Comparison of characteristics and limits of detection of pneumatic micronebulizers and a conventional nebulizer operating at low uptake rates in ICP-AES† Invited Lecture Jose�-Luis Todolý�,*‡a Vicente Hernandis,a Antonio Canalsa and Jean-Michel Mermetb aDepartamento de Quý�mica Analý�tica, Universidad de Alicante, 03071 Alicante, Spain. E-mail: jose.todoli@ua.es bLaboratoire des Sciences Analytiques (UMR 5619), Universite� Claude Bernard-Lyon, F-69622 Villeurbanne Cedex, France Received 21st January 1999, Accepted 4th March 1999 Three micronebulizers, the high-eYciency nebulizer (HEN), the microconcentric nebulizer (MCN) and the micromist (MM), were compared with a conventional pneumatic concentric nebulizer working at low liquid flow rates in ICP-AES.The gas back-pressure, the free liquid aspiration rate, the drop size distribution of primary and tertiary aerosols, the solvent and analyte transport rates, the emission intensity and the limits of detection were measured.The solvent evaporation inside the spray chamber proved to be a very important transport phenomenon when working at very low liquid flow rates. The micronebulizers produced finer primary aerosols, higher solution transport rates through the spray chamber and higher sensitivities than the conventional pneumatic concentric nebulizer. The HEN used in this work provided slightly lower ICP-AES limits of detection than the other two micronebulizers, but at the expense of a higher back-pressure. at low liquid flow rates, thus improving the nebulizer perform- Introduction ance in terms of surface generation.The HEN and MM are Nebulization of solutions is by far the most common means made of glass, whereas the MCN is manufactured of an of sample introduction in inductively coupled plasma atomic HF-resistant polymer. emission spectrometry (ICP-AES). Normally, the ICP-AES Liu et al.6,7 successfully used the HEN to analyze four conventional liquid sample introduction systems consist of a certified samples by ICP-AES and ICP-MS.This nebulizer did nebulizer, mainly of the pneumatic concentric type, a spray not suVer from tip blocking when working with biological and chamber and an injector tube.1 plant reference materials. Because of its characteristics, the Nowadays, the analysis of very small volumes of sample HEN has been used as an interface between liquid separation solutions is becoming one of the key research subjects in techniques and ICP-MS.9,10,14 Pergantis et al.9 employed the HEN with a microscale FIA-HPLC system and ICP-MS for atomic spectroscopy.This is explained by the great number of the determination of arsenic in very low sample volumes. areas in which the sample size may be limited: semiconductor, Good reproducibility was observed even for 0.5 ml sample clinical, geological, on-chip technology,2 etc. In addition, volumes.3,9 Because of its small dead volume, less band coupling ICP-AES with separation methods may require low broadening was observed in capillary electrophoresis with the liquid uptake rates.HEN than a conventional pneumatic nebulizer.10,11 Generally, although conventional nebulizers can be used at The MCN was satisfactorily applied to the determination rates under 1 ml min-1, their design is not optimized for this of rare earth elements in wine by ICP-MS,17 showing that the purpose, as regards either the dead volume or the liquid and matrix eVects were less severe than for a conventional pneu- gas injection areas.There is, therefore, a need for nebulizers matic nebulizer, which was explained by the lower liquid flow devoted to work eYciently at rates as low as 10 ml min-1.3–34 rates employed with the MCN. This nebulizer can withstand Among them, there are pneumatic micronebulizers that are NaCl solutions without tip blocking.16,18,22 The MCN was usually employed in conjunction with a spray chamber, such also applied, with good results, to As and Se speciation by as the high-eYciency nebulizer (HEN),3–15 the microconcentric employing a microbore column in ICP-MS.23 A recent paper25 nebulizer (MCN)16–26 and, more recently, the micromist described the features of the MCN coupled to a cyclonic spray (MM).27 All are concentric nebulizers and the main diVerence chamber as a capillary electrophoresis–ICP-MS interface.with respect to the conventional nebulizers is a conspicuous The MCN has also been adapted to a desolvation system.reduction in their critical dimensions (i.e., gas and/or liquid In this case, the aerosol is first heated by means of a convection/ exit cross-sectional areas and liquid capillary wall thickness). conduction mechanism and the generated vapor is then This reduction allows a more eYcient gas–liquid interaction removed by means of a porous membrane.3,20,26,28,29 This assembly is known as the MCN6000 and evidence was given that matrix (acids) eVects were reduced both in ICP-AES20 †Presented at the 1999 European Winter Conference on Plasma and ICP-MS.29 Spectrochemistry, Pau, France, January 10–15, 1999.No results on the characterization of the behavior of the ‡Work performed while on leave from the Universidad de Alicante, Spain. MM were found in the literature. J. Anal. At. Spectrom., 1999, 14, 1289–1295 1289One should also mention that two micronebulizer designs Filtration was eVected by forcing the solution through a 1.2 mm pore size filter (Millipore, Bedford, MA, USA).have been described in which the spray chamber was avoided (i.e., the aerosol was directly injected into the plasma): the The gas flow rate was controlled by means of a mass flow controller (5857 TR Series, Brooks Instruments, Veenendaal, direct injection nebulizer20,26,30–34 and the direct injection highe Yciency nebulizer (DIHEN).3,12,13 This leads, obviously, to The Netherlands). The optimized value in terms of the ICP-AES SBR was 0.70 l min-1 for all the nebulizers.an analyte transport eYciency close to 100%. Because of this, under the same conditions, the sensitivity is expected to Aerosol data were obtained by using a Model 2600c Fraunhofer laser diVraction system (Malvern Instruments, increase and the wash-out times to decrease with respect to the other nebulizers. However, as more solvent load is also Malvern, Worcestershire, UK) equipped with a lens of 63 mm focal length that allowed the measurement of drop diameters introduced with the concomitant analyte, the deterioration of the plasma excitation properties (in ICP-AES) and/or the in the range 1.2–118 mm. The software was the B.0D version.A model-independent algorithm was used to calculate the drop increase in the spectral interferences (in ICP-MS) must be taken into consideration when these nebulizers are to be size distribution from the energy data. In order to minimize the eVect of solvent evaporation on the characteristics of the selected.Together with the pneumatic micronebulizers mentioned aerosols generated by the nebulizer (primary aerosol ), the nebulizer tip was placed 5 mm away from the beam axis. Note above, some others have been proposed based on diVerent principles, such as the micro-ultrasonic nebulizer35 and the that the beam was around 10 mm in diameter. The aerosols emerging from the spray chamber (i.e. tertiary aerosols) were oscillating capillary nebulizer.14,36 According to previously published work, the main advantage measured by placing the outlet of the spray chamber 5 mm away from the laser beam center.of the micronebulizers is that they give rise to limits of detection (LODs) similar to those obtained with the conven- Unless stated otherwise, the surface mean diameter (D3,2) was selected to describe the aerosol mean size. A set of three tional nebulizers, but at liquid flow rates 10–40 times lower.3,4,6,12,15 Obviously, this is the result of increased trans- replicates was performed in each case.For primary aerosols, the D3,2 precision (RSD) was seen to be dependent on the port eYciency4 and/or enhanced signal stability.8,15 Low sample consumption rates also mean a decrease in the waste liquid flow rate (Ql) employed; the lower was Ql, the higher were the RSD valueently, the RSD changed from management costs.33 So far, no systematic comparison has been performed <1 to around 6% on decreasing Ql from 600 to 20 ml min-1.This was clearly a problem resulting from the very low liquid between these diVerent pneumatic concentric micronebulizers (PCMNs) in ICP-AES. Therefore, the aim of the present study volume of the aerosol. When it is too low (i.e. several microliters per minute), small variations in the background was to evaluate the behavior of these pneumatic micronebulizers (i.e., HEN, MCN and MM) in ICP-AES with aqueous may lead to significant shifts in the Malvern energy distribution and, consequently, in the aerosol drop size distribution, thus solutions, by reference to a conventional nebulizer.To this end, the drop size distribution of the aerosol, the solvent and giving rise to a degradation of the precision.19 The solvent transport rate (Stot) was measured by means of analyte transport rates, the signal-to-background ratio (SBR) and the LODs were determined. a direct method, i.e., by the adsorption of the aerosol in a Utube filled with silica gel during a 10 min period.37 By weighing it before and after the aerosol tube exposure, the Stot values Experimental were easily derived.The analyte transport rates were obtained by a direct Three diVerent pneumatic concentric micronebulizers were used in conjunction with a double pass spray chamber: a high- method, i.e., by collecting the aerosol on a glass-fiber filter (type A/E, 47 mm diameter, 0.3 mm pore size; Gelman Sciences, eYciency nebulizer (HEN) (Meinhard, Santa Ana, CA, USA), an MCN-100, microconcentric nebulizer (MCN) (CETAC Ann Arbor, MI, USA) placed above the spray chamber.38 A 500 mg ml-1 Mn solution was nebulized.The Mn retained Technologies, Omaha, NE, USA) and an AR30-1-FM005 micromist (MM) (Glass Expansion, Camberwell, Victoria, after a period of 10 min was extracted by washing the filters with 1.0% (m/m) hot nitric acid. The total solution volume Australia).A TR-30-A3 conventional pneumatic concentric nebulizer (PN) (Meinhard) was selected for comparison. was adjusted to 100 ml in a calibrated flask. Finally, the Mn concentration in each solution was determined by flame atomic Table 1 gives their relevant dimensions. In order to compare the behaviors of the nebulizers absorption spectrometry (FAAS). A Perkin-Elmer Optima 3000 DV ICP-AES system was employed, the same Ryton double-pass Scott-type spray chamber (120 ml inner volume) was used to transport and used in the axial viewing mode.Table 2 gives the experimental conditions. This instrument includes a 40 MHz free-running filter the aerosol towards the torch; this is the most widely employed spray chamber in ICP-AES. generator, a polychromator with an e� chelle grating of 79 lines mm-1 with a blaze angle of 63.4° and a cross-dispersing For the PN, HEN, MCN and MM, the liquid flow was varied in the range 5–160 ml min-1 by means of a peristaltic element for the UV range and a prism for the visible and a segmented-array charge-coupled device (SCD) detector that pump (Perkin-Elmer, Norwalk, CT, USA).Tygon capillaries of 0.25 mm id were used. allowed the simultaneous measurement of several line intensities in addition to the background signals. The resolution of Aqueous sample solutions were filtered in order to prevent capillary blocking especially with the HEN and the MCN. Table 2 Instrumental conditions of the ICP-AES spectrometer Table 1 Characteristics of the nebulizers used Gas outlet Liquid Capillary Rf power 1.2 kW Outer gas flow rate 15 l min-1 cross-sectional capillary inner wall Nebulizer area/mm2 diameter/mm thickness/mm Intermediate gas flow rate 0.5 l min-1 Nebulizer gas flow rate Variable Integration time 20 ms PN 0.028 0.40 0.06 HEN 0.011 0.10 0.03 Sampling time 1000 ms Injector id 2 mm MCN 0.017 0.10 0.03 MM 0.025 0.14 0.05 Plasma viewing mode Axial 1290 J.Anal. At.Spectrom., 1999, 14, 1289–1295Table 3 Elements, lines and Esum values a given gas flow rate the pressure is expected to follow the order: HEN>MCN>MM. The results agreed with the Element l/nm Esum/eVa expected trends. Thus, to keep the gas flow rate at 0.7 l min-1, the measured back-pressures were 120 psig (850 kPa), 36 psig Cr II 205.55 12.77 (250 kPa) and 20 psig (140 kPa) for the HEN, MCN and Mn II 257.61 12.25 Mg II 280.27 12.07 MM, respectively. For the conventional nebulizer the back- Sr II 421.55 8.64 pressure was 15 psig (110 kPa).According to these results, at K I 766.49 1.62 a given liquid and gas flow rate, the amount of energy available Li I 670.78 1.85 for the nebulization was larger for the HEN than for the Mg I 285.21 4.30 remaining nebulizers. Zn I 213.86 5.80 Pneumatic concentric nebulizers are normally able to aEsum=ionization potential+excitation potential. aspirate a solution at a given flow rate without any external pumping device.At a 0.7 l min-1 gas flow rate (Qg) the free liquid uptake rates were 40 and 90 ml min-1 for the HEN and the system was kept at its normal value, i.e., each pixel was MM, respectively. Although this magnitude was not measured considered as a whole. for the MCN, recent studies23 reported that the free liquid A standard solution containing Cr, Zn, Mn, Mg, Sr, Li and uptake rate for an MCN lies between 30 and 50 ml min-1 K (1 mg ml-1) was used. Table 3 lists the characteristics of the depending on the gas flow rate.The conventional nebulizer, lines employed. in turn, gave a liquid uptake rate of 160 ml min-1. These results can also be accounted for by considering that the Results and discussion smaller the inner diameter of the liquid capillary, the lower is the free aspiration rate (Table 1).41 Design considerations Fig. 1 displays the images of the tips corresponding to the Primary drop size distributions three PCMNs, showing that the HEN, MCN and MM are Few studies dealing with the drop size characterization of the diVerent in design.Thus, for the HEN [Fig. 1(a)], both liquid aerosols generated by the PCMN have been published3–6,12,19 and gas outlets are located on the same plane. For the MCN and they are mainly devoted to the comparison of the PCMN [Fig. 1(c)], the liquid capillary ends outside the nebulizer. In with conventional nebulizers. Fig. 2 shows that for the three the case of the MM, Fig. 1(b) indicates that the sample micronebulizers the primary aerosols become coarser (D3,2 capillary exhibits a recess with respect to the nebulizer tip. increases) as Ql increases. This widely described behavior is According to the pictures shown in Fig. 1, the gas–liquid typical for pneumatic concentric nebulizers and it is due to interaction for the HEN takes place just at the exit of the the increase in the energy per unit mass ratio, thus enhancing nebulizer. For the MCN this interaction is produced at a given the generation of liquid surface (i.e., finer aerosols).4–6,42–44 distance from the gas exit [approximate measurements indi- Fig. 2 also shows that the three micronebulizers generated cated that the portion of liquid capillary depicted in Fig. 1(c) primary aerosols with lower D3,2 values (i.e., finer) than the was around 400 mm in length]. In the case of the MM the conventional nebulizer. This is reasonably the combined liquid is subject to the action of the high velocity gas stream result of their smaller gas and liquid sections and thinner wall inside the nebulizer. As a result of these diVerences, it can be of the liquid capillary, that make the liquid–gas interaction predicted that the gas–liquid interaction will be more eYcient more eYcient.40,44 One should highlight the importance of the for the MM than for the MCN.This assessment is supported capillary diameter on the characteristics of the aerosol by the fact that for the MCN the aerosol generation is generated by comparing the data for PN and MM.produced when the gas stream has lost a fraction of its initial Among the micronebulizers, the HEN generated the finest kinetic energy.39,40 From this point of view, the HEN primary aerosols (i.e., lowest D3,2, Fig. 2). These expected represents an intermediate situation. data were the result ofhe higher gas back-pressure (i.e., Besides these considerations, the back-pressure is a very kinetic energy) for the HEN.40,42–44 However, the aerosols useful parameter to evaluate the potential nebulizer performgenerated by the MM are slightly finer than those generated ance.According to the gas section areas given in Table 1, for by the MCN. The reason for this behavior is probably that, as has already been pointed out for the MCN, the aerosol Fig. 2 Primary aerosol mean surface diameter (D3,2) versus liquid flow Fig. 1 Images of the nebulizer tip. (a) High-eYciency nebulizer; rate, Ql.(A) High-eYciency nebulizer; (B) micromist; (C) microconcentric nebulizer; (D) conventional pneumatic concentric nebulizer. (b) micromist; (c) microconcentric nebulizer; (d) conventional pneumatic concentric nebulizer. Qg=0.8 l min-1. J. Anal. At. Spectrom., 1999, 14, 1289–1295 1291Fig. 4 Volume drop size distribution curves for the tertiary aerosols. Fig. 3 Volume drop size distribution curves for the primary aerosols. (A) High-eYciency nebulizer; (B) micromist; (C) microconcentric (A) High-eYciency nebulizer; (B) micromist; (C) microconcentric nebulizer; (D) conventional pneumatic concentric nebulizer. Ql= nebulizer; (D) conventional pneumatic concentric nebulizer.Ql= 0.6 ml min-1; Qg=0.8 l min-1. 0.6 ml min-1; Qg=0.8 l min-1. comparison between the diVerent nebulizers led to similar conclusions as for primary aerosols. Again, the micronebulizers generation takes place at a given distance from the gas outlet. In this zone, the initial amount of kinetic energy available to generated finer tertiary aerosols than the conventional nebulizer.Liu and Montaser5 achieved a similar conclusion for the generate the aerosol is reduced because of the expansion of the gas stream produced at the nebulizer gas vent. Therefore, HEN in comparison with concentric and cross-flow standard nebulizers. On comparing the three micronebulizers, the HEN the amount of energy useful to spread out the liquid bulk is reduced.Hence the gas–liquid interaction is less eYcient, thus gave rise to the finest tertiary aerosols (Fig. 4). Droplets with diameters larger than 8 mm have been reported giving rise to coarser aerosols than expected according to the back-pressure values. not to be useful for analytical purposes. Olesik and Fister46 found that these droplets were not completely vaporized in Fig. 3 shows the volume drop size distribution curves, presented as the variation of the percentage of liquid volume the plasma measurement zone, leading to large plasma perturbation and noise.According to the data in Fig. 4, the percent- contained in droplets under a given drop diameter versus the drop diameter, D (i.e., undersize representation) for the four ages of liquid volume contained in droplets larger than 8 mm (V8) were 3, 11, 16 and 26% for the HEN, MM, MCN and nebulizers, showing the same trends as in Fig. 2. Hence, under the conditions employed in Fig. 3, the percentage of the aerosol PN, respectively.Another useful parameter supplied by the particle sizer is volume contained in droplets with diameters under 9.6 mm was 86, 63, 58 and 41% for the HEN, MM, MCN and PN, the so-called volume concentration (VC). This parameter is defined as the fraction of the active laser beam probe volume respectively. The results obtained for the HEN agree with previously published work by Olesik et al.4 Note that diVerent that is occupied by droplets.45 Therefore, this variable gives an indication of the amount of solution that is transported liquid and gas conditions are employed.Summarizing these results, the reduced dimensions of the sample capillary and out of the spray chamber. The examination of the VC for the tertiary aerosols indicated that (i ) VC increases on increasing the gas outlet with respect to a conventional nebulizer allow the micronebulizers to work eYciently at very low Ql values. the liquid flow rate (for the HEN, VC increased from 0.02 to 0.05% when Ql increased from 0.1 to 0.6 ml min-1) and (ii) the HEN provides the highest VC values (0.05, 0.04, 0.03 and Tertiary drop size distributions 0.02% for the HEN, MM, MCN and PN, respectively).The behavior of the tertiary aerosols as the liquid flow rate On considering simultaneously the results for V8 and VC, a was varied was close to that shown by the primary aerosols lower background noise for the PCMNs than for the MN (i.e., the lower the Ql value, the finer the tertiary aerosols). could be expected.This assessment is based on the fact that Nevertheless, the relative variations of the tertiary aerosol the absolute volume of the tertiary aerosol contained in characteristics as Ql was varied were less pronounced than droplets larger than 8 mm (i.e., proportional to V8×VC) was those of the primary aerosols because of the processes taking lower for the former. Hence, it can be easily derived that, at place inside the spray chamber (i.e., drop evaporation, coagu- 0.6 ml min-1 and 0.8 l min-1, this parameter took values of lation, losses of the coarsest droplets, etc.).These phenomena 0.15, 0.44, 0.48 and 0.52 for the HEN, MM, MCN and PN, tend to dampen the diVerences between the statistical param- respectively. eters of the tertiary aerosols with respect to those found for primary aerosols. In addition, several problems linked to the Solvent and analyte transport rates diYculty of measuring very low aerosol volumes appeared.4 Hence the examination of the D3,2 values did not provide clear In order to gain more insight into the processes taking place along the aerosol path inside the spray chamber, the total evidence of the Ql eVect, because no significant variations were found on decreasing this parameter. In these cases, some mass of solvent, Stot, and analyte, Wtot, transport rates were measured.additional aerosol parameters could be examined, e.g., the median of the volume drop size distribution (D50).45 Thus, for Fig. 5 shows the eVect of Ql on (a) Stot and (b) Wtot for the four nebulizers. The variations of these transport param- instance, at a 0.8 l min-1 gas flow rate, the D50 values of the tertiary aerosols generated by the HEN changed from 2.7 to eters with the liquid flow rate were concomitant with those of VC. 3.0 mm when the liquid flow was varied from 100 to 600 ml min-1. These results agreed with previously published On considering the diVerent nebulizers, for a given flow rate, the values of Stot and Wtot agreed well with the tertiary work.5 Fig. 4 shows plots of the volume drop size distributions of aerosol VC data discussed above. Hence, the HEN aVorded the highest transport values since it generated the finest the tertiary aerosols for the HEN, MCN, MM and PN. The 1292 J. Anal. At. Spectrom., 1999, 14, 1289–1295duced chiefly from the aerosol at locations near the nebulizer tip, 100% solvent transport eYciency would be expected at the lowest Ql.In fact, an almost 100% solvent transport eYciency for the HEN at 10 ml min-1 was obtained (Table 4). Nonetheless, at 120 ml min-1, the amount of water required to saturate the argon stream represents only between 12 and 18% of the solution volume nebulized. Note that the contribution of the evaporation from the chamber walls was neglected. The solvent transport eYciencies for the micronebulizers at the lowest liquid flow rates were fairly high (92% in the case of HEN at 10 ml min-1).Other researchers12,21 have found transport eYciencies in the range 50–100% for liquid flow rates below 20 ml min-1. Solvent transport eYciencies of up to 50% have also been reported for an MCN coupled to a double-pass spray chamber at 100 ml min-1 by Woller et al.23 for a solution of ammonium malonate. In this case, an indirect method was applied to measure Stot. As regards the conventional pneumatic nebulizer, few reports about the solvent transport at low liquid flow rates have been published.The results given in Table 4 are in agreement with previously published work.2,49 Thus, for a conventional nebulizer–double-pass spray chamber combination, Hettipathirana and Davey49 found 12% for es at Ql= 100 ml min-1. The same trends were observed for the analyte transport eYciency, en (Table 4). At 10 ml min-1, en reached values of Fig. 5 Solvent [(a) Stot] and analyte [(b) Wtot] transport rates versus 55, 43, 42 and 23% for the HEN, MM, MCN and PN, liquid flow rate, Ql, for a double-pass spray chamber.(A) Highrespectively. Liu et al.6 reported that, for the HEN operated eYciency nebulizer; (B) micromist; (C) microconcentric nebulizer; at 1.0 l min-1 and coupled to a double-pass spray chamber, (D) conventional pneumatic concentric nebulizer. Qg=0.7 l min-1. the analyte transport eYciency dropped from 50 to around 7% on increasing Ql from 11 to 220 ml min-1.Olesik et al.4 Table 4 Solvent (es) and analyte (en) transport eYciencies (Qg= found en values of 20% using the same set-up at 50 ml min-1. 0.7 l min-1) From the data presented in Table 4, it can be observed that Ql/ml min-1 Parameter HEN MM MCN PN the PN gives rise to transport eYciencies around half those obtained for the HEN. Olesik and co-workers48,50 found 10 es (%) 92 87 83 80 smaller diVerences between the analyte transport eYciencies 20 64 62 60 55 for an HEN and a PN (TR-30).In this case the former 40 41 36 34 33 nebulizer aVorded en values around 1.5 times higher than the 80 25 23 20 20 latter in the best of the cases.48 120 20 19 18 15 On comparing the es results with those for en, interesting 10 en (%) 55 43 42 23 conclusions can be drawn. First, as expected, es>en in all the 20 34 28 25 16 cases. Second, the relative improvement of es on switching 40 19 15 17 10 80 11 9 8 6 from PN to any of the PCMN was lower than that of en. 120 8 7 6 5 Thus, the solvent ratio (es)HEN/(es)PN was 1.2 at 10 ml min-1, whereas the analyte ratio (en)HEN/(en)PN reached a value of 2.4. These results can also be explained by considering that a large fraction of the solvent was transported through the spray primary aerosols and more solution mass was allowed to leave the spray chamber.38,43 chamber as a vapor. In addition, it is also apparent that the solvent mass evaporated from the inner walls of the spray The solvent transport eYciency (es) values in Table 4 indicate that es is a function of the liquid flow employed.The lower is chamber was not negligible. This factor is expected to be less significant as the liquid flow rate is increased. Therefore, it Ql, the higher are the es values.3,6,13,21 Several reasons could be put forward to explain this general behavior. The number can be concluded that the solvent evaporation inside the spray chamber is one of the predominating aerosol transport phen- of particles decreases as Ql drops.In this way, there will be less probability for the droplets to coagulate, thus generating omena when very low Ql are employed. As a consequence, parameters such as the environment temperature can modify larger droplets that could be more easily removed from the aerosol stream.47,48 In addition to this eVect, the solvent the values of both Wtot and Stot. The great importance of the solvent evaporation could also evaporation from the aerosol inside the spray chamber was, in relative terms, more significant as the liquid flow rate explain some data presented in a previous paper.19 Thus, for nitric and sulfuric acid solutions, a decrease in the emission diminished.This facilitated the aerosol transport through the spray chamber. It has been indicated that, at 25 °C, 20–30 mg signal with respect to plain water solutions was observed, this eVect being more pronounced at low (several tens of ml min-1) of water per liter of argon are required to saturate the gas stream.4,48 In our studies, a 0.7 l min-1 argon flow rate was than at high Ql (around 1 ml min-1). The extent of the solvent evaporation for acids is lower than for plain water since acid employed.Therefore, 14–21 ml of water could be evaporated per minute. The dotted line in Fig. 5(a) represents this value. solutions have a lower vapor pressure.51 Because of this, the analyte transport is expected to be more eYcient for water On comparing this line with the values of Stot, it can be concluded that a very important fraction of the solvent leaving than for acids. The (Wtot)acid/(Wtot)water ratio is expected to decrease as the liquid flow rate is reduced to a value for which the spray chamber is present in vapor form.Indeed, at the lowest values of Ql the tertiary aerosol was certainly not the solvent evaporation becomes one of the predominant aerosol transport phenomena. saturated with water.Assuming that the evaporation is pro- J. Anal. At. Spectrom., 1999, 14, 1289–1295 1293The LODs were calculated by employing the 3sb criterion, where sb is the standard deviation for 10 replicates of the blank. In general terms, the background noise produced by the PCMNs was lower than that produced by the PN, consistent with their smaller amounts of solution in droplets larger than 8 mm. Table 5 gives the LOD for the four nebulizers. As a result of the increased sensitivities and reduced background noise, the PCMNs achieved lower LOD values than the conventional nebulizer.Among the micronebulizers, the HEN gave the lowest LODs, since it exhibited the highest SBR values. The (LOD)HEN/(LOD)PN ratio was within the range 2.5–5. The background equivalent concentration values (BEC) Fig. 6 ICP-AES signal-to-background ratio (Mn 1 mg ml-1) versus Ql followed a similar trend. Table 5 indicates that the lowest BEC when using a double-pass spray chamber.(A) High-eYciency was for the HEN, followed by the MM and MCN and then nebulizer; (B) micromist; (C) microconcentric nebulizer; (D) conven- by the conventional nebulizer. These results can be accounted tional pneumatic concentric nebulizer. Qg=0.7 l min-1. for by the slightly lower background signals and the higher analyte emission signals obtained by the three PCMNs with ICP-AES analytical figures of merit respect to the PN. As expected, there is a correlation between the figures of merit and the data on aerosol characterization and solution trans- Conclusions port.As can be seen in Fig. 6, an increase in Ql led to increases From the data obtained in the present study, it can be in the SBR values. The rate of growth of the SBR became concluded that the use of micronebulizers as liquid sample smaller as the liquid flow rate increased. The comments made introduction systems for ICP-AES leads to good results when in the previous section about the ease of transport of an the sample volume or the solution rate is the limiting factor aerosol as the droplet density decreases should be applicable to perform the analysis.However, the conventional pneumatic here, also. This behavior was the same for the diVerent concentric nebulizer operated at several tens of microliters per elements and conditions tested, although, from Fig. 6, it minute has proved to be more useful than expected, since appears that a slight signal decrease was observed for the analyte transport eYciencies as good as 20% were reached.HEN at the highest liquid flow rate (i.e. 160 ml min-1). The Nevertheless, the high dead volume is still a limitation to deterioration of the plasma conditions as a result of the high apply the PN to the analysis of very low liquid sample volumes. solvent plasma load may be invoked to account for it. In Solvent evaporation was shown to be one of the major order to verify this hypothesis, the Mg II 280.27 nm to Mg I processes inside the spray chamber when working at low liquid 285.13 nm intensity ratio was calculated, as suggested else- flow rates.where.52,53 This parameter revealed that actually the plasma The best nebulizer in terms of primary aerosol features, conditions had slightly deteriorated. The ratio decreased from solution transport and ICP-AES analytical behavior was the 6.0 to 5.5 when Ql was increased from 10 to 160 ml min-1. high-eYciency nebulizer.Nevertheless, the high back-pressure Under the same conditions, the three micronebulizers required for the HEN may be a limitation for some aVorded higher SBR values than the conventional nebulizer. commercially available ICP-AES systems. This result was in agreement with the transport results (Fig. 5 The use of pneumatic concentric micronebulizers used in and Table 4). Furthermore, the signal improvement correthe direct injection mode (DIN), or incorporating desolvation sponded more or less to that of the analyte transport, i.e.at units (MCN6000), would further increase the ICP-AES sensi- 10 ml min-1 the (SBR)HEN/(SBR)PN ratio for Mn was 2.8, tivities and decrease the LODs. Nevertheless, the DIN has whereas the transport ratio was 2.4. The slight diVerences some important limitations (i.e., prone to blocking, high cost, between these two ratios could be explained by the smaller diYcult to use). More work is required in order to optimize mean sizes of the tertiary aerosols obtained when using the these systems further.One recent example is the DIHEN HEN, thus leading to a more eYcient analyte excitation developed by Montaser and co-workers.3,12,13 A comparison process. This explanation has also been suggested recently.48 among these two direct injection nebulizers would be very As happened for Wtot, the lower the Ql value, the higher was interesting. Also, a study to the response of the DIHEN to the improvement factor provided by the PCMNs.the matrix (acids, salts, etc.) eVects would be advisable. Table 5 Limits of detection (LOD) and background equivalent The authors thank Perkin-Elmer for the loan of the Optima concentrations (BEC) (Qg=0.7 l min-1; Q1=10 ml min-1) 3000 DV ICP-AES system. They are also grateful to Spin France and CETAC Technologies for the availability of the Element Parameter HEN MM MCN PN MCN. The MM was provided by courtesy of Glass Expansion Cr LOD/ng ml-1 12 17 15 30 (Switzerland).J.L.T. also thanks the European Commission Zn 10 12 14 26 for a Marie Curie Grant (Project No. ERBFMBICT961632). K 9 15 18 35 Mn 0.9 1.0 1.5 3.0 Mg II 0.6 1.4 2.2 4.0 References Sr 0.3 0.2 0.4 0.8 Li 0.4 0.8 0.5 1.3 1 Sample Introduction in Atomic Spectroscopy, ed. J. Sneddon, Elsevier, New York, 1990, p. 1. Cr BEC/ng ml-1 219 347 520 803 2 G. M. Hieftje, J. Anal. At. Spectrom., 1996, 11, 613. Zn 194 297 454 722 3 A. Montaser, M. G. Minnich, J. McLean, H.Liu, J. A. Caruso K 1444 4553 6141 9242 and C. W. McLeod, in Inductively Coupled Plasma Mass Mn 56 85 113 178 Spectrometry, ed. A. Montaser, Wiley-VCH, New York, 1998, Mg II 28 43 58 125 p. 83. Sr 31 52 70 204 4 J. W. Olesik, J. A. Kinzer and B. Harkleroad, Anal. Chem., 1994, Li 70 128 182 349 66, 2022. 1294 J. Anal. At. Spectrom., 1999, 14, 1289–12955 H. Liu and A. Montaser, Anal. Chem., 1994, 66, 3233. 29 CETAC Technologies Applications Bulletin, 003MCN1196U, CETAC Technologies, Omaha, NE, 1997. 6 H. Liu, R. H. CliVord, S. P. Dolan and A. Montaser, Spectrochim. 30 K. E. Lawrence, G. W. Rice and V. A. Fassel, Anal. Chem., 1984, Acta, Part B, 1996, 51, 27. 56, 289. 7 H. Liu, A. Montaser, S. P. Dolan and R. S. Schwartz, J. Anal. At. 31 D. R. Wiederin and R. S. Houk, Appl. Spectrosc., 1991, 45, 1408. Spectrom., 1996, 11, 307. 32 S. C. K. Shum, S. K. Johnson, H. Pang and R. S. Houk, Appl. 8 S. H. Nam, J. S. Lim and A. Montaser, J. Anal. At. Spectrom., Spectrosc., 1993, 47, 575. 1994, 9, 1357. 33 J. S. Crain and J. T. Kiely, J. Anal. At. Spectrom., 1996, 11, 525. 9 S. A. Pergantis, E. M. Heithmar and T. A. Hinners, Anal. Chem., 34 D. R. Wiederin, F. G. Smith and R. S. Houk, Anal. Chem., 1991, 1995, 67, 4530. 63, 219. 10 J. A. Kinzer, J. W. Olesik and S. V. Olesik, Anal. Chem., 1996, 35 M. A. Tarr, G. Zhu and R. F. Browner, Anal. Chem., 1993, 65, 68, 3250. 1689. 11 Q. Lu, S. M. Bird and R. M. Barnes, Anal. Chem., 1995, 7, 2949. 36 L. Wang, S. W. May, R. F. Browner and S. H. Pollock, J. Anal. 12 J. A. McLean, H. Zhang and A. Montaser, Anal. Chem., 1998, At. Spectrom., 1996, 11, 1137. 70, 1012. 37 D. D. Smith and R. F. Browner, Anal. Chem., 1982, 54, 533. 13 J. A. McLean, M. G. Minnich, L. A. Iacone, H. Liu and 38 R. F. Browner, A. Canals and V. Hernandis, Spectrochim. Acta, A. Montaser, J. Anal. At. Spectrom., 1998, 13, 829. Part B, 1992, 47, 659. 14 C. B’Hymer, K. L. Sutton and J. Caruso, J. Anal. At. Spectrom., 39 A.H. Lefebvre, Atomization and Sprays, Hemisphere, New York, 1998, 13, 855. 1989. 15 Meinhard Associates Web Site, http://www.meinhard.com, 40 B. L. Sharp, J. Anal. At. Spectrom., 1988, 3, 613. December 1998. 41 A. Kundt and E.Warburg, Ann. Phys., 1975, 155, 337. 16 M. Vanhaecke, M. Van Holderbeke, L. Moens and R. Dams, 42 A. Canals, V. Hernandis and R. F. Browner, J. Anal. At. Spectrom., 1990, 5, 61. J. Anal. At. Spectrom., 1996, 11, 543. 43 J. L. Todolý�, A. Canals and V. Hernandis, Spectrochim. Acta, Part 17 S. Augagneur, B. Medina, J. Szupnar and R. Lobinski, J. Anal. B, 1993, 48, 373. At. Spectrom., 1996, 11, 713. 44 D. E. Nixon, Spectrochim. Acta, Part B, 1993, 48, 447. 18 C. Dubuisson, E. Poussel and J. M. Mermet, J. Anal. At. 45 2600 Laser DiVraction User Manual, Malvern Instruments, Spectrom., 1997, 12, 281. Malvern, 1987. 19 J. L. Todolý�, J. M. Mermet, A. Canals and V. Hernandis, J. Anal. 46 J. W. Olesik and J. C. Fister, Spectrochim. Acta, Part B, 1991, At. Spectrom., 1998, 13, 55. 46, 869. 20 J. L. Todolý� and J. M. Mermet, J. Anal. At. Spectrom., 1998, 47 A. Canals, V. Hernandis and R. F. Browner, Spectrochim. Acta, 13, 727. Part B, 1990, 45, 591. 21 T. D. Hettipathirana and D. E. Davey, J. Anal. At. Spectrom., 48 J. W. Olesik, presented at the 1999 European Winter Conference 1998, 13, 483. on Plasma Spectrochemistry, Pau, France, 10–15 January, 1999. 22 M. De Wit and R. Blust, J. Anal. At. Spectrom., 1998, 13, 515. 49 T. D. Hettipathirana and D. E. Davey, Appl. Spectrosc., 1996, 23 A. Woller, H. Garraud, J. Boisson, A. M. Dorthe, P. Fodor and 50, 1015. O. F. X. Donard, J. Anal. At. Spectrom., 1998, 13, 141. 50 J. W. Olesik, J. A. Kinzer, J. Hartshorne and A. JeVery, presented 24 E. Debrah, S. A. Beres, T. J. Gluodenis, Jr., R. J. Thomas and at the 1998 Winter Conference on Plasma Spectrochemistry, E. R. Denoyer, At. Spectrosc., 1995, 16, 197. Scottsdale, AZ, January 5–10, 1998. 25 K. A. Taylor, B. L. Sharp, D. J. Lewis and H. M. Crews, J. Anal. 51 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998, 13, At. Spectrom., 1998, 13, 1095. 1249. 26 CETAC Technologies Web Site, http://www.cetac.com, 52 J. M. Mermet, Spectrochim. Acta, Part B, 1989, 44, 1109. November 1998. 53 E. Poussel, J. M. Mermet and O. Samuel, Spectrochim. Acta, Part B, 1993, 48, 743. 27 Glass Expansion Web Site, http://www.glassex.com.au, November 1998. 28 R. I. Botto and J. J. Zhu, J. Anal. At. Spectrom., 1996, 11, 675. Paper 9/00598F J. Anal. At. Spectrom., 1999, 14, 1289–129