Spectrochimica Acta Part A 78 (2011) 136–141
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Raman spectroscopic study of mixed carbonate materials W. Kaabar a,b,∗ , S. Bott a , R. Devonshire a a b
High Temperature Science Laboratories, Department of Chemistry, University of Shefﬁeld, Shefﬁeld S3 7HF, UK Département de Chimie, Faculté des Sciences Exactes, Université Mentouri Constantine, Constantine 25000, Algeria
a r t i c l e
i n f o
Article history: Received 16 May 2010 Received in revised form 4 August 2010 Accepted 8 September 2010 Keywords: Raman microscopy Carbonates Calcium Strontium Barium
a b s t r a c t The main parameters for precipitation of mixed carbonate materials have been studied by Raman microscopy. These carbonates are compounds of barium, strontium and calcium. It has been shown that the Raman spectrum of a sample is exclusively controlled by its composition, the precipitation parameters do not affect the crystal structure. Even at relatively low levels, the calcium content of a sample can dominate the vibrational frequencies as measured by Raman spectroscopy. Calcium contents greater than 17% show this effect to a considerable degree, and give the broadest or two Raman peaks and thus the least uniform unit cells. The analysis of the lattice modes demonstrates that each Raman shift observed for a mixed carbonate sample corresponds to a speciﬁc crystal structure. Some peaks lie within two or three shifts that are observed for different crystal structures. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Double and triple mixed carbonates are specially prepared emitter materials used for efﬁcient thermal electron emission of cathode structures. In the current triple oxide cathode manufacturing process, a matrix of barium, strontium and calcium carbonate layer is either painted or sprayed on a metal cathode base . Emission from an oxide cathode is dependent upon the ratio of alkaline earth carbonates present at the cathodes surface. The method by which each carbonate is made plays a key role in its function in a cathode. It has been found that the crystalline structure of the alkaline earth carbonates affects the life of oxide cathode tubes . The study presented here will serve to characterise the carbonate precursor materials using Raman microscopy. There have been numerous spectroscopic investigations of pure carbonates. Amongst the earlier experimental investigations are those of Couture  on aragonite, strontianite and witherite. In view of their isomorphism a comparative study of the Raman spectra of those three simple carbonates was given by Krishnamurti . Grifﬁth  published the ﬁrst compilation of FT Raman spectra of minerals, including several carbonates. Adler and Kerr  reported the IR spectra of a dozen of simple and double carbonates and showed that differences in the spectra were a function of cation size. Later, Scheetz and White  published a detailed Raman and IR study of alkaline-earth double carbonates of natural origin. Gillet et al.  studied the behaviour of simple carbonates includ-
ing calcite and aragonite with high temperature and high pressure. They showed that the measured relative pressure and temperature shifts of the Raman lines were greater for the lattice modes than for the internal modes of the CO3 groups. Alia et al.  studied synthetic aragonite–strontianite solid solution samples using IR and FT-Raman spectroscopy; they reported that positional disorder induced by the random cationic substitution resulted in strong increase of the halfwidth in several vibrational bands. Frost et al. [10–13] have recently reported a Raman and infrared spectroscopic study of carbonate minerals from different origins. In this paper we have used Raman microscopy as a diagnostic tool for mixed carbonate materials produced with different compositions, temperatures of precipitation and concentrations of the precipitation agent. 2. Experimental A standard preparation method  is to precipitate carbonates from an aqueous solution of the alkali earth nitrates with ammonium carbonate. Raw materials of high purity (99.99%) were used throughout these studies, sourced from the Aldrich Chemical Company. The powdered carbonate material is prepared by coprecipitation from aqueous solutions in respective nitrates with ammonium carbonate according to the following reaction: (Ba, Sr, Ca)NO3(aq) + NH4 HCO3(aq) → (Ba, Sr, Ca)CO3(s) + NH4 HNO3(aq)
∗ Corresponding author at: Département de Chimie, Faculté des Sciences Exactes, Université Mentouri Constantine, Constantine 25000, Algeria. Tel.: +213699854214. E-mail address: w [email protected]
(W. Kaabar). 1386-1425/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2010.09.011
Brieﬂy, the method for precipitation of the carbonate powders is as follows: Ba, Sr, Ca nitrate weighed out in the required ratios by mass and then dissolved in 50 ml deionised water. The solu-
W. Kaabar et al. / Spectrochimica Acta Part A 78 (2011) 136–141
of-plane bend (2 ), a doubly degenerate asymmetric stretch (3 ) and a doubly degenerate bending mode (4 ). The symmetries of these modes are A1 (R) + A2 (IR) + E (R, IR) + E (R, IR) and occur at 1064, 879, 1415 and 680 cm−1 respectively. The symmetric stretching vibration is very intense in the Raman spectrum, while the asymmetric stretching is weak. The asymmetric bending although Raman-allowed is very weak. The internal modes can be related to the vibrations of the free carbonate ion, by simply comparing the band frequencies. The site symmetry of the CO3 group is determined by the cation environment and modiﬁes the selection rules (i.e. the number of bands observed), and the frequencies to a small extent, when compared with the free group. In this study the relative Raman shifts and the full width at half maximum (FWHM) of the prominent peak corresponding to the 1 band are compared for the different samples.
16 8 20
1.00 1 0.00
Ba content Fig. 1. Ternary phase diagram showing the content of mixed carbonate powders prepared with different compositions.
tion is heated to required temperature with continuous stirring. Once nitrate solution is stable at required temperature, ammonium carbonate solution is added drop-wise at an approximate rate of 2 ml/min. On completion of the precipitation, heating is turned off and sample is left to cool to room temperature with rigorous stirring. Sample is ﬁltered and collected precipitate was oven dried at 40 ◦ C for 12 h. Additional cathode components may be included such as ZrO2 , which is known to extend cathode life . From the above, it can be observed that the main parameters which will affect the precipitation process are: the ratio of alkali earth nitrates, the temperature of precipitation and the concentration of the precipitating agent. Raman spectra of the carbonate powder samples prepared are acquired with a Renishaw RM1000 series, using 514.5 nm excitation from an Ar+ laser. Rayleigh scattered light rejection is with the aid of a holographic notch ﬁlter, the cut-off from which is approximately 190 cm−1 . Spectral manipulation such as base line adjustment, smoothing and normalisation are performed using the software package GRAMS (Galactic Industries Corporation).
126.96.36.199. Pure carbonates. Fig. 2 shows the 1 fundamental in the Raman spectra of the pure carbonates. The peaks from the three samples are well separated, with centres located at 1059.77 cm−1 , 1071.28 cm−1 and 1086.09 cm−1 , for pure barium (Emitter 17), strontium (Emitter 2) and calcium (Emitter 1) carbonates respectively. Farmer  showed that the (CO3 )2− symmetric stretching band varied according to the ionic radius of the cation. The higher the ionic radius the lower the wavenumber of the symmetric stretching mode. The ionic radii of Ba2+ , Sr2+ and Ca2+ ions are 149 pm, 132 pm and 114 pm respectively. Each Raman spectrum shows a single narrow peak with FWHM values of between 3.5 and 4.5 cm−1 . The results from these pure carbonate samples can be used for comparison to the double and triple component samples prepared. 188.8.131.52. Double component samples with similar ratio. These comprise Emitters 4, 12 and 13, which are composed of 50:50 Sr:Ca, 50:50 Ba:Ca, and 50:50 Ba:Sr respectively. Raman spectra of the 1 band for each of these are given in Fig. 3. Emitter 13 gives a narrow peak, FWHM of 6.5 cm−1 , at a Raman shift between that of pure barium and strontium carbonates. This is due to the fact that the Ba2+ and Sr2+ ions are of a similar size. The exchange of these two species in a lattice structure therefore causes only minor changes to the crystal structure, and the pres1086.09
3. Results and discussion
3.1.1. Carbonate anion internal modes The free ion, CO3 2− with D3h symmetry exhibits four normal vibrational modes; a symmetric stretching vibration (1 ), an out-
3.1. Change of composition
Raman microscopy is performed on 23 carbonate powder samples of varying compositions, the temperature of precipitation is 68 ◦ C and the concentration of ammonium carbonate solution is 0.25 mol/dm3 (Table 1). The various compositions are represented as a ternary diagram of the relative amounts of the three alkali earth carbonates present in each of the samples prepared. This is given in Fig. 1. The Raman spectra of well mixed powder samples acquired show clear differences between samples of differing composition. In general, the spectra may be separated into two regions. Those bands at wavelengths above 600 cm−1 are due to the internal motions of the molecular carbonate ion. Those below 600 cm−1 are due to motions involving the entire unit-cell usually referred to as lattice modes.
Wavenumber / cm-1 Fig. 2. Raman spectra of carbonate anion symmetric stretching band for single component carbonate powder samples.
W. Kaabar et al. / Spectrochimica Acta Part A 78 (2011) 136–141
Table 1 Raman shifts and FWHM for Emitter samples with different compositions. Name
Nominal ratio Ba:Sr:Ca
Raman shift prominent peak (cm−1 )
FWHM Prominent peak (cm−1 )
Emitter 1 Emitter 2 Emitter 17 Emitter 4 Emitter 12 Emitter 13 Emitter 6 Emitter 8 Emitter 9 Emitter 5 Emitter 7 Emitter 11 Emitter 3 Emitter 10 Emitter 21 Emitter 14 Emitter 18 Emitter 19 Emitter 20 Emitter 16 Emitter TP2 Emitter 15 Emitter 22Pure Ba + ZrO2
0:0:100 0:100:0 100:0:0 0:50:50 50:0:50 50:50:0 17:17:66 55:20:25 66:5:29 17:66:17 80:10:10 66:29:5 33:33:33 66:17:17 45:45:10 50:30:20 66:0:34 66:4:30 66:15:19 66:24:10 66:30:4 66:34:0 100:0:0
1086.09 1071.28 1059.77 1084.95 1087.23 1066.74 1084.95 1083.81 1084.95 1072.42 1062.16 1064.43 1078.11/1086.09 1063.31/1080.40 1067.83 1088.33 1060.92/1087.11 1062.05/1085.97 1063.19/1085.05 1064.32 1064.21 1059.77 1059.72
4.384 4.164 3.601 21.983 20.417 6.545 15.689 20.637 22.422 6.482 9.175 8.455 7.547 9.176 8.035 Broad 7.499 9.642 6.963 11.248 7.532 4.821 3.764
ence of the two bond types result in the observation of an “average” frequency. The Ca2+ ion however is considerably smaller, and along with the different structural variations of CaCO3 , this produces different results than in the Ba:Sr case. For both Emitters 12 and 4, Raman peaks are considerably broader, FWHM of approximately 21 cm−1 , and are close to the Raman shift observed for the single component CaCO3 sample. This remarkable increase in the band width has previously been reported in the Raman spectra of double carbonates [7,9]. This suggests that the size of the Ca2+ ion play a dominant role in the vibrational frequencies observed by Raman microscopy. 184.108.40.206. Triple component samples. Table 1 summarises the Raman shift and the full width at half maximum of the 1 band for samples with three components. The effect of sample composition on the observed Raman shifts may be explained as follows: Emitters 6, 8, and 9 contain over 20% calcium, and this result in the Raman peak being centred close to the single component CaCO3 sample. The Raman spectrum acquired from Emitter 5 is centred on the position of the single component SrCO3 sample. This is due to the very high Sr content in this sample, and the fact that SrCO3 is the strongest Raman scatterer of the three carbonate materials . 1084.95
Wavenumber / cm-1 Fig. 3. Raman spectra of carbonate anion symmetric stretching band for equimolar carbonate powders.
Carbonate samples Emitter 7 and 11 show Raman shifts between those of the single component BaCO3 and SrCO3 samples. These contain cumulative amounts of Ba and Sr of 90% and 95% of the total composition, with only small proportions of Ca making up the remainder. At these low levels Ca has only a small effect on the vibrational frequencies. The Raman spectra of Emitter 3 and 10 are different to all those above in the fact that they both show two maxima. Emitter 3 has equivalent amount of all three materials. Sr is the highest Raman scatterer, but the effect of Ca on the Raman spectrum to shift the peak towards its pure position has been observed. The presence of Ba serves to add to the Sr shift which reduces the domination of the spectrum of large amounts of Ca seen for the double component samples. The spectrum therefore represents a “compromise” of the system between these effects. Emitter 10 contains a lower proportion of Ca and an increased proportion of Ba and Sr combined. The spectrum is expanded in the direction of the pure Ba position as this is now the dominant component, but the second feature as a result of the presence of Ca is still clearly visible. The variation of Raman spectrum observed for samples with similar Ba composition (66%) is a result of the variation of Sr/Ca ratio, given all other parameters were kept constant. From these results it may be conﬁrmed that the proportion of calcium content in a triple component carbonate sample dominates the Raman spectrum. Calcium contents greater than 10% show a Raman spectrum with two maxima, the prominent peak is expanded in the direction of the pure Ba position as this is the dominant component. This effect is not observed when the proportion of Ca is 10% or less. Emitter 21 (45:45:10) shows a Raman shift similar to the 50:50 Ba:Sr sample this is due to the small Ca content of 10%. Emitter 14, with more important calcium content of 20%, gives a Raman peak centred close to the single component CaCO3 sample, similar to the results for Emitter 8. In summary, lower Ba content causes the Raman shift to move to higher wavenumbers, Ca contents greater than 17% causes the Raman shift to be close to that of 100% Ca. The Raman peak is only in the region of the 100% Sr shift when Sr content is greater or equal 66%. The presence of two peaks indicates that the proportions of Ba, Ca and Sr are similar, or Ba content is 66% while Ca and Sr contents are within 5% of each other, or Ba content is 66% and Ca content is ∼30–34%.
W. Kaabar et al. / Spectrochimica Acta Part A 78 (2011) 136–141
Intensity / counts
TP1 (48ºC) TP2 (68ºC)
Emitter 23 (71ºC fast)
25000 20000 15000
135.36 0 1600
Wavenumber / cm-1 Fig. 4. Raman spectra of lattice modes for single component carbonate powder samples.
A previous study, carried out by Murabayashi [17,18] using scanning electron microscopy, demonstrated the variation in carbonate powder morphology with composition for a large range of samples. The extension of the study using Raman microscopy has provided a further analysis of composition variations. Two samples, Emitters 17 and 22 were prepared with a composition, 100% BaCO3, but 2% zirconia was added to the 22 sample prior to precipitation. Analysis by Raman microscopy (Table 1) indicates that the addition of zirconia does not inﬂuence the precipitated Barium carbonate structure, and exists in the materials as a separate inclusion. 3.1.2. Lattice modes Fig. 4 shows the Raman low frequency spectral region corresponding to the pure carbonate substances. A pair of intense Raman bands is observed for each of the cases of CaCO3 , SrCO3 and BaCO3 . These bands are at 206.54 and 154.99 cm−1 in CaCO3 , 180.33 and 148.44 cm−1 in SrCO3 and 153.20 and 135.36 cm−1 in BaCO3 . These bands match perfectly with previous results for aragonite, strontianite and witherite [4,9]. The very weak intensity bands at 509.5 cm−1 in CaCO3 and at 510.77 cm−1 in SrCO3 are not in the literature. In relation to CaCO3 , the band at 281.83 cm−1 corresponds to a calcite structure. CaCO3 can exist in different structures under similar condition, the occurrence of both aragonite and calcite is very common. Calcite/aragonite intergrowth is almost always observed when carbonation is carried out in the temperature range 30–80 ◦ C . Table 2 gives the positions of the Raman vibrations for the various Emitters. As no crystal structure information is available on any of the emitter materials, one has to make use of correlations which have been established between Raman spectral features observed in the emitters and known crystal structures of related carbonate compounds. Raman shifts characteristic of various crystal structures were from the Database of Raman spectroscopy, X-ray diffraction and chemistry of minerals website  and other references [9,21–24]. Where possible, speciﬁc crystal structures were assigned to each Raman shift observed in the Emitters. Some peaks in Table 2 lie within two shifts that are observed for two different crystal structures. It may suggest that the crystal structure in the emitter is some way between the two extremes, and
Wavenumber / cm-1
Fig. 5. Raman spectra of carbonate samples prepared at different precipitation temperatures and speed.
each Raman shift can be calculated to arise from a mixing of two or more shifts from the individual pure carbonate substances (aragonite, strontianite and witherite) in their respective ratios. For example: Emitter 10 (17.2% Ca, 16.9% Sr, 66% Ba) has a peak at 235.06 cm−1 (nearest literature peak is aragonite at 272 cm−1 , strontianite at 248 cm−1 and witherite at 222 cm−1 ): (0.172 × 272) + (0.169 × 248) + (0.66 × 222) = 235.22 cm−1 . 3.2. Temperature and speed of precipitation Three samples were prepared to determine the effect of the temperature of precipitation on the Raman spectra. The composition (4% Ca, 30% Sr, 66% Ba) and concentration of the ammonium carbonate solution (0.25 mol/dm3 ) were kept constant. Raman microscopy was performed on each of the samples, and Fig. 5 shows the resulting spectra. The rising baseline is a result of the exact focus of the laser on the sample, and so the differences in
35000 CP4: 0.129 mol/dm 30000
CP5: 0.064 mol/dm
CP1: 0.881 mol/dm
Intensity / count
CP3: 0.257 mol/dm CP2: 0.512 mol/dm
3 3 3 3 3
Wavenumber / cm-1 Fig. 6. Raman spectra of carbonate samples precipitated with ammonium carbonate solutions of differing concentration.
W. Kaabar et al. / Spectrochimica Acta Part A 78 (2011) 136–141
Table 2 Observed Raman shifts in the Emitters. Emitter % Ca % Sr % Ba
2 100.0 – –
154.99 180.39 206.54 281.83 509.55 705.56 712.43 1086.09 1435.84 1462.07 Emitter % Ca % Sr % Ba
% Ca % Sr % Ba
Arag Arag Arag Cal Arag Carb Carb Carb Carb Carb
148.44 180.33 214.52 236.20 244.74 258.31 510.77 710.13 1071.28 1445.91
33.4 33.3 33.3
50.1 50.0 –
17.0 66.1 17.0
66.8 17.1 17.0
10.1 10.2 80.0
25.0 20.0 55.2
Stron 155.36 Stron 191.77 Stron 273.76 Stron 510.78 Stron 701.08 Stron 711.25 Stron 1078.11 Carb 1086.09 Carb 1451.84 Carb –
Arag 188.09 Alsto 268.01 Arag 509.98 Stron 692.95 Carb 707.94 Carb 710.19 Carb 871.22 Carb 1084.95 Carb 1397.12 – –
29.0 5.0 66.0
17.2 16.9 66.0
5.0 28.8 66.0
50.1 – 50.1
194.19 259.98 691.88 705.54 873.23 1084.95 1325.34 – – Emitter
– 100.0 –
Alsto 164.39 Mix 235.06 Carb 690.75 Carb 704.41 Carb 1063.31 Carb 1080.40 Carb 1432.64 – – – –
Baryt 189.10 Cal III 267.05 Stron 510.84 Carb 693.03 Carb 712.37 Carb 874.18 – 1087.23 – – – –
– – 100.0 135.36 153.20 179.67 224.90 690.25 699.48 1059.77 1420.23 –
Baryt 166.76 Cal 234.39 Carb 509.59 Carb 690.75 Carb 1064.43 Carb 1433.73 Carb – – – – –
34.2 – 66.0 With With With With Carb Carb Carb Carb –
138.49 155.05 184.47 229.64 691.02 700.92 1060.92 1087.11 1424.50
Cal III Arag Mix Baryt Carb Carb Carb Carb Carb
Arag 180.09 Cal III 246.50 Arag. Stron 507.55 Carb 696.45 Carb 1072.42 Carb 1447.25 Carb – Carb – Carb – – –
Arag 189.04 Stron 266.92 Arag 508.39 Carb 702.08 Carb 712.39 Carb 874.18 – 1084.95 – – – – – –
13 – 50.1 50.0 Baryt Cal III Arag Carb Carb Carb Carb – –
142.89 171.81 237.28 509.67 693.02 706.69 710.15 1066.74 1435.42
Arag 164.46 Stron 231.60 Arag 510.20 Carb 689.61 Carb 1062.16 Carb 1386.66 Carb – – – – – – –
20.0 30.0 50.0
– 34.3 66.7
10.3 24.2 66.6
Stron 163.10 Mix 272.48 Mix 692.94 Stron 1088.33 Carb 1443.84 Carb – Carb – Carb – Carb –
Baryt 138.52 Arag 160.75 Carb 205.17 Carb 691.50 Carb 698.59 – 703.12 – 1059.77 – 1428.77 – –
29.9 4.0 66.0
19.5 15.5 66.1
10.2 45.5 45.6
139.68 156.18 184.67 230.27 690.58 1062.05 1085.97 1425.83 –
Cal III Calcite Stron Baryt Carb Carb Carb Carb –
141.36 158.46 233.65 690.58 1063.19 1085.05 1429.02 – –
Arag Cal III Baryt Carb Carb Carb Carb – –
Baryt 183.81 Cal III 257.53 Stron 691.88 Carb 707.84 Carb 874.18 Carb 1083.81 – 1448.06 – – – – – –
142.01 164.11 239.55 693.08 1067.83 1442.66 – – –
Mix 144.78 Mix 160.21 Mix 234.80 Carb 690.58 Carb 1064.32 Carb 1430.18 Carb – Carb – – –
Stron Baryt Carb Carb Carb Carb Carb – – –
Alsto Arag Cal Carb Carb Carb – – –
22 – – 98.1 Arag Baryt Cal II Carb Carb Carb – – –
135.59 152.70 178.88 224.46 690.51 693.92 699.62 1059.72 1419.83
With With With With Carb Carb Carb Carb Carb
Arag = aragonite, Cal = calcite, Carb = carbonate, Stron = strontianite, Alsto = alstonite, Baryt = barytocalcite, With = witherite, Mix = mixture.
slope for the three samples do not represent differences in structure. All spectral features occur at the same position and the relative magnitudes of features in each spectrum are consistent across the three samples. Thus the use of different precipitation temperatures with the same composition does not affect the crystal structure. The speed of precipitation may also affect the Raman spectra of the carbonate materials, and the standard drop-wise addition of ammonium carbonate is thought to be essential to the precipitation process. To test this, one sample was prepared at a measured temperature with the same composition as the temperature samples above, but the ammonium carbonate solution was added in one step, rather than the usual drop-wise addition. The Raman spectrum acquired from the sample Emitter 23 at 71 ◦ C is given in Fig. 5, with the previous temperature samples for reference. It can be seen that the same spectral features are observed in the same positions with the same intensity ratios. The speed of precipitation, at a given composition, does not appear to alter the crystal structure.
3.3. Concentration of precipitation solution Five samples were prepared with the same composition (4% Ca, 30% Sr, 66% Ba) and precipitated at approximately same temperature, 70 ◦ C, but with differing concentrations of the ammonium
carbonate solution. Raman spectra were obtained for each sample, and these are shown in Fig. 6. The spectra are very similar indicating that the crystal structure is consistent throughout all ﬁve samples.
4. Conclusions The study of the main parameters for precipitation of alkali earth carbonate materials has uncovered several important features. The Raman spectrum of a mixed carbonate material is only composition dependent. The precipitation parameters are shown not to affect the crystal structure. It has been shown that, even at relatively low levels, the calcium content of a sample can dominate the vibrational frequencies as measured in the Raman spectrum. This is due to the discrepancy in size between the Ca2+ ion and the Sr2+ and Ba2+ which are signiﬁcantly larger and similar to each other. The analysis of the lattice modes demonstrates that the presence of calcium forces considerable changes in the formation of the carbonate lattice. Each Raman shift observed for a sample corresponds to a speciﬁc crystal structure. Some peaks lie within two or three shifts that are observed for different crystal structures. This work shows that with the Raman spectra used as a reference, one can analyse an unknown alkali earth carbonate sample
W. Kaabar et al. / Spectrochimica Acta Part A 78 (2011) 136–141
using Raman microscopy. If these studies could be extended to directly analyse oxide materials, structural information could be inferred non-destructively from Raman spectra acquired from a sample surface. References        
F. Poret, J.M. Roquais, Appl. Surf. Sci. 251 (2005) 31–41. D. Shafer, J. Turnbull, Appl. Surf. Sci. 8 (1981) 225–230. L. Couture, Ann. Phys. Ser. 12 (2) (1947) 5–94. D. Krishnamurti, Proc. Ind. Acad. Sci. A 31 (1960) 285–295. W.P. Grifﬁth, J. Chem. Soc. A (1970) 286–291. H.H. Adler, P.F. Kerr, Am. Mineral. 48 (1963) 839–853. B.E. Scheetz, W.B. White, Am. Mineral. 62 (1977) 36–50. P. Gillet, C. Biellmann, B. Reynard, P. McMillan, Phys. Chem. Minerals 20 (1993) 1–18.  J.M. Alia, Y. Diaz de Mera, H.G.M. Edwards, P. Gonzalez Martin, S. Lopez Andrés, Spectrochim. Acta A 53 (1997) 2347–2362.
 R.L. Frost, M. Dickfos, Polyhedron 26 (2007) 4503–4508.  R.L. Frost, M. Dickfos, J. Raman Spectrosc. 38 (11) (2007) 1516–1522.  R.L. Frost, M.C. Hales, D.L. Wain, J. Raman Spectrosc. 39 (1) (2008) 108– 114.  R.L. Frost, M. Dickfos, Spectrochim. Acta A 71 (1) (2008) 143–146.  A. Hermann, D. Wagner, The Oxide-Coated Cathode, vol. 1, Chapman and Hall, London, 1951.  E.F. Lowry, Illum. Eng. 48 (6) (1951) 288–294.  V.C. Farmer, Mineralogical Society Monograph 4: The Infrared Spectra of Minerals, 1974.  M. Murabayashi, Shin Nippon Denki Giho 4 (2) (1969) 169–176.  M. Murabayashi, H. Shirai, Shin Nippon Denki Giho 6 (1) (1971) 27–31.  D. Chakrabarty, S. Mahapatra, J. Mater. Chem. 9 (1999) 2953–2957.  Database of Raman Spectroscopy, X-ray Diffraction and Chemistry of Minerals Website, http://rruff.info/index.php.  M.Y. Fong, M. Nicol, J. Chem. Phys. 54 (2) (1971) 579–585.  L.G. Liu, T.P. Mernagh, Am. Mineral. 75 (1990) 801–806.  E.E. Coleyshaw, W.P. Grifﬁth, R.J. Bowel, Spectrochim. Acta A 50 (1994) 1909–1918.  R. Frech, E.C. Wang, Spectrochim. Acta A 36 (1980) 915–919.