JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 807 Evaluation of an Ultrasonic Nebulizer for Sample Introduction in Inductively Coupled Plasma Atomic Emission Spectrometry* Theresa M. Castillano Nohora P. Vela and Joseph A. Carusot Department of Chemistry University of Cincinnati Cincinnati OH 45221 -01 72 USA W. Charles Story Environmental Health Research and Testing 3235 Omni Drive Cincinnati OH 45245 USA A low cost ultrasonic nebulizer has been designed using a forced air cooling method. The performance of this ultrasonic nebulizer in inductively coupled plasma atomic emission spectrometry has been evaluated and compared with a concentric nebulizer and a glass frit nebulizer. Higher signal-to-background ratios have been achieved for the ultrasonic nebulizer compared with the concentric and glass frit nebulizers.Detection limits and sensitivity have been improved by approximately one and two orders of magnitude respectively. In addition the response time of the ultrasonic nebulizer is comparable to those of the other types of nebulizers studied; a relative standard deviation of 3-5% for the ultrasonic nebulizer. Tests were also conducted to determine the effects of a high matrix sample such as synthetic ocean water. Keywords Inductively coupled plasma atomic emission spectrometry; ultrasonic nebulizer; matrix effect Several techniques are available for sample introduction into the inductively coupled plasma (ICP). Among these methods the most commonly used is the direct nebuliza- tion of the sample. Many different types of nebulizers can be used for this purpose including the concentric or cross flow nebulizers the glass-frit nebulizer and the ultrasonic nebulizer.Each type has its own advantages and disadvan- tages. Recent investigations have been conducted to mea- sure the size of the droplets formed by these it was found that the ultrasonic nebulizer generates a broader size distribution of the primary aerosol than the pneumatic nebulizer but mass transport is still much higher owing to greater particle density of the aerosol produced by using the ultrasonic nebulizer. The pneumatic nebulizer has been commonly used because of its simplicity and ease of operation. However transport efficiency is extremely low (about 3-8O/0)~ owing to the very wide range of droplet sizes produced. The glass- frit nebulizer however produces a finer aerosol compared with the pneumatic nebulizer but can suffer from clogging of the fritted glass disk over time.4 The ultrasonic nebulizer also produces a fine aerosol and has high sample transport efficiency which results in better analyte sensitivity and detection limits.In contrast to the glass-frit nebulizer very little clogging occurs with the ultrasonic nebulizer. Another advantage of the ultrasonic nebulizer is that a separate optimization of nebulizer parameters can be carried out since the efficiency of nebulization is independent of the carrier gas flow rate. However the present applications are limited to non-viscous solutions. Ultrasonic nebulizers are available commercially but the high cost of the commercial instrumentation has yet to be justified.Work has been done to develop low-cost ultra- sonic nebulizers5v6 and the performance when coupled with ICP atomic emission spectrometry (ICP-AES) has been e ~ a l u a t e d . ~ - ~ A common feature in the design of most ultrasonic nebulizers is the use of a water-cooled system positioned behind the transducer. However several prob- lems arise with this method of cooling such as corrosion and electrical shorting.1° In an attempt to make a simple reliable and relatively inexpensive ultrasonic nebulizer that can be used for routine analysis in a laboratory an air- cooled system has been designed. Other air-cooled ultra- sonic nebulizers have been described in the literature." An evaluation of the performance of this nebulizer with ICP- AES was carried out in terms of the signal-to-background ratios sensitivity response factor detection limits and response in the presence of a high-salt content matrix.Experimental Design The ultrasonic nebulizer developed in this laboratory is a modification of one designed by Fassel and Bear.12 The first design incorporated a heat sink that was directly attached to the rear of the transducer. The drawbacks of this prelimi- nary design were the instability of the signal and the relatively short lifetime of the transducer. This short lifetime was mainly due to the inefficient cooling of the transducer by the heat sink. Also because the heat sink was attached directly to the back of the transducer the oscillation of the transducer was restricted.In order to compensate for this restriction the transducer had to be operated at a higher power which resulted in a diminished lifetime. In order to improve on the first design a forced-air cooling of the transducer was employed. The current design of the ultrasonic nebulizer is illustrated in Fig. 1. The heat sink was eliminated and two O-rings were used to hold the Sample Air in I c Aerosol *Presented in part at the 1992 Winter Conference on Plasma fTo whom correspondence should be addressed. Spectrochemistry San Diego CA USA January 6-1 1 1992. Fig. 1 I 51' 1 -Carrier gas 1 I 1 Drain i Air out Modified ultrasonic nebulizer with air-cooling system808 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 Table 1 Operating conditions Forward power/kW Gas flow rate/dm3 min-I of Ar Outer Auxiliary Sample uptake rate/cm3 min-I Nebulizer (carrier) gas flow rate/dm3 min-' of Ar Pneumatic Glass frit Ultrasonic Heating chamber temperature/"C Condenser temperature/"C Aerosol chamber temperature/'C Desolvation conditions (for ultrasonic nebulizer) 1 16 1 0.5 0.35 0.20 0.60 160 0 0 transducer in place.The area behind the transducer was enclosed so that air could be circulated in the hollow section. An air flow of approximately 1 dm3 min-l was sufficient for adequate cooling. Instrumentation The plasma instrument used for these experiments was the PlasmaTherm 2500 (PlasmaTherm RTE 73 Kresson NJ USA). The ICP conditions are shown in Table 1. The emission was passed through two focusing lenses to the 1.26 m monochromator (Spex Spectrophotometer Metu- chen NJ USA) with 700 V applied to the photomultiplier tube for detection. Data collection was obtained with a Spex Datamate system.A schematic representation of the ultrasonic nebulizer set-up is shown in Fig. 2. This laboratory-built ultrasonic nebulizer makes use of a cooled spray chamber (University of Cincinnati) cooled to a temperature of about 0 "C. The sample is delivered by means of a peristaltic pump. The solution flows on the surface of the transducer through a capillary tube with a tapered tip that is positioned in such a way that the tip is almost in contact with the surface of the transducer. The transducer used is a Channel Products (Chesterland OH USA) Model CPMT which resonates at a frequency of 1.35 MHz.The dimensions of the transducer are 1.06 in (diameter) and 0.135 mm (thickness). A Wavetek (Indianapolis IN USA) sweep-function generator Model 180 was used as the signal source. The signal was fed into a power amplifier (University of Cincinnati Electronics Shop) which drives the transducer at the resonant fre- quency. A desolvation apparatus is required because of the high transport efficiency of the ultrasonic nebulizer. The heating chamber consists of a quartz tube (0.6 cm i.d. 35 cm long) wrapped with heating tape and powered by a Variac power supply (Fisher Scientific Pittsburgh PA USA). A 20 cm long condenser was used to remove the solute vapour producing a dry aerosol which was carried Dry aerosol to plasma Function t generator 1 Power amplifier 1 Sample Condenser Coolant - in Fig.2 Schematic diagram of the set-up of the ultrasonic nebulizer to the plasma by the nebulizer gas flow. Desolvation conditions are also shown in Table 1. Solutions A 100 ppm multi-element solution containing barium cobalt copper chromium iron magnesium nickel and antimony was prepared in 2% nitric acid from 1000 ppm standard solutions (Fischer Scientific NJ USA). The serial dilution method was used to prepare multi-element solu- tions in concentrations ranging from 10 ppm to 1 ppb. Blanks for each element and concentration were also prepared in an effort to minimize interferences. In order to study matrix effects multi-element solutions containing barium cobalt copper and iron in synthetic ocean water (SOW) were prepared in concentrations rang- ing from 100 ppm to 100 ppb.The composition of the SOW is as described in ref. 13. Results and Discussion Comparison of Different Nebulizers Using the 10 ppm solution for both the pneumatic and glass-frit nebulizers and the 1 ppm solution for the ultrasonic nebulizer signal-to-background ratios were com- pared and the results are shown in Table 2. The values listed for the ultrasonic nebulizer must be multiplied by ten to normalize for the difference in the concentrations. These results illustrate that the signal-to-background ratio for the ultrasonic nebulizer is 20-100 times better than that of the pneumatic and glass-frit nebulizers. A comparison of the sensitivity of the nebulizers studied for each of the elements listed is shown in Table 3.The data ~~ Table 2 Comparison of signal-to-background ratios using a 10 ppm multi-element solution Signal-to-background ratio Element and line Wavelengthhm Pneumatic nebulizer Ba I1 c o I1 Cr I1 c u I1 Fe I1 Mg I Ni I1 Sb I 233.527 236.379 205.552 324.7 54 238.204 285.21 3 22 1.647 206.838 14.5 2.39 1.81 3.48 6.3 1 4.62 0.50 12.8 Glass-frit nebulizer 9.26 1.81 1.29 1.87 2.57 5.94 2.7 1 0.29 Ultrasonic nebulizer* 45.4 15.0 8.98 7.24 34.6 62.0 18.9 1.67 *Values were obtained using a 1 ppm multi-element solution.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 809 Table 3 Comparison of sensitivity Sensitivity (arbitrary units per ppm) Element and line Ba I1 c o I1 Cr I1 c u I Fe I1 Mg I Ni I1 Sb I Pneumatic nebulizer 3.77 x 10-3 9.31 x 10-4 2.25 x 10-4 6 .1 5 ~ 10-3 2.93 x 10-3 1 . 7 7 ~ 9 . 1 4 ~ - Glass-frit nebulizer Ultrasonic nebulizer 2.37 x 10-3 3.23 x 10-4 4 . 5 0 ~ 1 . 3 4 ~ 10-4 9.84 x 10-3 2.35 x 10-3 3 . 1 6 ~ 10-1 1.95 x 10-4 1.25 x lo-' 6 . 1 5 ~ 8 . 1 9 ~ 10-1 4.27 x 10-4 2.85 x lo-* 1.38 x 10-3 1.25 x 10-1 - r2 (for ultrasonic nebulizer) 0.9988 0.9998 0.9994 0.9996 0.9972 1 .oooo 0.9990 0.9998 Table 4 Comparison of detection limits Detection limit* (ppb) Element and line Wavelengthhm Pneumatic nebulizer Glass-frit nebulizer Ultrasonic nebulizer Ba I1 c o I1 Cr I1 c u I Fe I1 Mg I Ni I1 Sb I 233.527 236.379 205.552 324.754 238.204 285.2 13 221.647 206.838 30 135 1150 70 41 1 1 37 - 36 176 246 96 3 54 19 141 - 6 7 4 1 1 2 2 3 200 *Detection limits were calculated using three times the standard deviation divided by the slope of the calibration graph.show an increase in sensitivity by approximately two orders of magnitude using the ultrasonic nebulizer. No value is listed for the sensitivity of the pneumatic and glass-frit nebulizers for antimony because a signal was only seen for the 100 and 10 ppm solutions. Values for r2 (r=regression coefficient) are also shown in Table 3 which indicate that the calibration graphs are linear over at least three orders of magnitude. The detection limits for the three different nebulizers are given in Table 4. As is evident the ultrasonic nebulizer gives detection limits that are approximately 5-20 times better (excluding chromium) than the other two nebulizers studied. The detection limit for chromium however improved by a factor of 200 which is much higher than was reported in previous studies.*l The reason for this is still under investigation.The Mg I 285.2 13 nm line was used to study the response times of the different nebulizers and a 100 ppm solution was used. Once a constant signal was obtained data were collected for 1 min the solution was changed to a 2% nitric acid blank for about 2 min and then back again to the 100 ppm solution within a total time of 5 min. The same procedure was followed for the glass-frit and ultrasonic nebulizers except that a 10 ppm multi-element solution was used for the ultrasonic nebulizer. Representative plots are shown in Fig. 3 for percentage signal intensity versus time. Using the pneumatic nebulizer [Fig. 3(a)] a negative peak was observed before the decrease in the signal indicating the point where the air bubbles formed when the change in solutions reached the nebulizer.However when using the glass-frit nebulizer [Fig. 3(b)] this resulted in a positive peak. There was no indication of this point when using the ultrasonic nebulizer [Fig. 3(c)] as the volume of the air bubble is negligible and might have spread over the total volume of the system. Based on the slopes of the response curves the glass-frit nebulizer showed a slightly longer response time due to the lower nebulizer flow rate used. Although a higher carrier gas flow rate was used for the ultrasonic nebulizer results obtained show a similar response time to that of the pneumatic nebulizer. This could be associated with the desolvation step incorporated when using the ultrasonic nebulizer.Stability tests performed on each of the nebulizers by alternating the 10 ppm mutli-element solution and the blank solution every 5 min for 1 h are illustrated in Fig. 4. The relative standard deviation (RSD) ranges from approx- imately 1-3% for the pneumatic and glass-frit nebulizers [Fig. 4(a) and (b)]. The ultrasonic nebulizer [Fig. 4(c)] loo l-7 I I I loo l-7 50 t 0 1 2 3 4 5 Time/min Fig. 3 nebulizer and (c) ultrasonic nebulizer Response times using (a) pneumatic nebulizer (h) glass-frit810 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 Table 5 Effect of matrix on signal intensity Suppression or Detection limit (ppb) enhancement Element and line Wavelengthhm factor 2% HN03 sow Ba I1 c o I1 c u I Fe I1 233.527 236.379 324.754 2 3 8.204 0.0 1 0.67 I .30 0.68 3 636 14 14 I08 70 23 17 0.12 0.08 0.04 0 7 ( a ) I 1 I I I I 0 10 i 20 30 40 50 Time/mi n 60 Fig.4 Stability test for the (a) pneumatic nebulizer (b) glass-frit nebulizer and ( c ) ultrasonic nebulizer using a 10 ppm multi- element solution and monitoring the Mg 285.2 13 nm line Effect of SOW Matrix on the Performance of the USN The effect of matrix on the analytical signal can be expressed in terms of a suppression or enhancement factor. This factor is the ratio of the analyte signal in an SOW matrix to that in 2% nitric acid solution.13 The results obtained using a 10 ppm multi-element solution containing the four elements in SOW at a sample uptake rate of 0.13 cm3 min-’ are shown in Table 5.Signal suppression occurred for all elements except copper which exhibited an enhancement in the analytical signal. From Table 5 the detection limit for copper is slightly improved while that for barium is severely degraded. This is probably due to the extremely low solubility of barium sulfate especially in the presence of sulfuric acid. The insoluble barium sulfate might have been lost in the sample container or in the desolvation chamber. The other two elements cobalt and iron show no significant effect from the presence of the SOW matrix. The slopes of the calibration graphs generated for copper in SOW and in 2% nitric acid are 0.020 and 0.019 respectively using a sample uptake rate of 0.25 cm3 mind’. This slight increase in sensitivity is consistent with copper having a signal enhancement in an SOW matrix.However the linear range of copper decreased by an order of magnitude in SOW. The same procedure for the stability test was used to determine if the presence of the matrix in solution affects the stability of the signal generated by the ultrasonic nebulizer. The 10 ppm multi-element solution in SOW was used as the analyte solution while the blank consisted of SOW only. The Cu I 324.745 nm line was monitored and the results are given in Fig. 5. The RSD for the six runs ranged from 3 to 8% which shows that the effects due to the presence of the SOW are minimal. Conclusions The simplicity of the design and good performance of the ultrasonic nebulizer makes it a suitable device for sample introduction into the ICP.In addition the laboratory-built yields a slightly higher RSD 2-6%. This is thought to be due to the uneven introduction of the sample onto the transducer. Future work will concentrate on a more efficient means of delivering the sample to the transducer. 0 10 20 30 40 50 60 Timelmin Fig. 5 Stability test for the ultrasonic nebulizer using the I0 ppm Cu solution in an SOW matrixJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1992 VOL. 7 81 1 ultrasonic nebulizer is a very cost effective alternative to commercial ultrasonic nebulizers available at the present time. The total cost of the design is approximately $1000 utilizing the materials available in this laboratory; mainly the desolvation apparatus. The values for analytical figures of merit such as the detection limit and RSD (*/o) in this study are not as low as expected owing to instrument limitations.However the degree of improvement of these values using the ultrasonic nebulizer over the other types of nebulizers is of the same order of magnitude as those previously reported.ll The authors acknowledge the National Institute of Environ- mental Health Science for research grants ES-03221 and ES-04908 and J. Carey and L. Olson for their input and assistance. References Tarr M. A. Zhu G. and Browner R. F. Appl. Spectrosc. 1991 45 1424. Clifford R. H. Ishii I. Meyer G. A. and Montaser A. Anal. Chem. 1990,62 390. Routh M. W. Spectrochim. Acta Part B 1986 41 39. 4 Brotherton T. and Caruso J. A. J. Anal. At. Spectrom. 1987 2 695. 5 Clifford R. H. and Montaser A. Anal. Chem. 1990 62 2745. 6 Jin Q. Zhu C. Brushwyler K. and Hieftje G. M. Appl. Spectrosc. 1990 44 183. 7 Goulden P. D. and Anthony D. H. J. Anal. Chem. 1984,56 2327. 8 Olson K. W. Haas W. J. J. and Fassel V. A. Anal. Chem. 1977 49 632. 9 Taylor C. E. and Floyd T. L. Appl. Spectrosc. 1981,35,408. 10 Denton M. B. Freelin J. M. and Smith T. R. in Sample Introduction in At om ic Spectroscopy A nalyt ical Spectroscopy Library ed. Sneddon J. Elsevier New York 1990 vol. 4 ch. 4. 1 1 Chan S. Zechmann C. A. and Yanak M. M. ICPInf Newsl. 1991 16 436. 12 Fassel V. A. and Bear B. R. Spectrochim. Acta. Part B 1986 41 1089. 13 Wang J. Shen W. L. Sheppard B. S. Evans. E. H. and Caruso J. A. J. Anal. At. Spectrom. 1990 5 445. Paper 2/0 I092E Received March 2 1992 Accepted June 18 1992