Behaviour of a Single-bore High-pressure Pneumatic Nebulizer Operating With Alcohols in Inductively Coupled Plasma Atomic Emission Spectrometry 1 Journal of I Analytical Atomic Spectrometry JOSE L. TODOLI ANTONIO CANALS AND J;ICENTE HERNANDIS Departamento de Quimica Analitica Universidad de Alicante 03071 Alicante Spain The behaviour of a single-bore high-pressure pneumatic nebulizer (SBHPPN) with alcohols in (ICP-AES) was investigated. A standard Meinhard nebulizer was used for comparison. To this end the drop size distribution of the primary aerosols the analyte and solvent transport rates W and Stor the fraction of solvent transported in liquid and vapour forms SKq and Sap the excitation temperature of the plasma T, and the molecular emission intensity of the C band were determined for solvents of different physical properties.The effect of the physical properties of the solvent on the nebulization process was also studied. The results show that the SBHPPN gives rise for all the solvents studied to primary aerosols that have smaller mean diameters than those produced by the Meinhard nebulizer for the same gas and liquid flows. The relative decrease in volume median diameter (I)"*) when switching from the Meinhard nebulizer to the SBHPPN is more noticeable for alcohols than for water. W is significantly higher for the SBHPPN than for the Meinhard nebulizer particularly at high liquid flows. However the differences between their S values are less pronounced. Under similar conditions the SBHPPN gives rise for all the solvents studied to higher emission intensities than the Meinhard nebulizer.However the relative signal enhancements achieved by changing from the SBHPPN to the Meinhard nebulizer are lower than the corresponding analyte transport enhancements. The relative signal enhancements achieved by switching from water to alcohols are lower for the SBHPPN than for the Meinhard nebulizer. This behaviour can be explained in terms of the higher S values associated with the SBHPPN which lower T in comparison with the Meinhard nebulizer. Keywords Inductively coupled plasma; organic solvent; alcohols; pneumatic nebulization; single-bore high-pressu 1.e pneumatic nebulizer One of the most important applications of organic sol~ents in atomic spectrometry is as a means to increase the senisitivity of the determination.',* However in low-medium power (ICP- AES) the organic solvent load has a strong influence on the physical and chemical properties of the plasma and iii most instances the sensitivity improvements achieved have nc it been as significant as expected.Therefore much effort has been devoted in trying to understand fully the interactions between organic aerosols and analytical plasmas that could account for the observed behaviour. It has been observed that under the same conditions (ie. rf power viewing height and Glolvent plasma load) organic solvents tend to decrease the excitation temperat~re~-~ and the electronic density of the pla~ma.~ Boumans and Lux-Steiner' attributed these effects to the enhancement of the thermal conductivity of the plasma due to the presence of the pyrolysis products of the organic clolvent (ie.Cz CN CO CH C I etc). Blades and Caughlin7 suggested that such an enhancement in the thermal conductivity of the plasma was mainly due to the energy absorbed by [.he C2 molecules to dissociate and ionize since in a water-loaded plasma the predominant molecules i.e. OH require much less energy to dissociate. In addition solvent pyrolysis products cause a large increase in the background emission of the plasma. Kreuning and Maessen3 measured the spatial distribution of the species C2 CN and C I in chloroform- and toluene-loaded plasmas and observed that the emission maxima of these species appeared in different zones of the plasma. Thus whereas C2 appears at low observation heights in the inner zone of the plasma CN appears higher in the outer zone since its formation results from the interaction of C atoms with the surrounding air.Recent studies conducted by Weir and have shown that methanol and chloroform behave in a similar fashion. These distributions of the solvent pyrolysis products which depend on the experimental conditions (i.e. rf power nebulizer/carrier gas flow liquid flow solvent nature) cause the effect of the organic solvents on the emission intensity to have a marked spatial character in ICP-AES. As regards the nature of the solvent employed Botto'l concluded that the energy consumption from the plasma was higher for those solvents with the highest H:C ratio since they have higher atomization energies. Maessen et al.12 classi- fied the solvents into two groups one for solvents of low vapour pressure (e.g.water xylene or IBMK) that do not affect the stability of a plasma of 1.75 kW and another for solvents with high vapour pressure (e.g. chloroform methanol and ethanol) that clearly deteriorate the behaviour of the plasma. Later Kreuning and Maessen3 concluded that for a given solvent load to the plasma the most significant variables are the load of carbon and the relative molecular mass of the solvent. Recent studies conducted by Weir and Bladesg reveal that the C:O ratio markedly affects the appearance of the plasma. These workers make a distinction between solvents with a C 0 ratio close to 1 (e.g. methanol) and those with a C:O ratio higher than 1 (e.g. chloroform). When working with water-ethanol mixtures McCrindle and Rademeyer6 have reported that the excitation temperature increases with the percentage of ethanol up to 15% (v/v) and then decreases.All these detrimental effects associated with the use of organic solvents become more noticeable on increasing the solvent The amount of solvent that reaches the plasma (S,,,) when using pneumatic nebulization depends on the following factors (1) The nebulizing gas flow Q,. Increasing gas flow gives rise to a decrease in the mean drop size of the primary aeros01'~ and to a larger amount of evaporated solvent thus causing Stot to in~rease.'~ (2) The liquid flow Q1. Increasing Q1 causes S, to increase even though the mean drop size of the primary aerosol increase^.'^ The increase in S, is less pronounced at high Q values.I2 (3) Solvent nature.As regards the influence of the solvent nature on St, two factors have to be taken into consideration. Firstly the fraction of solvent transported in liquid form (Sliq) is a function of the mean drop size of the primary aerosol i.e. the smaller the mean drop size of the primary aerosol the Journal of .Analytical Atomic Spectrometry October 1996 Vol. 1 1 (949-956) 949higher the value of Sliq. Secondly the fraction of solvent transported in gaseous form (Svap) depends on the relative volatility of the solvent. These two factors are not completely unrelated. On the one hand the rate of solvent vaporization is proportional to the aerosol surface. In addition the vapour pressure of a droplet increases on decreasing its curvature radius.16 Hence a reduction in the aerosol mean drop size facilitates solvent evaporation.On the other hand solvent evaporation would cause Sliq to decrease; however this hypo- thetical decrease is compensated for at least partially by the reduction in the mean drop size of the aerosol caused by the solvent evaporation which in turn facilitates the transport of the analyte. Since in general organic solvents show lower surface tension and higher volatility than water organic aero- sols usually have smaller mean drop sizes than aqueous aerosols and Sto values are usually higher for the f~rmer.".''.'~ Particular attention has been paid to the rate of solvent transported in vapour form. Long and Browner17 compared the effects on the plasma of introducing water as a vapour and as a liquid aerosol.They noticed that the introduction of vapour causes the electronic density to increase compared with the introduction of the same amount of water as droplets. These droplets diffuse very little from the axis of the central channel and hence most of the power consumption corre- sponds to this zone of the plasma." With organic solvents most of the solvent load reaches the plasma as vapour its side diffusion rate then being much higher and therefore it occupies a greater zone of the plasma. Hence when using organic solvents in ICP-AES it is advisable to take into account the physical state in which they will reach the plasma." (4) Introduction system. For a given spray chamber and specific nebulization conditions the characteristics of the pri- mary aerosol and hence the values of Sliq and Stat depend on the nebulizer ernpl~yed.'~-~' The effects of the organic solvents described above have been observed when using c~ncentric,'*~*'~ cross-flowg-" or V- groove3 pneumatic nebulizers.However other nebulizers have shown a different behaviour. Nisamaneepong et a1." reported that the use of a glass-frit nebulizer enhanced the tolerance of the plasma to the organic solvent load as compared with a concentric nebulizer. They explained their observations in terms of the small mean drop size and monodispersion of the aerosols generated. Later Brotherton et ~ 1 . ~ ' observed the same behaviour with a double grid nebulizer. Other more efficient nebulizers such as ultrasonic21.22 or therm~spray'~ nebulizers need to be used in conjunction with a desolvating system to reduce the load of organic solvent which otherwise would prevent any analysis from being performed.These papers highlight the importance of nebulizer design on the interfering effect of organic solvents. Some suggestions have been proposed to alleviate the detri- mental effects of organic solvents on the characteristics of the plasma (1) Working at high rf power. Many studies have shown that increasing rf power gives rise to higher excitation temperatures7 and hence to higher emission intensity."." (2) Using low gas flow.14 This leads to a lower solvent load to the plasma and to a longer residence time of the sample in the plasma. However lowering Q reduces the amount of analyte transported to the plasma.Therefore Q should not be too low. (3) Using refrigerated spray chambers. Maessen et ~ 1 . ' ~ found that when using toluene chloroform or 1,2-dichloro- ethane as solvents the signal-to-noise ratio for many elements increased significantly when the spray chamber was thermostated between -5 and -10°C. With ethanol and methanol cooling of the spray chamber was indispensable for the stability of the plasma. Pan et al.' obtained similar results. Bottoll reported that the power consumption decreased when spray chamber cooling was used. (4) Working with a narrower injector tube. Boumans and Lux-Steiner' and Nobile et ~ 1 . ' ~ designed a torch with a narrower injector tube which was particularly well suited to organic solvents. Recently a single-bore high-pressure pneumatic nebulizer (SBHPPN) has been developed in our laboratories which is suitable for FAAS,26,27 ICP-AES" and ICP-MS.29 In compari- son with the standard Meinhard nebulizer the SBHPPN provides higher sensitivities and lower limits of detection (LODs) when working with aqueous solutions of metals. In addition it does not show any tendency to become clogged because of high salt content in the solutions.As was indicated above the type (i.e. positive or negative) and extent of the effects of organic solvents on the plasma is a function of the nebulizer employed among other variables. Therefore when a new nebulizer is designed and evaluated one should test its applicability for the introduction of organic solutions. Hence the aim of this work was to evaluate with the SBHPPN the type and extent of the effect of alcohols in ICP-AES. To this end the drop size distribution of the primary aerosol the analyte and solvent transport rates the fraction of solvent transported in liquid and vapour forms the exci- tation temperature of the plasma T,, and the molecular emission intensity of the C band were determined for four alcohols with different physical properties.The effect of the physical properties of the solvent on the nebulization process was also studied. EXPERIMENTAL The design and working principle of the SBHPPN has been described in previous The glass nozzle employed in this work has an outlet bore whose section is 5.92 x mm2. A Meinhard nebulizer (TR-30-A3) was employed for comparison (gas outlet section 28.3 x mm2; liquid outlet section 0.82 mm').The liquid flow was varied between 0.6 and 2.0ml min-' by means of an Iso-Chrom HPLC pump (Spectra-Physics San Jose CA USA). Table 1 lists the significant physical properties of the five solvents employed. The gas flow was varied by means of a Model FC260 mass flow controller (Tylan Torrance CA USA). Air was employed for the measurement of the aerosol drop size distributions (DSDs) and argon for the remainder of the experiments. Although the DSDs obtained with air and argon are different their behaviour is similar as regards the variables considered in this work.I3 The DSDs of the primary aerosols were measured at a distance of 11 mm from the nebulizer tip in all instances except when the variation of the mean drop size along the radial position was measured; in this instance the distance was 15 mm to allow the aerosol cone to broaden.The drop size characteriz- ation was performed by means of a Model 2600c Fraunhofer laser diffraction system (Malvern Instruments Malvern Worcestershire UK) equipped with a lens with a focal length of 63 mm that allowed for the measurement of drop diameters from 1.2 to 118 pm. The instrument was previously calibrated with a reticule.31 The DSDs were calculated using a model- Table 1 Physical properties of the solvents* Solvent a/dyn cm - t P/CPS 4 Water 70.4 1 .oo 0.09 Methanol 20.3 0.50 1.00 Ethanol 21.4 1.13 0.69 Propan-2-01 19.8 2.08 0.27 Butan-1-01 22.8 2.38 0.11 *Measured at 20 "C. Surface tension. $Viscosity. §Relative volatility defined as the relationship between the liquid volume of solvent i and the volume of methanol required to saturate a given empty pace.'^^^^ 950 Journal of Analytical Atomic Spectrometry October 1996 Vol.11independent algorithm which does not presume any distri- bution function. The following parameters were employed for characterizing the primary aerosols volume median diameter (Dv.so) Sauter mean diameter ( 0 3 . 2 ) and volume concentration (VC) which is the ratio of the liquid volume of aerosol contained in the measurement zone of the laser beam to the total volume of this measurement zone in the particle ~ i z e r . ~ ~ . ~ ~ The solvent transport rate (Stat) was measured by a continu- ous indirect m e t h ~ d . ~ . ' ' ~ . ~ ~ The analyte transport rate (wet) was measured by a direct method i.e.by capturing the aerosol on a glass-fibre filter (Type A/E 47 mm diameter) of 0.3 pm size ~ o r e ' ~ . ~ ~ (Gelman Sciences Ann Arbor MI USA). Solutions of Mn (100 pg ml-') in each of the solvents Gdudied were used for this purpose. The Mn retained was extracted with hot 1.0% (m/m) HNO and determined by ICP-AES. The emission intensity was measured with a Modcl 2070 ICP-AES instrument (Baird Bedford MA USA). Solutions of Mn (1 pg ml-') in each of the solvents studied were used for this purpose. Table 2 lists the instrumental conditions employed. The emission intensity of the C2 molecular band (h= 516.468 nm) and the plasma excitation temperature T,,,. were also measured so as to assess the effect of organic solvcnts on the characteristics of the plasma.For calculating Z, the emission intensity of five ionic Fe lines was measured using solutions of this element (50 pg ml-') in each of the s1)lvents employed with the assumption that the population of the species of a given element obeys the Boltzmann distri- bUtion.3.5.7.18.34.35 Table 3 lists the characteristics of these lines. A straight line is obtained by plotting ln(~qp~qp/gqAqp I versus E (see Table 3 for definition of the symbols) the slope of which is - 1/KT,,c,3,s-7,'8,34,36 where K is the Boltzmann constant. The standard deviation of this method is abtlut 200 K,3 which is not too large for qualitative studies as in o iir case. RESULTS AND DISCUSSION Nebulization Process The peculiarities of the SBHPPN (i.e. working at high press- ures and having just one bore for both streams gas and liquid) Table 2 Experimental conditions for the measurement of the emis- sion intensity Outer gas flow rate Intermediate gas flow rate Nebulizer gas flow rate Liquid flow rate Spray chamber Rf power Slit-width Wavelength Integration time Viewing height PMT gain Torch 16 1 min-' 1.7 1 min-' Variable Variable Double-pass Sccj tt type 1000 w 0.1 nm 257.610 nm 0.2 s 7 mm ALC 4 For use with orpanic (V= 100 cm3) solvent^^*^^ Table 3 Iron ionic lines employed to measure the plasma excitation temperature* Wavelength/nm g '4,,/10-8 s-' E 'eV 259.940 22.1 4 17 275.574 21.1 5 .48 274.648 11.7 5 59 256.253 12.8 5 82 266.664 24.1 8 07 make it advisable to carry out a preliminary study of the fundamentals of the nebulization process with this nebulizer.Fig. 1 shows the effective section of the gas outlet A oersus Qi at a constant gas pressure for water and the four alcohols. In order to calculate A an equation that expresses this parameter as a function of Q and the gas pressure required to reach this gas flow was used.37 In all instances A decreases when Q1 is increased since the width of the liquid film adhering to the inner walls of the nozzle increases. Fig. 1 also gives some information about the influence of the physical properties of the solvent on A,. Thus for instance when comparing water and ethanol i.e. solvents of similar viscosity but different surface tension (Table l) it appears that surface tension hardly affects A which is about the same for both solvents.As regards viscosity Fig. 1 shows that the higher the solvent viscosity the smaller the effective section. Therefore when the solvent viscosity is increased a higher gas pressure should be applied if one seeks to keep the gas flow constant. In pneumatic nebulization the efficiency of the energy transfer from the gas stream to the liquid stream increases with the turbulence of the gas stream,37 and hence the mean drop size of the aerosol decreases. For the flow values used in this work the Reynolds number (Re) for the SBHPPN is up to 20 times higher than for the Meinhard nebulizer. Hence aerosols with a smaller mean drop size are to be expected for the former. DSD of the primary aerosols Influence of gas and liquidjows Fig. 2(a) shows that for both nebulizers and for all the solvents studied Dv.50 values decrease when Q is increased as expected since the energy available for nebulization is also increa~ed.'~-'' As regards the effect of Qi Fig.2(b) shows a different behaviour for both nebulizers. With the Meinhard nebulizer Dv,50 increases when Qi is increased for all the solvents in agreement with previously reported results." With the SBHPPN Dv,so for alcohols behaves similarly but for water Dv,so decreases slightly when Q1 is increased. This fact can be explained by taking into account that the surface tension values for alcohols are usually very low and hence the viscosity becomes a more important factor than with water.37 Thus the liquid vein remains unbroken for a longer time before being disintegrated by the gas stream.Therefore when Qi is increased the rupture of the liquid vein takes place in positions increasing downstream from the nebulizer tip where the energy of the gas flow becomes lower. This effect seems to predominate over the reduction in A caused by the increase in Q (Fig. 1). Obviously the intensity of this effect will be greater for solvents of high viscosity. Thus for solvents of low viscosity such as methanol the increase in Dv,50 is less pro- nounced than for solvents of high viscosity such as butan- 1 0.5 1.0 1.5 2.0 Q /ml min-' *p fundamental state; q upper state; g, orbital degenerancy of the upper state; Aqp Einstein coefficient for spontaneous emissior (8 A values from ref. 3); E, excitation potential. Fig. 1 Effect of the liquid flow rate on the 'effective' outlet section A for the SBHPPN.A Water; B methanol; C ethanol; D propan-2-01; and E butan-1-01. Gas pressure = 14 atm Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1 95120.0 1 (a) I \ 3 0 a 5*0* 0.50 0.60 0.70 0.80 0.90 6 Qg /1 min-’ 0.50 0.70 0.90 1.10 1.30 Q /ml min-’ Fig.2 Variation of Dv,50 for the solvents and nebulizers employed. (a) Dvis0 uersus Qg; Q1= 1.2 ml min-’. (b) Dv,50 uersus Q1; Qg=0.57 1 min- . (dotted line) Meinhard nebulizer; (solid line) SBHPPN. A Water; B methanol; C ethanol; D propan-2-01; and E butan-1-01 1-01. For solvents with a high surface tension such as water viscosity plays a secondary role. Comparison of SBHPPN and Meinhard nebulizer Fig. 2 shows that the SBHPPN generates organic aerosols the Dv,50 of which are 1.5-3.0 times smaller than those generated by the Meinhard nebulizer under the same conditions.The decrease in Dv,50 when changing from the Meinhard nebulizer to the SBHPPN is more important for alcohols than for water. The excess of kinetic energy of the gas stream associated with this change gives rise when applied to liquids with a surface tension much lower than that of water to an excess of surface which is much more significant for alcohols than for water. Efects of surface tension and viscosity of the solvent The effect of the surface tension on Dv,50 can be seen by comparing the curves of butan-1-01 and water (Fig.2). For any given set of conditions butan-1-01 gives rise to aerosols with a smaller Dv,50 than those produced by water as expected from the surface tension values,14*32*38 although the viscosity of the former is twice that of the latter (Table 1).As regards the viscosity its effect can be seen by comparing the curves of methanol ethanol propan-2-01 and butan-1-01 (Fig. 2). These solvents have similar surface tension values but different viscosities (Table 1). When working with the Meinhard nebulizer the values of Dv,50 follow the same order as the solvent viscosities whatever the conditions employed. This behaviour can be explained in terms of the ability of the viscosity to dampen the instabilities of the liquid vein and to lengthen it.37 However with the SBHPPN the behaviour is not the same. Fig. 2 shows firstly that the differences between the Dv,50 values are smaller than with the Meinhard nebulizer and secondly that the Dv,50 values do not follow the order of the solvent viscosities exactly.This divergence can be attributed to the fact that as already mentioned for the SBHPPN but not for the Meinhard nebulizer the effective gas outlet section (As) decreases when changing from one solvent to another of higher viscosity. Thus a higher pressure is required to keep the gas flow constant which in turn means more kinetic energy for surface generation. The final result with these high viscosity solvents is that the aerosols obtained with the SBHPPN have smaller Dv,50 values than expected from their surface tension and viscosity alone. Eflect of the volatility of the solvent Although the mean drop size depends to some extent on the volatility of the ~olvent,~’ the total liquid volume is the aerosol property that is most influenced by the solvent volatility. This influence was monitored by means of the so-called ‘volume concentration’ VC.Table4 summarizes the VC values for the four alcohols employed under several Qg and Q1 conditions. It appears that for the same Q1 Q and nebulizer VC values follow the rev- erse order of the solvent volatilities i.e. VC( butan-1-01)> VC(propan-2-01) > VC(ethano1) > VC(methano1). Hence meth- anol is the solvent that has lost the most liquid volume due to volatilization as expected. Comparison between nebulizers (SBHPPN versus Meinhard) is not possible because VC is a function of the gas (aerosol) velocity and this velocity is higher for the SBHPPN than for the Meinhard nebulizer under the same Q and Q1 conditions.Variation of the mean drop size along the radial position From a visual inspection pneumatically generated aerosols have the shape of a cone the angle of which decreases as Q is increased. In order to optimize the design from the point of view of aerosol transport it is important to know how droplets of different sizes are distributed within the cone. Fig. 3 shows the variation of D3,2 in relation to the transversal distance from the cone axis for both nebulizers. In this instance D3,2 was plotted instead of Dv,50 because the variations correspond- ing to Dv,50 for the SBHPPN appeared to be masked by the large values of Dv,50 for the Meinhard nebulizer. Some differ- ences between the two nebulizers are apparent.With the Meinhard nebulizer the largest drops are usually found in the outer zone of the cone whatever the solvent employed whereas the aerosol contained in the inner zone has a higher proportion of small drops. Similar observations have been reported pre- viously for technical atomizers (i.e. devices designed to generate aerosols in non-spectroscopic applications such as mechanical Table 4 Volume concentration values of the primary aerosols obtained for the alcohols using the SBHPPN and the Meinhard nebulizer vc (Yo) Solvent Methanol* Ethanol* :Propan-2-01* Butan-1-ol* Methanol? Ethanol7 Propan-2-01? Butan-1-01? Methanol$ Ethanol$ Propan-2-011 Butan-1 -011 Meinhard nebulizer 0.0620 0.07 17 0.0756 0.0882 0.1380 0.141 1 0.1532 0.1651 0.0893 0.1031 0.1152 0.1228 SBHPPN 0.0301 0.0353 0.04 16 0.0469 0.0676 0.08 13 0.0946 0.0966 0.0482 0.0559 0.06 13 0.0634 *Qg=0.57 1 min-’; Ql=0.6 ml min-’.tQg=0.57 1 min-‘; Ql=1.2 rnl min-l. $&,=0.87 1 min-’; Q,= 1.2 ml min-’. 952 Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1,;*A Fig. 3 Variation of D3+2 in relation to the transversal distance from the cone axis for both nebulizers. A Water; B methanol; C:' butan- 1-01. (dotted line) Meinhard nebulizer; (solid line) SBHPPN. Q,=0.68 1 min-'; Q,=2 ml min-'; distance from nebulizer tip to laser beam= 15 mm engineering and combustion p r o c e s s e ~ ) ~ ~ - ~ ~ the working prin- ciples of which are the same as for the Meinhard nebulizer and for other pneumatic nebulizer^.^^-^^ The SBHPPN pro- vides a more homogeneous distribution of drop sizes particu- larly when working with water.The slight increase in observed with alcohols on the outside of the cone which does not appear with water might be the result of the preferential evaporation of small droplets circulating in this zonc. From Fig. 3 it is also apparent that under the same set of Q and Q1 the cone angle is larger for the Meinhard nebuliier than for the SBHPPN. These observations clearly show that the geometriec of the interaction between gas and liquid streams are different in the two nebulizers. In the Meinhard nebulizer the aerosol is generated from the surface of the cylindrical liquid vein since the gas stream flows externally to it,37 whereas in the SHHPPN the liquid vein flows as a film externally to the gas stream.These results support the coupling of Meinhard nebulizers to double-pass spray chambers so as to remove most of the large drops by impaction against the inner walls of the inner tube which would not contribute effectively to the final emission signal and would also contribute to the inciease in signal fluctuation^,^^ thus deteriorating the LOD. As regards the SBHPPN this coupling does not seem to be so beneficial since small droplets as well as large droplets are likely to be lost that otherwise would contribute to the signal. For the same reason the spray chamber should be lengthened for the SBHPPN since otherwise the high velocity of the aerosol jet causes many potentially useful droplets to be lost through impaction against the bottom walls of the spray chamber.In spite of all these considerations a conventional double-pass Scott chamber was used in this work for both nebulizers so as not to change more than one factor at a time. Transport Efficiency The transport rates W, and St, finally obtained as well as the DSD of the tertiary aerosol are the result of cornbining the characteristics of the primary aerosol and the processes that take place in the spray chamber. The interaction of the tertiary aerosol with the plasma determines the analytical signal i.e. the emission signal and the LOD. Under the same conditions organic solvents proviide Wot and S, values higher than those provided by water whereas organic tertiary aerosols have smaller average drop sizes than aqueous tertiary aerosols. These facts are expected lo have opposite effects on the plasma the net result on the signal being unpredictable.Solvent transport eflciency Under any set of conditions S, is expected to increase when Q is increased owing to two parallel factors. Firstly the mean drop size of the primary aerosols decreases thus facilitating the transport and secondly evaporation is favoured since there is more gas and a larger surface. Stot is plotted against Q in Fig. 4(a). The results confirm the expected trends. Several reports on pneumatic nebuli~ation'~*'~~'~ state that S, increases again when Q1 is increased. Since Stat= Q1 x E where E is the solvent transport effi~iency,'~ it is clear that the increase in the amount of aerosol predominates over the reduction of the transport efficiency caused by the increase in the mean drop size of the primary aerosol.Fig. 4(b) shows the same tendency as increasing Q in our case. When S, values are compared in terms of the solvent employed it appears that volatility (or vapour pressure) is the most significant physical property of the ~ o l v e n t . ~ * ' ~ * ' ~ ~ ' ~ Fig. 4 shows that the arrangement of the solvents according to their Stot values coincides with the arrangement according to their volatility (Table 1). Finally when Stot values are compared in terms of the nebulizer employed it is clear that the SBHPPN provides higher values than the Meinhard nebulizer as expected since the aerosols produced by the SBHPPN have smaller D,,50 and D3,2 values than those produced by the Meinhard nebulizer.Analyte transport eficiency Fig. 5 shows the influence of several variables on KO,. The trends shown here are very similar to those obtained for S,,,. As mentioned before the transport values are higher for the SBHPPN than for the Meinhard nebulizer. Table 5 gives the relative transport values (SBHPPN/Meinhard) for solvent (Stot)rel and for analyte (Wot)rel. It can be seen that ( ~ o t ) r e l values are significantly higher than (Stot)rel values which indi- cates that part of the solvent is transported in vapour form. The amount of solvent that reaches the plasma in vapour form Svap and in liquid form SIiq can be estimated from mass balance considerations for both solvent and ana1~te.I~ Table 5 shows that Sli is much higher for the SBHPPN than for the 4.0 -I 1 - @ 0.40 0.45 0.50 0.55 0.60 0.65 - 'Y) - 4.0 9 Y % 3.0 2.0 1 .o 0 Q,/I min-' 0.50 0.70 0.90 1.10 1.30 Q /ml min-I Fig.4 Variation of S, for the solvents and nebulizers employed. (a) S, versus Q,; QI = 1.0 ml min- ' . (b) S, versus Q1; Q =OSO 1 min- ' . (dotted line) Meinhard nebulizer; (solid line) SBHPPN. A Water; B methanol; C ethanol; D propan-2-01; and E butan-1-01 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11 953m 0 100- 80 - 60 - 40 - 20 - d x o 4 v) Q g / 1 rnin-' I - 0.40 0.45 0.50 0.55 0.60 0.65 I 3 1401 1 2 1201 (b) I 601 40 2ol * ..._._._........ * ..... rl *... ........... .* A O L I 0.5 0.7 0.9 1.1 1.3 Q /ml min-' Fig. 5 Variation of KO for the solvents and nebulizers employed. (a) Tot uersus Qg; QI = 1.0 ml min-'. (b) To versus Q,; Qg=0.50 1 min-'.(dotted line) Meinhard nebulizer; (solid line) SBHPPN. A Water; B methanol; C ethanol; D propan-2-01; and E butan-1-01 Meinhard nebulizer as expected from their DSDs whereas Svap is similar for both nebulizers. Obviously the solvent with the highest volatility i.e. methanol gives rise to the highest Svap values. Kreuning and M a e ~ s e n ~ ~ state that the solvent evaporation factor4 is the parameter that determines the frac- tion of solvent in vapour and in liquid form. According to these workers methanol should be the solvent with the highest Svap value whereas water should show the lowest value as is in fact the case (Table 5 ) . Sliq in turn depends on the DSD of the primary aerosol (the higher the proportion of small drops in the primary aerosol the higher the Sliq should be) and on solvent volatility (the higher the solvent volatility the lower the Sliq should be).Thus for instance Sliq for methanol is about the same as for ethanol. The smaller mean drop size of the primary aerosol for methanol is probably compensated for by the lower volatility of ethanol. Emission Intensity Fig. 6(u) shows the variation of the net emission intensity Znet uersus Qg for the nebulizers and solvents employed. For the SBHPPN [Fig. 6(u)(ii)] when Q is increased the emission intensity increases up to a maximum value and then decreases whatever the solvent. For the Meinhard nebulizer the behav- iour is similar but no maximum is found for water and butan- l-ol [Fig. 6(u)(i)]. These results can be explained on the basis of opposite effects that take place when Qg is increased.Since the mean drop size of the primary aerosol becomes lower and there is an increasing volume of dry gas the following occurs (i) W, increases; (ii) S, increases; (iii) the gas load increases; and (iu) the sample residence time in the plasma decreases. Factor (i) will contribute to a higher emission signal whereas factors (ii) (iii) and (iu) will act in the opposite direction. Fig. 6 also shows that the SBHPPN gives rise to higher emission signals than the Meinhard nebulizer since their W, values are higher (Fig. 5). However the relative emission enhancement on switching from the SBHPPN to the Meinhard nebulizer is lower than the corresponding analyte transport enhancement. Also the signal maxima are more noticeable for the SBHPPN than for the Meinhard nebulizer probably because the S, values are also higher for the former (Fig.4). Hence the plasma excitation temperature seems to be lowered when the SBHPPN is used. As regards the effect of the solvents in addition to the above-mentioned factors (i) and (ii) another factor should be borne in mind the nature of the solvent. Fig.6(u) shows that the position of the maxima as well as the intensity val- ues are dependent on the solvent employed. For the com- binations methanol-SBHPPN and propan-2-ol-SBHPPN [Fig. 6(u)(ii)] the I, drop after their maxima is more pro- nounced than for the remaining solvents. Moreover for propan-2-01 maxima are located at a lower Q value ( i e . 0.39 1 min-I). On comparing the net intensity values obtained for the different alcohols it can be observed that the increasing order of net intensity does not coincide with the increasing order of W,,,.Thus for both nebulizers ethanol is the solvent that affords the highest net intensity values. For the remainder of the solvents tested the increasing order of net intensity depends on the gas flow rate employed. For example at high Q values propan-2-01 gives net intensity signals even lower than those of water whereas at low Qg values water gives the lowest I, values. In order to explain these results it is necessary to take into account the interaction between the plasma and the solvent. When organic solvents are introduced into the plasma a factor to be considered is the concentration and nature of the Table 5 Relative values (SBHPPN/Meinhard) of the analyte and solvent transport rates and liquid and gas fractions of the solvent transport rate Solvent Water Methanol Ethanol Propan-2-01 Butan- l-ol Water Methanol Ethanol Propan-2-01 Butan-1-01 Q J 1 min-' 0.43 0.43 0.43 0.43 0.43 0.50 0.50 0.50 0.50 0.50 Qi/ ml min-' 1 .o 1 .o 1 .o 1 .o 1 .o 0.6 0.6 0.6 0.6 0.6 (Stot )re,* 1.44 1.32 1.27 1.42 2.10 ( Kot )re1 t 4.00 2.58 2.66 2.27 3.00 SBHPPN 220 181 1 926 595 373 Meinhard nebulizer 248 1716 1125 613 349 SBHPPN 213 910 926 709 610 Meinhard nebulizer 52 351 336 306 21 1 1.33 1.35 1.16 1.31 1.74 2.12 1.60 1.81 1.95 2.46 157 1796 938 579 317 168 1432 1094 599 280 176 819 83 1 740 649 82 509 428 363 27 5 *(Stot)rcl =(Stat )SBHPPN/(Stot)Meinhard- t(W tot ) re] =(W tot )SBHPPN/(W tot )Meinhard.$See ref. 15. 954 Journal of Analytical Atomic Spectrometry October 1996 Vol. 1 1120 100 m 80 2 x 60 - 0) %" 40 20 0 m I3 K Y s u 0.34 0.37 0.40 0.43 0.46 0.49 600 -1 I / 400 0.34 0.37 0.40 0.43 0.46 0.49 800 m 600 2 Ky 400 s u 200 0 0.34 0.37 0.40 0.43 046 0.49 0.34 0.37 0.40 0.43 0.46 0.49 8000 M C 6000 i5001 5000 .I 0.34 0.37 0.40 0.43 0 46 0.49 M h c Q,' 8000 7500 7000 6500 6000 5500 5000 (d 0) 0.34 0.37 0.40 0.43 0.46 0.49 1 min-' Fig.6 ICP emission parameters uersus Q for the different .,olvents and nebulizers employed. (a) Net emission intensity I,, for ionic line of Mn; (b) net emission intensity of C molecular band lcPc; (c) excitation temperatures. (i) Meinhard nebulizer; (ii) SBHPPN. A Water; B methanol; C ethanol; D propan-2-01; and E butan-1-01.Q,= 1.0 ml min-' solvent pyrolysis products C being the most significant 3*9*10*12 Fig. 6(b) shows C2 emission intensity I C X versus Q It can be observed that for a given nebulizer Ic-C increases with Qg and that for a given solvent the IC-c values obtained v ith the SBHPPN are higher than those for the Meinhard nebulizer. These results are in agreement with the behaviour of Stot (Fig. 4). On comparing the behaviour of I, and I uersus Qg it appears that both intensities increase with Q until the maximum value of I, is reached. From this point I, decreases whereas I increases even faster t lian at lower Q values thus indicating that the plasma exhitation temperature is lowered. As regards the nature of the -,olvent Fig. 6(b) shows that I decreases in the following order propan-2-01 > ethanol > butan-1-01 >methanol. This is the result of the amount of carbon transported to the plasma which is related to Stat and the number of C-C bonds that exist in each solvent molecule.For methaiiol the C species comes from recombination reaction^.^ Fig. 6(c) shows the influence of Qg on T, for all the :.olvents and nebulizers used. It can be seen that T, decreases when Q is increased and also when the SBHPPN is used These results can be attributed to the higher Stot values obtained (Fig. 4). As regards the nature of the solvent propan-2-01 shows the lowest T, values. In agreement with this it also provides the highest I values and the sharpest deciease in I, when Q is increased. Methanol also shows a 1( w T,, which probably highlights the negative effect of a solvrnt load that is too high (Fig.4). It can also be observed that ihe T, values for alcohols are lower than that for water. Thi; might explain why the net emission enhancement obtaintad with alcohols with respect to water is lower than that of the analyte transport rate as shown in Table 6. An additional result that can be derived from Table6 is that the enhancements in transport rate and emission signal achieved by the SBHPPN when switching from water to alcohols are less important than those observed for the Meinhard nebulizer. Table6 includes the LOD for Mn in each of the solvents employed showing that propan-2-01 and water provide the highest LOD values for both nebulizers and that the values obtained for the SBHPPN are lower than those for the Meinhard nebulizer by a factor of about two.The effect of Q1 on the emission intensity is similar to that of Qg although less detrimental since gas load and residence time do not change on varying Q1. CONCLUSIONS (1) The SBHPPN gives rise for all the solvents studied to primary aerosols with smaller Dv,50 values than those produced by the Meinhard nebulizer under the same Q and Q1 con- ditions. The relative decrease in Dv,50 when switching from the Meinhard nebulizer to the SBHPPN is more noticeable for alcohols than for water. In addition for the SBHPPN D3,2 and Dv,50 show little variation along the radial (transversal) position in the aerosol cone whereas for the Meinhard nebul- izer the largest drops are usually found in the outer zone of the aerosol cone.(2) The analyte transport rates Wtot are significantly higher for the SBHPPN than for the Meinhard nebulizer particularly at high liquid flows. However the differences between their S, values are less pronounced. Journal of Analytical Atomic Spectrometry October 1996 Vol. 11 955Table 6 Relative values (solvent i/water) of analyte transport rate and net emission intensity and limits of detection (LOD) for the solvents and nebulizers employed* ( Wot )Ire t (LOD/ng ml- ’)§ Meinhard Solvent 1 min-’ nebulizer SBHPPN Qd Water 0.43 1 .o 1 .o Methanol 0.43 7.6 4.9 Ethanol 0.43 7.0 4.6 Propan-2-01 0.39 6.0 3.4 Butan-1 -01 0.43 4.2 3.1 Meinhard nebulizer SBHPPN 1 .oo 1 .oo 1.64 1.52 3.40 3.14 2.42 2.16 1.64 1.40 Meinhard nebulizer SBHPPN 11.9 5.5 2.8 1.5 2.9 1.2 11.0 6.2 5.4 2.3 *Q1 = 1.0 ml min-’.t(kl(ot)’rel = ( ~ o t ) s o ~ v c n t i/(kl(ot)water. f(Znct)lrel =(Inet)solvent i/(Znet)water. §LOD calculated according to the 3sb criterion where s is the standard deviation from ten replicates of the blank. The experimental conditions were optimized for each solvent. (3) The SBHPPN gives rise for all the solvents studied to higher emission intensities than the Meinhard nebulizer under the same Q and Q conditions. However the relative signal enhancements achieved on changing from the SBHPPN to the Meinhard nebulizer are lower than the corresponding analyte transport enhancements. (4) The relative signal enhancements achieved on switching from water to alcohols are lower for the SBHPPN than for the Meinhard nebulizer. ( 5 ) Conclusions (3) and (4) can be explained in terms of the higher Stot values associated with the SBHPPN which lower the excitation temperature of the plasma in comparison with the Meinhard nebulizer.(6) Solvent viscosity has a small effect on the characteristics of the primary aerosol generated with the SBHPPN at least up to 2.3 cP. This makes the SBHPPN particularly promising for the analysis of high viscosity samples (e.g. wear metals in lubrication oils). This aspect is currently under study in our laboratories. The authors thank the DGICYT (Spain) for financial support of this work (Projects PTRI91-0029 and PB92-0336). REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Cresser M. S. in Solvent Extraction in Flame Spectroscopy Analysis Butterworths London 1978. Boorn A.W. and Browner R. F. in Inductively Coupled Plasma Emission Spectroscopy ed. Boumans P. W. J. M. Wiley New York 1987 p.151. Kreuning G. and Maessen F. J. M. J. Spectrochim. Acta Part B 1989 44 367. Boorn A. W. and Browner R. F. Anal. Chem. 1982 54 1402. Pan C. Zhu G. and Browner R. F. J. Anal. At. Spectrom. 1990 5 537. McCrindle R. I. and Rademeyer C. J. J. Anal. At. Spectrom. 1995 10 399. Blades M. W. and Caughlin B. L. Spectrochim. Acta Part B 1985 40 579. Boumans P. W. J. M. and Lux-Steiner M. Ch. Spectrochim. Acta Part B 1982 37 97. Weir D. G. and Blades M. W. J. Anal. At. Spectrom. 1994 9 1311. Weir D. G. and Blades M. W. J. Anal. At. Spectrom. 1994 9 1323. Botto R. I. Spectrochim. Acta Part B 1987 42 181.Maessen F. J. M. J. Kreunning G. and Balke J. Spectrochim. Acta Part B 1986 41 3. Canals A. Wagner J. Browner R. F. and Hernandis V. Spectrochim. Acta Part B 1988 43 1321. Browner R. F. Canals A. and Hernandis V. Spectrochim. Acta Part B 1992 41,659. Canals A. Hernandis V. and Browner R. F. Spectrochim. Acta Part B 1990 45 591. Cresser M. S. Prog. Anal. At. Spectrosc. 1982 5 35. Long S. E. and Browner R. F. Spectrochim. Acta Part B 1988 43 1461. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Pan C. Zhu G. and Browner R. F. J. Anal. At. Spectrom. 1992 7 1231. Nisamaneepong W. Haas D. L. and Caruso J. A. Spectrochim. Acta Part B 1985 40 3. Brotherton T. Barnes B. Vela N. and Caruso J. J. Anal. At. Spectrom. 1987 2 389.Wiederin D. R. Houk R. S. Royce K. W. and D’Silva A. P. Anal. Chem. 1990 62 1155. Botto R. I. and Zhu J. J. J.Ana1. At. Spectrom. 1994 9 905. Laborda F. de Loos-Vollebregt M. T. C. and de Galan L. Spectrochim. Acta Part B 1991 46 1089. Maessen F. J. M. J. Seeverens P. J. H. and Kreuning G. Spectrochim. Acta Part B 1984 39 1171. Nobile A. Jr. Shuler R. G. and Smith J. E. Jr. At. Spectrom. 1982 3 73. Todoli J. L. Canals A. and Hernandis V. Spectrochim. Acta Part B 1993 48 373. Todoli J. L. Canals A. and Hernandis V. Spectrochim. Acta Part B 1993 48 1461. Todoli J. L. Mufioz M. Valiente M. Hernandis V. and Canals A. Appl. Spectrosc. 1994 48 573. Hernandis V. Todoli J. L. Canals A. and Sala J. V. Spectrochim. Acta Part B 1995 50 985. Mora J. Hernandis V. and Canals A. J. Anal. At. Spectrom. 1991 6 573. Hirleman E. D. in Liquid Particle Size Measurement Techniques eds. Tishkoff J. M. Ingebo R. D. and Kennedy J. B. ASTM STP848 Philadelphia 1984 p. 35. Malvern Instruments 2600 Laser Diffraction User Manual Malvern 1987. Tarr M. A. Zhu G. and Browner R. F. J. Anal. At. Spectrom. 1992 7 813. Alder J. F. Bombelka R. M. and Kirkbright G. F. Spectrochim. Acta Part B 1980 35 163. Nakamura S. J. Anal. At. Spectrom. 1995 10 467. Walker A. L. Curry D. L. and Fannin H. B. Appl. Spectrosc. 1994,48 333. Sharp B. L. J. Anal. At. Spectrom. 1988 3 613. Farino J. and Browner R. F. Anal. Chem. 1984 56 2709. Boorn A. W. Cresser M. S. and Browner R. F. Spectrochim. Acta Part B 1980 35 823. Tishkoff J. M. Hammond D. C. Jr. and Chraplyvy A. R. J. Fluids Eng. 1982 104 313. Schulte G. Dannehl M. and Bauckhage K. Presented at Partec 89 Nurnberg 1989. Dodge L. G. Rhodes D. J. and Reitz R. D. Appl. Opt. 1987 26,2144. Faske A. J. PhD Thesis Georgia Institute of Technology USA 1986. Shum S. C. K. Johnson S. K. Pang H. and Houk R. S. Appl. Spectrosc. 1993 47 575. Liu H. and Montaser A. Anal. Chem. 1994 66 3233. Olesik J. W. and Fister J. C. 111 Spectrochim. Acta Part B 1991 46 851. Kreuning G. and Maessen F. J. M. J. Spectrochim. Acta Part B 1987 42 677. Paper 610321 5J Received May 8 1996 Accepted July 11 1996 956 Journal of Analytical Atomic Spectrometry October 1996 Vol. 11