Indirect fluorometric detention of arsenic and selenium oxyanions separated by capillary zone electrophoresis and ion ch...


Material Information

Indirect fluorometric detention of arsenic and selenium oxyanions separated by capillary zone electrophoresis and ion chromatography
Physical Description:
vii, 148 leaves : ill. ; 29 cm.
Gunshefski, Maryann, 1966-
Publication Date:


Subjects / Keywords:
Water -- Analysis   ( lcsh )
Trace analysis   ( lcsh )
Arsenic compounds -- Analysis   ( lcsh )
Selenium compounds   ( lcsh )
Chemistry thesis Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1993.
Includes bibliographical references (leaves 142-147).
Statement of Responsibility:
by Maryann Gunshefski.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001933252
oclc - 30784209
notis - AKA9319
System ID:

Full Text








During my time here at the University of Florida, I have met many people who

have helped me in countless ways. First, I would like to thank Steve Lehotay for his

help in learning about indirect fluorometric detection, and in setting up some of my

initial instrumentation. I want to thank Anthony Chuck, an undergraduate student who

worked with me during my 3"' year of graduate studies. His questions and input were

invaluable and the time he spent in lab was very enjoyable. I want to also thank Andi

Pless, a fellow graduate student who worked with me for the past year, for all of her

help with solutions, experimental runs, and questions which brought new insight to this

research in numerous ways.

I want to sincerely thank Dr. Bob Kennedy and Dr. Ben Smith. Their answers

to questions I had about CZE and spectroscopy were indispensable to me, and I am

grateful for their patience and understanding. I would also like to thank Dr. Jim

Winefordner for the opportunity to be a part of his group, and for his support during my

latter years at UF. In addition, I would like to thank the entire Winefordner group, and

others who I have met here in Gainesville, for making my time here an unforgettable

experience. I want to wish them all the best for the future. I appreciate the monetary

support granted to me by the state of Florida during the semesters which I was on

teaching assistantship, and from the Environmental Protection Agency (EPA) for funding

the research project which I was a part of.

I would also like to thank my parents for always believing in me and teaching me

the importance of education, and love.

Mostly, I want to thank my husband, Bobby, for his support, understanding,

patience, and love, especially during the stressful times. I am lucky to have him with

me, always.


ACKNOWLEDGEMENTS .................................. ii

ABSTRACT ......................................... vi

CHAPTER 1 INTRODUCTION ............................. 1
Environmental Significance ............................. 1
Arsenic and Selenium ............................ 2
EPA Regulations ................................ 8
Indirect Fluorometric Detection ......................... 14
Theory ................ .................... 14
Probe Species ................................. 19
Scope of Dissertation ................................ 20

Theoretical Aspects of Capillary Electrophoresis ................ 21
Definitions .................................. 22
Types of Analytes .............................. 26
Detection Methods .............................. 28
Using a Diode Laser as an Excitation Source ................. 33
Introduction to Diode Lasers ....................... 33
Experimental Section ............................ 35
Optimization of Parameters ........................ 41
Discussion ................................... 48
Using a He-Cd Laser as an Excitation Source ................. 48
Experimental Section ............................ 49
Optimization of Parameters ....................... 53
Results ..................................... 58
Discussion .................................... 80

Theoretical Aspects of Ion Chromatography .................. 82
Definitions .................................. 83
Ion Exchangers ................................ 86

Eluents .......................... ........ ...... 87
Using a Diode Laser as an Excitation Source ................. 90
Experimental Section ............................ 91
Optimization of Parameters ....................... 94
Discussion ...................................107
Using a He-Cd Laser as the Excitation Source ................. 107
Experimental Section ............................108
Results ................... .............. .. ..111
Discussion ...................................123


REFERENCE LIST ...................................... 142

BIOGRAPHICAL SKETCH ................................. 148

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



MaryAnn Gunshefski

August 1993

Chairperson: James D. Winefordner
Major Department: Chemistry

In order to satisfactorily assess water supplies for the presence of pollutants, it

is necessary to be able to detect many different chemical species which are present at

very low concentrations. The presence of arsenic and selenium is currently reported as

total arsenic and total selenium, which does not give an accurate assessment of the water

quality. It has been shown that the different species of these elements have different

toxic effects on humans. The purpose of this research was to investigate the use of

indirect fluorometric detection for the detection of arsenic and selenium oxyanions in

water samples. Both capillary zone electrophoresis (CZE) and ion chromatography (IC)

are utilized to separate these analytes, with a comparison of the results being presented.

Two different laser sources are evaluated with each separation technique. The

first is a diode laser-based system. This excitation source was chosen because of its low

noise characteristics. The dyes which can be used at the wavelengths available with

diode laser systems were found to react unfavorably with the capillary (in CZE), and the

stationary phase (in IC).

The second system which was evaluated is a helium-cadmium (He-Cd) laser-based

system. The selection of dyes which can be used with this source is much larger than

with the diode laser excitation, and several different dyes are evaluated with each

separation technique. Separation of four arsenic and selenium oxyanions is demonstrated

with CZE and IC under the conditions necessary for indirect fluorometric detection.

With a CZE separation, absolute detection limits in the femtomole range are achieved.

This corresponds to concentration detection limits in sub-ppm range. Utilizing an IC

separation, absolute detection limits are in the picomole range. This also corresponds to

concentration detection limits in the sub-ppm range.

This research demonstrates the viability of using indirect fluorometric detection

for the analysis of arsenate, arsenite, selenate and selenite in a simple water matrix with

either CZE or IC separation. The detection limits achieved with these systems are

comparable to those now published for these anions by other methods.


Environmental Significance

Our knowledge and understanding of the environment in which we live is

increasing every day. One of the concerns addressed by the Environmental Protection

Agency (EPA) is that of potable water quality. To be able to satisfactorily assess

drinking water supplies for the presence of pollutants and their associated risks to humans

and the environment, it is necessary to be able to detect the many different chemical

species of an element which are present. This is due to the varying toxicities of the

elements in their different oxidation states. The analytical chemist has been called upon

to develop methods to meet these demands. Currently, the presence of arsenic and

selenium in environmental samples is reported as total arsenic and total selenium. This

does not give an adequate assessment of the drinking water quality. Recently, methods

have been developed to separate and quantify the various species of arsenic and selenium

present in water samples. These techniques include capillary zone electrophoresis (CZE)

and ion chromatography (IC). Now that separation of these compounds is possible,

improved detection methods are necessary to be able to detect ultra-trace levels of these

compounds which may be present in water supplies. This would improve our ability to

accurately assess water quality, and health hazards which may exist in the community.

Arsenic and Selenium

Arsenic has been recognized as a poison since antiquity. In high doses, it can

cause nausea, vomiting, abdominal pain, brittle nails, dry flaky skin, hair loss, burning

and tingling of extremities, mental confusion, and shuffling locomotion. In humans, it

tends to concentrate in hair, nails and teeth, with only minor absorption into other body

tissues. In spite of this fact, arsenic has been used for its "therapeutic" effects during

the 1800s.' The most common formulation was named "Fowler's Solution" and

contained 10 mg/ml of As203. It was prescribed for symptomatic relief of many

conditions ranging from acute infections to epilepsy, asthma, psoriasis and eczema. For

obvious reasons, this practice was abandoned with the advent of safer remedies for these

conditions. Arsenic occurs naturally in nature, and is not often found as a free element

but rather as part of a compound. In water, arsenic is most commonly found in the +5

oxidation state in the form of arsenic acid, H3AsO4, but can also be found in the +3

oxidation state as arsenous acid, HAsO2. The acid dissociation constants for arsenic acid

are pK., = 2.20, pKa = 6.97, and pK, = 11.53:2

H3AsO4 H2AsO4'+ H*
H2AsO41 4 HAs04-2+ H+ (1-1)
HAsO042 4 AsO4-3+ H'

The pK, for arsenous acid is 9.18:2

HAsO2 4 AsO21+ H' (1-2)

Thus, in neutral water, As(V) would exist as a mixture of the singly protonated arsenate

dianion and the diprotonated arsenate anion, where As(II) would exist mostly in its


nonionic protonated form. Arsenic also exists in methylated forms due to animal and

plant metabolisms. The ratio of As(V):As(III) in a water system can vary greatly, for

example, the ratio in seawater ranges from 0.1:1 to 10:1. Figures 1-1 and 1-2 show the

effect of pH on the dissociation of H3AsO4 and HAsO2, respectively.

Selenium is noted for its unusual trait of being a poison at high levels, yet being

a necessary nutrient at lower levels.3 It has been found that a concentration of up to 0.1

ppm is beneficial to humans, but a concentration of 0.4 ppm is considered toxic.3 The

presence of a small amount of selenium in the human body helps to prevent damage by

oxygen to tissues, as does vitamin E.3 Too little selenium is associated with a

cardiovascular disease known as Keshan disease, and an osteoarthropathy disease named

Kashin-Beck disease. The effects of ingesting too much selenium include an upset

stomach, difficulty in breathing, and weakness. Over time, the symptoms include brittle

nails, loss of hair and nails, skin lesions, tooth decay, and, in the later stages,

convulsions, paralysis and motor disturbances.

Selenium is most commonly found as sodium selenite, NaSeO3, and sodium

selenate, NaSeO4. In water, selenium can exist in the Se(VI) oxidation state in the form

of selenic acid, H2SeO4, which has pK., = 0.3, and pK, = 1.92:2

H2SeO4 HSeO- + H* (1-3)
HSe04' e Se4O2+ H'

Se(IV) can also occur in water in the form of selenous acid, H2SeO3, which has pKI =

2.46 and pK.2 = 7.31:2

u IO

0 Q



o o3 CO C'J -D 0nc3

(9-3a0 SaLunL)
(W) uot-vJ.IIuzo 3




^ 8



cO r- co rl CO mn C\ C 2 o

(g-HO T sau:u.l)
(Wy) -Osv0 jo uoTlt'e.luaouoD

H2SeO3 a HSeO3-+ H" (1-4)
HSeO3'1 4 SeO'2+ H'

In a typical water sample, Se(VI) is in the form of the selenate dianion, and Se(IV) exists

as a mixture of deprotonated and singly protonated selenite anions. Selenium also exists

as methylated forms produced by metabolic processes in the environment. Figures 1-3

and 1-4 show the effect of pH on the dissociation of H2SeO4 and H2Se03, respectively.

The analysis of arsenic and selenium in environmental samples at the trace

concentrations levels necessary for and accurate assessment of water quality has been a

difficult problem. Various analytical techniques have been investigated including: flame

atomic absorption spectroscopy,4 graphite furnace atomic absorption spectroscopy,5'

theromochemical hydride generation-atomic absorption spectrometry, inductively-coupled

plasma-atomic emission spectroscopy,8 suppressed ion chromatography with conductivity

detection,9 capillary zone electrophoresis with UV detection,t0 ion chromatography with

atomic absorption spectrometric detection," among others. Without some type of initial

separation, most of these techniques can only determine total arsenic or total selenium

present in a water sample.

EPA Regulations

In the early 1970s, it was indicated that a significant number of water supplies did

not meet the established standards of water quality. It was at this point that legislation

was established which provided the EPA with the ultimate authority over all water

supplies. The Safe Drinking Water Act of 1974 (SWDA) established regulations for

" -


MN g0



O co r- cz k n ~

(W -aor s4uan1o)
(I~) uoq~~nqueouoj

~ ,



- __

w a-




cc r-O cm LI -14 C\2~

og-a oT saLUT O
(W) UOTWJ1U-4uG3uoj


contaminants in drinking water supplies. To assure that the water we drink will cause

no adverse effects on health, recommended maximum contaminant levels (RMCLs) were

instituted. Since its beginnings in 1974, many amendments have been adopted to the

SDWA. These new amendments have updated the SDWA with safer maximum

contaminant levels (MCLs) as well as providing guidelines for the analysis of these

contaminants. These guidelines, called "standard methods," allow all water samples

across the country to be analyzed in a similar fashion, and then to be compared to one

another, as well as with EPA standards.

Arsenic and selenium are both considered toxic metals, and appear on the list of

contaminants which are regulated under the SDWA. The MCL in drinking water for

arsenic has been set at 50 pg/L, and for selenium has been set at 10 /pg/L. It should be

noted that these levels are for the total arsenic and total selenium present in a water

sample. As the toxicological studies continue, it is becoming evident that the different

forms (and oxidation states) of these metals can cause different toxic effects. Because

of this, it may become necessary to regulate each form separately, which necessitates the

development of techniques to analyze these species in water samples. For routine

analysis, techniques which are reliable, reproducible and relatively simple to perform are

best. In addition, the instrumentation should be easy to maintain, and be rugged enough

for daily use.

Indirect Fluorometric Detection

Small and Miller" were the first to characterize Indirect Photometric Detection

(IPD) in 1982, although others had utilized indirect detection methods previously.13"1

IPD was first developed for the universal detection of non-absorbing analytes separated

by ion chromatography. This detection method eliminated the need for a suppressor

column after the analytical column. The performance of Indirect Fluorometric Detection

(IFD) is basically the same as for IPD except that it is more sensitive to the probe

molecule, and therefore more sensitive than IPD. IPD and IFD have been applied to

many separation techniques including ion chromatography,1'2,1 HPLC,23-24 thin-layer

chromatography,2 gas chromatography,26 and capillary electrophoresis.V27- The use of

indirect detection has been reviewed in several articles.3'33


Indirect fluorometric detection is based on a fluorescing probe species being

present in the solvent, producing a constant background signal. When an analyte ion is

present the probe is displaced and the fluorescence signal at the detector decreases.

Since it is the presence of the analyte which causes the signal, and not the characteristics

of the analyte, indirect detection is classified as a universal detection method. This

allows the analysis of many analytes that cannot be detected by standard chromatography

detectors, such as conductivity or UV-vis absorption. Figure 1-5 shows a schematic of

an indirect detection measurement. The upper portion shows a segment of the capillary

(or flow cell) which contains the probe species as part of the eluent. The analyte zone





0 is
0 0 0 i
0 0
0 00
* e* = .
9 w 28^\.
S0 "
* 0 0 i
0 0 1
00 0N
000 0 ___
r~ugis 93U33SJoflu

in the figure shows how the probe is displaced by the analyte in that portion of the

capillary. The lower portion illustrates a segment of the detector response. It shows

how the fluorescence signal decreases when the analyte zone passes the detector window.

For a chromatographic separation, the resulting chromatogram consists of a series of

dips, rather than peaks.

Several parameters are extremely important when developing an indirect detection

method. The first of these is the dynamic reserve (DR). This is related to the signal-to-

noise ratio of the system, and provides a measure of the ability to detect a small change

in a large fluorescence background signal.31 For a laser-induced fluorescence background

signal, the stability of the laser plays a major role in the value of DR. For a common

laser-based signal, a typical DR is between 102 to 103. This DR is low when compared

to some other indirect techniques: for indirect UV/vis absorption, DR is around 5 x 103;

for refractive index detection, DR is around 106; and for polarimetry, DR is greater than


The second important parameter is the transfer ratio (TR). This is defined as the

number of probe molecules displaced per analyte molecule.31 Experimentally, it can be

determined by measuring the signal produced by a known amount of probe species and

dividing by the change in signal produced in the presence of a known amount of analyte:

TR = slope of analyte calibration curve
slope of probe calibration curve

The displacement which occurs can be a charge displacement, a volume displacement,

a chromatographic partitioning, or a combination of all three. For ion chromatography

and capillary electrophoresis, a charge displacement mechanism prevails, based on the

laws of eletroneutrality. Theoretically, this mechanism asserts that one analyte ion

displaces one probe ion of the same charge, producing a TR of one.

Reversed-phase HPLC of neutral species relies on the solubility displacement

mechanism for IFD. The presence of an analyte will cause the partitioning of the probe

species between the stationary phase and the mobile phase to change. Usually analytes

with shorter retention times than the probe species will induce a solubility enhancement

of the probe into the mobile phase resulting in a positive peak. Those analytes with

longer retention times than the probe species induce a solubility decrease, and negative

peaks result.35 The TR for HPLC is typically 104."

The detection limit for indirect detection is based on DR, TR and the

concentration of the probe species, Cm. This relationship is expressed as:

C Cm (1-6)
(TR x DR)

where Cu is the theoretical concentration limit of detection.31 According to this equation,

the LOD can be decreased by lowering the concentration of the probe species while

keeping the values of TR and DR constant. This is usually limited, however, by the

flicker noise of the source. In most cases, when the concentration of the probe is

lowered, the TR and DR values also change. In one laser-based approach, the

fluctuations in the power of a He-Cd laser were compensated by special optics and a

reference cell.37 Another problem which can arise when conventional sources are used

is the fluorescence and/or absorption of an analyte at the detection parameters, which

would interfere with the basic detection principle for indirect detection.

Probe Species

The selection of the probe species to be used in an analysis is of crucial

importance. Many key factors must be considered. First, the probe must be able to

provide a high background signal for the type of detection being utilized. For indirect

fluorescence, the probe should have a high molar absorptivity and fluorescence quantum

yield, which helps to achieve the maximum sensitivity for the system. Another

consideration is the desired charge of the probe. The most efficient sensitivity occurs

when the transfer ratio is one, which occurs with a charge displacement mechanism

prevailing. This dictates that the probe species should have the same charge as the

analyte of interest. This works especially well with ion chromatography and

electrophoresis. A variety of probe species have been used with these separation

techniques for both indirect UV-vis absorption and indirect fluorescence including:

phthalate; 12,17,1923 salicylate; ,28,37-39 benzoate; 6 2,4-dihydroxybenzoate;" quinine sulfate;30

and methylene blue.4 One other factor to be considered is the compatibility of the probe

with the separation technique. It would be detrimental to use a probe which would

inhibit the separation from occurring by irreversibly attaching to the stationary phase or

the capillary wall. It is also necessary to determine if the probe is soluble in the chosen

solvent, and if the probe interacts with the analytes in unfavorable ways.

Scope of Dissertation

The purpose of this dissertation is to investigate the use of indirect fluorometric

detection for the detection of arsenic and selenium oxyanions in water samples. Both

capillary zone electrophoresis and ion chromatography are used to separate these

analytes, with a comparison of the results being presented. Two different laser systems

are used with each separation technique. The first is a diode laser-based system. This

type of laser system was chosen due to its low noise characteristics. Several near-IR

dyes will be evaluated for their use as probe molecules in this type of system. The

second is a He-Cd laser based system which utilized a commercial spectrofluorometer

as the detection device. A variety of new probe molecules are evaluated with this

system. This dissertation will show that this combination of CZE or IC with indirect

fluorometric detection can provide a sensitive and universal alternative for the analysis

of arsenic and selenium oxyanions in water samples.


Theoretical Aspects of Capillary Electrophoresis

Capillary zone electrophoresis (CZE) is a method which utilizes an electric field

to facilitate a separation of charged species. The advantages of CZE over the more

common technique of high performance liquid chromatography (HPLC) include the

smaller sample sizes needed, and the higher number of theoretical plates achievable.

Since it is a similar technique to HPLC, many of the terms used in CZE have been

adopted from HPLC. Mikkers, Everaerts and Verheggen41 were the first to perform CZE

in narrow bore tubes. They were able to achieve separation of 16 anions in less than 10

minutes. Jorgenson and Lukacs42 were the next to describe a system using open-tubular

glass capillaries for performing zone electrophoresis. CZE has been used for the analysis

of anions and cations spanning from small inorganic ions to large biomolecules. The

separation of neutral molecules has also been achieved by using micelles in the buffer

solution.43"* The neutrals partition between the micelles and the buffer solution, which

drives the separation. This type of analysis has been very effective for this type of



A basic CZE system consists of a fused silica capillary tube filled with a buffer

solution, and with each end of the capillary immersed in a vial also containing the buffer

solution. An electrode is placed in each vial, and a potential is applied across the

capillary. Detection is normally performed "on-line" by removing a small section of the

polyimide coating which coats the capillary. This window in the capillary coating then

becomes the detector "flow cell". When the potential is applied, each ion will then move

toward one of the electrodes, and pass the detector window, according to its effective

charge and size, hence its mobility. The velocity at which each ion moves is called the

migration velocity and can be described by the following equation:

L (2-1)

where v is the migration velocity (m/s), A is the mobility (mn/Vs), V is the total applied

voltage (V), and L is the total length of the capillary (m).45 The time required, tL, for

the analyte zone to migrate the entire length of the tube is:

L L2
tL = = LF (2-2)

Zone broadening is caused by molecular diffusion, and the spatial variance, a,, is given

by the Einstein equation:

a = 2Dt
= 2t (2-3)
where D is the molecular diffusion coefficient of the solute in the zone (m2/s). The

efficiency of the CZE separation can be expressed in terms of theoretical plates, as is

done in chromatography. The number of theoretical plates, N, can be defined as:

N 2D (2-4)

and can be calculated from peak profiles as in chromatography by using the formula:

tL (
N = 5.54 (2-5)

where w, is the width of the peak at half of its height (s), and tL is the migration time

(s). The equation for the height of a theoretical plate (H) is then given as:4S

N (2-6)

In CZE separations, hundreds of thousands of theoretical plates can be achieved. This

provides an enhanced separating ability to that of conventional HPLC.

In CZE, there is also another force which is affecting the separation. This is the

electroosmotic mobility, which is due to the induced bulk flow of the liquid buffer. The

inner wall of the fused silica capillary possesses a negative charge due to the silanol

groups of the silica. The positive ions in the buffer migrate toward these silanol groups.

When the electric field is applied, these positively charged ions move toward the negative

electrode, thus creating the electroosmotic flow. The apparent mobility of an ion, 1/,,

is the critical parameter in CZE; Ci, is the combination of the electroosmotic mobility

of the buffer ,C., and the electrophoretic mobility of that ion ,/p. This relationship is

given by:

/ = app +-L. e(2-7)

In the case where the electroosmotic mobility and the electrophoretic mobility of the

analyte are in the same direction, the values are added. If they are moving in opposite

directions, the values are subtracted. Many times the detection is performed on-line by

making a detector window at some designated length of the capillary. The apparent

mobility can be obtained in terms of easily measurable quantities and thus calculated by

the following equation:

t 1tV (2-8)

where L, is the length of the capillary from the injection end to the detector window, and

LT is the total length of the capillary. The electroosmotic mobility of the buffer can be

calculated in a similar fashion by injecting a neutral molecule, one that will only move

through the capillary due to the force of the flowing buffer, and recording its migration

time (t.):
o tVg (2-9)

By combining equations (2-6) and (2-7) it can be shown that:

LDL 1 1
P V t- (2-10)

It is important to report data in terms of electrophoretic mobility, instead of

migration time. There are several reasons for this: first, it is sometimes hard to control

the electroosmotic flow, making mobility values more reproducible than migration times;

second, migration times are dependent on many experimental conditions, which can vary

from day-to-day and run-to-run; third, by reporting results as electrophoretic mobilities,


direct comparisons can be made easily and effectively; and lastly, it allows you to

correlate the electrophoretic behavior of a solute with its structure.

The resolution (R) of two peaks can also be calculated in a similar fashion to that

which is used in chromatography:

R = APLP VVF (2-11)
4 u

where AI, is the difference of the two mobilities, ji, is the average of the two

mobilities, and N is the number of theoretical plates.

There are basically two types of injections which can be used in CZE. The first

of these is hydrodynamic injections. For this type of injection, one end of the capillary

is removed from the buffer vial and placed in a vial containing the sample. This vial is

then raised a predetermined height for a specified time. During this time, the sample is

drawn up into the capillary by siphoning action. The sample vial is then lowered, and

the end of the capillary is returned to the buffer vial. The separation is then performed.

The volume injected, Vi, (ml), can be calculated by the Poiseuille equation:

V pgAhTD4t, (2-12)
= 128qLr

where p is the density (g/cm3), g is the gravitational constant (cm/s2), Ah is the height

difference (cm), D is the inner diameter of the capillary (cm), t, is the injection time (s),

,q is the viscosity of the buffer solution (g/cm s), and L is the length of the capillary

(cm). Hydrodynamic injection provides the most precise injection of a sample because

it is based strictly on volume loading of the sample.

The second type of injection is called electrokinetic injection. With this technique

one end of the capillary, along with the corresponding electrode, are placed in the sample

vial. A voltage is then applied for a designated amount of time. During this time, the

electroosmotic flow and the individual electrophoretic mobilities determine the amount

of sample which is introduced into the capillary. After the allotted time, the end of the

capillary and the electrode are returned to the buffer vial, and the separation is

performed. The amount of analyte injected can be calculated by the equation:

Q rr2c(s p,+p/)Eti~ b (2-13)
Q inj x J

where Qnj is the amount of sample injected (mol), r is the radius of the capillary (m), c,

is the concentration of the sample (mol/L), Pep is the electrophoretic mobility of the

sample (cm2/V s), po is the electroosmotic mobility (cm2/V s), E is the electric field

strength (V/cm), ti is the injection time (s), k is the conductivity of the sample buffer

electrolyte, and X, is the conductivity of the sample. With electrokinetic injection, ions

with higher mobilities are concentrated in the injection volume, whereas ions with lower

mobilities are diluted in the injection volume. This characteristic can be useful when a

preconcentration step is desired, but can be detrimental to an analysis if this is not


Types of Analytes

As was mentioned earlier, a wide variety of analytes can be separated and

detected using CZE systems. These include inorganic anions,4649 inorganic cations,50

organic acids,51s53 organic bases," amino acids,4255-58 peptides,59 and proteins.59 Since

these analytes possess many different characteristics, in order to separate them all, some

basic experimental parameters need to be altered. In the conventional configuration, the

electroosmotic flow is toward the negative electrode with the sample being injected in the

end of the capillary near the positive electrode. With this setup, the cations will be

detected first (since they are moving with the electroosmotic flow toward the negative

electrode), the neutrals will be carried with the flow, and the anions (which are moving

against the flow) will be detected last, if at all.

Anions. In some cases, where the electroosmotic flow is not very fast, and the

analytes are small inorganic anions, they will stay at the injection end and not be

detected. One way to overcome this is to reverse the voltage. Since the injection end

would now be negative, the anions would migrate through the capillary, and pass the

detector window. There could still be a problem, however, since the anions continue to

move against the electroosmotic flow (which is moving toward the negative electrode).

In this case, a slow electroosmotic flow would be desired, so as not to "push" the

analytes back toward the negative electrode. This can be accomplished by treatment of

the capillary with a chemical derivitizing agent, such as trimethylchlorosilane. This

process, called silylation, changes the charge density of the capillary wall.28 The next

option is to reverse the electroosmotic flow. This can be accomplished by coating the

inner surface of the capillary which will change it to carry a positive charge. In this

configuration, the anions in the buffer will migrate toward the walls, and upon the

application of an electric field, the solution will flow toward the positive electrode.

Figure 2-1 shows a schematic of this flow reversal process. This reversal has been

accomplished by adsorbing a coating onto the capillary wall. The most common type of

coatings involve cationic surfactants which are added to the buffer system. Some of the

surfactants which have been shown to reverse the electroosmotic flow include:

dodecyltrimethylammonium bromide (DTAB),60 tetradecyltrimethylammonium bromide

(TTAB),60 hexadecyltrimethylammonium bromide (CTAB),60 and

cetyltrimethylammonium chloride.61 Waters Chromatography Division has even developed

their own "electroosmotic flow modifier" to be used with their CZE system for the

analysis of anions by indirect UV-absorption spectroscopy."" Their modifier is a

proprietary compound, and all they will reveal is that it is a type of alkylammonium

compound. Another type of compound which has been used to reverse the

electroosmotic flow is diethylenetriamine (DETA) by Dionex Corporation.62

Detection Methods

Due to its enhanced separating abilities, CZE is becoming a more popular

technique for analysis of ions. Analytical chemists are constantly trying to achieve lower

detection limits for more and more compounds. Many advances in detection methods

have been explored in conjunction with CZE. Initially, UV-vis absorption and

conductivity were the methods of choice.

Absorption. Absorption detection is easily adaptable to CZE, since a wide variety

of analytes do absorb in this region, and the detection can be performed on-line.'63"

The drawback to this method is its decreased sensitivity from the use of small inner


W c


+ I
+ +

+ +

+ +

+ +

+ +

+ +
+ +

+ +

+ +

+ +

+ +
+ +

+ +













diameter capillaries, giving a small pathlength (normally < 100 /lm). One novel

approach which was investigated to provide increased sensitivity is called axial-beam

absorbance detection." In this technique, the light beam is directed along the capillary

axis. The resulting pathlength is then equal to the analyte band width. A 60-fold

improvement in path length was achieved.

Conductivity. The use of conductivity detection can provide an increase in

sensitivity, but it also requires some type of sample flow cell other than the capillary

itself. Several designs have been published in which a capillary detector cell was

designed. The detector cell had a similar inner diameter as the capillary, and was

adapted to allow platinum electrodes to be embedded into the wall.51"' An on-column

conductivity detector was designed by Huang and co-workers consisted of two platinum

wires sealed into 40--m holes penetrating the walls of the fused-silica capillary." The

disadvantages of this technique are mainly baseline drift, and poor detection limits.

Fluorescence. Fluorescence detection has also been utilized with CZE. As with

absorbance detection, on-line detection is easily achieved. The use of fluorescence

provides an enhancement of sensitivity over absorption, but the range of analytes which

can be detected decreases. Initially, a broad source was used for excitation (for example,

a xenon-arc lamp),6 but more recent advances involve the use of lasers as excitation

sources. Currently, the most sensitive detection limits, in the attomole range, have been

achieved with laser-induced fluorescence (LIF).58'6869 Lasers which have been used as

excitation sources include the cw helium-cadmium (He-Cd),70 the cw argon ion (Ar+),71

and a cw semiconductor (diode) laser4. To extend the applicability of fluorescence

detection to non-fluorescence molecules, fluorescent labeling is often used. This type of

analysis has been performed with amino acids, with zeptomole detection limits for

fluorescein thiohydantoin derivatives of amino acids being reported by Wu and Dovichi.5s

Mass spectrometry. Mass spectrometry is potentially an ideal method for CZE.

It would allow direct identification of each analyte zone, as well as low detection limits.

Electrospray ionization was the first type of interface to be used.' The difference

between MS detection and the previously mentioned ones is that the detection end of the

capillary is not placed into a vial, but rather into the MS interface. A metal needle or

a thin film of metal deposited on the capillary end ensures electrical contact and allows

the appropriate voltage potential to be created along the capillary. One problem which

arises with MS detection occurs with aqueous buffers. Many times organic modifiers are

added to the buffer system to increase its volatility. Other researchers have developed

a fast-atom bombardment-mass spectrometry (FAB-MS) interface for a CZE system.7"74

Detection limits in the femtomole range have been reported with this type of system.

Indirect detection methods. Both indirect absorption and indirect fluorescence

have been used for detection with CZE. Kuhr and Yeung were the first to utilize and

indirect fluorescence detection scheme for the analysis of amino acids separated by

CZE.2 Since that time, several studies have been performed with indirect absorption

detection for such analytes as: organic acids,52-53 amino acids,52 inorganic anions,4849,-776

polyamines,56 and inorganic cations.7 Jones and Jandik have been successful in

separating 30 anions in 89 seconds using CZE with an electroosmotic flow modifier and

indirect absorption detection.48 Indirect fluorescence has also been investigated further.


Analytes detected by this method include: nucleotides,28 nucleosides,28 sugars,29 inorganic

cations,30'78 amines,30 amino acids,4 and inorganic anions.47"78 Several review articles

covering indirect detection have also been published.79s Indirect detection has the

advantage of being a sensitive and universal detector. It eliminates the need of pre- or

post-column derivitization, and offers the possibility of one calibration curve for a series

of analytes.

Using a Diode Laser as an Excitation Source

One way to reduce the detection limits in an indirect fluorometric detection system

is to increase the signal-to-noise ratio. This can be achieved by increasing the intensity

or decreasing the noise of the excitation source, which is most often a laser. Diode

lasers (also called semiconductor lasers) can offer this improved stability. For example,

the power stability of a He-Cd laser is about 0.1%, whereas the power stability of a

diode laser is around 0.004%. This characteristic alone makes diode lasers an interesting

option as an excitation source for indirect fluorometric detection.

Introduction to Diode Lasers

The first diode laser was demonstrated in 1962,81 but continuous wave (cw)

operation at room temperature was not reported until 8 years later." Diode lasers offer

several advantages over conventional lasers. They are compact, efficient and cheap. A

diode laser, complete with heat sink, power stabilizer and protective window can be as

small as 50 mm2, yet deliver up to 5000 mW of power. The efficiency of the diode laser


is typically 30%, as compared to 2% for the He-Ne laser. The cost of a diode laser

system can range from $100 to $10,000, depending on the stability, power and

wavelength desired. The lifetime of the diode laser is given at 106 hours. This is in

contrast to a He-Cd laser, whose plasma tube lifetime is approximately 4000 hours, and

costs over half of the original price of the system to replace. Diode lasers are also

extremely rugged. Since there are no plasma tubes or other fragile optical components,

the diode laser would be an ideal candidate for field instrumentation.

Diode lasers are solid state electronic emitters of electromagnetic radiation." The

are similar to light-emitting diodes (LEDs) except that the diode laser possesses a cavity

to induce lasing. They are commonly made out of gallium doped arsenide, which has

an energy gap that corresponds to wavelengths in the red to near-IR region of the

spectrum. The one major drawback of the diode laser is the wavelength availability.

Currently, they are only available at a few wavelengths between 670 nm and 3000 nm.

They are tunable (over a 20-30 nm range) but normally it is not possible to have

continuous tunability over this wavelength range, because of the tendency of diode lasers

to mode hop, or instantaneously jump from one wavelength to another.

One major drawback of the diode laser is the wavelength availability. Working

in the red and near-IR region of the spectrum decreases the background fluorescence due

to contaminants and impurities in a sample; however, it also limits the number of

compounds which can be detected by LIF. Recently, there has been research involving

the labelling of compounds with red and near-IR absorbing fluorophores. Several

different groups of chemicals possess the qualities necessary for this type of analysis.


The first group consists of thiazines and oxazines. These compounds contain amine

groups which are useful for labelling purposes. Representatives of this group include

Nile Blue, Methylene Blue and Oxazine 750. Another group consists of a class of

compounds called phycobiliproteins. They show high molar absorptivities and good

quantum yields, however, they are very large molecules, and therefore their applications

are limited. A third group of compounds consists of the cyanine dyes. These compounds

have high molar absorptivities and are strongly fluorescent, and for these reasons they

are commonly used as laser dyes and as fluorescent bioprobes." A disadvantage to the

cyanine dyes is that their fluorescence and solubility is rapidly decreased in aqueous

solvents.85 Nevertheless, cyanine dyes have been used as fluorescent labels in several

studies.8'9 Table 2-1 gives a list of some of the above mentioned dyes, along with some

of their characteristics.

Experimental Section

Electrophoresis system. A schematic of the instrumental setup is shown in Figure

2-2. The CZE system was built in the laboratory. It consists of a Bertan Model 205B

high voltage power supply (Hicksville,NY 11801), with a range of -50 to +50 kV. A

plexiglass box, with a safety interlock, was designed to house the vials of buffer in which

the capillary ends are immersed. Platinum wire was used for the electrodes, which are

also immersed in the buffer solutions. The fused silica capillary tubing was obtained

through Polymicro Technologies (Phoenix,AZ 85017). Due to the configurations of the

entire system, a capillary length of 100 cm was used throughout. Two different capillary

Table 2-1: Fluorescent near-IR dyes for possible use with diode laser systems.

Formula (nm) (nm) M-cm-1


Methylene blue

Nile blue

Oxazine 750






















































































I__ I I

Values in water from reference 89.
b Values in methanol from reference 85.
SValues in dimethylsulfoxide from reference 91.
d Values in methanol from reference 92.
' Values in methanol from reference 93.
















* 02



............... ..
........................ .
................. ..

. . . . .
. .

. . . .
. . . .
...................... ...........
..................... ..
....................... ..
....................... ..

..................... .


inner diameters, 50 l/m and 75 /m, were investigated during the experimentation. A

manual injection mode was utilized throughout. This was accomplished by placing the

injection end of the capillary into a vial containing the sample solution, raising it up to

a predetermined distance (normally around 10 cm), for a set amount of time (ranging

from 60 s to 5 s).

Excitation system. The excitation was accomplished by the use of diode lasers.

The first diode laser used was a Spectra Diode Labs SDL-2422-H1 diode laser controlled

by a Spectra Diodes SDL 800 laser driver with temperature control (San Jose,CA

95134). This laser emitted at 796 nm, with an output power of 200 mW being utilized.

Problems arose with its stability, so a second laser was tried. This laser was a

Mitsubishi 4402 diode laser (Tokyo,Japan) driven by the Spectra Diodes SDL 800 laser

driver, with an ILX LDT-5910 thermoelectric temperature controller. This laser emitted

at 780 nm and had a power output of 3 mW at an operating current of 49 mA. The laser

beam was focused onto the section of capillary which was being used at the detector

window. The beam diameter was approximately 150 pm.

Detection system. On-line detection was performed by removing a small portion

of the polyimide coating from the fused silica capillary. Two different techniques were

evaluated for the removal of this coating. The first, and easiest, was to burn the coating

off by heating a portion of the capillary with a match (or lighter). The burnt coating was

then removed by gently cleaning the capillary with methanol. While this is an effective

method, care must be taken, since the heating processes leaves the uncoated capillary

weak and brittle. The second method is to dip the portion of the capillary into hot


sulfuric acid. This too is an effective method, and while it does not cause the capillary

to become weak and brittle, the need to heat concentrated acid is a cause of concern.

The portion of the capillary which serves as the detector window was held on a

microscope stage by a specially modified holder. The fluorescence was collected at 90

by a Nikon 20x microscope objective. Cutoff filters were placed before the detector to

reduce the laser scatter from the capillary. Several filter combinations were evaluated,

with the best combination being a Corion RG 850 high-pass filter (Holloston,MA 01746)

and a Kodak #87C Wratten filter (Rochester,NY 14650). The detector was a Hamamatsu

HC210-3314 red-sensitive photodiode with amplifier (Bridgewater,NJ 08807). The

output was directed to a Keithley 182 Sensitive Voltmeter (Cleveland,OH 44169),

followed by a Fisher Series 500 chart recorder (Pittsburg,PA 15219). The voltmeter

settings included a 16.6 ms integration time, and the digital filtering on medium

response. The output offset was used to compensate for the large background

fluorescence signal, and to allow for close analysis of the background noise and the

analyte signals. The entire microscope, including the detector, was contained in a black

box. This helped to reduce scatter and external sources of light from reaching the


Chemicals. The analyte solutions were made from their sodium salts. Sodium

arsenite, sodium arsenate, sodium selenite, and sodium selenate were obtained from

Sigma Corporation (St.Louis,MO 63178). The near-IR dye, IR-125, was obtained from

Exciton Corporation (Dayton,OH 45431). The cationic surfactants,

dodecyltrimethylammonium bromide (DTAB), tetradecyltrimethylammonium bromide


(TTAB), and hexadecyltrimethylammonium bromide (CTAB), used as flow modifiers

were also obtained from Sigma Corporation. The phosphate buffer was prepared with

sodium phosphate monobasic and sodium phosphate dibasic from Fisher Scientific

(Pittsburg,PA 15219). The pH was adjusted, when necessary, with a dilute solution of

sodium hydroxide. Bamstead nanopure water (Dubque,IA 52004) was used to prepare

all of the solutions. Methanol and acetonitrile were HPLC grade and also obtained

through Fisher Scientific.

Optimization of Parameters

Probe species. The probe species which was chosen for these experiments was

IR-125. Its formal chemical name is anhydro-l,1-dimethyl-2-[7-[l,1-dimethyl-3-(a4-


benz(e)indolinium hydroxide sodium salt; its Chemical Abstracts Services (CAS) number

is [3599-32-4]. It is commonly known as indocyanine green, or ICG, which is associated

with its use as an indicator. The structure of IR-125 is given in Figure 2-3. Its X, is

at 784 nm, with its X,, at 805 nm. IR-125 was chosen for several reasons. First, IR-

125 is a zwitterionic salt, which has a slight negative charge at a pH above 3.27 (its pK,

in aqueous solutions). Common uses of IR-125 include its use as a laser dye, and as a

clinical indicator dye for testing in vivo blood flows and hepatic functions in animals and

humans." As was mentioned earlier, a displacement based on a charge mechanism is

desired, and since the analytes of interest are anionic in nature, it is necessary for the

probe species to also be anionic. Another reason for choosing IR-125 is its solubility










-o (






characteristics. It is one of the few near-IR dyes which is soluble in aqueous solvents.

Although it is soluble in water, it is not highly fluorescent in this type of solvent. Figure

2-4 shows the results of a study to determine the effects of increasing organic solvent on

the fluorescence signal of IR-125. This study was performed on a SPEX Fluorolog

spectrophotometer system, which utilized an R928 PMT as the detector. It shows that

there is a linear relationship between the percentage of organic solvent present, and the

fluorescence intensity of IR-125. Since organic solvents can be used in electrophoresis

to modify the electroosmotic flow, this could be worked into the electrophoretic

separation parameters.

The first objective was to determine if the CZE system was operating properly,

and to assure that the position of the capillary, an associated optics, were aligned. Using

a buffer solution containing 30% methanol and 1 mM phosphate buffer, injections of IR-

125 were performed. The background signal of the buffer solution was 12.8 mV with

a N, of 0.02 mV, giving a S/N ratio of 642. With this system, a LOD of 4 x 10-18

moles (4 attomoles) was achieved for IR-125. The reproducibility of the migration time

(u,) was less than 3%.

The next step was to evaluate the behavior of the IR-125 as a component in the

buffer system. Solutions containing varying amounts of IR-125, methanol, and phosphate

buffer were prepared and evaluated in the CZE system. Unfortunately, many

reproducibility problems arose. Inexplicable oscillations in the background fluorescence

signals would occur. Substantial drops in the background would also be observed. An

electroosmotic flow was created, and when water was injected at the positive electrode,

4- I


0 0 0 0 0 0
--C\-2 in



0 0 0 0 0 0 0 0 0 0 0

(sdo) (ajirTuo1aov .1oj) AS4suaxruI

an electroosmotic flow marker was observed, as was a dip in the baseline due at the .,

of IR-125. For a buffer solution consisting of 1.29 x 10' M of IR-125, 10% methanol

and 1 mM phosphate buffer, the electroosmotic mobility (j.) was calculated to be 4.6

x 104 cm2/Vs, and the mobility of IR-125 0(z-12) was calculated to be -1.2 x 104

cm2/Vs. For a buffer solution made of 1.29 x 106 M IR-125, 30% methanol and 1mM

phosphate buffer (pH 6.6), the A.o was calculated to be 2.83 x 104 cm2/Vs, and the aIt-s

was calculated to be -1.0 x 104 cm'/Vs. The decrease in the electroosmotic mobility was

due to the higher percentage of methanol present in the buffer solution. Although the

theoretical value for the mobility of IR-125 could not be found, this value does

correspond to the trends found for other ions.

Flow modifiers. To enhance the separations and to decrease the analysis time,

electroosmotic flow modifiers were investigated. The first to be tried was the cationic

surfactant dodecyltrimethylammonium bromide (DTAB). A cationic surfactant was

chosen because, due to its charge, it coats the inner surface of the capillary and changes

it from negative to positive. This type of surface modifier has been incorporated in other

electrophoretic separations for the analysis of anions. The addition of this modifier to

the IR-125 buffer system caused even more baseline drifts and instability. Upon further

investigation, it was found that IR-125 forms ion-pairs with cationic surfactants.9 This

is most likely the cause of the instabilities in the fluorescence signal as well as the

electroosmotic flow. Another surface modifier which has been used is diethylenetriamine

(DETA). The addition of DETA also caused instability in the system.


While the expectations for the use of IR-125 in this CZE system were high, many

problems developed during its progression. One of the most perplexing problems was

the irreproducible, and unpredictable, fluorescence background. Sudden changes in the

fluorescence signal along with long-term drifts and oscillations suggested an unstable

electrophoretic parameters. This could have been due to the low organic content of the

buffer, or some type of interaction with the IR-125 and the capillary walls.

Using a He-Cd Laser as an Excitation Source

The advantages which arise from the use of the cw helium-cadmium (He-Cd) laser

as an excitation source include the many different probe species which can be utilized

with indirect fluorometric detection. Unlike the large dye molecules which must be used

with the diode lasers, smaller, more mobile, probe species can be used. These probes

are typically water soluble, and have mobilities which are closer to those of the small

inorganic anions.

Introduction to the He-Cd Laser

The He-Cd laser involves transitions of cadmium atom excited by collisions with

metastable helium. The cadmium is present as a metal vapor. The He-Cd laser is a cw

(continuous-wave) laser, and is pumped by an electrical discharge. The most intense

wavelengths are 325 nm and 442 nm. The output power is normally around 10 mW for

the 325 nm transition, and around 15 mW for the 442 transition, but can be as high as

5000 mV. The stability of the output power is typically 0.1 %. The lifetime of the

metal vapor tube is about 4000 hours. This does limit its attractiveness as an excitation

source to be used for routine analysis.

Experimental Section

Electrophoresis system. A schematic of the instrumental setup is shown in Figure

2-5. The CZE system is the same as the one used in the previous section of this chapter.

One difference is that the outer diameter of the fused-silica capillary in these experiments

was 360 fm, verses an outer diameter of 150 j/m in the previous study. It was found

that the larger outer diameter of the capillary caused no decrease in the fluorescence

signal, and it was more sturdy than the capillary with the smaller outer diameter.

Excitation system. A Liconix He-Cd laser (Santa Clara,CA 94089) was used as

the excitation source. The laser optics allowed use of the 325 nm line for excitation.

The power output was approximately 5 mW. Several problems arose during the course

of the experimentation, including the need for a new tube, and the need for the internal

optics to be either cleaned, or replaced.

Detection system. On-line detection was performed, as before, by removing a

small portion of the polyimide coating from the fused silica capillary. The portion of the

capillary which serves as the detector window was held in the sample chamber of a

SPEX Fluorolog spectrophotometer by a specially modified holder. The fluorescence

was collected at 90 by a commercial 0.22 m double monochromator. A Corion P10-326





c *


cutoff filter was placed over the entrance portal of the sample chamber, to decrease the

amount of unwanted stray light from entering the sample chamber. A Corion LG-370

long-pass filter, along with a piece of plexiglass, were placed before the emission

monochromator to reduce the laser scatter from the capillary. It was found that the

background signal produced from the capillary alone, and noise of the system could be

drastically reduced by eliminating excess sources of light from reaching the detector.

The emission wavelength varied slightly, depending on the probe species used, but was

always around 420 nm. An R928 water-cooled PMT served as the detector. A SPEX

data station and associated software were used for data collection and analysis.

Chemicals. The analyte solutions were made from the same sodium salts

mentioned previously. The probe species were as follows: sodium salicylate from Sigma

Corporation (St.Louis,MO 63178); sulfosalicylic acid from Kodak (Rochester,NY

14650); 5-aminoisophthalic acid from Aldrich Chemical Company (Milwaukee,WI

53233); and m-hydroxybenzoic acid from Mallincrodt (Chesterfield,MO 63017). The

cationic surfactant flow modifiers were the same as mentioned previously. Barnstead

nanopure water (Dubque,IA 52004) was used to make all of the solutions.

Optimization of Parameters

Probe species. Several different probe species were evaluated with this system.

They are: salicylate; sulfosalicylate; m-hydroxybenzoate; and 5-aminoisophthalate. The

structures of these probe species are shown in Figure 2-6. The characteristics which

were considered when determining which species to investigate included the size, the










O)----l- C

a 0
' U

W l 4 i



I- 4 <
a a

S\O /


|B B
'a 4

i u id
^- t) 4 4




charge, the excitation wavelength, and the fluorescence intensity of the probe. Small

probe species with a negative charge were desired so that the mobility could be similar

to the mobility of the arsenic and selenium oxyanions. Probes with the potential of

possessing a -2 charge can have higher mobilities and should also allow better

displacement of the analytes, since three of the four anions also have a -2 charge. The

last requirement was that the probe absorb at 325 nm, which is the output of the He-Cd

laser. The fluorescence wavelength was not as crucial, as long as it was far enough

away from the excitation wavelength. This allows for the scatter from the laser to be

decreased by the use of filters before the emission monochromator.

Flow modifiers. As with the diode laser based system, flow modifiers were

investigated to enhance the separations and to decrease the analysis time. The first to be

tried was DTAB. This type of flow modifier has been incorporated in other

electrophoretic separations for the analysis of anions. It was shown that the addition of

varying amounts of DTAB did indeed slow down the electroosmotic flow, and eventually

reverse it. The effect of this surfactant on the electroosmotic flow is shown in Figure

2-7. Two other cationic surfactants were also found to reverse the electroosmotic flow.

They are: cetyltrimethylammonium bromide (CTAB); and tetradecyltrimethylammonium

bromide (TTAB). TTAB was found to give the lowest fluorescence noise and the best


PH. The pH of the system is important for a variety of reasons. First, the pH

will determine the form of the analyte which is present, based on its pK, values. The

pH in these experiments ranged from 3 to 12, depending on the desired form of the






- 0


CO l

0 0 0 0 0 0
I I I -+- I I
0 0 0 0 0 0
(SA! uIxo) 4JItlqoJA[ oqtoumsooJabooia

analyte or probe species. The pH also affects the electroosmotic flow of the system.

Figure 2-8 shows the dependence of the electroosmotic flow on the pH. For most of the

separations, a pH of 9.6 was used. This was to assure that all of the analytes were in

an anionic form, so that they might be detected by a charge displacement mechanism in

the indirect fluorometric detection. The other reason for the high pH was to have the

probe species with a -2 charge which aided in the detection and the separation of these



Salicylate probe species. Salicylate was the first probe to be used with this

system. It was chosen as a test probe, since it has been utilized with CZE/IFD

previously."28 Initially, the system was run in "conventional" mode, with a positive

power supply and injection end. With 2.0 mM salicylate at a pH of 9, /A, was 9.1 x 10

cm'/Vs. The pH was then decreased to 3.1, and the u, decreased to 5.1 x 104 cm2/Vs.

Neither of these conditions allowed separation or detection of the arsenic and selenium

oxyanions. During this time, problems with system noise were encountered.

Optimization of the emission monochromator slit widths, as well as a newly modified

capillary holder were performed. It was soon after discovered that the laser was not

functioning properly, and was sent in for repair. Upon its return, varying concentrations

of salicylate were used as the probe species, but no effective detection was achieved.

The next step was to reverse the electroosmotic flow. The first surfactant chosen

was DTAB. With this flow modifier, the electroosmotic flow was indeed reversed, and





o P



o 10 o 10 o 0 ~
I i I n i I I

(SA/!. xro) A'jITqoIy oTlouiosoojxoaig




separation of the oxyanions was attempted. The results, unfortunately, were erratic. The

fluorescence baseline would change irregularly, and the noise would also change without

warning; this occurred with varying concentrations of salicylate (although it was more

common at the lower salicylate concentrations), with a DTAB concentration of 2.0 mM.

The DTAB was kept at 2.0 mM to facilitate reversal of the electroosmotic flow.

After experiencing difficulty with the DTAB, a second surfactant, TTAB, was

tried. A buffer solution containing 2.0 mM salicylate and 0.5 mM TTAB at a pH of 9.7

showed promising results. Figure 2-9 shows an electropherogram in which HAsO42 and

SeO3"2 were separated. The signal-to-noise ratio (DR) for this system was calculated to

be 775. This is in the range of expected values for excitation with a He-Cd laser. The

j0, of this system was found to be -3.5 x 104 cm2/Vs. The individual mobilities were

calculated as the following: SeO42 -6.2 x 104 cm2/Vs; Se03-2 -5.7 x 104 cm2/Vs; HAsO42

-5.3 x 10' cm2/Vs; and AsO2-1 -2.8 x 10" cm2/Vs. In an attempt to lower the detection

limits (which were in the ppm range), a lower concentration of salicylate was evaluated.

This did not, however, help the separation or the detection. It was detrimental to the

detection of the same concentrations of analytes used above, as well as to the signal-to-

noise ratio of the system. The signal-to-noise ratio (DR) dropped to 490, from 775. It

appears that the separation and detection of these four analytes is possible with salicylate

as the probe molecule, but not ideal. Further optimization of the separation parameters

was not performed.

5-aminoisophthalate. The next probe which was tried was 5-aminoisophthalate

(AIP). While this probe did have a -2 charge and it was highly fluorescent, it was not

Figure 2-9: Electropherogram of arsenic/selenium oxyanion mixture with buffer
containing 2 mM salicylate and 0.5 mM TTAB at pH 9.7 at -20 kV.
(A) full scale view; (B) magnified view. Signals are designated as:
(a) 13 ppm SeO32-; (b) 14 ppm HAsO42; (c) salicylate (probe); (d)
possible impurity.


ab c











U .U U 1 0 0 0-0-0 2 0 2 0 2 0 2 0i o
100 120 140 160 180 200 220 240 260 280 300
Time (s)

very water soluble. The most concentrated solution which could be made was 0.25 mM.

With 0.25 mM AIP and 0.5 mM TTAB, a signal-to-noise (DR) of 1220. While the

baseline was very stable before an injection, the baseline became erratic a few minutes

into the run and then it re-stabilized. It is suspected that the amino group on the probe

has a detrimental effect on the indirect fluorometric detection (and possible on the

separation as well). Occasionally sharp dips in response were observed, but the large

drifts in the background signal hampered thorough evaluation of these dips.

Sulfosalicylate. With the difficulties encountered by the use of salicylate and AIP

as probes, sulfosalicylate was evaluated as a probe. Sulfosalicylate (SS) is also excited

by the 325 nm line of the He-Cd laser, and has a fluorescence at 422 nm. Again,

varying concentrations of sulfosalicylate were evaluated (from 1.0 mM to 0.1 mM), with

TTAB as the flow modifier. Figure 2-10 shows a separation of SeO4"2 and SeO32 with

a buffer consisting of 1.0 mM sulfosalicylate and 0.5 mM TTAB at pH 9.7. The

electroosmotic mobility was -2.0 x 10" cm2/Vs. The mobilities of two of the analytes

were as follows: SeO4-2 -7.0 x 10" cm2/Vs; SeO32 -5.8 x 104 cm2/Vs. The signal-to-

noise ratio (DR) for this system was 953, which is in the same range as the other probes

evaluated with the He-Cd laser based system. When a mixture of all four analytes was

injected, the background fluorescence signal became noisy at times, which hampered the

simultaneous separation of the arsenic and selenium oxyanions with this probe. As with

the salicylate, decreasing the concentration of the probe below 1.0 mM caused further

baseline instability, and no improvement in the signals was observed when the anions

were injected.

Figure 2-10: Electropherogram of arsenic/selenium oxyanion mixture with a buffer
containing 1.0 mM sulfosalicylate and 0.5 mM TTAB at pH 9.7 at
-30 kV: (A) full scale view; (B) magnified view. Signals are
designated as: (a) 14 ppm SeO42, (b) 13 ppm SeO32,(c) sulfosalicylate


750- a b
$ 650-
S 600-
I 550
0 350
o 250-
g 200
0 50 100 150 200 250
Time (s)



a 800

3 790

S 780


Time (s)

m-hydroxybenzoate. The last probe which was investigated was m-

hydroxybenzoate. The difference between this probe and the salicylate was simply the

location of the hydroxide group on the benzene ring. As with the others, varying

concentrations of this probe were evaluated. It was found that concentrations lower than

1.0 mM of m-hydroxybenzoate did not allow effective analysis of the arsenic nd selenium

oxyanions. At a buffer composition of 1.0 mM m-hydroxybenzoate and 0.5 mM TTAB

at pH 9.9, analysis of all four anions was indeed possible. Figure 2-11 shows this

separation. The mobilities for the analytes are as follows: SeO4-2 -7.0 x 104 cm2/Vs;

Se3-2 -6.0 x 104 cm2/Vs; HAsO42 -5.7 x 104 cm2/Vs; and AsO;- -3.0 x 104 cm2/Vs.

The p4, of the system was measured to be -1.3 x 104 cm2/Vs. The signal-to-noise ratio

(DR) was 892. Figures 2-12 through 2-16 show the calibration curves for these analytes,

as well as the calibration curve for m-hydroxybenzoate. The limits of detection for this

system are given in Table 2-2 and are based on the area of the dip observed. The linear

dynamic range for this system was around 1 order of magnitude. Part of the reason for

this was the limitations placed on the concentration of the analyte in the sample by the

concentration of the buffer solution. The best resolution should occur when the analyte-

to-buffer ratio is 1:500. When the analyte concentration reached that of the buffer

solution, the injected zone had a higher resistance than the buffer solution, hence a lower

current flowing through it. This caused it to move through the capillary as a plug, rather

than separating into the corresponding analyte zones. The lower end of the concentration

scale was determined by the limit of detection for the system. The transfer ratios for

these analytes were calculated as follows: SeO4-2 0.32; SeO3"2 0.46; HAsO4"2 0.34; and

Figure 2-11:

Electropherogram of arsenic/selenium oxyanion mixture with buffer
containing 1.0 mM M-hydroxybenzoate and 0.5 mM TIAB at pH 9.9
at -20 kV: (A) full scale view; (B) magnified view. Signals are
designated as: (a) 14 ppm SeO42; (b) 13 ppm SeO32; (c) 14 ppm
HAsO4"; (d) 11 ppm AsO2'; (e) m-hydroxybenzoate (probe); (f)

f b


100 200 300
Ime (s)

400 500


, 1.00-

0.90 -
; o.8o-

g 0.60 -
j 0.50-
E 0.30-

II irI i z






0 00

. II
O as

'2m 5


> I



0 rr



\DO O oOC 6
\0 00 d





r +





c M

I -I

O O O 0 C
(spuisnoi, L)

( .ajv) di(I


-4 C


IN 0
*- el
N I>


bO E

0 0 0 0 o
Ln M wO (

(e0ai) dITQ



C= C-)




C1 00


x W6
t i


bO r



(1r3a ) dlQ






4c .


r .0





\ -to


\ I3 IP

ir3O LO O L(O O cOi (?j ^ -
L(sCpC N s noTt ,

Table 2-2: Limits of detection using the CZE/He-Cd laser based system
with m-hydroxybenzoate as the probe species.


SeO4-2 4.5 0.64 12
SeO3-2 3.0 0.38 8
HAsO4-2 24.0 3.4 63
AsO21 4.2 0.45 10

a concentration limit of detection in jsmol/L.
b concentration limit of detection in ppm.
I absolute limit of detection in fmol.

AsO2,- 0.06. Similar values for the three anions with -2 charges suggested that a charge

displacement of the -2 charged probe species was occurring. The arsenite anion had a

lower transfer ratio, since the displacement of the probe by the anion was not as strong.


These experimental results have shown that the analysis of arsenic and selenium

oxyanions can be accomplished using CZE with indirect fluorometric detection. With

m-hydroxybenzoate as the probe species, separation of these analytes was possible in less

than 10 minutes. The detection limits for this system were comparable to those obtained

by indirect UV-vis absorption detection. One of the problems which prevented lower

detection limits from being achieved had to do with the collection optics. The SPEX

system was designed to collect the fluorescence from a cuvette which is placed in the


sample chamber. If the capillary was not placed in the exact location necessary for

optimum fluorescence collection, a loss of sensitivity was observed. While this

optimization was carefully performed with the fluorescent probe in the capillary, the

different positions of the capillary also changed the "blank" signal, ie. the signal

observed when no fluorescent probe was present. This made optimization more

complicated. Another possible source of limited sensitivity may have arisen from the

He-Cd laser noise. Addition of a laser power stabilizer could enhance the detection

limits by increasing the signal-to-noise ratio of the system.


Theoretical Aspects of Ion Chromatography

Ion chromatography (IC) is a form of high performance liquid chromatography

(HPLC) which is used for the separation of ionic compounds. This separation technique

combines chromatography with ion-exchange. Determination of exactly when the two

techniques were combined are debatable, but an important landmark was the ion-

exchange work of Adams and Holmes in 1935.95 From that time until the early 1970s,

the theory of ion-exchange chromatography, as well as its applications, were investigated.

With the development of automated detectors and advances in separation, the use of ion-

exchange chromatography blossomed. The types of analytes which were being separated

and detected included: amino acids, rare earth elements, proteins, dyes, pharmaceuticals,

synthetic polymers, and some inorganic ions. At this point in time, the most common

detector was the UV absorption detector.95 In the mid 1970s, Dow Chemical Company

granted a license to Dionex Corporation to manufacture and market instrumentation

which utilized suppressed conductivity detection for ion exchange chromatography. They

called this technique "ion chromatography", and applied it to the analysis of many

inorganic and organic ions which were undetectable by UV absorption or refractive index

detectors.95 Today, ion chromatography is applied to a wide variety of ionic species in


all types of matrices. The detectors which can be used with IC are also expanding, and

include: UV absorbance," indirect UV absorption,1221,22297-'00 and indirect



A basic IC system consists of a pump, and injection valve, a separator column,

and a detector. The main difference between HPLC and IC is the stationary phase in the

separation column. The mobile phase (eluent) of the system serves only to transport the

analytes through the chromatographic system. The separation occurs due to the different

affinities of the analytes for the stationary phase, which is an ion-exchanger.

The rate at which an analyte moves through the chromatographic system is

determined by its distribution constant, K, between the two phases:

K= Cs (3-1)

where Cs is the concentration of solute (analyte) in the stationary phase (mol/L), and CM

is the concentration of solute in the mobile phase (mol\L). The capacity factor, k, is

given by the ratio of the number of molecules in the stationary phase, ns, to those in the

mobile phase ,nM:

n. CV Vs
k n Cs = K (3-2)

where Vs is the volume of the stationary phase (L), and VM is the volume of the mobile

phase (L).

The volume of eluent which is required to remove all of the analyte from the

column is called the total retention volume, VR. The total free volume of the system,

including the injection loop, the column, and all connection tubing is called the "dead"

volume and is determined by the retention time of an unabsorbed analyte. The actual

retention volume of the analyte is given by:

Vi = V,-V. (3-3)

where V'R is the corrected retention volume (L), and Vo is the dead volume (L). Since

chromatograms are typically recorded as a function of time, it is the retention time, tR,

which is determined. From the tR (s) and the flow rate, F (L/s), the retention volume

(L) can be determined by the relationship:

VR = tF (3-4)

The capacity factor can be related to the retention time or the retention volume by the


k V = tR (3-5)
Vo to
The separation of two analytes is determined by their distribution constants, and

can be defined as a separation factor, a:

k2 (3-6)
ki KK

The efficiency of the separation is defined by the number of theoretical plates, which can

be defined as:

1 ~2
L = 5.54 t R (3-7)
h Wh

where L is the length of the column (cm), h is the height equivalent to a theoretical plate

(HETP) (cm), and wh is the width of the peak at half of its height (s). The resolution

of the chromatographic process is defined as:

Rs 1 (3-8)

where wb is the width of the peak at the base (s). The resolution of a separation can be

related to the capacity factor, and the separation factor by the equation:

R 1 ( 1) k (3-9)
Rs -4 1+kr

This equation is a fundamental relationship in chromatography and clearly indicates how

N, a, and k can affect the separation.

As was mentioned, the stationary phase in IC is an ion exchanger. The success

of IC as a separation technique depends on the relative affinities of the exchangeable

(analyte) ions toward the fixed (stationary phase) ions, and the co-ion solutee). The

theory of this affinity is not completely understood. It does, however, depend on the

electrostatic field strength around the fixed ions and the radius of the spherical (analyte)

ions. Solvation of the ions by the solvent also exerts considerable influence on the

electrostatic interactions. In general, the affinity of the ions for the stationary phase is


inversely proportional to the radius of the hydrated ion and directly proportional to the

ionic charge."'2

Ion Exchangers

The key to an IC separation is the stationary phase. This determines the

separation mechanism, and the separating ability of the entire system. There are 3

important elements which make up an ion exchange stationary phase. They are: an

insoluble matrix, which can be organic or inorganic; the fixed ionic sites, which are

attached to or a part of the matrix; and the equivalent number of ions of charge opposite

to that of the fixed sites. The attached groups are often called functional groups, and the

associated ions are called counterions. Other desired characteristics for an IC stationary

phase include the ability to exchange ions rapidly, chemical stability over a large pH

range, mechanical stability, and resistance to deformation during packing and when

subjected to the flow of the mobile phase."

Modern IC ion exchangers use 2 major matrices: silica and organic polymers

based on styrene. The most common way to produce an ion exchanger is to make a

styrene-based polymer and chemically modify the polymer to introduce the functional

groups. One type of organic polymer matrix is the gel-type polymer. The degree of

swelling depends on the amount of divinyl benzene (DVB) present. The DVB forms

crosslinking in the polymer. The degree of crosslinking is given as mole percent DVB.

These types of resins will readily admit small ions and molecules, but resist the intrusion

of larger species. Lower crosslinking improves the diffusion through the resin. Practical


limits on crosslinking are usually in the range of 4% to 12%.95 The most common

functional group for anion analysis is the quaternary amine group. A typical procedure

for low-capacity ion exchangers is the surface agglomeration method. With this method

the surface sulfonated styrene-DVB copolymer particles are contacted with colloidal anion

exchange particles. Resins produced by this method are called pellicular resins.

Pellicular resins have the advantages of an increased resistance to higher pressures,

stability in broader pH range, and stability in the presence of organic solvents. They

also produce a reduction of the diffusion path, which accelerates exchange of the eluent

and analyte ions which increases the separation efficiency.


A particular eluent is considered to be the optimum choice if it gives rapid,

selective and sensitive determination of the analyte ions. The first consideration is the

composition of the eluent. As a general rule, the retention properties of the eluent and

the analyte ions should be similar. For example, for the determination of weakly

retained analyte ions, a weakly retained eluent should be used. For the analysis of a

mixture of weakly and strongly retained ions, it is sometimes possible to use one eluent

to achieve an adequate separation. If not, the other option is to use gradient elution to

change the strength of the eluent so as to separate the analytes, and also decrease the

analysis time. Many different types of eluents have been incorporated into IC

separations, including: hydroxide, carbonate, bicarbonate, aminoacetate, glycinate,

benzoate, tartrate, oxalate, salicylate, aminosalicylate, and phthalate.101

The concentration of the eluent can be varied and optimized for a particular

separation. Variation of the concentration of the eluent changes the ion-exchange

equilibrium. Increasing the eluent concentration causes an increase in the eluting power

of the eluent. The pH of the system can also be used to control the selectivity of the

separation.An increase in pH can shift the equilibrium of the eluent species to produce

an anion with a larger negative charge and an increase in eluting power. Organic

additives can be added to the eluent to improve the separation of some ions. The

addition of acetonitrile tends to increase the eluting power and therefore decrease the

analysis time. Methanol has also been used as an organic additive.'o

An ion chromatogram may have two extra peaks. The first of these is called the

water, or injection peak, which is a result of the passage of the unretained zone of water,

containing the eluting anion in an amount equivalent to the anionic composition of the

sample. 3 The presence of this peak can hamper the analysis of anions which are very

weakly retained. The "dead volume" of the system can be determined by the retention

time of this peak. In a dual-column system (one which has a suppressor column as well

as a separation column), this peak can be reduced by using demineralized water, or

eluents which give a low signal at the detector. In a single-column system, there is no

effective way to eliminate this peak. The second "extra" peak is called the system peak.

It occurs due to the desorption of the molecular form of the eluent from the separator

column on sample injection. The system peak usually occurs at long retention times, and

can interfere with the analysis of some analytes. The time, height and area of this peak

depends on the pH, the concentration, and the volume of the injected sample. The most

efficient way to eliminate the second extra peak is to use eluents at high pH values so as

to not have any molecular form of the eluent present.'03

Detection Methods

A wide variety of detection methods have been used with IC. Because of the

desire to achieve lower detection limits, the evaluation of new detection methods for IC

continues. Detectors for IC must be: highly sensitive, have a small cell volume, have

a highly stable baseline signal, and the signal should remain stable when the flow-rate

of the eluent is changed.

Conductometric detection. Conductometric detection for IC sounds like an ideal

situation, since each ion will produce a signal at the detector. However, this detection

technique was hampered by the high ionic strength of the eluent itself. It was not until

the early 1970s when ion suppression was developed, that conductometric detection for

IC became a viable detection method.'4 In this case, the eluent is converted into a low-

conductivity compound, thus decreasing the background signal. Dionex Corporation has

developed a membrane suppressor (AMMS) which allows very efficient ion suppression,

allowing even gradient elution to be used with their system. Single-column IC with

conductivity detection has been evaluated by the use of organic acids as eluents.'15,' It

is, however, the suppressed IC technique which has allowed IC to become a more

common analytical technique for the analysis of ions.

Spectroscopic detection. The first spectroscopic detectors used for IC were UV-

visible absorption detectors. This can be used for the determination of organic ions,

especially those of aromatic compounds which absorb strongly in the UV region of the

spectrum. Direct UV absorption can be a more sensitive and selective detection method

for many ions." For those analytes which do not absorb, one option is to perform a

post-column reaction which will produce a species which will produce a signal at the

detector. This type of analysis is not all that common, but has been used to determine

rare earth metals and transition metals."3

Indirect detection methods. Indirect photometric chromatography was first

proposed by Small and Miller in 1982. 2 Because of its sensitivity and universal nature,

indirect photometric detection (IPD) has been utilized for those ions for which

conductivity or direct UV absorption could not be used.21'22,97" One reason IPD has

grown so rapidly is that many of the eluents already being used for IC with conductivity

detection, could be used with this detector. Many times, conductivity and IPD are used

in conjunction with each other to provide complete analysis of a sample. The technique

of indirect fluorometric detection (IFD) has also been used with IC.37,38s,~0 Since

fluorescence is inherently a more sensitive technique than absorption, IFD can extend this

to species which do not themselves fluoresce. This technique has been used to detect

common inorganic anions at the picogram level.10'

Using a Diode Laser as an Excitation Source

As was the case for the CZE separations, the use of a diode laser for excitation in

IC was evaluated.

Experimental Section

Chromatography system. A schematic of the instrumental setup is shown in

Figure 3-1. It consisted of a Dionex Series 4500 Chromatography system. This system

utilized an eluent degas module in conjunction with a gradient pump module which

allowed mixture of 4 different eluents, as well as gradient elution programs. A 50 tL

injection loop was used throughout. An AG-11 guard column preceded and AS-11 anion

separator column. A Dionex Advanced Computer Interface was used to interface the IC

system with a computer for data collection and analysis.

Excitation system. A Mitsubishi diode laser driven by a Spectra Diodes SDL 800

laser driver and an ILX LDT-5910 thermoelectric temperature controller was used as the

excitation source. The excitation wavelength of the laser was 780 nm at 20( C and 49

mA of current. The output power was 3 mW.

Detection system. A CDM-II conductivity detector, which was part of the Dionex

IC system (Sunnyvale,CA 94088) was used as a comparison to the indirect fluorometric

detector which is described below. An AMMS-II membrane suppressor was used when

suppressed conductivity detection was desired. The diode laser based system followed

the conductivity detector. A flow cell was made from a section of Polymicro

Technologies (Phoenix,AZ 85017) fused silica capillary about 8 cm long. The outer

diameter of the capillary was 360 u/m, and the inner diameter was 250 Jim. The

calculated volume excited by the laser was 10 nL. This flow cell was held on a

microscope stage by a specially designed holder. The fluorescence was collected at 90'

by a Nikon 20x microscope objective. Two cutoff filers were placed before the detector







............. ..........

\^^^ ^^ 9 ^ *' '*'**'*''*' ~ ~ .. ...... .'.*.* ."" :
^ ~ ~ ~ ~ ~ ~ ~ ... .....-i*:::::::: \1^:^ ::*::