Effect of Ascorbic Acid and Sucrose on Electrothermal Atomic Absorption Signals of Indium Journal of Analytical Atomic Spectrometry SHOJI IMAI Department of Chemistry Faculty of Integrated Arts and Sciences University of Tokushima Tojushima 770 Japan NORIYUKI HASEGAWA YASUKO NISHIYAMA YASUHISA HAYASHI Department of Chemistry Joetsu University of Education Joetsu Niigata 943 Japan KENGO SAITO Nissei Sangyo SI Center Sakamachi Shinjuku Tokyo 160 Japan A significant enhancement of sensitivity with the formation of an unresolved double peak signal was observed for indium when deposited in a pyrolytic graphite (PG) furnace pyrolysed with an organic matrix solution and when deposited in a bare PG furnace as a matrix-added solution such as ascorbic acid and sucrose. As the pyrolysis temperature increased the integrated absorbance increased owing to an increase in absorbance at the second peak.The temperature at the second peak remained nearly constant. Integrated across the maximum the integrated absorbance decreased at the decrease in absorbance at the second peak which was shifted to low temperature. The treated PG furnace gave better thermal stability for the effect of pyrolysis temperature for In atomization whereas the organic matrix additive did not. The double peaks were analysed using Arrhenius plots in the treated PG furnace with a low pyrolysis temperature. The same atomizing species smaller sized droplets of In(l) on either active sites or in the thermally stable amorphous carbon were attributed to the species in the rate-determining step for the first and second signals respectively in the treated PG furnace with a low pyrolysis temperature.With the matrix additive the two atomizing species In,O(g) and smaller sized droplets of In(1) on active sites were attributed to the first and second signals respectively. When the pyrolysis temperature increased the atomizing species in the matrix additive did not change to the larger sized droplets of In(1) except in the treated PG furnace. It was concluded that smaller sized droplets dispersed on the active sites vaporize easily before the larger sized droplets were formed because of a decreased probability of movement due to interaction with the active sites. Those dispersed in the amorphous carbon with porous morphology form larger sized droplets by collision and coalescence of the smaller sized droplets.Keywords Ascorbic acid; sucrose; chemical modifier; electrothermal atomic absorption spectrometry; indium; kinetic analysis Ascorbic acid has been widely used as an organic chemical modifier in the analysis of complex samples in ETAAS. It has been reported that the effectiveness of ascorbic acid is due to the formation of active carbon species and reductive gases.'-' Reductive gases such as H CH CO and CO affect the ETAAS signal for lead in the gas phase.' Schcherbakov et al. proposed that the addition of ascorbic acid results in a more uniform dispersion of the analyte on the furnace surface prior to atomization and the formation of reductive centres which are activated in the temperature range 950-1070 K.Vaporization of the active carbon species in the temperature range 970-1070 K was observed as a background absorption pulse in the analytical line for gold.6 Recently the character- istics of the pyrolysis of ascorbic acid were investigated by electrothermal vaporization inductively coupled argon plasma mass spectrometry (ETV-ICP-MS) and Raman spectrometry (i) formation of gaseous products such as hydrocarbons CO and COz below 580 K; (ii) formation and release of the active carbon species in the range 600-1100 K; (iii) formation of thermally stable amorphous carbon in the range 1000-1200 K and release of the amorphous carbon uia decomposition into active carbon species in the range 1200-2400 K; and (iv) for- mation of a lower oriented pyrolytic graphite (PG) phase with porous surface morphology after 100 times reported pyrolysis with 5% m/v ascorbic acid at a temperature above 2500 K.6-8 The effect of the two types of carbon produced by the pyrolysis of ascorbic acid on atomization of gold has been reported.6 Gold for which the metal-graphite interaction is as weak as that for indium is adsorbed on both the active carbon species and the thermally stable amorphous carbon.Smaller sized droplets of gold are formed on the active carbon species resulting in a low-temperature shift of the gold signal and larger sized droplets are formed on the thermally stable amorphous carbon resulting in a high-temperature shift of the gold signal. In the presence of 0.02% m/v ascorbic acid since a significant low-temperature shift of the gold signal was observed the less pronounced high-temperature shift the effectiveness of the formation of thermally stable amorphous carbon was less than that with an organic solvent such as methanol or ethan01.~ In the presence of 1% m/v ascorbic acid most of the gold is adsorbed in the thermally stable amorphous carbon resulting in the formation of larger sized droplets.Indium is one of the elements that exhibit a large loss of analyte in ETAAS particularly when using a PG furnace. A non-pyrolytic graphite (NPG) furnace with a porous and active surface gives better sensitivity." McAllister," using graphite furnace mass spectrometry and thermochemical pre- dictions proposed that the thermal dissociation of In,03 (s) to form In,O(g) during atomization cycles above 1000 K causes substantial losses in sensitivity This sensitivity loss reaction is catalysed by carbon atoms at active sites on the graphite wall In,O,(s) + 2C(s)+In,O(g) + 2CO where C(s) is the active carbon atom." A double peak signal with better sensitivity was observed in the NPG furnace and in the PG furnace a single peak signal with two unresolved Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 (601 -606) 601pulses was reported with identical mechanisms for both fur- naces."*" McNally and Holcombe" noted in kinetic investi- gations that two activation energies may arise due to the droplet size such as bulk metal (larger sized droplets) and smaller sized droplets.Gilmutdinov and c o - ~ o r k e r s ~ ~ * ' ~ how- ever described a gas-phase distribution of In atoms and oxides (pg levels of In) in graphite furnaces by the shadow spectral filming technique which is an imaging technique in atomic and molecular absorption spectrometry. They presented the following direct heterogeneous reduction scheme as an atomiz- ation mechanism In,O(g) + C(s) + 2In(g) + CO (3) where C(s) represents the carbon atoms of the hot graphite wall.Recently Imai et a1." proposed the following atomization process to be identical for the PG and NPG furnaces + ZC(S) + C(S) - 2co - co In,O(g) - 2In(g) (first peak) (4) InzO,(s) + 3C(S) - 2In(l) -- In2(g) -+ 2In(g) - 3CO (second peak) The temperature at the second peak in the NPG furnace was higher than that in the PG furnace. With gallium which is one of the elements exhibiting a large loss of analyte in the same way as indium the use of the NPG furnace or 3% m/v ascorbic acid additive gave a better sensitivity in a similar way to indium and it was concluded that this is due to the presence of active sites promoting the formation of a more easily atomized form of gallium rather than the volatile oxide form.15 In this work we found unresolved double peak signals with substantial sensitivity enhancement and thermal stability for the pyrolysis temperature effect when an indium sample solu- tion was deposited in a PG furnace after pyrolysis of the organic matrix solution above 1100 K.Although organic matrix additives gave a sensitivity enhancement the develop- ment of thermal stability was not observed. Further at low concentrations of organic matrix added a decrease in sensi- tivity was found.This work was undertaken to elucidate the effects of an organic matrix pyrolysed and added on the ETAAS of indium. EXPERIMENTAL Apparatus A Hitachi (Marunouti Tokyo Japan) Model 2-8000 flame and graphite furnace atomic absorption spectrometer equipped with a Zeeman-effect background corrector an optical temperature controller system (Hitachi Model 180-0341) and an automatic data processor was used. A 20 pl volume of sample solution was injected by an automatic sampler. The analytical wavelength and spectral bandwidth were 325.6 nm and 1.3 nm for In respectively. An Oki (Toranomon Tokyo Japan) if-800 Model 50 personal com- puter was used to record the absorbance signal profiles at 20 ms intervals.The output data from the optical temperature controller were acquired at 4ms intervals by the computer and stored on a diskette. Temperature data were calibrated using a Chino (Shinjuku Tokyo Japan) Model IR-AH1S radiation pyrometer. The pyrometer was calibrated with a Pt-Rh thermocouple. For this pyrometer the wavelength was 960 nrn and the uncertainties were 0.5% from 870 to 1500 K 1.0% from 1500 to 2300K and 2.0% from 2300 to 3300K. The standard atomizer conditions are given in Table 1. Each measurement was carried out five times. Table 1 Standard atomization conditions Inner gas Ramp Hold flowrate/ Stage Temperature*/"C time/s time/s ml m-' Drying 120 30 0 200 Pyrolysis 800 30 10 200 Cleaning 2900 0 3 200 Atomization? 2500 0 3 30 * Programmed for the atomizer unit.t The optical temperature controller was used. Raman spectra were measured at room temperature by means of a Jobin-Yvon (Atogo Bussan Shinbashi Tokyo Japan) Ramanor T64000 spectrometer based on a triple 0.64 m focal length monochromator equipped with three 1800 grooves mm-' gratings and a 1024 x 256 element CCD detector. A triple subtractive configuration was used with a spectral bandwidth of about 3.5 cm-' with 20 mW of laser power at the sample using argon radiation at 514.5 nm and 180" scat- tering. The wavenumbers of the observed Raman spectra were calibrated using argon plasma lines. Reagents An aliquot of a commercially available standard solution (Wako Chuoku Osaka Japan) was appropriately diluted with 0.1 mol dmP3 HNO before use. Ascorbic acid and sucrose were of laboratory-reagent grade (Wako).Distilled de-ionized water was purified with a Milli-QII system (Waters Milford MA USA). High-purity argon prepared by Takachiho Chemical Industry (Shibuya Tokyo Japan) was used. RESULTS AND DISCUSSION When investigating the pyrolysis products of ascorbic acid and sucrose it is useful to discuss the difference in the effectiveness of the two compounds. It has been reported that two types of carbon species are formed as pyrolysis products of an organic matrix one of which is an active carbon species vaporising in the temperature range 970-1070 K and the other a thermally stable amorphous carbon existing above 1100 K. The vaporis- ation of the active carbon species was observed as a back- ground absorption pulse the response of which corresponded to the organic matrix concentration in the range 970-1070 K.The value of the background absorption for 1% m/v sucrose solution (0.004) was less than that for 1% m/v ascorbic acid solution (0.050) indicating smaller production of active carbon species for the 1 YO m/v sucrose than the 1 YO m/v ascorbic acid solution. To investigate the degree of coating of the PG furnace wall by the thermally stable amorphous carbon Raman spectra were measured after pyrolysis of 5% m/v ascorbic acid and 5% m/v sucrose solutions at 1575 and 1394 K repsectively. The spectra observed with the bare PG furnace wall are shown in Fig. 1. For the PG furnace wall two Raman bands a sharp band corresponding to the E mode near 1584cm-' and a broad band with weak intensity at about 1361 cm-' (disorder mode) appeared.'.'' The broad bands (1603 and 1356 cm-') for the thermally stable amorphous carbon were observed as the laser beam was focused at the centre of the sample compartment of the PG furnace after pyrolysis of 5% m/v ascorbic acid and 5% m/v sucrose solutions at 1575 and 1394 K respectively.The E band for the PG surface was not observed for the ascorbic acid solution but only for the sucrose solution. The intensity of the disorder mode for sucrose was less than that for ascorbic acid. Hence the proportion of surface area coated by the thermally stable amorphous carbon relating to the area of laser beam spot for sucrose is less than that for ascorbic acid. It can also be found from Fig. 3 in ref. 8 602 Journal of Analytical Atomic Spectrometry August 1996 Vol.11Fig. 1 Raman spectra of a surface of the sample compartment of PG furnace. 1 Bare furnace wall; 2 after pyrolysis of 20 pl of 5% m/v ascorbic acid solution at 1575 K; 3 after pyrolysis of 20 pl of 5% m/v sucrose solution at 1394 K that as the bandwidth for the two Raman bands corresponding to the thermally stable amorphous carbon residues decreased with increase in the treatment temperature the amorphous carbon becomes organized with increasing temperature in the range 1200-1700 K. The amorphous carbon is graphitized above 2500 K.* Fig. 2 shows typical transient signals for the atomic absorp- tion observed under the standard atomizer conditions for only indium deposited in PG and NPG furnaces and a PG furnace treated with 1% m/v ascorbic acid solution at 1281 K.The time axis is labelled in such a way that zero time has elapsed before the atomization stage of the heating cycle. The atomiz- ation mechanism for the double peak signal in the NPG furnace with a porous surface was investigated. The single peak signal observed in the PG furnace also consists of two unresolved pulses whose atomization process is identical with that in the NPG furnace. In the PG furnace treated with ascorbic acid an unresolved double peak signal was also observed similar to that in the NPG furnace. When organic matrix solution was pyrolysed in the PG furnace above 1100 K a thin layer of thermally stable amorphous carbon was formed on the PG wall after the active carbon species had been released.The thin layer of amorphous carbon also has a surface with macropores which play an important role in the admission diffusion or transport into the inside of the thin layer producing a decrease in the tendency of smaller sized droplets to diffuse on the graphite A kinetic approach is a useful method although the inter- Time/s Fig.2 Atomic absorption signals for 6ng of indium deposited in bare NPG (1) and PG (2) furnaces and for 2 ng of indium deposited in a PG furnace treated with 20 p1 of 1% m/v ascorbic acid solution at 1281 K (3) with temperature profile. Temperature profile 4 the NPG furnace; and 5 the PG furnace mediate species involved are not observed but speculated upon. A kinetic approach including the rate controlling mechanism for the reaction has been proposed for the single peak signal in the absorbance-time Since unresolved double peak signals were observed for indium deposited in the PG furnace treated with the 1% m/v ascorbic acid at 1281 K the absorbance for the first peak involved in the rising side of the signal must be corrected for the contribution from the second peak to utilize a kinetic appr0a~h.l~ A kinetic approach based on the following equation was adopted to obtain information about the Arrhenius activation energy (E,) in the rate con- trolling mechanism in ETAAS 1n[g(a)Tp2] = -E,R-l T-' + ln[ARP(dT/dt)-'E,-l] (6) where g(a) is a function that depends on the process controlling the reaction rate a is the fractional conversion (a = AJA,, where A, and A are the absorbance at maximum absorption and time t respectively) dT/dt is the heating rate of the furnace wall A is the frequency factor P = 1-2X+6X2-24X3+120X4+*..,X=RT/E,and TandR have their usual meanings.E zpp T, and g(a) for the first and second peaks were defined as E, KPpl Tmaxl and g(a)l and Ea2 zpp2 Tmax2 and g ( ~ ) ~ respectively. A typical Arrhenius plot for 2ng of indium deposited in the PG furnace treated with 1% m/v ascorbic acid at 1281 K is shown in Fig. 3. A straight-line segment was observed in the Arrhenius plots above 1680 K. A first-order reaction (Fl) for the rate-con- trolling mechanism was chosen in order to give the best possible fits; the E value was obtained as 100 f 15 kJ mol-I. To correct the absorbance of the first peak for that of the second peak the absorbance corresponding to the second peak in the period of the first peak was evaluated by extrapolating the straight line for the second peak to each atomization time.For the corrected absorbances for the first peak the F mechanism was chosen and the E value was obtained as 110 & 10 kJ mol-'. When the absorbance correction was not carried out the E value was obtained as 130 f 20 kJ mol-' with the F mechanism. The KPp2 and Tmaxl values can also be estimated with the absorbance correction. In the treatment with 5% m/v ascorbic acid solution the absorbance corrections were carried out in the same way. The E g(a) zPp and Tma values are summarized in Table2 with those obtained at a pyrolysis temperature of 1620K in the PG furnace treated with 1% m/v ascorbic acid at 1360 K.The rate-controlling mechanism was chosen for each case. The process-controlling mechanism can be established from the specific dependence of the absorbance-time profiles on the analyte mass.,* When the -13.0 n 1 b I = t Y .!? -19.0 -2 I .o \ ' I -23.0 5.00 6.00 7.00 8-00 9.00 lo4 T-' I K" Fig. 3 Arrhenius plot with the first-order process-controlling mechan- ism for 2 ng of indium deposited in the PG furnace treated with 20 pl of 1% m/v ascorbic acid solution at 1281 K and calculation of the second signal absorbances in the period of the first signal using eqn. (5). 0 observed value; and 0 calculated value Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 603Table 2 Arrhenius activation energy (E,) appearance temperature ( zpp) temperature at maximum absorbance (T,,,) and process-con- trolling mechanism [g(a)] of indium atomic absorption signal in the PG furnace treated with ascorbic acid solution Concentration of ascorbic acid in the treatment solution E,,/kJ mol-' 1% m/v (1281 K)t 800 1190 k 13 1456 k 20t (1520 f 20) (130 k 20) F1 t 1340 f 151 1890 f 15 100 f 15 F1 110 f 101 5% m/v 1% m/v (1281 K) (1360 K) 800 1620 1190 f 20 1590 k 15t 67 f lot Fl 1 1450 Tt_ 151 1990 f 15 130 f 15 (75 f 10) 1740 & 15 1950 f 15 262 f 20 F1 Fl * Tpyr = pyrolysis temperature for indium atomization; qppl Tmaxl E and g(a)l =parameters for the first peak; ~ p p z TrnaxZ Ea2 and g ( ~ 1 ) ~ = parameters for the second peak.t Treatment temperature. 1 Estimated by using the absorption correction. atomic absorption signals were measured over various indium concentrations the time at the maximum absorbance and at the shoulder maintained a constant value.The dependence of the absorbance-time profiles on the analyte mass indicates that the atomization for both peaks occurs with the F1 mechanism. It was found that there is a decrease in E and an increase in Ea2 with an increase in xPp2 and Tmax2 with increase in the concentration of ascorbic acid in the treatment solution. The amount of active sites and thermally stable amorphous carbon increases with increase in the matrix concentration. When gold sample solution was deposited in the PG furnace treated with organic matrix solution low- and high-temperature shifts of the signal with a decrease and increase in E values respect- ively take place at the same The low-temperature shift was caused by the formation of smaller sized droplets of gold on active sites and the high temperature shift was caused by the formation of larger sized droplets of gold in thermally stable amorphous carbon.McNally and Holcombe12 obtained two sequential atomization energies 142 & 15 and 222 & 8 kJ mo1-l for E and Ea2 for indium deposited by an aerosol in a bare PG furnace without any absorbance correc- tion. The Eal value for the aerosol deposition is close to the Eal and Ea2 values for the treated PG furnace. The first and second atomizations for the aerosol deposition take place via vaporization from smaller sized droplets and bulk vaporization from the larger sized droplets respectively. Thus the observed decrease and increase in Eal and Ea2 values are attributed to the decrease and increase in indium droplet size in different situations such as on the active site and in the amorphous carbon respectively.The integrated absorbance for 2 ng of indium deposited in a PG furnace treated with 20 p1 of an organic matrix solution (0 0.01 1 and 5% m/v ascorbic acid and 1 2 and 5% m/v sucrose) at various treatment temperatures and atomized according to the standard atomization conditions is shown in Fig. 4. Fig. 5 shows typical transient signals for atomic absorp- tion with the PG furnace treated with 1% m/v ascorbic acid solution. At treatment temperatures above 1100 K at which the active carbon species has been released the absorbance imreased with increase in the ascorbic acid concentration indicating the sensitivity enhancement to be in response to the proportion of the area coated by the thermally stable amorph- ous carbon.The absorbance increased with increase in the 0.12 0.10 3 I 0.08 1 0.06 0.04 a CI 0.02 0.00 700 )# 1100 1300 ls00 1700 1900 2100 2300 ZSOI Pyrolysis temperature I K Fig. 4 Integrated absorbance for 2 ng of indium deposited in a PG furnace treated with 20p1 of organic matrix solution at various treatment temperatures. Ascorbic acid concentration 0 0; A 0.01; 0 1; and A 5% m/v. Sucrose concentration 0 1; 4 2; and x 5% m/v 0.20 a 0 a 1 .oo 2 .oo 3.00 s o L 0 2 0 a - 3273 273 0 1 .oo 2.00 3.00 Time/s Fig. 5 Atomic absorption signals for 2 ng of indium deposited in a PG furnace treated with 20 pl of 1% m/v ascorbic acid solution at various treatment temperatures.Treatment temperature 1,840; 2 1128; 3 1281; 4 1384; 5 1683; 6 1779; and 7 1882 K. 8 temperature profile treatment temperature especially for the second peak and reached a maximum at 1480 and 1740 K with 1 and 5% m/v ascorbic acid respectively. This increase in absorbance is due to the suppression of analyte loss according to eqn. (2). When the treatment temperature increased from 1281 to 1384 K the bandwidth of the first peak decreased and that of the second peak increased with increase in the integrated absorbance (lines 3 and 4 in Fig. 5). In the temperature range 1200-1700K with bandwidth for the two Raman bands for the pyrolysed ascorbic acid decreased with increase in the treatment temperature due to organization of the amorphous carbon.The decreases of the first and second peaks are ascribed to a decrease in the disordered sites and to ordering of the amorphous carbon respectively. Across the maximum the second peak was shifted slightly to low temperature with 604 Journal of Analytical Atomic Spectrometry August 1996 Vol. 11increase in the treatment temperature owing to thermal destruction of the amorphous carbon and the formation of active sites. At 1 and 5% m/v ascorbic acid the enhancement was still observed at 2000 and 2400 K respectively. At 2390 K the contribution of amorphous carbon to the gold signal was observed as a broadening of the signal.6 The sensitivity enhancement was less pronounced for sucrose than ascorbic acid at the same concentration of the matrix which is also in response to the Raman spectrometric data.Fig. 6 shows the effect of the pyrolysis temperature on the integrated absorbance for indium deposited in the PG furnace treated with the 1% m/v ascorbic acid at 1360 K deposited in the bare PG and NPG furnaces and for indium sample solution containing the organic matrix in the PG furnace. The absorbance was lost at temperatures below 1400 K in the bare PG and NPG furnaces in which indium atomizes via inter- mediate species In,O(g) and In(1) for the first and second peaks respectively. However in the treated PG furnace a higher absorbance was observed at 1360K above which the first peak disappeared and the signal pulse could be observed at 1800 K. No enhancement of thermal stability was observed for indium sample solution containing the organic matrix.As a single peak signal corresponding to the second peak was only observed at a pyrolysis temperature of 1620 K in the treated PG furnace the kinetic analysis was carried out without the absorbance correction (Table 2). As the Ea2 value is close to the enthalpy of bulk vaporization of indium 240 kJ mol-1,f9 and the E value for the aerosol deposition,12 the indium atomization mechanism is attributed to bulk vaporization from larger sized droplets. It was proposed for the aerosol deposition that the larger sized droplets are formed by the collisions and coalescence of the small particles of indium on the graphite surface. With gold a growth of droplet size in the macropores in a PG furnace treated with 5% m/v ascorbic acid solution at 1270 K was reported.6 Hence an acceptable explanation for the redistribution of indium for the second peak is that the smaller sized droplets dispersed in the thermally stable amorph- ous carbon probably move along the macropores with increas- ing temperature ultimately to form larger sized droplets in the amorphous carbon by collisions.Fig. 7 shows atomic absorption signals for 2 ng of indium in a 1% m/v organic matrix and those with variation in the matrix concentration. The kinetic approach was applied to the second peak for 1% m/v ascorbic acid and sucrose. The results are given in Table 3 with those for the first peak in 0.1% m/v ascorbic acid and 1% m/v sucrose. The Fl mechanism chosen - 0.m I 1 3273 0.01 0 (w( 1m 1m 1400 1600 1880 2mo 22m Pyrolysis temperature I K Fig.6 Effect of pyrolysis temperature on integrated absorbance for indium in the PG and NPG furnaces. PG furnace 0 absence of organic matrix 6 ng In (1.3 scale); A 1% m/v ascorbic acid additive 2 ng In; x 1 YO m/v sucrose additive 2 ng In; and A furnace treated with 20 pl of 1% m/v ascorbic acid solution at 1360 K 2 ng In. NPG furnace 0 absence of organic matrix 6 ng In (1.3 scale) 0.10 R 2.00 3.00 LI 1 .oo s o 42 9 O*'O [(b) 0 Timeis Fig.7 Atomic absorption signals for 2ng of indium dissolved in various concentrations of ascorbic acid (a) and sucrose (b). Concentration 1 0 2,O.Ol; 3,O.l; and 4,1% m/v. 5 temperature profile was supported by the results of the dependence of the profiles on the analyte mass.The Eal values correspond to those reported for CO additive rather than those for pure argon." The Ea2 values obtained correspond to the E and Ea2 values for indium deposited in the PG furnace treated with ascorbic acid solution which indicates the formation of smaller sized droplets of indium metal. The study of the effect of pyrolysis temperature indicated a lower thermal stability for the matrix additive than that for the treated PG furnace (Fig. 6). The absorbance was almost completely lost at a charring tempera- ture of 1400K similar to the absorbance at the first peak observed in the treated PG furnace and the integrated absorbance in the bare NPG furnaces. Hence indium deposited as a solution containing the organic matrix vaporizes before larger sized droplets are formed during the pyrolysis stage.This is due to a decreased probability of movement of the droplets due to interactions with the active sites. It has been reported that 3% m/v ascorbic acid additive gives a better sensitivity for gallium which is also one of the elements that exhibits a substantial sensitivity loss in ETAAS like indium due to volatile monoxide formation at temperatures above 1000K.15 This is due to the reduction of oxides to the metal at a lower temperature before volatile species such as the monoxide are formed.I5 If the reduction mechanism is the dominant mechanism for sensitivity enhancement with indium no decrease in absorbance due to an organic matrix additive should be observed. In this case significant sensitivity losses were observed with a low-concentration organic matrix addi- tive as shown in Fig.8. The sensitivity loss with sucrose is less pronounced than that with ascorbic acid owing to the strength of background absorption for the vaporization of the active carbon species. As it has also been reported that the reduction of In203(s) to In20(g) due to active sites on the graphite takes place the sensitivity loss occurs as a result of the mechanism in eqn. (2) due to active sites in the pyrolysed organic matrix. Above 0.1% m/v ascorbic acid and at 1% m/v sucrose the sensitivity was enhanced at both peaks. It has been reported Journal of Analytical Atomic Spectrometry August 1996 Vol. 11 605Table 3 Arrhenius activation energy (Eal and Ea2) and rate-controlling mtechanism [g(cx)l and g(cx)J for the first and second atomizations of indium First atomization Furnace Conditions Gas Ea,/kJ mol-' g(41 PG* Only indium Ar 226 f 10 Fl NPG* Only indium Ar 201 f 15 F1 PG* Only indium 1.01% co 175 f 10 Fl NPG* Only indium 1.01% co 150 f 15 F PG Ascorbic acid additive Ar 140 f 20t Fl PG Sucrose additive5 Ar 176 f 20 Fl Second atomization E,,/kJ mol-' g ( 4 2 103 f 10 75 f 10 103 f 10 l l O f 15$.F1 120 f 20 Fl * Ref. 10. t 0.1% m/v ascorbic acid. $ 1% m/v ascorbic acid. 0 1% m/v sucrose. that 1.01 YO CO additive gave a sensitivity enhancement due to the increase in the first peak by promotion of the gas-phase dissociation of In,O(g). The E values in the presence of 0.1% m/v ascorbic acid and 1% m/v sucrose correspond to those for the CO additive. The increase in absorbance at the first peak takes place as a result of the promotion of the dissociation of In,O(g).The increase in absorbance at the second peak is due to the increase in the amount of smaller sized droplets dispersed on the active sites of the amorphous carbon. The sensitivity enhancement is attributed predominantly to the reduction of In,O,(s) to In(1) due to active sites and trapping of the In(1) as smaller sized droplets on the active sites. CONCLUSIONS Pyrolysis of an organic matrix provides two types of carbon species. One is the active carbon species formed above about 600 K and released in the temperature range 950-1 100 K. The other is the thermally stable amorphous carbon formed during release of the active carbon species and existing above 1100 K.The active carbon species plays the role of a condensed-phase reductant to reduce In,O,(s) to In,O(g). The gas-phase reduc- tants formed from the organic matrix give a sensitivity enhance- ment owing to promotion of the gas-phase dissociation of In,O(g). The thermally stable amorphous carbon which forms a thin layer directly reduces In203(s) to In(1). Smaller sized droplets are formed on the active sites of the amorphous carbon and in the amorphous carbon. The smaller sized droplets dispersed on the active sites vaporize easily before the larger sized droplets are formed because of a decreased prob- ability of movement due to interaction with the active sites. The smaller sized droplets dispersed in the amorphous carbon layer form larger sized droplets by collision and coalescence of the droplets.Although larger sized droplets are not formed when indium sample solution containing the organic matrix is deposited in the bare PG furnace they are formed when only indium is deposited in the PG furnace treated with organic matrix solution at temperatures above 1100 K. The amorphous carbon layer plays an important role as an area for movement collision and coalescence of indium droplets and produces a decrease in the tendency of larger sized droplets to diffuse on the graphite surface. REFERENCES 1 2 Regan J. G. T. and Warren J. At. Absorpt. Newsl. 1978 17 89. Hydes J. D. Anal. Chem. 1980 52 959. 0.05 I -3 -2 1 0 1 log {[organic matrix](%)} Fig. 8 Integrated absorbance for 2 ng of indium dissolved in various concentrations of organic matrices.Organic matrix broken line none; 0 ascorbic acid; and A sucrose. Errors are represented by bars 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Tominaga M. and Umezaki Y. Anal. Chim. Acta 1982,139,279. Schcherbakov V. I. Belyaev Yu. I. Myasodov Y. B. F. Marov I. N. and Kalinichenko N. B. Zh. Anal. Khim. 1982 37 1717. Gilchrist G. F. R. Chakrabarti C. L. and Byrne J. P. J. Anal. At. Spectrom. 1989 4 533. Imai S. and Hayashi Y. Bull. Chem. SOC. Jpn. 1992 65 871. Imai S. and Hayashi Y. Anal. Chem. 1991 63 772. Imai S. Nishiyama Y. Tanaka T. and Hayashi Y. J. Anal. At. Spectrom. 1995 10 439. Imai S. Okuhara K. Tanaka T. Hayashi Y. and Saito K. J . Anal. At. Spectrom. 1995 10 37. Imai S. Hasegawa N. Hayashi Y. and Saito K. J. Anal. At. Spectrom. 1996 11 515. McAllister T. J. Anal. At. Spectrom. 1990 5 171. McNally J. and Holcombe J. H. Anal. Chem. 1991 63 1918. Gilmutdinov A. Kh. Zakharov Yu. A. Ivanov V. P. and Voloshin A. V. J. Anal. At. Spectrom. 1991 6 505. Gilmutdinov A. Kh. Zakharov Yu. A. Ivanov V. P. Voloshin A. V. and Dittrich K. J. Anal. At. Spectrom. 1992 7 675. Botha P. V. and Fazakas J. Anal. Chim. Acta 1984 162 413. Yoshikawa M. Muter. Sci. Forum 1989 52-53 365. Iwamoto E. Miyazaki N. Ohkubo S. and Kumamaru T. J. Anal. At. Spectrom. 1989 4 433. McNally J. and Holcombe J. H. Anal. Chem. 1987 59 1105. Handbook of Chemistry and Physics ed. Weast R. C. CRC Press Boca Raton FL 75th edn. 1994. Paper 61036626 Received May 28 1996 606 Journal of Analytical Atomic Spectrometry August 1996 Vol. 1 1