In-line application of electric fields in capillary separation systems

Faculty of Technology and Science C hem istry

D ISSE R TA TIO N

B jö rn E rik sson

In-line ap p lication of electric fi elds

in cap illary sep aration system s

B jö rn E rik sso n

In-line ap p lic atio n o f elec tric fi elds

in c ap illary sep aratio n system s

Björn Eriksson. In-line application of electric fields in capillary separation systems

D ISSER T A T IO N

Karlstad University Studies 2006:55 ISSN 1 4 03 -8 09 9

ISBN 9 1 -7 063 -08 6-0

© T h e auth or

D istrib ution:

Karlstad University

F aculty of T ech nolog y and Science C h emistry

SE-651 8 8 KA R L ST A D SW ED EN

+ 4 6 54 -7 00 1 0 00

w w w .kau.se

Abstract

The magnitude of an electric field possible to apply in a capillary separation system is limited, because a high electric field causes a too high current through the capillary. Application of the electric field in-line will give an increased conductivity in the column, further increasing the risk of too high currents. The conductivity changes were found to result from an overall increase in ionic strength within the electric field. The increase in ionic strength is caused by the increase in mobile phase ions with electrophoretic velocity against the flow, together with OH^- or H3O⁺ ions (depending on polarity) formed at the inlet electrode. Further it was found that the use of a pressurized reservoir or splitting of the flow at the inlet electrode could significantly limit the conductivity changes and thereby the maximum applicable electric field strengths could be increased.

List of abbreviations

A Cross section area (m

)

c Concentration (M)

CE Capil l ary el ectrophoresis

CEC Capil l ary el ectrochromatography CZE Capil l ary zone el ectrophoresis E El ectric fiel d strength (kV/cm) EOF El ectroosmotic fl ow

F Faraday constant (96485 C/mol )

HPLC High performance l iquid chromatography

HV High vol tage

I El ectric current (µA) I. D. Inner diameter (µm)

κ Conductivity (1/Ωm)

L Length (cm)

LC Liquid chromatography

MS Mass spectrometry

NMR Nucl ear magnetic resonance spectroscopy µ El ectrophoretic mobil ity (m

/Vs)

O. D. Outer diameter (µm)

P Effect (W )

PEEK Pol yetheretherketone R El ectric resistance (Ω)

U Vol tage (V)

UV Ul traviol et

v

El ectrophoretic vel ocity (m/s)

z Ion charge

List of papers

Deviation from Ohm’s law in electric field assisted capillary LC Björn O. Eriksson, Magnus B.O. Andersson, Lars G. Blomberg, J. Chromatography A 1010 (2003) 17-24.

Paper II

Changes in m obile phase ion distribution when combining pressurized flow and electric field

Björn O. Eriksson, Magnus Dahl, Magnus B.O. Andersson, Lars G. Blomberg, Electrophoresis 25 (2004) 3092-3097.

Paper III

Flow splitting at the inlet electrode as a m ethod for decreasing the electric current in electric field assisted liquid chrom atography Björn O. Eriksson, Magnus B.O. Andersson, Lars G. Blomberg, J. Chromatography A 1119 (2006) 170-175.

Paper IV

In-line application of electric field in capillary separation system s Part I: Joule heating, pH and conductivity

Björn O. Eriksson, Nicola Marlin, Magnus B.O. Andersson, Lars G. Blomberg, Submitted to J. Chromatography A.

Paper V

In-line application of electric field in capillary separation system s Part II: Separation param eters affecting the conductivity

Björn O. Eriksson, Inger Lill Skuland, Magnus B.O. Andersson, Lars G.

Blomberg, Submitted to J. Chromatography A.

Table of contents

1 Introduction...1

1.1 Liquid Separation Techniques ...1

1.1.1 Liquid Chromatography... 1

1.1.2 Capillary LC ... 3

1.1.3 Capillary Electrophoresis ... 3

1.1.4 Capillary Electro Chromatography... 5

1.2 Combination of Pressurized Flow and Electric Fields...5

1.2.1 In-line Application of Electric Fields... 6

1.3 Conductivity in Capillary Separation Techniques...7

2 Aim ...9

3 Results and Discussion ...10

3.1 Parameters Influencing the Conductivity...11

3.1.1 Electric Field Strength... 12

3.1.2 Flow Velocity... 13

3.1.3 Electrophoretic Mobility of the Buffer Ions... 14

3.1.4 Electrode Set-up... 15

3.1.5 Ionic Strength ... 18

3.2 Theoretic Model ...18

3.2.1 Flow Velocity and Electrophoretic Velocity of the Mobile Phase Ions ... 19

3.2.2 Electrolysis of Water ... 20

3.2.3 Increase in Ionic Strength... 21

4 Future Studies...25

5 Acknowledgements ...26

6 References ...28

1 Introduction

1.1 Liquid Separation Techniques

Separations are a cornerstone in modern chemical analysis. The widespread use of separation techniques is due to the complexity of many samples. Samples often contain a large number of substances and the substances of interest often must be separated from interfering substances before detection and quantification. Furthermore, separation is frequently needed when several chemically similar substances need to be quantified individually. The two major separation principles are chromatography and electrophoresis which will be discussed in the following sections.

1.1.1 Liquid Chromatography

Liquid chromatography (LC) commonly referred to as high performance liquid chromatography (HPLC) is one of the most widely spread instrumental analysis techniques. HPLC is frequently used for separation of nonvolatile organic substances in an aqueous or partly aqueous sample solution. Chromatography was first presented by the Russian botanist Mikhail Semenovich Tswett (1872- 1919) in a lecture at the W arsaw Society of Natural Sciences on the 8 March 1903 [1]. Tswett observed yellow and green rings or bands forming when rinsing plant extract through a glass cylinder packed with calcium carbonate [2- 4]. Tswett could thereby show that the chlorophyll of plants actually consists of several different chlorophylls. The word chromatography is composed of the words chroma (color) and graphein (to write) from Greek. Here chromatography can be interpreted as “color writing” referring to the colored bands observed by Tswett. Chromatography can however also be interpreted as

“Tswett’s writing” because the Russian word for color is “Tswett” [1].

The instrumentation and technique of HPLC has changed dramatically since Tswett’s days and the theory is also now well known and documented [5]. The basic set-up of an HPLC instrument is shown in Figure 1.

Figure 1. The basic set-up of an HPLC instrument. 1. HPLC pump (pumps), 2 . Injector v alv e w ith sample loop, 3 . Column, 4 . D etector.

The HPLC pumps produce a flow of mobile phase through the system. The separation is preceded by an introduction of the sample to the column through the inj ector valve. After the separation in the column the analytes are quantitatively (and/or qualitatively) determined by the detector. The fundamental principle of chromatography is the different distribution of analytes between a stationary and a mobile phase. An analyte with high affinity for the stationary phase will thus be retained longer in the column than an analyte with low affinity. Substances with different affinity to the stationary phase will therefore be detected at different times, respective to the sample inj ection.

The dominating mode of HPLC is the use of reversed phase columns. Such columns have packing materials with hydrophobic surfaces and the mobile phase compared to these stationary phases is relatively hydrophilic. The analytes will thereby be separated according to their hydrophobicity. Hydrophobic analytes will be retained longer in the column and they will thus be eluted lastly.

The HPLC columns are in general packed with more or less spherical particles having a diameter of 3-10 µm depending on the application. The most commonly used packing material is based on porous silica with a surface of C

groups and the mobile phases are usually mixtures of water and some organic

solvent such as acetonitrile or methanol. In addition, the mobile phase

commonly contains a buffer in order to obtain a stable pH.

1.1.2 Capillary LC

The difference between capillary LC and conventional HPLC is the smaller inner diameter, typically 100-500 µm, of the capillary columns compared to 4.6 mm in conventional HPLC columns. The main advantages from the use of capillary columns are higher concentration sensitivity and lower sample and mobile phase consumption [6, 7]. Capillary LC also offers, at least in theory, better compatibility with mass spectrometry (MS), because less mobile phase needs to be removed in the LC-MS interface. Further, capillary columns are, due to their low thermal mass, more suitable for temperature control and programming [8] and the higher electrical resistance of capillary columns makes them more suitable for application of an electric field.

The instrumentation used in capillary LC is generally the same as in conventional HPLC but each component such as pumps, mixers, tubing, unions, valves and detector flow cells, needs to be adapted to the lower flow rates.

1.1.3 Capillary Electrophoresis

Electrophoretic separation was introduced 1937 by Arne Tiselius (1902-1971) who showed that ions could be separated in an electric field due to their different electrophoretic mobilities [9, 10]. Tiselius student Stellan Hjertén improved the electrophoresis method by performing the separation in a capillary with inner diameter of 2 mm [11, 12]. Hjertén thereby eliminated some of the problems with heating experienced by Tiselius in the early methods.

Electrophoresis was developed further during the 1980s through the work by Jorgenson and Lukacs [13] using narrower capillaries. This technique, now named capillary electrophoresis (CE), is today an established separation technique.

The basic set-up of a CE instrument is shown in Figure 2.

Figure 2. The basic set-up of a CE instrument. A fused silica capillary (1) placed with each end in buffer vials (2). An electric field is created in the capillary by apply ing hig h voltag e to the Pt electrodes (3) from the hig h voltag e supply (4). The detection is usually performed on-line using U V detection (5 ).

In CE a fused silica capillary is placed with each end in buffer vials. A high electric voltage is applied over the vials creating a high electric field in the capillary in which ions can be separated. The sample is injected by switching the inlet vial to a vial containing the sample. Through the use of pressure or electric field, a sample plug is then introduced into the capillary. The analytes are separated within the capillary, and are usually detected using on-line UV detection through a window created by removing the capillary outer coating.

The fundamental separation mechanism in CE is, as mentioned above, that the different ions in a solution have different electrophoretic velocity in an electric field. The electrophoretic velocity (v

) is a function of the electric field strength (E) and the specific electrophoretic mobility (µ) according to:

Eq. 1

An important feature in CE is the electroosmotic flow (EOF). When a high

electric field is applied over the fused silica capillary the liquid spontaneously

moves towards the negative electrode. No pressure driven flow system is

needed in CE and the flow profile of the electroosmotic flow is more plug-like

than when a pressurized flow is applied. The plug-like flow profile of CE gives

less band broadening compared to HPLC and is one of the major advantages of the technique.

In the original mode of CE, capillary zone electrophoresis (CZE), the difference in electrophoretic velocity separates the charged analytes. In other modes with different additives to the running buffer it is also possible to separate uncharged analytes according to their hydrophobicity which then makes CE a complement to HPLC.

1.1.4 Capillary Electro Chromatography

There are a number of techniques that combine elements from chromatographic and electrophoretic separation. The most successful in this area is capillary electro chromatography (CEC) that combines the separation technique of LC with the more uniform flow profile of CE [14]. In addition, higher efficiencies can also be obtained in CEC with the use of smaller particle sizes which can be employed because of the lower back pressure. The common instrumental set-up for CEC is with a packed capillary column in a CE instrument.

1.2 Combination of Pressurized Flow and Electric Fields

Despite the advantages of an EOF driven flow a pressurized flow is sometimes used together with an electric field [15-48]. The main driving force behind combining an applied electric field with a pressurized flow is that such a system does not depend upon the generation of an EOF for transporting the mobile phase through the column. The separation can then be optimized by the mobile phase flow rate and electrophoretic migration separately [15-18, 23-25, 27, 30, 35, 45]. The magnitude or polarity of the electric field can thus be altered to achieve a good separation without changing the transport of the mobile phase.

The relative independence of the EOF also means that any pH or buffer composition can be used, that otherwise would not have generated an EOF [19, 33]. The application of pressure in combination with an electric field has also frequently been used in order to suppress bubble formation [15, 24, 29, 32, 33, 46], to give a more stable and reliable flow [23, 27, 33] or to obtain higher flow

rates and reduced separation times [19]. The use of HPLC pumps also simplifies the use of solvent gradients in CEC [27, 33, 34, 37-40, 43, 46, 47].

The application of an external pressure has also been showed to aid coupling to detection techniques such as MS [24, 46] and NMR [36, 42].

1.2.1 In-line Application of Electric Fields

There are two fundamental set-ups of a capillary separation system that can be utilized when a pressurized flow is combined with an electric field. These two set-ups are: CE-based systems with electric field and pressure applied to reservoirs and LC-based systems with a pressurized flow and an electric field applied in-line. The in-line approach has been used frequently [15-19, 21-30, 32- 40, 42, 43, 45-47] and when CE-based systems have been used the applied pressure has often been quite limited (~10 bar) [31 (10 bar), 41 (8 bar), 44 (12 bar),48 (12 bar)].

When high pressures are needed the electrodes are frequently placed in-line.

Here, stainless steel unions or injector valves can be used as electrodes [17-19, 21, 24, 25-29, 32, 35-40, 42, 43, 45, 46]. It is however recommended to use electrodes made from platinum rather than steel due to possible problems with corrosion, Paper III. One possible Pt-electrode set-up is to employ a PEEK-tee fitting as described in Figure 3.

Figure 3. Set-up utilizing a Pt-electrode in-line. 1. PEEK-tee fitting, 2. F low inlet, 3. F low outlet, 4. Pt wire inside a PEEK tube.

As an alternative to in-line application, highly pressurized reservoirs have been used by Behnke and Bayer [20 (200 bar), 22 (150 bar)] and by us in Papers I and III. This will be further discussed in Section 3.1.4.

1.3 Conductivity in Capillary Separation Techniques

High electric field strengths are often desirable in electric field assisted separation techniques, for example when an electric field is used to enhance the separation of ions in a chromatographic separation. As the ions are only affected by the electric field when they are in the mobile phase, the effective time for electrophoretic separation is very short. The only possibility to increase the contribution of the electrophoretic separation is to use a high electric field, but that will come at a cost of high currents (causing band broadening, noisy baseline, etc). In an in-line system, a high electric field will also cause an increased conductivity in the capillary, further increasing the current. (This will be shown in Section 3.)

For better understanding of the work some important relationships and theories regarding conductivity will be discussed.

The current (I) through a conductor at a specific voltage (U) is determined by Ohm´s law:

U =RI Eq. 2

where R is the electric resistance. The current in a solution consists of migrating ions, rather than electrons as in a metal conductor. The conductivity (κ^{) is the} ability of a specific ion solution to transport current. The electric resistance depends, apart from the conductivity, only on the dimensions of conducting medium:

A R L

=κ Eq. 3

where L is the length and A is the cross section area of the medium.

The conductivity depends on the charge of the ions (z), their el ectrophoretic mobil ity (µ) and their concentration (c) according to:

i ic z

F µ

κ ⁼

∑ Eq. 4

where F is the Faraday constant [ 49] . Highl y concentrated sol utions containing highl y charged, high mobil ity ions are therefore the best conductors.

W henever a current passes through an el ectrol yte, as in vol tage assisted separation systems, heat wil l be generated due to the resistance of the el ectrol yte. This heating is commonl y referred to as Joul e heating. W ith l arge appl ied effects (P) the temperature inside the capil l ary wil l increase due to the Joul e heating. The increase in temperature depends, besides on the appl ied effect, on the abil ity of the system to transfer excess heat to the surroundings through the capil l ary wal l .

It has been shown that the conductivity depends l inearl y on the temperature of

the sol ution [ 50] . The change in conductivity, caused by a change in

temperature, is approximatel y 2% per °C. For this reason conductivity is

frequentl y used to estimate the mean temperature increase in capil l aries caused

by Joul e heating [ 51-54] . The rel ation between the conductivity and the

temperature is, however, rel ativel y compl ex. An increase in temperature due to

Joul e heating causes an increase in conductivity which in turn increases the

current which resul ts in more Joul e heating and so on. The temperature and the

conductivity thus continue to increase until a steady state is reached due to the

heat transported through the capil l ary wal l . This is sometimes referred to as the

auto-thermal effect [ 51] . In this thesis non-Joul e heating rel ated conductivity

changes within in-l ine capil l ary separation systems wil l be discussed. A

contribution of Joul e heating to the conductivity changes is however

unavoidabl e.

2 Aim

High electric fields are often desired in separation systems utilizing an electric field. The maximum applicable electric field is limited by problems resulting from high electric currents. Joule heating and bubble formation are examples of such problems. The aims have been to describe the theory behind the conductivity changes observed in capillary separation systems when applying an electric field in-line, and if possible give suggestions how to limit these conductivity changes.

3 Results and Discussion

A typical Ohm’ s law plot, with current as a function of voltage, in an in-line system using a packed capillary column and an acetate buffer is presented in Figure 4. Two unique features generally observed in in-line systems are clearly illustrated in Figure 4. First, upon increasing voltages the current increases dramatically at high voltages. Further, this increase in current is highly dependent on the direction of the current. Large currents are a maj or limiting factor to any separation system using in-line applied electric field. It is therefore desirable, if possible, to minimize the conductivity changes which results in the observed high currents.

0 5 0 1 00 1 5 0 2 00 2 5 0

0 2 4 6 8 1 0 1 2

Voltage (kV)

Current (µA)

P o s itiv e v o lta g e N e g a tiv e v o lta g e

Figure 4. Current at given voltage depending on the direction of the current through a pack ed capillary colum n. Colum n: 1 5 cm , 1 5 0 µ m I.D . capillary pack ed w ith 3 .5 µ m Y M C B as ic. M ob ile phas e: 7 5 m M acetate, pH 4 w ith 1 % A CN . (F ig. 2 , P aper I)

Joule heating is known to cause conductivity changes in separation systems

utilizing an electric field, as discussed in Section 1.3. Conductivity changes

caused by Joule heating should, however, be of the same magnitude when the

polarity is shifted. The conductivity changes within in-line systems are

therefore, to a maj or part, believed to be of a different origin. The unavoidable

influence of the Joule heating on the conductivity was examined in Paper IV.

The relatively small influence of Joule heating on the conductivity is evident from the results presented in Figure 5. The large non-Joule heating related conductivity changes only observed in in-line systems have to the best of our knowledge never been studied until now.

0 0.02 0.04 0.06 0.08 0.1 0.1 2

0 0.1 0.2 0.3 0.4 0.5 0.6

P/L (W/m)

Conductivity (1

/ΩΩΩΩ

m)

N e g a tiv e p o la rity w ith J o u le h e a tin g N e g a tiv e p o la rity w ith o u t J o u le h e a tin g P o s itiv e p o la rity w ith J o u le h e a tin g P o s itiv e p o la rity w ith o u t J o u le h e a tin g

Figure 5. Conductivity as a function of effect per meter with voltag e applied in-line. O pen mark ers sh ow th e orig inal data, wh ile filled mark ers sh ow th e data wh ere th e contrib ution of J oule h eating to conductivity ch ang e h as b een sub tracted. F or furth er details, see P aper IV . (F ig . 2 b , P aper IV )

Determination of the influence of various parameters on the conductivity can potentially give information about the processes behind the conductivity changes. Such information can be used when choosing parameters with the aim to limit the current.

3.1 Parameters Influencing the Conductivity

The conductivity in in-line systems is influenced by several parameters, five of which will be discussed here: The electric field strength, the flow velocity, the

electrophoretic mobility of the mobile phase ions, the set up of the inlet electrode and the ionic strength.

3.1.1 Electric Field Strength

It is evident from the Ohm’ s law plot, Figure 4, that the conductivity increases with the applied voltage. The conductivity, however, depends upon the applied electric field rather than the voltage as shown in Paper V. It was found that the dramatic raise in conductivity often observed in in-line systems occurs at a specific applied electric field rather than of a specific applied voltage. At which electric field strength this raise in conductivity occurs depends on various parameters such as the polarity of the electric field, the flow velocity and the mobility of the mobile phase ions with electrophoretic velocity against the flow velocity (from hereon referred to as the decelerated ions). This will be commented further in the following sections.

A coefficient plot of the various factors influence on the conductivity reprinted

from Paper V further illustrates the relatively large influence of the electric field

on the conductivity, Figure 6a.

Figure 6. Response plot of the significant factors affecting a : the cond u ctiv ity and b : the w eighted cond u ctiv ity . A ll ex perim ents d one on a 5 1 m m , 5 0 µ m I.D . capillary w ith 5 µ L /m in (4 2 .4 m m /s) of 5 m M N aC l. In the w eighted cond u ctiv ity the k now n effect of the ion strength and the electrophoretic m ob ility of E q u ation 4 are su b tracted . T his is d iscu ssed fu rther in S ection 3 .1 .3 (F ig. 4 , P aper V )

Besides the large influence of the electric field (E) and its quadratic term (E*E) the results showed significant covariance with the ionic strength (Conc) and the electrophoretic mobility of the decelerated ion (dec).

3.1.2 Flow Velocity

A unique feature of an in-line system is that the conductivity is highly dependent on the flow velocity through the column. The conductivity as a result of increasing flow velocities at linearly programmed elevation of the electric field strength is presented in Figure 7. At a flow velocity range about 2 mm/s, which often is used in LC, the observed conductivity was about ten times larger than without pressurized flow. The conductivity reaches a maximum at a flow velocity of 8 mm/s. In these experiments the decelerated ions have an electrophoretic velocity that is in the same range as the flow velocity. This is not a coincidence, and will be further discussed in Section 3.2.1.

At larger flow velocities the conductivity gradually decreased approaching a common value independent of the electric field strength.

0.00 0.02 0.04 0.06 0.08 0.1 0 0.1 2 0.1 4 0.1 6 0.1 8

0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0

Flow velocity (mm/s)

Conductivity (1

/ΩΩΩΩ

m)

0.3 9 0.5 9 0.7 8 0.9 8 1 .2 1 .4 E (k V /c m )

Figure 7. Conductivity as a function of flow velocity at different electric fields. 5 m M N aCl in a 5 .1 cm , 5 0 µ m I.D . coated cap illary. T h e electric current was recorded at each flow velocity as th e voltag e was increased from 0 -7 k V at 2 k V /m in. (F ig . 3 from P ap er V )

It is evident from the result in Figure 7 that relatively low conductivities can be obtained through the use of high flow velocities. The possibility of decreasing the conductivity by means of using large flow velocities is however often not practical. Chromatographic separation is commonly best performed with flow velocities below 10 mm/s. W ith the use of decelerated buffer ions with low electrophoretic mobilities the maximum conductivity can however be obtained at lower flow velocities than in Figure 7 which will be discussed further in the next section and in Section 3.2.1.

3.1.3 Electrophoretic Mobility of the Buffer Ions

Similar to the flow velocity the influence of the mobilities of the mobile phase

ions on the conductivity in in-line systems differ significantly from the influence

expected in CE-like systems. From the coefficient plot, Figure 6, it is clear that

the mobilities of the decelerated buffer ions (dec) have a large influence on the

conductivity in in-line systems. The velocity of an ion in an electric field (the

electrophoretic velocity) is a product of the electric field strength and the

electrophoretic mobility. The relative importance of the electric field strength

and the mobility of the decelerated ions, as well as the product term (dec*E), indicates that the electrophoretic velocity of the decelerated ions is of great importance for the conductivity in in-line systems.

The mobility of the accelerated ions (acc) on the other hand was shown to have a minimal influence on the conductivity. In Paper V the known influence of the mobilities and the ionic strength on the conductivity, according to Eq 4, was subtracted from the conductivity data, Figure 6b. Here no significant influence of the mobility of the accelerated ions was observed suggesting that the velocity of the accelerated ions have no or minimal importance for the conductivity changes in in-line systems.

3.1.4 Electrode Set-up

The conductivity changes are, as mentioned above, never observed in CE-like set-ups. It was originally observed in Paper I that when the voltage at the inlet was applied to a pressurized reservoir rather than in-line, the conductivity was lower and larger electric field strengths could be applied. The use of pressurized reservoirs was more thoroughly studied in Paper III through the use of linearly programmed increasing voltages. Two different electrode set-ups were used in these experiments. A stainless steel reservoir or a PEEK tee fitting with a Pt wire, identical to Figure 1, was used either as inlet or outlet electrode, Figure 8.

The evident conclusion was that the use of the reservoir as inlet electrode significantly limited the increase in conductivity whereas large conductivity changes still were observed when the reservoir was used as outlet electrode, Figure 9.

Figure 8. Set-up for using a pressurized reservoir as one of the electrodes. 1. Connection line from the pump 2. Stainless steel reservoir. (empty pre-column) 3. L ow O .D. capillary 4 . PE E K tee with 5. Pt electrode. (Fig. 1b , Paper III)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Electric field strength (kV/cm)

Conductivity (1

/ΩΩΩΩ

m)

R e s e rv o ir a t th e in le t R e s e rv o ir a t th e o u tle t

Figure 9 . Conductivity at increasing electric field strength using a reservoir as inlet or outlet electrode. L inearly increasing voltage from 0-5 kV at 2 kV/min. Flow rate: 10 µL /min, solution: 10mM NaCl, capillary length 5.5 cm and capillary I.D. 50 µm. (Data from Fig. 3, Paper III)

Further, it was shown that the conductivity changes could be limited similarly

to the use of a reservoir when the flow was splitted at the inlet electrode using a

PEEK-cross fitting with a Pt wire as electrode, Figure 10. The Pt electrode was

mounted on the opposite side as the test capillary with the split-flow passing

between them. It was found that only a small split-flow (1: 0.5) was needed to

limit the increase in conductivity observed when no split-flow (splitless) was used, Figure 11. Further it was found, using the condition of Figure 11, that increasing the split-ratio over 1:4 had no further effect on the conductivity.

Figure 10. Set-up for flow splitting at the inlet electrode. A PEEK cross (1), originally with 7 50 µm I.D., drilled to 1/16 ” in one direction. Positions: 2. Inlet, 3. Pt electrode, 4. split outlet, 5. 5.5 cm, 50 µm I.D. coated capillary, 6 . Close-up of the center with the Pt electrode (top) and coated capillary (bottom). (Fig. 1a, Paper III)

0 0.05

0.1 0.1 5 0.2 0.2 5 0.3 0.3 5 0.4

0 0.2 0.4 0.6 0.8 1 1 .2 1 .4 1 .6 1 .8 2

Electric field strength (kV/cm)

Conductivity (1

/ΩΩΩΩ

m)

Split 1:4-1:44 Split 1:0 .5

Splitle s s

Figure 11. Conductivity at increasing electric field strength at different sp lit-ratios. L inearly increasing voltage from 0 -7 k V at 2 k V /m in. O ther conditions as in F ig. 9 . T he overlap p ing p lots “S p lit 1 :4 to 1 :4 4 ” contain 1 2 lines at 4 levels of sp lit-ratio (1 :4 , 1 :1 1 , 1 :2 4 , 1 :4 4 ).

(D ata from F ig. 4 , P ap er III)

The l arge i nfl uence of the i nl et el ectrode corresponds wel l to the l arge i nfl uence of the decel erated mobi l e phase i ons, whi ch have an el ectrophoreti c vel oci ty toward the i nl et el ectrode. Si mi l arl y, the l ow i nfl uence of the outl et el ectrode corresponds wel l to the l ow i nfl uence of the accel erated mobi l e phase i ons.

3.1.5 Ionic Strength

The i nfl uence of the i oni c strength on the conducti vi ty of a sol uti on i s wel l known and i s descri bed by Eq. 4. The response pl ot of Fi gure 6b descri bes additional effects of the i oni c strength on the conducti vi ty i n an i n-l i ne system.

Here the i nfl uence of the i oni c strength, as wel l as the el ectrophoreti c vel oci ti es, on the conducti vi ty accordi ng to Eq. 4 has been subtracted. The maj or i nfl uence of the i oni c strength on the conducti vi ty i n i n-l i ne systems cannot be expl ai ned by Eq. 4 whi ch i s evi dent when compari ng Fi gures 6a and 6b. The l arger i nfl uence of the i oni c strength i s bel i eved, as wi l l be di scussed further i n Secti on 3. 2. 1, to be a resul t of a l arger number of mobi l e phase i ons enteri ng the el ectri c fi el d wi th the fl ow.

The resul ts i n Paper V i l l ustrate the l arge potenti al of decreasi ng the conducti vi ty through the use of l ow i oni c strength mobi l e phases. The use of l ow i oni c strength mobi l e phases i s however often i mpracti cal because of the l ow buffer capaci ty. As wi l l be di scussed i n Secti on 3. 2. 2 hi gh buffer capaci ty i s even more i mportant i n i n-l i ne systems as compared to i n CE set-ups.

3.2 Theoretic M odel

W e suggest that an i ncrease i n i oni c strength i s the overal l expl anati on for the

conducti vi ty changes and the pH changes observed wi thi n i n-l i ne systems. Thi s

i ncrease i n i oni c strength i s a resul t of the bul k fl ow, the el ectrophoreti c

vel oci ty of the decel erated buffer i ons and the el ectrol ysi s of water. The

processes behi nd the i ncrease i n i oni c strength wi l l be di scussed i n the

fol l owi ng secti ons.

3.2.1 Flow Velocity and Electrophoretic Velocity of the M obile Phase Ions

The change in velocity experienced by the mobile phase ions when entering the electric field can cause a change in concentration of these ions (provided that the column can stay electroneutral, see Section 3.2.2). The increased velocity of the accelerated ions can result in a decreased concentration because of the shorter time spent in the electric field. The corresponding increase in concentration of the decelerated ions can, at high electric field strengths, become much larger. This is, in theory, easiest realized when considering decelerated ions with equal or larger electrophoretic velocity as the flow velocity. W hen these ions enter the electric field they would loose their down stream velocity. As new ions constantly enter the electric field the concentration would, in theory, become infinitely large. As a result the potential increase in mobile phase ion concentration would be largest when the electrophoretic velocity of the decelerated ion is in the same range or larger than the flow velocity. The electric field strength needed to obtain this is rather low when using, for chromatography, normal flow rates and mobile phase ions. For example, in a system with an average flow velocity of 2 mm/s and with sodium ions as decelerated ions only 0.39 kV/cm is needed for the sodium ions to obtain the same velocity as the flow, although with opposite direction.

If a reservoir is used at the inlet electrode, as in CE, an ion with velocity against the flow can pass the inlet of the capillary and be diluted in the relatively large volume of the reservoir. If the inlet electrode instead is placed in-line, an ion with velocity against the flow can not pass the inlet electrode. If the ion should pass the inlet electrode it would loose its electrophoretic velocity and the mobile phase flow would transport it back into the electric field.

Figure 12 illustrates the velocity of the mobile phase ions when affected by the pressure driven bulk flow and their electrophoretic mobility. Flow profile A illustrates the velocity of the ions outside the electric field. Flow profiles B and C in Figure 12 illustrate the sum velocity of the accelerated and decelerated ions in the electric field respectively. It can be observed that the decelerated ions close to the wall can obtain a net flow velocity upstream, even if the average flow velocity is larger than the electrophoretic velocity of the decelerated ions.

Therefore, the concentration of the decelerated mobile phase ions can increase under lower electric field strengths than if the flow had been perfectly plug like.

C E le c tric fie ld B

A

Figure 12. The velocities of ions in a capillary. Outside the electric field (A ) b oth positive and neg ative ions follow the velocity of the solution. Inside the electric field (shaded), the velocity of the ions w ith either positive or neg ative charg e is increased (B ), w hile the velocity of the oppositely charg ed ions is decreased (C ). (F ig ure 2 , P aper II)

3.2.2 Electrolysis of Water

The results in Paper V indicate that large pH changes could be expected within in-line systems. This is because OH

and H

O

ions (depending on polarity) formed at the inlet electrode will be diluted in the mobile phase within the small volume of the capillary column rather than in reservoirs as in CE based systems. With the 10 mM TRIS buffer used in Paper V significant pH changes were observed at only ~10 µA. A linear relation between current and the apparent addition of OH

or H

O

ions was suggested from the results in Paper V. Without sufficient buffer capacity, large pH changes could therefore be expected. Increasing the buffer concentration would however result in even larger currents which make the pH changes in in-line systems harder to prevent.

Further, the electrolysis of water, forming OH

and H

O

ions, is essential for

the increase in mobile phase ions discussed in Section 3.2.1. This increase in

equally charged mobile phase ions would have been impossible if not

oppositely charged OH

or H

O

ions would have been formed at the

electrodes from neutral water. The increase in mobile phase ions together with

the, at the inlet formed OH

or H

O

ions, will increase the ionic strength in

the column.

3.2.3 Increase in Ionic Strength

There is not enough data in this study to provide a complete description of how the ions are distributed in and after the electric field. There are, however, enough data to predict the overall changes in concentration, of decelerated and accelerated mobile phase ions as well as OH^- or H3O⁺ ions, within and after the electric field, Figure 13. In Figure 13, in-line applied voltage and high electric field strength (> 0.5 kV/cm, higher with low mobility ions) have been assumed.

The negative electrode is placed as inlet electrode. The contours of the concentration curves in and after the electric field are tentative and could be a matter of further studies, see Section 4. The large changes in concentrations of decelerated ions and OH^- ions around the electrodes are, however, more certain assumptions than the contours in and after the electric field.

Figure 13. Schematic illustration of the changes in ion concentration, within and after an electric field, in an in-line system. The contours for the concentrations curves are tentative, see Section 4 (Future studies). Further, the concentration curves will probably depend upon the radial position of the column.

Some general conclusions (which can be explained by Figure 13) can be drawn regarding the changes in ion concentration within in-line systems:

• As discussed above, the change in concentration of the decelerated ions will be much larger than the change in concentration of accelerated ions.

This difference will increase with larger electric fields.

• In the electric field there will be no change in ionic strength due to the accelerated ions. Any decrease in concentration of these ions will be canceled by a corresponding increase in OH^- or H3O⁺ ions. The change in concentration as a result of the decelerated ions, on the other hand, will together with OH^- or H3O⁺ ions increase the ionic strength and thereby the conductivity, Figure 13.

• The concentration of the OH^- or H3O⁺ ions will at each given position be the sum of the change in concentration of the decelerated and the accelerated ions, Figure 13. Applying the polarity as in Figure 13 the change in charge due to the increase in positive ions and the decrease in negative ions in the electric field will be neutralized by an equal increase in OH^- ions formed at the inlet electrode.

• An increase in number of mobile phase ions inside the electric field will result in an equally large decrease in number of those ions after the electric field and vice versa. Similarly, with the polarity as in Figure 13 an increase in the number of OH^- ions inside the field will result in an equally large increase in number of H3O⁺ ions after the electric field.

If a buffering mobile phase is used, some of the OH^- or H3O⁺ ions will transfer their charge to the buffer and form water. If a charged buffer ion is created in the reaction the only change will be on the average mobility of the ions and not the ionic strength of the solution. If instead a buffer molecule is created from a buffer ion in the reactions it is possible that this could have a limiting effect on the increased ionic strength and the conductivity. We have, however, never observed any significant effect of using such buffers. The buffer effect is too small to neutralize the large number of OH^- or H3O⁺ ions created at high currents. Further, it is not practical to increase the buffer concentration in order to decrease the conductivity.

As discussed above, the data of this study do not provide enough information about the distribution of the concentration changes inside and after the electric

field. It is however likely that concentration changes both in longitudinal and radial directions could be observed. The latter because of the parabolic flow profile.

In Paper III the decrease in concentration of the, in the electric field,

decelerated ions after the electric field was reported. In an unpublished

experiment the decrease in absorption, of a decelerated ion (with UV-

absorption) after the electric field, was monitored as a function of increasing

electric fields. At moderately high electric fields the absorption changes

followed a linear relationship as a function of electric field strength. At high

electric field strengths the absorption did not decrease further upon increasing

electric fields. A plug of water was injected and the absorption of the water plug

was in the same range as the absorption of the test solution with the high

electric fields. These results illustrate that the concentration of a decelerated ion

can after the outlet electrode decrease to virtually zero. If, for example, the

concentration of the decelerated ion inside the electric field increases to more

than two times the original concentration there can not be an equally large

decrease in concentration after the outlet electrode. The concentration can then

obviously only decrease to zero, but then in a larger volume than inside the

electric field. It will, however, always be the same number of ions increasing inside

as the corresponding decrease after the outlet electrode.

4 Future Studies

Experimental data of how the change in concentration of mobile phase ions as well as OH^- and H3O⁺ ions distributes within and after the column is desirable.

A set-up with a microscope and a CCD camera and the use of cresol red/TRIS solution as in Paper IV could give a picture of the pH changes in and after the electric field both radially and longitudinally. UV-absorbing or fluorescent mobile phase ions could be used as probes for changes in mobile phase ion concentration.

5 Acknowledgements

Jag skulle vilja rikta ett stort tack till följande personer som på ett eller annat sätt hjälpt mig under min tid som doktorand:

Lars Blomberg för att du alltid finns till hands för korta eller långa samtal.

Även om det både gått bättre och sämre i projektet har det alltid varit mycket god stämning mellan oss.

Magnus (Bror Olof) Andersson för alla praktiska trick om µ-LC och naturligtvis för alla timmar i telefon. Det har i ordets rätta bemärkelse varit ett nöje att ha dig som biträdande handledare.

Magnus Dahl för ditt fantastiska arbete med finita element simuleringarna.

Våra diskussioner har varit av största vikt för det här projektet.

Inger Lill Skuland för ditt envisa arbete inom detta projekt trots att det mesta gick emot oss.

Nicola Marlin for your large help during the later part of this project.

Lisa Elofsson för ditt utmärkta examensarbete inom projektet.

Product Analysis, Analytical Development, PAR&D, AstraZeneca, Södertälj e för ekonomiskt stöd till projektet.

Reidar Lyng för att du, för ett antal år sedan, fick mig att tänka på effekten av accelererade och bromsade joner som sedan blev fundamentet till hela projektet.

Lars Renman för att du har visat så stort förtroende för mig i samband med

kurserna i analytisk kemi. Forskningsprojektet och denna avhandling är bara en

del av det jag lärt mig under doktorandtiden. Det övriga har jag till stor del dig

att tacka för.

Mina närmaste doktorandkollegor Jeanette Olsson, Marcus Öhman, Maria Bohlin, Zeki Altun, Mårten Jönsson, Jan Bohlin och Christina Bohlin.

Utan er hade jag varken börjat eller fortsatt arbeta som doktorand så länge. Det har oftast varit ett rent nöje att gå till jobbet tack vare er.

Alla fantastiska kollegor vid Karlstads Universitet för all hjälp och roliga samtal.

Min familj för att ni alltid har stöttat och hjälp mig på alla tänkbara sätt.

Maria och Stina, för att ni påminner mig om vad som verkligen är viktigt för mig.

6 References

[1] L.S. Ettre, LC•GC North America 21 (2003) 458.

[ 2] M.S. Tswett, Trudy Varshavskogo Obshchestva Estestvoi spytatatel ei Otdel eni e Bi ol ogi i 14 (1905) 20.

[ 3] G. Hesse, H. W ei l , Michael Tswett’s First Paper on Chromatography, M . W oel m, Eschwege, 1954.

[ 4] V.G. Berezki n, Chromatographi c Adsorpti on Anal ysi s: Sel ected W orks of M.S. Tswett, El l i s Harwood, NewYork, 1990.

[ 5] L.R Snyder, J.J. Ki rkl and, J.L. Gl aj ch, Practical HPLC Method Development, Second edi ti on, John W i l ey & Sons Inc., New York, 1997.

[ 6] F.J. Yang i n P.A. Mi l l ner (Ed.), High Resolution Chromatography, Oxford Uni versi ty Press Inc., New York, 1999, pp. 23-46.

[ 7] F.J. Yang (Ed.), Microbore Column Chromatography, Marcel Dekker Inc., New York, 1989.

[ 8] T. Grei brokk, T. Andersen, J. Chromatogr. A 1000 (2003) 743.

[ 9] A. Ti sel i us, The moving boundary method of studying the electrophoresis of proteins, Thesis (Uppsal a Uni versi ty), Al mqvi st & W i ksel l s Boktryckeri AB, Uppsal a, 1937.

[ 10] A. Ti sel i us, Trans. Faraday Soc. 33 (1937) 524.

[ 11] S. Hj ertén, Free Zone Electrophoresis, Thesis (Uppsal a Uni versi ty), Al mqvi st

& W i ksel l s Boktryckeri AB, Uppsal a, 1967.

[ 12] S. Hj ertén, Chromatogr. Rev. 9 (1967) 122.

[ 13] J.W . Jorgenson, K.L. Lukacs, Anal Chem. 21 (1981) 498.

[ 14] S. Eel ti nk, G.P. Rozi ng, W .T. Kok, Electrophoresis, 24 (2003) 3935.

[ 15] T. Tsuda, Anal. Chem. 59 (1987) 521.

[ 16] T. Tsuda, Anal. Chem. 60 (1988) 1677.

[ 17] E.R. Verhei j , U.R. Tj aden, W .M .A. Ni essen, J. van der Greef, J.

Chromatogr. A 554 (1991) 339.

[ 18] T. Tsuda, LC•GC Int. 5 (1992) 26.

[ 19] M. Hugener, A.P. Ti nke, W .M .A. Ni essen, U.R. Tj aden, J. van der Greef, J. Chromatogr. A 647 (1993) 375.

[ 20] B. Behnke, E. Bayer, J. Chromatogr. A 680 (1994) 93.

[ 21] S. Ki tagawa, T. Tsuda, J. Microcol. Sep. 6 (1994) 91.

[ 22] B. Behnke, E. Bayer, J. Chromatogr. A 716 (1995) 207.

[ 23] T. Ei mer, K.K. Unger, J. van der Greef, Trends Anal. Chem. 15 (1996)

463.

[24] J.T. Wu, P. Huang, M.X. Li, D.M. Lubman, Anal. Chem. 69 (1997) 2908.

[25] S. Kitagawa, A. Tsuji, H. Watanabe, M. Nakashima, T. Tsuda, J. Microcol.

Sep. 9 (1997) 347.

[26] Y. Zhang, W. Shi, L. Zhang, H. Zou, J. Chromatogr. A 802 (1998) 59.

[27] B. Behnke, J.W. Metzger, Electrophoresis 20 (1999) 80.

[28] S. Kitagawa, H. Watanabe, T. Tsuda, Electrophoresis 20 (1999) 9.

[29] P. Huang, X. Jin, Y. Chen, J.R. Srinivasan, D.M. Lubman, Anal. Chem. 71 (1999) 1786.

[30] A. Apffel, H. Yin, W.S. Hancock, D. McManigill, J. Frenz, S.L. Wu, J.

Chromatogr. A 832 (1999) 149.

[31] E.F. Hilder, C.W. Klampfl, P.R. Haddad, J. Chromatogr. A 890 (2000) 337.

[32] S. Nagaraj, H.T. Karnes, Biomed. Chromatogr. 14 (2000) 234.

[33] Q.H. Ru, J. Yao, G.A. Luo, Y.X. Zhang, C. Yan, J. Chromatogr. A 894 (2000) 337.

[34] T. Adam, K.K. Unger, J. Chromatogr. A 894 (2000) 241.

[35] M.B.O. Andersson, L.G. Blomberg, J. Sep. Sci. 24 (2001) 304.

[36] P. Gfrörer, L.H. Tseng, E. Rapp, K. Albert, E. Bayer, Anal. Chem. 73 (2001) 3234.

[37] M. Stahl, A. Jakob, A. Von Brocke, G. Nicholson, E. Bayer, Electrophoresis 23 (2002) 2949.

[38] K. Zhang, Z. Jiang, C. Yao, Z. Zhang, Q. Wang, R. Gao, C. Yan, J.

Chromatogr. A 987 (2003) 453.

[39] C. Yao, R. Gao, C. Yan, J. Sep. Sci. 26 (2003) 37.

[40] K. Zhang, R. Gao, Z. Jiang, C. Yao, Z. Zhang, Q. Wang, C. Yan, J. Sep.

Sci. 26 (2003) 1389.

[41] A.R. Ivanov, C. Horváth, B.L. Karger, Electrophoresis 24 (2003) 3663.

[42] E. Rapp, A. Jakob, A.B. Schefer, E. Bayer, K. Albert, Anal. Bioanal.

Chem. 376 (2003) 1053.

[43] R. Nakashima, S. Kitagawa, T. Yoshida, T. Tsuda, J. Chromatogr. A 1044 (2004) 305.

[44] V. Szucs, R. Freitag, J. Chromatogr. A 1044 (2004) 201.

[45] C. Yao, S. Tang, R. Gao, C. Jiang, C. Yan, J. Sep. Sci. 27 (2004) 1109.

[46] Z. Liang, J. Duan, L. Zhang, W. Zhang, Y. Zhang, C. Yan, Anal. Chem.

76 (2004) 6935.

[47] T.P. Hennessy, M. Quaglia, O. Kornysova, B.A. Grimes, D. Lubda, K.K. Unger, J. Chromatogr. B 817 (2005) 127.

[48] B. Channer, G.G. Skellern, M.R. Eurby, A.P. McKeown, A.S. Rathore, J.

Chromatogr. A 1095 (2005) 172.

[49] I.N. Levine, Physical Chemistry, Fourth edition, McGraw-Hill, Inc. New York, 1995, p 477.

[50] T. Isano, J. Chem. Eng. Data 29 (1984) 45.

[51] W.A. Gobie, C.F. Ivory, J. Chromatogr. A 516 (1990) 191.

[52] D.S. Burgi, K. Salomon, R.-L. Chien, J. Liq. Chromatogr. 14 (1991) 847.

[53] W. Kok, Chromatographia 51(Suppl.) (2000) S24.

[54] C.J. Evenhuis, R.M. Guijt, M. Macka, P.J. Marriott, P.R. Haddad,

Electrophoresis 26 (2005) 4333.

Karlstad University Studies

In-line ap p lic atio n o f elec tric fi elds in c ap illary sep aratio n system s

T h e m ag nitude o f an elec tric fi eld p o ssib le to ap p ly in a c ap illary sep aratio n system is lim ited, b ec ause a h ig h elec tric fi eld c auses a to o h ig h c urrent th ro ug h th e c ap illary.

A p p lic atio n o f th e elec tric fi eld in-line w ill g ive an inc reased c o nduc tivity in th e c o lum n, furth er inc reasing th e risk o f to o h ig h c urrents. T h e c o nduc tivity c h ang es w ere fo und to result fro m an o verall inc rease in io nic streng th w ith in th e elec tric fi eld. T h e inc rease in io nic streng th is c aused b y th e inc rease in m o b ile p h ase io ns w ith elec tro p h o retic velo c - ity ag ainst th e fl o w , to g eth er w ith OH

o r H

O

io ns (dep ending o n p o larity) fo rm ed at th e inlet elec tro de. F urth er it w as fo und th at th e use o f a p ressuriz ed reservo ir o r sp litting o f th e fl o w at th e inlet elec tro de c o uld sig nifi c antly lim it th e c o nduc tivity c h ang es and th ereb y th e m ax im um ap p lic ab le elec tric fi eld streng th s c o uld b e inc reased.

Björn ErikssonIn-line application of electric fields in capillary separation systems