JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 799 Noise Characteristics of Aerosols Produced by Inductively Coupled Plasma Nebulizers* Shen Luan Ho-ming Pang Sam C. K. Shum and R. S. Houk? Ames Laboratory US Department of Energy and Department of Chemistry Iowa State University Ames IA 50011 USA The noise characteristics of aerosols produced by inductively coupled plasma nebulizers were investigated. A laser beam was scattered by aerosol and detected by a photomultiplier tube and the noise amplitude spectrum of the scattered radiation was measured by a spectrum analyser. Discrete frequency noise in the aerosol generated by a Meinhard nebulizer or a direct injection nebulizer was primarily caused by pulsation in the liquid flow from the pump. By use of a pulse-free pump such as a gas displacement pump or a dual piston pump for liquid chromatography the discrete frequency noise was eliminated.The configuration of the spray chamber affected the level of white noise. A Scott-type spray chamber suppressed white noise while a conical straight- pass spray chamber enhanced white noise relative to the noise seen from the primary aerosol. The noise in the aerosol from a continuous-flow ultrasonic nebulizer had a relatively high 1 / f component. Keywords Noise amplitude spectra; aerosol; nebulizer; inductively coupled plasma mass spectrometry; inductively coupled plasma atomic emission spectrometry The precision detection limits and dynamic range of instrumental measurements including inductively coupled plasma mass spectrometry (ICP-MS) and inductively coup- led plasma atomic emission spectrometry (ICP-AES) are generally influenced by noise in the signals measured.The optimization of experimental variables is often based on reducing the relative amount of noise or increasing the signal-to-noise ratio (S/N). However the S/N provides only an inclusive view of the performance of the system. A more fundamental understanding of noise characteristics re- quires a knowledge of the noise power spectrum (NPS) which may identify the types origins and frequency composition of the noise. Information on noise character- istics sometimes provides a sound rationale for improving the performance of the instrument. The NPS for ICP-MS1-3 and ICP-AES4-15 have been reported and the prevailing types of noise identified as (i) white noise (ii) l/f noise and (iii) interference noise.16 Noise in the ICP from large undissociated wet droplets and undesolvated dry particles has also been characterized exten~ively.~~-~* In ICP spectrometry the precision that can be achieved is often believed to be limited by the sample introduction s y ~ t e m .~ ~ - ~ ~ To improve the stability and the precision of the ICP sample introduction system the sources of noise should be isolated and characterized. Despite the numerous studies of noise behaviour of mass spectrometric or emission signals from the ICP no pub- lished results are known to exist that characterize the noise behaviour of the aerosols themselves in a thorough manner. This is the objective of the present work.Montaser et aLZ7 have described interferometric measurements of light scat- tered from aerosols produced by ultrasonic nebulizers. These results have not yet been published; however the main objective of Montaser's study was to measure droplet sizes and velocities rather than noise behaviour. Experimental A block diagram of the experimental set-up used for noise measurements is shown in Fig. 1. The instrumentation used is summarized in Table 1. The experiments were performed in a darkened room. Vibrations and air flows were minimized. *Presented at the 1992 Winter Conference on Plasma Spectro- chemistry San Diego CA USA January 6-1 1 1992. tTo whom correspondence should be addressed. Aerosol 543.5 nm ';$ He-Ne laser analyser Fig. 1 Block diagram of apparatus used for NAS measurements The spectrum analyser actually displayed the square root of the power spectrum sometimes called the amplitude spectrum.The ordinate of all plots was scaled in units of dBV [(l dBV=20 log A where A was the signal amplitude in volts root mean square (RMS)]. Note that this dBV unit was essentially an absolute measure of the noise level whereas the dB unit in our previous work' measured noise (at a specific frequency) relative to the d.c. level. The spectra shown were not normalized to the d.c. level so that the d.c. level (i.e. the scattering signal) observed for the various nebulizers and spray chambers could also be presented. Since the dBV scale is logarithmic the values can readily be converted into relative units (i.e. dB) by simply taking the difference between the noise amplitude at the frequency of interest and the dBV reading at the d.c.level (frequency = 0 Hz). The apparatus shown in Fig. 1 was similar to that used for laser Doppler spectroscopy (LDS) for particle size measure- m e n t ~ . ~ ~ In the present work aerosol droplet size was not measured. Instead noise amplitude spectra (NAS) were taken to characterize the noise sources in the aerosols generated by ICP nebulizers. The noise behaviour of the laser light scattered from the aerosol droplets was assumed to be related to noise processes in the production and transport of the aerosol itself. Spectra measured at several scattering angles had the same basic features as those measured at 15" which was used subsequently because it provided the highest scattering signal.A Meinhard nebulizer (Table 1) was studied because of its common use in ICP spectrometry. Distilled de-ionized water was the only sample used in this work. Different pumps were employed to deliver water to the Meinhard nebulizer as shown in Table 1. Self-priming aspiration (or natural uptake) was also investigated. The direct injection nebulizer (DIN) was similar to those described r e ~ e n t l y . ~ ~ - ~ ~ The internal diameter of the sample800 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 Table 1 Instrumental components Laser Component Manufact ureddescri ption Melles Griot GreNe green He-Ne cylindrical laser head and power supply (Carlsbad CA USA) Focusing lens Iris diaphragm Interference filter PMT Rolyn spherical plano-convex glass lens (Covina CA USA) Edmund precision iris diaphragm (Barrington NJ USA) Melles Griot He-Ne laser line interference filter (Carlsbad) Hamamatsu R 955 side-on type (Japan) PMT power supply Tennelec/Nucleus TC 952 high voltage power supply (Oak Ridge TN USA) (Loveland CO USA) Spectrum analyser Hewlett-Packard Model 3582A X-Y recorder Picoammeter Meinhard nebulizer Direct injection nebulizer Spray chamber Ultrasonic nebulizer Peristaltic pump Syringe pump Single head LC pump Dual head LC pump Houston Omnigraphic Series 2000 Keithley Model 485 autoranging Meinhard Type TR-30-C3 (Austin TX USA) picoammeter (Cleveland OH USA) borosilicate glass concentric nebulizer (Santa Ana CA USA) Construction as described r e ~ e n t l y ~ ~ - ~ ~ Scott-type double-pass spray chamber34 (Precision Glassblowing Englewood CO USA) chamber (Fig.7 of ref. 35 but without any mixing baffle) Continuous-flow similar to the designs of Fassel and Bear,35 Model CPMT transducer (Channel Products Chagrin Falls OH USA) Plasma-Therm Model UNPS- 1 r.f. power supply (Kresson NJ USA) Gilson Minipuls 2 HP-I single channel peristaltic pump (Middleton WI USA) Sage Model 34 1 B single channel syringe pump (Orion Research Boston MA USA) with B-D Plastipak syringe (Becton Dickinson Rutherford NJ USA) Conical straight-pass spray SSI Model 2221) digital HPLC pump (Scientific Systems State College PA USA) with pulse damper Varian 2010 LC pump (Varian Instrument Group Walnut Creek CA USA) with SSI Model LP-21 LO-PULSE pulse damper (Scientific Systems) *Should be free from alias contamination.28 Operating conditions/specifications Wavelength 543.5 nm Power 5 mW Continuous wave Scattering angle 8 15" Aperture 2.5 mm Wavelength 543.5 nm Full width at half maximum 10 f 2 nm Bias voltage -300 V Frequency range d.c.to ~ 2 5 kHz* Sensitivity selects the maximum input level that can be applied to the instrument without overloading Coupling d.c. Passband shape Hanning Averaging mode RMS Number of averages 4 for 1 Hz 16 for 10 Hz 64 for more than 100 Hz 1.d. of sample delivery capillary 30 pm Resonant frequency of transducer 1.36 MHz Incident r.f. power 40 W Reflected r.f. power 0-1 W Liquid flow rate 2.5 cm3 min-' Argon carrier gas flow rate 1.5 dm3 min-I capillary was 30 pm.33 A single head liquid chromatography (LC) pump (Table l ) a dual head LC pump (Table 1 ) and a laboratory-made gas displacement pump29 were used to deliver water to the DIN.Aerosol NAS were measured directly on the aerosol from the Meinhard nebulizer or DIN and after the primary aerosol passed through either a Scott-type double-pass spray chamber or a conical straight-pass spray chamberJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 80 1 (Table 1). Ideally the tertiary aerosol should also be probed at the exit port of the injector tube of the ICP t o r ~ h . ~ ~ . ~ ~ This was not feasible however because the wet aerosol condensed at the cold tip of the injector The aerosol that coalesced on the inside surface of the spray chamber eventually flowed to the chamber drain and then to a waste receptacle.Irregularities in this flow may have caused pressure fluctuations in the chamber as the liquid flow ran into the waste receptacle. In the present work appropriate steps were taken to suppress these pressure pulsation^.^^.^^-^^ Finally noise in the aerosol generated with an ultrasonic nebulizer (USN) (Table 1) was studied. Water was intro- duced into the USN using the dual head LC pump (Table 1). Again NAS were measured on the wet droplets leaving a straight-pass spray chamber. Desolvation was not em- ployed. The spatial position where the aerosol was pro- duced on the face plate of the transducer fluctuated with time which precluded reliable measurements of NAS of the primary aerosol issuing directly from the surface of the USN. Argon was used throughout the present work as nebulizer gas (for the pneumatic nebulizers) and carrier gas (for the USN).The flow rate was controlled with a Matheson 7600 series flowmeter equipped with tube No. 602 (Matheson Gas Products Secaucus NJ USA). The needle valve on this flowmeter was positioned at the outlet. This arrange- ment allowed accurate measurements of flow rate for any back-pressure (or downstream pressure) provided that the gas pressure fed to the flow meter stayed equal to the pressure for which the tube was calibrated. Each nebulization system was oriented horizontally on a movable platform such that the distance between the nebulizer tip or the outlet of the spray chamber and the point of interaction with the laser beam could be altered.Usually the distance was 10.0 mm. The vertical position was also adjusted so that the laser beam could be directed into the centre of the aerosol plume. Results and Discussion Control Experiments A series of experiments were performed without nebulizers to determine whether the measurement system (i.e. laser photomultiplier tube (PMT) and spectrum analyser) contri- buted any significant noise. The NAS were measured from (i) the d.c. current output of a battery connected directly to the input of the spectrum analyser (ii) the output of a flashlight directed onto the PMT (iii) the attenuated output of the laser sent directly onto the PMT and (iv) laser scattering from a piece of abrasive paper with 42 ,um diameter particles detected by the PMT. In each case the signal measured by the picoammeter was - 1 PA which was similar to the magnitude of the scattering signal seen subsequently from aerosols.For each control experiment the white noise was less than - 120 dBV. Some small discrete frequency peaks (< - 1 10 dBV) were also observed. All these ‘instrumental’ noise levels were insignificant compared with the noise levels in the light scattered from the aerosols. Furthermore the background current of the PMT was -3 PA which was also insignificant compared with the usual scattering signals of = 1 PA. Based on these results the apparatus did not contribute appreciable noise compared with that seen from the light scattered by the aerosols. Noise from Aerosol Generated by Meinhard Nebulizer The NAS of scattering from the aerosol generated by a Meinhard nebulizer with liquid delivered by a peristaltic 10 0 -10 -20 -30 -40 -50 2 -60 D 2 -70 \ .- 2 0 2 4 6 8 10 P E 10 - .- s o 0 -10 -20 -30 -40 -50 -60 0 20 40 60 80 100 FrequencyIHz Fig.2 (a) NAS of scattering from the primary aerosol generated by a Meinhard nebulizer with a peristaltic pump; and (6) NAS of the primary aerosol generated by a Meinhard nebulizer with self- priming aspiration (or natural uptake). No spray chamber was used for either (a) or (6). The liquid flow rate was 1.0 dm3 min-I. The argon nebulizer gas flow rate was 1.0 dm3 min-’ pump is shown in Fig. 2(a). Three types of noise are observed white noise ( i e . the asymptote at the higher end of the frequency axis) llfnoise (the gradual decline of noise amplitude as frequency increases from 0 Hz) and interfer- ence noise (the peaks at discrete frequencies).The fre- quency axis in Fig. 2(a) stopped at 10 Hz because no discrete noise peaks were seen at higher frequencies which was the case for all the nebulizers pumps efc. evaluated in this study. In Fig. 2(a) the interference noise peaks at 1.88 and 2.84 Hz were harmonics of the 0.96 Hz fundamental noise. The frequencies of these peaks did not change as the nebulizer gas flow rate and the distance between the nebulizer tip and the point of interaction with the laser beam (i.e. the ‘scattering position’) were changed. However the peak frequencies increased with liquid flow rate (Table 2). At each liquid flow rate studied the fundamental frequency of the pulsation measured from the NAS was essentially the same as the frequency with which a fresh roller touched the compressible pump tubing as the head of the pump rotated.Thus the pulsation in the liquid flow induced the discrete noise peaks observed in the NAS. Similar peaks have been seen in NPS from ICP s i g n a l ~ . ~ ? ~ J ~ Table 2 Pulsation induced by peristaltic pump* Liquid flow rate/ Frequency of Frequency of cm3 min-I pulsation/Hz fundamental noise/Hz 0.84 0.79 0.80 1 .o 0.95 0.96 1.5 1.36 1.36 *All frequencies in this table were measured three timcs cach measurement yielded the same value.802 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 Table 3 Pulsation induced by various pumps* Liquid flow rate/ Frequency of Pump cm3 min- ' fundamental noise/Hz Syringe 0.78 5.2 1.1 7.2 1.6 10.8 Single head LC 0.800 0.32 Single head LC + pulse damper 0.800 0.32t *All frequencies in this table were measured three times each measurement yielded the same value. ?The amplitude of the pulsation was less severe than when the single head LC pump was used alone (without the pulse damper).Also similar peaks at discrete frequencies were seen from NAS taken when the Meinhard nebulizer was fed by the syringe pump or the single head LC pump (Table 3). In each case the frequency of the fundamental noise peak was the same as the frequency of a fluctuation in the liquid flow rate caused by the mechanism that drove the pump.44-46 The amplitude of these noise peaks was lowest (= - 30 dBV) for the syringe pump but such peaks were still present which rebutted the occasional claim that syringe pumps deliver pulseless flow.The dual head LC p ~ m p ~ ~ ~ ~ ~ did not produce detectable interference noise peaks. As shown in Fig. 2(a) harmonic peaks were seen at both odd and even integer multiples of the fundamental fre- quency. The intensity of the harmonic peaks fell off as frequency increased. A similar pattern of harmonic peaks is obtained in the frequency domain when a full-wave rectified sine wave in the time domain is subjected to the Fourier transform process.48 Thus the liquid flow from the peristaltic pump the syringe pump and the single head LC pump is modulated by a full-wave rectified sine wave pattern which is physically reasonable. Fourier transforma- tion of a sawtooth wave also yields this pattern of R (=' -30 -90 1 I I I I I 0 0.2 0.4 0.6 0.8 1.0 Frequency/Hz Fig.3 NAS of DIN aerosol with a single head LC pump (a) without the pulse damper; and (b) with the pulse damper. The liquid flow rate was 50 mm3 min-' harmonics in the frequency domain but the flow output of a pump probably does not have a sawtooth character. Self-priming aspiration (or natural uptake) was also used to deliver liquid. As expected no discrete frequency peaks were present in the noise spectra [Fig. 2(b)]. Comparison of Figs. 2(a) and 2(b) showed that for unknown reasons the white noise level (ie. the baseline asymptote at high frequency) was higher by about 10 dBV when the Meinhard nebulizer was fed by natural uptake. Noise from Aerosol Generated by DIN The NAS from a DIN aerosol using a single head LC pump is shown in Fig.3. Without the pulse damper substantial interference noise was observed [Fig. 3(a)]. The frequency of the fundamental noise was found to be 0.020 Hz which again corresponded to the frequency of the pumpinghefill stroke at the very low liquid flow rate used (50 mm3 min-I). With the pulse damper the interference noise was elimi- nated [(Fig. 3(b)]. Pulse-free pumps including a dual head LC pump (with pulse damper) and a gas displacement pump,29 were investigated with the DIN. As expected no interference noise was present. Only the pulse-free pumps were used for subsequent experiments to suppress noise peaks so that the llfstructure in the NAS could be seen clearly. It is interesting that the NAS obtained with a single head LC pump (with pulse damper) a dual head LC pump (with pulse damper) and a gas displacement pump all gave basically identical llfprofiles and similar white noise levels (measured relative to the d.c.level). Thus the white noise and llfnoise in the NAS were not greatly influenced by the pumps used. Comparison of Noise from Meinhard Nebulizer and DIN The NAS of scattering from the aerosol from the Meinhard nebulizer is compared with that from the DIN in Fig. 4. These two spectra would be fairly similar if they were normalized to the same d.c. level. In each instance the white noise levels were -50 dBV below the d.c. level. The d.c. levels were related to the total scattering signals seen from the two nebulizers. These d.c. levels differed by - 15 dBV because the liquid flow rates and droplet size distribu- tions were different for the two nebulizers.The contribution of white noise to the relative standard deviation (RSD) of a measured signal can be estimated from the white noise level N (in dB) using the following equation:lJO RSDx 1 (1) 10 I I 0 2 -10 -g -20 9 CI .- - a -30 E -40 v) .- O -50 z -60 -70 1 I 1 I 0 20 40 60 80 100 FrequencyIHz Fig. 4 Comparison between the NAS of the aerosol from A Meinhard nebulizer (the liquid flow rate was 1 .O cm3 min-I) and B a DIN (the liquid flow rate was 50 mm3 min-I). A dual head LC pump was used in both instancesJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992. VOL. 7 803 The value of N is measured below (or relative to) the d.c. level. For example in Fig.4 B the white noise level is -65 dBV and the d.c. level is - 15 dBV so N in eqn. (1) is (- 65)-( - 15)= - 50 dB. This subtraction process is equi- valent to normalizing the noise amplitude at the chosen frequency to the d.c. level. In the present study white noise levels of - -48 to - 50 dB were observed for both the Meinhard nebulizer and the DIN. Thus white noise as the sole noise source would be expected to yield RSDs of 0.3-0.4%. These RSDs were comparable to those obtained from DIN experiments in ICP-MS at relatively high signal levels ( x 1 x los-1 x lo6 counts s - ' ) . ~ ~ In contrast this level of white noise is a minor contribution to the total signal instability of ~ 2 % RSD or worse that is typical of many different ICP-MS devices at high signal levels when a Meinhard nebulizer is used.49 -lo fi -30 c .- -50 I 5 % I \ I -70 tz -90 1 0 20 40 60 80 100 FrequencyjHz Fig.5 NAS from aerosol from a Meinhard nebulizer after passage through a Scott-type7 double-pass spray chamber. A dual head LC pump was used. The liquid flow rate was 1 .O cm3 min-I. The argon nebulizer gas flow rate was 1.0 dm3 min-' -40 -50 -60 F 3 -70 P 10 > c1 .- - 5 0 .- $ 0 -10 z -20 -30 -40 -50 -60 -70 t h I I I I 1 0 20 40 60 80 100 F requency/Hz Fig. 6 NAS from aerosols from a Meinhard nebulizer after passage through a conical straight-pass spray chamber. A dual head LC pump was used. The liquid flow rate was 1.0 cm3 min-'. The argon nebulizer gas flow rate was 1.0 dm3 min-I. The argon make-up gas flow rate was (a) 0; and (6) 0.5 dm3 min-I Effect of Spray Chamber on Noise from Aerosol The NAS from aerosol from a Meinhard nebulizer after it passed through a Scott-type double-pass spray chamber with the use of a relatively pulseless pump (the dual head LC pump) is shown in Fig.5. No discrete frequency noise was observed. Note that the white noise level N was x (-82)-(-22)= -60 dB (Fig. 5) which was x 10 dB lower than that present in the primary aerosol (Fig. 4 A). The -60 dB value would correspond to an RSD of -0.1%. Compared with the RSDs of 0.3-0.4% expected for the primary aerosol (as described in the preceeding paragraph) the Scott-type double pass spray chamber apparently improved precision by a factor of 3-4 The NAS from aerosols that were passed through a conical straight-pass spray chamber are shown in Fig.6. Again no discrete peaks were observed. Apparently neither spray chamber contributed noticeable interference noise. Thus the usual audible noise peaks at 200-400 Hz that were seen in either ICP-MS or ICP-AESloJs cannot be blamed on either the nebulizer or the spray chamber. Without argon make-up gas flow (introduced through the spray chamber) the llfnoise in Fig. 6(a) was very large as indicated by the slow decline in noise amplitude as frequency increased. Next an argon make-up gas was added through the drain tube of the spray chambeP at 0.5 dm3 min-l. The total gas flow rate ( i e . the nebulizer gas flow rate plus the make-up gas flow rate) was 1.5 dm3 min-I. Addition of the make-up gas suppressed the llf noise substantially [Fig. 6(b)].Thus the gas flow patterns through a spray chamber can affect the characteristics of noise in the aerosol leaving the chamber. Also it is interesting to note that the white noise level N in Fig. 6(b) is -(-43)-(-8)=-35 dB which is =15 dB higher than that present in the primary aerosol (Fig. 4 A). This indicated that the precision was poorer (by a factor of 5-6) after the aerosol passed through the conical straight-pass spray chamber. This degradation in precision agreed with previous experience with the use of a conical straight-pass spray chamber in ICP-MS experiments with the DIN29 and in other work in ICP-AES.sO The effect of the two spray chambers on llfnoise was illustrated by comparing the NAS from the primary aerosol produced by the Meinhard nebulizer with the NAS from the secondary aerosol produced by the same Meinhard nebu- lizer after passage through either a Scott-type double-pass spray chamber or a conical straight-pass spray chamber.In this instance only each of these spectra was normalized to 0.0 a -0 .- 2 -0.2 E" .- 2 - a -0.4 0 C -0.6 U E .- -0.8 E z -1.0 0 20 40 60 80 100 Frequency/Hz Fig. 7 Normalized NAS of the aerosols from A a Meinhard nebulizer alone; B the same Meinhard nebulizer with a Scott-type double-pass spray chamber; and C the same Meinhard nebulizer with a conical straight-pass spray chamber with make-up gas. Each spectrum is normalized to the same white noise level at - 100 Hz. A dual head LC pump was used and the liquid flow rate was 1.0 cm3 min-'804 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL.7 lo c 0 2 -10 4 -20 F 4- .- - a -30 Em -40 .- % 0 -50 Z -60 1 -70 1 0 20 40 60 80 100 FrequencyIHz Fig. 8 NAS of the aerosol generated by USN (after passage through a straight-pass spray chamber). A dual head LC pump was used. The liquid flow rate was 2.5 cm3 min-l and the argon carrier gas flow rate was 1.5 dm3 mind' the same white noise level (- I .O). As shown in Fig. 7 this normalization process facilitated visual comparison of the extent to which noise amplitude dropped off as frequency increased. Curves B and C were obtained with spray chambers. The Ilfportions of these two curves were similar and lay slightly below that for curve A. In other words the noise dropped off somewhat faster and approached the white noise limit at a lower frequency when either spray chamber was used. Thus use of either spray chamber apparently attenuated 1 lf noise slightly.Such information was hard to discern unless the spectra were normalized to the same white noise level. Noise from Aerosol Generated by USN The NAS of the aerosol generated by a USN with the usual straight-pass spray chamber is shown in Fig. 8.35 The d.c. signal (i.e. the point at 0 Hz) was 13 dBV higher than that for the Meinhard nebulizer at a comparable total gas flow rate [1.5 dm3 min-l Fig. 6(b)]. A higher scattering signal was expected from the USN partly because of its higher uptake rate (2.5 cm3 mine'). It can also be seen from Fig. 8 that a fairly large amount of llfnoise was observed from the USN. The noise amplitude still dropped as frequency increased up to frequencies of at least 100 Hz.The llfnoise profile was worse for the USN (than for the pneumatic nebulizers) regardless of whether or not the NAS were normalized to the d.c. level. The spray chamber for the USN was very similar to the conical straight-pass spray chamber described in the last section. The carrier gas for the USN was added through the same drain port as the make-up gas when the Meinhard nebulizer was used with this type of spray chamber. With the Meinhard nebulizer addition of make-up gas greatly attenuated llfnoise [i.e. compare Figs. 6(a) and 6(b)]. The USN itself was apparently more susceptible to llfnoise than the Meinhard nebulizer because llf noise was still substantial with the USN when the same spray chambers were used for either nebulizer.Discrete frequency noise was not noticeable; however it could have been masked by the llf noise. Conclusions The present study illustrates several points of practical interest for analysis by ICP-MS or ICP-AES (i) interference noise in nebulization is at relatively low frequencies (ie. 0.02-10 Hz) and is caused by pump fluctuations not by either the nebulizer or spray chamber; (ii) the configuration of and gas flow patterns through the spray chamber can affect levels of llfnoise and white noise. In particular the common Scott-type chamber suppresses both llf noise and white noise relative to that in the primary aerosol; (iii) the ultrasonic nebulizer suffers from additional 1 lf noise beyond that expected from the conical spray chamber usually employed; and (iv) the RSD of the scattering signal from the secondary aerosol can be as good as 0.1%.This value is comparable to the best precision commonly achievable for analytical signals from the ICP in cases where special care is taken to optimize precision. Examples include integrated intensity measurements with a thermo- stated spectrometer and Myers-Tracy corrections1 in ICP- AES or isotope ratio measurement by fast peak hopping in ICP-MS. While the sample introduction system can limit precision of spectrochemical measurments with the ICP this need not always be the case. The authors gratefully acknowledge R. Hendrickson of Department of Nuclear Engineering Iowa State University for the loan of the spectrum analyser and S. J. Weeks for the loan of the laser.The authors thank R. K. Winge D. E. Eckels F. G. Smith K. Hu and H.-C. Chang for their helpful discussion and assistance during the course of this work. Ames Laboratory is operated by Iowa State Univer- sity for the US Department of Energy under Contract No. W-7405-Eng-82. This research was supported by the Office of Basic Energy Sciences Division of Chemical Sciences. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 References Crain J. S. Houk R. S. and Eckels D. E. Anal. Chem. 1989 61 606. Furuta N. Monnig C. A. Yang P. and Hieftje G. M. Spectrochim. Acta Part B 1989 44 649. Furuta N. J. Anal. At. Spectrom. 1991 6 199. Walden G. L. Bower J . N. Nikdel S. Bolton D. L. and Winefordner J. D. Spectrochim. Acta Part B 1980 35 535.Belchamber R. M. and Horlick G. Spectrochim. Acta Part B 1982 37 17. Benetti P. Bonelli A. Cambiaghi M. and Frigieri P. Spectrochim. Acta Part B 1982 37 1047. Davies J. and Snook R. D. J. Anal. At. Spectrom. 1986 1 195. Davies J. and Snook R. D. J. Anal. At. Spectrom. 1987 2 27. Antanavichyus R. L. Serapinas P. D. and Shimkus P. P. Opt. Spektrosk. 1987 63 224. Winge R. K. Eckels D. E. DeKalb E. L. and Fassel V. A. J. Anal At. Spectrom. 1988 3 849. Sing R. L. A. and Hubert J . J. Anal. At. Spectrom. 1988 3 835. Montaser A. Ishii I. Tan H. Clifford R. H. and Golightly D. W. Spectrochim. Acta Part B 1989 44 1163. Montaser A. Clifford R. H. Sinex S. A. and Capar S. G. J. Anal. At. Spectrom. 1989 4 499. Goudzwaard M. P. and de Loos-Vollebregt M. T. C. Spectrochim.Acta Part B 1990 45 887. Furuta N. Anal. Sci. 1990 6 683. Ingle J. D. Jr. and Crouch S. R. Spectrochemical Analysis Prentice Hall Englewood Cliffs 1988. Olesik J. W. Smith L. J. and Williamsen E. J. Anal. Chem. 1989 61 2002. Olesik J. W. and Fister J. C. 111 Spectrochim. Acta Part B 1991 46 851. Fister J . C. 111 and Olesik J . W. Spectrochim. Acta Part B 199 1 46 869. Hobbs S. E. and Olesik J . W. Anal. Chem. 1992 64 274. Cicerone M. T. and Farnsworth P. B. Spectrochim. Acta Part B 1989 44 897. Winge R. K. Crain J. S. and Houk R. S. J. Anal. At. Spectrom. 199 1 6 60 1. Browner R. F. and Boorn A. W. Anal. Chem. 1984 56 786A and 875A. McGeorge S. W. and Salin E. D. Appl. Spectrosc. 1985 39 989.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 805 25 Hobbs P.Spillane D. E. M. Snook R. D. and Thorne A. P. J. Anal. At. Spectrom. 1988 3 543. 26 Sneddon J. in Sample Introduction in Atomic Spectroscopy ed. Sneddon J. Elsevier Amsterdam 1990 ch. 1 p. 1. 27 Montaser A. Clifford R. H. and Sohal P. paper presented at the 199 1 European Winter Conference on Plasma Spectroche- mistry Dortmund Germany January 14- 18 199 1. 28 Hewlett-Packard Operating Manual for Model 3582A Spec- trum Analyzer Hewlett-Packard Loveland 1 978. 29 Wiederin D. R. Smith F. G. and Houk R. S. Anal. Chem. 1991 63 219. 30 Wiederin D. R. Smyczek R. E. and Houk R. S. Anal. Chem. 1991,63 1626. 31 Smith F. G. Wiederin D. R. and Houk R. S. Anal. Chim. Acta 1991 248 229. 32 Wiederin D. R. and Houk R. S. Appl. Spectrosc. 1991 45 1408. 33 Shum S.C. K. Neddersen R. and Houk R. S. Anal-vst 1992 117 577. 34 Scott R. H. Fassel V. A. Kniseley R. N. and Nixon D. E. Anal. Chem. 1974 46 75. 35 Fassel V. A. and Bear B. R. Spectrochim. Acta Part B 1986 41 1089. 36 Hinds W. and Reist P. C. AerosolSci. 1972,3 501 and 5 1 5 . 37 Sharp B. L. J Anal. At. Spectrom. 1988 3 939. 38 Clifford R. H. Ishii I. Montaser A. and Meyer G. A. Anal. Chem. 1990 62 390. 39 Belchamber R. M. and Horlick G. Spectrochim. Acta Part B 1981 36 581. 40 Belchamber R. M. and Horlick G. Spectrochim. Acta Part B 1982 37 1075. 41 Boumans P. W. J. M. and Lux-Steiner M. Ch. Spectrochim. Acta Part B 1982 37 97. 42 Schutyser P. and Janssens E. Spectrochim. Acta Part B 1984 39 737. 43 Schutyser P. and Janssens E. Spectrochim. Acta Part B 1979 34 443. 44 Orion Research Incorporated Sage Model 34IB Syringe Pump Instruction Manual Orion Research Boston 1987. 45 Scientific Systems Incorporated User’s Manual for SSI Model 2220 Digital HPLC Pump Scientific Systems State College 1990. 46 Poole C. F. and Schuette S. A. Contemporary Practice of Chromatography Elsevier Amsterdam 1984. 47 Varian Associates Inc. 2010 Pump/2210 System Operators Manual Varian Instrument Walnut Creek 1984. 48 Sprott J. C. Introduction to Modern Electronics Wiley New York 1981. 49 Jarvis K. E. Gray A. L. and Houk R. S. Handbook of Inductively Coupled Plasma Mass Spectrometry Blackie Glas- gow 1992. 50 Fry R. C. personal communication 1992. 51 Myers S. A. and Tracy D. H. Spectrochim. Acta Part B 1983 38 1227. Paper 2/00184E Received January 9 1992 Accepted July 6 1992