JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 827 Boron Determination in Steels by Inductively Coupled Plasma Atomic Emission Spectrometry. Comparative Study of Spark Ablation and Pneumatic Nebulization Sampling Systems* Aurora G. Coedo Teresa Dorado Ester Escudero and Isabel G. Cob0 Centro Nacional lnvestigaciones Metalurgicas Consejo Superior de Investigaciones Cientificas Gregorio del Amo 8 28040 Madrid Spain An experimental study for the determination of boron in steels by inductively coupled plasma atomic emission spectrometry is presented. A comparison is made of spark ablation and pneumatic nebulization (after microwave digestion) sampling systems. A one-step microwave digestion procedure for total boron content using diluted aqua regia (HCI+HNO 3+1) and high pressure vessels was developed.The influence of microwave power and time on the dissolution of boron compounds is discussed. The strongest available conditions (0 Q 600 V 600 Hz) were required for spark ablation sampling. The stability of spark sampling during the spark ablation-ICP process was tested by plotting iron and boron emission versus sparking time. The iron content of collected and dissolved spark-produced particles was analysed and showed that the amounts of aerosol from different boron steels samples during 90 s sparking processes were fairly similar. The analytical performance of both systems was evaluated. Using pneumatic nebulization after microwave digestion of the sample a detection limit for boron of 2.6 ,ug g-1 and overall relative standard deviation (RSD) values of 1-3.5% were found.For spark ablation the detection limit for boron was 0.65pg g-l the overall RSD ranging from 0.5 to 1.5%. A comparison of the data for British Chemical Standards (BCS) Certified Reference Materials (CRMs) Carbon Steel Residual Series (Group B) and Spectroscopic Standard (SS) 456/1 to 460/1 indicated that the accuracy of both methods was satisfactory. Keywords Spark ablation; microwave digestion; inductively coupled plasma atomic emission spectrometry; boron determination; steel The importance of the effect caused by very low concentra- tions of boron on the physical properties of steels (hot workability hardenability creep resistance etc.) requires precise and accurate determination of this element. Com- pared with most other alloying elements the amount of boron added to steel is extremely small and commonly the boron content in boron treated carbon steels ranges from 0.0005% or less to about 0.005% (too little boron may be ineffective and too much can cause difficulties in rolling or forging).Boron was probably the first element in steelmak- ing to require analytical techniques capable of determining this element at the pg g-I level. Ambrose et aL2 have reviewed methods for the determina- tion of boron in steel. These workers also examined and commented on the method adopted by the European Committee for Iron and Steel Standardization/Technical Committee 20 (ECISS/TCZO) and presented analytical results for the boron content of some certified reference steel samples using different methods.The procedure suggested by Thiering,3 and the study mentioned above form the basis of the International Organization for Standardization (SO) standard method for boron determinati~n.~ This interna- tional standard is applicable to boron contents in steel of between 0.0005% and 0.012% (m/m) and specifies a spectrophotometric procedure using curcumin which is sensitive to experimental conditions during the dissolution steps in order to obtain quantitative dissolution of all the boron compounds and low blank values. The technique of microwave digestion in pressure vessels has progressed significantly over recent year^,^-^ and offers a very effective and fast dissolution system in addition to low blank values. These characteristics make the system very attractive for use in inductively coupled plasma (ICP) met hods.Over the past several years there has been growing interest in developing direct solid sample introduction *Presented at the 1993 European Winter Conference on Plasma Spectrochemistry Granada Spain January 10- 15 1993. systems for use with ICP spectrometry. Broekaert et aL9 reported a study of some techniques for direct solid sampling in plasma spectrometry. Although several s t u d i e ~ I ~ - ~ ~ have shown that the spark ablation-ICP combination is a convenient means of direct solid steel sample analysis it is difficult to ascertain the physical and chemical characteristics of the spark-produced aerosol and consequently the analyti- cal performance of the technique. Watters et af.14 presented a detailed study of the physical properties and the chemical composition of a spark-produced aerosol and corresponding erosion craters on brass certified reference materials (CRMs). In this paper a comparison is made between pneumatic nebulization (from microwave dissolved samples) and spark ablation sampling systems used with ICP spectro- metry for analysing boron in steels.In an effort to establish a microwave dissolution proce- dure a study was conducted to select the most important variables conditioning the microwave digestion process acid mixture digestion vessels proportion of sample to reagent required power time and number of samples. For spark ablation the first assays aimed to prove the similarity of the amounts of aerosol produced (from different steel samples) and the stability of its chemical composition.The aerosol particles produced for the spark- ing process were collected and dissolved in aqua regia (HCl+HN03 3+ 1) and the iron content was analysed by ICP spectrometry to determine the total mass eroded. The stability of the aerosol composition was verified by plotting the boron and the iron emission intensities versus spark erosion time curves. British Chemical Standards (BCS) CRMs Carbon Steel Residual Series Group B and Spectroscopic Standard (SS) 45611 to 460/1 were used for testing both methods. Experimental Instrumentation Inductively coupled plasma atomic emission spectrometry (ICP-AES) measurements were performed with a Jobin828 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL.8 Table 1 Selected working parameters Microwave oven Power1W 360 Time/min 30 Digestion vessels HPV 80 Voltage1V 600 Repetition rate (f)/Hz 600 Sampling spark Capacitance1pF 1 InductancelpH 20 Resistance (R)lQ 0 Gas flow rate/l min-l Permanent carrier gas 2.1; analysis gas 0.80 ICP-AES Power1W 950 Gas flow rate/l min-' 14 Observation height1mm 13 Yvon JY 24 plasma spectrometer purging with nitrogen. Spark ablation sampling was carried out with a Jobin Yvon JY-SAS sparking unit. Microwave digestion solutions were obtained using a Milestone MLS- 1200 microwave oven and HPV 80 high pressure vessels. Table 1 lists the selected operating conditions. Samples British Chemical Standard CRMs Carbon Steel Residual Series BCS-CRM and SS 456/ 1 to 460/ 1 are boron-carbon steels with certificate boron contents between 0.001 5 and 0.01 18%.The BCS-CRM samples come in chip form for wet analysis while SS samples come in disc form for direct solid analysis. The boron content values are similar in both types of sample and they were re-certified in 1988 in accordance with the study conducted by the European Committee for Iron and Steel Standardization (ECISS/TC 20). The Fe-0 sample with a boron content below 1 ppm was employed as a 'blank'. Results and Discussion Microwave digestion Boron is present in steel both in acid-soluble and non- soluble forms. The most significant non-soluble boron compound is BN. The dissolution step not only increases analysis time but can also cause errors when ICP-AES is used by increasing the spectral background and decreasing overall analytical sensitivity .In this paper a microwave digestion system was used to completely dissolve the sample. Tests were performed on different acid mixtures (containing different proportions of hydrochloric nitric hydrofluoric phosphoric and per- chloric acids) and different proportions of samples and reagents to find the simplest procedure for stable and quantitative dissolution of all types of boron compounds. After this selection different microwave digestion pro- grammes were applied varying the operational parameters of the oven (microwave power and time). Two types of pressure vessels were used low pressure SV-140 [Pmax= 18 bar ( 1 bar= 1 x 1 O5 Pa) V= 140 ml] and high pressure HPV- 80 150 bar V= 80 ml). When a closed vessel is used the pressure rises with an increase in microwave power and depending on how long the microwave acts.Consequently the temperature of the digestion mixture increases also enhancing efficiency. The BCS 46011 sample was used to test the influence of the parameters studied. Coedo and Dorado15 concluded that by using an ICP method this CRM with a certificate value for total boron of 28 ppm contains a high proportion = SV 140 (low pressure) 0 HPV 80 (high pressure) 240 360 500 600 PowerW Fig. 1 Effect of microwave power on boron dissolution of BCS- CRM 46011 for two types of pressure vessel for a time of 30 min = SV 140 low pressure 0 HPV 80 high pressure 3o I ,n 20 n n n I I I I I I U 5 10 20 30 45 Time/s Fig. 2 Effect of time on boron dissolution of BCS-CRM 46011 for two types of pressure vessel.Power 600 W for S V 140 and 360 W for HPV 80 of non-soluble boron particles ( ~ ~ 5 3 % of the total boron content) and so can be used to evaluate the efficiency of the dissolution procedures. The results obtained by varying the microwave power for the two types of vessels and for a fixed period of 30 min are shown in Fig. 1. The influence of time for a fixed power value is shown in Fig. 2. This power value was set at 600 W for SV-140 (the maximum value that could be applied with the selected reagents without risking vessel failure) and at 360 W for HPV-80 (this level of power was enough to recover all the boron content when this type of vessel was used). The BCS 460/1 sample was completely dissolved only when high pressure vessels were used.The volume of the high pressure vessels (HPV 80) is 80 ml the volume of the low pressure vessels (SV 140) being 140 ml. This difference means that under the same operating conditions (same portion of test sample reagents and oven parameters) the pressure reached in the HPV vessels is greater and conse- quently the temperature and the dissolution process effi- ciency is higher. It could be suggested that a high pressure is required to dissolve all the boron compounds in sample BCS 46011 owing to the presence of a high amount of boron nitride and of a high carbon content (0.45%) producing fine carbon particles that help to retain other products such as boron compounds. As a result of this study the following operating dissolu- tion procedure was adopted.A 0.250 g sample was dissolved in an HPV-80 vessel (maximum pressure 150 bar) with 5 ml of HC1+ 2 ml of HN03 + 10 ml of H 2 0 + 2 drops of HF by applying a one-step microwave programme of 360 W lasting 30 min. After cooling the solutions obtained were made up to 50 ml with water in graduated polyethy- lene flasks. The efficiency of the method was verified by simultaneous treatment of 2-6 samples.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 829 Spark Ablation Since the use of the spark to generate sample aerosol divides the sampling and excitation processes into two separate events systematic errors in each step should be examined in order to evaluate the results provided by the combined spark ablation-ICP technique. Most studies have centred on the erosion of the sample surface yet the analytical information is carried to the plasma by the particles that are formed.A high sample ablation and rate of analyte introduction into the plasma provide higher levels of sensitivity and higher detection capacity. The amount of analyte reaching the plasma increases both with the voltage and the repetition rate. However the particle size is increased by raising the voltage applied not by raising the repetition rate. A high repetition rate is preferable to a high voltage because of the particle size produced. An increase in the particle size results both in decreased stability and poor signal-to-background ratio of ICP signals (the finer the particles the better the stability in the plasma).Conse- quently source parameters resistance (R) voltage (V) and repetition rate (f) which have a clear influence on the analyte produced must be changed in line with the nature and characteristics of the samples to be analysed. The condition and hence the burn is stronger with decreasing R but increasing V or f Owing to the hardenability of boron steels and considering that fand R have a direct influence on the amount of material eroded but not on the particle size these parameters were set to obtain strong burns (R = 0 SZ and f=600 Hz). Tests were conducted using different voltage spark values from 350 to 600 V (maximum available voltage). The most repeatable and accurate results were obtained by applying the strongest conditions (600 V). As a result of this study the operating conditions listed in Table I were selected for spark sampling.As the samples heated up during the sparking process they were cooled in a carbon dioxide stream to allow consecutive measurements to be made under the same conditions with a view to achieving better sampling repeatability. Under the established conditions a spark ablation-ICP process lasted about 85 s (30 s for pre-integration and cleaning 15 s for transfer and 40 s for a ‘three-point’ mode measurement with background correction). To minimize sparking times peak intensities were measured employing a ‘three-point’ mode. In this mode the window includes three points with a distance between them of 0.0003 nm and the intensity value is a weighted average of these three points. The aerosols produced from 90 s spark ablation attacks of boron-carbon steels were trapped and dissolved in diluted aqua regia. These particles were removed from the end of the 0.75 m long tube (5 mm id.) which transports the analyte from the sparking chamber to the ICP torch with the aid of an argon gas flow of 0.8 1 min-l by introducing the end of the tube into the acid solution (10 ml of aqua regia+ 40 ml of water).After completing the dissolution by heating the iron content of these solutions was determined by ICP-AES in order to calculate the amount of analyte produced during the period of time from the beginning of the spark process to the acquisition of a spark ICP measurement ( ~ 8 5 s). The results show good repeatability of the amount of eroded material produced during sparking processes from the different boron-carbon steel samples tested.This amount was 0.700-+0.025 mg of iron for sparking periods of 90 s. The profiles of the boron and the iron emission intensities versus spark time show the signal stability of the aerosol produced under the selected spark conditions at least for the duration of the spark ablation-ICP measurement (Fig. 3). ICP Measurements Before measurement a preliminary study was conducted in order to choose the boron analytical line. Under the operating parameters selected for the JY 24 ICP instrumen- tation the lowest detection limit and the best values of relative standard deviation (RSD) were obtained for the line at 208.959 nm. The scans around the selected boron line are shown in Fig. 4. These scans correspond to the ‘blanks’ and to the calibration samples employed with both sampling systems.After having considered these scans values for background correction were measured at 208.925 nm (0.035 nm left of the boron peak). By using pneumatic nebulization blank values measured with background correction were approxi- mately half the total emission intensity. There are two components in the blank total emission a specific peak signal and an elevation of the background. Table 2 gives the analytical performance expressed in terms of background equivalent concentration (BEC) detection limit (DL) and precision. To calculate the DL (defined as the concentration of a solution which gives an absorbance equal to three times the SD of the blank) the following formula was used The BEC values (boron concentration giving a net analyte signal equal to the background signal) were 138 pg g-’ for pneumatic nebulization and 46 ,ug g-l for spark ablation.In the first system the BEC value was so high because of the 208.959 nm 0.01 nn i H 208.959 nm 208.959 nm 30 50 70 90 Time/s Fig. 3 Effect of sparking time on emission intensity of A Fe; B B using SS 45911; and C B using SS 45611 Fig. 4 Emission profiles. (a) Spark ablation 1 SS 45911; 2 SS 4561 1 ; and 3 Fe-0. (b) Pneumatic nebulization (0.5 g per 100 ml) I BCS-CRM 45911; 2 BCS-CRM 45611; and 3 Fe-0. (c) Solutions (100 ml) 1 0.5 g of Fe plus 0.7 ppm of B; 2 0.5 g of Fe plus 0.1 ppm of B; 3 0.5 g of Fe; and 4 blank830 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL.8 Table 2 Certified and found values for SS samples x=mean of six determinations; T=interval of uncertainty of the mean for a level of probability of 95%. T= t95,5 where t95,S = 2.57 1 Total Blpg g-I Samples SS 45611 SS 45711 SS 45811 SS 45911 SS 46011 CENIM * L=low calibration standard. i H = high calibration standard. Microwave digestion Spark ablation ~~ ~ Certified X T a 15 14.7 0.31 0.26 25 24.4 0.50 0.43 61 59.6 0.81 0.70 118 120.0 1.32 1.14 28 26.8 0.49 0.42 - 49.5 0.47 0.4 1 X T a Calibration L* 25.3 0.90 0.78 62.0 1.12 0.97 Calibration H t 27.3 0.85 0.74 48.7 0.90 0.78 background elevation produced by the iron matrix and of the elevation of the specific signal as a result of the dissolution process (as can be seen in Fig. 4). At levels five times higher than the corresponding DLs RSDs (n=6) of 1.3 and 3.5% were obtained using pneumatic nebulization and spark ablation respectively.Calibration After verifying the linearity of the boron emission intensi- ties versus boron concentrations within the interval of boron contents considered only two samples (‘low’ and ‘high’) were used to obtain the calibration graphs. Using pneumatic nebulization two calibration samples were prepared with 0.01 and 0.07 mg of boron added from a standard boron solution in the presence of 0.5 g of pure iron (Fe-0 sample) per 100 ml. These two calibration solutions are equivalent to steel samples containing 0.0020 and 0.0 14% of boron respectively. Two samples were used for calibration with spark ablation sample SS 456/ 1 (0.00 1 5% of boron) as ‘standard low’ and sample SS 459/1 (0.01 18% of boron) as ‘standard high’.The ICP calibration graphs obtained from pneumatic nebulization and spark ablation sampling systems are shown in Fig. 5. In these graphs the BCS-CRM and SS samples not used for calibration are interpolated by plotting their emission values against the certified concentrations. The correlation coefficients obtained with the linear regres- sion of all the data points were 0.9992 and of 0.9990 respectively. The results obtained from the above calibration graphs for the BCS-CRM and SS samples series and for a Centro Nacional de Investigaciones Metalurgicas (CENIM) sample are shown in Table 3. The mean values obtained with both methods for the boron content in the CENIM sample were compared by employing the Student’s t-test.This sample 1 !!? 1200 .- C 5. 1000 L 2 800 fJ 600 4- >. 400 200 c. .- - .- cn c I 1 I I I 1 0 0.002 0.004 0.006 0.008 0.010 0.012 0.014 - B (%I Fig. 5 Calibration graphs obtained using A pneumatic nebuliza- tion; x spark ablation; and U BSC-CRM (pneumatic nebuliza- tion) and SS (spark ablation) of samples was analysed first in disc form by spark ablation sampling and then by pneumatic nebulization of a dissolution of chips obtained from the disc sample. As the same number of determinations were carried out with the two methods (n = 6) mutual agreement was tested by the ratio XA -XB t= 0 2 + a2B JI1-l where x= mean; o= SD; v= degrees of freedom; ad ( 1 -a) is the probability level. For a probability level (1 -a)=0.95 and 10 degrees of freedom (v=2n-2= lo) in Student’s tables t = 2.228.The value calculated from figures in Table 3 t=2.026 is lower than the tabulated one. This means that the difference between the two means is statistically insignificant and can be explained by random errors alone. Conclusion The high pressure microwave digestion method is a valid system for complete dissolution of boron compounds in steels. The sample dissolution time is reduced from hours to 30 min. A further advantage was seen when six samples were dissolved simultaneously. As no evaporation occurred during the digestion process only a few millimetres of aqua regia were required as digestion reagent leading to low blank values. This allows lower DL to be attained than with conventional dissolution procedures. As a result of reduced sample handling and the impossibility of airborne contami- nation the risk of contamination is substantially reduced.By employing strong spark conditions (0 SZ 600 V 600 Hz) spark ablation can be used as a solid sampling system for the determination of boron in steels. The amounts of spark aerosol produced from boron steel surfaces prepared under the same conditions during a spark ablation ICP process are fairly similar and their elemental composition (B and Fe) remains stable and matches the bulk compo- sition reasonably well. Table 3 Analytical performance Sampling system BEC*/pg g-l DLtlpg g-l RSDS (0.5 g per 100 ml) 138 2.60 1.3 Microwave digestion Spark ablation 46 0.65 3.5 * Background equivalent concentration giving an emission signal equal to twice the total emission of the blank.t Detection limit producing an emission signal equal to three times the SD of the blank measured with background correction. $ Relative standard deviation (n=6) at five times the DL.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1993 VOL. 8 83 I The ICP-AES technique with the two tested sampling systems could replace traditional chemical methods for boron analysis in steel samples. The drawback to the microwave digestion method is its higher background values and consequently its lower signal-to-background ratios and higher DL (2.5 times that achieved with spark ablation) however the precision values are better. The main obstacle to spark ablation sampling is the need to use solid standard samples for calibration. The accuracy of both methods is demonstrated by the results obtained for the CRMs analysed.The assays concerning development of the microwave sample procedure form part of a research project financially supported by the European Community for Steel and Carbon (ECSC) No. 7210/GD/936 (E2. 3/89). References Steel and Its Heat Treatment ed. Thelning K. E. Butter- worths London 2nd edn. 1984 p. 405. Ambrose A. D. Harine M. Staats G. and Weichert E. Steel Res. 1989 60 363. Thierig D. Fresenius'Z. Anal. Chem. 1982 310 154. 4 5 6 7 8 9 10 11 12 13 14 15 IS0 10153 199 1. International Organization for Standardiza- tion P.O. Box 56 CH- 12 1 1 Geneva 20 Switzerland. Introduction to Microwave Sample Preparation eds. Kingston H. M. and Jassie L. B. ACS Professional Reference Book Washington D.C. 1988. Kammin W. R. and Brandt M. J. Spectrosc. Znt. 1989,1 50. Gilman L. B. and Engelhart G. Spectrosc. Znt. 1990 2 16. Progress of Analytical Chemistry in the Iron and Steel Industry Commission of the European Communities. EUR 141 13 Luxembourg 1992 pp. 293-295. Broekaert J. A. C. Leis F. Raeymaekers B. and Zaray Gy. Spectrochim. Acta 1988 39 339. Mandoki A. Boron in Low Alloy Steels; Analytical Report Instruments S.A. Jobin Yvon Longjumeau 199 1. Lemarchand A. Labarraque G. Masson P. and Broekaert J. A. C. J. Anal. At. Spectrom. 1987 2 481. Prell L. J. and Koirtyohann S. R. Appl. Spectrosc. 1988,42 1221. Coedo A. G. Dorado L. T. Seco J. L. and Cobo I. G. J. Anal. At. Spectrom. 1992 7 11. Watters R. L. Jr. DeVoe J. R. Shen F. H. Small J. A. and Marirenko R. B. Anal. Chern. 1989,61 1826. Coedo A. G. and Dorado L. Rev. Metal. Madrid 1985 21 87. Paper 2/05430B Received November 9 1992 Accepted February 8 1993