Introduction

The use of mixed gas plasmas in inductively coupled plasma atomic emission spectrometry (ICP−AES) and mass spectrometry (ICP−MS) is well known to offer signal enhancement [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16]. Mixed gas plasmas minimize the mass spectral interferences resulting from the production of oxides and polyatomic ions and the matrix effect besides improving sensitivity [2, 3]. Generally, N2 [4,5,6,7,8,9,10,11,12], O2 [5, 6], and H2 [8, 13] are prevalent additive gases, and some of them have been applied to laser ablation ICP−MS (LA−ICP−MS) [14, 15] and single-particle ICP−MS (sp ICP−MS) [16]. Although it offers an effective way to improve analytical sensitivity, it applied to the analysis of only particular elements, and not all signal intensities are enhanced. For instance, N2 gas addition to the outer Ar gas in the ICP enhances the signal intensities of Y, Zr, and As but has no effect on Sr [5].

Fundamental studies about the mixed gas effect have been conducted via pneumatic nebulizer (i.e., concentric nebulizer; CN). Recent results obtained using ultrasonic nebulizer (USN) should be noted [12]. Signal enhancement through aerosol desolvating the aerosol produced by USN is a common technique in ICP−MS. In addition, we observed that the combination of USN and N2 gas (2.2%) addition to the nebulizer gas enhanced the signal intensity of Sr 3.7 times compared with the use of USN alone [12]. Although Sr was used as an inert element toward the mixed gas effect in previous studies [5], our report [12] suggests that the combination of nebulization and the use of mixed gas plasma lead to a signal enhancement effect exceeding the individual impacts of the two approaches.

In this paper, we comprehensively evaluated the analytical figures of merit for 63 elements, such as the limit of detection (LOD), the slope of the calibration curve, and the background noise (BGN) intensity, when N2 gas (0–5%) addition to the nebulizer gas was combined with CN or USN. In addition, we examined the mechanism of the mixed gas effect on USN using the three-dimensional (3D) intensity distributions in the ICP provided by shifting an ICP torch box when 1% N2 and USN were used. Our study provides new insights into the role of nebulization (CN or USN) and the mixed gas effect in enhancing the sensitivity of the elemental analysis.

Experimental

Reagents

One gram per liter (or 100 mg L−1) of single-element standard solution for each of the 62 elements was used except for U. The standard solutions were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan), except Ba, Ru, Hf, Re and Ir. The solution for Ba was obtained from Kanto Chemical Co., Inc. (Tokyo, Japan), and the others were obtained from, AccuStandard, Inc. (New Haven, CT, USA). The solution for U was prepared from a multi-element mixture of 100 mg L−1 B, Th, and U (2% HNO3, PerkinElmer, Inc., Waltham, MA, USA) as the single-element standard solution of U was not available in Japan. High-purity, 15.1 mol L−1 HNO3 (68%, 1.4 g mL−1), was purchased from Tama Chemicals Co., Ltd. (Kanagawa, Japan). The ultrapure water (with a resistivity of 18.2MΩ cm) was obtained from the PURELAB Ultra water purification system (Organo Corporation, Tokyo, Japan).

Instrumentation

A single quadrupole ICP−MS (NexION 300S, PerkinElmer) combined with a U5000AT+ USN (Teledyne CETAC Technologies, Omaha, NE, USA) or quartz baffled cyclonic spray chamber with a micro-mist concentric nebulizer was used. A TruFlo sample monitor (Glass Expansion, Melbourne, Australia) was used to monitor the sample flow rate. The parameters used in the operation of ICP−MS are provided in Table S1 in Supplementary Information (SI). The 59 elements (7Li, 9Be, 11B, 23Na, 24 Mg, 27Al, 45Sc, 47Ti, 51 V, 55Mn, 59Co, 60Ni, 63Cu, 66Zn, 69 Ga, 74Ge, 75As, 85Rb, 88Sr, 89Y, 90Zr, 93Nb, 98Mo, 102Ru, 103Rh, 106Pd, 107Ag, 111Cd, 115In, 118Sn, 121Sb, 130Te, 133Cs, 138Ba, 139La, 140Ce, 141Pr, 142Nd, 152Sm, 153Eu, 158Gd, 159 Tb, 164Dy, 165Ho, 166Er, 169Tm, 180Hf, 181Ta, 184 W, 187Re, 192Os, 193Ir, 195Pt, 197Au, 202Hg, 205Tl, 208Pb, 209Bi, and 238U as monitored mass number) were measured without the use of a dynamic reaction cell (DRC) technique (i.e., STD mode). Four elements (39 K, 40Ca, 52Cr, and 56Fe as monitored isotopes) were measured via the DRC mode with 1 mL min−1 of ammonia gas to remove interferences. A nebulizer gas flow rate not exceeding 3% of the oxide ratio to singly charged ions (140Ce16O+/140Ce+) was obtained. It was 0.98 L min−1 in the only use of pure Ar and was 0.90 L min−1 of Ar in the case of Ar−N2 mixed gas.

The internal diameters of quartz tubes were 17.95 mm, 13.95 mm, and 2.00 mm for the outer, middle, and injector tubes, respectively. The abbreviation “d” is the distance between the tip of the ICP torch and the sampling cone, which was 5.5 mm at the initial position. The diameter of the orifice was 1.1 mm. N2 gas (0–50 mL min−1) was mixed with 0.80–1.20 L min−1 of the nebulizer gas through a Y-shaped connector in the forestage of the nebulization device. 0.1 or 1 μg L−1 of a single-element standard solution containing 62 elements (except U), and a mixture containing U were used. The HNO3 concentration was adjusted to 0.2 mol L−1.

Evaluation and calculations

The effect of two different types of nebulization (CN or USN) on the Ar−N2 mixed plasma was examined using three evaluation criteria, sensitivity enhancement ratio (SER), BGN, and LOD. The concentrations of 0.1 and 1 μg L−1 standard solutions (dissolved in 0.2 mol L−1 HNO3) and blank solutions (0.2 mol L−1 HNO3 solution) were measured via CN and USN, respectively; then, their sensitivities (i.e., cps/(μg L−1)) were calculated the net intensity after the subtraction of the BGN (cps) was divided by the concentration (μg L−1) of the standard solution. To avoid spectral interference, single-element standard solutions were used in this study, while U was used as a mixed solution due to legal regulation. In the acquisition of signal intensity, the average intensity was calculated from 25 measurements. The SERs were calculated as the ratio of the sensitivity in the presence or absence of N2 by the following equation:

$$SER\,\,\, = \,\,\,\frac{{Sensitivity_{{Ar - {\text{N}}_{2} }} }}{{Sensitivity_{Ar} }}$$
(1)

where SensitivityAr (cps/(μg L−1)) and \(\text{Sensitivity}_{Ar-N_{2}}\) (cps/(μg L−1)) represent the sensitivity of a pure Ar plasma and that in the presence of N2 in Ar gas, respectively.

The 3-sigma method based on the Gaussian distribution model is commonly used for the calculation of LODs; in the case of very low signal intensities, the Poisson distribution is suitable to determine the LOD [17]. In this study, LOD values were calculated using the upper limit (Lc) of the 95% one-sided confidence interval based on the Poisson distribution model [17].

$$LOD\,\,\, = \,\,\,2L_{c} \,\,\, \times \,\,\,Sensitivity$$
(2)

where Sensitivity (cps/(μg L−1)) represents the slope of the calibration curve. The BGN of the 95% confidence interval was calculated using free software R [18].

3D intensity distribution in the plasma

The 3D intensity distribution of target ions in the plasma was obtained by changing the distance between the ICP torch and the sampling cone. 9Be, 88Sr, and 208Pb were selected as the typical two model cases in the activated elements (9Be and 88Sr) and the inert elements (208Pb), and 0.1 μg L−1 of the single-element standard solution (Be, Sr, and Pb) in 0.2 mol L−1 HNO3 was measured via USN. The nebulizer gas flow rate was 1.04 L min−1 in the absence of N2. In contrast, the nebulizer gas flow rate was 1.03 L min−1 in the presence of 1% N2 (10 mL min−1). An interval distance (Z-axis) between the torch and the sampling cone (i.e., sampling depth) and an alignment (X- and Y-axis) were controlled by software for ICP−MS. The sampling depth was defined as an initial point at 5.5 mm, and distance variation was adjusted at 1.5 mm intervals in the region from 2.5 mm to 8.5 mm. The lateral signal intensity was measured in 0.6 mm increments between − 2.4 mm and + 2.4 mm to the position of the orifice.

In addition, the transition of intensity distribution by adding N2 into the nebulizer gas was investigated. 0.1 μg L−1 of the single-element standard solution (Be, Sr, Ce, and Pb) in 0.2 mol L−1 HNO3 was measured. The transition degree (T%) was calculated by the following equation:

$$T\% \,\,\, = \,\,\,\left( {\frac{{I_{{x\left( {Ar - N_{2} } \right)}} }}{{I_{{max\left( {Ar - N_{2} } \right)}} }} - \frac{{I_{{x\left( {Ar} \right)}} }}{{I_{{{\text{max}}\left( {Ar} \right)}} }}} \right)\,\,\,\, \times \,\,\,100$$
(3)

where \(I_{\text{x(Ar-N}_\text{{2}})}\) and \(I_{\text{max(Ar-N}_\text{{2}})}\) represent the signal intensity at the point and the maximum intensity at the depth in the presence of 1% N2, respectively. Ix(Ar) and Imax(Ar) represent the signal intensity at the point and the maximum intensity at the depth in the absence of N2, respectively.

Results and discussion

Effect of nebulization (CN or USN) on SER in Ar − N2 mixed plasma

Table 1 shows the sensitivity (in pure Ar plasma and Ar−N2 mixed plasma) and SER values of the 63 elements measured via CN or USN. In 53 elements, the sensitivity was improved by the Ar−N2 mixture via CN (SER > 1). In contrast, the sensitivity for 35 elements was improved by the Ar−N2 mixture via USN. The results are summarized in (Fig. 1), and each circle means the respective elements. The circle size and the color represent the mass number of each element. The error bars show the propagation error of SER calculated from the standard deviation of signal intensity in absence of N2 and presence of 1% N2. The circles above the equivalent line (y = x) show higher SER via USN than CN, and below show higher SER via CN than USN. The circles on the line indicate that there is no difference in SER measured via USN and CN.

Table 1 The comparison of the sensitivity enhancement ratio (SER)
Full size table
Fig. 1
figure 1

Relationship between sensitivity enhancement ratios (SERs) of USN and CN. The size and color of circles indicate m/z

Full size image

Among 18 of the improved 35 elements in USN (SER > 1), the values are higher than those in CN (7Li, 9Be, 11B, 23Na, 47Ti, 51 V, 55Mn, 59Co, 60Ni, 63Cu, 66Zn, 69 Ga, 74Ge, 75As, 85Rb, 88Sr, 89Y, and 103Rh). The gray-colored area having both SER values less than 1 for CN and USN represents the inert zone toward the addition effect of N2 into Ar plasma (8 elements: 56Fe, 111Cd, 130Te, 192Os, 193Ir, 195Pt, 197Au, and 202Hg). 106Pd had a SER > 1; nevertheless, the values were nearly 1. Many of them were platinum group elements, Cd and Hg. The solid linear line (y = x ± 5%) means that the equivalent SER values are obtained in the case of CN and USN. In other words, the elements on this line are those that do not dominance between CN and USN in the sensitivity enhancement obtained by the Ar−N2 effect (6 elements: 45Sc, 90Zr, 93Nb, 102Ru, 107Ag, and 118Sn). In addition, the circle size and its color indicate an intense tendency that use of USN is effective for elements lower than m/z 100 except 103Rh.

Figure 2 shows the impact of the concentration of N2 gas (0–5%) on the ratio of the resultant intensity (with N2) to original intensity (without N2). The ratio was calculated from each of the intensities obtained via CN and USN. The value 1 was defined as the same intensity as that without N2. N2 flow rate within 1% of nebulizer gas flow rate was optimum for most elements. The dependency of additive gas flow rate on sensitivity enhancement effect seems to have occurred because of shifting ionization region in the ICP, therefore, we investigated about this inference by plotting a 3D intensity distribution in the next chapter.

Fig. 2
figure 2

Impact of the concentration of N2 gas on the ratio of resultant intensity (with N2) to original intensity (without N2). Nebulizer gas flow rate not exceeding 3% of CeO/Ce was optimized. The intensities were obtained via concentric (CN, black bar) and ultrasonic nebulizers (USN, white bar). Sample: 0.1 μg L − 1 in 0.2 mo L − 1 HNO3

Full size image

3D relative intensity distribution in the ICP

The signal intensity distribution measured by changing the positional relationship of the ICP torch to the orifice represents the actual distribution of target ions in the ICP [19]. Namely, the place showing the maximum signal intensity means that the target analyte is the most ionized at the position. The 3D distribution of the relative intensity based on the background intensity is shown in (Fig. 3). It was observed by changing the ICP position axially and laterally to the orifice. 0.1 μg L−1 single-element standard solution of three elements (9Ba, 88Sr, and 208Pb) were measured via CN in (a) the absence of N2 and (b) the presence of 1% N2; furthermore, they were also measured via USN in (c) the absence of N2 and (d) the presence of 1% N2. 9Be and 88Sr had SERs activated by N2 in USN. In contrast, 208Pb was selected as a SER-inert element in N2 addition via USN. Moreover, the elements (typically, 9Be and 88Sr) with m/z lower than 100 showed little response to the addition of N2 at any distance via CN, the element above m/z 100 (typically 208Pb) was slightly activated at the distance via CN. Relatively, the elements with m/z lower than 100 were significantly activated via USN; in contrast, Pb was not activated via USN. While the distance between the torch and the sampling cone increased, the high intensity was even maintained in the case of USN. These 3D distributions indicate that N2 addition assisted the extension and stabilization of the ionized regions. Mixtures of Ar and N2 have a high thermal conductivity at temperatures between 5000 and 10,000 K [20]; therefore, the Ar−N2 mixture contributes to LOD improvement via signal enhancement.

Fig. 3
figure 3

3D distribution of the relative ratio of the intensity to the background intensity in the a absence of N2 via CN or b presence of 1% N2 via CN; c absence of N2 via USN or d presence of 1% N2 via USN. Each intensity was measured at different sampling depths (2.5–8.5 mm)

Full size image

Figure 4 shows the transition (T%) of the signal intensity via USN from the absence of N2 to the presence of N2 for four elements (9Be, 88Sr, 140Ce, and 208Pb). The red color shows that the signal intensity at a given point increased in the presence of 1% N2 compared to that in the absence of N2. The blue color shows that the signal intensity at a given point decreased in the presence of 1% N2 compared to that in the absence of N2. For the elements with m/z lower than 100, such as 9Be and 88Sr, the maximum intensities were expanded from the torch side (short distance of “d”) to the sampling cone side (long distance of “d”). In contrast, the positions of maximum intensities were shifted to the plasma side for the element with m/z larger than 100, such as 140Ce and 208Pb. Since an ordinary distance of “d” is 5.5 mm for detection, it seems difficult to find the optimum point of the maximum intensity resulting from the shift, which is caused by N2 addition, in the measurement of elements with m/z above 100. N2 effects via USN for elements with m/z above 100 might require special configurations for the position adjustment between the torch and sampling cone.

Fig. 4
figure 4

Transition (T%) of the position of the maximum intensity by adding N2 via USN. The standard detection face was the 5.5 mm distance between the plasma and the torch

Full size image

Impact of Ar−N2 mixed plasma on background noise intensity (BGN)

Table S2 in SI shows the impact of Ar−N2 mixed plasma on BGN (in the measurement of 0.2 mol L−1 HNO3 as a blank solution), and the interference ions reported in the literature [21,22,23,24] were also listed. In the presence of N2 in Ar plasma, both detrimental (i.e., increase of BGN) and beneficial (i.e., decrease of BGN) effects were confirmed. The detrimental effect is occurred due to the generation of polyatomic ions (containing N atoms) based on the charge transfer reaction between N2 and Ar [25]. For instance, the intensity of m/z 55 (position for 55Mn) was significantly increased by the generation of 40Ar15N+ and 40Ar14N1H+. The detrimental effects such as spectral interferences were observed on m/z 27, 45, 47, 51, 52, and 56 (position for 27Al, 45Sc, 47Ti, 51 V, 52Cr, and 56Fe, respectively). In contrast, the beneficial effect can be seen in the intensities of m/z 39, 40, and 66 (position for 39 K, 40Ca, and 66Zn, respectively) as the decrease of BGN. Typically, the interference ions toward 39 K, 40Ca, and 66Zn appear on those m/z positions in pure Ar plasma due to the generation of polyatomic ions containing Ar or O atoms. Specific examples of interfering ion are below: 38ArH+ on 39 K+, 40Ar+ on 40Ca+, and 34S16O2+ on 66Zn+ as shown in Table 2S. The Ar volume is relatively reduced by the N2 volume increase in the plasma; thus, the resultant the generation of Ar-derived interference ions decreased.

Moreover, Fig. 5A shows the BGN suppression efficiency in the case of CN or USN in the absence or presence of N2. Each circle means respective elements. In both USN and CN, the BGN in the presence of N2 was lower than that in the absence of N2 indicating that N2 addition suppresses BGN for both CN and USN. Figure 5B shows the relationship between m/z and the mean square of the improvement degree in BGN (√χ2). χ represents the root-square value that the difference value of BGNs obtained via CN and USN in Ar−N2 is divided by BGN of CN in Ar−N2. From the results, the suppression effect of USN (via Ar−N2) works for m/z lower than 120; however, it became relatively worse on higher m/z values. Although the suppression effect (by N2 addition) does not discriminate between CN and USN based on the result of (Fig. 5A), it contributes mainly to m/z less than 120.

Fig. 5
figure 5

The relationship of BGN in between absence of N2 and presence of 1% N2 via USN (red circle) and CN (green circle) (A) and the m/z dependence on BGN (B). The error bars in A show the standard deviation of BGN

Full size image

Limit of detection

Table 2 shows the LODs calculated from the upper limit of the 95% one-sided confidence interval of the observed background intensity and sensitivity. The LODs were improved 3.3-fold (median) by combination between Ar−N2 and USN as compared with the general method (i.e., the combination of pure Ar and CN). While LOD did not improve regarding 9 elements (Al, Mn, Ga, Ge, Te, Ce, Tm, Os, and Bi), the present method (i.e., the combination of Ar−N2 and USN) is of benefit to other elements. Figure 6 exhibits the LOD improvement factors on each m/z. The values were calculated by the ratio of the LODs in USN and CN via the presence or absence of N2. The LOD improvement depended on the character of the element. For some elements, the LOD was lower in the presence of N2, while for others, the LOD was lower in the absence of N2. Regardless of the nebulization method used, the effect of N2 addition was ineffective for the m/z range of 139–169 (lanthanides). The m/z 7(Li), 9(Be), 47(Ti), 75 to 89(As to Y), 102(Ru), and 103(Rh) showed a tendency for decreased LOD by the addition of N2 in USN. This is attributed to the effect of N2 in USN (i.e., increase in SER) and the low SD of the BGN. For the elements with m/z above 180, the SER improvement by N2 was observed in CN; thus, it would greatly improve the LOD of CN.

Table 2 The obtained LOD with or without N2 via CN or USN
Full size table
Fig. 6
figure 6

The LOD improvement factor. The blue and orange bars show the ratio of LODs between USN and CN via the absence or presence of 1% N2, respectively

Full size image

Conclusion

The improvement in analytical performance by combining Ar−N2 gas addition into the nebulizer with different nebulization methods (i.e., CN and USN) in ICP−MS for 63 elements was reported. We showed the effect of the N2 gas concentration range 0–5% in the nebulizer gas on the analytical performance for 63 elements via commercially available nebulization methods (CN and USN). The SERs, BGNs, and LODs were investigated. Our results showed there was a tendency of SER improvement for elements lower than m/z than 100 using USN and mixed gas effect. In addition, 3D intensity distributions in the plasma were discussed to characterize the mixed gas effect and the mechanism of the sensitivity improvement. The 3D distributions indicated that N2 addition assisted the extension and stabilization of the ionized regions due to their high thermal conductivity. In addition, the transition of the maximum intensity to the torch side was found, especially for elements lower than m/z 100. The use of USN as a desolvation tool helped in decreasing the BGN. Mixed Ar−N2 plasma caused the generation of N-derived polyatomic ions such as 40Ar15N+. On the other hands, the generation of Ar-derived interference ions decreased. LOD was calculated in each case. From those results, the improvement in analytical performance using USN and Ar−N2 as with CN and Ar−N2 is considered. The insights obtained in this study are as follows:

  • The use of USN was more effective for light elements (m/z < 100 excluding 103Rh). In contrast, the use of CN was more effective for heavy elements on signal enhancement effect with use of Ar−N2 (1% N2). However, most elements in the platinum group were inert.

  • The mass number dependence of the signal enhancement effect has explained by shifting the ionization region obtained from ion intensity distribution in the ICP.

  • The effect of background reduction (especially, for light elements having m/z < 120) caused by reduction of solvent-derived polyatomic ion interference was observed when Ar−N2 was combined with USN.

  • LOD was improved as a result of signal enhancement and background reduction effect, therefore, USN was advantageous for light elements and CN was advantageous for heavy elements in many cases.