Journal Archive

Six standard asbestos mineral samples, such as chrysotile, amosite, actinolite, anthophyllite, crocidolite, and tremolite (Health and Safety Laboratory, HSE, Harpur Hill, Buxton, Derbyshire, UK) and talc (a non-asbestos Mg-silicate mineral, Sigma-Aldrich) were selected for analysis. The bulk fibrous minerals were ground to smaller particulate sizes using a mortar and pestle. The ground particles were sieved (<20 μm) and collected on Ag foils in a well dispersed manner suitable for single particle analysis.

The low-Z particle EPMA was applied to obtain the information on morphology and elemental compositions of individual mineral particles. The EPMA measurements were carried out using a Jeol JSM-6390 SEM equipped with an Oxford Link SATW ultrathin window EDX detector. The resolution of EDX detector is 133 eV for Mn-Kα X-rays. X-ray spectra were recorded under the control of Oxford INCA Energy software. To achieve optimal experimental conditions, such as a low background level in the spectra and high sensitivity for low-Z element analysis, an accelerating voltage of 10 kV was used. The beam current was 1.0 nA for all measurements. More details of the measurement conditions can be found elsewhere (Ro et al., 1999). The net X-ray intensities of the elements were obtained by nonlinear least-squares fitting of the collected spectra using the AXIL program (Vekemans et al., 1994). The elemental concentrations of individual particles were determined from their X-ray intensities by the application of a Monte Carlo calculation combined with reverse successive approximations. (Ro et al., 2003). The quantification procedure provided results accurate within 12% relative deviations between the calculated and nominal elemental concentrations when the method was applied to various types of standard salt or mineral particles (Ro et al., 2001, 2000).

Principal component analysis (PCA) was performed, using a code built on MATLAB (version 7.11.0) platform (Jolliffe and Cadima, 2016; Jackson, 1991), on chemical composition results obtained for individual particles using low-Z particle EPMA to investigate whether the quantitative compositional results were sufficient to distinguish between the six asbestos mineral particles and the talc particles.

ATR-FT-IR imaging measurements of individual particles were performed using a Perkin-Elmer Spectrum 100 FT-IR spectrometer interfaced to a Spectrum Spotlight 400 FT-IR microscope. For ATR imaging, an ATR accessory employing a germanium hemispherical internal reflection element (IRE) crystal with a diameter of 600 μm was used. The ATR accessory is mounted on the X-Y stage of the FT-IR microscope, and the IRE crystal makes contact with the sample via a force lever. The IRE crystal is scanned laterally by an IR beam, and both the crystal and attached sample are moved under the beam as the X-Y stage is driven by a stepper-motor. Given that the crystal and the sample move together as they are scanned, the initial contact between the crystal and the sample is maintained during the imaging data acquisition. The maximum scan area is limited to 400×400 μm². The ultimate spatial resolution of the IR imaging is approximately equal to the wavelength of the incident IR radiation. However, the hemispherical IRE crystal is similar to a lens, in that the IR beam is condensed when the beam strikes the IRE. The extent of the condensation is proportional to the refractive index of the IRE material (4.0 for germanium) (Griffith and Haseth, 2007). Thus, a spatial resolution of 3.1 μm at 1,726 cm^-1 is achieved beyond the ultimate spatial resolution limit (Crane et al., 2007; Van Dalen et al., 2007). A 16×1 pixel mercury cadmium telluride (MCT) array detector was used to obtain FT-IR images with a pixel size of 1.56 μm. For each pixel, an ATR-FTIR spectrum, ranging from 720 to 4,000 cm^-1 with a spectral resolution of 4 cm^-1, was obtained from four interferograms that were co-added and Fourier-transformed. The position of the crystal on the sample was determined using a visible light optical microscope equipped with a light-emitting diode and a charge-coupled device camera. Additionally, the optical image was used to relocate single particles that had been analyzed using low-Z particle EPMA before the ATR-FTIR imaging measurement, as the optical microscopy provides the image of sufficient spatial resolution. Spectral data processing was performed using the Perkin-Elmer Spectrum IMAGE software. ATR-FTIR spectra for the bulk mineral fiber samples were also obtained using a universal ATR accessory of the Perkin-Elmer Spectrum 100 FT-IR spectrometer, where a diamond was used as the IRE crystal.

A quantitative single particle analytical technique, low-Z particle EPMA, was employed to examine six asbestiform and one non-asbestiform talc mineral samples. Fig. 1 shows SE image of those particles, where particles of the different samples do not show unique morphology depending on their types (except some rod-like particles of tremolite), indicating the micrometer size particles cannot be identified nor distinguished only from their morphology (Langer et al., 1972). Fig. 2 shows X-ray spectra and elemental concentrations of typical single particles of the six asbestos and talc minerals together with relative elemental concentrations of Mg, Ca, Fe, Na, and Al normalized to Si. The X-ray spectra of the minerals show that Si and Mg are present in all the mineral types with different relative atomic concentrations. Furthermore, minor Na, Ca, and/or Fe in some mineral types are present characteristically among the mineral types, which seems to be useful to distinguish between the different asbestos and talc minerals. The relative elemental concentrations normalized to Si obtained from the measured single particles for the mineral types are plotted in the right panel of Fig. 2. The relative elemental concentrations for all the mineral types determined by low-Z particle EPMA quite well match with their idealized chemical compositions. Just for an example, the relative atomic concentration of Mg normalized to Si for anthophyllite is 86.5%, which corresponds well to its stoichiometric value of 87.5% (its idealized formula is (Mg,Fe)₇Si₈O₂₂(OH)₂). The error bars in the panel are the min-max ranges which indicate the range of the relative elemental concentrations obtained from all the measured single particles for each mineral type. As shown in the right panel of Fig. 2, actinolite (Ca₂(Mg,Fe²⁺)₅Si₈O₂₂(OH)₂) can be easily distinguished from the other mineral types as its abundant contents of Mg, Ca, and Fe are unique among the mineral types of interest. Amosite and crocidolite ((Fe²⁺)²(Fe²⁺,Mg)₅Si₈O₂₂(OH)₂ and Na(Mg,Fe²⁺)₃(Fe³⁺)₂Si₈O₂₂(OH)₂) are also distinguishable from the others as Mg and Fe are significant with higher Fe than Mg. Amosite and crocidolite are different because of the significant presence and absence of Na in crocidolite and amosite, respectively. Chrysotile (Mg₃(Si₂O₅)(OH)₄) is unique among all the mineral types of interest as it is Mg/Si=1.46. Tremolite (Ca₂Mg₅Si₈O₂₂(OH)₂) and actinolite (Ca₂(Mg,Fe²⁺)₅Si₈O₂₂(OH)₂) are characteristic because of their significant Ca content, but tremolite is distinguishable from actinolite as tremolite has negligible Fe content compared to actinolite. Anthophyllite ((Mg,Fe)₇Si₈O₂₂(OH)₂) can be unique as it has a comparably high Mg content to Si with considerable Fe. The non-asbestiform talc mineral (Mg₃(Si₄O₁₀)(OH)₂) can be also unique as it has a comparably high Mg content to Si without Fe. Based on the low-Z particle EPMA results, the asbestiform and non-asbestiform minerals seem to be well differentiated from each other on a single particle level.

Fig. 1. 
Typical SE images of six asbestos (actinolite, amosite, anthophyllite, chrysotile, crocidolite, and tremolite) and a non-asbestos (talc) Mg-silicate mineral particles.

Fig. 2. 
X-ray spectra and elemental concentrations of six asbestiform and talc minerals on single particle basis and relative elemental concentrations normalized to Si obtained from all the measured single particles for each mineral type (error bars: Min-Max).

In order to visually examine the low-Z particle EPMA results regarding chemical composition characteristics of the minerals of interest, a multivariate modeling and analysis technique, i.e., principal component analysis (PCA), was performed for the low-Z particle EPMA X-ray data using a MATLAB program. Fig. 3 shows a cluster principal component (PC) scatter plot of the EPMA data for the total seven different groups of individual Mg-silicate minerals on the first two PC axes. The scatter plot shows that amosite, actinolite, crocidolite, and chrysotile minerals are clearly differentiated from the others. On the other hand, anthophyllite, tremolite, and talc belong to one group which cannot be differentiated among themselves on the PC scatter plot, which is strange as the raw EPMA data seem to be quite different between anthophyllite, tremolite, and talc. As the Ca content of tremolite is significant to be differentiated from anthophyllite and talc based on their EPMA X-ray data, somehow PCA did not take account of the Ca content when the EPMA X-ray data for all the minerals were processed. When PCA analysis was performed using only three data sets of anthophyllite, tremolite, and talc minerals, tremolite was clearly differentiated from anthophyllite and talc, but anthophyllite and talc were not yet separated well on the PC plot (not shown here), indicating that the low-Z particle EPMA technique can clearly distinguish between all the individual asbestos mineral particles except anthophyllite and talc.

Fig. 3. 
PC scatter plot of six standard asbestos and talc (non-asbestos) samples on PC1 and PC2 for data obtained from low-Z particle EPMA.

Anthophyllite ((Mg,Fe)₇Si₈O₂₂(OH)₂) is a mineral which is generated from the metamorphic reaction of talc (Mg₃(Si₄O₁₀)(OH)₂) and has similar chemical compositions with talc, except Fe in anthophyllite. The stoichiometric ratios of Mg/Si in anthophyllite and talc are 0.87 and 0.75 respectively, which are somewhat too close to be distinguished based on the EPMA data, i.e., the Mg/Si ratios are overlapping in the EPMA data as shown in the right panel of Fig. 2 (Mg/Si=0.701-1.031 and 0.608-0.808 for anthophyllite and talc, respectively). When the EPMA X-ray data of anthophyllite were carefully examined, among analyzed 21 standard anthophyllite particles, 20 particles were identified in the pure form of anthophyllite, where 11 and 9 particles were with and without Fe, respectively. The remaining one particle turned out to be Mg-vermiculite as an impurity. The X-ray spectral data can be useful to distinguish between Fe-containing anthophyllite and talc particles, but it is difficult to clearly differentiate non-Fe-containing anthophyllite particles from talc.

ATR-FTIR imaging technique was employed to distinguish between anthophyllite and talc. Fig. 4 shows secondary electron images (SEIs) before and after ATR-FTIR measurements, ATR-FTIR PCA image, and light optical image of the same 21 individual anthophyllite particles on Ag foil. The ATR-FTIR image of Fig. 4C was obtained by the application of principle component analysis (PCA) after the first differentiation of original ATR-FTIR spectra at all the pixels in the image. Fig. 4D is a light optical image obtained from an optical microscope in the ATR-FTIR imaging instrument, which is used for the location of particles of interest. Although the quality of SEIs and ATR-FTIR images differs due to the inherently different spatial resolutions of the images (e.g., ~100 nm for the SEI and 3.9 μm at 1,000-1,200 cm^-1 for the ATR-FTIR image), the same patterns of particle location among the images ensure that the same particles of micrometer size were dealt with. Their equivalent diameters range from 1.54 to 6.12 μm (determined from SEI before ATR-FTIR measurement), where the equivalent diameter is calculated by assuming that a particle with the same area for a particle on SEI is circular. The area of each image is approximately 150×110 μm². Considering that the pixel size of an ATR-FTIR image is 1.56×1.56 μm², the number of pixels for the ATR-FTIR image is ~6,780. All pixels in the image contain full IR spectra ranging from 4,000 to 680 cm^-1. Although the experimental ATR-FTIR imaging data were obtained on 1.56×1.56 μm² pixel size, the manufacture’s software interpolates ATR-FTIR imaging pixel data onto a display image with many more pixels such that the final display image looks better than the actual image.

Fig. 4. 
Secondary electron images (SEIs) (A) before and (B) after ATR-FTIR measurements, (C) ATR-FTIR PCA image, and (D) light optical image of anthophyllite particles on Ag foil. The same particles in the four images are denoted by particle number.

The SEIs clearly shows the morphology and locations of the 21 particles before and after ATR-FTIR imaging measurement. For the ATR-FTIR imaging measurement, the sample has to be in contact with the IRE crystal so that some force is applied to the sample during the contact. When they are in good contact and well pressed with the IRE crystal, particles are (partly or fully) embedded into the ductile Ag collecting foil. The nonconductive particles sitting on Ag foil (Fig. 4A) look bright because of high secondary and backscattered electron yields of insulating particles, whereas the particles on SEI taken after FTIR measurement look dark (Fig. 4B), as electrons can flow from the embedded particles into the metallic foil, resulting in low secondary and backscattered electron yields (Goldstein et al., 2003). By comparing Fig. 4A and 4B, most particles spread out and some particles were reoriented and/or broken off into small parts as the force by the IRE crystal was applied on the particles. Some particles were missed during ATR-FTIR imaging measurement (e.g., three particles with asterisk notation in Fig. 4B). This sort of modification made during ATR-FTIR measurements can be easily identified by comparing two SEIs collected before and after ATR-FTIR imaging measurements. As the locations and sizes of all the particles in the image field are known from the SEIs, their ATR-FTIR spectra can be extracted easily from raw ATR-FTIR image data.

The same type of low-Z particle EPMA and ATR-FTIR imaging experiments was carried out for the standard talc and tremolite samples, where 8 and 16 particles were analyzed, respectively. The most significant region of typical ATR-FTIR spectra (ranging from 700 cm^-1 to 1,200 cm^-1) for single particles of tremolite, anthophyllite, and talc is shown in Fig. 5, where spectra for standard bulk samples are also shown as an inset. In case of tremolite, the spectral pattern with various significant differences in peak positions is sufficient to distinguish from individual anthophyllite and talc particles. The pattern of the ATR-FTIR spectra obtained from ATR-FTIR imaging measurement are somewhat similar both for the individual particles of anthophyllite and talc (Fig. 5). A very sharp and intense peak for Si-O stretching vibration is observed almost in the similar region, i.e., 1,008-1,012 cm^-1, for the individual particles of anthophyllite and talc. However, some significant differences are present in the ATR-FTIR spectra of anthophyllite and talc minerals, i.e., a weak Si-O stretching peak observed at ~1,088 cm^-1 and ~1,056 cm^-1 region for anthophyllite and talc, respectively. An additional ATR-FTIR peak at ~912 cm^-1 region is for anthophyllite, which is absent in talc. For anthophyllite, an ATR-FTIR peak at ~776 cm^-1 in the silicate chain stretching vibration region is observed, which is absent in talc, indicating that individual anthophyllite and talc particles of micrometer size can be differentiated from each other based on their ATR-FTIR spectra. On the other hand, this ATR-FTIR technique fails to provide the information on morphology and elemental compositions of mineral particles, which is also necessary for detailed mineral characterization. Therefore, the combined use of EPMA and ATR-FTIR data can be used to clearly distinguish between regulated asbestiform minerals and talc on a single particle basis, which is not somewhat sufficiently carried out by employing either technique individually.

Fig. 5. 
Typical ATR-FTIR spectra of individual tremolite, anthophyllite, and talc particles. Spectra for the bulk sample of each mineral are shown as an inset. Data are shown only for the most significant spectral region of 700-1,200 cm^-1.