Interfacing microchip capillary electrophoresis with inductively coupled plasma mass spectrometry for chromium speciation

Abstract

A microchip based electrophoresis separation system was interfaced with inductively coupled plasma mass spectrometry to provide rapid elemental speciation capabilities. The feasibility of this hyphenated method for elemental speciation was demonstrated by the on-line electrophoretic separation of Cr^III and Cr^VI within 30 s using an 8 cm long separation channel etched in a glass base.

---

Introduction

Speciation studies are vital for the understanding of the toxicity, mobility and bioavailability of elements in environmental or biological samples. The coupling of capillary electrophoresis (CE) with inductively coupled plasma mass spectrometry (ICP-MS) or atomic emission spectrometry (ICP-AES) provides a very sensitive, element specific detection method in conjunction with high separation efficiency. However, interfacing these techniques remains challenging, primarily due to the flow incompatibility. Since Olesik¹ first proposed the idea of interfacing CE with ICP-MS and ICP-AES for elemental speciation, several papers with various interface designs have been published.^2–9 Most of these designs include the use of a make-up stream to overcome the flow incompatibility between CE and ICP-MS or ICP-AES. Despite this problem having been mostly overcome, the conventional CE separation typically takes several minutes or even longer to separate the analytes, during which the ICP-MS is idle. This reduces the overall efficiency of the hyphenated technology by decreasing the sample throughput and increasing the running cost per analysis. The obvious area where increases in efficiency can be made is to reduce the separation time considerably. This can be achieved either by a higher separation voltage or a shorter and therefore faster separation system. Higher separation voltages are limited by instrumental (maximum voltages from power supplies) and physical (heating) effects.

Microfabricated analytical systems using electrophoresis as a separation mechanism have been shown to drastically reduce separation times and to simplify instrumental design. A reduction in the dimensions of a capillary dramatically reduces the time required for the separation.¹⁰ Another advantage of microchip technology is that pre-etched channel networks can be used to provide several flow streams, omitting tube connectors and thus reducing dead volumes.

Sample introduction is a key issue for the success of electrophoretic analyses. In most microchip CE (µCE) systems, sample injection is achieved by the electrokinetic method.¹¹ This is because the majority of µCE systems were developed to separate biomolecules, such as proteins, peptides and complex compounds. However, for elemental analysis of highly charged, low-molecular-mass inorganic ions, the electrokinetic method is not always appropriate. The high mobility of these small ions would inevitably cause severe sample bias,¹²e.g., for a sample containing anions and cations, a positive injection potential would discriminate in favour of the anions. In such situations, injection with hydrodynamic flow is more appropriate since it eliminates the discrimination.

This paper describes how a microchip based electrophoresis system was interfaced to an ICP-MS. The sample was introduced into the separation channel by hydrodynamic injection with a gravity pump. A voltage applied along the length of the channel separated the ions of interest, but a hydrodynamic flow was also used to shorten the overall analysis time.

Experimental

The microchip used in this µCE-ICP-MS system consisted of two (30 × 25 × 3 mm) glass plates. The device was fabricated using a previously reported method.¹³ The channel network was produced in the base plate by photolithography and wet chemical etching (1% HF–NH₄F), to produce channels of 100 µm width and 20 µm depth. The length of the main channel was 10 cm and the effective length for the separation (after the sample injection channel) was about 8 cm. The last section of the channel (4 mm in length) was manually enlarged to accept the interface capillary. Holes were drilled into the top plate to form reservoirs (2 mm in diameter) in the designated positions. The two plates were then thermally bonded in a furnace at 570 °C for 3 h. In order to increase the capacity of the reservoirs, 200 µl pipette tips (with ends cut to fit tightly into the drilled holes) were inserted into the top plate. The design of the microchip, with the channel layout, is shown in Fig. 1.

Fig. 1 Schematic diagram of the microchip.

The microchip was interfaced to the ICP-MS (Thermo Elemental PQ2+) via a commercially available, low flow rate concentric nebuliser (Micromist, Glass Expansion, Switzerland). This was achieved by using a length of 40 mm encapsulated PEEK tubing (0.178 mm id, 1.59 mm od, CETAC); a 4 mm length of the sleeve was removed from one end and the inside capillary (0.47 mm od) was inserted into the exit port of the separation channel and held in place with a thin layer of epoxy glue (UHU Plus, Germany). The other end was then directly connected to the nebuliser by using a Teflon connector (EzyFit, Glass Expansion, Switzerland), as shown in Fig. 2. A miniature cyclonic spray chamber (Cinnabar, Glass Expansion, Switzerland) was used to obtain high sample introduction efficiency with the low flow rate nebuliser. Table 1 shows the operating conditions and measurement parameters for the ICP-MS.

Fig. 2 Schematic diagram of the interface.

Table 1 Optimised instrumental conditions and measurement parameters for the Thermo Elemental PQ2+ ICP-MS

Rf forward power/W	1350
Reflected power/W	1–3
Coolant gas flow rate/1 min^−1	14
Auxiliary gas flow rate/1 min^−1	1.2
Nebuliser gas flow rate/1 min^−1	0.890
Spray chamber	Glass, water cooled at 4 °C
Data acquisition mode	Peak jumping
Points per peak	3
Dwell time/ms	10.24
Detector mode	Pulse counting

---

The natural uptake flow rate of the nebuliser at the optimised conditions was about 42.5 µl min⁻¹, whereas the flow rate through the separation channel was controlled by a gravity pump at 0.585 µl min⁻¹ to give a suitable residence time within the channel for separation to occur. The make-up solution was delivered by peristaltic pump (Minipuls 3, Gilson) to the make-up reservoir. Make-up fluid was then introduced near to the end of the separation channel by a short channel that offered little hydrodynamic resistance (see Fig. 1). This arrangement prevented any suction effect from the nebuliser on the separation channel. A gravity pump was used to drive the carrier flow and introduce the sample plug. The length of the sample plug was determined by the injection time, which was manually controlled by a three-way injection valve. Platinum wires were inserted through small openings in the sidewall of the pipette tips to serve as electrodes. A Wallis Hivolt power supply system (Advance Hivolt, West Sussex, UK) was used to supply the high voltage for the separation. The platinum wire in the makeup reservoir was used as the cathode.

The makeup solution consisted of 1% nitric acid with 10 µg l⁻¹ In, and was used to tune the nebuliser for optimal instrument performance. A solution consisting of 15 mM of nitric acid was added to the carrier reservoir by a syringe. When necessary, 100 µg l⁻¹ of Pb could also added to the carrier solution as an internal standard to monitor the stability of the hydrodynamic flow in the microchip. The mixed standard of Cr⁶⁺ and Cr³⁺ used throughout the work was prepared from single standards of Cr⁶⁺ (1000 ± 3 µg ml⁻¹, in water) and Cr³⁺ (1000 ± 3 µg ml⁻¹, in 2% HCl) from Qmx Laboratory Limited (Thaxted, Essex, UK). Sample solutions were freshly diluted from corresponding speciation standards each day. High purity de-ionised water (Elgastat UHQ PS, Elga, High Wycombe, UK) was used throughout (18 MΩ cm⁻¹ resistivity).

Results and discussions

Stability of µCE-ICP-MS interface

Initially, the stability of the interface was assessed using the indium standard. A solution of indium (10 µg l⁻¹) was continuously pumped into the make-up reservoir. It was found that if a connection was made between the make-up delivery tubing and the make-up reservoir, pulses were observed leading to the deterioration of the baseline stability. To overcome this problem, the makeup solution was just pumped into the reservoir allowing natural uptake of the solution by the nebulizer. The highest counts were obtained when the nebulizer gas flow was tuned to 0.890 l min⁻¹ and the natural uptake flow rate was measured at about 42.5 µl min⁻¹. Under these conditions the counts for a 10 µg l⁻¹ indium solution were measured to be around 170,000 cps and the stability of 10 measurements over 10 min was better than 2%. This was about half the counts obtained by the conventional nebuliser (PFA DeGalen V-groove nebulizer), where the sample uptake was 1 ml min⁻¹. Considering the large difference in the sample uptake flow rates, this demonstrated that the sample transport efficiency of the test system was satisfactory for the work to be carried out.

Fig. 3 shows the control of the flow of the carrier stream and the stability of the system. The baseline was monitored via the indium signal, and the hydrodynamic flow in the separation channel was monitored by a lead internal standard solution. When the carrier flow was switched off, the counts for lead dropped sharply to the baseline until the carrier flow was turned on again (Fig. 2). This shows that the flow through the separation channel could be externally controlled by the flow injection valve. The slight increase in the indium signal was expected and is due to the increase of makeup flow when the carrier flow was stopped.

Fig. 3 Baseline stability and the hydrodynamic flow in the carrier stream. Makeup solution contains 10 µg l⁻¹ of In in 1% of nitric acid. Carrier solution consists of 15 mM nitric acid and 100 µg l⁻¹ of Pb.

Sample injection

The next step of the experiment was to evaluate sample injection into the microchip. As mentioned earlier, electrokinetic injection was not appropriate for small ion analysis. Hence, in this system, a simple flow injection mode driven by a gravity pump was employed. When the valve was switched to the “sampling” position, the sample in the reservoir was hydrostatically pushed into the carrier channel. After changing to the carrier position, the sample plug was moved through the separation channel by the carrier electrolyte. The length of the sample plug was controlled by the injection time. The main problem with the hydrodynamic introduction of the sample is the dispersion of the sample plug due to Poiseuille flow.¹⁴ High flow rates and the use of a serpentine design for the microchannel help to alleviate the dispersion.¹⁵ As long as the height of the gravity pump was fixed the elution time of the sample plug was very reproducible, and was 24 s when the carrier flow was controlled at 0.585 µl min⁻¹. Fig. 4 shows four repeat injections of the mixed standard without the application of the separation voltage to show the repeatability of the injection method.

Fig. 4 Peak profiles for four repeat hydrodynamic injections of 100 µg l⁻¹ Cr^III and 50 µg l⁻¹ Cr^VI without application of voltage. Injection time was 7 s.

When looking at quantitative analysis, the standard deviation of four repeated injections was 7.8% based on the peak areas and 3.6% based on the peak heights. These values would be radically improved by automation of the injection. As expected, peak broadening and tailing was observed when the carrier flow rate was reduced.¹⁵

Electrophoretic separation of Cr^III and Cr^VI

To evaluate the feasibility of electrophoretic separation of different species, a mixed standard solution consisting of 100 µg l⁻¹ of Cr³⁺ and 50 µg l⁻¹ of Cr⁶⁺ was added to the sample reservoir. The injection valve was switched to the on position for 5 s, and then switched to the carrier position. The voltage was applied between the carrier reservoir and the make-up reservoir. As can be seen in Fig. 5, the two species are clearly beginning to separate on application of 2000 V. As the voltage increased the resolution significantly improved and, as can be seen from the timescale in Fig. 5, the separation was completed within 30 s. Safety considerations with the prototype set-up meant that the voltage could only be increased up to 5000 V. A further increase in the carrier flow rate (by raising the gravity pump reservoir) would give a further reduction in the residence time in the separation channel; however, this would raise two problems. Firstly, as the sample injection was manually controlled in the present experiment, an injection time of less than 5 s would inevitably cause poor reproducibility. Secondly, reducing the overall time for which the voltage was applied would compromise the resolution. Currently a new design of the system is being developed which can work with higher voltages and has an automated sample injection process, giving a faster separation and better resolution.

Fig. 5 Separation of Cr^III with Cr^VI in a microchip based CE-ICP-MS system. The carrier electrolyte consists of 15 mM nitric acid. Sample solution consists of 100 µg l⁻¹ Cr^III and 50 µg l⁻¹ Cr^VI. Two replicate injections were carried out for each voltage. The duration of sample injection was 5 s.

Conclusion and future work

In this work, a microchip based capillary electrophoresis system was interfaced with ICP-MS. Its feasibility for rapid elemental speciation was demonstrated by the separation of Cr^III and Cr^VI in 30 s using an 8 cm separation channel. Future work will be focused on the optimisation of this system. This would include the application of higher voltages, automation of the sample injection process and optimisation of the carrier electrolyte and composition of the sample solution. Separation of species of other elements will also be carried out in the near future.

Acknowledgements

The authors would like to thank Shanks First for financial support for this project. We would also like to thank Robert Knight of the University of Hull for his technical assistance.

References

1. J. W. Olesik, J. A. Kinzer and S. V. Olesik, Anal. Chem., 1995, 67, 1–12.
2. E. Mei, H. Ichihashi, W. Gu and S. Yamasaki, Anal. Chem., 1997, 69, 2187–2192.
3. V. Majidi and N. J. Miller-Ihli, Analyst, 1998, 123, 803–808.
4. J. Li, T. Umemura, T. Odake and K. Tsunoda, Anal. Chem., 2001, 73, 5992–5999.
5. Y. Y. Chan and W. T. Chan, J. Chromatogr. A, 1999, 853, 141–149.
6. K. A. Taylor, B. L. Sharp, D. J. Lewis and H. M. Crews, J. Anal. At. Spectrom., 1998, 13, 1095–1100.
7. L. Bendahl, B. Gammelgaard, O. Jøns, O. Farver and S. H. Hansen, J. Anal. At. Spectrom., 2001, 16, 38–42.
8. C. B'Hymer, R. M. C. Sutton, K. L. Sutton and J. A. Caruso, Anal. Commun., 1999, 36, 349–353.
9. A. Tangen, W. Lund, B. Josefsson and H. Borg, J. Chromatogr. A, 1998, 826, 87–94.
10. A. Manz, J. C. Fettinger, E. Verpoorte, H. Lüdi, H. M. Widmer and D. J. Harrison, Trends Anal. Chem., 1991, 10(5), 144–149.
11. C. Zhang and A. Manz, Anal. Chem., 2001, 73, 2656–2662.
12. S. F. Y. Li, Capillary Electrophoresis: Principles, Practice and Applications, Elsevier, Amsterdam, 1993, pp. 38–40.
13. T. McCreedy, Anal. Chim. Acta, 2001, 427, 39–43.
14. J. Růžička and E. H. Hansen, Flow Injection Analysis, John Wiley, New York, 2nd edn., 1988, ch. 3, p. 99.
15. R. Tijssen, Anal. Chim. Acta, 1980, 111, 71–89.

---