Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Determination of on-capillary pH and proton and hydroxide mobility in capillary electrophoresis Duso, Angela B. 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0371.pdf [ 3.94MB ]
Metadata
JSON: 831-1.0061433.json
JSON-LD: 831-1.0061433-ld.json
RDF/XML (Pretty): 831-1.0061433-rdf.xml
RDF/JSON: 831-1.0061433-rdf.json
Turtle: 831-1.0061433-turtle.txt
N-Triples: 831-1.0061433-rdf-ntriples.txt
Original Record: 831-1.0061433-source.json
Full Text
831-1.0061433-fulltext.txt
Citation
831-1.0061433.ris

Full Text

D E T E R M I N A T I O N O F O N - C A P I L L A R Y P H A N D P R O T O N A N D H Y D R O X I D E M O B I L I T Y IN C A P I L L A R Y E L E C T R O P H O R E S I S by A N G E L A B. D U S O B . S c , University of British Columbia, 1994 A THESIS S U B M I T T E D FN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Chemistry We accept this thesis as conforming to the required standard The University of British Columbia July 2001 ©Angela B. Duso, 2001 U B C Special Collections - Thesis Authorisation Form Page 1 of 1 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y a v a i l a b l e for reference and study. I further agree that permission for extensive copying of th i s thesis for sch o l a r l y purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or pu b l i c a t i o n of th i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of CWeyVMSt^ The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada Date ^ept .3-1 lo\ http://www.library.ubc.ca/spcoll/thesauth.html 9/29/01 ABSTRACT The primary aim of this work is the measurement of on-capillary p H within the field of capillary electrophoresis. The first part of this thesis presents an investigation into the properties of the p H junction formed through a multi-electrolyte system within the capillary. For 1-unit p H differences, the p H character of a sample zone, sandwiched between zones of higher or lower p H was preserved, although the p H junctions themselves were observed to move. For 2-unit p H differences, an increase in time spent on the capillary, along with movement of the dye molecules being used to measure the pH, made it impossible to measure the p H of the sample zone. These experiments led to an interest in the mobility of proton and hydroxide ions within the capillary. While extensive measurements had been performed on bulk solutions of strong acids and bases under an electric field, no one as yet had measured the mobilities on a capillary. Using a p H indicator to measure the migration velocity of the ions, the mobilities of OH" and H + ions were determined, and found to be similar to empirical measurements made using bulk solutions. Finally, in a departure from p H measurement, method development for the assay of functionalized dendrimer compounds was performed. A qualitative assay was developed and evaluated for linearity, quantitation limit, and specificity. The assay was also applied to samples of the compound prepared in a plasma sample matrix. iii Table of Contents A B S T R A C T ii Table of Contents iii List of Tables v List of Figures vi 1 Introduction 1 1.1 Historical Background of Capillary Electrophoresis 1 1.2 Theory 7 1.2.1 Electroosmotic Flow 7 1.2.2 Electrophoretic mobility 11 1.2.3 Capillary Zone Electrophoresis 15 1.2.4 Performance Criteria 18 1.3 Aims of Research 20 2 On-capillary pH measurement 21 2.1 Introduction 21 2.2 Experimental 24 2.2.1 Chemicals and Reagents 24 2.2.2 Apparatus and Procedure 24 2.3 Results and Discussion 25 2.3.1 Choice of ratiometric pH probe 25 2.3.2 Multi-section electrolyte 28 2.3.3 Buffer vial replacement 30 2.3.4 Control experiments: pH junction 33 2.3.5 1-unit pH junction 36 2.3.6 2-unit pH junction 47 2.4 Conclusions 52 3 Proton and Hydroxide Mobility 54 3.1 Introduction 54 iv 3.2 Experimental Section 56 3.2.1 Chemicals 56 3.2.2 Apparatus and Procedure 56 3.3 Results and Discussion 57 3.3.1 Proton mobility using methyl red indicator 60 3.3.2 Proton mobility using alizarin red indicator 63 3.3.3 Hydroxide ion mobility 65 3.4 Conclusion 66 4 Method Development and Validation in the Analysis of Dendrimer Compounds 68 4.1 Introduction 68 4.2 Experimental 74 4.2.1 Chemicals 74 4.2.2 Instrumentation 75 4.3 Results and Discussion 76 4.3.1 Low pH phosphate buffer 76 4.3.2 SDS Gel Electrophoresis 81 4.3.3 Borate Buffer Systems and Method Validation 83 4.3.4 Biological Assay 95 4.4 Conclusion 100 5 References 101 List of Tables Table 1-1. Factors that influence the E O F 11 Table 2-1. Comparison of measured vs. calculated pH 28 Table 2-2. E O F mobility and current in 40 mM borate buffer 29 Table 2-3. Reproducibility of pH over 10 analyses using the same 1.6 mL buffer vials 32 Table 2-4. Reproducibility of pH over 5 analyses with new buffer vials each run 32 Table 2-5. Multi-section electrolyte experiments: controls 34 Table 2-6. 1 unit pH-junction experiments 37 Table 3-1. Mobilities (10"4 cm2/Vs) of Representative Ions in Water at 25°C 4 7 54 Table 3-2. Mobility data (10"4 cm2/Vs) for H + and OH" ions measured using methyl red (MR) and alizarin red (AR) 61 Table 4-1. Molecular weight, number of monomeric units (lysine), number of primary amino terminal groups, and hydrodynamic diameter for Denkewalter dendrimers 5 5 72 Table 4-2. Mobility data for dendrimer P-l in borate buffer 84 Table 4-3. Injection repeatability for P - l 88 Table 4-4. Linearity Upper Range 89 Table 4-5. Linearity over a Lower Concentration Range 92 Table 4-6. Linearity Values 93 Table 4-7. Stability Samples 94 Table 4-8. Optimization of plasma pre-treatment 99 vi List of Figures Figure 1-1. Diagram of electrical double layer occurring at the interface between capillary wall and electrolyte solution. Solvent ions are represented by grey dots surrounding central ion.. 8 Figure 1-2. Free zone capillary electrophoresis of ions and neutral molecules 16 Figure 1-3. Schematic of capillary electrophoresis apparatus 18 Figure 2-1. Structure of HPTS molecule. Two ionization forms of dye are shown 26 Figure 2-2. pH dependent spectra of HPTS 4 3 27 Figure 2-3. pH 8, pH 9 control experiments using multi-section electrolyte. Shading represents HPTS presence [a] sample region is blank, B G E contains HPTS [b] sample region contains HPTS 33 Figure 2-4. Control experiments, dye in sample zone. Detection wavelength = 454 nm 35 Figure 2-5. Diagram of 1 unit pH junction experiments. pH junctions exist where sample zone interacts with B G E regions (patterned). Dye was included in either B G E , or sample zone (Table 2-6) 36 Figure 2-6. Electropherogram of pH 8:9:8 junction. Detection wavelength = 454 nm 38 Figure 2-7. Proposed mechanism of HPTS movement and focussing in capillary, [a] pH 9 sample zone contains -4 charge HPTS. [b] electric field is applied, some HPTS migrates out of sample zone, net flow is to detector with EOF [c] focussing occurs at back pH junction 41 Figure 2-8. pH 9:8:9 electropherogram. Detection wavelength is 454 nm 42 Figure 2-9. Electropherogram of pH 8:9:8 junction, sample zone is blank. Wavelength shown is 454 nm 43 Figure 2-10. pH 8:9:8 multi-section electrolyte experiment, [a] dye is contained in shaded regions. Volume next to and including detector is still at pH 8, but contains no dye to prevent detector auto zero from occurring with dye in capillary window, [b] separation voltage applied [c] at detector, front portion of dye is focussed on front pH junction and has pH 9 character 44 Figure 2-11. Electropherogram of pH 9:8:9 junction. Detection wavelength = 454 nm. pH values of various regions are indicated on figure 45 Figure 2-12. [a] is the pH 8:9:8 experiment also shown in Figure 2-6. [b] is the same experiment, with a 2-unit pH junction (pH 7:9:7) 48 Figure 2-13. Electropherogram of pH 7:9:2 junction. Detection wavelength 454 nm shown 50 Figure 2-14. Electropherogram of pH 9:7:9 junction. Detection wavelength shown is 454 nm...51 Figure 2-15. Electropherogram of pH 9:7:9 junction. Detection wavelength shown is 454 nm... 52 Figure 3-1. Prototypic conduction process of proton and hydroxide ions, [a] Proton conduction under an electric field, [b] hydroxide ion conduction under an electric field 55 V l l Figure 3-2. Chemical structure and UV-spectra of methyl red prepared in 50 mM KC1 solution, [a] pH>6[b] pH<4 59 Figure 3-3. Chemical structure and UV-spectra of alizarin red prepared in 50 mM KC1 solution. a)pH>6b) pH<4 60 Figure 3-4. Electropherogram showing the conversion of methyl red from basic to acidic form, indicating migration time of proton, a) absorbance at 520 nm b) absorbance at 410 nm 61 Figure 3-5. Electropherogram showing the conversion of alizarin red from basic to acidic form, indicating migration time of proton, a) absorbance at 410 nm b) absorbance at 520 nm 63 Figure 3-6. Electropherogram of conversion of alizarin red from acidic to basic form, indicating migration time of hydroxide ion. a) 520 nm b) 410 nm. B G E contains 50 mM KC1 65 Figure 4-1. 2 n d generation P A M A M dendrimer. Generated from 3-directional ammonia core, with amide connectivity 69 Figure 4-2. Schematic of Denkewalter dendrimer synthesis. Protected amino acid, N,N'-bis(terf-butoxycarbonyl)-L-lysine is the monomeric building block 71 Figure 4-3. R group attachment to 1° amine termini on dendrimer 73 Figure 4-4.4th and 5 t h generation unsubstituted BHAlys dendrimers analyzed in low pH phosphate buffer 79 Figure 4-5. Electropherogram of P-l analyzed in low pH phosphate buffer 80 Figure 4-6. Effective mobility vs. borate concentration 84 Figure 4-7. Electropherogram of P-l analyzed in 40 mM borate buffer 85 Figure 4-8. P-2 analyzed in 40 mM borate buffer 86 Figure 4-9. Peak Area vs. Concentration over a range of 10-2000 ug/mL for compound P-l 89 Figure 4-10. Structure of impurity with RMT 1.05 95 1 Introduction 1.1 Historical Background of Capillary Electrophoresis Electrophoresis occurs when charged particles suspended in a liquid medium move under the influence of an electric field. In the late 1800's, physical chemists and electrochemists were involved in discovering the nature and properties of electrolyte solutions, and they used the information gained from electrophoresis experiments to define the properties of electrolytes l . In the 1870s, Kolrausch was the first to measure the conductivity of an ionic solution using a Wheatstone bridge 2 . In 1899, Hardy showed that proteins and enzymes could be characterized based upon their unique mobilities using electrophoresis. This led to the routine use of electrophoresis to characterize molecules based on mobility measurements. Michealis (1909) determined the isoelectric point (pi) of various proteins and enzymes, using a type of electrophoresis known as iso-electric focussing, when little else was known about the physical and chemical properties of these compounds 3 . Moving-boundary electrophoresis was used extensively in the early 1900's to determine transport numbers and hence, equivalent ionic conductance of several electrolyte solutions l. In 1937, Tiselius was the first to apply the moving-boundary method to separate a mixture of proteins. Tiselius's moving-boundary apparatus consisted of a U-shaped quartz tube that was thermostated in a water bath to reduce the convective currents that were a result of Joule heating. Joule heating occurs whenever a current runs through a material. In electrophoresis, the current carrying medium is the buffer or electrolyte solution, and the temperature in an unthermostated tube can reach the boiling point of the capillary contents. Ionic strength is inversely proportional to resistance, and therefore, from Ohm's 2 law, directly proportional to current. Since different zones of the tube can contain different ionic strengths (e.g. analyte zone vs. buffer zone), current may vary throughout the tube, resulting in local temperature differences. In addition, the solution next to the walls of the tube is able to dissipate heat more readily than the solution in the middle of the tube, resulting in a radial temperature gradient. Variations in temperature can result in convective currents that enhance the diffusion of analytes. To reduce the convective currents, Tiselius employed a rectangular, rather than round, tube to increase the surface area, which increased the amount of heat that could be dissipated, as well as ensuring that all points within the solution were close to a wall 2 . Another problem Tiselius faced was gravitational diffusion. His tube was U-shaped, and the inherent mobility of the proteins in the mixture would cause them to concentrate into bands in one vertical arm of the tube. Gravity caused the zones to diffuse downwards in the tube, mixing with zones that lay below. Because of these difficulties, further attempts at electrophoresis were mainly performed using anti-convective media such as gels, glass beads, and paper 4>5. Thermal convection currents were suppressed, but not without compromises caused by the interaction of analytes with the stabilizing media. The use of anti-convective media required only a simple electrophoresis apparatus. The paper or gel was placed in an isolated bath, connected to a power supply and electrophoresis could be performed. Simple detection systems, such as staining or even visual interpretation, were possible with these systems. In spite of the problems with design issues, the use of electrophoresis was quickly accepted in biochemical and molecular biological circles 4 Even while high performance liquid chromatography (HPLC) was gaining ground as a widely applicable and robust technique in the analytical laboratory, electrophoresis was favoured for its advantages in the analysis of large biopolymers, such as proteins and D N A fragments. Gel electrophoresis proved to be so popular that even today it is a cornerstone of biochemical laboratory procedures 6 . However, gel-electrophoresis and its related techniques are often time-consuming processes, involving relatively large amounts of sample and consumables, as well as a high degree of operator skill and experience in interpreting the results 1 . While further developments in gel electrophoresis were being made, other avenues of electrophoresis were being explored. Pretorius and Gould investigated the use of electroosmotic flow (EOF) as a pumping mechanism in liquid chromatography 8>9. The E O F originates from the electrical double layer that is formed at the solid-liquid interface of a charged surface in contact with an electrolyte solution. In H P L C , mechanical pumps provide a pressure-driven flow that results in a parabolic flow profile. The parabolic flow profile distorts the shape of the narrow sample band that was the injected plug, and is one cause of band broadening in H P L C . Frictional interaction between the H P L C solvent and the chromatography medium causes laminar flow. A flat flow profile is maintained in electrophoresis, because the driving force on the fluid is applied continuously throughout the length of the tube, column or gel. Electroosmotically driven systems are simpler when compared to the expensive pumps and high-pressure columns required for pressure-driven liquid chromatography systems. In 1967, Hjerten 1 0 published a comprehensive work on a type of electrophoresis that was conducted with no anti-convective media. The technique was termed zone electrophoresis and was conducted in tubes of quartz glass coated with methylcellulose. 4 Hjerten used rotational forces to eliminate the disturbance caused by convection. He developed an instrument that was able to axially rotate the electrophoresis tube in a temperature controlled water bath. Hjerten saw the electroosmotic flow as a detrimental force in electrophoresis, and he coated his quartz tubes with methylcellulose in order to remove its effects. A small band of sample was introduced to the quartz tube using a microlitre syringe as a sample application device 1 0. In 1979, Mikkers et al 5 , found that they could overcome the limitations of Joule heating by reducing the diameter of the tube in which the electrophoresis was conducted. A reduction in tube diameter leads to an increase in surface-to-volume ratio, allowing for more efficient dissipation of heat. Most commercial instruments available today implement some form of thermostating, to keep capillary temperatures constant and reduce variability. In Mikkers experiments, he used chemically and electrically inert polytetrafluoroethylene (PTFE) tubes of 200 um inner diameter (i.d.). Previously, other forms of electrophoresis (moving-boundary and isotachophoresis) had been conducted in this style of tubing 1 1 . Mikkers also contributed to the concept of focussing, which he termed the "concentrating capability" of the electrophoretic system 5 . He realised that when prepared in low conductivity water, the sample analytes would form sharper (narrower) peaks than when prepared in higher conductivity separation buffer. It was Jorgenson however, in 1981, who first applied open tubular glass capillaries to the technique of free-zone electrophoresis and the term "capillary electrophoresis" (in reference to the narrow flexible glass capillaries) was born. Jorgenson demonstrated the benefit of combining the effects of electroosmotic flow and the electrophoretic mobility of an analyte. The presence of an E O F in the glass capillaries allowed for rapid analysis of positive, neutral and negative analytes. Jorgenson 5 determined that maximum resolution was obtained when the magnitude of the E O F and the electrophoretic mobility of the analyte opposed one another 1 2 . The advances in capillary electrophoresis (CE) since Jorgenson's first publication have been numerous. In 1986, Jorgenson described 3 different modes of C E : free zone, isotachophoresis and isoelectric focussing 4 . The technique of isotachophoresis (ITP) has been used to determine the mobility of ions 1 3 . In ITP, substances are separated based on their different effective mobilities. The sample is sandwiched between a fast leading ion and a slow terminating ion, and the mobilities of the sample components must fall between those of the leading and terminating ions. A voltage is applied, and the sample ions will order themselves according to their individual mobilities and move through the capillary at a fixed speed, dictated by the leading ion n . Today, there are a multitude of modes of electrophoresis available to the researcher, and several advancements in instrumentation and detection have been made. In micellar electrokinetic chromatography ( M E K C ) a pseudo-stationary selectivity media with an affinity for the analytes of interest is included in the C E buffer 1 4 . The development of M E K C by Terabe allowed for the separation of neutral analytes, previously not possible using zone electrophoresis 1 4 In 1989, Guttman showed that the resolution of enantiomers was possible using chiral buffer additives such as cyclodextrins 1 5 . Previously, expensive enantioselective H P L C columns were required for the routine analysis of enantiomers. As many pharmaceuticals consist of enantiomers, this breakthrough led to interest in C E from the pharmaceutical industry. Burgi and Chien 1 6 -1 8 pioneered techniques in on-line sample concentration, improving the detection limit for U V detection. While the mass sensitivity in C E is actually very good when one examines the amount of sample reaching the detector, the concentration sensitivity is poor, due to small sample volumes, and the short optical path length dictated by the on-capillary detector window. Recent advances in the field of biotechnology, which have included such milestones as the sequencing of the human genome 1 9 , have greatly influenced the popularity of capillary electrophoresis as an information-gathering tool. Several C E methods have been applied to D N A sequencing 20-22j a n c i commercial capillary array D N A sequencers that are able to analyze 400 nucleotides per hour are available. There has also been an interest in using C E in the field of proteomics, in which the products of gene expression are characterized 2 3 . In summary, no single contribution can be credited with the development of capillary electrophoresis that exists in the many forms that are used today. From the early experiments with electrolyte solutions by Kolrausch, to miniaturized C E "lab on a chip" devices 2 4 that have been recently designed, the development path of C E has seen contributions from a broad array of fields, including analytical chemistry, physical chemistry, pharmacology, and biochemistry. The use of the technique in many different disciplines is a testament to the collaborative nature of scientific studies in the area of separation science. In addition, increased research in this field is leading towards greater acceptance of the technique within commercial enterprise, as the reproducibility and validity of C E methods are proven. 1.2 Theory 1.2.1 Electroosmotic Flow Electroosmosis is described as the flow of liquid, in contact with a solid surface, under the influence of a tangentially applied electric field. It is attributed to the formation of an electrical double-layer at the solid-liquid interface. Within a fused-silica capillary, a negatively charged layer exists at the capillary surface due to the presence of ionized surface silanol groups. The electrical double layer consists of an immobilized charge layer of unsolvated cations, termed the Stern layer, and a second layer of solvated cations that is termed the diffuse layer. The graduations of charge at the capillary surface result in a potential drop, which is linear over the Stern layer, and then decreases exponentially over the diffuse layer, Figure 1-1. The centres of the solvated ions found in the diffuse layer define the outer Helmholtz plane, and the potential measured at this point is termed the zeta (£) potential. 8 Stern layer Diffuse layer Electrolyte zeta potential *I3 <y-x> c>-i) ( © : : :a Distance from wall Figure 1-1. Diagram of electrical double layer occurring at the interface between capillary wall and electrolyte solution. Solvent ions are represented by grey dots surrounding central ion. When an electric field is applied, the cations of the diffuse layer move towards the cathode. Due to their hydration, they drag the bulk solvent with them, resulting in bulk fluid motion, termed the electroosmotic flow. At the wall, the flow velocity is 0, and it reaches a maximum at the interface between the diffuse layer and the bulk solution. The driving force behind the E O F is applied along the entire capillary, resulting in a uniform flow velocity and a flat flow profile. The mobility of the electroosmotic flow is described by; 47in (1.1) where £, is the zeta potential, e is the dielectric constant of the solvent, £o is the constant of permittivity in a vacuum, and r| is the viscosity 2 5 . The mobility of the E O F is directly proportional to the zeta-potential, and is therefore dependent on the number and nature of the surface charges at the capillary wall (o) and the thickness of the double layer ((3) 2 5 , as shown in Eqn. (1.2). « = (1-2) ee 0 The thickness of the double layer increases with decreasing ionic strength, due to the condition of electroneutrality. In low ionic strength buffer, fewer cations will occupy the Stern layer, requiring a larger diffuse layer to balance the negative charges at the capillary wall. The opposite holds true for high ionic strength buffers, or buffers containing multivalent cations. In these buffers, the double layer is smaller and the potential drop (or zeta-potential) will be steeper over the Stern layer. Because the first few layers of ions near the solid-liquid interface do not move, electroosmotic flow is dependent on the diffuse layer. The thicker the layer, the greater the magnitude of the E O F . The double-layer thickness is described by o £ £ 0 k T 2 N A e 0 2 I Where k is the Boltzmann constant, T is the temperature, N A is Avogadro's number, and eo is the fundamental charge on an electron. Ionic strength (I), is related to sum of the product of the charges (z) and concentrations (c) of all of the ionic species present in an electrolyte 2 5 . I = l / 2 £ Z i 2 C i (1.4) 10 Because electric field strength affects the mobility of cations in the diffuse portion of the double layer, the velocity of the E O F is dependent on the applied electric field (E) 2 6 . v = MeoE (1.5) At p H values below the p K a of the silanol groups (around 2.5), there will be a reduced charge density (a) at the wall, diminishing the thickness of the double layer. In the absence of the electrical double layer, an E O F is not generated. Hence, one technique in C E for greatly reducing the magnitude of the E O F is to use a low p H buffer. The E O F can also be reduced or negated through the use of neutral capillaries in which the silica surface is permanently coated with a polymer layer such as polyacrylamide. The capillary surface is uncharged, and no electrical double layer will form. The E O F can be reversed by switching the polarity of the power supply, placing the cathode at the capillary inlet. While the E O F is easily manipulated and provides a flat flow profile that reduces analyte band spreading 9 , it is often irreproducible, and leads to variation in analyte migration times over a series of analyses. Many things can affect the E O F ; minor changes in the charge density of the capillary wall, local concentration differences within the capillary, production of protons and hydroxide ions at the inlet or outlet vials are a few examples. Table 1-1 highlights the major factors that can influence the E O F . 11 Table 1 -1 . Factors that influence the E O F Modification Effect on E O F Reason p H of buffer T E O F with T p H Deprotonation of silanol groups at capillary wall buffer concentration T E O F with ^buffer strength Reduced ionic strength of buffer leads to an increase in thickness of electrical double layer ( T £, potential) temperature ^viscosity T E O F Less viscous solvents reduce frictional retarding force on ions organic content in buffer ^dielectric constant - l E O F Reduction in dielectric constant from organic solvents reduces £, potential cationic surfactant X or reverses E O F Adheres to capillary wall and masks silanol groups, causing E O F reversal by setting up positively charged layer neutral coating suppression of E O F Prevents electrical double layer formation by masking silanol groups 1.2.2 Electrophoretic mobility The mobility of ions in solution under an electric field is described as the limiting velocity (v) of an ion attained under a unit force (F), (1.6) The units of force are commonly defined as a potential gradient of one volt per centimetre acting on the ionic charge l , this is essentially the electric field strength, E ; 12 V E = — (1.7) where V is the applied voltage and Lt is the total length of the capillary. The frictional resistance of a particle moving through an ideal liquid medium can be calculated from the viscosity (rj) of the solvent and the dimensions of the particle. For spherical particles, Stokes 1 obtained the relationship FM=v6nrir (1.8) where r is the Stokes radius of the sphere and Ff„ the frictional forces acting upon it. The electric field force, F e i , opposes the frictional force and has a magnitude of F e, =qE (1.9) where q is the charge on the ion. The two forces oppose one another and can be equated qE = 67t7]rv (1.10) By solving for velocity, the absolute mobility (u0) of an ion is, " o = 7 ^ - (1.11) 6nnr However, Eqn. (1.11) only holds true for ideal spherical particles in an infinitely dilute solution. The absolute mobility is a characteristic constant of a particular ion in a specific solvent and it is related to equivalent ionic conductance (ko) at infinite dilution. The equivalent conductance (A) of a solution is the relative conductivity corrected for the equivalent concentration of the electrolyte. Conductivity is equal to the amount of current 13 passing through a solution, and is related to the current-carrying capacity of the electrolytes in the solution. Equivalent conductance can be divided into contributions from the positive and negative ions in a solution A 0 = A ; + A - (i.i2) Equivalent ionic conductance at infinite dilution (KQ) relates to absolute ionic mobility (u 0) through X0=u0F (1.13) where F is the Faraday constant. Absolute mobility cannot be measured directly, but it can be extrapolated from measurements on dilute solutions. Many physical chemists determined ion mobilities in the late 19 t h century using conductivity measurements 2 7 - 3 0 . Kolrausch 1 1 determined that for strong electrolytes measured in dilute solution (< 0.001 M ) , equivalent ionic conductance varies linearly with the root of concentration, Xc=XQ-a4c (1.14) where a is a constant. Equivalent ionic conductance at infinite dilution can be determined by finding the y-intercept of the relationship in (1.14). Using (1.13) it is then possible to determine un. While absolute ionic mobility is an important characteristic used to describe the properties of individual ions, the presence of other ions in non-dilute solutions will affect mobility . Debye and Hiickel determined that the presence of other ions in solution will create an ion atmosphere in which a central ion is surrounded by counter-ions of opposite 14 charge. Under an electric field, two effects occur that retard ion motion toward the cathode (anode for a negative ion). Relaxation or asymmetry effects result when the E -field is applied, and the central ion moves outside of its ion atmosphere, resulting in a distortion in the distribution of the counter-ions. The counter-ions tend to accumulate in the wake of the central ion, and their presence equates to a retarding force. A n electrophoretic effect also occurs, in which a counter-current of solvated ions heading in the opposite direction of the central ion results in a retarding force. It was also noted by Jorgenson 1 2 that partial dissociation of ions would result in a reduction in mobility. If an ion exists in equilibrium with its undissociated form, for example a weak acid, its effective mobility will be smaller than its absolute mobility. The free ions will migrate as a uniform substance in equilibrium with the undissociated form. The effective mobility is the sum of the fraction of each form present, multiplied by the mobilities of each form. where 0 is the degree dissociation of species /, and p is the mobility of species i. Hence, the mobility of an ion measured at any concentration other than infinite dilution is termed the effective mobility, u e f f ) and it will be affected by the following parameters; ionic radius, solvation, dielectric constant and viscosity of solvent, shape,, charge, pH, % dissociation and temperature. The effective mobility was related to the absolute mobility by Reijenga, using extensive computer simulations, based on conductivity measurements made in the early 19 t h century 3 1 II -II „(-0-5z'-78V7) , n Meff - u o e (1-16) 15 where z is the charge on the ion, and I the ionic strength of the solution. As seen from this equation, effective mobility decreases with increasing ionic strength. In summary, the mobility of an ion in solution is related to the conductivity of the solution, as the ions are the current-carriers in non-conducting media. Effective mobility is the mobility of the ion in the presence of other ions in solution, and is always smaller than absolute mobility, which is the mobility of a single ion at infinite dilution. 1.2.3 Capillary Zone Electrophoresis The two forces discussed, electrophoretic mobility (u e p) and electroosmotic flow (ue of) allow for the analysis of positive, neutral and negative analytes in one separation, Figure 1-2. 16 t=0 Apply voltage t=i <+> Q ( R ) Q t=2 (+) t. detector detector electroosmotic flow © ( n ) © * detector electroosmotic flow © 0 G 15 time (min) (-) (-) Figure 1-2. Free zone capillary electrophoresis of ions and neutral molecules Without an E O F , anions and neutral molecules will never reach the detector. The E O F affects the mobility of cations, causing them to move more quickly through the capillary. Neutral compounds move at the same speed as the E O F , and the mobility of anions is opposed by the E O F . As long as the magnitude of the E O F mobility is greater than that of the anion, the anion will reach the detector. 17 Ion mobility observed in the presence of the E O F is termed the apparent mobility (uapp) and it is a combination of the effective electrophoretic mobility of the ion (peff) and the mobility of the electroosmotic flow (ueof)-Mapp = U e f f + U e o f ( L 1 7 ) For two closely migrating cations, the E O F has a negative effect upon their resolution because it masks slight differences in the components' mobilities. This is a situation in which it becomes desirable to modify the E O F , although reduction of the E O F results in an increase in analysis time 1 2 . The apparent mobility of an ion can be measured from its velocity within the capillary. Migration velocity is the quotient of the distance traveled to the detector (L D ) and the time required for the ion to get there. Ion velocity is the product of electric field strength and the apparent mobility of the ion. v = u a p p E (1.18) A n electric field is applied across a narrow capillary filled with electrolyte, via electrodes placed in inlet and outlet vials, Figure 1-3. The fused-silica capillary is generally between 20 and 100 cm in total length. The capillary may be filled with buffer, rinse, or sample solutions by the application of pressure at the inlet or outlet vials. The high voltage power supply is normally grounded at the capillary outlet (detector side) for "normal polarity". When the ground is reversed, the instrument is said to be in "reverse polarity". There is a protective polyimide coating on the exterior of the capillary surface which is removed at a small region prior to the outlet in order to create a detector window. A lamp or laser as well as optics associated with the radiation source are trained on this spot, and a photomultiplier tube or other light-collection device is located adjacent 18 to the spectral source. Data collection and storage is achieved through a computer coupled to the light collection device. This computer may also be used to control the instrument and to analyze data. To perform electrophoresis, the capillary is filled with buffer and the sample is introduced at the capillary inlet. The sample vial is then replaced with a buffer reservoir and a separation voltage is applied. Current is monitored as the run progresses. Detection of analytes occurs on-line with optical and electrochemical detectors, or off-line using mass spectrometry. thermostated capillary detector data processing buffer reservoir high voltage power supply Figure 1-3. Schematic of capillary electrophoresis apparatus 1.2.4 Performance Criteria Terms such as selectivity and resolution are commonly used in separation science to describe the separation obtained between different analytes. Selectivity (a) refers to the ratio of migration rates (velocities) of two migrating components, A and B. 19 (1.19) The greater the difference in the migration rates, the larger the separation between the two migrating components in the capillary. Using a longer capillary will also increase the separation between two adjacent peaks, by maximizing the distance between them. Many methods are available to manipulate the migration rate, and hence selectivity, of two closely migrating compounds. Buffer additives that complex more favourably with one analyte over the other are a common way to manipulate selectivity. Resolution describes the two effects of band spreading and band separation between two adjacent peaks. Experimentally it can be determined from the relative differences in migration time (tmig r) and the summation of the base peak widths (w) of components A andB. Narrowing zone bandwidths and increasing band separation (improved selectivity) are two ways to improve the resolution of adjacent peaks. One method of narrowing analyte zone bandwidth is termed focussing. A n analyte is said to be focussed if its bandwidth at the detector is narrower than its injected bandwidth. Focussing is a way of improving the detection limit of methods as well, because it allows for large volume injections. Normally injection volumes > 1 % of the capillary volume are unfavourable, due to band spreading 2 6 . However, manipulation of R = (1.20) 20 the sample matrix with respect to the B G E can result in an on-line concentrating effect, known as focussing. 1.3 Aims of Research The aim of this work was to gain more understanding about the p H of solutions within the capillary, in the presence of a multi-section electrolyte, in which p H junctions exist. In addition, the influence of the migration rates of proton and hydroxide ions within the capillary was studied. The first portion of this project undertook to find out more about the nature of an on-capillary p H junction that was created using a multi-section electrolyte. The nature of 1 and 2 unit p H junctions were evaluated using pH-indicating dye molecules to visualize the p H junctions as they passed the capillary detector window. Secondly, pH-indicating dyes were used to determine the mobility of hydroxide ions and protons under an electric field. Theoretical values for proton and hydroxide ions are available, and the values obtained using C E were comparable. The influence of other factors such as ionic strength, and p H on the mobility of the proton and hydroxide ions was also evaluated. The last part of this project deals with the development of a robust and qualitative assay for the analysis of novel new compounds called dendrimers. Method development involved optimizing C E parameters in order to obtain the best resolution, reproducibility and simplicity of analysis. Information on the mobility of these compounds in electrophoresis gives some insight into their structure or shape when in solution. 21 2 On-capillary pH measurement 2.1 Introduction In C E , changes in p H can affect the ionization state of analytes, electrolytes and buffer additives, as well as modify the E O F through modification of the charge layer (a) at the capillary wall 2 6 . Thus p H plays a major role in manipulating the selectivity of capillary electrophoretic separations. However, not much is known about the p H of solutions inside the capillary itself. p H measurements of bulk electrolytes and sample solutions performed in the traditional manner using glass membrane indicator electrodes may not reflect the actual p H inside the capillary once a separation begins. Differences in the composition of sample and background electrolyte (BGE) zones, as well as the evolution of electrolysis products at the inlet and outlet ends of the capillary, may significantly affect the on-capillary pH. Because the volumes of liquid inside the capillary do not lend themselves to traditional p H measurement techniques, other methods must be employed. Ratiometric dyes are compounds whose excitation and/or emission maxima change in response to changes within the dye's surrounding environment. Within the field of analytical cytology, ratiometric dyes are used to measure the cytoplasmic concentration of hydrogen ions and free calcium ions in living cells 3 2 . Recently, ratiometric dyes have been applied to measure on-capillary p H in the field of capillary electrophoresis 3 3 . Ratiometric dyes make it possible to visualize the p H at the capillary detector window and thus get a representation of the p H within the capillary at different time points during the separation process, as components pass this window. Previously, Timperman et al performed on-column p H monitoring with an application to volume-limited outlet vials 3 3 . A fluorescent ratiometric molecule, carboxyseminaphthorhodafluor (C .SNARF) , was 22 used to measure the p H changes within the capillary that resulted from electrolysis occurring at the cathodic end of a low-volume capillary outlet. Common pH-indicators also have ratiometric properties, although they are generally non-fluorescent. Macka et a l 3 4 used the p H indicators xylenol blue and bromescrol green to determine the effects of electrolysis at the anode and cathode on the p H of the electrolyte in the capillary. They measured how the distance between the electrode and capillary inlet/outlet affected the uptake of proton and hydroxide ions into the capillary when using unbuffered electrolyte. They also visualized p H changes within the inlet and outlet vials using a special photometric cell. In another application Corstjens and co-workers evaluated on-capillary p H from the mobility, peak height and peak area of an indicator dye 3 5 . Until now, the focus of on-capillary p H measurements has been to measure the effect of the electrolysis products on the p H within the capillary. The p H within a self-prepared p H gradient or multi-section electrolyte system has not been evaluated, although Huang et al did use methyl red indicator and whole column imaging detection to measure the existence of a p H gradient created from the electrolysis of water 3 6 . While the existence of the gradient was confirmed visually by examining colour changes of the methyl red dye, the precise p H values within the capillary were not evaluated. Multi-section electrolyte systems involve filling the capillary with different zones or regions of electrolyte that differ from one another in terms of concentration, buffer additive or p H value. Palmer and Landers used high salt concentrations within the sample matrix to create a single boundary, discontinuous electrolyte system that caused stacking (focusing) of neutral analytes 3 7 . Quirino and Terabe pioneered the technique of sweeping, in which a large volume sample zone containing no surfactant is sandwiched 23 between a leading B G E , which contains micelles, and a trailing B G E that does not. This technique resulted in a million-fold sensitivity increase over non-sweeping forms of micellar electrokinetic chromatography ( M E K C ) 3 8 - 3 9 . The first documented use of a p H -junction to improve concentration sensitivity was by Aebersold and Morrison in 1990, who improved the concentration sensitivity of peptides by preparing them in a high p H sample matrix and analyzing them in a p H 2.5 citrate buffer 4 0 . Another application of a multi-section electrolyte system in which a pH-junction occurs is velocity-difference induced focussing (V-DIF), which was proposed by Britz-McKibbin et a l 4 1 . Britz-McKibbin et al. achieved an improvement in the concentration sensitivity of zwitterionic catecholamines and weakly acidic compounds by preparing the sample solution at a different p H than the background electrolyte (BGE) in order to exploit the mobility differences between the various ionization forms of the analytes. Large volume sample injections were performed, resulting in a multi-section electrolyte system containing two p H junctions where the sample and B G E zones met. Although this technique has been successfully applied to focussing a variety of compounds, the nature of the focussing mechanism is not well understood. It was proposed that when the sample matrix was prepared at low pH, hydroxide ions from the higher-pH surrounding-BGE would quickly enter the lower-pH sample zone, resulting in ionization of the analytes and resultant differences in analyte velocity between "invaded" and "non-invaded" regions of the sample zone. This mechanism suggests that focussing occurs due to stacking of analyte molecules at the mobile p H junction. The junction then dissipated, allowing for a normal separation of the analyte bands from one another. 24 In the work presented in this chapter, a ratiometric dye was added to a multi-section electrolyte system in which a 1 or 2-unit p H junction exists. The dye was then used to monitor the p H junction in order to determine how it behaves under an electric field, and to see if the mechanism predicted by Britz-McKibbin is plausible. 2.2 Experimental 2.2.1 Chemicals and Reagents The ratiometric p H indicator dye 8-hydroxypyrene-l,3,6-trisulfonic acid (HPTS or pyranine) was purchased from Molecular Probes (Eugene, OR; Catalog number H-348). The aqueous B G E consisted of 40 m M borate (Borax/Boric Acid, Sigma-Aldrich Canada Ltd. Oakville, Ont.) and was prepared at p H 8.0, 8.5 and 9.0 by varying the borax and boric acid amounts. Where necessary, the p H of the B G E and sample solutions was adjusted using 1.0 M HC1 (BDH, Toronto, Ont., Canada). 1.0 M N a O H (BDH) was used to condition the capillary wall. For ratiometric dye p H experiments, when dye was required in the sample or B G E zone, it was prepared at a concentration of 0.1 m M in the appropriate p H buffer. HPLC-grade methanol was purchased from Fisher Scientific (Nepean, Ont., Canada). Deionized, distilled water was used in all solutions. 2,2-dimethyl-1-phenol-1-propanol (DPP) was obtained from Sigma (Sigma-Aldrich Canada Ltd.). DPP was prepared in 2% MeOH/40 m M Borate (v/v) at a concentration of 0.4 m M . 2.2.2 Apparatus and Procedure Separations were performed on a P/ACE™ M D Q automated capillary electrophoresis system (Beckman-Coulter Inc., Mississauga, Ont., Canada). Uncoated capillaries (Polymicro Technologies, Phoenix, A Z ) were used with inner diameters of 75 pm, outer diameters of 375 pm, and length of 30 cm to detector (L d ), 40.2 cm total (Lt ) . New 25 capillaries were first rinsed with 1.0 M N a O H (2 minutes, 20 psi), deionized water (5 minutes, 20 psi) and then B G E (2 minutes, 20 psi) and used immediately. Each separation was preceded with a rinse of 1.0 M N a O H (1 min, 20 psi), deionized water (2 min, 20 psi), and B G E (2 min, 20 psi). The sample was introduced using a 0.5 psi pressure injection for 120 seconds. Separations were carried out at 20 k V under a controlled temperature of 2 0 ° C . The ramp time for the voltage to reach 20 k V was 0.17 minutes. Absorbance detection was performed with a Beckman M D Q photodiode array (PDA) detector over a wavelength range of 250-600 nm, with extracted electropherograms at 405 and 454 nm. Off-capillary bulk solution p H measurements were made with a Beckman O10-pH meter equipped with an Orion combination p H glass electrode (Thermo Orion, Beverly, M A P.N. 915600). 2.3 Results and Discussion 2.3.1 Choice of ratiometric p H probe The ratiometric dye HPTS (8-hydroxypyrene-l,3,6-trisulfonic acid) was chosen for these experiments because it is relatively inexpensive, water soluble, and has an excitation maxima that changes with H + ion concentration, enabling the use of a UV-vis absorbance detection system (405 and 454 nm using a P D A detector). The chemical structure of HPTS is shown in Figure 2-1. (-)0 3S- ^ ^SOa(-) pH < 7.6, net charge -3 (-)0 3S- ^ ^S03(-) pH > 7.6, net charge -4 Figure 2-1. Structure of HPTS molecule. Two ionization forms of dye are shown. The excitation maxima of HPTS dye shifts from 405 nm to 454 nm when the dye is deprotonated 4 2 . The excitation (absorbance) spectra of HPTS at various p H values are shown in Figure 2-2. The isobestic point at 417 nm is the wavelength at which the dye absorbance is independent of p H changes and dye concentration can be measured in the absence of p H effects. 27 0.01 M NaOH 350 400 450 Wavelength (nm) 500 Figure 2-2. p H dependent spectra of HPTS 4 3 Using absorbance measurements at two wavelengths, 405 nm and 454 nm, the p H value of a solution that contains the dye can be calculated 4 4 using Equation (2.1). pH = pKa + log (R-RA) (2.1) Where R is the ratio of dye absorbance at 405 and 454 nm ( A 4 0 5 / A 4 5 4 ) . R A and R B are the limiting values of the ratio at the acidic and basic endpoints of titration. R A and R B were determined to be 1.49 and 0.31 respectively at p H values of 6.1 and 9.2, when measured in 40 m M Borate buffer. The calculated and measured p H values of four, 40 m M Borate buffers, are shown in Table 2-1. The calculated p H differs slightly from the measured p H (positive bias of 0.3 p H units) but in general gives a fairly accurate representation of p H within the capillary. 28 Table 2-1. Comparison of measured vs. calculated p H p H (measured) Ratio A405/A454 p H (calculated) 7.3 0.95 7.5 7.6 0.68 7.9 8.3 0.41 8.6 9^ 2 033 9A 2.3.2 Multi-section electrolyte Burgi and Chien first identified a sample stacking effect that occurs when samples are prepared in a high resistivity (low ionic strength) environment 1 6> 1 8. In a multi-section electrolyte system prepared from components with different ionic strengths, the electric field strength will vary from section to section within the capillary. Because the p K a of borate is 9.2, the ionic strength of borate buffer will be lower at p H 8 than p H 9. Lower ionic strength is equated with greater resistivity, and hence higher electric field strength. The velocity of an ion in a region of higher electric field will be greater than the same ion in a lower electric field, from Eqn.(1.5). For the same reasons, the E O F will also be greater in regions of higher electric field 1 1 . In the p H experiments performed here, a 120 s, 0.5 psi injection into the 40.2 cm total length (L T ) capillary fills 47 % of the capillary volume. The % fill was calculated from 2 6 APr 2 t . % capillary filled = f- x 100 (2.2) Where AP is pressure, t i n t is the introduction time of the sample, T| the viscosity, r the radius of the capillary and L T the length to the detector. The presence of large volumes of different p H borate buffers within the same capillary results in field strength differences between sections. The net E O F in a multi-section electrolyte capillary system is a weighted average of the EOFs in the different regions of electrolyte within the capillary 1 1 . The presence of field strength differences will affect the migration velocity of H P T S molecules in this multi-electrolyte system. The E O F mobility and current (which is proportional to ionic strength) for p H 8 and p H 9 borate buffers are given in Table 2-2. Table 2-2. E O F mobility and current in 40 m M borate buffer Buffer p H E O F mobility ( 10~4cm2/Vs) 8 7.4 9 6.0 Current (uA) 5.4 32.7 For a singly charged ion in a multi-section electrolyte prepared from the buffers shown in Table 2-2, the ion velocity would be greater in the p H 8 sections, than the p H 9 sections, provided the charge (hence mobility) of the analyte did not change. However, p H differences can affect the ionization state of analytes. In summary, within a multi-section electrolyte (pH difference) system two forces affect the overall velocity of the analyte; 1. electric field strength differences will increase analyte velocity in regions of low ionic strength, decrease analyte velocity in regions of high ionic strength 2. ionization equilibria of analytes will affect their charge state (q) and hence mobility (Eqn. (1.11)) in regions of differing pH. 30 2.3.3 Buffer vial replacement As was evident from previous work by Corstjens and Macka 3 4 - 3 5 , the p H of the inlet and outlet vials in C E can be affected by electrolysis of water in contact with the anode and cathode. In order to avoid complications caused by such p H changes, preliminary experiments were performed to determine the best experimental procedure. During these preliminary experiments it was noted that the inlet vial, containing buffer prepared with dye, often changed colour from yellow (basic) to green (acidic) during a separation. This occurs because of proton production at the anode (inlet). Electrolysis at the cathode and anode will result in an increase in p H at the outlet and a decrease in p H at the inlet, when operating the system in "normal" polarity mode. The redox reaction occurring at the (inlet) anode is 2 H 2 0 ->02+4H++4e" (2.3) The reaction occurring at the (outlet) cathode is 2H 2 0+2e"^H 2 +20H" (2.4) The ability of the buffer to accommodate excess ions will depend on its concentration, p K a value and the p H at which it was prepared. Generally, to obtain reproducible results in C E analysis, buffer vials should be replenished frequently, but there are no specific guidelines as to how frequently, especially with low volume inlet/outlet vials. Low volume inlet/outlet vials are gaining popularity in C E because less solution is required to fill them. This is advantageous when expensive electrolytes and buffer additives such as substituted cyclodextrins are used. The number of runs that are performed from one pair (or set) of buffer vials varies from user to user. Some advocate switching buffer vials with every run, especially if the analysis time is long, or a low concentration buffer is used. The use of 500 u L inserts as buffer vials is widespread. Disposable inserts are cheap, widely available, and easy to prepare by cleaving the lid from 500 u L Eppendorf tubes. They can hold anywhere from 200-450 u L of buffer and still maintain contact with the capillary and electrode. Time consuming washing and rinsing of reusable vials is eliminated, as well as any potential for contamination from improperly cleaned vials. In order to determine the effect of proton production at the inlet in this buffer system over repeated runs, 10 analyses were performed using the same pair of 1.6 m L buffer vials. The buffer in the inlet vial contained 40 m M Borate, p H 9.2 and 0.1 m M of H P T S dye. There was no dye in the outlet vial, although dye from the outlet would not be expected to migrate back into the capillary due to the presence of the E O F . The borate was prepared at its p K a value for maximum buffering capacity. Each analysis was conducted at 20 k V for 30 minutes. The A405/A454 ratio was obtained at 15 minutes into the separation and the p H of the solution in the capillary was calculated using Eqn. (2.1). Results for these analyses are shown in Table 2-3. 32 Table 2-3. Reproducibility of p H over 10 analyses using the same 1.6 m L buffer vials Replicate Calculated p H 1 9.2 2 9.2 3 8.2 4 8.2 5 7.8 6 7.8 7 7.8 8 7.6 9 7.6 10 7.4 As seen from the results in Table 2-3, p H drops over a range of 1.8 units from the first to last analysis. In contrast, the same experiment was conducted using 400 u L inserts, in which the inlet and outlet buffer vials were changed for new ones with each run. The data are shown in Table 2-4. Table 2-4. Reproducibility of p H over 5 analyses with new buffer vials each run Replicate Calculated p H 1 8.6 2 8.6 3 8.4 4 8.5 5 8.6 33 The buffer p H is reproducible (range of 0.2 units difference) when the vials are changed with each run, which allows for the use of smaller buffer volumes. This data stresses the importance of changing buffer vials with each run, especially when using reduced-volume inlet and outlet vials. 2.3.4 Control experiments: p H junction Experiments were undertaken to evaluate the nature of the p H junction described by Britz-McKibbin et al 4 1 in a multi-section electrolyte system. A series of control experiments were performed, in which both the B G E and the sample zones were prepared at the same known pH. The sample zone was injected for 120 seconds at 0.5 psi and a 20 k V potential applied across the capillary. The sample was either blank (with no dye), or contained 0.1 m M dye. Figure 2-3 illustrates the control experiments [a] anode (+)  H sample zone electroosmotic flow [b] anode(+) a. CD o CL o CD a. o cathode (-) cathode (-) Figure 2-3. p H 8, p H 9 control experiments using multi-section electrolyte. Shading represents HPTS presence [a] sample region is blank, B G E contains HPTS [b] sample region contains HPTS 34 The calculated p H values for these experiments can be found in Table 2-5. Table 2-5. Multi-section electrolyte experiments: controls Measured buffer p H Dye located in Calculated p H pH8.0 sample zone 7.9 pH8.5 sample zone 8.7 p H 9.0 sample zone 9.6 pH8.0 background electrolyte 7.9a, 7.9 p H 8.5 background electrolyte 8.7, 8.8 p H 9.0 background electrolyte 9.4, 9.7 a first value for front portion of BGE, second for the back portion of BGE From Table 2-5, the calculated p H agrees with the p H measured using the p H meter. In the cases where the sample zone was left blank, and the B G E contained the dye, the back B G E region is the area of B G E found after (behind) the sample zone. The front B G E region is the part in front of the sample zone, it would be expected to reach the detector first. In the controls, there was no variation in p H within regions of dye, as shown in Figure 2-4, which contains the electropherograms from control experiments with p H 8 and 9 buffers. 35 0 2 4 6 . 8 10 Time (minutes) Figure 2-4. Control experiments, dye in sample zone. Detection wavelength = 454 nm. In Figure 2-4 the absorbance of the dye is greater in the p H 9 experiment, despite the two dye solutions being prepared at 0.1 m M . The difference in absorbance is due to the difference in absorbance between the two ionization forms of the dye. The dye moves more quickly to the detector in the p H 8 experiment, due to the greater magnitude of E O F in p H 8 40 m M borate buffer. From these experiments the HPTS molecules were observed to migrate out of the true sample zone due to their large negative charge. This property of the dye makes it difficult to measure the p H of the entire sample region. However, as the sample zone was very large (47% of capillary) the molecules did not have time to completely migrate free of the sample zone. Thus p H can still be measured, but it requires some effort in interpreting the results. The parameters of the true sample zone had to be evaluated using a neutral 36 compound, 2,2-dimethyl-l -phenol- 1-propanol (DPP). The true sample zone parameters were determined with an injection of neutral sample swept to the detector with the E O F . The results of these control experiments indicate that on-column p H measurements can be performed accurately using the ratiometric dye HPTS. 2.3.5 1-unit p H junction In the next set of experiments, a one-unit p H junction was created between the sample zone and the B G E . Figure 2-5 shows a diagram of the experiments conducted. Figure 2-5. Diagram of 1 unit p H junction experiments. p H junctions exist where sample zone interacts with B G E regions (patterned). Dye was included in either B G E , or sample zone (Table 2-6). H P T S was included in either the sample zone, or the B G E . When HPTS was included in the B G E , the sample zone was left blank. For ease of discussion, p H junction descriptions will be documented as p H 8:9:8, where the underlined number indicates that H P T S was included in that region. Table 2-6 describes the experiments. electroosmotic flow o. 2. anode (+) cathode (-) 37 Table 2-6. 1 unit pH-junction experiments Experiment Description Sample zone B G E A p H 8:9:8 p H 9, contains HPTS p H 8, blank B p H 9:8:9 p H 8, contains HPTS p H 9, blank C p H 8:9:8 p H 9, blank p H 8, contains H P T S D p H 9:8:9 p H 8, blank p H 9, contains H P T S Figure 2-6 shows the electropherogram obtained for experiment A . The results indicate that the first HPTS molecules to reach the detector remained at p H 9, with the remainder of the dye molecules at p H 8. 38 Figure 2-6. Electropherogram of p H 8:9:8 junction. Detection wavelength = 454 nm. Figure 2-6 shows that the HPTS region is no longer uniform in signal intensity or p H value, as it was during the control experiments (Figure 2-4). It also indicates that a portion of the multi-section electrolyte within the capillary is still at ~ p H 9. It is important to note that conclusions can only be drawn for the portion of the sample zone in which the HPTS still resides. Experiments conducted with neutral analyte DPP show that in this format of multi-section electrolyte, the true sample zone passes the detector window from 0.47 - 1.28 minutes. In Figure 2-6, 0.65 minutes marks the front of the measured p H 9 region. The back of the p H 9 region is at 1.32 minutes. The p H 8 portion of the HPTS zone extends back to 4.98 minutes. These numbers indicate that part of the original sample zone now contains no HPTS (0.47-0.65 minutes). From 0.65 - 1.32 39 minutes, the sample zone is p H 9 (a slight increase over the width of the true sample zone). Then, the rest of the HPTS at p H 8 has migrated out of the sample zone into the surrounding p H 8 B G E . The p H junction appears to be intact, and is quite sharp, judging by the slope of the transition between the p H 9 and 8 regions. The p H of the measurable portion of the sample zone has been retained, and the junction appears to have shifted slightly backwards, possibly due to OH" ion migration toward the anode. The differences in peak height between the p H 9 and p H 8 regions are real concentration differences, verified by evaluating the absorbance at 417 nm, the isobestic point of the dye. From 0.65-1.32 minutes, the signal height is 7.3 mAu, from 1.32-4.98 minutes, the signal height is 1.1 mAu (measured at 417 nm). When compared to the p H 9 control experiment, the dye in the "focussed" region of HPTS in Figure 2-6 has a signal height approximately four times greater than the control experiment. The changes in dye concentration can be explained by the ionization equilibrium of H P T S , which is dependent on pH, and by the differences in electric field strength within the different sections of electrolyte. The p K a of HPTS is 7.6. When prepared in p H 8 buffer, the HPTS will be 72% ionized; for example, in a 0.1 m M solution of the dye, 0.072 m M will have a -4 charge, while 0.028 m M will have a -3 charge. At p H 9, the HPTS will be 96% ionized, discounting any effects that the buffer composition will have on the p K a of H P T S 4 5 . Upon first consideration, one might expect that the -4 species would migrate more swiftly toward the anode (inlet) than the -3 species. However, it is also necessary to 40 examine the electric field in the respective regions of the capillary. Two effects are at work, the reduction in negative charge on the molecule in the p H 8 region, and the increase in electric field strength in the p H 8 region. It appears that the two effects contribute to both the focussing of the p H 9 region, as well as the long drawn out portion of dye molecules in the p H 8 region, as illustrated in Figure 2-7. Once the electric field is applied, the net movement of all species in the capillary will be towards the detector (cathode) due to the E O F . However, the effective mobility of the HPTS molecules is negative, due to their negative charge (q). The HPTS molecules will migrate towards the back of the sample zone, away from the detector, while the sample zone itself is moving forward, with the E O F . Some HPTS molecules will approach the p H junction at the back of the sample zone. Here, their charge state may change from -4 to -3, resulting in a reduction in effective mobility (pep), and causing a bottleneck of HPTS molecules at the p H junction. This bottleneck results in the focussing of the molecules trapped behind the p H junction. Once free of the bottleneck however, the majority of the HPTS molecules will still have a -4 charge, and within the p H 8 B G E , their velocity will increase due to the greater local E field strength, resulting in the long drawn out band of H P T S seen in Figure 2-6. 41 [a] [b] pH 9 sample zone mmmm pH 8 BGE 1 ^ pH junctions ' i [c] + -3 ^3" 1-4 4-4 --•* ^ ft .4 '"4 ,,N"4 "4 electroosmotic flow -3 , -3 o "4 -3-3 -3" -3 -3 electroosmotic flow pH 8 BGE 4 o i 4 Figure 2-7. Proposed mechanism of HPTS movement and focussing in capillary, [a] p H 9 sample zone contains -4 charge HPTS. [b] electric field is applied, some H P T S migrates out of sample zone, net flow is to detector with E O F [c] focussing occurs at back p H junction. In experiment B , the opposite to A (Table 2-5), a p H 9:8:9 junction was created. The electropherogram in Figure 2-8 indicates that the entire region of HPTS molecules is at p H 9 when it reaches the detector, because all of the HPTS molecules have left the sample zone. The area in which the HPTS molecules migrate is from 1.61-3.88 minutes, which is well outside of the true sample zone defined using a neutral analyte (DPP). In this case, the -3/-4 charged HPTS molecules migrate backwards out of the sample zone into a p H 9 environment, where all of them attain a -4 charge, which increases their negative mobility, allowing for free movement and no bottleneck of slower-moving molecules. Because the electric field strength is greater in the p H 8 zone, the migration out of this region occurs quickly, which should result in some focussing at the p H junction. Indeed, the peak is somewhat focussed when signal intensity is compared to the p H 9 control (Figure 2-4). The signal intensity for the junction experiment is 80 mAu, compared to 18 mAu for the control. 42 Absorbance (arbitrary) 1 1 1 1 J \ 1 i i i i 1 i i i i | i i i i 1 i i i i | i i i i 1 i i i i | i i i i 1 i i i i | i i i i 1 i i i i | D 2 4 6 8 10 •Time (minutes) Figure 2-8. p H 9:8:9 electropherogram. Detection wavelength is 454 nm. In experiments C and D, dye was included in the B G E , while the sample zone was left blank, in order to evaluate the sample zone p H as the dye moved into that area. The first p H junction tested in this manner was p H 8:9:8, experiment C . 43 pH = 9 pH = 8 —'l i i i r I r i i i I i i i I I I i i i | i i i i | i i i i | i i i i i i i i i | i i i i i i i i i | 0 2 4 6 8 10 T i m e ( m i n u t e s ) Figure 2-9. Electropherogram of p H 8:9:8 junction, sample zone is blank. Wavelength shown is 454 nm. The results indicate that the front portion of HPTS is at p H 9, and the back portion is at p H 8. The front portion of HPTS molecules was considerably focused, although the back portion was uniform in height (not focused). Figure 2-10 details the mechanism of focussing and indicates how the front portion of the dye has migrated into the p H 9 blank sample zone, indicating that the front of the sample zone has maintained its p H value upon reaching the detector window. 44 pH 8 BGE [a] -3" pH 9 sample zone J \ pH junctions i [b] pH 9 sample zone H 8 « pH junctions a. [C] + pH 9 sample zone Figure 2-10. p H 8:9:8 multi-section electrolyte experiment, [a] dye is contained in shaded regions. Volume next to and including detector is still at p H 8, but contains no dye to prevent detector auto zero from occurring with dye in capillary window, [b] separation voltage applied [c] at detector, front portion of dye is focussed on front p H junction and has p H 9 character. The front focussed portion of HPTS has a migration time of 0.59 minutes, which shows that those dye molecules are within the true sample zone. This confirms that the H P T S molecules have migrated into the blank sample zone and that the front of the sample zone is still at p H 9 in this multi-section electrolyte system. The focusing of this portion of the HPTS molecules is most likely due to a field-amplified stacking effect caused by differences in the electric field between the p H 8.0 and 9.0 electrolytes. The reason that a field-amplified stacking effect occurred here, and not in a similar situation with p H 9:8:9 junction, is due to the smaller size of the HPTS zone in the front B G E portion ( 1 minute injection instead of 2 minute). And, as it was injected ahead of the blank sample zone, the front HPTS portion spent less time in the capillary. There was not 45 enough time for the focussing gained by the field-effect to dissipate prior to the peak reaching the detector window. The final 1 unit pH-junction experiment involved the junction 9:8:9, experiment D . The electropherogram contained zones of dye at both p H 8 and p H 9, Figure 2-11 < | i i i i—| i i i i | i i—i i | i i—i i | i i i i | i i i i | i i i—i | i i i—i | i i i i | i—i i i | 0 2 4 6 8 10 Time (minutes) Figure 2-11. Electropherogram of p H 9:8:9 junction. Detection wavelength = 454 nm. p H values of various regions are indicated on figure. In Figure 2-11, the p H in the "gap" between the two "peaks" is ~8 (calculated 7.8). It appears that HPTS molecules have migrated completely into the blank sample zone, but only to a minor degree. The gap (of lower concentration) is from 0.71 to 1.42 minutes in length, shifted slightly backwards from the position of a neutral sample component (0.45 to 1.32 minutes). In front of the sample zone, the HPTS molecules remain at p H 9, and are not focussed. There is a large depletion of dye from - 2 - 4 minutes, at 417 nm, this is 46 a less concentrated zone of dye. The reason for this area of depletion is not known, although the effect is reproducible. It may have to do with other negatively charged ions from the sample zone migrating backwards. The results from this experiment indicate that the p H 8 sample zone also remains intact to some degree, along with the p H character of the surrounding B G E zones. It also looks like the p H junction is shifting backwards, which would be due to the backwards mobility of hydroxide ions. From the experiments conducted (A, B, C and D) the following observations have been made. Firstly, any experiment to measure the p H junction requires careful construction, as the HPTS tends to migrate out of the sample zone due to its negative effective mobility. Focussing effects were observed when the HPTS molecules moved from regions of high electric field (pH 8) to regions of low electric field (pH 9). This focussing was sometimes confounded by the fact that the dye was further ionized at p H 9. Dye velocity is greater in p H 8 buffer, due to differences in field strength, and the reduction (by ~ 20%) in the ionization of the dye that occurs in p H 8 buffer is not enough to counteract the increase in velocity due to field strength. There is evidence that for large volume ( > 47% of capillary volume) sample sizes, with a 1-unit p H gradient in 40 m M borate buffer, the p H of most of the sample area remains intact. From the changes in p H visualized using the dye ratio, it appears that all of the p H junctions that were seen were very "sharp", although the time frame within the capillary was < 2 minutes, and there was not much time for diffusion to occur. In the p H 9:8:9 case, the junction also appears to move backwards relative to the true sample zone defined by a neutral analyte. 47 2.3.6 2-unit p H j unction The V-DEF experiments conducted by Britz-McKibbin utilized a p H difference of > 1.9 units 4 6 . For this reason, the experiments outlined in the previous section (2.3.5) were also conducted using p H 7 and 9 borate buffers. 40 m M Borate buffer was adjusted to p H 7 using 1.0 M HC1. The E O F mobility in the p H 7 buffer is 5.2 x 10"4 cm 2 /Vs, compared to 6.0 x 10"4 cm 2 /Vs for p H 9 buffer. The p H 7 electrolyte has greater ionic strength, due to the contribution of H + and CI" ions from the added HC1, this was confirmed by measuring the current in p H 9 buffer and p H 7 buffer. The currents were 32.7 and 65.2 u A , respectively. Within the p H 7 electrolyte, the E O F would be reduced due to a reduction in ionization of capillary wall silanol groups, as well as the increased ionic strength of the electrolyte. A comparison of the electropherograms from two p H junctions, 1-unit and 2-unit, is shown in Figure 2-12. 48 4 6 Time (minutes) Figure 2-12. [a] is the p H 8:9:8 experiment also shown in Figure 2-6. [b] is the same experiment, with a 2-unit p H junction ( p H 7:9:7). From Figure 2-12b, the HPTS peak has a calculated p H of 7, and is concentrated into a narrow zone. The -4 charge HPTS molecules from the p H 9 sample zone have migrated to the back of the sample zone, where they encountered the p H junction. The p H 7 region is lower field than the p H 9 region, resulting in a reduction in velocity of the analytes. In addition, at p H 7, 98% of the HPTS molecules will have a -3 charge, which will reduce their effective mobility. These two effects result in focussing of the analyte zone, which will then move into the p H 7 B G E and as the separation proceeds. When the same experiment in. Figure 2-12 was conducted with a neutral sample, DPP, the sample zone migrates from 0.59 to 1.68 minutes, well before the HPTS molecules in Figure 2-12b. The neutral sample zone is much wider than the peak in Figure 2-12b (1.09 minutes vs. 49 0.31 minutes), indicating that focussing of the dye molecules has occurred. The net E O F in the capillary is slower than in previous experiments, leading to the late migration time of the H P T S peak. The p H character of the sample zone cannot be evaluated as no H P T S remains in the sample zone, however, the p H of the B G E in which the H P T S resides does not appear to have been affected by the large section of p H 9 B G E that precedes it. In the related experiment, p H 7:9:7, all areas containing dye in the electropherogram were at p H values < 7. The front dye region, at p H 7, from time measurements has migrated into the true sample zone, but its pH is 7. It appears that the p H junction has moved, or the presence of p H 7 buffer has influenced the p H of the sample zone. This result was the opposite of what was observed for a similar experiment with a 1-unit p H junction (pH 8:9:8, Figure 2-9), although the two electropherograms look similar, Figure 2-13. 50 Time (minutes) Figure 2-13. Electropherogram of p H 7:9:7 junction. Detection wavelength 454 nm shown. From Figure 2-13, there is also some shifting of dye concentration behind the sample zone, which was not seen in the electropherogram for the 1 unit p H junction of the similar experiment. In the p H 8:9:8 experiment, the dye signal intensity was uniform behind the sample zone. Potentially, the back p H junction has moved backwards with time, and has overtaken the HPTS molecules, which obtain a -4 charge and then run into -3 charge H P T S from the area which has not been overrun by the p H junction. Although, if this scenario were true, we would expect the dye to be at p H 9 in the focussed part, but it is at p H 7. In the next experiment, p H 9:7:9, junction, all of the dye is at p H 9, and well out of the sample zone, indicating that all HPTS has migrated free of the sample zone. The dye 51 is sharply focussed at ~2 minutes. After that point the dye extends back away from the focussed zone, still at p H 9. Moving from the p H 7 sample zone to the p H 9 B G E , the dye should have experienced an increase in velocity, due to ionization from -3 to -4, along with higher field strength in p H 9 B G E . From these effects, one would predict a long broad zone of sample with no focussing. In Figure 2-14, the analyte zone is long and broad, but there is an unexplained focussing of the dye at ~ 2minutes. Absorbance (arbitrary) 1 1 1 1 1 r i i i i I i l i i 1 i l i i | i i i i | i i i i | i i i i | i i i i | i i i i | i i i i 1 i i i i | 3 2 4 6 8 10 Time (minutes) Figure 2-14. Electropherogram of p H 9:7:9 junction. Detection wavelength shown is 454 nm. Finally, in the p H 9:7:9 experiment, all areas containing dye are at p H 9 once again, and the front portion of the HPTS has focussed into a sharp peak, unlike the 9:8:9 case, there is no dye in the area between the focussed region and the back dye region from the B G E . The electropherogram from this experiment is very similar to that of the 9:7:9 experiment shown in Figure 2-14. 52 i i i i i i i i I i i i i i i i i i I i i i i i i i i i I 0 2 4 6 Time (minutes) I i i i i | i i i i I i i i 8 10 Figure 2-15. Electropherogram of p H 9:7:9 junction. Detection wavelength shown is 454 nm. In all of the experiments using the 2-unit p H junction, the ratiometric p H calculated from the H P T S molecules tended to match the p H of the B G E , e.g. in the p H 7:9:7 experiments, all areas containing HPTS were p H 7, and in all cases, the dye had migrated free of the sample zone. In the p H 9:7:9 and 9:7:9 experiments, the lowest dye p H was 8.2, the greatest 9.2, and the HPTS was not found in the sample zone (defined using a neutral compound). It appears that the p H 9 buffer had some influence on the p H 7 buffer, possibly due to free hydroxide entering the p H 7 buffer zones and causing the p H to change, which was the mechanism of focussing proposed by Britz-McKibbin et al. 2.4 Conclusions The p H of a 1-unit p H junction was measured on-capillary using a ratiometric p H -indicating dye, H P T S . The junction was found to change position slightly in relation to its 53 original boundaries, but the two p H regions created with a multi-section electrolyte system remained distinct through the detection window within a time frame of < 2 minutes. With the 2-unit p H junction, it was not possible to measure the p H over actual junctions, as the dye molecules migrated free of those areas prior to being measured. The E O F of the p H 7 electrolyte was slow, resulting in increased migration times and more time spent on-capillary, which led to greater migration of HPTS from the sample zones. The experiments conducted by Britz-McKibbin employed 160 m M Borate buffer, the high concentration of borate resulted in a very slow E O F . As evidence of p H junction "movement" occurred in the experiments conducted here, and as more movement would occur with a longer residence time in the capillary, it is conceivable that a sample zone of different p H would have dissipated by the time it reached the detector window, given enough time. 3 Proton and Hydroxide Mobility 3.1 Introduction i Prior to the introduction of capillary electrophoresis (CE) in the early 1980s, ion mobilities were determined from electrical conductivity and ion transport number measurements. Extensive work in the early 19th century established absolute mobility values in aqueous solution for many common ions, such as sodium, potassium, acetate and perchlorate 2 7 ' 2 8 and these values are shown in Table 3-1. Table 3-1. Mobilities (IO -4 cm2/Vs) of Representative Ions in Water at 25°C 4 7 Cation Mo Anion Mo H + 36.25 OH" 20.50 L i + 4.01 cr 7.91 Na + 5.19 B f 8.10 K + 7.62 r 7.97 Rb + 8.06 N03" 7.40 Cs + 8.01 cio 4- 6.98 C H 3 N H 3 + 6.08 CH 3 COO" 4.24 M g " 5.50 CH 3 (CH 2 ) 2 COO" 3.38 From this table of data, the values for the proton and hydroxide ion are significantly larger than any other ion. The "prototypic process of conduction" of these specific ions explain the abnormally high mobilities Under an electric field, a proton will pass from one water molecule to a favourably oriented neighbouring water molecule, causing the water that just lost its proton to be unfavourably oriented for another transfer of the same proton 4 8 . A related mechanism can also be used to account for the high mobility of the hydroxide ion, Figures 3-la and 3-lb. 55 cathode (-) Figure 3-1. Prototypic conduction process of proton and hydroxide ions, [a] Proton conduction under an electric field, [b] hydroxide ion conduction under an electric field The proposed mechanism of proton conduction is based on the tetrahedral arrangement of hydrogen ions around each oxygen atom in aqueous solution. Two hydrogens will be close (<0.95 A away) and the other two, which belong to other oxygen atoms will be further away (<1.81 A). Ionization results when one of the hydrogens that is far away, moves a distance of 0.86 A closer to the central O atom, resulting in a central O with three hydrogens a distance of 0.95 A away surrounding it. The oxygen that just donated a hydrogen is left with one close, and three far away hydrogens. When a potential is applied, this process results in the rapid transfer of positive (or negative) charge 4 8 . Previous values of absolute proton and hydroxide ion mobility found in the literature are based on conductance measurements 1 . It is important to verify if these values apply in C E systems. In addition, the conductivity and mobility data in the literature refer to ionic strengths originating from the ions themselves. In C E , the analyte of interest is surrounded (generally) by a high concentration of buffer or electrolyte. While the mode of C E termed isotachophoresis (ITP) has been used to measure the mobility values of many different ions 1 3 , proton and hydroxide ions have not been evaluated using this 56 technique, due to their large mobility values, which makes it impossible to choose an appropriate leading electrolyte. From the experiments performed in the previous chapter, the p H junction within a multi-electrolyte system is mobile under an electric field. Knowing the rates of migration of proton and hydroxide ions may allow for the calculation of the mobility of the p H junction. This project demonstrates that the mobility of hydroxide ions and protons can be measured using a commercial C E instrument equipped with UV-visible absorbance detection. As proton and hydroxide ions have no inherent UV-vis absorbance, a secondary detection system was required. The ions were detected with the aid of a p H -indicating dye. Ion mobility was calculated from the migration time of the ion. 3.2 Experimental Section 3.2.1 Chemicals. The aqueous background electrolyte (BGE) consisted of 50 m M KC1 (Fisher Scientific, Nepean, Ont., Canada) containing either 0.5 m M methyl red ( B D H , Toronto, Ont., Canada) or 0.5 m M alizarin red sulphonate (BDH). The p H of the B G E and sample solutions was adjusted using 1.0 M HC1 (BDH) and 1.0 M N a O H (BDH). HPLC-grade methanol was purchased from Fisher Scientific. Concentrated H O A c (Fisher), and H2SO4 (BDH) were diluted to 1.0 M . Deionized, distilled water was used in all solutions. 3.2.2 Apparatus and Procedure Separations were performed on a P/ACE™ M D Q automated capillary electrophoresis system (Beckman-Coulter Inc., Mississauga, Ont., Canada). Neutral capillaries (Beckman-Coulter Inc., Mississauga, Ont., Canada) coated with polyacrylamide were used with inner diameters of 50 pm, and outer diameters of 375 pm. Capillary lengths 57 were measured with a ruler and found to be 29.30 cm to detector, and 39.70 cm overall. New capillaries were first rinsed with deionized water (5 minutes, 20 psi) and then with 50 m M KC1 B G E (2 minutes, 20 psi) and used immediately. Each separation was preceded with a rinse of deionized water (1 min, 0.5 psi), 50 m M KC1 (1 min, 0.5 psi), and followed by B G E (2 min, 0.5 psi). The sample was introduced using a 0.5 psi pressure injection for 5s, and separations were carried out at 10 k V under a controlled temperature of 2 5 ° C . The ramp time for the voltage to reach 10 k V was 0.17 minutes. The migration time for H + ions was measured by rinsing the basic form of the dye through the capillary, injecting acid and applying a voltage under normal polarity. The migration time for OH" ions was measured by rinsing the acid form of the dye through the capillary, reversing the polarity and injecting base. Hydroxide mobility measurements were made with alizarin red dye only, as the acidic form of methyl red is only sparingly soluble in aqueous solution. The absence of an E O F (due to the neutral surface on the capillary interior) was established by performing a 90-minute separation at 10 k V using 50 m M KC1 as the B G E . A 2 s, 0.3 psi injection of 50% v/v M e O H / H 2 0 was performed and no E O F marker peak was observed during 90 minutes of separation. Absorbance detection was performed with a Beckman M D Q P D A detector over a wavelength range of 300-600 nm (extracted electropherograms at 410 and 520 nm). UV-visible spectra from 300-600 nm for methyl red and alizarin red dyes were extracted from P D A measurements of dye in the capillary. 3.3 Results and Discussion Apparent ionic mobility (u a p p) can be calculated from the migration time of the ion: Where L D is the length to detector, L T is the total length of the capillary, V p r o g is the programmed separation voltage, t m i g r is the migration time of the ion, and t r a m p is the time required for the separation voltage to reach its programmed value 4 9 . Usually, t r a m p is not taken into account when determining u a p p , but because the migration times for these ions are so short (36-50s), the 10.2 s required for the separation voltage to reach its programmed value will have an effect on the migration time of the analyte. In addition, the actual length that the sample must travel to reach the detector becomes important when the sample is travelling so quickly. The length of the sample zone for a 5 s, 0.5 psi injection is 0.35 cm. The length to the detector window was measured to be 29.30 cm. Therefore, the front of the sample plug must travel 28.95 cm to reach the detector. The migration time of the analyte was the time at which the absorbance change in the dye first started, corresponding to the first ions reaching the start of the detector window. The change in absorbance of the dye is almost instantaneous. The effective mobility of the ion, peff, is the mobility of the ion in the absence of electroosmotic mobility, u e o f . Using a neutral capillary, the E O F can be eliminated, and the effective mobility can be measured directly from the apparent mobility. Various pH-indicating dyes have been used to measure the p H inside a capillary, mainly to determine the effects of electrolysis of water on the p H of the background electrolyte 3 3 " 3 5 . By measuring dye absorbance at two wavelengths, it is possible to determine p H values, regardless of dye concentration. Huang et al. used methyl red prepared in water to show that a pH-gradient could be created within the capillary from 59 the electrolysis of water 3 6 . The pH gradient was then used to separate proteins using capillary iso-electric focussing (CIEF). In the present study, both methyl red and alizarin red sulphonate were used to detect the presence of proton and hydroxide ions at the detector window. The spectral properties of the dyes cause them to shift their absorbance maxima when protonated or deprotonated. Figure 3-2 depicts the chemical structure of methyl red and its UV-spectra under acidic and basic conditions. 1 I I 1 1 1 1 300 350 400 450 500 550 600 Wavelength (nm) Figure 3-2. Chemical structure and UV-spectra of methyl red prepared in 50 m M KC1 solution, [a] p H > 6 [b] p H < 4 The acidic form of methyl red has a maxima at 525 nm, the basic form has a maxima at 440 nm. The other dye used in the proton and hydroxide measurements was alizarin red sulphonate, which is depicted in Figure 3-3. 60 "i 1 r 400 450 500 Wavelength (nm) Figure 3-3. Chemical structure and UV-spectra of alizarin red prepared in 50 m M KC1 solution, a) p H > 6 b) p H < 4 3.3.1 Proton mobility using methyl red indicator. Figure 3-4 depicts an electropherogram in which a 1.0 M HC1 injection has been made in B G E containing methyl red. At 0.723 minutes, there is a sharp increase in absorbance at 520 nm, and a decrease in absorbance at 410 nm. The absorbance changes correspond to the protonation of the dye. In control experiments (no injection), absorbance remained constant throughout the analysis. Proton mobility was found to be 31.6xl0" 4cm 2/Vs using methyl red indicator (Table 3-2). 61 Absorbance (arbitrary) 1 1 1 1 1 a) Absorbance (arbitrary) 1 1 1 1 1 r ' Absorbance (arbitrary) 1 1 1 1 1 i i i i i i D 1 2 3 4 5 Time (minutes) Figure 3-4. Electropherogram showing, the conversion of methyl red from basic to acidic form, indicating migration time of proton, a) absorbance at 520 nm b) absorbance at 410 nm. The data for all proton and hydroxide experiments are shown in Table 3-2. Table 3-2. Mobility data ( I O 4 cm 2 /Vs) for H + and OH" ions measured using methyl red (MR) and alizarin red (AR) H + i n H 2 S 0 4 H + i n H C l H + i n HO Ac OH" in NaOH b M R AR MR AR MR A R 10 kV 15 kV Mobility a 32.2 35.9 31.6 34.5 28.7 30.8 12.4 12.9 R S D % 2.5 0.7 2.5 0.8 1.1 0.8 3.1 2.0 a average of 6 replicates bmeasured in AR only The measured H + mobility is smaller than the theoretical value for u H + of 36.25xl0"4 cm 2 /Vs. This is expected, as the solution in which the protons were measured contained several different types of ions including: dye molecules, N a + and OH" ions from the base used to adjust the p H of the B G E , K + and CI" ions from the electrolyte, and H + and CI" 62 ions from the acid injected as the sample. The effects of relaxation and the electrophoretic effect are discussed in Section 1.2.2. In further experiments, the ionic strength of the B G E was decreased. In 0.5mM methyl red, 50 m M KC1, the average proton mobility was 31.6xlO" 4cm 2/Vs (RSD 2.5%) and in a more dilute B G E of 0.25mM methyl red, 25mM KC1, the average proton mobility increased to 32.8xl0" 4cm 2/Vs (RSD 1.8%). As expected, proton mobility is greater in the dilute solution. Three different acids were evaluated as proton sources in the experiments. The p K a values for H 2 S 0 4 , HC1, and H O A c are -5, -2.2, and 4.7 respectively 5 0 . The greatest average mobility in Table 3-2 was obtained when using H2SO4, the strongest of the three acids, and also the only diprotic acid evaluated. It is well known that the partial dissociation of an analyte can have a negative effect on its effective mobility. At the p H of the solutions in which the H O A c was measured, (> 5.4) the H O A c would have been more than 50% dissociated which may account for the differences in the proton mobility when using H O A c as a proton source. The acids were prepared in deionized water, so in the sample plug (6.9 nL) injected into the capillary, the H O A c would have been partly dissociated. As the sample mixed with the B G E , the p H would have increased and dissociation would have become complete. The time frame for the subsequent equilibria to occur may have accounted for the reduced migration of the protons in the H O A c sample. H 2 S 0 4 is a strong acid, and will completely dissociate in solution. Proton mobilities from H 2 S 0 4 would not be expected to have reduced mobilities due to partial dissociation. In addition, when the neutralization of protons at the front of the H 2 S 0 4 sample plug occurred (prior to the application of the electric field), these protons could be 63 replaced by the second dissociation step in from the diprdtic acid, thereby maintaining the protic strength of the sample plug. 3.3.2 Proton mobility using alizarin red indicator Over 50 different p H indicator dyes, spanning a p H range from 0.2-14.0 are commercially available 5 0 . To see if this measurement system is applicable using another pH-indicator, proton mobility was also measured using alizarin red sulphonate. Figure 3-5 depicts an electropherogram in which a 1.0 M HC1 injection has been made in B G E containing alizarin red. 410 nm 520 nm 1 T 2 3 Time (minutes) Figure 3-5. Electropherogram showing the conversion of alizarin red from basic to acidic form, indicating migration time of proton, a) absorbance at 410 nm b) absorbance at 520 nm. At 0.644 minutes, there is a sharp decrease in absorbance at 520 nm, and increase in absorbance at 410 nm. The absorbance changes correspond to protonation of the dye. In control experiments, the absorbance remained constant throughout the separation. Proton mobility using alizarin red indicator and 1.0 M HC1 as the proton source was found to be 34.5xl0" 4cm 2/Vs (Table 3-2). The proton mobility values are larger when using alizarin red than methyl red. If the alizarin red B G E had lower ionic strength than the methyl red B G E , then larger mobilities might be expected. However, both dyes were prepared at 0.5 m M in the same pH=10.5, 50 m M KC1 solution. When both dyes are prepared at the same concentration, alizarin red should have greater ionic strength, as the sodium salt form of the dye was used. A n examination of the current traces for the two different dye solutions confirms this point. The current reached a maximum value of 51.08 u A with alizarin red B G E , and 43.63uA with methyl red B G E , indicating that alizarin red had greater ionic strength. The respective ionic strengths of the two B G E ' s do not clarify why the effective proton mobility is greater when measured in alizarin red. However, the p H of the two dye solutions is different. When prepared in p H 10.5, 50 m M KC1, a 0.5 m M methyl red solution has p H of 7.26, while a 0.5 m M alizarin red solution has p H 5.41, reflecting the different p K a values of the two dyes. When a 1.0 M acid sample plug is injected into the capillary, the lower p H value of an alizarin red solution means that less H + ions will get neutralized in the sample zone. The front of the zone will remain intact and the sample plug, being longer, will have less distance to travel, resulting in a shorter migration time. The p H of the B G E has an effect on the measured ion mobility when using this measurement system. 65 3.3.3 Hydroxide ion mobility The mobility of hydroxide ions was measured using alizarin red. Figure 3-6 shows an electropherogram with a 1.0 M N a O H injection in B G E containing alizarin red. 0 1 2 3 4 5 Time (minutes) Figure 3-6. Electropherogram of conversion of alizarin red from acidic to basic form, indicating migration time of hydroxide ion. a) 520 nm b) 410 nm. B G E contains 50 m M KC1 At 1.63 minutes, there is a sharp increase in absorbance at 520 nm, and decrease in absorbance at 410 nm. The absorbance changes correspond to the dye being deprotonated. Hydroxide ion mobility using alizarin red indicator and 1.0 M N a O H as the hydroxide source was found to be 12.4xlO"4cm2/Vs (Table 3-2). The mobility values are again lower when compared to the theoretical value (u0) for hydroxide ions of 20.05xl0"4 cm 2 /Vs. However, as for protons, it can be expected that the ionic strength, and p H of the B G E will affect effective hydroxide ion mobility. The p H of the alizarin red solution was adjusted to 3.6 in order to obtain the acid form of the dye required as an indicator. Prior 66 to reaching the detector window, some of the OH" ions would be neutralized in the acidic environment of the capillary, resulting in longer migration times. The use of p H indicators such as Cresol-Red or Thymol-Blue, that shift from acid to base at higher p H values, should result in larger measured mobility values for OH" ions, as the electrolyte could be prepared at a higher p H value and more of the OH" sample zone would reach the detector window intact. The hydroxide mobility experiments were conducted at 10 and 15 kV, to show that this method for measuring mobility is voltage-independent (Table 3-2). Theoretically, from Eqn. (1.5), mobility is voltage independent. The data confirm that this measurement system is not affected by voltage changes. One area of concern was that the dye molecules possess charged groups and will move under an electric field. The electropherogram in Figure 3-6 shows the dye concentrating into peaks at approximately 4 minutes into the separation. However, the increase in absorbance is manifest at both wavelengths, indicating that this is not a p H change within the capillary; rather it is a dye concentration effect. In addition, this effect occurs well after any measurement of proton or hydroxide ion migration time. Despite the fact that the charged dye molecules will have greater mobilities at higher voltage values, the data obtained at 10 and 15 k V indicate that the hydroxide mobility measurements are voltage independent. 3.4 Conclusion The effective mobility of H + and OH" ions was measured indirectly using a P D A detector with pH-indicating dyes. This method of OH" and H + mobility determination is relatively simple and can be performed using commercially available instrumentation. Effective mobility measurements were dependent on the strength of the injected acid, and on the p H and ionic strength of the electrolyte in which the mobilities were measured. Values were within the range of the absolute mobility measurements from the literature. Measurements were independent of the applied separation voltage. 68 4 Method Development and Validation in the Analysis of Dendrimer Compounds 4.1 Introduction Dendrimers are a class of synthetic polymers, which are synthesized in a cascade of repetitive reactions, resulting in a branched molecule originating from a central core. They possess reactive end-groups, which allow for controlled synthesis extending radially as the molecule grows. The reactive end groups also allow for controlled branching topology and versatility in the design and modification of the terminal end groups. The modifications applied to terminal end groups will affect the functional properties of the dendrimer molecule. Many types of dendrimer have been synthesized to date. In 1985, Tomalia et a l . 5 1 reported the synthesis of an ammonia core dendrimer, with repeating methyl acrylate-ethylene diamine branches, which were termed PAMAJVI, or polyamidoamine dendrimers. They are also known commercially as Starburst® dendrimers, Figure 4-1. 69 o NH o Figure 4-1. 2 n a generation P A M A M dendrimer. Generated from 3-directional ammonia core, with amide connectivity. The term generation refers to the number of monomer units attached to the central core. In Figure 4-1, the P A M A M dendrimer is a described as 2 n d generation, as it branches twice from the central ammonia group. Two steps of monomer addition have occurred to create the molecule. In the late 70's and 80's, advances in the physical isolation and purification of macromolecules led to advances in dendrimer chemistry 5 2 . The dendrimers investigated here were created using a divergent synthesis pattern, termed the Denkewalter synthesis after R . G . Denkewalter, who patented this scheme in 1981 5 3 . Divergent synthesis involves successive rounds of reactions, in which repeating monomer sub-units are added, constructing the macromolecule from an initiator core. The core of a Denkewalter dendrimer is a benzhydrylamine (BHA) molecule. The repeating monomer unit is a tert-butoxycarbonyl (Boc) protected lysine; N,N'-bis(terf-butoxycarbonyl)-L-lysine. Repeating lysine units are attached to each successive "generation" of lysines, using peptide linkages. The nature of dendrimer synthesis leads to a number of potential by-products with differing degrees of substitution. The incidence of these by-products increases with increasing generation, and the three major causes of imperfections are; • defects in dendrimer synthesis leading to non-symmetry in the molecule, loss of the branching point, and non-uniform branching points; • incompletely reacted previous generation present in final product; • incompletely reacted terminal groups during terminal group modification. Dicyclohexylcarbodiimide (DCC) is used as a carboxy-activating agent to catalyze peptide bond formation in the Denkewalter dendrimer. For every lysine molecule added, two new connection points (amines) are created for fresh lysine addition. Boc is used to protect the amines on the excess monomer added to the growing macromolecule, ensuring that only one type of connection will result. Deprotection occurs under an aqueous acid treatment, which is mild enough to leave other peptide bonds untouched 5 4 . A schematic of the Denkewalter synthesis can be found in Figure 4-2. 71 NH-Boc O NH-Boc DCC NH-Boc benzytTydrylarrine MH NH-Boc 1.1-r i-ysj V * 1 ^ r f 2. excess protected lysine 2cnd Generation BHAlys dendrimer Z excess protected lysine NH-Boc Boc-HN Figure 4-2. Schematic of Denkewalter dendrimer synthesis. Protected amino acid, N , N ' -bis(ferr-butoxycarbonyl)-L-lysine is the monomeric building block. The result of the synthesis is a hydrophilic, monodisperse, asymmetrical, highly branched compound that has several amide bonds within its interior and primary amines at its terminus. Generation 0 refers to a B H A core with one lysine unit attached and is termed BHAlys . Generation 1 refers to a Generation 0 molecule with two additional lysines attached to the original lysine. The terminology for a Generation 1 Denkewalter dendrimer is BHAlysilys2, the core lysines named first, followed by the outermost lysines. Table 4-1 outlines the number of lysine molecules and primary amines involved in each successive generation of Denkewalter dendrimer. 72 Table 4-1. Molecular weight, number of monomeric units (lysine), number of primary amino terminal groups, and hydrodynamic diameter for Denkewalter dendrimers 5 5 Generation Terminology Monomer units Terminal amines Molecular weight Diameter (A)a 0 BHAlys 1 2 311 11.8 1 BHAlyslys 2 3 4 568 14.6 2 BHAlys 3lys 4 7 8 1029 18.2 3 BHAlys 7lys 8 15 16 2106 22.8 4 BHAlys 1 5lys, 6 31 32 3949 28.6 5 BHAlys 3ilys 3 2 63 64 7635 36.0 aby measurement of intrinsic viscosity The range of applications for these relatively new polymers is broad, they have been used as molecular size standards 5 6 , light-gathering antennae 5 7 , pseudo-stationary phases in micellar electrokinetic capillary chromatography ( M E K C ) 3 7> 5 8 , ion pair agents in ion-exchange chromatography 5 9 , drug delivery/solubilization systems 60-62 ; a n c j toners 6 3 , synthetic peptides 6 4 , and as radio-label tags for antibodies 6 5 . It can be seen from these applications that there is a need to characterize the dendrimers, especially for their use in biomedical applications. Knowledge of the purity of process intermediates is also necessary for the optimization and scale-up of dendrimer synthesis. The dendrimers investigated here were developed for biomedical applications. The terminal 1° amines of a 4 t h generation Denkewalter dendrimer (BHAlysislysi6) were capped with 32 negatively charged R groups. The R group is a 3,6-disulphonate-l-isothiocyanate naphthalene, and its thiourea 5 4 linkage to the 1° amine on the dendrimer is shown in Figure 4-3. 73 NCS Dendrimer t S (-)o3s S03(-) Figure 4-3. R group attachment to 1 ° amine termini on dendrimer The p K a of the sulphonate molecules attached to the R groups is approximately 1.85 5 0 , therefore each R group will have a -2 charge and the total charge on the outer shell of a capped dendrimer will be -64 (32 x -2). Within the p H ranges investigated, from p H 2-10, the interior of the dendrimer is most likely neutral, due to the relative neutrality of the amide bond 5 4 . Several analytical methodologies have been employed in the characterization of dendrimers, including size-exclusion chromatography (SEC) 6 6 " 6 8 , high-performance liquid chromatography (HPLC), nuclear magnetic resonance (NMR), and slab-gel electrophoresis 6 9 . The resolution using H P L C and S E C techniques is not adequate for the separation of closely related compounds 5 5> 7 ( ). Mass spectrometry (MS) has proven to be a useful tool for the analysis of dendrimers 7 0 , however, the characterization of a 74 complex polydisperse sample using M S alone is difficult, due to the presence of isomeric by-products 7 0 . Capillary electrophoresis has also been used to analyze the dendrimers, as a stand-alone technique, and when coupled to mass spectrometry systems. Brothers et al. used capillary zone electrophoresis to separate five generations of ammonia-core P A M A M dendrimers. They also successfully used C E to separate low-generation P A M A M dendrimers in which terminal amino groups had been modified to varying degrees with salicyaldehyde groups 6 9 . The goal of this work is the analysis of terminally functionalized, higher generation Denkewalter dendrimers, and the development of an assay that will be used to characterize dendrimers prepared for biomedical applications. The objective is to develop a simple, reproducible assay that is non-destructive of the sample, highlights the presence of impurities, is linear over a wide range of concentrations, and has a low limit of quantitation. Ideally, the method will be applicable to biological samples containing the dendrimer. 4.2 Experimental 4.2.1 Chemicals Dendrimers were obtained from BRI Inc., an Australian research organisation. Unsubstituted 4 t h and 5 t h generation Denkewalter dendrimers, BHAlysi5lysi6 and BHAlys3ilys32 were used to determine the migration behaviour of the unsubstituted molecules. 4 t h generation BHAlysi 5 lysi6 substituted with sulphated napthyl-isothiocyanate was investigated as the primary compound. Two batches of the substituted compound were analyzed, one of a "pure" compound termed P - l , and one in which the substitution reaction was "incomplete" leading to variations in the number of terminating groups, termed P-2. Various buffer solutions were prepared during the analytical method development process. Borate buffer, p H 9.2, was prepared at concentrations of 20, 40, 60, 80, and 100 m M using Borax (Sigma Chemical Co., St. Louis, M O ) . Where required, the p H of the borate separation buffer was adjusted using 1.0 M HC1 ( B D H Chemicals, Toronto, Ont., Canada) or 1.0 M N a O H (BDH). Phosphate buffer, 0.1 M , p H 2.7 was prepared from monobasic sodium phosphate and phosphoric acid (Fisher Scientific, Nepean, Ont., Canada). Different buffer additives were evaluated, including sodium dodecyl sulphate (SDS) and dodecyltrimethylammonium bromide (DTAB) , both from Sigma. 4.2.2 Instrumentation Separations were performed on a P/ACE™ M D Q automated capillary electrophoresis system (Beckman-Coulter Inc., Mississauga, Ont., Canada). Unless otherwise specified, fused silica capillaries with inner diameters of either 50 or 75 pm were used (Polymicro Technologies, Inc., Phoenix, Arizona). New capillaries were first rinsed with 1.0 M N a O H (5 minutes, 20 psi), deionized water (2 minutes, 20 psi) and then with separation buffer (5 minutes, 20 psi) and used immediately. Absorbance detection was performed with a Beckman M D Q U V detector at 214 nm. UV-spectra of compounds were obtained using a Beckman M D Q photo-diode array (PDA) detector. Separation details such as capillary length, separation voltage, capillary temperature and buffer systems are specified in the individual section dealing with each type of buffer system utilized. A l l data was evaluated using P / A C E M D Q software (Beckman, Coulter). 76 4.3 Results and Discussion The dendrimers used in this study have many properties that are suited to C E analysis, as they are water-soluble, stable in aqueous solution, absorb U V light, and are charged in solution. However, they also have some unfavourable properties. Due to their large size, asymmetry, and multiple charges, Denkewalter dendrimers deviate from the Debye-Hiickel-Henry theory of electrophoretic mobility used to calculate the absolute mobility of charged molecular species, Eqn. (1.11). The complicated branching patterns contribute to uncertainties as to the effective charge on these molecules In addition, the positively charged free amine moieties found on the uncapped dendrimers may cause the dendrimers to interact with the capillary wall. The use of low p H buffers is a solution to this problem; at low pH, the silanol moieties on the capillary wall will be protonated, and the attractive forces for the cationic dendrimer will be reduced. [SiOHl In a bare fused-silica capillary, the equilibrium at the surface is dictated by (4.1) and the p K a is around 2.5 2 6 , above p H 2.5 the walls will be negatively charged due to partial deprotonation. Accordingly, the BHAlys unsubstituted dendrimers (no capping groups) were first analyzed in low p H phosphate buffer in which the E O F is minimal and the silanol groups at the wall will be protonated. Brothers et al. when analyzing P A M A M dendrimers also used low p H phosphate buffer 6 9. 4.3.1 Low p H phosphate buffer The 75 um internal diameter (i.d.) capillary had a total length (L t) of 60.2 cm, with 50.2 cm to the detector (L d). The separation voltage was 15 k V over 15 minutes, and the 77 resultant current was measured to be 85 u A , with the capillary temperature set at 2 5 ° C . The buffer was 100 m M phosphate adjusted to p H 2.8 using phosphoric acid. Samples were prepared in deionized water at a concentration of 0.1 mg/mL and introduced using a 0.5 psi injection for 3 seconds. U V absorbance was monitored at 214 nm. The optimum detection wavelength was determined using a photodiode array (PDA) detector to obtain a spectrum of the analytes in their respective buffers. The BHAlys "unsubstituted" molecules were found to have maximum absorbance at 200 nm when measured in 100 m M phosphate buffer. The capped molecules, P - l and P-2, were found to have maximum absorbance at 232 nm, most likely due to the presence of the naphthalene capping groups. A wavelength of 214 nm was chosen as it is close to the maximum for both types of compound, and a 214 nm filter was available for the U V detector. Samples dissolved immediately on contact with deionized water, indicating that they are hydrophilic. The E O F is negligible due to the protonation of the silanol groups at the capillary wall in the p H 2.8 buffer, additionally, any free primary amines on the dendrimers are protonated and carry a +1 charge. From the data in Table 4-1, the charge-to-size ratios for the 4 t h and 5 t h BHAlys generations are 2.2 and 3.6 respectively, which indicates that higher generations should have a larger electrophoretic mobility and migrate to the detector first. The experimental data show a reversal of the expected trend, with the 4 t h generation having an electrophoretic mobility (u e p) of 2.86 x IO"4 cm 2 /Vs , and the 5 t h generation 2.81 x IO"4 cm 2 /Vs. Brothers et al. when analyzing P A M A M dendrimers with primary amines at the terminus also noted this trend 6 9 . As noted previously, the dendrimers are non-ideal analytes, being asymmetric and multiply charged. The effective charge may be less than the formal charge, especially with 78 increasing generation number and complexity of dendrimer architecture. Furthermore, ion-pairing of the dendrimers with phosphate ions in the buffer can also affect the electrophoretic mobility. The observed electrophoretic mobility may be a result of these factors. The electropherogram of the 5 t h generation dendrimer did not show unidentified signals. However, the 4 t h generation electropherogram showed a slight impurity, which is probably a by-product. This impurity had a greater mobility than the main peak, indicating that it is most likely a younger generation of the same dendrimer (Figure 4-4). 79 J 1 j J ^ 5 t h Generation (12.03 min) Time (minutes) Figure 4-4.4 t n and 5 m generation unsubstituted BHAlys dendrimers analyzed in low p H phosphate buffer. The phosphate buffer system was also used to analyze the capped dendrimers, P - l and P-2. The polarity was reversed, as the net charge on the dendrimers is negative, due to the sulphonate residues present on the capping groups. The mobility values for the substituted dendrimers were greater than the mobility values for the BHAlys unsubstituted molecules, due to the double negative charge on each capping group of the substituted dendrimers. The mobility of P - l was 3.41 xlO" 4 cm 2 /Vs, and there was a small 80 impurity peak observed migrating unresolved just after the main peak, Figure 4-5. Batch P-2 was also analyzed and a large peak appeared that was poorly resolved from other impurity peaks. The mobility of P-2 was 3.02 xlO" 4 cm 2 /Vs, indicating that there were less negative charges on this molecule, which was expected, as this is the "less" substituted batch. P - l peak at Abs 15.45 mm (Au) V Time (minutes) Figure 4-5. Electropherogram of P - l analyzed in low p H phosphate buffer. Surfactant molecules such as sodium dodecyl sulphate (SDS) can enhance the selectivity of a method by providing a "pseudo-stationary phase" when added to a C E 81 buffer. Once above their critical micelle concentration, or C M C , the surfactant molecules will aggregate into micelles, which comprise a hydrophobic hydrocarbon interior, surrounded by a charged hydrophilic perimeter. Selectivity enhancements are achieved when the analyte molecules partition into the micelles. Consequently the mobility of the analytes is impeded or increased, depending on the mobility of the micelle. Partitioning of analyte molecules into micelles will be affected by both the hydrophobicity of the analyte, and the charge repulsion or attraction experienced between the analyte and additive. Efforts were now focussed on improving the resolution of P - l from its impurity peak, and seeing if the resolution of the many compounds present in P-2, the incompletely substituted batch, could be improved. It was proposed that the addition of SDS to the separation buffer might improve the resolution through ion pairing of the SDS molecules with the free-amines present in the P-2 batch or by the interaction of the dendrimers with SDS micelles. No improvement in resolution was seen, and the failure of SDS to improve resolution in phosphate buffer systems was potentially attributed to its negative charge, and its similarity to the capping groups on the substituted dendrimer molecules. The positively charged surfactant dodecyltrimethylammonium bromide (DTAB) was also evaluated as a buffer additive, but once again, no improvements in resolution were observed. In conclusion, D T A B and SDS buffer additives used in low p H phosphate buffer did not enhance the resolution of the P - l and P-2 compounds from their impurities. 4.3.2 SDS Gel Electrophoresis Another separation system for the analysis of high molecular weight compounds is SDS-capillary gel electrophoresis (SDS-CGE). Most often used with proteins, this 82 technique was adapted from S D S - P A G E (polyacrylamide slab gel electrophoresis), commonly used in biochemistry to separate proteins based on molecular weight. Proteins are denatured with 5-mercaptoethanol, resulting in linear polypeptide chains. The anionic surfactant, SDS, is added, which combines with the proteins in a ratio of 1.4:1 (w/w). The separation is then performed in a sieving matrix and is based largely on molecular weight differences between the proteins, as they will all have the same charge-to-mass ratio when coated with SDS. For the dendrimers P - l and P-2, P - l is completely substituted with -2 charged sulphates, and is not expected to combine with SDS molecules, yet will travel through the gel based on its inherent negative charge. P-2, being incompletely substituted, carries some positive charge sites, and it is expected that these regions might combine with the SDS, and that the various compounds will separate based on size differences. The capillary used for S D S - C G E is coated to minimize the E O F and filled with a linear polyacrylamide gel matrix. A protein "test kit" was obtained from Beckman-Coulter, Inc., which contained a neutral, coated capillary and a polyacrylamide gel buffer. Unlike denatured proteins, the dendrimers will not be linear and their asymmetrical "fan" shape leads to excessive band broadening, with no concomitant increase in resolution. The BHAlys unsubstituted compounds were also analyzed with the S D S - C G E system. These compounds were observed at the detector window, indicating that there was some association with the negative SDS molecules. However, there was no improvement over regular buffer systems in terms of the resolution of P - l and P-2 and their impurities. 83 4.3.3 Borate Buffer Systems and Method Validation O f three common buffers evaluated; Tris, acetate, and borate, the borate buffer provided the best resolution of the main peak of P - l from its impurity, as well as better reproducibility of peak area and migration time. In C E , the mobility of the analyte will affect peak area, as low mobility analytes will have a greater transit time in front of the detector window. This effect can be offset through the use of corrected peak areas. A - j , peak area Corrected peak area = — (4.2) migration time Even with this correction, it is still important to develop a method that keeps migration times as reproducible as possible, in order to avoid discrepancies in peak area determination 7 1 . In order to determine the optimum borate buffer strength, the buffer concentration was varied from 20-120 m M while holding voltage and temperature (30 k V , 25 °C) constant. The capillary was 75pm i.d., 363 um o.d., 40.2 cm L D , and 50.2 cm L D . Table 4-2 contains data for effective electrophoretic mobility (corrected for E O F ) of P - l at varied borate buffer concentrations. 84 Table 4-2. Mobility data for dendrimer P - l in borate buffer Borate Current E O F Peak Meof M-app P-eff m M u A min min xlO" 4cm 2 /Vs 20 25 1.513 3.433 7.38 3.25 -4.13 40 47 1.746 4.500 6.40 2.48 -3.91 60 71 1.854 5.262 6.02 2.12 -3.90 80 92 1.970 6.117 5.67 1.83 -3.84 100 117 2.020 6.97 5.53 1.60 -3.93 120 134 2.040 7.606 5.47 1.47 -4.01 As expected, dendrimer mobility is greatest at the lowest ionic strength but with increasing ionic strength, where one would expect mobility to decrease the opposite effect occurs (Figure 4-6). 4.15 co 4.10 *j 4.05 t 4.00 o 5 - 3.95 1 3.90 o 5 3.85 3.80 0 20 40 60 80 100 120 140 Borate (mM) Figure 4-6. Effective mobility vs. borate concentration 85 From Figure 4-6, the mobility decreases, then increases again as the ionic strength increases. This effect may be due to differences in the radius of the dendrimer as ionic strength changes. The increased presence of counter-ions in higher strength buffer may effectively shield the negative charges at the ends of the dendrimer branches from one another, resulting in a more compact globular structure. Further method development for P - l was conducted in 40 m M Borate buffer as it provides good resolution (1.96) from the impurity peak, has a relatively short run time at 30 k V (< 5 minutes), and provides reproducible results with respect to peak area ratio against the internal standard p-nitrophenol (RSD 1.8%, n=6). For P-2, borate buffer alone does not provide any better resolution of the main peak from the many impurities surrounding it when compared to phosphate buffer systems, but the borate buffer does highlight the differences between the pure (P-l) and impure (P-2) batches, therefore will be a useful qualitative tool, Figure 4-7 and 4-8. A b s r I (Au) I P - l peak at 4.55 min \i — „ — ^ . . . t y -\ V.. Time (minutes) Figure 4-7. Electropherogram of P - l analyzed in 40 m M borate buffer 86 P-2 peak 3.45 -4.60 min Abs (Au) Time (minutes) Figure 4-8. P-2 analyzed in 40 m M borate buffer In order to evaluate the applicability of the method using 40 m M borate buffer to the analysis of P - l , some method validation parameters were evaluated. A full validation for the analysis of P - l was not possible, due to the lack of a qualified reference standard, but the validation parameters of linearity, precision and specificity were assessed. A preliminary assessment of sample stability (in solution) was also conducted. In addition, an attempt was made to identify the impurity peak in the P - l batch. Method validation is performed to provide proof that the analytical method meets its intended purpose. Guidelines for the parameters that should be validated can be found in regulations governing pharmaceutical, environmental and manufacturing industries. The United States Pharmacopoeia (USP) contains information on the Validation of Chromatographic Methods that can be applied to C E . However, as many industries have been slow to adopt C E technology 7 2- 7 3, there has not been a significant amount of 87 discussion with respect to the validation of C E methods among the various regulatory bodies. Several C E methods have been validated and reported in the literature 7 4> 7 5 . As no pure standard batch of P - l was available, p-nitrophenol (p-NP) was chosen as an internal standard. A n internal standard improves the precision of peak area determination and reduces injection error, which can occur in C E , especially when using small sample volumes over long periods. Internal standards also compensate for dilution errors during sample preparation and errors during pre-treatment of biological samples. A n ideal internal standard will have a migration time that is similar to that of the sample, but migrates ahead of the peak of interest, in order to reduce analysis times. For detection purposes the internal standard should have a similar chemical structure to the analyte. p-N P has a greater apparent mobility than P - l and migrates slightly ahead of it in 40 m M borate buffer. It is negatively charged, UV-active, and stable in aqueous solution 5 4 Peak area ratios of P - l to p-NP were used for calculation of injection repeatability, one measure of precision. Precision, in the form of injection repeatability was evaluated using a 0.12 mg/mL solution of P - l and 6 injections. Results are shown in Table 4-3. 88 Table 4-3. Injection repeatability for P - l ection Peak area (mAu) . Area % P - l ratio P - l p-nitrophenol 1 152080 106789 1.42 55.2 2 158098 110273 1.43 55.0 3 131695 95512 1.38 54.5 4 159525 111938 1.43 54.9 5 184486 129470 1.43 54.8 6 185736 129321 1.44 54.8 avg. 161937 113884 1.42 54.9 R S D % 12.7 11.7 1.5 0.4 Peak area, with an R S D > 12%, is not a precise measurement 2 6. Variations in the E O F from run to run can affect the peak mobility, hence peak area. When the peak area ratio against an internal standard is used, the R S D decreases from 12.7% to 1.5%, a large improvement in the precision of the measurement. Linearity is the ability, within a given range of concentrations, to obtain test results that are directly proportional to the concentration of analyte in the sample. Linearity was initially evaluated from 10-2000 pg/mL. Each concentration was analyzed in duplicate and peak area was plotted against concentration. Data are shown in Table 4-4. The linearity plot is shown in Figure 4-9. 89 Table 4-4. Linearity Upper Range Concentration Peak area ug/mL Au*sec 10 9549 50 52219 100 136804 500 642904 1000 1224706 2000 2432250 3.00E+06 2.50E+06 -I « 2.00E+06 -1 < (0 1 .50E+06 -I 0) i§ 1.00E+06 Q. 5.00E+05 -! 0.00E+00 500 1000 1500 Concentration (ug/mL) 2000 Figure 4-9. Peak Area vs. Concentration over a range of 10-2000 ug/mL for compound P - l The plot is linear with an R 2 value of 0.9997 (Table 4-6). 9 0 The Quantitation Limit (QL) of a method is the lowest concentration that can be detected accurately and reproducibly. The detection limit (DL) is the lowest concentration that can be detected above the background noise of the detection system. The Q L is affected by any noise associated with the measurement, the physical and chemical properties of the analyte and the composition of the sample matrix. Normally, the Q L is estimated from the linearity data and then evaluated by preparing a number of samples at the Q L concentration and analyzing them for Accuracy and Precision. A n analyte concentration that produces a signal ten times the standard deviation of the blank (a) signal is commonly defined as the Q L concentration 7 6 . From the calibration curve of an analyte, Q L concentration will be the concentration at which a signal of 10a + y-intercept (b) is produced. From the slope equation of the regression line of the calibration plot, the concentration can be calculated as (4.3) Where S is the slope of the regression line. Sigma (a) can be determined from 7 7 from the standard deviation of the blank. Once estimated, the Q L must be verified by the analysis of several samples prepared at the Q L concentration. Using the data from the linearity plot, the Q L was calculated to be 77 ug/mL (Table 4-6), this value is greater than the lowest value detected and integrated in the linearity experiment. It is possible that the expanded range of concentrations over which the linearity was evaluated led to greater deviation in the y-intercept, the value used for sigma (a). For this reason, and because triplicate measurements of the linearity experiment were required to evaluate injection reproducibility over the range of concentrations tested, the experiment was repeated over a lower concentration range. A range from 0.05 to 50 ug/mL was evaluated. The peaks present in the electropherogram for the 0.05 and 0.1 ug/mL concentrations were too small to correctly integrate, therefore data are shown for the 1-50 ug/mL range (Table 4-5). 92 Table 4-5. Linearity over a Lower Concentration Range Concentration Injection # Peak area Peak Height ug/mL Au*sec uAu 1 1 2 3 R S D % 1233 1046 1026 10.4 260 261 156 26.7 5 1 2 3 R S D % 6405 5743 5406 8.7 1184 1003 773 20.9 10 1 2 3 R S D % 11592 11832 12476 3.8 1915 1978 2078 4.1 50 1 2 3 R S D % 43633 . 44374 45662 2.3 7314 7699 7968 : 4.3 The Q L was determined from the lower concentration range and found to 12 ug/ml, which is in keeping with the visual examination of the data. The slope, y-intercept, standard deviation of the y-intercept, correlation coefficient R 2 and Q L can be found in Table 4-6. 93 Table 4-6. Linearity Values Parameter Concentration Range Upper Lower Slope (Au*sec*mL*mg1) 1215.5 865.7 Y-intercept (Au*sec) 8276 1585 Standard Error Y-intercept 9408 1022 R 2 0.9997 0.9958 QL (Calculated) ug/mL 77 12 From Table 4-5, the 10 and 50 pg/mL concentrations provide acceptable precision 7 7. The Q L must be evaluated further once a reference standard is available. Without a reference standard, accuracy cannot be determined at the Q L , but 12 pg/mL can be tentatively assigned. Specificity is intended to show that the method is specific to the analyte being tested, and that the method will highlight differences between the analyte and any other compounds in a test sample. Specificity can also be used to demonstrate that the method is stability indicating, by showing that degradation products don't interfere with the analysis of the main compound. In order to evaluate specificity, a solution of P - l was prepared in deionized water at 60 ug/mL. The solution was analyzed and then stored for 12 days at under the following conditions; - 2 0 ° C under darkness, 4 ° C under darkness, room temperature under light (12 hour daylight), and room temperature under darkness. The greatest amount of degradation occurred in the sample stored at room temperature and exposed to light. The corrected peak areas of P - l after various storage conditions are shown in Table 4-7. 94 Table 4-7. Stability Samples Storage Condition Corrected Peak Area (Au*sec) Control 9740 - 2 0 ° C , 12 days 8872 4 °C , 12 days 9303 Room temperature dark, 12 days 7337 Room temperature light, 12 days 4951 The room temperature light storage also had the greatest incidence of new impurities. Relative migration time (RMT) is used to identify the impurities relative to the main peak (P-l), which is defined as the reference in this case. The main impurity ( R M T 0.96) that was always present in the P - l batch did not increase in area, but a new impurity formed ( R M T of 1.05). The resolution of P - l from its impurities was compromised under room temperature light storage, but the resolution was greater than 1 for all other storage conditions. The manufacturer of P - l suspected that the impurity with an R M T of 0.96 is the compound shown in Figure 4-10. A sample of the compound in Figure 4-10 was analyzed alone and when spiked into a solution of P-1. It was found to have a R M T of 1.05, and its U V spectra is very similar to that of P - l , the only difference being a slight increase in absorbance at 210 nm in the P - l spectra, over the impurity. This impurity is present in all of the stability samples, indicating that lysine groups are hydrolysed from the main compound on storage. R M T = migration time peak (4.4) mibration time reference 95 ( - ) O 3 S ^ N ^ ^ NH NH • ( - ) 0 3 S ^ s riS II II x ^ S 0 3 ( - ) J (-)o3s 0 Figure 4-10. Structure of impurity with R M T 1.05 The 40 m M borate buffer assay of P - l was found to be linear over a range of 10-2000 ug/mL with a correlation coefficient of 0.9997. Precision, in the form of injection repeatability, was evaluated and found to have an R S D < 2% when peak area ratios were used. The Q L was calculated to be 12 ug/mL, and this value is within the linear range of the assay. p-NP was chosen as an internal standard, although the assay was not evaluated for accuracy (quantitation) as no reference standard of P - l was available to analyze the sample against. The method was only evaluated on the two available batches of material, but it did show specificity between the completely and incompletely substituted forms of the compound. 4.3.4 Biological Assay As a compound with potential pharmaceutical applications, there is a need to evaluate the concentration of P - l in biological fluids, mainly plasma. Plasma is the supernatant remaining after whole blood is treated with an anticoagulant (EDTA) , and centrifuged to remove the red and white blood cells. Plasma contains proteins, ions, and any pharmaceutical agents or their metabolites that reside in the blood stream. Pharmacokinetics is the study of the routes and rates of elimination of pharmacological 96 compounds from living beings. In order to establish the time frame and route of removal, it is necessary to have a robust, sensitive method for the determination of drug concentration in biological fluids. Most biological assays require some sort of sample pre-treatment step, mainly for the removal of proteins. Plasma consists of 90 % water and 10 % solutes. O f the plasma solutes, 70% are comprised of plasma proteins such as serum albumin, lipoprotein, immunoglobulin, fibrinogen, prothrombin, and specialized transport proteins such as transferrin. Ten percent of plasma solutes are inorganic components including NaCl , KC1, bicarbonate, phosphate, CaCL; , MgCi2, and Na2S04. The remaining 20% of plasma solutes consist of organic metabolites and waste products such as glucose, amino acids, lactate, pyruvate, ketone bodies, citrate, urea, and uric ac id 1 . Further work was conducted to determine if the borate buffer assay for P - l was a good candidate for the detection and quantitation of the compound in plasma. As no biological samples were available, human plasma was obtained from Canadian Blood Services and spiked with 400 pg/mL of P - l in order to evaluate the drug in a plasma matrix. Two routes of development were available: 1. To assay the plasma directly using a method that will prevent the proteins from being absorbed onto the capillary wall (e.g. low p H phosphate buffer) 2. To remove the proteins from the plasma in order to assay the drugs without their interference 7 8 . Due to the large size of the dendrimers, there is a danger that they will also be removed with the proteins. For instance, one method of protein removal involves the use of low molecular weight cut-off ultrafiltration devices. These devices are small centrifuge 97 tubes that contain a mass-selective membrane. Under centrifugal force, the sample filters through the medium, leaving anything with a molecular weight greater than the cut-off on top of the membrane. The molecular weight of P - l is 14998 g/mol or 15 kiloDaltons (kD). Therefore a filter with a cut-off greater than 15 kD must be employed. When using higher p H methods, such as borate buffer, protein removal or treatment is a necessity, as the cationic proteins will otherwise associate with the capillary wall. Three methods of protein removal or treatment were assessed; 1. Protein coating with SDS. SDS to a concentration of 100 m M was added to the plasma samples, which were then vortexed and analyzed. Proteins will acquire a negative charge with the binding of SDS molecules. The negatively charged proteins will not be attracted to the capillary wall, and migrate to the detector later, due to their large negative charge. 2. Protein precipitation with acetonitrile (ACN). Acetonitrile (50% v/v) was added to a plasma sample spiked with P-1, followed by centrifugation and analysis of the supernatant. 3. Filtration of plasma with molecular weight cut off ultra centrifugation device. Plasma spiked with P-1 was applied to a M i c r o c o n ® 30kD ultracentrifuge filter device, and centrifuged for 8 minutes at 10 000 rpm. The filtrate was removed and analysed. Two buffer systems were used to analyze the resultant samples from the protein removal experiments; 40 m M borate buffer at p H 9.2; 40 m M borate buffer containing 98 100 m M SDS at p H 7.5. SDS was included as an additive in order to maintain a negative charge on any proteins that remain after sample pre-treatment. From the three sample treatment methods evaluated, protein coating with SDS, followed by the analysis of the sample in SDS buffer provided the best results. The compound P-1 was not recovered from the protein precipitation with A C N , although in a control sample, in which a standard solution of P-1 was mixed with A C N , centrifuged and analyzed, the P - l peak was visible. This indicates that A C N itself does not affect the P - l assay. It is possible that the large size of the dendrimer led to the co-precipitation of P - l with the protein mass. The 30 kD molecular weight cut-off filter provided acceptable results, when the filtrate was analyzed in 40 m M borate buffer containing 100 m M SDS. Some plasma solutes do pass through this type of filter, but their migration time is in the range of 2-7 minutes, which does not interfere with the migration time of P - l (13-14 minutes). When the filtrate was analyzed in regular 40 m M borate, the plasma solutes interfered with the detection of the P - l peak. The same results were obtained when SDS was added to the plasma sample. Table 4-8 shows a summary of the results. 9 9 Table 4-8. Optimization of plasma pre-treatment Pre-treatment type Sample type Buffer 1 Buffer 2 Acetonitrile (1) Controls precipitation (a) 50:50 v/v Peak at 4.20 Not evaluated ACN:P-1 (prepared minutes (P-l) in d H 2 0 ) (b) 50:50 v/v Several peaks at Not evaluated ACN:plasma 1.8-3.0 minutes (2) Spike 50:50 v/v Electropherogram Not evaluated ACN:plasma spiked looks same as (2). with 4 0 0 p g / m L P - l P - l peak not visible SDS addition (3) Controls (a) Blank + E O F marker at 1.83 E O F marker at 3.07 methanol for E O F minutes minutes measurement (b) P - l standard P- l peak at 3.85 P - l peak at 13.91 containing 100 m M minutes minutes SDS (c) plasma Large region of Large region of containing 100 m M peaks from 2-7 peaks from 8-12 SDS minutes minutes (4) spike Plasma spiked with Same as 3c P-1 peak visible at P - l containing 100 14.21 minutes, after m M SDS plasma solutes 30 k D cut-off filter (5) spike Plasma spiked with No P-1 peak visible, Plasma components P - l , centrifuged plasma solutes from from 8-12 minutes, 2-7 minutes P - l peak at 14.68 min. While the results were promising with the 30 kD filtration method, no control samples were analyzed due to a shortage of filtration devices. In addition, results for the filtration device were similar to those obtained using the addition of SDS, and the filtration devices incur a significant cost to the assay. Also, studies would be required to determine that P - l 100 does not bind to the filter membrane. Results were also promising for the assay in which 100 m M SDS was added to the buffer and plasma sample. In regular borate buffer, the P-1 peak was overwhelmed by the presence of plasma components that had not been eliminated with the various removal techniques. The presence of SDS in the borate buffer had the effect of reducing the mobility of P - l causing it to appear after all of the major plasma components had left the capillary. The p H of this buffer was lower (7.5) and the ionic strength greater (due to contribution of N a + ions and sulphate ions from SDS). This would result in a slower E O F , retarding the migration of the micelles toward the detector, which in turn would slow down any P - l molecules associated with the micelles. One negative aspect of the SDS buffer, is that it reduces the resolution of P - l from its major impurity, which will make it difficult to quantitate P - l in plasma. Further investigation of the biological assay is required in order to evaluate the detection limit, and to improve the resolution of the impurity from the main P - l peak, but the SDS addition technique is promising. 4.4 Conclusion A simple and robust assay for the analysis of pharmacologically active Denkewalter dendrimer compounds has been developed. This method will discern between incompletely and completely substituted compounds. For completely reacted compounds, the use of para nitrophenol as an internal standard improves the method precision. The method is linear over a concentration range of 10-2000 ug/mL and has not been evaluated for accuracy, as no standard reference compound was available. 101 5 References 1) Robinson, R.; Stokes, R. Electrolyte Solutions; 1st ed.; Butterworths Scientific Publishers: London, 1955. 2) Simpson, C ; Whittaker, M . Electrophoretic Techniques; Academic Press: London, 1983. 3) Bier, M . Electrophoresis: Theory, Methods, Applications; 1st ed.; Academic Press: New York, 1959. 4) Jorgenson, J. Analytical Chemistry 1986, 58, 743A-760A. 5) Mikkers, F.; Everaerts, F.; Verheggen, T. Journal of Chromatography 1979,169, 11-20. 6) Andrews, A . Electrophoresis: theory, techniques, and biochemical and clinical applications; Clarendon Press: Oxford, 1981. 7) Lehninger, A . ; Nelson, D.; Cox, M . Principles of Biochemistry; 2 ed.; Worth Publishers: New York, 1997. 8) Gould, D.; Synge, R. Analyst 1952, 77, 964. 9) Pretorius, V . ; Hopkins, B.; Schieke, J. Journal of Chomatography 1974, 99, 23-30. 10) Hjerten, S. Chromatographic Reviews 1967, 9, 122-219. 1 l)Everaerts, F.; Beckers, J.; Verheggen, T. Isotachophoresis; theory, instrumentation, applications; Elsevier Scientific Publishing Company: Amsterdam, 1976; Vo l . 6. 12) Jorgenson, J.; Lukacs, K. Analytical Chemistry 1981, 53, 1298-1302. 13) Pospichal, J.; Gebauer, P.; Bocek, P. Chemical Reviews 1989, 89, 419-430. 14) Terabe, S.; Otsuka, K. ; Ando, T. Analytical Chemistry 1985, 57, 834-841. 15) Guttman, A . ; Paulus, A . ; Cohen, A . ; Grinberg, N . ; Karger, B. Journal of Chromatography 1988, 488, 41. 16) Burgi, D.; Chien, R. Analytical Chemistry 1991, 63, 2042-2047. 17) Chien, R.; Helmer, J. Analytical Chemistry 1991, 63, 1354-1361. 18) Chien, R.; Burgi, D. Analytical Chemistry 1992, 64, 1046-1050. 19) Meldrum, D . Science 2001, 292, 515-517. 102 20) Zhang, J.; Voss, K. ; Shaw, D.; Roos, K. ; Lewis, D.; Yan, J.; Jiang, R.; Ren, FL; Hou, J.; Fang, Y . ; Puyang, X . ; Ahmadzadeh, FL; Dovichi, N . Journal of Nucleic Acids Research 1999, 27, 24. 21) Venter, J. Science 2001, 291, 1304. 22) Chen, D.; Swerdlow, FL; Harke, H . ; Zhang, J.; Dovichi, N . Journal of Chromatography 1991, 559, 237-246. 23) Manabe, T. Electrophoresis 1999, 20, 3116-3121. 24) Dolnik, V . ; Liu , S.; Jovanovich, S. Electrophoresis 2000, 21, 41-54. 25) Kuhn, R.; Hofstetter-Kuhn, S. Capillary Electrophoresis: Principles and Practice; Springer-Verlag: Berlin, 1993. 26) Landers, J. P. Handbook of Capillary Electrophoresis; 2 ed.; Landers, J. P., Ed. ; C R C Press: Boca Ratan, 1997. 27) Darken, L . S.; Meier, H . F. Journal of the American Chemical Society 1942, 64, 621-623. 28) Saxton, B.; Langer, T. L . Journal of the American Chemical Society 1933, 55, 3638-3645. 29) MacInnes, D.; Shedlovsky, T.; Longsworth, L . Chemical Reviews 1933, 33, 29-46. 30) MacInnes, D.; Shedlovsky, T.; Longsworth, L . Journal of the American Chemical Society 1932, 54, 2758-2762. 31) Reijenga, J.; Kenndler, E . Journal of Chromatography A 1994, 659, 403-426. 32) Kohen, E . ; Hirschberg, G . Cell Structure and Function by Microspectrofluorometry; Ploem, J . , Ed. ; Academic Press: San Diego, 1989. 33) Timperman, A . ; Tracht, S. E . ; Sweedier, J. V . Analytical Chemistry 1996, 68, 2693-2698. 34) Macka, M . ; Andersson, P.; Haddad, P. Analytical Chemistry 1998, 70, 743-749. 35) Corstjens, FL; Billiet, H . A . FL; Johannes, F.; Luyben, K. C . A . M . Electrophoresis 1996,17, 137-143. 36) Huang, T.; W u , X . Z . ; Pawliszyn, J. Analytical Chemistry 2000, 72, 4758-4761. 37) Palmer, C . Journal of Chromatography A 1997, 780, 75-92. 38) Quirino, J.; Terabe, S. Analytical Chemistry 1999, 71, 1638-1644. 103 39) Quirino, J.; Terabe, S. Analytical Chemistry 2000, 72, 1023-1030. 40) Aebersold, R.; Morrison, H . Journal of Chromatography 1990, 516, 79-88. 41) Britz-McKibbin, P.; Chen, D . D. Y . Analytical Chemistry 2000, 72, 1242-1252. 42) Zhujun, Z . ; Seitz, W . Analytical Chimie Acta 1984,160, 47. 43) Haughland, R. Handbook of Fluorescent Probes and Research Chemicals; 6 ed.; Molecular Probes Inc.: Oregon, 1996. 44) Owen, C . Analytical Biochemistry 1992, 204, 65-71. 45) Albert, A . ; Serjeant, E . Ionization Constants of Acids and Bases; Methuen and Company: London, 1962. 46) Britz-McKibbon, P.; Kranack, A . ; Paprica, A . ; Chen, D. Journal of Chromatography A 1999, 853, 535-540. 47) Eisenberg, D.; Crothers, D. Physical Chemistry with Applications to the Life Sciences; 1st ed.; Benjamin/Cummings Publishing Company: Menlo Park, 1979. 48) Glasstone, S.; Laidler, K. J.; Eyring, H . The Theory of Rate Processes; 1st ed.; McGraw-Hil l Book Company: New York, 1941. 49) Williams, B . A . ; Vigh, G . Analytical Chemistry 1996, 68, 1174-1180. 50) Weast, R. C . CRC Handbook of Chemistry and Physics; 68 ed.; C R C Press: Boca Raton, 1987. 51) Newkome, G . R.; Moorefield, C . N . ; Vogtle, F. Dendritic Molecules: Concepts, Synthesis, Perspectives; 1st ed.; V C H Publishers Inc.: New York, 1996. 52) Tomalia, D.; Baker, H . ; Devald, J.; Hall, M . ; Kallas, G . ; Martin, S.; Roech, J.; Ryder, J.; Smith, P. Polymer Journal 1985,17, 117-132. 53) Denkewalter, R. G . ; Kolc, J.; Lukasavage, W . J. Macromolecular highly branched homogenoeous compound based on lysine units; All ied Corporation (Morris Township, Morris County, NJ): US; 4289872; Sept 15 54) Vollhardt, C ; Peter, K. Organic Chemistry; W . H . Freeman and Company: New York, 1987. 55) Tomalia, D . A . ; Naylor, A . M . ; Goddard, W . A . Agnew. Chem. Int. Ed. Engl. 1990, 29, 138-175. 56) Dubin, P. L . ; Edwards, S.; Kaplan, J.; Mehata, M . ; Tomalia, D.; Xia , J. Analytical Chemistry 1992, 64, 2344-2347. 104 57) Fox, M . ; Jr., W . J.; Watkins, D. Chemical Engineering News 1993, 71, 34-48. 58) Tanaka, N . ; Tanigawa, T.; Hosoya, K.; Kimata, K. ; Araki, K . ; Terabe, S. Chemical Letters 1992, 959-962. 59) Cherestes, A . ; Engel, R. Polymer 1994, 35, 3343-3344. 60) Newkome, G . ; Moorefield, C . Unimollecular Micelles and Method of Making the Same; University of Florida: U S A ; 5154853; Oct 3, 1992 61) Peppas, N . ; Nagai, T.; Miyamiji, M . Pharmaceutical Tecnology Japan 1994,10, 611-617. 62) Sakthivel, T.; Toth, I.; Florence, A . Pharmaceutical Research 1998,15, 776-782. 63) Winnik, F.; Davidson, A . ; Breton, M . Ink with Dendrimer Colorants; Xerox Corporation: U S A ; 5098475; March 24, 1992 64) Tam, J. Proceedings of the National Academy of Science 1988, 85, 5409-5413. 65) Roberts, J.; Adams, Y . ; Tomalia, D.; Mercer-Smith, J.; Lavallee, D. Bioconjugate Chemistry 1990,1, 305-308. 66) Mengerink, Y . ; Mure, M . ; Bradander, E . M . M . d.; Wal, S. V . D. Journal of Chromatography A 1996, 730, 75-81. 67) Pesak, D.; Moore, J.; Wheat, T. Macromolecules 1997, 30, 6467-6482. 68) Young, J.; Baker, G . ; Newkome, G . ; Morris, K. ; Johnson, C . Macromolecules 1994, 27, 3464-3471. 69) Brothers, H . M . ; Piehler, L . T.; Tomalia, D. A . Journal of Chromatography A 1998, 814, 233-246. 70) Stockigt, D.; Lohmer, G . ; Belder, D. Rapid Communications in Mass Spectrometry 1996.10, 521-526. 71) Watzig, FL; Degehardt, M . ; Kunkel, A . Electrophoresis 1998,19, 2695-2752. 72) Hettiarachchi, K . ; Cheung, A . P. Journal of Pharmaceutical and Biomedical Analysis 1993.11, 1251-1259. 73) Hettiarachchi, K . ; Cheung, A . Journal of Chromatography A 1995, 717, 191-202. 74) Galli, V . ; Olmo, N . ; Barbas, C . Journal of Chromatography A 2000, 894, 135-144. 75) Sabbah, S.; Scriba, G . Journal of Chromatography A 2001, 907, 321-328. 105 76) Willard, H . ; Merritt, L . ; Dean, J.; Settle, F. Instumental Methods of Analysis; Wadsworth: Belmont, California, 1988. 77) F D A "Guidance for Industry: Analytical Procedures and Methods Validation: Draft document," Food and Drug Administration;Rockville, M D , 2000. 78) Taylor, R.; Toasaksiri, S.; Reid, R. Electrophoresis 1998,19, 2791-2797. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061433/manifest

Comment

Related Items