Hydrophilic SeparationMaterials for LiquidChromatography

Petrus Hemström

Umeå University, Department of ChemistryUmeå, 2007

– II –

© Copyright, Petrus Hemström, 2007. All rights reserved. ISBN 978-91-7264-406-9

Distribution: Department of Chemistry [Kemiska institutionen] Umeå University, 901 87 UMEÅ, Sweden. Tel: +46 (0)90-786 50 00 E-mail: petrus.hemstrom@chem.umu.se

Printed by: Print och Media : 2003591 Umeå University, UMEÅ.

– III –

Till Elin

– IV –

Abstract

The main focus of this thesis is on hydrophilic interaction chromatography (HILIC) and the preparation of stationary phases for HILIC. The mechanism of HILIC is also discussed; a large part of the discussion has been adapted from a review written by me and professor Irgum for the Journal of Separation Science (ref 34). By reevaluating the literature we have revealed that the notion of HILIC as simply partitioning chromatography needed modification. However, our interest in the HILIC mechanism was mainly inspired by the need to understand how to construct the optimal HILIC stationary phase. The ultimate stationary phase for HILIC is still not found. My theory is that a non-charged stationary phase capable of retaining a full hydration layer even at extreme acetonitrile (> 85%) concentrations should give a HILIC stationary phase with a more pure partitioning retention behavior similar to that of a swollen C18 reversed phase. The preparation of a sorbitol methacrylate grafted silica stationary phase is one of our attempts at producing such a stationary phase. The preparation of such a grafted silica has been performed, but with huge difficulty and this work is still far from producing a column of commercial quality and reprodicibility.

This thesis also discusses a new method for the initiation of atom transfer radical polymerization from chlorinated silica. This new grafting scheme theoretically results in a silica particle grafted with equally long polymer chains, anchored to the silica carrier by a hydrolytically stable silicon-carbon bond. The hydrolytic stability is especially important for HILIC stationary phases due to the high water concentration at the surface.

– V –

Table of Contents

1. Introduction ..................................................................................................... 1

2. Liquid Chromatography .................................................................................. 2

3. Why Hydrophilic Separation Materials? ........................................................ 4

4. Hydrophobic Interaction Chromatography .................................................. 5

5. Separation of Hydrophilic Compounds ......................................................... 65.1. Hydrophilic Interaction Chromatography ...................................................... 7

5.1.1. History .......................................................................................................... 8

5.1.2. Mechanism ................................................................................................... 8

6. How to Make Hydrophilic Separation Materials ......................................... 126.1. Polymeric monoliths ...................................................................................... 12

6.2. Silica based ..................................................................................................... 14

7. Controlled Polymerization ............................................................................ 167.1. Atom Transfer Radical Polymerization – ATRP .......................................... 16

8. Expanding the Scope of HILIC ..................................................................... 188.1. Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) ....................... 19

8.2. LC-ICP-MS .................................................................................................... 19

9. Concluding Remarks and Future Aspects .................................................... 20

– VI –

List of Papers

I. Hemström, P., Nordborg, A., Irgum, K., Svec, F., Fréchet, J.M.J., “Polymer-based monolithic microcolumns for hydrophobic interaction chromatography of proteins”; Journal of Separation Science 29 (1), 2006, 25-32.

II. Persson, J., Hemström, P., Irgum, K., “Preparation of a sorbitol methacrylate grafted silica based stationary phase for Hydrophilic Interaction Chromato-graphy”; Manuscript.

III.Hemström, P., Szumski, M., Irgum, K., “Atom-transfer radical graft polymerization initiated directly from silica applied to functionalization of stationary phases for high-performance liquid chromatography in the hydrophilic interaction chromatography mode”; Analytical Chemistry 78 (20),2006, 7098-7103 .

IV. Hemström, P., Nygren, Y., Björn, E., Irgum, K., “Evaluation of Alternative Organic Solvents for Hydrophilic Interaction Chromatographic (HILIC) Separation of Cisplatin Species with On-line Inductively Coupled Plasma Mass Spectrometric (ICP-MS) Detection”; Journal of Separation Science, Submitted.

My contribution to the included papers: I. All lab work, writing most of the paper. II. Supervising the lab work, writing most of the paper. III. Everything from idea to finished paper. IV. Most of the lab work, writing most of the paper.

Relevant papers referred to but not included in Thesis:

Hemström, P., Irgum, K., “Hydrophilic interaction chromatography”; Journal of Separation Science 29 (12), 2006, 1784-1821.

Nygren, Y., Hemström, P., Åstot, C., Naredi, P., Björn, E., “HydrophilicInteraction Liquid Chromatography (HILIC) Coupled to Inductively Coupled Plasma Mass Spectrometry (ICPMS) Utilizing a Mobile Phase with a Low-Volatile Organic Modifier for the Determination of Cisplatin and its mono-Hydrolyzed Metabolite”; Analytical Chemistry, Submitted.

– VII –

List of Abbreviations

ATRP Atom Transfer Radical Polymerization

CRP Controlled Radical Polymerization

DMF Dimethylformamide

DVB Divinylbenzene

ESI-MS Electrospray Ionization Mass Spectrometry

GC Gas Chromatography

HIC Hydrophobic Interaction Chromatography

HILIC Hydrophilic Interaction Chromatography

HPLC High Performance Liquid Chromatography

ICP-MS Inductively Coupled Plasma-Mass Spectrometry

IMAC Immobilized Metal Affinity Columns

LC-ICP-MS Liquid Chromatography Inductively Coupled Plasma-Mass Spectrometry

NPC Normal Phase Chromatography

RAFT Reversible Addition-Fragmentation Chain Transfer (Polymerization)

RF Radio Frequency

RP-LC Reversed Phase High Performance Liquid Chromatography

SEC Size Exclusion Chromatography

SFRP Stable Free Radical Polymerization

TFA Trifluoroacetic acid

ZIC-HILIC Zwitterionic Hydrophilic Interaction Chromatography

– 1 –

1. Introduction

Chromatography is a technique for the separation of components by their difference in affinity for, or solubility in, two different phases. One phase is stationary, usually housed in a tube of glass, plastic or metal (the column) and the other phase (the mobile phase or eluent) is moving through the pore space of the stationary phase. When the sample components (or analytes) are associated with the stationary phase they also are rendered immobile and when they are in the mobile phase they are moving along with the stream of liquid or gas at the speed with which it is moving in the column. Separation is accomplished by compounds spending different time on the stationary phase. In practice this means that compounds will emerge at the column outlet as separate “peaks” at “retention times” that are characteristic of each compound under the current experimental conditions.

The concept of liquid chromatography is more than 100 years old. In March 1903 M.S. Tswett presented a lecture “On a new category of adsorption phenomena and their application to biochemical analysis” to the Biological Section of the Warsaw Society of Natural Sciences. This lecture is normally seen as the birth of chromato-graphy [1, 2]. Using a glass tube packed with calcium carbonate and flushing it with mixtures of organic solvents, Tswett could separate different forms of plant pig-ments into colored bands in the column. He named the technique chromatography, partly some claim after himself refering to the translation of his Russian name. This, the earliest form of chromatography, is still widely practiced today as flash chro-matography [3] mainly by organic chemists for purification of products on in-organic oxides and salts. Since this was the normal way of doing chromatography for so long, it became “normal phase” chromatography. This chromatographic mode had, however, some major drawbacks in that only relatively unpolar sub–stances could be separated since they had to be soluble in the organic solvent, hexane being among the most common nowadays. By coating silica with a hydro-phobic polymer [4] or by covalently attaching an aliphatic hydrocarbon chain (a polyethylene graft) a “reversed phase” was created [5, 6]. These reversed phases, featuring a hydrophobic stationary phase that was eluted with a relatively polar mobile phase, proved to be exceptionally versatile and soon took over as more or less the standard method of separation in liquid chromatography, in particular for small organic molecules such as pharmaceuticals, their metabolites and degradation products.

– 2 –

2. Liquid Chromatography

The modern set up of a High Performance Liquid Chromatography (HPLC) system is outlined in Figure 1. It comprises an eluent or a mixture of eluents and a high pressure pump that drives the eluent through the system. The sample to be separated is introduced by an injector where a pre-filled sample loop is brought online by turning the injector from the load to the inject position. The column where the separation is performed is placed downstream the injector. Finally a detector is needed to record the different substances as they emerge from the column. There is a wide variety of detectors available on the market for use in liquid chromatography, from simple absorbance detectors to tandem mass spectrometers costing millions of SEK.

Eluent Pump

Injector

Column

Detector

Figure 1: Schematic representation of a HPLC set-up.

The difference in retention on the stationary phase in the HPLC column determines how well two compounds are separated in liquid chromatography. There are two different mechanisms that will result in retention. The first case assumes a direct interaction or binding between analyte and stationary phase. In order to detach from the stationary phase the analyte has to be displaced from the surface by the eluent. In the second case the analytes retention is considered as being caused by a partitioning between the bulk eluent and a stationary “liquid” phase. Here the retention time is determined by how well solvated the analyte is in the two phases, the better solvation offered in the stationary phase compared to the eluent, the longer the retention time will be.

– 3 –

The time between the injection of a compound and its appearance at the detector is called its retention time (tR). Included in this is the time the compound spends on the stationary phase (tR') and the time it spends in the mobile phase (tm). For the chromatographist it is the tR' or adjusted retention time that is of interest, this has to be calculated by tR'= tR–tm. In order to compare columns and analytes in different HPLC set-ups it is more convenient to use the retention factor (k) and the number of effective theoretical plates per meter (N/m) of a chromatographic column. The retention factor is given by the equation k = tR'/ tm and the number of plates (N) is calculated as N= a (tR'/w)2, where w is the peak width (in the same unit as tR) at either the base (then a=16) or at half the peak height (then a=5.54). The more plates per meter, the more efficient the column.

The chromatographic properties of a stationary phase are mainly dependent on two factors; selectivity and efficiency. Surface chemistry determines how substances can interact with the stationary phase, i.e., what selectivity the stationary phase has. Efficiency is determined by the number of interactions between a solute and the stationary phase and the axial dispersion of the solute molecules. There is a number of different classes of stationary phases offering different ways for interaction with analytes. The reversed phase has already been mentioned, a non-polar stationary phase (typically an eighteen carbon aliphatic hydrocarbon bound to silica) that is eluted with a more polar eluent, usually a mixture of methanol or acetonitrile with an aqueous buffer. Analytes are separated by their partitioning between the swollen hydrocarbon layer which is enriched in the organic eluent component, and the bulk eluent. Stationary phases for normal phase separations are polar and are eluted with a less polar eluent, often consisting of a mixture of hexane, and methanol or ethyl acetate. The analytes are bound at the surface and elution is taking place by a dis-placement exchange between analyte and solvent [7, 8]. Ion exchange phases con-tain charged groups to which analytes carrying the opposite charge can be attracted by electrostatic interaction. Elution is achieved by competitive displacement using buffered salt solutions.

These are the three most common separation techniques and they exemplify two fundamentally different retention modes in chromatography; partitioning and adsorption. There is also a third separation mode, size exclusion chromatography, where there is (ideally) no enthalpic interaction between the analytes and the stat-ionary phase. Separation is instead based on the fraction of the pore space in the stationary phase that can be accessed by the molecules to be separated. The more pores that are accessible by a molecule, i.e., the smaller it is, the longer time it will take for it to emerge from the column.

– 4 –

3. Why Hydrophilic Separation Materials?

This thesis is dedicated to hydrophilic separation materials, in particular materials for hydrophilic interaction chromatography. The first question to address is; what is the purpose of using hydrophilic stationary phases? Today there is a wide variety of separation materials on the market and a number of these are hydrophilic, so why are they used, and for what?

Ion exchange chromatography of proteins is usually performed either on soft gels based on polysaccharide matrixes, or on rigid particles made of hydrophilized poly-styrene-co-divinyl benzene (Mono-S and Mono-Q), surface modified silica, or on methacrylate based particles (TSKgel BioAssist Q). Dextran, agarose, methacrylates (TSKgel), and polymer-coated silica (Poly LC) are dominating hydrophobic inter-action chromatography. These columns also contain a low amount of hydrophobic ligands with butyl- and phenyl groups being most commonly used. Columns for size exclusion chromatography of water-soluble compounds, especially biomacro-molecules, are either diol modified silica (Zorbax GF-250 and GF-450, TSKgel SW, and others) or crosslinked dextran beads (Sephadex).

Immobilized metal affinity chromatography (IMAC) is a specialized technique for separation of protein or peptides, based on their affinity for an immobilized metal. The metal affinity is often used for selective capture of genetically modified proteins containing a “His-tag” (six histidines in a row at the peptide chain terminal), al-though it is useful for many specialized protein separations, including determinat-ion of post-translational modifications, in particular phosphorylations. In order to minimize secondary interactions, the carrier matrix is generally very hydrophilic. The polyacrylamide gels used for gel electrophoresis of proteins, DNA and other large biomolecules are intended to act only as a sieving medium, obstructing the macromolecules movement through the gel. The result is a separation based on the physical size and charge of the molecules.

All examples listed above are separation techniques in which the driving force be-hind the use of a hydrophilic stationary phase it is not hydrophilicity per se but the minimization of nonspecific binding. Chromatographic modes where the separat-ion mechanism is based on interaction between a hydrophilic stationary phase and the solute are normal phase and hydrophilic interaction chromatography (HILIC). Normal phase chromatography is now a mature discipline where the stationary phases used are neat silica, aminopropyl-, cyano-, or diol-modified silicas. Columns used for HILIC are mainly neat silica (Atlantis) but also silicas carrying polymeric coatings like zwitterionic methacrylate (ZIC-HILIC) and poly(2-hydroxyethyl aspartamide) (PolyHydroxyethyl A) are available. A polymer-based version of the zwitterionic methacrylate material is also available.

– 5 –

Two fundamentally different reasons for wanting to produce hydrophilic separation materials can thus be defined:

Either it should act as an inert solid matrix where a chromatographically active species is anchored, or where size exclusion chromatography can take place; or

The hydrophilicity of the stationary phase is utilized in the chromatographic process.

4. Hydrophobic Interaction Chromatography

As seen in the overview above, one of the dominating applications of hydrophilic separation materials are for protein separation. This is perhaps the most challenging task in liquid chromatography, especially difficult is chromatographic separation of proteins in their native state. Chromatography of proteins can be performed by a number of techniques. Among these, ion-exchange chromatography is the most common, but hydrophobic interaction, size exclusion and reversed phase chroma-tography are also routinely used.

There are, however, some significant differences between the chromatographic be-havior of small molecules and that of large macromolecules like proteins. The large size translates into small diffusion coefficients. Because of this the number of inter-actions with the stationary phase possible per unit of time is lower. Since diffusion is the primary mass transfer mechanism in chromatography, the efficiency in the se-paration of biological macromolecules is inherently lower than that of small mole-cules, under similar conditions. Protein separations therefore rely on selectivity. On a macroscopic scale the selectivity can be explained by the steep rentention curves (log k' vs. log [eluent]) of biomacromolecules. The reason for the steepness of the curves may be rationalized by the high number of “binding points” between solute and surface. In reality this is far more complex because of solvation of both contact surfaces is involved. Thus, if a protein binds to a surface it will not detach until the eluting power of the eluent passes through the concentration where a steep decrease in k' is seen. This is why all chromatography of proteins (except SEC) are per-formed using gradient elution. Once a protein has detached from the stationary phase it will brought along by the eluent gradient at a concentration where the elu-ting strength is high enough to prevent the protein from re-binding to the column [9]. As a consequence of the few interaction events between proteins and stationary phase, columns for protein separation should ideally be short [10] and have wide pores, and there is no need for a large surface area unless intended for preparative chromatography.

– 6 –

Under conditions close to physiological, proteins in their native state can be con-sidered as kinetically stable but thermodynamically instable, and the exposure to the surface of a stationary phase can influence the forces acting on the protein to such an extent that unfolding takes place. That conformational changes occur as a result of exposure to silica particles has, e.g., been shown by circular dichroism [11] and NMR spectroscopy [12]. This propensity for unfolding therefore often pre-cludes prolonged contact with the stationary phase. In methods where long sep-aration times are used, such as in preparative separation on soft gels, the chroma-tography must usually take place in cold rooms.

The work presented in Paper I was carried out during a secondment in the group of prof. Svec at University of California at Berkeley. The task was to prepare a mono-lithic capillary column capable of separating proteins in hydrophobic interaction (HIC) mode. As discussed above the demands on columns for protein separations make the polymeric monolith (further discussed in section 6.1) an ideal choice, especially for micro scale chromatography.

HIC is a “mild” separation technique, this makes it possible to separate proteins while preserving their biological activity [13-15]. It is routinely used in many pro-tein purification protocols as a preparative technique, its analytical use is somewhat limited principally by the difficulty of achieving sharp separations [16]. The pro-teins are loaded in a high salt eluent (usually 1-2 M [NH4]2SO4) and are eluted by decreasing the eluent salt concentration. Increasing the surface tension of water by adding large quantities of cosmotropic salts makes adsorption of the proteins to the stationary phase entropically favorable [17]. The entropy of the water released from the hydration shells of both the protein and the stationary phase contributes to the “binding strength”. The amount of interaction is moderated by having a hydro-philic stationary phase with a low concentration of hydrophobic ligands in order to allow elution by a low salt buffer.

5. Separation of Hydrophilic Compounds

The chromatographic separation of highly hydrophilic compounds has traditionally been regarded as difficult. Gas chromatography has not been an option due to the low volatility, stability and high reactivity inferred by polar functional groups. In liquid chromatography much attention has been focused on how to create retention in RP-LC for compounds with no or very low partitioning into the hydrophobic layer on the stationary phase. When retention has been achieved at all, it has often been through the use of eluents with very low organic solvent contents, an approach that usually leads to inadequate phase wetting and the expulsion of eluent from the

– 7 –

pores [18]. Highly aqueous eluents are therefore known to cause problems like non-reproducible retention times and low separation efficiencies.

Retention of compounds containing one or more charged functional groups can be established by electrostatic interaction. Ion-exchange chromatography is possible for practically all charged solutes, from small inorganic ions to proteins and other biological macromolecules. An alternative to ion exchange is ion pairing, using RP columns that are less expensive and often have better separation efficiency than ion exchange columns. Retention in ion pair chromatography is established by the for-mation of temporary ion pairs, although the separation can also be regarded as taking place in a dynamically coated ion exchanger. Ion pairing reagents typically added to the mobile phase are sodium heptane sulfonate or trifluoroacetic acid (TFA), but ion pairing agents have been shown to significantly reduce the signal in-tensity in electrospray ionization mass spectrometry (ESI-MS) [19].

The lack of viable alternatives has for long forced chromatographers working with highly hydrophilic compounds into using techniques based on RP columns, in spite of them being badly suited for the purpose. Many column manufacturers have ad-dressed the problem of phase collapse by so called polar embedded or polar end-capped phases, especially designated for use with high water content eluents. This has increased the usefulness of RP-LC for hydrophilic substances somewhat, but without addressing the conceptual polarity mismatch.

Ideally very hydrophilic and uncharged compounds should be separated in HPLC by a “reversed reversed phase”, i.e., a separation mode where polar solutes are parti-tioned into a polar stationary phase, while the eluent at the same time offers reason-able solvent properties to provide a fast and linear distribution between the two phases. Hydrophilic interaction chromatography (HILIC) is such a technique.

5.1. Hydrophilic Interaction Chromatography

Hydrophilic interaction chromatography is a variant of normal phase chromato-graphy, where the retention mechanism is believed to be partitioning of the analyte between a water-enriched layer of semi immobilized eluent on a hydrophilic statio-nary phase and a relatively hydrophobic bulk eluent, usually consisting of 5-40 % water in acetonitrile [20]. The boundary between HILIC and normal phase chrom-atography (NPC) is somewhat blurred, especially since some authors have adopted the term “aqueous normal phase” in parallel with HILIC. However the definition proposed by Alpert [20] seems to be gaining acceptance, i.e., the term HILIC should be used a) if the strongly eluting solvent is water and b) the retention mechanism is by partitioning.

HILIC has a number of advantages compared to conventional normal phase chro-matography; HILIC is often more reproducible and eluent preparation is simpler

– 8 –

since there is no need for absolute control over a low water content in the solvents. The solubility of polar compounds is usually better in acetonitrile/water mixtures than in hexane-based eluents and interfacing with electrospray mass spectrometry works very well with the typical low salt acetonitrile/water HILIC eluents, whereas efficient ionization is not as easily achieved with normal phase solvents.

The elution order in HILIC is more or less orthogonal to that seen in reversed phase separations [20], which means that HILIC works best for solutes that are the most problematic in RP. That the typical HILIC eluent is high in acetonitrile, which also gives it two additional advantages over RP-LC; higher sensitivity in ESI-MS [21-23], and faster separations due to the lower viscosity [24].

5.1.1. History

The acronym HILIC was suggested by Alpert in 1990 [20] to describe a chromato-graphic technique employing a hydrophilic stationary phase and a relatively hydro-phobic eluent in which water is the stronger eluting member. Separations operated according to the HILIC principle were introduced in 1975 for the analysis of sugar and oligosaccharides [25, 26], but separations using hydrophilic stationary phases eluted with organic solvent containing low amounts of water were described as early as in the 1950’s [27]. In the ground-breaking work of Alpert [20] he con-cluded that although HILIC is a variant of normal phase chromatography, what dis-tinguishes HILIC from traditional normal phase separations is that it is mainly based on partitioning between the bulk eluent and a partially immobilized layer enriched with water at the stationary phase surface. Support for this theory was mainly drawn from the literature. Prior to that, Rabel et al. [26] in 1976 suggested that the separation of sugars on silica columns using high acetonitrile eluents was a variant of normal phase chromatography. The existence of a water-enriched layer on the surface of hydrophilic stationary phases like ion exchange resins exists was first described by Gregor et al. [28], and the uptake of non-electrolytes in ion exchange resins by means of this stagnant water layer was suggested by Rückert and Samuelson in 1954 [29].

5.1.2. Mechanism

An adsorptive retention model was considered most likely for peptides in HILIC in a recent review by Yoshida [30]. He found linear relationships between the log k' vs. the logarithm of the water content of a series of peptides [31] (Figure 2). The analogy in elution patterns was indicative of a retention mechanism more similar to the surface adsorption established for normal phase separations. Yoshida also sug-gested hydrogen bonding as the principal interaction mode in HILIC [30].

Guo and Gaiki [32, 33] investigated the effect of column temperature on retention by means of van't Hoff plots and were unable to find any strong specific inter-

– 9 –

actions between the solutes and four HILIC stationary phases of very different na-ture. This also supports a partitioning mechanism. The increase in retention time with salt concentration also points indirectly at partitioning as the prevailing mech-anism.

Figure 2: Plots of log k’ vs. log water content for a number of peptides reproduced from ref 31.

In our recent HILIC review [34] we extended the analogy concept used by Yoshida. We based the discussion of the HILIC retention mechanism on plots of retention times published in the literature for different solute classes and stationary phases. These plots were compared with the established retention equations for reversed phase chromatography (where retention is ideally controlled by partitioning only) and normal phase (where retention is by adsorption).

The relationship established for partitioning separations is,

Skk W'log'log [1] ,

where k'W is the retention factor for the weaker eluent component only as mobile phase, is the volume fraction of the stronger member of a binary mobile phase mixture, and S is the slope of log k' vs. when fitted to a linear regression model [35].

– 10 –

In conventional normal phase chromatographic systems, where retention is based on surface adsorption and subsequent displacement of the analyte by eluent mole-cules [7, 8], the relationship between the retention and the mole fraction NB of the stronger solvent B in the eluent should adhere to the following expression [8, 36]:

BB

SB NnAkk log'log'log

[2] ,

where k'B is the solute retention factor with a pure weak eluent, AS and nB are the cross-sectional areas occupied by the solute and the eluent molecules on the surface, and NB is mole fraction of the stronger member in the eluent. Plots of log k' vs. the linear and logarithmical functions of the water contents in the eluents should then ideally give one straight and one bent curve, revealing whether the dominating re-tention mechanism is partitioning or adsorption. There are examples of both types of plots in the literature data, even using the same columns and mobile phases but most plots were inconclusive. It is, however, obvious that something else is involved aside from a straightforward partitioning mechanism, which was previously the generally accepted view. Whether this is a surface adsorption phenomena or due to some other mechanism still remains to be elucidated.

Figure 3: Phase diagram of the acetonitrile-water-NaCl ternary system at 298.2 ± 0.3

K as a function of mole fractions of acetonitrile, xAN

, water, xW

, and xNaCl

. The symbols

, , ×, and represent (1) ternary mixture, (2) phase separation, (3) precipitation

of NaCl, and (4) phase separation with precipitation of NaCl. The solid line repre-

sents the border between ternary mixture and phase separation, and the broken line

shows the solubilities of NaCl in acetonitrile-water mixtures. Reprinted from refer-

ence 37 with permission.

– 11 –

Since it is likely that some form of partitioning is involved in the HILIC retention mechanism, there has to be two at least partially resolved liquid phases in the column. Acetonitrile and water are miscible at any ratio but cooling or addition of salt can induce a phase separation. The phase diagram of acetonitrile and water in the presence of NaCl has been studied by Takamuku [37]. In eluents containing more than 85% acetonitrile no NaCl is soluble, any salt addition results in preci-pitation and phase separation (Figure 3). This is also the composition where the retention in HILIC starts to deviate from the theory for a pure partitioning mechanism.

Takamuku also measured the respective volumes of the two phases in a 1:1 mixture of acetonitrile and water at different salt concentrations. As seen in Figure 4 an in-crease in salt concentration leads to an increased expulsion of acetonitrile from the water rich phase [37], and since NaCl is only solvated by the water molecules the effect of decreasing the water content should be the same.

Figure 4: Volumes for acetonitrile-rich ( ) and water-rich ( ) phases after separation as a

function of mole fraction of total NaCl, xNaCl,tot

. Reprinted from ref 37 with permission.

These findings should be directly translatable to the phase separation induced by a stationary phase that is preferentially solvated by water. It therefore seems likely that the non-linearity seen when plotting log k' vs. % water in the eluent for some HILIC separations arises from a non-linear change in composition of the retained water rich phase on the stationary phase.

Currently the use of buffered eluents is recommended by all column manufacturers to reduce electrostatic interactions between charged analytes and dissociated silanol groups or other charged species on the stationary phase. These mixed mode inter-actions play an important role when separating charged molecules on HILIC media.

– 12 –

Ion exchange effects for basic or positively charged species and electrostatic repul-sion of negatively charged species is expected to lead to expulsion from the pore space, especially in silica based separation materials. The most commonly recom-mended buffer salts are ammonium salts of formate or acetate due to their high solubility in acetonitrile rich eluents, suitable buffering range, and compatibility with MS detection.

6. How to Make Hydrophilic Separation Materials

In this thesis hydrophilic separation media have been prepared by two main routes. Hydrophilic macroporous monoliths and grafted monodisperse porous silica mic-rospheres.

6.1. Polymeric monoliths

Chromatographic separation media cast in one piece, directly in the column, is an attractive idea for several reasons. Pumping the mobile phase through pores in the stationary phase decreases the dependence on diffusion for mass transfer since there is also convective mass transfer. Having just one large particle has the added ad-vantage of making column packing and retaining frits unnecessary.

Figure 5: Showing a methacrylate monolith that has detached from the wall

of a fused silica capillary. Reprinted from reference 38 with permission.

– 13 –

It is, however, essential that the stationary phase is firmly anchored to the column wall or large voids can appear or the monolith can be pushed out of the column (Figure 5) [38]. The entire stationary phase is in principle a single entity and the mobile phase is pumped through this porous body. Thus, in order for a reasonable flow rate through and pressure drop over the column, there has to be a set of trans-secting and relatively large through pores (macropores) in the material. The con-tinuous polymer bed introduced by Hjertén [39], the rigid polymer rods of Svec [40], and the monolithic silica of Nakanishi and Tanaka [41] initiated an explosive development of this new field of separation materials over the last 15 years [42]. These materials have been known under a large number of names but are now commonly referred to as monolithic separation materials or simply monoliths [43]. The Svec type polymeric monoliths have mainly been prepared using acryla-tes/methacrylates or styrene/DVB copolymers resulting in hydrophobic stationary phases. The main applications for these columns have been for reversed phase sep-arations of large molecules, notably proteins and peptides. Such molecules have low diffusion coefficients, whereby separation becomes more dependent on selectivity than on efficiency. Maximum advantage is thus taken of the convective mass transfer component and limiting the main disadvantage of common polymeric monoliths, which is low surface area.

Monolithic polymer columns are prepared from a homogenous solution of mon-omers, porogens and an initiator, typically polymerized directly in the final separation column. The monomer part is usually a mixture of mono- and divinyl containing (meth)acrylates or styrenics. Development of a new polymer monolith is an iterative process where a number of factors have to be optimized. In Paper I,hydrophilic monoliths for hydrophobic interaction chromatography were prepared. The surface chemistry of the final monolith is mainly determined by the monomers used in the preparation but surface modification (e.g., grafting) of the monolith may also be used. The next step to perform after having decided on monomers is to search for a suitable porogen or mixture of porogens. The porogens have to be miscible with the monomers but should not solubilise the growing polymer chains too well, otherwise the phase separation and aggregation necessary for macropore formation will not take place. A more detailed look at the pore formation and struc-ture has been published elsewhere [44]. As shown in Paper I, different porogen systems can yield monoliths with similar pore sizes; in this case water/1-propan-ol/1,4-butanediol and 1-dodecanol/cyclohexanol. My initial hypothesis was that, although prepared from the same monomers, the distribution of these monomers inside the polymer backbone should be affected because of the large difference in polarities of the chosen porogen systems (in particular the presence of water in the system first mentioned). The similarity in chromatographic performance seen for the two different monoliths was therefore quite surprising. It is, however, possible to interpret the slightly more tailing lysozyme peak for the column prepared using

– 14 –

the more hydrophobic porogen mixture (figure 6 pek 1) as a slightly higher hydro-phobicity for this monolith. It is admitted that this is an extremely circumstantial piece of “evidence”, but there are, as far as I know, no other investigations on the effect of the porogenic solvent on stratification of monomers in the final monolith.

In Paper I only relatively hydrophilic proteins could be separated on the monolithic HIC column; the more hydrophobic proteins tested (BSA, ovalbumin) could not be eluted after binding. These polymeric monoliths are most likely not well suitable for HIC applications and the main reason for failing to produce a good HIC material is believed to be the need to prepare a homogeneous polymerization mixture, which precludes the use of monomers with widely differing polarity. This leads to a hydro-phobicity that will be too evenly distributed on the material with an increased risk of protein unfolding.

6.2. Silica based

Particulate porous silica is by far the most common starting material for the pre-paration of chromatographic stationary phases. Unmodified silica is polar and has successfully been used as both NPC and HILIC stationary phases. However, the most commonly used route to produce silica-based stationary phases for chromate-graphy is through surface modification [45]. The preparation of reversed bonded phases by attachment of tri-chloro or tri-methoxy octadecylsilanes to porous silica particles set the standard and almost all surface modification of silica is performed using silane coupling reactions. In Paper II the objective was the surface modifi-cation of porous silica with sorbitol methacrylate (Figure 6).

Figure 6: Sorbitol methacrylate.

Today all commercially available HILIC columns are either neat silica or contain some form of charged species (or are kept secret by the manufacturer), making them in essence mixed mode materials. By covering the porous silica beads with highly hydrophilic carbohydrate containing methacrylate brushes, a stationary phase with a thick, non-charged water-retaining layer should result. Such a thick interactive layer might allow for HILIC with a chromatographic behavior more similar to the partitioning of solvent-swollen C18 reversed phase stationary phases.

– 15 –

A layer of polymer-bound sorbitol should provide for a highly water retaining stationary phase since it binds water strongly (it is, e.g., used to retain moisture in toothpaste) and does not crystallize if wet (it is also used to prevent the crystal-lization of sugar). In Paper II we prepared such a stationary phase by polymerize-tion from a surface-bound initiator, which results in a polymer brush attached at one end to the particle and extending into solution. The preparation of a surface-bound initiator can be either by in-situ synthesis on the surface [46-48], or by for-ming a self-assembled monolayer of initiator on the substrate [49-51]. The attach-ment of tert-butyl hydroperoxide and the subsequent initiation of polymerization from this initiator has previously been used successfully in our laboratory [52]. Using standard silane chemistry to attach pre-formed oligomers of sorbitol meth-acrylate on silica is not feasible since the silanes will react just as well with the hy-droxyl groups on the oligomers as with the silanol groups on the silica surface and hence form a crosslinked network of polymer on the particle surface. One weakness shared by both the silanization and the tert-butyl hydroperoxide reactions is the hydrolytically labile Si-O-C bond that will link the polymer brush to the silica support. It is known that water-repelling bonded phases are more stable toward hydrolytic attack [53]. For example long alkyl chains protect silica bonded phases to some extent against hydrolysis [35]. In HILIC, where the retention mechanism is dependent on retaining large amounts of water at the surface of the stationary phase, the susceptibility to hydrolysis should hereby also be increased. Anchoring the polymer brushes by a hydrolytically stable bond is thus more important when making stationary phases for HILIC and other highly polar stationary phases.

Preparing a polymer brush stationary phase by conventional free radical polymer-ization will yield a highly polydisperse polymer coating due to the inherent poly-dispersity of normal radical polymerizations. The zeroth order decomposition kine-tics of radical initiators means that only the first few polymer chains to be initiated by the surface tethered initiator will have good access to monomer and be suffici-ently sterically unhindered to reach a high degree of polymerization. Relatively low monomer concentrations have to be used to avoid total clogging of the pore system, and the monomer inside the pore will rapidly be consumed. Chains growing in the pore entrance will have a higher likelihood of incorporating fresh monomer diffu-sing into the pore space, leading to a bottleneck effect with decreased effective pore size and limited functionalization of the inner parts of a pore [54]. The solution to this problem is to reduce the polymerization rate drastically, which would allow for a diffusion of monomer from outside the particles into its entire pore space. It is also possible to limit the monomer consumption in a particle by using a low poly-merization temperature thus having a slow initiation rate, in combination with a low monomer concentration limiting the amount of available monomer for incorporation by each initiation event, it is thus possible to produce good polymer grafted stationary phases [55]. As described in Paper II, sorbitol methacrylate was

– 16 –

not polymerizable in any good solvent (solvent capable of dissolving the monomer) and hence neither the monomer concentration nor the polymerization temperature were really accessible as tuning parameters for producing a homogeneous grafted layer. Despite these problems and limitations, a sorbitol methacrylate grafted stationary phase remains a priority target due to the interesting chromatographic selectivity in HILIC mode shown in Paper II. So far our attempts at increasing the sorbitol methacrylates polymerizability through derivatization have been fruitless.

7. Controlled Polymerization

Paper III describes an entirely new way of preparing polymer grafted silica-based stationary phases, a procedure that should address both the low hydrolytic stability and the uneven surface coverage of such particles. Slowing the rate of polymeriza-tion to such an extent that diffusion can hold the monomer concentration inside the pores constant during the polymerization should result in a more uniform film thickness and decrease the risk of blocked pores. There are numerous types of controlled polymerizations described in the literature. The living ionic poly-merization [56] was the first but nowadays stable free radical polymerization (SFRP) [57], reversible addition-fragmentation chain transfer (RAFT) [58], and atom transfer radical polymerization (ATRP) [59] are the most used. The rate of these polymerizations is controlled by trapping the growing radicals in a “dormant” (inactive) state or by transferring them to a different polymer chain. The possibility of initiating polymerization from a halide atom sets ATRP apart from the other controlled polymerizations and was the reason for our interest in ATRP for grafting from silica.

7.1. Atom Transfer Radical Polymerization – ATRP

ATRP is catalyzed by a transition metal complex (often CuBr complexed by two 2,2'-bipyridine molecules) but has been shown to work for a large number of other metals like titanium [60], molybdenum [61], rhenium [62], iron [63], ruthenium [64], osmium [65], rhodium [66], cobalt [67], nickel [68], and palladium [69]. A halide atom of the initiator or growing polymer chain is homolytically cleaved, generating a radical at the site where it was attached, in the procee oxidizing the Cu+

ion to Cu2+. The generated radical then reacts with monomer present by the stan-dard radical polymerization mechanism [70-75] (Figure 7). The resulting Cu(II) complex can act as a radical scavenger, combining with the radical on the chain ext-ended growing polymer, thereby deactivating it. The observed polymerization rate (kATRP), is thus a balance between the rate of activation (kact), the rate of deactivation (kdeact), and the radical polymerization rate (kp). The relationship between structure

– 17 –

Figure 7: Initiation of polymerization from a standard bromo-

isobutyrate type ATRP initiator grafted on a silica substrate.

and reactivity in ATRP are complex but has been excellently reviewed by Braun-ecker and Matyjaszewski [76]. In Paper III ATRP was initiated directly from a silica surface, where the silanol groups had been replaced with chlorine atoms (Figure 8).

Figure 8: Scheme showing the initiation of ATRP from chlorinated silica.

This novel way of initiation has a number of benefits compared with the traditional silane coupling approach where the ATRP initiator is attached by a silanol conden-sation reaction [77-80]. The initiation rate from such an initiator should be very high due to the high reactivity of this chlorine atom, a high initiation rate compared to the propagation rate yields low polydispersity index of the formed polymer. The high reactivity can, however, be a problem since chlorinated silica reacts violently with water and alcohols, but the reactivity seems to drop markedly upon the addi-tion of the ATRP catalyst and ligands. Polymerization can thus be performed in

– 18 –

presence of a few percent water, if added after the ATRP catalyst and ligands. Initiating ATRP from a chlorinated silica particle also should infer another large advantage over attaching a traditional ATRP initiator by silylation; it forms a direct covalent bond between the silica substrate and the carbon polymer chain (a Si-C bond) instead of a silicon-carbon ether (Si-O-C). The preparation of chlorinated silica is simple and fast compared with the four step synthesis of (11-(2-Bromo-2-methyl)propionyloxy)undecyl trichlorosilane [80] modified silica.

The conclusions that can be drawn from Paper III are that the synthesis procedure works for a number of monomers, but so far we have not been successful in using this method to polymerize sorbitol methacrylate. The reproducibility of the grafting using this method was not excellent. Initially, of two duplicate samples processed in parallel, one would yield grafted silica and the other would not. As the polymerization scheme was developed, mainly by the use of a glove box, these pro-blems diminished but still duplicate samples prepared on different days would yield different results. The most likely explanation to this is the sensitivity of the initiator to moisture, hydrolysis of the chlorine to a silanol or reactions producing other group could occur during any of the steps in the procedure.

It is also difficult to evaluate the grafting, mainly since the anchoring Si-C bond leaves no easy way to cleave the polymer brushes from the silica substrate. Dis-solving the silica in NH4HF2 (aq) [81] is a possibility, but when this approach was applied to a glycidyl methcrylate grafted silica, a white slime insoluble was produced in all solvents tested and the attempts were discontinued.

8. Expanding the Scope of HILIC

Although this thesis deals with hydrophilic separation materials in general, my main interest has been the development and advancement of HILIC. It has been extremely rewarding to see this technique evolving, in a time span as short as this thesis work, from an obscurity practiced only by a few visionaries [82-87] into one of the hottest topics discussed at the HPLC 2007 conference in Ghent, Belgium. It is still far from a mature field; the mechanism is still not fully established and many of the columns vigorously marketed for HILIC separations are in reality conventional normal phase silica or ion exchange columns engaged as a stop gap solution to capture market shares in this rapidly emerging market. The emergence of post-genomic science areas such as metabolomics and glycomics have, in combination with the advantageous high sensitivity in ESI–MS, expanded the applicability of HILIC from only being used solely for carbohydrate analysis, via peptides, to small molecules and even whole proteins [88]. In a recent paper [89] we also introduced

– 19 –

HILIC coupled to inductively coupled plasma mass spectrometry (ICP-MS). This hyphenation is problematic because the HILIC eluents must contain large amounts of organic solvents, which interfere with the plasma processes and produce deposits in the mass spectrometer inlet.

8.1. Inductively Coupled Plasma-Mass Spectrometry (ICP-MS)

ICP-MS is an analytical technique where solutes are atomized and element ions are generated in an argon plasma and then analyzed by a standard mass spectrometer. It is the most sensitive technique for trace elemental analysis (detection limits in fg range are possible), and has a wide linear range (eight orders of magnitude in fav-orable cases). Any element with an atomic weight above 7 can be analyzed, but it is most commonly used for metal and metalloid elements.

The plasma is formed at the tip of a torch where a stream of argon is passed through the space surrounded by an induction coil, connected to a continuous wave RF generator with 1-2 kW output at 27.12 or 40.68 MHz. Once ignited by a Tesla coil, the high frequency induced magnetic field couples with the charged species in the gas converts the argon into a plasma with gas temperatures of up to 8000 °C. The high temperature in combination with a relatively long (~ 2 ms) residence time of the sample in the plasma zone leads to almost complete atomization and high degree of ionization [90]. Sample is usually introduced in the center of the plasma in the form of a liquid aerosol formed in a pneumatic nebulizer. The ions generated are introduced through two conical metal discs with narrow orifices (the sampler and skimmer cones) into the mass spectrometer, by pressure difference (the plasma is at atmospheric pressure and the mass is at 10-4 Pa).

8.2. LC-ICP-MS

The preference for liquid samples and an ability to accommodate (need for) rela-tively large sample flow rates compared to other LC-MS interfaces makes coupling of conventional HPLC to ICP-MS relatively straightforward. Ion exchange, size exclusion, reversed phase, ion-pair, and chiral chromatography are LC subtech-niques that have previously been used in combination with ICP-MS [91]. The only problems generally encountered are salt encrusting using high salt eluents in ion exchange chromatography, and deposition of carbon on ICP sampler and skimmer cones at high organic load on the plasma [92]. Removal of organic solvents from the eluent prior its introduction into the plasma is normally performed using chilled spray chambers, membrane-based or cryogenic desolvation systems. Any remaining carbon in the plasma can usually be combusted by the addition of oxygen to the nebulizer gas making, LC-ICP-MS a reasonably stable system.

– 20 –

However, the sensitivity and instrument wear will often increase with eluents rich in organic solvent, to which HILIC eluents belong.

By using a high boiling organic solvent, dimethylformamide (DMF), instead of acetonitrile as the organic member of the HILIC eluent we could eliminate the carbon deposition on instrument parts without adding oxygen. This led to a ten-fold increase in instrument sensitivity enabling the detection of cisplatin (a potent anti-cancer agent) in cells treated with concentrations close to those administered clinically [89]. The chromatographic performance of DMF and a number of other high boiling organic solvents for the separation of cisplatin species by HILIC are evaluated in Paper IV. Of the tested solvents, 1,4-dioxane and 1-propanol yielded good chromatographic behavior for cisplatin and showed no signs of overloading the plasma with carbon. However, the use of high boiling eluents for chroma-tography is generally not a good idea due to the inevitable increase in viscosity with boiling point. When switching from acetonitrile to 1,4-dioxane or 1-propanol we found a 50 % loss of column efficiency for a 100x2.1 mm ZIC-HILIC column when operated at the recommended flow rate of 0.1 ml/min (Paper IV). Paying this price in efficiency obviously makes sense only if there is something to be gained from switching solvent. In the case of cisplatin there were two benefits; cisplatin reacted with acetonitrile forming new species during the sample preparation and separa-tion, and secondly an almost complete removal of carbon from the plasma lead to increased sensitivity. This increase in sensitivity should be generic in all LC-ICP-MS applications where eluents of high organic contents are used.

9. Concluding Remarks and Future Aspects

Hydrophilic separation materials have a long and successful history, from the early work of Tswett via “gel filtration” and the acrylamide “slab” gels for gel electro-phoresis. Still I believe that the golden age of hydrophilic separations is just about to start. With the emergence of HILIC it is for the first time possible to get truly effi-cient separation of hydrophilic small molecules giving a formidable boost to the “metabolomics” field. The possibility for fast separation of glycosylated and phos-phorylated peptides, complex carbohydrates and organic acids using eluents perfect for electrospray mass spectrometry will make for breakthroughs few would even have dared dream about just ten short years ago. In the words of Tom Petty “the future is wide open” [93].

– 21 –

References

[1] Ettre, L.S., LC GC Mag. 2003, 31, 458-467. [2] Engelhardt, H., J. Chromatogr. B 2004, 800, 3-6. [3] Still, W.C., Kahn, M., Mitra, A., J. Org. Chem. 1978, 43, 2923-2925. [4] Martin, A.J.P., Synge, R.L.M., Biochem. J. 1941, 35, 1358-1368. [5] Kirkland, J.J., DeStefano, J.J., J. Chromatogr. Sci. 1970, 8, 309-314. [6] Majors, R.E., Anal. Chem. 1972, 44, 1722-1726. [7] Snyder, L.R., Anal. Chem. 1974, 46, 1384-1393. [8] Snyder, L.R., Poppe, H., J. Chromatogr. 1980, 184, 363-413. [9] Dubinina, N.I., Kurenbin, O.I., Tennikova, T.B., J. Chrom. A 1996, 753, 217-225. [10] Kalghatgi, K., Horvath, C., J Chrom. 1987, 398, 335-339. [11] Lundqvist, M., Sethson, I., Jonsson, B.H., Langmuir 2004, 20, 10639-10647. [12] Lundqvist, M., Sethson, I., Jonsson, B.H., Biochemistry 2005, 44, 10093-10099. [13] Hjertén, S., J Chromatogr 1973, 87, 325-331. [14] Regnier, F., Science 1987, 238, 319-323. [15] Fausnaugh, J.K., Kennedy, L.A., Regnier, F.E., J. Chromatogr. 1984, 317, 141-155. [16] Scopes, R.K., Protein Purification, Springert-Verlag, 1994.[17] Esquibel-King, M.A., Dias-Cabral, A.C., Queiroz, J.A., Pinto, N.G.,

J. Chromatogr. A 1999, 865, 111-122. [18] Walter, T.H., Iraneta, P., Capparella, P., J. Chromatogr. A 2005, 1075, 177-183. [19] Gustavsson, S.Å., Samskog, J., Markides, K., Långström, B., J. Chromatogr. A 2001,

937, 41-47. [20] Alpert, A.J., J. Chromatogr. 1990, 499, 177-196. [21] Grumbach, E.S., Wagrowski-Diehl, D.M., Mazzeo, J.R., Alden, B., Iraneta, P.C.,

LC GC North America 2004, 22, 1010-1023. [22] Naidong, W., J. Chromatogr. B 2003, 796, 209-224.[23] Shou, W.Z., Naidong, W., J. Chromatogr. B 2005, 825, 186-192.[24] Shou, W.Z., Chen, Y.L., Eerkes, A., Tang, Y.Q., Magis, L., Jiang, X.Y., Weng, N.D.,

Rapid Commun. Mass Spectrom. 2002, 16, 1613-1621. [25] Linden, J.C., Lawhead, C.L., J. Chromatogr. 1975, 105, 125-133.[26] Palmer, J.K., Anal. Lett. 1975, 8, 215-224. [27] Samuelson, O., Sjöström, E., Sven. Kem. Tidskr. 1952, 64, 305-314. [28] Gregor, H.P., Collins, F.C., Pope, M., J. Colloid Sci. 1951, 6, 304-322. [29] Rückert, H., Samuelson, O., Sven. Kem. Tidskr. 1954, 66, 337-344. [30] Yoshida, T., J. Biochem. Biophys. Meth. 2004, 60, 265-280. [31] Yoshida, T., J. Chromatogr. A 1998, 811, 61-67.[32] Guo, Y., Gaiki, S., J. Chromatogr. A 2005, 1074, 71-80.[33] Guo, Y., Huang, A.H., J. Pharm. Biomed. Anal. 2003, 31, 1191-1201. [34] Hemström, P., Irgum, K., J. Sep. Sci. 2006, 29, 1784-1821. [35] Poole, C. F., The Essence of Chromatography, Elsevier, Amsterdam 2003. [36] Nikitas, P., Pappa-Louisi, A., Agrafiotou, P., J. Chromatogr. A 2002, 946, 33-45.

– 22 –

[37] Takamuku, T., Yamaguchi, A., Matsuo, D., Tabata, M., Kumamoto, M., Nishimoto, J., Yoshida, K., Yamaguchi, T., Nagao, M., Otomo, T., Adachi T., J Phys. Chem. B 2001, 105, 6236-6245.

[38] Courtois, J.; Szumski, M.; Byström, E.; Iwasiewicz, A.; Shchukarev, A.; Irgum, K.J. Sep. Sci. 2006, 29, 14-24

[39] Hjertén, S., Liao, J., Zhang, R., J. Chromatogr. 1989, 473, 273-275. [40] Svec, F., Frechét, J.M.J., Anal. Chem. 1992, 64, 820-822. [41] Minakuchi, H., Nakanishi, K., Soga, N., Ishizuka, N., Tanaka, N., Anal. Chem.1996,

68, 3498-3501. [42] Svec, F., J. Sep. Sci. 2004, 27, 1419-1430. [43] Viklund, C., Svec, F., Frechet, J.M.J., Irgum, K., Chem Mater 1996, 8, 744-750. [44] Viklund, C., Monolithic Columns for Analytical Flow Applications, PhD Thesis,

Umeå University 2001.[45] Unger, K.K., Porous silica, J. Chromatogr. Lib. Vol. 16.; Elsevier: Amsterdam; 1979.[46] Frey, N., Laible, R., Hamann, K., Angew. Makromol. Chem. 1973, 34, 81-109. [47] Carlier, E., Guyot, A., Revillon, A., React. Polym. 1992, 16, 115-124. [48] Tsubokawa, N., Ishida, H.J. Polym. Sci Part A: Polym Chem 1992, 30, 2241-2246. [49] Prucker, O., Ruhe, J., Langmuir 1998, 14, 6893-6898. [50] Prucker, O., Ruhe, J., Macromolecules 1998, 31, 592-601. [51] Prucker, O., Ruhe, J., Macromolecules 1998, 31, 602-613. [52] Jiang, W., Irgum, K., Anal. Chem. 2002, 74, 4682-4687. [53] Hetem, M.J.J., de Haan, J.W., Claessens, H.A., van de Ven, L.J.M., Cramers, C.A.,

Kinkel, J.N.J., Anal. Chem. 1990, 62, 2288-2296. [54] Revillon, A., Leroux. D., React. Funct. Polym. 1995, 26, 105-118. [55] Jiang, W., Fischer, G., Girmay, Y., Irgum, K., J Chromatogr A 2006, 1127, 82-91. [56] Szwarc, M., Levy, M., Milkovich, R., J. Am. Chem. Soc 1956, 78, 2656-2657. [57] Georges, M.K., Veregin, R.P.N., Kazmaier, P.M., Hamer, G.K., Macromolecules

1993, 26, 2987-2988. [58] Chiefari, J., Chong, Y.K., Ercole, F., Krstina, J., Jeffery, J., Le, T.P.T.,

Mayadunne, R.T.A., Meijs, G.F., Moad, C.L., Moad, G., Rizzardo, E., Thang, S.H., Macromolecules 1998, 31, 5559-5562.

[59] Wang, J.S., Matyjaszewski. K., J Am. Chem. Soc. 1995, 117, 5614-5615. [60] Kabachii, Y.A., Kochev, S.Y., Bronstein, L.M., Blagodatskikh, I.B., Valetsky, P.M.,

Polym. Bull. 2003, 50, 271-278. [61] Brandts, J.A.M., van de Geijn, P., van Faassen, E.E., Boersma, J., van Koten, G.,

J. Organomet. Chem. 1999, 584, 246-253. [62] Kotani, Y., Kamigaito, M., Sawamoto, M., Macromolecules 1999, 32, 2420-2424. [63] Matyjaszewski, K., Wei, M.L., Xia, J.H., McDermott, N.E., Macromolecules 1997,

30, 8161-8164. [64] Kato, M., Kamigaito, M., Sawamoto, M., Higashimura, T., Macromolecules 1995,

28, 1721-1723. [65] Braunecker, W.A., Itami, Y., Matyjaszewski, K., Macromolecules 2005, 38, 9402-

9404.[66] Percec, V., Barboiu, B., Neumann, A., Ronda, J.C., Zhao, M.Y., Macromolecules

1996, 29, 3665-3668.

– 23 –

[67] Wang, B.Q., Zhuang, Y., Luo, X.X., Xu, S.S., Zhou, X.Z., Macromolecules 2003, 36, 9684-9686.

[68] Granel, C., Dubois, P., Jerome, R., Teyssie, P., Macromolecules 1996, 29, 8576-8582. [69] Lecomte, P., Drapier, I., Dubois, P., Teyssie, P., Jerome, R., Macromolecules 1997,

30, 7631-7633. [70] Haddleton, D.M., Crossman, M.C., Hunt, K.H., Topping, C., Waterson, C.,

Suddaby, K.G.Macromolecules 1997, 30, 3992-3998. [71] Matyjaszewski, K., Macromolecules 1998, 31, 4710-4717. [72] Lutz, J. F., Neugebauer, D., Matyjaszewski, K., J. Am. Chem. Soc. 2003, 125, 6986-

6993.[73] Matyjaszewski, K., Paik, H., Shipp, D.A., Isobe, Y., Okamoto, Y., Macromolecules

2001, 34, 3127-3129. [74] Singleton, D.A., Nowlan, D.T., Jahed, N., Matyjaszewski, K., Macromolecules 2003,

36, 8609-8616. [75] Wang, A.R., Zhu, S.P., Macromolecules 2002, 35, 9926-9933. [76] Braunecker, W.A., Matyjaszewski, K., Prog. Polym. Sci. 2007, 32, 93-146. [77] Huang, X., Wirth, M., Anal. Chem. 1997, 69, 4577-4580. [78] Ejaz, M., Yamamoto, S., Ohno, K., Tsujii, Y., Fukuda, T., Macromolecules 1998, 31,

5934-5936.[79] Perruchot, C., Khan, M.A., Kamitsi, A., Armes, S.P., von Werne, T., Patten, T.E.,

Langmuir 2001, 17, 4479-4481. [80] Matyjaszewski, K., Miller, P.J., Shukla, N., Immaraporn, B., Gelman, A.,

Luokala, B.B., Siclovan, T.M., Kickelbick, G., Vallant, T., Hoffmann, H., Pakula, T., Macromolecules 1999, 32, 8716-8724.

[81] Titirici, M.M., Sellergren, B., Anal. Bioanal. Chem. 2004, 378, 1913-1921. [82] Zhu, B.Y., Mant, C.T., Hodges, R.S., J. Chromatogr. 1991, 548, 13-24. [83] Alpert, A.J., Shukla, M., Shukla, A.K., Zieske, L.R., Yuen, S.W., Ferguson, M.A.J.,

Mehlert, A., Pauly, M., Orlando, R., J. Chromatogr. A 1994, 676, 191-202. [84] Oyler, A.R., Armstrong, B.L., Cha, J.Y., Zhou, M.X., Yang, Q., Robinson, R.I.,

Dunphy, R., Burinsky, D.J., J. Chromatogr. A 1996, 724, 378-383. [85] Yoshida, T., Anal. Chem. 1997, 69, 3038-3043. [86] Mant, C.T., Kondejewski, L.H., Hodges, R.S., J. Chromatogr. A 1998, 816, 79-88. [87] Strege, M.A., Anal. Chem. 1998, 70, 2439-2445. [88] Lindner, H., Sarg, B., Meraner, C., Helliger, W., J. Chromatogr. A 1996, 743, 137-

144.[89] Nygren, Y., Hemström, P., Åstot, C., Naredi, P., Björn, E., Analytical chemistry

Submitted. [90] Montaser, A., (Ed.), Inductively coupled plasma mass spectrometry, Wiely-VCH

1998.[91] Wang, T.B., J. Liq. Chromatogr. Relat. Technol. 2007, 30, 807-831. [92] Michalke, B., TrAC-Trends Anal. Chem. 2002, 21, 154-165. [93] Lynne, J., Petty, T., Cambpell, M., MCA records 1991.

– 24 –

Acknowledgements Först och främst måste jag tacka Knut för den enorma frihet jag haft att göra precis det jag har velat och dessutom fått betalt för det. Visst ibland har jag känt mig som Peter Dalle när jag kommit farande med ”jag har en idé!!!” lika ofta som inte slutade det också med ett ”-tänkte inte på det”. Jag har dock lärt mig otroligt mycket kemi men också att ta vara på och tro på mig själv.

Ett stort tack också till Einar för din uppmuntran och allmänna positivism, utan det hade jag nog aldrig börjat doktorera. Alla andra på SeQuant (som alla var doktorander när jag började) Camilla, Wen och Tobias för att ni tog hand om mig och gjorde att jag trivdes och speciellt Patrik som nog mer än någon annan är ansvarig för att jag hamnade där jag är. Det finns även ett par gamla rävar till som är ansvariga för att jag blev doktorand Martin, Göran, Fredrik, Pelle, Andreas (vart du nu tog vägen)

Tack alla doktorander nya och gamla på analytisk kemi Erika, Mai, Emil, Fredrik, Julien, Tom, Lars, Yvonne, Johanna, Dong, Daniel, Sofi, James och naturligtvis min ständige vapen-dragare i vått och torrt (och Berkeley) Anna. Alla andra som passerat avdelningen under åren men speciellt Jonas, Jeroen, Michal, Gerd, Nhat, Duc, och Thuy.

A great big thanks also to the gang in Berkeley; Frank, Emily, Kelly, Bas, Dewey, Tim, Dean, Vincenzo and of course Professor Fréchet it was a great experience and I learned a lot from the long discussions about stuff and other stuff with Emily.

Sen måste jag väl, även om det bär emot att tacka ett gäng som sitter och snackar skit och super var och varannan torsdag. Tack till Whisky klubben.

Sist men mest naturligtvis tack till min Elin trots att du fått mig att bli ett allmänt åtlöje pga. mitt konstanta mobil pratande de senaste två åren, även din side-kick som tvingat upp mig ur sängen varje morgon genom att stå på mig och vråla MJAU i mitt öra eller slicka mig på näsan.