3., . E.
239a.” . q ‘ I 1‘
II: .
x: 2... JP: .05....9aw. «02:51..
a "P Kin-ﬂ
w.

15.1.35.
3.1.55.2
t.
.
I

: :55 . ....

3.. a s 5.. $43.72.: 0!: .43..

v. .y... a, L.
y

Y . If. ‘ . 3. 1‘ 3! ~
, V ,1 ~33.“ {g
.. \f-zv ‘ . , . 1 . .
a. , ,. , . fa.....,,i . V. .,. .
. 7v, )‘ .‘lcf« .1 0-.
, , t ‘ B. . av.
, , , ‘73.. hr .. L...
r. :..:.I
2. L $.53... it
‘31.... (.12. . «I.
.=%.>n...nu.. z. 3.;
x: i113:
t

. i2
. -H I.

It
. Ii; . x

t 5-2333..-’

, I... {xiii p I

3...“:
a.“ 2;: .m
. w
.s. has“. .0":- d .-

.99.“ B

:lrxu.

. ,
4....
a

3 :5; A.-
ai I...

DA
is} :
I‘VE...“
ESQ? . .0.

Mun,

Ix .v. . r... 55‘! 1.
11.5.. .1? E 9?.
1.. i .51....“ . viz.
. . :v. p.
. 1% .6......;i|.i..§|
35“»...tndw8‘
$.59L1: ,
(4.1. .

. 1",}? .l. . .v
til}! (1,).

3" .. 3.. 1b
3.....(i: 2341.5. 5
ovirlltzilskln:
b‘llrht. it. VYH
ﬂ

, Arc-$.19.
2a 5 lvr‘i".
, VnrQltF.‘ .t

06131 '6‘... it.
.3331;

 

l'HESlS

4..
fOOCN

 

LIBRARY

Michigan State
Unlverslty

 

 

 

This is to certify that the

dissertation titled ’

A Mmitvm‘bed “Mm? MIA/RMQLQWCQ Dom/tor
owd om Automatic Cam‘vlﬁ Imtvodubh‘on 9L Sktw
m (aftllowv $0M But/haphoregtr

presented by
CH Uh) H 0 N6 f6 NE.

has been accepted towards fulﬁllment
of the requirements for

_El(l - D degree in ( hi Mkltgrnj

WKM

ajor professor

 

Dateﬂdg 5; ZWJ

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0- 12771

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m mus-p.14

A MINIATURIZED CHEMILUMINESCENCE DETECTOR
AND AN AUTOMATIC SAMPLE INJECTION SYSTEM IN
CAPILLARY ELECTROPHORESIS

By

Chunhong Pen g

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1999

ABSTRACT
A Miniaturized Chemiluminescence Detector and an Automatic Sample Introduction
System in Capillary Zone Electrophoresis
by

Chunhong Peng

A miniaturized, high collection-efﬁciency, post-column, sheath-ﬂow
chemiluminescence (CL) detector for use with capillary zone electrophoresis (CZE) is
described. The stainless steel detector body signiﬁcantly cuts down the electrical noise
associated with AC power supply. The inner wall of this CL detector was carefully
machined to a high smoothness to increase retro-reﬂection. The other advantage of this
novel CL detector arises because the PMT is placed extremely close to the separation
capillary to increase the solid angle for light collection. This simple detector requires no
lens or concave reflection mirror for enhancement of collection efﬁciency. Calculations
predict that the collection efﬁciency of this chemiluminescence detector is z 58%.

The miniaturized CE-CL detector is characterized using transition metal cations as
the sample species. The major problem for the separation of transition metal cations is
their strong adsorption to the surface of the fused silica capillary. We are able to achieve
a detection limit of 1.5 x 10'” M for Coz“, 5.0 x 10‘8 M for Cr“, 3.0 x 10'8 M for Cu“,
and a theoretical plate number of 340,000 for 1.0 x 10'11 M Co“. This is by far the best
separation efﬁciency that has been achieved with post-column reaction detector in

capillary zone electrophoresis.

An automatic sample injection device is also described. This novel sample injector
requires neither the relocation of separation capillary nor the disconnection of the high
voltage during sample injection. Sample throughput is greatly enhanced compared to that
of traditional sample injection. This automatic sample injection system includes a
mechanical system, an electronic control circuit and a computer. The whole system is
mounted on a 10” x 18” x ‘/2” home-made aluminum optical breadboard.

In this sample introduction system, the separation capillary is ﬁxed on a capillary
holder, which is mounted on a optical translation stage. The sample capillary is held on
the sample platform mounted on a rotary alignment system. A computer controls the
electromagnet to pull the platform to a precise position. The rotary alignment system
assures the precise alignment of sample capillary and separation capillary during sample
injection. A damping device is used to minimized the shock when the sample capillary
joins the separation capillary. The new injection system is characterized in chapter 5. The
automatic injection system shows a better reproducibility of injection (10.9% RSD) than

the manual injection system (27.0% RSD).

Acknowledgements

I would like to thank all the people who were helpful in the preparation of this thesis.
No one can do everything by himself although I am the sole author of this dissertation.

I would like to express the greatest gratitude to my adviser, Dr. Stanley R. Crouch.
Stan is not only an adviser, but also a friend who understands his students. I learned many
things from the course he taught and from regular conversations with him. He is not
someone who tells you what to do, but someone who makes you to think what you should
do. Special thanks go to Dr. Gary Blanchard for serving as my second reader and for
providing suggestions during the writing of this dissertation. I would also like to thank
the rest of the committee members, Dr. Dantus and Dr. Hollingsworth for their support.

My special thanks go to Russell Geyer and Richard Menke in the chemistry machine
shop for teaching me various machining skills. Also thanks to Ron Haas, Scott, Glen in
the Electronic shops for their excellent service whenever help is needed.

Finally I would also want to thank some former group members, David Short, Dana
Spence, Yi Shi, Tom Cullen and Dan Severs (the Karate man) for the help during my

Ph.D journey at Michigan State University.

TABLE OF CONTENTS

CHAPTER ' PAGE
LIST OF TABLES .................................................................................... 1
LIST OF FIGURES ................................................................................. v

1. INTRODUCTION TO CHEMILUMINESCENCE AND

CAPILLARY ZONE ELECTROPHOREIS .................................................. 1
1.1 Introduction to Chemiluminescence (CL) ............................................ 1
1.1.1 Theoretical Aspects of Chemiluminescence Reaction ...................... 1

1.1.2 Chemiluminescence Reagents ................................................... 3

1.1.3 Historic Background of Chemiluminescence Detection 5

1.2 Introduction to Capillary Zone Electrophoresis ....................................... 6
1.2.1 Historical Background of Electrophoresis .................................... 6
1.2.2 Advantages of Capillary Zone Electrophoresis .............................. 7
1.2.3 Electroosmosis ................................................................... 7
1.2.4 Theoretical Aspects of Capillary Zone Electrophoresis .................... 9

1.2.4.1 Migration Time ........................................................ 9
1.2.4.2 Separation Efﬁciency ................................................. 9
1.2.4.3 Resolution ............................................................. 10
1.2.5 Sample Injection ................................................................ 10
1.2.5.1 Electrokinetic Injection .............................................. 10
1.2.5.2 Hydrodynamic Injection ............................................ 11
1.2.6 Detection in Capillary Zone Electrophoresis ................................. 12
List of References ..................................................................... 18

2. CONSTRUCTION OF MINIATURIZED CHEMILUMINESCENCE

DETECTOR IN CAPILLARY ZONE ELECTROPHORESIS .......................... 26
2.1 Design Considerations for Chemiluminescence Detector ............................ 27
2.2 Construction of Miniaturized CE-CL Detector ........................................ 28

2.2.1 Outer Body of the CL Detector ............................................... 28

2.2.2 Etching of the Separation Capillary .......................................... 28

2.2.3 Reaction Capillary ............................................................. 32
2.2.4 Installation of the Miniaturized Chemiluminescence Detector ........... 32
2.2.4.1 Installation of Chemiluminescence Flow Cell ....................... 32
2.2.4.2 Installation of CL Flow Cell to CL Outer Body ..................... 33
2.3 Instrumentation for Chemiluminescence Detection in Capillary

Electrophoresis ............................................................................ 34
2.4 Electrical Noise ............................................................................ 35

2.5 Noise Level of the Miniaturized CL Detector and the Conventional
CL Detector ................................................................................. 37
2.6 Factors Affecting the Collection Efﬁciency ............................................ 39
2.7 Collection Efﬁciency of the Miniaturized CL Detector .............................. 40
2.8 Conclusion .................................................................................. 43
List of References .............................................................................. 44

. Detection of Transition Metal Ions Using Miniaturized Chemiluminescence

Detector in Capillary Zone Electrophoresis ................................................ 46

3.1 Preparation of Stock Solution ............................................................ 46

3.1.1 Preparation of 0.01 M Luminol Stock Solution (pH = 12.00) ......... 47

3.1.2 Preparation of 0.05 M H202 Solution ........................................ 47

3.1.3 Preparation of Separation Buffer ............................................. 47

3.1.4 Preparation of Sample Stock Solution ....................................... 48

3.2 Instrumentation ............................................................................ 48

3.2.1 Instrumentation for Lifetime Measurement ................................. 48

3.2.2 Instrumentation for Electrophoretic Separation ........................... 50

3.3 Procedures .................................................................................. 50

3.3.1 Capillary Conditioning ........................................................ 50

3.3.2 Sample Injection ............................................................... 53

3.4 Results and Discussion ................................................................... 53
3.4.1 Lifetime and Intensity of Luminol/ H202 Chemiluminescence

at Various pH Values .......................................................... 53

3.4.2 Selection of Mixing Mode .................................................... 56

3.4.2.1 Mixing Mode 1: Mix H202 in Separation Buffer 56
3.4.2.2 Mixing Mode 2: Mix Luminol with H202 as
Post-Column CL Reagent ....................................................... 57

vi

3.4.2.3 Mixing Mode 3: Luminol in the Separation Buffer 58

3.4.3 Optimization of pH for Post-Column Chemiluminescence

Detection ....................................................................... 60
3.4.4 Optimization of pH for the Separation Buffer .............................. 66
3.4.5 Signal Enhancement by Complexation ....................................... 69
3.4.6 Effect of pH on Sensitivity for Buffer with HIBA ......................... 79
3.4.7 Reproducibility of Sample Injection ......................................... 84
3.4.8 Number of Theoretical Plates with Post-Column
Chemiluminescence Detection ................................................. 85
3.4.9 Calibration for Transition Metal Cations .................................... 87
3.4.10 Conclusions .................................................................... 91
List of References .............................................................................. 92

. CONSTRUCTION OF AUTOMATIC SAMPLE INJECTION SYSTEM

WITHOUT THE INTERRUPTION OF HIGH ELECTRIC FIELD IN

CAPILLARY ZONE ELECTROPHORESIS ............................................. 94
4.1 Instrumentation of Automatic Sample Injector ....................................... 95
4.2 Electronic Circuit ........................................................................... 97
4.3 Damping Device ........................................................................... 99
4.4 Rotary Alignment System ............................................................... 100
4.5 Holder for the Separation Capillary .................................................... 102
4.6 Solenoid Support .......................................................................... 104
4.7 Capillary Polishing Tools ............................................................... 105
4.8 Tubing Connector and Pinch Valve ................................................... 106
. CHARACTERIZATION OF THE AUTOMATIC SAMPLE INJECTION
SYSTEM IN CAPH.LARY ZONE ELECTROPHORESIS ........................... 108
5.1 Reproducibility of Timing .............................................................. 109
5.2 Reproducibility of Peak Height ........................................................ 110
5.3 Function of Pinch Valve ................................................................ 113

5.4 Comparison between the Automatic Injection System and Other
Injection Systems ......................................................................... 114

vii

5.5 Conclusions ............................................................................... 115

List of References .............................................................................. 115

6. Summary and Conlusions .................................................................... 116
7. FUTURE WORK ............................................................................. 117
7.1 The Chemiluminescence Detection System .......................................... 117
7.2 Automatic Injection System in Capillary electrophoresis .......................... 119
7.2.1 Linear Motion Device for Alignment ...................................... 119

7.2.2 Timing Device ................................................................ 120

List of References ................................................................................. 121

viii

LIST OF TABLES

TABLE PAGE
3.1 Solubility of luminol in buffers of different pH values ................................... 58
3.2 Stability of the buffers containing various amount of HIBA at pH 3.80 ................ 70
3.3 Reproducibility of 10 electrokinetic injections for Co2+ ................................... 85
3.4 Reproducibility of 10 electrokinetic injections for Cr3+ .................................... 85
3.5 Detection limits for some transition metals using CL and other detectors ............. 87

5.1 Reproducibility of the injection system without sample injection ...................... 110
5.2 Reproducibility of 10 manual injections for 4 3 injection ............................... 114

LIST OF FIGURES

FIGURE PAGE
1.1 Schematic representation of capillary electrophoresis .................................... 6
1.2 Schematic representation of silica-solution interface .................................... 8
2.1 Design considerations for chemiluminescence detector ................................ 27
2.2a Front view: miniaturized post-column chemiluminescence detector in

capillary electrophoresis ..................................................................... 29
2.2b Side view: miniaturized post-column chemiluminescence detector in

capillary electrophoresis ..................................................................... 30
2.2c Detailed view of the ﬂow cell in miniaturized CL detector in CB ..................... 31
2.3 Instrumental set—up for the miniaturized post—column CE-CL detection .............33
2.4 Electrical noise of the miniaturized chemiluminescence detector made of

2.5

2.6

3.1

3.2a

3.2b

black Delrin without proper grounding. PMT power suppy voltage: 800 V.
ampliﬁer: 108 WA; sampling rate: 1000 points/s ....................................... 36

Comparison of noise level between the miniaturized CL (lower trace) and
conventional CL detector (upper trace) in capillary electrophoresis. PMT
power supply: 800 V; ampliﬁer: 108 WA; sampling rate: 1000 points/s ........39

Three dimensional diagram for the calculation of solid angle. a is the
length of PMT window, 25 mm; b is the width of PMT window, 12 mm;
c is the distance from stray screen to the reaction capillary, 4 mm ................... 41

Instrument for lifetime measurement of luminol/H202 chemiluminescence
ampliﬁer: 108 V/A; response time: 1 ms; sampling rate: 1000 points/s .............. 49

Electropherogram of the ﬁrst run of 1.0 x 10’8 M Co”, 1.0 x 1045 M Cr“,

1.0 x 10'6 M Cu“.

Sample: 1.0 x 10'8 M Co“, 1.0 x 1045 M Cr“, 1.0 x 10'6 M Cu2+

in deionized distilled water; buffer: 2.0 x 10'4 M luminol, 0.01 M

NaAc/HAc, 0.005 M HIBA, pH = 3.53; post-column CL reagent:

0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate:

6 uIJmin; electrophoretic voltage: 10 kV; injection voltage: 10 kV;

injection time: 10 s; fused silica capillary: 50 um x 363 um o.d. x 50 cm ............ 51

Electropherogram of the ninth run of 1.0 x 10'8 M Co”, 1.0 x 10‘6 M Cr“,
1.0 x 10'6 M Cu“.

3.3

3.4

3.5

3.6

3.7

3.8

3.9

Sample: 1.0 x 10‘8 M Co“, 1.0 x 10’6 M Cr”, 1.0 x 10*5 M Cu2+

in deionized distilled water; buffer: 2.0 x 10'4 M luminol, 0.01 M
NaAc/HAC, 0.005 M HIBA, pH = 3.53; post—column CL reagent:

0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate:

6 uUmin; electrophoretic voltage: 10 kV; injection voltage:

10 kV; injection time: 10 s; fused silica capillary:

50 um x 363 um o.d. x 50 cm ........................................................

Lifetime of luminol-H202 chemiluminescence at various pH values
PMT power supply: 600 V; Ampliﬁer: 108 WA; sampling rate: 1000
points/s; A: pH = 12.50; B: pH = 13.00; C: pH = 12.00; D: pH = 11.50;

...... 52

E: pH = 11.00; F: pH = 10.50 ............................................................... 54

Intensity of luminol-H202 chemiluminescence at various pH values
PMT power supply: 600 V; Ampliﬁer: 108 V/A; sampling rate: 1000 points/s
Reagent A: 200 uL 0.025 M H202 at various pH ---> Reagent B: 20 1.1L of

3.0 x 10'4 M luminol, 0.01 M NaAc/HAc, pH: 4.65 ..................................... 55

Mixing mode 1 for chemiluminescence detection in capillary electrophoresis
Mixing mode 2 for chemiluminescence detection in capillary electrophoresis
Mixing Mode 3: luminol in the Separation Buffer ................................
Final pH value of the post-column CL reagents ...................................

Peak height of 1.0 x 10'8 M Co2+ at various pH value of CL reagent.

Sample: 1.0 x 10‘8 M Co2+ in deionized distilled water; separation buffer:

2 x 10“t M luminol, 0.010 M NaAc/HAC, pH: 4.06; post-column CL
reagents: 0.025 M H202, 0.025 M phosphate, pH = 10.98, 11.90, 12.34;

ﬂow rate: 6 [IL /s; electrophoretic current: 0.8 11A; PMT voltage: 800 V;
ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;
sample injection: 10 s at 10 kV; separation capillary: 50 um i.d. 363 um

o.d. x 50 cm long .......................................................................

3.10 Peak height of 1.0 x 10‘9 M Co2+ at various pH value of CL reagent.

3.11

Sample: 1.0 x 10'9 M Co“ and 5.0 x 10'7 M Cr3+ in deionized distilled
water; separation buffer: 2 x 10“1 M luminol, 0.010 M NaAc/HAc, pH: 4.74;
post-column CL reagents: 0.025 M H202, 0.050 M phosphate, pH = 11.29,
12.06, 12.49; flow rate: 6 ILL ls; electrophoretic current: 2.2 11A; PMT
voltage: 800 V; ampliﬁer: 108 VIA; ampliﬁer rise time: 3 ms; separation
voltage: 10 kV; sample injection: 10 s at 10 kV; separation capillary: 50 um
i.d. 363 um o.d. x 40 cm long. .......................................................

Peak height of 5.0 x 10'7 M Cr3+ at various pH value of CL reagent
Sample: 1.0 x 10'9 M Co2+ and 5.0 x 10'7 M Ci3+ in deionized

xi

...... 56

...... 57

...... 59

...... 62

...... 63

...... 64

distilled water; separation buffer: 2 x 104 M luminol, 0.010 M

NaAc/HAC, pH: 4.74; post-column CL reagents: 0.025 M H202,

0.050 M phosphate, pH = 11.29, 12.06, 12.49; ﬂow rate: 6 [LL ls;

electrophoretic current: 2.2 uA; PMT voltage: 800 V; ampliﬁer:

108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;

sample injection: 10 s at 10 kV; separation capillary: 50 um i.d. 363 um

o.d. x 40 cm long ............................................................................. 65

3.12 Peak height of 1.0 x 108 M Co2+ at various pH values
buffer: 2.0 x 10’4 M luminol, 0.01 M N aAc/HAC; post-column CL
reagent: 0.05 M H202, 0.1 M phosphate, pH = 12.49; flow rate:
6 ILL/min; electrophoretic voltage: 10 kV; injection voltage:
10 kV; injection time: 10 s; capillary: 50 um x 363 um o.d. x 40 cm ................ 67

3.13 Migration time of 1.0 x 10'8 M Co2+ at various pH values
buffer: 2.0 x 10“ M luminol, 0.01 M NaAc/HAc; post-column CL
reagent: 0.05 M H202, 0.1 M phosphate, pH = 12.49; flow rate:
6 uL/min; electrophoretic voltage: 10 kV; injection voltage: 10 kV;
injection time: 10 s; capillary: 50 um x 363 um o.d. x 40 cm ......................... 68

3.14 Peak height and migration time of Co2+(1.0 x 10'8 M), 0.000 M HIBA
buffer: 2.0 x 10“ M luminol, 0.01 M NaAc/HAc, 0.000 M HIBA,
pH = 3.78; post-column CL reagent: 0.05 M H202, 0.1 M phosphate,
pH = 12.49; ﬂow rate: 6 ILL/min; electrophoretic voltage: 10 kV;
injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363
um o.d. x 40 cm ............................................................................... 71

3.15 Peak height and migration time of Co2+ (1.0 x 10‘8 M) with 0.002 M HIBA
buffer: 2.0 x 10'4 M luminol, 0.01 M NaAc/I-IAc, 0.002 M HIBA,
pH = 3.79; post-column CL reagent: 0.05 M H202, 0.1 M phosphate,
pH = 12.49; ﬂow rate: 6 ILL/min; electrophoretic voltage: 10 kV;
injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363 um
o.d. x 40 cm .................................................................................. 72

3.16 Peak height and migration time of Co2+ (1.0 x 10'8 M) at 0.004 M HIBA
buffer: 2.0 x 10“ M luminol, 0.01 M NaAc/HAc, 0.004 M HIBA,
pH = 3.80; post-column CL reagent: 0.05 M H202, 0.1 M phosphate,
pH = 12.49; ﬂow rate: 6 ILL/min; electrophoretic voltage: 10 kV;
injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363 um
o.d. x 40 cm .................................................................................. 73

3.17 Peak height and migration time of Co2+ (1.0 x 10'8 M) at 0.006 M HIBA
buffer: 2.0 x 104 M luminol, 0.01 M NaAc/HAc, 0.006 M HIBA,
pH = 3.78; post—colurrm CL reagent: 0.05 M H202, 0.1 M phosphate,
pH = 12.49; ﬂow rate: 6 ILL/min; electrophoretic voltage: 10 kV;

xii

injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363 um
o.d. x 40 cm .................................................................................... 74

3.18 Background signal of the buffer containing zero HIBA, pH = 4.00.
Buffer: 2.0 x 10“ M luminol, 1.0 x 10'8 M Co“, 0.01 M NaAc/HAc,
0.050 M phosphate, pH = 13.00; CL reagent ﬂow rate: 6 pl. ls;
electrophoretic current: 1.2 11A; PMT voltage: 800 V; ampliﬁer:
108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV ......................... 76

3.19 Background signal of the buffer containing 0.004 M HIBA, pH = 4.00.
Buffer: 2.0 x 10‘4 M luminol, 1.0 x 10'8 M Co“, 0.01 M NaAc/HAc,
0.004 M HIBA, pH: 4.00; post-column CL reagents: 0.025 M H202,
0.050 M phosphate, pH = 13.00; CL reagent ﬂow rate: 6 11115;
electrophoretic current: 1.2 ILA; PMT voltage: 800 V; ampliﬁer:
108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV ......................... 77

3.20 Background signal of the buffer containing 8mM HIBA, pH = 4.00
Buffer: 2.0 x 10“ M luminol, 1.0 x 10'8 M Co”, 0.01 M NaAc/HAc,
8 mM HIBA, pH: 4.00; post-column CL reagents: 0.025 M H202,
0.050 M phosphate, pH = 13.00; CL reagent ﬂow rate: 6 11115;
electrophoretic current: 2.3 11A; PMT voltage: 800 V; ampliﬁer:
108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV .......................... 78

3.21 Migration time and sensitivity of Co2+ for buffer with 5 mM HIBA, pH 3.50
Sample: 1.0 x 10'8 M Co“, 5.0 x 10'7 M Cuz"; buffer: 2.0 x 10'4 M luminol,
0.01 M NaAc/HAC, 0.005 M HIBA, pH = 3.50 post-column CL reagent:
0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 III/min;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time:
10 s; capillary: 50 mm x 363 um o.d. x 50 cm ........................................... 80

3.22 Peak height and migration time of Co” for buffer with 5 mM HIBA, pH 4.00
Sample: 1.0 x 10'8 M Co“, 5.0 x 10'7 M Cu“; buffer: 2.0 x 10“ M luminol,
0.01 M NaAc/HAC, 0.005 M HIBA, pH = 4.00; post-column CL reagent:
0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 uIJmin;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time:
10 s; capillary: 50 mm x 363 um o.d. x 50 cm ........................................... 81

3.23 Effect of pH (3.85) on the background signal
Buffer: 2.0 x 10“ M luminol, 1.0 x 10'8 M Co”, 0.01 M NaAc/HAc,
0.004 M HIBA, pH 3.85; post-column CL reagents: 0.025 M H202,
0.050 M phosphate, pH = 13.00; CL reagent flow rate: 6 “L /s;
electrophoretic current: 1.2 ILA; PMT voltage: 800 V; ampliﬁer:
108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV; capillary:
50 um x 363 um o.d. x 40 cm long ........................................................ 82

3.24 Effect of pH (4.10) on the background signal

xiii

3.25

3.26

3.27

3.28

4.1

4.2

4.3

Buffer: 2.0 x 10'4 M luminol, 1.0 x 10'8 M Coz+, 0.01 M NaAc/HAc,

0.004 M HIBA, pH 4.10; post-column CL reagents: 0.025 M H202,

0.050 M phosphate, pH = 13.00; CL reagent ﬂow rate: 6 [.11. Is;

electrophoretic current: 2.0 11A; PMT voltage: 800 V; ampliﬁer:

108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV; capillary:

50 um x 363 um o.d. x 40 cm long ........................................................ 83

Theoretical plates for 1.0 x 10'“ M Co2+ 0,: 188.9 s; peak width: 1.3 s)

Sample: 1.0 x 101' M in deionized distilled water; separation buffer:

1.0 x 10‘4 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-column CL

reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 ILL ls;
electrophoretic current: 2.0 LIA; PMT voltage: 800 V; ampliﬁer: 108 V/A; ampliﬁer
rise time: 3 ms; separation voltage: 10 kV; sample injection: 10 s

at 10 kV; capillary: 50 um i.d. x 363 um o.d. x 40 cm long ............................ 86

Calibration curve for Co2+

Sample: 1.0 x 10'11 M ~ 1.0 x 10'8 M Co2+ in deionized distilled water;

separation buffer: 2 x 10“ M luminol, 0.010 M NaAc/HAc, pH: 4.74;

post-column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.49;

ﬂow rate: 6 uL /s; electrophoretic current: 2.0 11A; PMT voltage: 800 V; ampliﬁer:
108 V/A; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;

sample injection: 10 S at 10 kV ............................................................. 88

Calibartion curve for Cr 3+

Sample: 5.0 x 10'8 M ~ 1.0 x 10'5 M Cr3+ in deionized distilled water;

separation buffer: 2 x 104 M luminol, 0.010 M NaAc/HAc, pH: 4.74;

post-column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.06;

flow rate: 6 [IL ls; electrophoretic current: 2.0 11A; PMT voltage:

800 V; ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage:

10 kV; sample injection: 10 s at 10 kV .................................................... 89

Calibration curve for Cu2+

Sample: 5.0 x 10’8 M ~ 1.0 x 10'5 M Cu2+ in deionized distilled water;

separation buffer: 2 x 10“1 M luminol, 0.010 M NaAc/HAc, pH: 4.74;

post-column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.06;

ﬂow rate: 6 11L ls; electrophoretic current: 2.0 ILA; PMT voltage:

800 V; ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage:

10 kV; sample injection: 10 s at 10 kV .................................................... 90

Automatic sample introduction system in capillary electrophoresis ................... 96

Electronic circuit for automatic injection system in capillary electrophoresis
MTP30NO6EL: 0.05 (2 turn on resistance; maximum current: 15 A;
Response time: 0.7 us ....................................................................... 98

Dashpot for damping the Shock between sample and separation capillary 100

xiv

4.4

4.5

4.6

4.7

4.8

5.1

5.3

5.4

Rotary alignment system for automatic sample injection .............................. 101

Separation capillary holder in automatic injection system ............................ 103
Support for the solenoid and the shock—dampin g dashpot ............................. 104
Polishing tools for the sample and the separation capillary ........................... 105
Pinch valve holder for automatic injection system ..................................... 107
Reproducibility of peak height for 10 consecutive injections using

automatic injection system for 0.96 S.

Sample: 1.0 x 10'8 M C02+; buffer:

1.0 x 104 M luminol, 0.01 M acetic acid, pH = 4.75. Capillary:

50 um i.d. x 363 um o.d. x 50 cm long fused silica capillary.

Separation voltage: 10 kV. Ampliﬁer: 108 WA, 3 ms response time.

PMT voltage: 800 V. Post-column reagents: 0.05 M H202, 0.1 M

phosphate (pH = 12.50), flow rate: 3 [.111an per channel.

Height of sample injection: 76 cm. Injection interval: 30 S.

RSD of the peak height: 10.7% ........................................................... 111

Reproducibility of peak height for 10 consecutive injections using automatic
injection system for 5 S.

Sample: 1.0 x 10'8 M Co“: buffer: 2.0 x 10“ M luminol, 0.01 M acetic acid,

pH = 4.75. Capillary: 50 um i.d. x 363 um o.d. x 50 cm long fused

Silica capillary. Separation voltage: 10 kV. Ampliﬁer: 108 WA, 3 ms

response time. PMT voltage: 800 V. Post-column reagents: 0.05 M H202,

0.1 M phosphate (pH = 12.50), flow rate: 3 ILL/min per channel. Height of

sample injection: 76 cm. Injection interval: 30 3. Mean of the peak

height: 2.42). RSD of the peak height: 10.9% .......................................... 112

Reproducibility of peak height for 10 consecutive injections using

automatic injection system with pinch valve for 5 5.

Sample: 1.0 x 10'8 M Co“;

buffer: 2.0 x 104 M luminol, 0.01 M acetic acid, pH = 4.75. Capillary:

50 um i.d. x 363 um o.d. x 50 cm long fused silica capillary. Separation

voltage: 10 kV. Ampliﬁer: 108 WA, 3 ms response time. PMT voltage:

800 V. Post-column reagents: 0.05 M H202, 0.1 M phosphate (pH = 12.50),

ﬂow rate: 3 uUmin per channel. Height of sample injection: 76 cm.

Injection interval: 30 8. Mean of the peak height: 2.22. RSD of the peak

height: 13.4% ............................................................................... 113

XV

Chapter 1

Introduction to Chemiluminescence and Capillary Zone Electrophoresis

This chapter describes some basic principles of chemiluminescence, such as the
mechanism of chemiluminescence and the factors affecting the emission intensity.
Several popular chemiluminescence reagents are listed. Chemiluminescence detection in
static and ﬂow systems is also described. The historical background of capillary
electrophoresis is also presented. The advantages of capillary zone electrophoresis over
conventional electrophoresis are described. Some basic principles of capillary zone
electrophoresis, including electroosmosis, sample injection, and various detection
techniques including chemiluminescence are given. The interface between
chemiluminescence and capillary electrophoresis is summarized in detail. Sample
injection techniques such as electrokinetic injection and hydrodynamic injection in CE

are also discussed.

1.1 Introduction to Chemiluminescence
1.1.1 Theoretical Aspects of Chemiluminescence Reactions

Chemiluminescence (CL) is the emission of light from a chemical reaction (1).
Chemiluminescence reactions generally yield a product in an electronically excited State.
Visible light is produced when the excited state relaxes to the ground state. The typical
CL reaction can be represented as:

A+B---->C*---->C+hv (1.1)

where A and B are the reactants, C* is the excited intermediate, and C is the ﬁnal
product. For CL to occur, three general conditions must be met. First, there must be
sufﬁcient energy to produce an excited state. Thus, the reaction must be sufﬁciently
exothermic that

he
—AG = 1.2
9. ( )

8X

 

Here AG is the free-energy change (kcal/mol) of the reaction, h is Planck ’S constant, c is
the speed of light, and 71.." iS the long-wavelength limit in nm for excitation of the
luminescent species. For CL emission in the visible region, -AG must be 40 to 70 kcal /
mol (1). Second, there must be a favorable reaction pathway to produce the excited state.
Third, photon emission must be a favorable deactivation process.

The emission intensity ICL, deﬁned as photons emitted per second, depends on the
chemiluminescence quantum yield, (00,, deﬁned as the number of photons emitted per

molecule reacting and the rate of reaction, dC/dt (2):

dC
ICL = (bu —d_t- (1.3)

the], depends on how efﬁciently the excited states are generated from the molecules
reacting (ogx, excitation efﬁciency) and on how efﬁciently the excited states luminesce
((01,, luminescence efﬁciency) (4):

¢CL =¢ex in (1.4)

Combine Equations 1.3 and Equation 1.4, we have:

dC
ICL =¢Ex ¢L Et- (1'5)

Equation 1.5 indicates that the intensity of chemiluminescence depends on the rate of
chemical reaction, the efﬁciency of production of the excited state, and the efﬁciency of
light emission from the excited state. For most CL reactions, the emission intensity
quickly increases to a maximum after mixing and then slowly decays to the background
level as the CL reagents are consumed, giving rise to a peak-shaped transient signal.

The advantages of chemiluminescence detection include high sensitivity, wide linear
dynamic range, and simple instrumentation (1). Chemiluminescence is considered a dark
ﬁeld technique because it generates a signal against a dark background. No external light
source is needed as in UV absorption, and thus the background Signal is very low. Unlike
ﬂuorescence, the CL background signal is usually determined by reagent purity rather
than by Raman scattering. Small changes in a small background are more easily detected
than small changes in a large background. The problem with chemiluminescence
detection is the lack of selectivity. CL emission depends on a variety of environmental
factors, such as pH, temperature, ionic strength, solvent and species other than that of
interest. CL emission is transient; the emission vs. time proﬁle can be different from one
compound to another, making the detection more complicated. Furthermore, CL reactions

with complex kinetics can give rise to nonlinear calibration curves (3).

1. 1.2. Chemiluminescence Reagents

Several CL reactions have been used for analytical purposes, including the
peroxyoxalate, acridinium, ﬁreﬂy luciferase, lucigenin, and luminol (5-amino-2,3-
dihydro-l,4-phthalazine-dione) reactions. Luminol is one of the most widely used CL

reagents. In aqueous alkaline solution, luminol is oxidized by H202 or other oxidant such

as Mn04', C10' and I2 (5) to 3-aminophtha1ate in the presence of catalysts (peroxidase,
hemin and transition metal ions). The CL intensity is proportional to the concentrations of
luminol, H202 and catalysts (2). Thus, measurements of CL intensity can be used to
quantitate luminol or derivatives of luminol, H202 or species that can be converted into
H202, and catalysts. The mechanism of the luminol-H202 reaction can be described
below. Luminol is ﬁrst ionized to an anion carrying two electrons in alkaline media,
which then reacts with active 02* to generate an excited state intermediate. When the
excited state of 3-aminophthalate returns to the ground state, blue light is emitted at 425

nm.

11 if

\ \
NH N'
| + 20H" """‘ l + 2H20
n/NH ll/N
NH2 0 NH2 0
Luminol 3-aminophthalate
catalyst at
H202 H20+02
It 1.
N' t O-
|_ + 02 ""’ |_ + N2
. l .

- 0'
O (|)_ ' ' ' ' " O I _ + hv (425nm)
/0 /0

1.1.3 Historical Background of Chemiluminescence Detection

Many transition metal ions are known to activate or inhibit the chemiluminescence of
luminol in basic solution. Seitz and Neary (6), and Isacsson and Weterrnark (7) have
reviewed analytical applications. When the reagents are in excess, the CL emission can be
related to the concentration of metal ions (8). In 1962, Babko and coworkers (9-12) did
the earliest determination of cobalt, copper and iron based on their catalysis of the
luminol reaction in static mode. By Simply mixing the reactants while exposing the
container to a photographic ﬁlm and measuring the exposure as a function of
concentration, they found detection limits of 1 ppb, 3 ppb and 10 ppb for cobalt, copper
and iron, respectively. In 1972, Seitz and Hercules developed speciﬁc methods for the
determination of Cr3+ (13) and Fe2+ (14). The CL responds more sensitively to Co2+ than
to other metal ions, and the detection limit for Co2+ has been estimated to be 101' M (8).
The early work laid the foundation for on-line CL detection of transition metal ions
separated by ion chromatography (15, 16). CL has been used as a sensitive means of
detection in capillary zone electrophoresis; a detailed description is given in section 1.2.6

of this chapter.

1.2 Introduction to Capillary Zone Electrophoresis
1.2.1 Historical Background of Capillary Electrophoresis

Electrophoresis is the separation of charged species based on differential migration
under an applied electric ﬁeld. It was ﬁrst described by Arne Tiselius in 1937 (17). In
1967, Hjerten presented the pioneering demonstration of capillary electrophoresis (18). In
1974, Virtanen described electrophoretic separations in 0.2 - 0.5 mm id. glass tubes with
potentiometric detection (19). In 1979, Mikkers and Everaerts performed the separation
of organic and inorganic anions in Teﬂon tubes (20, 21). However, high separation
efﬁciency was not achieved due to sample overloading. In 1981, Jorgensen and Lukacs
demonstrated the ﬁrst high efﬁciency separation of more than 400,000 theoretical plates
using a narrow bore glass capillary (22). This work has become the landmark in modern
capillary zone electrophoresis. The literature of capillary electrophoresis has been
reviewed in detail by several workers (23-26). In this section, the important theory and

various detection techniques, including chemiluminescence detection, are reviewed.

 

Figure 1.1 Schematic Representation of Capillary Electrophoresis

In capillary zone electrophoresis, a narrow plug of sample is introduced into the
buffer-ﬁlled capillary from the high voltage end, as illustrated in ﬁgure 1.1. When a high
electric ﬁeld is applied across the capillary, different species in the sample migrate

differentially, according to ionic mobility.

1.2.2 Advantages of Capillary over Conventional Electrophoresis

The major limitation in conventional electrophoresis is the temperature rise induced
by Joule heating which results in temperature gradients across the capillary. The
subsequent convection increases zone broadening. Severe Joule heating can lead to
boiling of the electrophoretic buffer. The unique advantage of capillary electrophoresis is
the enhanced heat dissipation. A narrow bore capillary provides better heat dissipation
due to the large surface area / volume ratio, and permits the use of very high electric
ﬁelds which are required for fast and efﬁcient separations. Other advantages of using a
capillary include minimal convection due to the small radius, sampling from a

microenvironment like a single cell and Simple automation.

1.2.3 Electroosmosis
Electroosmosis is the ﬂow of bulk solvent caused by the electrical double layer
adjacent to the inner surface of a capillary under the inﬂuence of an electric ﬁeld. This

phenomenon can be explained by the interaction of silanol groups with aqueous solution

at the inner wall of the fused-silica capillary. The surface Si-OH group is ionized to Si-O'
in alkaline and slightly acidic media; thus, the inner wall of the capillary is negatively

charged and positive counter ions are present in the stagnant double layer and diffuse

layer as shown in ﬁgure 1.2. In strongly acidic media (pH < 2), electroosmotic ﬂow is
very small. The potential across the double layer is termed the zeta potential, C, and is

given by: .

 

4nnu
C: CO
e

(1.6)

Here n is the viscosity of the operating buffer, pa, is the electroosmotic ﬂow coefﬁcient,

and 8 is the zero frequency dielectric constant of the carrier buffer.

 

8 g
z 17
“'60 4 It 11 ( )

The thickness of the double layer is usually a few nanometers to a few hundred
nanometers. When an electric ﬁeld is applied across a narrow bore capillary, the solvated
counterions in the diffuse layer migrate with a ﬂat ﬂow proﬁle toward the cathode. The
direction and rate of electroosmotic ﬂow depend on the polarity and magnitude of the
zeta potential, respectively. The direction of electroosmotic ﬂow is normally toward the
cathode in aqueous solutions, but can be reversed by introducing a cationic surfactant
such as cetyltrimethylammonium bromide (CT AB) into the separation buffer due to its

adsorption to the negatively charged capillary wall (27).

 

Figure 1.2 Schematic Representation of Silica-Solution Interface

1.2.4. Theoretical Aspects of Capillary Zone Electrophoresis
1.2.4.1 Migration Time

In capillary zone electrophoresis, the migration time tr is the time for a charged
species to travel the distance L’ from the beginning of capillary to the detector (22).

L’L

(1.8)
Vep (uep + lies)

 

tr:

where L is the entire length of capillary, pep is the electrophoretic mobility of the sample,
and ch is the electrophoretic voltage. The length of capillary has a profound inﬂuence on

migration times as can be seen from Equation 1.8.

1.2.4.2 Separation Efﬁciency
Separation efﬁciency in terms of the total number of theoretical plates, N, is given by
(22):

__ (Hep + Heo) Vep
2D

N (1.9)

 

Here D is the solute 's diffusion coefﬁcient. Equation 1.4 indicates that the separation
efﬁciency has nothing to do with the length of capillary. This is somewhat different from
chromatography where the theoretical plate number increases with the length of column.
Ideally, high efﬁciency separation is best performed with high voltage across a short
capillary. However, there are practical limits to this approach. As the capillary is made
Shorter, the surface area available for heat dissipation is decreased; on the other hand,

more Joule heat is produced due to the reduced electric resistance of the capillary.

1.2.4.3 Resolution

Resolution R8 of two zones in capillary zone electrohporesis is given by (22):

 

 

 

l
2 ll "11
Rs =_1_[(ll+lleo)V] M (1.10)
4 2D l-l'H’reo
v
Rs =0.177(tte 1 ‘11s 2) 6" (1.11)
p p 1301 +l'leo)

where Item and pep2 are the electrophoretic mobilities of the two species and u is their

average mobility. From this equation, it is clear that resolution is poor if there is a large
electroosmotic ﬂow in the same direction as the electrophoretic migration. Resolution can
be improved by suppressing the electroosmotic ﬂow at the cost of a longer migration

time.

1.2.5. Sample injection

Electrokinetic and hyrdrodynamic injection are best suited for reproducible delivery
of nanoliters or even picoliters of sample into 25 pm to 75 um internal diameter
capillaries (28, 29). Several other injection techniques, such as a split injector (31), a

rotary injector (32), and a micro injection system (3-34) have also been designed.

1.2.5.1 Electrokinetic injection
In electrokinetic injection, the anodic end of the capillary and the anode are placed
into the sample vial together. An injection voltage (typically 2 kV — 10 kV) is applied

brieﬂy, causing electromigration of small plug of sample into the capillary. The capillary

10

and the anode are then placed into the buffer reservoir, and the separation voltage
(typically 10 kV - 20 kV) is applied. The quantity of sample injected into the capillary,

Qinjection can be calculated (28)

2

t C

_ ("6P + ”60) Vinjection 7‘ ’
injection _ L

injection

 

(1.12)

where tinjection is the time of injection at injection voltage Vim-cam, r is the inner

radius of the capillary, and C is the sample concentration. The amount of sample injected

can also be determined by substituting equation 1.8 into equation 1.12.

i 2 '
Q _ L 1! r C Vinjection tinjection
in‘ection _
J Vep tr

 

(1.13)

The drawback of electromigration injection is the two types of discrimination (27, 29)
that result. The ﬁrst one occurs as a result of the differences in mobility of the sample
species. Charged species with higher mobility are introduced into the capillary faster and
in larger quantities than species with lower mobility. The second type of discrimination is
related to the differences in the conductivity between the sample and the buffer (29).
Dasgupta and colleagues (35) attempted to use a small wire loop to reduce differential
injection of species with different mobilities. They reported that the mobility induced bias
is dramatically lowered. However, the injection device is very sophisticated and difﬁcult
to construct. Linhares and coworker (36) fabricated an on-column fracture for sample
injection to reduce discriminative injection. In this method, an on-column fracture was
made on the capillary. The fracture allows ions, but does not allow a signiﬁcant quantity
of buffer solution to pass through. During injection, a potential is applied between the

fracture and the outlet of the capillary. Electroosmotic ﬂow (EOF) pulls sample ions into

11

the capillary through the inlet of the capillary. The amount of sample introduced is
proportional to the EOF. It is assumed that no electromigration of analytes occurs and

introduction of sample is on longer biased by using the EOF as pump.

1.2.5.2 Hydrodynamic Injection

In hydrodynamic injection, the end of the capillary is placed into the sample reservoir,
followed by lifting the sample reservoir to a Speciﬁed height Ah above the level of the
ground buffer reservoir (29). This height difference causes hydrostatic pressure, pushing
the sample into the capillary without any discrimination for ionic components. The

quantity of sample injected can be calculated by

4
P g TE r Ah C tinjection

Qinjection = 8 n L (1.14)

 

where p is the solution density, g is the gravitational force constant, and n is the viscosity

of the sample solution.

1.2.6. Detection in Capillary Zone Electrophoresis

Capillary zone electrophoresis is a powerful technique for the separation of biological
macromolecules such as oligoneucleotides, proteins and DNA. One of the major
limitations is the lack of sensitive detection. Several reviews have emphasized the
necessity for enhanced detection (23, 24, 37, 38). Currently commercial CE instruments
provide either UV/visible absorbance or ﬂuorescence detection. UV-visible absorbance
is the most popular on-column detection technique, although the detection sensitivity is

rather poor (detection limit 10'5 ~ 10'7 M) due to the short optical pathlength. Several

l2

attempts have been made to enhance sensitivity by increasing the optical path length (39-
47). Sensitivity can be increased using a capillary of bigger internal diameter, but a bigger
diameter also means more Joule heating, which is not good for efﬁcient separation.
“Bubble cells”, a pipet-like structure made directly from capillary, offer longer path
length with minimal loss of resolution when using a reduced detection slit width (39). Z-
cells (a short section of bends in the capillary) offer a longer pathlength but cause a loss
of separation efﬁciency (40-41). Also, the Z—cell structure is fragile. A Z-cell design is
commercially available from LC Packings Inc. Rectangular capillaries give longer
pathlength but higher Joule heating (42). Internal reﬂection cells offer longer pathlength
by reﬂecting the light through the section of the capillary several times, but cause a loss in
resolution (43-44).

Fluorescence detection is by far the most sensitive detection technique routinely used
in CE. Chen and coworkers have reported the detection of 30 molecules of Rodamine
using a green He-Ne laser (48) and the detection of 6 molecules of sulforhodamine 101
using a yellow He-Ne laser (49). Only a small number of chemicals are naturally
ﬂuorescent. For those non-ﬂuorescent analytes, such as oligosaccharides (48-50), the
primary amine group of amino acids (51-53), peptides (54), biological thiols (55), and
phosphate containing compounds (56), precapillary or post-capillary Chemical
derivatization is required. Fluorescence derivatization sometimes results in poor yield and
several impurities which can increase background. Other detection techniques include
radioisotopes (57-58), Raman spectroscopy (59), mass spectrometry (60-64), conductivity

(65—72), amperometry (73-88), thermoptical (89-91) and chemiluminescence (92-115).

13

Chemiluminescence (CL) detection is based on the reaction of the analyte with the CL
reagent to produce light in the detection cell. Many post-column CL reactors have been
successfully interfaced with CE capillaries. The coaxial CL ﬂow cell is popular for CE
detection in which the separation capillary is inserted into a reaction capillary held in
position by a TEE connector (93). Since the CL signal is transient, mixing should take
place in front of the collection optics. In a ﬂow system like CE, the ideal CL-CE reactor
should have no turbulence, and diffusion should be the dominant mixing mechanism.
This requires the matching of the ﬂow rate between the separation buffer and the CL
reagents. If the CL reagent ﬂow rate is too high, sample zone can be swept from the ﬂow
cell before the reaction is complete; if the ﬂow rate is too low, band broadening will
occur due to a long residence time in the ﬂow cell (93). A sheath ﬂow cell allows the CL
reagents to carry the sample zone toward the detection window undisturbed, minimizing
the turbulent ﬂow which often occurs in conventional cross mixing.

Hara and colleagues reported the ﬁrst peroxyoxalate CL reaction detection for CE
(92, 94—97). However, the stability of peroxyoxalate CL reagents can be affected by the
high electric ﬁeld in CE. The low solubility and instability of oxalate derivatives in
aqueous solution are another problem. The use of an organic solvent as the CE buffer
affects the migration behavior of the analytes. Wu and Huie (98) overcame the problem
of incompatibility between mixed aqueous—organic solvents and electrically driven
separations by switching off the CE power supply at an appropriate time and connecting

the CE capillary to a syringe pump to effect dynamic elution.

l4

Ruberto and Grayeski (99-100) reported the acridinium CL reaction for CE detection
using a post-column CL reactor. A detection limit of 650 attomoles of acridinium ester
was reported. Separation produced ~ 4,000 theoretical plates.

In 1992, Dadoo and coworkers (101) reported a detection limit of 3 x 10’9 M for
luminol and 7 x 10'9 M for N-(4-aminobutyl )-N-ethylisoluminol using the luminol CL
reaction for CE detection with a coaxial CL reactor. The detection limits are about two or
three orders of magnitude lower than UV detection. In this work, parabolic mirror was
used to reﬂect the light generated from the CL reaction onto the PMT for high collection
efﬁciency. Due to the turbulent mixing of the analyte with the CL reagent, separation
efﬁciency was very poor (< 20,000 theoretical plate number). To simplify the instrument
design, Dadoo and coworkers also developed an end-column CL detector in which the
signal is generated at the capillary outlet (102). In this new design, the outlet end of the
separation capillary is immersed into the buffer reservoir containing the CL reagent. The
eluted analytes react with CL reagents in the buffer reservoir to produce light, and then
the light is transported through a single ﬁber to a photomultiplier tube. The new design
Showed a less effective collection efﬁciency due to the small solid angle of the ﬁber
optics for light collection. A detection limit of 2 x 10'8 M for luminol and 5 x 10'9 M for
ATP was reported. Zhao and coworkers (103) separated the isoluminol thiocarbamyl
derivatives of amino acids using a sheath ﬂow cuvette and were able to improve the
separation efﬁciency up to ~ 100,000 plate number although the complete separation of
twenty isoluminol derivatized amino acids was not accomplished.

Huang and coworkers (104-106) applied the catalytic effect to detect transition metal

ions in CE by a sample stacking injection. They claimed that mixing luminol in the

separation buffer rather than in the post-colurrm CL reagents could signiﬁcantly enhance
the sensitivity of detection. The separation of metal ions is best preformed in acidic media
to avoid hydrolysis in basic media. They claimed an average theoretical plate number of
460,000 for the separation of Coz“, Cu“, Ni”, Fe3+ and Mn”, which is much higher than
the common theoretical plate number with post-column detection in CE (~ 50,000). A
post-column CL detector induces additional zone broadening due to the CL reaction.
There are quite a few things that need to be questioned in their most recent article (106).
For example, the workers were able to dissolve luminol to 1 mM in a buffer of pH 4.54.
The same experiment was repeated here, but the solubility of luminol in the identical
buffer was found to be less than 0.3 mM. Based on the electropherograrn given in this
paper (106), we calculated the theoretical plate number for each metal ion. Results
indicate that the theoretical plate number is ~ 34,596 for (302+, 21,232 for Cu“, 213,444
for Ni“, and 150,544 for Fe3+ respectively. The average theoretical plate number is
104,954, much lower than the average theoretical plate number claimed. Huang and
coworkers (106) also claimed the detection limits of 5.0 x 10'13 M for Co“, 1.0 x 10'10 M
for Cu“, 5.0 x 10'9 M for Ni”, 5.0 x 10‘8 M for Fe”, and 8.0 x 10'6 M for Mn“.

Indirect CL detection has also been used for the detection of species that produces
interference or suppression to a CL reaction where the analytes are detected as an inverted
peak. Liao and colleagues (110-111) applied indirect CL detection for ﬁve amino acids
using a Cu2+-catalyzed CL system. The formation of a Cu2+-amino acid complex reduced
the catalytic activity of Cu(II) considerably, producing a negative peak. Indirect CL
detection can be applied to more species although it is less sensitive than direct CL

detection.

16

Several reviews have also covered the use of chemiluminescence as a method of
detection (116-120). The previously described CE systems with CL detection are bulky
and complicated. Campana and coworkers, (117) who presented the most recent review
on the application of chemiluminescence to capillary electrophoresis in 1997 suggested,
“Future advances should focus on developing new detectors that are instrumentally
simpler than current ones”. Chapter 2 describes the construction of the miniaturized
chemiluminescence detector. The CL detector is mounted on the side of a high voltage
Plexiglas box so that the whole system is a single unit, unlike those used in previous work
in which the CL detector is separated from the CE system. Chapter 3 describes the
characterization of the miniaturized CE-CL detector. The major problem is previous work
in which the CL detector is separated from the CE system. Chapter 3 describes the
characterization of the miniaturized CE-CL detector. The major problem is the strong
adsorption of metal ions to the surface of the fused silica capillary. Chapter 4 describes
the construction of an automatic sample injection system for CE. This new injection
system is characterized in chapter 5. Chapter 6 summarizes brieﬂy two parts of the work
(chemiluminescence detection and the automatic sample injection system for CE) and
draws some important conclusions. Chapter 7 describes the future work for CL detection
for CE and the automatic sample injection system in which I suggest the adoption of a

linear motion control system for more precise sample injection.

17

List of References

l.

10.

ll.

12.

13.

14.

15.

Ingle, J. D. Jr. and Crouch, S. R., Spectrochemical Analysis, Prentice Hall,
Englewood Cliffs, New Jersey, 1988, p 479—485.

Birks, John W, Chemiluminescence and Photochemical Reaction Detection in
Chromatography, VCH Publishers Inc., 1989, p99.

Nieman, T. A., Luminescence Techniques in Chemical and Biological Analysis,
Practical Spectroscopy Series, Vol. 12, Marcel Dekker, New York, 1991, pp523.
Burr, John G, Chemiluminescence and Bioluminescence, Marcel Dekker Inc., New
York, 1985, p 470.

Seitz, W. R., CRC Crit. Rev. Anal. Chem, 1981, 13, l-58.

Seitz, W. R. and Neary, M. P., Contemp. Topics Anal. Clin. Chem, 1977, 1, 49-
125.

Isacsson, U. and Wetterrnark, G., Anal. Chim., Acta., 1974, 68, 339-362.

Seitz, W. R., and Hercules, D. M, Chemiluminescence and Bioluminescence,
Plenum, New York, 1973, p427-449.

Babko, A. K and Lukovskaya, N. M. Zh. Anal. Khim., 1962, 17, 50.

Babko, A. K. and Lukovskaya , N. M. Zavod. Lab., 1963, 29, 404.

Babko, A. K. and Dubovenko, L. I. Zh. Anal. Khim., 1964, 200, 428.

Babko, A. K. and Kalinichenko, I. E. Ukr. Khim. Zh., 1965, 31, 1316.

Seitz, W. R.; Suydam W. W.; Hercules, D. M. Anal. Chem, 1972, 44, 957-963.
Seitz, W. R., and Hercules, D. M., Anal. Chem, 1972, 44, 2143-2149.

Neary, M. P., Seitz, W. R., and Hercules, D. M., Analytical Letters, 1974, 7, 583-

590.

18

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32

33

34.

Delumyea, R. and Hartkopf, A. V., Anal. Chem, 1976, 48, 1402-1405

Tiselius, A. Trans. Faraday Soc., 1937, 33, 524.

Hjerten, S. Chromatogr. Rev., 1967, 9, 122-219.

Virtanen, R. Acta Polytech. Scand, 1974, 123, 1.

Mikkers, F. E. P.; Everaerts, F. M.; Verheggan, Th. P. E. M. J. Chromatogr., 1979,
I69, 11.

Mikkers, F. E. P.; Everaerts, F. M.; Verheggen, Th. P. E. M. J. Chromatogr., 1979,
I69, 1.

Jorgenson, J. W.; Lukacs, K. D. Anal. Chem, 1981, 53, 1298.

Kuhr, W. G., Anal. Chem, 1990, 62, 403R-414R.

Kuhr, W. G.; Monnig, C. A. Anal. Chem, 1992, 64, 389R-407R.

Clarie, R. L. 8., Anal. Chem, 1996, 68, 569-586R

Beale, Stephen 0, Anal. Chem, 1998, 70, 279R-300R.

Li, S.F.Y. Capillary Electrophoresis-Principles, Practice and Applications;
Journal of Chromatography Library, Vol. 52; Elsevier: Amsterdam, 1992; pp31-
53.

Lukacs, K. D.; Jorgenson, J. W.; J. High Res. Chromatogr. Commun, 1985, 8, 407.
Huang, X.; Gordon, M. J .; Zare, R. N. Anal. Chem. 1988, 60, 375-377.

Rose, D. J.; Jorgenson, J. W. Anal. Chem, 1988, 60, 642.

Deml, M.; Foret, F.; Bocek, P. J. Chromatogr. 1985, 320, 159.

Tsuda, T.; Mizuno, T.; Akiyama, J. Anal. Chem. 1987, 59, 799.

Wallingford, R. A.; Ewing, A. G. Anal. Chem. 1988, 60, 1972.

Wallingford, R. A.; Ewing, A. G. Anal. Chem. 1987, 59, 678.

19

35.

36.

37.

38.

39.

40.

41.

42.

43

45.

46.

47.

48.

49.

50

51

Dasgupta, P. K.; Surowiec, Kazimierz. Anal. Chem. 1996, 68, 4291-4299.
Linhares, M. C.; Kissenger, P. T. Anal. Chem. 1991, 63, 2076-2078.

Goodall, D. M.; Lloyd, D. K.; Williams, S. J. LC-GC 1990, 8, 788-799.

Novotony, M. V.; Kobb, K. A.; Liu, J. J. Electrophoresis, 1990, 11, 735-749.
Heiger, D. N., “High Performance Capillary Electrophoresis ”, an introduction.
Hewlett Packard, Waldbronn, Germany, 1992, p.101

Chervet, J. P.; Van-Soest, R. E. J .; Ursem, M. J liq. Chromatogr. 1991, 543, 439.
Moring, S. E.; Reel, R. T.; Van-Soest, R. E. J ., Anal. Chem. 1993, 65, 3454.

Tsuda, T; Sweedler, R. N.; Zare, R. N. Anal. Chem. 1990, 62, 2149.

Wang, T.; Hartwick, A. Anal. Chem. 1992, 64, 1745. P

Wang, T.; Aiken, J. H. Huie, C. W.; Hartwick, R. A. Anal. Chem. 1991, 63,
1372.32.

Djordjevic, N. M.; Widder, M.; Widder, M.; Kuhn, R. J. High Resolut.
Chromatogr. 1997, 20, 189-192.

Delonge, T.; Fouckhardt, H. J. J. Chromatogr., A 1995, 716, 135-139.

Djordjevic, N. M.; Ryan, K. J. liq. Chromatogr. Relat. Technol. 1996, 19, 201-206.
Chen, D. Y.; Dovichi, N. J. J. Chromatogr., B: Biomed. Appl. Chem. 1994, 657,
265-9.

Chen, D. Y.; Adelhelm, K.; Cheng, X. L.; Dovichi, N. J. Analyst 1994, 119, 349-
352.

Chen, F.-T. A.; Evangelista, R. A. Anal. Biochem. 1995, 230, 273-280.

Lada, M. W.; Kennedy, R. T. Anal. Chem. 1996, 68, 2790-2797.

20

52.

53

54

55

56

57.

58.

59.

60.

61.

62.

63.

65.

66.

67.

Engstroem, A.; Anderson, P. E.; Josefsson, B.; Pfeffer, W. D. Anal. Chem. 1995,
67, 3018-3-22.

Chen, P.; Novotny, M. V. Anal. Chem. 1997, 69, 2806-2811.

Beijersten, I.; Westerlund, D. J. Chromatogr., A 1995, 716, 389-399.

Orwar, 0.; Fishman, H. A.; Ziv, N. E.; Scheller, R. H.; Zare, R. N. Anal. Chem.
1995, 67, 4261-4268.

Wang, P.; Giese, R. W. Anal. Biochem. 1995, 230, 329-332.

Tracht, S. E.; Cruz, L.; Stobba-Wiley: C. M.; Sweedler, J. V. Anal. Chem. 1996, 68,
3922-3927.

Klunder, G. L.; Andrews, 1. E., Jr.; Grant, P. M.; Anderson, B. D.; Russo, R. E.
Anal. Chem. 1997, 69, 2988-2993.

Walker, P. A., III.; Kowalchyk, W. K.; Morris, M. D. Anal. Chem. 1995, 67, 4255-
4260.

Ramsey, R. S.; Mcluckey, S. A. J. Microcolumn Sep. 1995, 7, 461-469.
Mazereeuw, M.; Hofte, A. J. P.; Tjaden, U. R.; van der Greef, J. Rapid Commun.
Mass Spectrom. 1997, 11, 981-986.

Cao, R; Moini, M. J. Am. Soc. Mass Spectrom. 1997, 8, 561-564.

Severs, J. C.; Smith, R. D. Anal. Chem. 1997, 69, 2154-2158.

Lamoree, M. H.; Tjaden, U. R.; van der greef, J. J. Chromatogr., A 1997, 777, 31-
39.

Huang, X.; Pang, T-K J; Gordon, M. J .; Zare, R.. N. Anal. Chem. 1987, 59, 2747.
Huang, X.; Luckey, J. A.; Gordon, M. J .; Zare, R. N. Anal. Chem. 1989, 61, 766.

Huang, X. H.; Zare, R. N. Anal. Chem. 1991, 63, 2193-2196.

21

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

Zemann, A. J .; Schnell, Erhard; Volgger, Dietrnar; bonn, G. K. Anal. Chem. 1998,
70, 563-567.

Mayrhofer, K.; Zemann, A. J .; Schnell, Erhard; Bonn, G. K. Anal. Chem. 1999, 71,
3828 — 3833.

Mueller, D.; Jelinek, 1.; Opekar, F.; Stulik, K.; Electrophoresis 1996, 8, 722-725.
Harrold, M.; Stillan, J .; Bao, L.; Rocklin, R.; Avdalovic, N. J. Chromatogr., A 1995,
717, 371-383.

Haber, C.; Jone, W. R.; Soglia, J.; Surve, M. A.; McGlynn, M.; Caplan, A.;
Reineck, J. R.; Krstanovic, C. J. Capillary Electrophor. 1996, 3, 1-11.

Weber, P. L.; Lunte, S. M. Electrophoresis 1996, 17, 302-309.

Wen, J .; Cassidy, R. M.; Anal. Chem. 1996, 68, 1047-1053.

He, P.; Ye, J .; Fang, Y.; Suzuki, 1.; Osa, T. Anal. Chim. Acta 1997, 337, 217-223.
Swanek, F. D.; Chen, G.; Ewing, A. G. Anal. Chem. 1996, 68, 3912-3916.

Lin, H.; Xu, D.-K.; Chen, H.-Y. J. Chromatogr., A 1997, 760, 227-233.

Liu, J .; Zhou, W.; You, T.; Li, R; Wang, E.; Dong, S. Anal. Chem. 1996, 68, 3350-
3353.

Jin, W.; Weng, 0.; Wu, J. Anal. Chim. Acta 1997, 342, 67-74.

You, T.; Wu, M.; Wang, E. Anal Lett. 1997, 30, 1025-1036.

Jin, W.; Weng, Q.; Wu, J. Anal. Lett. 1997, 30, 753-769.

Hu, S.; Wang, Z.-L.; Li, P.-B.; Cheng, J .-K. Anal. Chem. 1997, 69, 264-267.
Matysik, F.-M. J. Chromatogr., A 1996, 742, 229-234.

Park, S.; Lunte, C. E. Anal. Chem. 1995, 67, 4366-4370.

22

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

Zhong, M.; Zhou, J.; Lunte, S. M.; Zhao, G.; Giolando, D. M.; Kirchholl, J. R.
Anal. Chem. 1996, 68, 203-207.

Voegel, P. D.; Zhou, W.; Baldwin, R. P. Anal. Chem. 1997, 69, 951-957.

Ferrnier, A. M.; Gostkowski, M. L.; Colon, L. A. Anal. Chem. 1996, 68, 1661-
1664.

Zhong, M.; Lunte, S. M. Anal. Chem. 1996, 68, 2488-2493.

Waldron, K. C.; Li, J. J. J. Chromatogr,, B: Biomed. Appl. 1996, 683, 47-54.
Seidel, B. S. Steinle, E.; Faubel, W.; Ache, H. J. Proc. SPlE-int. Soc. Opt. Eng.
1996, 2836, 283-292.

Ren, J .; Li, B.; Deng, Y.; Cheng, J. Talanta 1995, 42, 1891-1895.

Hara, T.; Okarnura, S.; Kato, J .; Yokogi, J.; Nakajima, R. Anal. Sci. 1991, 7, 261-
264.

Rose, D. J. and Jorgenson , J. W. J. Chromatogr., 1988, 447, 117.

Hara, T.; Kayama, S.; Nashida, H.; Nakajima, R. Anal. Sci. 1994, 10, 223-25.

Hara, T.; Kayama, S.; Nashida, H.; Nakajima, R. Anal. Sci. 1994, 10, 823-25.
Tsukagoshi, K.; Tanaka, A.; Nakajima, R.; Hara, T. Anal. Sci. 1996, 12, 525-28.
Tsukagoshi, K.; Akasaka, H.; Nakajima, R.; Hara, T. Chem. Lett. 1996, 467-468.
Wu, N.; huie, C. W. J. Chromatogr. 1993, 634, 309-15.

Ruberto, M. A.; Grayeski, M. L. Anal. Chem. 1992, 64, 2758-62.

Ruberto, Michael A. and Grayeski, M. L. J. Microcol. 1994, Sep. 6, 545-550.
Dadoo, R.; Colon, L. A.; Zare, R. N. J. High Resolution Chromatography, 1992,
15, 133.

Dadoo, R.; Seto, A. G.; Colon, Luis A.; Zare, R. Z. Anal. Chem, 1994, 66, 303.

23

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

Zhao, J. Y.; Labba, J .; Dovichi, N. J. J. of Microcolumn Separation, 1993, 5, 331.
Huang, b.; Li, J ., and Cheng, K., “Highly Sensitive Chemiluminescence Detection
for Capillary Ion Analysis”. Sepu, 1995, 13, 430.

Huang, 3.; Li, J. J .; Cheng, J. K. Chemical Journal of Chinese Universities, 1996,
17, 528-530.

Huang, Bo; Li, Jian-jun; Zhang, Le; Cheng, Jie-ke, Anal Chem, 1996, 68, 2366-
2369.

Dasgupta, P. K.; Genfa, Z.; Li, J .; Boring, C. B.;Jambunathan, S., and Al-Horr, R.,
Anal. Chem, 1999, 71, 1400.

Tsukagoshi, K.; Fujimura, S. and Nakajima, R. Anal. Chem, 1997, 13, 279.

Lee, Y.-T, and Whang, C.-W. J. Chromatogr., A, 1997, 771, 379.

Liao, S.-Y., and Whang, C.-T., J. High. Resolut. Chromatogr., 1995, 18, 667.

Liao, S.-Y., and Whang, C.- Y., J. Chromatogr., 1996, 736, 247.

Hara, T.; Yokogi, J.; Okamura, S. J. Chromatography, 1993, A652, 361.

Gilman, S. D.; Silverman, C. E.; Ewing, A. G. J. of Mirocolumn Separation, 1994,
6, 97.

Hara, T.; Nishida, H.; Kayama, S.; Nakajima, R. Bull. Chem. SOC. Japan, 1994,
67,1193.

Gilman, S. D.; Silverman, C. E.; Ewing, A.G. J. Microcolumn Separation, 1994, 6,
97.

Baeyens, W.R.G.; Ling, B. L.; Imai, K.; Calokerino, A.C.; Schulman, S.G. J.

Microcolumn Separation, 1994, 6, 195.

24

117.

118.

119.

120.

Garcia Campana, Anna M; Baeyens, W. R.G. ; Zhao, Y. N. Anal. Chem. 1997,
83A.

Sanders, M. G.; Andrew, K. N., and Worsford, P. J., “Trends in
Chemiluminescence Detection for Liquid Separation, Anal. Commun., 1997, 34, 13.
Baeyens, W. R. G.; Schuman, S. G.; Calokerinos, A. C.; Zhao, Y., Garcia
Compana, A. M.; Nakasima, K., and De Keuleleire, D., J. Pharm. Biomed. Anal,
1998, 17, 941.

Soper, S. A.; Warner, I. M., and McGown, L. B., Anal. Chem, 1998, 70, 477R.

25

Chapter 2
Construction of a Miniaturized Post-Column Chemiluminescence Detector for Capillary
Electrophoresis

Chemiluminescence (CL) has been applied as a sensitive means of detection in
capillary electrophoresis (1-23) since the early 90’s. As discussed in this chapter, most
CE-CL systems use a post-column sheath ﬂow cell for the detection. These CL detectors
are usually enclosed in a large, dark box. The box is bulky, heavy, and inconvenient for
transportation. It is not convenient for adjustment on the bench-top since the
chemiluminescence detector and the CE system are not held together. Furthermore, since
only one end of the reaction capillary is held by a TEE connector, the reaction capillary
may be tilted slightly to a different angle after another installation, which can affect the
solid angle for light collection and cause a variation in detection sensitivity. A large dark-
box with many joint edges may increase the chances of room-light leakage which can
signiﬁcantly increase the background noise level.

In this chapter, some major design considerations are listed for chemiluminescence
detection in CE. The source of noise and factors affecting collection efﬁciency are also
discussed. Finally, the collection efﬁciency of the miniaturized CL detector is calculated
based on the solid angle from the light source, the chemiluminescence reactor (24). Our
goal was to construct a miniaturized CE-CL system with low noise, but high collection

efﬁciency.

26

2.1 Design Considerations for Chemiluminescence (CL) Detector

The design considerations for the CL detector are listed in ﬁgure 2.1.

 

Design Considerations

 

 

 

 

 

 

 

 

high sensitivity

10‘" “0136 level high collection efﬁciency

 

 

 

 

 

 

 

 

 

 

 

 

 

increase the solid angle

good grounding light tight for light collection

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

reduce the number of interfaces

. . miniaturization
make each Interface more efﬁCIent

 

 

 

 

 

 

Figure 2.1 Design considerations for chemiluminescence detector

27

2.2 Construction of the Miniaturized CE-CL Detector
2.2.1 Outer Body of the CL Detector

The detector shown in ﬁgure 2.2a & b consists of three parts. The left part in ﬁgure
2.2a is used to hold the body to the CE high voltage box. It also allows for visual
inspection of the capillary. The central part was used to ﬁx the separation and reaction
capillaries. The right part was used to mount the photomultiplier tube. The hole on the
right part is carefully machined to ﬁt the PMT power supply socket; a good ﬁt efﬁciently
blocks room light. Since the PMT casing diameter is bigger than this hole, the PMT
casing further block the room light. A l/2”-20 hole is drilled on the central part to allow
the mounting of a front-coated mirror which serves to reﬂect the light back to the PMT

tube for enhancement of collection efﬁciency.

2.2.2 Etching of the Separation Capillary

A Bunsen burner is used to burn off the polyimide coating about 5 cm from the tips of
twenty 50 um i.d. x 363 um o.d. x 51 cm or 50 um i.d. x 363 um o.d. x 41 cm long fused-
silica capillaries (Polymicro Technologies, Inc., product number TSP050375). The tips
were dipped into the liquid wax of a burning candle for sealing. This coating-removed
segment of the capillary was then carefully immersed into a 49 % HF solution inside a
fumeth for 72 minutes exactly; the etching rate is z 2.86 um / min. The rate of etching
was found to be constant during the etching process. The residual HF from the etched
portion was rinsed off with distilled water. The outside diameter of the etched segment
was 116.8 um, measured by a micrometer. The capillary tip was trimmed with a capillary

cutter so that the overall length of capillary is 40 cm or 50 cm.

28

/ 0.363 mm o.d. separation capillary

/ 0.380 mm i.d. x 1/16” PEEK Sleeve
[
3 -] / TEE dead volume: 566 nL

.1
'11
i

 

   
    
  

/- W“—W 0.05 M 11202
stainless steel ,‘\-". ‘
1 ~- phosphate
ground
:: o.
.. in“. T"
2.0 Fwd .35 power

Signal

 

reaction capillary

_ buffer

 

<-—— painted black

Figure 2.2a Front view: miniaturized post-column chemiluminescence detector in

capillary electrophoresis

29

/ 0.363 mm o.d. separation capillary

    
 
    
  
 
 

/ 0.380 mm i.d. x 1/16" PEEK sleeve

TEE dead volume: 566 nL

0.05 M H202

1.112" PMT phosphate

1.25"

1.375"

2.0"/

    
 

reaction capillary

3),))’))

buffer

3”)’}’$))D}

*— painted black

Figure 2.2b Side view: miniaturized post-column chemiluminescence detector
in capillary electrophoresis

30

50 um i.d. x 363 um o.d.
separation capillary

 

 

320 um id x 450 um o.d.
reaction capillary

 

 

 

 

Figure 2.2c Detailed view of the ﬂow cell in miniaturized CL detector in CB

31

2.2.3 Reaction Capillary

A 320 um i.d. x 435 um o.d. x 8 cm (Polymicro Technologies, Inc., product number
TSP320450) capillary was chosen as the reaction capillary to ﬁt with the etched segment
of the separation capillary. A Bunsen burner was used to burn off the polyimide coating

about 3 cm from one end for a length of 2 cm to make a window for CL detection.

2.2.4. Installation of the Miniaturized Chemiluminescence Detector
2.2.4.1 Installation of Chemiluminescence Flow Cell

A ferrule (Upchurch Scientiﬁc, product number F142) was pushed onto the cavity of
a screw (Upchurch Scientiﬁc, product number F300). A 500 um i.d. x l/16” o.d. PEEK
sleeve (Upchurch Scienticﬁc, product number F230) was then pushed through the ferrule
and tightened against a TEE connector (Upchurch Scientiﬁc, product number F727) to be
locked. The screw was set loosely on the TEE connector. The unetched end of the
separation capillary was inserted through the PEEK sleeve into the TEE connector. The
capillary was then pulled carefully until the beginning of the etched segment could be
seen from the side-arm of the TEE connector. A screw with ferrule-locked PEEK sleeve
was then inserted into the separation capillary and fastened onto the TEE connector to
hold the separation capillary in place. The next part turned out to be the most challenging
work for the overall installation process. Extreme care was required to insert the reaction
capillary into the etched segment of the separation capillary through the PEEK sleeve.

Finally the reaction capillary was tightened to the TEE connector.

32

2.2.4.2 Installation of CL Flow Cell to CL Outer Body

The PMT was mounted onto the right part in ﬁgure 2.2a The right part is then
mounted to the central part in ﬁgure 2.2a. The interface of the CL ﬂow cell described in
section 2.2.4.1 to the CL outer body is tricky. First, the outer body must be laid on a

bench-top, without using a hand to hold it since it can be broken easily. Secondly, the

phosphate buffer 0.05 M H202

separation capillary

 

 

 

 

+10kV

 

 

 

 

1\

safety interlock

 

 

 

 

00:33..

 

 

 

 

sample —-#2 2

 

 

 

 

 

 

 

 

 

 

 

|

Figure 2.3 Instrumental set-up for the miniaturized post-column CE-CL detection

33

reaction capillary was inserted through the two ‘41” holes on the top and bottom of the
outer body carefully. Then the cap for the ground buffer reservoir was inserted into the
reaction capillary and tightened to the outer body. The front-coated mirror was mounted
to the l/2”-20 hole on the outer body until it just touched the reaction capillary. Finally,
the left part in ﬁgure 2.2a was mounted onto the rest of the outer body, and the whole
installation process was complete after it is screwed onto the high voltage box. For an

experienced person, this process takes about 20 minutes.

2.3 Instrumentation for Chemiluminescence Detection in Capillary Electrophoresis

The instrumental set-up for the miniaturized post-column CE—CL detector is shown in
ﬁgure 2.3. The output cord of the high voltage power supply (Bertan Associates, Inc.,
Syosset, New York, USA, Model 205 A-lOP) was inserted through the top of the
Plexiglas into the buffer reservoir. The power supply was set at 10 kV for all the
electrokinetic injections and separations. The light generated in the chemiluminescence
reaction was detected by a PMT (Hamamatsu, 1P28A), powered by a High Voltage
Power Supply (Heath model EU-42A, Benton Harbor, M1) at 800 V at all times.
Photocurrent from the PMT is ampliﬁed by a current ampliﬁer (model 427A, Keithley
Instruments Inc., Cleveland, Ohio). The response time of the ampliﬁer was set to 3 ms at
an ampliﬁcation of 108 V/A for all the electrophoretic operations. Finally, the ampliﬁer
output was recorded by a microcomputer (33MHz, 486, Gateway, Inc., North Sioux, SD
57049) through a data acquisition board (National Instruments, Austin, TX) at a sampling

rate of 1000 points/s.

34

The instrumentation for CL detection is very simple, just a sample cell, PMT, and
data acquisition system. Since there is no light source required, a wavelength selection
device is not necessary. The CL signal is often very weak, so efﬁcient light collection is
critical. The sample cell is placed as close to the detector as possible. A front-coated
plane mirror is glued on the l/2”-20 threaded rod and then carefully adjusted into the
reaction capillary until a soft touch is felt. The front-coated mirror reﬂects light more

efﬁciently than a back-coated mirror in this application.

2.4 Electrical Noise

The sources of noise were not clearly understood early in this research. Our second
version of the CL detector was made of black Delrin (26), which picks up a signiﬁcant
amount of electrical noise associated with the power lines due to the poor shielding (see
ﬁgure 2.4). To study the source of this noise, an oscilloscope was used to monitor the
output from the current ampliﬁer. Surprisingly, there was a large AC noise of ~ 60 Hz.
The p-p amplitude of this AC noise was about 0.66 V, making it the dominant source of
noise in the CL detection. However, if the black Delrin body was covered with grounded
aluminum foil, this power line noise was eliminated. We postulated that the PMT tube
and housing is not only a good detector for light, but also a good antenna for receiving all
sorts of electrical interference. A poorly grounded CL detector body can pick up a
signiﬁcant amount of electrical noise associated with the power lines. A stainless steel
body was used for the ﬁnal version of the CL detector. The ground buffer reservoir was
glued with metal tape to remove the electrical interference as well. This modiﬁcation to

the early CE-CL detector turned out to be efﬁcient in shielding the ubiquitous AC

35

electrical interference from the power lines. A BNC coaxial cable was used to isolate the

electrical noise during the signal transport.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
0.9-777-7----7---iiimiiié’
0.8 1%“ W * ~ ,7 - ~7- -- -4-- 7 AﬁA- a- #4 A..- __-. _._ ._ a
A 0.7 3* —-— — -— — —- — — — .— ._ _._ _._ ._..__ ﬂ
3
0..-- - - I- I- I I I I I-
a
E
8 0.5 ~V 7 - 7 7 7 7 7 m V _-_L
'5
Z
8 0.4-- a 4 4 44444 +—------~—
'5
§
m 0.3 ' l
£ . l
11 L l I l 1
0.1 I 1 l l
0.0 g l 1 l l l l, 7 1 ll 1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 100 200 300 400 500 600 700 800 900 1000

Time (ms)

Figure 2.4 Electrical noise of the miniaturized chemiluminescence detector made of
black Delrin without proper grounding. PMT power supply voltage: 800V;
ampliﬁer: 108 V/A; sampling rate: 1000 points/s.

36

2.5 Noise Level of the Miniaturized CL Detector and the Conventional CL Detector

The leakage of room light into the CL detector increases the noise level dramatically,
but can be reduced by minimizing the number of joint interfaces while making each
interface more light-tight. In the ﬁrst version of the CL detector (25), we enclosed the
PMT in a large dark box (10” x 10” x 18”), which showed a serious leakage of room light
through the front door. We then explored the construction of the miniaturized
chemiluminescence detector using black Delrin (26) and stainless steel (27) as materials
for the outer casing.

The miniaturized CL detector solved many problems of the traditional
chemiluminescence detector. The noise levels between the miniaturized and the
traditional CL detector are compared in ﬁgure 2.5. The miniaturized CL detector reduced
the average noise level by a factor of 20 compared to the traditional dark-box CL
detector. The p-p amplitude of the miniaturized CL detector is also much lower than that
of the traditional CL detector. The lower trace in ﬁgure 2.5 is the noise of the
miniaturized CL detector (average noise level is 0.0044 V). The upper trace is the noise
of the conventional CL detector (average noise level is 0.0885 V). Clearly, the
miniaturized CL detector is considerably more light tight than the conventional CL
detector. I believe the major source of light leakage in the conventional detector is from
the door. In the miniaturized CL detector, the “door” does not exist. The only possibility
of light leakage is through the separation capillary; experimental results show little

difference of background signal when there is light or no light in the room.

37

0.5

 

0.4-» - - ~-

 

 

 

 

Noise (V)

0.2 .3. a. . . m. ..__. . _ I

 

 

 

 

 

 

   

 

0 100 200 300 800 900 1000

Figure 2.5 Comparison of noise level between the miniaturized CL (lower trace) and

conventional CL detector (upper trace) in capillary electrophoresis. PMT
power supply: 800 V; ampliﬁer: 108 WA; sampling rate: 1000 points/s.

38

2.6 Factors Affecting the Collection Efﬁciency

Our second version of CL detector was made of black Delrin. With this material, the
interior wall can absorb some of the chemiluminescence emission. Ideally, all the light
generated from the chemiluminescence reaction should go to the PMT and the rest should
be reﬂected back to the PMT for higher collection efﬁciency. To improve the reﬂection
of light, we replaced the black Delrin with machined stainless steel. The latter material,
when grounded properly, is very efﬁcient in shielding against electrical interference from
nearby sources.

The high collection efﬁciency can be achieved by increasing the solid angle for light
collection. Collection efﬁciency, in terms of solid angle, is deﬁned as the amount of light
reaching the optical element (28):

Q:A (2.1)

52
Here A is the effective or limiting area of the optical element as determined either by the
element itself or by an aperture stop, and S is its distance from the source. In the present
CL detector, A is the sensing area of the photocathode, and S is its distance from the tip
of the separation capillary. Equation 2.1 indicates that the solid angle can be increased by
placing a lens of high numerical aperture between the PMT and the reaction capillary or
by reducing the distance from the PMT to the reaction capillary. However, there is little
space inside the miniaturized CL detector for inserting a lens. On the other hand, the lens
needs a support, making the internal conﬁguration of the CL detector more sophisticated.
It is better to place the reaction capillary close to the PMT. With the above-mentioned
concepts, a miniaturized CL detector was constructed as Shown in ﬁgure 2.2a and ﬁgure

2.2b.

39

2.7 Collection Efﬁciency of the Miniaturized CL Detector

When the sample zone exits the separation capillary, the chemiluminescence reaction
starts immediately. The emitted light is collected by the PMT in the miniaturized CL
detector. It is assumed that the sample zone (~ 2 mm long) is a point source. A screen in
front of the photocathode is internally built to scatter light and enhance its collection.
Light reaching the screen is reﬂected to the photocathode. By calculating the solid angle
of the sample zone to the screen, the collection efﬁciency of the miniaturized CL detector
can be calculated. Figure 2.6 is the schematic diagram of the inner conﬁguration of the
CL detector. The screen surface is divided into four equal parts as labeled in ﬁgure 2.6.
Area A is further divided into two parts, area A1 and area A“. Q. and On denote the solid
angles of area A] and area A" respectively. Since the solid angle of the four equal parts is
identical, the overall solid angle, Q, spanned by the rectangular stray screen can be
calculated as

(2:4(52, +9”) (2.2)
Q; can be represented as a two dimensional integral over the surface area of Al:

o,=ﬂdo=”sine d6d¢
A1 A1

— ¢Od 19’ 'nl9 d0
‘ 0 ¢ 10 5‘ (2.3)

=I (fodtb [l—cosHI]

Here the upper limit (1)0 of the (1) integration and the upper limit 91 of the 0 integration are

deﬁned in ﬁgure 2.6. The solid angle Q" can be calculated in the same way.

40

 

 

 

Figure 2.6

 

 

 

>"V

sample zone

Three dimensional diagram for the calculation of solid angle

a is the length of PMT window, 25 mm; b is the width of the PMT
window, 12 mm. c is the distance from stray screen to the reaction
capillary, 4 mm.

41

Q” =J‘Jdﬂ =JJsin6d6 (105
I l

E 6
=I 2d¢ I ”sin6d6 (2'4)
¢0 0
2:
:J 2d¢ [l—cosHH]
¢o

The integration limits involved in equation 2.3 and 2.4 can be evaluated as:

 

 

(110 = tan—1(2) (2.5a)
a
_ -l a
19, — tan [2ccos¢] (2.5b)
6,, tan [2c Sin 4) J (2.5a)

Substituting expressions 2.3 - 2.5 into expression 2.2, we have

 

 

2:
(2:4 I §0d¢ [l—cosBI]+J 3001(1) jl—cosﬁuj (2.6)
Since tane = a (2 7)
I 2ccos¢ .
1
c0319, = (2.8)

‘/1+tan2 6,

Based on the same strategy, we can derive

l

‘/1+tan26”

Substituting equation 2.8 and 2.9 into equation 2.6, we have

 

cos 6” = (2.9)

42

 

 

 

 

aldi’ 1- +I2 ]d¢ 1— 1 >

O 2 t -1 2 b
\[I+[ “ J an [“ \/1+[2csin¢]
_ 2ccos¢ - ' ‘ 1

With a = 12 mm; b = 25 mm and c = 4 mm, the solid angle (2 was calculated to be 3.66

 

 

 

 

 

 

 

steradians, and the collection efﬁciency was thus found to be:

77 = 9— : 0.29(29%)
47:

The reﬂection plane mirror (10 mm diameter) was placed very close to the separation
capillary. The inner wall of the stainless steel detection chamber was carefully machined
to be smooth. If we assume the chemiluminescence toward the mirror is totally reﬂected
back to the PMT photocathode, the total collection efﬁciency of the miniaturized

chemiluminescence detector is doubled to 58%.

2.8 Conclusion
In this work, a miniaturized chemiluminescence detector was designed for capillary
zone electrophoresis. This detector signiﬁcantly reduces the electrical noise and leakage

of room light, and gives a collection efﬁciency of z 58%.

43

List of reference

1.

2.

10.

11.

12.

13.

14.

15.

l6.

17.

18.

Hara, T.; Kayama, S.; Nashida, H.; Nakajima, R. Anal. Sci. 1994, 10, 223-25.

Hara, T.; Kayama, S.; Nashida, H.; Nakajima, R. Anal. Sci. 1994, 10, 823-25.
Tsukagoshi, K.; Tanaka, A.; Nakajima, R.; Hara, T. Anal. Sci. 1996, 12, 525-28.
Tsukagoshi, K.; Akasaka, H.; Nakajima, R.; Hara, T. Chem. Lett. 1996, 467-468.
Wu, N.;' huie, C. W. J. Chromatogr. 1993, 634, 309-15.

Ruberto, M. A.; Grayeski, M. L. Anal. Chem. 1992, 64, 2758-62.

Ruberto, M. A.; Grayeski, M. L. J. Microcol. 1994, Sep. 6, 545-550.

Dadoo, R.; Colon, L. A.; Zare, R. N. J. High Resolution Chromatography, 1992, 15,
133.

Dadoo , R.; Seto, A. G.; Colon, L. A.; Zare, R. N.; Anal. Chem, 1994, 66, 303.
Zhao, J. Y.; Labba, J .; Dovichi, N. J. J. of Microcolumn Separation, 1993, 5, 331.
Huang, B.; Li, J ., and Cheng, K. Sepu, 1995, 13, 430.

Huang, R; Li, J. J.; Cheng, J. K. Chemical Journal of Chinese Universities, 1996,
17, 528-530.

Huang, Bo; Li, Jian-jun; Zhang, Le; Cheng, Jie-ke, Anal Chem, 1996, 68, 2366-
2369.

Dasgupta, P. K.; Genfa, Z.; Li, J.; Boring, C. B.;Jambunathan, S., and Al-Horr, R.,
Anal. Chem, 1999, 71, 1400.

Tsukagoshi, K.; Fujimura, S. and Nakajima, R. Anal. Chem, 1997, 13, 279.

Lee, Y.-T.; Whang, C.-W. J. Chromatogr., A, 1997, 771 , 379.

Liao, S.-Y.; Whang, C.-T., J. High. Resolut. Chromatogr., 1995, 18, 667.

Liao, S.-Y.; Whang, C.- Y., J. Chromatogr., 1996, 736, 247.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Hara, T.; Yokogi, J. ; Okamura, S. J. Chromatography, 1993, A652, 361.

Gilman, S. D.; Silverman, C. E.; Ewing, A. G. J. of Mirocolumn Separation, 1994, 6,
97.

Hara, T.; Nishida, H.; Kayama, S.; Nakajima, R. Bull. Chem. SOC. Japan, 1994,
67,1193.

Gilman, S. D.; Silverman, C. E.; Ewing, A.G. J. Microcolumn Separation, 1994, 6,
97.

Baeyens, W.R.G.; Ling, B. L.; Irnai, K.; Calokerino, A.C.; Schulman, S.G. J.
Microcolumn Separation, 1994, 6, 195.

Ingle, J. D.; Crouch, S. R., Spectrochemical Analysis, Prentice Hall, Englewood
Cliffs, New Jersey, 1988, p 479-485.

Peng, C. H.; Crouch, S. R., “High Sensitivity Post-Column Chemiluminescence
Detection in Capillary Zone Electrophoresis”, presented at the ANACHEM/SAS
Meeting, October 2, 1996, Dearbom, Michigan.

Peng, C. H.; Crouch, S. R., “Miniaturized Post-Column Chemiluminescence
Detector in Capillary Zone Electrophoresis”, presented at the ANACHEM/SAS
Meeting, Nov. 3, 1997, Dearbom, Michigan.

Peng, C. H.; Crouch, S. R. “Miniaturized post-column chemiluminescence detector
in capillary zone electrophoresis”. Manuscript submitted.

Ingle, J. D.; Crouch, S. R., Spectrochemical Analysis, Prentice Hall, Englewood

Cliffs, New Jersey, 1988, p 53.

45

Chapter 3

Detection of Transition Metal Ions Using Miniaturized Chemiluminescence
Detector in Capillary Zone Electrophoresis

Chemiluminescence (CL) has been used as a sensitive method of detection in
capillary electrophoresis (CE) since the early 90’s when Hara et al.(l) and Dadoo et al.(2) .
reported their pioneering work. Luminol is the most popular CL reagent in CL detection.
Detection of metal ions in CE has been explored extensively using several kinds of
detection as described in chapter 1. In this work, metals are separated by CE and detected
by chemiluminescence because they catalyze the luminol/H202 chemiluminescence. The
reaction scheme of luminol/H202 chemiluminescence is described in chapter 1.

This chapter describes the separation of several transition metal cations. The
importance of daily capillary conditioning is emphasized. It was found that a regular
conditioning procedure in CB (3), is not enough for the separation of transition metal
ions. Because of this, we developed a special conditioning procedure that involves 1 hour
of rinsing with 0.1 M NaOH and 12 hours of rinsing with 0.1 M HCl, based on the work
of Lu and Cassidy (4).

The lifetime of luminol/H202 chemiluminescence at different pH values was
measured to aid in the pH optimization of the post-column CL reagents. Some important
factors such as the pH of the separation buffer, the buffer composition and adsorption are
discussed. Ten consecutive injections gives a reproducibility of 8 - 18% under various
conditions (in term of RSD for peak height), which is poorer than that for on-column

detection (about 2 - 10%) (3, 5).

46

3.1. Preparation of Stock Solutions
3.1.1 Preparation of 0.010 M Luminol Stock Solution (pH = 12.00)

Weigh accurately 1.7716 g (0.010 mole) 3-Aminophthalhydrazide (luminol) (Aldrich
Chem. Co., product # 12,307-2) and dissolve in a 1000 mL beaker. About 990 mL of
deionized distilled water is added to the beaker. Lumniol tends to form small chunks in
water; the ultrasonic bath is turned on to crack these small chunks. The solubility of
luminol is very low in acidic solutions but much higher in basic solutions. A 5 M NaOH
solution is added to adjust the luminol solution to pH 12.00. The luminol solution is then
ﬁltered through a 0.45 um pore Size membrane ﬁlter (Millipore Inc.). Finally, the luminol
solution is diluted to 1000 mL with deionized distilled water and transfered into a 1L
glass bottle. The bottle cap must be tightened to prevent the CO2 from entering.
Introduction of C02 can change the pH of the luminol stock solution dramatically over
long periods of time. Finally, the luminol solution is stored in the refrigerator. The pH of

the luminol solution was found to be stable for at least one year.

3.1.2 Preparation of 0.050 M H202
Pipet 491 [AL of 30.9 % H202 (density 1.12 g/mL, J. T. Baker, product # 2186-01) into
a 100 mL volumetric ﬂask, and dilute to the mark with deionized, distilled water. All

0.050 M H202 solutions are prepared daily following this procedure.

3.1.3 Preparation of Separation Buffer (0.01 M acetic acid, 2 x 10’4 M luminol)
Pipet 1 mL of 1 M acetic acid, various amounts of 2-Hydroxyisobutyric acid (HIBA)

(Aldrich Chem. Co., product # 32,359-4) if needed, and 2 mL of 0.010 M luminol stock

47

solution, then dilute to 100 mL. The pH of the buffer is adjusted to the desired value with
2 M or 5 M NaOH. The buffer solution is stored in refrigerator since the luminol solution

is not stable at room temperature.

3.1.4 Preparation of Sample Stock Solution

Metal solutions were prepared from metal nitrate salts, Co(NO3)2 . 6H20 (Spectrum
Quality Inc.), Cu(NO3)2. 2.5 H20 (J. T. Baker Inc.), Cr(N03)3 . 9H20 (Spectrum Quality
Inc.) without further puriﬁcation. All sample solutions at different concentrations were

serially diluted daily and stored in plastic bottles.

3.2 Instrumentation
3.2.1 Instrumentation for Lifetime Measurement

The experimental set-up for measuring the CL lifetime is shown in ﬁgure 3.1. The
light generated by the chemiluminescence reaction is detected by the PMT (Hamamatsu,
1P28A PMT Photomultiplier tube) placed about 1 cm in front of a 1 mL glass vial. The
PMT is powered at 600 V (Heath Power Supply, model EU-42A). The PMT photocurrent
is ampliﬁed by a current ampliﬁer and then recorded by a computer at a sampling rate of
1000 points/s.

A solution of H202 (200 11L of 0.025 M, various pH values) was injected into a 20 UL
solution of separation buffer (0.01 M acetic acid and 3 x 104 M luminol, pH = 4.65). The
pH of the mixed solution remained unchanged due to the large volume of H202. For
precise CL measurement, an automatic syringe is often employed to achieve a more

reproducible injection speed and volume (6). Ideally, the mixing process should be fast

48

enough that the emission vs. time curve is affected solely by the rate of

chemiluminescence reaction rather than by the rate of reagent mixing.

200 “L H202 at various pH

I
M (- PMT: 600 V

20 uL 0.3 mM luminol

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4'\

I .

i
“IiWWEet—ig

1000 points/s

 

"5

 

 

 

 

Figure 3.1 Instrument for lifetime measurement of luminol-H202 chemiluminescence
Ampliﬁer setting: 108 WA; rise time: 0.1 ms; sampling rate: 1000 points/s

49

3.2.2 Instrumentation for Electrophoretic Separation

The instrumentation used for CE separation is described in chapter 2.

3.3 Procedures
3.3.1 Capillary Conditioning

Regular conditioning for fused silica capillaries was done by ﬂushing the capillary
with 0.1 M NaOH for about one-half hour, and then with 0.1 M HCl for another one-half
hour. However, this is not enough for the separation of transition metal ions due to their
strong adsorption to the capillary wall. Thus, we developed a special conditioning
procedure. The capillary was ﬁrst ﬂushed with 0.1 M NaOH for 1 hour at a ﬂow rate of
10 tLIJmin, then ﬂushed overnight with 0.1 M HCl at a ﬂow rate of 3 ulJmin. The
capillary must be conditioned by this procedure after each working day or the
reproducibility of the peaks can be very poor. Figure 3.2a and ﬁgure 3.2b Show the
Electropherograms of the ﬁrst run and the ninth run of 1.0 x 10'8 M Co“, 1.0 x 10'6 M
Cr“, 1.0 x 10'6 M Cu2+ under identical conditions. The ﬁrst run shows a higher peak but
a longer migration time. The peak area and migration time for the same sample under the
same separation conditions decreases during a series of electrophoresis runs. This may
result from the strong adsorption of the transition metal cations to the capillary walls (4).
During the electrophoretic runs, the active sites on the capillary wall are continuously
covered by the transition metal cations, which can Slow down the electroosmotic ﬂow and
reduce the amount of sample injected into the separation capillary. This explains the

decreased peak height during the separation of the same sample under identical

50

 

 

 

3.0
2
2.5 —
g peak 1, Co2+ (1.0 x 10‘8 M)
a peak 2, Cr3+ (1.0 x 1045 M)
2+ -6
E 2.0 _ peak 3, Cu (1.0 x 10 M)
E
8
s 1
,5 1.5 -
8
t:
8
8
_C.’
E 1.0 -
:51 3
E
.C:
U
0.5 -

 

 

    

 

0.0 l I I l l I I 1 T T

300 350 400 450 500 550 600 650 700 750
Time (s)

Figure 3.2a Electropherogram of the ﬁrst run of 1.0 x 10'8 M Co“, 1.0 x 10'6 M Cr“,
1.0 x 1045 M Cu“.
Sample: 1.0 x 10‘8 M Co“, 1.0 x 10‘6 M Cr“, 1.0 x 10*5 M Cu2+ in
deionized distilled water; buffer: 2.0 x 104 M luminol, 0.01 M NaAc/HAC,
0.005 M HIBA, pH = 3.53; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 uUmin; electrophoretic voltage: 10 kV;
injection voltage: 10 kV; injection time: 10 s; fused silica capillary: 50 um x
363 um o.d. x 50 cm.

51

3.0

 

l"
1
1
1

A.__.__.___.._.___—_ _.______._ ______—__.— _._— . ._—.1

peak 1, Co2+ (1.0 x 10'8 M)
peak 2, Cr3+ (1.0 x 10“ M)
peak 3, Cu2+ (1.0 x 10‘6 M)

 

l"
o
1
1
1

 

Chemiluminescence Intensity (arbitrary unit)
'2 II.
p 1
y 1
1
1
1
|
1 1

I 3
054‘ ; -A #

 

 

 

 

 

   

0.01 l T I l I I I I I I I
300 350 400 450 500 550 600 650 700 750

 

Time (s)

Figure 3.2b Electropherogram of the ninth run of 1.0 x 10'8 M C02+, 1.0 x 10’6 M Cr”,
1.0 x 10'6 M Cu’r.
Sample: 1.0 x 10'8 M Co“, 1.0 x 10*5 M Cr”, 1.0 x 10‘6 M Cu2+ in
deionized distilled water; buffer: 2.0 x 10'4 M luminol, 0.01 M NaAc/HAc,
0.005 M HIBA, pH = 3.53; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 III/min; electrophoretic voltage: 10 kV;
injection voltage: 10 kV; injection time: 10 s; fused silica capillary: 50 um x
363 um o.d. x 50 cm.

52

conditions as shown in ﬁgures 3.2a and ﬁgure 3.2b. Most likely, the migration time
decreases because the capillary wall becomes less negatively charged due to the
accumulation of the cations. This reduces the adsorption of metal cations to some extent
and allows the sample zones to migrate faster than in the early runs. The increase in the
electrophoretic is larger than the decrease in the electroosmotic ﬂow, making the overall
rate of electrophoretic ﬂow faster. The exact mechanism for the adsorption of metals was
not studied in this work, but experimental results Show that strict conditioning is crucial

in order to achieve reproducible separation of transition metal cations.

3.3.2 Sample Injection

Electrokinetic injection was used for all the CE separations in this chapter. The
capillary and the Pt electrode were placed into the sample solution, and a 10 kV injection
voltage was applied for 5 - 10 seconds. Then the capillary is placed into the buffer,

solution and electrophoresis begun.

3.4 Results and Discussion

3.4.1 Lifetime and Intensity of Luminol-H202 Chemiluminescence at Various pH values
Figure 3.3 indicates the lifetime of chemiluminescence is less than 60 ms between pH

12.50 and 13.00. The chemiluminescence reaches a maximum at pH 12.50. The CL

emission is ﬁnished within 300 ms under any solution pH as shown in ﬁgure 3.3. For CL

detection of CE separations, it is important to calculate how long the sample zone is

observed by the PMT. The linear ﬂow rate of the CL reagents is 1.24 mm/s, given the

volume ﬂow rate of the post-column CL reagent (6 uUmin) and the internal diameter of

53

 

6.0

5.0 -

3.0 1

2.0 -

Chemiluminescence Intensity (arbitrary unit)

 

 

1 A: pH = 12.50
B: pH = 13.00

C: pH = 12.00

\ D: pH = 11.50
E: pH = 11.00

3 F: pH = 10.50

 

 

Figure 3.3

 

 

Lifetime of luminol-H202 chemiluminescence at various pH values
PMT power supply: 600 V; Ampliﬁer: 108 WA; sampling rate: 1000
points/s

54

7.0

 

Chemiluminescence Intensity (arbitrary unit

 

 

 

0.0 T I I I I
10.50 11.00 11.50 12.00 12.50 13.00 13.50

 

pH

Figure 3.4 Intensity of luminol-H202 chemiluminescence at various pH values
PMT power supply: 600 V; Ampliﬁer: 108 WA; sampling rate: 1000 points/s
Reagent A: 200 11L 0.025 M H202 at various pH ---> Reagent B: 20 1.1L of 3.0
x 10“4 M 1uminol,0.01 M NaAc/HAc, pH: 4.65

55

the reaction capillary (320 um i.d.). The exposure time of the sample zone in front of the
PMT is approximately 3 seconds; this is sufﬁcient for completion of the CL reaction
catalyzed by the transition metal ions.

The maximum emission at various pH values is shown in ﬁgure 3.4. The CL emission
increases with rising pH, and reaches a maximum at pH 12.50, and decreases in more
basic solutions.3.4.2 Selection of Mixing Mode

For the separation of transition metal ions in CB, there are three possible modes of
mixing, as described in ﬁgure 3.5, ﬁgure 3.6 and ﬁgure 3.7. These are studied in order to
select an optimum mode. The transition metal ions are best separated in acidic media, but

detection is best performed in basic media as shown in ﬁgure 3.4.

3.4.2.1 Mixing Mode 1: Mix H202 in Separation Buffer

In mode 1 as shown in ﬁgure 3.5, 0.05 M H202 is added into the separation buffer
(0.01 M NaAc/HAC, pH = 4.50). This mode of mixing produces severe decomposition of
H202. The Pt electrode turned out to be a good catalyst for the decomposition of H202,
which generates bubbles around the Pt electrode surface. These bubbles frequently disrupt
the high electric ﬁeld. Decomposition of H202 is much faster when the 10 kV high
voltage is applied across the separation capillary. Routine electrophoresis is not feasible

with this mixing mode. Therefore, we gave up this mode of mixing early in this work.

56

separation capillary

10 mM luminol
phosphate buffer
pH: 12.00

    

high voltage

 

reaction capillary—a

0.01 M M NaAc/HAc buffer
0.050 M H202

Figure 3.5 Mixing mode 1 for chemiluminescence detection in capillary
electrophoresis

3.4.2.2 Mixing Mode 2: Mixing Luminol with H202 as Post-column CL Reagents

In mode 2 as shown in ﬁgure 3.6, 0.05 M H202 is mixed online with an equal amount
of 0.01 M luminol using a syringe pump (Harvard Apparatus, model 22). The lifetime of
the luminol/H202 chemiluminescence is less than 400 ms at any pH value as shown in
ﬁgure 3.3. It takes about 30 minutes for the merged stream of CL reagents to reach the
post-colunm ﬂow cell for detection, thus signiﬁcant amount of luminol/H202
chemiluminescence can occur within the Teﬂon tube and the mixing TEE before the
separated metals can catalyze the luminol/H202 chemiluminescence. The experimental
result showed poor sensitivity. The detection limit for cobalt, which has the greatest
catalytic effect on the luminol/H202 chemiluminescence is about 1045 M. Mode 2 did not
demonstrate the high sensitivity of chemiluminescence detection. Consequently, no

further effort was attempted for mixing mode 2.

57

separation capillary

0.010 M luminol
0.1 M phosphate buffer
pH = 12.50

reaction capillary \

0.050 M H202

 

0.010 M NaAc/HAc buffer

Figure 3.6 Mixing mode 2 for chemiluminescence detection in capillary
electrophoresis

3.4.2.3. Mixing Mode 3: Luminol in the Separation Buffer

In mode 3 as shown in ﬁgure 3.7, luminol is added into the separation buffer.
peroxide and phosphate buffer are mixed from two independent streams and merged as
one stream ﬂowing down the TEE connector to react with the eluted transition metal ions.
Unfortunately, the solubility of luminol is poor at the low pH necessary for good
separation of the transition metal ions. A solubility test for luminol in the buffer at
different pH values was done and is listed in Table 3.1. The buffer containing 3.0 x 104
M luminol is not stable in the pH 3.50 - 5.00 range. The buffer containing 0.01 M acetic
acid and 2.0 x 10'4 M luminol did not Show a luminol precipitate at pH 3.50 - 5.00 one
month after preparation and was chosen for all the CE separations. The sensitivity of this

mode is much higher than that of mixing mode 1 and mixing mode 2.

58

Table 3.1 Solubility of luminol in buffers of different pH values

 

 

 

Buffer pH = 3.50 pH = 4.50 pH = 5.00
0.01 M acetic acid, soluble soluble soluble
1.0 x 10'4 M luminol
0.01 M acetic acid, soluble soluble soluble

2.0 x 10“ M luminol

 

 

 

 

 

 

0.01 M acetic acid, precipitate precipitate precipitate in one or
3.0 x 104 M luminol two hours
separation capillary
phosphate buffer
pH: 12.50

 

   

high voltage
reaction capillary

 

2.0 x 10'4 M luminol
0.010 M NaAc/HAc buffer
pH: 3.50 ~ 5.00

Figure 3.7 Mixing mode 3: luminol in the separation buffer

59

\

0.05 M H202

 

3.4.3 Optimization of pH for Post-Column Chemiluminescence Detection

The luminol CL reaction is optimized at alkaline conditions, depending upon the
catalyst. The transition metal ions are best separated by CE at low pH. Consequently, a
post-column pH change is required when applying the luminol CL reaction for detection
in CE. Generally, the mixing process reduces the number of theoretical plates (7). A
smaller internal diameter capillary enhances the separation efﬁciency, and allows easier
and faster pH switching. In our work, a 50 um i.d. fused silica capillary was selected for
all the electrophoretic operations; a 320 um i.d. x 435 um o.d. fused silica capillary was
chosen as the reaction capillary. The volume ﬂow rate of the CL reagents is about 41
times that of the separation capillary under the same linear ﬂow rate. Considering the
strong buffer capacity of the post-column CL reagents, it should be possible to switch
rapidly to the more basic pH.

In a ﬂow system like CE with CL detection, the post-column pH switching occurs
through diffusion. When the sample zone exit the separation capillary, the pH of the zone
is quickly raised toward that of the post-column CL reagents. The rate of this pH ramping
is determined by many factors such as the pH, the buffer capacity and viscosity of both
the separation buffer and post-column CL reagent. For a rapid pH switching, a low
concentration of separation buffer and a CL reagent with high buffer capacity is essential.
Only a few sample injections were performed each day for the freshly conditioned
capillary due to the limited loading capacity of the fused silica capillary. An ideal buffer
should have strong buffer capacity and Should not interact with the sample.

In the detection of transition metals in CB, one of the most important factors is the pH

of the post-column CL reagents. The post-column CL reagent comes from two separate

60

streams of 0.05 M H202 and phosphate buffer. The H202 and the basic phosphate should
be mixed on-line with the phosphate buffer (pH 11.50 ~ 13.00) to reach the optimum pH
range for post-column CL detection, since H202 decomposes Slowly in alkaline solution.
Figure 3.8 shows the various combinations and the resulting pH of the mixtures.

The peak heights for 1.0 x10’8 M Co2+ from three electrokinetic injections are plotted
versus pH in ﬁgure 3.9. As the buffer solution exits the separation capillary, there is a
large pH change from pH 4.06 toward that of the post-column CL reagents. The peak
height increases with rising pH of the post-column CL reagent, since the pH of the
sample zone has not reached the optimized pH value (~12.5) for maximum
chemiluminescence emission.

The peak heights for Co2+ and Cr3+ from three electrokinetic injections are plotted
versus pH in ﬁgures 3.10 and 3.11 respectively. The pH (4.74) of the separation buffer
shown in ﬁgure 3.10 and 3.11 is higher than that of the buffer in ﬁgure 3.9 (pH 4.06). The
buffer capacity of the post-column CL reagents is also higher due to the higher
concentration of phosphate (0.1 M), and thus gives a more basic pH after switching. The
optimized pH for maximum chemiluminescence is reached earlier in ﬁgure 3.10 than in
ﬁgure 3.9. This explains the bending of the curve in ﬁgure 3.10.

Figures 3.10 and 3.11 Show similar trends for Co2+ and Cr“. The highest sensitivity is
achieved around pH 12.00 for both Co2+ and Cr“. The optimized pH for the post column
CL detection of Co“ and Cr3+ is somewhere between 12.00 and 13.00. If more pH values
were selected for pH optimization, the peak height v.s. pH may give a sharper maximum

and allow a more precise choice of pH.

61

pH 10.98

0.05 M H202

 

 

pH 11.90

pH 12.34

pH 11.29

pH 12.06

 

 

0.05 M phosphate, pH = 11.50

005MH202

 

0.05 M phosphate, pH = 12.50

0.05 M H202

 

1
|

0.05 M phosphate, pH = 13.00

0.05 M H 202

 

1
1

0.1 M phosphate, pH = 11.50

0.05 M H 202

 

 

pH 12.49

 

 

0.1 M phosphate, pH = 12.50

0.05 M H 202

 

 

 

 

0.1 M phosphate, pH = 13.00

Figure 3.8 Final pH value of the post-column CL reagents

62

 

1.80

1.60 a

1.40 ~-

l.00 -

Peak height (arbitrary unit)

1.20

0.80 . -

0.60

 

 

 

 

0.00

Figure 3.9

I I I T I

10.80 11.00 1 1.20 11.40 11.60 11.80 12.00 12.20 12.40 12.60

pH of post-column CL reagents

Peak height of 1.0 x 10'8 M Co2+ at various pH values of CL reagent.
Sample: 1.0 x 10'8 M Co2+ in deionized distilled water; separation buffer: 2
x 10“1 M luminol, 0.010 M NaAc/HAc, pH: 4.06; post-column CL reagents:
0.025 M H202, 0.025 M phosphate, pH = 10.98, 11.90, 12.34; ﬂow rate: 6
1.1L ls; electrophoretic current: 0.8 uA; PMT voltage: 800 V; ampliﬁer: 108
WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV; sample injection:
10 s at 10 kV; separation capillary: 50 um i.d. 363 um o.d. x 50 cm long.

63

 

7.00

6.00 4

5.00 -

4.00 -

3.00 4

Peak height (arbitrary unit)

2.00 .

1.00 ~

 

0.00

 

 

11.20

Figure 3.10

T 1' I I

11.40 1 1.60 11.80 12.00 12.20 12.40

pH of post-column CL reagents

Peak height of 1.0 x 10’9 M Co2+ at various pH values of CL reagent.
Sample: 1.0 x 10'9 M Co2+ and 5.0 x 10' M Cr3+ in deionized distilled
water; separation buffer: 2 x 104 M luminol, 0.010 M NaAc/HAc, pH: 4.74;
post-column CL reagents: 0.025 M H202, 0.050 M phosphate, pH = 11.29,
12.06, 12.49; ﬂow rate: 6 1.1L ls; electrophoretic current: 2.2 [J.A; PMT
voltage: 800 V; ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation
voltage: 10 kV; sample injection: 10 s at 10 kV; separation capillary: 50 um
i.d. 363 um o.d. x 40 cm long.

12.60

Peak height (arbitrary unit)

Figure 3.11 Peak height of 5.0 x 10'7 M Cr3+ at various pH values of CL reagent

7.00

 

6.00 -

5.00 a

4.00 a

3.00 -

2.00 ~

1.00 -.

 

0.00

 

 

11.20

11.40

11.60 11.80 12.00

pH of post-column CL reagents

12.20

12.40

12.60

Sample: 1.0 x 10'9 M Co2+ and 5.0 x 10’7 M Cr3+ in deionized distilled
water; separation buffer: 2 x 10“ M luminol, 0.010 M NaAc/HAc, pH: 4.74;
post-column CL reagents: 0.025 M 11202, 0.050 M phosphate, pH = 11.29,
12.06, 12.49; ﬂow rate: 6 1.1L ls; electrophoretic current: 2.2 11A; PMT
voltage: 800 V; ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation
voltage: 10 kV; sample injection: 10 s at 10 kV; separation capillary: 50 um
i.d. 363 um o.d. x 40 cm long.

65

3.4.4 Optimization of Separation Buffer pH

The selectivity of the metal ion separation can be predicted from the equivalent ionic
conductivities (8), A.;, which are directly related to the mobilities, 11,-, of the ions:

HI = 9»; / F (3-1)
where F is the Faraday constant. The velocity of migration for a certain ion Valpp is given
by:

Vapp = (lli + HEOP) E (32)
Here 1.1, is the electrophoretic mobility of the ion, 11501: is the electroosmotic ﬂow mobility
and E is the electric ﬁeld. Based on the mobility of Co2*(5.7 x 10“ mst), Cr3+(6.9 x 10'
4 mZN S), and Cu2+(5.6 x 104 m2/Vs), the predicted elution order should be Cr3+ ---> Coz+
---> Cuz”. However, the observed elution order was Co2+ ---> Cr3+ ---> Cu2+, due to the
adsorption of Cr3+ to the capillary wall and to complexation of the metals by the buffer
anions. Group IA metal ions like Li+, Na”, K+ and Rb”, exhibit no signs of adsorption, so
their peak proﬁles are always fronting (9).

Adsorption of transition metal ions to the capillary wall is one of the most difﬁcult
problems. Although there are several ways to minimize adsorption, such as lowering the
pH of the buffer, and addition of a-hydroxyisobutyric acid (HIBA) into the buffer. None
of these is a perfect solution. The best way is to use a coated fused silica capillary.
Actually we have explored this idea, but the result was not very successful due to our
inability to make an inert coating. Coated capillaries are discussed further in chapter 6.

It is very important to establish the optimum pH for the separation buffer in CB since
pH affects the electroosmotic ﬂow rate and the number of surface negative charges

dramatically as described in chapter 1. Maintenance of a constant pH is essential for

66

 

8.00

7.00“

6.00 a

8

Peak Height (arbitrary unit)
5» A
8 8

2.00 a

1.00 *1

 

 

 

 

0.00
3.50

3.75 4.00 4.25 4.50 4.75 5.00
pH

Figure 3.12 Peak height of 1.0 x 10'8 M Co2+ at various separation buffer pH values

buffer: 2.0 x 10“1 M luminol, 0.01 M NaAc/HAc; post-column CL reagent:
0.05 M H202, 0.1 M phosphate, ﬁnal pH = 12.49; ﬂow rate: 6 ill/min;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time: 10
s; capillary: 50 um x 363 um o.d. x 40 cm.

67

500

 

450 -

Migration Time (s)

300‘

 

250

350 - .-

 

 

 

3.50

Figure 3.13

3.75 4.00 4.25 4.50 4.75 5.00
pH

Migration time of 1.0 x 10'8 M Co2+ at various separation pH values

buffer: 2.0 x 104 M luminol, 0.01 M NaAc/HAc; post-column CL reagent:
0.05 M H202, 0.1 M phosphate, ﬁnal pH = 12.49; ﬂow rate: 6 uIJmin;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time: 10
s; capillary: 50 um x 363 um o.d. x 40 cm.

68

assuring reproducible migration times. This is particularly important for the separation of
transition metal ions since most of the transition metal ions undergo hydrolysis at pH
values greater than 5 (10, 11). Transition metal ions Show an increasing adsorption to the
inner wall of the capillary at higher pH; therefore, their separation is accomplished in
acidic solution (pH< 5) (12).

Figure 3.12 shows the peak height of 1.0 x 10'8 M Co2+ at various separation pH
values. Note that the peak height increases with increasing buffer pH. With higher buffer
pH, post-column pH switching can be accomplished faster; therefore, more light is
generated within the time of CL detection (28 ~ 3s). Figure 3.13 shows the migration time
of the sample zone at various buffer pH values. The migration time is decreased at higher

pH due to the increased electroosmotic ﬂow rate.

3.4.5 Signal Enhancement by Complexation

The adsorption of transition metals to the capillary wall is one of the major factors
that affect migration time, peak height, etc. Even at pH values lower than 4, adsorption
cannot be ignored, particularly at high metal ion concentrations. Certain applications
(e.g., separation of transition metal cations), require the addition of a weak complexing
agent such as hydroxyisobutric acid (HIBA) (13, 14), citric acid (15) or lactic acid (16) to
the background electrolyte in order to provide greater differences in mobility.

We studied the stability of the buffer containing 2 x 104 M luminol and various
amounts of HIBA. The results shown in table 3.2 indicate that the buffer (pH = 3.80) with
0.006 - 0.010 M HIBA showed poor stability. The precipitation of luminol in the buffer

causes a day-to-day signal variation.

69

Table 3.2 Stability of the buffers containing various amount of HIBA at pH 3.80

 

 

 

 

 

 

 

 

Buffer Preparation date Date of Stability period
luminol
precipitation
2 x 10“1 M luminol, 11/1/1999 not seen > 1 month
0.010 M NaAc/HAc
0 HIBA
2 x 101M luminol, 11/1/1999 not seen > lmonth
0.010 M NaAc/HAC
0.002 M HIBA
2 x 104 M luminol, 11/1/1999 not seen > 1 month
0.010 M NaAc/HAc
0.004 M HIBA
2 x 104 M luminol, 11/1/1999 11/4/1999 < 3 days
0.010 M NaAc/HAc
0.006 M HIBA
2 x 10“ M luminol, 11/1/1999 11/2/1999 < lday
0.010 M NaAc/HAc
0.008 M HIBA
2 x 10’4 M luminol, 11/1/1999 11/2/1999 < 1 day

0.010 M NaAc/HAC

0.01 M HIBA

 

 

 

 

 

 

55
A
3::
t:
3
E 4‘“
is
V
>~.
3::
VJ
t:
8
.s 3-
d)
0
t:
0
0
Ch
Q)
.g
2. 2:
d)
..C:
U

 

Figure 3.14

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time (s)

Peak height and migration time of C02+(1.0 x 10'8 M), 0.000 M HIBA
buffer: 2.0 x 104 M luminol, 0.01 M NaAc/HAc, 0.000 M HIBA,
separation pH = 3.78; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 III/min; electrophoretic voltage: 10
kV; injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363
um o.d. x 40 cm.

71

 

Chemiluminescence Intensity (arbitrary unit)

 

250

Figure 3.15

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l ‘t ,1 ,.,1 . h’. M“

.h. . -4..“r...o A: ‘5. .--.l. ‘—. ‘ a-“M-.bn_.d.- ‘AJ “1“ “Ad

I I

255 260 265 270 275 280 285 290 295 300
Time (s)

Peak height and migration time of Co2+ (1.0 x 10’8 M) with 0.002 M HIBA
buffer: 2.0 x 10'4 M luminol, 0.01 M NaAc/HAc, 0.002 M HIBA,
separation pH = 3.79; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 III/min; electrophoretic voltage: 10
kV; injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363
um o.d. x 40 cm.

72

 

Chemiluminescence Intensity (arbitrary unit)

 

O ~L:-_‘.n‘

270

Figure 3.16

 

 

 

 

 

 

 

 

 

 

 

' _ l ‘
I ,1. ' 1‘
ﬂ. U}, .l‘W‘i’w“. kibﬂ K». R, “A

’-|‘_.--4~-~.—Ao J l-‘A.-A“4 t

I I I I

280 290 300 310 320
Time (s)

Peak height and migration time of Co2+ (1.0 x 10'8 M) at 0.004 M HIBA
buffer: 2.0 x 104 M luminol, 0.01 M NaAc/HAc, 0.004 M HIBA,
separation pH = 3.80; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 uljmin; electrophoretic voltage: 10
kV; injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363
um o.d. x 40 cm.

73

 

Chemiluminescence Intensity (arbitrary unit)

 

270

Figure 3.17

A--.~-EA..‘_.. a A‘A. .L‘L‘

_._-”W.
J—

 

 

 

 

 

 

 

 

 

. ’1
‘. U}. .1‘ whit-0’11. hr t‘.‘ ‘. n. ." rt; 3“

:-

I I T I

280 290 300 3 10 320
Time (8)

Peak height and migration time of Co2+ (1.0 x 10'8 M) at 0.006 M HIBA
buffer: 2.0 x 10“ M luminol, 0.01 M NaAc/HAc, 0.006 M HIBA,
separation pH = 3.78; post-column CL reagent: 0.05 M H202, 0.1 M
phosphate, pH = 12.49; ﬂow rate: 6 uUmin; electrophoretic voltage: 10
kV; injection voltage: 10 kV; injection time: 10 s; capillary: 50 um x 363
um o.d. x 40 cm.

74

The apparent mobilities of metal ions are affected by the complexation of the metal
cations and HIBA. The effect of HIBA on the migration time and peak height was
investigated with a buffer (0.01 M NAc/HAc, 2 x 10‘4 M luminol at pH 3.80) containing
various amounts of HIBA. As the HIBA concentration is increased, the apparent
mobilities of the metal ions are decreased, which leads to a longer migration time as
shown in ﬁgure 3.14 - ﬁgure 3.17. The peak widths are also increased due to the longer
time for diffusion and the distribution of mobilities due to the differing degrees of
complexation. The peak height increases with increasing HIBA in the buffer. At high
concentrations of HIBA (0.006 M), the peak height decreases due to the inhibition of
chemiluminescence as the result of the strong complexation of the metal cations and
HIBA.

HIBA can form a weak complex with the transition metal ions, which reduces the
adsorption of metal ions to the inner wall of the capillary. However, can formation of the
complex affect the chemiluminescence detection reaction?

The interaction between the metal ion, M“, and the complexing agent, HIBA’, is

expressed by the following equilibrium equation (17):
M“ + nHIBA" .—_: [M(HIBA)n 12‘" (3.3)

[M (HIBA),, 12‘"
K = . .
w“) [HIBA’ 1" (3 4)

 

Here K is the stability constant, and n is the number of ligands. The total concentration of

metal ion, [M+]T, is expressed by:

[M“]T = [M“] + [M(HIBA)]Z'l + [M(HIBA)2]Z'2 + ...... + [M(HIBA),,]Z‘“ (3.5)

75

 

1.0

.C
so
1

1
1
1

1
1
1
1
1

.O
O\
l

 

p
b
l

Chemiluminescence Intensity (arbitary unit)
:5 c
1» Ln

9:
N

0.1

0.0 r

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.18

I T I I

100 200 300 400
Time (8)

Background signal of the buffer containing zero HIBA, pH = 4.00.

Buffer: 2.0 x 10“1 M luminol, 1.0 x 10'8 M Co“, 0.01 M NaAc/HAc, 0
mM HIBA, pH: 4.00; post-column CL reagents: 0.025 M H202, 0.050 M
phosphate, pH = 13.00; CL reagent ﬂow rate: 6 ILL ls; electrophoretic
current: 1.2 ILA; PMT voltage: 800 V; ampliﬁer: 108 V/A; ampliﬁer rise
time: 3 ms; separation voltage: 10 kV.

76

1.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.9 —»— —~-—~ , v ,i, f __ ___._-,___
A 0.8 -— ~ . If - _-_______,_
'E
:1
S 0.7 ._ +- _ --#_ _.
.2
.§‘ 0.6 ‘"—"‘ ——_._- ._
E
8
,5 0.5 ~ - - ,-., __,. _ -.
8
8
a 0.4 — _- IL . __
d.)
.C‘.
E
a 0.3
.8
U

0.2

0.1 A _ _ ..

0.0 l I I I I

0 100 200 300 400
pH

Figure 3.19. Background signal of the buffer containing 0.004 M HIBA, pH = 4.00.
Buffer: 2.0 x 104 M luminol, 1.0 x 10'8 M (202+, 0.01 M NaAc/HAc, 0.004
M HIBA, pH: 4.00; post-column CL reagents: 0.025 M H202, 0.050 M
phosphate, pH = 13.00; CL reagent ﬂow rate: 6 “Us; electrophoretic
current: 1.2 11A; PMT voltage: 800 V; ampliﬁer: 108 WA; ampliﬁer rise
time: 3 ms; separation voltage: 10 kV.

77

1.0

 

0.9 *-

0.8 a:

0.6 a

0.7 +

0.5

 

Chemiluminescence Intensity (arbitrary unit)

 

 

 

 

 

 

 

Figure 3.20

100 200 300 400
pH

Background signal of the buffer containing 8mM HIBA, pH = 4.00

Buffer: 2.0 x 10“1 M luminol, 1.0 x 10'8 M (302+, 0.01 M NaAc/I-lAc, 8
mM HIBA, pH: 4.00; post-column CL reagents: 0.025 M H202, 0.050 M
phosphate, pH = 13.00; CL reagent ﬂow rate: 6 “Us; electrophoretic
current: 2.3 uA; PMT voltage: 800 V; ampliﬁer: 108 V/A; ampliﬁer rise
time: 3 ms; separation voltage: 10 kV.

78

HIBA acts as a carrier for the metal ions, and thus reduces the adsorption of metal ions to
the capillary wall.
The apparent mobilties of the metal ions are the sum of the mobilities of free metal

and each complex:
papp=0tuMz+ +[3umBAZ_1 +...+uE0F (3.6)

Here or, B are the mole fractions of each species containing the metal ion, and 1150}: is the
electroosmotic ﬂow mobility. The uncomplexed metal ions have a higher mobility than
the background electrolyte and the peaks Show various degrees of fronting (18, 19). The
complexed cations possess lower apparent mobilities, which more closely match the
mobility of the electrolyte. The peak shapes in Figure 3.14, therefore, exhibit better peak
symmetry.

A static test was done to study if the metal-HIBA complex can affect post-column
chemiluminescence. A separation buffer of 2 x 10“ M luminol, 1.0 x 10'8 M Co“, 0.01
M acetic acid buffer with different amounts of HIBA was prepared. Under 10 kV high
voltage, the COB-HIBA complex in the separation buffer migrates toward the ground
buffer end. No signiﬁcant difference in background signal can be seen as shown in ﬁgure
3.18 - ﬁgure 3.20. This indicates that the metal-HIBA complexes can release the free

metal ions quickly upon mixing with the post-column CL reagents.

3.4.6 Effect of pH on Sensitivity for Buffers with HIBA
The effect of buffer pH on peak height and migration time of Co2+ is shown in ﬁgure
3.21 and ﬁgure 3.22. As the buffer pH is increased from 3.50 to 4.00, migration times

decrease due to the increased electroosmotic ﬂow; the peak height increases signiﬁcantly

79

3.0

 

1
1
1

1
1
1
1

 

1
1
I

Chemiluminescence Intensity (arbitrary unit)
kl!

 

 

 

 

 

     

 

0.5 1 :~- - ——— ~ -- -----_-.._
1
0'0 M I I I I T
400 450 500 550 600 650 700

Tune (8)

Figure 3.21 Migration time and sensitivity of Co2+ for buffer with 5 mM HIBA, pH 3.50
Sample: 1.0 x 10'8 M Co“, 5.0 x 10'7 M 012+; buffer: 2.0 x 10“t M luminol,
0.01 M NaAc/HAc, 0.005 M HIBA, pH = 3.50 post-column CL reagent:
0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 Warm;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time: 10
s; capillary: 50 um x 363 um o.d. x 50 cm.

80

 

 

 

C\
L
I
1

1

 

 

1

:
1

 

 

 

 

 

 

 

 

Chemiluminescence Intensity (arbitrary unit)

N
1
1

 

 

 

 

 

 

 

Time (s)

Figure 3.22 Peak height and migration time of Co2+ for buffer with 5 mM HIBA, pH
4.00
Sample: 1.0 x 10'8 M Co“, 5.0 x 10'7 M Cu“; buffer: 2.0 x 10“ M luminol,
0.01 M NaAc/HAC, 0.005 M HIBA, pH = 4.00; post-column CL reagent:
0.05 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 uUmin;
electrophoretic voltage: 10 kV; injection voltage: 10 kV; injection time: 10
s; capillary: 50 um x 363 um o.d. x 50 cm.

81

h—I
C

 

.9
\o
l

p
m
l

p
\l
1
r

P
05
l

.C
In

53
rs
l

p
U)
l

 

Chemiluminescence Intensity (arbitrary unit)

9
N

9
ﬂ

  

 

. “MIN" “1 U

 

0.0 t

Figure 3.23

I I I I I 7 I

100 200 300 400 500 600

Time (5)

Effect of pH (3.85) on the background signal

Buffer: 2.0 x 10“ M luminol, 1.0 x 10‘8 M Co“, 0.01 M NaAc/HAc, 0.004
M HIBA, pH 3.85; post-column CL reagents: 0.025 M H202, 0.050 M
phosphate, pH = 13.00; CL reagent ﬂow rate: 6 ILL ls; electrophoretic
current: 1.2 ILA; PMT voltage: 800 V; ampliﬁer: 108 VIA; ampliﬁer rise
time: 3 ms; separation voltage: 10 kV; capillary: 50 um x 363 um o.d. x 40
cm long.

82

0.20

 

0.18

0.16 -

0.14 -

0.12 .

0.10

 

Chemiluminescence Intensity (arbitrary unit)

Figure 3.24

 

 

 

100 200 300 400 500 600
Time (s)

Effect of pH (4.10) on the background signal

Buffer: 2.0 x 10'4 M luminol, 1.0 x 10'8 M Co“, 0.01 M NaAc/HAc, 0.004
M HIBA, pH 4.10; post-column CL reagents: 0.025 M H202, 0.050 M
phosphate, pH = 13.00; CL reagent ﬂow rate: 6 ILL ls; electrophoretic
current: 2.0 uA; PMT voltage: 800 V; ampliﬁer: 108 WA; ampliﬁer rise
time: 3 ms; separation voltage: 10 kV; capillary: 50 um x 363 um o.d. x 40
cm long.

for Co“. However, no peak can be detected at pH 4.25. Why? As the buffer pH increases,
the mole fraction of A‘ increases (pKa of HIBA = 3.79)(20) as illustrated by equation
3.7:
HA .-—‘ H” + A" (3.7)

This reduces the mole fraction of free metal ions by forming more complex. The
increased stability of the metal-complex at higher buffer pH makes it more difﬁcult for
the post-column CL reaction to compete for the metal ions from the complex in the short
period of time (about 3 s) available; therefore, no peaks can be detected.

The pH affects both the electroosmotic ﬂow and the equilibrium of the complexation
as discussed above. To further understand the effect of pH on the sensitivity, a static test
was done to study the effect of buffer pH on the background Signal in the presence of
0.005 M HIBA. Figure 3.23 (buffer pH 3.85) shows a higher background signal than
ﬁgure 3.24, because the metal-HIBA complex can still release the free metal ions in ~ 2 s
at pH 3.85. However, at pH 4.00, the mole fraction of the HIBA anion (pKa HIBA = 3.65)
increases so that it take longer than 2 - 3 s for the complex to release the free metal

cations.

3.4.7 Reproducibility of Sample Injection
The reproducibility of 10 electrokinetic injections is listed in Tables 3.3 and 3.4. The
reproducibility is not as good as achieved with on-column detection (2 - 10%), perhaps

due to the extra variance associated with post-column CL detection.

84

Table 3.3 Reproducibility of 10 Electrokinetic Injections of Co2+

 

 

 

Peak Height Mean Standard Deviation RSD
1.0 x 10'9 M Co2+
10 injections 5.64 0.67 1 1.89%

 

 

 

 

 

Sample: 1.0 x 10'9 M Coz“; buffer: 2.0 x 104 M luminol, 0.01 M acetic acid, pH = 4.75.
Capillary: 50 um i.d. x 363 um o.d. x 50 cm long fused silica capillary. Separation
voltage: 10 kV. Ampliﬁer: 108 V/A, 3 ms response time. PMT voltage: 800 V. Post-
column reagents: 0.05 M H202, 0.1 M phosphate (pH = 12.50), ﬂow rate: 3 ulJmin per
channel. Voltage of injection: 10 kV. Time of injection: 4s.

Table 3.4 Reproducibility of 10 Electrokinetic Injections of Co2+

 

 

 

Peak Height Mean Standard Deviation RSD
1.0 x 10‘6 M Cr3+
10 injections 3.45 0.62 17.98%

 

 

 

 

 

Sample: 1.0 x 10'9 M Co”, 1.0 x 10'6 M Cr3+ in deionized distilled water; separation
buffer: 2 x 104 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-column CL reagents:
0.025 M H202, 0.050 M phosphate, pH = 12.49; ﬂow rate: 6 1.1L ls; electrophoretic
current: 2.0 11A; PMT voltage: 800 V; ampliﬁer: 108 VIA; ampliﬁer rise time: 3 ms;
separation voltage: 10 kV; sample injection: 10 s at 10 kV; capillary: 50 um i.d. x 363 um
x 40 cm long.
3.4.8 Number of Theoretical Plates with Post-Column Chemiluminscence Detection

The number of theoretical plates is inﬂuenced by many factors such as sample, and
separation conditions. We estimated the number of theoretical plates for 1.0 x 10'll M
Co2+ based on ﬁgure 3.25. The peak showed a migration time of 188.9 s and a peak width
of 1.3 s at the baseline. The number of plates was calculated from the equation 3.8:

2
N = 16 [IL] (3.8)

W

Here tr is the migration time of the sample zone, and w is the peak width at the baseline.

85

 

1.0

S3 .O .O
\l 00 \O
l 1 l

P
at
1

.O
A
l

Chemiluminescence Intensity (arbitrary unit)
o o
i» u:

.0
N
l

 

1.0 x 10'11 M Co2+

 

 

 

 

 

 

.9
y—n

 

 

 

 

 

0.0 ﬂ

Figure 3.25

 

Theoretical plates for 1.0 x 10'11 M Co2+ (t,: 188.9 s; peak width: 1.3 s)
Sample: 1.0 x 10'11 M in deionized distilled water; separation buffer:

2 x 104 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-column CL
reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.49; ﬂow rate: 6 [IL ls;
electrophoretic current: 2.0 M; PMT voltage: 800 V; ampliﬁer: 108 WA;
ampliﬁer rise time: 3 ms; separation voltage: 10 kV; sample injection: 10 s
at 10 kV; capillary: 50 um i.d. x 363 um o.d. x 40 cm long.

86

We were able to achieve 340,000 plates for 10'11 M Co“, which is superior to the
200,000 plates for isoluminol as reported by other workers (7) using CL detection in CE.
This plate number matches those with on-column detection, meaning that there is little
additional contribution to peak broadening from the post-column CL reaction. Figure 3.2a

gives 189,225 plates for 1.0 x 10‘6 M Cr3+ and 116,64 plates for 1.0 x 1043 M Cu”.

3.4.9 Calibrations for Transition Metal Cations

A series of standard solutions containing various levels of Co”, Cr“, and Cu2+ was
prepared using deionized distilled water. All these standard solutions were used under the
same experimental conditions.

The peak areas for Co)”, Cr“, and Cu2+ were plotted against concentrations as shown
in ﬁgures 3.26, 3.27, and 3.28 respectively. The curves deviate from linearity at high
concentrations because of the increased adsorption to the capillary wall. Based on a S/N
ratio of 3, the detection limit is estimated in table 3.5. The detection limit for Co2+ with
CL detection is 6 orders of magnitude lower than that with UV detection and
electrochemical detection (ECD); the detection limit of Cr3+ with CL detection is 3 orders
of magnitude lower than that with ECD. The detection limit of Cu2+ with CL detection is

two orders of magnitude lower than that with UV detection and ECD.

Table 3.5 Detection limits for some transition metals using CL and other detectors

 

 

 

Cations Detection Limit (UV) Detection Limit Detection Limit (CL)
Ref. (21) (ECD) ref (4)
Co2+ 110 ppb or 1.8 x106M 5.1x 10'6M 1.5 x 10'12M
Cr3+ 3.8 x 10'5 M 5.0 x 10'8 M
Cu2+ 206 ppb or 3.2 x 10‘6 M 5.0 x 10*5 M 3.0 x 10'8 M

 

 

 

 

 

87

 

45

40*

35 -*

30~

25‘

Peak Area (arbitrary unit)

l5

10 m

20 a»

,_ ——v r————-

 

 

 

 

0 I I I I I I

 

 

 

0.00E+00 2.00E-09 4.00E-09 6.00E-09 8.00E-09 1 .00E-08 1 .20E-08

Concentration of Co“ (M)

Figure 3.26 Calibration curve for Co“.

Sample: 1.0 x 10'11 M - 1.0 x 10'8 M Co“ in deionized distilled water;
separation buffer: 2 x 10’4 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-
column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.49; ﬂow
rate: 6 ILL Is; electrophoretic current: 2.0 uA; PMT voltage: 800 V;
ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;
sample injection: 10 s at 10 kV.

88

 

80

70--

60 _ _._.-- -_ , -7, ,.,, ,-L-, 7,, WV, .,7 W . __

50---

40 - -- ------ ~~~ , ----—-- eeeee - , ---~———————-—~ ,

 

30

 

 

Peak Area (arbitrary unit)

20 - ----- -, , e~--«---~

10‘
0. r

0.00E+00 2.00E-06 4.00E-06 6.00E-06 8.00E-06 1.00E-05 1 .20E-05

 

 

 

 

I I

Concentration of Cr3+ (M)

Figure 3.27 Calibration curve for Cr“.
Sample: 5.0 x 10’8 M - 1.0 x 10'5 M Cr3+ in deionized distilled water;
separation buffer: 2 x 104 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-
column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.06; ﬂow
rate: 6 1.1L Is; electrophoretic current: 2.0 ILA; PMT voltage: 800 V;
ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;
sample injection: 10 s at 10 kV.

89

 

Peak Area (arbitrary unit)

 

 

 

 

 

O I | l I l

0.00E+OO 2.00E-06 4.00E-06 6.00E-06 8.00E-06 l .OOE-OS l .20E-05

Concentration of Cu“ (M)

Figure 3.28 Calibration curve for Cu“.
Sample: 5.0 x 10‘8 M - 1.0 x 10'5 M (:u2+ in deionized distilled water;
separation buffer: 2 x 10“1 M luminol, 0.010 M NaAc/HAc, pH: 4.74; post-
column CL reagents: 0.025 M H202, 0.1 M phosphate, pH = 12.06; ﬂow
rate: 6 uL ls; electrophoretic current: 2.0 uA; PMT voltage: 800 V;
ampliﬁer: 108 WA; ampliﬁer rise time: 3 ms; separation voltage: 10 kV;
sample injection: 10 s at 10 kV.

90

3.10 Conclusions

This work demonstrates the sensitivity of chemiluminescence detection compared to
other techniques such as electrochemical and UV detection. The post-column CL reactor
demonstrates a high separation efﬁciency of 340,000 plates, which is even competitive
with those with on-column detection in capillary zone electrophoresis. Many factors such
as the pH of the post-column CL reagents and separation buffer were optimized.
Reproducibility of post-column CL detection is not as good as that with on-column
detection due to the extra variance introduced by the post-column mixing and reaction.
The calibration curve showed deviations from linearity at high concentration due to the
increased adsorption of metal ions to the active sites on the capillary surface. It is
necessary to reduce the adsorption of transition metal cations, particularly at high

concentrations.

91

List of References

1

10.

ll.

12.

l3.

14.

15.

Hara, T.; Okamura, S.; Kato, J .; Yokogi, J.; Nakajima, R. Anal. Sci. 1991, 7, 261-
264.

Dadoo, R.; Colon, L.A.; Zare, R. N. J. High Resolution Chromatography, 1992, 15,
133.

Evans, C. E. Anal. Chem. 1997, 69, 2952-2954.

Lu, W. Z.; Cassidy, R. M. Anal. Chem. 1993, 65, 1649-1653.

Fuller, R. R.; Sweedler, J. V. Anal. Chem. 1999, 71 , 4014-4022.

Ingle, J. D.; Crouch, S. R., Spectrochemical Analysis, Prentice Hall, Englewood
Cliffs, New Jersey, 1988, p 479-485.

J. Y. Zhao, J. Labba, and N. J. Dovichi, J. of Microcolumn Separation, 1993, 5,
331.

CRC Handbook of Chemistry and Physics, 72th edition, 195.

Peng, C. H., “Application of Micro-Machining Technology for the Construction of
End-Column Conductivity Detector in Capillary Zone Electrophoreis”, M. Sc.
Thesis, p 57, University of Alberta, 1993.

Riviello, J M.; Harrold, M. P. J of Chromatography A, 1993, 652, 385-392.

Kuban, P.; Karlberg, B. Anal. Chem. 1998, 70, 360-365.

Weston, A.; Brown, P. R.; Jandik, P.; Jones, W. R.; Heckenberg, A. L. J.
Chromatogr. 1992, 593, 289-295.).

Foret, F.; Fanali, S.; Nardi, A.; Bocek, P. Electrophoresis 1990, 11, 780-783.

Chen, M.; Cassidy, R. M. J. Chromatogr. 1993, 640, 425—431.

Lin, T. 1.; Lee, Y. H.; Chen, Y. C. J. Chromatogr. 1993, 654, 167-176.

92

l6.

l7.

18.

19.

20.

21.

Shi, Y.; Frits, J. S. J. Chromatogr. 1993, 640, 473-479.

Swaile, D. F.; Sepaniak, M. J. Anal. Chem. 1991, 63, 179.

Mikkers, F. E. P.; Everaerts, F. M.; Verheggen, Th. P. E. M J Chromatogr., 1979,
169, l-lO.

Foret, F.; Fanali, L.; Ossicini, L.; Bocek, P. J Chromatogr., 1989, 470, 299.

Perrin, D. D. Stability Constants of Metal Ion Complexes, Part B; IUPAC Chemical
Data Series 22; Pergamon Press: Oxford, 1982; p187.

Weston, A.; Brown, P. R.; Jandik, P.; Jones, W. R.; Heckenberg, A. L. J.

Chromatogr. 1992, 593, 289-295.

93

Chapter 4
Automatic Sample Injection System with No Interruption of High Voltage

in Capillary Zone Electrophoresis

Sample introduction is one of the very challenging aspects of capillary zone
electrophoresis. Electrokinetic introduction and hydrodynamic introduction are the most
commonly used sample introduction methods. In electrokinetic introduction, an electric
ﬁeld is applied across the capillary for a certain period of time to move the sample into
the capillary. In hydrodynamic introduction, the sample is forced into the capillary using
the siphoning effect. Electrokinetic introduction causes differential injection of species
due to difference in their mobilities. Hydrodynamic injection causes a parabolic proﬁle of
sample zones due to frictional forces at the capillary walls.

The common drawback of these two injection methods is the relocation of the
separation capillary during sample injection, which is not convenient for automation.
Another drawback is the low sample throughput limited by the migration time. In this
chapter, we describe an automatic sample injection system, which requires neither the
relocation of separation capillary nor the disconnection of the electric ﬁeld during sample
injection. In this automatic sample injector, the sample is stored in a reservoir 76 cm
higher than the high voltage reservoir. There are two identical capillaries, one for sample
injection, one for separation. The two capillaries are accurately aligned. Upon injection,
the sample capillary mounted on a sample platform rotates toward the separation capillary

and meets the separation capillary for a time period controlled by computer. During this

94

process, the sample ﬂows from sample capillary into the separation capillary driven by
gravity. After injection, the sample capillary moves away from the separation capillary.
The automatic sample injector allows sample injection without the interruption of high

voltage, unlike traditional electrokinetic or hydrodynamic injection.

4.1 Instrumentation for the Automatic Sample Injector

The automatic injection system includes a mechanical system, an electronic control
circuit and a computer. A 10” x 18” x 1/2 “ aluminum plate with lAl”-20 holes spaced 1”
apart was used as a breadboard for mounting various parts. The computer can give a i 12
V voltage output with 2 uA of current. This current running through the electromagnet is
too low to attract the steel plate glued to the Plexiglas rotary platform shown in ﬁgure 4.1.
As we can see from ﬁgure 4.1, the pinch valve and the solenoid are synchronized. The
pinch valve is normally closed, which prevents the leakage of sample solution. As the
solenoid is actuated to pull the sample platform, the pinch valve opens in 30 ms and
sample solution is ready to be delivered into the separation capillary. It takes 1.04 s for
the sample platform to move into position to join the separation capillary. This period of
time must be reproducible or else the time required for joining the sample and the
separation capillary will not be consistent.

This automatic sample injection system uses two segments of fused silica capillaries
as sample capillary and separation capillary respectively, and requires the accurate
alignment of the two segments for reproducible sample injection. When the computer

sends out a trigger signal (4V), the DC power supply gives 12 V to the solenoid and the

95

 

 

$3853.58? aha—ammo E 889?. 8:0an 01:83 638832 :d oSmE

 

 

 

 

 

 

 

 

 

 

nouasm Jaynq punosﬁ G—l I I

BBQ ~0on
/ Bozo—8
. /v
Sheba—m ES mammxo—mllv _
my.
.522 3598/ e
- l! 8an ace—=30
\a/ - _ w \J o..\ ml ..
C a _ e K e
.m m. m
u
w boar—amok 3.55 >3:an 9383 I! m. m.
m. omﬁm noun—ﬂab
0
m.
m1 033 as u l
33> semi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

£0388 0388 b

 

96

pinch valve in less than 1 us, through the electronic circuit shown in ﬁgure 4.2. Here the
ﬁeld effect transistor (MTP30NO6EL, Motorola) acts as a switch. A 3%” x 1” x 1/64”
thick steel plate is glued to the left side of the sample platform for electromagnetic
attraction. The sample capillary is held on a Plexiglas sample platform sitting on a Mi”
shaft. The separation capillary is held on another Plexiglas platform mounted on a
translation stage. Upon activation of the electronic circuit in ﬁgure 4.2, the sample
platform with the steel plate is attracted by the electromagnet. The sample capillary on the
sample platform slowly meets the separation capillary with the assistance of the damping
device described in ﬁgure 4.3. It is very important to align the sample capillary to the
separation capillary for a good reproducibility. The alignment of the sample capillary is
assured by the two precision ball bearings described in ﬁgure 4.4.

The normally closed pinch valve opens upon its activation. The sample capillary
delivers the sample and is detached from the separation capillary upon the repelling of the
spring and the attraction from the permanent magnet mounted on the solenoid support as
shown in ﬁgure 4.5. The pinch valve is closed and the injection is over. This is how the

automatic sample injection system works.

4.2 Electronic Circuit

The power supply to the sample injection system should have a short turn on/off time
(< 1 ms). Any delay will cause inaccurate sample injection time. The 12 V DC power-
supply can not be turned on and off directly due to the delay. The electronic circuit was
thus designedto solve the problem of delay as shown in ﬁgure 4.3. The ﬁeld effect

transistor has a very fast response time (< 1 us). The electronic circuit is able to give 12 V

97

.mi pd 6:5 8.838 3. 2 ”E830 E=E_§E "85358 no ES G 3.0 ”AmoOZOmmHE
88838820 E:E~o E 889?. screen: 88883 88 :58? 28:85 mé 8=mE

 

 

 

 

 

 

> N— .
xiv Z—
.lil. u- .888
Am—cOZOmmH—Z Illv_ AMwOZOmmh—Z 111v.
..IL 111
G 3
we; 7: ﬁw 33> coca we? 7: 2828 WW we; Z—
- > S +

 

 

 

98

DC output in less than 1 us to the solenoid and the pinch valve. The solenoid is made of
the coated copper wire (28 gauge or 0.013” diameter, 8054 copper wire, Belden Corp.,
Chicago, IL 60644). The copper wire was wound around a 1/2” x 10” diameter steel rod
for about 7 inches. The total resistance of the copper wire is 4.4 Q. A 15 .0 power resistor
is connected in series with the copper coil to prevent the coil from excessive current. On
the other hand, this resistor limits the strength of the electromagnetic ﬁeld so that the
attraction of the sample capillary toward the solenoid is not too strong otherwise impact

damping can be difﬁcult.

4.3 Damping Device

A dashpot (Airpot Corp., series 1603) for damping the shock which results when the
sample and separation capillary meet is shown in ﬁgure 4.3 The dashpot was mounted on
the solenoid support described in section 4.5. Maximum damping or time delay occurs
when the adjustment screw is turned clockwise until it seats. More turns make the screw
slip toward the open position which results in a slight decrease in damping. Another
clockwise turn will restore maximum damping. To reduce damping, the screw is turned
counter-clockwise.

The dashpot cannot be used directly in the sample injection system. The threaded ball
joint was cut off and a stainless steel tube was glued on the tiny piston rod. The piston
retainer was replaced with a homemade Plexiglas retainer which ﬁt exactly into the glass
cylinder. A l/ 16” hole was drilled on the Plexiglas retainer to allow the stainless steel
tubing to go through. The retainer is then screwed onto the l/2”-13 threaded hole on the

right side Plexiglas plate.

99

 
 

lil'llllllili‘llllllll"

Adjustment Knob / a gladius. Shcliplder
ountmg ut

a — Belleville Lockwasher

 

 

 

 

 

 

 

Adjustment Screw
Mounting Stud Ultra Low Friction
_._— x linder
\ / Pym Cy
IT!
Ultra Low Friction
Graphite Piston

\‘ \“ ‘ ‘ l“\“““\“ ‘

 

 

Piston Retainer -—-E|
Threaded Ball Joint
\$ Split Lockwasher
./

\

 

 

#2-64 Hex Nut

CID

Figure 4.3 Dashpot for damping the shock between sample and separation capillary

4.4 Rotary Alignment System

Ideally, when the sample capillary and the separation capillary are joined, the
alignment should be excellent and the contact slow and gentle. The most difﬁcult work is
the accurate alignment of the sample capillary and separation capillary due to their small

dimensions. Some commercially available linear motion devices give repeatability up to

1 pm. However, the cost is too expensive (~ $9000) for sample injection in CB. Perhaps it

100

 

 

 

 

 

O <—— 4-40 countersunk hole
<- round groove
0 363 um diameter
A
4-40 tapped hole —- ->Qf
sample capillary l
l 1.5"
—— 0.25"
O
l!

 

 

 

0.25" diameter stainless steel shaﬁ—>

313

 

 

 

 

 

 

 

 

0.25" i.d. x 0.625" o.d. ——~ 3 l 0-25 ---'-1:l 2-0
precision ball bearing '8.
'-' 1.0" thick Plexiglas
g sample platform
0.625" i.d. x 0.8125" o.d.—+1 v
aluminum tube 4 2 0" :

 

0.8125" i.d. x 2.0" o.d.—1
Plexiglas tube

2.0"

F————————dp—————————1

 

 

 

 

 

Q-_____..__.._________,

 

 

 

ll

a 3.0" :

 

 

 

 

 

Figure 4.4 Rotary alignment system for automatic sample injection

101

is not necessary to use such expensive equipment if we can construct an alternative.
Attempts to build a linear motion device were not successful. We then constructed the
rotary system for alignment shown in ﬁgure 4.4, and the result was satisfactory. The
precision of the two ball bearings (Barden Inc., product No. SR4SS) determines the
quality of alignment. The aluminum tube holding the two precision ball bearings is glued
to the Plexiglas with epoxy. The sample capillary is held on the sample platform with a

special capillary holder described in section 4.4.

4.5 Holder for the Separation Capillary

Figure 4.5 shows the schematic diagram of the capillary holder made of V2” thick
Plexiglas. The machining of the groove to hold the capillary is somewhat tricky. A short
segment of the capillary (same dimension of those used for separation) was ﬁxed on the
centerline using scotch tape between the two holes of the capillary holder. Another piece
of Plexiglas was used to sandwich the capillary. The Plexiglas block was then placed in
the chuck of the milling machine; the chuck was then tightened, so that pressure carved a
round groove on the centerline of the capillary holder. The round groove was carefully
examined under the microscope, and it was found to be smooth and uniform. The
capillary was then mounted onto the Plexiglas support block. The Plexiglas support block
is then ﬁxed onto a right-angle bracket (Newport Inc.) which is screwed onto the top of a

two-dimensional translation stage, for precise position adjustment.

102

6.0"

 

 

° ‘7 4-40 taped hole 0-5"

/ O 0.25" hole G

1.0.. o capillary 0.5"

 

 

 

 

Material: 0.5" thick Plexiglas plate 3-0"

 

o <-—- 4-40 countersunk hole

<—— round groove 0 O

o 0.363 mm diameter

 

 

 

 

 

 

 

 

Clamp : e
3.0”

 

Figure 4.5 Separation capillary holder in automatic injection system

103

4.6 Solenoid Support

The solenoid support is shown in ﬁgure 4.6. It is made of three pieces of Plexiglas, a
bottom piece and two side pieces. A 1/2 hole is drilled on each of the side Plexiglas plate
in order to hold the solenoid in position with a set-screw. Two ‘A” holes were drilled on

the bottom to hold the solenoid support in position on the aluminum breadboard. The side

pieces were connected to the bottom piece with six I/4-20” screws.

 

 

 

 

 

 

0.5" 0.5"
solenoid
0.5" ]
32:! 1/4"-2o x 0.5" deep hole ’ 1.0" :22:

 

 

 

 

 

 

 

 

 

 

 

1/2"-13 threaded hole \\ 1.0”
O 0.5" hole
7.0"
2
O
:1 :5, I: 0.5" thick Plexiglas plate 6.0"
b
:1 3' O O I: 4.0"
o
‘1‘
:1 3 l: 0.25" hole
"‘ O O O
4.0"
bottom side

Figure 4.6 Support for the solenoid and the shock-damping dashpot

104

4.7 Capillary Polishing Tools

The end of the sample and separation capillary must be very flat in order to have the
good contact required for sample introduction. A home-made polishing tool, shown in
ﬁgure 4.7, is used to do the polishing work. About 20 capillaries (50 um i.d. x 363 um
o.d.) are clamped in the 1/:z” wide x 0.150 mm deep slot of the polishing tool. The end of
the capillaries protrudes out of the bottom for about 0.5 mm. The tool is slowly moved in
a circular motion on a piece of wet ﬁne sand-paper for about 5 minutes. The end of the
capillaries is then inspected under the microscope, and the polishing repeated until a
smooth surface is obtained. A few capillaries were clogged by the particles from
sandpaper. The clogged capillaries were then placed in the ultrasonic bath and a syringe

ﬁlled with water was used to clear the clogged capillaries.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

sideview front view of part B

9°°angle . ............ ,,

‘:::::: g ::::::

l:::::: g- ::::::,

*:::::: ~19, ::::::

£:::::: a ::::::

. ...... a ......

partA partB ....... O ......

‘l ’ {2221212222223 3.0"

l ...... o ......

J _____________________ ’053'. . :-.:..
___g.2_5: ____________ .—.-.. . 3 .O.
1.5"» ::::l:: E ::::::
:pp-ijz' m ::::::

’::::.:: ° ::::::

., ::::g:: :::::: .,
i 1.5" ; ‘ 2.0" ,

Figure 4.7 Polishing tools for the sample and the separation capillary

105

4.8 Tubing Connector and Pinch Valve

Sample solution is stored in a plastic syringe. The sample capillary enters the buffer
reservoir under high voltage upon the activation of the solenoid. A short piece of silicon
rubber tubing is used to connect the sample syringe with the sample capillary. The regular
disposable needle (Becton Dickinson & Co.) should not be used directly since the
stainless steel needle may cause interruption of the high voltage. The needle is pulled out
and then a l/ 16” hole is drilled on the plastic socket. A 500 um i.d. x 1/ 16” o.d. PEEK
sleeve (Upchurch Scientiﬁc Inc., product No. F230) is forced into the hole. One end of a
silicone rubber tubing 0.015” i.d. x 1/32” o.d. (Manostat Corp., product # 75-300-015) is
connected to the PEEK sleeve, and the other end is connected to the sample capillary (50
um i.d. x 363 um o.d. x 20 cm long). The silicon tubing is installed on the pinch valve,
which is normally closed to prevent the leakage of sample solution. The end of the silicon
tubing is ﬁxed onto the high voltage box. 5

The pinch valve is screwed on the four small holes (1/2” spacing) of the Plexiglas
pinch valve holder as shown in ﬁgure 4.8. The pinch valve holder is inserted into a V2”
aluminum rod through the V2” center hole shown in top-view of ﬁgure 4.8 and held in
position about 65 cm above the breadboard with a lA8520 set-screw. The sample syringe is

mounted 76 cm above the breadboard.

106

0.25"
t

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ 2'0 /o.125" hole
0.25" - 0.5"
t /
Front-view O O 1.0"
o
T
I
4-40 hole
0.75"
0.25"
11 L__.___ 8-32 taped hole
. l/4"-20 ,,
top-View taped hole ‘3 2.0
0.5" hole
1: 8-32 taped hole
2.0"
side-view i:
0.5"

 

 

 

 

 

 

Figure 4.8 Pinch valve holder for automatic injection system

107

Chapter 5
Characterization of the Automatic Sample Injection System in Capillary

Zone Electrophoresis

Sample injection in CE is different from that in LC, GC or FIA where a loop of ﬁxed
volume is used for injection. Since the capillary bore is very small, a loop injector is not
technically practical at the moment. The amount of sample injected into the separation
capillary in CE is time-controlled; therefore human error usually plays a big part in the
overall error. The injection error in CB includes timing error and the transfer error.
Timing error is one of the most important sources of error in capillary electrophoresis in
manual injection. It is not uncommon to see 10% or higher relative standard deviation
(RSD) in peak height due to irreproducible injections in CE.

Transfer error occurs when the capillary is moved from the sample to the buffer; this
period of time can last about 2 - 4 seconds during which there is no electric ﬁeld across
the separation capillary. During this time, the sample zone can diffuse back and forth in
the capillary, making the zone wider. Any motion of the separation capillary during this
transfer process can further broaden the sample zone. Unfortunately, no CE system,
including those commercially available, is able to eliminate diffusion occurring during
this transfer. If we can reduce the timing error to some extent, we can certainly enhance
the reproducibility of sample injection in capillary electrophoresis.

Sample throughput is deﬁned in ﬂow injection analysis as the number of sample

injections performed per unit time. In CE with conventional injection, the manual

108

injections usually limit the sample throughput to ~10 injections/hour. In our automatic
injection system, 10 consecutive sample injections were accomplished in 7 minutes,
which converts to about 85 injections/hour. The higher throughput is attributed to the
continuous injections during the electrophoretic run.

Many other factors such as sample composition, and detection scheme, may affect the
reproducibility of injections in CB. In general, the RSD of injections can vary from 3% to
10% for on-column detection (2, 3). The RSD for post-column detection is higher due to
the additional source of error. In this chapter, an automatic sample injection system is
characterized. The linear flow rate v, of sample in the sample capillary (open tubular) can

be estimated by the Hagen-Poiseuille equation:

V:3LUAP

r2

(5.1)

Here AP is the pressure drop along an open tubular capillary of length L with radius r

containing a ﬂowing solution of viscosity T].

5.1 Reproducibility of Timing

We repeated 10 tests to calculate the reproducibility of this time period using a
stopwatch and obtained a mean value of 1.04 s with a travel time standard deviation of
0.04 s. The measured RSD of the automatic injection system is 3.7%. This is perhaps an
upper limit because of the irreproducibility of the manual time measurement. Knowing
the 0.04 5 travel time standard deviation, we measured the precision of the injection time

under various preset values of injection time. The measured time includes the travel time

and the injection time (tmeals = tinj + ttravel); hence the injection time is that measured time

109

minus the 1.04 5 travel time. The means are reported for 10 injections. The RSD

decreases with longer injection time.

Table 5.1 Reproducibility of the Injection System without Sample Injection

 

 

 

 

 

Nominal injection time
23 SS 105
Measured time (tum) 2.95 i 0.04 6.01 i 0.05 11.00 i 0.03
Injection time (tinj) 1.91 i 0.06 4.97 i 0.06 9.96 i 0.05
RSD 3.1% 1.2 % 0.5%

 

 

 

 

5.2 Reproducibility of Peak Height

Figure 5.1 shows the reproducibility of 10 consecutive injections using the automatic
injection system and a 0.96 3 injection (nominal injection: 2 s). The relative standard
deviation (RSD) of the peak height is 10.7%. The automatic injection system can
efﬁciently reduce the timing uncertainty. The advantage of automatic injection is seen
clearly here. For a 1 8 manual injection in CE, accurate timing is difﬁcult. The fourth
peak is relatively low probably due to the vibration between the sample capillary and the
separation capillary when they were joined. As we discussed earlier, good pressure
damping is essential for reproducible results. Figure 5.2 shows the reproducibility of 10
consecutive injections using the automatic injection system for 3.96 5 injections. The
RSD of the peak height is 10.9%. In manual sample injection, the RSD is larger for
shorter injection due to timing error. In automatic injection, the timing error is reduced;
consequently the RSD for the two sets of sample injections under different injection time

(0.96 s and 3.96 3) shows no signiﬁcant variation (F-test at 95% conﬁdence level).

110

 

3.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25~——— —~ —— V *# a
E
3
E. 2.0 — — — __
..D
E.
a»
8
Q)
g 1.5 e— — — — — —
8
C
8
a”:
E
2 1.o~ e— M 4 4
E
.8
U
0.54 ,ﬁ
0.0
0 100 200 300 400

Time (s)

Figure 5.1 Reproducibility of 10 consecutive injections using the automatic injection

system for 0.96 3. Sample: 1.0 x 10'8 M Co“; buffer: 2.0 x 104 M luminol,
0.01 M acetic acid, pH = 4.75. Capillary: 50 um i.d. x 363 pm o.d. x 50 cm
long fused silica capillary. Separation voltage: 10 kV. Ampliﬁer: 108 WA, 3
ms response time. PMT voltage: 800 V. Post-column reagents: 0.05 M H202,
0.1 M phosphate (pH = 12.50), flow rate: 3 uI/min per channel. Height of
sample injection: 76 cm. Injection interval: 30 3. Mean of the peak height:
1.00. RSD of the peak height: 10.7%.

111

 

3.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0 +-- _ _

 

 

 

Chemiluminescence Intensity (arbitrary unit)
U!

0.5 ~ ~ _W _

I:

0.0 M I I I I I

0 50 100 150 200 250 300 350 400
Time (s)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

     

 

Figure 5.2 Reproducibility of 10 consecutive injections using automatic injection system
for 3.96 s. Sample: 1.0 x 10'8 M Co”; buffer: 2.0 x 10“4 M luminol, 0.01 M
acetic acid, pH = 4.75. Capillary: 50 um i.d. x 363 um o.d. x 50 cm long fused
silica capillary. Separation voltage: 10 kV. Ampliﬁer: 108 WA, 3 ms response
time. PMT voltage: 800 V. Post-column reagents: 0.05 M H202, 0.1 M
phosphate (pH = 12.50), flow rate: 3 til/min per channel. Height of sample
injection: 76 cm. Injection interval: 30 3. Mean of the peak height: 2.42. RSD
of the peak height: 10.9%.

112

5.3 Function of Pinch Valve

The sample could continuously leak out in our automatic injection system. To solve

 

3.0

 

 

I

 

 

 

N

O
i
|

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Chemiluminescence Intensity (arbitrary unit)
UI

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

   

1.0 * ~ ﬂee H g} ._ _ *_ -__.
0.5 -, . .
0.0 ' r ' T . .

0 50 100 150 200 250 300 350 400

Time (s)

Figure 5.3 Reproducibility of 10 consecutive injections using automatic injection system
with pinch valve for 3.96 3. Sample: 1.0 x 10'8 M Co“; buffer: 2.0 x 10'4 M
luminol, 0.01 M acetic acid, pH = 4.75. Capillary: 50 um i.d. x 363 um o.d. x
50 cm long fused silica capillary. Separation voltage: 10 kV. Ampliﬁer: 108
WA, 3 ms response time. PMT voltage: 800 V. Post-column reagents: 0.05 M
H202, 0.1 M phosphate (pH = 12.50), ﬂow rate: 3 pJJmin per channel. Height
of sample injection: 76 cm. Injection interval: 30 5. Mean of the peak height:
2.22. RSD of the peak height: 13.4%.

113

this problem, a pinch valve was installed as described previously. Figure 5.3 shows the
reproducibility of automatic injection when a pinch valve is engaged. RSD for ﬁgure 5.3.

(13.4%) differs only slightly from that for ﬁgure 5.2 (10.9%). The mean of the peak
height in ﬁgure 5.2 (2.42) is close to that in ﬁgure 5.3 (2.22). Without the pinch valve, the
sample is driven out continuously by the siphoning effect and a tiny growing droplet of
sample remains at the tip of the sample capillary between sample injection. When the
sample injection is actuated, the sample capillary joins the separation capillary and brings
the tiny droplet of sample into the buffer reservoir. This small contamination of the buffer
seems to affect neither the peak height nor the baseline as shown in ﬁgure 5.2 and 5.3.
However, if the injection height is increased and the dr0plet is growing at higher rate, this

could be a problem. Therefore, it is still desirable to install the pinch valve.

5.4 Comparison between the Automatic Injection System and other Injection Systems
The RSD in CB injections vary between 2 ~ 10% (2, 3, 4, 5) for on-column detection.
In manual injection, the RSD may be higher when the injection time is short. The RDS

under a 4 5 injection is reported in table 5.2. The mean of the peak height is 1.74, and the

Table 5.2 Reproducibility of 10 Manual Injections for 4 s Injection

 

Peak Height Mean Standard Deviation RSD

 

 

I 4 3 Manual injection 1.74 0.47 27.0%

 

 

 

 

Sample: 1.0 x 10'9 M (302+; buffer: 2.0 x 10“t M luminol, 0.01 M acetic acid, pH = 4.75.
Capillary: 50 um i.d. x 363 um o.d. x 50 cm long fused silica capillary. Separation
voltage: 10 kV. Ampliﬁer: 108 WA, 3 ms response time. PMT voltage: 800 V. Post-
column reagents: 0.05 M H202, 0.1 M phosphate (pH = 12.50), flow rate: 3 pJJmin per
channel. Voltage of injection: 10 kV. Time of injection: 43.

114

 

standard deviation is 0.47. The RSD of the 10 manual injections with a 4 3 injection is
27.0%, substantially larger than the RSD (10.9%) of the10 automatic injection shown in

ﬁgure 5.2.

5.5 Conclusions

This work demonstrates several advantages of the automatic sample injection system
over manual injection in capillary zone electrophoresis. Several sources of injection error
are reduced signiﬁcantly in the new injection system. As a result, the new injection
system presents a unique approach to sample introduction in CB. With some instrumental
improvements in alignment and in the damping device, the RSD should be improved. The
new injection system also gives higher sample throughput. This unique advantage can be
particularly useful in some CE separations where long separation times are required so
that a series of injection can be performed before the previous run is ﬁnished. It saves

time and operation cost.

List of References

1. Evans, C. E. Anal. Chem. 1997, 69, 2952-2954

2. Fuller, R. R.; Sweedler, J. V. Anal. Chem. 1999, 71 , 4014-4022.

3. Bachmann, K.; Boden, J.; Haumann, I. J Chromatogr. 1992, 626, 259-265.

4. Weston, A.; Brown, P. R.; Heckenberg, A. L.; Jandik, P.; Jone, W. R. J.

Chromatogr. 1992, 602, 249-256.

115

Chapter 6

Summary and Conclusions

In this work, we have constructed a miniaturized chemiluminescence (CL) detector
for capillary electrophoresis (CE) which provides high collection efﬁciency (== 58%). This
CL detector is well grounded and signiﬁcantly cuts down the electrical noise associated
with AC power supply. The miniaturized CL detector is characterized using a few
transitional metal cations (Coz+, Cr“, and Cu2+) as the sample species. We found these
metal cations are all strongly adsorbed to the wall of fused-silica capillary at high
concentrations. We obtained the best theoretical plate number, 500,000, for a post-
column reaction detector in capillary electrophoresis. We were able to achieve a detection
limit of 1.5 x 10'12 M for Co”, 5.0 x 10'8 M Cr“, and 3.0 x 10'8 M Cu”. These limits of
detection are signiﬁcantly lower than those obtained by other methods of detection such
as electrochemical and UV detection.

We also constructed an automatic sample injection system for CE. This new injection
system greatly enhances the sample throughput in CB, compared to manual injections.
The automatic sample injection system shows a better reproducibility of injection
(10.9%) than the manual injection system (27.0%) in our CL-CE system for a 43

injection, based on 10 consecutive injections.

116

Chapter 7

Future Work

7.1 The Chemiluminescence Detection System

The fused silica capillary is so far the most widely used capillary in capillary
electrophoresis; the majority of CE separations have been performed in untreated fused-
silica capillaries. The walls of untreated fused-silica capillaries are chemically reactive.
They can adsorb solutes by electrostatic and/or chemical interactions, especially in the
case of transition metal cations and large molecules like proteins. The adsorption results
in band broadening, and in some cases the peak is lost completely (1).

Inert capillaries such as Teﬂon, and PEEK (Polyether Ether Ketone) are ideal for the
separation of transition metal cations if their mechanical property matches that of the
fused-silica capillary. PEEK tubing of 63 um i.d. x l/ 16” o.d. is commercially available
(Upchurch Scientiﬁc Inc), but the outside diameter (1/ 16”) is so large that it cannot be
inserted into the reaction capillary. A portion of the PEEK capillary must be machined to
a smaller outside diameter in order to ﬁt the reaction capillary; so far we have not found a
practical way of doing that.

A number of approaches for suppressing the electroosmosis and/or minimizing wall
adsorption for fused-silica capillaries have been explored. These methods include
adjustment of the buffer pH (2-5), addition of high concentrations of ionic salts (6—8) or
zwitterions (9, 10), addition of buffer modiﬁers (11, 12), application of an external

electric ﬁeld (13-15), and permanent wall coatings (16-37).

117

Several groups have tried to eliminate adsorption by covalently bonding organic
phases to the wall, or physically coating the wall. Chemical modiﬁcation of the inner
capillary wall is the most practical method, and it includes coating the capillary inner wall
with methylcellulose (l6), chitosan (17), acrylamide (18, 19), trialkoxysilane (11, 20-23),
epoxydiols (24), maltose (25), polyethylene glycol (26, 27), poly(vinylpyrrolidine) (1),
arylpentaﬂuoro (28), polyethyleneimine (29), polyelectrolyte multilayers (30), OV-l (31),
Carbo-wax 20M (31, 32), and Plexiglas (33).

Generally speaking, capillaries coated with cross-linked polymers which are bonded
to the inner wall are stable, and provide high efﬁciency separations (34, 35). Chemical
bonding with trialkoxysilane involves blocking the silanol groups with an uniform
addition of an organic monolayer, which subsequently serves as a link for attachment of a
polymer layer. The primary purpose of the latter is to provide a “comfortable” medium
for the separated solutes, with minimum interaction.

Physical coating provides a new alternative for achieving long and narrow-bore
capillaries. Shao and coworkers from Lee ‘5 group (36) used Plexiglas to physically
modify fused silica capillary. The procedure was carried out using a dynamic coating
method at room temperature. Greatly suppressed electroosmosis was observed at pH 3-
10. Plexiglas coating shields the surface Si-OH group so that the adsorption of metal ions
can be greatly depressed. The electroosmotic ﬂow shows little dependence on the pH of
the separation buffer. The adsorption of transition metal cations has been a serious
problem at high pH (pH > 4.50). Plexiglas is resistant to alkali, dilute acids, and aqueous
solutions of inorganic salts (37). With Plexiglas-coated capillaries, strict control of buffer

pH is not necessary. Separation of transition metal cations can be performed at higher

118

buffer pH, which allows a faster post—column pH switching. No capillary preconditioning
is necessary. These approaches should be explored in the future for use with the metal
ions, CL CL system. We anticipate that adsorption to the walls of the capillary can be

eliminated or minimized.

7.2 Automatic Injection System in Capillary Electrophoresis

Linear injection devices can be a good alternative for automatic sample injection in
CB. A linear motion device sweeps lesser area than a rotary alignment device. The linear
motion device for sample injection in CE should present high repeatability, appropriate
travel distance (~ 10 mm) and travel speed (~5 mm/s). Newport Corp. provides a variety
of motorized precision translation stages and drive modules that meet these requirements.
The mid-range travel translation stage (UTM25CC1DD) gives the following features:
travel distance 25 mm, repeatability to 0.5 pm, maximum speed > 20 mm/sec, load
capacity to 20 kg. Its weakness is the high cost. The translation stage costs about $3000,
and the drive module costs $2000-3000. For easy position manipulation, we need free
adjustment in three dimensions, which can be done by mounting a motorized translation
stage via a right-angle bracket on top of a two—dimensional translation stage ($450).

Therefore, the motorized translation stage and drive module are the major cost.

7.2.1 Linear Motion Device for Alignment
In order to make the linear motion device cost-efﬁcient for use in CE as a sample
injector, it is necessary to modify the drive system by replacing the precision screw with

other methods such as an air cylinder or electromagnetic coils. In such cases, a damping

119

device (i.e., airpot from Airpot Corp.) is needed to reduce the shock upon contact of the
sample and separation capillaries. The cost of the linear injection system can be
signiﬁcantly reduced if the number of translation stages for alignment can be reduced
from three dimensions (XYZ directions) to one dimension. The translation stage for the Z
direction can be removed if the sample and the separation capillary can be set at the same
height on two different plates. A right angle V-groove can be machined so that the two
plates are in precise alignment. The sample capillary and the separation capillary can be
placed in the right angle V-groove to achieve good alignment, and a capillary holder can
be used to hold the capillary in position. In this way, only one translation is needed to
bring the sample and separation capillary into contact with good alignment for sample

injection.

7.2.2 Timing Device

The timing device described in chapter 4 counts from the moment of rotation of the
rotary alignment system, rather than the moment of contact between the sample and the
separation capillary. Therefore, we have to offset the traveling delay. It is better to time
from the moment of contact in order to reduce the timing error caused by the variation of
traveling time of the rotary alignment system. This can be accomplished by installing a

timer which can be actuated when the sample and separation capillaries are met.

120

List of References

l.

10.

ll.

12.

l3.

14.

15.

16.

17.

18.

19.

20.

Chen, M.; Cassidy, R. M. J. Chromatogr. 1992, 602, 227.

McCormick, R. M. Anal. Chem. 1988, 60, 2322.

. Lauer, H. H.; McManigill, D. Anal. Chem. 1986, 58, 166.

Walbroehl, Y.; Jorgenson, J. W. J. Microcolumn Sep. 1989, I, 41.

. Vinther, A.; Bjorn, S. E. Sorenson, H. H.; Soeberg, H. J. Chromatogr. 1990, 516,

175.
Green, J. S.; Jorgenson, J. W.; J. Chromtgr. 1989, 478, 63.
Lauer, H. H.; McManigill, D. Trent. Anal. Chem. 1986, 5, ll.
Bushey, M. M.; Jorgenson, J. W. J. Chromatgr. 1989, 480, 301.
Stover, F. S.; Haymore, B. L.; McBeath, R. J. J. Chromatogr. 1989, 470, 241.
Yeung, K. K.-C.; Lucy, C. A. Anal. Chem. 1997, 69, 3435.
Hjerten, S. J. Chromatgr. 1985, 347, 191.
Corradini, D.; J. Chromatogr. B 1997, 699, 221.
Lee, C. S.; Blanchard, W. C.; Wu, C. T. Anal. Chem. 1990, 62, 1550.
Lee, C. S.; McManigill, D.; Wu, C. T.; Patel, B. J. Chromatogr. 1991, 63, 1519.
Lee, C. 8.; Wu, C. T.; Lopes, T.; Patel, B.; J. Chromatogr. 1991, 559, 133.
Hjerten, S. J. Chromatgr. Rev., 1967, 9, 122.
Sun, P.; Landman, A.; Hartwick, R. A. J. Microcol. Sep. 1994, 6, 403-407.
Cobb, K. A.; Donik, V.; Novotny, M. Anal. Chem. 1990, 62, 2478.
Huang, X. Y.; Doneski, L. J .; Wirth, M. J. Anal. Chem. 1998, 70, 4023-4029.

Hjerten, S.; Kiessling-Johansson, M. J. Chromatogr., 1991, 550, 811.

121

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Huang, M. X.; Dubrovcakova-Schneiderman, E.; Novotny, M. V.; Fatunmbi, H.;
Wirth, M. J. J. Microcol. Sep., 1994, 6, 571-576.
Jorgenson, J. W.; Lukacs, Science (Washington, D. C. ), 1983, 222, 266.

Shao, X. W.; Shen, Y. F.; O’Neil, K.; Lee, M. J. Chromatogr. A, 1999, 830, 415.
Bruin, G. J. M.; Huiden, R.; Kraak J. C.; Poppe, H. J. Chromatogr. 1989, 480, 339.
Bruin, G. J. M.; Huiden, R.; Kraak J. C.; Poppe, H. J. Chromatogr. 1989, 480, 339.
Bruin, G. J. M.; Chang, R. H.; Kuhlman, R. H.; Zegers, K.; Kraak, J. C.; Poppe, H. J.
Chromatgr. 1988, 471, 429.)

Herren, B. J.; Shafer, S. G.; Alstine, J. V.; Hams, J. M.; Snyder, R. S. J. Colloid
Intelface Sci. 1987, 115, 46.

Ma, Y.; Hyner, K. J .; Swedberg, S. A. J. High Resolut. Chromatogr. 1991, 14, 65.
Towns, J. K.; Regnier, F. E.; J. Chromatogr. 1990, 516, 69.

Graul, T. W.; Schlenoff, J. B. Anal. Chem. 1999, 71 , 4007-4013.

Lux, J. A.; Yin, H.; Schomberg, G. J. High Resolut. Chromatogr. 1990, 13, 145.
Chen, M.; Cassidy, R. M. J. Chromatogr. 1992, 602, 227.

Shao, X. W.; Shen, Y. F.; O’Neil,K.; Lee, M. J. Microcolumn Separations, 1999, 11
(5), 325.

Gilges, M.; Kleemiss, M. H.; Schomburg, G. Anal. Chem. 1994, 66, 2038.

Ren, X.; Shen, Y; Lee, M. J. Chromatogr. A 741, 1996, 115.

Shao, X. W.; Shen, Y. F.; O’Neil,K.; Lee, M. J. Microcolumn Separations, 1999, 11
(5 ), 325.

Harrison, D. J.; Flury, K.; Seiler, K.; Fan, 2.; Effenhauser, C. S.; Mans, A. Science

1993, 261, 895.

122