JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 393 Characterization of Automotive Catalysts Using Inductively Coupled Plasma Mass Spectrometry Sample Preparation James A. Brown Jr. Frank W. Kunz and Ronald K. Belitz Ford Motor Company Research Analytical Sciences Department P. 0. Box 2053 Dearborn MI 48121 USA Several methods utilizing microwave energy induced dissolution were developed for rapid digestion and subsequent analysis of automotive catalyst material by inductively coupled plasma mass spectrometry (ICP- MS). Dissolution was accomplished within 2-3 h rather than days as with the widely accepted ashing methods. A comparison was made with several other ICP-MS techniques and the conventional wavelength dispersive X-ray fluorescence method. Keywords Micro wa we energy; acid dissolution; inductively coupled plasma mass spectrometry; automotive catalyst material; wavelength dispersive X-ray fluorescence A catalytic converter an attachment to the exhaust pipe is fitted to almost every new automobile in the USA in order to reduce pollution and to meet the requirements of a 1975 amendment to the Clean Air Act.The platinum group metals some of which are shown in Table 1 are used extensively in these catalytic systems and are often coated onto an alumina-based ceramic monolith or a metal monolith substrate. The over-all performance of these catalytic converters is highly dependent upon several factors namely the physical location within the exhaust system the surface area of the monolith and of most importance the precious metals loading ratio.The economical dispersion and subsequent recovery of these precious metals depends on their accurate determina- tion in the ceramic material at levels ranging from ppb upwards.' Inductively coupled plasma mass spectrometry (ICP-MS) is a fairly new technique for elemental and isotopic analyses. The technique combines the characteristics of the ICP for atomizing and ionizing injected material with the sensitivity of mass spectrometry. There are two unique advantages in the use of ICP-MS firstly the spectra are simple and interferences from inter-elemental effects are largely absent and secondly isotope abundance informa- tion is inherent in the m e t h ~ d . ~ ~ ~ Sample preparation is a critical step in any trace analysis procedure; frequently it is responsible for establishing the lower detection limit of an analysis through its influence on the analytical blank.This is particularly important in ICP- MS as the technique requires that the sample be in liquid form. Wet-digestion methods have gained widespread acceptance however they require close and constant attention by the analyst to avoid unintentional boiling over or evaporation of the sample to dryness. Digestion can often require 48 h to ensure complete dissolution. Thus a more rapid and effective digestion method for trace metal analysis has long been Microwave energy can be used to dissolve solid samples more safely and efficiently than traditional methods (i.e. 2-3 h). Oven-like heating devices are utilized instead of hot-plates for dissolving material for elemental analysis at trace levels i.e. ppm and lower. Microwave technology has two main advantages it is quick and it is relatively contamination free. These attributes are essential for reliable elemental trace analy~is.~-~ The objectives of this study were (i) to develop methods for the dissolution of monolithic ceramic automotive catalysts by microwave energy and subsequently to analyse the material quantitatively for trace elements by ICP-MS; (ii) to compare results from ICP-MS with wavelength dispersive X-ray fluorescence (WDXRF); and (ziz] to com- pare scanning with peak-jumping ICP-MS results. Experimental Reagents Elemental standards were prepared from reagents pur- chased from Spex Industries Edison NJ USA. Standard 1000 pg ml-l stock solutions were diluted in 1% HN03 to 10 pg ml-l working stock solutions which were then used for further appropriate dilutions.Aqueous calibration standards of 20,50 and 100 pg ml-l were prepared from the working stock solutions of precious metals. Table 1 shows the elemental components found in an automotive catalytic converter. The elements listed on the left represent ele- ments used to make the device whereas elements on the right of the table represent possible environmental contami- nants found in spent catalytic converters. Indium (100 pg 1-l) was used as an internal standard. All acids were Ultrex I1 Ultrapure reagents purchased from J. T. Baker Phillipsburg NJ USA. De-ionized water of 18 Mi2 cm-' resistivity was always used and was prepared using equip- ment obtained from Millipore Bedford MA USA.Instrumentation The ICP mass spectrometer used for this work was a VG PlasmaQuad PQ2+ system purchased from VG Instruments Danvers MA USA. Standard instrument parameters for the operating state are listed in Table 2. The WDXRF spectrometer used in this experiment was a fully automated Siemens Model SRS 303 with a ten- position sample changer. Matrix corrections for absorption and enhancement of the element-characteristic X-rays were made using the Artz method.* The operating parameters for the instrument are listed in Table 3. Apparatus All glassware was cleaned with a neutral detergent and rinsed with tap water to remove all of the detergent. It was then soaked in 50% HN03 for at least 24 h rinsed thoroughly with de-ionized water air dried and stored in sealed plastic bags in a closed cupboard.Microwave oven Digestion was performed in an MDS-81D system pur- chased from CEM Matthews NC USA.394 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 Table 1 Elemental components found in an automotive catalytic converter Environmental contaminants in Primary constituents spent devices Element Relative atomic mass Element Relative atomic mass Pt Rh Ce Ni Ba Pd La 195 103 140 58 137 106 139 Fe* 56 Pb 208 S* 32 P* 31 Zn 65 Mn 55 Ca* 40 * Problems for elemental ICP-MS analysis either because of interference or unfavourable ionization potential. Table 2 Standard instrument parameters for the operating state of the PlasmaQuad PQ2+ Parameter Setting ICP system- Forward power 1300 W Plate current 650 mA PA grid current at 1300 W 200 mA Reflected power <I0 w PA ioltage at 1300 W Filament voltage Cool gas flow Auxiliary gas flow Nebulizer gas flow Nebulizer gas pressure Sampling depth Sample cone orifice Skimmer cone orifice Skimmer spacing ring Extraction setting Collector setting L1 setting L2 setting L3 setting L4 setting Front plate setting Pole bias setting Low mass resolution Coarse resolution Fine resolution Cathode current at 0 u Cathode current at 300 u Grid current at 300 u Pulse discriminator Pulse (HT2) multiplier Pulse background Analogue multiplier Analogue offset Ex pan si on Intermediate Anal yser In teduce- Ion lenses- Quadrupole- Detector- vucuum- 4 kV 7.8 V 0.6 1 min-I 0.7 1 .min-' 138 kPa (20 psi) 12 mm 13 1 min-' 1.0 mm 0.7 mm 3.0 mm 1.0 7.7 7.7 5.4 5.0 3.8 8.0 6.0 5.00 3.00 5.00 10.00 mA 75.00 mA 0.0 mA 20.0 7.0 4.5 15.0 Hz 50.0 Hz 200 Pa 1 x Pa 1 x Pa Digestion vessel and capping station The digestion vessels used were made of PFA-Teflon [(perfluoroalkoxy)ethylene] purchased from CEM and fitted with relief valves.The vessels were rotated on a PFA- Teflon turntable during the digestion period to ensure Table 3 X-ray fluorescence operating parameters and sample preparation technique X-ray tube X-ray tube voltage X-ray tube current Sampling Sample size Sample grinding Sample containment Sample press Calibration Number of elements determined Cr 60 kV 50 mA Uses 3 samples (inner middle and 6.0 g 5 rnin ball mix 30 mm Somar No. 330 cap 18 tons in-2 for 30 s Single element standards using substrate material and outlet side of catalyst) 24 uniform heating of each sample.The caps were tightened in a capping station purchased from CEM to provide the uniform tightness required so that the relief valves vented at the same pressure in each vessel viz. 828 kPa (1 20 psi). The technique involves microwave heating of the materials when mixed with concentrated mineral acids (i.e. HN03 HCl and HF) so as to dissolve the material rapidly for suitable application to analysis by ICP-MS. The digestion vessels were cleaned by washing with detergent and rinsing with tap water then heating with 20 ml of concentrated HN03 in the microwave oven for 10 min at 100% power and finally rinsing thoroughly with deionized water. Method 1 Approximately 5.0 g of catalyst sample were ball-mixed in a Spex Model 8000 mixer-mill for 5 min using a Spex Model 8004 tungsten carbide vial-ball set producing a particle size of (400 mesh.This ball-mixed catalyst sample was dried in an oven at 100 "C weighed into a 120 ml PFA-Teflon digestion vessel containing 5 ml of HC1 and 5 ml of HF. The relief valves were placed on the PFA-Teflon vessels which were then sealed in the capping station. All of the samples were digested in the microwave oven for 5 min at 100% power (600 W) and left to stand for 10 rnin without power. They were then twice put through a cycle of cooling to room temperature digesting for 10 rnin at 100% power followed by standing for 10 rnin without power. After that the samples were cooled to room temperature transferred into poly- tetrafluoroethylene or quartz evaporation dishes and 5 ml of HC104 were added to each dish.The evaporation dishes were heated to HC1O4 fume on a medium temperature hot-plate cooled and 5 ml of aqua regia [HN03-HC1 (1 +3)] were added. All of the samples were then diluted to an appropriate volume and assayed by ICP-MS.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 199 1 VOL. 6 395 Table 4 Comparative analysis of automotive catalyst material using WDXRF and ICP-MS techniques. Results given are the mean or mean +standard deviation (in ppm) for n= 18 determinations WDXRF Scanning Peak-jumping Element Laboratory 1 Laboratory 2 ICP-MS ICP-MS IMPd 1917 1 773 5 6 1938f31 1980+21 lo3Rh 373 366 f 7 431 + 14 397 k 5 140Ce 30 528 3 1 793 f 404 35471 +517 31 653f324 58Ni 22713 21 2492230 21 01 15656 23 953 f 385 56Fe 2 670 2 889 f 31 - 5 564 2 332 13'Ba 1951 1062k 13 1003f15 1 1402 15 Table 5 Comparative analysis of automotive catalyst materials using different microwave digestion methods.Results shown are the mean or mean 5 standard deviation (in ppm) for n = 18 determinations WDXRF ICP-MS Element Laboratory 1 Laboratory 2 PA* BAT lMPd 880 857 k 8 799 +_ 55 756k42 Io3Rh 180 185+5 189 5 6 141 + 3 I4OCe 28762 32 116f245 35 532 5401 20410 f 677 58Ni 20894 20275f 116 22 625 + 393 21 221 f 9 4 5 56Fe 2 693 3 132f 17 2 926 f 955 2 190+40 13'Ba 1905 991 f 9 1122+30 788 5 18 * PA Analysis after perchloric acid (HC104) digestion. t BA Analysis after boric acid (H3BO3) digestion. Method 2 Approximately 5.0 g of catalyst sample were ball-mixed and dried as outlined in Method 1.From the 5.0 g sample an aliquot of 0.5 g was transferred into a 120 ml PFA-Teflon digestion vessel and 6 ml of HN03 3 ml of HC1 and 2 ml of HF were carefully added. Relief valves and caps were placed on the vessels and each vessel was sealed in the capping station. The samples were digested for 3 min at 100% power and 15 min at 65Oh power. They were then cooled uncapped within the capping station neutralized with 89 ml ofH3B03 re-capped and digested at 100% power in the microwave oven for 10 min. The samples were cooled filtered and diluted as necessary for assay by ICP-MS. Results and Discussion In order for a method to be acceptable it must be demonstrated that the method produces results with good precision and with no evidence of significant bias.This work shows that the reproducibility of ICP-MS determina- tions is very good for the precious metals however there appear to be problems for some of the other metals especially Ce Ni Fe and Ba (see Tables 4 and 5). Precision is easily assessed but proof that the results are unbiased is not possible as the true result is unknown. It is possible however to provide evidence for the absence of significant bias in well controlled measurement systems which allow indvidual evaluation sources of systematic error.Z Good correlation was obtained between the scanning and peak- jumping ICP-MS results and the WDXRF analysis results (see Table 4). Good correlation was also obtained between the HClO. digestion and WDXRF results as can be seen in Table 5.The peak-jumping method detected Fe whereas scanning did not even after 18 determinations. Although ICP-MS is a sensitive technique for trace and ultra-trace elemental analysis singly charged elemental analyte ions are only measured with the quadrupole mass spectrometer. A large number of elements however pro- duce singly charged elemental analyte ions (M+) many form monooxides (MO+) some form doubly charged ions (M2+) and a few form hydroxide species (MOH+). Other polyions such as ARN+ and Am+ might also be formed. These analyte species are important because serious inter- element interferences can occur owing to spectral overlap.' This might account for discrepancies seen in Tables 4 and 5 in the values for Ce Fe and Ba.Boric acid digestion ICP- MS results (shown in Table 4) are consistently low suggesting that the sample was not completely digested. We believe that this can be corrected by adjusting the digestion time. This will be investigated further in the future. Conclusions The rapid relatively safe acid digestion methods devel- oped in this study are capable of complete dissolution of monolithic ceramic automotive catalytic material. The material is sufficiently digested to allow trace element analysis by ICP-MS. The microwave energy utilized in these procedures can completely digest this material in 2-3 h rather than 1-2 d as for the widely accepted ashing methods. The ICP-MS results especially for precious metals compared favourably with those obtained from WDXRF. The ICP-MS peak-jumping technique appears to be the method of choice as the operation time is halved. Development work is being carried out to produce digestion procedures that are even more efficient. References Westwood L. C. and Lu Y. T. personal communication. Houk R. S. and Thompson J. J. Mass Spectrom. Rev. 1988 7 425. Okamoto K. and Fuwa K. Anal. Chem. 1984,56 1758. Nadkarni R. A. Anal. Chem. 1984,56 2233. Kingston H. M. and Jassie L. B. in Introduction to Microwave Sample Preparation Theory and Practice American Chemical Society Professional Reference Book Washington DC 1989. Kingston H. M. and Jassie L. B. Anal. Chem. 1986 58 2534. Vaughn M. A. and Horlick G. Appl. Spectrosc. 1986 40 434. Artz B. A. X-ray Spectrom. 1977 6 165. Paper 0/04 781 C Received October 23rd. I990 Accepted March 20th. I991