Influence of particle size on fractionation with nanosecond and femtosecond laser ablation in brass by online differential mobility analysis and inductively coupled plasma mass spectrometry

Abstract

A differential mobility analyzer (DMA) passes laser ablation (LA) particles and agglomerates within a narrow range of electrical mobilities to the inductively coupled plasma mass spectrometer (ICP-MS). No particle collection or offline particle analysis is required. Fractionation is assessed with the Cu⁺/Zn⁺ signal ratio. Results presented in this paper support previous findings that ns LA provides many small Zn-rich particles and some much larger Cu-rich particles and that fs LA produces large agglomerates of small particles. The composition of the aerosol produced by fs LA falls between the relatively Zn-rich and Cu-rich extremes of ns LA. Femtosecond LA provides elemental ratio measurements that remain more stable with respect to time, which allows a greater degree of confidence in LA results. Even though some parameters besides pulse length differ between the lasers compared in this study, the suppression of particle size related fractionation with fs LA can be attributed to pulse length.

---

Introduction

LA-ICP-MS is an appealing technique for the fast, sensitive, and spatially resolved trace elemental analysis of a wide variety of solids.^1–3 However, elemental fractionation remains a problem for the technique, especially for ns LA.^4–7 Elemental fractionation in LA has been used to describe the changes in elemental ratios as a function of various experimental parameters. Unfortunately, the same term has come to refer to different effects. One common definition of fractionation describes the time-dependent change of elemental ratios during a single LA run; this phenomenon is only briefly addressed in this work. The term fractionation can also refer to deviations between the overall determined elemental ratio and the actual stoichiometry of the sample, with no regard for time-dependent changes. This work primarily examines fractionation as the change in elemental ratio as a function of particle size.

Many causes of fractionation have been investigated in the different stages of LA-ICP-MS, and various methods have been explored to compensate for, reduce, or eliminate fractionation. Fractionation due to the LA process is affected by many laser parameters, including wavelength,^8–12 energy density,^12–15 and pulse duration,^11,16–21 which can affect the size distribution and composition of the aerosol. Fractionation has also been attributed to transport and ICP effects, which are also affected by particle size.^22–29 Smaller particles generally lead to less ICP-induced fractionation. Still, thermal effects of the LA process, such as melting, differential vaporization among elements, and re-ablation of deposited particles, can cause fractionation even if large particles are not transported effectively to the plasma or are filtered from the aerosol before reaching the ICP. Relative to ns lasers, fs lasers have been reported to reduce or eliminate these thermal effects, providing both a more nearly stoichiometric aerosol and smaller particles with a narrower size distribution.^5,16

To investigate particles resulting from LA, techniques such as optical particle sizing, particle impaction, and differential mobility analysis have been demonstrated.^30,31 Optical particle sizing can determine size distributions of LA aerosols, but it cannot transmit a selected range of particle size for direct, online analysis with ICP-MS. Particle impaction can be used to analyze and measure a wide range of particle and agglomerate sizes and shapes and can be used as a low pass size filter for online analysis. However, to analyze a particle size range defined by both a low and a high cutoff, the technique is limited to offline analysis, which is tedious and time consuming. Only differential mobility analysis offers the flexibility to directly select and examine particles and agglomerates in a specific electric mobility diameter range during online elemental analysis by ICP-MS.

In this work, particles of a range of sizes are produced during LA, and a range of particles and agglomerates centered around a desired electrical mobility diameter (d_m) is selected by a DMA and passed directly to the ICP-MS for elemental analysis. As noted elsewhere,³¹d_m is not equivalent to particle volume or mass since the structure of the particle or agglomerate contributes to its electrical mobility. Therefore, particles are inseparable from agglomerates at each d_m value of the DMA. Still, the DMA provides valuable insights into the effect of particle size on fractionation because it allows direct, online selection and variation of specific electrical mobility fractions of the total LA aerosol. While not the primary focus of this work, the time stability of elemental ratio measurements during rastering and single spot LA is also briefly investigated.

Because of their different volatilities, Cu and Zn in brass were chosen as the primary elements to examine particle size and laser effects on fractionation. Obtaining the “true” Cu⁺/Zn⁺ ratio is not the focus of this work. Since the experimentally determined Cu⁺/Zn⁺ ratio may be dependent on factors such as sampling position, ionization efficiency, differential atomization, and mass bias, this work is mainly concerned with changes or trends in the Cu⁺/Zn⁺ signal ratio with d_m. Any observed change in elemental signal ratio with respect to d_m is evidence of particle size related fractionation at some d_m. Otherwise, the same Cu⁺/Zn⁺ signal ratio would be obtained at all d_m values.

Experimental

Laser ablation

A CETAC Technologies LSX-500 LA system was used for ns LA. The laser is a Q-switched, beam-homogenized, frequency-quadrupled (266 nm) Nd:YAG laser. The sample is placed on a movable xyz-translation stage, and ablation can be visually monitored in real time with a charge-coupled device (CCD) and a zoom lens. A pulse width of <6 ns, repetition rate of 10 Hz, spot size of 100 µm, and energy specified to be ∼9 mJ per pulse were used.

A fiber chirped pulse amplification (FCPA) fs laser was used for fs LA (FCPA µJewel, IMRA America, Inc., Ann Arbor, MI, USA). This laser provides pulses of approximately 3.7 µJ and 370 fs at a wavelength of 1045 nm and a repetition rate of 100 kHz. The spot size was determined by the focus of the laser; after the experiments it was measured to be approximately 50 µm. The samples were placed in a home-made ablation cell, mounted on an xyz-translation stage, and viewed with a CCD camera.

Both lasers used cylindrical ablation cells with the gas entering the ablation region at the bottom and exiting at the top. For the ns laser, the standard ablation cell from CETAC was used: 25 mm high × 60 mm diameter, total volume 70 cm³, inlet 2 mm id. For the fs laser, a home-made, double-walled cell was used: 45 mm high × 50 mm diameter, total volume 90 cm³, inlet 4 mm id. The same Tygon tubing (4 mm id × ∼1 m long) was used to conduct the aerosol to the ICP in both cases. The same brass sample was ablated with the ns laser using either cell; the signal levels and noise behavior were similar. Thus, the cell geometry does not greatly affect the signal or noise behavior described below.

DMA

An electrostatic classifier using a DMA (Model 3080L, TSI Inc., Shoreview, MN, USA) selectively passed ablated particles and agglomerates centered around a specific d_m^30,32 to the ICP-MS device.³¹

ICP-MS

A Finnigan Element 1 ICP-MS device was used.^33–35 This device consists of a magnetic sector and electrostatic analyzer in a reverse Nier–Johnson arrangement. Low resolution mode (m/Δm = 300) was used throughout these experiments since higher resolution was not necessary for the isotopes observed.

Samples and conditions

NIST brass standards C1100 and C1102 were analyzed. Table 1 summarizes the ICP-MS operating conditions and the isotopes measured. Argon was used as a carrier gas, and sample gas flow rate was optimized in each experiment for maximum ICP-MS signal stability. This flow rate is approximately 0.15 L min^–1 lower than that which provides maximum mass transport from the ablation cell and sacrifices approximately one order of magnitude of signal. However, this lower sample gas flow rate leads to far more stable LA signal response from the ICP-MS device. Experiments were carried out in rastering and single spot modes on both samples. No sample pretreatment was performed.

Table 1 ICP-MS operating conditions

ICP-MS device	Finnigan Element 1
Outer gas	14 L min^–1
Auxiliary gas	0.70–0.75 L min^–1
Sample gas	1.05–1.10 L min^–1
Forward rf power	1200 W
Detector mode	Analog
Mass window	150%
Sampling time	0.01 s
Samples per peak	10
Isotopes measured	^63Cu, ^64Zn, ^65Cu, ^66Zn

---

For the rastering experiments, the samples were moved under the stationary laser beam at approximately 30 µm s^–1. Rastering times of 171 and 341 s (100 and 200 ICP-MS scans) were used for the ns and fs experiments, respectively.

For the ns single spot ablation experiments, a spot was continuously ablated at 10 Hz for approximately 120 s (70 ICP-MS scans through the mass range). For the fs single spot experiments, the laser very quickly drilled into the surface deep enough to be out of focus and below the ablation threshold. Therefore, limited signal was obtained for the fs single spot ablation experiments.

For each set of experiments (rastering and single spot ablation with both lasers on the samples, for a total of eight sets of experiments), data were acquired with the DMA for particles with a d_m ranging from 50 nm to 1000 nm, in 25-nm increments. For the fs experiments, data acquired at d_m = 100 nm and below were not included in the final plots because of low signal level. The sample gas flow was kept constant within each experiment so that changes in measured signals would only be attributable to d_m.

Owing to the high voltage applied to the electrodes in the DMA, arcing can be a problem when using Ar as a carrier gas. Under the conditions of these experiments, an abrupt signal change, likely caused by arcing, was sometimes observed at d_m > 1000 nm. In two experiments, single spot fs LA on NIST C1100 and rastering ns LA on NIST C1102, arcing may have been responsible for some artifacts at d_m < 1000 nm (data not shown in figures). The data trends prior to the possible arcing are consistent with those in the other experiments. Since the trends were consistent between samples and for brevity, only the figures describing the results from one standard are presented for each laser (ns, NIST C1100; fs, NIST C1102). All data, including those not presented in the figures, are included in the summary in Table 2.

Data analysis

Three types of plots were generated from the raw ICP-MS data. First, the data were background subtracted, and the peaks from each isotope were integrated for each ICP-MS scan. Integrated count rates from two isotopes each of Cu and Zn were summed and normalized to 100% of the respective element. Plots of normalized signal count rate versus d_m, elemental ratio versus d_m, and a 3-dimensional plot of elemental ratio versus d_mversus time were generated for each pair of elements and for each experiment. For the single spot ablation experiments, the signals from each spot were integrated for the duration of the experiment. No standard deviations are reported for the single spot ablation experiments because each reported count rate is the integrated signal from a single spot.

Results and discussion

System noise analysis from isotope ratios

To characterize any experimental or data analysis artifacts in the data plots, plots of ⁶³Cu⁺/⁶⁵Cu⁺ and ⁶⁴Zn⁺/⁶⁶Zn⁺ signal ratios were prepared for each experiment (Figs. 1 and 2). No upward or downward trends in these intraelemental ratios are observed with respect to d_m, with the possible exception of single spot ablation with the fs laser at very small d_m, which is attributable to the low signal level discussed previously. These data help confirm that any trends observed in the Cu⁺/Zn⁺ signal ratio with respect to d_m are due to particle size related elemental fractionation and not experimental or analysis artifacts.

Fig. 1 Isotope ratio plots for nanosecond LA of NIST C1100 brass: (a) rastering (1s error bars) and (b) single spot. For clarity, only the upper or lower halves of error bars are shown in (a).

Fig. 2 Isotope ratio plots for femtosecond LA of NIST C1102 brass: (a) rastering (1s error bars) and (b) single spot.

Nanosecond LA

Fig. 3 shows the Cu⁺/Zn⁺ signal ratio plots for rastering and single spot ns LA of NIST C1100 brass. Both rastering and single spot ablation show similar trends with respect to d_m. The Cu⁺/Zn⁺ signal ratio steadily increases with d_m at values between approximately 300 nm to 600 nm, which supports previous observations by Niemax et al.¹⁴ and Russo et al.²⁵

Fig. 3 Normalized elemental ratio plots for nanosecond LA of NIST C1100 brass: (a) rastering (1s error bars) and (b) single spot.

The signal plots shown in Fig. 4 show a maximum in signal for both Cu⁺ and Zn⁺, centered at d_m = 200 nm. The transfer function of the DMA, which is related to the width of the d_m transmission window, increases with d_m,^31,32 which would bias the signal toward large d_m. In addition, larger particles contain more analyte atoms. Since the plots show the largest signal count rate for smaller d_m values, from 100 nm to 300 nm, more particles and/or agglomerates are produced in this range than at larger d_m values. These findings are consistent with previous work.^4,14,25

Fig. 4 Signal plots for nanosecond LA of NIST C1100 brass: (a) rastering (1s error bars) and (b) single spot.

While the DMA cannot directly provide information about the nature of the particles being passed (i.e., whether they are single particles or agglomerates of smaller particles), some such information can be gleaned from the ICP-MS signal itself. There is a dramatic increase in the relative standard deviation of the ICP-MS signal with d_m during the rastering experiment (4% for d_m ≤ 200 nm to 18% for d_m ≥ 600 nm), beginning at d_m ∼ 400 nm (Fig. 4). This noise is likely caused by larger, single particles that are less completely vaporized, atomized, and ionized in the plasma, causing localized pockets of excess ions.³⁶ At smaller particle sizes, all the particles and agglomerates are more extensively vaporized, atomized, and ionized, leading to a much more stable and less noisy ICP-MS signal.²² It is still likely, though, that agglomerates of smaller particles exist at all d_m values included in this study.³⁷

Femtosecond LA

Fig. 5 shows plots from rastering and single spot fs LA of NIST C1102 brass. In these experiments, there was some variation of elemental ratios with d_m, but the aerosol had a more uniform Cu⁺/Zn⁺ signal ratio (a smaller difference between low d_m and high d_m) throughout the entire d_m range than aerosol from ns LA. Contrary to the trend observed in the ns LA experiments, the Cu⁺/Zn⁺ signal ratio decreased slightly with higher d_m values in the fs LA experiments.

Fig. 5 Normalized elemental ratio plots for femtosecond LA of NIST C1102 brass: (a) rastering (1s error bars) and (b) single spot.

The curves in Fig. 6 show that the majority of the ICP-MS signal comes from particles and agglomerates with d_m > 200 nm. Previous work^14,17 has shown that fs LA of brass produces an aerosol composed of agglomerates of very small particles. The current study supports this previous work: since most of the signal is found at higher d_m values, the aerosol is composed of either a distribution of large particles or large agglomerates of small particles. However, the small relative standard deviations observed in the signal and ratio plots imply that there is not a distribution of large particles causing excessive ICP-MS signal fluctuation, but rather a size distribution of agglomerates composed of similar-sized, small particles.^14,17

Fig. 6 Signal plots for femtosecond LA of NIST C1102 brass: (a) rastering (1s error bars) and (b) single spot. Note that error bars in (a) are smaller than the symbols.

As described elsewhere,^14,38,39 the particle re-condensation model describes a likely mechanism for production of these small particles. Since Zn is more volatile than Cu and would nucleate later in a saturated vapor, it is reasonable that the smallest condensed particles are relatively Zn-rich compared with the slightly larger condensed particles. Since smaller particles tend to agglomerate more readily than larger particles,¹⁷ it is expected that larger agglomerates would tend to be slightly Zn-rich, relative to smaller agglomerates.

Assuming the particle re-condensation model as the dominant particle formation mechanism, this qualitative explanation is consistent with the Cu⁺/Zn⁺ signal ratios shown in Fig. 5(a). Different Cu⁺/Zn⁺ signal ratios in particles of different sizes formed during fs LA are expected,¹⁴ but the overall Cu⁺/Zn⁺ signal ratio is still much more uniform throughout the d_m range in fs LA than in ns LA.

Plasma loading effects

Mass loading in the ICP has recently been reported to have a large effect on experimental ratios in the LA-ICP-MS of silicates under certain conditions.^7,11 However, the magnitude of this effect in LA analyses of brass has not been fully investigated. For silicate matrices, this mass loading effect causes the Cu⁺/Zn⁺ signal ratio to increase with increased mass loading; the signal from Zn, which is more difficult to ionize than Cu, would be more suppressed than that from Cu due to the cooling of the plasma by higher mass loading.¹¹ The matrix of the brass standards used in this study is composed of well over 99% Cu and Zn, but typical silicates have only a small fraction of these elements. It is unclear whether the mass loading effect observed during LA of silicates would be present when the sample is composed of almost 100% of the analytes of interest. It is prudent, however, to assume that there may be some mass loading effect on the Cu⁺/Zn⁺ signal ratio in LA of brass. If mass loading had a dominant effect on the Cu⁺/Zn⁺ signal ratios in these experiments, the Cu⁺/Zn⁺ signal ratios would trend in the same direction as the signal intensity with changes in d_m. This is not observed. Although mass loading effects may be present in these experiments, the observed effects are more consistent with particle size related fractionation due to differences in the LA aerosols.

Comparison of trends

Table 2 summarizes trends in the Cu⁺/Zn⁺ signal ratio (weighted by the Cu⁺ signal) and signal levels in different d_m regimes for the eight sets of experiments. Data are included for all samples and conditions, including those not shown in other plots. The weighting by Cu⁺ signal relates the data to contributions by each d_m size regime to the overall Cu⁺/Zn⁺ signal ratio that would be measured using these lasers and conditions without the DMA.

Table 2 Weighted average Cu⁺/Zn⁺ signal ratios and percentage of total signal from various size regimes (1s error given for rastering experiments)

		C1100				C1102
		Single spot		Rastering		Single spot		Rastering
	d m/nm	ns	fs^a	ns	fs	ns	fs	ns^a	fs
a Possible arcing was observed for dm > 600 nm.
Cu^+/Zn^+ signal ratio	<300	3.77	6.55	2.83 ± 0.13	6.54 ± 0.48	4.72	7.66	3.57 ± 0.20	8.68 ± 0.36
	300–575	4.07	6.41	5.15 ± 0.60	5.97 ± 0.60	5.52	7.75	6.29 ± 0.60	7.87 ± 0.25
	≥600	5.47	5.24	7.56 ± 1.4	5.80 ± 0.32	10.46	7.44	12.53 ± 1.9	7.25 ± 0.22
	All	4.21	6.17	5.25 ± 0.72	5.89 ± 0.42	6.38	7.58	9.24 ± 1.2	7.55 ± 0.24
Percentage of total signal	<300	41	14	37	5	39	8	18	5
	300–575	41	64	24	32	37	41	27	38
	≥600	19	22	39	63	24	51	55	58

---

Different trends in the Cu⁺/Zn⁺ signal ratio and signals with respect to d_m are evident between ns and fs LA. Nanosecond LA aerosols in all experiments show a substantial increase in the Cu⁺/Zn⁺ signal ratio in larger d_m regimes, while aerosols in the fs LA experiments show a gradual decrease in the Cu⁺/Zn⁺ signal ratio in larger d_m regimes. In the ns LA experiments, even the least pronounced ratio difference between the lowest and highest d_m regimes (single spot, NIST C1100) is larger than that in any of the fs LA experiments. In addition, the Cu⁺/Zn⁺ signal ratios for fs LA in all d_m regimes fall between the Zn-rich and Cu-rich extremes of ns LA. This is not immediately clear for the case of ns single spot LA on C1100; the rise of the Cu⁺/Zn⁺ signal ratio is found at higher d_m than in other experiments (Fig. 3), which skews the average in the d_m ≥ 600 nm regime to a somewhat lower value (maximum Cu⁺/Zn⁺ signal ratio, 7.03). Another trend observed is that the percentage of total signal obtained with ns LA was generally higher in lower d_m regimes, while the opposite was normally true for fs experiments.

These trends, which are independent of sample and sampling style, along with the previous discussion regarding the differences in relative standard deviations of the ICP-MS signals or ratios, reinforce the predominant view that ns LA produces many small particles (d_m < 300 nm) and some large, single, Cu-rich particles, while fs aerosols are predominantly agglomerates of small particles.

Since this work assesses particle size related fractionation through trends in the Cu⁺/Zn⁺ signal ratio with d_m, obtaining the “true” Cu⁺/Zn⁺ value is not required. While an in-depth study of the overall stoichiometry of the LA aerosols would obfuscate the purpose of this paper as a fast, online, qualitative assessment of particle size related fractionation in readily available ns and fs lasers, it is still of some interest to assess the stoichiometry of the LA aerosols. To that end, the ratio of the Cu⁺/Zn⁺ signal ratios in the two standards was calculated [Cu⁺/Zn⁺_(C1102)/Cu⁺/Zn⁺_(C1100)] for all of the experiments to assess this type of fractionation. In the absence of fractionation, the true ratio, 1.284, would be expected. Table 3 shows a summary of this ratio determined for three d_m regimes and for the overall aerosol. There is some deviation from the true value in the various d_m regimes for experiments with both lasers, but the largest particles from ns LA demonstrate the highest degree of fractionation. Femtosecond LA gives an overall more stoichiometric aerosol than ns LA for both single spot and rastering LA. This method of looking at the overall stoichiometry of LA aerosols is not intended to be a comprehensive analysis, but rather a rudimentary assessment, of overall fractionation.

Table 3 Ratio of the Cu⁺/Zn⁺ signal ratios for C1102/C1100 in various d_m regimes (calculated from certified values: 1.284)

	Nanosecond LA		Femtosecond LA
d m/nm	Single spot	Rastering^a	Single spot^a	Rastering
a Possible arcing was observed for dm > 600 nm.
<300	1.25	1.26	1.17	1.33
300–575	1.36	1.22	1.21	1.32
≥600	1.91	1.66	1.42	1.25
All	1.43	1.44	1.28	1.28

---

Time stability

Three-dimensional plots of elemental ratio versus d_mversus time were made to assess time dependence on elemental ratio measurements while using rastering or single spot LA. Fig. 7 shows that for both rastering and single spot ns LA, the Cu⁺/Zn⁺ signal ratio is stable at small d_m values but fluctuates wildly at large d_m values. Fig. 8 demonstrates that rastering with the fs laser provides more stable elemental ratio measurements with respect to time for all d_m compared with ns rastering LA. The spikes at small d_m values are largely due to counting statistics at the low signal level. Since there was a very short signal transient when using single spot ablation with the fs laser, this type of elemental ratio versus d_mversus time plot is not informative for those data.

Fig. 7 Time stability of nanosecond LA: (a) rastering and (b) single spot.

Fig. 8 Time stability of femtosecond rastering LA. Note that the d_m axis is reversed as compared with Fig. 7.

Conclusions

The fast, online measurement of elemental compositions of selected aerosol sizes in this study demonstrates a feasible qualitative complement to time-consuming offline studies involving other particle analysis techniques. The Cu⁺/Zn⁺ signal ratios obtained throughout the d_m range from both single spot and rastering fs LA fall between the ratios found from the relatively Zn-rich and Cu-rich extremes in the ns LA experiments. Aerosols from fs LA provide much more stable elemental ratios with respect to d_m and time, which allows greater confidence in fs LA results.

The two lasers were operated in their “out-of-the-box” state to compare ns and fs LA with readily available laser systems. The two lasers used in this study have obvious differences, including wavelength, repetition rate, and pulse duration, and these parameters could not be matched for the two systems. However, many of the parameters that differ between these two lasers, with the notable exception of pulse duration, would be expected to enhance particle size related fractionation for the fs laser, especially the longer wavelength and faster repetition rate. Since the fs laser suppressed particle size related fractionation, it is reasonable, based on previous research,^{5,14,16,17,25} to attribute this suppression largely to the much shorter pulse length.

Future studies will include more in-depth examination of the sub-100 nm regime for both rastering and single spot fs LA to gain further insight into the elemental composition of the particles throughout this particle size range. Other results using fast data acquisition are currently in preparation.⁴⁰

Acknowledgements

TSI Inc., Shoreview, MN, USA, provided the electrostatic classifier/DMA. IMRA America Inc., Ann Arbor, MI, USA, provided the FCPA µJewel fs laser. Ames Laboratory is operated for the US Department of Energy (USDOE) by Iowa State University of Science and Technology under Contract No. DE-AC02-07CH11358. This work was supported by the USDOE, Office of Defense Nuclear Nonproliferation, Office of Nonproliferation Research and Engineering, NA-22.

References

1. A. L. Gray, Analyst, 1985, 110, 551–556.
2. R. E. Russo, X. Mao and O. V. Borisov, Trends Anal. Chem., 1998, 17, 461–469.
3. S. F. Durrant, J. Anal. At. Spectrom., 1999, 14, 1385–1403.
4. H. R. Kuhn and D. Günther, Anal. Chem., 2003, 75, 747–753.
5. J. Koch, H. Lindner, A. von Bohlen, R. Hergenröder and K. Niemax, J. Anal. At. Spectrom., 2005, 20, 901–906.
6. M. Guillong and D. Günther, J. Anal. At. Spectrom., 2002, 17, 831–837.
7. I. Kroslakova and D. Günther, J. Anal. At. Spectrom., 2007, 22(1), 51–62.
8. R. E. Russo, X. L. Mao, O. V. Borisov and H. Liu, J. Anal. At. Spectrom., 2000, 9, 1115–1120.
9. M. Guillong, I. Horn and D. Günther, J. Anal. At. Spectrom., 2003, 18, 1224–1230.
10. J. Gonzáles, C. Liu, X. Mao and R. E. Russo, J. Anal. At. Spectrom., 2004, 19, 1165–1168.
11. J. Koch, M. Wälle, J. Pisonero and D. Günther, J. Anal. At. Spectrom., 2006, 21, 932–940.
12. T. Iizuka and T. Hirata, Geochem. J., 2004, 38, 229–241.
13. J. H. Yoo, O. V. Borisov, X. Mao and R. E. Russo, Anal. Chem., 2001, 73, 2288–2293.
14. J. Koch, A. von Bohlen, R. Hergenröder and K. Niemax, J. Anal. At. Spectrom., 2004, 19, 267–272.
15. S. H. Jeong, O. V. Borisov, J. H. Yoo, X. L. Mao and R. E. Russo, Anal. Chem., 1999, 71, 5123–5130.
16. R. E. Russo, X. Mao, J. J. González and S. S. Mao, J. Anal. At. Spectrom., 2002, 17, 1072–1075.
17. C. Liu, X. L. Mao, S. S. Mao, X. Zeng, R. Greif and R. E. Russo, Anal. Chem., 2004, 76, 379–383.
18. J. González, S. H. Dundas, C. Y. Liu, X. Mao and R. E. Russo, J. Anal. At. Spectrom., 2006, 8, 778–784.
19. Q. Bian, C. C. Garcia, J. Koch and K. Niemax, J. Anal. At. Spectrom., 2006, 21, 187–191.
20. J. Koch and D. Günther, Anal. Bioanal. Chem., 2007, 387, 149–153.
21. F. Poitrasson, X. Mao, S. S. Mao, R. Freydier and R. E. Russo, Anal. Chem., 2003, 75, 6184–6190.
22. D. B. Aeschliman, S. J. Bajic, D. P. Baldwin and R. S. Houk, J. Anal. At. Spectrom., 2003, 18, 1008–1014.
23. C. Y. Liu, X. L. Mao, J. González and R. E. Russo, J. Anal. At. Spectrom., 2005, 20, 200–203.
24. H. R. Kuhn and D. Günther, J. Anal. At. Spectrom., 2004, 19, 1158–1164.
25. C. Liu, X. Mao, S. S. Mao, R. Greif and R. E. Russo, Anal. Chem., 2005, 77, 6687–6691.
26. B. Hattendorf, C. Latkoczy and D. Günther, Anal. Chem., 2003, 75, 341A–347A.
27. H. R. Kuhn, M. Guillong and D. Günther, Anal. Bioanal. Chem., 2004, 378, 1069–1074.
28. L. Yang, R. E. Sturgeon and Z. Mester, Anal. Chem., 2005, 77, 2971–2977.
29. D. Bleiner, P. Lienemann and H. Vonmont, Talanta, 2005, 65, 1286–1294.
30. E. O. Knutson and K. T. Whitby, J. Aerosol Sci., 1975, 6, 443–451.
31. H. R. Kuhn, J. Koch, R. Hergenröder, K. Niemax, M. Kalberer and D. Günther, J. Anal. At. Spectrom., 2005, 20, 894–900.
32. C. Hagwood, Y. Sivathanu and G. Mulholland, Aerosol Sci. Technol., 1999, 30, 40–61.
33. R. S. Houk, V. A. Fassel, G. D. Flesch, H. J. Svec, A. L. Gray and C. E. Taylor, Anal. Chem., 1980, 52, 2283–2289.
34. L. Moens and N. Jakubowski, Anal. Chem., 1998, 70, 251A–256A.
35. R. S. Houk, ‘Elemental Speciation by ICP-MS with High Resolution Instruments’, in Handbook of Elemental Speciation, eds. R. Cornelis, J. Caruso, H. Crews and K. Heumann, John Wiley & Sons, 2002.
36. D. C. Perdian and R. S. Houk, unpublished work.
37. H. R. Kuhn and D. Günther, Anal. Bioanal. Chem., 2005, 383, 434–441.
38. B. S. Luk’yanchuk, W. Marine and S. I. Anisimov, Laser Phys., 1998, 8, 291–302.
39. R. Hergenröder, J. Anal. At. Spectrom., 2006, 21, 1016–1026.
40. D. C. Perdian, S. J. Bajic, D. P. Baldwin and R. S. Houk, J. Anal. At. Spectrom. , submitted for publication.

---