CLASSIFICATION OF WATER TYPES IN STRUCTURAL CHANNEL OF BERYL BY MEANS OF
RAMAN SPECTROSCOPY

LÊ THỊ THU HƯƠNG^1,2, WOLFGANG HOFMEISTER ²

¹Faculty of Geology, Hà Nội University of Sciences, Hà Nội, Việt Nam
²Centre of Gemstone Research, Johannes Gutenberg University, D-55099 Mainz, Germany

Abstract: Two types of water molecules in the structural channel of beryl are classified by Raman spectroscopy under room temperature. The type I is that occuring alone and with a typical orientation in the way that the symmetry axis of water molecules is perpendicular to the c axis of the emerald crystal; the type II is water molecules which are associated with nearby alkalis and the water molecule symmetry axis is parallel to c axis of the host crystal as the result of interaction with alkali ions. Under low temperatures (198-123^o K) the splitting of the band of water type II is present and possible reasons for the band splitting are discussed.

I. INTRODUCTION

The types of water molecules in structural channels of beryl (Be₃Al₂Si₆O₁₈) have been classified by infrared spectroscopic studies in mixed powdery beryl and KBr pellets since more than 40 years ago [6, 7]. This article brings out the possibility of classifying water types by non-destructive Raman spectroscopy. As related to alkali amount, water molecules exist differently in type and “density” in beryl of different origins that can be detected by Raman spectroscopy, and since the method does not require taking material from samples it could be very useful in the determination of gem origin.

According to Wood & Nassau [6, 7], the water of type I is that which occurs alone and with a typical orientation in the way that the symmetry axis of the water molecule is perpendicular to the c axis of the emerald crystal; the type II is a water molecule which is associated with nearby alkalis, and the water molecule symmetry axis is parallel to c axis of the host crystal as the result of interaction with alkali ions. In their studies concerning the distinction of natural and synthetic emeralds, Schmetzer [4] and Schmetzer and Kiefert [5] supplemented that water molecules can either exist alone or with nearby alkalis (mainly sodium). In the case, that there is a nearby alkali ion, the water type II may be associated with this ion in two sequences: H₂O-Na-OH₂ or H₂O-Na-, in which  represents the vacancy of water (no water) in the channel sites. Therefore, they divided the water type II into two subtypes: the subtype IIa is water molecules like H₂O-Na-OH₂, and the subtype IIb is water molecules like H₂O-Na-. Furthermore, they found that also hydroxyl ions (OH^-) can be associated with alkali ions in a sequence like HO-Na-. The study of Aurisicchio et al. [1] confirmed the existence of two types of water and also suggested the presence of OH groups associated with alkali ions in the channels (Figure 1). According to Aurisicchio et al. (1994) and Brown and Milis [2] Na⁺ together with other alkali ions such as Cs⁺, Rb⁺ and K⁺ can occupy the channel positions of the beryl structure.

II. MATERIAL AND EXPERIMENTAL METHODS

Single crystals of natural emeralds from Brazil (Carnaiba, Capoeirana, Itabira, Santa Terezinha, Socoto), Colombia (Chivor), Austria (Habachtal), Russia (Ural), Madagascar (Mananjary), South Africa (Transvaal), Zambia (Kafubu), Nigeria (Gwantu), China (Malipo) and faceted synthetic “emeralds” of Tairus, Biron (hydrothermally-grown), Gilson, Chatham, Lennix (flux-grown) producers were studied.

The presented Raman spectra were done in the confocal measuring mode by the Jobin Yvon (Horiba group) LabRam HR 800 spectrometer. The system was equipped with an Olympus BX41 optical microscope and a Si-based CCD (charge-coupled device) detector. The spectra were excited by the Ar⁺ ion laser emission of 514 nm as a green laser with a grating of 1800 grooves/mm. Due to these parameters and the optical path length of the spectrometer, a resolution of 0.8 cm^-1 was achieved. The spectra acquisition time was a set of 240 seconds for all measurements. The range of water was measured from 3500 cm^-1 to 3700 cm^-1 under room to low temperature (from 300 K to 78 K) in steps of 25 K. The peak analysis was performed with an Origin-lab 7.5 professional software package. The single or overlapping peaks were fitted using Gauss-Lorentz function.

The alkali concentrations were analyzed by LA-ICP-MS. The ablation was achieved with a New Wave Research UP-213 Nd:YAG laser ablation system, using a pulse repetition rate of 10 Hz and 100 µm crater diameters. Analyses were performed on an Agilent 7500 ce inductively coupled plasma-mass spectrometer in pulse counting mode (one point per peak and 10 ms dwelling time). Data reduction was carried out by the software “Glitter”. The amount of material ablated in laser sampling is different for each spot analysis. Consequently, the detection limits are different for each spot and are calculated for each individual acquisition. The detection limits generally range between 0.001 and 0.5 ppm (µg/g). ²⁸Si was used as the internal standard. Analyses were calibrated against the silicate glass reference material NIST 612 using the values of Pearce et al. [3], and the US Geological Survey (USGS) glass standard BCR-2G was measured to monitor accuracy.

III. RESULTS AND DISCUSSION

1. Room temperature measurements

Under room temperature, all samples of this investigation were measured for obtaining the spectra in the range of OH^- and water molecule vibration (from 3500 to 3700 cm^-1). All measurements were conducted with E normally to the c axis. Two bands may be seen in this range (although there is always not equal among all samples): one at about 3608 cm^-1 and the other at 3598 cm^-1 (Figure 2). All flux synthetic “emeralds” do not show any Raman bands in this range differing from the signal of luminescence. This elucidates the fact that, there is no water in flux grown “emeralds”. Hydrothermally grown synthetic “emeralds” show one band at 3608 cm^-1.

Considering chemical data, it is obvious that the appearance of the band at 3598 cm^-1 and also the intensity ratio of the two bands (3598 and 3608 cm^-1) depend on the amount of alkali ions. The band 3598 cm^-1 is detectable only in alkali-bearing samples; in alkali-free samples (hydrothermal synthesis) this band absolutely disappears while the band 3608 cm^-1 truly exists. This band increases in intensity as the alkali content of emerald increases, and the more alkali ions are present in sample, the higher is the ratio I₃₅₉₈/I₃₆₀₈.

In those samples with the amount of alkali ions higher than 1,1 wt% (emeralds from Brazil, Russia, Austria, Madagascar, Zambia, South Africa) the intensity of the band 3598 cm^-1 is higher than that of the band 3608 cm^-1 (I₃₅₉₈/I₃₆₀₈ > 1), and on the contrary; in those samples with the amount of alkali ions lower than 1.1 wt% (emeralds from Colombia, Nigeria, China and synthetic “emeralds”) the intensity of the band 3608 cm^-1 is higher than that of the band at 3598 cm^-1 (I₃₅₉₈/I₃₆₀₈ < 1) Figure 3). According to this observation and to the classification of Wood & Nassau [6, 7], it can be stated here, that the Raman band at 3608 cm^-1 is generated by the water type-I (those water molecules having not any presence of nearby alkali ions) and the Raman band at 3598 cm^-1 is generated by water type-II (water molecules with nearby alkali ions).

In all natural stones, the band at 3598 cm^-1 (water type II) shows to be much broader than the band at 3608 cm^-1 (water type I). The FWHM values of band 3598 cm^-1 range from 11.2 to 14.8 cm^-1 while those of band 3608 cm^-1 (water type I) range from 1.6 cm^-1 to 2.8 cm^-1. The reason that the Raman band of water type II is broader than that of type I may be due to the vibrational characteristics of each water type itself. This means, since there is the appearance of nearby alkali ions (in cases of water type II), the orientation of water molecules changes; therefore, the vibration energy and/or state of vibration of water molecules actually changes. This explains why the water type-II molecules generate the Raman band at different positions and have different FWHMs than the water type-I molecules. Nevertheless, it may be also the case that, the broadening of the band 3598 cm^-1 is caused by the combination of more than one band, since its width is not normal for a single typical Raman band. According to the study of Schmetzer & Kiefert [5], the water type II can be subdivided into the subtype IIa (H₂O-Na-OH₂) and subtype IIb (H₂O-Na-ÿ). Therefore, it could be supposed here, that the band of 3598 cm^-1 is not a single band, but in fact a combination of at least two bands, one due to the type-IIa water and the other due to the type-IIb.

2. Low temperature measurements

To investigate the behaviour of this broad band, the Raman measurements were carried out under different temperature levels. One sample from Madagascar (with high alkali content) was chosen for low temperature experiment; the position of the sample was kept constant (in the way that the crystallographic c axis is normal to the vibration of the laser beam) during measurement. Figure 4 shows 10 spectra measured from 300 K down to 78 K, i.e. 27 ^oC to -195^oC.

From 300 to 223 K the spectra show to maintain 2 peaks, although the FWHM values and their position changes under different temperatures. In addition, the lower the temperature is, the closer the two peaks come together (from 11.6 cm^-1 when measured under room temperature, to 10.3 cm^-1 when measured at 223 K). The FWHM of the band 3596 cm^-1 increases, while that of band 3608 cm^-1 decreases. From 198 to 123 K, the spectra have clearly 3 peaks (Figure 5); only lower than 198 K the splitting of the band of the water type II are present. Below 98 K, again only two peaks are detectable, one at 3593 cm^-1 with lower intensity and another at 3604 cm^-1 with higher intensity.

Figures 6 and 7 illustrate the diagrams showing the appearance of the additional peak between 123 and 198 K, and the change of band positions and FWHMs of two types of water bands according to the change of temperature. There are some proposals for the following appearance of the additional peak. However, it has to underline, that the presented proposals need further experimental confirmation. We need a strictly accurate polarizer filter 1. Following the study of Schmetzer & Kiefert [5], in which 3 types of water have been assigned, the type I, type IIa and type IIb by IR spectroscopy, the additional peak is seen in Raman spectroscopy, therefore it could be assigned for the third type of water. And, the peak which is detected at room temperature at 3598 cm^-1, is the overlap of two water types IIa and IIb. At low temperatures, the silicate structure of emerald is slightly constricted. Accordingly, the deformation or reorientation of the water type II molecules may take place (remember that the water type II molecules are oriented in such a way that their two-fold symmetry axis is parallel to the crystallographic six-fold axis of the emerald structure). This means, the parallel position changes slightly to the diagonal position. This new position generates a new energy which is recordable as an additional peak in the spectrum.

IV. CONCLUSIONS

In this article, the types of water have been studied and classified by non-destructive Raman spectroscopy. The type I are those molecules which occur without nearby alkali, and the type II are those which are associated with nearby alkali ions. The singular broadening of the band at 3598 cm^-1 (water type II) and the splitting of this band under low temperatures between 198 and 123 K suggest that, this band may be the overlap of two subtypes of water (IIa, IIb) in the sequences like H₂O-Na-OH₂ and HO-Na-. It needs nevertheless further study to confirm if the splitting of the band of water type II under low temperatures is caused by different sequences or simply by the deformation and reorientation of the water type II, as the result of a constriction of the channel structure.

Acknowledgments: This work has been financed by the Johannes Gutenberg - University Research Fund for Gemstone Research. Analytical facilities have been provided by the Institute of Geology at Johannes Gutenberg University (Mainz, Germany). The authors are grateful for these supports.

REFERENCES

1. Aurisicchio C., Grubessi O., Zecchini P., 1994. Infrared spectroscopy and crystal chemistry of the beryl group. The Canad. Mineralogist, 32 : 55-68.

2. Brown G.E., Milis B., 1986. High-temprature structure and crystal chemistry of hydrous alkali-rich beryl from the Harding pegmatite, Taos County, New Mexico. Amer. Mineralogist, 71 : 547-556.

3. Pearce N.J.G., Perkins W.T., Westgate J.A., Gorton M.P., Jackson S.E., Neal C.R., Chenery S.P., 1997. A compilation of new and published major and trace element data for NIST SRM 610 and NIST SRM 612 glass reference materials. Geostandards Newsl., 21 : 115-144.

4. Schmetzer K., 1989. Types of water in natural and synthetic emerald. Neues Jahrb. für Miner. Monatshefte, p. 15-26.

5. Schmetzer K., Kiefert L. 1990. Water in beryl: A contribution to the separability of natural and synthetic emeralds by infrared spectroscopy. J. of Gemmology, 22/4 : 215-223.

6. Wood D.L. and Nassau K., 1967. Infrared spectra of foreign molecules in beryl. The J. of Chem. Physics, 47/ 7 : 2220-2228.

7. Wood D.L. and Nassau K., 1968. The characterization of beryl and emerald by visible and infrared absorption spectroscopy. Amer. Miner., 53 : 777-800.