furanones
Homofuraneol has an ethyl group as one of the side groups and this additional
increase in molecular weight and boiling point is probably the reason for its later elution
compared to the other furanones.
Sotolone (fig.5-7), a 2(5H) furanone, also has a molecular weight of 128 and
combined with the different position of the R groups could be responsible for the
decrease in its volatility.
H3C 0 0
Figure 5-7 Structure of sotolone, a 5(2H) furanone
The use of a dissimilar column such as the wax column is an excellent choice in
situations like these, as the order of elution will be different. By applying the "like
attracts like" principle, polar compounds would be retained for a longer time on the
polar wax column.
Caramel Compounds Found in Grapefruit Juice
Furaneol and sotolone were detected by GC-O while 5-HMF was detected by GC-
FID. Furaneol has previously been reported in grapefruit juice and has been known to
mask the flavor of orange juice (4;11). Its intensity was over 50% higher in the two
heat treated samples as opposed to the unheated reconstituted juice (table 5-3).
Furaneol is formed from rhamnose through the 2,3-enolization pathway of the Amadori
compound followed by dehydration and molecular rearrangement (22). The amino acid
arginine has been shown to be a necessary substrate for its formation under acidic
conditions. Sotolone has recently been reported as an off-flavor to citrus sodas
containing ethanol and ascorbic acid (28). Sotolone imparted a caramel, cooked flavor
to grapefruit juice. The 5-HMF has been reported in citrus juices and has been
suggested as a marker for thermal abuse. This compound is formed from the 1,2-
enolization pathway and can be polymerized to form colored pigments or other flavor
compounds. Unlike the other furanones, 5-HMF has a high threshold and consequently
does not have a direct aroma impact. However, it produced substantial FID response.
Furfural was detected by GC-MS. It has a unique fragmentation pattern with the major
ion found at mass 95, indicating the cleavage of a hydrogen atom. By using this
knowledge as well as selecting mass peaks 96 and 95, the spectra was deduced and then
matched with that of the standard at the expected retention time.
3-Methyl 2(5H) Furanone
A maltol degradation product, 3-methyl-2(5H) furanone (fig. 5-8), (43) not
previously reported in grapefruit juice was tentatively identified on both polar and non-
polar columns. Its LRI on a the non-polar column was 982 on and 2153 on the polar
column. It had a weak-moderate caramel aroma. It has been reported in mammee
3-methy 2(5H) furanone
Figure 5-8 Structure of the 3-methyl-2(5H) furanone
When the standard was analyzed on GC-MS, the retention time and linear
retention index suggested that it eluted with peak # II (fig. 5-9).
Commercial FCGJ undergoes at least two heat treatments, once during the
evaporation step and second after reconstitution. However, the untreated juice, which
was reconstituted grapefruit juice, underwent only the first heat treatment, which
involves heating temperatures as high as 93C. A high temperature such as this is
sufficient to induce the formation of these furanones. The heating of the juice in both
the glass and stainless steel containers may have increased the formation or
concentration of these compounds, but due to their extremely low threshold the aroma
remains quite intense.
The Maillard reaction is initiated by high temperature treatment of foods, and its
products are increased with extreme heat treatment and long-term storage. The first
step involves the reaction of a sugar with an amine and subsequent loss of water. There
is only a slight loss of reactants since sugars are relatively plentiful in grapefruit juice.
Identification of Volatiles Using GC-MS
Since MS is an identification tool, it was used to identify the volatiles based on
their specific fragmentation spectra as well as their linear retention indices. Most
compounds fragment in a unique pattern and by comparing this pattern with standards
analyzed under the same conditions or by comparing to library databases with known
spectra, compounds can usually be identified. Fig. 5-9 shows the total ion
chromatograph of grapefruit juice from 10-33 minutes while table 5-5 lists the identities
of a few of the major components in grapefruit juice. The areas were normalized to
internal standard #2.
Matching linear retention indices was used for a secondary confirmation. This is
particularly important in the detection of terpenes. Terpenes are C10H16 hydrocarbons
and have a molecular weight of 136. It is very difficult to identify these compounds by
solely matching mass spectra as they tend to exhibit similar fragmentation patterns
particularly for the major mass peaks.
Total Ion Chromatogram
v. 11 IX
10 16 20 24 28 32
Retention Time (min)
Figure 5-9 Segmented MS total ion chromatogram using ZB-5 column for unheated reconstituted grapefruit juice (A), juice heated in
glass (B) and juice heated in stainless steel (C).
glass (B) and juice heated in stainless steel (C).
Components identified in grapefruit juice by GC-MS
and analyzed on a DB-5
Peak Retention LRI Component Unheated Glass Stainless
Number Time Name reconstituted Steel
%Area %Area %Area
I 12.47 941 a-pinene 15.68 2.63 0.74
II 13.68 987 myrcene 45.93 7.52 4.66
III 14.93 1035 limonene 2512.60 611.60 364.36
IV 15.14 1043 E-ocimene 2.25 0.74 0.73
V 15.92 1073 linalool oxide 6.72 27.23 18.83
VI 16.31 1089 linalool 3.79 16.42 11.51
VII 19.05 1201 c-terpineol 10.57 20.67 11.74
VIII 23.19 1385 a-copaene 6.80 0.32 3.39
IX 24.24 1435 P- 30.21 1.73 12.61
X 25.55 1499 a-humulene 12.40 7.43 7.75
XI 26.00 1522 bisabolene 11.22 1.02 6.49
XII 26.14 1529 y-cadinene 1.43 1.77 0.96
XIII 31.44 1823 nootkatone 80.76 43.69 41.95
An example of this can be seen in peaks number I and III, which were identified as a-
pinene and myrcene respectively. Both these compounds are terpenes having piney and
musty aroma respectively. From the spectra of these two compounds, fragments 121,
91 and 77 (fig. 5-10) were common to both and appeared to have the same intensity.
Mass peaks at 92, 93 and 79 had slightly different intensities and may be used as
distinguishing fragments. However, by using the linear retention indices of 939 and
989 respectively, these two compounds could be easily distinguished.
I,~ .. .. I l Il
40 60 80 100 120 14(
Mass/Charge (m/z) ratio
Figure 5-10 Mass spectra of myrcene (top) and a-pinene (bottom) taken from a sample
analyzed on a DB-5 column.
Limonene (peak III) was present in the highest amount as expected. Limonene is the
most abundant terpene present in citrus juices. Peak VII was identified as a-terpineol.
Figure 5-11 is an expanded view of the chromatogram indicating the differences among
the three samples. Alpha terpineol content of juice heated in glass was higher than both
the stainless steel (metal) and unheated reconstituted from concentrate juice. It 63%
, I J I ,i
higher than the unheated reconstituted juice and 43% higher than the juice heated in
a-terpineol levels in
18.95 19.05 19.15 19.25
Retention Time (min)
Figure 5-11 Alpha-terpineol levels in the three samples, unheated reconstituted juice,
juice heated in glass and juice heated in metal
Alpha-terpineol can be formed from either (or both) the acid-catalyzed reaction of
limonene or linalool (fig. 5-12) in the presence of heat (45). Alpha terpineol has also
been shown to negatively correlate with the flavor of citrus juices (4) and has been
suggested as an indicator for over-processed or stored orange juices.
Figure 5-12 Formation pathway for a-terpineol in the presence of limonene and linalool
in grapefruit juice
HPLC RESULTS AND DISCUSSION
Unlike GC, HPLC is a non-thermal method of analysis used in the identification of
both volatiles and non-volatiles. HPLC is useful in situations where polar volatiles have
strong UV chromaphores and are difficult to extract using typical GC extraction
procedures. For example, Furaneol in grapefruit juice is usually found at levels below the
detection threshold of the flame ionization detector, making detection and quantification
difficult. Solid phase extraction employing C-18 chemically bonded silica, can trap polar
volatiles onto the modified silica for later analysis on reversed phase HPLC.
Since furans are thought to be responsible for heated off-flavor, HPLC was used to
identify and measure furans in grapefruit juice. Over the years reversed phase HPLC has
been favored in the isolation of furans in several fruit juices (19;21;46). Reversed phase
LC is the most popular liquid chromatographic method used today (47). Compounds
with increasing hydrophobicity or non-polar behavior will be retained on the column
more strongly. Water is the weak solvent and a methanol or acetonitrile gradient is
employed to selectively elute the materials off a reverse phase column.
In citrus juices several mobile phase compositions have been used ranging from
methanol or acetonitrile/water, acetonitrile/methanol/buffer, methanol/acetate buffer.
These solvent systems either did not produce a good separation or did not exhibit good
reproducibility. Acetate buffer has an absorbance maximum at 235nm, which interferes
with the spectra of sotolone whose maximum absorbance is 237nm. Although previous
studies (19) reported that methanol/phosphate buffer condition could not resolve furfural
and Furaneol, this mobile phase composition was re-examined since the phosphate buffer
had a lower UV absorbance background. It had been reported that 30%
methanol/phosphate buffer could not resolve furfural and Furaneol. Thus several
gradients were evaluated until a suitable separation was between 5-HMF, furfural, and
Furaneol was achieved. These three compounds had been previously been identified in
grapefruit juice using RP-HPLC (19). In order to increase retention time and promote
resolution, the amount of modifier, methanol was decreased to 5% with a slow increase
up to 30% in 27 minutes. It was found that at approximately 20% methanol Furfural was
eluted and Furaneol close to 24% methanol. Additionally, the temperature of the column
was kept at 250C and the sample 240C. Since the three compounds were completely
resolved, other Maillard reaction products, (mesifurane, homofuraneol, sotolone, and
maltol) were evaluated under similar conditions and the result shown in fig. 6-1. The
small humps at the end of the late eluting peaks were apparently due to a small void at the
head of the HPLC column. These standards are all products of the Maillard reaction and
possess caramel or burnt sugar aroma that could possibly be formed in grapefruit juice
during heating. These seven compounds along with their retention times and maximum
absorbance are listed in table 6-1.
Table 6-1 Retention time and maximum absorbance of furans standards on a C-18
column and using a methanol/phosphate buffer at flow rate of ImL/min
Furan Standards Retention Time Max Wavelength
5-HMF 13.04 283
Furfural 14.95 276
Maltol 17.41 274
Furaneol 18.86 287
Sotolone 20.24 235
Mesifurane 27.63 278
Homofuraneol A 30.40 287
Homofuraneol B 31.80 287
278nm a p
15 20 25 30
Retention Time (min)
Figure 6-1 HPLC spectra of the seven standards (5-HMF, furfural, Furaneol, sotolone,
mesifurane and homofuraneol) monitored from 230-380nm.
Homofuraneol has been known to exist in two tautomeric forms accounting for the
presence of an A and B form (48) and produces two distinct HPLC peaks. As can be
seen from fig. 6-2, homofuraneol has an ethyl group that can be interchanged with the
methyl group so that it can be either in the 2 or 5 position.
H3C 0 CH3CHCH CH3CH2
Figure 6-2 Homofuraneol tautomers, A- 2-ethyl-4-hydroxy-5-methyl-3(2H) furanone; B-
The value of the PDA detector lies in its ability to distinguish compounds based on
their spectra as well as retention time. This feature can be useful in identifying unknown
peaks and determining if there are co-eluted compounds in a single HPLC peak. With the
exception of sotolone, which has a maximum absorbance at 235nm, most furans exhibit
maximum wavelength absorbance between 270-290nm. For this reason dual wavelength
monitoring at 235nm and 280nm was carried out. Additionally, monitoring was done at
335nm as these furans had minimum absorbance at that wavelength. This would be
useful in distinguishing furans from other co-eluting compounds. A comparison of the
spectra and the retention times of components in the sample with those of standards,
suggest the presence of 5-HMF, furfural, and Furaneol in the grapefruit juice (fig. 6-3).
As discussed earlier, these three compounds are heat induced hexose sugar degradation
products that have been previously identified in citrus juices.
I I I I I I
13 14 15 16 17 18 19
Retention Time (min)
Figure 6-3 Segment of HPLC chromatogram showing overlay of unheated, juice heated
in glass and juice heated in metal and monitored at 290nm
By looking at the spectra of each of these compounds in the sample and comparing
it with that of the standard, it was found that 5-HMF was the only compound of interest
that was well resolved from other grapefruit juice components (fig. 6-4).
I I CO CO
220 240 260 280 300 320 340 360
5000- 0 g
'-' I I I '_ I I I0 I
220 240 260 280 300 320 340 360
Figure 6-4 Spectra of 5-HMF in standard (top) and in the grapefruit juice sample
(bottom) analyzed on C-18 column and monitored at 290nm.
Spectral data for both Furaneol and furfural peaks suggested the presence of
coeluted interfering compounds. Figure 6-5 shows the spectra of furfural and it can be
seen that the peak characteristics are different from the standard. An absorbance
maximum of 276nm was seen for the standard while a maximum of 326nm was seen for
the sample. Between 260 nm and 300nm, there appears to be another peak indicating that
co-elution might have occurred. It was not possible to obtain an absorbance "difference"
chromatogram using output from the two wavelengths (326nm and 278nm) at this time.
Previous studies had used this technique to identify and characterize Furaneol in a
grapefruit juice matrix (21).
240 260 280 300 320
S20000 / \
220 240 260 280 300 320 340 360
Figure 6-5 Spectra of furfural standard (top) and sample (bottom) showing a difference in
the wavelength maxima at 15.15 min
Grapefruit juice furan and furanones were quantified using the external standard
method, wherein different concentrations of standards were injected onto the LC in
triplicate. Area responses were then plotted against the amounts injected and linear
responses were obtained. The 5-HMF amounts ranged from 1.07ng to 21.4ng to ensure
that the range found in samples was covered. From calibration plot (fig. 6-6), the r2 value
which shows how well the data fit a linear regression was 0.9993, indicating good
Calibration Curve for 5-HMF
0 5 10 15 20 25
Figure 6-6 Linear calibration curve for 5-HMF analyzed in triplicate.
Based on the linear equation, the amount of the each furan in grapefruit juice samples
could be calculated. Juice heated in the presence of stainless steel had a 5-HMF
concentration of 0.47ng/ul, while the same juice heated in glass had 0.34ng/ul (table 6-2).
Unheated juice contained the least amount of 5-HMF (0.03ng/ul).
Table 6-2 Concentrations of 5-HMF, furfural and Furaneol in the three grapefruit juice
5-HMF (ng/ul) Furfural (ng/ul) Furaneol (ng/ul)
Unheated 0.03 0.40 0.22
Glass 0.34 0.25 0.08
Metal 0.47 0.25 0.08
A similar plot (fig.6- 7) was conducted for Furaneol with amounts ranging between 1.8ng
to 18ng and an r2 value of 0.92 was found. Unheated reconstituted juice had a
concentration of 0.22ng/ul while the two heated juice samples had levels of 0.08ng/ul.
Furaneol Calibration curve
5 10 15 20
Figure 6-7 Linear calibration curve for Furaneol at 290nm.
The standard curve (fig. 6-8) for furfural also had a good fit with an r2 value of 0.99
using a range of 2ng to 200ng. Concentration of furfural in the unheated reconstituted
juice was 0.40ng/ul while juices heated in glass and metal had 0.25ng/ul.
Furfural Calibration Curve
0 50 100 150 200 250
Figure 6-8 Linear calibration curve for furfural standard at 290nm.
All three compounds, 5-HMF, furfural and Furaneol are sugar degradation
products. Formation of 5-HMF occurs through the 1,2 enolization at low pH values in
the presence of a hexose sugar. The pH of grapefruit juice is usually between 2.9- 3.2,
which is highly acidic, and provides an explanation for the increase in the 5-HMF
formation in the two heated samples. Studies have shown when grapefruit juice that
under similar conditions the rate of formation of 5-HMF was far greater than that of
furfural (9). Furfural can be formed from acid catalysis of ascorbic acid in the presence
of an amine. Since it can be formed from ascorbic acid, which is known to degrade with
heating, it was expected that this compound would have higher concentrations in the two
heated samples. However, it is a very reactive compound and, in the presence of acids
can react with aldehydes, ketones and amino acids (49). This could possibly be a reason
for the lower levels in the heated samples.
Furaneol is also a sugar degradation product and its concentration was expected to
increase as a result of the heating. Its formation is favored at high pH and goes through
the 2,3-enolization pathway, however at low pH values it is formed in the presence of
alanine. Studies have shown that Furaneol is unstable and can be degraded at high
temperatures (24) to form highly reactive dicarbonyls, ketones and alcohols. In
equilibrium, Furaneol can exist in its open chain form, thereby facilitating attack of sulfur
groups such as thiols on the carbonyl (50). It can also react with sulfur compounds, such
as cysteine at low pH, and hydrogen sulfide during thermal processing (51;52).
Consequently, at high temperatures, the concentration of Furaneol would be lower than
No significant flavor difference was detected between juices heated 10 min in
contact with stainless steel and the same juice heated in glass. However, this excessive
heating induced a flavor change that was significantly different from the unheated
reconstituted grapefruit juice at the 95% confidence interval. Thermal processing
produced heated, cooked, pineapple and metallic off-flavors.
Similar volatiles were detected in both heat-treated samples, suggesting that high
processing temperatures and long times appeared to be major contributing factors for
cooked off flavor formation. The quality of the juice heated on glass surface was not of
superior quality to that of metal surface, therefore, the current practice of using stainless
steel is agreeable.
Intensities of compounds such as a-pinene and myrcene that are known to
positively contribute to fresh citrus juice character were diminished during the heating
process, while aroma intensity of heat induced compounds such a methional and Furaneol
increased during heating. Sugar degradation products such as sotolone, and 3-methyl-2
(5H) furanone were detected by GCO. The 3-methyl-2 (5H) furanone was tentatively
identified in grapefruit juice for the first time. Concentrations of 5-hydroxy methyl
furfural, another sugar degradation product, increased with heating.
LIST OF REFERENCES
1. Dugo, G.; Giacomo, A. D. Citrus, The Genus Citrus; Taylor & Francis: New York,
2. Offenbacher, E. G.; Pi-Sunyer, F. X. Temperature and pH effects on the release of
chromium from stainless steel into water and fruit juices. Journal ofAgricultural
and Food Chemistry 1983, 31, 89-92.
3. Woodroof, L. Commercial Fruit Processing; The AVI Publishing Company, Inc:
Westport, Connecticut, 1975.
4. Tatum, J. H.; Nagy, S.; Berry, R. E. Degradation products formed in canned single-
strength orange juice during storage. Journal of Food Science 1975, 40, 707-9.
5. Naim, M.; Rouseff, R. L.; Zehavi, U.; Schutz, O.; Halvera-Toledo, E. Chemical and
sensory analysis of off-flavors in citrus products. ACS Symposium Series 1998,
6. Nagy, S.; Randall, V. Use of furfural content as an index of storage temperature
abuse in commercially processed orange juice. Journal ofAgricultural andFood
Chemistry 1973, 21, 272-5.
7. Fellers, P. J.; Carter, R. D. Effects of thermal processing of chilled orange juice on
flavor quality. Food Processing 1993, 436-441.
8. Orange juice: How far from fresh? Consumer Reports, 1995.
9. Lee, H. S.; Nagy, S. Quality changes and nonenzymic browning intermediates in
grapefruit juice during storage. Journal ofFood Science 1988, 53, 168-172.
10. Kanner, J.; Harel, S.; Fishbein, Y.; Shalom, P. Furfural accumulation in stored
orange juice concentrates. Journal ofAgricultural and Food Chemistry 1981, 29,
11. Lee, H. S.; Nagy, S. Chemical degradative indicators to monitor the quality of
processes and stored citrus products. ACS Symposium Series 1996, 631, 86-106.
12. Scarpellino, R.; Soukup, R. J. Key flavors from heat reactions of food ingredients.
Flavor Sci. [Discuss. Flavor Res. Workshop] 1993, 309-35.
13. Rodin, J. O.; Himel, C. M.; Silverstein, R. M.; Leeper, R. W.; Gortner, W. A.
Volatile flavor and aroma components of pineapple. I. Isolation and tentative
identification of 2,5-dimethyl-4-hydroxy-2,3-dihydro-3-furanone. Journal ofFood
Science 1965, 30, 280-5.
14. Baek, H. H.; Cadwallader, K. R.; Marroquin, E.; Silva, J. L. Identification of
predominant aroma compounds in muscadine grape juice. Journal of Food Science
1997, 62, 249-252.
15. Krammer, G. E.; Takeoka, G. R.; Buttery, R. G. Isolation and identification of 2,5-
Dimethyl-4-hydroxy-3(2H)-furanone glucoside from tomatoes. Journal of
Agricultural and Food Chemistry 1994, 42, 1595-7.
16. Pickenhagen, W.; Velluz, A.; Passerat, J. P.; Ohloff, G. Estimation of 2,5-dimethyl-
4-hydroxy-3(2H)-furanone (FURANEOL) in cultivated and wild strawberries,
pineapples, and mangoes. Journal of the Science of Food and Agriculture 1981, 32,
17. Mayerl, F.; Noef, R.; Thomas, A. F. 2,5-Dimethyl-4-hydroxy-3(2H)-furanone
glucoside: isolation from strawberries and synthesis. Phytochemistry 1989, 28, 631-
18. Tonsbeek, C. H. T.; Plancken, A. J.; Weerdhof, T. v. d. Components contributing to
beef flavor. Isolation of 4-hydroxy-5-methyl-3(2H)-furanone and its 2,5-dimethyl
homolog from beef broth. Journal ofAgricultural and Food Chemistry 1968, 16,
19. Lee, H. S.; Nagy, S. High performance liquid chromatography of 2,5-dimethyl-4-
hydroxy-3(2H)-furanone in pineapple and grapefruit juices. Journal of Food
Science 1987, 52, 163-5.
20. Pyysalo, T.; Honkanen, E.; Hirvi, T. Volatiles of wild strawberries, Fragaria vesca
L., compared to those of cultivated berries, Fragaria ananassa cv Senga Sengana.
Journal ofAgricultural and Food Chemistry 1979, 27, 19-22.
21. Walsh, M.; Rouseff, R.; Naim, M. Determination of furaneol and p-vinylguaiacol
in orange juice employing differential UV wavelength and fluorescence detection
with a unified solid phase extraction. Journal ofAgricultural andFood Chemistry
1997, 45, 1320-1324.
22. Pisarnitskii, A. F.; Demechenko, A. G.; Egorov, I. A.; Gvelesiani, R. K.
Methylpentoses are possible precursors of furanones in fruits. Applied Biochemistry
and Microbiology 1992, 28, 97-100.
23. Romero, C.; Manjon, A.; Bastida, J.; Iborra, J. L. A method for assaying the
rhamnosidase activity of naringinase. Analytical Biochemistry 1985, 149, 566-71.
24. Shu, C. K.; Mookherjee, B. D.; Ho, C. T. Volatile components of the thermal
degradation of 2,5-dimethyl-4-hydroxy-3(2H)-furanone. Journal ofAgricultural
and Food Chemistry 1985, 33, 446-8.
25. Guichard, E.; Pham, T. T.; Etievant, P. Quantitative determination of sotolon in
wines by high-performance liquid chromatography. Chromatographia 1993, 37,
26. Blank, I.; Lin, J.; Fumeaux, R.; Welti, D. H.; Fay, L. B. Formation of 3-Hydroxy-
4,5-dimethyl-2(5H)-furanone (Sotolone) from 4-Hydroxy-L-isoleucine and 3-
amino-4,5-dimethyl-3,4-dihydro-2(5H)-furanone. Journal ofAgricultural andFood
Chemistry 1996, 44, 1851-1856.
27. Hognadottir, A.; Rouseff, R. L. Identification of aroma active compounds in orange
essence oil using gas chromatography-olfactometry and gas chromatography-mass
spectrometry. Journal of Chromatography, A 2003, 998, 201-211.
28. Koenig, T.; Gutsche, B.; Hartl, M.; Huebscher, R.; Schreier, P.; Schwab, W. 3-
Hydroxy-4,5-dimethyl-2(5H)-furanone (sotolone) causing an off-flavor: elucidation
of its formation pathways during storage of citrus soft drinks. Journal of
Agricultural andFood Chemistry 1999, 47, 3288-3291.
29. Lin, J.; Rouseff, R. L.; Barros, S.; Naim, M. Aroma composition changes in early
season grapefruit juice produced from thermal concentration. Journal of
Agricultural and Food Chemistry 2002, 50, 813-819.
30. Friedrich, J. E.; Acree, T. E. Issues in gas chromatography-olfactometry
methodologies. ACS Symposium Series 2000, 756, 124-132.
31. Mistry, B. S.; Reineccius, T.; Olson, L. K. Gas chromatography-olfactometry for
the determination of key odorants in foods. Food Science and Technology 1997, 79,
32. Mussinan, C. J. Instrumental analysis in the flavor industry. In Flavor Sci.
[Discuss. Flavor Res. Workshop], 1993; pp 169-224.
33. Marcy, J. E.; Rouseff, R. L. High-performance liquid chromatographic
determination of furfural in orange juice. Journal ofAgricultural andFood
Chemistry 1984, 32, 979-81.
34. Lee, H. S.; Rouseff, R. L.; Nagy, S. HPLC determination of furfural and 5-
hydroxymethylfurfural in citrus juices. Journal of Food Science 1986, 51, 1075-6.
35. Huber, L. Optimization of diode array detection for sensitivity, selectivity,
wavelength resolution, and linearity. Chromatographic Science Series 1993, 62,
36. Martin, B.; Etievant, P. X.; Le Quere, J. L.; Schlich, P. More clues about sensory
impact of sotolon in some flor sherry wines. Journal ofAgricultural andFood
Chemistry 1992, 40, 475-8.
37. Girard, B. Retention index calculation using Kovats constant model for linear
temperature-programmed gas chromatography. Journal of Chromatography, A
1996, 721, 279-88.
38. Meilgaard, M.; Civille, G.; Carr, B. T. Sensory Evaluation Techniques, 3rd ed.;
CRC Press: Boca Raton, FL, 1999.
39. Ott, R L; Longnecker, M. An Introduction to Statistical Methods and Data
Analysis, 5th ed.; Duxbury Thompson Learning: Pacific Grove, CA, 2001.
40. Buettner, A.; Schieberle, P. Characterization of the most odor-active volatiles in
fresh, hand-squeezed juice of grapefruit (Citrus paradisi Macfayden). Journal of
Agricultural and Food Chemistry 1999, 47, 5189-5193.
41. Ahmed, E. M.; Dennison, R. A.; Shaw, P. E. Effect of selected oil and essence
volatile components on flavor quality of pumpout orange juice. Journal of
Agricultural and Food Chemistry 1978, 26, 368-72.
42. Lin, J.; Rouseff, R. L. Characterization of aroma-impact compounds in cold-
pressed grapefruit oil using time-intensity GC-olfactometry and GC-MS. Flavour
and Fragrance Journal 2001, 16, 457-463.
43. Kim, M.-O.; Baltes, W. On the role of 2,3-Dihydro-3,5-dihydroxy-6-methyl-4(H)-
pyran-4-one in the Maillard reaction. Journal ofAgricultural and Food Chemistry
1996, 44, 282-9.
44. Morales, A. L.; Duque, C. Free and glycosidically bound volatiles in the mammee
apple (Mammea americana) fruit. European Food Research and Technology 2002,
45. Clark, B. C., Jr.; Chamblee, T. S. Acid-catalyzed reactions of citrus oils and other
terpene-containing flavors. Developments in Food Science 1992, 28, 229-85.
46. Yuan, J.-P.; Chen, F. Separation and identification of furanic compounds in fruit
juices and drinks by high performance liquid chromatography photodiode array
detection. Journal ofAgricultural and Food Chemistry 1998, 46, 1286-1291.
47. Shugar, G. J.; Dean, J. A. The Cheini\'s Ready Reference Handbook; McGraw-
Hill: New York, NY, 1990.
48. Huber, U. A. Homofuronol: a powerful tool to prepare and improve sophisticated
high quality flavors. Perfumer & Flavorist 1992, 17, 15-19.
49. Kanner, J.; Fishbein, J.; Shalom, P.; Harel, S.; Ben-Gera, I. Storage stability of
orange juice concentrate packaged aseptically. Journal of Food Science 1982, 47,
50. Haleva-Toledo, E.; Naim, M.; Zehavi, U.; Rouseff, R. L. Effects of L-cysteine and
N-acetyl-L-cysteine on 4-hydroxy-2,5-dimethyl-3(2H)-furanone (furaneol), 5-
(hydroxymethyl)furfural, and 5-methylfurfural formation and browning in buffer
solutions containing either rhamnose or glucose and arginine. Journal of
Agricultural andFood Chemistry 1999, 47, 4140-4145.
51. Shu, C. K.; Ho, C. T. Effect of pH on the volatile formation from the reaction
between cysteine and 2,5-dimethyl-4-hydroxy-3(2H)-furanone. Journal of
Agricultural and Food Chemistry 1988, 36, 801-3.
52. Van den Ouweland, G. A. M.; Peer, H. G. Components contributing to beef flavor.
Volatile compounds produced by the reaction of 4-hydroxy-5-methyl-3(2H)-
furanone and its thio analog with hydrogen sulfide. Journal ofAgricultural and
Food Chemistry 1975, 23, 501-5.
Wendy Bell was born in Westmoreland, Jamaica, and graduated from the Wolmer's
High School for Girls in Kingston. In 1995, she moved to the United States and
embarked upon her undergraduate studies at Barry University, Miami, Florida. She then
transferred to the University of South Florida, Tampa, Florida, where she completed her
Bachelor of Arts in Chemistry in May 1999. In pursuance of her career, Wendy gained
employment at Pasco Beverage Company and Florida Department of Citrus.
Wendy will receive her master's degree in December 2004 and will be joining
Cadbury Adams, Morris Plains, NJ, upon graduation.