Chemical Vapor Generation Atomic Spectrometry Using

No evidence of hydride formation was observed from the Se(VI) and Te(VI). ...... According to the literature data,38 it can be assumed that Fe(III) ox...
0 downloads 0 Views 137KB Size
Anal. Chem. 2003, 75, 2591-2600

Chemical Vapor Generation Atomic Spectrometry Using Amineboranes and Cyanotrihydroborate(III) Reagents Alessandro D’Ulivo*,† Valeria Loreti,† Massimo Onor,† Emanuela Pitzalis,† and Roberto Zamboni‡

Laboratorio di Chimica Analitica Strumentale, C.N.R., Istituto per i processi chimico-fisici, Area della Ricerca di Pisa, Via G. Moruzzi, 1, 56124 Pisa, Italy, and Dipartimento di Chimica e Chimica Industriale, Universita` di Pisa, Via Risorgimento, 35, 56126 Pisa, Italy

Amineboranes of the type L-BH3 (L ) NH3; tert-BuNH2; Me2NH; Me3N) and sodium cyanotrihydroborate(III) (NaBH3CN) have been tested as derivatization reagents in the generation of volatile hydrides and elemental mercury following aqueous phase reaction with ionic species of Hg(II), As(III), As(V), Sb(V), Sb(III), Bi(III), Se(IV), Se(VI), Te(IV), and Te(VI). Continuous flow generation atomic absorption spectrometry coupled with a flameless quartz tube atomizer (T ) 25 °C) and a miniature argon-hydrogen diffusion flame atomizer were employed for the detection of mercury vapors and volatile hydrides, respectively. All of the reductants were able to reduce Hg(II) to the elemental state, giving sensitivities comparable to NaBH4 reduction. Under reaction conditions giving maximum sensitivity for hydride generation with NaBH4, only some amineboranes are able to produce volatile hydrides from all the elements. No evidence of hydride formation was observed from the Se(VI) and Te(VI). In general, the reducing power decreased in the order NaBH4 > H3N-BH3 > tert-BuNH2-BH3 > NaBH3CN g Me2HN-BH3 > Me3N-BH3. In comparison with THB, amineboranes and NaBH3CN allowed, in general, a better control of interference effects of Fe(III), Ni(II), Co(II), and Cu(II). Application to determination of mercury in certified reference material is reported. The most likely mechanism of reaction of borane complexes in chemical vapor generation is based on the direct action of hydrogen bound to boron. Chemical vapor generation coupled with atomic spectrometry represents one of the most powerful analytical tools for ultratrace determination of elements such as Hg, As, Sb, Bi, Se, Te, Ge, Sn, and Pb.1 Derivatization of the aqueous ions of the analyte to their corresponding volatile species, the hydride or the element, can be performed with the aid of different reaction systems. For mercury, the classical reduction scheme with stannous chloride represents one of the oldest chemical vapor generation tech* To whom correspondence should be addressed. E-mail: dulivo@ icas.pi.cnr.it. † C.N.R.. ‡ Universita ` di Pisa. (1) Deˇdina, J.; Tsalev, D. L. Hydride Generation Atomic Spectrometry; Wiley: Chichester, 1995; Chapter 2. 10.1021/ac020694p CCC: $25.00 Published on Web 04/24/2003

© 2003 American Chemical Society

niques.2 The other elements can be converted into the corresponding volatile hydrides by sodium or potassium salts of tetrahydroborate(III) (THB) under different acidic conditions or a metal/acid system.1,3 The use of sodium tetrahydroborate dramatically increased in popularity since the first reports published on 1973 by Schmidt and Royer,4 Pollock and West,5 and Fernandez.6 At present, THB is almost exclusively used for determination of hydride forming elements, and more recently its use has been extended to the generation of volatile species of Cd,7,8 In,9 Tl,10,11 Cu,12 Au,13 Ni,14 Ag13,15, and Zn.13,16 The major disadvantage of THB solutions is represented by its sensitivity to protic environments, which results in fast decomposition in aqueous solution. The acid hydrolysis of THB is a second-order process with rate constants of 1.6 × 106 and 2.03 × 106 L mol-1 s-1 at 25 and 30 °C, respectively.17,18 This high decomposition rate implies that slow hydrolysis of aqueous THB solutions occurs also in the presence of NaOH or KOH added as stabilizers.19 Using the THB reductant may entail lower selectivity because of its side reactions with matrix components (mainly transition metals) and some other hydride-forming elements.1,20 (2) Puk, R.; Weber, J. H. Appl. Organomet. Chem. 1994, 8, 293-302. (3) Nakahara, T. Prog. Anal. At. Spectrosc. 1983, 6, 163-223. (4) Schmidt, F. J.; Royer, J. L. Anal. Lett. 1973, 17, 17-23. (5) Pollock, E. N.; West, S. J. At. Absorpt. Newsl. 1973, 12, 6-9. (6) Fernandez, F. J. At. Absorpt. Newsl. 1973, 12, 93-97. (7) Sanz-Medel, A.; Valde´s-Hevia y Temprano, M. C.; Bordel Garcı´a, N.; Fernandez de la Campa, M. R. Anal. Chem. 1995, 67, 2216-2223. (8) Guo, X. W.; Guo, X. M. J. Anal. At. Spectrom. 1995, 10, 987-991. (9) Busheina, J. S.; Headridge, J. B. Talanta 1982, 29, 519-520. (10) Yan, D.; Yan, Z.; Cheng, G. S.; Li, A. M. Talanta 1984, 31, 133-134. (11) Ebdon, L.; Goodal, P.; Hill, S. J.; Stockwell, P.; Thompson, K. C. J. Anal. At. Spectrom. 1995, 10, 317-320. (12) Sturgeon, R. E.; Liu, J.; Boyko, V. J.; Luong, V. T. Anal. Chem. 1996, 68, 1883-1887. (13) Luna, A. S.; Sturgeon, R. E.; Campos, R. C. Anal. Chem. 2000, 72, 35233531. (14) Guo, X. M.; Huang, B.; Sun, Z.; Ke, R.; Wang, Q.; Gong, Z. B. Spectrochim. Acta, Part B 2000, 55, 943-950. (15) Matousˇek, T.; Deˇdina, J.; Vobecky´, M. J. Anal. At. Spectrom. 2002, 17, 52-56. (16) Sun, H.; Suo, R.; Lu, Y. Anal. Chim. Acta 2002, 457, 305-310. (17) Gardiner, J. A.; Collat, J. W. J. Am. Chem. Soc. 1965, 87, 1692-1700. (18) Agterdenbos, J.; Bax, D. Anal. Chim. Acta 1986, 188, 127-135. (19) Narsito; Agterdenbos, J.; Santosa, S. J. Anal. Chim. Acta 1990, 237, 189-199. (20) Smith, A. E. Analyst 1975, 100, 300-306.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003 2591

Cyanoborohydride (CBH), in the form of alkaline salts (MBH3CN; M ) Li, Na, K), is a reducing agent milder than THB, and it is well-known and widely employed in organic chemistry for its slow decomposition in protic media.21 Acid hydrolysis of lithium cyanoborohydride is ∼8 orders of magnitude slower than THB at room temperature.22 CBH has been employed for selective reduction, oximes, enamines, reductive amination of aldehydes and ketones, and reductive alkylation of amines and hydrazines.23 It is also used as a reducing agent for transition metal cations, such as, for example Ag(I) and Cu(II), that can be reduced to the metallic state.24 The only analytical application of this reagent to trace element determination is given by Brown et al.25 They demonstrated the capacity of CBH to convert traces of selenium and arsenic to the corresponding hydrides and the capacity of CBH to mask interferences from some transition metals. Since that time, the use of CBH has been not reported in the literature, probably because of the hazard related to the evolution of toxic HCN during its acid hydrolysis. The resistance to acid hydrolysis, typical of CBH, can be found in many other borane complexes belonging to the class of amineboranes, R3N-BH3 (RdH, alkyl). These are selective reducing agents well-known among organic chemists and characterized also for their slow decomposition in acidic media.26,27 Some amineboranes (L-BH3) have been proven to be very stable in the solid state (L ) NH3; Me2NH; Me3N; tert-BuNH2), even if heated at temperatures around 110 °C, and to present suitable solubility in water.26,27 The applications of amineboranes are essentially in the same field as CBH and are mostly concerned with organic chemistry.27-30 Their reducing power has also been employed in electroless plating solutions, since these compounds are able to slowly reduce many transition metals cations to the pure metallic form without production of metal borides.26 The reducing power as well as the resistance to acid hydrolysis of amineboranes can be modulated by an appropriate choice of the amino group.26,27 To the best of our knowledge, there are no studies reporting on the behavior of amineboranes as reagents for the chemical generation of volatile species of mercury and the elements forming volatile hydrides. This paper reports a study, performed by atomic absorption spectrometry (AAS), on the generation of mercury vapors and volatile hydrides following the reaction of their aqueous ions at trace levels with ammonia borane (AB), tert-butylamineborane(TBAB), dimethylamineborane (DMAB), and trimethylamineborane (TMAB). The use of CBH has been extended to mercury, antimony, bismuth, and tellurium after the first report of Brown et al.25 on arsenic and selenium. Comparison with the performances obtained by THB are reported, and the potentialities of (21) Borch, R. F.; Durst, H. D. J. Am. Chem. Soc. 1969, 91, 3996-3997. (22) Kreevoy, M. M.; Hutchins, E. C. J. J. Am. Chem. Soc. 1969, 91, 43294330. (23) Lane, C. F. Synthesis 1975, 135-146. (24) Berschied, J. B.; Purcell, K. F., Jr. Inorg. Chem. 1970, 9, 624-629. (25) Brown, R. M.; Fry, R. C., Jr.; Moyers, J. L.; Northway, S. J.; Denton, M. B.; Wilson, G. S. Anal. Chem. 1981, 53, 1560-1566. (26) Lane, C. F. Aldrichim. Acta 1973, 6, 51-58. (27) Hutchins, R. O.; Learn, K.; Nazer, B.; Pytlewski, D. Org. Prep. Proced. Int. 1984, 16, 335-372. (28) Moorman, A. E. Synth. Commun. 1993, 23, 789-795. (29) Cabacungan, J. C.; Ahmed, A. I.; Feeney, R. E. Anal. Biochem. 1982, 124, 272-278. (30) Billman J. H.; McDowell J. W. J. Org. Chem. 1962, 27, 2640-2643.

2592 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

Figure 1. Schematic representation of the experimental apparatus for CVG.

amineboranes and CBH in diagnostic applications are discussed. The method has been tested to determine trace mercury in certified reference materials. EXPERIMENTAL SECTION Instrumentation. Atomic absorption experiments were performed with a Perkin-Elmer model 2380 atomic absorption spectrometer. Electrodeless discharge lamps (EDL System II, Perkin-Elmer) were employed for all the investigated elements and were operated at the current recommended by manufacturer. Absorbance measurements were performed at the following wavelengths and spectral bandwiths: As, 193.7/0.7 nm; Sb, 217.6/ 0.2 nm; Bi, 223.1/0.2 nm; Se, 196.0/2 nm; Te, 214.3/0.2 nm; and Hg, 253.6/0.7 nm. A quartz T tube (130 mm long, 8-mm i.d.) operated at room temperature was employed as the atomic cell for mercury. For hydride-forming elements, the atomizer was a miniature argon-hydrogen diffusion flame31 supported on a quartz tube (6.5-mm i.d.). Atomic fluorescence determination of mercury was performed by a continuous flow nondispersive atomic fluorescence spectrometer described previously32 and using the same argonhydrogen miniature flame atomizer used for AAS experiments with hydride forming elements. The continuous flow generator (Figure 1) was realized with a peristaltic pump (Ismatec pump head MS/CA4-12 fitted on a Masterflex L drive (H-7519-25). Ismatec Tygon microtubings of appropriate diameter were used for propelling sample, reductant, and waste solutions. Sample and reductant flow rates were 4 and 2 mL min-1, respectively. The solutions merged in the first mixing T junction, T1 (Kel-F, 0.8-mm i.d, Ismatec), and were allowed to react in the reaction loop (0.1-2.5 mL, Teflon or Teflon PFA). The reaction mixture coming from the reaction loop merged with Ar stripping gas in a second mixing T-junction, T2 (Kel-F, 0.8-mm i.d, Ismatec), and the stripping of gaseous products continued in the stripping loop (Teflon PFA, 1.5-mm i.d., 30 cm long). The reaction mixture was then delivered to the gas liquid separator (borosilicate glass, 60 mm long, 10-mm i.d.). An auxiliary flow of gas was introduced into the gas liquid separator. The composition and the flow rates of gases varied with the atomizer used: for the quartz T-tube atomizer, 300 mL min-1 of Ar (added through T2) and 180 mL min-1 of Ar as auxiliary gas added through gas liquid separator; for the miniature flame atomizer, 125 mL min-1 (31) Deˇdina, J.; D’Ulivo, A.; Lampugnani, L.; Matousˇek, T.; Zamboni, R. Spectrochim. Acta, Part B 1998, 53, 1777-1990. (32) Bramanti, E.; D’Ulivo, A.; Lampugnani, L.; Raspi, G.; Zamboni, R. J. Anal. At. Spectrom. 1999, 14, 179-185.

of Ar (added through T2) and 100 mL min-1 of Ar and 175 mL min-1 of hydrogen, both added through the gas liquid separator as auxiliary gas. Gas flow rates were controlled by calibrated ball rotameters. Sample digestions were performed with pressurized Teflon bombs (Parr, USA), heated in a microwave oven (Supratronic, Miele). Chemicals. Stock aqueous solutions of amineboranes, CBH, and THB were prepared by using the following solid reagents (Aldrich): borane-ammonia complex (Aldrich, assay 90%), borane-tert-butylamine complex (Aldrich, pellets, assay 97%), borane-dimethylamine complex (Aldrich, assay 97%), boranetrimethylamine complex (Aldrich, assay 97%), sodium cyanoborohydride (Aldrich, assay 97%), sodium borohydride (BDH, pellets, reagent for AAS, assay 96%). The solubility in water of amineboranes varies in a wide range of concentrations. The least soluble are TMAB (∼0.12 mol dm-3) and TBAB (∼0.2 mol dm-3), followed by AB and DMAB, for which a solution more concentrated than 0.2 mol dm-3 could be prepared. AB and THB solutions required stabilization by addition of 0.1 mol dm-3 NaOH and microfiltration on a 0.45-µm membrane. For safety reasons, also, the 1 mol dm-3 CBH solution was made 0.1 mol dm-3 in NaOH in order to prevent evolution of HCN. Solutions of TBAB, DMAB, and TMAB are clear and colorless and did not required filtration. All of the prepared solutions are stable for at least 2 weeks at room temperature, except 1.3 mol dm-3 THB solution, which is stable for 1 week only if stored in a refrigerator at 0-4 °C. Stock solutions of 1000 µg mL-1 of the analytes Hg(II), As(III), Sb(III), Bi(III), Se(IV), and Te(IV) and of 10 000 µg mL-1 for the interfering elements Fe(III), Ni(II), Co(II), and Cu(II) were prepared by dilution of Merck concentrates and a stock solution of 1000 µg mL-1 of As(V) (Carlo Erba). A stock solution of 100 µg mL-1 of Sb(V) was freshly prepared by oxidation of Sb(III) stock solution with KBrO3. Se(VI) and Te(VI) were prepared by dissolving Na2SeO4‚10H2O (BDH) and K2TeO4 (Heraeus) in water. All of the other chemicals were of analytical grade or higher. Ultrapure chemical reagents, 10 mol dm-3 HCl, 14 mol dm-3 HNO3, and 9.5 mol dm-3 H2O2 were employed for sample digestion and solution acidification. Water purified with a Milli-Q system (Purelab Pro, USF) was used in all the operations. Dogfish Muscle Certified Reference Material (DORM-2) was obtained from the Institute of National Measurements Standards (NRCC, Canada). Estuarine Sediment Certified Reference Material (CRM580) was obtained from the Institute for Reference Materials and Measurements (IRMM, Belgium). Quantitative Determination of Amineboranes. The general method used for the determination of amineboranes by iodine titration26 was adopted for TBAB.

R3N-BH3 + 3I2 + 3H2O f R3NH+ + 6I - + 5H+ + B(OH)3 (1)

The TBAB solution was buffered with acetic acid and sodium acetate at pH 5, then the titration was directly performed with iodine solution in the presence of starch as end-point indicator. The end-point of the titration is sharp, and the reaction is fast and quantitative.

The assay of AB solution was determined by iodimetric method as previously reported for THB titration.17 An excess of KIO3 was added, and the solution was stirred for 3 min,

H3N-BH3 + IO3- f I - + NH3 + B(OH)3

(1a)

then the solution was acidified, and an excess of KI was added. The evolved iodine was titrated with thiosulfate solution. Determination of Mercury in Certified Reference Materials. Between 0.15 and 0.2 g of DORM-2 was accurately weighed directly into Teflon bombs, and 1 mL of 9.5 mol dm-3 H2O2 and 3 mL of 14 mol dm-3 HNO3 were added. The samples were allowed to react overnight, then the bombs were capped and put in the microwave oven. The heating cycle was 1 min at 850 W followed by 1 min at 0 W for 10 cycles. After cooling, the digests were transferred into a 50-mL calibrated flask and diluted with water to the mark. For CRM580, 0.9-1.0 g of sample was accurately weighed directly inside PARR Teflon bombs, treated with 8 mL of aqua regia (2 mL of 14 mol dm-3 HNO3 + 6 mL 12 mol dm-3 HCl), and allowed to react overnight at room temperature. The bombs were capped, and the same microwave heating cycle described for DORM2 was then performed. Subsamples for determination of mercury were prepared by further dilutions with either 1 mol dm-3 HCl or 0.1 mol dm-3 HCl according to the reductant used for determination. The analyte addition technique was used for Hg(II) quantification in subsamples. Safety Precautions for Cyanoborohydride. Acid hydrolysis of CBH led to formation of toxic HCN. Considering that this reaction, also, is relatively slow in acid conditions, the unreacted CBH will be present in acid waste solution, from which the evolution of toxic gases still continues. For this reason, two precautions were adopted. The former is the use of an efficient fume extractor above the atomizer compartment, thus eliminating volatile toxic compounds formed during CBH hydrolysis. The latter is the collection of the waste solutions in a tank containing concentrated sodium hydroxide. For the destruction of cyanides and CBH, an excess of NaClO was added to alkaline waste solutions, and they were brought to dryness by gentle heating under a fume hood. RESULTS AND DISCUSSION Reactions and Stability of Borane Complexes. According to the electrochemistry of boron compounds33 the oxidation halfreactions of the borane complexes considered in the present work are

BH4- + 8OH- f BO2- + 6H2O + 8e-

(2)

BH3CN- + 3H2O f B(OH)3 + 5H+ + HCN + 6e- (3)

The oxidation half-reaction for amineboranes, R3N-BH3, would follow the same scheme as reaction 3.

R3N-BH3 + 3H2O f B(OH)3 + 5H+ + R3NH+ + 6e- (4) (33) Morris, J. H.; Gysling, H. J.; Reed, D. Chem. Rev. 1985, 85, 51-76.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2593

Table 1. Rate Constants for Acid Hydrolysis of Some Borane Complexes in Water at 25 °C borane complex

k (L mol-1 s-1)a

time for total hydrolysis (s)b

kBH4/k

ref

NaBH4 NH3-BH3 tert-BuNH2-BH3 LiBH3CN Me2NH-BH3 Me3N-BH3

1.6 × 106 6.0 0.87 1.7 × 10-2 9.7 × 10-3 1.2 × 10-4

4.3 × 10-6 1.1 7.9 406 712 5.7 × 104

1 2.7 × 105 1.8 × 106 9.4 × 107 1.6 × 108 1.3 × 1010

17 35 36 22 36 35

a Rate constant for second-order reaction. b Time calculated for 99.9% decomposition at 1 mol dm-3 H+ in a pseudo-first-order process.

Then, for one mole of reductant, 25% fewer moles of electrons will be produced in the oxidation of both CBH and amineboranes in comparison with the oxidation of THB. The borane complexes considered in the present work, BH3X- (X ) H -, CN -) and L-BH3 (L ) NH3, tert-BuNH2, Me2NH, Me3N), undergo acid hydrolysis according to the general reaction schemes17,23,34 -

+

BH3-X + H + 3H2O f HX + 3H2 + B(OH)3

(5)

L-BH3 + H+ + 3H2O f LH+ + 3H2 + B(OH)3

(6)

Table 1 reports the available literature data on kinetic constants for the acid hydrolysis in water. The time for total hydrolysis and the ratio of the rate constants relative to THB are calculated and reported in Table 1 because they are useful for discussion of the results. The reaction rates for acid hydrolyis decisively depend on the nature of complexing group. The withdrawing effect of the complexing group affects the hydridic character of hydrogen atoms bound to boron. For aliphatic amineboranes, both the reducing power and the reaction rates for acid hydrolysis decrease when the substitution of the amino group increases: H3NBH3 > RH2NBH3 > R2HNBH3 > R3NBH3.27,35,36 In general, the rate of hydrolysis also gives an indication of the reducing capacity of the borane complex.27 An important characteristic of amineboranes is that their reducing properties are dramatically enhanced in acid media.26,27,37 It is well-known that TMAB behaves as a reducing agent only with reluctance, but the reduction rates of organic and inorganic substrates, for example, ketones27 and HNO2,37are dramatically enhanced in the presence of H+ and Lewis acids, which activate the subtrates for the hydride attack. The resistance to acid hydrolysis allows preservation of aqueous solutions of amineboranes for long storage periods without particular precautions. Titration of neutral aqueous solutions of 0.21 mol dm-3 TBAB, stored at 25 °C, showed that after 1 and 4 weeks, the fraction decomposed was as low as 4 and 9%, respectively. On the basis of the data reported in Table 1, comparable or better stability should be obtained for CBH, DMAB, (34) Ryschkewitsch, G. E. J. Am. Chem. Soc. 1960, 82, 3290-3294. (35) Kelly, H. C.; Marriott, V. B. Inorg. Chem. 1979, 18, 2875-2878. (36) Kelly, H. C.; Marchelli, F. R.; Giusto, M. R. Inorg. Chem. 1964, 3, 431437. (37) Bell, K. E.; Kelly, H. C. Inorg. Chem. 1996, 35, 7225-7228. (38) Sanz, J.; Martinez, M. T.; Galban, J.; Castillo, J. R. J. Anal. At. Spectrom. 1990, 5, 651-655.

2594

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

Figure 2. Reduction curves for mercury obtained using AAS with a quartz tube atomic cell. Hg(II), 5 × 10-7 mol dm-3 in 1 mol dm-3 HCl solution. Reaction loop, 0.1 mL. The absorbance given by the reduction with 10-3 mol dm-3 THB solution was taken as reference signal. THB ) NaBH4; AB ) NH3-BH3; TBAB ) tert-BuNH2-BH3; CBH ) NaBH3CN; DMAB ) Me2NH-BH3; and TMAB ) Me3N-BH3.

and TMAB. Stability of a 0.19 mol dm-3 AB solution, prepared in 0.1 mol dm-3 NaOH and stored at 25 °C, was verified by titration, and the fraction decomposed was only 2 and 4% after 2 and 5 weeks, respectively. Under similar conditions, THB solutions decompose quite fast19 and must be kept cold, around 0 °C or below. CVG of Mercury and Antimony. Mercury and antimony are among the elements that are more easily converted to a volatile form using THB reduction. According to literatura data, the maximum sensitivity for Hg(II) in 1 mol dm-3 HCl was obtained for THB concentrations greater than 10-6 mol dm-3,32 and for Sb(III) in 3 mol dm-3 HCl, a volatilization of 88 and 97% was obtained for THB concentrations of 2.6 × 10-4 and 2.6 × 10-3 mol dm-3, respectively.19 The first experiments were, therefore, performed with mercury and antimony, and the results are reported in Figures 2 and 3, respectively. Parallel experiments were performed with THB reduction in order to assess the signal level corresponding to the maximum achievable sensitivity. The signal levels obtained with THB concentration of 10-3 and 0.25 mol dm-3 for mercury and antimony, respectively, were considered suitable indicators of the maximum sensitivity achievable, and are indicated with dotted lines in Figures 2 and 3. The concentration of reductant solutions was varied in the range of 10-9-10-1 mol dm-3 for mercury, and in the range of 10-5-10-1 mol dm-3 for antimony. This allowed us to obtain reduction curves, which are useful for the description of reducing capacity of the reagents investigated. The use of reductant solution more concentrated then 0.1 mol dm-3 was hindered in some cases by the solubility and in other cases, by excessive foaming produced by the alkylamineboranes. All of the reagents are able to produce volatile mercury vapors by reaction with 5 × 10-7 mol dm-3 solution of Hg(II), and stibine, by reaction with 8 × 10-7 M solution of Sb(III), but considerable differences in their respective reducing capacities are evident. From Figures 2 and 3, it is possible to obtain, by interpolation, an estimation of reductant concentration, CR, giving 50% of the maximum sensitivity. For mercury, the values of CR (in mol dm-3) are THB, 3 × 10-7; AB, 1.5 × 10-6; TBAB, 5 × 10-6; DMAB, 1.5

Table 2. Relative Signal Intensities Obtained by Reaction of Hydride Forming Elements with Different Reagents Reductant (0.1 mol dm-3) elementa As(III) As(V) Sb(III) Sb(V) Bi(III) Se(IV) Se(VI) Te(IV) Te(VI)

HCl (mol dm-3) 1 1 1 1 1 4 4 4 4

THB (0.25 mol dm-3) b

S0

75 ( 2 26 ( 1 34 ( 1 55 ( 3 42 ( 2

THB

AB

TBAB

CBH

DMAB

R

R

R

R

R

93 ( 2 23 ( 1 101 ( 2 15 ( 2 93 ( 2 96 ( 4

78 ( 2 15 ( 1 92 ( 3 44 ( 2 86 ( 2 58 ( 3 nd 29 ( 3 nd

65 ( 2 4(2 88 ( 4 30 ( 5e 63 ( 2 32 ( 4 nd 9(4 nd

6(1 nd 82 ( 3 2(1 nd 11 ( 2 nd nd nd

2(1 nd 74 ( 2 nd nd nd nd nd nd

Rc 100 ( 2 47 ( 1 100 ( 3 35 ( 2 100 ( 3 100 ( 5 nd 100 ( 5 nd

58 ( 2

TMAB R ndd nd 14 ( 1 nd nd nd nd nd nd

a 100 ng mL-1; reaction loop, 0.5 mL. b Signal in arbitrary units. c S/S × 100; uncertainties are standard deviations, n ) 3. d Not detectable, e 0 0.5. e Double peak signal.

Figure 3. Reduction curves for antimony obtained using AAS with a miniature diffusion flame atomizer. Sb(III), 8 × 10-7 mol dm-3 in 1 mol dm-3 HCl solution. Reaction loop, 0.5 mL. The absorbance given by the reduction with 0.25 M THB solution was taken as reference signal. THB ) NaBH4; AB ) NH3-BH3; TBAB ) tert-BuNH2-BH3; CBH ) NaBH3CN; DMAB ) Me2NH-BH3; and TMAB ) Me3N-BH3.

× 10-4; CBH, 2 × 10-4; and TMAB, 2 × 10-2. For antimony, the values of CR are THB, 7 × 10-5; AB, 2 × 10-4; TBAB, 5 × 10-4; CBH, 2 × 10-2; and DMAB, 4 × 10-2, but for TMAB, the value of CR cannot be estimated from the present data. There is a good linear correlation between CR and the rate constant of the borane complexes investigated, as showed in Figure 4. This behavior confirms the general trend reported in the literature about the relation between the structure and the reducing power of amineborane complexes (see previous paragraph). The only exception is THB, which appears not to belong to the family of the borane complexes investigated. This discrepancy could be due to the fact that both the decomposition rate of THB and its reaction rate with the element are much faster than the mixing rate of the solutions in the T junction. In this case, the reactions are controlled not by their respective kinetic constants, but rather, by the mixing rate of solutions. Another limitation of correlation is that the concentration of reductant cannot be lower than the limit determined by the reaction stoichiometry. Therefore, the correlation plots possess some meaning only for those reductants whose reaction rates are slower than the mixing rates of reagents.

Figure 4. Correlation between the rate constant for acid hydrolysis of borane complexes, k, and their concentration, CR, giving 50% of the maximum sensitivity.

CVG of Hydride-Forming Elements. The study was extended to other hydride-forming elements, present in different oxidation states, using the same experimental setup employed for Sb(III) studies. The relative signals obtained with 0.1 mol dm-3 reductant are reported in Table 2. The signals obtained with 0.25 mol dm-3 THB are taken as the reference, indicating the maximum achievable sensitivity. Sample acidities were arbitrarily chosen, taking into account the range of acidity, giving maximum yield in generation of hydrides using THB reduction.1 The reducing capacity of the reductants follows the same order as those observed for mercury and antimony, with the extreme case of TMAB, which is not able to produce detectable signals, except for Sb(III). The ease of reduction for individual elements appears to decrease in the order Sb(III), As(III), Bi(III), Se(IV), Sb(V), Te(IV), and As(V). No signals were detected for Se(VI) and Te(VI), as was also observed for THB reduction. Some interesting results were obtained in the reduction of the trivalent state of arsenic and, in particular, of antimony. In some cases, the selectivity appears to be more pronounced versus the THB reduction, and in the case of antimony, the sensitivity ratio Sb(III)/Sb(V) observed using DMAB reduction is ∼150 or greater. These results indicate that, apart from the individual reducing capacity of different borane complexes, the general reactivity of amineboranes and CBH toward trace element reduction is very Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2595

Table 3. Effect of Reaction Loop Volume on Relative Signal Intensities for Mercury and Hydride-Forming Elementsa

elementb

loop volume (mL)

HCl (mol dm-3)

Hg(II)

THB (0.25 mol dm-3)

AB

0.1 0.5 2.5

1 1 1

100 ( 3 100 ( 2 100 ( 5

98 ( 4 100 ( 2 87 ( 4

0.1 0.5 2.5

1 1 1

100 ( 3 95 ( 2 94 ( 5

37 ( 4 95 ( 3 96 ( 4

0.5 0.5

1 5

101 ( 2 96 ( 3

92 ( 3 100 ( 4

0.1 0.5 2.5 0.5

1 1 1 5

95 ( 2 100 ( 3 93 ( 5 90 ( 3

32 ( 2 78 ( 2 75 ( 3 95 ( 3

0.1 0.5 2.5 0.5

1 1 1 5

100 ( 2 90 ( 2 90 ( 3 90 ( 2

92 ( 2 86 ( 2 85 ( 3 30 ( 2

0.1 0.5 2.5 0.5

4 4 4 10

100 ( 2 93 ( 2 98 ( 4 90 ( 2

44 ( 3 58 ( 2 57 ( 4 45 ( 2

0.1 0.5 2.5 0.1

4 4 4 10

100 ( 3 87 ( 2 75 ( 3 93 ( 2

33 ( 2 29 ( 2 19 ( 3 32 ( 1

Sb(III)

As(III)

Bi(III)

Se(IV)

Te(IV)

reductant and concentration TBAB CBH DMAB (10-5

dm-3)

mol 5(2 37 ( 2 100 ( 4 (10-3 mol dm-3) 5(3 nd 71 ( 3 nd 100 ( 4 37 ( 3 (0.1 mol dm-3) 88 ( 4 82 ( 3 93 ( 2 100 ( 3 (0.1 mol dm-3) 10 ( 3 nd 65 ( 2 6(1 97 ( 4 26 ( 1 85 ( 3 92 ( 2 (0.1 mol dm-3) 10 ( 3 nd 63 ( 2 nd 97 ( 4 26 ( 4 10 ( 1 nd (0.1 mol dm-3) 10 ( 2 nd 32 ( 3 11 ( 1 36 ( 4 19 ( 1 58 ( 2 60 ( 2 (0.1 mol dm-3) 10 ( 1 nd 9(4 nd nd nd 35 ( 2 63 ( 3

69 ( 3 101 ( 3 100 ( 4

TMAB

4(2 35 ( 3 95 ( 4

ndc nd 8(3

nd nd 20 ( 4

nd nd nd

74 ( 2 95 ( 2

14 ( 1 7(3

nd 2(1 16 ( 4 50 ( 3

nd nd nd nd

nd nd nd nd

-

nd nd nd 51 ( 3

nd nd nd nd

nd nd nd 21 ( 2

nd nd nd nd

a AAS measurements. The signal obtained with THB 0.25 mol dm-3 under the same conditions of acidity and reaction loop volume was taken as the reference signal. b 100 ng mL-1. Uncertainties are standard deviations, n ) 3. c Not detectable, e 0.5.

similar to that of THB and can be summarized as follows: (i) The reactivity of amineboranes and CBH with Hg(II) is largely independent of acidity (see section Determination of Mercury), and this has been also demonstrated for THB32 in our previous work; (ii) for all of the reagents, As(III) and Sb(III) gave better sensitivity than did As(V) and Sb(V); (iii) only Se(IV) and Te(IV), but neither Se(VI) nor Te(VI), are reduced to hydrides by all of these reagents. Effect of Reaction Time. The effect of reaction time was tested by varying the contact time between sample and reagent in the reaction loop (see Figure 1). Three different volumes of the reaction loop, 0.1, 0.5, and 2.5 mL, were tested. Some experiments were also performed at different acidities using a 0.5-mL reaction loop. The results are reported in Table 3. The effect of reaction loop volume is not critical for THB reduction and, as expected, this reagent quickly derivatizes the analytes. Reaction times longer than necessary may result in sensitivity losses that are more evident for thermally unstable hydrides, such as, for example hydrogen telluride. For amineboranes and CBH reduction, in general, the sensitivity substantially improves with an increase in reaction time, except for tellurium, for the same reason discussed above for THB. This points out that amineboranes and CBH are slower than THB in the reaction with analytes. For TMAB, only weak signals are obtained in the cases of mercury and antimony, and there was no detectable signal for the other elements. 2596 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

In addition to reaction time and reagents’ concentrations, acidity represents an important parameter for the conversion of the element in the corresponding volatile derivatives. For amineboranes and CBH, the increase in acidity resulted in a dramatic change of sensitivity. Strong positive effects are evident for As(III), Se(IV) and Te(IV) while negative effects are observed for Bi(III) (see Table 3). The effect of acidity appears to be more pronounced for those reductants that are more resistant to acid decomposition, such as, for example, CBH and DMAB. In particular, the sensitivity achievable by CBH reduction of Se(IV) and Te(IV) at elevated acidity (10 mol dm-3 HCl) is better or comparable to those obtained by the other amineboranes. Interferences. In CVG techniques employing THB as derivatizing agent, a number of interferences have been reported.1,3 The most serious interfering effects take place in the liquid phase during the reaction of THB with the sample solution and are observed in the presence of some transition metal species, such as, for example, Fe(III), Ni(II), Co(II), and Cu(II), and precious metals, such as Au(III), Pt(IV), Pd(II), and Ag(I), and the hydride forming elements. The magnitude of a given interference effect is related to the identity of both analyte and interfering element and to the experimental conditions. The formation of finely dispersed metal or metal boride colloids, which are able to adsorb and decompose the hydrides, is considered the most important mechanism of action of transition metal interferences.1,39 In comparison with THB, the use of CBH was reported to remove

Table 4. Tolerance Limits for Fe(III), Ni(II), Co(II) and Cu(II) in CVG-AAS

analyte

interfering element

Hg(II)c

THB

AB

Fe(III) Ni(II) Co(II) Cu(II)

>1000 >1000 >1000 600

>1000 >1000 >1000 >1000

Fe(III) Ni(II) Co(II) Cu(II)

>3000 160 300 400

>3000 >3000 >3000 350

Fe(III) Ni(II) Co(II) Cu(II)

1 140 150 20

1.5 >1000 >1000 >1000

Fe(III) Ni(II) Co(II) Cu(II)

>1000 7 20 0.2

>1000 >1000 >1000 2

As(III)d

Sb(III)e

Bi(III)f

tolerance limits (mg L-1)a,b TBAB 10-3

dm-3

reductant mol >1000 >1000 >1000 >1000 reductant 0.1 mol dm-3 >3000 >3000 >3000 250 reductant 0.1 mol dm-3 1.5 >1000 >1000 >1000 reductant 0.1 mol dm-3

CBH

DMAB

>1000 >1000 >1000 >1000

>1000 >1000 >1000 >1000

>3000 >3000 >3000 250

>3000 >3000 >3000 300

1 >1000 >1000 >1000

1.5 >1000 >1000 >1000

>1000 >1000 650 100

a Concentration of interfering element giving 10% signal depression. b AAS measurements, 0.5 mL reaction loop. c 100 ng mL-1 Hg(II) in mol dm-3 HCl for AB and TBAB, and in 0.1 mol dm-3 for CBH and DMAB. d 200 ng mL-1 As(III) in 5 mol dm-3 HCl. e 200 ng mL-1 Sb(III) in 5 mol dm-3 HCl. f 200 ng mL-1 Bi(III) in 1 mol dm-3 HCl.

the interference of Ni(II) in arsenic determination and to allow better control of Cu(II) interference in selenium determination.25 There is no information about the interferences affecting CVG when amineboranes are used as derivatizing reagents. However, the use of amineboranes in electroless plating solution is reported to slowly reduce transition metal cations to pure metals without the formation of borides.26 This is a good general indication of their potentialities in the control of interferences in CVG technique. To have a comparative evaluation of the capacities of THB, CBH, and amineboranes to control the interferences, the effect of concomitants Fe(III), Ni(II), Co(II), and Cu(II) on the sensitivity of several analytes was determined using CVG-AAS. These transition metal cations can be considered as “priority offenders”, taking into account, also, their relative abundance in many samples. The analytes Hg(II), As(III), Sb(III), and Bi(III) were selected because almost quantitative conversion could be approached with at least one of the reductants by using a relatively short reaction time (0.5-mL reaction loop). For mercury, experimental conditions obtained by experimental design optimization were employed (see section on Determination of Mercury). The results are reported in Table 4. The appearance of interference effects can be correlated with the formation of a finely dispersed precipitates in the reaction system, which has been often observed for THB. With amineboranes and CBH, the formation of precipitates was not observed along reaction system tubings, T-junctions, and sometimes even in the gas liquid separator. In many cases, the formation of precipitates was observed only in the vessel collecting the waste solution. The only case in which all of the reagents fail to control the interference is the reduction of Sb(III) in the presence of (39) Deˇdina, J. In Flow Analysis with Atomic Spectrometric Detection; Sanz-Medel, A., Ed.; Elsevier: Amsterdam, 1999; Chapter 8.

Fe(III). In addition, no formation of precipitate was observed in this case. According to the literature data,38 it can be assumed that Fe(III) oxidizes part of the antimony to the pentavalent state, from which the conversion to hydride is less efficient (see Table 2). The substantial improvement of tolerance limits obtained by amineboranes and CBH reduction is of great importance. It indicates that a fast separation of gaseous analytes from a reaction mixture is not as critical as for THB reduction. It can be concluded that if amineboranes and CBH are slower than THB in the conversion of the analytes in the corresponding volatile species, they are even slower in reacting with metal ions to form interfering species. In other words, they are more selective, which is a typical feature of amineboranes and CBH chemistry. Determination of Mercury. The dependence of sensitivity from both the acidity of Hg(II) solution and reductant concentration was better investigated in the case of AB, TBAB, CBH, and DMAB. No further studies were performed with TMAB because of its poor reducing capacity. The acidity was varied in the range of pH 0-7, and the concentration of reductant was varied in the range of 10-2-10-8 mol dm-3. A reaction loop of 0.5 mL was employed in all experiments. To minimize the number of experiments, the statistic model of the cross-centered experimental design was adopted, and the nine couples of experimental points used are reported in Table 5. The results are given as contour plots in Figure 5. The contour plots, even if obtained with a limited number of experimental points, are useful for the description of the general characteristic of the reaction systems. The first observation is that the sensitivity is largely independent of the acidity. This is completely true for AB and TBAB, while for CBH and DMAB, the dependence on acidity is more pronounced at the highest reductant concentration investigated. For application to real sample analysis and for interference studies, it was decided Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2597

Figure 5. Contour plots obtained with central composite experimental design reporting the dependence of relative AAS signals for Hg(II) reduction as a function of different reducing agents. Conditions: quartz tube atomic cell; Hg(II), 5 × 10-7 mol dm-3; reaction loop, 0.5 mL. Table 5. Couples of Points Used for Cross-Centered Experimental Design point H+ mol dm-3 a reductant mol mol-3

1 1

2 1.8 × 10-2

3 1.8 × 10-2

10-5

3.2 × 10-4

3.2 × 10-4

4 3.2 × 10-4 10-2

5 3.2 × 10-4 10-5

6 3.2 × 10-4 10-8

7 5.6 × 10-6

8 5.6 × 10-6

9 10-7

3.2 × 10-4

3.2 × 10-4

10-5

a Acidity controlled by HCl for points 1-3, Acetic acid-sodium acetate buffer for points 4-8, NaH PO -Na HPO buffer for point 9. For points 2 4 2 4 4-9, sample solutions contain 0.05 mol dm-3 NaCl.

Table 6. AAS Calibration Graphs for Mercury Using Different Reductantsa reductantb

intercept (absorbance) A × 104

slope (mL ng-1) B × 104

r2

THBc ABc TBABc CBHd DMABd

2.9 ( 3 6.2 ( 3 -1.7 ( 2 0.5 ( 2 -1.8 ( 4

8.52 ( 0.07 9.32 ( 0.06 8.79 ( 0.04 8.55 ( 0.04 9.17 ( 0.09

0.99972 0.99981 0.99988 0.99988 0.99964

a QT-AAS; linear fitting for six concentrations in the range 0-100 ng mL-1. b 10-3 mol dm-3. c Sample acidity, 1 mol dm-3 HCl. d Sample acidity, 0.1 mol dm-3.

to use a sample acidity of 1 mol dm-3 HCl for AB and TBAB and 0.1 mol dm-3 HCl for DMAB and CBH. In all cases, the concentration of reductant was 10-3 mol dm-3. The calibration curves obtained under these experimental conditions were compared with that obtained by the reduction of Hg(II) in 1M HCl 2598 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

with 10-3 mol dm-3 THB. The comparison is shown in Table 6. In any case, using quartz tube atomic cell and AAS detection (QTAAS), the shape of the calibration graphs and the slope of the linear portion did not differ significantly. The detection limits were in the range of 1-1.5 ng mL-1 for all reductants employed, and linearity was up to 100 ng mL-1 of Hg(II). Under the same experimental conditions, better analytical performance could be obtained using miniature diffusion flame atomizer and nondispersive atomic fluorescence detection (MDF-AFS), with a 0.08-0.1 ng mL-1 detection limit and linearity up to 500 ng mL-1. Determination of mercury in certified reference materials was performed by QT-AAS for CRM580 river sediment, and by MDFAFS for DORM2 dogfish muscle tissues. The results are reported in Table 7. For DORM-2, THB, AB, and TBAB gave results that were in good agreement with the certified values. Both CBH and DMAB gave results lower than the certified value. The Student t-test at 95% confidence level confirmed that all results were not significantly different from the certified value, except for the

Table 7. Determination of Mercury in Certified Reference Materials Using Different Reductantsa Hg found (µg g-1)b

sample

Hg certified (µg g-1)c

THB

AB

TBAB

CBH

DMAB

DORM-2d CRM580e

4.64 ( 0.26 132 ( 3

4.6 ( 0.7 (5) 132 ( 14 (5)

4.6 ( 0.5 (6) 136 ( 5 (6)

4.3 ( 0.6 (5) 133 ( 17 (4)

3.9 ( 0.8 (5) 133 ( 6 (6)

3.5 ( 0.4 (6) 217 ( 39 (6)

a Sample acidity, 0.1 mol dm-3 HCl for CBH and DMAB and 1 mol dm-3 HCl for the other reagents. Reagent concentration, 10-3 mol dm-3; reaction loop, 0.5 mL. b Reported as mean ( SD on the number of replicates given in brackets. c Reported as mean ( half width of 95% confidence interval. d MDF-AFS apparatus. e QT-AAS apparatus.

DMAB reagent, which use resulted in an underestimation of mercury. For CRM580, THB, AB, TBAB, and CBH gave results in good agreement with the certified value. Also in this case, DMAB gave an inaccurate result, an overestimate of mercury by more than 64%. The inaccurate results obtained with DMAB are most probably related to the evolution of molecular species from the reaction matrix, which could generate interferences in the atomization/detection step. The effect produced by the interferences on the magnitude of the analytical signal will depend on the type of detector that is used, that is, an MDF-AFS for DORM-2 and a QT-AAS for CRM580. The molecular species could be generated by reaction of DMAB with some components present in the mineralized sample, or they could be impurities arising from a commercial reagent. Other explanations based on analyte losses or sample contamination seem unlikely, because the digestion solutions analyzed for mercury were the same for all of the reductants employed. In particular, the same six replicates were employed for determination of mercury in CRM580 by DMAB, AB, and CBH reduction and in DORM-2, by DMAB and AB reduction. Reaction Mechanism Consideration. The mechanism of reaction of trace elements with THB is still matter of debate, as recently illustrated by Deˇdina.39 Several authors supported the “nascent hydrogen” mechanism, in which the trace element is converted to the corresponding volatile species by monatomic hydrogen, the “nascent hydrogen”, formed during acid decomposition of THB. However, quite recently, Pergantis et al.,40 in their studies on arsine generation using deuterated reagents and mass spectrometry, found that the reaction of both As(III) and As(V) with NaBD4 in HCl and H2O produces AsD3 as the main product. When the reaction was performed using NaBH4 in DCl and D2O, the main product was AsH3. This experimental evidence supports the nonnascent hydrogen mechanism, in particular, all those mechanisms in which the direct transfer of hydrogen atoms from boron to arsenic take place. In the case of Hg(II) reduction, the only product is Hg0, which precludes studies based on the use of deuterated reagent and mass spectrometry. However, there is much experimental evidence indicating the occurrence of a nonnascent hydrogen mechanism. For THB reduction, the maximum yield is almost independent of acidity in a pH range of 0-13.32 Under alkaline conditions, the hydrolysis of THB cannot take place during the short reaction time ( AB > TBAB > CBH g DMAB > TMAB, whereas the ease of reduction for the analyte species decreases in the order Hg(II) > Sb(III) > As(III) > Bi(III) > Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2599

Sb(V) > Se(IV) > Te(IV) > As(V). Owing to the good correlation between the reduction capacity and the rate of acid hydrolysis of amineboranes and CBH, the rate constant of acid hydrolysis can be employed for a first evaluation of the reduction capacity of a given amineborane complex. Amineboranes can be considered promising reagents for trace element determination by CVG atomic spectrometry. In comparison to THB, amineboranes present many interesting features, such as improved selectivity and good control of interferences, sometimes combined with sensitivities comparable with those achievable with THB derivatization. Furthermore, the resistance to acid hydrolysis, which is typical of CBH, allowed the long-term storage of aqueous solution of amineboranes at room temperature and, in many cases, without the aid of alkaline stabilization. In comparison with CBH, the acid decomposition of amineboranes generates virtually nontoxic substances (boric acid, ammonium salts, and hydrogen). A disadvantage of some amineboranes is

2600

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

represented by the production of excessive foaming at concentrations approaching 0.1 mol dm-3. The evidence collected through the present work, complemented with literature studies on reaction mechanism of borane complexes, supports the hypothesis that in CVG, the reaction mechanism of these reagents is based on the direct action of hydrogen bound to boron rather than the action of any other hydrogen species generated by borane complex decomposition in acid media. ACKNOWLEDGMENT The authors thank Prof. Dr. D. L. Tsalev for helpful discussions. Received for review November 12, 2002. Accepted February 17, 2003. AC020694P