Investigation of Automated Determination of Germanium by Hydride Generation Using in situ Trapping on Stable Coatings in Electrothermal Atomic Absorption Spectrometry* HERMANN 0. HAUG AND LIAO YIPINGt Forschungszentrum Karlsruhe Institut f u r Technische Chemie Postfach 3640 0-76021 Karlsruhe Germany Sequestering and in situ concentration of Ge hydride in the graphite furnace can be automated by using a highly stable trapping reagent to replace the Pd modifier. In a systematic study two groups of trapping reagents which require only a single application i.e. carbide-forming elements (Zr Nb Ta or W) and noble metals (Ir Pd-Ir) were investigated and trapping temperature curves were measured. It was shown that effective trapping of germane is possible on Zr-coated tubes and platforms at trapping temperatures of 550-750 and 600-800 OC respectively. Trapping temperatures should not exceed 650 "C (the 'critical temperature') because at temperatures higher than 650 "C errors in absorbance values could occur owing to an adsorptive 'carry-over effect'.Good signal stability was observed over more than 400 complete trapping and atomization cycles and a precision of better 3% was obtained. Comparatively small signals were observed for the Nb- Ta- and W-coatings. Ir-coated graphite tubes allowed trapping of germane at lower temperatures (400-500 "C) but the signals were small and of low stability compared with those for the Zr coating. Characteristic masses of about 54 pg of Ge on Zr-coated graphite tubes (peak height) and 108 pg of Ge on Zr-coated platforms (integrated absorbance) were observed and the calibration graphs were linear up to 4 ng of Ge on both tubes and platforms.The detection limit was 18 pg of Ge for a 1 ml sample volume using flow injection hydride generation. The method was tested by applying it to the determination of Ge in sediment geological and low-alloy steel certified reference materials. Keywords Germanium hydride; in situ hydride trapping; graphite furnace atomizer; stable trapping reagent; atomic absorption spectrometry The determination of volatile hydride-forming elements by atomic absorption spectrometry (AAS) using hydride gener- ation combined with the graphite furnace as both the hydride trapping and preconcentration medium and the atomization cell has been reviewed by Yan and Ni.' The technique has been described in a number of and can lead to significantly enhanced sensitivities and effectively eliminates interferences in the atomizer.In Fig. 1 an attempt is made to summarize and illustrate the working ranges for the determi- nation of Ge by the application of different AAS detectors including hydride generation; it can be seen that an inherent preconcentration of the analyte can be achieved by the in situ trapping of the hydride. * Presented in part at the Colloquium Analytische t On leave from Peking University Beijing People's Republic Atomspectroskopie Konstanz Germany April 1995. of China. Journal of Analytical Atomic Spectrometry Most of the hydride techniques published have used batch or continuous hydride generation followed by transfer of the gaseous hydrides into the pre-heated graphite tube via a quartz capillary which was manually inserted into the injection h ~ l e .~ - ~ Continuous hydride generator? as well as flow- injection system^^.^ have been used to automate the Ge hydride generation step. The transfer of the hydride into the graphite tube via a quartz capillary mounted on the autosampler arm was reported by Li et al.,' while Tao and Fang' used a separate mechanism for moving and inserting the quartz capillary into the furnace. It is known that Pd on the surface of the graphite tube or platform acts as an efficient adsorber for the volatile hydrides of the elements of Groups IVA-VIA of the Periodic Table leading to an effective deposition of the analyte element on the Pd-treated s ~ r f a c e .~ . ~ - l ~ A catalytic dissociation on the Pd was suggested by Sturgeon et aL8 Pd significantly improved the sensitivity and precision of the in situ trapping of the hydrides in the graphite furnace. Pd-treated graphite tubes and platforms were successfully applied to the technique of in situ trapping for germane (GeH,) by Doidge et ~ l . ~ Zhang et al. Tao and Fang,' and Ma et a1.,6 and in our previous work.2 The application of the Pd solution could be carried out automatically as part of the furnace cycle by the PTFE capillary of the sample dispenser and the quartz capillary for the introduction of the hydride was manually inserted and removed.' Alternatively an automatic insertion and removal of the quartz capillary by the autosampler could be used after applying the Pd solution manually.A fully automated process combining hydride generation plus graphite furnace has been limited by the requirement of a high and constant trapping efficiency over many firings. i l l l l l 7 I / 1 / 1 ' / / I I l i l l l l l I 1 1 1 1 1 1 1 1 I I I Solution flame Solution ETAAS HG-flame t---. HG-ETAAS 0.0 1 0.1 1 10 100 1000 Concentration/ng ml-' Fig. 1 Concentration ranges for the determination of Ge by AAS techniques and detection limits (DL) (calculated for 1 ml of solution) Solution in flame AAS x=DL (ref. 20); solution ETAAS (vol. 20 pl) (ref. 2) x=DL (ref. 21) Pd modifier; HG-flame AAS (vol. 1 ml) x = DL (ref. 15); and HG-ETAAS (vol. 1 ml) x=DL (this work) Zr-coated tube Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1069It appears that the Pd trapping reagent must be applied to the graphite surface before each hydride trapping cycle.Pipetting of the Pd solution by the quartz capillary was disregarded by Shuttler et al." because of the danger of catalytic decomposition of the hydrides by any residual Pd. A second mechanical device for the movement of the quartz capillary as used by Tao and Fang' could be avoided by trapping the hydrides on a more durable coating of the graphite surface. A simple and elegant approach has been the appli- cation of the highly stable mixed Pd-Ir trapping reagent for As Se Bi which had to be applied only once when a new graphite tube or platform was installed." Stable signals for at least 300 complete hydride generation and furnace cycles were reported for Se." For trapping of germane on Ir-Mg- coated platforms ( Perkin-Elmer transversly heated furnace) Schlemmer and Feuerstein12 reported about 80-100 trapping cycles before the Ir-Mg coating had to be re-applied.Ni et ~ 1 . ' ~ and Yan and Nil4 used Zr-coated graphite tubes for trapping Sn and Pb hydrides and found tube lifetimes of 150-200 firings. Insertion of the quartz capillary into the graphite tube and its removal were performed automatically. The Zr-coated graphite tube was preferred by Ni et ~ l . ' ~ 'because the Zr remained on the graphite tube and the enhanced trapping effect for Sn lasted for more than 100 firings'. Ge hydride generation using acetate buffer solutions was reported by Castillo et a1.15 and by Halicz.16 Sodium acetate- acetic acid at pH 4-5 allowed the difficulties with the narrow acidity range for the generation of germane and interferences from foreign ions to be circumvented.The acetate buffer was also used in our previous investigation of the in situ collection of Ge hydride on Pd modifiers. As the reaction medium 0.2moll-' phosphoric acid was used by Jin et al.17 in batch hydride generation and by Nakahara and Wasa18 in continuous hydride generation. Interferences from transition metals were reduced by the addition of 0.1 moll-' thiourea" for the determination of Ge in iron meteorites. Brindle et ~ 1 . ' ~ suggested 0.02 moll HNO + 0.4% L-cysteine for reduction of interferences from transition elements e.g. for Ge in copper and iron samples.In the present work a comparative search was made for trapping reagents for Ge hydride that only needed to be applied once to a new graphite tube or platform. In addition hydride trapping on coatings with carbide-forming elements (Zr Nb Ta W) as well as noble metals (Ir Ir-Mg and Pd-Ir) was studied. EXPERIMENTAL Instrumentation Measurements were performed using a Varian SpectrAA-800 Zeeman atomic absorption spectrometer with a GTA-100 graphite furnace and programmable sample dispenser (Varian Techtron Mulgrave Australia). The Ge wavelength of 265.2 nm was selected from a Ge hollow cathode lamp (Varian) operated at 5 mA. Pyrolytic graphite coated partitioned graphite tubes (Varian Part No. 63-100012-00) and pyrolytic graphite coated graphite tubes with an integrated pyro- lytic graphite platform (Varian Part No.63-100123-90 Schunk Kohlenstofftechnik Germany) were used; henceforth these will be referred to as 'graphite tubes' and 'platforms' respectively. Both peak heights and peak areas were usually recorded. Hydride generation was accomplished using a VGA-76 continuous vapour generation accessory (Varian Techtron) as well as a FIAS-400 flow injection unit and AS-90 autosampler ( Perkin-Elmer Uberlingen Germany) operated by a separate computer. Reagents All chemicals were of analytical-reagent grade or better. De-ionized water was obtained from a Milli-Q System (Millipore Eschborn Germany). A Ge standard solution (Alfa-Johnson Matthey Karlsruhe Germany) with a Ge concentration of 1000 pg ml-' in 5% HNO was diluted with 1 mol 1-' HNO to provide a working stock solution of 10 pg ml-'. Working standard solutions were prepared daily by further dilution e g . in 0.1 mol I-' acetate buffer for the parameter studies.The 0.1 mol 1-' acetate buffer pH 4 was a 5 + 1 mixture of 0.1 mol 1-' acetic acid (Suprapur Merck Darmstadt Germany) and 0.1 mol I-' sodium acetate solution. For analysis of samples with complex matrices 0.02 moll-' HNO +0.4% L-cysteine (Merck) or 0.02 mol I-' phosphoric acid (Suprapur Merck) + 0.4% L-cysteine was used to reduce interferences. Mixed palladium nitrate-magnesium nitrate solution was used as a chemical modifier and was obtained by dilution of appropriate aliquots of a palladium nitrate solution (Merck) and a solution of Suprapur Mg(N0,),.6H20 (Merck) in de-ionized water.The mass applied on both graphite tubes and platforms was 15 pg of Pd+8 pg of Mg(NO,),. A 0.6-1.0% m/v NaBH solution was prepared by dissolving the appropriate amount of NaBH (Merck) in 0.1% NaOH (Suprapur Merck) and was used without filtration. The solutions used for coating with Zr and W were 0.02 moll-' ZrOC1 and 0.02 mol 1-' Na2W04 respectively and were prepared from their salts (Merck). For Nb- and Ta-coating commercial standard solutions of 10 mg ml-' (in 2% HF; Johnson Matthey) were used; for Ir-coating a 1 mg ml-' standard solution (Johnson Matthey) was used. Sample Preparation Sample solutions were prepared from the geological certified reference materials G-2 (granite) and AGV-1 (andesite) and the NIST Standard Reference Material (SRM) 1646 Estuarine Sediment.Microwave-heated high-pressure dissolution (MLS System Buchi Goppingen Germany) was applied 0.500 g of the sample was weighed into the PTFE liner of the pressure vessel and 2.8 ml of concentrated HNO,+0.5 ml of H202+ 1.7 ml of HF were added. The closed vessels were heated in three steps i.e. 5 min at 500 W; cooling and pressure release addition of 1 ml of concentrated HNO,; 4min at 600 W and 1 min without heating; 3 min at 600 W. The solution was diluted to 25 ml with 0.1 moll-' HNO and filtered after allowing it to stand overnight so that any non-silicate residues could settle. Low-alloy steel reference materials (NIST SRMs 361 and 363) were digested with aqua regia by the following procedure 0.200g of the steel sample was weighed into a 100ml PFA calibrated flask and 5 ml of aqua regia (1.2 ml of concentrated HNO + 3.6 ml of concentrated HC1) were added.The flasks were heated in a water-bath to 90°C for about 3 h. The solutions were diluted to 100 ml with 1 moll-' HNO and filtered after the residues had settled. The final solutions for all determinations were obtained by diluting the above sample solutions in 0.02moll-' HNO3-O.4O/0 L-cysteine or 0.02 moll- ' H,PO,-O.4% L-cyst- eine solution and were then used for the hydride generation step. Coating of the Graphite Tubes and Platforms Coating with Zr Nb Ta or W A simple coating method was proposed by Iwamoto et which did not involve soaking the whole graphite tube but rather injecting a 0.01 moll-' sodium tungstate solution into the graphite tube followed by a complete drying/ashing/ 1070 Journal of Analytical Atomic Spectrometry December 1995 Vol.10atomization cycle for liquid samples. We modified the coating procedure by injecting a 50 pl aliquot of the coating reagent solution (for Zr Nb Ta or W) and lop1 aliquots on to a platform respectively. After drying the residual reagent was run through a heating cycle (pyrolysis at 1200 "C for 20 s; atomization at 2500 "C for 3 s) and the procedure was repeated to build-up a certain mass of coating i.e. 2-3 pmol on graphite tubes and 1-2 pmol on platforms. The metal carbides are assumed to be formed during the heating procedure. Coating with Pd-Ir Ir or Ir-Mg According to the procedure of Shuttler et al.," 50pl of a mixture of 0.05% m/v Pd and 0.05% m/v Ir in 0.1 mol I-' nitric acid were injected into the graphite tube or 10 pl on to the platform slowly dried and heated in a reduction step at 1200°C for 20 s and 2000°C for 3 s.The complete injection and heating cycle was repeated so as to obtain a total mass of 50 pg of Pd and 50 pg of Ir on the graphite tube and 20 pg of Pd and 20 pg of Ir on the platform. The analogous procedure using a 0.1 YO m/v Ir solution or 0.1 YO Ir + 0.02% Mg afforded 100 pg of Ir or 100 pg of Ir+20 pg of Mg on the graphite tube and 40 pg of Ir or 40 pg of Ir + 8 pg of Mg on the platform. Hydride Generation Hydride generation was accomplished using two different generators the VGA-76. continuous unit and the FIAS-400 flow injection system. Using the Varian VGA-76 generator a PTFE sample injection valve (six-port with sample loop; Latek Heidelberg Germany) was introduced into the sample channel of the peristaltic pump into which a carrier solution of the required acid concentration (e.g.the acetate buffer solution) was pumped. The calibrated sample loop of 1 ml was filled by a second pump. As is shown in the flow diagram in Fig. 2 the acidic carrier/sample stream merges with the NaBH flow then with an Ar carrier gas flow to strip the liberated hydrides from the reaction mixture as it is flowing through the reaction coil to the gas-liquid separator. A second Ar gas flow was introduced into the separator to avoid condensation in the Sample -1 n n Carrier NaBH4 W -1- - ETAAS II Ar Fig. 2 Basic flow diagram for the sample injection-hydride generator system used for GeH generation and in situ trapping in the graphite furnace Table 1 Ge hydride generation conditions transfer tubing.23 All tubing downstream of the peristaltic pumps was of PTFE.For studying the signal stability on the coated tubes or platforms (over 400 cycles) the VGA-76 generator was run in a continuous mode with the Ge standard added to the carrier solution. In this instance the hydride was introduced into the furnace and trapped for 30 s. In a later stage of this study (during analysis of samples) hydride generation was accomplished using the Perkin-Elmer FIAS-400 flow injection system and AS-90 autosampler which has been described by Welz and Sch~bert-Jacobs,~~ and was controlled by a separate computer. The standard and sample solutions were injected from the automatically filled sample loop (1 ml) into the carrier solution stream at selected intervals. With this unit condensation occurred in the transfer tubing connected to the gas-liquid separator after about 30-50 runs.The hydrides and reaction gases were transferred into the graphite furnace and the hydrides were thermally decomposed on the pre-heated coated graphite surface where the analyte element was collected in situ and concentrated before the atomization step. The tip of a piece of quartz capillary of about 1 mm id (1.8 mm od) which was connected via a suitable length of PTFE tubing (2.0 mm id) to the outlet of the gas- liquid separator was inserted through the sample hole of the graphite tube and was held in close contact with the opposite wall or platform surface respectively. The end of the quartz tube had two bevels to both sides of the tip to provide an unobstructed outlet for the gases when placed in the graphite tube.At trapping temperatures >700"C a very small H2 flame was observed at the hot tip of the quartz capillary when it was lifted out of the graphite tube which extinguished after a few seconds. The advantage of a closed low-volume capillary system and gas-liquid separator ( 1 ml) compared with some batch hydride designs" is that entrained air is flushed out within a few seconds even during start-up. Progress with respect to practical use was the fixation of the platform within the graphite tube in the form of the 'forked platform' or the similar 'integrated platform' now available.A manually introduced quartz capillary used for some of the initial measurements could be simply rested on the platform in its fixed position in the same way as on the tube wall. An important step towards automation was the hydride transfer into the graphite tube uia the quartz capillary mounted in place of the PTFE capillary tubing of the sample dispenser arm. Changes in the AAS software allowed controlled hot-injection of the gaseous hydrides. The Ge hydride generation parameters were adjusted to a flow rate of the carrier solution of about 3-4mlmin-' and the NaBH concentration was varied between 0.5 and 1% at a flow rate of 1 mlmin-' for the VGA-76 unit and up to 2.5 ml min-' for the FIAS-400 unit. The aim was to achieve a low hydrogen production rate so as to increase the residence time of the hydrides in the furnace and a small back-pressure in the gas-liquid separator.The routinely selected conditions are summarized in Table 1. Solution For parameter studies (VGA-76) Carrier Reductant Total Ar gas flow Carrier For analysis of sediment and geological samples (FIAS-400) Reductant Ar gas flow Concentration Flow rate/ml min - ' 0.1 moll-' acetate buffer (pH 4) 0.6% NaBH in 0.1% NaOH 3 1 90 0.1 mol 1-' HN0,-0.4% L-cysteine or 0.2 mol 1-' 4 2.5 H,PO,-0.4% L-cysteine 1 Yo NaBH in 0.1 YO NaOH 50 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1071Procedure for in situ Trapping Before the first hydride generation cycle the graphite tube or platform was coated with the selected adsorber for hydride trapping as described above.The sequence of operations describing the hydride generation collection and atomization is as follows. The sample loop was filled with the sample (or standard) solution after which the hydride generator was started according to the conditions given in Table 1 as was the automatic furnace cycle with the temperature programme given in Table 2. After the graphite tube reached the selected trapping temperature (e.g. 600 "C) the quartz capillary was inserted automatically and the sample injection valve was then switched for the sample volume being carried into the generator by the acidic carrier solution. After throughput of the sample volume the reaction of the acidic carrier and reductant flow was continued for an additional 30 s to ensure the complete flushing of the reaction gases into the graphite furnace.The quartz capillary was removed automatically from the furnace at the end of the collection time and the furnace programme was continued with an Ar flush of the graphite tube followed by the atomization step (with gas stop) and a final cleaning step. a 0.2 c 5 2 0.1 a RESULTS AND DISCUSSION Use of Pd-treated Graphite Surfaces for in situ Trapping of the H ydrides In the first series of experiments the hydride generation and trapping cycle was started with injection of Pd-magnesium nitrate solution into the graphite tube or on to the platform which was then dried and thermally pre-treated at the hydride collection temperature. This application of Pd was performed before each hydride trapping/atomization cycle because the Pd was evaporated at the high atomization and cleaning temperatures of each cycle.In this work the trapping of Ge as its hydride on the Pd-Mg pre-treated graphite tubes or platforms is considered as the reference case when compared with trapping on other reagent-coated graphite tubes or platforms. An efficient collec- tion of Ge on the Pd-Mg pre-treated surface of the pyrolytic graphite tube or platform was obtained over a wide range of trapping temperatures from below 200°C to higher than 800°C. More detailed results on the Pd-Mg applications will be discussed in the context of the other investigations. - - Test of Long-term Stability of Zr- Nb- Ta- W- Ir- Ir-Mg- and Pd-Ir-coated Graphite Tubes and Platforms for in situ Trapping of Ge Hydride Pre-treatment of the graphite surface with metals that form thermally stable carbides has been proposed to overcome difficulties in the determination of Ge because of losses caused by the formation of volatile compounds (with 0 S Cl etc.) during thermal pre-treatment steps.Coating with elements such as Zr ( platforms") W ( t ~ b e s ; ~ ~ ~ ~ ~ plat- f o r m ~ ~ ~ ) Mo ( t ~ b e s ; ~ ' ~ ~ platforms33) and Ta34-36 improved the sensitivity and precision for Ge in aqueous and organic extractant solutions. Ma et aL6 reported the trapping of germane using La- Ta- W- and Zr-coated normal graphite tubes compared with pyrolytic carbon coated (pyrC) graphite tubes. These workers used Pd modifiers on the coated surfaces and obtained a good response on Ta- W- and Zr-coated tubes but even higher sensitivity on the pyrC tubes.Trapping temperature curves In a second series of experiments the trapping of the Ge hydride was studied on new pyrolytic graphite tubes as well as on platforms that had only been coated once with Zr Nb Ta W Ir Ir-Mg or Pd-Ir according to the coating procedures described above. The influence of the deposition temperatures for in situ collection of germane on the coated tube wall is shown in Fig. 3 and for sequestering on the coated platforms in Fig. 4. Each set of trapping temperature curves is compared with the trapping achieved with a pre-injected Pd-Mg modifier which was applied before each trapping cycle. The Pd modifier treated surfaces obviously resulted in the highest response for trapping of germane. Trapping and atomization from the coated tube walls (Fig.3) showed relatively good efficiency for Zr between 550 and 750 "C (atomization temperature 2500-2600 "C); the signals decreased below 550°C. With Nb- Ta- and W-coated tubes only a very low response was observed. Typical signals are shown in Fig. 5. At trapping temperatures higher than 800"C the signal stability deteriorated. Trapping temperature curves on coated platforms are sum- marized in Fig. 4; the absorbance signals from the Zr-coated platform reached a maximum at trapping temperatures between 700 and 800°C and the efficiency decreased rapidly 0.3 I 10.9 0.6 a V z 3 0.3 2 0 200 300 400 500 600 700 800 900 Trapping temperaturePC Fig. 3 Trapping temperature curves for Ge hydride (2 ng Ge) on graphite tubes coated with 3 Ir-Mg; 4 Ir; 5 Nb; 6 Zr; 7 W; compared with graphite tubes with 2 pre-injected Pd-Mg modifier or 1 without any coating or modifier.Peak height mode Table 2 Graphite furnace parameters for Ge hydride trapping and atomization on the coated surfaces used in subsequent parts of this work Temperature/"C Ir- (or Pd-Ir-) coating Zr- (or Nb- Ta- W-) coating Step tube/platform tube/platform Ramp/s Hold time/s Ar sheath gas/l min-' 1 Hydride introduction 400/500 600/700 5 50* 31- 2 Ar flush 400/500 600/700 10 3 3 Gas stop 400/500 600/700 2 0 4 Atomization 2200/2400 2 5OO/2500 1 2-4 0 5 Cleaning 2200/2400 2550/2600 - - - 2 3 * Depending on the volume of the sample loop (e.g. 1 ml). 7 Proprietory adjusted for hot-inject mode; total gas flow through the furnace.1072 Journal of Analytical Atomic Spectrometry December 1995 Vol. 100.20 8 0.16 0 m e 0.12 5 n 2 0.08 2 CJ c 0.04 - 0 /x 3 1 I I I 4 * 100 200 300 400 500 600 700 800 900 Trapping temperature/"C Fig. 4 Trapping temperature curves for Ge hydride (2 ng Ge) on pyrolytic graphite platforms coated with 3 Ir-Mg; 4 Ir; 5 Zr; 6 W compared with 2 pyrolytic graphite platforms with pre-injected Pd-Mg modifier; or 1 without any coating or modifier. Peak area mode 0 1 Time/s 2 Fig. 5 Signal graphs obtained after in situ trapping of Ge hydride (2 ng Ge) and atomization on graphite tubes coated with 1 Zr (6OO/26OO "C trapping/atomization temperatures) and 2 Nb (7OO/25OO "C) 0 1 2 Timels 3 Fig. 6 Signals obtained after in situ trapping of Ge hydride (2 ng Ge) and atomization on platforms coated with 1 Zr (700/26oO "C trapping/ atomization temperatures) and 2 W (600/2500 "C) below 600°C.The Ge signals (Fig. 6) from the Zr platform showed a small shoulder. Less effective GeH trapping on W- Nb- and Ta-coated platforms was observed in the range between 600 and 750"C with a double peak for the Ta-coated platform. The temperature curves for graphite tubes with Ir Ir-Mg or Pd-Ir coatings show that Ge trapping occurs at lower temperatures compared with the Zr coating. However the signals (Fig. 7) were small and showed considerable tailing and the absorbance decreased after a relatively small number 0.2 Q C e 5 s 0 0.9 Q C e 5 2 0 0 1 2 3 Time/s Fig. 7 Signals obtained after in situ trapping of Ge hydride (2 ng Ge) and atomization on graphite tubes coated with 1 Pd-Mg (500/2600 "C trapping/atomization temperatures) and 2 Ir (5OO/22OO "C) of firings eg.during recording of the trapping temperature curve (even at atomization temperatures below the limiting value of 2400 "C). Very small Ge signals with considerable tailing were observed for the trapping of germane on Ir- Ir-Mg- and Pd-Ir-coated platforms. The signals were again not stable for more than a small number of firings. The carbide-forming elements were found to shift the Ge absorption peaks towards shorter atomization times where- as the noble metal coatings shifted the signals to longer atomization times with the longest delay by the Pd modifier. Hydride generation and in situ trapping of Ge hydride was most effective with Zr-coated tubes or platforms with respect to fully automatic sample processing.This coating was investi- gated further for long-term stability and suitability for analysis of samples. Signal stability on Zr-coated surfaces After only a single coating with the trapping reagent the reproducibility of the absorbance and integrated absorbance signals of Ge was tested in 400 complete in situ trapping/ atomization cycles on the coated graphite tubes and platforms the hydride being generated in a continuous mode (from a 2 ng ml-' solution) and trapped for 30 s (equivalent to a mass of 2.8 ng of Ge). In Fig. 8 the stability of the Ge absorbance signal and the precision is shown for hydride trapping on a Zr-coated tube (3 pmol Zr; 750/2600 "C trapping/atomization temperature).The curves indicate that the stability of this coating as a trapping reagent was sufficient for at least 400 complete hydride trapping and atomization cycles. The precision of the absorbance signals (n = lo) showing some variations is better than 3%. 0.20 0.15 Q 0 2 0.10 5 n 0.05 0 -12 ~ 19- h I E 100 200 300 400 No. of firings Fig. 8 Signal repeatability and precision of 400 complete cycles of Ge hydride (2.8 ng Ge) in situ trapping and atomization on graphite tubes coated with 3 pmol Zr as trapping reagent (750/26oO "C trapping/ atomization temperature) Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1073The stability of the Ge signal was also followed for the in situ trapping of germane on a Zr-coated platform over 400 firings and a precision of better than 4% was obtained for the integrated absorbance signals.According to our experience during the subsequent analysis of samples some changes in sensitivity (of 10-20%) occurred with some Zr-coated plat- forms between about 200 and 300 firings and a more frequent reslope of the calibration graph beyond 180 firings is therefore recommended. Calibration graphs and carry-over efSect Calibration graphs were established for in situ trapping of GeH on Zr-coated graphite tubes and platforms. There was only a very small adsorptive ‘carry-over effect’ for Ge hydride collection compared with the ‘carry-over effect’ that was observed at higher trapping temperatures in our with Sn hydride trapping. By ‘carry-over effect’ it is understood that a very small fraction of the Ge is deposited at the tip of the quartz capillary and that this small fraction is carried over to the next measurement of Ge sample or blank.The ‘carry-over effect’ for Ge trapping is very small or negligible at trapping temperatures of 600-650 “C and increases at higher tempera- tures. Thus Ge trapping should not exceed the ‘critical tem- p e r a t ~ r e ’ ~ ~ of 650-680 “C. The calibration graphs for Ge were linear up to about 4 ng of Ge on both the Zr-coated tubes (peak height) and Zr-coated platforms (peak area). Analytical Figures of Merit Trapping efJiciency The efficiency of the in situ trapping on the coated surfaces was compared with that obtained by direct injection of an aqueous Ge standard solution on the same coated surface. Trapping efficiencies for germane (based on peak area) on the Pd-Mg (modifier) treated tube and platform of about 70 and 87% respectively were observed. On the Zr-coated surfaces the trapping efficiencies were 61 and 60% on the tube and platform respectively.Characteristic mass The characteristic mass m as determined from the slopes of the lower linear part of the calibration graphs is summarized in Table 3. A comparison of the sensitivity suggests that the atomization of Ge from the Zr-coated tubes or platforms is considerably lower than from the Pd modifier treated surfaces. A characteristic mass (based on peak height at A = 0.0044) of 12 pg of Ge using a Pd modifier on graphite tubes was reported by Doidge et aL3 and of 16 pg of Ge using a Pd-Mg modifier in our previous work.’ Using platforms Tao and Fang’ found a characteristic mass (at Aint = 0.0044 s) of 49 pg of Ge with a Pd modifier while a value of 38 pg of Ge was reported in our previous paper’ with a Pd-Mg modifier.Table3 Characteristic mass m0 for Ge hydride trapping on Zr-coated graphite tubes and platforms Graphite tube Zr-coated Pd-Mg modifier Pd modifier Zr-coated Pd-Mg modifier Pd modifier Platform m0 m o (peak height)/pg (peak area)/pg Ref. 54 170 This work 16 55 2 3 12 - 32 110 This work 10 38 2 49 5 - However in terms of the automated in situ collection of germane in electrothermal AAS the Zr-coated tubes or plat- forms can both be recommended because of the good long- term trapping stability of the single coating which is effective for > 400 firings. Detection limits and precision Calculated detection limits for in situ trapping of Ge hydride generated from acetate buffer solution based on the relative standard deviation of the reagent blank (3s) and a sample volume of 1 ml were 9 pg of Ge on a Zr-coated tube (peak height) and 18 pg of Ge on a Zr-coated platform (peak area).The precision of the determination using a Zr-coated graphite tube is better than 2% for determinations at the 1 pg I-’ level. The reagent blank was very small and is estimated to contain 0.1 ng of Ge using a 1 ml sample volume. Determination of Ge in Geological and Sediment Samples There is a lack of reference materials with a certified Ge content. The geological reference standards G-2 (granite) and AGV-1 (andesite) as well as the NIST SRM 1646 Estuarine Sediment were analysed for their Ge content using both the 0.02 mol 1-1 HNO3-0.4% L-cysteine and the 0.02 mol 1-1 H3P04-0.4% L-cysteine solutions as reaction medium for the same sample solution.The results of the sample analyses using standard additions are summarized in Table 4. It is interesting that the Ge concentration values obtained for the geological and sediment samples using the 0.02 moll-’ HNO,-0.4% L- cysteine solution were higher than with the 0.02 mol I-’ H3P04-0.4% L-cysteine solution as reaction medium. Com- paring standard additions with the slopes of the calibration graphs there was no significant interference from the sample matrix and a Ge spike added before the dissolution procedure was recovered quantitatively.The difference between the two media was an additive error which could not be corrected by standard additions. The Ge concentration values obtained for the AGV-1 and G-2 standards are lower than the ‘recommended concentration’ values in the compilation of geostandards by Gladney et However our values show better agreement with the ‘method mean’ values measured by atomic absorption listed in the same compilati~n.~~ Determination of Ge in Steel Samples Steel samples in the form of the NIST SRMs 361 and 363 were digested with aqua regia. After dilution the concentration of Ge in the final solutions was in the range 0.8-1.2 ng ml-’ which resulted in a peak absorbance A of about 0.2. The measured Ge concentration values are also listed in Table4.The Ge values from the hydride generation from 0.02 moll- HN03-0.4% L-cysteine and 0.02 moll-’ H,PO,-0.4% L-cyst- eine media show satisfactory precision but are low compared with the non-certified values of Ge in the steel SRMs. The Ge content found for SRM 361 was confirmed by an indepen- dent determination using inductively coupled plasma mass spectrometry. Further investigation of the factors influencing the Ge hydride generation process is required and additional optimiz- ation for complex matrices such as minerals and alloys is necessary. CONCLUSION A single addition of Zr solution as trapping reagent and thermal pre-treatment of the carbide-forming element on the pyrolytic graphite surface allows the efficient in situ trapping 1074 Journal of Analytical Atomic Spectrometry December 1995 Vol.10Table 4 Determination of Ge in geological sediment and low alloy steel certified reference materials by standard additions using flow injection hydride generation in situ collection on Zr-coated graphite tubes. (All values in pg g-’). Values in parentheses indicate number of determinations Material Geological samples AGV- 1 via peak height via peak area via peak height via peak area NIST SRM 1646 via peak height via peak area NIST SRM 3619 via peak height via peak area NIST SRM 363 via peak height via peak area G-2 Sediment sample Low-alloy steels Found Hydride generation using 0.02 moll-’ HNO + 0.4% L-cysteine 1.23 & 0.02 (4) 1.17f0.03 (4) 1.01 kO.01 (4) 1.01 kO.01 (4) 1.9f0.2 (2) 1.9f0.1 (2) 35k2 (6) 34k1 (6) 26k1 (2) 25$1 (2) 0.02 moll - ’ H3P04 t 0.4% L-cysteine 1.07rt0.01 (2) 1.03 & 0.01 (2) 0.97 f 0.01 (4) 0.98 f 0.01 (4) 1.6k0.1 (2) 1.6k0.2 (2) 35f1 (2) 3 5 f l (2) 3 1 f l (2) 2 8 f l (2) Recommended concentration3’ 1.25f0.13* 1.14 k 0.15* AA ‘mean valuey3’ 1.06t 0.95-t * ‘Recommended concentration’ values in ref.38. t Atomic absorption ‘method mean’ values in ref. 38. $ ‘Non-certified‘ concentration value by NIST. 9 A value of 36 kg g-’ of Ge was obtained for this SRM by ICP-MS. of Ge hydride (with subsequent atomization) for more than 400 complete trapping/atomization cycles. Instrument and software requirements for automating the hydride generation and trapping of the gaseous hydrides in the graphite furnace are simplified by the application of such a long-lasting trapping reagent for coating of a new graphite surface and repetitions of such coatings during the lifetime of the graphite tube or platform do not seem to be necessary.Zr-coated graphite tubes as well as Zr-coated platforms can be recommended for the effective in situ trapping and concen- tration of Ge hydride because of the reasonably high sensitivity and good long-term stability of the absorbance signals. The trapping of germane on surfaces coated with other carbide- forming elements or with noble metals (Ir and Pd-Ir) was very poor and the Ge signals decreased after a relatively small number of firings. The coating from a carbide-forming element allows higher atomization and clean-out temperatures to be used. The technique of Ge hydride generation and trapping in the graphite furnace provides very low limits of detection of Ge because reagent blanks are usually very low for this element.A wide range of concentrations can be analysed by employ- ing different volumes of the sample loop (and different Ge wavelengths). Further investigation of the factors influencing the hydride generation process which was beyond the scope of this work would be required to enable Ge hydride generation to be used for samples with complex matrices such as minerals and alloys. The authors thank Varian Techtron Australia for providing the software modifications for sample introduction at relatively high temperatures. REFERENCES 1 2 3 Yan X.-p. and Ni Z.-m. Anal. Chim. Acta 1994 291 89. Haug H. O. and Ju C. J. Anal. At. Spectrom. 1990 5 215. Doidge P.S. Sturman B. T. and Rettberg T. M. J. Anal. At. Spectrom. 1989 4 251. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Zhang L. Ni Z.-m. and Shan X.-q. Spectrochim. Acta Part B 1989 44 751. Tao G. and Fang Z. J. Anal. At. Spectrom. 1993 8 577. Ma Y. Han Y. and Ariguli R. Fenxi Huaxue 1994 22 586. Li S. Wang R. Ni Z.-m. Zhang L. Yan X. He B. Shan X. Jing S. and Han H. Guangpuxue Yu Guangpu Fenxi 1992 12 (4) 117. Sturgeon R. E. Willie S. N. Sproule G. I. and Berman P. T. Spectrochim. Acta Part B 1989 44 667. Zhang L. Ni Z.-m. and Shan X.-q. Spectrochim. Acta Part B 1989 44 339. Sturgeon R. E. Willie S. N. and Berman S. S. Anal. Chem. 1987 59 2441. Shuttler I. L. Feuerstein M. and Schlemmer G. J. Anal. At. Spectrom. 1992 7 1299.Schlemmer G. and Feuerstein M. paper presented at the Colloquium Analytische Atomspectroskopie CANAS’95 Konstanz Germany April 1995. Ni Z.-m. Hang H.-b. Li A. He B. and Xu F.-z. J. Anal. At. Spectrom. 1991 6 385. Yan X.-p. and Ni Z.-m. J . Anal. At. Spectrom. 1991 6 483. Castillo J. R. Lanaja J. and Aznarez J. Analyst 1982 107 89. Halicz L. Analyst 1985 110 943. Jin K. Terada H. and Taga M. Bull. Chem. SOC. Jpn. 1981 54 2934. Nakahara T. and Wasa T. Microchem. J. 1994 49 202. Brindle I. D. Le X.-c. and Li X.-f. J . Anal. At. Spectrom. 1989 4 227. Welz B. Atomic Absorption Spectrometry VCH Verlag Weinheim 2nd edn. 1985 pp. 284 and 285. Varian Manufacturer’s Literature Varian Australia 1994. Wamoto E. Shimazu H. Yokota K. and Kumamaru T. J. Anal. At. Spectrom. 1992 7 421. Sturman B. T. Appl. Spectrosc. 1985 39 48. Welz B. and Schubert-Jacobs M. At. Spectrosc. 1991 12 91. Zheng Y. and Zhang D. Anal. Chem. 1992 64 1656. Kolb A. Mueller-Vogt G. and Stoebel W. Spectrochim. Acta Part B 1987 42 951. Gao Y.-q. and Ni Z.-m. Huaxue Xuebao 1982,40 1021. Gao Y.-q. and Ni Z.-m. Xiyou Jinshu 1982 1 57. Zheng Y. and Zhang D. Fenxi Ceshi Tongbao 1988 7(6) 41. Benzo Z. Cecarelli C. Carrion N. Alvarez M. A. Rojas C. and Rosso M. J. Anal. At. Spectrom. 1992 7 1273. Bao C. Cheng X. Li Y. Wei Y. and An Z. Guangpuyue Yu Guangpu Fenxi 1991 11(5) 56. Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 107532 Bao C. Cheng X. Liu C. and Wei Y. Fenxi Huaxue 1992 20 429. in the press. 33 Criaud A. and Fouillac C. Anal. Chim. Acta 1985 167 257. 34 Mino Y. Shimomura S. and Ota N. Anal. Chim. Acta 1979 37 Haug H. O. and Liao Y. Spectrochim. Acta Part B 1995 50 38 Gladney I!. S. Jones E. A. Nickell E. J. Roelandts I. Geostand. Newsl. 1992 16 111. 107 253. 35 Hoquellet P. and Lebeyrie N. Analusis 1975 3 505. 36 Sohrin Y. Isshiki K. Kuwamoto T. and Nakayama E. Talanta Paper 5104463 D Received July 10 1995 1987 34 341. Accepted August 22 1995 1076 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10