ANALYTICAL SCIENCES APRIL 2017, VOL. 33 537

Introduction

Silicon carbide (SiC) in powder, sintered, and single-crystal forms is used in various products.1 Owing to its hardness, heat resistance, and chemical stability, sintered SiC has recently been used for semiconductor manufacturing equipment. Moreover, the use of single-crystal SiC for semiconductor devices has been investigated because the bandgap, dielectric breakdown field strength, and thermal conductivity of SiC are higher than those of silicon.2,3

If single-crystal SiC is to be used in semiconductor devices, the concentrations of dopants in the p-n junction formed on the surface must be controlled; various analytical methods have been reported for this purpose. For example, secondary ion mass spectrometry has been used to evaluate techniques for implanting dopants. This method can be used to determine the depth profile of the dopants, but the oblique impact angle and the primary ion energy must be optimized for each element.4–6 In addition, secondary ion mass spectrometry is not appropriate for quantitative analysis, because matrix effects and variations in the accuracy and precision are larger than those of laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS), as well as those of glow discharge mass spectrometry.7–9 Glow discharge mass spectrometry is generally used for bulk analysis,10–13 and has detection limits ranging from 0.0005 to 0.1 μg g–1.14 However, the detection limits are

influenced by the sample matrix and the sample conductivity, and the detection limits for trace elements in SiC have been reported to range from 1 to 100 μg g–1.15 Therefore, this technique is usually used only for the qualitative analysis of bulk samples. The use of electrothermal vaporization inductively coupled plasma optical emission spectrometry for analysis of SiC has also been reported.16 In addition, LA-ICPMS has been used to analyze SiC powder with non-matrix-matched calibration;17 the measured concentrations of B, Ti, Cr, Mn, Fe, Ni, and Cu in a reference material (BAM-S003) were found to be in good agreement with the certified values for the material. Slurry nebulization ICPMS analysis of SiC powder has been reported, but CRMs are necessary for calibration.18

When SiC is used as a semiconductor substrate, the purity of the surface-layer (to several micrometers deep) is important for the performance of wiring patterns formed on the substrate. For  example, if impurities are present in the surface-layer, semiconductor failures, such as reversal of the p-n junction, can occur. In silicon semiconductor manufacturing, surface analysis can be performed by means of acid digestion ICPMS. However, because some types of SiC crystals are hard to digest, acid digestion ICPMS cannot be used for the quantitative analysis of single-crystal SiC surfaces.

LA-ICPMS permits multielement analysis with minimal sample preparation, and is widely used for direct analyses of solid samples. However, this technique requires a calibration standard that has a matrix similar to that of the sample to minimize the effects of elemental fractionation related to matrix effects.19–21 Because the preparation of a matrix-matched calibration standard is laborious, many alternative calibration

2017 © The Japan Society for Analytical Chemistry

† To whom correspondence should be addressed.E-mail: nfuruta@chem.chuo-u.ac.jp

Determination of Trace Elements in Sintered and Single-Crystal Silicon Carbide by Laser Ablation in Liquid Inductively Coupled Plasma Mass Spectrometry

Ryo MACHIDA,*,** Rina NISHIOKA,** Masahide FUJIWARA,* and Naoki FURUTA*†

* Faculty of Science and Engineering, Department of Applied Chemistry, Chuo University, 1-13-27 Kasuga, Bunkyo, Tokyo 112–8551, Japan

** Chiba Laboratory, Sumika Chemical Analysis Service, Ltd., 9-1 Kitasode, Sodegarura, Chiba 299–0266, Japan

Laser ablation in liquid (LAL) sampling method transformed hard-to-digest materials to soluble particles, and thus allowed for smooth decomposition by acid digestion. LAL sampling is useful to generate nanoparticles from samples with less contamination. After acid digestion, trace elements in the LAL-sampled particles were analyzed by solution nebulization inductively coupled plasma mass spectrometry (ICPMS). For the first time we demonstrated that LAL-ICPMS can be used to determine trace elements in hard-to-digest samples; sintered SiC and single-crystal SiC. Results obtained by laser ablation ICPMS and LAL-ICPMS were compared in terms of accuracy and detection limits. The detection limits of LAL-ICPMS were 0.04 – 0.4 μg g–1 for Al, Ti, Cr, Mn, Fe, Co, Ni, Cu, Sr, Y, Zr, and W. LAL-ICPMS is expected to be used to control contamination in the manufacturing of semiconductor devices.

Keywords Single-crystal silicon carbide, sintered silicon carbide, laser ablation in liquid, laser ablation, ICPMS, quantitative analysis

(Received December 15, 2016; Accepted February 21, 2017; Published April 10, 2017)

538 ANALYTICAL SCIENCES APRIL 2017, VOL. 33

strategies have been proposed for the quantitative analysis by LA-ICPMS. For example, commercially available NIST reference glasses (SRM 610 – 617) have been widely used for LA-ICPMS applications with non-matrix-matched calibration.22–24 Methods involving standard solution calibration,25,26 standard addition,27 and online isotope dilution have also been reported.28–30

As mentioned above, highly sensitive and accurate quantification of impurities is required to control surface-layer contamination of SiC substrates. Therefore, laser ablation in liquid (LAL)-ICPMS was demonstrated for highly sensitive analysis. In addition, LAL-ICPMS can be expected to provide reliable data because solution calibration can be used. LAL-ICPMS has been reported for the analysis of glass standards and high-purity iron metal.31–33 In this method, the sample is set in ultrapure water, the sample surface is laser ablated through the water, and the ablated particles are collected in the water and then analyzed by acid digestion ICPMS. The nanoparticles produced by LAL sampling decompose more readily than the particles produced by cracking of samples with a hammer. In our previous study, we found that the recoveries of Ag, As, Rb, Sb, Ga, Tl, Bi, Pb, Cd, Re, Sn, Cu, W, Mn, Ni, Ba, Ni, Ba, Zn, Al, Cr, Gd, Yb, Y, Sr, Zr, Hf, U, and Th from a CRM NIST 610 glass standard ranged from 95 to 108%.31 To date, LAL-ICPMS has not yet been used in practice. Herein, we report on the first use of LAL-ICPMS to quantify trace impurities in sintered SiC and single-crystal SiC, which are hard-to-digest materials.

Experimental

Samples and reagentsThe following SiC powder CRMs were used for this study:

NMIJ 8001a and NMIJ 8002a from the National Metrology Institute of Japan (Ibaraki, Japan) and JCRM R021, JCRM R022, and JCRM R023 from the Ceramic Society of Japan (Tokyo, Japan) (Table 1). The crystal forms of NMIJ 8001a and NMIJ 8002a are hexagonal(4H)-SiC and cubic-SiC, respectively. The crystal forms of the JCRM CRMs are not known.

As sample materials, sintered SiC with the crystal forms of hexagonal(6H)-SiC and single-crystal SiC with the crystal forms of hexagonal(4H)-SiC were provided by a SiC equipment manufacturer and a SiC wafer supplier, respectively. Photographs

of the sintered SiC and the single-crystal SiC samples are shown in Figs. 1a and 1b, respectively.

Glass standard of NIST 612 (National Institute of Standards and Technology, Maryland, USA) was used for tuning the LA-ICPMS system.

Ultrapure water (>18.2 MΩ cm, Milli-Q, Merck Millipore, Molsheim, France) was used to prepare all the solutions used in this work. Hydrofluoric acid (Tamapure-AA-10, 38%, Tama Chemicals Co., Kawasaki, Japan), HNO3 (Tamapure-AA-10, 68%, Tama Chemicals Co.), and H2SO4 (Tamapure-AA-100, 98%, Tama Chemicals Co.) were used for acid digestion.

Calibration standard solutions were prepared from the following SPEX CertiPrep (Metuchen, NJ, USA) multielement standards for ICPMS: XSTC-1 (Y), XSTC-8 (Ti, Zr, and W), and XSTC-13 (Al, Cr, Mn, Fe, Co, Ni, Cu, and Sr). The standards were diluted with 0.1 M HNO3. A  single-element standard solution of Rh (Kanto Chemical Co., Tokyo, Japan) was used as an internal standard. Calibration curves were prepared by measuring standard solutions, in which the concentrations of each element were 0, 5, 10, 100, 500, and 5000 pg mL–1.

InstrumentationA LA system (UP213, Electro Scientific Indutries (ESI),

Portland, OR, USA) and an ICPMS instrument (Agilent 7500ce, Agilent Technologies, Tokyo, Japan) were used in this study (Table 2).

Fig. 1　Photographs of (a) sintered SiC and (b) single-crystal SiC. In Fig. 1b, the yellow circle with a diameter of 5 cm is single-crystal SiC. Dotted lines indicate LAL sampling areas (2 × 4 mm).

Table 1　Certified values (μg g–1) for NMIJ and JCRM SiC powder CRMsa

ElementCRMs of NMIJ CRMs of JCRM

8001a 8002a R021 R022 R023

Al 83.2 ± 7.2 (8) 189 ± 19 (10) 390 ± 20 (5) 580 ± 20 (3) 30 ± 3 (10)Ti 6.37 ± 0.68 (11) 47 .7 ± 3 (6) 100 ± 4 (4) 30 <10 Cr 1.98 ± 0.52 (26) 61 .9 ± 9.4 (15) 40 ± 3 (8) 60 ± 5 (8) 10 Mn 0.53 ± 0.09 (17) 1 .6 ± 0.34 (21) <10 10 <10 Fe 46.7 ± 7.8 (17) 130 ± 7.4 (6) 180 ± 10 (6) 510 ± 10 (2) 150 ± 9 (6)Co —b —b —b —b —b

Ni 1.52 ± 0.22 (15) 4 .43 ± 0.8 (18) 10 10 10 Cu 0.47 ± 0.17 (36) 11 .5 ± 2.6 (23) —b —b —b

Sr —b —b —b —b —b

Y 0.31 ± 0.07 (21) 0 .58 ± 0.07 (12) —b —b —b

Zr —b —b —b 10 <10 W —b —b —b —b —b

a. The values in parentheses are percent relative standard deviations. The values without uncertainty are given for information only.b. Concentration is not certified.

ANALYTICAL SCIENCES APRIL 2017, VOL. 33 539

A pressure vessel with a stainless-steel jacket (HU-50, San-ai Kagaku, Aichi, Japan) was used for the acid digestion of SiC powders and sintered SiC cracked with a hammer.

SiC particles were observed by scanning electron microscopy (SEM; S-4300, Hitachi High Technologies, Tokyo, Japan). For SEM measurements, the surfaces of all the samples were coated with a Pt/Pd alloy by means of an ion sputtering device (E-1045, Hitachi High Technologies, Tokyo, Japan).

Dynamic light scattering (ELSZ-2000ZS, Otsuka Electronics Co., Osaka, Japan) was used to measure the diameter distribution of LAL-sampled particles. A  semiconductor laser with a wavelength of 660 nm was used.

The LAL-sampled particles of sintered SiC and single-crystal SiC were placed on a glass slide by means of a needle, and their  chemical compositions were measured by laser Raman microscopy (RAMAN-11, Nanophoton Corporation, Osaka, Japan; laser wavelength, 532 nm). This method enabled us to analyze individual particles.

Procedures for LA-ICPMS analysisFor LA-ICPMS, disk-shaped pellets (diameter, 10 mm; height,

2 mm) were prepared by pressing 300 mg of each of the five SiC powder CRMs at 10 MPa for 60 s.

Laser ablation was conducted under a He atmosphere, and the ablated particles were introduced into inline-type cascade impactors with a cutoff of <1.0 μm (type NL-1-1A, Tokyo Dylec, Tokyo, Japan) to select ablated particles smaller than 1.0 μm, which were then introduced into the ICPMS. This protocol suppressed the influence of elemental fractionation due to incomplete vaporization and ionization, which tend to occur when large particles are introduced into the ICP.19 Argon gas was added between the ablation cell and the impactor. A schematic diagram of the LA-ICPMS system with the ablation cell and the cascade impactor is shown in Fig. 2.

Using a NIST612 glass standard, the LA-ICPMS system was

tuned to provide the highest sensitivity for 59Co, 139La, and 232Th. For the determination of elements in both the sintered SiC and the single-crystal SiC, LA was carried out in the line-scanning mode under in-focus conditions. The crater diameter was 100 μm, the line width was 100 μm, the repetition rate was 20 Hz, the scanning speed was 10 μm s–1, and the scanning time was 30 s, resulting in a single laser line with a length of 0.3 mm. The concentrations of Al, Ti, Si, Cr, Mn, Fe, Co, Ni, Cu, Sr, Y, Zr, and W were measured by LA-ICPMS with 29Si as an internal standard. When the two NMIJ CRMs were analyzed, the three JCRM CRMs were used for calibration; and when the sintered SiC and the single-crystal SiC were analyzed, the two NMIJ CRMs were used for calibration. The integrated signal intensities of the elements were normalized by the integrated signal intensity of 29Si, because the amounts ablated from different samples depended on the interaction between the laser light and the sample surface. LA-ICPMS was performed in the time-resolved analysis mode, and element concentrations were calculated from Eq. (1):

[ ] ( )M /M Si= ∑ ∑ −I I bs

(1)

where [M] is the concentration of element M in the sample, ∑IM is the integrated signal intensity of element M during laser ablation, ∑ISi is the integrated signal intensity of 29Si during laser ablation, s is the sensitivity for element M, and b is the intercept of the calibration curve.

To obtain depth profiles of elements in the sintered SiC, LA in the single-site mode under in-focus conditions was conducted. The crater diameter was 100 μm, and the repetition rate was 5 Hz. The depth profiles of Al, Y, and Zr were measured by LA-ICPMS with 29Si as an internal standard.

Procedures for high-pressure acid digestion and solution nebulization ICPMS

Approximately 1 mg each of the NMIJ 8001a and 8002a SiC powders were weighed out and subjected to high-pressure acid digestion with a mixture of 2 mL of HF, 4 mL of HNO3, and 1 mL of H2SO4 in a polytetrafluoroethylene pressure vessel. The pressure vessel was placed in a stainless-steel jacket, and then heated in an oven at 210°C for 24 h.17 The digested solution was evaporated on a hot plate until dryness, and the resulting residue was diluted to 2 mL with 0.1 M HNO3. The concentration of H2SO4 was adjusted to <1% for analysis by ICPMS because spectral interference associated with S could not be corrected for accurately when the H2SO4 concentration exceeded 1% even if a collision cell was used.

Fig. 2　Schematic diagram of the LA-ICPMS system with a laser cell and a cascade impactor to select particles with sizes of <1.0 μm.

Table 2　Operating conditions for LA and ICPMS

Laser model UP213

Laser type Nd:YAGWavelength 213 nmPulse width 4 nsLaser energy 0.9 mJFluence 11 J cm–2

Diameter of crater 100 μmCarrier gas flow (He) 1.0 L min–1

ICPMS model Agilent 7500ce

Plasma conditionSolution nebulization Laser ablation

Hot plasma Cool plasma Hot plasma

RF power 1500 W 700 W 1500 WIntegration time 0.1 s 0.1 s 0.1 sCarrier gas flow (Ar) 0.8 L min–1 0.8 L min–1 0.9 L min–1

Auxiliary gas flow (Ar) 0.3 L min–1 0.7 L min–1 —Collision gas (He) 3.0 mL min–1 — 2.0 mL min–1

Internal standard 103Rh 103Rh 29SiIsotopes measured 27Al, 47Ti,

52Cr, 59Co, 60Ni, 63Cu, 88Sr, 89Y, 90Zr, 184W

55Mn, 56Fe 27Al, 47Ti, 52Cr, 55Mn, 56Fe, 59Co, 60Ni, 63Cu, 88Sr, 89Y, 90Zr, 184W

540 ANALYTICAL SCIENCES APRIL 2017, VOL. 33

Sintered SiC was cracked into particles of less than 1.0 mm with a hammer; approximately 10 mg of the particles was then subjected to high-pressure acid digestion for 96 h (4 days) in an oven at 210°C. Single-crystal SiC could not be decomposed, even when 10 mg of the cracked sample was subjected to high-pressure acid digestion for 336 h (2 weeks) at 210°C.

For the quantitative analysis by solution nebulization ICPMS, two plasma conditions were used: hot plasma in a He-collision mode was used for determination of Al, Ti, Cr, Co, Ni, Cu, Sr, Y, Zr, and W; and a cool-plasma mode was used for determination of Mn and Fe to reduce background noise. 103Rh was measured as an internal standard under both plasma conditions. The operating conditions for the ICPMS instrument are summarized in Table 2.

Procedures for LAL samplingPrior to LAL sampling, sintered SiC and single-crystal SiC

were deposited in a mixture of HF(19%) and HNO3(34%) and heated for 15 min to clean the surface. Two open-top chambers made of a tetrafluoroethylene-perfluoroalkyl vinyl ether copolymer were used for LAL sampling.31 The outer diameter, inner diameter, and height of the chamber used for the sintered SiC were 36, 32, and 10 mm, respectively. The outer diameter, inner diameter, and height of the chamber used for the single-crystal SiC were 60, 52, and 7 mm, respectively. Each sample was placed in a specified chamber, and then 5.6 mL (for the sintered SiC) or 6.4 mL (for the single-crystal SiC) of ultrapure water was added; as a result, the surface of the sample was 3 mm below the surface of the water. The sample surface was irradiated through the water. Forty laser lines, each 2 mm long, were ablated at 100 μm intervals. The line width was 100 μm, the repetition rate was 20 Hz, and the scanning speed was 10 μm s–1, resulting in a 2 × 4 mm ablation area. Ablation of the laser lines took approximately 2.5 h. After ablation, the LAL-sampled particles suspended in the water were transferred to a polytetrafluoroethylene pressure vessel by micropipette. To  ensure that all of the ablated particles were collected, the chamber was rinsed with 1 mL of ultrapure water, and then the rinse solution was transferred to a pressure vessel.

The sampled amounts of sintered SiC and single-crystal SiC were measured with a microbalance (MT5, Mettler Toledo, Zürich, Switzerland). The amounts of ablated particles were calculated from the difference between the weights of the samples before and after LAL sampling. Samples were dried by using a blow-dryer. The amounts of ablated particles were calculated five times; the standard deviation was ±0.002 mg. The sampled amounts of sintered SiC and single-crystal SiC were 0.23 and 0.19 mg, respectively. The LAL-sampled particles were subjected to high-pressure acid digestion by means of the procedures described for the SiC powder CRMs.

Results and Discussion

SEM observation of a SiC powder standard and LAL-sampled particles of sintered and single-crystal SiC

One of the SiC powder standards (NMIJ 8002a) and LAL-sampled particles of sintered SiC and single-crystal SiC were placed on silicon wafers for SEM observation at a magnification of 5000× (Fig. 3). The SEM images indicated that the diameters of the NMIJ 8002a particles ranged from 0.1 to 2.0 μm (Fig. 3a). LAL sampling of the sintered SiC gave spherical particles and angular particles (Fig. 3b), whereas LAL sampling of the single-crystal SiC gave mainly spherical particles (Fig. 3c). SEM image indicated that the diameters of the LAL-sampled sintered and single-crystal SiC particles ranged from 0.1 to 5.0 μm, whereas dynamic light-scattering measurements indicated that the major diameters were 0.2 and 0.3 μm for the sintered and single-crystal SiC, respectively.

The difference in the particle shape may have been due to differences in how the particles were generated: specifically, the spherical particles were likely generated in the heat-affected zone by means of melting-congelation, whereas the angular particles were likely generated mechanically by the laser-induced plasma shock wave. Since sintered SiC is polycrystalline and has defects, angular particles were generated by shock waves. On the other hand, single-crystal SiC has few defects and the surface is flat. Therefore, angular particles were less generated by shock waves.

Verification of changes in chemical composition by means of laser Raman microscopy

Raman spectra of a glass slide, sintered SiC, single-crystal SiC, and LAL-sampled particles of sintered SiC and single-crystal SiC are shown in Fig. 4. The peaks at 771, 793, and 973 cm–1 in the spectrum of the sintered SiC (Fig. 4b) correspond to the hexagona(6H)-SiC polytype, and the peak at 780 cm–1 in the spectrum of the single-crystal SiC (Fig. 4c) corresponds to the hexagonal(4H)-SiC polytype.34 In the Raman spectra of the LAL-sampled particles obtained from the sintered SiC and the single-crystal SiC (Figs. 4d and 4e, respectively), the peak at 521 cm–1 corresponds to silicon. In the Raman spectra of the LAL-sampled particles, none of the typical peaks for the SiC polytypes were detected. Broad peaks at 1369 cm–1 (D band) and 1591 cm–1 (G band) were observed in the spectrum of the LAL-sampled particles from the sintered SiC. These peaks indicate the presence of amorphous carbon.35

The Raman spectroscopy results indicate that LAL sampling transformed hard-to-digest sintered SiC and single-crystal SiC into soluble particles composed of silicon and amorphous

Fig. 3　SEM images of (a) commercially available SiC powder (NMIJ 8002a) without pretreatment, (b) sintered SiC particles collected by LAL sampling, and (c) single-crystal SiC particles collected by LAL sampling.

ANALYTICAL SCIENCES APRIL 2017, VOL. 33 541

carbon. This transformation, which probably occurred by means of melting-congelation during laser ablation of the sample surface, allowed solution nebulization ICPMS analysis after high-pressure acid digestion.

Analysis of the NMIJ 8001a and NMIJ 8002a CRMsThe NMIJ 8001a and NMIJ 8002a CRMs were analyzed by

LA-ICPMS and acid digestion ICPMS (Table 3). For acid digestion ICPMS, a standard solution was used for calibration. ICPMS was performed in the spectrum analysis mode.

The elements Co, Cu, Sr, Y, and W could not be quantified by LA-ICPMS, because their concentrations in the JCRM reference

materials were not certified. The measured concentrations of Al, Ti, Cr, Fe, and Ni in NMIJ 8001a and NMIJ 8002a were in good agreement with the certified values. The recoveries of these five elements were 100 ± 20%. The Mn recoveries from NMIJ 8001a and NMIJ 8002a were 74 and 69%, respectively. The uncertainty was not given for the concentration of Mn in JCRM R021 and JCRM R023. Therefore, the Mn concentration was less reliable than that for the other elements. These results indicate that LA-ICPMS analysis is useful for the quantitative analysis of SiC powders without any sample pretreatment, but is inappropriate for elements for which the concentration is less reliable.

In the case of acid digestion ICPMS, the recoveries of all the measured elements from both CRMs were 100 ± 20%. It is noticed that the elemental sensitivity is the same even if the crystal forms are different; hexagonal(4H)-SiC and cubic-SiC for NMIJ 8001 and NMIJ 8002a, respectively. The recoveries of Mn and Cu from NMIJ 8002a were 81 and 88%, respectively; that is, they were a little lower than the recoveries of the other elements, although the measured values of Mn and Cu were within the range of certified values.

In the case of acid digestion ICPMS, the recoveries of Ti, Mn, Fe, and Ni from NMIJ 8001a were close to 100%, that is, better than the recoveries obtained by LA-ICPMS.

Analysis of sintered SiCSintered SiC is used for semiconductor manufacturing

equipment. Therefore, it is important to measure any impurity on the surface. The surface analysis was done by LA-ICPMS and LAL-ICPMS on an intact sintered SiC sample, whereas acid digestion ICPMS was carried out on cracked sintered SiC (Table 4). The NMIJ CRMs were used as standard reference materials for LA-ICPMS. Because the certified values for the NMIJ CRMs have large variation, the concentrations obtained by acid digestion ICPMS (Table 3) were used for the calibration of LA-ICPMS.

The surface-layer concentrations of Al, Ti, and Cr determined by LA-ICPMS and by LAL-ICPMS were in good agreement (within 32%). The concentrations of Fe and Ni obtained by LA-ICPMS were higher than those obtained by LAL-ICPMS, and those of Y and Zr obtained by LA-ICPMS were lower than

Fig. 4　Raman spectra of (a) a glass slide, (b) sintered SiC, (c) single-crystal SiC, (d) LAL-sampled particles of sintered SiC, and (e) LAL-sampled particles of single-crystal SiC.

Table 3　Analysis of NMIJ 8001a and NMIJ 8002a SiC CRMsa

Element

NMIJ 8001a NMIJ 8002a

LA-ICPMS Acid digestion ICPMS LA-ICPMS Acid digestion ICPMS

Caliblated by JCRM CRMsb Solution calibration Caliblated by JCRM CRMsb Solution calibration

Found/μg g–1

Recovery, %

Found/μg g–1

Recovery, %

Found/μg g–1

Recovery, %

Found/μg g–1

Recovery, %

Al 79.8 ± 0.6 (1) 96 81.7 ± 0.8 (1) 98 184 ± 2.9 (2) 97 183 ± 2.1 (1) 97Ti 7.3 ± 0.2 (2) 114 6.6 ± 0.4 (6) 103 56 .3 ± 3.4 (6) 118 47 .8 ± 3.2 (7) 100Cr 2.0 ± 0.1 (4) 101 1.9 ± 0.2 (8) 96 62 .3 ± 1.4 (2) 101 61 .6 ± 1.4 (3) 100Mn 0.39 ± 0.005 (1) 74 0.53 ± 0.05 (9) 100 1 .1 ± 0.1 (11) 69 1 .3 ± 0.02 (2) 81Fe 47.5 ± 0.3 (1) 102 46.5 ± 1.5 (3) 99 119 ± 6.6 (6) 92 126 ± 3.9 (3) 97Co —c —c 0.03 ± 0.002 (8) —c —c —c 0 .15 ± 0.01 (7) —c

Ni 1.7 ± 0.1 (8) 110 1.5 ± 0.05 (3) 99 4 .32 ± 0.8 (18) 97 4 .31 ± 0.1 (2) 97Cu —c —c 0.51 ± 0.1 (12) 108 —c —c 10 .1 ± 0.5 (5) 88Sr —c —c 0.02 ± 0.01 (40) —c —c —c 0 .04 ± 0.02 (43) —c

Y —c —c 0.33 ± 0.03 (8) 107 —c —c 0 .54 ± 0.01 (1) 92Zr 0.31 ± 0.01 (5) —c 0.28 ± 0.02 (6) —c 16 .1 ± 0.6 (4) —c 9 .2 ± 0.1 (1) —c

W —c —c 0.30 ± 0.02 (6) —c —c —c 10 .6 ± 0.4 (1) —c

a. The values in parentheses are percent relative standard deviations (n = 3). b. JCRM R021, JCRM R022, and JCRM R023. c. No data.

542 ANALYTICAL SCIENCES APRIL 2017, VOL. 33

those obtained by LAL-ICPMS. However, the concentrations of these four elements determined by these two techniques had the same order of magnitude.

The precision of the results obtained by means of LA-ICPMS was good among LA-ICPMS, LAL-ICPMS, and acid digestion ICPMS. Surface-layer data obtained by means of LAL-ICPMS had larger variation than those obtained by acid digestion ICPMS for bulk cracked sintered SiC even though ICPMS with solution calibration was used for both analyses. This difference was due to the fact that the sampled amount obtained by LAL-ICPMS (0.23 mg) was considerably smaller than that used for acid digestion ICPMS (13.5 mg), which in turn resulted in dispersion of the LAL-ICPMS data.

The concentrations obtained for bulk cracked SiC can be taken as the average concentrations of impurities in sintered SiC. The concentrations of Al, Ti, Cr, Fe, Ni, and Zr were higher than those of other elements. Al, Ti, Cr, Fe are known to be present in the raw materials used to prepare the sintered SiC. The presence of these elements as a result of contamination from a growth furnace for sintering SiC powders has been reported.36 It is considered that Ni and Zr were also contaminated in a manufacturing process.

The surface-layer data (obtained by LAL-ICPMS) and bulk data for Ti, Cr, Fe, Co, Ni, Sr, and W were in good agreement, indicating that these elements were homogeneously distributed in the sintered SiC. In contrast, the surface-layer concentrations of Al, Mn, Cu, Y, and Zr were higher than the concentrations in the bulk. Depth profiles, obtained by LA-ICPMS in the single-site mode, of Al, Y, and Zr in sintered SiC are shown in Fig. 5. Aluminum was enriched at a depth of approximately 0 – 5 μm from the sample surface, and partial enrichment was observed at a depth of 18 and 67 μm from the sample surface in the first measurement of Al. Yttrium was enriched at a depth of approximately 0 – 5 μm from the sample surface. Zirconium was enriched at a depth of approximately 0 – 1 μm from the sample surface and partial enrichment was observed at a depth of 18 μm from the sample surface in the first measurement of Zr. The trend in elemental enrichment was the same as that in elemental concentrations shown in Table 4. Depth profiles for Mn and Cu could not be obtained by LA-ICPMS, because the concentrations of these elements were below the detection limits.

Analysis of single-crystal SiCSingle-crystal SiC is used for a semiconductor substrate.

Therefore, it is important to measure any impurity on the surface. However, it is more difficult to decompose by acids than sintered SiC. Table 5 lists the concentrations of elements in the single-crystal SiC determined by means of LA- and LAL-ICPMS, along with the corresponding detection limits calculated as 3-times the standard deviation of repeated measurements of an operational blank (n = 5). Using LA-ICPMS, only Al and Ti concentrations could be determined; the concentrations of the other elements were below the detection limits.

In contrast, using LAL-ICPMS, the concentrations of Al, Ti, Cr, Mn, Fe, Ni, Cu, and Zr could be determined. The Al concentration obtained by LAL-ICPMS was higher than that obtained by LA-ICPMS. The discrepancy between the values measured by these two techniques may have been due to the fact that the measured values were much lower than the calibration curve range of the NMIJ 8001a and NMIJ 8002a CRMs (Table 1). In contrast, because the determined Ti concentration was within the calibration curve range (Table 1), the Ti concentrations obtained by LA- and LAL-ICPMS were in good agreement. The detection limits for LAL-ICPMS analysis of the twelve elements listed in Table 5 ranged from 0.04 to 0.4 μg g–1. To quantify Co, Sr, Y, and W, the detection limits needs to be lowered by using a longer laser ablation time to increase the sampled amount.

The presence of Al, Ti, Cr, Mn, Fe, Ni, Cu, and Zr is reportedly due to contamination from graphite susceptors and thermal insulation materials used in chemical vapor deposition.36,37 Because cutting blades and surface-polishing materials are made of metals, these also could be possible source of contamination.38

Conclusions

In this study, we evaluated the use of LA-ICPMS and LAL-ICPMS for the determination of trace elements in sintered SiC and single-crystal SiC. LA-ICPMS was found to be useful for the quantitative analysis without any sample pretreatment. However, data obtained by LA-ICPMS were less reliable than

Table 4　Surface-layer and bulk analysis of sintered SiCa

Sampling area Surface Surface Bulk cracked SiC

Element

LA-ICPMS LAL-ICPMS Acid digestion ICPMS

— Sampling amount: 0.23 mg Sampling amount: 13.5 mg

Caliblated by NMIJ CRMsb

Found/μg g–1

Solution calibrationFound/μg g–1

Solution calibrationFound/μg g–1

Al 518 ± 16 (3) 460 ± 35 (8) 261 ± 6 (2)Ti 146 ± 12 (8) 193 ± 25 (13) 204 ± 11 (5)Cr 38 ± 3.0 (8) 32 ± 1.2 (4) 38 ± 2.7 (7)Mn <1 0 .4 ± 0.08 (22) 0 .07 ± 0.003 (5)Fe 162 ± 10 (6) 100 ± 11 (11) 118 ± 10 (8)Co 0 .4 (9) 1 .1 ± 0.2 (19) 1 .3 ± 0.1 (6)Ni 37 ± 3 (8) 24 ± 3.4 (14) 30 ± 1.9 (6)Cu <0.4 0 .48 ± 0.06 (12) 0 .05 ± 0.02 (41)Sr <0.2 0 .17 ± 0.02 (9) 0 .16 ± 0.01 (9)Y 1.2 ± 0.2 (19) 5 .4 ± 0.7 (13) 0 .4 ± 0.04 (10)Zr 40 ± 6.2 (15) 98 ± 2.5 (3) 29 ± 2.5 (9)W <0.2 0 .48 ± 0.07 (14) 0 .35 ± 0.03 (10)

a. The values in parentheses are percent relative standard deviations (n = 3). b. NMIJ 8001a and NMIJ 8002a.

ANALYTICAL SCIENCES APRIL 2017, VOL. 33 543

LAL-ICPMS when the concentration of an element in a sample was outside the calibration curve range of CRMs. In addition, if the concentration of an element was not certified in the CRMs, the concentration of that element in the CRMs had to be determined by a reliable analytical method such as acid digestion ICPMS. In contrast, LAL-ICPMS could be used to determine

trace elements concentrations without CRMs. Furthermore, LAL-ICPMS data were more reliable than LA-ICPMS data because solution nebulization was used for the former technique. The use of LAL sampling permitted local ablation of the surface of the hard-to-digest samples; sintered SiC and single-crystal SiC to give digestible particles for measurement by ICPMS. We found that LAL sampling was useful for nanoparticulation of sample surfaces and that it transformed hard-to-digest crystals to soluble particles. On the basis of our analysis of the impurities in the surface-layer of sintered SiC and single-crystal SiC, we  suggest that possible sources of contamination include raw materials, the growth furnace, graphite susceptors, thermal insulation materials, cutting blades, and surface-polishing materials.

For the analysis of single-crystal SiC, a highly sensitive method is required. The detection limits of LA-ICPMS and LAL-ICPMS were 0.2 – 6 and 0.04 – 0.4 μg g–1, respectively. Moreover, increasing the sampled amount by increasing the LAL sampling time can be expected to improve the detection limits.

Although LAL-ICPMS must be investigated further for use in situations in which high sensitivity and throughput are required, this technique should become a powerful tool for the control of contamination in the manufacturing of semiconductor devices.

Acknowledgements

This research was supported through a joint research project with Sumika Chemical Analysis Service and the Institute of Science and Engineering of Chuo University and through a

Fig. 5　Depth profiles of Al, Y, and Zr in sintered SiC obtained by LA-ICPMS measurements, which were carried out three times for each element.

Table 5　Surface-layer analysis of single-crystal SiC by LA- and LAL-ICPMS, with detection limits (DLs)a

Element

LA-ICPMS LAL-ICPMS

Caliblated by NMIJ CRMsb Solution calibration

Found/μg g–1 DL/μg g–1 Found/μg g–1 DL/μg g–1

Al 6 ± 1 (17) 2 22 ± 0.89 (4) 0.4Ti 13 ± 5 (36) 2 19 ± 3.3 (18) 0.4Cr <1 1 0 .4 ± 0.06 (14) 0.2Mn <1 1 0 .2 ± 0.02 (9) 0.04Fe <6 6 5 .3 ± 0.9 (17) 0.4Co <0 .4 0.4 <0 .1 0.1Ni <4 4 0 .7 ± 0.1 (15) 0.2Cu <0 .4 0.4 3 .3 ± 0.3 (8) 0.2Sr <0 .2 0.2 <0 .1 0.1Y <0 .2 0.2 <0 .04 0.04Zr <0 .2 0.2 1 .1 ± 0.2 (18) 0.04W <0 .2 0.2 <0 .04 0.04

a. The values in parentheses are percent relative standard deviations (n = 3). b. NMIJ 8001a and NMIJ 8002a.

544 ANALYTICAL SCIENCES APRIL 2017, VOL. 33

Grant-in-Aid for Scientific Research (C) (No. 26410160) funded by the Ministry of Education, Culture, Sports, Science, and Technology, Japan.

References

1. H. Tanaka, J. Ceram. Soc. Japan, 2011, 119, 218. 2. H. Okumura, Jpn. J. Appl. Phys., 2006, 45, 7565. 3. H. Matsunami, Jpn. J. Appl. Phys., 2004, 43, 6835. 4. K. Wittmaack, Surf. Interface Anal., 1998, 26, 290. 5. A. Besmehn and P. Hoppe, Geochim. Cosmochim. Acta,

2003, 67, 4693. 6. M. P. McCann, W. F. Calaway, M. J. Pellin, I. V. Veryovkin,

I. Constantinides, A. Adriaens, and F. Adams, Surf. Interface Anal., 2002, 33, 394.

7. J. S. Becker and H. J. Dietze, Int. J. Mass. Spectrom., 2003, 228, 127.

8. J. Pisonero, B. Fernández, and D. Günther, J. Anal. At. Spectrom., 2009, 24, 1145.

9. M. J. Whitehouse, Geostand. Geoanal. Res., 2004, 28, 195.10. R. Matschat, J. Hinrichs, and H. Kipphardt, Anal. Bioanal.

Chem., 2006, 386, 125.11. Y. Xing, L. Xiaojia, and W. Haizhou, Int. J. Mass Spectrom.,

2007, 262, 25.12. W. Vieth and J. C. Huneke, Spectrochim. Acta, 1991, 46B,

137.13. M. Kasik, C. Venzago, and R. Dorka, J. Anal. At. Spectrom.,

2003, 18, 603.14. A. I. Saprykin, J. S. Becker, and H. J. Dietze, Fresenius J.

Anal. Chem., 1999, 364, 763.15. R. Jäger, A. I. Saprykin, J. S. Becker, H. J. Dietze, and J. A.

C. Broekaert, Mikrochim. Acta, 1997, 125, 41.16. U. Schäffer and V. Krivan, Anal. Chem., 1999, 71, 849.17. H. Zhou, Z. Wang, Y. Zhu, Q. Li, H. J. Zou, H. Y. Qu, Y.

Chen, and Y. Du, Spectrochim. Acta, Part B, 2013, 90, 55.18. Z. Wang, D. Qiu, Z. Ni, G. Tao, and P. Yang, Anal. Chim.

Acta, 2006, 577, 288.19. R. Machida, T. Nakazawa, and N. Furuta, Anal. Sci., 2015,

31, 345.20. H. P. Longerich, D. Günther, and S. E. Jackson, Fresenius

J. Anal. Chem., 1996, 355, 538.21. K. P. Jochum, B. Stoll, K. Herwig, and M. Willbold, J.

Anal. At. Spectrom., 2007, 22, 112.22. M. Ohata, D. Tabersky, R. Glaus, J. Koch, B. Hattendorf,

and D. Günther, J. Anal. At. Spectrom., 2014, 29, 1345.23. D. B. Aeschliman, S. J. Bajic, D. P. Baldwin, and R. S.

Houk, J. Anal. At. Spectrom., 2003, 18, 872.24. D. Tabersky, N. A. Luechinger, M. Rossier, E. Reusser, K.

Hametner, B. Aeschlimann, D. A. Frick, S. C. Halim, J. Thompson, L. Danyushevsky, and D. Günther, J. Anal. At. Spectrom., 2014, 29, 955.

25. C. O’ Connor, B. L. Sharp, and P. Evans, J. Anal. At. Spectrom., 2006, 21, 556.

26. L. Halicz and D. Günther, J. Anal. At. Spectrom., 2004, 19, 1539.

27. F. Claverie, J. Malherbe, N. Bier, J. L. Molloy, and S. E. Long, Anal. Chem., 2013, 85, 3584.

28. S. F. Boulyga and K. G. Heumann, Int. J. Mass Spectrom., 2005, 242, 291.

29. C. K. Yang, P. H. Chi, Y. C. Lin, Y. C. Sun, and M. H. Yang, Talanta, 2010, 80, 1222.

30. M. Tibi and K. G. Heumann, J. Anal. At. Spectrom., 2003, 18, 1076.

31. R. Machida, T. Nakazawa, Y. Sakuraba, M. Fujiwara, and N. Furuta, J. Anal. At. Spectrom., 2015, 30, 2412.

32. S. Okabayashi, T. D. Yokoyama, Y. Kon, S. Yamamoto, T. Yokoyama, and T. Hirata, J. Anal. At. Spectrom., 2011, 26, 1393.

33. D. N. Douglas, J. L. Crisp, H. J. Reid, and B. L. Sharp, J. Anal. At. Spectrom., 2011, 26, 1294.

34. S. Nakashima and H. Harima, Phys. Stat. Sol. (a), 1997, 162, 39.

35. A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, and U. Pöschl, Carbon, 2005, 43, 1731.

36. J. Nishio, M. Kushibe, K. Masahara, K. Kojima, T. Ohno, Y. Ishida, T. Takahashi, T. Suzuki, T. Tanaka, S. Yoshida, and K. Arai, Mat. Sci. Forum, 2002, 389–393, 215.

37. J. C. Viala, F. Bosselet, V. Laurent, and Y. Lepetitcorps, J. Mater. Sci., 1993, 28, 5301.

38. L. Zhou, V. Audurier, P. Pirouz, and J. A. Powell, J. Electrochem. Soc., 1997, 144, L161.