JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 815 Vesicle-mediated Hig h-performance Liquid Chromatography Coupled to Hydride Generation Inductively Coupled Plasma Atomic Emission Spectrometry for Speciation of Toxicologically Important Arsenic Species* Yi Ming Liu Maria Luisa Fernandez Sanchez Elisa Blanco Gonzalez and Alfred0 Sanz-Medel? Department of Physical and Analytical Chemistry University of Oviedo C/Julian Cla veria 8 33006-Oviedo Spain The use of vesicles as mobile phases can provide a synergic combination for high-performance liquid chromatographic (HPLC) separation when coupled to hydride generation inductively coupled plasma atomic emission spectrometric (HG-ICP-AES) detection as illustrated here for arsenic speciation. The more toxic arsenic species including arseneous arsenic monomethylarsonic and dimethylarsinic acids can be separated within 10 min by using didodecyldimethylammonium bromide (DDAB) vesicles in phosphate buffer containing 0.5% methanol and a c18 reversed-phase column which had previously been modified by DDAB soluton.In addition the sensitivity of detection on-line by HG-ICP-AES of the four separated arsenic species was enhanced by the presence of DDAB vesicles. The normalized detection limits for arsenic with the proposed speciation method were down to the sub-ng level and the observed precisions were always better than *!job for the four forms of arsenic at the 10 ng level of the element in water. Arsenic recoveries (full procedure) were found to range from 93 to 108% for tap water and human urine samples by this vesicle-mediated HPLC-HG-ICP-AES technique for arsenic speciation.Keywords Vesicles; high-performance liquid chromatography; arsine generation; inductively coupled plasma atomic emission spectrometry arsenic speciation Element speciation at real-life concentration levels presents a formidable challenge to modern analytical chemistry. Although electroanalytical and ultraviolet-visible mole- cular spectroscopic methods can be used the most popular powerful and reliable analytical tools for element speciation today would appear to be the so-called 'hyphenated' techniques particularly the combination of a powerful separation technique [mainly gas chromatography (GC) or high-performance liquid chromatography (HPLC)] with an adequate element specific detector (e.g.atomic detectors).' Plasma atomic spectroscopy methods including induc- tively coupled plasma and microwave-induced plasma atomic emission spectrometry (ICP-AES and MIP-AES respectively) and recently ICP mass spectrometry (MS) are presently believed to be the most effective for element speciation owing to their high sensitivity simultaneous multi-element monitoring ability and element specific In spite of present availability of fast automatic impe- dance-matching networks that enable good coupling of the r.f. field to the ICPs into which organic solvents are injected one of problems with coupling high-performance liquid chromatography (HPLC) to ICP-AES (or MS) detec- tion is associated with the low tolerance of plasmas to the commonly used organic solvents.6 Their introduction to the ICP detector results in higher plasma background de- creased stability and increase in noise and even eventual extinction of the plasma owing to excessive organic solvent loading.Therefore the search for alternative HPLC mobile phases is obviously worthwhile because it could guarantee satisfactory performance of HPLC-ICP-AES (or MS) hybridizations for speciation purposes. Suyani et al. described the application of a micellar mobile phase of sodium dodecyl sulfate (SDS) for HPLC separation i.e. micellar chromatography of alkyltin compounds followed by ICP-MS dete~tion.~ However the high SDS concen- tration in the micellar HPLC mobile phase which was *Presented at the 1993 European Winter Conference on Plasma tTo whom correspondence should be addressed.Spectrochemistry January 10- 15 1993. necessary for achieving acceptable alkyltin separation could also cause problems such as salt deposition at the nebulizer tip and/or mass spectrometer sampling cone. On the other hand research on arsenic speciation is receiving a great deal of attention because of the high toxicity of certain arsenic compounds and the widely different toxicological effects of several arsenic species. Whereas inorganic arsine arseneous (As*l1) and arsenic (AsV) acids are highly toxic monomethylarsonic (MMAs) and dimethylarsinic (DMAs) acids are only moderately toxic. Moreover arsenobetaine and arsenocholine are considered to be non-toxic to living organisms.8 Numerous approaches have been applied to the determination of arsenic species in various matrices mainly using hyphe- nated techniques based on HPLC separation coupled to atomic spectrometric d e t e ~ t i o n .* ~ ~ ~ ~ ~ ~ ~ ~ - * ~ Separations by HPLC including ion chromatography ion-exchange and ion pair reversed-phase chromatography are most frequently used for arsenic speciation. Sheppard et a1.13 resolved As"' AsV DMAs and MMAs on a single ion chromatography column in 15 min but a slightly alkaline mobile phase (carbonate buffer pH 7.5) had to be used which is not advisable for silica based HPLC columns. Hakala et a l l 4 separated the same mixture of arsenic species in 4 min on a c18 reversed-phase column with tetrabutylammonium hydroxide ( 10 mmol dme3) in the mobile phase as the ion-pairing reagent.However the chromatographic resolution observed between AsIII-DMAs and DMAs-MMAs appeared to be inadequate. More- over Morin et a l l 8 claimed that octadecyl-bonded silica columns were not able to resolve arsenic species such as As1I1 or arsenobetaine by ion-pair reversed-phase chromatography whatever the ion-pairing reagent concen- tration used owing to their limited pH stability range. So far anion-exchange chromatography has been most frequently employed for separation of arsenic species because of their widely varying first pK values. However problems associated with the limited lifetime of these ion-exchange columns and what is more the need to resort to organic solvents and high buffer concentrations for shortening the elution times have been noted by various w o r k e r ~ .~ ~ J ~ J ~816 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Previous work in this laboratory has shown that surfac- tant organized assemblies (e.g. micelles and vesicles) can enhance chemical generation processes of volatile species [e.g. hydride generation (HG)]. In other words ‘ordered media’ ability to organize reactants at a molecular level and thus favourably to modify the equilibria or kinetic constants,20 can be exploited to enhance the performance of analytical HG-ICP-AES. It has been shown that the detection limits of arsenic determination by vesicle-en- hanced HG-ICP-AES were improved by a factor of two with a significant improvement in the precisions of the ICP emission signals and in the tolerance to matrix interfer- ences.,’ This favourable effect of vesicles could be com- bined with their potential in HPLC separations.22 Although the utilization of surfactant-based organized assemblies in HPLC separations has greatly increased in recent years only micellar mobile phase systems have been investigated extensively.The use of vesicles as mobile phases seems so far to be largely neglected owing to their relatively high viscosities. This paper reports on the first vesicle-mediated coupling of HPLC separation to HG-ICP-AES detection and its detailed evaluation for speciation of toxicologically impor- tant arsenic species which happen to form volatile hydrides on-line with the plasma detector after direct reaction with sodium tetrahydroborate solutions. Experimental Apparatus A Knauer Model 6400 HPLC pump with an attached sample injection valve equipped with a 100 mm3 loop was used for eluent delivery and sample introduction.The analytical column was an LKB cartridge (240 x 4 mm id.) packed with 10 pm Lichrosorb RP 1 8 which had previously been modified by didodecyldimethylammonium bromide (DDAB) solution as described below. A four-channel peristaltic pump (Minipuls-2 Gilson) a 100 cm poly(tetra- fluoroethylene) (PTFE) mixing coil and a laboratory-made gas-liquid separator23 constituted the continuous hydride generator. A mass flow controller (Sho-Rate Brooks Veemendaal The Netherlands) was used for introducing an argon stream into the gas-liquid separator. An S.M. Model 500 W high-intensity ultrasonic processor equipped with a microtip (Sonics & Materials) was used for DDAB vesicle preparation.An Amicon Model 52 (W.R. Grace Danvers USA) was used for ultrafiltration experiments. A Perkin-Elmer ICP 5000 spectrometer interfaced with a microcomputer (Perkin-Elmer Model 3 500) was used for ICP emission measurements. A computer program for transient ICP emission data acquisition and processing was written in BASIC. Chemicals Stock solutions (1000 mg dm-3 of As) of arsenite DMAs and MMAs were prepared by dissolving appropriate amounts of As203 (Merck) in 25 cm3 of 0.5 mol dm-3 NaOH solution and then diluting the solution to 1 dm3 with 0.6 mol dm-3 HC1; however (CH3),AsO2Na-3H,O (Sigma) and CH3AsO(ONa)2-6H20 (Carlo Erba) were dissolved directly in water. An AsV stock solution (1000 mg dm-3) was obtained from Merck.The stock solutions were stored in polyethylene bottles at 4 “C. Working standards were freshly prepared daily by dilution in ultrapure Milli-Q water. The DDAB vesicular solution ( 1 x mol dm-3 DDAB) was prepared by adding 0.23 g of DDAB (Fluka) to 50 cm3 of water and sonicating the solution until all added Vesicular Sample (100 mm7 mobile phase - 1 cm’ min-’ HPLC pump To ICP KI+ Gas-liquid separator vesicle 1 cm3 min-’ Mixing coil Waste NaBH Peristaltic pump Fig. 1 Schematic diagram of the vesicle-mediated HPLC-HG- ICP-AES hybridization for arsenic speciation DDAB was dissolved (about 10 min with a power output of 60 W). This solution was used for preparing by simple dilutions all the other vesicular solutions mentioned in this work.Sodium tetrahydroborate solution (2% m/v) was pre- pared by dissolving NaBH powder (Probus Barcelona) in 0.5% m/v NaOH solution. Filtration of the solution through a Whatman grade 4 filter-paper before use was carried out. All other chemicals were of analytical-reagent grade and distilled and de-ionized (Milli-Q system Millipore) water was used throughout the work. Procedures Column modijkation The C ,-bonded silica reversed-phase column was modified by passing through 500 cm3 of a DDAB solution (1 x mol dm-3) in 50% methanol-50% water at a flow rate of 1 cm3 min-l. Water was then passed through the column. The modified column was kept in water when not in use. Ultrajiltration for partition experiments Single-component arsenic solutions containing 50 pg dm-3 of As were prepared for each form of arsenic species in the mobile phase.Ultrafiltration was performed by using cellulose ultrafiltration membranes (Diaflo Amicon Divi- sion). Arsenic contents of the filtrate and of the original solution were determined by the HG-ICP-AES method. Arsenic speciation Typically 100 mm3 of the working standards of the arsenic species mixture were injected directly into the HPLC system to study the experimental conditions and to evaluate the analytical performance characteristics of the hyphe- nated procedure. The mobile phase was prepared as follows 200 cm3 of NaH,PO (10 mmol dm-3) buffer solution (pH 5.75) containing 0.5% (v/v) methanol was de- gassed with helium and then 200 mm3 of mol dm-j DDAB vesicle solution were added.The HPLC eluate was first mixed with the HCl solution through the mixing coil and then mixed with sodium tetrahydroborate solution for arsine generation (see Fig. 1). A continuous stream of argon carries the generated arsine directly into the ICP injector tube that is the nebulizer was removed and the exit of the gas-liquid separator was directly connected to the plasma. The HPLC-HG-ICP-AES hybridized assembly is shownJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 - 817 - " 0 Table 1 Experimental conditions for arsenic speciation by the vesicle-enhanced HPLC-HG-ICP-AES hybridization L .- \ .- i CI .- CI c.. ;* G 4 K Chromatography- Column i x-x-x-x-x-x D Column temperature Sample loop size Mobile phase HG- (a) HCI+ Kl+ vesicle (b) NaBH Argon carrier flow Analytical line R.f.forward power Refected power Viewing height I CP-A ES- C,,-bonded silica 10 pm particle size 30 "C 100 mm3 10 mmol dm-3 NaH2P04 buffer (pH 5.75)+ 240 x 4 mm i.d. 1 x methanol mol dme3 vesicle of DDAB+0.5% Flow rate 1 cm3 min-' Flow rate 1 cm3 min-I I 10% m/v 0.1% m/v 1 x mol dm-3 DDAB 2% NaBH (stabilized by 0.5% m/v NaOH) Flow rate 1 cm3 min-I 0.70 dm3 mind' 193.69 nm 1 kW < 5 w 15 mm schematically in Fig. 1. Peak heights from the chromato- gram were used in all arsenic quantifications. Experimental conditions finally selected for operation after preliminary investigations are summarized in Table 1. Tap water and urine samples Tap water and urine samples were analysed for arsenic. These samples contained no detectable natural arsenic.They were therefore spiked with the four arsenic com- pounds at different concentration levels. Aliquots ( 100 mm3) of the spiked samples filtered through a nylon 0.45 pm filter were immediately injected into the HPLC system. Results and Discussion Chromatographic Separation The column was previously modified with DDAB as detailed above. The DDAB coating formed proved to be very stable and resistant to water and to the mobile phase passing through. Although organic solvents e.g. methanol could remove the coating the modified column was very durable in the recommended vesicular operation. In fact no significant column behaviour indicating degradation was observed after several months of daily usage. This great advantage of the modified column proposed here versus anion-exchange c01umns~~J~ could be reasonably attributed to continuous stationary phase renewal by the continuous dynamic exchanges of single DDAB molecules between the mobile phase and the modified stationary phase.The pH of the vesicular mobile phase had important effects on the arsenic separation as can be seen in Fig. 2. In the pH range of 5.5-7.0 the two inorganic arsenic species As111 and AsV could always be resolved and separated from the methylated species. Arseneous acid as HAsO (pKa=9.3) which is completely hydrophilic and not ionized in this pH range was eluted with the void volume. On the other hand arsenic acid as H3As0 (pKa=2.3 6.9 11.4) was always negatively charged in the pH range studied; that is the strongest electrostatic interaction was with the positively charged stationary phase and thus was the last to elute.For the two methy- lated arsenic species i.e. (CH,),AsO,H (pKa= 6.3) and CH3As03H2 (pKa=2.6 8.2) their elution sequence could be 5.0 5.5 6.0 6.5 7.0 pH of the mobile phase Fig. 2 Retention times of the arsenic species A As"; B MMAs; C DMAs; and D As"' as a function of the pH of the mobile phase ( I0 mmol dm-3 NaH,P04); experimental conditions as described in Table 1 changed by altering the pH of the mobile phase. At a pH lower than 6 dimethylarsinic acid is mainly in its neutral form and hence it is eluted before monomethylarsonic acid. However at pH higher than 6 it starts to dissociate and so it has a higher apparent negative charge. This increased electrostatic interaction in addition to its natural possible hydrophobic interaction with the stationary phase results in longer retention in the column (now close to that of monomethylarsonic acid).As is schematically shown in Fig. 3 in this vesicle-mediated HPLC separation process arsenic separation is probably achieved by competition of the arsenic species and HzP04- ionic components for interaction sites on the stationary phase and the mobile vesicle pseudophase. The partition coefficients between DDAB vesicles and water for the four arsenic species at the pH 5.75 of the vesicular mobile phase were estimated using an ultrafiltration method. It should be borne in mind that because the stationary phase and the mobile vesicle pseudophase probably possess a similar chemical nature (see Fig.3) the partition of the solute between water and DDAB vesicular pseudophase resembles that between water and the DDAB-modified stationary phase. Results of ultrafiltration experiments for the four arsenic species have been plotted in Fig. 4. These ultrafiltration818 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Stationary phase Fig. 3 Simplified representation of interactions of the solute in the bulk aqueous phase (water) with the stationary phase and the mobile vesicles Aslll DMAs MMAs AsV Fig. 4 Partition estimations as obtained by ultrafiltration of the arsenic species between DDAB vesicles and water results showed that the largest vesicular partition coefficient (ie. minimum arsenic concentration in the filtrate) was obtained for the solute with higher apparent charge AsV.Neutral As1]* species exhibited the smallest partition coeffi- cient. Thus these results also suggested the involvement of ion-exchange or an ion chromatography mechanism in this HPLC separation process. However it does not appear to be a single interaction mechanism. The DMAs and MMAs showed almost the same vesicle-water partition values although they bear completely different apparent charges at pH 5.75. Moreover they can be well separated with this vesicle-mediated HPLC procedure at lower pH (Fig. 2). Obviously these solutes should not only be affected by electrostatic effects but also by hydrophobic interactions. The dependence of the solute retention time on the concentration of DDAB vesicles in the mobile phase was different to that observed in micellar liquid chromato- graphy (MLC).It is commonly assumed in MLC that an increase in the surfactant concentration in the mobile phase results in a decrease in the retention time of the solutes. The magnitude of this decrease will depend upon their partition coefficients between micelles and the stationary phase.22 It should be realized that the surfactant modified C18 station- ary phase and the vesicles allow for electrostatic interac- tions of ionic solutes simultaneously with possible hydro- phobic interactions. The observed retention behaviours in this study suggest that for the four arsenic species studied the partition between the bulk solution in the mobile phase and the modified stationary phase seems to be the dominant mechanism for separation.Possible partition between the Table 2 Effect of DDAB vesicle concentration on the retention times Concentration of DDAB vesicle/ mol dm-3 Retention time/min As111 DMAs MMAs AsV O* 2.82 4.55 5.96 9.80 1 x 10-6 2.83 4.54 5.98 9.82 5x 10-6 2.83 4.55 5.96 9.82 1 x 10-5 2.83 4.55 5.94 9.82 I x 1 0 - 4 2.83 4.56 5.95 9.81 *The column behaviour was not reproducible owing to the continuous DDAB desorption and release from the stationary phase in the column. U 5 10 15 [N aH *PO I/mmol dm” Fig. 5 Effect of buffer concentration on retention times for A AsV; B MMAs; C DMAs; and D As111; experimental conditions as described in Table 1 bulk aqueous solution and the vesicles in the mobile phase does not seem to play a significant role. As shown by the results in Table 2 it follows that the four arsenic species could still be separated even if no vesicle was added to the mobile phase.It should be pointed out however that the column behaviour was by no means reproducible in the absence of vesicles in the mobile phase owing to the continuous DDAB desorption and release to the passing buffered aqueous solution. Further studies to probe the mechanism involved in this vesicle-mediated liquid chro- matography are under way in this laboratory. The effect of NaH,PO buffer concentration in the mobile phase was most significant in the elution of AsV. As expected the retention time of this solute decreased with the increase in the buffer concentration as can be seen in Fig. 5. When the NaH,PO concentration in the mobile phase was below 5 mmol dmq3 the retention of AsV in the column was too long and what is even worse a great peak broadening was observed.A typical chromatogram obtained in the conditions detailed under Procedures is shown in Fig. 6. The four arsenic species separation was accomplished in less than 10 min. The resolution factors (R) between each of these arsenic species [R = 1.172 (t2 - t l ) / ( w1 + w2) where t is the retention time and w is the peak width at half the peak height] were always far greater than 1 which is usually considered acceptable. l 3 HG-ICP-AES Detection In order to obtain efficient HG and transport to the ICP along with minimum peak broadening at the exit of the HPLC column a gas-liquid separator with the smallestJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL.8 819 I 4.5 min I Time - Fig. 6 Typical chromatogram showing arsenic speciation in water A Asi1'; B DMAs; C MMAs; and D AsV. Experimental conditions as described in Table 1. Arsenic content As111 and MMAs 10 ng; DMAs and AsV 20 ng t a) From HPLC 50 CI .- 40 E e 2 20 .= 30 Y ul .- 2 lo Y 0 MMAs ASU' DMAs Fig. 8 Vesicular enhancement of arsenic detection by HG-ICP- AES following HPLC separation. Arsenic content as for Fig. 6 Table 3 Figures of merit of the vesicle-mediated HPLC-HG- ICP-AES method As species RT*/min DLt/ng RSDS (n = 10) (%) As1'' 2.9 0.5 3.6 DMAs 4.5 1 .O 5.0 MMAs 5.8 0.6 3.0 AsV 9.4 1.2 2.0 *Retention time. TDetection limit calculated as 3 times the baseline noise. /Relative standard deviation 10 ng of arsenic for each species was injected. ( b ) Argon FromHPLC 1 -3 To ICP Fig.7 Gas-liquid separators tested (for comparison) (a) U-type; and ( b ) Browner-type possible void volume should be aimed at. In fact most of the HG-based HPLC-ICP-AES couplings described so far utilize a U-tube type gas-liquid separator. The U-type separator depicted in Fig. 7(a) was tested but a laboratory- made glass bead filled separator [Fig. 7(b)] described p r e v i o ~ s l y ~ ~ * ~ ~ provided much better ICP detection charac- teristics than the U-type separator. The analytical signal was higher and the base-line noise was greatly depressed while no significant broadening in the chromatographic peaks was observed (in spite of its relatively bigger size). The enhancing effect of vesicles2' on the ICP on-line signal of the separated arsenic species was verified.As can be seen in Fig. 8 in the presence of DDAB vesicles ( 1 x 1 0-3 mol dm-3) the ICP emission signal increased substantially for all of the four arsenic species under study. The relatively higher concentration of DDAB vesicles needed to produce the satisfactory enhancement in the ICP detection as compared with that for the separation was introduced with the arsine producing reagents after the HPLC separation (see Fig. 1). Unnecessary build-up of DDAB in the HPLC column (resulting in increased pressure drops across the column) from the mobile phase passing through can be easily avoided in this way. Characteristics of the Method The reproducibility of the arsenic separation by this novel chromatographic system was initially evaluated. The col- umn efficiency and the resolution values were studied over a period of 6 months.The column modification procedure was also performed three times (each time the old DDAB coating on the column was removed by passage of a large volume of methanol). No substantial fluctuation in reten- tion times and resolution values was observed from the data collected over these months. The combination of versatil- ity stability efficiency and low cost should make this vesicle-mediated HPLC procedure an attractive choice not only for arsenic separation but also for separations of other anions and possibly neutral species. The precision of the determinations of the arsenic species by this hyphenated method was investigated using ten injections of a mixed arsenic solution containing all of the four arsenic species each present at 100 pg dm-3 each.The relative standard deviations (RSDo/o) of the peak height results calculated for each form of arsenic were always better than 5% (Table 3). Typical calibration graphs for each of the arsenic compounds investigated are shown in Fig. 9. These were obtained by injecting the mixed standard arsenic solutions of different concentrations and plotting peak height for each arsenic species against the amount injected onto the column. As can be seen the four arsenic species showed different sensitivities. This result could be partly due to the different kinetics of the on-line continuous hydride generation of each arsenic compound and also due to the chromatographic peak broadening observed for DMAs and AsV.However the sensitivities would be more similar if peak areas instead of peak heights were used for calibration. The detection limits obtained for each species are shown in Table 3 and were estimated as three times the signal-to-noise ratio and referred to total arsenic injected in water and are seen to be 0.5-1.2 ng of arsenic. Real Samples Arsenic speciation in tap water and human urine by using this hybrid technique was also evaluated. As real samples of820 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 ~~ ~ Table 4 Recoveries of arsenic species added into tap water and human urine samples Experiment 1 Experiment 2 Samples Arsenic Tap water As1i1 DMAs MMAs AsV Human urine As111 DMAs MMAs AsV Spikedlpg dm-3 I50 250 150 250 I50 250 150 250 Recovered* (O/o) 95 98 102 103 96 I05 102 100 Spikedlpg dm-3 50 100 100 100 50 100 50 I00 Recovered* (O/O) 97 I02 100 104 101 I08 99 93 *Mean of two analyses. 60 1 1 0 10 20 30 40 50 Arsenic content/ng Fig.9 Calibration graphs for the four toxic arsenic species under study A As111; B MMAs; C DMAs; and D AsV the local tap water and control human urine available did not contain detectable All1 DMAs MMAs and AsV they were spiked with the four targeted arsenic compounds so as to produce various known arsenic concentrations in the real samples. The spiked samples were filtered through a 0.45 pm nylon filter and analysed immediately. The recovery values (mean of two injections) for all of the four arsenic species are shown in Table 4 and indicate that the proposed method can be used for arsenic speciation in such types of real samples.Conclusion The HPLC separation with HG-ICP-AES detection can be a synergic combination through the use of vesicles as mobile phases for the speciation and determination of As11' DMAs MMAs and AsV. These four arsenic compounds can be separated by using a DDAB vesicular mobile phase on a previously modified C,,-bonded silica column within 10 min. The vesicle-mediated HPLC procedure and column have proved to be fairly robust. Chromatographic behav- iour was most durable and reliable in fact no degradation behaviour was noticed after months of daily usage. On the other hand the sensitivity of the HG-ICP-AES detection of the separated arsenic compounds was enhanced by resort- ing to a post column for further addition of DDAB vesicles with the HG reagents.Moreover with a laboratory-made glass bead filled Browner-typez4 gas-liquid separator a smooth gas-liquid separation and thus reduced plasma noise and increased signals were achieved without signifi- cant chromatographic peak broadening. In brief the vesicle-mediated HPLC-ICP-AES technique proposed here provides a comparatively low-cost efficient and robust separation along with improved ICP perform- ance. Therefore this approach should be able to be extended to the speciation of other elements forming volatile species whose generation can be enhanced by surfactant-based organized media such as Pb Cd and Hg.zo Financial support from Comision Interministerial de Ciencia y Technologia (CICYT) (Project Ref.PB9 1-0669) is gratefully acknowledged. Y. M. L. thanks the Spanish Ministry of Education and Science for a postdoctoral fellowship. References 1 2 3 4 5 6 7 8 9 10 I 1 12 13 14 15 16 17 18 19 20 21 22 23 24 Trace Element Speciation Analytical Methods and Problems ed. Batley G. E. CRC Press Boca Rabon FL 1989. Environmental Analysis Using Chromatography Interfaced With Atomic Spectroscopy eds. Harrison R. M. and Rapso- manikis S. Ellis Horwood Chicester 1989. Chau Y. K. Analyst 1992 117 571. Ebdon L. Hill S. and Ward R. W. Analyst 1987 112 I . Van Loon J. C. and Barefoot R. R. Analyst 1992 112 563. Boom A. W. and Browner R. F. Anal. Chem. 1982 54 1402. Suyani H. Heitkemper D. Creed J. and Caruso J. Appl.Spectrosc. 1989 43 962. Hodgson E. Mailman R. B. and Chambers J. E. Dictionary of Toxicology Macmillan London 1988. Ebdon L. Hill S. Walton A. P. and Ward R. W. Analyst 1988 113 1159. Beauchemin D. Siu K. W. M. Mclaren J. W. and Berman S. S. J. Anal. At. Spectrom. 1989 4 285. Heitkemper D. Creed J. Caruso J. A. and Fricke F. L. J. Anal. At. Spectrom. 1989 4 279. Spall W. D. Lynn J. G. Andersen J. L. Valdez J. G. and Gurley L. R. Anal. Chirn. Acta 1986 58 1340. Sheppard B. S. Caruso J. A. Heitkemper D. T. and Wolnik K. A. Analyst 1992 117 971. Hakala E. and Pyy L. J. Anal. At. Spectrom. 1992 7 191. Rauret G. Rubio R. and Padro A. Fresenius' J. Anal. Chem. I99 1,340 157. Violante N. Petrucci F. La Torre F. and Caroli S. Spectroscopy 1992 7 36. Colon L. A. and Barry E. F. J. High Resolut. Chromatogr. 1991 14 608. Morin Ph. Amran M. B. Lakkis M. D. and Leroy M. J. F. Chromatographia 1992 33 58 1. Chana B. S. and Smith N. J. Anal. Chim. Acta 1992 197 177. Sanz-Medel A. Fernandez de la Campa M. R. Valdes-Hevia M. C. Aizpun Fernandez B. and Liu Y. M. Talanta in the press. Aizpun Fernandez B. Valdes-Hevia M. C. Fernandez de la Campa M. R. Sanz-Medel A. Talanta 1992 39 I 5 17. Ordered Media in Chemical Separations eds. Hinze W. L. and Armstrong D. W. American Chemical Society Washing- ton DC 1987. Menendez Garcia A. Sanchez Uria J. E. and Sanz-Medel A. J. Anal. At. Spectrom. 1989 4 581. Pyen G. S. Long S. and Browner R. F. Appl. Spectrosc. 1986 40 246. Paper 3 / O 1 004 J Received February 19 1993 Accepted April 30 1993