Determination of Trace Amounts of Boron by Microwave Plasma Torch Atomic Emission Spectrometry Using an On-line Separation and Preconcentration Technique QUN JIN HANQI ZHANG FENG LIANG WENJUN YANG AND QINHAN JIN* Department of Chemistry Jifin University Changchun 130023 Chinu A new system for the determination of trace amounts of boron by flow injection microwave plasma torch atomic emission spectrometry is proposed. In the system a strongly acidic cation-exchange column is connected in series with a strongly basic anion-exchange column. By the combined function of the two columns interferences (e.g. iron) in the sample solution are removed and the analytical performance improved. Optimization studies illustrate the dependence of the analytical signals of boron on various parameters including observation height microwave forward power flow rates and acidities of the sample solution and the eluents in both columns.The effect of concomitant ions on the relative emission intensity of boron is also examined. Detection limits of 0.0055 mg 1 - ' at 45 samples h-' and 0.0018 mg I-' at 20 samples h-' with a precision of 4.2% at the 0.020 mg I-' level of boron were achieved. With sample columns of 1.25 and 5.0 ml enrichment factors of 22 and 88 were obtained respectively. The proposed system has been applied to the determination of trace amounts of boron in steel samples and the results obtained were satisfactory. Keywords Boron; microwave plasma torch; atomic emission spectrometry on-line separation; preconcentration Several atomic spectrometric techniques including capacitively coupled microwave plasma atomic emission spectrometry (CMP-AES),' microwave-induced plasma (MIP) AES,2 flame atomic absorption spectrometry ( FAAS),3*4 direct current plasma (DCP) AES,' inductively coupled plasma (ICP) AES,6-" ICP mass spectrometry (MS),13 have been developed for the determination of boron.The determination of trace amounts of boron by AAS is not satisfactory as boron is a refractory element. Although ICP-AES and ICP-MS are suit- able for such determinations the purchasing and operating costs are relatively high. In comparison with an ICP micro- wave plasma (MWP) AES is relatively low cost. However very little has been carried out so far on the determination of trace amounts of boron by MWP-AES. Murayama et al.' employed a CMP to determine boron directly.They used a 400 W discharge in argon as an excitation source and the nebulized sample was introduced tangentially into the coaxial waveguide through a sample inlet. Lichte and Skogerboe2 used a low powered (100 W) argon MIP with a desolvation system to determine boron where a design modification to the Evensen cavity was made an impedanse matching device consisting of 20 gauge copper wire was attached to the fine tuning stub insulated with glass tubing to prevent shorting to the cavity wall and extended out from the cavity at nominally 45" to its axis thus the tuning problem was reduced; on the other hand * To whom correspondence should be addressed. Journal of Analytical Atomic Spectrometry the plasma was run in a quartz tube axially through the cavity instead of using the more conventional transverse configur- ation thus ignition and maintenance of the plasma (even when the argon support gas was saturated with water vapour) were possible.An end-on optical arrangement was utilized for both plasmas. However both methods are not sufficiently sensitive (the detection limits of the former and the latter were 0.030 and 0.010 mg 1-' respectively) for the determination of boron and no practical samples especially not those containing large amounts of iron have been analysed using these two tech- niques. The presence of iron in the samples can cause analytical difficulties in the determination of boron because iron has emission lines at 249.77 (Fe 11) 249.65 and 249.70 nm (Fe I) which produce spectral interferences to the more sensitive lines of boron at 249.77 and 249.68 nm. More recently a new excitation source the microwave plasma torch (MPT) has been reported by Jin et al.14 The MPT discharge possesses an ICP torch-like configuration. This configuration greatly improves its tolerance to the intro- duction of wet aerosols'' compared with that of traditional MIPS.However matrix effects are still significant. Matrix interferences in the atomic spectrometric determi- nation of trace amounts of boron have been studied pre- v i ~ u s l y . ~ - ~ * ' ~ These interferences include spectral interferen~e,~ interference from fluorides' and those from other elements present in steeL9.l6 To eliminate the matrix interferences a pre- separation procedure with a column filled with a chelating ion exchanger (Amberlite IRA-743 ion-exchange resin) was employed.16 However this type of resin does not fully meet the requirements as an ideal packing material for the determi- nation of boron in steel samples because a higher pH (5-6) of the sample solution is required for the chelating resin to complex the metals completely and so hydroxide precipitation of Fe3+ cannot be avoided when a large amount of iron is present.Since the concentration ratio of iron to boron in steel samples can exceed lOoOO:l the determination of trace amounts of boron will definitely be affected and this is a well known problem to be solved in the determination of boron in steel because of the spectral interference from iron. Other separation methods employed in ICP-AES for the determi- nation of boron in steels include di~tillation,~ extraction' and pyrohydroly~is.~ Although these methods succeeded in elimin- ating iron interference virtually none of them can be easily automated.In the present work a non-selective strongly acidic cation-exchange column was employed and the separation was performed under acidic conditions (pH 2) in an on-line flow injection (FI) mode. Therefore almost all types of cations were separated from boron and the precipitation of some metal hydroxides was prevented. This is particularly valuable for separating trace amounts of boron from large amounts of iron. To improve the sensitivity of the determination of boron by Journal of Analytical Atomic Spectrometry May 1996 Vol.I I (331 -337) 331ICP-AES anion-exchange chromatography has been employed to adsorb boron as (BF,)- . However the process was an off- line one and the precision was not satisfactory.'2 Boric acid is a weak acid and can exist in various forms ( H2B03- HB03,- BO3,- and H,BO,). Which form dominates amongst them depends on the respective fraction existing in solution (6). The expressions of 6H3R0 6H2BO3- hHBO32- and 6B033- are as follows (1) &,BO3- =ka,CH+ I'IJ (2) 8HB032- = kalka2[Hf1/J (3) 6 R 0 3 3 - = kalka2ka3/J (4) 6 H m 3 = CH + i 3 / ~ where k,,(= k (= 10-'2.74) and ka3(= 10-'3.80) are acid dissociation constants for the first second and third ionizations of boric acid respectively; J = [H+]3+kal[H+]2+kalka2[H+] +k,,k,,k,,; and [H'] is the hydrogen ion concentration.It is indicated in eqns. (1-4) that when [H'] is around 1 x lo-'' (pH 11) boric acid will exist mainly as H2B03- and when [H'] is about 1 x lo- (pH 2) it will exist mainly as H3B03. In the present work a micro- column packed with Type VS-I1 strongly basic anion exchange fibre was added following the cation-exchange column. Because the fibre is a network polymer in which the distance between active groups is larger than that in common anion exchangers and the active groups are on the surface of the fibres the fibre possesses a quicker exchanging and desorbing capability than that of common anion exchangers and the resulting peak shapes of the emission signals (plots of intensity uersus time) are sharper. Thus since peak height was used to measure the emission intensity the enrichment factor ( E F ) for boron is enhanced.Boron is adsorbed on the fibre in the H2B03- form in a basic medium (pH 11.2) and then eluted in the H3B03 form by HCl (1 moll-'). In addition an FI technique was also implemented in the system to improve the sample through- put and absolute sensitivity and to save on sample consump- t i ~ n . " . ' ~ The system has been applied to the determination of boron in some steel samples and the results are satisfactory. EXPERIMENTAL Instrumentation The plasma was sustained with an MPT assembly fabricated in this laboratory the design of which has been described previo~sly.'~ The plasma was operated with argon as both carrier and support gases and was viewed in a side-on mode. The sample introduction system consists of an FI system (Fig.1) with a peristaltic pump (LZ-1010) and two sign rotary valves a pneumatic nebulization (PN) system with a pneumatic nebulizer and a nebulizing chamber (both from the Beijing Vast Time Scientific Manufacturing Laboratory) and a desolv- ation system (DS) with a condenser and a desiccator containing concentrated H2S04. Other equipment used includes a micro- wave generator (WB-WC 2450 MHz Haiguang Instrument) a monochromator (Model WDG 500-11 Beijing Second Optical Instrument) a photomultiplier tube (PMT) (R456 Hamamastu) and its power supply (FH-4268 Factory No. 261 of China) and a chart recorder (Model 056 Hitachi). The operating conditions are as listed in Table 1 unless stated otherwise. Reagents All reagents used were of analytical-reagent grade and prepared with de-ionized water. All solutions were stored in polyethylene bottles.The purity of argon used as both carrier gas and support gas was 99.99%. A stock standard boron solution (loo0 mg 1-') was prepared m~ m i d ml min" Load 0 Elution Fig. 1 FI separation and preconcentration system. S sample; R1 reagent 1 (HCl solution pH2); R2 reagent 2 (1 mol I-' ammonia solution) E l eluent 1 (4 moll-' HCI); E2 eluent 2 (1 mol I - ' HCI); C1 column 1 (separation column); C2 column 2 (preconcentration column); V1 valve 1 (multifunctional valve); V2 valve 2; P pump; W waste by dissolving 2.858 g of anhydrous boric acid ( H3B03) in 0.5 1 of de-ionized water. A Type VS-I1 strongly anionic basic exchange fibre (Liaoyang Institute of Science and Technology China) was employed for preconcentration purposes and a Type 732 strongly cationic acidic exchange resin (Shanghai Resin Factory China) which is similar to Dowex 50W-X8 cation exchanger was used for separation purposes.Procedure The FI separation and preconcentration system consists of an eight-channel multifunctional valve (V 1) and a two-channel valve (V2) a peristaltic pump (P) a separation column (Cl) and a preconcentration column (C2) (Fig. 1). The correspond- ing flow rates in individual channels are also shown in Fig. 1. Column preparation An appropriate amount of anion-exchange fibre was soaked in 2 moll-' HC1 for no less than 6 h taken out and washed with water until no C1- could be detected (with AgNO,) and then fitted into a column [Fig.2(a)]. The preconcentration column was placed in position C2 of Fig. 1. The cation- exchange resin was also soaked in 2mol1-' HCl for about 8 h then washed with water and packed into a column [Fig. 2(b)]. The separation column was placed at position C l of Fig. 1. Separution and preconcentration Whenever a new run was started the multifunctional valve (Vl) was first set at the load position [Fig. l(u)J valve 2 (V2) was set to enable the sample to be pumped and to flow through 332 Journal of Analytical Atomic Spectrometry May 1996 Vol. 1 1Table 1 Operating conditions MPT Monochromator FI system Carrier and support gas Microwave frequency MHz Forward power W Support gas flow rate ml min-l Carrier gas flow rate ml min-' Plasma viewing mode Plasma viewing position Wavelength nm Slit-width pm Slit height mm HCl concentration of eluent 1 moll-' Acid concentration of sample moll-' Flow rate of sample ml min-' HCl concentration of reagent 1 moll-' Flow rate of reagent 1 ml min-' Ammonia concentration in reagent 2 moll-' Flow rate of reagent 2 ml min-' HCl concentration of eluent 2 mol I-' Flow rate of eluent 2 ml min-l Ar 2450 60 500 700 Side-on 6-8 mm above the top of the torch 249.68 25 2 4.0 0.01 2.5 0.01 2.5 1.0 2.5 1.0 1.5 1 Fig.2 (a) Type VS-I1 strongly basic anion fibre conical micro-column with push-fit connections. Conical bottom id 4 mm; od 6 mm. Conical top id 2mm; od 4mm; length 40mm. 1 PTFE tubing; 2 sorbent packing; 3 Tygon tubing; 4 plastic foam; 5 thick walled silicon rubber tubing. (b) Type 732 strongly acidic cation exchange uniform-bored column with threaded-fitting connections.Inner diameter (id) 5 mm; outer diameter (od) 8 mm; length 120 mm the separation column (Cl) where interfering cations were adsorbed. After separation the sample was mixed with reagent 2 (R2 1 moll-' ammonia solution) to reduce the acidity to a pH of 11.2 and then passed through the preconcen- tration column (C2) where boron was enriched by the fibre (the preconcentration time was usually 30 s). At the same time eluent 2 (E2 1 moll-' HCl) was pumped into the MPT through the PN and DS systems. Elution After a certain load time (30 s) V1 was rotated [Fig. l(b)] to allow eluent 1 (El 4mol1-1 HC1) to flow through C1 and the interfering cations to be eluted from C1 into the waste flask.At the same time E2 flowed through C2 and boron was eluted from C2 into the MPT through the PN and DS systems (the elution time was 30 s). Column washing After the signal had been recorded V1 was again rotated to the load position and V2 was rotated [Fig. l(c)] to allow reagent 1 (Rl HCl of pH 2) to flow through C1 then R2 was mixed with R1 and passed through C2 into the waste flask (the washing time was 20s). This stage ensured that both columns were under the same acid conditions as they were at the loading stage and thus guaranteed the RSDs of the method. Emission measurement The emission intensity was recorded with a chart recorder and measured as peak height. In a set of conditioning experiments the most intensive emission signal was indicated as 1.0 and then other emission signals were taken as relative emission intensities relative to the most intensive signal.Sample digestion Steel samples were dissolved in concentrated HCl and concen- trated HN03 (3 + 1 v/v) in a quartz beaker and gently heated on a sand bath until all of the steel had dissolved. Concentrated H,S04 and concentrated H3PO4 (2 + 1 v/v) were then added to the beaker and the sample solutions heated continuously until large amounts of sulfuric acid mist were produced. The residue was then transferred into a calibrated flask and diluted to the mark with de-ionized water. RESULTS AND DISCUSSION Ion-exchange Column Anion-exchange fi bre column The anion-exchange fibre column was designed to meet the requirements for high efficiency of boron enrichment and rapid elution with a minimum volume of eluent.The detailed con- struction of the column is given in Fig. 2(a). Because the adsorbing capacity of the fibre (0.67 mmol g-' of dry fibre) is large enough to eliminate a breakthrough situation a micro- conical cartridge containing about 30mg of dry Type VS-I1 strongly basic-anion exchange fibre was employed. The break- through appears when the concentration of boron exceeds 40 mg 1-l with an EF of 22 (preconcentration time is 30 s) which is much better than when a common strongly basic anion-exchange column was used (4 mg 1-' for the break- through point with an EF of 2) and which indicates why the fibre does not degrade the EF as most strongly basic anion exchangers do and does actually improve the sensitivity.Cation-exchange resin column As a strongly acidic cation exchanger Type 732 cation- exchange resin is not selective thus it can exchange many types of cations. The experimental results indicate that under appropriate acidic conditions the separation column can Journal of Analytical Atomic Spectrometry May 1996 Vol. 11 333adsorb high concentration of the interfering cation Fe3+ (lo00 mg 1-') owing to the higher adsorbing capacity offered by the larger volume of the column [Fig. 2(b)]. Boron Analytical Line Of the few strong emission spectral lines in the wavelength range 200-600nm the strongest boron lines observed from the MPT are at 249.77 and 249.68 nm which are similar to those observed from an ICP." In the present work the spectral line at 249.68nm was chosen as the analytical line.At this wavelength the background emission was shown to originate mainly from an NO molecular band emission." Hence it is clear that measurements at this wavelength will be affected by compounds present in the sample solutions that contain nitro- gen. Thus the ammonia solution remaining in C2 during the preconcentration stage and introduced into the MPT during the elution stage by the eluant would affect the determination of boron. So keeping the concentration of the ammonia solution used at the concentration stage constant is necessary to maintain good precision of the measurements. Plot 2 in Fig. 3(a) was obtained when blank sample solutions were introduced. Because the blanks did not contain the analyte the emission signals recorded were taken as the background emission signals. It is evident from this plot that background signals increase with ammonia concentration. Plot 3 in Fig.3(a) was obtained when sample solutions contain- ing the analyte were introduced. The emission signals (plot 3 ) recorded were taken as the signals for boron emission plus background emission. Plot 1 in Fig. 3(a) was obtained by subtracting signals in plot 2 from signals in plot 3. The 1.0 1 R O.,....,....,....0 . 9 5 ' ~ " " " ~ ~ ' ~ ~ " " ' " ' " ' ~ " " ' ~ ' ~ I . . _ . 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Ammonia concentrationlmol r' Fig. 3 (a) Effect of ammonia concentration on the relative emission intensity of boron. 1 Boron at 249.68 nm; 2 Background emission at 249.68nm; 3 Boron plus background at 249.68 nm.(b) Effect of ammonia concentration on the fraction of H,BO,- present in solution (&,BO -) detection limit for boron is limited by the signal to background ratio (S/B). A lower S/B value is preferred over high S/B values because the lower the S/B the lower the detection limit for boron. Hence the concentration of ammonia solution in R2 should be chosen to be as low as possible. However too low an ammonia concentration can cause a decrease in aHZBO3 - [Fig. 3 ~41. Optimization of the Experimental Conditions Eflect of observation height The effect of observation height on the emission intensity is shown in Fig. 4. The maximum values of the emission intensit- ies as well as the S/B were obtained at an observation height of 6-8 mm above the top of the torch which had been shown to be the best observation zone in previous work.20 Eflect of microwave forward power Microwave power is usually one of the most important param- eters which influences plasma performance such as plasma shape stability and capability of atomization and excitation of analytes. In the present work the effect of microwave forward power on the emission intensity of boron and the S/B was studied (Fig.5). At lower microwave power (< 55 W) the emission intensities of the analytical line increase with an increase in microwave power and at higher power (>55 W) an increase in forward power causes little increase in the emission intensity with slight decrease in the S/B. This result is different from that obtained with an MIP. In the latter case 3-0 1 0 2 4 6 8 1 0 1 2 Observation heighvrnrn Fig.4 Effect of observation height on the relative emission intensity of boron 0. relative intensity; 0 signal to background 2.5 2.0 x v) e .- 5 1.5 e .- c 0 .- .- % 1.0 E W 0.5 0 10 20 30 40 50 60 70 80 Microwave forward powerMl Fig.5 Effect of microwave forward power on the relative emission intensity of boron 0 relative intensity; 0 signal to background 334 Journal of Analytical Atomic Spectrometry May 1996 Vol. 11an increase in the forward power produced a proportionate increase in the net intensity of the spectral emission.2 The reason for this is probably that in the case of an MPT the plasma is formed between the intermediate and the central tubes near the top of the torch and extends into the surrounding air so that the volume of the plasma is enlarged with an increase in the forward power and the energy density in the plasma does not increase sharply.Acid conditions for the separation system In the proposed procedure before being mixed with R2 (ammonia solution) the sample solution has to be passed through the separation column to separate boron from cationic interferences. Of the concomitant elements that affect the emission intensity of boron in steel samples iron is the most troublesome interfering element so the optimum acidity of sample solutions was selected according to the ability of the column to adsorb iron. In the present work to eliminate the interference from iron a separation column (cation-exchange column) was used before the preconcentration of the boron analyte.To examine the sorbing-xtraction of iron on the separation column the eluate from the column was introduced into the plasma and the emission of iron was measured. In the study C2 in Fig. 1 was removed and a 1000 mg 1-’ iron sample solution was pumped through C1 into the PN system directly for about 120s without mixing with ammonia solution. The effect of sample acidity on the relative emission intensity of iron is shown in Fig. 6(a). Too low an HCl concentration (high pH values) were not attempted because in steel samples iron to boron ratios can exceed 1oooO 1 and thus when 0.1 mg 1-’ of boron was determined 1000 mg 1- of iron in the sample would precipitate with hydroxide ion under such acidic con- ditions. It is also indicated in Fig.6(a) that HC1 concentrations higher than 0.1 moll-’ (pH c 1) will cause the break-through of iron this result is consistent with Samuelson’s conclusion:” ‘The break-through capacity is diminished when the acidity of the solution is raised’. Pitts and Beamish” selected pH 1.5 as the sample acidity that would lead to the adsorption of large amounts of Fe3+ on a Dowex 50W-X8 cation-exchange column. In the present work an HCl concentration of 0.01 moll-’ (pH 2) was chosen in order to reduce the con- sumption of ammonia solution. The effect of the acidity of the El on the emission signal of iron is shown in Fig. 6(b). In the study a 1000 mg 1-’ iron solution (pH 2) was preconcentrated on C1 for 120 s and then eluted by El directly into the PN system. It was shown that a concentration of HC1 exceeding 2 moll-’ in El ensured that the emission of iron reached a maximum. This is because the high concentration of HCl facilitates the formation of chloro- complex anions ([FeClJ) which results in the desorption of al Concentration of HCI in 0 1 2 3 4 5 6 Concentration of HCI in sample solution/mol r1 eluent 1/m01 I-’ Fig.6 Effect of acidity of solution in separation system. (a) Effect of concentration of HCI in sample solution on the separation of matrix (iron); (b) effect of concentration of HCI in eluent 1 on the elution of matrix (iron) iron from the cation exchanger. In addition a higher concen- tration of HCl also took less elution time (30 s for 4 moll-’ HCI and 60 s for 2 mol I-’ HC1) and therefore favours an increase in the sample throughput.This result is also fairly similar to that observed with Dowex SOW-XS cation exchanger.23 Thus 4 moll-’ HC1 was selected as the eluent. Acidic conditions for the preconcentration system In the present work the pH value of the sample solution to be preconcentrated was adjusted by R2 (ammonia solution) (Fig. 1). When the concentration of ammonia solution is in the range 0.25-4.00 moll-’ the pH of the sample solution can be adjusted to 10.6-11.8 respectively. The effect of the concentration of the ammonia solution in a mixed sample- reagent solution on the emission intensity of boron and the fraction of H2B03- that exists in solution are shown in Fig. 3(a and b) [constructed according to eqn. (2)]. The similarity between curve 1 in the Fig.3(a) and the curve in Fig. 3(b) indicates that a pH lower than 11.2 (corresponding to 1 moll-’ ammonia solution) will decrease the emission signals because of the incomplete convertion of boric acid into H2B03-. However pH values higher than 11.2 are also not satisfactory because of the higher background emission resulting from the higher concentration of ammonia solution (curve 3) while the net emission intensity of boron will not increase with an increase in the concentration of ammonia solution (curve 1). Hence a pH value of 11.2 was selected for the mixed solution i.e. 1.0 moll-’ ammonia solution was chosen as R2. A study of the effect of acidity of E2 on the relative emission intensity of boron (Fig. 7) indicates that 1.0 moll-’ is the optimum concentration of HC1 for E2.It seems that from this acidity and higher boron exists in solution mainly in the form of H3BO3. In addition the introduction of HCl into the plasma between samples can also result in a rapid clean-up of the system because the memory effects within the desolvation system are greatly reduced through the use of an acid solution.24 Flow rates of sample reagent 2 and eluents The effect of total flow rate of sample and R2 on the relative intensity of boron was investigated while keeping the sample to R2 flow rate ratio (1 1) and the loading time (30 s) constant (Fig. 8). Higher ratios were not beneficial because under such circumstances a higher concentration of R2 was required which would deteriate the precision of the method. Ratios lower than 1 1 were also not attempted in order to avoid excessive dilution of the sample.Increasing the flow rate resulted in an increase in the relative intensity of boron. 1 .o m h 0.8 - 0 u) a e .- e .C 0.6 c v) .- ‘E OA Q) .- i;j 0.2 a - i? d 0.4 0.8 1.2 1.6 2.0 2.4 Concentration of HCI in eluent 2/ml I-’ Fig. 7 of boron Effect of the concentration of HC1 in eluent 2 on the elution Journal of Analytical Atomic Spectrometry May 1996 Vol. 1 1 335; 0.8 L .- v) a C 0 .- E 0.6 .g 0.4 .- E 3 0.2 P) - al LI 0 1 2 3 4 5 Total flow rates of sample and Rdml min-' Fig. 8 Effect of total flow rates of sample and reagent on the elution of boron However owing to the introduction of larger sample volumes at high flow rates the mechanical properties of the anion- exchange fibre would deteriorate as well as large amounts of sample being consumed. Thus a total sample and R2 flow rate of 5.0 ml min-' (at sample to reagent ratio 1 1) was employed.The eluent flow rate is an important parameter in column preconcentration. The effect of the flow rate of El on the relative intensity of iron and E2 on the relative intensity of boron are shown in Fig. 9. The optimum eluent rate for E2 is fairly similar to the normal sample introduction rate of 1-2ml min-' for ICP spectrometric systems." Too high a flow rate of E2 causes a decrease in the emission signal of boron. This could be for kinetic reasons that is the flow rate of E2 is too fast for the boron remaining on the fibre to be eluted into the solution as it is as fast as the rate under thermodynamic conditions will allow.A flow rate change for El within 1-5 ml min-' has no significant effect which proves that for the strongly acidic cation exchanger the flow rate can be chosen over a wide range. Reverse elution was adopted between the loading and elution stages of the preconcentration system to avoid insufficient contact between the fibre and the sample solution which could in turn influence the flow rate and even produce a break- through problem. Effect of Concomitant Ions A continuous sampling mode was adopted without C1 and C2 to investigate the effects of various concomitant ions including easily ionized elements (EIEs) on the emission intensity of 1 .o >r 0.8 c .- v) 0) c. .- 0.6 .- v) '$ % 0.4 .- c. - a LI o.2 I t . . I . . . . I 1 . . . 1 .. . . 1 . . . . 1 0 1 2 3 4 5 Flow rate of eluent 1 (for Fe) and eluent 2 (for B)/ml min" Fig. 9 separation of boron (0) from iron (0) Effect of flow rate of eluent 1 (El) and eluent 2 (E2) on the Table2 Effect of some concomitant ions on emission intensity of boron; emission intensity of boron (0.4 mg I - ' ) is taken as 100 when no other concomitant ions are added Emission intensity of boron Concomitant Concentration/ ion mg I-' None -. K + 1000 Na + 1000 Mg2+ 1000 Ba" 1000 Ca2 + 1000 Fe 200 Fe loo0 Fe 4000 - 1000 PO,^ - lo00 Cr04'- 1000 Without columns C1 and C2 100 140 140 180 124 108 720 1600 2400 90 95 85 With columns C1 and C2 100 100 100 100 100 100 100 103 105 100 100 98 boron (Table2). The most significant effect was shown to come from iron because of its spectral interference (at 249.78 and 249.65 nm).However when the separation column (Cl) and preconcentration column (C2) were used the effects of iron and other elements can be eliminated satisfactorily by the cation-exchange column (C1 ) provided that the concentration of Fe3+ and other ions are not too high (Table 2). The emission intensity of boron is not significantly affected by the existence of EIEs (including Na' K + Ca2+ Mg2+ and Ba2+) and some anions (including CrO,'- SO,'- and at the 1000 mg 1-' level. This situation is not really specific for the MPT. Hoare and MostynZ5 observed that lOOOrngl-' of Na did not affect the emission intensity of boron when determined by ICP-AES. Murayama et al.' and Lichte and Skogerboe2 also proved that the emission signal of boron was little affected by addition of Na into a CMP and MIP.Murayama et al. concluded that the continuum and analytical lines (which are less sensitive to the addition of Na) of boron radiate from the central core of the plasma. Detection Limit and Linear Dynamic Range In the present study the standard deviation of the blank measurement was used to calculate the DL. The DLs (30) for boron obtained by this method as well as others including the DL obtained with MPT-AES without preconcentration are as follows FAAS,3 1.5; ICP-AES (with an electrothermal vaporiz- ation method)," 0.0024; ICP-AES (with a pyrohydrolysis method),' 0.0035; ICP-MS,13 0.0004; MIP-AES,2 0.01; CMP- AES,' 0.03; and from the present work MPT-AES in a continuous sampling mode 0.03 and including separation and preconcentration 0.0018 mg I-'.It is clear that the proposed method is comparable with conventional ICP as far as the DL for boron is concerned. The linear dynamic range was shown to be over three orders of magnitude (0.02-40 mg 1-l). Precision The relative standard deviation (n = 11) for the determination of boron with the proposed method was shown to be 4.2% (at the 0.02 mg I-' boron level) and 0.15% (at 0.10 mg 1-'). Practical Sample Analysis Some standard reference samples of alloy steels were analysed in order to examine the applicability and accuracy of the method. The results obtained are listed in Table 3. 336 Journal of Analytical Atomic Spectrometry May 1996 Vol. 11Table 3 Determination of boron in steel samples using FI-MPT-AES with the on-line separation and preconcentration technique Certified value Sample Found (YO) (”/.I RSD (n = 5) (YO) Steel 1 0.0044 0.0049 3.6 Steel 2 0.0017 0.0015 3.9 Steel 3 0.0098 0.0100 3.8 Steel 4 0.0076 0.0080 4.1 Steel 5 0.0010 0.0012 4.7 CONCLUSION The present study suggests that FI-MPT-AES with on-line separation and preconcentration can greatly improve the per- formance of MPT-AES and successfully eliminate matrix effects on the spectrometric determination of trace amounts of boron even when there is an excess of iron present.The results compare well with those obtained with an ICP except for the slightly more complex mode of operation. However the cost of an MPT is more attractive and thus makes its shortcomings less important. REFERENCES Murayama S.Matsuno H. and Yamamato M. Spectrochim. Acta Part B 1968 23 513. Lichte F. E. and Skogerboe R. K. Anal. Chem. 1973 45 399. Instrumentation Laboratory Data Sheet 1979 09111 1/79. Pickett E. E. and Koirtyohann S . R. Anal. Chem. 1969,41,28A. Barch R. F. Adams D. M. Soloway A. H. Mechetner E. B. Alam F. and Anisuzzaman A. K. M. Anal. Chem. 1991,63,890. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Coedo A. G. Dorado T. Escudero E. and Cobo I. G. J. Anal. At. Spectrom. 1993 8 827. Molinero A. L. Ferrer A. and Castillo J. R. Talanta 1993 40 1397. Donaldson E. M. Talanta 1981 28 825. Ciba T. and Smolec B. Fresenius’ J. Anal. Chem. 1994,348,215. Hu B. Jiang Z. and Zeng Y. Fresenius’ J. Anal. Chem. 1991 340 435. Hlavacek I. and Hlavackova I. Microchim. Acta. 1989,111 309. Yamada K. Kujirai O. and Hasegawa R. Anal. Sci. 1993,9,385. Date A. R. and Gray A. L. Spectrochim. Acta Part B 1985 40 116. Jin Q. Zhu C. Boer M. W. and Hieftje G. M. Spectrochim. Acta Part B 1991 46 417. Aauilar J. F. C. Garcia R. P. Uria J. E. S. and Medel San A. Spectrochim. Acta Part B 1994 49 6 545. Sekerka I. and Lechner J. F. Anal. Chim. Acta 1990 234 199. Fang Z.-L. Flow Injection Atomic Spectroscopy ed. Buguera J. L. Marcel Dekker New York 1989 ch. 4. Tyson J. F. Spectrochim. Acta. Rev. 1991 169 14. Jin Q. Huang M. and Hieftje G. M. Microwave Plasmas in Analytical spectrometry Jilin University Press 1993 p. 190. Jin Q. Zhang H. Liang F. Jin Q. J. Anal. At. Spectrom. 1995 10 875. Samuelson O. Ion Exchange Separations in Analytical Chemistry Wiley New York 1963 p. 111. Pitts A. E. and Beamish F. E. Anal. Chem. 1969 41 1107. Govindaraju K. Anal. Chem. 1968 40 1. Veillon C. and Margshes M. Spectrochim. Acta Part B 1968 23 503. Hoare H. C. and Mostyn R. A. Anal. Chem. 1967,39 1153. Paper 5/05563F Received August 8 1995 Accepted January 2 1996 Journal of Analytical Atomic Spectrometry May 1996 Vol. 11 337