JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 199 I VOL. 6 215 Evaluation of a 13.56 MHz Capacitively Coupled Plasma as a Detector for Gas Chromatographic Determination of Organotin Compounds Degui Huang and Michael W. Blades* Department of Chemistry The University of British Columbia 2036 Main Mall Vancouver British Columbia V6T IY6 Canada A demountable parallel plate capacitively coupled plasma source operating at 13.56 MHz has been developed as a gas-chromatographic detector for the determination of organotin compounds. The effect of operating power and gas flow-rate on analyte emission intensity has been measured and spatial emission characteristics of the plasma have been evaluated. The detection limits for Me,Sn Me,SnCI and Pr,Sn were 0.079 0.190 and 0.168 ng s-l re- spectively.Keywords Capacitiwely coupled plasma; gas chromatography; organotin compounds In recent years metal and non-metal speciation studies have increasingly attracted the interest of analytical chemists because of the importance of trace elements in toxicology and environmental science. One means of obtaining species speci- fic information is through the use of chromatographic separa- tions. A variety of atomic emission sources have been developed as chromatographic detectors since these sources can provide element specific information about each eluting peak. Plasma sources in particular the inductively coupled plasma (ICP) direct current plasma (DCP) and microwave- induced plasma (MIP) have been extensively investigated for this purpose. The coupling of gas chromatography (GC) with an ICP was described by Windsor and Denton' for simultaneous multi- elemental analysis of organic and organometallic compounds. Microwave-induced plasmas have been investigated for many years.The first coupling of an MIP with GC was reported by McCormack et al. in 1965.* Since then many examples of the use of this methodology have been developed including the use of capacitively coupled microwave plasmas (CMP)3 and surface wave sustained plasmas (~urfatron).~ Various types of samples have been determined such as pesticide residues,' haloforms in drinking water6 and organometallic compounds.' Organotin compounds have been used widely as biocides cat- alysts and polymer stablizers and their effects on the environ- ment are causing concemR A knowledge of the concentration chemical form and distribution of these compounds provides important information on the origin and transport mechanisms.Several GC and liquid chromatography approaches incorporat- ing plasma-based detection have been developed for the deter- mination of organotin compounds. Krull and Panaro' used a system whereby the organotin compounds were separated using high-performance liquid chromatography (HPLC) followed by continuous on-line hydride generation with a DCP emission spectrometer being used to detect the effluent. Suyani et al."'de- scribed the use of helium microwave-induced plasma mass spectrometry for capillary gas chromatographic detection of or- ganotin compounds. Uchida et al." recently reported a capaci- tively coupled helium microwave plasma as an excitation source for the determination of organotin compounds. They used a CMP in which microwaves were generated using a magnetron and conducted through a coaxial waveguide to the CMP excita- tion source.A tubular tantalum electrode sample injector was employed for the CMP in order to achieve high sensitivity and a more stable discharge. The analytical performance of this plasma source for the determination of inorganic tin and butyltin was evaluated by interfacing the helium CMP to a gas chromato- graph. The analytical merit compared well with helium MIP systems although electrode contact with the plasma introduces the possibility of contamination. * To whom correspondence should be addressed. One of the problems of using an MIP as a GC detector is that materials deposit on the walls of the discharge tube.Noticeable deposits can be found for long-chain hydrocarbons and other oxygen-free compounds. This problem is even more acute with samples containing inorganic compounds that are prone to forming refractory species. Besner and Huberti2 in- vestigated the effect of dopants on tin emission in a helium MIP. The helium plasma was doped with various liquid and gaseous materials and sulphur hexafluoride was found to give the best results. Recently a novel parallel plate capacitively coupled plasma (CCP) which can be used as an emission spectrometric detector for gas chromatography has been described.I3 Using this CCP a helium or argon plasma can be generated at atmospheric pres- sure at frequencies of 0.20 or 27.18 MHz and at carrier gas flow- rates as low as 20 ml min-'.The present paper describes the further development of the CCP as a GC detector operated at 13.56 MHz outlines some of the spectral and operational char- acteristics and characterizes its application to the determination of some environmentally important organotin compounds. Experimental Power Sup p I y An Advanced Energy Model RFX 600 13.56 MHz (Fort Collins CO USA) radiofrequency (r.f.) generator equipped with an Advanced Energy Model ATX-600 automatic impe- dance matching system was used to supply power to the CCP torch. The ATX-600 tuner was modified to include a 4-5 pH inductor in series with an output line to improve matching effi- ciency. When operated at 200 W forward power the reflected power could be maintained at less than 3 W.Spectrometric System The monochromator photomultiplier tube current amplifier and chart recorder used in this study were as described in a previous publication.'.' Gas Chromatograph The gas chromatograph was the same as that used in reference 13 except that a Supelco Model SPB-I fused silica capillary column (15 m x 0.53 mm 0.d. with a film thickness of 0.50 pm) was used rather than a packed column because of its chemical inertness and high column efficiency. For organotin compounds if a solid column support is insufficiently covered by the stationary liquid phase (e.,?. 2-5%) sample adsorption on the exposed siliceous sites becomes significant with polar solutes and peak tailing occurs.'3 The injector block was maintained at 280 "C for all experiments.216 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 Plasma Torch Fig.1 shows a schematic diagram of the CCP and holder used for this work. The torch was fabricated from a section of fused silica rectangular in cross-section with external dimensions of 6 x 4 mm and internal dimensions of 4 x 2 mm. The total length of the rectangular portion of the torch was 6cm. One end of the torch from which emission was observed was open to the atmosphere and the other end was connected to a T- shaped arrangement of quartz tubes (0.d. 4 mm i.d. 2 mm) for the introduction of sample and make-up gas (see Fig. 1). The effluent from the GC was introduced into the torch through a quartz capillary tube which was sealed to the torch at one end.An additional gas inlet allowed the use of a make-up gas. The torch was placed in a vice-like clamping device and 40mm long 1.8 mm thick wafers of boron nitride were placed between the torch and the stainless-steel electrodes which were 4cm in length. These were clamped in place using Delrin and an aluminium holder which was tightened using an adjustable clamp as depicted in Fig. 1. Using this torch mount the CCP discharge could be easily assembled and different sizes of quartz tubing and electrodes could be tested. The torch was operated both with and without the presence of the boron nitride insulator between the quartz and the elec- trodes. With the former the intensity of the H I line at 686.13 nm at 150 W was almost the same as that at 100 W with the latter.It would appear that 50% of the power is lost with the former structure. However without the boron nitride insulator the electrodes became fairly hot and the Delrin sof- tened which made it difficult to maintain the integrity of the torch and holder. For this reason the torch was operated with the boron nitride strips for all the experiments described in this paper. Helium was used for both the carrier and make-up gases. Transfer Interface For the transfer line between the chromatograph and the CCP the capillary column was enclosed in a copper tube (70 x 0.32 cm o.d.) around which was wound heating tape and envel- oped with glass wool and cotton tape. The temperature was controlled by using a Variac rheostat to adjust the voltage to the heating tape.The transfer line was maintained at a temper- ature of 280 "C for all experiments. Support rod Plasma torch I Make-up gas / Adjustable screw db inlet Boron nitride clamp- Top view Stainless steel 0 *' - Delrin End-on view Fig. 1 Schematic diagram of the discharge tube and support structure Data Acquisition Except when indicated otherwise the working conditions were as follows the r.f. power supply operated at a forward power of 150 W with a reflected power of 2 W in auto-matching mode. The make-up helium gas flow-rate was 200 ml min-I. The monochromator wavelength setting for Sn I at 284.0 nm was made by using a tin hollow cathode lamp. After the sample was injected the gas chromatographic column was maintained at the initial temperature. The solvent began to elute at a retention time of 0.61 min.As soon as the solvent was eluted the r.f. power was applied; the plasma self-ignited and the chromatograph was operated in isothermal or tempera- ture programme mode and data were collected on the chart recorder. Chemicals Ferrocene (98%) tetramethyltin (Me,Sn 99%) and trimethyl- tin chloride (Me,SnCI 99%) were purchased from Aldrich di- chloromethane (CH,C12 99.9%) from BDH and tetrapropyltin (purity unknown) was obtained from Alpha Inorganics. All the chemicals were used without further purification. Results and Discussion Helium Plasma Background A wavelength scan of the background emission from the 13.56 MHz atmospheric pressure helium CCP between 200 and 600 nm is shown in Fig. 2. The identification and assign- ment of the molecular bands were made using data obtained from reference 15.The most prominent features are those orig- inating from OH NH NO N and Nz+ similar to those ob- served in an MIP.Ih However one of the differences is that a Q,-branch with 308.9 nm (O,O) and 282.90 nm (1,O) for OH ( ? ~ + - ? I I ) was found to be more intense for the CCP compared with the R,-branch at 306.36 nm (0,O) and 28 1.13 nm (1,O) for the MIP.Ih This may be related to the difference in gas temper- ature and suggests that the gas temperature in the CCP is prob- ably lower than that for an MIP. However further experiments must be completed in order to verify this suggestion. The background features observed at different positions (vertically and axially to the discharge tube) did not change significantly.From this observation it can be concluded that the OH NH NO N and N,' molecular bands arise f!om impurities in the helium gas supply. Estes et al.' used a 5 A molecular sieve im- mersed in liquid nitrogen in order to investigate whether the molecular bands observed in an MIP were caused by back dif- fusion of air. Their results confirmed that no air entrainment occured and the bands resulted primarily from the presence of impurities. Spatial Emission Characteristics The spatial emission characteristics for He I at 447.15 nm He I at 504.77 nm and H I at 486.13 nm were measured for the helium CCP and are shown in Fig. 3. These measure- ments were made by forming an end-on image of the plasma at the entrance slit of the monochromator and translating the image across the entrance slit by moving the torch assembly.As can be seen from Fig. 3(a) the spatial distribution for all three species shows a maximum near the walls and a minimum at the centre. The spatial distributions of emission from Fe at 371.99 nm and Sn at 284.0 nm introduced as organic compounds were also measured. For these lines a relative maximum was observed at the centre as shown in Fig. 3(h). The Sn spatial distribution was obtained using the output from the gas chromatograph whereas the Fe signal was collected with continuous introduction of sublimed fer- rocene in the headspace of a sampling vial. A similar distri-JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 L 217 200 250 300 350 400 450 500 550 Wavelengthlnm Fig.2 flow-rate 200 ml min-'; and camer gas flow-rate 0 ml min-' Background emission spectrum of the 13.56 MHz helium plasma from 200 to 600 nm. Input power; 100 W; reflected power 2 W; make-up gas c .r 0 1 2 .- $40 + - P) a 30 20 10 I I 1 0 1 2 Distance from left side wall/mm Fig. 3 (a) Spatial distribution of emission from A H 1 486.13 nm; B He I 504.77 nm; and C He I 447.15 nm without sample. Input power I50 W; make-up gas flow-rate 200 ml min-' and carrier gas flow-rate 0 ml min-'. (h) Spatial distribution of A Sn 284.0 nm; and B. Fe 371.99 nm for a helium CCP with organic sample introduced. Input power 150 W and make-up gas flow-rate 200 ml min-' bution has been observed in a surface wave plasma (surfa- tron),".'' in which cylindrical plasma tubes were used.Richard et d . l X explained that if one-step excitation through electron collision with an atom in the ground level is assumed the distribution of emission is dependent on the distribution of the total electric field intensity ET and the electron density [n(r)] as a function of radial position 1' through the relation (1) where n,(r) is the population density of the excited atoms in level j A is a constant independent of position and k is a value dependent on the plasma medium and excited-state parameter and can be determined from theory. The magnitude of both the electric field and the electron density are spatially dependent. The electric field intensity is higher at the walls whereas the electron density decreases near the walls as a result of recom- bination losses.Although the appropriate measurements have not been made it is possible that the spatial distributions observed for the CCP are similar in origin to those for the sur- fatron. n,(r) = A n(r) E,"(I') Effect of Input Power In order to study the effect of changes in r.f. input power several injections of 20 ng of Sn (0.1 p1 of a solution of 100 ppm of Me,SnCI in CH,CI solvent) were made into the gas chromatograph and the emission intensity for Sn was measured. The intensity was not corrected dynamically for background however it was found that the background was relatively constant for these experiments. The peak intensity for the Sn I line at 284.0 nm as a function of input power is shown in Fig. 4. As stated earlier about 50% of the power is lost with the electrodes isolated from the discharge tube but this structure prevented the electrodes from becoming too hot and produced a more stable plasma From Fig.4 it can be seen that the intensity of Sn increased almost linearly with an increase in r.f. power. It is probable that the total applied r.f. power is not all delivered to the plasma since there are losses in the output inductance the dielectric ma- terial and the electrodes themselves. Therefore the actual power consumed in the discharge is less than is indicated in Fig. 4. As a result of the heating of the electrodes it was difficult to operate at powers higher than about 400 W since the heat would soften the Delrin insulators. For these reasons a torch with water cooled copper electrodes has been developed and future work will be carried out using this new design.218 .$ 50 E 40- .- f 30- .- c 3 20 a 10 J0URNA.L OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL.6 - - - 0 100 200 300 400 500 Power/W Fig. 4 Effect of input power on emission intensity for Sn at 284.0 nm. Make-up gas flow-rate 200 ml min-I and carrier gas flow-rate 10 ml min-' 30 CI d I /x 10 lk 2bo 3 k &I 5& sbo Make-up gas flow-rate/ml min-' Fig. 5 284.0 nm. Input power 100 W and carrier gas flow-rate 10 ml min-' Effect of make-up gas flow-rate on emission intensity for Sn at 70 I /x I 0 1 I 1 I I 5 10 15 20 25 Carrier g a s flow-rate/ml min-' Fig. 6 284.0 nm. Input power. 100 W and make-up gas flow-rate 200 ml min-' Effect of camer gas flow-rate on emission intensity for Sn at Effect of Gas Flow-rate The effect of the helium make-up gas flow-rate on the Sn peak intensity is shown in Fig.5. The Sn signal increased with the gas flow-rate from 50 to 350 ml min-' reaching a maximum between 350 and 400 ml min-1 after which a decrease was ob- served. At flow-rates greater than about 250 ml m i d turbu- lence could be observed in the gas flow through the torch resulting in instability of the emission signals. Therefore a make-up gas flow-rate of 200 ml min-' was used for further experiments. The response of the Sn emission intensity as a function of gas chromatograph carrier gas flow-rate over the range 8-20 ml min-l is depicted in Fig. 6. Although an in- crease in carrier gas flow-rate caused an increase in the Sn emission intensity the chromatographic resolution degrades at flow-rates higher than about 10 ml min-I.Therefore a 10 ml min-' carrier gas flow-rate was used for all further studies. CG-CCP System Performance The stock solutions (loo0 ppm of Sn for each compound) were prepared by dissolving the relevant organotin compounds Me,Sn I R.f. o n \ R.f. off Retention time/min Fig. 7 Chromatogram of mixture of Me,Sn Me,SnCI and Pr,Sn using temperature programming. Input power 150 W; make-up gas flow-rate 200 ml min-I; and camer gas flow-rate 10 ml min-' in dichloromethane and working solutions were prepared by appropriate dilution with the same solvent. Fig. 7 shows a typical chromatogram of a mixture of Me,Sn Me,SnCI and Pr,Sn. A 0.1 pl aliquot of a 10 ppm solution of each com- pound containing 1 ng of Sn was injected.The signal was collected at 284.0 nm without background correction. The r.f. voltage was switched on after the solvent was eluted. The chromatograph was used in the temperature programme mode i.e. it was maintained at 45 "C for 1 min then raised to 260 "C at the rate of 50 "C min-I and then held at this temperature for another 2 min. As can be seen in Fig. 7 the sensitivity of the peak signal for Me,Sn is 2.8 times better than that for Me,SnCl. Serious tailing is exhibed by Pr,Sn which is be- lieved to be due to condensation of Pr,Sn on the walls of the discharge tube. Since the heating tape could not be brought close to the electrodes because of possible discharge between the electrodes and the heating tape a 'cold gap' existed in this region.Therefore condensation of Pr,Sn which has a high boiling point (222 "C) is a strong possibility. Another reason for tailing may be the deposition of tin oxide on the walls. Doping with some reagent materials can minimize this problem. I? The detection limits for Me,Sn Me,SnCI and Pr,Sn were 0.079 0.190 and 0.168 ng s-' respectively. For these values the definition of minimum detectable level used by Sullivan19 has been used. This is the mass of analyte required to produce a peak that is twice the height of the peak-to-peak noise divided by the full width at half height of the peak in seconds. Chromatographers usually (but not always) measure the peak- to-peak base line variation which is considered to be 60 as a measure of the noise.z0 For this report the peak-to-peak noise measurement was averaged over a time of period of 30 s.Conclusions The parallel plate CCP torch is a potential alternative to the MIP or CMP as a detector for GC for the determination of organotin compounds. The CCP can be operated at an r.f. input power of up to 400 W and make-up gas flow-rates ranging from 50 to 400 ml min-I. The current system is an improvement over that previously described" in that a de- mountable torch structure has been used and an automatic impedance matcher has been coupled to a 600 W 13.56 MHz r.f. oscillator. Detection limits for organotin compounds for the CCP indicate that the detection limit is superior to those obtained using a CMP" but inferior to those obtained using an MIP."' However the CCP is still at an early stage of de- velopment and there is scope for improvement and further exploration into torch geometries operating frequencies power sources and sample introduction strategies.Using theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 219 present configuration there is some power dissipation in the r.f. electrodes which leads to power loss and heating of the electrodes. A torch incorporating water cooled copper elec- trodes is currently being tested in an effort to overcome this problem. The authors thank Professor M. Fryzuk for the use of the gas chromatograph. Acknowledgement is made to the donors of The Petroleum Research Fund administered by the American Chemical Society the Natural Sciences and Engineering Re- search Council of Canada and the University of British Colum- bia for partial support of this research.References Windsor D. L.. and Denton M. B. J. Chromatogr. Sci. 1978 32 366. McCormack A. J.. Tong S. C. and Cooke W. D. Anal. Chem. 1965,37 147. Hanamura S . Smith B. W.. and Winefordner J. D. Anal. Chem. 1983,55,2026. Hubert J. Moisan M. and Richard A. Spectror*him. Aria Part B 1979.34. 1. Bache C. A. and Lisk D. J. Anal. Chem. 1967,39,786. Quimby B. D. Delaney M. F. Uden P. C. and Barnes R. M. Anal. Chem. 1967,51,875. Estes S . A. Uden. P. C. and Barnes R. M. Anal. Chem. 1981,53 133. 8 9 10 1 1 12 13 14 15 16 17 18 19 20 Thompson J. A. J.. Sheffer M. G.. Pierce R. C. Chau Y. K. Cooney J. J. Cullen W. R. and Maguire R. J. Orgonotin Com- pounds in the Aquatic Enrironment Scientific Criteria for Assessing Effects on Environmental Quality NRCC Report No. 22494 National Research Council. Ottawa Canada 1985. Krull I. S. and Panaro K. W. AppI. Spectrosc.. 1985,39,960. Suyani. H. Creed J. and Caruso J. and Satzger. R. D. J. Anal. A t . Spectrom. 1989,4777. Uchida H. Johnson P. A. and Winefordner J. D. J . Anal. At. Spec- from.. 1990,5 8 1. Besner. A. and Hubert J. J. Anal. At. Spectrom. 1988,s. 381. Huang D. Liang D. C. and Blades M. W. J. Anal. At. Spectrom. 1989,4789. Crompton T. R. Comprehensive Organometallic Analysis Plenum New York 1981 p. 487. Pearse R. W. B. and Gaydon A. G. The Identification of Molecular Spectra Chapman and Hall London 4th edn. 1976. Zander A. T. and Hieftje. G. M. Anal. Chem. 1978,50. 1257. Proud J. M. and Lussen L. H. Radiative Processes in Discharge Plasmas Plenum New York 1986 p. 38 1. Richard A. Barbeau A. Besner J. Hubert J.. Moisan M. and Sauve G. Can. J. Phys. 1988.66 740. Sullivan J. J. in Modern Practice of Gas Chromatography ed. Grob R. L. Wiley New York 1977. Quimby B. D. and Sullivan. J. J. Anal. Chem. 1990,62 1027. Paper 0103 7350 Received August 14th 1990 Accepted December 8th 1990