The laboratory of the National Gold & Diamond Testing Center (NGDTC) encountered some plagioclase (moonstone) beads with blue adularescence. Fifteen of the 22 moonstones fluoresced moderate to strong bluish white to long-wave UV, with the fluorescence visible in fissures. Electron microprobe analysis of one bead and micro-infrared reflectance spectra of all 22 samples indicated a composition nearly identical to albite. The specimens with strong fluorescence exhibited 3053 and 3038 cm–1 peaks in their direct transmission infrared spectra, confirming impregnation by a material with benzene structure. This treatment can be detected with a standard gemological microscope by observing characteristics such as relief lines.In identifying gemstones from the Chinese market over the last five years, the National Gold & Diamond Testing Center (NGDTC) found that some treatments usually applied to top-grade colored stones such as emerald (Johnson et al., 1999) or jadeite jade (Fritsch et al., 1992) had also been used to enhance other materials. Impregnated aquamarine, tourmaline, and kyanite have all been encountered. Li et al. (2008) examined the characteristics and identifying features of filled aquamarine. Wang and Yang (2008) reported on a filling technology applied to carvings, beads, and faceted gems from the jewelry market of Guangdong Province. They also researched the identification of these filled gemstones.
A few months ago, the NGDTC laboratory received from a client six bracelets of plagioclase (moonstone) beads with blue adularescence. The bracelets were reportedly from Donghai County in Jiangsu Province, the trading center for crystal quartz. They displayed moderate to strong bluish white fluorescence in an irregular curvilinear pattern, which caused suspicion. Observing the beads from different directions showed that the fluorescence was confined to the fractures, and the authors deduced the presence of some foreign material. In addition to determining the mineral composition of the samples, we collected infrared spectra to confirm the existence of the filling material and examine its composition.
The samples came from a strand of 22 moonstone beads (table 1) that showed beautiful blue adularescence (figure 1). We examined the samples’ standard gemological properties using an Abbe refractometer, an ultraviolet fluorescence lamp, and a microscope.
MATERIALS AND METHODS
The chemical composition of sample 1 was first determined by electron microprobe analysis at the Chinese Academy of Geological Sciences (CAGS). The sample was removed from the strand, and a flat surface was polished oblique to the lamellae of polysynthetic twinning. After the electron microprobe analysis we could still see the strongest adularescence of this sample and collect its infrared spectra for further tests. CAGS used a JXA-8230 electron microprobe with an accelerating voltage of 15 kV, a beam current of 20 nA, and a beam diameter of 5 µm. Jadeite was used as the Na standard, and Na was run before the other elements to avoid undercounting sodium. The standard materials for this test were natural minerals and synthetic oxides, and the detection limit was about 100 ppm.
The 22 samples, including sample 1, were also tested at NGDTC with a Nicolet Nexus 470 Fourier transfer infrared spectrometer. To collect the microscopic reflective infrared spectra, we used an MCT/B detector. A total of 32 sample scans were taken at a resolution of 8.0 cm–1 and a background gain of 4.0. The Omnic 6.1a software recommends a scanning wavenumber range of 4000–650 cm–1, and the infrared spectrometer extended that range to 7800–400 cm–1.
Given the test requirements of the functional group (4000–2000 cm–1) and the fingerprint region of silicate minerals in reflective infrared spectroscopy, the scanning wavenumber range was set at 1300–500 cm–1.
Thompson and Wadsworth (1957) used infrared spectroscopy to determine albite and anorthite proportions in plagioclase. Li Jianjun et al. (2007) showed that the infrared spectra will vary when the samples are tested in different orientations. Thus the authors sought to obtain infrared spectra from a consistent crystal orientation to determine whether the samples had the same composition. We used a simple orientation method: With the light source directed from the viewpoint, we looked for the area where the blue adularescence was the strongest and recorded the micro-infrared reflective spectra of each sample from the same orientation. Because the chemical composition of sample 1 was determined by both EPMA and microscopic reflective infrared spectroscopy, comparing the spectra of all other samples to that of sample 1 allowed us to determine whether they had the same composition.
Direct transmission was then applied to each whole bead to test the existence of the filling material using a DTGS KBr detector. A total of 32 sample scans were taken at a resolution of 8.0 cm–1, a background gain of 1.0, and a scanning range of 7000–400 cm–1. With air as the background, we collected the spectra of infrared rays through each whole bead.
RESULTS AND DISCUSSIONGemological Properties. The samples’ spot refractive index (RI) was approximately 1.53. The RI of the polished surface on sample 1 was 1.530–1.535. Because each sample contained a hole for stringing, specific gravity was not measured due to the possible complication caused by the holes. Most samples fluoresced weak to moderate blue-white to long-wave UV (table 1; figure 2 left). Only one sample was inert to long-wave, while three displayed strong fluorescence. Under short-wave UV their fluorescence was weaker or inert (figure 2, right). Because the fluorescence was visible along the fissures, we deduced that there might be some foreign material within them. Large fissures would contain more foreign substance, producing stronger fluorescence while the beads with no fissures were inert under UV fluorescence lamp.
Microscopic observation with brightfield illumination revealed a fine, closely woven needle-shaped schistose structure (or inclusions) in all samples (figure 3), while parallel twin layers were visible from certain directions (figure 4). There was a clear relationship between twinning planes and adularescence intensity: Adularescence was the strongest when the lighting and viewing directions were approximately perpendicular to the twinning planes. To keep a constant viewing direction, we collected micro-infrared reflectance spectra of all samples with the incident infrared rays perpendicular to the twinning planes. Observing these samples under the microscope with overhead illumination, we saw many veins on their surfaces, which appeared similar to the relief lines on filled aquamarine described by Li et al. (2009). Nevertheless, it was difficult to find the fractured reflective surfaces we would expect to accompany such veins; cracks were visible on the surface but barely penetrated the moonstone (figure 5). Meanwhile, an unusual residual flat high-relief area (again, see figure 5) was observed in sample 2, but not in any other moonstone. Similar high-relief areas are common microscopic features in filled aquamarine (Li et al., 2008) and thought to be products of incomplete filling. In other words, they were holes or gas bubbles.
Electron Microprobe Analysis. Complete electron microprobe data from six analytical points on sample 1 are listed in table 2. Based on the calculation method of Brandelik (2009), the three components of sample 1 are albite (Ab), orthoclase (Or), and anorthosite (An). The calculated composition of sample 1 was Ab91.01Or1.92An7.07.
Infrared Spectroscopy Analysis. As figure 6 demonstrates, the 22 samples had very similar micro-infrared reflection spectra when they were collected at the strongest iridescence area (perpendicular to the polysynthetic twinning plane). This means the samples had identical mineral composition. As figure 6 demonstrates, the 22 samples had very similar micro-infrared reflection spectra when they were collected at the strongest iridescence area (perpendicular to the polysynthetic twinning plane). This means the samples had identical mineral composition.
1200–900 cm–1: This region shows the Si-O stretching vibration bands in SiO4 tetrahedral polymers (Zhang et al., 1986). The 22 samples generally shared the same peaks or shoulders: 1187, 1040, and 1007 cm–1 peaks; a 1140 cm–1 shoulder; and a shoulder developing to a peak in the 1076 cm–1 region.
800–700 cm–1: This region shows the Si-O bending vibration bands in SiO4 tetrahedral polymers, as well as the Al-O stretching vibration bands in polyhedral polymers (Zhang et al., 1986). There were four peaks in all 22 samples.
Below 700 cm–1: These are the stretching vibration bands of Al-O (and/or Si-O) and the bending vibration bands of O-Si-O (and/or O-Al-O), producing sharp peaks at 652 and 589 cm–1 and the shoulders between them (Zhang et al., 1986).
As figure 7 shows, the direct transmission infrared spectra of the beads with moderate or strong fluorescence collected from three orthogonal directions presented absorption peaks at 3053 and 3038 cm–1, which is due to the cumulative frequency involved in the stretching vibration of C-H in benzene and the bending vibration of the benzene ring (Johnson et al., 1999a,b). The 4344 cm–1 peak was due to the combined frequencies of the stretching and bending vibrations of C-H in CH3 and CH2 (Zhang et al., 1999), but the peak at about 4065 cm–1 was associated with the combined frequencies of the stretching vibrations of C-H and C-C bands from organic material. Interestingly, an earlier study of filled jadeite jade found a 4060 cm–1 absorption peak, confirming the filler material as epoxy or a similar substance (Zhang et al., 1999). Meanwhile, the infrared spectra of untreated moonstones from NGDTC’s database showed no peaks at 4344, 4065, 3053, or 3038 cm–1 (again, see figure 7).
Most of the beads showed absorption peaks at 2969, 2927, and 2869 cm–1, associated with the stretching vibration of CH2. Three strong absorption peaks at 2962, 2926, and 2872 cm–1 were frequently found by Johnson et al. (1999a) in a study of emerald filled by epoxy. The untreated moonstones did not present these three peaks (again, see figure 7).
From the above tests, we confirmed that all the beads were filled by a material with the structure of benzene.
There was a clear difference in the 2927–2869 cm–1 range between the strongly and weakly fluorescent samples. The strongly fluorescent moonstone had a strong absorption band, and the weakly fluorescent samples showed weak absorption (figure 8). This suggests that samples with stronger fluorescence contain more filling.