Comparison of ultrasonic and thermospray systems for high performance sample introduction to inductively coupled plasma atomic emission spectrometry

Abstract

This paper describes detailed work done in our lab to compare analytical figures of merit for pneumatic, ultrasonic and thermospray sample introduction (SI) systems with three different inductively coupled plasma-atomic emission spectrometry (ICP-AES) instruments. One instrument from Leeman Labs, Inc. has an air path echelle spectrometer and a 27 MHz ICP. For low dissolved solid samples with this instrument, we observed that the ultrasonic nebulizer (USN) and fused silica aperture thermospray (FSApT) both offered similar LOD improvements as compared to pneumatic nebulization (PN), 14 and 16 times, respectively. Average sensitivities compared to PN were better for the USN, by 58 times, compared to 39 times for the FSApT. For solutions containing high dissolved solids we observed that FSApT optimized at the same conditions as for low dissolved solids, whereas USN required changes in power and gas flows to maintain a stable discharge. These changes degraded the LODs for USN substantially as compared to those utilized for low dissolved solid solutions, limiting improvement compared to PN to an average factor of 4. In general, sensitivities for USN were degraded at these new conditions. When solutions with 3000 μg/g Ca were analyzed, LOD improvements were smaller for FSApT and USN, but FSApT showed an improvement over USN of 6.5 times. Sensitivities compared to solutions without high dissolved solids were degraded by 19% on average for FSApT, while those for USN were degraded by 26%. The SI systems were also tested with a Varian Instruments Liberty 220 having a vacuum path Czerny-Turner monochromator and a 40 MHz generator. The sensitivities with low dissolved solids solutions compared to PN were 20 times better for the USN and 39 times better for FSApT, and LODs for every element were better for FSApT. Better correlation between relative sensitivities and anticipated relative analyte mass fluxes for FSApT and USN was observed with the Varian instrument. LOD improvements averaged 18 times lower than PN with FSApT while with USN values averaged 8 times lower. When solutions with high dissolved solids were studied it was found that FSApT still offered 5.5 times better LODs than PN and USN offered 4.6 times better LODs than PN. Sensitivities for FSApT averaged 20 times better, while those for USN were 13 times better compared to PN. Finally, background RSDs on the Varian system were generally higher for FSApT than for the USN for similar sample types. A third instrument used for a small set of elements was a Perkin-Elmer model 5500 ICP-AES. This system has a 27 MHz generator with a N 2 purged Czerny-Turner monochromator. LOD trends, background RSDs, and sensitivities were similar to those with the Leeman instrument. However, matrix effects more closely resembled those seen with the Varian instrument for both SI systems. To compare performance and recoveries on a real sample, a National Institute of Standards and Technology, Standard Reference Material 1643c trace elements in water, was analyzed using the Varian system and it was found that both SI systems offered similar recoveries.