Electrophoresis in open capillaries : some fundamental aspects · ELECTROPHORESIS IN OPEN CAPILLARIES Some fundamental aspects PROEFSCHRIFT ter verkrijging van de graad van doctor

Electrophoresis in open capillaries : some fundamentalaspectsCitation for published version (APA):Ackermans, M. T. (1992). Electrophoresis in open capillaries : some fundamental aspects. Eindhoven:Technische Universiteit Eindhoven. https://doi.org/10.6100/IR377599

DOI:10.6100/IR377599

Document status and date:Published: 01/01/1992

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 08. Jul. 2020

https://doi.org/10.6100/IR377599

https://doi.org/10.6100/IR377599

https://research.tue.nl/en/publications/electrophoresis-in-open-capillaries--some-fundamental-aspects(133ecf7d-1361-4649-bbab-5696374c68fd).html

ELECTROPHORESIS IN OPEN CAPILLARIES

Some fundamental aspects.

MARIETTE THEODORA ACKERMANS


Some fundamental aspects

PROEFSCHRIFT

ter verkrijging van de graad van doctor a.an de Technische Universiteit Eindhoven, op gezag van de Rector Magnificus, prof.dr. J.H. van Lint, voor een commissie aangewezen door het College van Dekanen in het openbaar te verdedigen op vrijdag 4 september 1992 te 16.00 uur

door MARIEITE THEODORA ACKERMANS

geboren te Eindhoven

Druk: Boek eq Offsetdrukkerij Letru, Helmond, 04920-37797

Dit proefschrift is goedgekeurd door de promotoren: prof.dr.ir. F.M. Everaerts

en prof.dr.ir. C.A.M.G. Cramers

co-promotor: dr.ir. J.L. Beckers

Learning witho'Ut thinking is useless, Thinking witho'Ut learning is dangerous.

Confucius

Aan mijn ouders

CONTENTS i

CONTENTS

INTRODUCTION 1 Brief history 1 Scope of this thesis 2 References 4

CHAPTER 1 PRINCIPLF.S OF ELECTROPHORESIS 5

1.1 What Is electrophoresis 5 1.2 Concept of mobility 5

1.2.1 Mobility at infinite dilution 6 1.2.2 Mobility at finite dilution and effective mobility 7

1.3 Electroosmosis 9 1.4 Different modes of electrophoresis 10

1.4. l Zone electrophoresis 10 1.4.2 Moving boundary electrophoresis 11 1.4.3 Isotachophoresis · 12 1.4.4 Isoelectric Focusing 12 1.4.5 Micellar electrokinetic capillary chromatography 12

1.5 Instrumentation 14 References 16

CHAPTER2 QUALITATIVE ASPECTS 17

2.1 Peak recognition in CZE 17 2.1.1 Introduction 17 2.1.2 The mobility of the electroosmotic flow 18 2.1.3 Measurement of the effective mobility 24 2.1.4 The effective mobility as parameter for screening 26

2.2 Determination of physlco--chemical parameters 28 2.2.1 Introduction 28 2.2.2 Theory 29 2.2.3 Calculation of pK values and mobilities at infinite dilution 30

2.3 Peak recognition in MECC 32 . 2.4 Conclusions 35 References 35

CHAPTER3 SEPARATION POWER 37

3.1 Introduction 37 3.2 Peak broadening 38

3.2.1 Introduction 38 3.2.2 Contribution due to injection 39 3.2.3 Contribution due to diffusion 39 3.2.4 Contribution due to Joule heating 40

ii CONTENTS

3.2.5 Contribution due to electrodispersion 40 3.2.6 Contribution due to electroosmosis 41 3.2. 7 Contribution due to interaction of the analytes with the

capillary wall 42 3.3 Estimation of the contributions of injection, diffusion and Joule

heating to peak broadening 42 3.3.1 Injection 42 3.3.2 Diffusion and Joule heating 44

3.4 The separation number 47 3.5 Conclusions 53 References 54

CHAPI'ER4 QUANTITATIVE ASPECTS 55

4;1 Introduction 55 4.2 Statistics 55

4.2.1 Fundamental concepts 55 4.2.2 Outliers 56 4.2.3 Calibration graphs in instrumental analysis 56

4.3 Method validation 60 4.4 Conclusions 64 References 64

CHAPTER 5 CHARACTERISTICS OF CZE 65

5.1 Introduction 65 5.2 Kohlraush regulation function 65 5.3 Quantitative analysis in CZE with conductivity and indirect UV

detection 67 5.3.1 Introduction 67 5.3.2 Theoretical 67 5.3.3 Experimental 71 5.3.4 Results and discussion 72

5.4 Conclusions 78 References 78

CHAPI'ER6 DETERMINATION OF PHARMACEUTICALS 79

6.1 Introduction 79 6.2 Determination of sulfonamides in pork meat extracts 80

6.2.1 Introduction 80 6.2.2 Experimental 81 6.2.3 Results and discussion 82 6.2.4 Conclusions 90

6.3 Determination of some P2-agonists in pharmaceuticals 90 6.3.1 Introduction 90 6.3.2 Experimental 90

CONTENTS iii

6.3.3 Results and discussion 93 6.3.4 Conclusions 99

6.4 Determination of some pharmaceuticals by MECC 100 6.4.1 Introduction 100 6.4.2 Experimental 100 6.4.3 Results and discussion 101 6.4.4 Conclusions 106

6.5 Determination of aminoglycosides with hyphenated CZE with indirect UV detection and MECC 107 6.5.1 Introduction 107 6.5.2 Experimental 108 6.5.3 Results and discussion 109 6.5.4 Conclusions 115

References 116

APPENDIX ISOTACHOPHORESIS IN OPEN CAPILLARIES 117

A.1 Introduction 117 A.2 The isotachopboretic model 119 A.3 Instrumentation 123 A.4 Variations in the EOF during ITP experiments 124 A.5 Examples of the various modes of ITP in open systems 131 A.6 Problems in quantitative analysis 134 A. 7 Conclusions 140 References 141

Summary 143

Samenvatting 145

List of symbols and abbreviations 147

Dankwoord en CutTiculum Vitae 149

Author's publications on electrophoresis 150

INTRODUCTION

BRIEF HISTORY

The term electrophoresis covers a wide group of separation methods, based on the electrophoretic principle, where charged particles move under the influence of an electric field. Due to differences in the effective mobilities of the species, separations can be achieved.

The first electrophoretic experiments were carried out by Von Reuss [l] and, in the middle of the nineteenth century, Wiedeman [2,3] and Buff [4] reported on the phenomenon that charged particles migrate in a solution when an electric field is applied. Later experiments, carried out by Lodge [5) and Whetham [6,7], were the basis on which Kohlraush [8] developed a theory of ionic migration. With the equation he derived, all electrophoretic methods can be described, including zone electrophoresis, moving boundary electrophoresis, isoelectric focusing and isotachophoresis. The interest in the electrophoretic work was greatly stimulated by the discoveries of Hardy [9,10], that many biocolloids, such as proteins, show characteristic mobilities that depend largely on the pH of the electrolyte system used for their analysis. Owing to the major interest in compounds such as proteins and enzymes, and because of the lack of high-resolution detectors, most attention was paid to one of the basic techniques already described by Kohlraush [8], viz., zone electrophoresis. So far, flat-bed and slab gel techniques, used with the two-dimensional method of O'Farell [11], provide for high resolutions and are the most commonly used methods in the analysis of biomacromolecules. For the physico-chemical characterization of proteins, mainly disc-electrophoresis in poly acrylamide gel rods [12] was used. However, regardless of the good separation results achieved, these electrophoretic techniques in supporting media are far removed from the instrumental developments common in chromatography, their main drawbacks being the laborious, multi-stage handling of gel beds and insufficient long-term reproducibility of results.

The first successful electrophoretic analyses in free solution were performed by Hjerten [13], introducing rotating tube electrophoresis, using a 1-3 mm I.D. tube as separation channel. Another successful attempt in the field of capillary zone electrophoresis was that of Virtanen [14,15,16]. In this experiments glass capillaries of 0.2-0.5 mm I.D. were used, and in his work, the theory of capillary zone electrophoresis was treated both theoretically and experimentally.

2 INTRODUCTION

A further progress in capillary zone electrophoresis was connected to the development of capillary isotachophoresis, where in 1979 Mikkers [17] published the experimental evaluation of the non-diffusional model of concentration distributions in free zone electrophoresis using a laboratory made isotachophoresis equipment, and showed that rapid and highly efficient separations; reaching the theoretical predictions as made by Giddings [18], were possible [19]. In 1981 Jorgenson and Lukacs [20,21,22,23] used open glass capillaries with an I.D. of only 75 µm as separation column. On applying voltages up to 30 kV across a capillary of 1 m equipped with a fluorometric detector, they separated and detected fluorescamine derivatized amino acids and peptides. The efficiency of the separations reached the limits of the theoretical model where diffusion was considered as the only dispersion effect. These recent papers, combined with the fact that commercially apparatus for capillary zone electrophoresis were available, probably initiated the great attention which is now being paid to capillary zone electrophoresis.

SCOPE OF TIIIS THESIS

With the growing interest in capillary zone electrophoresis several companies and institutes are interested in the possibilities of this new separation technique for their specific applications. The research described in this thesis is based upon the question of the Dutch State Institute for Quality Control of Agricultural Products, asking what are the screening possibilities of capillary electrophoresis for the determination of veterinary drugs? As the group of veterinary drugs includes charged as well as uncharged components, with and without UV absorption, capillary zone electrophoresis (CZB), as well as micellar electrokinetic capillary chromatography (MECC) are investigated.

If components in complex matrices have to be determined, the following questions have to be answered: 1. Does the component migrate in a chosen electrolyte system, and what parameter

can be used to recognize it? 2. Is the separation capacity of the method sufficient to separate the component

adequately from the other components of the complex matrix? 3. Is a quantitative determination possible and at what level can the component be

detected?

In this thesis, several aspects of capillary electrophoresis, which have to be taken into account to answer these questions, are discussed. Chapter 1 deals with the principles of electrophoresis. Because the concept of mobility plays an important part in electrophoresis, this concept and some effects, affecting the mobility, such as the retardation and relaxation effects and the influence of the pH are described. Further, some fundamentals with respect to the electroosmotic flow (BOF) are described. Different modes of electrophoresis are discussed, and finally the instrumentation used for the CZE and MBCC experiments from this thesis, is described.

In chapter 2 the qualitative aspects of capillary electrophoresis are described. It is shown that migration times or apparent mobilities can never be used for the identification of ionic species in capillary electrophoresis if an electroosmotic flow is present, because the velocity of this flow varies considerably with the "state" of the

SCOPE OF THIS THESIS 3

capillary. The effective mobility, however, which can be calculated from the migration times of the ionic species and the BOP, is constant and nearly independent of the concentration of the ionic species, and can be used as parameter for screening. From the effective mobilities of a component in two different electrolyte systems, at which the degree of dissociation of the component differs sufficiently, the mobility at infinite dilution and the pK value of the component can be calculated. For MECC, a pseudoeffective mobility, analogous to the effective mobility of CZE can be used as parameter for screening.

The most important phenomena causing peak broadening in capillary electrophoresis are Joule heating, electroosmosis, diffusion, electrodispersion, injection and interactions of the solutes with the capillary wall. In chapter 3 these phenomena are described briefly and an estimation is made of the magnitude of each of these effects causing peak broadening in a CZE experiment. Further, a separation number,

· analogous to the separation number in gaschromatography (GC) is defined for CZE, showing that the use of this separation number gives a good indication whether components can be separated or not.

On applying a new analytical method such as CZE, the accuracy and precision of the method have to be evaluated, and the method must be validated. In chapter 4 a concise summary is given of the statistics, dealing with the accuracy and precision of a separation method, the validation of the method, calibration graphs and limits of detection, and a method validation of CZE with respect to HPLC and ITP is given.

In chapter 5, some characteristics of capillary electrophoresis are given. In 1897 Kohlraush formulated his so-called "Beharrliche Funktion", which prescribes that during the electrophoretic separation, at any point the sum of the concentrations divided by the absolute values of the mobilities must be constant. Using this equations, it is shown that for CZE with indirect UV detection or conductivity detection, the use of an internal standard allows the accurate determination of a concentration in a mixture without separate calibration of the response for each component.

In chapter 6, finally, the determination of several pharmaceuticals with different modes of capillary electrophoresis is given. To investigate the applicability of capillary electrophoresis for the determination of drugs in several matrices, groups of components with different properties were selected to show the applicability of CZE with UV and indirect UV detection and MECC. Sulfonamides were determined in pork meat extracts with CZE in the cationic mode with UV detection. Qualitative and quantitative aspects of CZE in the cationic mode with UV detection are compared with those of HPLC and ITP, using some iSragonists in pharmaceutical dosage forms as model components. Some typical neutral drugs are determined using MECC and, finally, some aminoglycosides and neutral components in combined pharmaceuticals are analyzed in an electrolyte system for hyphenated CZE with indirect UV detection and MECC.

One of the rnain problems of capillary electrophoresis with UV detection at this moment is the limit of detection which can be reached. For most applications this limit of detection is too high, which means that a sample pre-concentration is required. In isotachophoresis, a sample concentration always occurs using samples with low concentrations, as can be deduced from Kohlrausch's regulation function. Bearing in mind that isotachophoresis can be used possibly as an on-line sample pre-concentration method prior to CZE, the possibilities of isotachophoresis in open capillaries are studied. The results are given in the appendix. In open systems, if an electroosmotic

4 INTRODUCTION

flow is present, four modes can be defined for ITP, viz., the anionic, the cationic, the reversed anionic and the reversed cationic mode. The applicability of each of these modes depends upon the mobilities of the ions of the electrolyte systems used and the mobility of the EOF which is present. Because the velocity of the electroosmotic flow depends also on the composition of the electrolyte in the capillary, the electroosmotic flow will change when the terminating electrolyte migrates into the capillary during ITP experiments. This makes quantitative analysis with ITP in open systems troublesome. Generally, closed systems without EOF or open systems with fully suppressed EOF are to be preferred for ITP analyses.

References

1. F. von Reuss, Comment. Soc. Phys. Univ. Mosquensem, 1(1808)141. 2. G. Wiedeman, Popp. Ann., 99 (1856) 197. 3. G. Wiedeman, Popp. Ann., 104 (1858) 156. 4. H. Buff, Ann. d. Chemie und Phar., 105 (1858) 168. 5. O. Lodge, Brit. Ass. Advan. Sci. Rep., 56 (1886) 389. 6. W.C.D. Whetham, Phil. Trans. Ruy. Soc. London, Ser. A, 184 (1893) 337. 7. W.C.D. Whetham, Phil. Trans. Roy. Soc. London, Ser. A, 184 (1893) 507. 8. F. Koblraush, Ann. Phys. (Leipzig), 62 (1897) 209. 9. W.B. Hardy, Proc. Ruy. Soc. London, 66 (1900) 110. 10. W.B. Hardy, J. Physiol. (London), 33 (1905) 251. 11. P.H. O'Farell, J. Biol. Chem., 250 (1975) 4007. 12. N. Castimpoolas, in: Electrophoresis, A survey of Techniques and Applications, Pan I:

Techniques, Z. Deyl (Ed.), Elsevier, Amsterdam, The Netherlands, 1979, p. 167. 13. S. Hjerten, Arkiv. Chem.,13 (1958) 151. 14. R. Virtanen and P. Kivalo, Suomen Kemistilehti, B (1969) 182. 15. R. Virtanen and V. Nint<S, Scand. J. Clin. Lob. Invest., 21 (1971) 27. 16. R. Virtanen, Acta Polytech. Scand., 123 (1974) 1. 17. F.E.P. Mikkers, F.M. Everaerts and Th.P.E.M. Verheggen, J. Chromatogr., 169 (1979) 1. 18. J.C. Giddings, Separ. Sci., 4 (1969) 181.. 19. F.E.P. Mikkers, F.M. Everaerts and Th.P.E.M. Verheggen, J. Chromatogr., 169 (1979) 11. 20. J.W. Jorgenson and K.D. Lukacs, Anal. Chem., 53 (1981) 1298. 21. J.W. Jorgenson and K.D. Lukacs, J. Chromatogr., 218 (1981) 209. 22. J.W. Jorgenson and K.D. Lukacs, Clin. Chem., 27 (1981) 1551. 23. J.W. Jorgenson and K.D. Lukacs, J. High Res. Chromatogr. Chromatogr. Commun.,

4 (1981) 230.

CHAPTER 1

PRINCIPLES OF ELECTROPHORESIS

The concept of mobility plays an important part in electrophoresis. Therefore the concept of mobility and some effects, affecting the mobility, such as the retardation and relaxation effects and the influence of the pH are described. Further, some fundamentals with respect to the electroosmotic flow are described. Different modes of electrophoresis are discussed and finally the instrumentation used for the CZE and MECC experinients of this thesis is described.

1.1 WHAT IS ELECTROPHORESIS?

If an electric field is applied over a solutioQ in which ions are present, these ions will start to move to the electrode with the opposite charge, with a velocity which is proportional to the electric field. The velocity per unit field strength is called the mobility:

v m== E

(1.1)

where m is the mobility, v is the velocity and E the electric field strength. The mobilities are taken positive for cations and negative for anions. The mobility is characteristic for each component in a certain solvent, and depends upon several factors such as charge, radius and degree of dissociation. The separation principle by · which charged components are separated in an electric field is called electrophoresis.

1.2 CONCEPT OF MOBILITY

The concept of mobility plays an important part in electrophoretic techniques, because differences in the effective mobilities determine whether ionic species can be separated or not.

6 CHAPTER 1: PRINCIPLES OF ELECTROPHORESIS

1.2.1 Mobility at infinite dilution

If an electric field is applied to an electrolyte solution, the charged particles will start to move and finally, a steady state will be reached in which the velocity of the charged particles, in the direction of the field, is constant in the time. At infinite dilution, two different forces act on the particle. The first force, Fi , is a force exerted on the charge of the particle. It can be denoted by:

(1.2)

where 'Lis the charge of the particle and E is the electric field strength. The second force, F2 , is a friction force, which Stokes determined for a rigid spherical particle as:

(l.3)

where v is the electrophoretic velocity, r is the radius of the particle, 71 is the viscosity of the solvent and fc is the friction factor. For non-spherical particles a correction factor has to be introduced, to allow for size shape and orientation of the particle.

In the steady state, the sum of the forces must be zero:

(1.4)

and combination of eqn. 1.1 and 1.4 gives for the mobility at infinite dilution:

(1.5)

The mobility is related to the molar ionic conductivities. The molar conductivity, denoted by A, is the conductivity of a molar weight of an electrolyte, measured in a conductance cell of which the distance between the electrodes is 1 cm, and the volume contains exactly one mole of the electrolyte. Kohlrausch showed that at a fixed temperature, the relationship between the molar conductivity of an electrolyte and the square root of the concentration is nearly linear, especially at very low concentrations and for strong electrolytes. At infinite dilution, the equivalent conductivities can be interpreted in terms of ionic contributions, whereby the contribution of an ion is independent of the other ionic species of the electrolyte:

(1.6)

where >-i and >..? are the molar ionic conductivities of the cations and the anions respectively, v+and v_ are the numbers of cations and anions per formula unit of electrolyte, and A 0 is the molar conductivity, all at infinite dilution.

At infinite dilution, the relationship between the molar ionic conductivity and the mobility for 1:1 electrolytes is given by:

1.2 CONCEPT OF MOBILITY 7

(1.7)

where Fis Faraday's constant. It can be concluded from eqn. 1. 7 that mobilities at infinite dilution can be

calculated by dividing the molar ionic conductivities at infinite dilution by the Faraday constant. The molar ionic conductivities can be obtained by measuring the transport numbers. As the transport numbers give the fractions of the total current carried by each ion:

(1.8)

where t is the. transport number. For several ions in several different solvents, values of the equivalent ionic conductivity and/or transport number can be found in literature [1,2,3].

1.2.2 Mobility at finite dilution and effective mobility

At finite dilution, due to the presence of op~itely cE.arged particles, forming a so called ionic atmosphere, two more forces, F3 and F4 act on the particle. In Fig. 1.1 a schematic illustration is given of the electrophoretic and reJaxation effects.

b

Fig. I.I: Schematic illustration of the electrophore1ic and relaxation effects. (a) In the absence of an applied field, the ionic atmosphere is spherically symmetric, but {b) when an elet;tric field is present it is distoned and the centres of negative and positive charge no longer coincide. This retards the motion of the central ion.

The central cation will be surrounded by an ionic atmosphere having an average charge opposite to the charge of the cation. When no external field is applied to the


solution, the ionic atmosp]Jere has a sphericaJ.. symmetry around the cation (Fig. l. la). When the external field E is switched on, F1 acts on the ion. The central ion and its ionic atmosphere move in opposite directions towards the electrodes (Fig. 1. lb). This counter current of ions produces a decrease in the effective velocity of the cation due to two effects. The first of those effects, the electrophoretic effect, is due to the fact that the electric field exerts a force on the ions of the ionic atmosphere, which is transferred to the molecules of the solvent. The particle considered does not move through a stationary ~lvent, but thro~h a solvent moving in the opposite direction resulting in a force, F3 , opposite to f 1. The second effect, represents the relaxation effect, which is the result of the fact that when an electric field is applied the particle moves away from the centre of the ionic atmosphere. To rebuild this atmosphere in its proper place talces a finite time, called the relaxation time. Thus the centre of the particle constantly lags behin_q the centre of the particle in the stationary state, resulting in an electric force, F4 , on the charge of the particle.

For the quantitative formulation of these effects, Debye, Hiickel and Onsager [4] derived an expression for the correction of the conductivity for relaxation and retardation effects:

(1.9)

where Arel and Aret are corrections for the decreasing effects on the conductivity due to relaxation and retardation effects respectively.

When using the equations for Arel and Aret as .described [4] and rewriting eqn. 1.9 the following equation for A is derived for 1: 1 electrolytes:

A=A0 -(A+BA°'){C (1.10)

where

A = z2ep2 (~) j and B = qz3ep2 [-2-J j 311"17 eRT 24TeRT 7reRT

(l.11)

and c is the concentration of the species and q=0.59 for 1:1 electrolytes. If for a certain solvent 11 and e are known, these factors can be calculated. With these factors, a correction factor, 'Y· can be calculated, which relates the mobility at finite dilution to the mobility at infinite dilution, according to:

(1.12)

In practice, one is not working with mobilities at infinite or finite dilution, but with effective mobilities. The effective mobility can be calculated from the mobility at infinite dilution, correcting for relaxation and retardation effects, and correcting for the degree of dissociation of the component. Tiselius [5] pointed out that the effective mobility, m, was the summation of all products of the degree of dissociation and the mobilities at finite dilution:

1.3 ELECTROOSMOSIS 9

(1.13)

where ai is the degree of dissociation ..

1.3 ELECTROOSMOSIS

One of the fundamental aspects one has to deal with using capillary electrophoresis in open tubes, is electroosmosis. Electroosmosis is the phenomenon that liquids in a capillary start to flow in an electric field, which is a consequence of the surface charge of the capillary. The positive ions in the diffuse double layer move towards the cathode, but as these ions are solvated, they will drag along the solvent molecules. The result is a flow of the electrolyte solution towards the cathode. In Fig. 1.2 a schematic representation of a capillary wall and the double layer of the electrolyte solution is given.

Fig. 1.2: Schemaric representation of a capillary wall and the double layer of the electrolyte solution.

If in a fused silica capillary the wall is in contact with a buffer solution, the capillary wall is negatively charged, due to the ionization of functional surface groups. This negatively charged wall attracts from the solution positively charged ions and as a result a positively charged double layer is formed. For the double layers, several models are set up, viz., the Helmholtz model, which is to suppose that the solvated ions range themselves along the surface of the capillary and are held away from it only by the presence of their hydration spheres, and the Gouy-Chapman model, in which a diffuse double layer is emphasized. Neither the Helmholtz nor the Gouy-Chapman model is a good representation of the structure of the double layer, both taking only one part of the double layer into account. Stem formulated a model in which the ions closest to the wall are constrained into a rigid plane as in the Helmholtz model, called


the Stern layer, while outside that plane the ions are dispersed in a diffuse double layer, as in the Gouy-Chapman model, consisting of moving ions, but with a net positive charge (see Fig. 1.2). The potential at the boundary between those two regions is called the r potential. The r potential can be calculated according to the Helmholtz-Smoluchowski equation:

(1.14)

The value of the S- potential depends on factors such as surface material, pH and concentration of the electrolyte solution, solvent and surface active additives.

The value of the electroosmotic velocity depends on the S- potential as can be seen from eqn. 1.14, the higher the r potential, the higher the electroosmotic flow. Analogous to the mobilities for ions, a mobility for the electroosmotic flow m50p can be defined, using vEOF as given in eqn. 1.14:

(1.15)

If the diameter of the capillary is much bigger than the thickness of the double layer and no hydrodynamic flow and radial temperature gradient are present, the flow profile of the electroosmotic flow will be like a plug profile.

1.4 DIFFERENT MODES OF ELECTROPHORESIS

On applying the electrophoretic principle, four main techniques can be distinguished: 1. Zone electrophoresis: which can be compared with the elution principle in

chromatography; 2. Moving boundary electrophoresis: as the analogue of chromatographic frontal

analysis; 3. Isotachophoresis: the electrophoretic displacement principle; 4. Isoelectric focusing: comparable with chromatofocusing. In Fig. 1.3 the different separation techniques are given schematically.

Combinations of these principles can also be used. Beckers and Everaerts [6, 7] described a way to combine isotachophoresis and zone electrophoresis in such a way that in one run some components migrate isotachophoretically, and other zone electrophoretically. Micellar electrokinetic capillary chromatography is a hybrid technique, which is in fact a combination of the separation principle of chromatography and electrophoresis. In this section the various modes of electrophoresis and micellar electrokinetic capillary chromatography will be discussed.

1.4.1 Zone electrophoresis

In zone electrophoresis (see Fig. l.3a) the components are allowed to migrate in spatially separated zones. Such a separation is achieved by filling all compartments

1.4 DIFFERENT MODES OF ELECTROPHORESIS

a b c

0 0 0 0 0 0

AS AB

AB I j B

A A

0 0 0 0 0 0

d

0 0

A8

B

l 0 0

11

Fig. 1.3: Schematic representation of the separation techniques of (a) 1.0ne electrophoresis, (b) moving boundary electrophoresis, (c) isotachophoresis and (d) isoelectricfocusing.

of the equipment with one electrolyte, the so called background electrolyte. The sample containing the components is introduced as a small band in the background electrolyte. Applying a high voltage the components will move, according to their effective mobility, with their own velocity depending on the local experimental conditions. After an appropriate time the components will move in different zones, spatially separated by zones of the background electrolyte. Each zone moves with its own velocity in a superimposed configuration on the background electrolyte. Zone electrophoresis can be carried out on anticonvective media such as paper, gels of agar and poly acrylamide, cellulose acetate and also in capillaries. If zone electrophoresis is carried out in capillaries with an electroosmotic flow, positively as well as negatively charged components can be separated in one run. This electrophoretic technique is comparable with elution techniques in chromatography and is probably the most popular electrophoretic technique in use.

1.4.2 Moving boundary electrophoresis

In moving boundary electrophoresis (see Fig. 1.3b), the sample is introduced at the beginning of the separation compartment, which is filled completely with one electrolyte, the so-called leading electrolyte. To guarantee a constant sample feed the sample. compartment should be large. The leading electrolyte consists of a leading component, which is an ion with the same charge sign as the components which have to be separated and a higher effective mobility, and a counter ion with the opposite charge, to preserve electroneutrality. The polarity of the electric field should be chosen in such a way that the components move in the direction of the leading electrolyte. Applying a voltage the moving boundary process begins, characterized by the separation of the most mobile component and mixed zones for all other components. As an analytical method moving boundary electrophoresis has only limited value, but

12 CHAPTER 1: .PRINCIPLES OF ELECTROPHORESIS

it should be born in mind that the moving boundary principle is present in almost every electrophoretic process in the initial phase.

1.4.3 Isotachopboresis

In isotachophoresis (ITP)(see Fig. l.3c) a discrete amount of sample is applied at the interface between two electrolyte systems: the leading electrolyte and the terminating electrolyte. In ITP in one run only anions or cations can be separated and in its most simple configuration, both the leading and the terminating electrolyte contain only one ionic constituent of the same charge sign as the sample components and a counter constituent to preserve electroneutrality and to obtain a buffering capacity. The effective mobility of the leading ion should be higher than any of the sample components, and that of the terminating ion should be lower. If the steady state is reached, all sample components will be arranged in consecutive zones, generally in order of their mobilities. Provided that the current is constant, all zones will migrate with equal and constant velocity. The velocity of each sample component zone is equal to that of the leading electrolyte zone, and as a result the electric field strength in each zone is inversely proportional to the effective mobility of the sample components. Since mobilities are component dependent properties, the measurement of the electric field strength or its related parameters can be used for the detection. According to the Kohlrausch regulation function, the concentrations in the zones are adapted to the concentration of the leading electrolyte and the zone boundaries have self-correcting properties against convective disturbances. Within a zone the concentration is constant and the zone length provides the quantitative information. The step height, which is determined by the effective mobilities of the leading· and the sample ion, contains qualitative as well as quantitative information [8,9].

1.4.4 IsoeJectric Focusing

In isoelectrlc focusing (see Fig. 1.3d) the column contains a buffer solution, that creates a pH gradient in the tube. When a sample, consisting of a mixture of arnphiprotic molecules, with particular pi values, is introduced, the particles will move until they reach the pH in the tube equal to their pi value. At this point the effective mobility of the component will be zero and the component will no longer move. In the steady state the particles will be separated if their pi values differ, according to their pi value. During the separation process the effective mobilities of the components decrease as the approach the pH value of their pi point. As a result often rather long analysis times are needed. The main field for the application of isoelectric focusing is in the separation of proteins.

1.4.5 Micellar electrokinetic capillary chromatography

Since the introduction of micellar electrokinetic capillary chromatography (MECC) by Terabe and co-workers [10,11], it has proven to be a highly efficient separation method. Amongst others, many compounds of pharmaceutical interest have been separated by MECC, such as vitamins, cephalosporins, penicillins, antipyretic

1.4 DIFFERENT MODES OF ELECTROPHORESIS 13

and analgesic preparations, barbiturates and optical isomers of drugs [12,13, 14,15,16, 17, 18].

In MECC a surfactant is added to the background electrolyte, in a concentration above its critical micelle concentration (CMC), which means that micelles are formed. The interior of the micelles is hydrophobic, because of the presence of the hydrophobic tails of the surfactant molecules and the exterior is hydrophylic due to the charged functional groups of the surfactant molecule. These micelles can be considered as dynamic structures, which are in· equilibrium with the monomers of the surfactant, which are present in the background electrolyte. So far, the negatively charged sodium dodecyl sulphate (SDS) is the most common used surfactant for MECC. The principle of a MECC separation is shown in Fig. 1.4.

capillary wall

l solute

0l'v

~.,_- esolut•_ ~ 8 ~$rfa3t EOF

molecule

Fig. 1.4: The principle of a MECC separaiion for negatively charged (e. g. SDS) micelles.

The separation system in MECC consists of two mobile phases viz., an aqueous phase and a micellar phase, of which the interior is lipophilic. The surface of the micelles is highly negatively charged, through which the micelles exhibit an effective mobility in the direction of the anode. In most cases using fused silica capillaries, the electroosmotic flow is directed to the cathode.· If the mobility of the EOF exceeds the absolute value of the effective mobility of the micelles, the micelles will have a net · velocity in the direction of the cathode, resulting in a fast moving aqueous phase, and a slower moving micellar phase, both in the direction of the cathode.

Uncharged solute molecules will move in the direction of the cathode, with the velocity of the EOF if they are in the aqueous phase, and with the net velocity of the micelles, when they are solubilized in the micelles. The separation between different solutes is achieved if the partition coefficient of those two over the aqueous and the micellar phase is different. Totally unsolubilized components will migrate with the velocity of the EOF and will reach the detector first. Totally solubilized components will migrate with the velocity of the micelles and will reach the detector last. All other components will migrate with a velocity between those two and reach the detector in a time between that of the EOF and the micelles. Remember that this is only the case


for uncharged components. If charged particles are present they can migrate of course somewhere outside this time window (see section 6.5).

1.5 INSTRUMENTATION

Generally, electrophoretic equipment consists of six modules, viz., an anode and a cathode compartment connected to each other by a separation unit, an injection module, a high voltage power supply, and a detector. Injection can be achieved hydrodynamically, with pressure or by electromigration. In Fig. 1.5 a schematic representation of the equipment is given.

power

supply

"'-capillary

detector

electrolyte vessel

Fig. 1.5: Schematic representation of electrophoresis equipment.

For all electrophoretic experiments performed in open systems, described in this thesis, the Beckman P/ ACE 2000™ HPCE system (Beckman Instruments, Palo Alto, California, U.S.A.) is used [19]. This is a fully automated electrophoresis equipment, which contains a temperature regulated capillary column, a high voltage power supply, an automated sample-injection system, and a UV absorbance detection system with associated data collection equipment. In Fig. 1.6 a schematic representation of the P/ ACE 2000 HPCE system is given.

Temperature control The fused silica capillary tube (20 to 100 cm long) is located within a

temperature regulated cartridge (15 to 500 C).Temperature regulation is accomplished by surrounding the capillary tubing with a thermostated heat transfer fluid [20] that is recirculated through the cartridge.

High voltage power supply The system is equipped with a high voltage power supply that is capable of

delivering up to 250 µ.A at potentials ranging from 1 to 30 kV. The high voltage

1..S INSTRUMENTATION .

mirror ---

photomultiplier tube

deuterium lamp

mirror

vial trays

Fig. 1.6: Schematic representation of the PIACE 2000 HPCE system

l.S

power supply is connected to the separation channel by means of platinum electrodes positioned adjacent to the capillary inlet and outlet. The polarity of the electrodes is easily reversible.

Sample tra.ys and sample introduction Sample solutions, electrolyte solutions and capillary conditioning solutions are

brought to the capillary inlet and outlet by randomly accessed fluid trays that rotate just below the thermostated capillary cartridge. One rotating tray is assigned to the inlet and can contain 24 vessels, the other is assigned to the outlet and can contain 10 vessels. Once the trays have been rotated in the appropriate positions, the vials are pneumatically lifted and the capillary tips and platinum electrodes are immersed in the appropriate solutions.

Two modes of sample introduction are possible. A small plug of sample solution may be driven into the capillary either by applying a slight head pressure to the sample for a carefully controlled period of time, or by simple electromigrating a small portion under the influence of a relatively low voltage for a short time interval.

Specially designed vial caps serve to seal and mate the vials with the capillary cartridge interface and to reduce evaporative loss of sample solutions.

Sample detection A selectable wavelength UV absorbance detector monitors a short section of the

separation medium near the capillary outlet. Light emitted from the deuterium lamp source is directed through an eight-position filter wheel and onto a small aperture located on the capillary cartridge. A well-defined beam of filtered light is then passed through the capillary and collected by an optical fiber coupled to a photomultiplier tube. Alignment of the capillary detection volume with the source and the detector is accomplished automatically as the capillary cartridge is installed to the unit. The PI ACE 2000 HPCE is standard equipped with a 214, 254 and 280 nm filter.


Data collection and handling The capillary electrophoresis system is provided with both analog and RS232

output, and may be connected to a stripchart recorder or integrator for stand-alone operation, or to a microcomputer for data acquisition. Automatization is accomplished by using the windows-based PI ACE control software. Data collected with this software were analyzed with the laboratory-written data analysis program CAESAR, which was developed by ir. B.J. Wanders.

References

1. Landolt-BOmstein, Zahlenwerte und Funktionen aus Physik, Chemie, Astronomie, Geophysik und Technik, K.H. Hellwege und A.M. Hellwege, (Eds.), Springer Verlag, 1959.

2. R. Femandez-Prini, in: Physical chemistry of Orgatlic solvents. A.K. Covington and T. Dickinson (Eds.), Plenum Press, London aod New York, 1973, p. 525.

3. M. Spiro, in: Physical chemistry of Organic solvents. A.K. Covington and T. Dickinson (Eds.), Plenum Press, London and New York, 1973, p. 615.

4. H. Falkenhagen, Elektrolyte, Verlag von S. Hirzel, Leipzig, 1932. S. Tiselius, Nova Acta Reg. Soc. Sve. Sd, Upsal., 4,7 no 4 (1930). 6. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 508 (1990) 3. 7. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 508 (1990) 13. 8. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 470 (1989) 277. 9. M.T. Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 595 (1992) 327. 10. S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchiya and T. Ando, Anal. Chem., 56 (1984) 111. 11. S. Terabe, K. Otsuka and T. Ando, Anal. Chem., 57 (1985) 834. 12. S. Fujiwara, S. Iwase and S. Honda, J. Chromatogr., 447 (1988) 133. 13. D.E. Burton, M.J. Sepaniak and M.P. Maskarinek, J. Chromatogr. Sci., 26 (1988) 406. 14. H. Nishi, N. Tsumagara and S. Terabe, Anal. Chem., 61 (1989) 2434. 15. Y. Miyashita, S. Terabe and H.Nishi, Chromatogram 11 (1990) 7. 16. S. Fujiwara and S. Honda, Anal. Chem., 59 (1987) 2773. 17. H. Nishi, T. Fukuyama, M. Matsuo and S. Terabe, J. Microcolumn., Sep. 1(1989)234. 18. A. Wainright, J. Microcolumn. Sep., 2 (1990) 166. 19. V.P. Burolla, S.L. Pentoney and R. Zare, Am. Biotechnol. Lab., 7(10) (1989) 20. 20. P. Gozel E. Gassman, H. Michelsen and R.N. Zare, Anal. Chem., 59 (1987) 44.

CHAPTER2

QUALITATIVE ASPECTS

Migration times or apparent mobilities can never be used for the identification of ionic species in capillary zone electrophoresis if an electroosmotic flow is present, because the velocity of this flow can vary considerably with the "state• of the capillary. From the migration times of the EOF and the ionic species, the effective mobilities can be calculated. These effective mobilities are nearly independent of the concentrations of the sample ionic species. Although a large excess of one of the sample components can cause different values of the calculated effective mobility owing to a distorted zone electrophoretic process, they are reproducible if the matrix has a constant composition and in this way effective mobilities can be used for screening purposes. In the determination of effective mobilities the use of a "true• EOF marker is extremely important.

If effective mobilities are measured in two different.electrolyte systems at different pH values, at which the degrees of dissociation differ sufficiently, the mobilities at infinite dilution and pK values of ionic species can be calculated. On this way values obtained for mobility and pK are compared with data obtained isotachophoretically, showing a good agreement.

In micellar electrokinetic capillary chromatography, two parameters can be used for peak recognition. The first one is a capacity factor analogous to the capacity factor from chromatography, the second is a pseudo-effective mobility, analogous to the effective mobility in capillary zone electrophoresis. In the last section the use of both parameters is discussed and the advantages of the use of the pseudo-effective mobility over the use of the capacity factor is shown.

2.1 PEAK RECOGNITION IN CZE

2.1.1 Introduction

If in CZE the applied voltage, V, and the length of the capillary, le, are known, the electric field strength is:

v E=lc

(2.1)

If the distance from the injection point to the detector, id, and the migration time, t, of a component are known, the velocity, v, and the apparent mobility, mapp• can be calculated using [l]:

18 CHAPTER 2: QUALITATIVE ASPECTS

(2.2)

In CZE an EOF can act, generally in the direction of the cathode using silica capillaries. The velocity of the EOF can be determined using the "migration" time, tEOF• of an uncharged substance and because this velocity shows a linear relation with E the mEoF can be defined as:

Id lcld mEoF= --=-- (2.3)

tEoFE tEoFV

and the effective mobility of a component can be obtained by:

(2.4)

If the absolute values of the effective mobilities of anionic species are smaller than mEOF• they can be determined in the Upstream Mode (US) simultaneously with cations migrating in the Downstream Mode (DS).

2.1.2 The mobility of the electroosmotic flow

From eqn. 2.4 it can be concluded that the effect of the EOF is extremely important in the determination of effective mobilities. In Fig. 2.1 the calculated

ggffi~g~~o

1000

r 0

mobility (10""cm•tvsi injection

Fig. 2.1: Calculated relmionship between migration time and effective mobilily far several values of mEoF applying an E gradient of 25 kV!m. 1he different lines are marked with a number, representing the mobilily meop·IO' of the EOF (cm2Ns). At large EOF velocities negative ions can be analyzed in the Upstream Made. From this relationship electropherograms can be deduced, as shown for an mEOF of 30·UI1 cm2Ns.

2.1 PEAK RECOGNITION IN CZE 19

relationship between migration time and effective mobility for an E gradient of 25 kV /m and an le and Id of 1 m is given for several values of mEOF· It can be clearly seen that at low EOF only cations can be determined whereas at high EOF simultaneously anions in the US can be determined with Im I <mEoF· A disadvantage at high EOF is, however, that the separation power for cations diminishes (see section 3.4).

Choice. of the EOF marker For the calculation of effective mobilities from the apparent mobility and the

EOF, the velocity of the EOF has to be precisely known. There are several ways to measure the EOF. In Fig. 2.2 some possibilities for the cationic mode are shown schematically. The real EOF displacement is indicated with an arrow. Using a neutral EOF marker, it is possible that this marker indicates the EOF displacement (2). If the marker, however, meets a power of attraction from the capillary or if it is partially negatively charged by complexation with negative ions of the background electrolyte, it will be too slow (1), or it will be too quick if it is positively charged by complexation (3).

EOF displacement

i 1 mi!

2

3 'ml EOF marker

I A-+ UV dip s--r c-+

I

Fig. 2.2: Several possibilities of measuring EOF indications. For junher expkmation see text.

Because many electrolyte systems absorb UV light at a wavelength of 214 nm, sample solutions with a lower buffer concentration than the background electrolyte show a dip in the UV signal, because the local concentration of the buffer is lower. If an aqueous sample solution is introduced, the original concentration (UV) dip can indicate the correct EOF displacement (B), but because the shape of the concentration dip can change owing to diffusion effects, the shape can become asymmetric [2,3] to one of the sides (A or C), depending on the mobilities of the background ionic species, indicating the wrong EOF.

In first instance (UV detection at 214 and 254 nm) acetone, benzene, crotonaldehyde, mesityl oxide (MO) and paracetamol were compared as EOF marker with a background electrolyte 0.02 M TRISIMOPS at pH 8.2. The best results (high


a A v b

c

----> time (arb. units)

Fig. 2.3: Measured UV signal/or a sample solution of (a) O. 0001 M mesityl oxide in 100 % buffer, (b) a mixture of 1 % water and 99 % buffer solution and (c) 50 % water and 50 % buffer solution. Background electrolyte was 0.02 M HIST/MES at pH 6.2 • Capillary from SGE, I.D. 72 µm, lc""'S6.55 cm and l4=49.60 cm. Pressure injection time, 5 s.

TABLE2.I

MEASURED MIGRATION TIME, 1 (min) AND MOBILITY OF THE EOF, mmp·la5 (cm21Vs), USING A 0.02 M HIST/MES BACKGROUND ELECTROLYTE AT pH 6.2

Capillary from SGE, I.D. 72 µm, le = 56.55 cm and Id = 49.60 cm. Pressure lltjection time, 5 s. Applied voltage, 25 kV.

No. 0.0001 M MO in 100% background 1 % water and 99 % background electrolyte electrolyte

mEOF mEOF

1 3.55 52.67 3.55 52.67 2 3.54 52.82 3.55 52.61 3 3.56 52.53 3.56 52.53 4 3.56 52.53 3.55 52.67 5 3.56 52.53 3.55 52.67

absorbance and symmetrical peaks) could be obtained using MO as EOF marker. In all further experiments of this chapter a wavelength of 214 nm was used.

In Fig. 2.3 the measured UV signal for a sample solution of (a) 0.0001 M MO in 100 % buffer, and mixtures of (b) 1 % water and 99 % buffer and (c) 50 % water and 50 % buffer are shown. In Table 2.I the measured migration times and calculated mEOF are given for the 0.0001 M MO solution in 100% buffer and the mixture of 1 % water in 99% buffer. In the latter instance the observed UV dip is used for the


determination of the EOF. The background electrolyte was 0.02 M HIST/MES at pH 6.2. It can be concluded that the reproducibility of the experimental values is good and MO can be used as a true EOF marker in this system.

In Fig. 2.4 (a) the UV dip injecting water as a sample and (b) the UV signal injecting an aqueous solution of 0.0005 M MO are shown for the system 0.01 M KOH/MES at pH 6.2. In this instance the EOF marker lies behind the water dip. In Table 2.II the mEoF and effective mobilities of clenbuterol and benzoic acid are given, using for the calculation of the mEoF (1) the beginning of the UV dip, (2) the lowest point of the UV dip (with MO present). (3) the middle of the UV dip (without MO) and (4) the UV peak of MO. The importance of the use of a "true" EOF marker will be clear considering the differences in the effective mobilities. The experiments with the system KOH/MES were carried out applying 10 kV, in order to avoid temperature effects as this system shows much higher electric currents than HIST/MES owing to the higher conductivity of the system. If an EOF marker was used, it was carefully checked whether the migration times of the water dip and EOF marker were identical.

4 I I

(/) ...... c ::::l

.ri .... ~ Q) 0 c m

..Q ..... g

..0 m


Fig. 2.4: Measured UV signal for a background electrolyte of 0.01 M KOH at pH 6.2 adjusted by adding MES.for a sample consisting of (a) 100 % water and (b) 0.0005 M MO in water. (1) Beginning and (3) the middle of the UV signal/or 100% water and (2) lowest point and (4) top of the UV signal for 0.0005 M MO in water. Capillary from SGE, I.D. 72 µ.m, l., = 56.43 cm and ld = 49.83 cm. Pressure injection time, ls.

Dependence of the electroo&motic flow on applied voltage and pH In Fig. 2.5 the measured velocity of the EOF, 1-EOF• as a function of the applied

voltage is given for the apparatus used at two different times. As expected a linear relationship is obtained, although the values differ in time. The background electrolyte was a 0.02 M TRIS/MOPS system at pH 8.2. In both instances the EOF marker was dissolved in both water and buffer respectively and identical values were obtained in each case.


TABLE 2.11

MEASURED MIGRATION TIME, t (min) AND EFFECTIVE MOBILITY, m·loS (cm2Ns), USING A 0.01 M KOH/MES BACKGROUND ELECTROLYTE AT pH 6.2

Capillary from SGE, I.D. 72 jtm, le= 56.43 cm and ld= 49.83 cm. Pressure injection time, S s. Applied voltage, 10 kV.

15

c

~ _g 10

l >

5 10 15 20 25

V (kV)

Fig. 2.5: Measured relationship between the velocity of the EOF and the applied voltage at two different times. Background electrolyte, 0.02 M TRISIMOPS at pH 8.2. Capillary from SGE, 1.D. 72 jtm,

ld = 56.95 cm and le = 50.05 cm. Pressure injection time, 5 s.

In order to measure the effect of the pH on the EOF, measurements were carried out in several background electrolytes. In Table 2.m the composition of all background electrolytes is given. In Table 2.IV the mEoF values for some electrolytes are given for the original Beckman capillary cartridge (I). Series II and m and series IV and V were measured with two different SGE capillaries. As can be concluded from Table 2.IV the mobility of the EOF generally increases with the pH, as expected from the theory. Nevertheless not always the titration curve shape of the curve is found, possibly due to the fact that for the different pHs different buffer and counter ions are used. For the phosphate buffers, the increment of the mobility of the EOF with the pH shows the expected path.


TABLE2.ill

COMPOSITIONS OF BACKGROUND ELECTROLYTES AT DIFFERENT pH VALUES.

All buffer solutions were prepared by adding the buffering counter species to the cations until the desired pH was reached. All phosphate buffers were prepared by adding o-phosphoric acid to 0.01 M KOH until the desired pH was reached.

Cation Buffering counter pH Cation Buffering counter pH species species

0.02 M f1·alanine formic acid 3.5 0.02MHIST MES 6.2 0.02 M f1·alanine formic acid 3.8 O.OIMKOH MES 6.2 0.02 M f1·alanine acetic acid 3.9 O.o2MTEA MOPS 7.0 0.02MEAC formic acid 4.0 0.04MIMID MOPS 7.5 0.02MEAC acetic acid 4.4 0.02MTRIS MOPS 7.9 0.02 M {1-alanine acetic acid 4.7 0.02MTRIS MOPS 8.2 0.02MEAC acetic acid 5.0 0.02MDEA BI CINE 9.0 0.02MHIST MES 6.1

TABLE 2.IV

m60p·ta5 (cm2Ns) AS A FUNCTION OF pH FOR THE ORIGINAL BECKMAN STANDARD CAPILLARY (I) AND A SCIENTIFIC GLASS ENGINEERING CAPILLARY (II-V) AT DIFFERENT TIMES

For composition of background electrolytes, see Table 2.ill. V: Phosphate buffers.

pH n m IV v

moop mEOF mEOF pH msop pH mEOP

3.8 33.8 30.2 28.0 3.8 27.8 2.5 15.1 4.4 36.2 37.1 35.0 5.0 52.4 3.0 26.8 5.0 47.5 6.1 45.3 4.0 36.1 6.2 55.6 56.2 53.1 1.0 44.5 5.0 47.1 1.S 61.7 61.8 60.4 1.9 62.4 6.0 56.9 8.2 69.9 72.1 73.1 9.0 59.4 7.0 65.0

8.0 70.6 9.0 73.1

On running ultracentrifuged serum samples a dramatic change in the EOF resulted. For a pH of 3.8, the migration time of the EOF changed from about S.6 minutes to 21.2 minutes(!) for a capillary of SGE. In order to examine what happens with time the migration times of a mixture of amprolium, levamisol, clenbuterol (all positive ions), mesityl oxide (EOF marker) and benzoic acid (negative ion) were measured 10 times. After each run the capillary was rinsed repeatedly: 10 min with 0.1 M KOH, 10 min with water and 10 min with the background electrolyte. The results of the measurements are given in Table 2. V. Although there appears a dramatic. course in EOF with time, all the effective mobilities of the sample components were nearly constant.


TABLE2.V

MEASURED MIGRATION TIMES, t (min) AND EFFECTIVE MOBILITIES, m·loS (cm2Ns), FOR A SAMPLE OF AMPROLIUM; LEV AMISOL, CLENBUTEROL, MESITYL OXIDE AND BENZOIC ACID WITH 0.02 M /l·ALANINE/FORMICACID AT pH 3.8 AS BACKGROUND ELECTROLYTE

Capillary from SOE, I.D. 72 µm, Z,,=57,0 cm and ld-50,0 cm. Pressure injection time, 5 s. Applied voltage, 25 kV.

No. amprolium levamisol clenbuterol EOF (MO) benzoic acid t m m t m t m t m

1 4.21 36.22 5.44 25.95 6.78 19.07 21.23 8.95 2 4.01 36.56 5.13 26.27 6.30 19.36 17.59 10.80 3 3.86 36.80 4.88 26.45 5.94 19.52 15.23 12.47 4 3.80 36.79 4.77 26.SO 5.78 19.56 14.26 13.33 5 3.74 36.84 4.69 26.55 5.67 19.59 13.62 13.96 6 3.60 36.92 4.45 26.76 5.32 19.81 11.93 15.93 7 3.50 36.89 4.31 26.69 5.11 19.75 10.90 17.43 21.14 -9.22 8 3.43 36.84 4.21 26.61 4.98 19.68 10.27 18.51 20.57 -9.27 9 3.39 36.75 4.14 26.59 4.88 19.67 9.84 19.30 18.92 -9.26 10 3.34 36.77 4.06 26.58 4.77 19.67 9.44 20.18 17.39 -9.26

It can be concluded from Table 2. V that migration times. (or apparent mobilities) can never 'be used for the identification of sample components without problems. Although in· the above experiments severe changes in EOF occurred and hence also in the apparent mobility of sample ionic species, it was noticed that the effective mobilities were fairly constant.

2.1.3 Measurement of the effective mobility

To investigate the reproducibility with time of the effective mobility several experiments were carried out with a sample consisting of the positive ions procaine, clenbuterol and fenoterol and the negative ions of uric, p-hydroxyphenylacetic and benzoic acid. As EOF marker always mesityl oxide was used. All measurements were always carried out several times on different days, and the variance is given in parentheses (see Table 2. VI). In order to study the effect of the samfle concentrations three concentrations were measured, viz. 1·10'4, 5·10·5 and 1·10· M. Further, the sample components were measured dissolved both in water and in background electrolyte. Between all measurements the capillary was rinsed only with background electrolyte for 5 minutes, except if the values are indicated with (R). In that case there was an extra rinsing step with 0.1 M KOH for 5 min and 5 min with water.

In Table 2. VI all effective mobilities, calculated from the apparent mobilities, are given. Although sometimes the EOF differs the effective mobilities are remarkably constant as can be concluded from Table 2. VI.

In Table 2. VII all effective mobilities (in duplicate, calculated from the measured apparent mobilities) for the same components as used for Table 2. VI are given for several background electrolytes at different pHs (See Table 2.IlI for the electrolyte compositions). Remember that all electrolyte systems have a different ionic strength

TABLE2.VI p ,_

EFFECTIVE MOBILITIES, m· loS (cm.2/Vs). FOR SEVERAL IONIC SPECIES IN A BACKGROUND ELECTROLYTE 0.02 M HIST/MES AT pH 6.2 "t1 tI1

The components were dissolved in water (W) and background electrolyte (B). If the capillary was rinsed with 0.1 M KOH the measurements are indicated as WR ~ and BR. SGE capillary, I.D. 721&m. 10 = S6.S5 cm and ld = 49.60 cm. Pressure injection time, 5 s. Applied voltage, 25 kV. Variances are given in parentheses.

~ Compound Concentration (M) W (5 expts) B (25 expts) WR (5 expts) BR (5 expts) Average

8 procaine 0.0001 20.87 (0.07) 20.68 (0.02) 20.74 (0.06) 20.82 (0.09) 20.72 (0.17) s 0.00005 20.96 (0.00) 20.82 (0.11) 20.87 (0.06) 20.86 (0.05) 20.84 (0.10)

0.00001 20.82 (0.00) 20.80 (0.19) 20.94 (0.05) 20.86 (0,27) 20.83 (0.18) 0 z clenbuterol 0.0001 18.13 (0.07) 18.30 (0.17) 18.15 (0.04) 18.10 (0.09) 18.06 (0.15) z 0.00005 18.21 (0.00) 18.10 (0.09) 18.12 (0.08) 18.11 (0.06) 18.12 (0.11)

0 0.00001 18.09 (0.00) 18.10 (0.14) 17.93 (0.43) 18.07 (0.07) 18.07 (0.20) ~

fenoterol 0.0001 16.12 (0.10) 16.03 (0.14) 16.06 (0.03) 16.06 (0.08) 16.0S (0.13) o.oooos 15.91 (0.3S) 16.08 (0.10) 16.14 (0.08) 16.10 (0.06) 16.08 (0.17) 0.00001 16.05 (0.00) 16.09 (0.16) 16.19 (0.09) 15.92 (0.15) 16.07 (0.15)

mesityl oxide 0.0001 52.17 (0.07) 51.04 (0.92) 48.83 (0.49) 51.89 (0.19) 51.01 (1.17) 0.00005 52.09 (0.00) S0.64 (o.95) 47.68 (0.18) 51.40 (0.06) 50.33 (l.38) 0.00001 Sl.94 (0.00) SO.SS (0.96) 47.36 (0.05) s 1.34 (0.16) 50.43 (1.47)

uric acid 0.0001 -22.40 (0.06) -22.S3 (0.10) -22.3S (0.03) -22.53 (0.06) -22.49 (0.11) 0.00005 -22.48 (0.02) -22.58 (0.10) -22.43 (0.03) -22.s2 (0.04) -22.54 (0.09) 0.00001 -22.58 (0.02) -22.70 (0.11) -22.62 (0.04) -22.66 (0.10)

p·hydroxyphenylacetic acid 0.0001 -26.41 (o.60) -26.89 (0.54) -26.57 (0.04) -26.55 (0.27) -26. 73 (0.15) o.oooos -26.81 (0.02) -26.88 (0.10) -26. 70 (0.04) -26.82 (0.03) -26.84 (0.10) 0.00001 -26.95 (0.02) -27.02 (0.09) -26.94 (0.05) -26.99 (0.08)

benzoic acid 0.0001 -29.28 (0.05) -29.41 (0.10) -29.13 (0.04) -29.31 (0.04) -29.35 (0.13) 0.00005 -29.41 (0.01) -29.53 (0.12) -29.51 (0.11) 0.00001 -29.59 (0.03) -29.69 (0.11) -29.66 (0.11)

N VI


TABLE 2.VII

EFFECTIVE MOBILITIES, m·toS (cm21Vs), FOR SEVERAL IONIC SPECIES IN SEVERAL BACK(JROUND ELECTROLYTES AT DIFFERENT pH VALUES.

Capillary from SGB, I.D. 72 µm, le = 56.55 cm and Id = 49.60 cm. Pressure injection time, 5 seconds. Applied voltage, 25 kV. For the composition of the background electrolytes see Table 2.III. 1 =procaine; 2=clenbuterol; 3=fenoterol; 4=mesityl oxide; S~uric acid; 6=p-hydroxyphenylacetic acid; 7 = benzoic acid

pH 1 2 3 4 s 6 7

3.9 22.49 19.25 17.42 29.45 -0.81 -23.24 -11.35 22.04 19.01 17.13 26.15 -0.85 -11.07

5.0 21.39 18.87 17.02 52.23 -7.60 -27.17 -26.93 21.09 18.57 16.99 52.53 -7.79 -27.29 -27.05

6.1 20.57 17.90 16.03 45.28 -20.48 -26.46 -29.06 20.69 18.00 15.92 45.39 -20.55 -26.59 -29.25

7.0 19.96 17.19 14.28 44.52 -24.96 -26.42 -29.17 19.84 17.30 14.39 44.42 -24.75 -26.30 -29.08

7.9 18.62 17.24 9.87 62.33 -25.30 -26.09 -28.94 18.41 17.03 9.66 62.54 -25.44 -26.23 -29.03

9.0 11.20 14.55 -1.11 59.36 -26.15 -26.38 -29.20 11.20 15.43 -1.11 59.36 -26.09 -26.33 -29.20

and equivalent concentration. The negative ions show smaller absolute values of the effective mobilities at low pHs (not fully ionized) and so do the positive ions at high pHs. Fenoterol is even negative possibly owing to the ionization of phenolic groups at high pH.

2.1.4 The effective mobility as parameter for screening

For a qualitative screening the most important question is, whether the component of interest can be separated from the matrix and its peak can be recognized from the electropherogram. As already shown, the effective mobility, which can be calculated from the mobility at infinite dilution and pK value, can be used as parameter. A complicating factor in the analysis of complex matrices is often the presence of an excess of one of the components such as sodium chloride in urine or serum. Beckers et al. [2,3] showed that this can lead to a different migration behaviour during the analysis. Schoots et al. [4] showed that, for closed systems, if the composition of the sample (uraemic serum samples) is nearly constant, reproducible migration times are found, although the migration times differ considerably compared with those of the pure components.

To investigate the effect of the presence of a sample component in excess, the effective mobilities of the mixture in Tables 2. VI and 2. VII were determined (all


components were 104 M, both in water or dissolved in buffer) and increasing amounts of sodium chloride were added. In Table 2. VllI all effective mobilities, calculated from the measured apparent mobilities, are given. It can be concluded from Table 2. VIII that up to about 0.01 M NaCl the effective mobilities are nearly constant, except for potassium and sodium, as they are not migrating in a proper CZE way. At higher concentrations of NaCl the effective mobilities decrease, although these values are reproducible. Using higher background concentrations this effect will, of course, diminish.

TABLE 2.VIII

EFFECTIVE MOBILITIES, m·loS (cm2Ns), DETERMINED INJECTING A SAMPLE WITH AN INCREASING AMOUNT OP SODIUM CHLORIDE DISSOLVED IN WATER AND BUFFER SOLUTION. WITH A HIST/MES BACKGROUND ELECTROLYTE AT pH 6.2

Capillary from SGE, I.D. 72 pm, l., = 56.43 cm and ld = 49.83 cm. Pressure i.rtjection time, 5 s. Applied voltage, 25 kV. !=procaine; 2=clenbuterol; 3=fenoterol; 4=mesityl oxide; 5•uric acid; 6=p·hydroxybenz.oic acid; 7•benzoic acid

Solution NaCl K Na 2 3 .4 s 6 7 (M)

Water 0 66.97 47.52 20.96 18.27 16.18 45.28 -21.70 -26.96 -29.53 0.0001 68.00 47.65 20.86 18.15 16.26 45.61 -21.82 -27.11 -29.74 0.0005 67.54 46.50 20.85 18.16 16.27 45.39 -21.78 -27.08 -29.69 . 0.001 67.32 48.12 20.86 18.15 16.26 45.61 -21.76 -27.09 -29.68 0.005 67.32 48.12 20.86 18.15 16.26 45.61 -21.70 -27.03 -29.62 0.01 47.19 20.63 17.94 16.05 45.61 -21.64 -26.94 -29.53 0.05 50.74 19.48 17.10 15.47 45.39 -21.33 -26.53 -29.09 0.075 52.36 18.92 16.79 15.19 45.28 -21.22 ~26.38 -28.92 0.1 54.01 18.16 16.29 14.72 45.17 -21.20 -26.31 -28.81

Buffer 0 68.44 47.17 20.84 18.16 16.09 45.17 -21.77 -27.02 -29.63 0.0001 70.44 47.52 20.96 18.27 16.18 45.28 -21. 79 -27 .06 -29.68 0.0005 71.99 47.63 20.84 18.16 16.29 45.17 -21.71 -27.01 -29.58 0.001 69.02 48.45 20.96 18.27 16.38 45.38 -21.70 -26.97 -29.55 0.005 47.87 20.85 18.16 16.07 45.39 -21.75 -27.03 ·29.62 0.01 48.81 20.62 17.94 16.07 45.39 -21.63 -26.94 -29.50 o.os 51.24 19.48 17.10 15.28 45.39 -21.42 -26.57 -29.lS O.o75 53.38 18.92 16.79 lS.00 45.28 -21.31 -26.46 -28.98 0.1 54.65 18.48 16.40 14.83 45.06 -21.21 -26.30 -28.79

l 0-fold diluted urine spiked with components 1-7 69.99 47.08 20.29 17.82 15.94 45.72 -21.34 -26.71 -29.25

An interesting point in these experiments was that on adding larger amounts of NaCl in the sample in first instance an UV dip was obtained, but at a certain NaCl concentration the UV signal of the EOF marker increased rapidly. The explanation is that if a high concentration of NaCl is present, at the point of the sample injection (note: the EOF · position)· the local ionic strength is very high, and according to


Kohlrausch an adaptation to this original concentration takes place. This means that at the point of the EOF later in the analysis a higher background concentration will be found, giving a high UV signal if the background electrolyte shows UV absorption.

In Fig.·2.6 this effect is shown for samples of aqueous NaCl solutions (without EOF marker). For higher concentrations of NaCl there is an increasing UV signal at the point of the EOF. The consequence of this effect for complex matrices can be that uncharged components, migrating at the EOF position are covered by this effect. The choice of a non-UV absorbing background electrolyte can be important.

(j) ..., 0.10 M NaCl E :::>

.ci 0.07 M NaCl '-~ 0.05 M NaCl Q)

g 0.03 .M NaCl Ill .0

£ 0.01 M NaCl

.0 Ill


Fig. 2. 6: Measured UV signal for the background electrolyte HIST/MES at pH 6. 2 for different aqueous solutions of NaCl without EOF marker. For farther explanation see text. Capillary from Siemens, J.D. 50 µm, lc=77.!J3 cm and ld=70.53 cm. Pressure injection time, 1 second.

As an example of screening possibilities the same components (0.0001 M, see Fig. 3.9) were added to 10-fold diluted ultrafiltered human urine (with about 0.015 M NaCl) and the effective mobilities of the components were calculated. These values are also given in Table 2. VIII, and it can be. seen that the components can be easily recognized from the effective mobilities. Potassium and sodium are indicated by negative UV dips in the electropherogram. Using capillaries with small diameters (50 µm) cations showed increasing tailing peaks owing to wall attraction forces.

2.2 DETERMINATION OF PHYSICO-CHEMICAL PARAMETERS

2.2.1 Introduction

From effective mobilities, pK values and mobilities at infinite dilution can be calculated. If those physico-chemical parameters of a component are known, this component is characterised for electrophoresis. Thus these parameters are important

2.2 DETERMINATION OF PHYSICO-CHEMICAL PARAMETERS 29

in the choice of a suitable electrolyte system for the separation of various components in complex matrices.

So far, mobilities and pK values have often be.en determined by ITP [5,6,7,8,9,10,11,12]. The calculation of mobilities and pK values in ITP is laborious, however, as in ITP all zones have different parameters such as pH, concentration and temperature, through which the data have to be calculated in an iterative way. The correction of the concentration dependence of mobilities and for activities is often troublesome for mixtures of ionic species with different charges. Further, in ITP the choice of the pH of the electrolyte systems is limited to about 3-11. Low pHs can not be applied in the separation of cations owing to the great influence of hydrogen ions on the zone conductivity and high pHs can hardly be used in the separation of anions owing to the disturbance by carbonate. Especially at low pHs major problems can be expected in finding an appropriate slow terminator in the· separation of cations because hydrogen ions can act as terminator with relatively high effective mobilities [13,14]. Generally, the determination of mobilities of weak acids and bases with low mobility is difficult as well as these of the subspecies from multivalent acids and bases.

In CZE, on the other hand, many of these limitations are not present. Background electrolytes at low and high pHs can be used easily, there is no need for a terminator, and corrections for several effects are relative easy because all parameters in the background electrolyte, such as ionic strength, pH, temperature and electric field strength, can be considered as to be nearly constant.

2.2.2 Theory

If the effective mobility of a component is known for two different electrolyte systems at different pHs, at which the component shows a different degree of dissociation, both its pK value and its mobility at infinite dilution can be calculated. For a monovalent acid the calculation is as follows. The thermodynamic equilibrium constant for the equilibrium:

is defined as:

with the assumption that for HZ the activity coefficient 'Y = 1. Hence, with

pH=-log("'(~[H1)

(2.5)

(2.6) .

(2.7)


pKrh = -log'Yz +pH - log ~~~ (2.8)

Because the effective mobility m = a.me, where me is the ionic mobility at a specific concentration c, it follows that

(2.9)

The value of me can be calculated from the molar ionic conductivity >.. e using Faraday's constant, F. For the molar ionic conductivity in mixed-solutions the expression according to Bennewitz, Wagner and Kuchler as described by Falkenhagen [15] was used:

Ac=>..0 -(0.229/..0 +30.1)./C (2.10)

where Ac and >..0 are the molar ionic conductivities at concentration c and infinite dilution respectively. This relation can be used for very diluted solutions of a particle in a bulk of the background electrolyte.

If the effective mobilities are known in two different electrolyte systems this means:

The activity coefficients 'Y can be calculated by:

-log')'= 0.5085z2{;

1 +0.3281 a{µ

(2.11)

(2.12)

where z is the valency of the component, µ, is the ionic strength of the solution as determined by the background electrolyte and a is the effective hydrated diameter of the ion in A. If the effective hydrated diameter was unknown, 5 A is assumed.

If the ionic strengths and the concentrations of two electrolyte systems are known, all activity coefficients can be calculated with eqn. 2.12. Using Faraday's constant and eqn. 2.10 them~ and mi can be replaced by the mobility at infinite dilution and by this the only unknown parameter in eqn. 2.11 is the mobility at infinite dilution and this can be calculated. With eqn. 2.8 the pKtll can be obtained. Analogous derivations can be given for multivalent anions and cations.

2.2.3 Calculation of pK values and mobilities at infinite dilution

In Table 2.IX the results of the calculations of pK values and mobilities at infinite dilution for several acids are given using the effective mobilities obtained in

2.2 DETERMINATION OF PHYSICO-CHEMICAL PARAMETERS 31

the systems at pH 4 and 6.2 (HIST/MES) using CZE. The results are compared with data given by Hirokawa et al. [7] and with data obtained in isotachophoretic way using the concept of the isoconductor [11] and specific zone resistance respectively [16J. The CZE experiments are carried out using 50 µ.m capillaries. Compared with 75 µm capillaries the obtained peaks for negative ions were much more gaussian, due to the repulsive forces between anions and negative wall charge. For cations, smaller diameters led to strongly tailing peaks. The ITP experiments (see section 6.3.2 for a description of the instrumentation) were carried out with a leading electrolyte of 0.01 M HCl with e-aminocaproic acid at a pH of 4 in combination with a terminator of 0.01 M pivalic acid, whereas for the system of pH 6 a leading electrolyte of 0.01 M HCl with histidine and the terminator 0.01 M MES was used.

TABLE 2.IX

CALCULATED pKV ALUES AND MOBILITIES AT INFINITE DILUTION, m0• tcP (cm2Ns), FOR SEVERAL ACIDS USING EXPERIMENTAL DATA FOR 1WO DIFFERENT ELECTROLYTE SYSTEMS WITH (I) ISOTACHOPHORESIS AND (II) OPEN CAPILLARY ZONE ELECTROPHORESIS versus (III) UTliRATURE VALUES.

Capillary from Siemens, I.D. 50 µm, Zc=77.33 cm and Zd=70.53 cm. Pressure injection time, 1 s. Applied voltage, 25 kV.

Compound (I) (II) (III) ITP CZB Lit. [7]

pK mo pK moa pK mo

m-aminobenzoie acid 4.79 -31.49 4.74 -31.64b benz.oic acid 4.18 -33.26 4.16 -33.40 4.19 -32.9 hippuric acid 3.63 -27.50 3.60 -27.77 2.70 -25.3 p-methoxypbenylacetic acid 4.37 -28.75 4.37 -29.03 4.36 -29.7 nicotinic acid 4.85 -33.71 4.82 -33.44 4.82 -34.6 p-nitrobenz.oic acid 3.38 -31.92 3.49 -31.94c 3.52 -32.3 a.-dinitropbenol 4.01 -32.33 4.04 -32.39 4.02 -31.3 2,6-dinitropbenol 3.65 -33.96 3.73 -33.99C 3.71 -31.3 phenylacetic acid 4.29 -30.84 4.28 -31.10 4.41 -31.7 propionic acid 4.89 -37.41 4.87 -37.1 sulfanilic acid 3.12 -33.81 3.27 ·33.93c 3.23 -33.7 uric acid S.55 -31.08 5.41 ·29.99'

a Marked values were determined using electrolyte systems at pH values of b4,7 and 6.2, c3.s and 6.2 and 45.o and 6.2.

From Table 2.IX it can be concluded that mobilities at infinite dilution and pK values can be obtained in this way, provided that a good set of effective mobilities is available. Remember that for this reason, for some ionic species (see Table 2.IX) other pHs than 4 were chosen to obtain larger differences between the degrees of dissociation (effective mobilities) for these components in the two electrolyte systems. For m-aminobenzoic acid pH 4. 7 as lowest pH was chosen in order to avoid the


possibility that this component partially dissociates to a positive ionic form. For uric acid pH S as lowest pH was chosen in order to obtain a real electrophoretic migration.

2.3 PEAK RECOGNITION IN MECC

MECC is a separation technique based on the partitioning of the components over two phases, just as in chromatographic techniques. However, now two mobile phases are used viz., an electroosmotically pumped aqueous mobile phase and the hydrophobic interior of micelles. Often for the peak recognition, an analogue of the capacity factor k is handled, which can be calculated according to the equation [17,18]:

(2.13)

where nMc and 1lw are the total moles of solute in the micelles and in the aqueous phase respectively, and t8, tEoF and tMc are the migration times of the solute, an appropriate marker for the determination of the EOF (insolubilized component) and the micelles, respectively.

Application of this concept of k in MECC shows that for higher values of k very low values of the resolution R [18] are the result. Further, a slight inaccuracy in the determination of t5 leads to a large difference in the calculated k value, especially for ts values near the tMc· This means that the use of k values for screening purposes is limited whereas the suggestion that a large difference between higher values of k leads to a separation is false.

Moreover, the EOF can change with time and therefore tEoF and tMc have to be measured in each experiment in order to calculate the appropriate k. The tMc can be obtained applying a micelle marker such as Sudan III or anthracene [18, 19] or by methods such as extrapolation with iteration or frontal analysis [20]. However, especially when organic co-solvents are used, it is still a subject of discussion how to obtain the "true tMc". If the velocity of the micelles is small or even negative, tMc can not be measured and, according to eqn. 2.13, the k can not be calculated, even when the components show a normal migration behaviour. For all these reasons k is often not suitable to handle in MECC.

Another way to describe the migration behaviour of components in MECC is in terms of mobilities [21,22]. The velocity of the aqueous mobile phase is determined by the mobility of the EOF, mEOF• and that of the micelles by the apparent mobility of the micelles, mapp,MC' defined as:

(2.14)

where mMc is the effective mobility of the micelles. For micelles, the mMc will be strongly dependent to the composition of the micellar phase. For a given surfactant concentration in a specific electrolyte system, a constant composition of the micellar

2.3 PEAK RECOGNITION IN MBCC 33

phase can be expected, through which· the effective mobility of the micelles and the k for the components must be constant.

In MBCC often uncharged particles, with no electrophoretic mobility are analyzed. As they are solubilized in the charged micelles for kl(k+ 1) part of the time, they will acquire a net velocity of:

(2.15)

or

(2.16)

or

(2.17)

where v5 , "Mc and vEoF are the velocities of the solubilized sample component, the micelle marker and the EOF marker, respectively.

From eqn. 2.17 it will be clear that for uncharged particles pseudo mobilities can be defined, similar to the mobilities of charged particles (see eqn. 2.2), satisfying the conditions:

with

and

ps k ms =--mMc

l+k

(2.18)

(2.19)

(2.20)

A great advantage working with pseudo-effective mobilities compared with k is · that for the calculation of the pseudo-effective mobilities the tMc or mMc are not required as can be seen from eqn. 2.21:

where le and Id are the total length of the capillary and the length of the capillary from injection to detection respectively, and V is the applied voltage.

To demonstrate the effect of a small inaccuracy in the determination of migration times on k and pseudo-effective mobilities, the values of k and m~8 for several values


TABLE2.X

CALCULATED CAPACITY FACTORS, k, AND PSEUDO-EFFECTIVE MOBILITIES, m§5·loS (cm2Ns). FOR SEVERAL VALUES OF THE MIGRATION TIMES (min) FOR THE EOF MARKER, 'EOF• A SAMPLE COMPONENT, 's· AND THE MICELLE MARKER, 'Mc·

The accuracy in the determination of t8 is taken 0.5% (le = 21 cm, ld = 20 cm, applied voltage 10 kV)

Fixed values ts k m§"

t8-0.5% ts ts+0.5% t8-0.5% ts t5 +0.5%

tEOF 2 min; tMc = 10 min 8.00 14.61 15.00 15.41 -33.69 -33.75 -33.81 8.50 20.93 21.67 22.44 -34.36 -34.41 -34.46 9.00 33.28 35.00 36.88 -34.95 -35.00 -35.05 9.30 48.58 52.14 56.21 -35.27 -35.32 -35.37 9.50 68.06 75.00 83.40 -35.48 -35.53 -35.67 9.80 155.64 195.00 259.90 -35.77 -35.82 -35.86 9.90 262.56 395.00 787.08 -35.86 -35.91 -35.95

tEOF = 10 min; /Mc = 20 min 11.0 0.21 0.22 0.24 -0.78 -0.82 -0.86 12.0 0.48 0.50 0.52 -1.46 -1.50 ·L54 13.0 0.83 0.86 0.88 -2.04 -2.08 -2.11 14.0 1.29 1.33 1.37 -2.54 -2.57 -2.60 15.0 1.94 2.00 2.06 -2.97 -3.00 -3.03 16.0 2.90 3.00 3.10 -3.35 -3.38 -3.40 17.0 4.48 4.67 4.86 -3.68 -3.71 -3.73 18.0 7.57 8.00 8.47 -3.97 -4.00 -4.02 19.0 16.26 18.00 20.10 -4.24 -4.26 -4.29 19.5 31.47 38.00 47.69 -4.36 -4.38 -4.41

'EoF = 10 min; 'Mc = so min 20.0 1.64 1.67 1.69 -4.48 -4.50 -4.52 25.0 2.96 3.00 3.04 -5.38 -5.40 -5.42 30.0 4.93 S.00 5.08 -5.98 -6.00 -6.01 35.0 8.18 8.33 8.49 -6.42 -6.43 -6.44 40.0 14.61 15.00 . 15.41 -6.74 -6.75 -6.76 45.0 33.28 35.00 36.88 -6.99 -7.00 -7.01 46.0 42.28 45.00 48.05 -7.03 -7.04 -1.05 47.0 56.82 61.67 67.33 -7.08 -7.09 -7.09 48.0 84.29 95.00 108.64 -7.12 -7.13 -7.13 49.0 155.64 195.00 259.90 -7.15 -7.16 -7.17

of tEOF• ts and 'Mc with an accuracy in the determination of the ts of 0.5 % were calculated. In Table 2.X all calculated values are given.

From Table 2.X it can be concluded that especially for ts values near these of tMc a dramatic change in k is the result, whereas the pseudo-effective mobilities are nearly constant, for small differences in t5•

2.4 CONCLUSIONS 35

2.4 CONCLUSIONS

In open tubular capillary zone electrophoresis with electroosmotic flow for lowmolecular-weight substances, migration times or apparent mobilities can never be used for the identification of components. The effective mobility, however, which can be calculated from the migration time and the EOF velocity, can be used as a parameter for identification. The choice of a "true" EOF marker is extremely important.

If the ionic strength of the matrix is high compared with that of the background electrolyte, differences in effective mobilities can be expected, although they are reproducible if the matrix is of constant composition. Hence, effective mobilities can be used for screening purposes.

From the effective mobilities, measured in two different electrolyte systems at pH values where the degrees of dissociation of a component differ sufficiently, the mobility at infinite dilution and pK value of this component can be calculated.

In MECC, the capacity factor, k, provides fundamental information concerning the distribution coefficient over the aqueous and the micellar phase, which can give a guide to improving the resolution according to the resolution equation of MECC. For screening purposes, pseudo-effective mobilities are to be preferred to capacity factors because they can be calculated even if tMc is unknown, and because they are less sensitive to inaccuracies in the determination of the migration time. Moreover, pseudoeffective mobilities give a better indication whether components can be separated or not.

References

1. J.L. Beckers, F.M. Everaerts and M.T. Ackermans, J. Chromatogr., 531(1991)407. 2. 1.L. Beckers and F.M. Everaerts, J. Chromiltogr., 508 (1990) 3. 3. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 508 (1990) 19. 4. A.C. Schoots, Th.P.E.M. Verheggen, P.M.J.M. de Vries and F.M. Everaerts, Clin. Chem.,

36 (1990) 435. 5. T. Hirokawa and Y. Kiso, J. Chromatogr., 252 (1982) 33. 6. T. Hirokawa, M. Nishino and Y. Kiso, J. Chromatogr., 252 (1982) 49. 7. T. Hirokawa, M. Nishino, N. Aoki, Y. Kiso, I. Sawamoto, T. Yagi and J.-1 Akiyama,

J. Chromatogr., 271 (1983) Dl-D106. 8. 1. Pospichal, M. DemJ, z. ZemJova and P. Boeek, J. Chromatogr., 320 (1985) 139. 9. J. Pospichal, M. Demi and P. B~, J.Chromatogr., 390 (1987) 17. 10. I. Hoffmann, R. Muenze, I. Dreyer and R. Dreyer, J. Radioanal. Chem., 74 (1982) 53. 11. J.L. Beckers, J.Chromatogr., 320 (1985) 147. 12. T. Hirokawa, T. Tsuyoshi and Y. Kiso, J. Chromatogr., 408 (1987) 27. 13. P. BoCek, P. Gebauer and M; DemJ, J. Chromatogr., 217 (1981) 209. 14. P. BoCek, P. Gebauer and M. Demi, J. Chromatogr., 219 (191H) 21. 15. H. Falkenhagen, Elektrolyte, Hirzel, Leipzig, 1932. 16. 1.L. Beckers and F.M. Everaerts, J. Chromatogr., 470 (1989) 277. 17. S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchiya and T. Ando, Anal Chem., 56 (1984} 111. 18. S. Terabe, K. Otsuka and T. Ando, Anal Chem., 51 (1985) 834. 19. 1. Gorse, A.T. Balchunas, D.F. Swaile and M.1. Sepaniak, J. High Res. Chromatogr.

Chromatogr. Commun., 11 (1988) 554. 20. M.M. Bushey and J.W. Jorgenson, J. Microcolumn. Sep., 1 (1989) 125; 21. M.T. Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 585 (1991) 123. 22. K. Ghowsi, 1.P. Foley and R.J. Gale, Anal. Chem., 62 (1990) 2714.

36

CHAPTER.3

SEPARATION POWER

In CZE the most important phenomena causing peak broadening are Joule heating, electroosmosis, diffusion, elecnodispersion, injection and interactions of the analytes with the capillary wall. The dispersive effects of these phenomena decrease the resolving power of the technique, and thus their minimi:zation is of great importance. In this chapter, the resolving power of CZE is discussed and the most important contributions to peak broadening are studied. An estimation is made for the influence of the different dispersive effects of these phenomena. In the last section a separation number, analogue to the separation number from GC is defined for CZ.E showing that the use of separation numbers gives a good indication whether components can be separated or not.

3.1 IN1RODUCTION

During zone electrophoresis, the sample ions migrate in the solution of the background electrolyte with a concentration much higher than that of the sample ions, which conducts almost all electric current. The composition of the background electrolyte is constant along the separation path and does not change with time. If two sample ions have to be separated, their mobilities should be different, resulting in different migration velocities of the two ions. For the efficiency of the zone electrophoretic separation the number of theoretical plates (N) can be used as a parameter, in analogy with the number of theoretical plates in chromatography [1]:

z2 N=- (3.1)

ri'-

where l is the length of the capillary and ri'- is the total spatial variance of the concentration profile of a zone. This variance is the result of several dispersive effects, such as Joule heating, electroosmosis, diffusion, electrodispersion, injection, interaction with the capillary wall and peak broadening stemming from the instrumentation used. In the following section, these effects will be discussed. If the detected peaks are symmetric and have a Gaussian profile, the following equation can be used to calculate the plate number from the analysis record [2]:

38

N=S.54 [-t ] 2 W1h

CHAPTER 3: SEPARATION POWER

(3.2)

where t is the migration time and w,h is the peak width at half height on a temporal basis.

For the quantitative description of the separation of the separands, the resolution, R, is used defined as the ratio of the distance between the two peaks maxima, Ax, to the mean peak width at the baseline (taken 4 ti, where (i is the mean standard deviation of their concentration distributions):

Ax R=-

4a (3.3)

The resolution is related to the number of theoretical plates by the simple equation [l]:

(3.4)

where Amlm is the relative difference in mobilities of the both separands.

3.2 PEAK BROADENING

3.2.1 Introduction

In electrophoresis the separation of substances is based on their different speeds of migration in an electrolyte system under influence of an electric field. During this process, the zones of the various substances are not only separated, but they are also continually broadened due to a series of dispersive factors. For the optimization of a separation, it is important to study the dispersive factors, and to minimize them.

In capillary electrophoresis, the total variance of a peak can be assumed to be the sum of the variances due to the particular sources of peak broadening [3,4J:

(3.5)

where the right hand term represents the contributions of injection, diffusion~ temperature profiles due to Joule heating, electrodispersion, electroosmosis and other effects, respectively. Except for the ot. which is determined by the initial situation, all a2 values are proportional to the analysis time, the constants of proportionality being the respective dispersion coefficients, D. The total variance of the sample distribution can therefore be expressed as:

3.2 PEAK BROADENING 39

C1~1 =C1f +2tED11 (3.6) n

where t is the migration time. The first four terms on the right hand side of eqn. 3.5 represent effects inherent

of the principle of the method and they can never be eliminated; however, their influence upon the separation efficiency can be controlled by appropriate design of the instrument and selection of the working conditions. In the following sections, several contributions to the peak broadening are discussed in more detail.

3.2.2 Contribution due to injection

The injection term is related to the shape and composition of the initial sample pulse. For an ideal case (if the sample is introduced to the capillary as a rectangular pulse of width w), the input variance can be calculated as:

2 w2 CT· =-

1 12 (3.7)

However, if the electrical resistance of the sample plug is different from that of the background electrolyte, which is often the case in capillary electrophoresis, a deviating electrical field is present in the injected plug, and other effects such as sample stacking and peak broadening due to Joule heating can an play important part in the determination of aT· In section 3.3 it will be shown that aT usually difers from w2t12.

3.2.3 Contribution due to diffusion

Diffusion is the motion of a substance under the influence of a gradient of chemical potential, or in a simplified way, under the influence of a gradient of the concentration· of this substance. Provided that the initial sample pulse is a b-shaped, after a certain time t, due to diffusion, the result is a Gaussian concentration profile, which can be characterized by its variance aa. which is related to the diffusion coefficient Dd of a substance by the Einstein relation:

(3.8)

In diluted solutions, the Nernst-Einstein equation holds for the relation between the diffusion coefficient and the mobility at infinite dilution of a substance:

m0RT D,=-zF (3.9)

Therefore diffusion can not be eliminated from electrophoretic experiments, and its dispersive effects must always be taken into account.

40 CHAPTER 3: SEPARATION POWER

3.2.4 Contribution due to Joule heating

During the passage of an electric current I through a separation capillary of cross section A filled with an electrolyte with conductivity u, the Joule heat produced per unit volume. per second is:

IE J2 P'=-·=-A oA2

(3.10)

where P is the dissipated electric power. Due to the Joule heating, the mean temperature inside the capillary increases,

which results in a change in the effective mobility of the migrating ions. When the capillary is cooled, the Joule heat produced will be conducted away at the capillary wall and a radial temperature profile is the result. This radial temperature profile can be approximated by a parabolic function for tubular capillary columns [5]:

PR2

(3.11)

where· TR is the temperature at the inner wall of the cooled capillary, Tr is the temperature at a distance r of the longitudinal axis (r=O) of the tube, )..T is the thermal conductivity of the electrolyte used, and P is the Joule heat produced per second, in the unit volume of the electrolyte.

Due to the Joule heating, a time-dependent peak broadening results, and in analogy with the diffusion it can be expressed as:

(3.12)

where Dr is the respective dispersion coefficient due to Joule heating. Virtanen [5] derived an expression for the dispersion coefficient for the peak broadening due to the Joule heating for a parabolic temperature profile:

tlclm2E6R6 Dr=----

3072Dd)..~ (3.13)

where o is the thermal coefficient of the mobility, u is the conductivity of the background, m is the effective mobility of the component, E is the electric field strength, R is the radius of the capillary and A.r is the thermal conductivity of the background electrolyte.

3.2.S Contribution due to electrodispersion

Electrodispersion is a characteristic effect, which takes place in systems, where the moving boundaries do not show the self-sharpening effect, and therefore it is very

3.2 PEAK BROADENING 41

important in CZE [6]. It is known that in such systems, the starting sharp zone boundaries are dispersed not only due to diffusion, but also due to a dispersion factor inherent in the electromigration itself: the so-called electrodispersion. For a quantitative description of the electrodispersion, it is necessary to solve the system of transport equations. This can be done in a telling, explicit and useful way for the system of univalent strong electrolytes if diffusion is neglected [7]. When also considering diffusional transport in the continuity equation, an explicit mathematical description of the separation process is no longer possible. In such case, the concentration profiles of the migrating zones can be obtained by computer simulation [8,9].

If the concentration of the sample zone is more than two orders of magnitude lower than the concentration of the background electrolyte, or if the difference in mobility between the sample component and the background electrolyte is minimized, the electrodispersion is negligible.

3.2.6 Contribution due to electroosmosis

The peak broadening of zone due to electroosmosis is closely related to the instrumentation used, since either laminar or turbulent flow can arise if 6pen or closed systems are used. Though some attempts have been made to determine peak broadening due to electroosmosis in open capillaries [10,11], lamimir electroosmotic flow in open capillary is usually considered as non-dispersive, aioF=O.

However, in capillaries with fixed boundaries (closed systems) the resulting turbulent flow can contribute substantially to the peak broadening of the migrating zones, and this peak broadening can be considered as time-dependent in analogy with diffusion:

(3.14)

where ~OF is the dispersion coefficient due to electroosmosis. For the calculation of the DEoF Virtanen [5] derived for closed systems the following equation:

2 2 1 R VEOF l R2e2f E2

DEoF= 48 D4

= 48 D4

11 (3.15)

where E is the dielectric constant of the background electrolyte, f is the zeta potential . and '1 is the viscosity of the background electrolyte. The same sort of contribution due to electroosmosis is found if a 'soft wall' is present. This is the case if two electrolytes are present in the capillary with a different V:EoF• causing a mismatch in the ~F [12].


3.2.7 Contribution due to interaction of the analytes with the capillary wall

In order to minimize tire contribution to peak broadening due to interaction of the analytes with the capillary wall, generally, the charge oh the capillary wall can be diminished or reversed by·adding surface active additives to the background.

3.3 ESTIMATION OF THE CONTRIBUTIONS OF INJECTION, DIFFUSION AND JOULE HEATING TO PEAK BROADENING

From the foregoing section it is clear that the main dispersive effects, using capillary zone electrophoresis for the analysis of small ionic components in open capillaries, are injection, diffusion and Joule heating. In this section an estimation is made of the contributions of these effects to the total peak broadening.

3.3.1 htjection

As already described in section 3.2.2 usually for the at in CZE the value of w2/12 is used. This equation is based on the idea that the injection pulse is rectangular with width w. In CZE, according to Kohlrausch' regulation function, tbe sample can be concentrated after injection. This is the case if diluted samples are injected. As can be calculated using the Kohlraush regulation function (see section 5.1), the concentration of the sample will adapt to the value of the regulation function of the background electrolyte. This means that when diluted samples are used, so-called 'sample stacking' occurs, which means that the sample zone is automatically compressed into a sharp pulse as it crosses the plane at the position of the original boundary between the sample plug and the background electrolyte. This process can be very fast, resulting in a very narrow injection plug.

As an example of this process, in Fig. 3.1 the simulated electropherograms are given, for a background electrolyte consisting of a cation with mobili~ 30· 10·5 cm2/Vs at concentration 0.01 M, and an anion with a mobility of -30· 10· cm2/Vs and the same concentration. As a sample a component with a mobility of 50· 10·5 cm2/Vs at a concentration of 0.001 M was introduced as a plug as can be seen in the electropherogram at 0.00 s. The simulation was carried out as described by Beckers et al. [13]. Sample ions in the backside of the sample plug move forward with a much higher velocity due to the high voltage gradient in the sample plug. This causes a sharpening up of the sample plug as can be seen in the figure. If the sample plug is completely sharpened up (after 0.05 s, its maximum height is reached) the normal dispersive processes start to act on the sample zone and peak broadening will start. As can be seen from the simulations, the sharping up effect takes place in a very short time and in fact it seems that a very small highly concentrated sample plug is injected rather than a very broad sample plug at low concentration. It will be obvious that the calculation of or with w2t12 will be only useful if almost no sample stacking occurs.

3.3 ESTIMATION OF THE CONTRIBUTIONS OF .... 43

i o.oo. I O.Ot • ·; j l ,§ i

I I x Wb. ..,,..., IC Cllti. ....itt)

I 0.02 $ I 0.03.

l l j I I

IC (lutl. .... .., IC Cllti. uriitlil

i 0.()4 $ 1 0.()5. ·;

1 j .6 .&

I !

I x (lib. ..... .., IC Cllti. ....ittl

'I o.oe • ! 0.07. ·; ·;

1 j

I - t IC <iutl. lritel x Wb. lriW

Fig. 3.1: Simulation of the electrophoretic sample stacking of a sample consisting of a cation with mobility 50·luS cm21Vs and concentration 0.001 M, in a background electrolyte with a cation of mobility 30· 10'5 cm2 /Vs and an anion with mobillty-30· 10'5 cm2 /Vs both at a concentration of 0.01 M. The original sample plug can be seen in the electropherogram at 0.00 s.


3.3.2 Diffusion and Joule beating

As can be concluded from the foregoing, or can be neglected if very diluted samples are used. In order to estimate the contributions of diffusion and Joule heating on the peak broadening, oa and O'f are calculated, using eqn. 3.8, 3.9, 3.12 and 3.13 for several applied voltages. In Table 3.1 the numerical values of the parameters used are given.

TABLE 3.I

NUMERICAL VALUES OF THE PARAMETERS USED FOR THE CALCULATION OF PLATE NUMBERS AND ,;. VALUES

Parameter Numerical value Unit

Buffer: 0.01 M TRIS!Formate pH 4.0

" 0.00079 110.cm

Capillary le 57 cm

ld 50 cm 1.D. 0.0075 cm

Sample (creatinine) mo 37.H0-5 cm2/Vs

mapp 55.1·10-5 cm2/Vs

Dd 9·10-6 cm2/s

Remaining & 0.025 1/K A-r 0.006 W/cm·K

In Table 3.11 the calculated t, aa. of and N are given using these parameters. As can be concluded from Table 3.11 for this experimental conditions, Joule heating plays only an important part at very high field strength. In the normal range of the used field strengths (up to about 450..500 V/cm) diffusionis the main dispersive factor. This can be seen also in Fig. 3.2 where the calculated plate height is given as a function of the applied voltage. As can be concluded, at a voltage of about 50 kV the deviation due to the Joule heating from the straight line is seen. From Table 3.11 it can be seen that a deviation occurs if uf. is less than about 100 times smaller than ~. Of course, the voltage at which the Joule heating starts to take part in the peak broadening depends upon the system parameters. The influence of the Joule heating is observed at much lower voltages if e.g. the cooling of the capillary is not sufficient, the background electrolyte is concentrated to 0.1 M, le is shortened, or the internal diameter is increased.

If the operating conditions are chosen in such a way that Joule heating does not play an important part, the Dd can be calculated from a linear graph of ~ot versus the

3.3 ESTIMATION OF THE CONTRIBUTIONS OF .... 45

TABLE 3.11

CALCULATED t (min). ~. of AND oi01 (cm2) AND PLATE NUMBERS N (I) WITH DIFFUSION AND JOULE HEATING AND (II) ONLY WITH DIFFUSION, FOR SEVERAL APPLIED VOLTAGES V (kV) OR ELECTRIC FIELD STRENGTIIS E (V/cm), USING THE SYSTEM PARAMETERS AS GIVEN IN TABLE 3.l

v E ~ of oiot N(I) N(ll)

2 35.09 42.64 4.61·10·2 3.23·10·12 4.61·10·2 105S3 70S53 3 S2.63 28.43 3.ono-2 2.45·10·11 3.ono-2 105830 105830 4 70.18 21.32 2.30·10-2 1.03-10-10 2.30-10-2 141107 141107 s 87.72 17.06 1.84·10-2 3.lS•l0-10 1.84·10-2 176383 176383 10 17S.44 8.53 9.21·10"3 1.0Ho-8 9.21· 10-3 352766 352767 15 263.16 S.69 6.14·1Cr3 7.66·10-S 6.14·10"3 529143 529150 20 350.88 4.26 4.6M0-3 3.2no·1 4.61-10-3 70S484 705S33 25 438.60 3.41 3.68·10-3 9.85·10"7 3.69·10-3 881681 881917 30 526.32 2.84 3.07·10-3 2.4S·IO'° 3.07·10'3 1057456 1058300 3S 614.04 2.44 2.63·10"3 S.30·10'° 2.64·10-3 1232202 1234683 40 701.7S 2.13 2.30·10"3 1.03·10"5 2.31·10-3 1404764 1411067 4S 789.47 1.90 2.0S·l0"3 U6·10"5 2.07-10-3 1573141 1S87450 so 877.19 1.71 1.84·10"3 3.1S·l0"5 1.87'10-3 1734152 1763833 SS 964.91 1.55 1.67·10·3 S.08·10·5 l.73· 10-3 1883118 1940217 60 10S2.63 1.42 l.S4·HT3 7.84·10"5 1.61·10-3 2013686 2116600 6S 1140.3S 1.31 1.42·10"3 1.17-lo-4 1.SH0-3 2118006 2292983 70 1228.07 1.22 l.32·10-3 l.70·10-4 1.49· I0-3 2187462 2469367 1S 1315.79 1.14 1.23·10-3 2.39·1o-4 1.47·10"3 2214094 264S750 80 1403.51 1.07 l.lS·l0-3 3.3Mo-4 l.48· 10-3 2192538 2822133 85 1491.23 l.00 1.08· 1()"'3 4.48·lo-4 1.S3· 10-3 2121896 2998517 90 1578.95 0.95 1.02·10'3 5.96·1o-4 1.62·10-3 2006709 3174900 95 1666.67 0.90 9.69·1o-4 7.8l·lo-4 1.75-10-3 1856438 3351283 100 1754.39 0.85 9.21-lo-4 1.01·10"3 l.93'10-3 1683526 3527667

migration time, because in that case orot is given by:

2 2 2 2 0'101="f +O'd=u; +2Ddt (3.16)

In Fig. 3.3 the linear graphs are given, as measured for creatinine, Girard Reagent P, levamisol and halofuginon. As can be seen from this figure, linear graphs · are obtained, but some deviation from linearity can be seen at high voltages (low t). The shape of the curve in this region shows similarities with the shapes of the curves, which are obtained if Joule heating plays a part in peak broadening. Probably the cooling of the capillary is not sufficient at these high voltages to avoid Joule heating. Another remarkable fact from this figure is that the intercept is not zero, indicating that there must be an initial zone width. In this case it can possibly be a oT because the sample components are mixed up and the concentration in the sample plug is not extremely low with respect to the background electrolyte. The regression coefficient, intercept, slope and calculated Dd and m!Dd for these graphs are given in Table 3.ID


v (kV.)

Fig. 3.2: Calculated relationship between plate height and the applied voltage taking into account ( +) only diffusion and (0) diffusion and Joule heating .

. 0.40

0.32

Iii' E 0.24 .g

., 0.16 c. E b

0.08

0.00 0 200 400

t (min}

Fig. 3.3: Measured linear graphs of ~01 versus t for (0) creatinine, (+) Girard Reagent P, (<>) halofuginon and (O) levamisol (sample concentrations 2·10-3 M). Capillary: Siemens: lc=l47.21 cm, ld=l40.51 cm, pressure Injection 2 s. Background electrolyte 0.01 M TRIS/fonnate pH 4.0

3.4 THE SEPARATION NUMBER 47

According to Einsteins relation (eqn. 3.9) the quotient of m!Dd must be constant at constant temperature. Therefore in Table 3.m also this quotient is given. As can be seen from Table 3.m this quotient is rather constant, the mean value (34.65) being, however, lower than the one theoretically expected (38.93). Nevertheless, this method for the determination of diffusion coefficients often leads to values deviating from theoretically ones [14,15]. Better results were obtained by Kenndler and Schwer [16] using CZE without EOF and the stopped flow method to determine diffusion coefficients.

TABLE3.ill

REGRESSION COEFFICIENT, r, INTERCEPT (cm2), SLOPE (cm:2/min) AND CALCULATED DIFFUSION COEFFICIENT Dd (cm2/s) FOR THE LINEAR GRAPHS FROM FIG. 3.3. AND CALCULATED m!Dd.

Component r intercept slope Dd

creatinine 0.9985 0.0275 0.000961 8.008·10-6

Girard Reagent P 0.9980 0.0256 0.000999 8.325·10-6

halofuginon 0.9986 0.0193 0.000597 4.975-Io-6

levamisol 0.9989 0.0252 0.000918 7.650·10-6

3.4 THE SEPARATION NUMBER

In gas chromatography the separation number (SN) [2]:

SN= tR(t.+1)-tR(i) -1 W11,i(i) +W1h(.t+l)

m!Dd

32.59

37.27

35.10

33.63

(3.17)

where tR is the retention time and W11,i the peak width at half height, is used in order to calculate the number of component peaks which can be placed between peaks of two consecutive homologous standards with z and z + 1 carbon atoms with a resolution of R5 = 1.177.

In CZE an analogous expression can be used as parameter for the separation power [17]. If the separation number SNm is defined as:


(3.18)

this number indicates how many components can be separated at an effective mobility m within a unit of effective mobility. Using the absolute value, this equation can be used for both cations and anions independent of the direction of the EOF.

Using eqn. 2.2 and 2.3 the migration time can be calculated from the effective mobility and the mEoF and using for " the expression:

or (3.19)

the SNm can be calculated. In the calculations other effects of zone broadening are neglected.

E

z 10 (J)

0'--' ......... --'--'--'---'---'-~~~~__.___,___, -70 0 70

Fig. 3.4: Calculated relationship between SNm values and effective mobilities/or mE01,.Jol (cm2Ns) of (a) 0, (b) 20, (c) 40 and (d) 60. For further ~lanation see teJCt.

In Fig. 3.4 the calculated relationship between SNm values and effective mobilities is given for several values of EOF (le and ld are 1 m, Eis 25 kV/m). For the diffusion constant Dd in the calculations the Einstein's expression was used:


(3.20)

It can be seen from Fig. 3.4 that if the mobility is tending to zero the SNm is strongly increasing because Dd is decreasing to zero. Very low diffusion constants only will act for very large molecules such as DNA fragments causing a high peak capacity. For small molecules this will be not true as, generally, a very low mobility means that an ionic species for the greater part will be present as a neutral molecule with a rather large diffusion constant. Therefore the SNm was recalculated as a function of the effective mobility, under the assumption that the D d is determined by the Stokes-Einstein relation:

D - kT (3.21) d- 6?r11a

taking arbitrarily an average value for Dd of S· to·10 m2/s.

20~~~....-~.....--.....~~~~~~

d c b a

e w 10

0 ........_.__.__,___.___.__,__,__.__,__..._...__..__..__,

-70 0 70

mobility (10-acm2/Vs)

Fig. 3.5: Calculated relationship between SNm values and effective mobilitiesJ,or mEoF·l<f (cm2tvs) of(a) 0, (b) 20, (c) 40and (d) 60, assuming a diffusion constantDd of5·1<1°m21s.

In Fig. 3 .S this relationship is shown (le and Id are 1 m . , E is 2S ·kV /m ) . It can be clearly seen that the separation number increases at lower effective mobilities owing to the fact that the difference in migration time is increasing at an


equal diffusion constant. In Fig. 3.6 the relationship between the separation number SN20 (at an effective mobility of 20· 10-5 cm2/Vs, le and ld are 1 m) as a function of the applied voltage is given showing that an increasing separation power is obtained applying larger voltages.

10

9

8

7

6 e z 5 (/)

4 a 3 b 2

c d

0 0 5 10 15 20 25

E (kV)

Fig. 3.6: Calculated relationship between SNm values and E §'adient for mwF·l<f (cm21Vs) of (a) 0, (b) 20, (c) 40 and (d) 60 at an effective mobility of 20·1U cm2tvs.

Measurements of separation numbers In screening the separation number is an important aspect because this

number indicates how many peaks can be distinguished within a unit of effective mobility. As already indicated (see Fig. 3.6) this number can theoretically be about 4-8, for cations with mobilities between 20·10·5 and 30·10·5 cm2Ns and different EOF. In practice this number will be much smaller because the total variance will be affected by the variances of the injection and detection and by several other effects causing zone broadening. To get an impression of the order of magnitude in pra~tice an electropherogram of a mixture of 19 ionic species in a HIST/MES


electrolyte system at pH 6.2 was measured and the separation number for the effective mobilities was calculated by:

SN = J 'm-o.s-tm+o.s t m 4cr m

(3.22)

In Fig. 3. 7 the electropherogram for this separation is given and in Table 3.IV all data are presented. In Fig. 3.8 the relationship between the separation numbers (both according to eqn. 3.22 and the theoretical ones, assumin§ a Dd of 5· 10·10 m2/s) and the effective mobilities is given for a mEoF of 47.97· 10-cm2/Vs. From Fig. 3.8 it can be concluded that the practical obtained separation numbers are smaller than the theoretical values owing to several zone broadening effects, especially for cations. In Fig. 3.9 the electropherogram of 10-fold diluted human urine, spiked with levamisol, procaine, clenbuterol and fenoterol is shown. The migration times and calculated effective mobilities are given in Table 3.IV. Using the effective mobilities, the components (I) potassium, (2) sodium, (3) levamisol, (4) procaine, (5) clenbuterol, (6) fenoterol, {7) creatinine, (9) uric acid and (10) hippuric acid can easily be recognized.

0.0050

5 ~ (J)

~ 0.0000 ..0 '-0 (/)

..0 n:I

-0.0050 0

4 3

5

6

a

eo=

9

10 15 20 25

time {min)

Fig. 3.7: The electropherogram of a mixture of 19 components in a background electrolyte HIST/MES al a pH of 6.2. See Table 3.W for the composition of the sample. Capillary from Siemens, I.D. 50 pm, lc=77.33 an and ld=70.53 an. Pressure injection time, 5 s.


15 ....------.-----.---~----,

0'---'----''"---' ........ ~~~~_.__.__,___.__._~

-70 0 70

mobility (10-scm2/Vs}

Fig. 3.8: Relationship between theoretical separation numbers (solid line) and experimentally determlned separation numbers (dashed line) and effective mobility. The experimentally detennlned numbers are calculated from the electropherogram of Fig. 3. 7.

5 ~ QI

~ -e g .0 <U

0.0020

0.0000

-0.0020

-0.0040 0

··-

56 7 4

3

.......

I

eo:: 2

5

g 10

11

a l 12 I . .

10 15 20 25

time (min)

Fig. 3.9: Electropherogram of 10-fold diluted human urine, spiked with levamisol, procaine, clenbuterol and fenoterol (0.()()()1 M) in a HIST/MES background electrolyte at pH 6.2. See Table 3.N for the composition of the sample. Capillary fonn Siemens, I.D. 50 p.m, lc=77.33 cm, ld=70.53 an. Pressure injection time, 1 s.

3.S CONCLUSIONS 53

TABLE3.IV

MIGRATION TIMES, t (min), EFFECTIVB MOBIUTIES, m·toS (cm2Ns), PEAK WIDTH AT HALF HEIGHT, wv., (min), SEPARATION NUMBER, SNm, AND THEORETICAL PLATE NUMBERS, N, FOR SEVERAL COMPONENTS IN A BACKGROUND ELECTROLYTE HIST/MES AT pll 6.2 AND MIGRATION TIMES, t (min), AND EFFECTIVB MOBILITIES, m· loS (cm2Ns), OF THE COMPONENTS IN SPIKED HUMAN URINE.

A and B: peak numbers of the components in Figs. 3. 7 and 3.9, respectively. Peaks 8, 11 and 12 in Fig. 3.9 are unknown. Capillary from Siemens, I.D. SO µm, le - 77.33 cm and ld • 10.53 cm. Applied voltage, 2S kV.

Component A B Sample mixture Human urine

m W11.r SNm N (·I0·5) t m

potassium 1 1 3.19 66.01 0.045 0.37 0.278 3.16 {;5.26 sodium 2 2 3.85 46.47 0.024 1.00 1.42 3.86 44.39 levamisol 3 3 S.02 24.46 0.019 2.15 3.86 4.91 24.25 procaine 4 4 5.27 21.03 0.018 2.49 4.74 S.14 20;93 clenbuterol s 5 5.49 18.26 0.017 2.87 5.77 5.35 18.15 fenoterol 6 6 5.66 16.27 0.020 2.59 4.43 5.51 16.18 creatinine 7 7 7.36 1.43 0.030 2.92 3.33 7.09 1.48 EOF 7.58 47.97 0.050 - 7.30 49.81 o-nitrophenol 8 8.07 -2.91 0.034 3.10 3.12 bromothymol blue 9 11.21 -15.53 0.043 4.73 3.76 uric acid 10 9 13.87 -21.75 0.065 4.79 2.52 12.94 -21.71 hippuric acid 11 10 15.21 -24.06 0.074 S.06 2.34 14.09 -24.00 p-methoxyphenyl· 12 15.84 -25.01 0.081 5.02 2.11

acetic acid p·hydroxyphenyl- 13 17.26 -26.90 0.089 5.42 2.08

acetic acid phenylacetic acid 14 17.45 -27.13 0.092 S.36 1.99 p-nitrobenzoic acid 15 18.44 -28.25 0.094 5.86 2.13 orotic acid 16 19.12 -28.95 0.110 5.38 1.67 benzoic acid 17 19.69 -29.50 0.123 S.11 1.42 sulfanilic acid 18 20.50 -30.23 0.129 S.28 1.40 aspirin 19 22.02 -31.46 0.150 S.24 1.19

3.5 CONCLUSIONS

Theoretically, several effects contribute to peak broadening during a CZE experiment. An estimation of the magnitude of all these factors shows that the main contribution to peak broadening is diffusion, provided that the applied voltage is not too high and the injection plug is narrow. In practice, however, the main contribution is often due to injection. The separation number, which indicates how many components can be separated at an effective mobility m within a unit of effective mobility is a good parameter to indicate the separation power. The experimental values of this separation power are, however, much smaller than the theoretical values.


References

1. J.C. Giddings, Separ. Sd., 4 (1969) 181. 2. C.F. Poole and S.K. Poole, Chromatography today, Elsevier, Amsterdam, The Netherlands,

1991. . 3. F. Foret, M. Deml and P. Boeek, J. Chromatogr., 452 (1988) 601. 4. F. Foret. and P. Boeek, Capillary Electrophoresis, in: Advances of Electrophoresis, Vol. 3.,

A. Chrambach, M.J. Dunn and B.J. Radola (Eds.). VCH Verlaggesellschaft, Weinheim, Germany, 1989. .

5. R. Virtanen, Acta Polytech. Scand., No. 123 (1974) 11. 6. F.E.P. Mikkers, F.M. Everaerts and Th.P.E.M. Verheggen, J. Chromatogr., 169 (1979) 1. 7. H. Weber, Die Partiellen Differential Gleichungen der Mathemalischen Physic, Vol. I,

Braunschweig, 1910. 8. J. Vacik:, in: Electrophoresis, A survey of techniques and Applica1ions: Part A, Techniques,

z. Deyl (Ed.) Elsevier, Amsterdam, The Netherlands, 1979, p. 23. 9. M. Bier, O.A. Palusinsky, R.A. Mosher and D.A. Saville, Science, 219 (1983) 1281. 10. T. Tsuda, K. Nomura and 0. Nakagawa, J. Chromatogr., 248 (1982) 241. 11. H. Bauer and B. Ebner, in: Electrophorese Forum '87, B.J. Radola (Ed.), Technische

Universitit, Miinchen, Germany, 1987, p. 556. 12. D.S. Burgi and R.-L. Chien, Anal. Chem., 63 (1991) 2042. 13. J.L. Beekers, Th.P.E.M. Verheggen and F.M. Everaerts, J. Chromatogr., 452 (1988) 591. 14. S. Terabe, 0. Shibata and T. Isemura, J. High Res. Chromatogr., 14 (1991) 52. 15. H.K. Jones, N.T. Nguyen and R.D. Smith, J. Chromatogr., 504 (1990) 1. 16. E. Kenndler and C. Schwer, Anal. Chein., 63 (1991) 2499. 17. J.L. Beckers, F.M. Everaerts and M,T. Ackermans, J. Chromatogr., 531(1991)407.

CHAPrER4

QUANTITATIVE ASPECTS

In this chapter a concise summary is given of some statistics, dealing with accuracy and precision of a separation method, validation of the method, calibration graphs and limits of detection. Further, a method validation of capillary zone electrophoresis with respect to high-performance liquid chromatography and isotachophoresis is given.

4.1 INTRODUCTION

On applying a new instrumental analysis method, the accuracy and the precision of the method have to be known, and the method must be validated. For the quantification, calibration graphs with several points per calibration graph are used and several readings from the unknown sample are made, in order to determine the concentration. In this chapter, the statistics used to validate the method and to evaluate the accuracy and precision, and the calibration graph and limits of detection, are discussed.

4.2 STATISTICS

4.2.1 Fundamental concepts

Once it is accepted that quantitative studies play a dominant part in any analytical laboratory, it must be accepted that the errors that occur in such studies are of supreme importance. In these errors, a distinction can be made between gross, random and systematic errors. Gross errors are readily described: they may be defined as errors, which are so serious that there is no real alternative to abandoning the experiment and making a completely fresh start. Random errors are errors that cause the individual results to fall on both sides of the average, and they are said to affect the precision of an experiment. Statistics can be used to evaluate these errors. Systematic errors are errors that cause all the results to be in error in the same sense (for instance, all results are too high), and they are said to affect the accuracy, i.e. the proximity to the true value. Mostly, the precision of an analytical method can be

56 CHAPTER 4: QUANTITATIVE ASPECTS

evaluated in a relatively easy way, the accuracy is much more difficult to determine. In the following part only precision will be discussed, assuming only random errors with a Gaussian distribution.

For an analytical method, the precision can be split up in two parts: the within-run or within-day precision or repeatability and the between-run precision or reproducibility. The repeatability of an experiment is defined as the precision on short terms, thus the precision measured if one experiment is replicated ;within a short time under the same experimental conditions. The reproducibility, also called the long-term precision, is the precision measured if an experiment is replicated within a longer time, with other experimental conditions.

In the comparison of results of experiments, normally the average value x, and the standard deviation, s, defined as:

- "{"' x,.· X= ,,t_.,-

i n and

. <xrX>2 { }

llz

S= L-i (n-1)

(4.1)

are used, where n is the number of experiments. Also widely used is the coefficient of variation, also known as the relative standard deviation (RSD, %), given as:

RSD = 100 !... Xt

(4.2)

4.2.2 Outliers

If an experiment is performed several times it can occur that one (or possibly more) of a set of results appears to differ unreasonably from the others in the set. As a method to test whether a certain value is an outlier or belongs statistically to the measured series, the Dixons Q test can be used. From a suspect value the Q value is calculated as:

Q = I suspect value - nearest value I largest value - smallest value

(4.3)

The critical values of Q can be found in literature [l,2]. If the calculated value of Q exceeds the critical value, the suspect is rejected.

4.2.3 Calibration graphs ln instrumental analysis

Usually in instrumental analysis, a calibration graph is set up using several standard concentrations, covering a wide range. Once ·the calibration graph is established, the analyte concentration in any test sample is obtained by interpolation [3]. This general procedure raises several important statistical questions: 1. Is the calibration graph linear? If it is a curve, what is the form of the curve?

4.2 STATISTICS 57

2. Bearing in mind that each of the points on the calibration graph is subject to errors, what is the best straight line (or curve) through these points?

3. Assuming that the calibration plot is actually linear, what are the estimated errors and confidence limits for the slope and the intercept of the line?

4. When the calibration plot is used for the analysis of a test sample, what are the errors and confidence limits for the determined concentrations?

5. What is the limit of detection of the method? That is, what is the least concentration of the analyte that can be detected with a predetermined level of confidence?

Before tackling these questions in detail, a number of aspects of the calibration graph must be considered. First, it is essential that the calibration standards cover the whole range of concentrations required in the subsequent analyses. Secondly, it is crucially important to include the value of a 'blank' sample in the calibration curve, and finally it should be noted that the calibration curve is always plotted with the instrumental response on the ordinate and the standard concentrations on the absciss. This is because many of the procedures for the statistical treatment of the data assume that all errors are in they-values, and that the x-values are error-free.

The product-moment· correlation. coefficient Assuming a straight line, the algebraic form will be:

y =bx+ a (4.4)

where b is the slape and a is the intercept. The individual pc>ints of the curve are referred to as (xi, Yi) and the mean values of the x and y as x and y. In order to estimate how well the experimental points fit a straight line, the product-moment correlation coefficient, r, is used. This statistic is often referred to as correlation coefficient because in quantitative sciences it is by far the most commonly used type of correlation coefficient. The value of r is given by:

:E {(xriHYrY)} r = ___ i _______ _

(4.5)

The Une of regression of y on x Assuming a linear relationship between the analytical signal (y) and the

concentration (x), the best straight line through these points can be found using the method of_ t!!_e least squares. Following this method, the straight line passes the centroid (x,y) ofthe points, and the slope and intercept are found by minimizing deviation in the y-direction.

The slope and the intercept are calculated according:


(4.6)

a= y-bx (4.7)

In order to calculate the errors in the slope and the intercept the statistic parameter Sy/x• defined as:

(4.8)

is used, where Yi is the fitted y-value corresponding to the individual x-values. Armed with this value Sytx the standard deviations, sb and s~. for the slope and the intercept can be calculated according to: ·

(4.9)

(4.10)

Calculation of a concentration Once the slope and the intercept of the regression line have been determined, it

is very simple to calculate an x-value corresponding to any measured y-value. A. more complex problem arises when it is necessary to estimate the error in the concentration calculated by using the regression line, because both the slope and the intercept are used and as shown in the previous section, both are subject to errors. For the calculation of the standard deviation, sX-O, of the concentration calculated from the measured response y0, the following equation can be used:

4.2 STATISTICS 59

(4.11)

where m is the number of readings made for the detennination of Yo and n is the number of calibration points.

limits of detection In general terms the limit of detection is defined as the lowest concentration

which gives an instrument signal (y) significantly different from the blank or background signal. A commonly used definition in the literature for analytical chemistry is that the limit of detection is the analyte concentration giving a signal equal to the blank signal, Ye· plus three times the standard deviation of the blank, sB, as shown in Fig. 4.1. In practice, as an estimate for YB• the intercept, and as an estimate for sB, Sytx is taken.

(i.j 20

..... ·c: :J

£ 15 .!)!

(I) (.)

~ 10 .... 0 {/)

~ 5

y 9+3s9 - LOD Ye

0 0 2 4 6 8 10 12

concentration (pg/ml}

Fig. 4.1: Regression line for a caUbration curve. The limit of detection is given by Ys + 3ss, where the ordinate intercept is an estimate for y8 , and s

11x is an estimaJe for s8•


4.3 METHOD VALIDATION

For CZE as rather new analytical tool, it is of interest to compare its performance with that of HPLC and ITP, which are more established separation techniques. In this comparison, validation of the methods is an important task. The validation requires a demonstration of the specificity, sensitivity, calibration linearity, precision and accuracy of the methods. For this reason the within-day precision (repeatability) and between-day precision (reproducibility) are considered and the calibration graphs obtained for the different methods are compared.

Within-day precision In order to obtain an impression of the within-day precision of the methods,

replicate separations (n= 10) were made of sample mixtures of five components with the apparatus for ITP, HPLC and CZE (for a precise description of the CZE instrumentation used see section 1.5, for the HPLC and ITP instrumentation see section 6.2.2).

With the ITP apparatus separations of a mixture of chloric, malonic, pyrazole-3,5-dicarboxylic, acetic .and glutamic acid (all at a concentration of 8· 104 M) applying a leading electrolyte of 0.01 M HCl adjusted at pH 6.0 by adding histidine and a terminating electrolyte of 0.01 M MES were performed. The electric current was 25 µ.A. In Table 4.I the average values and relative standard deviations are given for the relative step heights as a percentage of the step height of pyrazole-3,5-dicarboxylic acid and the zone lengths measured with the conductivity detector.

TABLE4.I

AVERAGE VALUES (AV) AND RELATIVE STANDARD DEVIATIONS RSD (%) OF THE RELATIVE STEP HEIGHTS, RSH (AS % OF THE STEP HEIGHT OF PYRAZOLE-3,5-DICARBOXYLIC ACID) AND ZONE LENGTHS, 7L (s), WITH ITP

Component RSH RSD 7L RSD (AV) (AV)

cbloric acid 63.6 1.35 13.23 1.18 malonic acid 83.6 0.54 24.23 0.90 pyrazole-3,5-dicarboxylic acid 100.0 27.02 0.65 acetic acid 141.6 0.71 15.93 1.57 glutamic acid 237.1 1.31 19.58 0.62

With the HPLC apparatus a mixture of toluene (5.9· 10-5 M), eth/1- (5.1 ·10-5 M), propyl- (4.5·10-5 M), butyl- (8·10-5 M) and pentylbenzene (7.3·10- M) in methanol was analysed, applying methanol-water (80:20, v/v) as eluent. The flow-rate was 1.0 ml/min. Wavelength UV detector 209 nm. In Table 4.Il the average values and relative standard deviations for the retention times and peak area are given.

With the CZE apparatus a mixture of salbutamol (1.75·10·5 M), creatinine (2· 10-5 M), aniline (l · 104 M), benzoic acid (l · 10-5 M) and m-aminobenzoic acid (2· 10-5 M) was analysed, applying a background electrolyte of 0.01 M TRIS at pH 5.0

4.3 METHOD VALIDATION 61

TABLE4.II

AVERAGE VALUES (AV) AND RELATIVE STANDARD DEVIATIONS, RSD (%), OF THE RETENTION TIMES, tR (min) AND PEAK AREA, A (AUs} WITH HPLC

Component tR RSD A RSD (AV) (AV)

toluene 3.58 0.11 5.47 0.92 ethylbenzene 4.43 0.15 5.36 1.00 propylbenzene 5.91 0.25 4.18 0.90 butylbenzene 8.19 0.26 6.52 1.01 pentylbeDz.ene 11.66 0.34 7.15 1.42

TABLE4.ffi

AVERAGE VALUES (AV) AND RELATIVE STANDARD DEVIATIONS, RSD (%),OF THE MIGRATION TIMES, t (min), EFFECTIVE MOBILITIES, m·loS (cm2Ns), AND PEAK AREA, A (mAUs) WITH CZE

Component t RSD m RSD A RSD (AV) (AV) (AV)

salbutamol 5.54 1.01 19.07 0.72 12.29 1.51 creatinine 6.12 1.02 12.61 0.90 6.72 2.08 aniline 6.43 1.05 9.56 1.15 26.37 1.21 m-aminobenzoic acid 12.96 1.71 -20.20 0.95 21.29 0.88 benzoic acid 17.19 2.19 -27.42 0.80 22.63 0.93

adjusted by adding acetic acid. In Table 4.ill the average values and relative standard deviations are given for the migration times, calculated effective mobilities (4] and peak area.

The within-day precision for the techniques is about 1-2 %.

Between-day precision To establish the between-day precision, peak area (n=5) for CZE and HPLC and

zone lengths (n=3) for ITP (both with the conductivity and UV detector) of salbutamol sulphate are measured and the average values and RSDs for several different concentrations of the solute are calculated. This series was repeated with freshly prepared electrolyte solutions after 1 week. The results are given in Table4.IV. From Table 4.IV it can be concluded that the reproducibility of the HPLC experiments is by far the best. The high RSD of 6.21 % in the second series of the CZE experiments is due to one bad value, that could not be considered statistically as an outlier, however. The ITP experiments are carried out only three times in order to be able to measure the complete calibration graph in 1 day (see section 6.3.2 for the electrolyte systems).


TABLE4.IV

BETWEEN-DAY PRECISION: AVERAGE VALUES OF PEAK AREA A (A Us) MEASURED WITH HPLC AND CZE AND ZONE LENGTHS ZL (s) MEASURED WITH lTP AND THE RELATIVE STANDARD DEVIATIONS, RSD (%) FOR SEVERAL CONCENTRATIONS, c (mg/ml) OF SALBUTAMOL SULPHATE, WITH REGRESSION PARAMETERS (r = REGRESSION COEFFICIENT, b = SLOPE AND a = INTERCEPT)

c HPLC (n=S) CZE (n=S}

Day 1 Day2 Day 1 Day2

A RSD A RSD A RSD A RSD

0.100 24.08 0.46 24.89 0.19 59.43 0.18 58.36 6.21 0.075 18.75 0.09 18.58 0.08 45.78 0.58 44.15 0.54 0.066 16.Sl 0.13 16.44 0.11 39.44 0.20 39.06 1.14 0.050 12.39 0.14 12.28 0.13 29.38 1.09 30.16 1.55 0.033 8.24 0.14 8.23 0.06 19.33 0.17 19.90 0.29 0.025 6.13 0.04 6.16 0.09 14.27 0.28 14.48 0.20 0.010 2.51 0.02 2.51 0.01 5.12 0.04 5.60 0.15

r 0.99943 0.99996 0.99967 0.99976 b 242.92 248.70 604.85 586.10 a 0.20 -0.027 -0.54 0.11

c ITP (COND)(n=3) lTP (UV)(n=3)

Day 1 Day2 Day 1 Day2

ZL RSD ZL RSD ZL RSD ZL RSD

I.OOO 124.93 0.66 126.37 0.34 124.60 1.00 126.60 0.47 0.750 94.15 0.97 95.42 1.05 94.00 0.98 95.80 1.30 0.625 76.43 0.69 77.52 1.10 76.20 0.79 78.20 1.17 0.500 64.22 0.72 64.55 0.60 64.60 1.07 63.63 0.95 0.375 46.95 0.32 46.62 0.54 46.20 1.30 46.60 1.49 0.250 31.33 0.24 31.40 0.55 31.60 1.10 31.40 1.10 0.100 12.65 0.40 12.85 l.56 12.38 1.22 12.37 0.71

r 0.99967 0.99970 0.99950 0.99988 b 124.57 126.44 124.42 127.31 a 0.32 -0.065 0.24 -0.53

Comparison of CZE, I'IP and HPLC For the comparison of the separation methods the peak area respectively zone

lengths for samples of salbutamol sulphate from 1 mg/ml to 0.001 mg/ml for HPLC and CZE and to 0.01 mg/ml for ITP were measured, and the linear regressiort lines were compared. All zone lengths and peak area were recalculated as percentages of the highest values for each method. In Fig. 4.2A, B, C and D the values are presented

4.3 METHOD VALIDATION 63

A B

ZL in% <ITPI ZL in % (ITP}

c D

A In % IH='l.C) ZL in % o..M

Fig. 4,2: Regression lines for the measured peak area A respectively zone lengths ZL as percentages of the maximum valuafor (A) HPLC versus JTP (conductivity signal), (B) CZE versus JTP (conductivity signal), (C) CZE versus HPLC and (D) C-Onductivity versus UV signal/or JTP.

graphically (logarithmic scale) for HPLC versus ITP (COND), CZE versus ITP (COND), CZE versus HPLC and conductivity versus UV signal for ITP, respectively.

In Table 4. V the slope, intercept, regression coefficient and limit of detection calculated for the regression lines (non-logarithmic scale) are given. The obtained linear relationships, with a slope of nearly 1 · and nearly zero intercept, validate the techniques. As can be seen from Table 4. V the limit of detection of ITP is significantly higher than that of CZE and HPLC. On applying e.g. a column-coupling system or lower electric currents this LOD can however be lowered (see section 6.3.3).


TABLE4.V

SLOPE, INTERCEPT, REGRESSION COEFFICIBNT AND LIMIT OF DETECTION FOR THE INDIVIDUAL REGRESSION LINES OF PEAK AREA RESPECTIVELY ZONE LENGTHS FOR SALBUTAMOL SULPHATE MEASURED WITH ITP, HPLC AND CZE AND CORRELATIONS BETWEEN THEM.

Regression line of Slope Intercept Regression Detection coefficient Limit (/.lg/ml)

lTP (COND) 139.21 0.1298 0.998929 41.20 lTP (UV) 11S.4S 0.8537 0.998508 58.38 CZE 744.97 -0.8805 0.999972 6.0S HPLC 283.72 0.3685 0.999960 7.17

HPLC-ITP (COND) 1.004 -0.096 0.999872 CZE-ITP (COND) 0.999 -0.592 0.999804 CZB-HPLC 0.991 -0.296 0.999871 ITP (COND)-ITP (UV) 1.005 -0.604 0.999759

4.4 CONCLUSIONS

· On applying a new analytical method, the accuracy and the precision of the method have to be known, and the method must be validated. In this chapter the method validation of CZE, compared to HPLC and ITP is given. The overall withinday precision and the between-day precision of the techniques are ea. 1-2 % • Comparing the results obtained for analysis of standard samples with the three techniques, linear relationships with a slope of nearly 1 and nearly zero intercept ·were obtained, validating the technique. The use of calibration graphs and the calculation of limits of detection with these three techniques is described in section 6.3 for the determination of some /3z-agonists in several pharmaceutical dosage forms.

References

1. E.P. King, J. Am. Statist. AssQc., 48 (1958) 531. 2. A.I. Bosch and H.J.L. Kamps, (Eds.), Statistic CQmpendium, Eindhoven University of

Technology, Eindhoven, The Netherlands. 3. J.C. Miller and I.N. Miller, Statistics for analytical chemistry, Ellis Horwood, Chichester,

West Sussex, England, 1988. 4. J.L. Beckers, F.M. Everaerts and M.T. Ackermans, J. Chromatogr., 531(1991)407.

CHAPTERS

CHARACTERISTICS OF CZE

Jn 1897, Koblraush formulated bis •Bebarrliche Funktion", which prescribes that, in an electrophoretic process, at any point the sum of the concentrations divided by the absolute values of the mobilities must be constant. Using this relation, an interesting point in quantitative CZE, on applying conductivity or indirect UV detection with non-UV absorbing components, is the existence of a relationship between effective mobilities and peak area, independent of the kind of ionic species. This relationship is considered theoretically for fully ionized monovalent ions resulting in a linear relationship, passing through the origin, between temporal peak area and the product of a correction factor (dependent only on the effective mobilities of the ionic species) and migration time for an equimolar sample composition. A good correlation between theory and practice could be established by applying measured data.

5.1 INIRODUCTION

In 1897 Kohlrausch formulated his "Beharrliche Funktion", which prescribes that at any point the sum of the concentrations divided by the absolute value of the mobilities must be constant. Using this regulation function, quantitative analyses in CZE with conductivity and indirect UV detection have some interesting properties. In this chapter, the regulation function of Kohlraush is described briefly and the special effects in quantitative analysis in CZE with conductivity and indirect UV detection are discussed.

5.Z KOHLRAUSH REGULATION FUNCTION

The electrophoretic process in capillary electrophoresis can be described by the continuity equation. Under the assumptions that no chemical reaction takes place and that in the current transport diffusion can be neglected with respect to migration, the continuity equation for a one dimensional case, for strong monovalent ions is:

66

ac. a a2c. _1 =--Em1·C1·+Dd._1 at ax '1

OX2

CHAPTER 5: CHARACTERISTICS OF CZE

(5.1)

where t is the time, x is the place, Eis the voltage gradient, and c, m and Dd are the concentration, mobility and diffusion coefficient of component i, respectively.

Starting from this equation Kohlraush [l] derived his regulation function, assuming that diffusion can be neglected, thus:

(5.2)

Assuming that Eis constant (which is the case for pure electrophoretic processes) and summation of all equations for all ions of all substances present in the solution gives:

(5.3)

or after dividing by I mi I : a c, a mi at~ lmil =-E ox~ lmil ci

(5.4)

(Note that m/ I lni_ I is in fact the sign of the charge of component l). Assuming electroneutrality in the form

gives, applying eqn. 5.4:

a ci -I:-=O at j lmd

and thus:

E C; i lmil =w

(5.5)

(5.6)

(5.7)

is constant in time. This equation is known as the regulation function ofKohlraush [l]. Later, the regulation function was also derived for multivalent strong electrolytes and univalent weak electrolytes [2]. The importance of the regulation function stems from its ability to describe the systems in a telling way.

In CZE, the regulation function enables one to describe in a telling way the phenomena which take place e.g., at the point of sample injection. Here, the discontinuity formed by sample injection into the electrophoresis capillary filled with

5.3 CONDUCTNITY AND INDIRECT UV DETECTION IN CZE 67

background electrolyte persists even after the sample ions have migrated away and this place is occupied by the background ions. Thus two stationary concentration boundaries are formed at the point of injection and the concentration between them is adjusted to the value of the regulation function of the sample before the experiment. In the same way, the concentration of the ionic sample species is adapted to the value of the regulation function of the background, once they have passed the concentration boundary. This mechanism can be advantageously used for the sharpening of the injection pulse [3]. In this case, the sample is injected as a highly diluted solution, and its zone is automatically compressed into a sharp pulse as it crosses the plane at the position of the original boundary between sample plug and background electrolyte (see also section 3.3.1).

5.3 QUANTITATIVE ANALYSIS IN CZE WITH CONDUCTIVITY AND INDIRECT UV DETECTION

5.3.1 Introduction

Huang et al. [4] reported on the unique advantage of quantitative capillary zone electrophoresis using conductivity detection, that use of an internal standard allows accurate determination of a concentration in a mixture without separate calibration of the response for each component and they found a direct relationship between peak area and migration times. Further consideration of the principle of the measured conductivity shows, however, that although a relationship exists between peak area and migration time it is nearly linear only over a small mobility range. In this section, the relationship between peak area and effective mobility for conductivity and indirect UV detection in CZE is considered for fully ionized monovalent ions.

5.3.2 Theoretical

Assuming only the presence of fully ionized monovalent ionic constituents, some effects in the electrophoretic separation mechanism can approximately be described by Kohlrausch's regulation function:

(5.8)

The numerical value of the Kohlrausch function&> is locally invariable with time [S]. If a volume element of the capillary is originally filled with a carrier electrolyte AB {consisting of a co-ion A and counterion B) at a concentration cA, it will contain after some time a mixture of sample components and carrier electrolyte when sample components pass, but finally the original situation will be restored again. If a mixture of a component i and the carrier electrolyte AB passes through such a volume element, the following equation is valid:

68 CHAPTER 5: CHARACTERISTICS OF CZE

(5.9)

with

(5.10)

The superscripts C and S refer to the composition of the pure carrier electrolyte AB zone and the sample zone, respectively. The concentration of the counterion B is determined by the electroneutrality condition.

For the zone conductivity er can be derived:

(5.11)

where

(5.12)

Applying a conductivity detector in capillary electrophoresis, a detector response, directly related to a5-<F, can be expected being linear proportional to bici and hence the spatial [6] peak area will be proportional to the product of bi and the injected amount Qinj·

Generally the measured peak area will be expressed on a temporal basis [6] and it can be expected that for CZE both without and with EOF the measured peak area Ai will be proportional to:

(5.13)

As the migration time ti is reversely proportional to I mi I or I mi+mEoF I {without and with EOF), at a given voltage the measured peak area Ai will be proportional to:

or (5.14)

The relationship between measured peak area Ai and b/ I mi I , b/( I mi +mEoF I ) or biti must be linear passing through the origin, whereas the products Ai I mi I !bi, Ai( I mi+mEoF I )/bi or A/bli should be a constant for all different ionic species for an equimolar sample composition. It must be remembered that generally in chromatographic techniques one has to work with spatial peak area as the components move with equal speed through the detector.

In Fig. 5.1 the peak areas (arbitrary units), calculated according. to eqn. 5.14 (without EOF) for a given Qini , are given as a function of the migration times for ionic species with effective mobilities varying from -80·10"5 to -20·10"5 cm2/Vs and assuming effective mobilities for mA and m~ of (a) -30· 10"5 and 30· 10-5, (b) -80· 10·5

and 30· 10·5 and (c) -55· 10"5 and 30· 10·5 cm /Vs. It can be clearly seen from Fig. 5.1

5.3 CONDUCTIVITY AND INDIRECT UV DETECTION IN CZE 69

that the peak area changes sign at a mobility m;, equal to m A and increases with larger differences between m;, and mA- Further, the relationship is not linear, although for a fairly small mobility range (the arrows indicate broadly the mobility range of formate to hexanoate) it is nearly linear.

1.0

0.0

"iii -1.0 a ..., ·c: :;) -2.0

j -3.0

(\) -4.0 ~ c

111

.:<: -5.0 111

~ -6.0

-7.0 20 b

-8.0·. o.oo 0.05

time (arb. units)

Fig. 5.1: Calculated relationship between temp_oral peak area and migration time (arbitrary units) far ionic species with ejfeCJive mobilities of-80· I<ls to ·20· JUs cm2 /Vs assuming effective mobilities for coand counter-ions of (a) -30·JU5 and 30·I<T5, (b) -80·10·5 and 30·1<T5 and (c) -55·JU5 and 30·1U5

cm2/Vs. The numbers refer to the effective mobilities ·m·ltY cm21Vs of the ionic species.

In Table 5.I the measured peak area and migration times for an equimolar sample composition, according to Huang et al. [4] and the calculated ratios c/cst, with hexanoate considered as standard (st), using the equation

(5.15)

as used by Huang et al. [4] are given. It can be concluded from Table 5.I that this equation can not be used in general, as the quotients have to be unity for an equimolar sample composition.

In Fig. 5.2A and 5.2B the relationships between peak area, as measured by Huang et al. [4], and the calculated biti or ti are given. Eqn. 5.13 (and not eqn. 5.14) was applied, because it was unknown whether the EOF was fully suppressed by the addition of TI AB. It can be clearly seen that the linear relationship between temporal peak area and migration time (dashed line), as measured by Huang et al. changes into a linear relationship, nearly passing through the origin, on applying the correction factor bi (solid line), as expected from theory. The results in Fig. 5.2B are better than those in Fig. 5.2A (large intercept on the ordinate) because the accuracy of the effective mobility of MES is very critical to the calculated value of bi. In the calculation of the factor bi the mobilities have been corrected according to the DebyeHiickel-Onsager theory.


TABLE 5.I

MEASURED PEAK AREAS, A; (ARBITRARY UNITS), MIGRATION TIMES, t; (s), AND RATIOS c/cSt CALCULATED WITH THE EQUATION ACCORDING TO REF. 4

Background electrolyte: (I) 0.01 M MES/HIST at pH 6; (II) 0.005 M Cbloride/TRIS at pH 7.1.

Ionic species II

A; ti qlcs1; A; t; c/cst

formate 53.l 113 1.83 10.5 280 0.14 acetate 37.0 148 1.67 26.7 338 0.42 propionate 29.4 164 1.46 33.8 369 0.58 butyrate 25.5 175 1.36 39.1 390 0.70 pentanoate 20.6 185 1.16 42.4 405 0.79 hexanoate 16.9 195 1.00 51.0 425 1.00

A B 60

1! ·5

j "'

30

;

t o......_.~~~........_~ .......... ......_.

0 llO 100 1150 :200 :2!!50

b,*t. or t. larb. U'litsl

Fig. 5.2: Relationship between measured peak area [4} and bit (solid line) and t; (dashed line) for the electrolyte system (A) 0.01 M MES adjusted at pH 6 by adding HIST and (B) 0.005 M HCI adjusted at pH 7.1 by adding TRIS. The sample consisted of a mixture of formate, acetate, propionate, butyrate, pentanoate and he:xanoate. Applying the correction/actor b; a linear relationship passing through the origin is obtained according to eqn. 5.H.

It is obvious that a relationship between peak area Ai and retention time ti, not linear and not passing through the origin, is not practical to be used for an internalstandard, in contrasted with the use of the relationship between peak area Ai and the product hli·

5.3 CONDUCTIVITY AND INDIRECT UV DETECTION IN CZE

For a UV detector the measured absorbance A will be

A=ecl

71

(5.16)

where E is the molar absorption coefficient and 1 is the effective path length in the detector. For the carrier electrolyte this means:

For a sample zone the absorbance will be:

A 8 = (eA +e8 )ciz +(Es +et)c;8 l

The UV signal of a sample zone, using eqn. 5.9, will be

A =Ac _As =ci8

l[(eA +cs)ki-(t:1+es)]

(5.17)

(5.18)

(5.19)

For non-UV-absorbing counter ions and sample ions, applying indirect UV detection with UV-absorbing co-ions, the UV signal is proportional to cA

Analogously to conductivity detection, the spatial peak area will be proportional to kiQinj and the measured peak area Ai on a temporal basis to kiQiili· Further, the expression Ailkiti has to be a constant in a given electrolyte system for all components at an equimolar sample composition. ·

5.3.3 Experimental

Instrumentation For all quantitative CZE experiments with a conductivity detector a laboratory

built capillary electrophoresis system with an on-column conductivity detector . as described previously [7] was used. As the apparatus is a closed system, the EOF is fully suppressed. The sampling takes place into a broadened part of the capillary tube (0.55 mm I.D.) connected with two feeders (0.4 mm diameter), perpendicular to the capillary tube. A constant d.c. power supply with a maximum potential of 20 kV was used. Peak areas were determined using the integration program CAESAR. The detector electronics were connected with an IBM XT PC via a LabMaster (Scientific Solutions, Solon, U.S.A.).

All quantitative CZE experiments with a UV detector were performed with the PI ACE System 2000 HPCE applying UV detection at 254 nm. All experiments were carried out at 25 °C using an original Beckman capillary of 57 cm, with a distance between injection and detection of 50 cm and an I.D. of 75 µm. For all zone electrophoretic separations the injection took place at the inlet side. In the anionic mode the cathode was placed at the inlet and the anode at the outlet side, and vice versa for the cationic mode.

Reagents and samples All chemicals were of analytical-reagent grade. Before preparing the sample

solutions, all chemicals were dried at 105 °C.


5.3.4 Results and discussion

In order to check the relationship between temporal peak area and mobilities for both conductivity ( eqn. 5 .14) and indirect UV detection ( eqn. 5 .19) the temporal peak area with conductivity detection and indirect UV detection, both in the anionic and cationic mode [8] were measured.

TABLE5.Il

EFFECTIVEMOBILmES,-mi·lo'(cm2/Vs),CALCULATEDFACTORS,1'j(ARBITRARYUNITS), MEASURED PEAK AREAS, Ai (ARBITRARY UNITS) AND CALCULATED VALUES OF K ( =~ I mi I !bi) FOR DIFFERENT BACKGROUND ELECTROLYTES IN CLOSED SYSTEMS

Component mi 1'j Ai K

0.01 M MES+ imidazole pH 7 chloride 74.45 79.30 218.47 205.1 chlorate 62.61 65.00 209.08 201.4 fluoride 53.22 52.15 201.93 203.7 formate 52.43 51.68 200.34 203.3 acetate 38.54 30.50 163.76 206.9 propionate 33.36 20.94 136.66 217.7 benzoate 29.93 13.75 112.66 245.2

0.01 M HAc + imidaz:ole pH 7 chloride 74.17 56.84 122.71 160.1 chlorate 62.35 40.93 103.74 158.0 fluoride 52.97 26.87 78.77 155.3 formate 52.18 25.62 72.51 147.7 propionate 33.13 -12.19 -53.10 145.9 benzoate 29.71 -21.61 -100.35 137.9

0.01 M HCl + imidaz.ole pH 7 chlorate 62.35 -20.14 -31.74 98.3 fluoride 52.97 -38.88 -68.82 93.8 formate 52.18 -40.60 -72.95 93.8 acetate 38.31 -77.67 -180.98 89.3 propionate 33.13 -96.60 -253.93 87.1 benzoate 29.71 -111.75 -316.45 84.2

In Table 5.II, the effective mobilities mi, calculated factors ~. temporal peak areas Ai (measured with a closed CZE apparatus with on-line conductivity detector [7]) and calculated values of K are given. The factors K (=Ai I mi I /bi) are virtually constant in the three different electrolyte systems although a disadvantage of the sample injection used in our apparatus is that although linear calibration curves are obtained for both ITP and CZE experiments, the effective injection volumes for the different components are not completely identical [7] and moreover the separation power of the apparatus used was too small to separate the whole sample mixture in a


single experiment. Of course the values of the factors K for the different systems are different due to different circumstances. The three different electrolyte systems consisted of the co-ions MES, acetate and chloride at pH 7 adjusted by adding imidazole (anionic mode, constant electric current 10 p.A). The sample components were chloride, chlorate, fluoride, ·formate, acetate, propionate and benzoate at a concentration of 5· 104 M.

In Fig. 5.3 the measured peak areas of Table 5.II are given as a function of time. The similarity with Fig. 5.1 is obvious. In Fig. 5.4 the measured peak areas of one of the electrolyte systems of Table 5.II are given as a function of both (a) the migration time ti and (b) hh. It can be clearly seen that relationship (a) changes into (b) a linear relationship passing through the origin.

To check the relationship between peak area and mobilities for the indirect UV detection with non-UV-absorbing components and counter ions, experiments were carried out in open systems, in both the anionic and cationic mode. In the anionic mode three background electrolytes were used, viz., 0.01 M benzoic acid, 0.01 M nicotinic acid and 0.01 M sulphosalicylic acid adjusted at pH 8 by adding TRIS. The sample mixture consisted of chloride, chlorate, fluoride, acetate, propionate and MES (5· 104 M), applying pressure injection times of 5, 10 and 15 s. All experiments were carried out with a constant voltage of 25 kV. In order to suppress the EOF for the greater part, 0.05 % MHEC was added to all solutions.

In Table 5.III the calculated effective mobilities and calculated factors ki. the measured migration times ti, measured peak area Ai and calculated values of K ( =A/kh) are given. The factor K should be a constant for all components in the same carrier electrolyte. It can be concluded from Table 5.III that the factor K is indeed a constant for all components and is linearly related to the amount of the components injected. In Fig. 5.5 an example of the separation of a mixture (in the anionic mode, pressure injection time 15 s) with indirect UV detection is given. The carrier electrolyte was 0.01 M benzoic acid at pH 8. The sample consisted of (1) chloride, (2) chlorate, {3) fluoride, (4) acetate, (5) propionate and (6) MES at a concentration of 5· 104 M. In order to obtain an impression of the velocity of the EOF (remember that the EOF is suppressed for the greater part by the addition of 0.05 % MHEC to all solutions) an EOF peak directed from outlet to inlet by injecting water at the outlet side by electromigration injection at 10 kV for 5 s was created (the length from injection to detection is only 7 cm for this EOF marker; the mEOF is about 14· 10·5

cm2/Vs; without MHEC the mEOF is about 67· 10·5 cm2/Vs). Note the fronting respectively tailing shapes of the peaks with high and low effective mobilities respectively. ·

In the cationic mode an electrolyte system consisting of 0.01 M benzoic acid adjusted at pH 8.5 by adding TRIS was applied. The applied voltage was 25 kV. The sample consisted of formate, acetate, propionate, butanoate, pentanoate and hexanoate at a concentration of 5· 104 M, applying pressure injection times of 5, 10 and 15 s. In order to study the repeatability of the method, all experiments were carried out five times. In Table 5 .IV the effective mobilities and calculated factors ki, the measured migration times ti, measured peak area Ai and calculated factors K ( =A/'fG.ti) are given. The effective mobilities of the components are calculated from the migration times of the EOF and of the components as described previously [9]. It can be concluded from these data that the repeatability is good for the migration times and the calculated

74 CHAPTERS: CHARACTERISTICS OF CZE

300 ...------------

2 100

·5 e .!!

-100

~ ,,;. «l

~ -JOO

-500 '---~---'---~---' 0.00 0.()4

t, (arb. units}

Fig. 5.3: Relationship between measured peak area and t1for conductivity detection in a closed system, for three different electrolyte systems. The sign of the peak area changes at an effective mobility equal to that of the co-ion confirming the theory (see Table 5.11/or data). Background electrolyte (a) MES, {b) acetic acid and (c) hydrochloric acid ail at pH 7 adjusted by adding lmidazole.

200 ----~------

!.il 100

~ € .!!!

0 \ «l

QJ \ ... «l \

Jt. \ Si -100 Q.

b \

\ -200

-3 -1 3 5

b,*t, or t, (arb. units)

Fig. 5.4: Relationship between measured peak area and (a) t1 and (b) bit for the electrolyte system 0.01 M acetic acid adjusted at pH 7 by adding imidazole. The sample consisted of a mixture of chloride, chlorate, fluoride, formate, acetate, propionate and benzoate (5· Jcr4 M). It can be clearly seen that the relationship between peak area and t1 (dashed line) changes in a linear relationship passing through the origin if the correction factor bi Is used (solid line).


TABLE5.m

EFFECTIVE MOBILITIES, -mi·toS (cm2Ns), CALCULATED FACTORS ~ ACCORDING TO EQN. 5.10, MEASURED MIGRATION TIMES, ti (min), MEASURED TEMPORAL PEAK AREA, Ai (ARBITRARY UNITS) AND CALCULATED VALUES OF K (=Ail~t;) FOR THREE DIFFERENT PRESSURE INJECTION TIMES

Wavelength UV detection: 254 nm; capillary length 57 cm; distance between injection and detection 50 cm; constant voltage 25 kV

Component Pressure injection time

5s

K

0.01 M benzoic acid adjusted to pH 8 with TRIS chloride 74.17 0.722 3.13 3.85 1.70 chlorate 62.35 0. 757 3.87 4.83 1.65 fluoride 52.97 0.796 4.67 5.90 1.59 acetate 38.31 0.896 7.34 11.02 1.68 propionate 33.13 0.952 9.09 14.25 1.65 MES 24.24 1.104 15.96 26.95 1.53

Average value Standard deviation

1.63 0.063

0.01 M nicotinic acid adjusted to pH 8 with TRJS chloride 74.17 0.733 3.46 3.51 1;38 chlorate 62.35 0.769 4.38 4.87 1.45 fluoride 52.97 0.808 5.41 6.23 1.42 acetate 38.31 0.909 9.34 12.58 1.48 propionate 33.13 0.966 12.37 16.92 1.42 MES 24.24 1.121 27.21 43.37 1.42


1.43 0,034

0.01 M sulphosalicylic acid adjusted to pH 8 with TRIS chloride 74.17 0.890 3.59 1.32 0.41 chlorate 62.35 0.934 4.68 1.70 0.39 fluoride 52.97 0.982 5.95 2.34 0.40 acetate 38.31 l.105 11.61 5.00 0.39 propionate 33.13 1.174 17.16 7.91 0.39


0.40 0.0089

10.s

Ai

3.17 7.42 3.92 9.85 4.73 11.64 7.38 21.73 9.11 28.26

15.76 55.51

3.39 6.98 4.26 9.45 5.21 11.94 8.62 23.0S

11.07 30.63 23.07 72.24

3.34 3.16 4.24 4.01 5.22 5.09 8.93 9.49

11.93 13.36

15 s

K

3.24 3.18 11.11 4.84 3.32 3.92 14.29 4.81 3.09 4.71 17.78 4.74 3.29 7.26 32.30 4.97 3.26 8.90 41.15 4.86 3.19 14.97 79.90 4.83

3.23 4.84 0.082 0.075

2.81 3.36 10.99 4.46 2.89 4.20 14.18 4.39 2.84 5.12 17.80 4.30 2.94 8.30 33.43 4.43 2.86 10.50 44.10 4.35 2.79 20.34 98.11 4.30

2.85 4.37 0.055 0.066

1.06 3.25 4.51 1.56 1.01 4.09 S.42 1.42 0.99 4.98 7.14 1.46 0.96 8.18 12.45 1.38 0.95 10.64 17.33 1.39

1.00 1.44 0.044 0.073

76 CHAPTER S: CHARACTERISTICS OF CZE

effective mobilities and factors k1• The repeatability of K values is poorer (SD ea. 1-2 %) owing to the inaccuracy of the measured peak areas, possibly owing to the injection method and/or inaccuracy of the peak area determination.

0.001

0.000

3 -0.001

ID 12 ~ -0.002 3 .c g -0.003

6 .c 4 IU

-0.004 EOF 5

-0.005 0 5 10 15 20

time (min)

Fig. 5.5: Electropherogram for the separation of (1) chloride, (2) chlorate, (3) fluoride, (4) acetate, (5) propionate and (6) MES in the indirect UV mode (5·1a4 M, pressure injection time 15 s). Carrier electrolyte, 0.01 M benz;oic acid adjusted at pH 8 by adding TRIS. Wavelength UV detector, 254 nm. Anionic mode, applied voltage 25 kV.

0.005

5 0.000

s 6

<Lt u 5 c -0.005 1 2

'- 4 0 fJl

2 .c -0.010 111

EOF 3

-0.015 0 7 14

t (min)

Fig. 5.6: Electropherogram for the separation of (1) hexanoate, (2) pentanoate, (3) butanoate, (4) propionate, (5) acetate and (6)/ormate in the indirect UV mode (5·1Ci4 M, pressure injection time 10 s). Carrier el.ectrolyte was O.OJ M bem:oic acid at a pH of 8.5 adjusted by adding TRIS. Wavelength UV detector, 254 nm. Cationic mode, applied voltage 25 kV.


In Fig. 5.6 an example of the separation of the mixture of Table 5.IV (in the cationic mode, pressure injection time 10 s) in the indirect UV mode is given. The carrier electrolyte was 0.01 M benzoic acid at pH 8.5 adjusted by adding TRIS. The strongly tailing effect for formate due to the absence of a self-correcting effect of the zones in CZE can be clearly seen.

TABLES.IV

CALCULATED EFFECTIVE MOBIUTIES, -mi·loS (cm2Ns), CALCULATED FACTORS 1q ACCORDlNG TO EQN 5.10, MEASURED MIGRATION TIMES, t; (min), MEASURED TEMPORAL PEAK AREA A1 (ARBITRARY UNITS) AND CALCULATED VALUES OF K (=A/k;t1) FOR THREE DIFFERENT PRESSURE INJECTION TIMES (n=S, STANDARD DEVIATIONS ARE GIVEN IN PARENTHESES).

Wavelength UV detection: 254 run. 10 = S1 cm, l.i = SO cm. Applied voltage, 25 kV

Component m; 1q t; ~ K

Pressure injection time 5 s hexanoate 26.09 (0.115) 1.064 (0.002) 4.69 (0.017) 11.54 (0.181) 2.31 (0.038) pentanoate 27 .65 (0.112) 1.03S (0.002) 4.88 (0.015) 11.72 (0.181) 2.32 {0.03S) butanoate 29.Sl (0.119) 1.003 (0.002) S.12 (0.019) 12.20 (0.089) 2.37 (0.019) propionate 32.58 (0.099) 0.959 (0.001) 5.58 (0.021) 13.34 (0.200) 2.49 (0.037) acetate 37.31 (0.102) 0.905 (0.001) 6.49 (0.029) 15.12 (0.133) 2.S7 (0.025) formate 51.67 (0.112) 0.803 (0.001) 12. 71 (0.127) 24.01 (0.625) 2.35 (0.054)

Average value 2.40 (0.105)

Pressure injection time 10 s hexanoate 26.50 (0.050) 1.056 (0.001) 4.61 (0.032) 23.53 (0.262) 4.83 (0.074) pentanoate 28.05 (0.035) 1.028 (0.001) 4. 79 (0.032) 23.97 (0.292) 4. 87 (0.076) butanoate 29.73 (0.068) 1.000 (0.001) 5.00 (0.039) 24.94 (0.306) 4.99 {0.080) propionate 32.59 (0.064) 0.959 (0.001) 5.41 (0.046) 26.86 (0.374) 5.18 (0.100) acetate 37.21 (0.064) 0.906 (0.001) 6.23 (0.060) 29.96 (0.370) S.31 (0.078) formate 51.38 (0.082) 0.804 (0.000) 11.64 (0.224) 48.20 (0.731) 5.15 (0.162)

Average value 5.06 (0.189)

Pressure injection time 15 s hexanoate 26.76 (0.109) 1.051 (0.002) 4.63 (0.010) 35.61 (1.480) 7.31 (0.302) pentanoate 28.28 (0.088) 1.024 (0.002) 4.81 (0.010) 36.08 (1.004) 7.33 (0.203) butanoate 29.84 (0.092) 0.998 (0.001) 5.01 (0.013) 37.49 {0.985) 7.50 (0.190) propionate 32.56 (0.113) 0.959 (0.001) 5.40 (0.013) 40.02 (0. 760) 7.73 (0.141) acetate 37.17 (0.116) 0.907 (0.001) 6.21 (0.017) 44.90 (1.423) 7.97 (0.250) formate 51.54 (0.097) 0.804 (0.001) 11. 71(0.083) 72.40 (3.469) 7 .69 (0.323)

Average value 7 .59 (0.256)


5.4 CONCLUSIONS

Some effects in the electrophoretic process. in CZE can be described by the Kohlraush regulation function. On applying this equation, for conductivity and indirect UV detection (for non UV-ab5orbing components, applying a non UV-absorbing counter ion and a UV-absorbing co-ion) there is a defined relationship between measured temporal peak areas and effective mobilities, independent of the kind of ionic species. Data measured for several components in several electrolyte systems confirmed the derived relationship.

The relationship between temporal peak area and the product of a correction factor (bi for conductivity. detection and ki for indirect UV detection) and migration time ti is linear, passing through the origin. Applying an internal standard, this relationship can be used for quantitative CZE analysis with calibration graphs being superfluous.

References

1. F. Kohlrausch, Ann. Phys. Chem., 62 (1897) 209. 2. E.B. Dismukes and R.A. Alberty, J. Amer. Chem. Soc., 16 (1954) 191. 3. F.E.P. Mikkers, F.M. Everaerts and Tb.P.E.M. Verheggen, J. Chromatogr., 169 (1979) 11. 4. X. Huang, J.A. Luckey, M.J. Gordon and R.N; Zare, Anal. Chem., 61 (1989) 766. 5. F.E.P. Mikkers, Thesis, University of Technology Eindhoven, 1980. 6. X. Huang, W.F. Coleman and R.N. Zare, J. Chromatogr., 480 (1989) 95. 7. Tb.P.E.M. Verheggen, J.L. Beckers and F.M. Everaerts, J. Chromatogr., 452 (1988) 615. 8. M.T. Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 549 (1991) 345. 9. J.L. Beckers, F.M. Everaerts and M.T. Ackermans, J. Chromatogr., 537 (1991) 407.

CHAPTER6

DETERMINATION OF PHARMACEUTICALS

To investigate the applicability of capillary electrophoresis for the determination of drugs in several matrices, experiments are performed for the separation of several drugs. As the group of drugs consists of many different components, various modes of capillary electrophoresis have to be used to analyse them. Sulfonamides are determined in pork meat extracts with CZE in the cationic mode. Qualitative and quantitative aspects of CZE are compared with HPLC and rrP, using some J3:2-agonists in pharmaceutical dosage forms as model components. Some typical neutral drugs are determined using MECC and, finally, some aminoglycosides are analysed in an electrolyte system for hyphenated CZE with indirect UV detection and MECC.

6.1 INTRODUCTION

In the previous chapters of this thesis, several aspects of capillary electrophoresis are discussed, such as peak recognition using effective or pseudo-effective mobilities and quantitative aspects of capillary electrophoresis with direct as well as indirect UV detection. In this chapter, experiments are described, where the various modes of capillary electrophoresis are used for the determination of various drugs in several matrices. The group of veterinary drugs includes charged as well as uncharged components, with and without UV absorbing properties. In order to obtain an impression of the applicability of capillary electrophoresis for the determination of drugs, some groups of components with different properties were selected to. study applicability of the various modes.

In section 6.2 the determination of sulfonamides in pork meat extracts is discussed. The components of this group absorb UV light and are usually, depending on the pH, negatively charged. For the quantification, CZE in the cationic mode (anode at the inlet, cathode at the outlet) is used, with UV detection at 254 nm. As these components are often seen at injection sites in pork, the sulfonarnides are determined in pork meat extracts.

Section 6.3 deals with the determination of some ~ragonists. These components also have UV absorbing properties, and are, generally, positively charged. The members of this group can be analysed easily by CZE, ITP and HPLC, and they are· very commonly used in pharmaceutical dosage forms. Three members of this group,

80 CHAPTER 6: DETERMINATION OF PHARMACEUTICALS

in several pharmaceutical dosage forms, are used to compare the qualitative and quantitative possibilities of CZE, lTP and HPLC.

Many veterinary drugs absorb UV light and are uncharged at common pH values. Therefore in section 6.4 the applicability of MECC for the determination of various uncharged drugs is shown.

Finally, in section .6.5 the determination of aminoglycosides is described. At most common pH values members of this group are highly positively charged, and as they do not absorb UV light, CZE with indirect UV detection is used. In many pharmaceutical dosage forms, a member of this group is present in combination with an uncharged component and therefore, the possibility of an electrolyte system for hyphenated CZE with indirect UV detection and MECC is demonstrated.

6.2 DETERMJNATION OF SULFONAMIDES IN PORK MEAT EXTRACTS

6.2.1 Introduction

. Many veterinary drugs in use for food~producing animals are available as injectable preparations. Depending on the formulation, the use of these preparations may give rise to residual amounts of the drug at the site of injection [1,2]. Moreover, use of these preparations may lead to considerable local irritation of the tissue, leading to visible injection sites when treated animals are slaughtered. According to the Dutch law, visible injection sites must be removed from the carcass in. For regulatory purposes it is often of interest to know which drug has been used, e.g., when the use of an illegal drug is suspected. In order to identify the drugs present at injection sites of slaughtered animals, a selective and routinely applicable analytical method is required. As often rather high concentrations of drugs will be found at injection sites, owing to incomplete [3] or slow [4] absorption of the drugs, there are rio special requirements regarding the sensitivity of the method. Most recently devel<:>ped methods allow the selective detection of a single component, or a select group of components, in very low concentrations [5,6,7], albeit that multi screenings methods have been reported [8]. Because often an extensive purification and derivatization is required and because of the limited applicability of the methods used so far, the group of sulfonamides was chosen as subject for the study of the applicability of CZE as separation method. This group of drugs consists of a relatively large number chemically related compounds. Studying this group of the drugs will yield information especially concerning the matrix effects of pork meat extracts. For the components studied, the effective mobilities at different pHs are determined, and the pK value and mobility at infinite dilution is calculated. Moreover, the optimum pH for the separation with respect to the number of compounds that can be separated and with respect to matrix interferences is determined. Finally for five components, the calibration graph is set up and the LOD is calculated [9].

6.2 SULFONAMIDES IN PORK MEAT EXTRACTS 81

6.2.Z Experimental

Chemicals and reagents The structural formulae of the compounds studied are given in Fig. 6.1.

Trimethoprim, sulfamethoxazole, sulfanilamide, sulfadimidine-Na, sulfathiazole, sulfamerazine, sulfadiazine, sulfadoxine, sulfamethoxypyridazine, sulfadimethoxine, sulfatroxazole, sulfaquinoxaline, sulfamethoxydiazine, sulfaguanidine, sulfachloropyrazine-Na and sulfachloropyridazine are supplied by the State Institute for Quality Control for Agricultural Products (RIKILT, The Netherlands). From these components, standards stock solutions of 1 mg/ml are prepared by weighing exactly 10.0 mg of the component and dissolving in 10.0 ml solvent. For the analysis these stock solutions are diluted to the desired concentrations by adding distilled water. All other chemicals were of analytical grade.

oo-i.

1: HJ:O~"N-1, HJ:O""'- ""'-.N

N-1,

0

-o-~ Ii/I II 0 .

2: -er· ?:

i) 12:

R= R= Ra-cc

3: 8: N 13: N)-R• - H R=~ R= --<-' oo;,

Ho= OCH, 4.:

i~°' 9: 14: ,. ~ R~ ~R=·-<:. R=

°'> 5:

-0 10:-<:J:OCH, 15.·-oc.t

R= R= v R= V ~

OCH,

6: -{)a;, 11~· 16: R= R= ~ R= -0-CI

-0

Fig. 6.1: Structural fomwlae of (1) trimethoprim, (2) sulfamethoxazole, (3) sulfanilamide, (4) sulfadimidine, (5) sulfathiazole, (6) sulfamerazine, (7) suljadiazine, (8) suljadoxine, (9) sulfamethoxypyridaz/ne, (10) sulfadimethoxine, (11) suljatroxazole, (12) sulfaquinoxaline, (13) sulfamethoxydiazine, (14) sulfaguanidine, (15) suljachloropyrazine and (16) sulfachloropyridazine

Separation conditions All experiments were carried out using the P/ ACE System 2000 HPCE. Working

temperature was 25 °C. Two capillaries were used, viz., an original Beckman Standard Capillary of length 57 cm, distance between injection and detection 50 cm, I.D. 75 µm, where 1 s pressure injection equals about 4.9 nl injection volume, and a


capillary from Polymicro Technologies (Phoenix, AZ, USA) of length 116.35 cm, distance between injection and detection 109.75 cm, I.D. 50 µm, where 1 s pressure injection equals about 0.45 nl injection volume. The sample was introduced by means of pressure injection, and for all experiments a constant voltage was applied. The wavelength of the UV detector was 254 nm for all experiments.

Sample pretreatment For the pork meat extracts, 10 g pork meat was homogenized in a household

food processor and extracted in a stomacher apparatus for 5 min with 100 ml acetonitrile. The sample was centrifuged for 10 min at 4000 g and filtrated through a 0.45 µm filter. If the sample had to be spiked, this was done between the centrifugation and the filtration steps. The filtrate was used for the injection.


Determination. of effective mobilities For the determination of the effective mobilities the background electrolytes and

separation conditions as given in Table 6.1 were used. The effective mobilities were determined with both the Polymicro Technologies and the Beckman Standard Capillary. Table 6.II gives the effective mobilities ·loS (cm2/Vs) for the sixteen components for both capillaries at four pH values. The voltage applied in the experiments with capillary II is lower in order to avoid a large difference between the electric currents. At pH 3.2 effective mobilities in the long Polymicro Technologies Capillary could not be determined because the velocity of the EOF was so low that the analysis time became about 80 min, causing very broad peaks through which precise migration times could not be obtained. From Table 6.II it can be concluded that the effective mobilities determined with both capillaries agree. The migration times of components, not completely separated from the EOF marker, are often difficult to determine resulting in imprecise effective mobilities. In Fig. 6.2 the effective mobilities as a function of the pH value of the background electrolyte for the Beckman Standard Capillary are given.

As can be seen in Fig. 6.2 the differences between the effective mobilities are optimum at pH 7. At this pH most of the sulfonamides can be separated from each other. Nevertheless two pairs of sulfonamides can not be separated, viz., sulfanilamide and sulfaguanidine, which are both uncharged at this pH (m=O) and can be used as EOF marker, and sulfathiazole and sulfamethox:fyridazine which have about the same m (average-6.42· 10·5 cm2/Vs and -6.71·10·5 cm /Vs resp.). Sulfadiazine, sulfadoxine and sulfadimethoxine are not completely separated, however, three tops of peaks can be determined if all three are present. As expected the components are better separated in the Polymicro Technology Capillary which has a longer separation length.

In Fig. 6.3 the electropherograms are given using the Beckman Standard Capillary for (A) and (B) a mixture of thirteen sulfonamides dissolved in 90% water and 10% acetonitrile using a background electrolyte of 0.02 M imidazole/acetate and 0.02 M phosphate/0.02 M borate respectively and (C) and (D) a mixture of thirteen sulfonamides dissolved in the pork meat matrix using a background electrolyte of 0.02 M imidazole/acetate and 0.02 M phosphate/0.02 M borate respectively. As can be seen in Fig. 6.3 (A) and (B) the separation of the standard mixture is good in both


TABLE6.I

BACKGROUND ELECTROLYTES AND SEPARATION CONDITIONS FOR THE DETERMINATION OF THE EFFECTIVE MOBILITIES OF THE SULFONAMIDES.

I: Polymicro Technologies Capillary l0 = 116.45 cm, ld = 109.75 cm, I.D. = SO µ,m. II: Beckman Standard Capillary le = 51 cm, Id = 50 cm, I.D. = 75 µ,m. The background electrolytes were prepared by adding the acid to the cationic component until the desired pH was reached, except for the phosphateborate buffer, where KOH is added to the mixture of the acids until the desired pH was reached. The observed current (p.A) for the applied voltages is. given in parentheses.

background electrolyte pH Applied voltage (kV) Injection time (s)

I II I II

0.02 M TRIS/acetate 8.2 30 (3.7) 15 (8.S) s 2 0.01 M imidazole/acetate 7.0 30 (4.9) 10 (8.4) s 2 0.02 M phosphate/0.02 M borate 7.0 30 (16.7) 10 (26.S) s 2 0.01 M TRIS/MES 6.S 30 (2.9) 10 (4.6) s 2 0.01 M TRIS/acetate 5.0 30 (3.7) 10 (S.7) s 2 0.01 M TRIS/formate 3.2 30 (S.2) 10 (8.1) s 2

(i) > ........ ('I

E ~ Ill

0 0

T""

* E

-25 ,__ _ __,__ _ __,__ _ __, __ ...._ _ __,__ __

3 4 5 6 7 8 g

pH

Fi.g. 6.2: Effective mobility as a junction of the pH of (solid lines): (+) 1, (A) 2, (O) 3, (+) 4, (i..) S, (•) 6, (v) 7 and (0) 8, and (doued lines):(+) 9, (A) 10, (O) 11, (+) 12, (i..) 13, (•) 14, (v) JS and (0) 16. For the names of the components see the legends of Fig. 6.J.


TABLE 6.II

EFFECTIVE MOBILITIES ·loS (cm2Ns) FOR THE SIXTEEN SULFONAMIDES AT DIFFERENT pH VALUES OF THE BACKGROUND ELECTROLYTES.

I: Polymicro Technology Capillary le = 116.45 cm, l4 = 109.75 cm, I.D. = 50 µm. II; Beekman Standard Capillary le = 57 cm, ~ = SO cm, I.D. = 75 p.m.

Component Cap. effective mobilities

pH 8.2 pH7.o• pH5.0 pH3.2

trimethoprim I 1.SO 11.25 18.76 II 1.87 11.73 19.08 20.26

sulfamethoxazole I -21.55 -20.48 -4.02 II -21.96 -21.15 -4.30 0.00

sulfanilamide I 0.00 0.00 0.00 II 0.00 0.00 0.00 1.30

sulfadimidine I -13.91 -2.92 0.00 II -14.26 -2.91 0.00 1.65

sulfathiazole I -19.41 -6.47 0.00 II -19.67 -6.36 0.00 1.21

sulfamerazine I -18.83 -9.51 -1.63 II ·18.98 -9.49 0.00 1.41

sulfadiazine I -21.01 -16.03 -1.41 II -21.29 -16.39 -0.96 0.89

sulfadoxine I -18.54 -16.81 -2.43 II -18.65 -17.24 -2.59 0.31

sulfamethoxypyridazine I -17.77 -6.61 -0.23 II -17.83 -6.81 0.00 1.05

sulfadimethoxme I -19.16 -17.15 -2.19 II -19.33 -17.37 -2.31 0.47

sulfatroxazole I -20.31 -19.22 -3.61 II -20.49 -19.65 -3.69 0.00

sulfaquinoxaline I -19.49 -18.20 -2.38 II -19.60 -18.59 -3.84 0.45

sulfamethoxydiazine I -19.38 -12.49 -0.58 II -19.35 -12.83 -0.76 0.49

sulfaguanidine I 0.00 0.00 0.00 II 0.00 0.00 0.00 2.21

sulfacbloropyrazine I -21.18 -21.21 -10.46 II -21.36 -21.73 -9.50 0.32

sulfacbloropyridazine I -20.85 -20.26 -5.65 II -20.89 -20.64 -4.96 0.41

•for pH 7.0 the imidazole/acetate system is used


A B 10

"! 710 5 ·5 2 ·5 2

i i 13 16

I .. ~ i

15

0 e 16 24 32 0 6 12 1e 24

timla (mini time lmi<V

c D 10

"! ~ 2 5 2

i 7 10 ~ 13 16

I I 15

0 8 16 24 32 0 6 12 ,. 24

time (mini time !mini

Ff.g. 6.3: Electropherograms of the separation of twelve sulfonamides (c=0.01 mg/ml) in the Beckman Standard Capillary (length between injection and detection 50 cm, pressure injection time 2 s = about JO nl, applied voltage 10 kV) of (A) and (BJ standard mixture in the background electrolyte 0.02 M imidazole-acetate at pH 7 and 0.02 M phosphate/0.02 M borate at pH 7 respectively and (CJ and (D) the standard mixture dissolved in the pork meat matrix using as background electrolyte 0. 02 M imidazole-aceJate at pH 7 and 0.02 M phosphate/0.02 M borate at pH 7 respectively. For the names of the components see the legends of Fig. 6.1.

background electrolytes. Nevertheless, the effective mobilities for some components are different in both background electrolytes (possibly owing to a slight difference in the pH and the nature of the two electrolyte solutions) but, as further experiments showed, the effective mobilities are reproducible and constant in each background electrolyte. As can be seen in Fig. 6.3 (C) and (D) the separation of the mixture in the matrix is much better in the phosphate/borate buffer than in the imidazole buffer, showing sharper peaks. For the quantitative analysis therefore the 0.02 M phosphate/0.02 M borate buffer is used.

Calculation of pK values and mobilities at infinite dilution If the effective mobility of a component is known for two different electrolyte

systems at different pH values, at which the component shows a different degree of dissociation, both its pK and its mobility at infinite dilution can be calculated [10]. In the calculations the best values are obtained if the pK value of the component lies between the pH values of the background electrolytes and if the effective mobilities


are not too small. For components with pK values of about 7-8 background electrolytes at pH 8.2 and 7 are chosen. For components with pK values of about 5-7 both background electrolytes with pH 7 and 5 and 8.2 and 5 are chosen.

The pK values of sulfadiazine and sulfamethoxydiazine are calculated from the background electrolytes at pH of 8.2 and 6.5. Table 6.m gives the pK values and mobilities at infinite dilution for fourteen sulfonamides calculated from effective mobilities in the different background electrolytes, and the pK value as found in literature [11, 12]. Sulfanilamide and sulfaguanidine are not mentioned in Table 6.m because they have very small effective mobilities in all electrolyte systems. From Table 6.m it can be concluded that the calculated pK values and mobilities at infinite dilution obtained from the effective mobilities of both capillaries with the same background electrolyte systems agree.

Matrix effects Because for a routine method the pretreatment should be very simple, the effects

for a pork meat extract were studied, where the sample pretreatment consisted only of extraction of the homogenized meat with acetonitrile and centrifugation. This acetonitrile sample was directly used for injection. In Fig. 6.4 the electropherograms of (a) the mixture of thirteen sulfonamides dissolved in water and (b) the pure matrix in the background electrolytes (A} 0.02 M TRIS/acetate at pH 8.2, (B) 0.02 M phosphate/0.02 M borate at pH 7.0, (C) 0.01 M TRIS/acetate at pH 5.0 and (D) 0.01 M TRIS/formate at pH 3.2 are given. It can be clearly seen that not only for the separation, but also for the matrix effects the background electrolyte at pH 7.0 gives the best results. The velocity of the EOF can vary from experiment to experiment and from this it seems that the components in Fig. 6.4 (D) migrate in front of the matrix components. Further experiments showed, however, that this is only due to the change in the velocity of the EOF. The determination of sulfaguanidine and sulfanilamide in the matrix is problematic because these components are uncharged at pH 7.0 and migrate at the position of the EOF in the acetonitrile plug.

Table 6.IV gives the effective mobilities m· 105 (cm2/Vs) for twelve sulfonamides determined separately, in the mixture and in the matrix spiked with the mixture, applying a background electrolyte at pH 7. For each component the effective mobilities determined in the mixture and in the matrix spiked with the mixture are comparable, and by this, the effective mobility can be used for the peak recognition.

Quantification To evaluate the quantitative abilities of the method, the calibration graphs for·

five sulfonamides in acetonitrile and in the matrix were constructed using the 0.02 M phosphate/0.02 M borate buffer pH 7.0 and the Polymicro Technologies Capillary. For the determination of the calibration graphs 10.0 mg sulfadimidine-Na, sulfamerazine, sulfadoxine, sulfatroxazole and sulfamethoxazole were weighed accurately and dissolved in 10.0 ml acetonitrile or matrix. From these solutions dilutions were made at the concentrations of 0.1, 0.01 and 0.001 mg/ml. Each of the solutions was used for injection with pressure injection times of 3, 5, 10, 15, 20, 25 and 30 seconds (1 s pressure injection equals ea. 0.45 nl injection volume; sometimes the dimension µg.s/ml is used which means the product of the pressure injection time and the concentration; 1 µg.s/ml = 0.45 pg). Peak area were determined using the


TABLE 6.III

CALCULATED pK VALUES AND MOBILITIES AT INFINITE DILUTION m0·to5 (cm2Ns) FOR THE SULFONAMIDES USING THE EFFECTIVE MOBILITIES OF THE COMPONENT IN BACKGROUND ELECTROLYTES WITH THE INDICATED pH VALUES, AND LITERATURE VALUES FOR pK VALUES.

I: Polymicro Technology Capillary le = 116.4S cm, ld = 109. 7S cm, I.D. = SO µm. II : Beckman Standard Capillary 10 = 57 cm, Id = SO cm, J.D. == 7S µm.

Component Cap. pK pH 8.2-7.o• pH 8.2-5.0 pH 7.0*-5.0 (lit)

pK mo pK mo pK mo

trimethoprim I 6.6 7.11 23.S9 7.13 22.53 II 7.26 21.75 7.23 22.83

sulfamethoxazole I 5.73 -25.40 5.72 -25.05 II 5.70 -25.82 5.70 -25.73

sulfadimidine I 7.4 7.82 -22.36 II 7.83 -23.09

sulfathiazole I 7.2 7.48 -26.25 II 7.50 -26.73

sulfamerazine I 7.0 7.13 -23.90 II 7.lS -24.12

sulfadiazine ...

I 6.4 6.55 -25.18 II 6.52 -25.4S

sulfadoxine I 6.1 5.91 -22.34 S.89 -21.49 II 5.88 -22.4S S.87 ·21.90

sulfamethoxy· I 6.7 7.37 -23.81 pyridazine II 7.39 -23.77

sulfadimethoxine I 6.2 5.98 ·22.99 S.9S ·22.04 II S.96 -23.16 S.93 ·22.22

sulfatroxazoJe I 5.8 5.75 -24.13 S.74 ·23.76 II 5.75 ·24.32 S.74 -24.12

sulfaquinoxaline I 5.5 5.94 -23.32 S.94 -23.15 II 5.70 -23.40 5.68 -22.98

sulfamethoxydiazine*• I 6.75 -23.70 II 6.73 -23.76

sulfachloropyrazine I 5.1 5.10 -24.97 5.10 -25.12 II S.18 -25.16 5.20 -25.72

sulfacbloropyridazine I 5.52 -24.66 5.51 -24.47 II 5.59 -24.71 S.60 -25.00

• For pH 7 .0 the imidazole/acetate system is used •• Calculated with background electrolyte systems at pH 8.2 and 6.S


A B

~ ! ·; j j a

I a I b

5 10 111 6

time lminl

C D

! i ~

j j

I II I

a l ! a

b ... b

IS 10 115 0

time <mW

I

10

11S

time fmin)

l

20

time (mjn)

24

I

Fig. 6.4: Electropherograms of (a) the mixture of thirteen sulfonamides (c=0.01 mg/ml) and (b) the pure matrix in the background electrolytes (A) 0.02 M TRJS-acetate pH 8.2, (B) 0.02 M phosphate!0.02 M barate pH 7.0, (C) 0.01 M TRJS-acetate pH 5.0 and (D) 0.01 M TRJS-formate pH 3.2. Beckman Standard Capillary 57150 cm, 75 pm, applied voltage JO kV, pressure injection time 2 s = about 10 nl.

laboratory-written data analysis program CAESAR. From the calibration graphs, the limit of detection for the method was evaluated. Table 6. V gives the regression coefficient and LOD (µg.s/ml) for the calibration graphs of the five sulfonamides. The regression coefficients are calculated with the data of the three concentration decades and for the determination of the LOD the calibration graph the sample concentration of 0.001 mg/ml was used. As can be concluded from Table 6.V the regression coefficients are fairly good (> 0.999) and the limit of detection for the method lies between 2 and 9 µg.s/ml. For the given sample pretreatment and an injection of time


of 10 seconds the value of the LOD in µg.s/ml in the extract equals the value of the LOD in ppm in the meat sample assuming 100% recovery.

TABLE6.IV

EFFECTIVE MOBILITIES, m·toS (c.fill/Vs), FOR TWELVE SULFONAMIDES DETERMINED SEPARATELY, IN TIIB MIXTURE AND THE MATRIX SPIKED WITH THE MIXTURE.

Background electrolyte 0.02 M imidazole/acet.ate at pH 7.0, Beckman Standard Capillary 57/50 cm, 75 µ,m, applied voltage 10 kV.

Component Separately Mixture Matrix with mixture

trimethoprim 11.49 11.70 11.19 sulfadimidine -2.92 -3.01 -3.40 sulfathiazole -6.42 -6.35 -6.52 sulfamerazine -9.53 -9.38 -9.44 sulfamethoxydia.zine -12.66 -12.45 -12.53 sulfadia.zine -16.21 -15.88 -15.89 sulfadimethoxine -17.26 -16.92 -17.02 sulfaquinoxaline -18.40 -18.12 -18.28 sulfatroxazole -19.43 -19.08 -19.22 sulfachloropyridazine -20.45 -20.10 -20.35 sulfamethoxazole -20.82 -20.54 -20.70 sulfachloropyrazine -21.47 -21.40 -21.28

TABLE 6.V

REGRESSION COEFFICIENT AND LIMIT OF DETECTION FOR TIIB CALIBRATION GRAPHS OF SULFADIMIDINE, SULFAMERAZINE, SULFADOXINE, SULFATROXAZOLE AND SULFAMETHOXAZOLE DISSOLVED IN (A) ACETONITRILE AND (B) THE PORK MEAT MATRIX.

For the determination of the LOD the calibration graph in the lowest concentration decade is used. The regression coefficient is calculated with the data over the three concentration decades.

Regression coefficient LOD (µ,g.s/ml)

A Sulfadimidine 0.9996 4.8 Sulfamerazine 0.9995 1.9 Sulfadoxine 0.9994 8.8 Sulfatroxazole 0.9993 6.3 Sulfamethoxazole 0.9994 3.7

B Sulfadimidine 0.9997 5.5 Sulfamerazine 0.9994 4.4 Sulfadoxine 0.9991 3.9 Sulfatroxazole 0.9991 6.8 Sulfamethoxazole 0.9991 2.8


6.2.4 Conclusions

Effective mobilities can be used for the identification of components, and they are measured for sixteen sulfonamides in two different capillaries as a function of pH. From these experiments it can be concluded that a pH 7 of the background electrolyte is the optimum pH for separation. At this pH fourteen sulfonamides can be determined and twelve of them can be identified using the effective mobilities. For some drugs, however, a further confirmation such as UV absorbance ratios or diode array detection should be used. From the effective mobilities the pK value and the mobility at infinite dilution for the sulfonamides have been calculated.

For the determination of sulfonamides in pork meat a very simple pretreatment consisting of an extraction with acetonitrile and centrifugation can be applied. For five sulfonamides the calibration graphs are set up both in acetonitrile and in the matrix. Linear calibration graphs were obtained with regression coefficients of at least 0.999 and limits of detection in the range 2-9 ppm in the sample for a pressure injection time of 10 seconds (about 4.5 nl) using a Polymicro Technology Capillary of length 116.35 cm, distance between injection and detection 109.75 cm and I.D. 50 µm.

6.3 DETERMINATION OF /3i-AGONISTS IN PHARMACEUTICALS

6.3.1 Introduction

Salbutamol, terbutaline and fenoterol are 132-adrenoceptor agonists, widely used for the treatment of bronchial asthma. They are commercially supplied as amongst others, tablets, syrups and injectables. In the past, several methods have been reported on the determination of such components in pharmaceuticals and plasma, including gas chromatography-mass spectrometry [13,14], ion-pair liquid chromatography [15] and reversed-phase high performance liquid chromatography [16,17]. As capillary zone electrophoresis is a rather new separation method, it is of interest to compare the qualitative and quantitative abilities of this technique for the determination of those J3i-agonists in pharmaceuticals with those of the conventional methods [18].

In this section, CZE is compared with isotachophoresis and reversed-phase high performance liquid chromatography for the determination of salbutamol, terbutaline sulphate and fenoterol hydrobromide in pharmaceutical dosage forms. In Fig. 6.5 the structural formulae of the compounds are given.

6.3.2 Experimental

Instrumentation The HPLC equipment (Pharmacia-LKB, Bromma, Sweden) consisted of a Model

2150 pump, a Model 2152 controller, a low-pressure mixer, a Model 2156 solvent conditioner and a VWM 2141 dual-wavelength UV detector. Chromatographic separation was obtained with a LiChrospher 100 RP-18 endcapped column (125 x4 mm, 5 µm) from E. Merck (Darmstadt, Germany). Injections were made with a Model 7125 universal loop injector (20 µl) (Rheodyne, Berkeley, CA, USA).

For all CZE experiments the PI ACE System 2000 HPCE was used. The capillary was an original Beckman capillary cartridge (capillary length 57 cm, distance between

6.3 PrAGONISTS IN PHARMACEUTICALS 91

HO

~· HO

Sa!butarnol Terbutaline

Fenoterol

Fig. 6.S: Structural formulae of salburamol, terbutaline andfenoterol

injection and detection 50 cm and 75 µm I.D.). Wavelength of the UV detector was 214 nm in all experiments and the operating temperature was 25 °C. All experiments were carried out in the cationic mode (anode placed. at the inlet and cathode at the outlet) applying a constant voltage of 12.5 kV. The sample introduction was achieved by pressure injection for 5 seconds (about 25 nl).

For all ITP experiments a laboratory-built apparatus [19], with a conductivity and an UV detector (254 nm) was used. In this apparatus a closed system is obtained by shielding the separation capillary from the open electrode compartments with semipermeable membranes. A PTFE capillary tube (0.2 mm I.D.) was used, in contrast to the fused silica capillary in the Beckman apparatus. The sample was introduced with a syringe and the sample volume was 3 µI, unless stated otherwise.

All data obtained from the chromatograms and electropherograms were handled using the laboratory•written program CAESAR.

Chemicals Salbutamol sulphate, terbutaline sulphate and fenoterol hydrobromide were

donated by the State Institute of Quality Control for Agricultural Products (RIKILT, Wageningen, The Netherlands). All salbutamol pharmaceuticals are Ventolin products from Glaxo B. V. (Nieuwegein, The Netherlands), the terbutaline pharmaceuticals are Bricanyl products from Astra Pharmaceutica B. V. (Rijswijk, The Netherlands) and the fenoterol pharmaceuticals are Berotec products from Boehringer Ingelheim (Alkmaar, The Netherlands).

Standard solutions Standard solutions of 1 mg/ml of salbutamol, terbutaline sulphate and fenoterol

hydrobromide were prepared by weighing accurately 50.0 mg of the standards and dissolving them in 50.0 ml of distilled water. From these solutions appropriate


dilutions were made so that the concentration of each sample solution approached the concentration of that in the middle of the standard solution range.

Sample preparation All tablets and the capsules were mixed with 10 ml water and after

ultrasonication for about 30 min, the sample solution was centrifuged. The clear supernatant solution was used for the analysis after dilution with distilled water to the desired concentration. All liquid pharmaceuticals were diluted to the desired concentration with distilled water.

Separation conditions for HPLC Reversed-phase HPLC was performed at ambient temperature. Several

experiments were carried out to select a proper mobile phase and a mixture of watermethanol (60:40, vfv) containing 0.002 M KOH and 0.01 M hexanoic acid as an ion-pair reagent was found suitable for the analysis of the pharmaceuticals. In Fig. 6.6 the capacity factor k is given as a function of the percentage water in the watermethanol mixture for salbutamol sulphate, fenoterol hydrobromide, clenbuterol hydrochloride and terbutaline sulphate. The mobile phase was degassed by vacuum filtration through a 0.22 µm filter and sparging with helium. The column was equilibrated with mobile phase at a flow-rate of 0.4 ml/min for about an hour.

20

16

12

.:t:.

8

4

0 40 50 eo 70 80 90 100

% water

Fig. 6. 6: Relationship between capacity factor k and percentage of water in the water:methanol mobile phase containing 0. 002 M KOH and 0.01 M hexanoic acid far (.1.) terbutaline sulphate, (0) salbutamol sulphate, ( +) fenoterol lrydrobromide and ( ._) the analogue antibiotic clenbuterol hydrochloride.

Separation conditions for ITP For the ITP experiments two electrolyte systems were used. System A consisted

of a leading electrolyte of 0.01 M histidine adjusted at pH 4.75 by adding acetic acid with the terminator acetic acid at pH 3.5. System B consisted of a leading electrolyte

6.3 ,82-AGONISTS IN PHARMACEUTICALS 93

of 0.01 M KOH adjuste<l at pH 4. 75 by adding acetic acid with the terminator acetic acid at pH 3.5.

With electrolyte system A, ionic species present in the sample solutions with high effective mobilities {such as sodium) will migrate in a zone electrophoretic manner through the leading zone of histidine. The drugs migrate in an ITP way between the leading ions histidine and the terminating hydrogen ions. Applying electrolyte system B, the drugs migrate behind a large zone of the sample ions with high effective mobility. Nevertheless, identical results were obtained for test samples both with system A and B. As an example, the isotachopherograms (both the UV and conductivity signal) ofVentolin syrup are given in Fig. 6.7 applying both {a) system A and (b) system B. It can be clearly seen that in system B, salbutamol migrates behind a large amount of a quick sample cation, whereas in system A that sample cation migrates in the leading zone of histidine. For the determination of the drugs in the pharmaceuticals system B was applied.

UV

Qi a ..... ·5 Cond L x

..ci UV ... ~ T

ii $

6i b '(ii

x Cond L

0 15 30 45

time (min)

Fig. 6.1: Jsotachopherograms for the analysis of VemoUn syrup by JTP applying (a) el.ectrolyte system A and (b) electrolyte system B. The z.one of Salbutamol is indiCflled with S. The unknown sample component X migrates isotachophoretically in system B and zane electrophoretically in the leading zone histidine in system A. L - leading electrolyte, T = terminating electrolyte.

Separation conditions for CZE All CZE experiments were carried out with the background electrolyte 0.01 M

TRIS adjusted to pH 5.0 by adding acetic acid.


F/fect of sample composition The composition of the sample can strongly influence the quality of the

separation. Especially sample components with high effective mobilities present at a high concentration in a sample will affect the migration behaviour in electrophoresis.


In CZE this can create an ITP system with two leading ions [20,21] leading to very sharp peaks and high plate numbers (this effect must be distinguished from stacking effects due to the injection of very dilute samples).

A typical difference between electrophoretic and chromatographic techniques is that in electrophoresis at any point the situation is determined by the initial conditions, as Kohlrausch formulated by his "regulation function" in 1897. This means that in electrophoresis the concentration of the injected sample adapts to the initial concentration of the background electrolyte migrating in the separation capillary. If one of the sample components is present at a high concentration the length of the injected sample zone elongates during this adaptation process and the separation capacity of the system can be insufficient to separate all sample components. A way to solve this problem is to inject smaller amounts of the sample. However, the amount of the sample component of interest must be sufficient in order to detect and quantify that component. This is often at a disadvantage in ITP because the sample components migrate in consecutive zones, after the separation process, at a concentration adapted to that of the leading ions, which generally means at a concentration of about 0.005 to 0.01 M. Very small amounts of a sample component lead to very short indetectable zones.

In ITP, the response factor RF [22], defined as the slope of the calibration graph of the product ZIA (As) versus the amount of the sample Q (mole), can be utilized for quantitative determinations:

ZL·l RF=

Q

(6.1)

where ZL is the zone length (s) and l the electric current (A). This RF value is a constant (about 2· IOS C/mole for monovalent ions) and indicates that the minimum detectable amount can be decreased by applying lower values of the electric current, assuming that a minimum zone length is required. Of course, this results in longer analysis times. This principle is applied in the determination of salbutamol, fenoterol hydrobromide and terbutaline sulphate with ITP (see Tables 6.VI-VIlI). In first instance deviating values were obtained. After diluting the sample solutions ten-fold, injecting 1 µl sample solution and applying 7 µA in stead of 25 µA good results were obtained for previously too low values. Another way to solve this problem is to use a column-coupling system with a higher separation capacity.

Detennlnation of drugs in phannaceuticals In the Tables 6. VI-YIU all results for the quantification of salbutamol, terbutaline

sulphate and · fenoterol hydrobromide are given. For the calibration graphs, the concentration decade, CD, indicated with its highest concentration, the regression coefficient, r, and calculated limit of detection, LOD (µg/ml), the determined quantity of the drugs, Q, and the calculated relative standard deviation, RSD, are given. For all pharmaceuticals the number of measurements, m, is 3.

For HPLC and CZE standard solutions in the decade 0.01-0. l mg/ml were used, the calibration graph determined twice at a wavelength of 214 nm and each calibration graph was applied for the quantification ofall sample solutions.

6.3 Pi-AGONISTS IN PHARMACEUTICALS 95

For ITP in first instance standard solutions with concentrations in the concentration decade of 0.1-1 mg/ml were used and an electric current of 25 µA was applied. With the calibration graph, both using the conductivity and UV detector signals (injection volume 3 µl) the amount of the drugs in all sample solutions was determined and too low values for the liquid samples containing a very large amount of a sample component with a high effective mobility (probably sodium) were obtained. After diluting these samples ten-fold and working in the concentration decade 0.01-0.1 mg/ml (electric current 7 µA) the results covered the labelled values. In some instances (marked with asterisks in the tables) even 1 µl had to be injected in order to obtain a complete separation and in these cases the RSD was calculated with the three lowest points (n=3) of the concentration decade in order to get the measured zone length in the centroid of the regression line.

For the peak area in CZE the temporal and not the spatial peak area was determined because the migration time in all experiments was fairly constant.

Detenninati.on of salbutamol The determination of salbutamol is studied in Ventolin tablets (labelled value

4 mg/tablet), Ventolin solution for intravenous infusion (labelled value 1 mg/ml) and Ventolin syrup (labelled value 0.4 mg/ml). In Table 6.VI all results are given.

A B c

11 1 a llOF

; -; 11 l j b 'S b

l

I i c l d

°""" 0 0 10 0 3G

time tninl t1 .... *"'"' -~

Jig. 6.8: (A) HPLC chromatograms, (B) zone electropherograms and (C) isotachopherograms of (a) a standard solution of salbutamol, (b) the sample solution of Vento/in tablet, (c) the sample solution of the Ventolin solution for intravenous infusion and (d) the sample solution of Ventolin syrup. Salbutamol is indicated with Sand all unknown components are indicated with X. In the isotachopherograms (c) and (d) the electric current is decreased to 7 µA at the time of detection of the zone X.

In Fig. 6.8 examples are given of the (A) HPLC (B) CZE and (C) lTP experiments on (a) a standard solution of salbutamol, (b) a sample solution prepared from the Ventolin tablet, (c) the Ventolin solution for intravenous infusion and (d) the Ventolin syrup. For the HPLC and CZE experiments the UV signal and for the ITP


experiments the conductivity and UV signals are given. The salbutamol zones are marked with S. All other unknown sample components are indicated by X.

Comparing the chromatograms and electropherograms of Fig. 6.8A, Band C some remarks have to be made. In all cases the salbutamol could be easily separated from other sample components without any pretreatment and the values obtained cover the labelled values, although for ITP a low electric current density had to be applied to obtain a complete separation from the matrix with sufficiently large zone lengths of salbutamol. For the Ventolin tablet, all techniques show only the salbutamol component. For the Ventolin solution for intravenous infusion there is present at least one extra non-UV absorbing sample component with a high effective mobility (probably sodium, see Fig. 6.8C), that could be present at about tr,O in the HPLC chromatogram and is invisible in CZE. For the Ventolin syrup CZE shows only the salbutamol peak, ITP shows one non-UV absorbing component and HPLC two extra UV absorbing components. The time of analysis for HPLC and CZE is about 6 and 9 minutes (until EOF marker); respectively, and increases to about 30 minutes for the ITP analyses with samples containing a large amount of component X with high effective mobility .. An advantage of the electrophoretic methods is that the analysis can be stopped after detection of the desired sample component, after which a new experiment can be started. A disadvantage of the HPLC method is the rather long equilibration time of the system and the fact that all components, including those not of interest, must pass the detector before a new run can be started.

TABLE6.VI

AMOUNTS OF SALBUT AMOL, Q, IN VENTOLIN TABLETS (mg/tablet), VENTOLIN SOLUTION FOR INTRA VENOUS INFUSION (mg/ml) AND VENTOLIN SYRUP (mgfml) AND CALCULATED RELATIVE STANDARD DEVIATION(%) DETERMINED WITH HPLC, CZE ANDITP. All sample concentrations are determined three times (m=3). For further information see text.

Method Calibration graph Tablet Infusion Syrup (4 mg/tablet) (1 mg/ml) (0.4 mg/ml)

CD r LOD Q RSD Q RSD Q RSD

HPLC 214nm 0.1 0.99985 1.76 4.05 1.02 1.04 0.81 0.40 1.03 214nm 0.1 0.99977 1.80 4.00 1.04 1.02 0.83 0.40 1.03

CZE 214nm 0.1 0.99945 2.23 3.94 1.35 0.99 1.16 0.41 1.31 214nm 0.1 0.99947 2.73 3.72 1.70 1.00 1.31 0.39 1.61

lTP UV 1 0.99933 37.0 3.96 2.20 0.90 3.06 0.34 2.62 COND 1 0.99939 35.1 3.95 2.10 0.89 2.94 0.35 2.47 COND 0.1 0.99931 3.12 4.07 1.77 1.00 1.91* 0.40 2.s2•

6.3 ,S2-AGONISTS IN PHARMACEUTICALS 97

TABLE 6.VII

AMOUNTSOFTERBUTALINESULPHATE,Q,INBRICANYLTABLETS(mg/tablet),BRICANYL AMPOULES FOR INJECTION (mg/ml) AND BRICANYL SYRUP (mg/ml) AND CALCULATED RELATIVE STANDARD DEVIATIONS(%) DETERMINED WITH HPLC, CZB AND ITP. All sample concentrations are determined three times (mm3). For further information see text.

Method Calibration graph Tablet Ampoules Syrup (5 mg/tablet) (0.5 mg/ml) (0.3 mg/ml)

CD r LOD Q RSD Q RSD Q RSD

HPLC 214nm 0.1 0.99995 1.03 4.92 0.48 0.51 0.46 0.31 0.83 214nm 0.1 0.99974 2.32 4.89 1.09 0.51 I.OS 0.30 1.91

CZE 214nm 0.1 0.99895 4.62 4.68 2.28 0.52 2.05 0.30 3.89 214nm 0.1 0.99952 3.11 4.67 1.54 0.50 1.43 0.30 2.62

ITP UV 1 0.99985 17.31 4.84 0.82 0.41 2.26 0.27 1.63 COND 1 0.99990 14.01 4.87 0.66 0.42 1.78 0.27 1.32 COND 0.1 0.99952 3.12 4.95 1.45 0.50 2.80'" 0.31 2.55

TABLE 6.VIII

AMOUNTS OFFENOTEROLHYDROBROMIDE, Q, IN BEROTECTABLETS (mg fenoterol/tablet), BEROTEC ROTACAPS (mg fenoterol hydrobromide/capsule) AND BEROTEC RESPIRATOR SOLUTION (mg fenoterol hydrobromide/ml) AND CALCULATED RELATIVE STANDARD DEVIATIONS(%) DETERMINED WITH HPLC, CZE AND ITP. All sample concentrations are determined three times (mm3). For further information see text.

Method Calibration graph Tablet Rotacaps Respirator (2.5 mg/tablet) (0.2 mg/capsule) (S mg/ml)

CD r LOO Q RSD Q RSD Q RSD

HPLC 214nm 0.1 0.99983 1.54 1.84 0.76 0.123 2.27 4.98 0.71 214nm 0.1 0.99982 1.92 1.86 0.94 0.125 2.78 4.99 0.89

CZE 214nm 0.1 0.99986 l.70 1.88 0.82 0.121 2.57 4.89 0.80 214nm 0.1 0.99935 3.64 1.97 1.68 0.117 5.73 4.91 1.71

ITP UV 1 0.99985 17.53 1.91 0.84 0.133 1.68 4.14 0.99 COND l 0.99994 10.94 1.89 0.53 0.136 1.02 4.24 0.60 COND 0.1 0.99995 1.05 1.83 0.52 0.154 1.19 4.96 0.49

98 CHAPTER 6: DEIERMINA TION OF PHARMACEUTICALS

a EOF

2! x s x

.e u 3 Q) s u

b c EOF ro .0

i c s OF

x

0 10 20 30

time (min)

Fig. 6.9: Zone electropherograms of VentoUn syrup applying as background electrolyte (a) TRJS!acetate at pH 5.0, (b) TRIS/acetate at pH 5.0 after rinsing Jhe capillary tube with 0.1 M KOH to obtain a higher velocity of the EOF and (c) HIST/acetate at pH 5.0. for further explanation see text.

In order to obtain more information about the sample composition of Ventolin syrup an experiment was performed with the backgroµnd TRIS/acetate at pH 5.0, see Fig. 6.9(a), where only one sample component salbutamol could be observed. After carefully rinsing the capillary tube with 0.1 · M KOH, in order to obtain a higher velocity of the EOF this separation was repeated, see Fig. 6.9(b). With the high EOF also negative ions can be detected and two UV absorbing negative ions are present in the electropherogram. This separation was again repeated with the background electrolyte 0.01 M histidine adjusted at pH 5.0 by adding acetic acid. Non-UV absorbing positive ions can now be made visible by indirect UV detection and in Fig. 6.9(c) it can be clearly seen that also a non-UV absorbing positive ion with high effective mobility is present in the sample solution.

Applying a background electrolyte with UV absorbing ions affects the UV signal of salbutamol [23] as can be observed in Fig. 6.9(c). The results indicate that the sample ofVentolin syrup contains at least four components viz., salbutamol, a non-UV absorbing positively charged component and two UV-absorbing negatively charged components. The CZE experiments were repeated with several different background electrolytes at pH 4 to 8 giving good separations, showing that the choice of the background electrolyte is not critical.

Determination of terlmtaline sulphate For the determination of terbutaline sulphate Bricanyl tablets (labelled value

5 mg/tablet), Bricanyl ampoules for injection (labelled value 0.5 mg/ml) and Bricanyl syrup (labelled value 0.3 mg/ml) were used. In Table 6. VII all results for terbutaline sulphate are given.

As for salbutarnol, terbutaline could easily be separated from all other sample components without any pretreatment. The obtained values for the Bricanyl tablet with

6.3 P2-AGONISTS IN PHARMACEUTICALS 99

CZE are a slightly lower than those for the other methods. The liquid Brica.nyl samples also contain some other components, visible with HPLC and lTP.

Detennination of fenoterol hydrobromide Fenoterol hydrobromide was determined in Berotec tablets (labelled value

2.5 mg fenoterol/tablet, all measured values were recalculated to fenoterol/tablet), Berotec rotacaps (labelled value 200 µg fenoterol hydrobromide/capsule) and Berotec respirator solution (labelled value 5 mg fenoterol hydrobromide/ml). In Table 6. VIII all results are given.

The results of the respirator solution covered the labelled values. All methods show comparable values for the amount of fenoterol per tablet, although much lower than the labelled value. In the sample preparation much of the tablet did not dissolve and probably fenoterol partially adsorbs on the insoluble components. In the comparison of the techniques a sample preparation with 100 % recovery was not persued.

A similar problem occurred with Berotec rotacaps. In the isotachopherograms a slow UV absorbing component could be observed migrating between fenoterol and the terminating hydrogen zone, probably in an enforced way [24,25]. This component is partially mixed with fenoterol. For this reason the determined values are higher than those of HPLC and CZE. On diluting the sample solution the determined amount of fenoterol increases because fenoterol is completely separated from the unknown sample component with high effective mobility and the zone is enlarged owing to the presence of the unknown UV absorbing sample component with low effective mobility.

6.3.4 Conclusions

In the analysis with ITP, CZE and HPLC for all components a linear relationship between measured peak area or zone length and concentration of the components is obtained with regression coefficients better than about 0.999 and RSD values up to about 2 % • ITP and CZE seem to be more sensitive for irreproducibilities in amongst others the injected amounts, through which the repeatability of the HPLC values seem to be slightly better than those of ITP and CZE, although a disadvantage of the HPLC technique is the decreasing column quality.

Application of the techniques to the determination of salbutamol, terbutaline sulphate and fenoterol hydrobromide in several pharmaceutical dosage forms gave comparable results covering the labelled values although, especially in the electrophoretic techniques, other sample components present at high concentrations, can disturb the separation. In CZE with high EOF, both anions and cations can be observed, in contrast to ITP. For most pharmaceuticals a very simple pretreatment is sufficient to obtain sample solutions .. This procedure is, however, not adequate to desorb fenoterol from the Berotec tablets and rotacaps. Because the aim of this investigation was to compare ITP, CZE and HPLC a procedure to desorb completely the fenoterol was not persued in this case. Combined application of these techniques provides more information about the sample composition.


6.4 DETERMINATION OF SOME PHARMACEUTICAlS WITH MECC

6.4.1 Introduction

As already stated in chapter 1, MECC is a separation technique which can be used for uncharged and in water almost insoluble components. In order to study the qualitative and quantitative abilities of this technique, eight drugs were selected, the structural formulae of which are given in Fig. 6.10. As an example of a quantification the amount of dapsone in a tablet is. measured.

For the preparation of a sample mixture, water cannot be used as solvent and therefore the effect of the presence of methanol in the sample on the separation was examined.

0 I

ro~ ·1 0

Fig. 6.10: Structural formulae of drugs used (11icarbazin is a 1: 1 mixture of the two given components).

6.4.2 Experimental

Instromentation All experiments were carried out using the P/ ACE System 2000 HPCE, in a

fused silica capillary from Polymicro Technologies (Phoenix, AZ, USA), 50 µm I.D., total length 27.65 cm, distance between injection and detection 20.85 cm. The capillary was treated with a solution of .10 M HCl for 5 hours at a temperature of 160 °C in order to obtain a high mEOF· The wavelength of the UV detector was 214 nm. All experiments were carried out applying a constant voltage oflO kV with

6.4 SOME PHARMACEUTICALS WITH MECC 101

the anode at the inlet and the cathode at the outlet side. Data analysis was performed using the laboratory-written data analysis program CAESAR.

Separati,on conditions For all analyses an electrolyte system of 0.02 M TRIS with 100 mM SDS at

pH 8.5 adjusted by adding boric acid was used. It must be noted however, that due to difficulties in preparing reproducible electrolyte compositions, differences can occur in k or m~5 values using different batches. In all experiments the sample was introduced by pressure injection for 5 seconds.

If the capillary· tube is filled with a solution of SDS, the UV· signal of a solution of 0.001 M mesityl oxide in SDS introduced by pressure injection can be obsexved after 192 seconds. This means that at a separation volume of about 410 nl, the injected volume for 5 seconds pressure injection is about 11 nl. Although the minimal injection time with our apparatus is 1 second, a pressure injection time of 5 seconds was chosen for the sake of reproducibility.

Chemicals All drugs were donated by the State Institute for Quality Control for Agricultural

Products (RIKILT, The Netherlands). The dapsone tablets (OPGFarma 89c08-90067) were obtained at a local pharmacy.


Effect of the presence of methanol in the sample In order to study the effect of methanol in the sample on tEoF and tMc three

sample solutions of creatinine (which proved to ·be uncharged and an insolubilized component in this electrolyte system) and Sudan III as micelle marker, in a 100 mM SDS solution were prepared, and methanol was added to concentrations of 0, 10 and 20 % methanol. Three separations were carried out, injecting twice in each separation. The first injection was the sample (a) without, (b) with 10% and (c) with 20% methanol. After a separation for 3 min at 10 kV the sample mixture with 0 % methanol was injected in all the three instances, whereafter the separation was completed. The three electropherograms are given in Fig. 6.11. As can be seen in Fig. 6.1 la the time difference between tEoF and tMc of the first injection (t1) and that of the second injection (t'J) are equal. By the addition of only 10 % methanol to the sample of the first injection (case b) t1 decreases whereas t2 is nearly constant, and in case c a strong decreasing effect on t1 can be seen at a constant t2•

The fact that the time intervals between the two tEoF values are nearly constant in all instances means that by the addition of methanol in a sample the velocity. of the EOF is not influenced. The sample components, however, show a different migration behaviour. They tend to remain in the methanol plug (EOF) for a longer time, resulting in shorter migration times, due to a strong solubility effect and a local breakdown of the micelles. The results of this effect on the efficiency of the separations are demonstrated in Fig. 6.12, where the electropherograms are given for the separation of a mixture of phenol, p-cresol and 2,6-xylenol (all at lo-4 M, in all sample solutions) with the micelle marker Sudan III, with an increasing percentage of methanol in the sample.


t, I I

c s s c

a I J -

t, I I t.

s c c s

b J I J '--

" I t, t.

[ I I

c 0Ut c

-V- . I I

t.

0 6 12

t (min)

Fig. 6.11: Electropherograms for the analysis of creatinine (CJ and Sudan III (S) with and without methanol in the sample. In each experiment two injections occurred, viz;., the first ilflection (a) without, (b) with 10% and (c) with 20% methonolin the sample and the second ilflection without methanol in the sample for all instances. A separate step for :t mln at 10 kV was performed between the two injections. In all cases the sample contained 0.15 mg/ml creatinine and 0.035 mg/ml Sudan Ill and the injected volume was about 11 nl. For further explanation see text.

It can be clearly seen that, although the tEoF (methanol) is nearly constant for all electropherograms, the migration times of· all components strongly decrease whereas the separation efficiency declines dramatically. The addition of methanol to sample solutions for solvation effects or as EOF marker generally must be avoided, as it can greatly affect the separation as shown before. A better way to dissolve water-insoluble components is to use an SDS solution.

In Table 6.IX the calculated values for k according to eqn. 2.13 are given for the components of the sample used for the electropherograms of Fig. 6.12 (up to 70 % methanol). It will be clear that the k values strongly depend on the presence of methanol in the sample. In Table 6.X the average k values (standard deviation) from ten experiments are given for a sample of eight drugs dissolved in SDS without methanol and with 20% methanol. Although the k values are affected by the presence of methanol in the sample, the repeatability is fairly good at a given methanol concentration through which the k can be handled for screening purposes in such case.

In section 2.3 some advantages in the use of the ~s over k are already discussed. The most important advantage was that ~s can also be calculated if tMc values are unknown. To demonstrate this advantage eight experiments were carried out

6.4 SOME PHARMACEUTICALS WITH MECC

i

I

i

I

i

I

o.ooe 0.000 - 0

10% Methanol

• •

_l 2

30% Methanol

• A

8

o.oHr-----------. Q.01!C>

0.016

0.010

o.ooe 0.000 1----'l>"''-' '"" ~-- ,.,___

-o.oos'---~-~~-~--' 0 8

t !mW

o.025

®'20

(1()15

0010 ' .. o.ooe 0.000

2 .. 6 • I lmir\I

70% Methanol

0,026

o.020

0.016

CI010

o.oo&

0.000 - 0 2 ' a • tlmnl

90% Methanol

o.025

o.oeo 0.0t6

CIOtO

0000

0.000

2 .. • 6

t !mini

103

20% Methanol

"""" i °"120 ••

I at>•• at>•O

o.oo&

0.000 -0 2 .. 6 • t lminl

40% Methanol

0.025

i at>20

I O.O•&

0.0•0 •• o.ooe 0.000

2 .. 6 8

t....., 60% Methllnol

OOH

i 60®

I GO•&

(1<)10 • •• GOOS

0.000

11 4 8 8

t lminl

80% Methanol

OOH

i (),020

I 0011'

0.010

o.oo&

0.000

2 6 a tlminl

100% Methanol

OOH

i 0020

I 0.015

0.010

o.ooe 0.000

-<>.006 0 2 4 6 a

tlminl

Fig. 6.12: Electropherograms for the separation of (1) phenol, (2) p·cresol, (3) 2,6-xylenol and (4) Sudan Ill with increasing amounts of methanol in the sample. Concentration of all sample components was O.<XXJI M and the injection volume was about 11 nL


TABLE6.IX

CALCULATED CAPACITY FACTORS, k, FOR PHENOL, p-CRESOL AND 2,6-XYLENOL FOR SAMPLES WITH INCREASING AMOUNTS OF METHANOL IN THE SAMPLE. kMeoH=O; ksudanm= oo

Methanol k (%)

phenol p-cl'e591 2,6-xylenol

10 0.98 2.S9 4.72 20 0.92 2.43 4.38 30 o.ss 2.20 4.06 40 0.81 2.0S 3.76 so 0.77 1.SS 3.46 60 0.78 1.72 3.14 70 0.79 1.60 2.98

TABLE6.X

AVERAGE VALUES OF THE CAPACITY FACTORS, k, FOR THE SAMPLE COMPONENTS WITH STANDARD DEVIATION (IN PARENTHESES) FOR TEN EXPERIMENTS.

k

Component Without methanol With methanol

nicarbazin 0.375 (0.002) 0.398 (0.037) dimetridazole 0.565 (0.004) 0.563 (0.007) sulfadimidine 0.852 (0.005) 0.797 (0.009) sulfadiazine 1.833 (0.025) 1.570 (0.026) carbadox 2.137 (0.012) 1.905 (0.023) furaltadone 3.061 (0.012) 2.787 (0.023) dapsone 5.378 (0.018) 4.754 (0.038) fenbendazole 00 00

with a samfle mixture of the eight drugs for different mEoF varying between 50· 10"5

and 40· lo- cm2/Vs. In order to change the mEoF the capillary was rinsed extensively with 1 M HCI and/or KOH, after which a rinsing step with distilled water was carried out. Fenbendazole was used as micelle marker. In Table 6.XI the average calculated values (standard deviation) of k and the m~s are given for the five experiments in which the tMc could be measured (high EOF) and three experiments (low EOF) without a tMc· The obtained m~s values from both series of experiments cover each other, whereas in the latter case no k values could be calculated.

6.4 SOME PHARMACEUTICALS WITH MECC 105

TABLE6.XI

AVERAGE CALCULATED VALUES AND STANDARD DEVIATIONS (IN PARENTHESES) FOR k AND m§*· HP (cm2/Vs) FOR EIGHT DRUGS IN EXPERIMENTS WITH VARYING mroF·

Compound k m§* m§" (n=5) (n=5) (n=3)

EOF 0 45.71 (2.87) 41.13 (0.47) Nicarbazin 0.271 (0.005) - 8.06 (0.20) - 8.25 (0.04) Dimetriclazole 0.447 (0.006) -11.68 (0.24) -11.94 (0.07) Sulfadimidine 0.714 (0.004) -15.75 (0.22) -16.09 (0.03) Sulfadiazine 1.615 (0.052) -23.35 (0.56) -22.72 (l.11) Carbadox l.890 (0.031) -24.73 (0.42) -24.60 (0.70) Furaltadone 2. 756 (0.021) -27.75 (0.36) -28.10 (0.03) Dapsone 4.875 (0.070) ·31.38 (0.43) -31.81 (0.04) Fenbenclazole QO -37.81 (0.44)

TABLE 6.XII

AVERAGE MIGRATION TIME, t (min), m§"·l05(cm2/Vs), SLOPE AND INTERCEPT (ARBITRARY UNITS) AND CORRELATION COEFFICIENT OF CALIBRATION GRAPHS FOR THE DIFFERENT SAMPLE COMPONENTS (STANDARD DEVIATIONS IN PARENTHESES).

Compound m§' Slope Intercept Correlation Coefficient

nicarbazin 2.57 (0.020) -11.73 (0.096) 78.96 -0.40 0.999 dimetriclazole 2.86 (0.026) -15.48 (0.078) 123.70 -0.63 0.999 sulfadimidine 3.17 (0.043) -18.71 (0.194) 133.85 -1.43 0.998 sulfadiazine 4.37 (0.058) -27.03 (0.196) 195.33 -0.24 0.999 carbadox 4.64 (0.065) -28.33 (0.130) 92.43 -1.36 0.997 furaltadone 5.36 (0.081) -31.10 (0.125) 119.19 -1.50 0.999 dapsone 6.66 (0.127) -34.69 (0.141) 429.19 -2.54 0.999 fenbendazole 11.30 (0.342) -40.60 (0.137) 258.87 -2.62 0.997

Quantiiative analysis To study the quantitative possibilities of MECC, experiments are carried out with ·

a sample mixture consisting of 0.30 mg/ml of nicarbazin, dimetridazole, carbadox and furaltadone, 0.15 mg/ml sulfadimidine, sulfadiazine and dapsone and 0.030 mg/ml fenbendazole dissolved in a 100 mM SDS solution. This sample was diluted 1.2, 1.5, 2, 3, 6 and 10 times. All these dilutions were measured three times and with the measured peak area calibration curves were set up. In Table 6.XII all measured migration times and calculated m~8 for the components, the slopes, intercepts and the correlation coefficients of the calibration curves are given (standard deviations are given in parentheses). From Table 6.XII it can be concluded that linear calibration curves (nearly through the origin) are obtained. In Fig. 6.13 the calibration curves are shown applying the average values of the three experiments for all dilutions.


(j) :::> 300 <( E .......

t\1

~ ~ IU 150 Q) a.

% of original sample cone.

Fig. 6.13: Calibration curves for ( +) nicarbazin, (.i.) dimetrlda:wle, (0) suifadimidine, ( +) sulfadiazine, (•) carbadox, (e) furaltadone, (v) dapsone and (0) fenbendazole. For further explanation see text.

The detennination of dapsone in tablets with MECC As an application the amount of dapsone in a tablet (mass 202.8 mg, containing

100 mg dapsone/tablet) was determined. The tablet was pulverized and 15.4 mg was dissolved in 50 ml of 100 mM SDS solution containing about 0.1 mg fenbendazole as tMc marker. As duplicate a 2 times diluted solution was used. Applying the calibration curve, (at the 95% confidence interval) respectively 104.5 (S.D. = 1.2 mg) and 105.8 (S.D. = 1.4 mg) dapsone in a tablet were found, showing that MECC can be used in an appropriate way for the quantitative determination of dapsone in tablets. In Fig. 6.14 the electropherogram of the standard sample mixture with the eight drugs and the electropherogram of the sample mixture from the tablet are given. Although the migration times for peak 7 and 8 differ considerably, due to a small difference in tEOF• the calculated m§5 are nearly identical.

6.4.4 Conclusions

If uncharged and in water practically insoluble compounds have to be analysed, MECC can be used as a highly efficient separation method. Nevertheless special attention should be paid to the sample preparation. It appeared that methanol can not be used as solvent for the sample components, because a high concentration of

6.5 AMINOGLYCOSIDES WITH HYPHENATED CZE/MECC

~

I Ill

0.030

0.025

0-020

0.015

0.010

0.005

0.000

-0.005 0

A 2 4

1

3

6 5

I I! I

EOF

5

B 7 I

0.030 7

0.025

~ 0.020

8 I 0.015

0.010

0.005 ECF

L........A.....W 0.000

-0.005 i 10 15 0 5

t (min) t (min)

107

a

10 15

Fig. 6.14: Electropherogram of (A) a standard sample of (1) nicarbazin, (2). dimetridazole, (3) sulfadimidine, (4) sulfadiazine, (5) carbadox, (6)furaltadone, (7) dapsone and (8)/enbendazole, and (B) a sample of a dapsone tablet with fenbendazole added as tMc marker. Injected volume was about 11 nl (see te.xt for composition of the sample).

methanol in the sample affects the separation dramatically. A better way to prepare the sample is to use a micelle solution to dissolve the sample.

When samples are prepared in the proper way linear calibration graphs can be obtained for all the components studied, and the results for the determination of dapsone in tablets without any sample pretreatment were satisfactory.

6.5 DETERMINATION OF AMINOGLYCOSIDES WI1H HYPHENATED CZE WITH JNDIRECT UV DETECTION AND MECC

6.5.1 Introduction

Since the introduction of MECC [26,27] and CZE [28,29] many components of pharmaceutical interest have been determined [9,18,30,31,32, 33] using these techniques. No attention has been paid, however, to the analysis of arninoglycoside antibiotics. So far, the non-UV absorbing arninoglycoside antibiotics are determined by e.g. ion-pair RP-HPLC using a refractive index detector [34,35] and spectrometric methods using derivative agents [36].

Combined pharmaceuticals often contain both charged and neutral compounds, whether or not with UV. In this section, the possibilities of CZE with indirect UV for the determination of aminoglycoside antibiotics and hyphenated CZE with indirect UV and MECC for the determination of aminoglycoside antibiotics and neutral components in combined pharmaceuticals is studied. In Fig. 6.15the structural formulae of some representatives of aminoglycoside antibiotics are given.


Fig. 6.15: Structuralfonnulae of some amill{)glycoside antibiotics.

6.5.2 Experimental

Instrumentation For all experiments the P/ ACE System 2000 HPCE was used. All experiments

were carried out in a fused-silica capillary from Siemens (Siemens AG, Karlsruhe, Germany) 50 µm I.D., total length 27 cm, distance between injection and detection 20 cm, or total length 67 cm, distance between injection and detection 60 cm. The wavelength of the UV detector was 214 nm. Data analysis was performed using the laboratory-written data analysis program CAESAR.

Chemicals Amikacin dihydrate, gentamycin sulfate, streptomycin sulfate and tobramycin

were obtained from Fluka (Buchs, Switserland), Butirosin disulfate salt, dibekacin sulfate salt, dihydrostreptomycin sesquisulfate salt, kanamycin B sulfate salt, lividomycin sulfate salt, neomycin sulfate, paromomycin sulfate, ribostamycin sulfate salt and sisomycin sulfate salt are obtained from Sigma (St. Louis, USA), paracetamol was obtained from Merck-Schuchardt (Hohenbrunn, Germany), dexamethason was obtained from 'De onderlinge pharmaceutische Groothandel (nr. 85G30-50971, Utrecht, The Netherlands), dapsone was donated by the State Institute for Quality Control of Agricultural Products (Wageningen, The Netherlands) and Otosporin eardrops were obtained from The Wellcome Foundation (London, England). The fluorochemical surfactant FC 135 was obtained from Fluorad/3M (Leiden, The Netherlands). Hydrocortisone was obtained from Aldrich (Brussels, Belgium).

Standard solutions For the calibration graphs, standard solutions of the aminoglycoside antibiotics

were prepared weighing accurately 50.0 mg of the standards and dissolving them in 50.0 ml of a 100 mM CTAB solution. For the calibration graphs, dilutions of this stock solution were used at concentrations of 1.0, 0.8, 0.6, 0.4, 0.2 and 0.1 mg/ml.

6.5 AMINOGLYCOSIDES WITH HYPHENATED CZE/MECC 109

For the determination of neomycin and hydrocortisone in Otosporin eardrops, a stock solution of 0.2 mg/ml hydrocortisone and 0.1 mg/ml neomycin was prepared and six dilutions were prepared spread between the 1- and ten-fold dilution, so that the concentration of the sample is near the centroid of the calibration graph.

Sample preparation Otosporin eardrops labelled value 10 mg/ml hydrocortisone and 5 mg/ml

neomycin, were diluted IOQ..fold with distilled water. This dilution was used for the injection without further pretreatment.


Detenninafion of aminoglycosides by CZE with indirect UV Aminoglycoside antibiotics are non-UV-absorbing components, positively

charged in their protonated form at a pH between 3-8. ITP experiments showed that they migrate at intermediate pH with effective mobilities of 20· 10·5 to 50· 10·5 cm2/Vs with positive charges of 2 + to 5 + as could be concluded from their response factor [21]. In first instance the arninoglycoside antibiotics were determined in CZE in the cationic mode (cathode at detection side) with indirect UV mode. Very bad peak shapes, probably due to strong attraction forces between the highly positively charged components and the negatively charged capillary wall, and a low resolution were the result. Because higher separation numbers [10] can be obtained at low apparent mobilities and to suppress the attraction forces between analytes and capillary wall, experiments are carried out in the anionic mode (anode at detection side) with reversed electroosmotic flow by the addition of FC 135 to the background electrolyte [37]. All arninoglycoside antibiotics now migrate in the upstream mode. By the addition of FC 135 easily mEoF values can be obtained down to -90· 10·5 cm2/Vs. In Fig. 6.16 the mEoF as a function of the pH value of the background electrolyte is given for the background electrolytes with FC 135. In Table 6.XIII the compositions of all background electrolytes are given. As can be seen from Fig. 6.16, the absolute values of m00p increase with decreasing . pH, in contrast to the mEoF in background electrolytes without FC 135, where the m00p increases with increasing pH. This can be easily understood as follows. In fused silica the negative charge of the capillary wall increases with increasing pH (higher t-potentail, higher m00p). An adsorbing layer of FC 135 molecules shields this negative charge. At high pHs this shielding is less effective, resulting in a lower mEoF· Owing to this effect, the peak shape of the aminoglycosides in the reversed mode will also be the best at low pH, and will get worse at higher pH values. In Fig. 6.17 an example of the separation of a mixture of amikacin, dihydrostreptomycin, kanamycin, lividomycin, sisomycin and tobramycin (all at a concentration of 0.1 mg/ml) is given applying a background electrolyte of 0.01 M imidazole at pH 5.0 adjusted by adding acetic acid with the additive FC 135 (50 µ.g/ml) [38]. Effective mobilities of the aminoglycoside antibiotics were determined as a function of pH. In Table 6.XIV all determined effective mobilities of the aminoglycoside antibiotics are given. From Table 6.XIV and Fig. 6.17 it can be concluded that aminoglycosides easily can be determined in the anionic CZE mode with reversed EOF by the addition of FC 135 with indirect UV at an optimum pH of about 5, although not all components can be separated at this pH.


-60 (j) > ;;:-. E -70 g

"' 0 ..-

* -80

~ E

-90

-100'--~_,___,__....__,_....._~_.__.__...___,___.___.....__,

3 4 5 6 7 8 9 10

pH

Fig. 6.16: mEOF as afimction of pH.for several background electrolytes with 50 µ.glml FC 135.

5 5 (I) 0 c (\) .a ,_ 0 (/)

fd

0.001

0.000

-0.001

-0.001

-0.002

-0.003 0

I -·

5

;"'

2 1

4 6 3

EOF 5

10 15

time (min)

Fig. 6.17: Electropherogram for the separation of (1) · dihydrostreptomycin, (2) lividomycin,. (3) kanamycin, (4) tobramycin, (5) sisomycin and (6) amikacin (all 0.1 mg/ml) in the anionic mode with reversed EOF applying a background electrolyte 0.01 M imida;.ole acetate at pH 5.0 containing the additive FC 115 (50 µ.llml), l., = 61 cm~ applied voltage 12.5 kV, pressure injection, time 2 s.


TABLE 6.XllI

COMPOSIDONS OF BACKGROUND ELECTROLYTES AT DIFFERENT pH VALUES

All buffers are prepared by adding the buffering counter species to the cations until the desired pH was reached. To all buffers, FC 135 was added at a concentration of SO µ.g/ml..

Cation

0.01 M imidazole 0.01 M imidazole 0.01 M imidazole 0.01 M imidazole 0.02 M imidazole 0.02 M imidazole 0.02 M benzylamine

TABLE6.XIV

Buffering counter species

formic acid formic acid acetic acid MES acetic acid acetic acid acetic acid

pH

3.3 4.0 s.o 6.0 7.0 7.9 9.0

CALCULATED EFFECTIVE MOBILITIESm· la5 (cm2Ns) OF AMINOGLYCOSIDEANTIBIOTICS AT DIFFERENT pHs

For the composition of the background electrolytes see Table 6.Xlll

component pH 3.23 4.03 4.99 6.01 7.02 7.90

am.ikacin 42.76 42.31 41.68 40.01 34.15 30.11 butirosin 43.08 42.74 41.0S 37.85 34.27 32.07 dibekacin 52.35 51.46 48.45 46.34 40.26 34.33 dihydrostreptomycin 35.31 35.00 34.58 35.26 32.Sl 31.47 gentamycin 50.39 49.03 46.39 44.81 40.34 35.93 kanamycin 50.31 49.31 46.18 43.30 35.37 27.90 lividomycin 42.73 42.05 40.03 36.34 28.91 23.78 neomycin 51.28 50.22 47.99 46.39 39.98 32.53 paromomycin 47.52 46.94 44.51 41.50 34.68 28.00 ribostamycin 46.05 44.85 41.10 39.14 34.52 28.54 sisomycin 51.41 50.71 48.15 45.53 39.94 33.86 streptomycin 34.99 34.94 34.70 34.92 33.22 32.36 tobramycin 51.34 50.67 47.51 45.04 38.21 30.73

Hyphenated CZB and MBCC Pharmaceuticals often contain both neutral and charged components. In order to

determine simultaneously both charged and neutral components a micelle-forming surfactant has to be added to the background electrolyte.

Applying a hyphenated CZE and MECC system, both negative, positive and neutral components can migrate in any order. As an illustration, a schematic representation of the different migration possibilities is given in Fig. 6.18. In

a 0

b 8

c

7

L

us

I i 4

I 6.i k=oo

:

! i

----~I I ! :5

i

I 31

I m_

2. k=O

lt MECC window i ' MC jtEOF

MS OS -----time

©

©

Fig. 6.18: Schematic representation of several migration modes in hyphenated CZE with indirect UV in the anionic mode with reversed EOF and MECC with a cationic surfactant. (a) original situation, (b) separation after some time and (c) electropherogram of components migrating in different modes. For further explanation see text.

...... ...... N


background electrolyte containing a cationic surfactant. The cathode is placed at the injection site (i). In Fig. 6.18b the situation is given after some time where Fig. 6.18c shows the corresponding electropherogram. In this electropherogram component 1 is negatively charged and migrates in the down-stream mode (DS) in front of a waterdip (midstream mode, MS), that can act as EOF marker. Also non solubilized neutral component can act as EOF marker (with a Ca.pacity factor k=O), if the component absorbs UV light. The completely solubilized component 6 acts as micelle marker (k= oo ). The time-window for neutral components migrating in the MECC mode is demarcated by tMc and tEoF and e.g., oomponent 4 migrates in the MECC mode. Component 3, negatively charged but partially solubilized, migrates behind the EOF marker. Component 5, without UV, is a positive component with a mobility.smaller than that of the micelles, whereas the positive component 7, without UV, with a mobility larger than that of the micelle marker migrates behind the micelle marker.

For the determination of neutral components simultaneously with aminoglycosides antibiotics in the anionic mode with reversed EOF, in first instance sodium dodecyl sulfate (SDS) was used as micelle forming surfactant. Owing to the fact that probably the additive FC 135 solubilized in the SOS-micelles, the mobility of the reversed EOF strongly decreases through which the aminoglycoside antibiotics could no longer be detected in the anionic mode. For that reason the cationic surfactant cetyl trimethyl ammonium bromide (CTAB) was tried as micelle-forming surfactant causing, moreover, a reversed EOF. Good results in the separation of aminoglycoside antibiotics and several neutral components could be obtained with the addition of CTAB and FC 135.

In Fig. 6.19 an example is given of the separation of a mixture of the UV absorbing neutral components paracetamol and dapsone (0.02 mg/ml) and dexamethason and the aminoglycoside antibiotics dihydrostreptomycin, kanamycin, tobramycin and sisomycin (all 0.1 mg/ml). The background electrolyte consisted of 0.01 M imidazole at pH 5.0 adjusted by adding acetic acid with the additives 50 µg/ml FC 135 and 100 mM CTAB. In order to check quantitative aspects of separations in a hyphenated CZE/MECC system calibration graphs of peak area versus injected concentration (pressure injection 5 sec) were set up for the aminoglycoside antibiotics dihydrostreptomycin and sisomycin and the neutral components paracetamol and dapsone in the same background electrolyte. In Fig. 6.20 the calibration graphs are presented and in Table 6.XV all regression parameters are given showing a linear relationship between peak area and injected concentration for both charged and neutral components.

As an application neomycin an hydrocortisone were determined in Otosporin eardrops. The sample was measured four times. In Table 6.XVI the regression coefficients for the calibration graphs of the two components, and the labelled and measured concentration of the components in the sample are given.


0.004

3

5 0.002 4

~ Q) u c 0.000 m e g .0

-0.002 (ll

2

EOF -0.004

0 5 10 15 20

time (min)

Fig. 6.19: Electropherogram for the separation of (1) paracetamol, (2) dihydrostreptom:ycin, (3) dapsone, (4) de.xamethaso11, (5) kanam:ycin, (6) tobramycin and (7) sisomycin applying a hyphenated CZE-MECC system consisting of 0.01 M imidazole acetate at pH 5.0 containing FC 135 (50 µJlml) and 100 mM CTAB. Capillary length 67 cm, applied voltage 15 kV, pressure injection time, 5 s.

360

(i) :J <(

.s 270

ro (]) .... ro

.::<:. 180

ro <l.l a.

90

concentration (mg/ml}

Fig. 6.20: Calibration graphs for peak area (mAUs) versus injected concentration (mg/ml) for (v) dihydrostreptom:ycin, (.i.) sisom:ycin, (+)paracetamol and (o) dapsone. For separation conditions see Fig. 6.19.

6.5 AMINOGLYCOSIDES WITH HYPHENATED .CZE/MECC 115

TABLE6.XV

REGRESSION COEFFICIENT, r, AND LIMIT OF DETECTION, LOD (p.g/ml), FOR THB CAUBRA TION GRAPHS OF DIHYDROSTREPTOMYCIN, SISOMYCIN, PARACETAMOL AND DAPSONE

Component r LOD

dibydrostreptomycin 0.9998 23.38 sisomycin 0.9995 35.89 paracetamol 1.0000 9.87 dapsone 0.9997 29.84

TABLE 6.XVI

REGRESSION COEFFICIENT, r, FOR THE CALIBRATION GRAPHS OF NEOMYCIN AND HYDROCORTISONE, AND LABELLED AND MEASURED CONCENTRATIONS (mg/ml) IN THB OTOSPORIN EARDROPS OF THESE COMPONENTS

Component

neomycin hydrocortisone

6.5.4 Conclusions

r

0.9997 0.9990

labelled

S.00 10.00

measured

S.42 10.56

The determination of aminoglycoside antibiotics in the cationic mode is troublesome owing to attraction forces between the positively charged aminoglycoside antibiotics and the negatively charged capillary wall. By addition of FC 135 to the background electrolyte, resulting in a reversed wall charge, the aminoglycoside antibiotics easily could be determined in the anionic mode with reversed EOF. The effective mobilities of thirteen aminoglycoside antibiotics were measured as a function of pH. Charged and neutral components can be determined simultaneously applying hyphenated CZE and MECC. By the application of an electrolyte consisting of 0.01 M imidazole at adjusted at a pH of 5.0 by adding acetic acid and the additives FC 135 (50 µg/ml, for reversed EOF) and CTAB (100 mM, as micelle-forming surfactant) the aminoglycoside antibiotics (in the anionic mode with reversed EOF with indirect UV) and neutral components (reversed MECC mode) could simultaneously be determined. In this way obtained values for neomycin and hydrocortisone in eardrops cover the labelled values.


References

1. J.F .. M. Nouws, A. Smulders and M. Rappalini, Vet. Quart., 12 (1990) 129. · 2. D.G. Pope and J.D. Baggot, in: Formulation of Veterinary Dosage Forms (Drugs and the

Pharmaceutical Sciences, Vol. 17; J. Blodinger Ed.) Marcel Dekker, New York, 1983. 3. H.D. Mercer, J.C. Garg, J.D. Powers and T.W. Powers, Am. J. Vet. Res., 38 (1977) 1353. 4. J.D. Baggot, Principles of Drug Disposition in Domestic animals: The Basis of Veterinary

Clinical Pharmacology, Saunders, Philadelphia, 1977 . 5. A.J. Manuel and W.A. Steller, J. Assoc. Off. Anal. Chem., 64 (1981) 794. 6. N. Haagsma and C. van de Water, J. Chromatogr., 333 (1985) 256. 7. J.D. Henion, B.A. Thomson and P.H. Dawson, Anal. Chem., S4 (1982) 451. 8. R. Malish, Z Lebensm. Unters. Forsch., 182 (1986) 385. 9. M.T. Ackermans, J.L. Beckers, F.M. Everaerts, H. Hoogland and M.J.H. Tomassen,

J. Chromatogr., 596 (1992) 101. 10. J.L. Beckers, F.M. Everaerts and M.T. Ackermans, J. Chromatogr., 537 (1991) 407. 11. The Merck Index, loth ed., Merck and Co. Rahway, N.J., U.S.A., 1983. 12. M.M.L. Aerts, Thesis, Free University of Amsterdam, The Netherlands, 1990. 13. L.E. Martin, J. Rees and R.J.N. Tanner, Biomed. Mass Spectrom., 3 (1976) 184. 14. J.G. Leferink, J. Dankers and R.A.A. Maes, J. Chromatogr., 229 (1982) 217. 15. D.A. Williams, E.Y.Y. Fung and D.W. Newton, J. Pharrn. Sci., 71(1982)956. 16. V. Das Gupta, J. Liquid Chromatogr., 9 (1986) 1065. 17. J.E. Kounterellis, C. Markopoulou and P. Georgakopoulos, J. Chromatogr., 502 (1990) 189. 18. M.T. Ackernliins, J.L. Beckers, F.M. Everaerts and I.G.J.A. Seelen, J. Chromatogr.,

590 (1992) 341. 19. F.M. Everaerts, J.L. Beckers and Th.P.E.M. Verheggen, Isotaclwphoresis, Theory;

Instrumentation and Applications. Elsevier, Amsterdam, 1976. 20. J.L. Beckers and F.M, Everaerts, J. Chromatogr., 508 (1990) 3. 21. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 508 (1990) 19. 22. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 470 (1989) 277. 23. M.T Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 549 (1991) 345. 24. P. Gebauer and P. Bocek, J. Chromatogr., 267 (1983) 49. 25. P. Boeek and P. Gebauer, Electrophoresis, 5 (1984) 338. 26. S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchiya and T. Ando, Anal. Chem., 56 (1984) 113. 27. S. Terabe, K.Otsuka and T. Ando, Anal. Chem., 51 (1985) 834. 28. F.E.P. Mikkers, F.M. Everaerts and Th.P.E.M. Verheggen, J. Chromatogr., 169 (1979) 1. 29. J.W. Jorgenson and K.D. Lukacs, Anal. Chem., 53 (1981) 1298. 30. M.T. Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 585 (1991) 123. 31. A. Wainright, J. Microcolumn. Sep., 2 (1990) 166. 32. H. Nishi, N. Tsumagara and S. Terabe, AnaL Chem., 61 (1989) 2434. 33. S. Fujiwara and S. Honda, Anal. Chem., 59 (1987) 2773. 34. G. Inchauspe and D. Samain, J. Chromatogr., 303 (1984) 277. 35. G. Inchauspe, P. Delrieu, P. Dupin,M. Laurent and D. Samain,J. Chromatogr.,404 (1987) 53. 36. S.S. Sampath and H.H. Robinson, J. Pharrnaceut. Sci., Vol. 79, No. 5, May 1990. 37. A. Emmer, M. Jansson and J. Roeraade, J. Chromatogr., 547 (1991)544. 38. M.T Ackermans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., accepted for publication.

APPENDIX

ISOTACHOPHORESIS IN OPEN CAPILLARIES

So far, isotachophoretic (ITP) analyses have been carried out in commercially available or laboratory-made ITP instruments, generally in ~closed systems. At present several instruments are commercially available, originally designed for capillary zone electrophoresis with open capillaries, and these instruments can also be used for ITP. If ITP experiments are carried out using such apparatus, however, an electroosmotic flow will act on the ITP system, and four modes can be distinguished, viz., the anionic, cationic, reversed anionic and reversed cationic mode. The applicability of these modes strongly depends upon the velocity of the electroosmotic flow. The velocity of the electroosmotic flow, however, varies with the choice of the leading and terminating electrolyte and the c-0m.position of the sample. In this appendix examples are given showing some typical features of these modes. Detailed consideration of the mathematical model for isotachophoresis with electroosmotic flow showed that this model is identical with that of isotachophoresis without electroosmotic flow.

Due to the variation in the velocity of the electroosmotic flow, the reproducibility in quantitative analysis is a serious problem. For quantitative analysis at least an internal standard must be used to correct for undesirable fluctuations in the electroosmotic flow and irreproducible injections. Better results can be obtained by effectively suppressing the electroosmotic flow by using additives such as methylhydroxyethylcellulose. Results of quantitative experiments of isotachopboresis in open capillaries are given, showing some of the problems.

A.1 INTRODUCTION

Generally, electrophoretic equipment can be used for all electrophoretic modes, viz., for isotachophoresis, zone electrophoresis, moving boundary and isoelectricfocusing. The choice of the electrolytes in the capillary and electrode compartments determines which electrophoretic·mode is used, e.g., for the ZE mode a background electrolyte and for ITP a leading and terminating electrolyte will be used.

So far, ITP has usually been carried out in laboratory•made [1] or commercially available apparatus with closed systems, i.e., working with a suppressed EOF. At present, commercial apparatus for CZE is available, generally with open capillaries. Because this equipment can also be used for ITP it is of interest to investigate the possibilities of ITP in open systems, especially on bearing in mind that ITP can be used possibly as a selective sample preconcentration method prior to CZE.

118 APPENDIX: ITP IN OPEN CAPILLARIES

The first ones to carry out ITP in open systems were Hjerten et al. [2] who carried out displacement electrophoresis experiments in open glass capillaries, which were coated to suppress the EOF and adsorption of the solutes on the tube wall. Udseth et al. [3] reported experiments with tandem ITP/MS and showed that the separation was not disturbed by the EOF. They recognized that during the analysis the velocity of the EOF is not constant and is first determined by the composition of the leading electrolyte and finally by that of the terminating solution. Thormann [4] studied the impact of electroosmosis on zone formation and displacement and described anionic and cationic separations in open capillary systems.

Different modes of ITP in open systems If ITP experiments are performed in open capillaries, four different modes can

be distinguished. In Fig. A. la, the situation is given for the cationic ITP mode (CM). In this instance the capillary will be filled with the leading electrolyte L, while the

INLET OUTLET

ITP eOF d&teOtO< a I - -CM +TI s I L l

b ITP EOF I --AM -TI s I L I +

c ITP EOF --RAM +L s T

d

ACM +

Fig. A.I: Schematic represemation of the four modes in ITP experimems with EOF. (a) CM represents the caiionic, (b) AM the anionic, (c) RAM the reversed anionic and (d) RCM the reversed cationic mode. For further explanations see text.

terminator solution T will be present on the inlet side of the apparatus. The sample solution S will be introduced between L and T. The detector is placed at the outlet side. The EOF will generally act (using silica capillaries), in the direction from anode to cathode through which the cationic ITP system will be pushed forwards with an extra velocity of the EOF compared with cationic ITP experiments in closed systems.

In Fig A.lb the situation is given for the anionic ITP mode (AM). Hence the cathode must be placed at the inlet side and the anode at the outlet. The capillary will be filled with the leading electrolyte L and the terminator T must be present at the inlet. This procedure can be compared with ITP with counterflow [1] and is only

A.2 THE ISOTACHOPHORETIC MODEL 119

useful if the velocity of the leading ions is larger than that of the EOF during the whole experiment. Remember that in this instance anionic species with mobilities smaller than that of the EOF must also migrate to the anode according to the isotachophoretic ·condition.

If the velocity of the EOF is greater than that of the anionic ITP system, there is a net migration into the direction of the cathode. The only way to carry out the experiments is to place the anode at the inlet side and the cathode at the outlet side (see Fig. A. le). The capillary must be filled with the terminating electrolyte T and the leading electrolyte L must be present at the inlet. Although the ITP separation takes place into the direction of the anode, there will be a net velocity of the ITP system in the direction of the cathode (detector-side) and the components will be detected in a reversed order compared with a normal anionic ITP system, viz., first the terminator, then all sample components with increasing mobilities and finally the leading ions. This is called the reversed anionic ITP mode (RAM). If the velocity of the leading ions is greater than that of the EOF the ITP system will move to the anode and the components cannot be detected at the outlet.

Analogous to the RAM a reversed cationic mode (RCM) can be considered, that can be applied if there is a reversed EOF (e.g., using coated capillaries and/or additives to the electrolyte system) from cathode to anode, with a velocity greater than that of the cationic ITP system (see Fig. A.Id). Here the cathode must be placed at the inlet and the anode at the outlet. The capillary must be filled with the terminator solution and the inlet vessel with the leading electrolyte. As in the RAM the components will be detected in a reversed order. The RCM shall not be considered,

· as this mode will generally not be useful in practice. It will be clear that the velocity of the EOF is extremely important in the

migration behaviour of ITP systems. A complicating factor is, moreover, the fact that in ITP experiments the velocity of the EOF continuously will change. In the beginning of the experiments the EOF will be determined by the composition of the electrolyte in the capillary, i.e., the leading electrolyte for a normal ITP system. During the analysis the capillary will be filled more and more with terminator solution, with a concentration and pH different from those of the leading electrolyte and hence with a different zeta potential and with an E gradient much larger than that of the leading electrolyte according to the isotachophoretic condition. Generally, this will lead to a higher velocity of the EOF. The variations in the EOF during ITP experiments will be discussed in section A.3. For non-compressible solutions it can be assumed, however, that at a certain time the velocity over the whole capillary will be constant, and by this. the EOF displacement in a time span will be constant over the whole capillary. This is used in the mathematical model for ITP with EOF.

A.2 THE ISOTACHOPHORETIC MODEL

If the steady state in ITP is reached, all zone properties such as concentration, conductivities and pH values of the zones and the effective mobility of the components can be calculated. The general equations used are the equilibrium equations, the mass balance of the buffer, the principle of electroneutrality, the modified Ohm's law and the isotachophoretic condition. In this model the reduced number of parameters is


always four in all zones, viz., [L]1 or [A]1, CBlv E and pH. For all zones always four known parameters and/or equations are necessary, by means of which all parameters can be calculated. For the leading zone the known parameters are generally [L]t and· [B]t and the equations are Ohm's law and the electroneutrality (EN). For all other zones the four available equations are the EN, Ohm's law, the buffer equation and the isotachophoretic condition (IC).

For the calculation of parameters of the different zones in ITP with EOF, an ITP model, already published by Beckers et al. [5,6] is modified.

Effective mobility Tiselius [7] pointed out that a substance that consists of several forms with

different mobilities in equilibrium with each. other will generally migrate as a uniform substance with an effective mobility given by [5,6]:

(A.1)

me can be calculated from m0 correcting for the Debye-Hiickel-Onsager effects in an iterative way.

Although in the general· descriptions ·of equilibria and effective mobility of a substance, no differences exist between the leading, sample, terminating and buffet ionic species, there is distinguished between them using the symbols L, A, T and B. For the description of the "steady state" in ITP with EOF, further the mass balance of the buffer, the principle of electroneutrality, the modified OHM's law and the isotachophoretic condition are needed.

Mass balance of the buffer With the mass balance of the buffer (OHM's law and the EN must also be

obeyed) the leading zone determines the conditions of the proceeding zones. The mass balance of the buffer can be derived as follows for a cationic ITP system (see Fig. A.2a).

The zone boundary LI A moves in a unit of time over a distance of EL I mL,L I +s or EA I mA,A I +s, assuming that the contribution of displacement by the EOF is s at a certain moment. For cationic species the displacement is increased with s, while for anions, moving to the opposite direction, the net displacement will be decreased with s. The anionic buffering counter ions at time t=O present at the zone boundary L/ A will reach point D at t= 1. The distance from U A to D then will be EA I mB,A I -s. The buffer ionic species at t=O present in point C, will just reach the boundary UA at t=l. The distance from C to UA is then EL I mB,L I -s.

A.2 THE ISOTACHOPHORETIC MODEL

INLET 6 1 OUTLET ~------------------------!>;

VIS'J/lt- 0

DI l IA/L : t I t l I I re------............... .;"'f------------~~E-------------31" l E.lm,..1- S E,.lrnuJ + S I E,.lrn,,J - s

a

+

·~-----li-:2---~:~~~~~~~~~~~~:L-----------~ A!Li c!t=O

i+ b

I

D' I IAIL ;t= l ~;.. .... ----------.;J.f----..................... ..:_.:E ...... ---------~ I E.lm.J + s E,.l"\J - s i E,.lrn,,J + s L~-----------t,.-;r-------.;-

C cr-----1L1-~~r-------£.laj-~~---------'. t = 0

+ I

:: : iuA oit~ I

I ~~-----------.;of I : .. - E,.l"\J : : ~:.-------------------------..:..!-E-·-~-------~ • l:,.lrn,,J + s . t::. 2

121

Fig. A.2: Migration paths of the buffering counter ionic species over a zone boundary between a leading zone and a sample zone (LIA), for the (a) cationic, (b) anionic and (c) reversed anionic mode. For funlu?r explanation, see text.

This means that all buffer ionic particJes present in the leading zone between U A and C with a concentration of [B]1.i at time t=O (41) will be present in the zone A with a concentration of [BJ,,A between UA and D at t=l (42). Therefore, the buffer mass balance will be as follows:

[B]1,A(EAlm8 .Al-s+ELlmL,LI +s)=[B]1,i(ELlm8 ,il-s+ELlmL,LI +s) (A.2)

or

(A.3)

(A.4)

or


(A.5)

The buffer balance in ITP with 'EOF is identical with that in closed systems, because the contributions of the EOF for anions and cations cancel each other. Following this procedure for anionic separations with an ITP. velocity greater than that of the EOF (see Fig. A.2b) and for a reversed anionic ITP system (see Fig. A.2c) the same equation is al.ways obtained.

The principle of electroneutrality In accordance with the principle of electroneutral.ity (EN), the arithmetic sum of

all products of the concentration of all forms for all ionic species and the corresponding valences, present in each zone, must be zero. For the electroneutral.ity of a zone can be written:

n11 n8

[H30J - [OHl + L (Z-l)[Az-i] + L (z-1)[B·i:-i] = 0 (A.6)

i=O i=O

Modified Ohm's law Working at a constant current density:

ELuL = EAuA (A.7)

or the function

RFQ = (A.8)

must be zero. The overall electric conductivity, u, of a zone is the sum of the values

c(Em+s)zFIE and consequently for a cationic separation:

n,i

[H30j(EmH+s)-[0Hl(Em0H+s)+ I: [Az-i](Emz-i(z-l)+s(z-l))+ i=O (A.9) ns I: [Bz-i](Emz-i(z-z)+s(z-i)) i=O

in all zones is constant, under the condition that mobilities for negative ions are indicated with a negative sign. ·

Elaboration shows that for all cases the effects of the displacement of the EOF cancel each other. The consequence of this is that for electrolyte systems with an equal zone conductance, equal electric currents must be measured on applying the same electric field gradient, for different velocities of the EOF. To check this, the electric

A.3 INSTRUMENTATION 123

TABLE A.I

MEASURED ELECTRIC CURRENT I (µA), AND MIGRATION TIME, trop (min), AND MOBIUTY OF. THE EOF, mrop· 1<>5 (cm2Ns), FOR SEVERAL ELECTROLYTE SYSTEMS WITH EQUAL CONDUCTIVITY AT DIFFERENT pH APPLYING 2S kV

lc=51 cm, ld=SO cm, pressure injection, 5 s of the. EOF marker 0.001 M mesityl oxide; UV detection at2S4nm.

Buffer solution pH of the mixed Measured I 1EOF mrop solution

0.01 M HCl + EAC 4.70 18.7 7.26 26.17 0.01 M HCl + EAC 5.2S 18.7 5.14 33.10 0.01 M HCI + HIST 6.05 18.8 4.72 40.25 0.01 M HCI + IMID 6.89 18.8 3.74 so.so 0.01 M HCI + TRIS 7.81 18.8 3.06 62.09 0.01 M HCI + DEA 8.80 18.7 2.69 70.63

current and mEoF were measured applying 25 kV, filling the capillary with solutions of equal conductivity at different pH. For the latter 99 ml of a solution of 0.010 M sodium chloroacetate with 1 ml of a buffer (ionic strength also 0.01) were mixed and the pH of the mixture was measured. The results are given in Table A.I.

Although the mEoF varies from about 25· 10"5 to 70· 10·5 cm2Ns the measured electric current is nearly constant, indicating that the conductivity of a zone is independent of the EOF.

lsotachophoretic condition In the steady state all zones move with a velocity equal to that of the leading

zone, therefore:

(A.10)

The contributions of the displacement of the EOF cancel each other.

Calculation procedure As the effect of the displacement of the EOF is cancelled in all equations, the

mathematical model of ITP without EOF can also be used for ITP with EOF. The calculation procedure is described completely previously [5,6].

A.3 INSTRUMENTATION

For all ITP experiments in open systems the P/ ACE System 2000 HPCE was used, applying UV detection .at 254 nm for all cationic and 214 nm for all anionic analyses. All experiments were carried out at 25 °C applying a constant direct current,. using an original Beckman capillary of 57 cm, with a distance between injection and


detection of 50 cm and an I.D. of 75 µm. For further information concerning the apparatus see section 1.5.

For ITP experiments in the cationic mode (CM) the capillary is filled with the leading electrolyte L, while the terminator solution T must be present on the inlet of the apparatus. The sample solution will be introduced between L and T by pressure injection. The detector and cathode are placed at the outlet and the anode at the inlet. For ITP experiments in the anionic mode (AM) the cathode is placed at the inlet side and the anode at the outlet. The capillary is filled with the leading electrolyte and the terminator is present at the inlet.

For all ITP experiments with closed systems, a laboratory-built apparatus, with conductivity and UV detector, as described previously [l] was used. In this apparatus, the closed system is obtained by shielding the separation capillary from the open electrode compartments with semipermeable membranes. Note that a PTFE capillary tube is used (0.2 mm I.D.), in contrast with the fused silica capillaries used in the P/ACE System 2000 HPCE.

A.4 VARIATIONS IN EOF DURING ITP EXPERIMENTS

On applying ITP in closed systems, all ITP zones move with an equal velocity if the steady state is reached. As all zone concentrations are adapted to that of the leading zone, according to the Kohlrausch condition, a linear relationship is obtained between zone length and amount of sample injected. Another approach to quantitative determinations is to use the response factor [8], RF (C/mole), representing the slope of a calibration graph of the product of the· zone length l (s) and applied constant electric current I (A) versus the amount of the sample (mole).

When using ITP in open systems an EOF acts on the ITP system. Generally, the velocity of the EOF changes during ITP analyses because the capillary is filled with different electrolytes during the analyses as the terminator is migrating into the capillary. Thus, the zones do not have the same velocity passing the detector and the detected zone length depends on the velocity of the EOF at the moment of detection. Non-linear calibration curves are the result. In order to. work quantitatively in ITP in open systems the behaviour of the EOF must be understood.

The velocity of the electroosmotic flow To obtain an impression of the velocity of the EOF the capillary can be fill, by

pressure injection, alternately with the chosen electrolyte and a mixture of this electrolyte and a UV-absorbing component, without electrophoretic mobility, such as mesityl oxide (MO). Monitoring the absorbance of the equidistant EOF marker bands in the chosen electrolyte during the ITP experiment gives an idea of changes in the EOF velocity.

As an example to demonstrate the change in EOF during ITP experiments, in Fig. A.3 the electropherograms are given for the migration in the cationic mode of (a) the leading electrolyte 0.01 M NaOH at pH 5.4 adjusted by adding nicotinic acid, (b) the terminating electrolyte of 0.01 M GABA at pH 5 adjusted by adding nicotinic acid and (c) the ITP experiment with this leading electrolyte sodium nicotinate and terminator GABA/nicotinate. The three experiments were carried out at a constant

A.4 VARIATIONS IN EOF DURING ITP EXPERIMENTS

a

0

T+MO I

5

L+MO I

T'+MO T'

L+MO 1 r

T

15

time (min)

125

20

Fig. A.3: Part of the electropherograms obtained by applying a constant electric current of 1.5 µA.for (a) a leading electrolyte 0.01 M sodium nicotinate at pH 5.4, (b) a tenninating electrolyte of 0.01 M GABA!nicotinate at pH 5 and (c) an ITP system with a leading electrolyte L of0.01 M sodium nicotinate at pH 5.4 and terminating electrolyte T of 0.01 M GABA-nicotinate at pH 5. The terminating solution T' is the adapted terminating solution acccrding to Kohlrausch 's law. The electrolyte in the capillary was alternately mixed up with equidistant bands of mesityl oxide (MO) in order 10 indicate variations in the velocity of the EOF.

electric current of 1.5 µA. To visualize the change of the velocity of the EOF, the · capillary was filled alternately with sodium nicotinate or respectively GABA/nicoti.nate

(22 s, pressure injection) and a mixture of sodium nicotinate (or GABA/nicotinate) with 0.001 M mesityl oxide (3 s, pressure injection). From the number of MO peaks over the separation length from inlet to detector the distance between the equidistant peaks can be calculated and using the differences in time between the adjacent peaks · the average velocity at that time can be calculated.

In Fig. A.4 the calculated velocities of the EOF as function of time are given for the situations of Fig. A.3. It can be clearly seen that if the capillary is filled with one electrolyte (a and b) the velocity is constant, although this velocity is much greater with GABA/nicotinate because the voltage gradient is higher at the same electric current and the mobility of the EOF is larger due to the low ionic concentration of the terminating solution at pH 5 (pK of GABA is 4.03).

If an ITP experiment is carried out (Fig. A.4c) in first instance the velocity of the EOF is comparable to that of the original leading electrolyte L. During the analysis the terminator T is migrating into the capillary behind the leading electrolyte and the velocity of the EOF strongly varies with time until a value is reached that is


0.2

b

c (/)

' E .g 0.1

~ >

a

0.0 0 5 10 15

time {min)

Fig. A.4: Calculated velocities of the EOF, using the time intervals between the MO peaks, for the electropherograms in Fig. A.3 as a function of time. The EOF velocity of (c) the ITP system varies between the values of (a) the leading and (b) terminating electrolyte, although not linearly.

comparable to that of the original terminating solution if the whole capillary if filled with that solution T. The velocity of the EOF varies during the ITP experiment, between the values of the pure leading and terminator solutions, this relationship is not linear. Between leading and terminating solution the adapted terminating.solution T' according Kohlrausch' law (Fig. A.3c) can be seen. The last MO peak coincides with the concentration boundary between adapted terminating T' and original terminating T solution.

TABLEA.Il

VOLTAGE, V (kV), MIGRATION TIMES, t (min}, VELOCITIES, v (emfs) AND MOBILITIES OF THE EOF, mi::oF·lo-5 (cm2Ns}, FOR THE LEADING ELECTROLYTE SODIUM NICOTINATE AND SEVERAL TER.\fINATING SOLUTIONS

Electrolyte pH Counter ion v v mEOF

0.01 MNaOH 5.4 nicotinic acid 2.24 80.64 0.010 26.30 O.OlMGABA 5.0 nicotinic acid 23.44 5.23 0.159 38.75 0.01 MGABA 3.5 formic acid 2.37 119.05 0.007 16.84 0.01 MHIST s.o nicotinic acid 3.41 59.52 0.014 23.40 0.01 MHIST 6.7 MES 13.89 6.46 0.129 52.94

In Table A.Il the applied voltages and migration times, velocities and calculated mobilities of the EOF are given for the leading electrolyte sodium nicotinate and

A.4 VARIATIONS IN EOF DURING ITP EXPERIMENTS 127

several terminating solutions applied, with a constant electric current of 1.5 µA. In fact the mobility of the EOF does not differ so much for sodium nicotinate and GABA nicotinate but it must be remembered that the voltage gradients in the ITP system differ considerably between leading and terminator solution so that the velocity of the EOF increases very rapidly if the terminating solution fills the capillary.

From Figs. A.3 and A.4 it can be concluded that the velocity of the EOF changes during ITP experiments because the capillary contains more than one electrolyte. Because it is essential that the EOF is constant during the detection to carry out quantitative ITP, the influence of both the effect of the composition of terminating and sample solutions and the effect of an EOF suppressing additive is considered, in order to choose an optimum ITP system.

The effect of different tenninators The variation in EOF is caused by the presence of more than one solution

migrating in the capillary. To demonstrate what the effect is of different terminator solutions, in Fig. A.5 the electropherograms (CM) are given, applying a leading electrolyte of 0.01 M sodium nicotinate at pH 5.4, alternately mixed up with 0.001 M

T

T

d L :r' 1EJ] 13 1&

T ....

0 13 26 39 52 65

time (rnin)

Fig A.5: Electropherograms for a leading electrolyte sodium nirotinate at pH 5.4 applying as terminating solutions (a) 0.01 M HIST!nirotinate at pH 5, (b) 0.01 M HJSTIMES at pH 6. 7, (c) 0.01 M GABA/formate at pH 3.5 and (d) 0.01 M GABA/nicotinate at pH 5. Electric current 1.5 µA. The enlarged window shows a part of the electropherogram D from 13 to 18 minutes. It can be clearly seen that different terminating solutions in ITP experiments can cause strong variations in the velocity of the EOF (note the varying time intervals between the MO bands) resulting in different times of analysis • .


MO, applying as terminator (a) 0.01 M histidine at pH S adjusted by adding nicotinic acid, (b) 0.01 M histidine at pH 6.7 adjusted by adding MES, (c) 0.01 M GABA at pH 3.5 adjusted by adding formic acid and (d) 0.01 M GABA at pH 5 adjusted by adding nicotinic acid. In Fig. A.6 all calculated velocities of the EOF as a function in time are given.

It can be clearly seen that by applying (a) a terminating solution of histidine (pK=6.04) with relatively high mobility at a pH at which a high ionic concentration is present (pH< pK) a nearly constant velocity of the EOF can be obtained during the ITP experiment (see also Fig. A.6a). For (b) the same terminator at low ionic concentration (pH> pK) the voltage gradient in the original solution will be very high causing a high end-velocity of the EOF and by this a strong variation in the velocity of the EOF during the analysis (see also Fig. A.6b). On carrying out similar experiments with the terminator GABA (pK=4.03) at pH respectively (c) 3.5 (pH <pK) and (d) 5 (pH>pK) similar effects can be obtained as with histidine as the terminator. On applying the terminator GABA/formate at pH 3.5, a step in the T' zone could be observed. Repeating this experiment applying GABA/nicotinate at pH 3.5 also gave a do1,1ble T' zone. Although not understandable, it might be a moving pH boundary between the adapted T' zone and the original terminating solution T as suggested by Hirokawa et al. [9].

0.2 ~---------------~

(i) d

e .g 0.1

~ >

a

0.0 0 10 20 30 40 50 60

time (min)

Fig. A.6: Calculated velocities of the EOF, using the time intervals between the MO peaks, for the electropherograms in Fig.A.Sas a function of time. The (b,d) EOF velocity of the 1TP system varies strongly if the pH of the terminator is higher than the pK value of the terminator (low ionic concentrations and thus a high EOF) compared with (a,c) terminating solutions at a pH <pK.

From Fig. A.5 clearly can be concluded that using terminators with a low effective mobility and/or at low ionic concentration strongly varying values of the EOF are the result. It should be remembered that sample ionic species will be found between the L en T' zones, and that EOF velocities at that moment determine whether reproducible quantitative determinations are possible. The time of analysis also varies strongly depending on the terminator applied.

A.4 VARIATIONS IN EOF DURING ITP EXPERIMENTS 129

The effect of sample solutions To demonstrate the effect of sample solutions on the velocity of the EOF during

ITP experiments, the leading electrolyte 0.01 M sodium nicotinate at pH 5.4 and the terminating electrolyte 0.01 M HIST/nicotinate at pH 5 were used, for which a fairly constant EOF can be expected.

5 II

10 s

15 s Q)

20 II 0 c n1 30 s

..0

~ 40 s

..0 n1

T'

18 24 30

time {min)

Fig. A.7: Electropherograms with a leading electrolyte (L) of0.01 M sodium nicotinate at pH 5.4, a terminator of O. 01 M HIST!nicotinate at pH 5 for several pressure injection times of a solution of O. 0025 M lithium nitrate (Li). Increasing zone lengths of the introduce,d sample resuhs in decreasing times of analysis due to an increasing velocity of the EOF. Electric current, 1.5 µA.

In Fig. A. 7 the electropherograms (CM), injecting a sample of 0.0025 M lithium nitrate for increasing pressure injection times are given. It is clear that the injection of large sample zones of diluted solutions causes strongly increasing EOF velocities, by which the migration times decrease substantially and quantitative analysis are erroneous. This effect will be much greater using sample solutions at a lower ionic strength. It should be remembered that the velocity of the EOF increases strongly because the mobility of the EOF is larger at low ionic strength and the voltage gradient over the original diluted sample zone is much larger.

Suppressed EDF From the foregoing, it can be concluded that the velocity of the EOF strongly

varies with composition of terminating and sample solutions. Varying EOF velocities cause irreproducible migration times and zone lengths and hence quantitative determinations are erroneous. To.work quantitatively with ITP in open systems in an appropriate way, the velocity of the EOF must be controlled or eliminated.

It is well-known that the EOF can be suppressed by adding surface-active substances to the electrolyte system, such as methylhydroxyethylcellulose (MHEC). A problem is that if MHEC is used as EOF suppressor, the MO peaks in the leading electrolyte cannot be used as EOF indicator as they no longer move. However, the migration times for a leading electrolyte in ITP should be constant, independent

130

(J) (.) c .2 .... 0 (/)

~

a

b

c

d

0

APPENDIX: ITP IN OPEN CAPILLARIES

L

T' L

T'

T'

L

T

L

10 20 30

time (min)

Fig. A.8: Electropherograms for a leading electrolyte sodium nicotinate at pH 5.4 applying as terminating solutions (a) 0.01 M HIST/nicotinate at pH 5, (b) 0.01 M HIST/MES at pH 6. 7, (c) 0.01 M GABA/formate at pH 1.5 and (d) 0.01 M GABA/nlcotinateat pH 5. Toallsolutions0.05 % MFmCwas added. It can be clearly seen that, compared with the electropherograms from Fig. A.5, different terminating solutions in 1TP experiments with suppressed EOF show a fairly constant velocity of the EOF resulting in reasonably constant times of analyses. Electric current, 1 µA.

A B

25 25

20 20

5 5 <!< 15 <!< 15

J 10 J 10 a g g Q a 5 a

b

0 0 0 13 25 39 52 IS! 0 13 26 39 52 ""

time !mil'll tim<r> (mjl'l)

Fig. A.9: Relationship between voltage drop over the capillary tube and time of analysis (A) for the electrolyte systems used In Fig. A.5, without MHEC, and (B) for the same electrolyte systems with MHEC as used in Fig. A.8. All data are recalculated to an electric current of J.5 µA. With the additive MHEC linear relationships are obtained in all instances.

A.5 EXAMPLES OF THE VARIOUS MODES 131

of the choice of the terminator. In Fig. A.8 the electropherograms for the same electrolyte systems as used in Fig. A.5 are given, but 0.05 % MHEC was added to all solutions. Although the migration times are not "identical" they are fairly equal, indicating that the EOF is nearly suppressed, i.e., the velocity of the EOF must be constant resulting in, however, longer times of analysis (Note that only the adapted T' zones are detected).

Another possibility to check the effect of MHEC on the EOF is the relationship between the voltage over the capillary and time of analysis. These relationships are given in Fig. A.9A for the electrolyte systems used for Fig. A.5 (different terminators without MHEC) and in Fig. A.9B for the electrolytes systems used for Fig. A.8 (the same terminators with MHEC). It can be clearly seen that with the addition of MHEC (Fig. A.9B) linear relationships are obtained in contrast to those obtained without MHEC (Fig. A.9A), especially for the cases b and d. The decrease in case c (Fig. A.9A) corresponds to the migration of the unadapted terminating solution into the capillary.

A.5 EXAMPLES OF THE VARIOUS MODES OF ITP IN OPEN SYSTEMS

Cationic mode As an example of the cationic mode in Fig .. A. I 0 the isotachopherogram is given

for the separation of sodium and lithium in a system with the leading electrolyte 0.01 M KOH at a pH 5 by adding nicotinic acid. The terminator was 0.01 M histidine at pH 5.5 by adding nicotinic acid. Using a UV absorbing counter ion, nicotinic acid,

0.05

L

5 0.00 1

~ 2 T'

!ll -0.05

~ -w ..0 -0.10 i ~ 1 2

.... T 0

(J) .0 111 -0.15 -- T' --UMUUUUM«Wl ·- 3

-0.20 0 5 10 15 20 25 30

t (min)

Fig. A.JO: lsotachopherogramfor the separation of (1) sodium and (2) lithium (pressure injection time 5 s; concentrations of the sample ions 0. 005 M) using a leading electrolyte L of potassium nicotinate at pH 5 and terminator histidine nicotlnate at pH S.5. Position (J) indicates the concentration boundary between (I") adapted and (T) originalterminating solution, which moves under the il'/fluence of the EOF. Electric current, 4 µA. UV detection at 254 nm. The sample zones can be clearly seen in the enlarged inset.


the different zones show different UV absorbances because the concentrations in the sample zones are different. The leading zone L, the sample zones of (1) sodium and (2) lithium, the (T') adapted terminating zone ofhistidine and (T) the original histidine solution can be clearly seen. The (3) UV dip indicates the original sampling spot and determines the concentration boundary. Cationic ITP experiments can easily be carried out, although the separation power is diminished compared with ITP in closed systems.

Reversed anionic mode The requirement in order to obtain a RAM is that Er I mEoF,T I > Er I mT I .

In that case the net migration velocity will be Er I mEoF,T I - Er I mT I , in first instance. Finally the net velocity will be EL I mEoF,L I - EL I mL I and if this net velocity is larger than zero this mode can be used for separations.

0.20 .------------..,-1------. r

proP ac

0.10

Na• T 0.00 ~-- !

EOF

-0.10 '--~~~~_.___.___.___.___.___.___.__.__.__._~

0 5 10 15

t (min)

Fig. A.11: Isotachopherogram for a separation in the reversed anionic mode. Leading electrolyte 0. 0025 M histidine acetate at pH 6. 4. Terminator MES at pH 6 adjusted by adding histidine. Pressure injection 10 s of0.0025 M sodium propionate. Electric current 4 µA. Wavelength UV detector 214 nm.

As an example of the reversed anionic mode, the isotachopherogram using as leading electrolyte 0.0025 M acetic acid at pHL of 6.4 by adding histidine and the terminator 0.0025 M MES at pH 6 by adding histidine is given in Fig. A.11. As sample a solution of 0.0025 M sodium propionate (pressure injection 10 s) was introduced. In Fig. A.12 the original position and (a) the zones formed after some time in a closed system and (b) the situation expected in open systems is shown. The zones will further be detected in the reverse order. This situation can be recognized in Fig. A.11 where the order of detection is the sodium dip of the sample solution in the original T zone, the transported concentration boundary between original terminating solution and adapted T' zone (this transition can be marked with an EOF marker) and finally the propanoate zone and leading acetate zone. In practice, the T'' zone, adapted to the original sample solution, will generally be short and masked by e.g., the UV water dip or EOF marker.

A.5 EXAMPLES OF THE VARIOUS MODES 133

INLET OUTLET

original position detector

+ L Na T prop

a

+ T

b

+ L T Na T -

Fig. A.12: Original position of the different electrolytes for the reversed anionic mode for ITP • The formed zones during the experiment are shown/or lTP (a) without and (b) with EOF. Note that the separation proceeds into the direction of the anode. Without EOF the concentration bcundaries T'IT" and T"IT can never be detected, but as a result of the EOF they can reach the detector.

Anionic mode The greatest problems arise in finding a good ITP system for the anionic mode.

At pH 4 (low VsoF) a migrating anionic mode could be carried out, using chloride as leading ion and formate, an ion with a high mobility as terminating ion. In most instances, however, the voltage drop increased until a certain value far below that theoretically expected, indicating that the ITP system moved in the anionic mode but after a certain moment no longer moved. How can this be explained?

TABLE A.ill

CALCULATED pH AND SZR25 (Om) FOR SOME TERMINATING ZONES IN ELECTROLYTE SYSTEMS AT A pHL OF 6 AND 4 AND MEASURED m80p.la5 (cm2Ns) AND VOLTAGE DROP (kV) AT 15 pA FOR THESE SOLUTIONS

Electrolyte pH SZR25 mEOF Voltage drop at 15 µA

0.01 M HCI + histidine 6.00 10.33 39.6 17.86 0.00654 M MES 6.43 42.99 50.94 73.53 0.025 M HCl + EAC 4.00 4.33 18.54 7.75 0.0225 M formate 4.29 7.62 24.4 12.6 0.0200 M acetate 4.69 16.47 35.85 25.33

In Table A.III the calculated pH, SZR25 and concentrations are given for MES as terminator using the leading electrolyte 0.01 M HCl at a pH of 6 by adding


histidine and for formate and acetate zones for the leading electrolyte 0.025 M HCl at pH 4 by adding EAC. All these solutions were carefully prepared and the ~OF and voltage drop for these solutions were measured and recalculated to an electric current of 15 µA.

20

15

5 ~ 10

> 5

0 0 5 10 15 20 25

t (min)

Fig. A.13: Measured voltage drop as a function of time during ITP experiments with EOF in the anionic mode using a leading of 0.025 M HCI at pH 4 by adding EAC and terminating solutions of 0.01 M sodium formate and 0.01 M sodium acetate respectively. Electric current, 15 µA.

In Fig. A.13 the measured voltage drop as a function of time is given for acetate and formate as terminating solution in the system Q.025 M HCl and EAC at pH 4. It can be seen that forformate the expected voltage drop of 12.6 kV is fairly reached and in the isotachopherogram the formate step could be seen. For acetate· the expected voltage drop of about 25.33 kV was not reached, but the voltage drop stopped at about 15 kV, which means that the ITP system came to a standstill.

The explanation is as follows. In first instance the net velocity of the ITP system is determined by EL I mL I - EL I mEoF,L I . Finally the net velocity is Er· I mr I - Er I mEoF,T I . According to the ITP condition ELmL = ErmT. Generally, Er I mEoF,T I will be much larger than EL I mEoF,L I because of the higher voltage gradient and generally larger mEoF (see Table A.ill). This means that the net velocity of the ITP system will get zero during the experiment and the system will not migrate from that time ..

A.6 PROBLEMS IN QUANTITATIVE ANALYSIS

Although the UV detector of the P/ ACE System 2000 HPCE is not a universal detector, it can be used reasonably in a universal way in ITP applying a UV absorbing counter ionic species because the concentration of the sample ions differ considerably according to Kohlrausch 's law. Owing to the electroneutrality equation the concentration of the buffering counter ions varies from zone to zone, causing different

A.6 PROBLEMS IN QUANTITATIVE ANALYSIS 135

UV absorbances. The same principle can be used in CZE applying a UV absorbing carrier ion [10,11].

In Fig. A.14 the electropherograms (cationic mode) are given obtained using a laboratory-made ITP apparatus with (a) conductivity and (b) UV detector and (c) using the P/ ACE System 2000 HPCE with UV detector. The leading electrolyte was 0.01 M sodium nicotinate at pH 5.4 and the terminating electrolyte was 0.01 M GABA/nicotinate at pH 5. These electropherograms show clearly that generally the UV detector is applicable for non-UV-absorbing components. Note the similarity between both UV signals.

Ill

8 c c: j!l < "' Ill .iii O' f "' 1 0 u L

, ·c g t) ::I

n (!) (I) Qi

i fr

~time

Fig. A.14: Electropherograms obtained with (a) a conductivity detector and (b) UV detector with a laboratory-made ITP apparatus with closed system and (c) UV detector of the PIACE System 2000 HPCE apparatus. The leading electrolyte was 0.01 M sodium nicotinate at pH 5.4 and the terminator was 0.01 M GABA/nicotinate at pH 5. The electric current was 1.5 p.Afor the PIACE and 5.5 p.Afor the laboratory-made ITP instrument. Note the similarity between both UV signals. The sample composition was 0.00125 M of (1) lithium, (2) GirP, (3) TRJS, (4) HIST, (5) CREAT, (6) OPDA and (7) EAC. Injection 1 p.lfor the laboratory-made lTP instnmumJ and 5 s pressure injection/or the PIACE System 2000 HPCE.

Effect of suppressed EOF in quandtative 1TP In order to study the effect of suppressing EOF in ITP in open systems, the zone

lengths versus the samr le amount by varying the pressure injection time for a concentration of 2.5· 10- and l · 10·2 M of histidine with the leading solution 0.01 M sodium nicotinate at a pH 5.4 using the terminators 0.01 M GABA/formate and 0.01 M GABA/nicotinate at pH resp. 3.5 and 5, without and with the addition of 0.05 % MHEC to all electrolyte solutions were measured [12]. All experiments were carried out with an electric current of 1.5 µ.A.

In Fig. A.15A all measured zone lengths as a function of the sample amount (in Ms units, i.e., molarity multiplied by injection time in seconds) are given without the addition of MHEC. The zone lengths using the terminator GABA-nicotinate at pH of 5 are much shorter than those of GABA/formate at pH 3.5 owing to its higher EOF


A B

191.5 19!1

:'!! $ ii ~ I 130 130 .!!

~ ~ 65 65

0. 1 ().2 o.3 ().4 o.s 0.1 0.2 0.3 0.4 0.5

a.niou-it injected !Msl

Fig. A.15: Relationship between measured 1.0ne length (s) and injected amount histidine (Ms) using the tenninator 0.01 M GABA at a pH 3.5 by adding Jonnie acid and a sample of {D) 0.01 Mand (0) 0.0025 M histidine and the tenninator 0.01 M GABA at pH 5.0 by adding nlcotlnlc acid with a sample of (.11.) 0.01 Mand(+) 0.0025 M hlstidinefor (A) without and (B) with the addition o/0.05 % MHEC to all electrolyte solutions.

velocity (effect of the low ionic concentration in the original terminating solution). For both terminating solutions the zone lengths decrease extensively for the low concentration of histidine applying longer injection times, due to the sample solution effect on the EOF velocity. It can be concluded that the influence of the composition of both the terminating solutions and the sample solutions on the EOF velocity makes quantitative determinations difficult.

In order to compare the effect of the EOF versus the suppressed EOF in quantitative ITP all experiments were repeated, after adding 0.05 % MHEC to all electrolyte solutions. In Fig. A.15B all measured zone lengths as a function of the sample amount are given for these experiments.

A comparison between Fig. A.15A and A.15B shows that the zone lengths for l · 10·2 M histidine for both terminators with MHEC are much longer than those for the systems without MHEC. The zone lengths for 2 .5 • 1 o·3 M histidine are longer than those for the systems without MHEC, although for GABA at pH 5 not quite a linear relationship could be obtained, identical with those for the l • 10-2 M solutions. From Figs. A.15A and B it can be concluded that although the effect of the composition of terminating solutions on the EOF by the addition of MHEC can be suppressed for the


greater part, the influence of the ionic strength of the sample solution can not be eliminated completely.

Reproducibility in quantitative analysis In order to study the reproducibility of quantitative ITP in open systems, a

calibration curve (CM) of lithium, TRIS, HIST and EAC was determined five times, in the system sodium nicotinate at pH 5.4, applying three different terminators viz., {A) GABA/nicotinate at pH of 5, (B) GABA/formate at pH 3.5 and (C) acetic~acid at pH 3, injecting six different amounts of the mixture of the sample components by pressure injection times of 5, 10, 15, 20, 30 and 40 s. To all solutions 0.05 % MHEC was added. In Table A.IV the data for the calibration curves are given.

Although all calibration curves were fairly linear (the average regression coefficient is 0.9989) the absolute values of the slopes differ up to about 12 % , whereas large differences exist for the systems with different terminator. If the slope was calculated relative to that of histidine the differences are much smaller. For example, the first two values of the slope of the calibration curve of lithium in system A are 870.94 and 934.11 while the relative slopes are 1.263 and 1.264.

Two important conclusions can be drawn from Table A.IV. First, by applying ITP in open systems absolute values of zone lengths can never be handled and it is essential to work with at least an internal standard. An extra advantage is that, by applying an internal standard, the fluctuation in injected volume is eliminated. Second, the use of different terminators can give totally different relative slopes although reproducible with time. For further experiments as terminator acetic acid at pH 3 was used.

Quantitative analysis with ITP in open capilla.ries In order to study the usefulness of the relative slopes in quantitative ITP the

relative slopes of several ionic species in both a cationic and anionic ITP system were determined. Further, the determined relative slopes in open systems were compared with those in closed systems and theoretical values.

In Table A. V all values are given for. cationic species determined with a leading electrolyte of 0.01 M KOH at pH 5.4 adjusted by adding nicotinic acid and the terminator acetic acid at pH 3 both with 0.05 % MHEC. MHEC was also added to all sample solutions. The calibration curves were measured by injecting five different amounts of the sample components by pressure injection times of 10, 20, 30, 40 and 50 s, and an applied direct current 3 µA.

From Table A. V it can be concluded that ITP in open systems can be applied in · a reproducible way for cations, although special care must be taken in the choice of the electrolyte system. Also numerous experiments with the terminators GABA at pH 5 and 3.5 were carried out and with these terminators the relative slopes varied much more in time, although the regression coefficients were good.

More problems with the reproducibility were encountered on applying an anionic system, because the systems in the AM seem to be much more sensitive for fluctuations in EOF.

In Table A. VI all values are given for anionic species with the leading electrolyte 0.01 M HCI at pH 6 adjusted by adding histidine and the terminator 0.01689 M HIST at pH 6.43 adjusted by adding MES, both containing 0.05 % MHEC. With the


TABLE A.IV

SLOPES OF CALIBRATION GRAPHS FOR LITHIUM, TRIS, HIST AND EAC AND AVERAGE SLOPE AND STANDARD DEVIATION OF THESE VALUES AND THOSE FOR SLOPES RELATIVE TO HISTIDINE.

Leading electrolyte 0.01 M Sodium nicotinate at pH 5.4, terminating electrolyte (A) 0.01 M GABA/nicotinate pH 5.0, (B) 0.01 M GABA/formate pH 3.5, (C) acetic acid pH 3. To all electrolytes 0.05 % MHEC was added. Constant current 1.5 µ.A. All calibration curves were determined by pressure injection of 5, 10, 15, 20, 30 and 40 s of the sample solution (all ionic concentrations were 0.002 M).

Slope (arbitrary units) Relative slope

A B c A B c

Lithium 870.94 442.50 464.03 1.263 0.833 0.861 934.11 435.44 477.18 1.264 0.854 0.850 911.06 458.32 454.26 1.279 0.839 0.846 946.03 461.53 459.68 1.360 0.860 0.872 876.85 456.56 452.03 1.320 0.860 0.854

average value 907.80 450.87 461.44 1.297 0.849 0.857 standard deviation 29.90 10.10 8.92 0.038 0.011 0.009

TRIS 730.82 509.18 530.38 1.060 0.959 0.984 789.47 499.03 544.62 1.068 0.979 0.970 754.09 525.41 507.03 1.059 0.962 0.944 769.82 501.53 491.59 1.106 0.934 0.932 707.29 518.09 505.50 1.065 0.975 0.955


HIST 689.47 531.06 538.82 1.000 1.000 1.000 739.18 509.79 561.32 1.000 1.000 1.000 712.38 546.12 537.24 1.000 1.000 1.000 695.79 536.82 527.41 1.000 1.000 1.000 664.12 531.15 529.24 1.000 1.000 1.000

average value 700.19 530.99 538.81 1.000 I.OOO 1.000 standard deviation 24.92 11.93 . 12.09

EAC 678.79 540.2.1 563.38 0.985 1.017 1.046 723.06 505.94 577.24 0.978 0.992 1.028 690.62 569.32 574.59 0.969 1.042 1.070 681.91 540.44 550.06 0.980 1.007 1.043 657.74 561.35 551.68 0.990 1.057 1.042



TABLEA.V

THEORETICAL VALUES OF SLOPES AND EXPERIMENTALLY DETERMINED VALUES FOR ITP IN CLOSED SYSTEMS AND OPEN SYSTEMS (IN DUPLICATE) FOR SEVERAL CATIONIC COMPONENTS IN COMPLEX MIXTURES RELATIVE TO HISTIDINE.

Leading electrolyte: 0.01 M KOH at pH 5.4 adjusted by adding nicotinic acid, terminator : acetic acid at pH 3, both with 0.05 % MHEC; constant current 3 µA; wavelenght UV detector, 254 nm.

Sample THE OR CLOSED OPEN (1) OPEN (2)

barium 1.377 1.321 1.349 1.315 CREAT 0.888 0.888 0.897 0.881 EAC 1.084 I.066 1.025 1.006 GABA 1.135 1.000 1.019 1.039 HIST 1.000 1.000 I.OOO 1.000 lithium 0.836 0.863 0.866 0.855 OPDA 0.940 0.906 0.922 sodium 0.740 0.713 0.685 0.643 TRIS 1.005 0.993 0.986 0.976

TABLE A.VI

THEORETICAL VALUES OF SLOPES AND EXPERIMENTALLY DETERMINED VALUES FOR ITP IN CLOSED SYSTEMS AND OPEN SYSTEMS (IN DUPLICATE) FOR SEVERAL ANIONIC COMPONENTS (A) MIXED WITH THE STANDARD AND (B) IN COMPLEX MIXTURES AND (C) IN COMPLEX MIXTURES WITH MHEC RELATIVE TO FORMATE.

Leading electrolyte 0.01 M HCI at pH 6.0 adjusted by adding histidine + 0.05 % MHEC, Terminating electrolyte 0.01689 M HIST at pH 6.43 adjusted by adding MES + 0.05 % MHEC. Constant current 3 µA, wavelength UV detector 214 nm.

Sample Theoretical Closed Open

A A B B c c

acetate 1.12 1.16 1.16 1.08 1.06 1.06 0.97 1.01 adipate 2.09 2.06 1.94 1.98 2.16 2.16 2.20 2.o7 benma.te 1.25 1.28 1.34 1.35 1.37 1.49 1.28 1.28 benzylaspartate 1.44 1.58 J.77 1.78 2.03 2.29 1.79 1.71 chlorate 0.95 0.99 0.99 1.01 1.15 1.19 0.99 0.99 enanthate 1.37 1.41 1.40 1.54 1.42 1.68 1.47 l.55 formate 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 malonate 1.80 1.80 1.82 1.89 1.90 2.02 1.88 1.54 propionate 1.19 1.21 1.26 1.29 1.09 0.98 1.14 1.18

addition of MHEC the anionic species could be measured in the normal AM. It is already discussed previously [13] that without the addition of MHEC most anionic systems do not move in the anionic mode due to the counteraction of the EOF.

To illustrate the importance of suppression of the EOF caused by the sample composition, in Table A. VI the theoretical relative slopes, the experimentally


determined relative slopes in closed systems, and those for open systems (in duplicate) for (A) an anion mixed up with the standard without MHEC, (B) mixtures of several anions without MHEC and (C) mixtures of anions with MHEC are given. The calibration graphs were measured by injecting five different amounts of the sample components by pressure injection times of 10, 20, 30, 40 and 50 s.

From Table A. VI it can be concluded that most values for the closed system fit the theoretical values reasonably well. The measured values for anionic components mixed with the standard formate (A) are not too bad, although that for benzylaspartate is too high. The measured values for mixtures of several anionic components (B) are often too high. It must be remembered that although 0.05 % MHEC is added to leading and terminating electrolyte in order to suppress the EOF, no MHEC was added to the sample mixtures. For small injected amounts it can be expected that the effect of the sample to the EOF will be minor effect but especially long injection times with diluted samples can cause serious variations in the EOF. Because the EOF is oppositely directed to the ITP migration velocity longer sample zones can be expected. The measured values of mixtures with MHEC (C) are better.

A. 7 CONCLUSIONS

The mathematical model for ITP without EOF (closed systems) is also valid for ITP with EOF (open systems). Four modes can be distinguished for ITP with EOF. The cationic mode can be applied without problems although the available separation length is not used optimally, compared with closed systems, due to the extra velocity of the EOF. Some precautions must be taken in the choice of the terminator to avoid strong fluctuations in the net migration velocity of the ITP system. The reversed cationic mode is practically of no value and can only be applied using additives or coated capillaries in order to obtain a strong reversed EOF. The reversed anionic mode can be used at high pH (with a high vEoF) and by the "counter current" of the EOF, this mode can be of interest in some cases. The anionic mode is only of interest if the vEOF is low, as e.g., in the use of non-aqueous solvents. Otherwise problems can be expected concerning the choice of the terminator solution. Too high E gradients in the terminator solution can lead to a standstill of the ITP system during the experiments.

A first comparison between ITP in open and closed systems, shows that ITP in closed systems is preferable, because no special precautions have to be taken in the choice of the electrolyte systems concerning the EOF. Unless the EOF can be strongly reduced, problems can be expected applying ITP in open systems.

The velocity of the EOF changes continuously if the capillary contains more than one electrolyte. Terminating and sample ions with low mobility, especially at a low ionic concentration, accelerate the EOF in an extensive way, through which quantitative analyses are without any sense. The addition of MHEC to the electrolyte solutions suppresses the EOF for the greater part, by which linear relationships between sample amounts and zone lengths can be obtained. In spite of the addition of MHEC, the reproducibility of the zone lengths in time is poor and depends on the "state" of the capillary and an internal standard is necessary for quantitative determinations. Different terminating electrolytes can cause differences in relative slopes.

A. 7 CONCLUSIONS 141

It can be state.cl that quantitative ITP in open systems is only possible if precautions are taken to suppress the EOF in an effective way. The addition of, e.g .• MHEC to also the sample solution is very important, especially in the AM. In the AM the reproducibility seems to be more troublesome. The use of other EOF suppressing agents can probably give better results. The presence of an EOF, however, facilitates hyphenated techniques, making the use of a "make-up" stream superfluous.

References

1. P.M. Everaerts, J.L. Beckers and Th.P.E.M. Verheggen, Isotachophoresis, Theory, Instrumentation and Applications (Journal of Chromatography Library, Vol. 6) Elsevier, Amsterdam, 1976.

2. S. Hjerten, K. Elenbring, P. Kilar, J.L. Liao, A.J. Chen, C.I. Siebert and M.D. Zhu, J. Chromatogr., 403 (1987) 47.

3. H.R. Udseth, J.A. Loo and R.D. Smith, Anal. Chem., 61 (1989) 228. 4. W. Thormann, J. Chromatogr., 516 (1990) 211. 5. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 480 (1989) 69. 6. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 508 (1990) 3. 7. A. Tiselius, Nova Acta Reg. Soc. Sci. Ups., Ser. 4, (1930) 4. 8. J.L. Beckers and F.M. Everaerts, J. Chromatogr., 470 (1989) 277. 9. T. Hirokawa, K. Nakahara and Y. Kiso, J. Chromatogr., 463 (1989) 51. 10. J.L. Beckers, P.M. Everaerts and M.T. Ackennans, J. Chromatogr., 537 (1991) 407. 11. M.T. Ackermans, F.M. Everaerts and I.L. Beckers, J. Chromatogr., 549 (1991) 345. 12. M.T. Ackennans, F.M. Everaerts and J.L. Beckers, J. Chromatogr., 545 (1991) 283. 13. J.L. Beckers, F.M. Everaerts and M.T. Ackermans, J. Chromatogr., 537 (1991) 429.

142

SUl\IMARY

This thesis reports on a study of some fundamental aspects of the separation process in capillary electrophoresis in open systems with UV detection. The applicabilty of this technique is considered with the help of pharmaceuticals, which are used in veterinary practice.

If capillary electrophoresis is used for the determination of components in complex matrices, important aspects which come up for discussion are (1) the separation of the component from the matrix and the identification of this component and (2) the quantitative determination, including the linearity of the calibration graph, the reproducibility and the limit of detection. If no good separation can be achieved or if a reliable quantification cannot be obtained, a sample pretreatment must be studied.

The group of drugs, in which the Dutch Institute for Quality Control of Agricultural Products is interested, can roughly be divided according to charge, viz, positive, negative or uncharged, and UV absorbing properties. To achieve a separation of these components, different modes of capillary electrophoresis must be used. Charged components with and without UV absorbing properties can be separated with capillary zone electrophoresis with UV or indirect-UV detection. To separate neutral components, micellar electrokinetic capillary chromatography can be used, naturally only in the UV mode.

In capillary zone electrophoresis, the effective electrophoretic mobility can be used as a parameter for screening, in contrast to the migration times or apparent mobilities, which can fluctuate considerably if the electroosmotic flow varies. In micellar electrokinetic capillary chromatography for neutral components, a pseudoeffective mobility can be defined. This pseudo-effective mobility is the electrophoretic mobility, which the components have in the period that they carry an apparent charge, i.e., in the time that they are solubilized in the charged micelles. Just as the effective mobility in capillary zone. electrophoresis, the pseudo-effective.mobility in micellar electrokinetic capillary chromatography can be used as a parameter for screening.

To investigate the quantitative applicability of capillary electrophoresis for the determination of drugs in several matrices, some examples were worked out. UV absorbing positively charged 132-agonists in several pharmaceutical dosage forms were determined with capillary zone electrophoresis and the results were compared with those obtained with isotachophoresis and high-performance liquid chromatography.

144 SUMMARY

Negatively charged UV absorbing sulfonamides were determined with capillary zone electrophoresis in pork meat extracts and some neutral UV absorbing drugs were determined with micellar electrokinetic capillary chromatography. Finally, some aminoglycoside antibiotics, which do not have UV absorbing properties, were determined using capillary zone electrophoresis with indirect UV detection. These antibiotics are often used together with neutral components in combined pharmaceuticals. Therefore the simultaneous determination of aminoglycosides and neutral components with hyphenated capillary zone electrophoresis with· indirect UV detection and micellar electrokinetic capillary chromatography is studied. The method is demonstrated using Otosporin eardrops, containing both an aminoglycoside and a neutral component.

Quantitative analyses showed that the limit of detection for most components is ea. 1·10"5 M. To lower this detection limit several sample concentration techniques can be used. The use of isotachophoresis as on-line sample preconcentration method is obvious. As a primer to the on-line coupling of isotachophoresis to capillary zone electrophoresis, the possibilities of isotachophoresis in open capillaries is studied, both qualitative and quantitative.

Besides these applications, the determination of separation numbers and some physico-chemical parameters, such as mobilities at infinite dilution and pK values, with capillary zone electrophoresis are taken into account.

SAMENV A'rflNG

In bet onderzoek dat aan dit proefschrift ten grondslag ligt, zijn een aantal fundamentele aspecten van bet scbeidingsproces bij capillaire electroforese in open systemen met UV detectie bestudeerd. Samenhangende met de vraag van bet Rijksinstituut voor Kwaliteitscontrole van Land- en Tuinbouwproducten (RIKIL1) of capillaire zone electroforese gebruikt kan worden als screeningsrnethode, is tevens de toepasbaarheid van deze analytische scheidingsmethode bekeken. Daarom loopt de bepaling van geneesmiddelen die toegepast worden in de veterinaire praktijk als een rode draad door dit proefschrift.

Als men capillaire electrof orese toepast bij de bepaling van componenten in een complexe matrix zijn belangrijke aspecten die aan de orde dienen te komen (1) de scheiding van de component uit de matrix en identificatie van deze component en (2) de kwantitatieve bepaling hetgeen omvat lineariteit van de ijklijn, reproduceerbaarheid en detectie limieten. Als geen goede scheiding c.q. betrouwbare kwantificering verkregen kan worden dient eventueel een monsteropwerkingsprocedure in ogenschouw genomen te worden.

De groep drugs, waarin het RIKILT interesse heeft, kan in grote lijnen onderscheiden worden naar lading, d. w.z. positief, negatief of ongeladen, en naar UV absorberende eigenschappen, hetgeen impliceert dat verschillende uitvoeringsvormen van capillaire electroforese gehanteerd dienen te worden om tot een scheiding te kunnen komen. Zo kunnen positieve en negatieve componenten, zowel metals zonder UV absorberende eigenschappen met capillaire zone electroforese in de directe resp. indirecte UV mode gemeten worden. Voor de scheiding van neutrale cornponenten kan . micellaire electrokinetiscbe capillaire chromatografie gebanteerd worden, uiteraard alleen in de UV mode.

Uit het onderzoek is naar voren gekomen dat de effectieve electroforetische rnobiliteit zinvol gehanteerd kan worden als parameter ter piekherkenning bij capillaire zone electroforese, dit in tegenstelling tot de migratietijden die sterk kunnen fluctueren bij een variabele electroosmotische flow. Bij micellaire electrokinetische capillaire chromatografie kunnen pseudo-effectieve mobiliteiten worden gedefinieerd voor neutrale componenten, die uiteraard geen effectieve electroforetische mobiliteit bezitten. Deze pseudo-effectieve mobiliteit is de electroforetische mobiliteit die de componenten hebben in de periode dat ze een schijnbare lading bezitten, dus in de tijd dat ze in de micellen zitten. Net als de effectieve electroforetische mobiliteit in

146 SAMENV AlTING

capillaire zone electroforese kan de pseudo-eff ectieve mobiliteit in micellaire electrokinetische capillaire chromatografie gebruikt worden voor de piekherkenning.

Orn de toepasbaarheid van capillaire electroforese met betrekking tot kwantitatieve aspecten te bekijken van drugs in de verschillende ·matrices werden een viertal applicaties uitgewerkt. De UV absorberende positieve '32-agonisten werden met capillaire zone electroforese, isotachoforese en hoge druk vloeistof chromatografi.e bepaald in verschillende farmaceutische preparaten en.de resultaten werden vergeleken. Negatieve sulfanomiden met UV absorberende eigenschappen werden bepaald met capillaire zone electroforese in vleesextracten en verschillende neutrale componenten met micellaire electrokinetische capillaire chromatografie waarbij bet gehalte van dapson in tabletten werd bepaald. Tenslotte werden enkele aminoglycosiden, die geen UV absorberen en positief geladen zijn, bepaald met capillaire zone electroforese met indirect UV detectie. Omdat aminoglycosiden veel voorkomen in combinatie preparaten, waarin ze gecombineerd zijn met een ongeladen geneesmiddel, werd tevens de simultane bepaling van de aminoglycosiden en ongeladen componenten met gekoppelde capillaire zone electroforese met indirect UV detectie en micellaire electrokinetische capillaire chromatografie bekeken, en werd de toepasbaarheid van deze analyse gedemonstreerd aan de hand van Otosporin oordruppels.

Bij de kwantitatieve analysen blijkt de detectiegrens van de componenten te liggen in de orde van grootte van ea. 10-s M. Orn deze grens te verlagen kunnen verschillende concentreringstechnieken gehanteerd worden. Het gebruik van lTP als on-line concentreringsmethode dringt zich hierbij op. In eerste instantie is het gebruik van ITP in open capillairen, zowel kwalitatief als kwantitatief, bestudeerd.

Naast de bepaling van geneesmiddelen in diverse matrices is nog gekeken naar de bepaling van scheidingsgetallen en van verschillende fysico-chemische parameters, zoals mobiliteiten bij oneindige verdunning en pK waarden uit de effectieve mobiliteiten, met behulp van capillaire zone electroforese.

LIST OF SYMBOLS AND ABBREVIATIONS 147

List of symbols and abbreviations.

Symbols a amount mole A area m'- value of the Dixons Q test

peak area A Us q charge of a particle ( =ze) c Debye-Hiickel-Onsager numerical constant

constant m7n/mole3n·D R radius m ionic species A gas constant (8.314) J/K·mole

a intercept of calibration graph r regression coefficient effective hydrated diameter m radius m

B Debye-Hiickel-Onsager I standard deviation constant m312/molelll displacement by the EOF m/s

ionic species B T temperature K b slope of a calibration graph terminating ionic species T b; calculated constant for C·m2/mole·Vs T' adapted terminating zone

conductivity detection T" adapted terminating zone c concentration mole/m3 t time s D dispersion coefficient m2/s transport number E electric field strength V/m v voltage v e elementary charge v velocity m/s

(1.602· 10'19> c w sample pulse width m F Faradays constant ~h peak width at half heigth s

(9.64·1o4) C/mole :x mean value of x F force N z mean value of y fc friction factor kg/s y calculated value of y I electric current A belonging to an x K dissociation constant z charge number

calculated value for number of carbon atoms indirect-UV detection a degree of dissociation

~ thermodynamic equilibrium "( activity coefficient constant correction factor for Debye-

k Boltzman constant Hiickel-Onsager effects (1.381·10'23) JIK 0 thermal coefficient

capacity factor of mobility 1/K

~ calculated constant for f. molar extinction coefficient m3/mole·cm conductivity detection dielectric constant C2/J·m

L leading ionic species L r . zeta potential v le length of the capillary m tJ viscosity Ns/m2

ld length of the capillary from A molar conductivity ml ID· mole injection to detection m Ao molar conductivity at

m effective mobility m2/Vs infinite dilution m2/0·mole number of readings of a >,O molar conductivity of an ionic

sample concentration species at infinite dilution m2/0-mole mo ionic niobility at infinite >.r thermal conductivity W/m·K

dilution m2/Vs " ionic strength me ionic mobility at finite " number of anions or cations per

dilution m2/Vs formula unit of electrolyte N plate number 11 conductivity 1/D·m NA Avogadro constant U1- variance m2 ors'-

(6.022·to23) 1/mole ll value of Kohlraush n number of experiments regulation function mole·Vs/m'

total mole of solute number of protolysis steps

p electric power per unit volume W/m3==VA/m3

148 LIST OF SYMBOLS AND ABBREVIATIONS

First subscript Abbreviations A,B,T and L according to substance A,B,T BICINE N,N-bis(2-hydroxyethyl)glycine

i +!-

z

and L CMC critical micelle concentration total COND conductivity belonging to substance i CREAT creatinine belonging to positve or negative CTAB cetyl trimethyl ammonium bromide

substance CZE capillary zone electrophoresis maximum charge DEA diethanolamine

DS downstream mode Second subscript EAC e-aminocaproic acid A,B, T and L in the zone of sub$tance A,B, T

andL EN electroneutrality EOF electroosmo.tic flow

Superscript rel ret z

ci Examples CBJi.A

lmB,AI

relaxation retardation maximum charge of an

species to the ith degree

GABA 'Y·aminobutyric acid GC gaschromatography GirP Girard Reagent P HIST histidine

ionic HPLC high performance liquid chromatogrsphy IC isotachophoretic condition I.D. internal diameter (m) IMID imidazole ITP isotachophoresis

total concentration of sub!;tance B LOD limit of detection in zone A MC micelle

absolute value of the effective MECC micellar electroldnetic capillary mobility of subtance B in the chromatography zone of subtance A MES 2-[N-morpholino]-ethanesulfonic

MHEC MO MOPS MS OPDA PROP PTFE RF RSD SD SDS SGE SZR25 TEA TRIS TIAB us ZL

acid methylhydroxyethylcellulose mesityl oxide morpholinopropanesulfonic acid midstream mode o-phenylenediamine propanoic acid polytetrafluoroethylene response factor (C/mole) relative standard deviation standard deviation sodium dodecyl sulfate Scientific Glass Engineering specific zone resistance (O·m) triethanolamine tris(hydroxymethyl)aminomethane tetradecyltrimethylammonium bromide upstream mode zone length (s)

DANKWOORD EN CURRICULUM VITAE 149

DANKWOORD

Een promotieonderzoek wordt zelden of nooit door de promovendus al.leen uitgevoerd. Veel dank ben ik dan ook verschuldigd aan mijn eerste promotor, Frans Everaerts en verder aan al.len die op enigerlei wijze hebben bijgedragen aan de totstandkoming van dit proefschrift. Een speciaal woord van dank wil ik richten tot Jo Beckers voor de wijze waarop hij mij van zeer nabij tijdens mijn promotieonderzoek geleid en begeleid heeft, Irma Seelen voor de prettige samenwerking en de vele analyses die zij nauwgezet heeft uitgevoerd, en ten slotte mijn ouders, die het mij mogelijk hebben gemaakt een academische studie te volgen en die door hun interesse mij daarin te al.len tijde gesteund en gestimuleerd hebben.

5 december 1963 juni 1982

sep. 1982- sep. 1987

sep. 1987 - sep. 1989

sep. 1989 - sep. 1992

Curriculum Vitae

Geboren · te Eindhoven. Eindexamen Gymnasium {j, Scholengemeenschap Augustinianum, Eindhoven. Studie Scheikundige Technologie, Technische Universiteit Eindhoven. Onderzoekersopleiding Scheikundige Technologie, Instituut Vervolgopleidingen, Technische Universiteit Eindhoven. Prornotieonderzoek in de vakgroep Instrumentele Analyse, Technische Universiteit Eindhoven.

150 AUTHOR'S PUBLICATIONS ON ELECTROPHORESIS

Author's publications on electrophoresis

1 M.T. Ackermans, Het profileren van cervixslijm met behulp van verschillende analytische technieken. Afstudeerverslag, Technische Universiteil Eindhoven, 1987.

2 M.M. Gladdines, M.T. Ackerma11s, F.M. Everaerts, P.J.Q. van der Linden, M.A.H.M. Wiegerinck and H.L. Vader, Analysis of the aqueous phase of human cervical mucus by reversed phase high performance liquid chromatography and capillary isotachophoresis. J. Chromatogr. Biomed. Appl., 431 (1988) 317-325.

3 H.A. Claessens, P.J.M. Hendriks, M.T. Ackermans, R.W. Sparidans and F.M. Everaerts, Sample pretreatment of body fluids by isotachophoresis prior to chromatographic analysis. Phamiac. Weekblad, Scientific Edition, 11 (1989) 6.

4 M.T. Ackermans, Isotachophoresis innon•aqueoussystems: some fundamentalsandapplications. Eindverslag, onderzoekers opleiding, lnstituut Vervolg Opleidingen, Eindhoven, 1989, ISBN: 90-5282-D23-6.

5 J.L. Beckers, F.M. Evetaerts and M.T. Ackermans, Determination of absolute mobilities, pK values and separation numbers by capillary zone electrophoresis; effective mobility as parameter for screening. J. Chromatogr., 531(1991)407-428.

6 J.L. Beckers, F.M. Everaerts and M.T. Ackermans, Isotachophoresis with electroosmotic flow: open versus closed systems. J. Chromatogr., 537 (1991) 429-442.

7 M. T. Ackermans, F.M. Everaerts and J.L. Beckers, Isotachophoresis in open systems: Problems in quantitative analysis. J. Chromatogr., 545 (1991) 283-297.

8 M.T. Ackermans, F.M. Everaerts and J.L. Beckers, Quantitative analysis in capillary zone electrophoresis with conductivity and indirect UV detection. J. Chromatogr., 549 {1991) 345-355.

9 M.T. Ackermans, F.M. Everaerts and J.L. Beckers, Determination of some drugs by micellar electrokinetic capillary chromatography: The pseudo effective mobility as parameter for screening. J. Chromatogr., 585 (1991) 123-131.

10 M.T. Ackermans, J.L. Beckers, F.M. Everaerts and I.G.J.A. Seelen, Comparison of isotachophoresis, capillary zone electrophoresis and high-performance liquid chromatography for the determination of salbutamol, terbutaline sulphate and fenoterol hydrobromide in pharmaceutical dosage forms. J. Chromatogr., 590 (1992) 341-353.

11 M.T. Ackermans, F.M. Everaerts and J.L. Beckers, Relationship between zone length and step height in isotachophoresis. J. Chromatogr., 595 (1992) 327-333.

12 M.T. Ackermans, J.L. Beckers, F.M. Everaerts, H. Hoogland and M.J.H. Tomassen, Determination of sulfonamides in pork meat extracts by capillary zone electrophoresis. J. Chromatogr., 596 (1992) 101-109.

13 M. T. Ackermans, F .M. Everaerts and J .L. Beckers, Hyphenated capillary zone electrophoresis with indirect UV and micellar electrokinetic capillary chromatography; determination of aminoglycoside antibiotics in pharmaceuticals. J. Chromatogr., accepted for publication.

STELLINGEN

behorende bij het proefschrift


Some fundamental aspects

van

Mariette Theodora Ackermans

De uitdrukking 'ITP on top of CZE' wordt door Vinther et al. ten onrecbte gebruikt daar bet in dit geval niet bandelt om een superpositie van ITP op CZE maar om een eenvoudige stacking procedure, terwijl bovendien bet verschijnsel waarschijnlijk samenhangt met electrode-processen.

A. Vinther, F.M. Everaerts and H. Soeberg, J. High Resolut. Chromatogr., 13 (1990) 639.

2 Met de huidige stand van zaken in capillaire zone electroforese en de daarbij behorende detectiemethoden is het gebruik van deze techniek als screenings-methode voor antibiotica in land- en tuinbouwproducten niet zinvol.

Dit proejschrift.

3 Het uitvoeren van een dimensie-analyse bij het hanteren van formules moet voor iedere wetenschapper een tweede natuur zijn. Hierdoor . kunnen onjuistheden in nieuw afgeleide ingewikkelde formules direct aan het licht komen en worden onnodige fouten vermeden.

G.C. Cecchi, Maes. Sci. Technol. 2 (1991) 1127.

4 De stelling van J.G.M. Janssen dat de slechte reproduceerbaarheid van de electroosmotische stroming in capillaire zone electroforese op dit moment nog een belangrijk obstakel vormt voor de brede toepasbaarheid van de techniek geldt alleen voor zover het volledig geautomatiseerde routine analyses betreft.

J.G.M. Janssen, Proefschrift, Technische Universiteit Eindhoven, 1991. Dit proefschrift, hoofdstuk 2 en 6.

5 De productie van benzodiazepine-achtige verbindingen in de koemaag is enerzijds een plausibele verklaring voor het doorgaans gezapige gedrag van deze herkauwers en is anderzijds een argument patienten met neiging tot portosystemische encephalopathie een beperking op te leggen wat betreft de consumptie van koemelk en rundvlees.

J.H. Medina, J.L. Danelon, C. Wasowski, M. Levi de Stein and A.G. Paladini, Biochem. Biophys. Res. Commun. 181 (1991) 1048.

E.A. Jones, Hepatology, 14 (1991) 1286.

6 Het feit dat bij capillaire zone electroforese vrlj eenvoudig electroferogrammen verkregen kunnen worden die een zekere gelijkenis vertonen met chromatogrammen heeft ertoe geleid dat veel chromatografici te makkelijk de stap na.ar electroforese hebben gezet. Daar echter bet scheidingsprincipe geheel verschillend is, zal men de theorie van electroforese eerst terdege moeten bestuderen, teneinde te voorkomen dat reeds bekende feiten in publicaties worden gepresenteerd als nieuwe ideeen.

E. V. Dose and G.A. Guiochon, Anal. Chem., 63 (1991) 1063. K.D. Altria and C.F. Simpson, Chromatographia, 24 (1987) 527. T. Tsuda, J. High Resolut. Chromatogr. Chromatogr. Commun., 10 (1987) 622. J. Vindevogel and P. Sandra, J. Chromatogr., 541 (1991) 483.

7 Aangezien micellaire electrokinetische capillaire chromatografie een techniek is waarvan het scheidingsmechanisme zowel op chromatografische als electroforetische principes berust, kan het met evenveel recht een chromatografische als een electroforetische techniek genoemd worden. Het plaatsen van capillaire · zone electroforese onder de noemer van electrochromatografische technieken is onjuist, gezien het feit dat bij · deze techniek bet scheidingsprincipe puur electroforetisch is.

J.H. Knox and l.H. Grant, Chromatographia, 24 (1987) 135.

8 Vrouwenbesnijdenis dient beschouwd te worden als het opzettelijk toebrengen van lichamelijk letsel aan kinderen en dient derhalve bij wet verboden te worden.

H. Crul, Elsevier, 17 (1992) 98. A. Simona, Elsevier, 22 (1992) 8.

9 Een beleid waarbij steeds lagere subsidies worden toegekend aan Kunst en Cultuur, zonder een daarmee gepaard gaande numerus fixus aan de hieraan geli.eerde beroepsopleidingen, kan als a-sociaal bestempeld worden.

10 In het licht van een overaanbod aan publicaties dienen de verschillende tijdschriften afspraken te maken met betrekking tot het vermijden van dubbelpublicaties.

11 Als in een wetenschappelijk tijdschrift discussie ontstaat over een publicatie, zou bij verschil van mening niet moeten worden volstaan met bet publiceren van de "letter to the editor" en een commentaar van de oorspronkelijke auteur, maar zou een onathankelijke deskundige gevraagd moeten worden zijn mening met betrekking tot bet discussiepunt te geven.

K. Shimao, Electrophoresis, 7 (1986) 121. P. Gebauer and P. BoCek, Electrophoresis, 8 (1987) 253.

12 Detective romans lijken wel voor mannen geschreven te zijn. De positie van de vrouw in deze lectuur is vaak te rolbevestigend, de taal vaak vrouw-onvriendelijk.

I. Fleming, Goldflnger. Charteris, De Saint en de groene papagaai; Leve de Saint. J. Bruce, Agonie and patagonie.

Mariette Theodora Ackermans Eindhoven, 4 september · 1992.

Documents

Electrophoresis in open capillaries : some fundamental aspects · ELECTROPHORESIS IN OPEN CAPILLARIES Some fundamental aspects PROEFSCHRIFT ter verkrijging van de graad van doctor