Role of impurities in gypsum–bassanite phase transition: A comparative Raman study

P. S. R. Prasad, N. Ravikumar,
A. S. R. Krishnamurthy* and L. P. Sarma

Mineral Physics Group, National Geophysical Research Institute, Hyderabad 500 007, India

*Indian Institute of Chemical Technology, Hyderabad 500 007, India

A comparative study on two naturally-occurring gypsum crystals has been carried out using Raman spectroscopy, in the temperature range 300–550 K, to understand the effect of alkali and alkaline halide impurities such as Na and Ba, on its transition sequence. New Raman modes emerged at around 388   5 K, at 1026, 1032 cm–1, respectively, indicating the onset of bassanite phase in both the samples in which the relative concentration of sodium impurities varies by about 4%. The role of impurities is also reflected in the thermal evolution of the characteristic n 1 (SO4) mode at 1008 cm–1 during the gypsum–bassanite transition.

THE dynamics of water molecules in hydrous minerals assume significance as they are responsible for the
interactions among the other constituents of the host matrix (mineral)1. A detailed knowledge of these interactions is essential in unravelling the crystallographic arrangement and order–disorder phenomena leading to the structural transitions. The dehydration process, associated with or without a structural change also produces variations in the thermodynamic properties, a knowledge of which is essential in delineating the geological processes of the earth’s crust2,3. The vibrational spectroscopic (Raman and IR) techniques are versatile in probing the structural changes and dehydration in minerals. The spectral variations of some soft and hard modes, with temperature and/or pressure, provide a definite information about the structural variation in minerals4,5.

Gypsum, one among the simple hydrous minerals, has a transition sequence of gypsum–bassanite–anhydrite, and these phase transitions are believed to be controlled by the dehydration and rehydration processes3. In the natural samples it is known that the gypsum coexists with anhydrite along with minor amounts of alkali and alkaline halide impurities6. It is known that the equilibrium temperature of gypsum–anhydrite phase is lowered due to the addition of impurities like Na2SO4 and H2SO4 (ref. 6). Furthermore it has also been reported that the presence of alkali (Na) and/or alkaline (Ba) ions prevent the free movement of water molecules of the gypsum7. Thus impurities could play an active role in the activation and deactivation of water molecules during the phase changes and thus, the onset temperature of gypsum–bassanite depends on the impurities. The water released due to dehydration of gypsum may result in the local solution, redistribution and deposition of the soluble salts such as halite and anhydrite of the surrounding evaporites8. The water molecules in gypsum are highly interactive with CaO and SO4 groups and are also essential constituents in its crystallographic arrangement7. A more detailed crystallographic arrangement proposed by Pederson and Semmingsen9 revealed that, two sheets of sulphate (SO4 tetrahedra) ions are closely bonded by calcium ions so as to form a stronger double sheet. The water molecules hold these sulphate double sheets by sharing two of its atoms with calcium in an octahedral coordination. In this arrangement the sulphate ions are distorted from (free) tetrahedral form, which is also evident from observed factor group splitting in vibrational modes of SO4 group. The dehydration process thus perturbs the crystal structure. In other words, the phase transitions are associated with a structural change. Different phases in this transition sequence, namely gypsum, bassanite and anhydrite have been characterized by a variety of experimental techniques like, X-ray diffraction7, inelastic neutron scattering10, neutron diffraction9, infrared11,12 and Raman spectroscopy13,14. A commonality in all these studies has been mainly to characterize the structure of individual phases and to understand the dehydration process. Further, the existence of an intermediate bassanite phase with H2  0.5, has been investigated in some of the studies7,15. The essence of all these studies is that stepwise dehydration of gypsum, by 1.5 and 2.0 H2O, is responsible for the formation of bassanite and anhydrite phases respectively. It has already been established in our earlier studies16, that gypsum to bassanite transition is associated with a rearrangement of sulphate ions. However, such changes are marginal during bassanite–
anhydrite transition. We have undertaken a systematic in situ Raman spectroscopic study to understand the structural phase transition sequence with the help of some thermosensitive modes. In our recent Raman spectroscopic investigation on a natural gypsum, the phase transition sequence has been studied using the thermal evolution of spectral parameters of sulphate ions16. In this paper we report some of our results on the transition sequence in two naturally-occurring gypsum samples with different percentages of alkali (Na) and alkaline (Ba) impurities.

A small crystal, 4   4   2 mm3, is cut from a large transparent crystal of gypsum, collected from an unknown geological location in India and obtained from, M/s Hindustan Minerals and Natural History Specimen Supply Co., Calcutta. The cut crystal is thoroughly polished before the Raman spectroscopic study.

Raman spectra has been recorded on a SOPRA, DMDP2000, double-monochromator in a single pass mode. Scattered radiation is captured in usual 90 geometry and has been analysed in photon counting mode with an air-cooled photomultiplier tube from Hamamatsu, Japan. Maximum scanning range for this monochromater is 1700 cm–1, and thus we have chosen some thermosensitive mode that falls within the spectral range of monochromator. Excitation source has been 514.5 nm radiation (80 mW), from a SPECTRA PHYSICS 2020, 5W argon ion laser. An in-house fabricated variable temperature cell, which could go up to 700 K, with an accuracy better than  5 K, is used for the Raman studies. More experimental details are given elsewhere16. The sample is irradiated along its x-axis and the scattered radiation is collected along y-axis and thus one expects to observe the Raman vibrations that are active in a zz and a zx geometries.

The samples used in this investigation have been characterized using vibrational spectroscopy (IR and Raman), X-ray fluorescence (XRF) and powdered X-ray diffraction (XRD) techniques. Hereafter we abbreviate the two samples as I and II for simplicity. The XRF studies indicated nearly equal amount of total alkali (Na) and alkaline (Ba) impurities in both the samples. However, the barium to sodium impurity ratio is lower by about 10% and the relative concentration of sodium is higher by about 4% in II than in I. The XRD studies resulted in identical spectra and the fitted cell
parameters are a = 5.68; b = 15.18; c = 6.51  and b  = 118.4 with Z = 4. The space group is I2/a (C62h) which is in agreement with that of Pedersen and

1411.gif (9548 bytes)


Figure 1.  Infrared spectra of gypsum I and II recorded at ambient conditions.

1412.gif (23208 bytes)

Figure 2.  Thermal evolution of the Raman spectra (a) and (b) in n 1 (SO4) (symmetric stretching) mode region of gypsum I and II respectively.

Figure 3.  a, Observed peak positions of the Raman mode in n 1 (SO4) mode region at various temperatures. b, Thermal-induced variations in the normalized peak intensity of the Raman modes at 1008 cm–1 in gypsum I. Filled and open squares indicate the spectral variations of Raman bands at 1008 and 1026 cm–1, respectively.

Semmingsen9. The infrared spectra of I and II in the range 400 to 4000 cm–1 has been shown in Figure 1. It is evident that the IR spectra of I and II are identical in this range. All the infrared modes of water were observed in our study at 3406, 3495 (n 1), 1622, 1688 (n 2) and 3555 cm–1 (n 3). The same for sulphate tetrahedra were at 1114, 1151 (n 3) and 601, 668 cm–1 (n 4). Additionally the higher order modes observed in our case are also comparable with previous reports11,12,15. It is thus clear that there is no observable shift or appearance of new bands in I and II, indicating that the crystallographic arrangement in gypsum is broadly the same. However, the temperature dependence of the Raman modes in n 1 (SO4) region for I and II, is vastly different. In the following paragraphs we describe these variations.

A systematic thermal evolution of the Raman modes in the region 1060–980 cm–1 is shown in Figure 2 for I and II. At ambient temperature a characteristic symmetric stretching Raman band is observed at 1008 cm–1 in both the samples. On increasing the temperature this mode in I is systematically decreased and downshifted by about 3 cm–1 as shown in Figure 3. Around 388   5 K, a new Raman mode emerged at 1026 cm–1, and its reduced intensity showed an Arrhenius type increase in the temperature range 388–420 K, after which it started diminishing16. From this thermal evolution, the activation energy associated with the two transitions namely gypsum–bassanite, and bassanite–anhydrite
has been estimated and found to be 92.25 and 32.94 kJ mol–1, respectively16.

However, the spectral variations of the modes in this region for sample I and II are different. Observed spectral variations in the peak shift and intensity of n 1 (SO4) mode for sample II have been plotted in Figure 4. In sample II, as is seen from Figure 2 b and Figure 4 b, the intensity of the mode at 1008 cm–1 is almost unchanged up to 381   5 K and after which it sharply decreased. From Figure 4 a, it is clear that this mode showed an upward shift with temperature and a new mode also appeared at around 1032 cm–1 around 381   5 K. All these observations broadly corroborate the phase change of gypsum to bassanite. However, distinctly different thermal evolution of spectral parameters of sulphate mode in gypsum indicate the variations in crystallographic environment at higher temperature in I and II, though the spectra of starting materials are identical.

1413.gif (10917 bytes)

Figure 4.  Thermal-induced variations in (a) peak position and (b) normalized reduced intensity of the Raman mode at 1008 cm–1 in gypsum II. Filled and open symbols represent the spectral variations for the two Raman bands.

A closer look into the atomic arrangement has revealed that two sheets of sulphate ions are closely bonded together by calcium ions so as to form a stronger double sheet. The water molecules hold these sulphate double sheets by sharing two of its oxygen atoms with calcium in an octahedral coordination9. In this crystallographic arrangement the sulphate ions are distorted from its (free) tetrahedral form. In natural gypsum, co-existence of anhydrite as secondary mineral along with other halide impurities is well known8,9. In the bassanite phase, gypsum loses its one and half molecules of crystalline water15. It is also known that the released water, because of the phase change, localize into the layers and then extincts from the crystal structure3,6. The larger cationic impurities (like Ba and Sr) may block the layers which would result in retaining the water molecules in the channels7,17,18 and thus activating the halite impurities to interact with the primary mineral of gypsum, leading to a change in the onset temperature for the phase transition6. In other words, the released water stays in the channels of gypsum and could interact with the impurities like NaCl and secondary mineral, anhydrite to form a local solution. In either case, it might show spectral variations in the water modes. On the other hand, the presence of soluble impurities such as NaCl and CaSO4, could result in the local distortion of SO4 tetrahedra. Recent studies of Toulkeridis et al.17, indicate that the presence of elements like Sr, Gd, Yb, etc. have a profound influence on the enrichment of S isotope and on the formation of gypsum itself in nature. Earlier studies on some of the gypsum samples by Kushnir18, have reported that the amount of Sr ion decreases the transition temperature of gypsum to bassanite. While other ions like Mg and K have a minor role on the transition. Further, Lager et al.7, reported that motion of water molecules in the channels of gypsum is restricted by the presence of ions like Ba. Thus a detailed spectroscopic investigation on gypsum with differing impurities is essential to understand their role and mechanism on the transition sequence. However, we are not aware of any such detailed studies especially on natural samples. We plan to undertake a systematic vibrational spectroscopic study to understand the role of impurities on phase transition sequence with more natural samples collected from different geological formations. Our present results on two gypsum samples which are collected from different geological locations that vary in alkali and alkaline impurities, indicate that the evolution of Raman mode at 1032 cm–1 in sample II, is higher by 6 cm–1 compared to sample I, and this could tentatively be attributed to the larger distortions of SO4 tetrahedra.

In summary, it is clear from our Raman spectroscopic studies that the temperature evolution of the characteristic gypsum mode at about 1008 cm–1 is vastly different in two natural samples that vary in halite impurities by about 4%. Onset of the bassanite phase in these samples is evident from the emergence of new Raman modes at 1026 and 1032 cm–1, respectively, around 388   5 K. However, a small variation in the onset temperature has been observed in these samples. In order to get more insight into the mechanism, a further systematic study on several samples with varying alkali impurities need to be undertaken.


  1. Winkler, B., Phys. Chem. Miner., 1996, 23, 310–318.
  2. Newton, R. C., Annu. Rev. Earth Planet Sci., 1989, 17, 385–412.
  3. Popp, T. and Kern, H., Earth Planet Sci. Lett., 1993, 120, 43–57.
  4. Gillet, P., Phys. Chem. Miner., 1996, 23, 263–275.
  5. Prasad, P. S. R., J. Phys. Chem., 1996, 100, 888–890.
  6. Hardie, L. A., Am. Miner., 1967, 52, 171–200.
  7. Lager, G. A., Armbruster, T., Rotella, F. J., Jorgensen, J. D. and Hinks, D. G., Am. Miner., 1984, 69, 910–918.
  8. Deer, W. A., Howie, R. A. and Zussman, J., An Introduction to the Rock forming Minerals, Longman Scientific and Technical, 1992, pp. 612–616.
  9. Pedersen, B. F. and Semmingsen, D., Acta Cryst., 1982, B38, 1074–1077.
  10. Winkler, B. and Hennion, B., Phys. Chem. Miner., 1994, 21, 539–545.
  11. Seidl, V., Knop, O. and Falk, M., J. Can. Chem. Soc., 1969, 47, 1361–1368.
  12. Bensted, J. and Prakash, S., Nature, 1968, 219, 60–61.
  13. Krishnamurthy, N. and Soots, V., Can. J. Phys., 1971, 49, 885–896.
  14. Berenblut, B. J., Dawson, P. and Wilkinson, G. R., Spectrochim Acta., 1973, 29, 29–36.
  15. Putnis, A., Winkler, B. and Fernandez-Diaz, L., Min. Mag., 1990, 54, 123–128.
  16. Sarma, L. P., Prasad, P. S. R. and Ravikumar, N., J. Raman Spectrosc., 1998, 29, 851–856.
  18. Toulkeridis, T., Podwojewski, P. and Clauer, N., Chem. Geol., 1998, 145, 61–71.
  19. Kushnir, J., Geochim. Cosmochim. Acta, 1982, 46, 433–446.

ACKNOWLEDGEMENTS.  We thank Dr H. K. Gupta, Director, NGRI for his encouragement and keen interest in this program, and for permission to publish the paper. Thanks are due to Dr T. N. Gowd for his coordination in carrying out this work. We sincerely acknowledge our colleagues Dr P. K. Govil and Dr K. Ravi Kumar for extending their experimental facilities.

Received 21 April 1998; revised accepted 22 September 1998