JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 839 Analysis of Glasses From the V,O,-As,O,-BaO System Using Inductively Coupled Plasma Atomic Emission Spectrometry* S. Del Barrio and R. Benito lnstituto Tecnologico Geominero de Espafia Rios Rosas 23 Madrid 28003 Spain F. J. Valle lnstituto de Ceramica y Vidrio (CSIC) 28500 Arganda del Rey Madrid Spain Glasses from the V20,-As,O,-BaO phase equilibrium system are interesting for applications in the production of thermistors and electron multiplier channels. The high volatilization rates of arsenic during the melting process make it necessary to carry out chemical analysis of the glasses in order to locate the final composition in the glass-formation area in the corresponding phase equilibrium system. Inductively coupled plasma atomic emission spectrometry is proposed for determining V205 As~O and BaO.A sodium hydroxide procedure was used to separate the main components from the impurities. The only spectral interference produced is due to the V I emission at 193.682 nm which overlapped partly with 193.698 nm of As 1. The analyses were carried out by using an internal standard Sc II at 361.384 nm. The determination of analytical short-term precision and precision of the method was also carried out. The relative standard deviations were 0.53,1.19 and 0.80 for V205 As,O and BaO respectively. The results obtained were compared with those obtained using gravimetry for As and Ba and spectrophotometry for V. Keywords Inductively coupled plasma atomic emission spectrometry; glasses; determination of V,O As203 and BaO The As203 introduced into a glass chemical composition can act either as a refining agent or network former,' and is present in the compositions of the V20,-As203-BaO phase equilibrium system for the latter purpose.A research project developed at the Spanish Instituto de Cerdmica y Vidrio (CSIC) since 1978 has been to determine precisely the glass building area within the above mentioned system (Fig. 1) and to study the future application of such materials to the production of thermistors electron photomultiplier channels and low-valency vanadium oxides2 During the research the necessity of knowing the composition of glasses in order to predict their properties has arisen. The percentages of the components expressed as oxides shift significantly from the theoretical values because during the fusion stage volatilization takes place that is difficult to control particularly for As203 and less appreciably for V205.Thermogravimetrically arsenic behaves in different ways A Glass building p area Mol(%) Fig. 1 Tentative glassy state area in the V20,-Asz0,-Ba0 system; 0 glass and e devitrified glass ~ ~ ~~~~~~~~ ~ ~~ *Presented at the 1993 Winter Conference on Plasma Spectro- chemistry Granada Spain January 10- 1 5 1993. 1 I I 1 0 2 0 0 1 0 0 6 0 0 8 0 0 VC Fig. 2 Thermogravimetric curves for A As203 and B M-5 glass (69% V2O5-23% A~~O3-896 BaO). The total volatilization of the As203 sample (0.17023 g) cannot be recorded as a single line because of the loss-in-mass scale used (100 mg).At 405 "C the top of the scale is reached depending on whether it is present as raw material (As~O~ As2S3 As2S5 etc.) or if it is inside the glass network. In the thermogravimetric plots presented in Fig. 2 both behav- iours can be observed. Line A corresponds to As~O~; a sample with 0.1 7023 g of As203 is fully volatilized at 600 "C with a heating rate of 6 "C min-I. Line B corresponds to a V205-As203-BaO glass containing 24.3 mol-Oh As203; a 0.14389 g sample of this material under the same heating conditions starts to volatilize only at temperatures >750 "C. This behaviour has great importance in the chemical analysis of V20S-As203-BaO glasses because it enables its disgregation with melting reagents when the working temperature is less than 750 "C e.g.using sodium hydroxide as a fusion reagent (attack temperature between 400 and 500 0C).3 The aqueous solution of the attacked sample contains the major components (V as V03- or V0.,H2 As as As2- or As03- Ba as Ba(OH) soluble) and840 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 Table 1 Specification of the ICP and operating conditions Plasma source FrequencyIMHz Induced power/W Reflected power/W Argon plasma gas flow rate/l min-' Argon aerosol camer gas flow ratell min-l Cooling water for coils/l min-I Sample introduction Sample delivery Measurement time Observation heigh t/mm Analytical lines/nm v I1 As I Ba I1 s c I1 27.12 MHz 1000 t 5 17 0.75 1.5 Pneumatic cross-flow nebulizer Gilson Miniplus 11 peristaltic pump 2 ml min-I 10 integrations of 5 stages per step 16 292.402 193.696 4 9 3 -409 36 1.384 the impurities (Al Fe Mn Ti Cr etc.) are separated as precipitated hydroxides or basic salts in the alkaline medium.The objectives of this work are to use inductively coupled plasma atomic emission spectrometry (ICP-AES) for the determination of the macroconstituents of glasses belonging to the V20,-As203-Ba0 system after sample fusion with sodium hydroxide; and to compare the results with those obtained using other well-known techniques for these materials (As gravimetry as Mg2As20,; Ba gravi- metry as BaSO,; V spectrophotometry as VOZ3+). Experimental Instrumentation A Jarrell-Ash Model ICAP-6 1 multichannel spectrometer with an ICP source was used equipped with a polychroma- tor (1 5 10 grooves mm-l and 32 photomultiplier tubes).The specifications of the ICP source and operating conditions are given in Table 1. Reagents Standard stock solutions for V As Ba and Y were used (Inorganic Venture). Sodium hydroxide (NaOH-H,O Suprapur Merck) was used for the sample fusion. High- purity HN03 (65%) and de-ionized water ( > l 8 mQ cm) were used for all preparations of standard and sample solutions. Sample Preparation and Procedure All theoretical compositions of glasses used in the present work are reported in Table 2. The sample prepared for analysis was ground in a tungsten carbide vibrating mortar to a particle size of less than 63 pm. A 0.2000 g of the ground sample was placed in a zirconium crucible and mixed with 0.50 g of sodium hydroxide (carbonate free).The mix was covered with a further 0.50 g of sodium hydroxide and transferred to a muffle furnace. The temperature was gradually raised until any water present was driven off and then held at 400 "C for 15 min until complete fusion occurred. The crucible was removed from the furnace and allowed to cool and 20 ml of de-ionized water were added. The crucible was transferred to a steam-bath to facilitate lixiviation of the melt which was filtered through a No 58g3 Scheleicher blue ribbon paper. The precipitate (impurities) was transferred to the filter with a jet of hot water and washed thoroughly with hot water. The filtrate was completely transferred into a 200 rnl calibrated flask acidified with 5 ml of HN03 (65%) and diluted to volume with de-ionized water.Standard Preparation and Calibration Five multi-element standards of 100 50 25 10 and 5 pg ml-I of V and 20 10 5 2 and 1 pg ~ m - ~ of As and Ba were prepared from 1 mg ml-l stock solutions and the blank of the reagents. These concentration ranges cover all three element contents in the analysed materials after diluting the original solutions 1 +9 and 1 + 19. Scandium (50 pg ml-l) was added to all diluted sample and standard solutions as an internal standard. - ~~~ ~ ~ ~ Table 2 Composition of selected glasses Theorical composition Glass sample Oxide M-2 v205 As203 BaO M-3 v2os As203 BaO M-4 v 2 0 5 A m 3 M-5 v 2 0 5 M-6 v 2 0 5 A N 3 BaO AS203 BaO BaO Mol (Oh) 85.0 8.5 6.5 70.5 12.0 17.5 80.0 12.0 8.0 69.0 23.0 8.0 60.0 28.0 12.0 Mass (%) 84.50 12.98 2.5 1 71.71 13.27 15.0 80.16 13.08 6.75 68.48 24.82 6.69 59.65 30.28 10.05 Table 3 Precision for V As and Ba with regard to sample dilution Glass sample Element Dilution M-2 V 1+9 As 1 +9 Ba 1 +9 M-6 V 1 +9 As 1 +9 Ba 1+9 1+19 1+19 1+19 1+19 1+19 1+19 RSD 0.18 0.30 0.32 0.49 0.18 0.27 0.23 0.29 0.36 0.54 0.20 0.27JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 84 1 ~ ~~ Table 4 Analytical results for macroconstituents in glasses of the V205-As203-Ba0 system Mass* (O/O) Spectrophotometry Gravimetry Gravimetry Glass Oxide ICP-AES V023+(H202) As207Mg2 BaSO 85.10 k0.35 85.61 k 0.80 - - - 9.95 k 0.26 - 10.17 k 0.14 71.35k0.39 71.60k0.85 - - - 10.87+0.30 - 11.04k0.14 M-2 v2°5 BaO 4.2 I k 0.05 - - 4.59k 0.19 As205 M-3 VZOS As205 BaO 16.90t0.10 - - 17.2 1 t0.28 M-4 v2os M-5 VZ05 AS205 BaO As205 BaO M-6 v 2 0 5 AS205 BaO 80.73 k 0.35 11.10k0.17 7.98 k0.06 69.87 t 0.40 20.69 k 0.2 1 8.90 t 0.07 59.86 k 0.42 24.31 k0.19 14.98It0.11 8 1.03 k 0.91 - - - 10.93-tO.36 - - - 8.43 k0.16 - 20.60 k 0.4 1 - - - 9.30 k 0.12 59.95 t 0.65 - - - 23.85 +- 0.40 - - - 15.29 k 22 - - 70.13 t 0.79 *Mean k standard deviation for the results of the chemical analysis of ten different samples taken from the same original specimen.1395 1 Time - Fig. 3 Wavelength scans in the vicinity of As I 193.696 nm A As 12.5 pg ~ m - ~ ; and B V 100 pg ~ m - ~ The linear calibration curves obtained were sufficiently accurate as shown by the correlation coefficients for the regression lines V r=0.9997; As r=0.9996; and Ba r=0.9995. Results and Discussion Spectral Interferences The proposed fusion process with sodium hydroxide as it suppresses metallic element impurities eliminates the spectral interferences of the impurities on the analyte emission lines.One exception is the V line at 193.682 nm which partially overlaps the As line at 193.696 nm (Fig. 3). The spectral interference was corrected on the basis of the equation where c is the real concentration of As caAs the apparent concentration of As cv the concentration of V and kAs,v the interelement correction coefficient of V for As. The kAs,v was determined by measuring the emission intensities at the specified wavelength for As using the single element solutions 10 pg ml-I of As; 50 pg ml-I of V; and 100 pg ml-I of V. Thus the value for the kAs,v was found to be 0.09.Effect of Dilution on the Precision of Results The optimum dilution working conditions for the determi- nation of V As and Ba that led to the best precision are summarized in Table 3 which also includes the relative standard deviation (RSD) values obtained for 20 ratio intensity measurements (short-term precision). It is in- ferred that the highest precision (lowest RSD) is obtained by diluting the sample solution (1+9) for the three elements. Analytical Results The chemical analysis was carried out for each glass on ten samples taken from the same original material following the optimum working conditions reported in Table 3. Table 4 shows the results of ICP-AES compared with the reference techniques for the five analysed glasses including the standard deviation (a) corresponding to the precision of ten analysed samples (precision of the method).The analytical results obtained using the three techniques are fairly acceptable and permit the location of five com- positions within the glass formation area in the V20J- As203-BaO phase equilibrium system. The method precision is significantly better using ICP-AES. By comparing data in Tables 2 and 4 it becomes clear that during glass production As203 volatilizes consider- ably V205 volatilizes to a lesser extent but BaO is not volatilized. If V205 were not volatilized its content in the glass would be higher than the theoretical value because of the need for compensation of As203 losses. Conclusions During the production process of glasses belonging to the V205-As20,-Ba0 system some volatilization of As203 and V205 takes place which needs to be controlled in order to determine precisely the glass building area. To reach this objective ICP-AES offers results comparable with those of other accepted techniques for the determination of V As and Ba.842 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 References 3 Voinovitch I. A. Debras-Guedon J. and Louvrier J. L 'Analyse des Silicates Hermann Paris 1968. 1 2 Fernandez Navarro J. M. El Vidrio Consejo Superior de Investigaciones Cientificas Madrid 1991. Jurado. J. R. Thesis Universidad Complutense de Madrid Paper 3/01 7426 Received March 26 1993 1978. Accepted May 14 1993