Alternatives for coupling sequential injection systems to commercial capillary electrophoresis–mass spectrometry equipment

Journal of Chromatography A, 1127 (2006) 278–285

Alternatives for coupling sequential injection systems to commercial capillary electrophoresis–mass spectrometry equipment B. Santos, B.M. Simonet 1 , B. Lendl 2 , A. R´ıos 3 , M. Valc´arcel ∗ Department of Analytical Chemistry, University of Cordoba, Edificio Marie Curie Anexo, Campus de Rabanales, Annex Building C-3, 14071 C´ordoba, Spain Received 13 January 2006; received in revised form 19 May 2006; accepted 29 May 2006 Available online 18 July 2006

Abstract On-line coupling of an automated flow system with a commercially available capillary electrophoresis (CE) system with an electrospray interface (ESI) for mass spectroscopic (MS) detection is described. The peculiarities of CE–ESI-MS interfaces, in which a high electrical field must be applied to the capillary end where the sample is provided by the flow system, introduce significant difficulties for the appropriate work of the entire arrangement. Experimental strategies are proposed for achieving stable conditions for on-line sample pre-treatment, conditioning of the separation capillary, sample injection, as the proper separation. The versatility and robustness of the proposed arrangement is discussed, taken as example the separation of a variety of amines. Connection of the CE system’s pressure to the automated flow system enables hydrodynamic introduction of sample with high precision. The developed hyphenated system is of practical relevance as it opens an avenue for the simplification and automation of the whole analytical process required when using powerful CE–ESI-MS equipments. © 2006 Elsevier B.V. All rights reserved. Keywords: Sequential injection (SI); Capillary electrophoresis–mass spectrometry (CE–MS); Hybridation; Automatization

1. Introduction Capillary electrophoresis (CE) has evolved as a powerful separation technique [1] with distinct advantages compared to other separation methods such as high-performance liquid chromatography (HPLC). These are high separation efficiency, short analysis time and low sample consumption. A further important advantage of these methodologies is their flexibility which allows the development of a variety of operational modes very well adapted to solve particular analytical problems. Electrophoretic techniques can be focused to increase sensitivity, for example, the incorporation of sweeping or stacking techniques;

∗

Corresponding author. Tel.: +34 957218616; fax: +34 957218616. E-mail address: [email protected] (M. Valc´arcel). 1 Present address: Department of Chemistry, University of Balearic Islands, 07122 Palma de Mallorca, Spain. 2 Present address: Institute of Chemical Technology and Analytics, TU Vienna, Getreidemarkt 9-164/AC, 1060 Vienna, Austria. 3 Present address: Department of Analytical Chemistry and Food Technology, University of Castilla-La Mancha, Campus de Ciudad Real, 13004 Ciudad Real, Spain. 0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.05.077

or to increase resolution, as, for example, the use of pressure assisted electrochromatography (pCEC) or micellar electrokinetic chromatography (MEKC). In fact, CE has probably moved to the forefront of analytical methodology. However, along with the above mentioned attractive properties two main deficits of CE methodologies can be identified when it is intended to solve difficult analytical problems on a routine basis. Among these challenges, lack of sensitivity [2] and the need for sometimes labor-intensive sample preparation steps in case of difficult sample matrices can be the principal practical problems. The first shortcoming was in part solved with the introduction of capillary electrophoresis–mass spectrometry (CE–MS) by the group of Smith [3,4] in 1987. This coupling combines the advantages of CE with those of mass spectrometry (a highly sensitive and selective detector), but the sensitivity achieved is not enough for many applications and it did not provide a solution to complicated and time consuming sample preparation. For example, due to the extremely small inner diameter of the capillary, solid particles present in the samples must be removed, in order to prevent clogging of the capillary. Also, for the analysis of samples containing large amounts of proteins or other macromolecules, it is frequently necessary to remove these components prior to sample

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285

injection, as macromolecules tend to adsorb on capillary walls which results in significant changes in migration times and the appearance of interfering peaks in the electropherogram [5,6]. These limitations of CE methodologies are widely recognised and responsible for hindering their spread and use in routine applications. A possible way to overcome these disadvantages is automation of the required operational steps prior to injection of the sample into a CE system. Apart from simplifying analysis, automation of sample treatment also would allow analytical methods to be transferred to applications in routine laboratories with greater ease. To this end, flow systems have been widely used for this purpose for a broad variety of different applications. In case of commercial CE systems this coupling has only recently been achieved at-line when dealing with commercial instruments (CE–UV and CE–MS) or on-line when dealing with laboratory-made CE systems employing optical or electrochemical detection. In at-line hyphenation of CE–MS with automated flow systems, the treated sample is directed located in a vial of the autosampler of the CE–MS system [7]. The atline coupling of sample treatment device to CE–MS equipments results in an important reduction of the volume of analysed sample. However, this volume is in part limited by the volume of treated sample that must be located into the vial, as this volume must be enough to carry out the sample introduction into the capillary. This limitation can be avoided by using an on-line coupling, which will permit to additionally reduce the sample volume required for analysis. Other important practical advantages of this coupling can be obtained in terms of analysis time, robustness, ruggedness and a greater flexibility, which allows to cover a wide range of applications. Laboratory-made CE systems with split flow interfaces for on-line coupling sample treatment devices have already been described [8,9]. Application of these systems for determination of pseudoephedrine in human plasma [10], priority phenol pollutants in water [11], acidic drugs in biological samples [12] or for the monitoring of glutamate and aspartate in vivo using microdialysis on-line [13] demonstrates the practical relevance of on-line coupling. In all these systems the capillary end interfacing to the flow systems was connected to earth and the high voltage applied at the side of the detector. In case of CE–electrospray interface (ESI)–MS systems this approach is not possible due to the characteristics of the ESI interface, in which a high electrical field must be applied to the capillary end to be interfaced with the flow system. Electrospray ionization is actually one of the most commonly used atmospheric pressure interfacing technique for coupling CE techniques to MS [14]. Fig. 1 compares the different configurations for coupling flow systems to CE equipments with optical and MS detectors. As can be seen, two different ESI interfaces can be distinguished depending on the way in which the ESI voltage is applied. However, as it was indicated before, in all the situations a high voltage must be applied in the inlet region of the CE capillary. This situation produces a difference of voltage between the flow system and the interface which can result in a non-desirable electrical current. In the literature, only one attempt to perform FIA–CE–ESI-TOF-MS (flow injection analysis–capillary electrophoresis–electrospray ionization time-of-flight mass spectrometry) has been described

279

Fig. 1. Typical configuration of commercial CE–ESI-MS equipment with the: (A) ESI needle connected to an earth wire, (B) ESI needle connected to an earth wire and (C) ESI chamber connected to an earth wire.

[15]. The authors did not describe problems with residual currents toward the flow system. Probably it was due to the buffer used in this work. In fact both, electrophoretic buffer and the stream of flow system, consisted of 50% of formic acid/ammonia (pH 3, 15 mM) and 50% acetonitrile. The low ionic strength and the high percentage of organic solvent of this buffer results in a low conductivity. However, when aqueous solutions or solutions with higher ionic strength are used important problems with residual currents toward the flow system are observed. For this reason, the objective of this work was to provide efficient strategies for reliable on-line coupling of sequential injection (SI) systems to commercial CE–MS equipment. 2. Experimental 2.1. Apparatus An Agilent HP3D capillary electrophoresis system (Waldbronn, Germany) was used to separate the analytes. CE separations were conducted in a 70 cm × 50 ␮m I.D. × 375 ␮m O.D. fused-silica capillary. The CE equipment was coupled to an Agilent 1100 Series LC/MSD system (equipped with a quadrupole

280

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285

analyser) via an electrospray atmospheric pressure ionization (API) interface. The make-up flow of sheath liquid was delivered by the Agilent 1100 isocratic pump, which was operated at a split ratio of 1:100. Control of the CE–MS system, data acquisition and processing was performed using the Agilent ChemStation software. The automated flow system was constructed using a Cavro (Sunnyvale, CA, USA) XP 3000 syringe pump equipped with a 5 ml syringe, one 10-port and one 6-port selection valve (Vici Valco, Schenkon, Switzerland) and PTFE tubing with 0.3, 0.5 and 0.8 ␮m of inner diameter. Alternatively to the Valco selection valve a Cavro six-port selection valve has been tested as well. The set-up was controlled by a laboratory-written MS Visual Basic 6.0 (Microsoft) based software (Sagittarius, Version 3.0.) running on personal computer under Windows XP. 2.2. Materials and reagents Histamine, ethanolamine, isopropylamine, isoamylamine, phenylethylamine, heptylamine and butylamine (Sigma, St. Louis, MO, USA) and 1,3-diaminopropane (Fluka, Buchs, Switzerland) were the analytes. Citric acid monohydrate (Merck, Darmstadt, Germany) and sodium hydroxide (Prolabo, Paris, France) were used to prepare the background electrolyte and electrophoretic solutions. Standard solutions of the amines, background electrolyte and sheath liquid were made in purified water (18 M
) from a Millipore Milli-Q water purification system. 2.3. Interfacing of the automated flow system to the CE–ESI-MS system The automated flow system was coupled to a capillary electrophoresis–mass spectrometry equipment using a

Fig. 2. Interface used for coupling CFS to CE–MS equipment.

laboratory-made polyether ether ketone (PEEK) T-piece (Fig. 2). To achieve this coupling the hardware configuration of the Agilent HP3D CE system had to be slightly modified. The reason for this modification lies in the analysis protocol of the commercial system. Following this protocol the instrument needs to detect a vial in the sample position that is lifted by an elevator to immerse the capillary end together with the surrounding electrode into the inlet vial. Only after this movement has been accomplished high voltage may be applied. Unfortunately, due to spatial constraints this process cannot be accomplished if the capillary end is connected to the automated flow system. However, in addition to the elevator which lifts the sample or buffer vials, the commercial system counts with a second elevator that allows to immerse the other capillary end into a outlet vial when only UV/vis detection but not MS detection is used. To solve this conflict the cabling of the two elevators was interchanged. As a consequence the ChemStation detects a vial at the inlet position which in fact is located at the other position. In this way for

Fig. 3. Manifold of the sequential injection system coupled to the CE–ESI-MS equipment.

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285

application of high voltage the system lifts the vial located at the outlet position instead of the vial located at the sample position. Thus, the collision of the elevator with the flow system–CE interface (T-piece) is avoided and a high voltage may be applied to the CE–ESI-MS system being on-line coupled to the automated flow system. The automated flow system manifold coupled to the CE–MS equipment is shown in Fig. 3. 2.4. CE and ESI-MS operating conditions The separation background electrolyte consisted of 50 mM citric acid, pH 2.0, and was prepared on a weekly basis. The applied voltage for electrophoretic separation was +15 kV, which generated a current of 10 ␮A, and the temperature of the capillary was kept constant at 20 ◦ C. The coaxial sheath liquid consisted of a 50:50 (v/v) methanol–water mixture containing 1% formic acid and was used at a flow rate of 0.3 ␮l/min. Fresh liquid was prepared weekly to ensure stable electrospraying in the ionization source. The operating conditions for ESI and MS were as follows: fragmentor voltage, 70 V; nebulized pressure, 10 psi; N2 (drying gas) flow rate, 7 l/min; N2 (drying gas) temperature, 175 ◦ C; and capillary voltage, 4 kV. For data acquisition, the ions with m/z values corresponding to [M + H]+ for each amine (viz. 60, 62, 74, 75, 88, 112, 116 and 122 u for isopropylamine, ethanolamine, butylamine, 1,3diaminopropane, isoamylamine, histamine, heptylamine and phenylethylamine, respectively) were monitored. Each individual amine was quantified from its corresponding ion. 3. Results and discussion The study of electrical stability of the system and the usefulness of the automatic SI system to perform clean-up and capillary conditioning, as well as hydrodynamic and electrokinetic sample introduction into the capillary, are discussed below. 3.1. Strategies for stable on-line coupling of flow systems with a CE–MS equipment In the flow injection (FI)–CE systems reported so far the interface has always been grounded. The challenge for on-line coupling an automated flow system to a capillary electrophoresis equipment with mass spectrometric detection results from the fact that a high voltage needs to be applied at the interface. If a set-up as described for FI–CE would be used for on-line coupling of a CE–MS equipment, uncontrolled current would be diverted by the background electrolyte. This situation, apart from presenting a hazard to the equipment and the operator, needs to be avoided as uncontrolled separation condition may result. In CE systems the current obtained throughout the separation is monitored and taken as a measure for stability of the achieved separation conditions. In the CE–MS equipment under use this current is calculated by the energy that the power supply needs to provide a constant voltage across the capillary. The electrical current diverted through the flow system thus impedes

281

from knowing the current generated in the capillary, which is responsible for the electrophoretic separation. To this end, the electrical capillary current must be recorded because it is recommended to be lower than 50 ␮A, in order to avoid problems with the MS interface. Therefore, the high potential applied at the interface needs to be isolated from the attached flow system. In addition, a dissipation line needs to be added for safety of equipment and operator. In this work, the automated flow system at the side of the interface consisted of a selection valve and a syringe pump. Valco and Cavro selection valves were tested. When using the Valco valve the dissipation (safety) line could be added between the syringe pump and the selection valve, as the Valco valve itself did not dissipate the applied potential. The safety line consisting of a platinum electrode and grounded with a copper cable was inserted to the flow system via a simple T-piece. The platinum was used to establish the electrical connection with the solution and to avoid electrochemical reactions. With such a safety line in place the on-line coupling could be studied in detail. To achieve stable conditions for electrophoretic separation, background electrolyte needs to be driven through the interface after sample injection. Whereas in the FI–CE systems reported so far this does not present any problem because the electrode at the interface is grounded, a residual current was measured in our system. Depending on the buffer system used, currents up to 10 times of those measured when disconnecting the flow system were obtained. These results show that considerable current is generated between the interface and the safety line. It is not recommended working under these conditions as changes in the composition of the BGE may occur, as well as due to the low reproducibility of the capillary current. In consequence, low reproducibility of migration time will be obtained. In this paper two different strategies to eliminate the dissipative current have been tested, as it is described below. 3.2. Use of deionized water plugs for avoiding dissipative currents The first approach of interrupting the current comprised insertion of deionized water between the interface and the automated flow system. For that purpose, the coil connecting the valve and the interface had to be sufficiently long as to accommodate both the separation buffer and the water. The volume of water required for this purpose is considerable due to the high dispersion in the sequential injection system. If the background electrolyte is used as carrier, 300 ␮l of water needs to be inserted in the reaction coil when working at a flow rate of 50 ␮l/min. If the water is used as the carrier in the SI system, the water contained in the reaction coil 1 (Fig. 3) is enough for insulating the interface from the security line. Using water as carrier and placing the BGE for electrophoretic separation in the coil 2 (Fig. 3) between valve and interface several background electrolytes which are frequently used in CE–MS were tested to check the stability and robustness of the experimental set-up. The use of formic acid, ammonium acetate, citric acid and ammonium borate as background electrolytes at concentrations ranging from 20 to 100 mM revealed that the intensity was stable and similar to

282

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285

Fig. 4. Typical evolution of the electrophoretic capillary current using the proposed configuration and pure water to avoid residual currents in the CFS system.

that obtained working with the same capillary but without the automated flow system. The reproducibility of the current was in all cases lower than 0.8%. As an example, Fig. 4 shows typical recordings of electrical capillary current, in which the high stability and reproducibility of the system can be observed. For this configuration the amount of background electrolyte located between the water and the interface must be enough for a complete electrophoretic separation. Considering the small volume of the interface, a constant flow of the BGE must be maintained to avoid possible changes in its composition. We observed that citric acid and formic acid buffers were the most stable as changes in the current were avoided being necessary to remove the buffer in the interface each 2 min. In case of using ammonium acetate a minimum constant flow of 50 ␮l/min was necessary. Therefore, we only recommend such a configuration when the electrophoretic separation time is low (see the dimensions of the coils in Table 1). It should be taken into account that long separation times require long coils; hence, the time to fill the coil with fresh buffer will also be long. The BGE must be introduced into the coil 2 in several steps to avoid that the BGE reaches the automatic burette due to dispersion in the coils. Table 2 shows a typical sequence used for the flow system. Due to the pressure drop inside a 0.5 mm I.D. reaction coil the maximum length of both the reaction coil and coil connecting the valve to the interface may not by higher than 300 cm each one. The length of the coil 2 (Fig. 3) is determined by the following criteria: (i) separation time, (ii) regeneration of the BGE in the interface (flow required during the electrophoretic separation) and (iii) water required to cut the electrical contact, which is a function of the nature of the BGE and the flow used. Typically we also use a coil of 300 cm for separation times of 20 min. To facilitate introduction of a new buffer and to reduce the analysis time, it is interesting to use the same BGE electrolyte as the carrier in the flow system. In this case, the cleaning and conditioning of the coil 2 will be faster.

3.3. Use of air plugs for avoiding dissipative currents An interesting alternative to insulate the electrode from the safety line is insertion of an air plug between the interface and the valve of the flow system in the reaction coil 2 (Fig. 3). We found that insertion of 100 ␮l of air is enough; however, for sake of robustness and safety we recommended to introduce a segment of 200 ␮l. Also, in this configuration, a safety line needs to be introduced into the system which was placed in the position depicted in Fig. 3. If we compare this system with the first one, the most remarkable characteristic is the dramatic reduction in the length of the reaction coil from 300 cm (minimum) to 50–100 cm, which permits an increase in the length of the coil between the valve and the interface, but in turn resulting in Table 1 Length of coil located between the multi-selection valve and the CFS–CE interface as a function of the background electrolyte composition and flow rate used to renovate the BGE from the interface Background electrolyte

Length (cm)

Flow rate (␮l/min)

Separation time (min)

Citric acid at pH 2.0

100 150 150 200

10 0 10 10

20a 5b 30a 40a

Formic acid at pH 2.0

150 150

0 10

7b 30a

Ammonium acetate at pH 5.0

150 150 150

0 10 50

1b 4b 10a

Ammonium acetate at pH 9.0

150 150 150

0 10 50

2b 5b 10a

a b

Time to observe a interruption of the electrical current. Time to observe a diminution of 5% in the electrical current.

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285 Table 2 Typical sequence used to fill the coil 2 with BGE

283

Table 3 Dimensions of the waste draining tube used to perform the rinsing step as a function of dimensions of electrophoretic capillary

Using water as carrier Steps to introduce the BGE in coil 2 (1) Aspirate of 300 ␮l of BGE in coil 1 (2) Propulsion of 100 ␮l through the coil 2 (3) Propulsion of 600 ␮l through the waste II (see Fig. 3) (4) Coil 2 was filled with BGE by repetition of this whole sequence Using BGE as carrier Steps to introduce the BGE in coil 2 (1) Propulsion of 1000 ␮l of carrier through the waste I (see Fig. 3)

longer separations. We limited our studies to the use of PTFE tubes with I.D. of 0.5 mm. Tubings with an I.D. of 0.8 mm were discarded due to the high dispersion generated in them. Tubes of smaller inner diameters (e.g. 0.3 mm) resulted also not practical due to the excessive pressure build up in the system, which results in some of the flow being redirected though the capillary. With tubes of 0.5 mm of diameter, the maximum separation time was established according to the flow, coil length and the nature of the BGE. As can be seen in Table 1 the length of the coil must be longer than 150 cm when working at 10 ␮l/min in order to achieve electrophoretic separation in 30 min. 3.4. On-line capillary conditioning In capillary electrophoresis the regeneration or clean-up of the capillary between runs is commonly crucial for achieving reproducible migration times. Even though most studies using on-line coupling of flow systems to capillary electrophoresis have eliminated rinsing as capillary clean-up step, it remains a desirable system capability in order to make all the batch separations directly transferable to our system. To implement this capability, a selection valve was located between the interface and the waste tube (Fig. 3) to select between two alternative drainage tubings. The first tube of 30 cm and an inner diameter of 0.5 mm is selected when performing electrophoretic separations. The other tube with an inner diameter of 0.3 mm and a length between 50 and 80 cm is selected during the capillary rinsing step. By using this second waste tube and by applying a high flow rate (50–100 ␮l/min), high pressure can be achieved in the system, which results in some flow being deviated though the capillary. In Table 3 we recommend lengths for different capillaries and for a flow rate of 50 ␮l/min. With this simple modification of the manifold, the rinsing step could be fully automated. For reduced analysis time, the whole rinsing process

Dimensions of capillary

Length of waster draining tube (cm)

35 cm × 50 ␮m I.D. 35 cm × 75 ␮m I.D. 60 cm × 50 ␮m I.D. 60 cm × 75 ␮m I.D. 100 cm × 50 ␮m I.D. 100 cm × 75 ␮m I.D.

40 25 80 60 150 100

can be conducted in one single step. For example, a standardized and generally accepted rinsing protocol for the conditioning of new 57 cm long capillaries consists of 10 min of 1 M HCl, 5 min of 0.1 M NaOH, 5 min of water and 5 min of BGE (pressure of 2 bar). In our system, a tube 80 cm long tube with an I.D. of 0.3 mm was located as the second waste tube for this conditioning. Afterwards, 250 ␮l of 0.1 M NaOH, 200 ␮l of water and 250 ␮l of 1 M HCl were aspirated at 20 ␮l/min and located in the reaction coil. All the reagents were then pumped through the interface at 30 ␮l/min and the conditioning of the capillary was easily achieved. In addition, we observe that with such a procedure the reproducibility of migration is generally better than for batch procedures. This is probably due to the partial mixing of the rinsing reagents by dispersion, which results in less dramatic pH changes in the capillary. A simple rinse with sodium hydroxide was enough to achieve reproducible migration times between runs. 3.5. Hydrodynamic and electrokinetic introduction of sample Most on-line continuous flow–CE studies reported so far use electrokinetic sample introduction despite the fact that hydrodynamic injection is the generally preferred mode as it avoids discrimination problems. For long separations also a long coil located between the CFS and interface is needed to maintain BGE at the capillary end. As a consequence, high dispersion of a sample is observed when injected into the flow system and transported to the capillary end which results in reduced sensitivity of analysis. In order to avoid this problem and to speed up the analysis sequence an auxiliary line for sample introduction has to be implemented (Fig. 3). Also, in this case, 300 ␮l air was introduced for insulation purposes.

Table 4 Comparison of hydrodynamic and electrokinetic introduction of sample into the capillary using the proposed methodologies Hydrodynamic introduction Using flow system

Histamine Phenylethylamine Butylamine

Electrokinetic introduction Using CE system’s pressure

% RSD (area)

% RSD (time)

% RSD (area)

% RSD (time)

5.4 6.7 5.0

0.8 0.7 1.1

1.5 1.9 1.1

0.5 0.6 0.5

Values correspond to six measurements.

% RSD (area)

% RSD (time)

2.3 3.1 3.4

0.8 1.5 1.3

284

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285

Fig. 5. Typical electropherogram obtained with the hybridized CFS–CE–ESI-MS system. The peaks were: (1) 1,3-diaminopropane; (2) histamine; (3) ethanolamine; (4) isopropylamine; (5) butylamine; (6) isoamylamine; (7) phenylethylamine; (8) heptylamine.

For the hydrodynamic introduction of the sample, one of the positions of the selection valve located between the interface and the waste tube was connected to the equipment’s air system. To do this the connection of the prepuncher was disconnected and directly attached to the valve. A sample plug was located at the interface through the auxiliary injection line (Fig. 3) to prepare for injection. The valve was then switched and the sample was hydrodynamically introduced into the capillary. Afterwards, the selection valve was switched to the waste position and the sample was removed from the interface by BGE. Once the separation channel and the secondary injection channel were isolated using air, the potential was applied to the equipment to perform the electrophoretic separation. The sample must be introduced when the channels are filled with solution as otherwise compression of the air will result in the sample volume introduced into the capillary being irreproducible. We also studied the possibility of performing the sample introduction using a system similar to that for rinsing the capillary. However, as the sample introduction must be performed with a high precision system, low pressure of 0.5 psi, with a precision of ±0.01 psi (2%), is recommended. Although the results obtained show that it is possible to introduce the sample at high pressure with the syringe pump through a waste tube, the precision of the results expressed as relative standard deviation are higher then 5%. In Table 4 the precision obtained in terms of peak areas and migration times for the three modalities of sample introduction are compared. For this study, 50 mM citrate at pH 2.0 was used as BGE. In order to study the reliability of the modalities and to avoid problems of discrimination, which can produce problems of reproducibility, individual aqueous standards solution of histamine, phenyethylamine and butylamine were used as samples. These three analytes were selected on the basis of their different structures, which permit

the precision of the system at low and higher migration times to be studied. 3.6. On-line preparation of standards Within this study the capability of the automated flow system to perform on-line sample pre-treatment has been used for Table 5 Figures of merit of the proposed method for the determination of target analytes Y = a + bx

r

Sy/x

LOD

LOQ

Isopropylamine

a = −0.023 b = 0.4031

0.997

0.091

25

83

Ethanolamine

a = 0.0021 b = 0.6451

0.999

0.020

30

100

Butylamine

a = 0.0003 b = 0.7865

0.998

0.018

23

76

1,3-Diaminopropane

a = −0.0032 b = 0.4317

0.999

0.021

51

162

Isoamylamine

a = −0.0021 b = 0.7612

0.998

0.015

19

66

Histamine

a = −0.0003 b = 0.6541

0.999

0.013

20

66

Heptylamine

a = 0.0021 b = 0.5472

0.999

0.017

31

103

Phenylethylamine

a = 0.0011 b = 0.6420

0.999

0.011

17

57

Calibration graphs were established form eight points. a: intercept; b: slope; r: correlation coefficient; Sy/x : standard deviation of residual; LOD: limit of detection (␮g/l); LOQ: limit of quantification (␮g/l).

B. Santos et al. / J. Chromatogr. A 1127 (2006) 278–285 Table 6 Analysis of synthetic water samples using the proposed method

4. Conclusions

Sample

Analyte

Amount added (␮g/l)

Amount found (␮g/l)

Recovery (%)

1

Isopropylamine Ethanolamine Butylamine 1,3-Diaminopropane Isoamylamine Histamine Heptylamine Phenylethylamine

250 250 250 250 250 250 250 250

245 261 247 259 250 253 254 247

98.0 104.4 98.8 103.6 100.1 101.2 101.6 98.8

2

Isopropylamine Ethanolamine Butylamine 1,3-Diaminopropane Isoamylamine Histamine Heptylamine Phenylethylamine

300 300 300 300 300 300 300 300

302 297 298 306 301 299 296 294

100.7 99.0 99.3 102.1 100.3 99.7 98.7 98.0

Isopropylamine Ethanolamine Butylamine 1,3-Diaminopropane Isoamylamine Histamine Heptylamine Phenylethylamine

400 400 400 400 400 400 400 400

396 399 405 405 394 403 393 410

99.2 99.8 101.2 101.3 98.5 100.8 98.3 102.5

3

285

on-line preparation of standards, as well as for on-line filtering of samples. For on-line filtering, microfilters were incorporated in the tube of sample aspiration (see Fig. 3). Standards at concentrations ranging from 0.25 to 15 mg/l have been prepared by aspirating segments of water and the most concentrated standard at different ratios into the reaction coil. A total volume of 500 ␮l has been prepared for each diluted standard. After flow reversal, the mixture was pumped to the interface via the injection line at 50 ␮l/min. After flushing the interface with buffer injection was performed hydrodynamically. This system was applied to the separation of the eight selected amines. Fig. 5 shows typical electropherograms and Table 5 describes the analytical features of the calibration curves. As can be seen, these results demonstrate the high efficiency of the proposed alternative for coupling automatic flow systems to CE–MS equipment. The reliability of the proposed method was evaluated by applying it to several synthetic water samples. As can be seen from Table 6, the recoveries of target analytes ranged from 98 to 105%. A paired t-test revealed the absence of statistical difference between the concentrations added and those found.

The robust on-line coupling of an automated flow system with a CE–MS system, as described in this contribution, introduces a new powerful integration of different analytical techniques. Complementarily, advantages may be achieved and exploited to approach complex analytical problems. Sample pre-treatment, including reaction and sample clean-up steps, can be performed in an automated fashion. The worked sample can then be separated in its components taking advantage of the high resolving power of modern capillary electrophoresis. Finally, important advantages in the detection stage in terms of selectivity and sensitivity can be achieved by mass spectrometry. Therefore, it may be expected that this hyphenation can be applied for the solution of a broad variety of different analytical problems. Acknowledgements The authors are grateful to the Spanish Ministry of Education and Science for the support through the project CTQ200401220. One of the authors (B.S.) also wishes to thank the Spanish Ministry of Education and Science for award of a fellowship of the FPI program. References [1] R. Weinberger, Practical Capillary Electrophoresis, Academic Press, London, 1993. [2] B.M. Simonet, A. R´ıos, M. Valc´arcel, Trends Anal. Chem. 22 (2003) 605. [3] J.A. Olivares, N.T. Nguyen, C.R. Yonker, R.D. Smith, Anal. Chem. 59 (1987) 1230. [4] R.D. Smith, C.J. Barinaga, N.T. Nguyen, H.R. Udseth, Anal. Chem. 60 (1988) 1948. [5] J.R. Veraart, H. Luingeman, U.A.Th. Brinkman, J. Chromatogr. A 856 (1999) 483. [6] M. Dankova, D. Kaniansky, S. Fanali, F. Ivanyi, J. Chromatogr. A 838 (1999) 31. [7] B. Santos, B.M. Simonet, A. R´ıos, M. Valc´arcel, Electrophoresis 25 (2004) 3231. [8] P. Kuban, B. Karlberg, Trends Anal. Chem. 17 (1998) 34. [9] Z.L. Fang, H.W. Chen, O. Fang, Q.S. Pu, Anal. Sci. 16 (2000) 197. [10] H.W. Chen, Z.L. Fang, Anal. Chim. Acta 394 (1999) 13. [11] P. Kuban, M. Berg, C. Garc´ıa, B. Karlberg, J. Chromatogr. A 912 (2001) 163. [12] J.R. Veraart, M.C.E. Groot, C. Gooijer, H. Lingeman, N.H. Velthorst, U.A.Th. Brinkman, Analyst 124 (1999) 115. [13] M.W. Lada, T.W. Vickroy, R.T. Kennedy, Anal. Chem. 69 (1997) 4560. [14] J. Abian, J. Mass Spectrom. 34 (1999) 157. [15] J. Samskog, S.K. Bergstr¨om, M. J¨onsson, O. Klett, M. Wetterhall, K.E. Markides, Electrophoresis 24 (2003) 1723.