Applications of organized media in chromatography and flow injection analysis

MISSING IMAGE

Material Information

Title:
Applications of organized media in chromatography and flow injection analysis
Physical Description:
xvii, 242 leaves : ill. ; 28 cm.
Language:
English
Creator:
Brooks, Stephen Houghton, 1961-
Publication Date:

Subjects

Subjects / Keywords:
Surface active agents   ( lcsh )
Micelles   ( lcsh )
Liquid chromatography   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1988.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Stephen Houghton Brooks.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001116919
notis - AFL3704
oclc - 19915168
System ID:
AA00003362:00001

Full Text











APPLICATIONS OF ORGANIZED MEDIA IN
CHROMATOGRAPHY AND FLOW INJECTION ANALYSIS












BY

STEPHEN HOUGHTON BROOKS


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

UNIVERSITY OF FLORIDA

1988


.r WSITY OF FLORIOX LUBRA1U

























This dissertation is dedicated to my grandfather,

Alvin Newman Dugar,

whose love and direction were such important factors

in the completion of this work.







Nemo vir magnus sine aliquo adflatu divino umquam fuit.

(No man was ever great without some portion of divine inspiration.)



Cicero

De Natura Deorum

Book ii, Chapter 66, Section 167




























To Maruja:


There is much satisfaction in work well done;

praise is sweet;

but there can be no happiness equal to the joy

of finding a heart that understands.



Victor Robinson

William Godwin, The Truth Seeker

6 January, 1906
















ACKNOWLEDGMENTS


There are so many people whom I need to thank for making this

achievement possible. Most immediately, I wish to thank my mentor,

John Gordon Dorsey, for his patience, direction, insight and exquisite

taste in food and wine. They have all greatly contributed to my

educational experience. My only regrets at the end of my tenure are

that John still insists upon pledging his allegiance to that semblance

of a baseball team which calls Fulton County Stadium its home and that

despite substantial trial and error (certainly the trials have far

outweighed the errors), he was unable to eclipse the NBS Depth Charge

standard of 1985 points. I did, however, have the privilege of

witnessing a 1948 (bringing to mind an ancient Chinese proverb) and

certainly, MGK once again breathes a sigh of relief.

Professionally, I wish to thank Alain Berthod, Maria A.

Hernandez Torres, Dan Leff, Rick Williams and Barbara Kirsch for their

contributions to this dissertation and wish them continued success in

their lives (especially Maruja!). Thank you goes to my typist, Cindy

Zimmerman, for her prompt and accurate preparation of this work.

Special thanks go to Danny Coffman for his outstanding draftsmanship

throughout the years.

I feel that this accomplishment would have been impossible

without the guidance and contribution of Mr. Sibson, Mr. Parr,

Mr. Boothby, Mr. Bolduc, Mr. Pinette, Mr. Thibault, Mr. Sharp,










Mr. Jackson, Mr. Weatherbie and the rest of the outstanding public

high school teachers who provide the foundation for achievements such

as this.

Special thanks are offered to the Dorsey group members, those

associated with the Analytical Division, residents of the Thunderdome,

Florida Pest Control, Dub Thomas, Stranger, Dorian Gray, Scarlet, the

Purple Porpoise, the Gator Sports programs, Jack Hairston, Skeeter's

flying biscuits, bouncing babies, warm winters, air conditioning, and

Walt Disney for making my four years in Gainesville a thoroughly

enjoyable experience.

Thanks go to Maruja for her love and support during these

years. Thanks go to all the members of my family who were so

instrumental in my continued perserverance, Gertrude Dugar, Bertha

Brooks, Rich, Kathy, Aunt Betty, Uncle Bill and Aunt Mary.

In keeping with tradition, I have chosen to save the greatest

thanks for last. I wish to thank my parents, Mr. and Mrs. Richard H.

Brooks, for their tireless contribution to my health and welfare from

day one until the present. In an everchanging world, all too often

dominated by things beyond our control, it has always been uplifting

to realize that there was the constancy of their love and support for

me to lean on.
















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS....................................................ii

LIST OF TABLES.............................................. ......viii

LIST OF FIGURES...................................................xi

ABSTRACT ..................... ....... .............. ........... .....xvi

CHAPTERS

I INTRODUCTION.......... ..................... ............. 1

Surface Active Agents...................................
Micelles and Micellization.......................4
Analytical Applications of Surfactant Containing
Solutions ............................. ... .... .... .. 11
Flow Injection Analysis ...............................13
Micellar Reversed Phase Liquid Chromatography............24

II FLOW INJECTION ANALYSIS DETERMINATION OF CRITICAL
MICELLE CONCENTRATIONS OF IONIC AND NONIONIC
SURFACTANTS ...... ...................................31

Background...................... .......... ..... .31
Theoretical Study of Critical Micelle Concentration
Determination by Flow Injection Analysis...............33
Experimental Section......................................48
Results and Discussion...................... ......... 68

III COMPARISON OF AQUEOUS, MICELLAR AND MICROEMULSION
CARRIERS IN FLOW INJECTION ANALYSIS: THE BASE
HYDROLYSIS OF ACETYLSALICYLIC ACID......................95

Introduction............................. ........... 95
Experimental.................................. .....96
Results and Discussion..................................97

IV THE SECOND MOMENT AS A DIRECT MEASURE OF DISPERSION
IN FLOW INJECTION SYSTEMS.............................. 132

Introduction...........................................1 32










Experimental ..... ........ .... ....................... .... 139
Results and Discussion..................................145

V EFFICIENCY AND SOLVENT STRENGTH IN MICELLAR LIQUID
CHROMATOGRAPHY: EFFECTS OF MEDIUM CHAIN LENGTH
NORMAL ALCOHOLS....................................... 167

Introduction........................................... 167
Experimental..... ............................... 169
Theory ................................................ 175
Results and Discussion..................................177

VI CONCLUSIONS AND FUTURE WORK ...........................214

APPENDICES

A TABULATION OF CHROMATOGRAPHIC FIGURES OF MERIT FOR
MICELLAR MOBILE PHASES.................. ............. 223

B BINDING CONSTANTS MEASURED FROM EQUATIONS 5.2 AND
5.3 FOR SDS/ALCOHOL MOBILE PHASES WITH A C18
COLUMN................. .................... ..... 230

REFERENCES.............. ........................... ...... 233

BIOGRAPHICAL SKETCH................................................ 242
















LIST OF TABLES


Table Page

I Typical surfactants with their CMC values and
aggregation numbers.......................................3

II Solubility data for arenes in water and micellar
solutions................................................ 12

III Numerical value of the parameters used to draw the
curves of Figures 8, 9, 11 and 12.........................45

IV Variation of relative viscosity (vs. 200C) and slopes
of log K vs. time as the weight percentage of ethylene
glycol is varied from 0-48% in the injected solution......73

V Determination of the efflux time for various solutions
in a #75 Ostwald capillary viscometer......................75

VI Investigation of the sample injection process..............82

VII A comparison of the theoretical and empirical sample
injection processes, vi = 10 uL .......................... 84

VIII Values of CMC for various surfactants.....................92

IX Reaction rate of acetylsalicylic acid (ASA)
hydrolysis in various media.............................. 117

X Figures of merit for the base catalyzed hydrolysis
of acetylsalicylic acid: wavelength of detection,
298 nm; volume of injection, 20 VL; flow rate,
1.0 mL/min; temperature, 250C..........................120

XI Dispersion values in aqueous and microemulsive
systems............................... .............125

XII Examination of peak broadening under reactive and
nonreactive conditions: wavelength of detection,
298 nm; volume of injection, 20 iL; injected sample,
4.92 x 10 M .................................... ...... 127


viii










XIII The effect of flow rate upon dispersion (D), area,
% deviation of calculated areas (% Dev.), second
moment (M2) and standard deviation of the underlying
Gaussian peak (oG) in a flow injection system............149

XIV Reproducibility of the second moment determination.......152

XV Effect of changing concentration and changing
sensitivity on the determination of the second moment....153

XVI Normalized desorption rate constants determined
for the investigated solutes in a 0.10 M SDS and a
0.25 M SDS mobile phase in the absence and presence
of 3% normal alcohols..................................181

XVII Adsorption rate constants determined for the
investigated solutes in a 0.10 M SDS and a
0.25 M SDS mobile phase in the absence and presence
of 3% normal alcohols.................................. 184

XVIII Calculated diffusion coefficients of investigated
solutes .......,** ... .................. ........ .......... 185

XIX A comparison of chromatographic efficiencies for
the investigated solutes in a 0.10 M SDS mobile phase
in the absence and presence of 3% normal alcohols and
a 60:40 methanol:water mobile phase.....................188

XX Strength of micellar mobile phases......................192

XXI A comparison of retention volumes for investigated
solutes in 0.10 M and 0.25 M SDS mobile phases in
the presence and absence of 3% normal alcohols...........194

XXII A comparison of the number of solute molecules and
micellar aggregates present in a peak volume for a
0.25 M SDS solution with 3% pentanol....................196

XXIII Physicochemical properties of SDS solutions in
the absence and presence of 3% normal alcohols...........198

XXIV Sodium dodecyl sulfate (SDS) concentrations used
in previous applications of the pseudophase
retention model...........................................200

XXV Solute-micelle binding constants per surfactant
monomer for the investigated solutes determined
from the three-phase retention model....................203










XXVI A comparison of the variation of K2, Pw and Ps
for acetophenone in 3% pentanol as a result of
changes in the value of stationary phase volume
in equation 5.2.........................................204

XXVII A comparison of the micelle/water (P ) and
stationary phase/water (Psw) partition coefficients
for investigated solutes in 3% alcohol solutions.........206

XXVIII The five principal vibronic bands in pyrene
monomer fluorescence........ .......................... 208

XXIX Study of wavelength of maximum emission (max)
and fluorescence intensity (I) of vibronic bands
I and III in various media.............................. 209
















LIST OF FIGURES


Figure Page

1 The relation between monomeric concentration in
solution and total added concentration of surfactant
with micelle formation.....................................6

2 A two-dimensional representation of the regions
of a spherical ionic micelle...........................,...7 7

3 Changes in various physicochemical properties of
an aqueous solution containing sodium dodecyl sulfate
as the concentration of surfactant in solution is
varied...................... ............... ............. 9

4 A calibration curve illustrating the change in slope
of the conductance versus concentration curve observed
at the CMC for hexadecyltrimethylammonium bromide (CTAB)...10

5 The dispersion of a sample plug in an FIA system............19

6 Velocity profiles, shape of injected sample bolus
and types of transport in closed tubes: (a) laminar
flow, parabolic velocity profile; (b) sample dispersion
due to laminar flow without diffusion; (c) convective
transport of sample molecules parallel to direction
of flow; (d) diffusional transport in the axial and
radial directions; (e) sample dispersion due to
laminar flow with diffusion...............................23

7 The two principal solute equilibria in a micellar
chromatographic system, k+ and k are,_ respectively,
the entrance and exit rate constants (s ) of the
solute to and from the micelle; k and k are,
respectively, the adsorption and lesorpt'on rate
constants (s ) of the solute to and from the
chromatographic stationary phase..........................26

8 Theoretical curves obtained for SDS: (1) conductimetric
detection (equations 2.13 and 2.14); (2 and 3)
spectroscopic measurements with a dye in the carrier
stream, P = 100 and 10,000, respectively (equations
2.19 and 2.20); (4 and 5) spectroscopic measurements










with the dye dissolved in the injected micellar solution,
P = 100 and 10,000, respectively (equations 2.7, 2.19
and 2.20) .................. .. .... .. ................42

9 Theoretical conductimetric curves obtained with
different time constants....................... ...... 43

10 Variation of the determined CMC of CTAB (using
equation 2.24) with flow rate..............................44

11 Theoretical absorbance curves with the dye
solubilized in the carrier stream with partition
coefficients equal to 100 and 10,000.......................46

12 Theoretical absorbance curves with the dye
solubilized in the injected surfactant solution............47

13 Ionic surfactants (with molecular formulas) under
investigation ........................... .... ......49

14 Structural formula of Coomassie Brilliant Blue R 250
(CBBR) *.. *....*.....* ....... *............................ 51

15 Visible spectra o 4.34 x 10-5 M CBBR in (A) aqueous
and (B) 2.5 x 10 M Brij 35 solution......................53

16 Visible spectra o5 4.57 x 10-5 M CBBR in (A) aqueous
and (B) 4.2 x 10 M Triton X-100 solution.................55

17 Nonionic surfactants (with molecular formulas) under
investigation.............................................. 56

18 (A) FIA manifold for the determination of the CMC
of ionic surfactants and (B) the analogous FIA manifold
for the determination of the CMC of nonionic
surfactants with CBBR dissolved in the carrier stream......57

19 Gradient chamber with arrows indicating the carrier
stream direction of flow: (A) teflon 0-ring, (B)
connecting screw, and (C) magnetic stirring bar............59

20 Detector response to the injection of an ionic
surfactant (CTAB) recorded at a chart speed of
(A) 1 cm/min and (B) 10 cm/min...........................63

21 Detector response to the injection of a nonionic
surfactant (Triton X-100) into an aqueous carrier
stream containing Coomassie Brilliant Blue
R 250 (1.3 x 10 ) .................. ....................67










22 Verification of the exponential character of the
FIA manifold by injection of 2.61 x 10 M Eriochrome
Black T with 437 nm detection............................71

23 Effects of increasing solution viscosity upon the
exponential character of the FIA manifold..................72

24 A plot of t vs. 0.693/Q for the determination of
1/2
the effective mixing volume of the conductance FIA
manifold: t2 (min) is the time required for any
1/2
concentration of an exponential concentration profile
to decrease to one half of its initial value, Q is
the flow rate (mL/min) ..................................... 77
-3
25 The UV-visible spectra of 1.25 x 10 M Eriochrome
Black T after aqueous dilution by factors of 10, 50
and 100.................................................... 80

26 The sample injection process: (A) a theoretical
sample plug injection, (B) the actual injection
profile corresponding to (A), and (C) the plug with
the t. value actually used in equation 2.23...............86

27 Computer-generated curves modelling conductance
detection in an FIA system................................ 89

28 Computer-generated curves modelling the use of a dye
as a micelle-tracer in an FIA system: (A) the dye is
solubilized in the carrier stream and (B) the dye is
solubilized within the injected surfactant solution........91

29 The base hydrolysis of acetylsalicylic acid...............100

30 Absorbance versus wavelength (nm) in aqueous media
following the progress of the acetylsalicylic acid
hydrolysis.............................. .............. 102

31 Absorbance versus wavelength (nm) in a 3.38 x 10- M
CTAB micellar system following the progress of the
acetylsalicylic acid hydrolysis..........................104

32 Absorbance versus wavelength (nm) in 99.28% (by
weight) water:0.36% CTAB (9.88 x 10 M):0.36% butanol
microemulsion system following the progress of the
acetylsalicylic acid hydrolysis..................... .........106

33 Change in maximum absorbance (298 nm) versus time (s)
in aqueous media for the detection of salicylic acid......109


xiii










34 Change in maxim m absorbance (298 nm) versus time (s)
in a 3.38 x 10 M CTAB micellar system for the
detection of salicylic acid..............................111

35 Change in maximum absorbance (298 nm) versus time (sJ
in a 99.28% (by weight) water:0.36% CTAB (9.88 x 10 M):
0.36% butanol microemulsion system for the detection
of salicylic acid .........................................113

36 Log (A A ) versus time (min) for the detection
of salicylic acid in an aqueous system...................114

37 Log (A A ) versus time (min for the detection of
salicylic acid in a 3.38 x 10 M CTAB micellar system....115

38 Log (A A ) versus time (min) for the detection of
salicylic a id in a 99.28% (by weight) water:0.36% CTAB
(9.88 x 10 M):0.36% butanol microemulsion system........116

39 The flow injection manifold employed for the
acetylsalicylic acid determination....................... 119

40 Response signals for the base-catalyzed hydrolysis
of acetylsalicylic acid in an aqueous (A) and
microemulsion (B) carrier stream (98.8% (by weight)
water:0.6% CTAB:0.6% butanol), NaOH 1 x 10 M. ..........123

41 A comparison of acetylsalicylic acid sample profiles
for plug dispersion in an aqueous (---) and
microemulsion (98.8% (by weight) water:0.6% CTAB:0.6%
butanol) carrier stream.......................... ....... 129

42 Measurement of peak width, W, and asymmetry factor,
b/a, at 10, 25, 50, and 75% of peak height on an
FIA response curve........................................142

43 Plot of dispersion versus flow rate for a 20 yL
sample of 0.10 M Nal injected into a 100 cm straight
tubing manifold.................. ................... 146

44 Plot of peak area (as determined by equations 4.5-4.8)
versus flow rate for a 20 uL sample of 0.10 M Nal
injected into a 100 cm straight tubing manifold...........147

45 Plot of second moment (as determined by equation 4.9)
versus flow rate for a 20 UL sample of 0.10 M Nal
injected into a 100 cm straight tubing manifold...........151










46 Plot of Gaussian contribution as a percent of the
total peak variance (as determined by equations 4.12)
versus flow rate for a 20 uL sample of 0.10 M Nal
injected into a 100 cm straight tubing manifold...........154

47 Plot of second moment versus flow rate for the
t-amyl iodide hydrolysis (straight manifold)...............156

48 Plot of second moment versus flow rate for the
t-amyl iodide hydrolysis (coiled manifold)................157

49 Plot of response curve shapes as a function of
elution time and flow rate in a straight tubing flow
injection system with the direction of + indicating
direction of increase in parameter........................161

50 Plot of second moment versus flow rate for a 20 uL
sample of 0.10 M Nal injected into an aqueous carrier
stream through a 100 cm manifold..........................164

51 Plot of second moment versus flow rate for a 20 yL
sample of 0.10 M NaI injected into a high viscosity
80:20 ethanol:water (v/v) carrier stream through
a 100 cm manifold........................................ 165

52 Excitation (a) and emission (b) spectra of pyrene
in a 60:40 methanol:water (v/v) solution..................172

53 Emission spectra of pyrene in a 0.10 M SDS solution
containing 3% (v/v) propanol (a), butanol (b),
pentanol (c) and hexanol (d)..............................212















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

APPLICATIONS OF ORGANIZED MEDIA IN
CHROMATOGRAPHY AND FLOW INJECTION ANALYSIS

BY

STEPHEN HOUGHTON BROOKS

August, 1988

Chair: John G. Dorsey
Major Department: Chemistry

The precision, speed and instrumental simplicity of a flow

injection analysis (FIA) system are combined with a gradient chamber

and flow-through conductance and absorbance detection to produce a

system for the rapid, accurate determination of the critical micelle

concentrations (CMCs) of ionic and nonionic surfactants. The

theoretical basis of the method is presented and the validity of the

technique is verified by a determination of the CMC values for both

ionic (CTAB, CTAC, SDS) and nonionic (Brij 35, Brij 56, Brij 99,

Triton X-100) surfactants.

The application of surfactant solutions as carrier streams in

flow injection analysis is examined. The reaction rates are

determined for the base catalyzed hydrolysis of acetylsalicylic acid

in aqueous, micellar and microemulsion solutions. The performance of

an aqueous and microemulsion carrier stream is compared. The loss in










relative sensitivity when employing the microemulsion can be traced to

the increased dispersion in the more viscous carrier stream.

Dispersion (D) is the most popular peak descriptor in FIA.

Unfortunately, D yields no direct information describing peak shape

and no information in the time domain. Using previously derived

equations, we examine the second moment (variance) of FIA peaks and

use this as a fundamental descriptor of the FIA response curves.

Unlike dispersion, the second moment is shown to obey a linear

relationship with respect to flow rate and to yield valuable

information in the presence of a chemical reaction.

The effects of the addition of 3% medium chain length (propanol-

hexanol) normal alcohols to SDS micellar mobile phases upon

chromatographic efficiency (N) and solvent strength (E) are

investigated. The solutes examined indicate that n-butanol and n-

pentanol are far superior as mobile phase additives with respect to

both N and eo when compared to n-propanol. The validity of the

pseudophase retention model of micellar liquid chromatography is

addressed and is employed for a direct examination of changes in

physicochemical properties (binding constants, partition coefficients)

with variation of alcohol chain length. Application of the random

walk model yields solute adsorption/desorption rate constants across

the stationary phase-mobile phase interface and confirms that

inefficient mass transfer across this interface is responsible for

decreased efficiencies observed in micellar liquid chromatography.


xvii
















CHAPTER I
INTRODUCTION


Surface Active Agents

A surface active agent or surfactant is an amphiphilic molecule,

containing distinct hydrophobic and hydrophilic regions. When

surfactant is present in a system, it adsorbs on surfaces or

interfaces of the system. An interface is defined (Rosen, 1978) as a

boundary between any two immiscible phases while a surface

specifically refers to a boundary where one phase is gaseous, in most

cases the gaseous phase is air. Interfacially adsorbed surfactants

thus have the ability to change the free energies of surfaces and

change the minimum work required to create or expand interfaces.

Measurement of the surface tension of surfactant containing solutions

allows the determination of the interfacial free energy per unit area

of a system. Most commonly, the interfacial free energy of a system

is decreased by the addition of surfactants.

Surfactants are generally classified as ionic or nonionic,

according to the nature of the hydrophilic headgroup. In general the

hydrophobic portion of the surfactant displays much less structural

variation than the polar headgroup. However, the nonpolar tail may be

a straight or branched alkyl chain, may contain unsaturated or

aromatic portions and may include perfluorinated or polysiloxane

groups (Rosen, 1978). Typically, the hydrophobic tail contains eight






2



to twenty hydrocarbon units in length. Shorter chain lengths are too

soluble to display surface activity while chain lengths of greater

than twenty carbon units are too insoluble for aqueous use. Table I

provides a list of common surfactants categorized by charge type.

Ionic surfactants can be further classified by headgroup as cationic,

anionic or zwitterionic.

Cationic surfactants have the general formula R X Y with R
n
representing the hydrocarbon chain. Nitrogen containing surfactants
+ -
(RnN Y ) are the most common cationic surfactants, due to their ease

of preparation and long term stability. Cationic surfactants

represented approximately 63% of the total 1985 U.S. output of 5.4

billion pounds (Chem. Eng. News, 1987).

The most common anionic surfactants are alkali or alkaline earth

metal salts of carboxylic, phosphoric, sulfonic or sulfuric acids

containing a saturated or unsaturated hydrocarbon moiety. These

surfactants have the general formula R X Y and are typically
n
manufactured by the hydrolysis of fats followed by neutralization with

a hydroxide to form the desired salt. Anionic surfactants accounted

for 8% of the 1985 U.S. output (Chem. Eng. News, 1987). Zwitterionic

surfactants contain both an anionic and a cationic group on the

hydrocarbon chain. Depending upon the pH of the solution and the

structure of the surfactant, the hydrophobic chain can possess

anionic, cationic or neutral characteristics. The most common

zwitterionic surfactants are members of the alkyl betaines and alkyl

sultaines which have the general structures of R N R2CO2 and

R1N R2S03, respectively.









Table I. Typical surfactants with their CMC values and aggregation
numbers.


Aggregation
Surfactanta CMC (MM) Number


Anionic
Sodium dodecyl sulfate (SDS) 8.1 62
CH (CH2)11 OS03Na

Potassium perfluoroheptanoate 30 b
C7F15C00 K

Cationic
Cetyltrimethylammonium bromide (CTAB) 1.3 78
CH (CH2)15N (CH )3Br

Cetylpyridinium chloridec 0.12 95
C16H 33N C5H5C1

Nonionic
Polyoxyethylene(6)dodecanol 0.09 400
CH (CH2)11(OCH2CH2 6OH

Polyoxyethylene(23)dodecanol (Brij 35) 0.1 40
CH3 (CH) 11(OCH2CH )23OH

Zwitterionic
N,N-dimethyl-N-(carboxymethyl)octyl- 250 24
ammonium salt (Octylbetaine)
CgH17N (CH )2CH2COO

N-dodecyl-N,N-dimethylammonium-3- 3 55
propane-1-sulfonic acid (SB-12)
CH (CH) 11N (CH ) SO


a Values for aqueous solution at 250C.

bNot available.
Not available.


c In 0.0175 M NaCl.











Nonionic surfactants accounted for approximately 29% of the 1985

U.S. commercial output (Chem. Eng. News, 1987). The most common

nonionic surfactants are polyoxyethylene and polyoxypropylene

derivatives. In the commercial preparation of these polymeric

surfactants (Garrett, 1972), typical reactants are an alcohol and

ethylene oxide. The initial reaction of the alcohol with ethylene

oxide produces a small amount of the monoether of glycol:


R OH + H COCH2 -> R 0 CH2 CH2 OH.


The resulting primary alcohol from this reaction has equivalent

reactivity toward additional ethylene oxide as the original alcohol,

R OH. Therefore, as the reaction proceeds, some monoalkyl diglycol,

monoalkyl triglycol, etc. will be produced. As a result, the final

product contains a series of monoalkyl glycols; the proportion of each

present is described by the Poisson distribution series. Variations

in the chain length of the final surfactant product are minimized by

the choice of pH, temperature and catalyst used to decrease the

activation energy of the reaction but they cannot be completely

eliminated. Therefore, the number of polyoxy-n-ene units for a given

nonionic surfactant represents the most commonly occurring chain

length and the physicochemical properties of the surfactant will vary

by batch and manufacturer.



Micelles and Micellization

When small amounts of surfactant are added to a totally aqueous

solution, the surfactant molecules are predominately present in the










unassociated monomeric form, although small concentrations of dimers,

trimers, etc. may be present. As the concentration of surfactant in

solution is increased, the individual surfactant monomers aggregate to

form structures known as micelles (from the Latin micella meaning

small bit). Micellization occurs over a narrow range of concentration

known as the critical micelle concentration (CMC). The continued

addition of surfactant to solution above the CMC results in the

formation of additional micelles, with the amount of free surfactant

monomer present in solution remaining approximately constant and

equivalent to the CMC. Figure 1 shows the effect of changing

surfactant concentration upon the concentration of monomer and

micellar aggregates present in solution. Micellar aggregates are in

dynamic equilibrium with the surrounding solution; individual monomers

are exchanged with the bulk solution on the microsecond to millisecond

time scale, while the entire micelle is exchanged on the order of

milliseconds to seconds.

In aqueous solutions at surfactant concentrations slightly above

the CMC and without additives present that have the ability to

solubilize within the micelle, the ionic micelle is considered to be

roughly spherical (Dill and Flory, 1981) in shape. Typically the

structure is 2-6 nm in diameter and contains 60-100 individual

surfactant monomers. As shown in Figure 2, the hydrophobic tails of

the surfactant are directed toward the interior of the structure in an

attempt to minimize contact with the surrounding aqueous solution.

The microviscosity of the hydrocarbon core is greater than that in

pure hydrocarbons. Surrounding the interior of the micelle is a





















E 1.5-

SAmphiphile present
4) as monomer




monomer conc
E =cmc
I/ / Amphiphile
S0.5 present in micellar
Form

cmc
0
0.5 1.0 1.5 2.0
Total amphiphile concentration



Figure 1. The relation between monomeric concentration in solution
and total added concentration of surfactant with micelle formation.
Dashed lines represent empirical procedures for determining the CMC;
the point at which the monomer concentration is equal to the CMC is
also shown (adapted from Tanford, 1980).

















Aqueous
bulk
phase


Range of
shear Core
surface -28
-H -Stern layer,
up to a few A

Gouy-Chapman--
double layer, up to
several hundred A


Figure 2. A two-dimensional representation of the regions of a
spherical ionic micelle (adapted from Fendler and Fendler, 1975).










region several angstroms in width, the Stern layer, containing the

polar headgroups, water and slightly more than one half of the

counterions associated with the micelle (Rosen, 1978). The

hydrocarbon chains do, however, have significant contact with the bulk

solvent (Dill et al., 1984). The remainder of the counterions

associated with the micelle are contained in the Gouy-Chapman portion

of the electrical double layer which extends further into the bulk

solution. The resulting structure can be viewed as a liquid-like

hydrocarbon droplet in a pool of water.

There are three primary forces which control micellization of an

ionic surfactant in aqueous solution. The hydrophobic effect or the

hydrophobic repulsion between the hydrocarbon tails of the individual

surfactant monomers and the surrounding aqueous medium is the dominant

factor controlling micellization. The attainable size of the

aggregate is limited by the charge repulsion between ionic headgroups

located in the Stern layer. To a lesser degree, the micelle is

stabilized by Van der Waals attraction between the alkyl chains within

the structure.

At the onset of micellization, many bulk solution properties are

significantly modified. Therefore, the CMC can be experimentally

determined by monitoring a physicochemical property of the solution

and noting the changes in that property as the concentration of the

surfactant in solution is varied (Figure 3). A calibration curve of a

physicochemical property vs. surfactant concentration commonly

exhibits a change in slope at the CMC. Figure 4 illustrates such a
















































Figure 3. Changes in various physicochemical properties of an aqueous
solution containing sodium dodecyl sulfate as the concentration of
surfactant in solution is varied (adapted from Rosen, 1978).










El
25-


20-


4 15-
0
E cmc
10 -


5-


0 E-3
0 2 3 4 5 6 7 8 9
[CTAB]


Figure 4. A calibration curve illustrating the change in slope of the
conductance versus concentration curve observed at the CMC for
hexadecyltrimethylammonium bromide (CTAB).










calibration curve for the determination of the CMC of hexadecyltri-

methylammonium bromide in aqueous solution using conductance

detection.

Changes in temperature, concentration of surfactant, additives in

the bulk solution and structural groups in the surfactant may cause

changes in the size, shape and aggregation number of the micelle

(Berezin et al., 1973; Fisher and Oakenfull, 1977; Menger, 1979).



Analytical Applications of Surfactant Containing Solutions

Analytical applications of micellar solutions arise from the

ability of the micelle to solubilize, compartmentalize and concentrate

ions and molecules. The utility of micelles in analytical chemistry

has been demonstrated in a variety of techniques including absorption

and luminescence spectroscopy, electrochemistry and chromatography.

Several reviews on the use of surfactants in analytical chemistry have

appeared (Cline Love et al., 1984; Pelizzetti and Pramauro, 1985;

McIntire, 1986).

Micellar aggregates have the ability to solubilize, in aqueous

solution, compounds which are normally insoluble or slightly

soluble. Table II shows a comparison of the solubility (Almgren

et al., 1979) of some aromatic hydrocarbons in purely aqueous and

aqueous micellar solutions. The observed increase in solubility can

be a direct result of the hydrophobic interaction between the solute

(possessing a certain degree of hydrophobic character) and the exposed

hydrocarbon chains of the micelle and/or electrostatic interaction of

the solute with the polar headgroups of the organized assembly.














Table II. Solubility data for arenes in water and micellar solutions.


Solubility (M)

Compound Water SDS (0.01-0.06 M) CTAB (0.01-0.05 M)

-2
Benzene 2.3 x 10-2 2.5 12.3

Naphthalene 2.2 x 10 0.38 1.11

Anthracene 2.2 x 107 0.63 x 10-2 3.3 x 10-2

Pyrene 6 x 10-7 7.0 x 10-2 0.41


Data adapted from Almgren et al. (1979).











Solute partitioning to micelles is characterized by a partition

coefficient analogous to any two-phase equilibrium. Micellar

solubilization is a dynamic equilibrium process which is dependent

upon temperature, nature of the solute, concentration of surfactant

and the type of micellar aggregate under consideration (Berezin

et al., 1973).

Micellar organization of ions and molecules on a molecular level

increases the proximity of reagent(s) and analyte in solution. This

may result in a modification of equilibria, acid-base and redox

properties, reaction rates and chemical pathways. Additionally,

micellar aggregates may change spectral distribution and/or

intensities. In order to take full advantage of a micelle's ability

to organize reactants it is desirable to have the species involved

preferentially associate with the micelle over the bulk solvent.

Using the example of accelerated reaction rates, this

compartmentalization is most easily accomplished by the judicious

choice of reactants which have hydrophobic character (resulting in

partitioning to the micelle) and/or are electrostatically attracted to

the polar headgroups of the aggregate. As a result of partitioning,

the micelle concentrates the interacting species in a much reduced

volume about the micelle, greatly increasing the local concentration

of reactants and thus the rate of reaction.



Flow Injection Analysis

Flow injection analysis (FIA) is firmly established as a rapid,

accurate, precise and versatile analytical technique. As an










indication of the growing acceptance and application of the technique,

the number of FIA publications has increased exponentially (Ruzicka

and Hansen, 1986) since its inception in 1975 with numerous reviews

(Ruzicka and Hansen, 1978, 1986; Betteridge, 1978; Ruzicka, 1983) and

several monographs (Ruzicka and Hansen, 1981; Ueno and Kina, 1983;

Valcarcel and Luque de Castro, 1987) appearing.

Flow injection analysis (FIA) is based on the injection of a

known, reproducible volume of a liquid sample into a moving,

nonsegmented carrier stream (Ruzicka and Hansen, 1981). The injected

sample forms a zone which disperses in a known, reproducible manner as

it flows through the system toward eventual detection. The simplest

FIA analyzer consists of a pump which propels the carrier stream, an

injection port for the introduction of a precise, reproducible volume

of sample and a reaction volume where the sample disperses as it

advances toward a flow-through detector, producing a recorder

response. The resulting FIA manifold requires no elaborate

instrumentation and the strengths of the system include low cost,

simplicity of operation, high sample throughput without carryover,

compatibility with all conventional flow-through detectors and high

precision stemming from the reproducibility of the sample injection

and dispersion processes.

The major obstacle to be overcome in the development of FIA as an

analytical technique was the prevention of intermixing between

adjacent samples. Skeggs (1966) first introduced the concept of air

segmentation in which adjacent samples were separated by air

bubbles. Air segmented continuous flow analysis served as the basis










for the Technicon AutoAnalyzer, the most popular automatic analyzer

ever marketed. Air segmentation, as a method to prevent sample

carryover, has many drawbacks (Ruzicka and Hansen, 1981). Among the

most deleterious are that the carrier stream has the tendency to

pulsate, due to the compressibility of air; the air bubbles must be

removed from the carrier stream prior to detection; control of the

size of the air bubbles is necessary; and that the movement of the

carrier cannot be instantaneously stopped or restarted (as in stopped-

flow FIA methods). With the many disadvantages of segmented flow

analysis as a driving force, the technique of FIA was developed by the

simultaneous work of Ruzicka and Hansen (1975) in Denmark and Stewart,

Beecher and Hare (1976) in the United States. The Danish group

developed the nonsegmented technique by using instrumentation

primarily associated with segmented flow analyzers while the Americans

developed the technique employing instrumentation which is typically

associated with high performance liquid chromatography (HPLC). As a

result of this common historical background, the technique of flow

injection analysis is often mistakenly considered as a hybrid of the

two technqiues.

The similarities between the two techniques are essentially

limited to the instrumentation and certain operational parameters used

for each method. Both techniques require the use of a pump, sample

injection valve, flow-through detector and a chart recorder or data

acquisition device. Additionally, injected sample volumes and flow

rates in the two methods are of the same order of magnitude. This,

however, is the extent of the similarities between the two techniques.










The general aim of HPLC is to separate several solutes which are

present in the injected sample as it is introduced into the flowing

system. The HPLC column contains two immiscible phases: the mobile

phase which flows through the system and the stationary phase which

consists of small diameter particles (3-20 pm) which fill the interior

volume of the column. The HPLC pump typically operates at pressures

in excess of 1000 psi, which is necessary to force the mobile phases

through the interstitial volume of a tightly packed column.

Separation of the solutes occurs due to differences in the

concentration distribution coefficients (the ratio of the mole

fractions of each solute in the two phases) between the various

solutes. Due to the retention of solutes in a chromatographic system,

the various chromatographic zones move at a velocity which is only a

fraction of the mobile phase velocity.

In FIA, however, the general aim is to control the dispersion and

the time allowed for a reaction to occur in the flow manifold,

ensuring detectable levels of product while achieving maximum

throughput (samples per hour) of the system. The peristaltic pump

propels the carrier stream through short lengths (50-300 cm) of

narrow, open teflon tubing (typically 0.5 mm id). Open tubing

translates to low backpressure; therefore, the peristaltic pump

typically operates at pressures well below 10 psi. The flow profile

in FIA is essentially laminar, ensuring that the solute front moves at

the same velocity as the carrier stream.

Flow injection analysis is classified as a method of continuous

flow analysis, whereby the analyte concentration is continuously










monitored in a moving stream of liquid (or gas). Typically, one

reactant is constantly pumped through a length of teflon tubing which

constitutes the FIA reaction manifold. The other reactant is

intercalated at one or more fixed points along the tubing by way of a

sample injection valve. As the sample moves through the manifold

toward eventual detection, it is dispersed and interacts with the

carrier stream, resulting in dilution. As a result of the dispersion

process, the concentration at the peak maximum is some fraction of the

concentration of sample which was initially introduced into the

manifold (Figure 5).

The dispersion coefficient, D, is the most common descriptor of

dilution in an FIA system and is defined as the ratio of the

concentration injected into the system to the concentration at peak

maximum. Experimentally, the value of D is determined by introducing

a standard sample of known concentration directly into the flow-

through detector to obtain a steady-state response height, H The

FIA manifold is then operated with the carrier stream of choice and

the standard is introduced into the continuously moving stream in the

form of an injected sample. This ensures that the sample molecules

will interact with the surrounding carrier stream, resulting in

dilution of the sample and a subsequent peak height response, Hmax

If the linear region of the calibration curve includes both C and
o
C (the concentrations corresponding to H and H ):
max o max


D = H/Hmax
0 max


(eq. 1.1)






























x
cO
Ch








0






03
N
















bo
E


















m


0
L.



17)





























bo
E

































U-4
0.

a,

0.

E


c4-
0










a,



a,




to
-4







--- 3SNOdS3d


NOI-LLV8dN3JDNO:D











and is equivalent to the dispersion of the flow injection manifold

under investigation. Dispersion is generally classified as limited

(D = 1-3), medium (D = 3-10) and large (D > 10) with the class of

dispersion chosen being dependent upon the application. The concept

of dispersion, however, is of limited utility in conveying information

about the FIA system since it only accounts for dilution which affects

peak height. In the presence of chemical reactions, the dispersion

coefficient loses its significance, since it conveys little

information related to peak width (an indication of the throughput of

the system). Coiling of the tubing introduces additional mechanisms

of transport which further obscure the significance of D and will be

more fully considered in Chapter IV. Additionally, we will introduce

an alternative measure of dispersion which gives direct information

concerning the variance of the FIA peak and allows for the

deconvolution of certain factors contributing to dispersion in an FIA

manifold.

There are two mechanisms contributing to dispersion in an FIA

system. The flow profile in FIA is essentially laminar in character

(Van den Berg et al., 1980; Reijn et al., 1981b; Vanderslice et al.,

1981), resulting in convective transport in the axial direction. For

conditions of laminar flow, in a straight length of tubing, it is well

known that molecules near the x axis (in the tube center) will move at

a greater mean velocity than those at the tube wall. The velocity of

any stream path, v at any radial position, rx, in a tube of radius r

can be given by










v = 2v[1 (rx/r)2], (eq. 1.2)


where v is the average flow velocity of the system;


v = [(PI P )r ]/8nL, (eq. 1.3)
1 o

where PI and Po are the system inlet and outlet pressures, n is the

viscosity (g/cm s) of the solvent and L is the tube length in cm.

Therefore, the molecules at the center of the tube will move at

maximum velocity being equivalent to twice the mean velocity of the

system. Theoretically, the molecules in direct contact with the

tubing wall will be stationary. These flow characteristics will

produce a parabolic velocity profile (Figure 6a). If laminar flow

were the only factor contributing to dispersion in the system, a

sample bolus introduced into the system would have an infinitely long

tail (Figure 6b) resulting in carryover between adjacent samples and

decreased throughput of the system. In addition to convective

transport, however, there is diffusional transport in the system which

is a result of concentration gradients in the convective transport

regime (Figure 6c). Diffusional transport in the axial direction

(Figure 6d) arises due to horizontal concentration gradients at the

leading and trailing edges of the sample bolus and only minimally

contributes to the overall dispersion in the system (Valcarcel and

Luque de Castro, 1987). Diffusional transport in the radial direction

arises from concentration gradients perpendicular to the flow vectors

(Figure 6d). The overall effect of radial diffusion is that sample

molecules at the tubing center diffuse toward the walls while sample






























Figure 6. Velocity profiles, shape of injected sample bolus and types
of transport in closed tubes: (a) laminar flow, parabolic velocity
profile; (b) sample dispersion due to laminar flow without diffusion;
(c) convective transport of sample molecules parallel to direction of
flow; (d) diffusional transport in the axial and radial directions;
(e) sample dispersion due to laminar flow with diffusion (adapted from
Valcarcel and Luque de Castro, 1987).












-"- tube wall


CONVECTIVE TRANSPORT

LAMINAR FLOW

/- y


DIFFUSIONAL


AXIAL


TRANSPORT


RADIAL


--


IlU


Velocity


profiles


and shapes of injected


sample bolus










molecules near the walls diffuse toward the center of the tubing. In

fact, this radial transport opposes convective transport in the axial

direction, maintaining the integrity of the sample plug (Figure 6e).

It is this radial diffusion, perpendicular to the direction of flow

which eliminates the need for air bubbles in FIA, resulting in low

carryover and cross contamination between adjacent samples.

Initially, application of FIA was amenable only to relative

measurements with quantitation being dependent upon calibration with

standards. The capability of the technique has now expanded to

include absolute measurements of diffusion coefficients (Gerhardt and

Adams, 1982, 1983; Betteridge et al., 1983; Vanderslice et al., 1984),

solution viscosity (Betteridge and Ruzicka, 1976; Betteridge et al.,

1981, 1983), reaction rates (Hungerford et al., 1985), acid constants

(Fossey and Cantwell, 1985) and complex stability constants (Yoza

et al., 1984a, 1984b).



Micellar Reversed Phase Liquid Chromatography

Armstrong and Henry (1980) first effectively demonstrated the use

of aqueous micelles as the sole mobile phase modifier in a reversed

phase liquid chromatographic system. The practical advantages of

micellar mobile phases include enhanced selectivity, low cost and low

toxicity when compared to typical hydro-organic mobile phases. While

these benefits are certainly desirable qualities for any alternative

chromatographic technique, they are not sufficiently great to cause

the laboratory practitioner to abandon the time tested methods

associated with hydro-organic mobile phases in HPLC. Expanded










application of micellar liquid chromatography (MLC) will result only

if academic investigators continue to exploit the unique capabilities

of MLC to perform separations which are not achievable with hydro-

organic mobile phases. Several review articles (Armstrong, 1985;

Dorsey, 1987; Khaledi, 1988) and a symposium series volume (Hinze and

Armstrong, 1987) have recently appeared on the subject.

Micellar liquid chromatography incorporates secondary equilibria

to adjust retention and selectivity in the chromatographic system.

The "primary" equilibrium in a chromatographic system describes only

the distribution of solute between the mobile phase and stationary

phase. Additional equilibria which occur in the mobile phase,

stationary phase or both are considered "secondary" in nature. As

shown in Figure 7, the solute present in MLC partitions both to the

hydrophobic environment of the micelle and to the hydrophobic

stationary phase. Other examples of secondary equilibria include the

use of acid-base equilibria and ion-pairing methods to provide unique

chromatographic capabilities.

Potentially, MLC has a much greater range of applicability than

other types of secondary equilibria. Any compound which partitions to

the micelle is a candidate for micellar liquid chromatography. This

includes all compounds with hydrophobic character which are

preferentially solubilized in the hydrophobic environment of the

micelle. Additionally, compounds which are electrostatically

attracted to the surfactant polar headgroups within the micellar

structure are potentially separable by MLC. As reversed phase is

generally the method of choice for the separation of hydrophobic
































BULK WATER


kd ka




STATIONARY PHASE


Figure 7. The two principal solute equilibria in a micellar chromato-
graphic system, k 1and k_, are, respectively, the entrance and exit
rate constants (s ) of the solute to and from the micelle;
k and k are respectively, the adsorption and desorption rate
constants (s ) of the solute to and from the chromatographic
stationary phase (adapted from Yarmchuk et al., 1984).










compounds, MLC has the potential to be applied to a large percentage

of reversed phase separations.

The capacity factor, k, in MLC is defined (as for any

chromatographic system) as the total amount of solute in phase X

divided by the total amount of solute in phase Y. In chromatographic

applications, the stationary phase is defined as phase X. The total

amount of solute in a phase is defined as the concentration of solute

in that phase times the volume of that phase. Arunyanart and

Cline Love (1984) were the first to show that, by considering the two

principal equilibria in MLC, the capacity factor could be given by


k [solute in the stationary phase]
solutee completed + solutee in the (eq. 1.4)
with the micelle] bulk mobile phase]


where 4 is the chromatographic phase ratio and is given by the ratio

of the volume of the stationary phase to the volume of the mobile

phase; solutee .] indicates the concentration of solute in the

indicated phase. This expression serves as the basis for the

pseudophase retention model which allows for the determination of

valuable physicochemical information about the structure of micelles

and the factors controlling efficiency in MLC. Further investigation

of the applicability, validity and information yielded by the

pseudophase retention model will be presented in Chapter V.

In secondary chemical equilibria separations, the strength and

selectivity of the mobile phase are at least partially controlled by

the concentration of the component affecting the secondary

equilibrium. As previously mentioned and shown in Figure 1, the










amount of monomeric surfactant present above the critical micelle

concentration in a micellar solution is approximately constant and

equivalent to the CMC. Therefore, MLC provides the experimenter with

the ability to change the strength of the mobile phase without

changing the bulk solvent composition. It is this property of

constant bulk mobile phase composition which results in many of the

unique capabilities of micellar mobile phases.

Gradient elution chromatography is the most popular method for

solving the general elution problem in HPLC. By increasing the

strength of the mobile phase over the course of the separation, the

highly retained compounds can be eluted more quickly and better

detection limits resulting from sharper peaks can be achieved (Snyder,

1980; Quarry et al., 1986). The major disadvantage of the gradient

elution technique is that during the course of the gradient, the

stationary phase solvation structure undergoes significant

modification. Therefore, the nature of the stationary phase at the

completion of the gradient is different from the initial conditions.

In order to perform a subsequent analysis, it is necessary to

reequilibrate the chromatographic column with many column volumes of

the initial mobile phase. As a result, the advantages gained in

separation time due to the gradient program are often negated in terms

of the total analysis time due to the time spent in the reequilibrium

step. Quarry et al. (1984) have also shown that modification of the

stationary phase composition leads to solvent demixing (preferential

uptake of a mobile phase component) and variation in column dead time,

further complicating the theoretical model of gradient elution.










The strength of the mobile phase in MLC is dependent upon the

concentration of surfactant in solution. An increase in mobile phase

surfactant concentration results in the formation of additional

micellar aggregates but the amount of free monomeric surfactant in

solution is approximately constant and equal to the CMC (Figure 1).

The direct result of this is that a micellar concentration gradient

may be used to speed the elution of strongly retained compounds

without altering the structure of the stationary phase. Landy and

Dorsey (1984) and Dorsey et al. (1984) have shown the constancy of the

stationary phase structure from both the standpoint of consistent

retention times for an early eluting compound and adsorption isotherms

of SDS onto reversed phase materials. Berthod et al. (1986) also

measured the adsorption isotherms of various cationic and ionic

surfactants on five chromatographic stationary phases and reported a

constant stationary phase composition with changing surfactant

concentration above the CMC. It is clear that micellar mobile phases

have the potential to greatly reduce analysis time and solvent

consumption in gradient MLC techniques.

Micellar mobile phases also offer unique advantages for detection

when compared to hydro-organic mobile phases. A majority of common

surfactants possess fully saturated hydrocarbon tails, making them

optically transparent at typical HPLC wavelengths of detection.

Micelles have the capability to yield enhanced detectability stemming

from their ability to organize solutes on a molecular level. This

property has led to detection schemes employing micellar solutions to










enhance analytical fluorometry (Hinze et al., 1984) and micelle-

stabilized room temperature phosphorescence (Cline Love et al., 1980,

1981; Skrilec et al., 1980). Khaledi and Dorsey (1985) have

demonstrated improved gradient compatibility with amperometric

electrochemical detection using micellar eluents.

The preceding discussion indicates that micellar liquid

chromatography can provide unique chromatographic capabilities and

that micellar mobile phases do offer solutions to problems which

cannot be resolved by conventional means. Two major obstacles to the

widespread acceptance of MLC are that the chromatographic efficiency

and solvent strength achievable with aqueous micellar mobile phases

are much less than those available with traditional hydro-organic

mobile phases. In Chapter V, we will present experimental results

examining the addition of low concentrations of normal alcohols to

micellar mobile phases and subsequent effects upon efficiency and

solvent strength in micellar liquid chromatography.
















CHAPTER II
FLOW INJECTION ANALYSIS DETERMINATION OF
CRITICAL MICELLE CONCENTRATIONS OF
IONIC AND NONIONIC SURFACTANTS


Background

It has been sixty years since the initial observation (Ekwall,

1927) linking changes in surfactant concentration with changes in

physicochemical properties of a solution. The concept of the critical

micelle concentration (CMC) has arisen to describe the concentration

range in which surfactant monomers assemble to form micelles. The CMC

varies with the intensive parameters of the solution in which the

micelle is formed. With ionic surfactants, it is observed that

variations of temperature, pressure (to a lesser extent) and amount of

electrolyte or nonelectrolyte added to a solution all influence the

observed CMC. These variations in CMC values in response to changes

in solution properties have proven useful both for the study of the

driving forces of micellization and for allowing inferences about

micelle shape and structure to be made. Therefore, when using

surfactants, the experimenter finds it necessary to determine the CMC

characteristic of the system which has been defined.

The value of the CMC is experimentally obtained by monitoring the

change of a physicochemical property of the solution with changing

surfactant concentration. Extrapolation of the responses obtained at

high concentrations (above the CMC) and at low concentrations (below










the CMC) yields an intersection whose concentration is equivalent to

the CMC. The intersection concentration obtained depends both upon

the physicochemical property chosen to determine the CMC and the

graphical method selected to represent the data. Numerous

physicochemical properties have been monitored to determine the CMC.

Measurements of surface tension, electrical conductance and spectral

properties are among the most popular techniques employed. Mukerjee

and Mysels (1971) address the reasons for methodical differences among

CMC determinations.

A search of the literature for the past five years reveals that

researchers have undertaken numerous projects in an attempt to develop

novel methods for the determination of the CMC of surfactants in

solution. The great majority of proposed determination methods rely

upon spectrophotometric measurements of a probe molecule which has

been introduced into a surfactant solution. Pyrene has been utilized

as a fluorescence probe molecule (Kapoor et al., 1982; Ohyashiki and

Mohri, 1983; Goddard et al., 1985) and numerous examples of dye

incorporation to aid in the spectrophotometric determination of the

CMC are found (Rosenthal and Koussaie, 1983; Kawashima et al., 1985;

Panda and Behera, 1985). Interactions between aggregates of dye

molecules and surfactant monomers can result in the formation of mixed

micelles (Kali et al., 1980) at surfactant concentrations which are

markedly below the true CMC (Robinson et al., 1973). The

determination of mixed micelles includes the effect of the dye-

surfactant interaction and does not yield an accurate determination of

the CMC. The subject of dye-surfactant interactions has been recently










reviewed (Diaz Garcia and Sanz-Medel, 1986). Additionally, the CMC

has been determined employing methods based upon kinetic dialysis

(Lake, 1982), emulsion polymerization (Al-Shabib and Dunn, 1981),

bubble pressure (Kretschemer et al., 1982), foaming power properties

(Bazhenov et al., 1983) and calculations from electrical capacitance

measurements (Deinega et al., 1983). The method proposed by Taylor

and Nieman (1984) based upon conductance detection with exponential

dilution flow circumvents the problem of repetitive measurement but

the requirement for an in-house constructed, computer-controlled

bipolar pulse conductance apparatus has served to limit the practical

application of the technique. Other proposed CMC determination

methods which will never receive widespread application due to their

instrumentation-limited nature are those based on photon correlation

spectroscopy (Roe and Barry, 1983) and an ultrasonic interferometric

technique (Kabachnyi et al., 1982). It is concluded, therefore, that

none of the above work effectively addresses the paramount problems

with existing methods for CMC determination. The need to acquire many

data points as a prerequisite for accurate extrapolation to the CMC

requires both extensive solution preparation and repetitive

measurements which exhaust both the material resources and the time of

the experimenter.



Theoretical Study of Critical Micelle
Concentration Determination by Flow Injection Analysis

The incorporation of a gradient chamber (Pardue and Fields,

1981a, 1981b; Pardue and Jager, 1986) as the reaction volume in FIA










produces an exponential concentration gradient as described by the

tank-in-series model (Levenspiel, 1972). Here we summarize a

theoretical model (Berthod et al., 1988) to explain the experimental

results of this chapter and to allow optimization of experimental

parameters. The present work includes (1) the set of equations

representing the surfactant concentration evolution during an FIA run,

(2) the conductimetric detection equations, (3) the equations

representing the light absorption change observed with a micelle-

solubilized dye and (4) theoretical curves illustrating the equations

proposed which allow discussion of the feasibility of the method.

Surfactant Solutions in Flow Injection Analysis

The FIA manifold consisted of a pump which propelled the carrier

stream at a constant flow rate, F, through a chromatographic injection

valve and an exponential dilution chamber (volume V). A detector

continuously monitored the output of the dilution chamber.

Injection. By using the injection valve, a volume, v., of

micellar solution with a surfactant concentration, ci, was injected

into the flow stream with a step-concentration shape. After a delay

time, to, corresponding to the sweeping time of the inevitable dead

volume, the micellar solution reached the dilution chamber. During a

time ti = v /F, the surfactant concentration in the dilution chamber

increased according to


dC = (ci C)Fdt/V (eq. 2.1)


in which C is the surfactant concentration, at time t, in the dilution










chamber (volume V). At t = to, C = 0, the integrated form of equation

2.1 is


C = o [1 exp(-(F/V) x (t to))]. (eq. 2.2)


The maximum concentration, Cmax, was reached at time t + t ; then


C m c [1 exp(-v /V)]. (eq. 2.3)


In all experiments vi << V; then equation 2.3 can be approximated by


C m = c v/V (eq. 2.4)


(Ruzicka and Hansen, 1981). For example, with a dilution chamber

volume of 1 mL and an injected volume of 30 1L, equation 2.4 gives the

concentration C with an error lower than 1.5%.
max
For this method of CMC determination, the C concentration must
max
be higher than the CMC, which means the injected surfactant

concentration must be higher than CMC x V/vi. With an actual value of

35, for the ratio V/vi, the injected surfactant concentration must be

at least 50 times the CMC.

Dilution. After the injection step, dilution occurred. During a

time dt, a volume, dv = Fdt, of pure carrier (without surfactant

molecules) enters the chamber. The same volume, with a concentration,

C, of surfactant molecules, exits the chamber. The mass-conservation

equation is


VC(t + dt) = VC(t) C(t)dv.


(eq. 2.5)










By using a differential form, equation 2.5 is


VdC = -CFdt.


With the boundary condition C = C

form of equation 2.6 is


max at t = ti + to, the integrated


C C eaxe[-(F/V) x (t (t + to))].


If the concentration C is higher than the CMC, the micelle

concentration, Cm, is C CMC:


C = (ci v/V)exp[-(F/V) x (t (ti + to))] CMC.
m ii l


(eq. 2.7)


(eq. 2.8)


When Cm 0, the CMC is reached, and the time t is tCMC. Equation 2.8

gives


t CMc (V/F)ln[(c iv)/(CMC x V)] + ti + t .
CMC 1 1 1 o


(eq. 2.9)


After the time tCMC no more micelles are present in the dilution
CMC
chamber. The monomeric surfactant concentration obeys equation 2.7

rewritten as


C CMC exp[(-F/V) x (t tCC)].


(eq. 2.10)


This set of equations shows that it will be possible to determine

the CMC of a surfactant, in only one run, if the micelle concentration

can be monitored by the sensor.

The micelle concentration can be monitored by conductance

measurements or by spectrophotometric measurements using a micelle-

soluble dye.


(eq. 2.6)










Conductance Measurements

The conductivity, K, of an electrolytic solution depends on the

electrolyte concentration and on the molar conductance A of the ions

(Monk, 1961). The conductivity of a micellar solution is given by


K A CMC + A x (C CMC) (eq. 2.11)
s m

in which the subscripts s and m refer to surfactant monomers and

micelles, respectively. The molar conductances are not constant; they

obey the Onsager-Kohlrausch law,


A A BC1/2 (eq. 2.12)


in which A is the molar conductance at infinite dilution (C 0) and
0
B is a constant (Smedley, 1980). For example, at 250C, the molar

conductance of sodium dodecyl sulfate (SDS) was given to be 66.8 S cm2

mole-1 at CMC (8.2 x 10-3 mole liter-1) and 72.5 S cm2 mole-1 at

infinite dilution (Kay and Lee, 1986). The molar conductance of SDS

in micellar form lies between 25 and 41 S cm2 mole-1 according to the

method used (Kay and Lee, 1986). As a first approximation, we will

assume that A and A are constant. So, by using equations 2.7, 2.8
s m
and 2.11, the conductivity of solutions above the CMC is


K = Am(c /V) x exp[(-F/V) x (t (t + to))]


+ (A A ) x CMC (eq. 2.13)
s m

and, with equations 2.10 and 2.11, the conductivity of solutions below










the CMC is


K = A CMC exp[(-F/V) x (t t )]. (eq. 2.14)


The conductimetric detector will produce two exponential decays

according to equations 2.13 and 2.14.

Spectrophotometric Measurements

The use of a dye, whose absorbance is modified by the presence of

micelles, to determine surfactant CMCs, has been known for many years

(Corrin and Harkins, 1947) and is still currently used (Rosenthal and

Koussaie, 1983; Berthod and Georges, 1985). When a hydrophobic dye is

dissolved in the hydrophobic core of a micelle, its absorption

spectrum is often modified (bathochromic effect). By recording the

absorbance of the dye in the mobile phase at a well-chosen wavelength,

it will be possible to determine the CMC of a surfactant.

Care must be taken when choosing the dye because it has been

pointed out that dyes could modify the CMC of ionic surfactants

(Berthod and Georges, 1985). Premicellar aggregates can form with

ionic dyes and surfactant molecules that greatly lower the observed

CMC (Sato et al., 1983). This problem is minimized with nonionic

surfactants, but still exists (Nemoto and Funahashi, 1981a, 1981b).

There are two possible ways to obtain the CMC of a surfactant:

(i) the dye is dissolved in the mobile phase and a concentrated

solution of the studied surfactant is injected in the colored mobile

phase, or (ii) the mobile phase is pure water and a concentrated

solution of the surfactant, in which the dye has been solubilized, is











injected, the dye being used as a micelle tracer. Both cases are

studied theoretically.

Monitoring of spectral changes. The affinity of a dye for the

micellar phase can be quantified by using a partition coefficient, P,

that is the ratio of the dye concentration in the micellar phase,

D in moles per volume of micellar phase, over the dye concentration

in the aqueous phase, Da in moles per volume of aqueous phase


P = D D D (eq. 2.15)
m a


Introducing V, the molar volume of the surfactant in micellar

form, one finds that the product C V is the volume fraction of the
m
micellar phase and 1 C V is the volume fraction of the aqueous
m
phase. The dye concentration, D, in the mobile phase is


D D (1 C ) + (Dm C ). (eq. 2.16)
a in mm

From equations 2.15 and 2.16, we derived


D D/(1 C V + PC V) (eq. 2.17)
a m m
and

D = DP/(1 C V + PC V). (eq. 2.18)
m m m


When micelles are present in the dilution chamber, the measured

absorbance, A, is, using equations 2.16-2.18,


A = [D x (A PC V + A (1 C V))]/
m m a m


(1 C V + PC V).
m m


(eq. 2.19)










The symbols A and A represent the absorptivity of the dye
m a
solubilized in micelles or in aqueous phase, respectively.

After tCMC, no micelles are present in the dilution chamber, and

the absorbance, A, is


A = DA (eq. 2.20)
a

To ensure that the dye concentration is constant during the FIA

analysis, it is necessary to inject a concentrated surfactant solution

containing the dye at the same concentration, D, as that in the mobile

phase. That is not very convenient. If the injected solution does

not contain the dye, it is easy to derive the equations of the time-

dependent dye concentration, D(t). During the injection time

(t << t t + t ), the dye concentration decreases according to


D(t) = D exp[(-F/V) x (t t )]. (eq. 2.21)


The minimum of the dye concentration is reached at time to + t As

already expressed, the ratio v./V is small, and the approximation used
1
in equations 2.3 and 2.4 is valid. With a dilution chamber of 1 mL

and an injected volume of 30 pL, the minimum of the dye concentration

is 97% of the initial dye concentration. After time t. + t the dye

concentration increases according to


D(t) = (D(t + t ) D) x exp[(-F/V) x (t (to t ))]. (eq. 2.22)


Although these dye concentration variations are negligible,

equations 2.21 and 2.22 were used in the computer program generating

the theoretical spectrophotometric data.










Illustration of the Theoretical Models

Conductimetric detection. Figure 8 shows the curves obtained by

using equations 2.2, 2.7, 2.10, 2.13 and 2.14 and the numerical values

listed in Table III, which correspond to SDS surfactant. The effect

of a time constant of 1, 2, 4 and 8 s is illustrated by Figure 9. A

break, corresponding to tCMC, is observed on the conductance curve,

which allows calculation of the CMC value if the injected volume and

concentration are known (equations 2.9 and 2.23). The time constant

introduces an error to the tCMC determination: the longer the time
CMC
constant, the higher the error in the tCMC determination.
CMC
We have experimentally verified the effect of variation in the

time constant on the CMC value determined using equation 2.24 by

varying the flow rate from 0.46-13 mL/min. Figure 10 shows that the

error in the CMC determined for CTAB (CMC = 0.92-0.98 mM, see Table

VII) increases as the flow rate in the system is increased (higher

time constant). Results of this study also indicate that the CMC

value calculated is approximately constant from 0.42-2.0 mL/min,

verifying that the flow rates used for the CMC determinations in this

work (1.000.05 mL/min) are valid.

Spectrophotometric measurements. Figure 8 shows the curves

obtained by using equations 2.19 and 2.20 with the numerical values

listed in Table III. Figures 11 and 12 show the theoretical curves

obtained with different time constants. In both cases, the dye was in

the mobile phase or the dye was previously solubilized in the injected

micellar solution; the CMC determination will not be possible with a


















tCMC
120 '100


\
100 \ '. 80



CU) 80 \

C \
3 60 0
o N .

0 40 '
20
o 40




0 ......... .. .... .... ....--- 0
20



0 0
0 20 40 60 80 100 120 140 160 180
time (s)










Figure 8. Theoretical curves obtained for SDS: (1) conductimetric
detection (equations 2.13 and 2.14); (2 and 3) spectroscopic
measurements with a dye in the carrier stream, P = 100 and 10,000,
respectively (equations 2.19 and 2.20); (4 and 5) spectroscopic
measurements with the dye dissolved in the injected micellar solution,
P = 100 and 10,000, respectively (equations 2.7, 2.19 and 2.20).
Numerical values are listed in Table III.







43
















120



100



- 80



60


C





20


0

0 20 40 60 80 100 120 140 160 180
time (s)







Figure 9. Theoretical conductimetric curves obtained with different
time constants. Time constants: dashed line = 0 s; full lines = 1,
2, 4 and 8 s.


















































8 10 12


Flow Rate (mL/min)


Figure 10. Variation of
2.24) with flow rate.


the determined CMC of CTAB (using equation


1 10


1.05




1 00


0.95


U
U


4 6











Table III. Numerical value of the parameters used to draw the curves
of Figures 8, 9, 11 and 12.


Parameter Symbol Value Unit


Dye absorptivity in water


Dye absorptivity in micelles


Injected concentration

CMC of SDS

Dye concentration in the
carrier stream

Dye concentration in the
injected solution

Flow rate


Dye partition coefficient

Dead time

Injection time

CMC time

Dilution chamber volume

Injected volume

SDS molar volume

Dead volume

Monomer molar conductivity


Micellar molar conductivity


2 x 106


ci

CMC


D


1

8.2 x 103


Di 2.5 x 10-4

F 0.98
16.27

P 100 and 10,000

to 7.4

ti 1.8

tMC 81.3

V 765

vi 30

7 0.246

-- 120

A 6.38 x 102


2.01 x 0-2
2.01 x 10


Absorb. unit
-1
mole liter

Absorb. unit
-1
mole liter

moles liter-1

moles liter-1


moles liter-
-1

moles liter
m~t -1

mL min
pL s

Dimensionless

S

3
s
s

S

iL

IL

liters mole-

pL

S mole-1 liter

-1
cm1

S mole-1 liter
cm-1
GE


Note: These parameters correspond to our setup and to SDS surfactant.
































0 60






1 P=1000

20
CC



0 20 40 60 80 100 120 140 160 180
TIMES)





Figure 11. Theoretical absorbance curves with the dye solubilized in
the carrier stream with partition coefficients equal to 100 and
10,000. Time constants: dotted line = 0 s; full lines = 1, 2, 4 and
8 s.






47

















100




80 60 80 100 1
S1 2, 000



U 60
z



40

tcmc

20




0 20 40 60 80 100 120 140 160 180
TIMES)




Figure 12. Theoretical absorbance curves with the dye solubilized in
the injected surfactant solution. Time constants: dashed line = 0 s;
full lines = 1, 2, 4 and 8 s.










partition coefficient equal to or lower than 100. With partition

coefficient values of 10,000 or higher, the CMC determination seems

possible by determining tCMC at the inflection point of the

experimental curves (Figures 11 and 12) and using equation 2.23, as

long as the time constant is lower than approximately 4 s. For

partition coefficient values lying between 100 and 10,000, the CMC

determination should be possible if the spectrum of the dye is greatly

modified by the micelle solubilization (Figures 15 and 16). To give

one an idea of typical values in SDS micellar solution, the partition

coefficients of methyl red, bromphenol blue and neutral red are 8,000,

5,000 and 28,000, respectively (Berthod and Georges, 1985).



Experimental Section

The following work then experimentally verifies the findings of

the preceding theoretical treatment of CMC determinations by FIA and

associated detection. It has been shown previously (Baxter-Hammond

et al., 1980) that a conductance vs. concentration curve for

exponentially diluted surfactants displays a change in slope at the

CMC. Knowledge of the time at which this change in slope occurs

yields the CMC through equation 2.23 given below. The validity of the

presented method for ionic surfactants is verified by a determination

of CMC values for the following surfactants: hexadecyl-

trimethylammonium bromide (CTAB), hexadecyltrimethylammonium chloride

(CTAC) and sodium dodecyl sulfate (SDS). Figure 13 shows the

molecular formulas of the ionic surfactants investigated in the

present work.













IONIC SURFACTANTS


Anionic

SDS: C12 H25 OS0 Na+


Cationic

TTAB : C14 H29N+Br-(CH3)3
CTAB : Ci6 H3 N Br- (CH3)3
CTAC : Ci6 H3 N+CI- (CH3)3


Figure 13. Ionic surfactants (with molecular formulas) under
investigation.










The absorbance spectrum of many dyes exhibits a shift in the

presence of micelles, and for many years this has been exploited to

determine the CMC of surfactant solutions (Corrin and Harkins,

1947). Coomassie Brilliant Blue G 250 (CBBG) has had several

applications as a micelle-tracer for the determination of CMCs

(Rosenthal and Koussaie, 1983; Samsonoff et al., 1985). The use of

CBBG, however, requires the preparation of an acidic dye solution

(Bradford, 1976), introducing compounds into solution which may

potentially modify the CMC. Coomassie Brilliant Blue R 250 (CBBR) is

the dye selected for the present work (Figure 14). This dye requires

no pretreatment and is dissolved directly in the aqueous carrier

stream at a concentration of 1.3 x 10-5 M. Additionally, in nonionic

micellar solutions, the wavelength of maximum absorbance of the dye is

red-shifted by ca. 50 nm, regardless of surfactant. Figures 15 and 16

show the bathochromic shift of CBBR when present in the micellar

environment of Brij 35 and Triton X-100, respectively. The

application of the present technique to nonionic surfactants is

introduced with the determination of CMC values for Brij 35, Brij 56,

Brij 99 and Triton X-100 (Figure 17).

Apparatus

A block diagram of the FIA manifold employed for the present work

is shown in Figure 18. An Isco (Lincoln, NE) Tris model peristaltic

pump propelled the aqueous carrier stream. Samples were introduced by

a Rheodyne (Cotati, CA) model 7125 sample injection valve. All

connecting tubing was teflon (1/16" od, 0.5 mm id) supplied by the

Anspec Co., Inc. (Ann Arbor, MI). The gradient chamber was machined
















H3 C


HN





0 CH3


Coomassie


Brilliant


SO3 Na











N "CH


S038


Blue R 250


MW = 826.0


g/mol


6104 -59- 2]

Figure 14. Structural formula of Coomassie Brilliant Blue R 250
(CBBR).




























-4

S.



0






x

LC)











00



m
a











CC)








-7
(Y)



0
E









ca









(n
0

















a)
cfl






















a,-4






L -
bo

-4

tx. I


















_._ I "J_ ~ E.
1m


-4










-j



ii


J I I.
iit








. 1.


T- 1,7 t i


!i!!!! 1 1,14l : i l t~ !~ ~ l !


* 4


SI iI I


______ '-- ~ IiI.:i...- I i, .r5


iii


S i i I 'Ii i I I


i l I I I


~r Ij '

I~~~ 'L j






I~ I


. t7t


'0


In
II




x
(a
n


n
*I
-4





o


IC


i en
i i
II
x



0
N


4


I I


I I I I'


I f I ;J


I I I ,1, =


i


--- ----L II~~


1 1 -I-- I -- 1 +t,,I II I


cir1"


C.e


I I I


l rI


-6U.B


!!1 I


*^s-L i


I i :i


' I .


: I
















u)
OH.

0




CDr

0~
cr7






CD

C
a,

0
c-
5







9,





x


0
U,




















.0
cD


















0
r-


















CD
a,

z

w

























x
0

ul

























C-t
0




0
0















4 .Zt I'


















II--


14'. 1 E1 i I I ___ilo.e


____ I L,'A




I I L I
ITT_

,1 I





I lK
i : t i :





::::ii:: i ti:::::: ::::
L______^.,,--'
_ __^_-, --z-_


I ___ J-Th


I ; 1 1.1- I


I -
II







__ I I __ __






i4~4 I 1i


' -- I I -, -- -. .-


I : '


I


.pE''


..e


I I I














NONIONIC


SURFACTANTS


Triton X-100: C8 H,7 -- (OCH2 CH2 )95. OH


Brij


35 : C2 H25 (OCH2 CH2 )23 OH


Brij 56 : Cs H33(OCH2 CH2 )o OH


Brij 99: Ce H7 (OCH2 CH2 )2 OH


Figure 17. Nonionic
investigation.


surfactants (with molecular formulas) under












(A)

S ~-------~FC ---^
















-5
1. 3 X 1 Q
CBBR




Figure 18. (A) FIA manifold for the determination of the CMC of ionic
surfactants and (B) the analogous FIA manifold for the determination
of the CMC of nonionic surfactants with CBBR dissolved in the carrier
stream. Symbols: Q = flow rate of aqueous carrier stream (mL/min), S
= sample injection, FC = flow-through detector cell, and W = waste.










from stainless steel, consisting of two parts (Figure 19). The lower

portion is circular (0.478 in. id) and contains a 2 mm x 7 mm teflon

covered magnetic stirring bar (Anspec). The upper part has a dome-

shaped inner cavity designed to avoid entrapment of air. The upper

and lower portions are held together by six equally spaced screws. A

teflon 0-ring is placed between the upper and lower parts and the

screws are tightened, creating a leak-free seal. A hole drilled at

the base of the lower chamber accommodates the teflon tubing which

carries the incoming solution. To avoid leakage from this connection,

a larger hole may be drilled and threaded to accommodate a 1/16" male

nut into the assembly. A hole drilled in the center of the upper part

provides the outlet for the solution and the teflon tubing may be

push-fitted at this connection. The chamber was placed on a Precision

Scientific (Chicago, IL) Mag-Mix model magnetic stirrer and mixing was

found to be equally effective in the range of 60-240 rpm. The ionic

surfactants were directly detected by an LDC/Milton Roy (Riviera

Beach, FL) conductoMonitor III conductivity detector. In the

determination of the nonionic surfactants, the absorbance of the

dissolved dye in the carrier stream was monitored at 599.6 nm by a

Kratos (Ramsey, NJ) Spectroflow 757 absorbance detector. The

resulting output signals were recorded on a Fisher Recordall Series

5000 recorder. All least squares calculations were performed by

Interactive Microware, Inc. (State College, PA) Curve Fitter program

run on an Apple II Plus microcomputer.


















A B








C




____ \ )__ __- --\ -





Figure 19. Gradient chamber with arrows indicating the carrier stream
direction of flow: (A) teflon 0-ring, (B) connecting screw, and (C)
magnetic stirring bar.










Chemicals

Brij 35 (polyoxyethylene 10 cetyl ether, m.w. 680), Brij 99

(polyoxyethylene 20 oleyl ether, m.w. 1150), Coomassie Brilliant Blue

R 250 (C.I. 42660, m.w. 826) were from Sigma Chemical (St. Louis, MO);

purum grade CTAB (cetyl trimethylammonium bromide, m.w. 364.5) and SDS

(sodium dodecyl sulfate, m.w. 288.3) were from Fluka Chemical

(Hauppauge, NY); Triton X-100 (polyoxyethylene 11 p-octyl benzyl

ether, m.w. 690) and Eriochrome Black T (C.I. 14645, Mordant black 11)

were from Fisher Scientific (Fairlawn, NJ); Brij 25 (polyoxyethylene

23 dodecyl ether, m.w. 1200) was from Aldrich Chemical (Milwaukee,

WI); reagent grade CTAC (cetyl trimethylammonium chloride, m.w. 320)

was from Eastman Kodak (Rochester, NY); methyl red sodium salt (C.I.

13020, m.w. 301) was from Matheson, Coleman and Bell (Norwood, OH);

ethylene glycol and sodium chloride were from Mallinkrodt (St. Louis,

MO); TTAB (tetradecyltrimethylammonium bromide, m.w. 336.4) was from

Chemical Dynamics Corporation (South Plainfield, NJ). All were used

as received. Water used for the preparation of solutions and the

carrier streams was deionized.

Procedure

Initially, a concentrated surfactant solution (approx. 100X the

CMC) of known concentration must be prepared. A known volume of this

standard solution is then injected via a fully loaded sample loop (see

"Practical Considerations") such that the post-dispersion peak height

is a minimum of 2 times the CMC of the surfactant. This initial

injection is performed to determine the response height equivalent to

the CMC. Visual inspection of the resulting response curve recorded










at 1 cm/min yields the response height equivalent to the CMC. With

ionic surfactants, this response is recognized by noting an abrupt

change in curvature on the descending edge of the response curve which

occurs at the intersection of the micellar and monomeric exponential

response gradients as shown in Figure 20. When employing a dye as a

micelle-tracer, the response height equivalent to the CMC is defined

by the inflection point of the response curve as shown in Figures 11

and 21.

Having determined the response height equivalent to the CMC, a

second injection is performed with the detector response being

monitored at a chart speed of 10-20 cm/min in an effort to provide

greater time resolution. The value obtained (at a chart speed of

1 cm/min) for the response height equivalent to the CMC is then

transposed to the response curve obtained at 10-20 cm/min in order to

accurately determine the postinjection time at which the concentration

of surfactant is equivalent to the CMC.

Knowledge of the time at which the concentration of the injected

surfactant solution is equal to the CMC allows the experimenter to

determine the CMC through


CMC = (c.vi)/V exp[-(F/V)(t CC- to ti)] (eq. 2.23)


where c. is the molar concentration of the injected solution
S (mol/L),

v is the volume (uL) of the injected solution,

V is the volume (pL) of the gradient chamber,

F is the flow rate (pL/s),






























Figure 20. Detector response to the injection of an ionic surfactant
(CTAB) recorded at a chart speed of (A) 1 cm/min and (B) 10 cm/min.
The abrupt change in curvature in (A) allows for determination of the
response height equivalent to the CMC (R ) by visual inspection.
cmc
The error intervals included represent t plus and minus 3 s.
cmc





























0 60 120 180 240


Time


( sec )


180

150


E
0
-C
E
z4%
































































V
U)
Cr
-4
.4.
c
0
C,


C





~L







0
0C



0
LO



o0
C\j

o
00*
+

.. O
I


0







m 00 0000
W/ 0
ro





0 0 0 0 0 0 0

(wuo/O1lW/) >






























Figure 21. Detector response to the injection of a nonionic
surfactant (Triton X-100) into an aqueous carrier stream containing
Coomassie Brilliant Blue R 250 (1.3 x 10 ). The inflection point on
the descending portion of the curve allows determination of the
response height equivalent to the CMC (R ) by visual inspection.
The error intervals included represent t plus and minus 5 s.
cmc










0.875



0.750


0.625



0.500 -

C)
-Q-
0.375 -
-Q
O\ R
0.250- cc



0.125 --



0.000

0 60 120 180 240300360
Time (sec)












tCMC is the time (s) on the response gradient which corresponds
to the concentration at the CMC,

t is the experimentally determined time between injection and
the initial appearance of the peak, and

t is the theoretical time of the sample injection process
(Berthod et al., 1988).

This time, ti, during which the surfactant concentration

increases in the gradient chamber as a result of the sample injection

process is equal to v /F. Once the experimenter has defined the

constant parameters of the system (C, ci and vi), it is only necessary

to perform two injections and to measure the flow rate (using a buret

located at the outlet of the manifold) to determine the CMC of

surfactant solutions. The total time necessary for the determination

of CMCs of either ionic or nonionic surfactant solutions is less than

30 minutes and requires a minimal amount of both solution preparation

and surfactant, greatly conserving the material and personal resources

of the experimenter.



Results and Discussion

Considerations for Practical Operation

When attempting to design, characterize and effectively operate

an FIA system for the determination of CMCs of surfactant solutions

there are certain guidelines which should be adhered to in order to

ensure the accuracy and precision of the determination.

Exponential character. The exponential character of the system

was verified by the injection and subsequent absorbance detection of a










20 uL sample of 2.61 x 10 M Eriochrome Black T solution. A plot

(Figure 22) of log Absorbance vs. time (s) yielded a straight line

(average r = 0.998 for 4 trials). A 50 iL sample volume (maximum

syringe capacity) of 11.72 mM NaCl solution was then injected to

investigate the effects of increased sample volume on the exponential

response gradient. An exponential least squares treatment of the plot

2
of response vs. time again yielded a straight line (r = 0.999),

verifying the exponential character of the concentration gradient

produced by the mixing chamber for the range of sample volumes

employed.

Solutions of high surfactant concentration (e.g. >100X the CMC)

possess a viscosity which greatly exceeds that of solutions with

concentrations of the same order of magnitude as the CMC. Thus, the

effect of variations in the viscosity on the exponential character of

the gradient chamber was examined. The desired viscosities were

provided by varying the weight percentage of ethylene glycol from 0 to

48% in aqueous solutions of NaC1. This varied the relative viscosity

(CRC Handbook of Chemistry and Physics, 1980a) of the solution from

1.00 to 3.54 (versus water at 200C), adequately spanning the range of

viscosities encountered when employing this method (Figure 23). A

Q-test examining the exponential least squares slopes of a plot of log

conductance vs. time showed no statistical difference in the slopes

from 0-48% ethylene glycol, verifying that the viscosity of the

injected solute has no effect on the exponential concentration

gradient (Table IV). For comparison purposes, the average efflux time

































0

(D
C)







.0
V


0,








CIS
















4-4 0


0-4
LO



















0
I-4





















x

0
CX3






















0E
0

40
0

E13


am
cOL


















C x
LO
























Cj
L.





















C0
cx




























P-
a0)


oLE
i-






/ 31-

P co o ~0 -- nrq-n r-w m- -
I I I -I b O
I u qI I I I I I I I I I I
Iuoqiosqv boI
























IQ-

O4





O .--------



LQ I I I I I

25 50 75 100 125 150

Time (Sec)


Figure 23. Effects of increasing solution viscosity upon the
exponential character of the FIA manifold. The relative viscosity of
the injected solution increases (see Table VI), from left to right,
from 1.00 to 3.54 times that of water at 200C.












Table IV. Variation of relative viscosity (vs. 200C) and slopes of
log K vs. time as the weight percentage of ethylene glycol is varied
from 0-48% in the injected solution.


% Ethylene Glycol [NaCl] (M) n/n 200C Slope (x 103)


0 0.469 1.00 -10.14

10 0.426 1.27 -9.97

20 0.383 1.66 -10.03

28 0.347 2.04 -10.15

40 0.293 2.83 -10.09

48 0.256 3.54 -10.00


Data adapted from CRC Handbook of Chemistry and Physics (1980a).










in a #75 Ostwald capillary viscometer (Atkins, 1982) at 250.10C for

an aqueous, 40% ethylene glycol and 300 mM CTAB solution are given in

Table V. These results indicate that the 40% ethylene glycol solution

has approximately the same viscosity as the 300 mM (300X the CMC) CTAB

solution.

Effective mixing volume. In order to use equation 2.23, the

effective mixing volume of the FIA manifold must be experimentally

determined. The presence of varying amounts of preflow cell tubing in

different detectors ensures that the mixing volume of the system will

vary with the flow-through detector employed. Therefore, the

effective mixing volume of the system must be determined for each

different type of detector used. The time during which any initial

concentration of a purely exponential concentration profile decreases

by one half of its initial value (t 12) is defined as


t/2 = 0.693 V/F. (eq. 2.24)


The effective mixing volume of the system can be experimentally

determined by examining the variation of t/2 with the flow rate. A

plot of t/2 vs. 0.693/F yields a straight line with a slope equal to

the effective mixing volume of the system.

For conductance detection, a constant volume (10 pL) of a 0.47 M

NaC1 solution was repetitively injected into the manifold as the flow

rate was varied from 0.49-1.66 mL/min. The plot of t2 vs.

0.693/F yielded a straight line (r2 = 0.999) of the form

y = 0.765X + 3.28 x 10-3, resulting in an effective mixing volume

equal to 765 pL (Figure 24).














Table V. Determination of the efflux time for various solutions in a
#75 Ostwald capillary viscometer.


Solution Efflux Time (s) Average Efflux Time (s)


H20 105.9 105.7
105.6
105.7

40% Ethylene Glycol 257.6 258.0
258.0
258.5

300 mM CTAB 253.3 254.2
255.0
254.2


a Temperature 250.10C.























o) 0-4
~CE
4-) 0C


OQ) 0

E a
~xE
r-4 a).
0
>cO)

bo cl
Cb~ L
-0
x 3
4 C 0
r= 0 -4
C4







0 C




"-4 a


) (D4.)
O 0d





0aa






r=r



cuil
0
02 0










O
0
0 -4 r-




co"












03r
LQ)0







.) -'-4

C) 4-)



Ca 0


4C)
'OEO






bo c r

>-)- 0
N ..
'-t 0
4-H~.



OO


~ Cc



d0
02-4





77














O C


coo
E








-- D \ ODr
(0




(UIl)


I
LL.


-0










When using absorbance detection, an accurate value of V is best

obtained by monitoring the absorbance of a compound which has a

spectral region (wavelength of maximum absorbance or absorbance

plateau) which does not shift upon dilution by the carrier stream.

This ensures that changes in response are due only to concentration

changes and not spectral shifts (i.e. a 50% decrease in the response

curve truly corresponds to a 50% decrease in concentration). The

effective mixing volume for absorbance detection was determined by the

injection of 10 iL of a 2.62 x 10-3 M aqueous solution of Eriochrome

Black T and subsequent 437 nm detection. This compound exhibits an

absorbance plateau from 433-441 nm (Figure 25) and an accompanying

linear calibration curve over the concentration range of

2.38-95.2 x 10" M. The plot of tl/2 vs. 0.693/F for flow rates

ranging from 0.52-1.91 mL/min yielded an effective mixing volume of

814 uL.

Sample injection process. An HPLC sample injection valve is a

convenient way to introduce samples into an FIA manifold. In

chromatographic separations, it is a common practice (Bakalyar and

Spruce, 1983) to use partially filled sample loops for sample

introduction. After injection, the sample is concentrated at the head

of the column, with no loss in sensitivity or reproducibility due to

the injection process.

In FIA, however, dispersion of the injected sample (and

accompanying loss of sensitivity) occurs as a result of its

interaction with the surrounding carrier stream. This interaction

between the sample and the carrier stream can also occur in a
































































































-4-
I cl





f-4




U') 0
C\j
0)




-4c
E-4















700.0







600.0







500.0

E





400.0







300.0


0 0 0
Dilution factor
Dilution factor










partially filled sample loop before the injection is made. Therefore,

it is necessary to use a completely filled sample loop to ensure the

greatest sensitivity and a reproducible introduction of both a

constant amount and concentration of sample at the point of injection.

In order to quantify this phenomenon, a comparison of a 10 yL

injection via a fully loaded 10 UL loop (20 UL load) and a partially

loaded 100 pL loop (10 yL load) was undertaken. For this

investigation of the sample injection process, the sample loop and

pregradient chamber tubing were connected directly to the absorbance

detector flow cell. The injected compound, 2.39'x 10-5 M aqueous

methyl red, was detected at 430 nm. As shown in Table VI, the loss of

sensitivity ranges from approximately 21-26% in the partially filled

loop when compared to the fully loaded 10 yL loop. In all trials, the

time from completion of sample load to sample injection never exceeded

5 s. In the partially filled loop, an increase in the length of time

between sample loading and injection results in an increase in

dispersion, further decreasing sensitivity. The results of Table VI

also indicate that the average relative standard deviation of the

sample injection process is less for the completely filled loop (1.3%)

compared to the partially filled loop (2.0%). Therefore, to ensure

the greatest possible reproducibility of the technique, a completely

filled sample loop is again recommended.

Practical use of equation 2.23. The theoretical treatment we

derived earlier (Berthod et al., 1988) assumes that the surfactant

solution is injected with a plug profile whose duration, ti, is














Table VI. Investigation of the sample injection process.


Maximum absorbancea (x 102)
Average Ratio of
Flow Rate 100-UL Loop RSD(%) 10-yL Loop RSD(%) Responses
(mL min-1) (10-IL load) (20-yL load) (100 uL:10 PL)


0.25 3.68 2.6 4.76 1.3 77.3

0.52 3.85 2.4 5.05 1.3 76.2

0.97 4.48 1.7 5.73 1.5 78.2

1.45 4.37 1.3 5.70 1.8 76.7

1.93 4.05 2.1 5.44 0.8 74.4


a Injected
430 nm.


-5
compound, methyl red (2.39 x 10-5 M);
The number of observations was >12 in


detection at
all cases.










v./F. As shown in Table VII, this is never the case; the injection

has an exponentially modified Gaussian profile as shown by Figure

26. Then, the injection time, ti, is greatly underestimated when

calculated as v /F. For example, when 10 yL were injected with a

20 VL/s flow rate (1.2 mL/min), the theoretical value of t is 0.5 s

(Figure 26). The actual value may lie between 5 and 15 s, that is

between 10 and 30 times higher than the theoretical value. To use our

theoretical treatment (Berthod et al., 1988), it is possible to

describe the injection as a plug of lower concentration, c ', and

higher volume, vi', such that the product c 'vi" is equal to civi,

as shown by Figure 26. This replaces the theoretical value, ti, of

the injection time by an empirical one, t '. It is not possible to

mathematically quantify this effect because the band broadening,

occurring during the injection process, depends on the connecting

volume, on the flow rate and on the viscosity and concentration of the

injected solution. To minimize the error on the CMC determination,

the connecting tubing volume must be as low as possible, and the

injected surfactant concentration must be as high as possible.

Indeed, the higher the surfactant concentration, the longer the

time, tCMC and the smaller the relative influence of t. on the CMC
CMC 1
value (equation 2.23). It must be noted that there is an upper limit

to tCMC; the observed break must occur in the middle of the range

used. If the break occurs too close to the baseline (i.e.

t CMC is too long), it will be unnoticed.

Empirically, we found that the maximum value of the recorded

signal (conductance or absorbance) could be used to obtain an