| Home | AmMin | GMR | RiMG | Collectors Corner | Directory | Short Courses | Contacts |
|
|
|||||||
|
|
Volume 64, pages 160-168, 1979 Coloring mechanisms in celestite LAWRENCE R. BERNSTEIN Department of Geological Sciences, Harvard University Abstract The common pale-blue color of celestite bleaches at about 200°C and reappears upon X-ray irradiation. A thermoluminescence maximum for blue celestite also occurs around 200°C. The thermal stability of the blue color (as measured by the bleaching time at 190°C) was found to be proportional to potassium content. Electron spin resonance data indicate the presence of SO 4-, SO3-, SO2-, and O- in blue celestite crystals. These paramagnetic hole centers presumably arise by the action of ionizing radiation on SO42-, which releases energetic electrons that can decompose other SO42- groups. The centers SO3-, SO2-, and O- absorb in the visible, producing the blue color, while SO4- absorbs in the ultraviolet. The color-producing centers are stabilized in the celestite lattice by the presence of trace components, primarily K+ substituting for Sr2+.The color in two specimens of orange celestite was correlated to the presence of copper, probably as Cu+, in the order of 50 ppm. Introduction The cause of the common pale-blue color of celestite, SrSO4, has long been disputed, and has been the subject of many investigations. In this report some new data are presented on blue and orange celestite, and possible models are derived for the causes of these colors. Previous hypotheses The numerous hypotheses that have been proposed for the coloring mechanism in blue celestite can be summarized as follows: 1. Organic inclusions (Stromeyer, 1821, p. 203-211; Weinhold, 1864). The mechanism whereby these inclusions produce the blue color was not described. 2. Vivianite inclusions (Wittstein, 1856). 3. Colloidal sulfur inclusions produced by radium radiation (Doelter, 1915, p. 65-66). The color is presumably caused by Tyndall scattering. 4. Colloidal gold inclusions (Friend and Allchin, 1939, 1940). The color is produced by Tyndall scattering and related optical effects. 5. Fe2+ - Fe3+ charge transfer (Vendl and Mandy, 1958). 6. Color centers A. F-centers (electrons filling anion vacancies) i. Alkali ions substitute for strontium, producing anion vacancies and interstitial alkali ions, which both act as electron traps (Schulman and Compton, 1962, p. 256-273, 281-282). ii. Caused by ionization of Sr to Sr+ + e-, with the electron filling an SO42- vacancy (Isetti, 1970). B. PO4 substitutes for SO4- (proposed for blue barite by Bershov and Marfunin, 1967). C. Presence of monovalent sulfur-containing radicals plus O- (Samoilovich, 1971; Bakhtin et al., 1973). The color is due to absorptions by orbital transitions within these radicals. Chemical analyses Experimental methods Emission spectroscopy was used on several specimens (LB 1, LB 6, LB 8, and LB 10) to obtain semi quantitative analyses of all the minor and trace elements. This was followed by quantitative atomic absorption analyses of all the specimens for the principal minor and trace elements. Routine methods were used for the analyses. Results and discussion The chemical analyses are presented in Table 1. Except for Ba and Ca, which substitute for Sr, the samples are remarkably low in minor and trace elements. An interesting result was the detection of copper in the two orange specimens (LB 5 and LB 10). As this is the only element not also found in the blue specimens in measurable amounts, it may be responsible for the orange color. The absence of detectable gold (< 10 ppm) tends to argue against the colloidal gold theory for the blue color, but does not entirely rule it out. The blue portion of specimen LB 11 contains notably more Na and K than the colorless portion. This is possible evidence that the presence of Na and K in natural crystals may cause the same sensitization to radiation-induced coloration that Schulman et al. (1952) observed in artificially alkali-doped crystals. It is also interesting that LB 6, the darkest blue specimen, has the lowest Na content but a high K content.
Thermal studies Experimental methods The thermal stability of the color in seven specimens of blue celestite and two specimens of orange celestite was investigated. Homogeneously colored crystal fragments without visible impurities, about 20 x 10 x 1.5 mm and weighing about 0.5 g each, were selected. These were broken roughly in half, one half being kept at room temperature as a reference for the original color. The experimental specimens were heated in a small furnace having a Pyrex window and a chromel-alumel thermocouple sensor. Each specimen was observed continuously during the heating, and observations were recorded at every 10°C rise in temperature. The specimens were heated to 850°C, the maximum for this furnace, at which point the specimens were incandescent. A heating rate of 5°C/ min could be maintained without decrepitation of the samples. In addition, the specimens were heated as a group at 190°C for six days in another furnace. The time required for the crystals to become completely colorless was recorded. To study specimens colored by ionizing radiation, specimens were first bleached at 300°C for 30 minutes and irradiated with X-rays as described in the following section. The irradiated specimens were then heated and observed.
Numbers in parentheses are estimated standard deviations. The high Fe, Si, Mg, and Al content of LB 8 is at least partly due to admixed impurities. Looked for but not detected by emission spectroscopy: Ag, As, Au, B, Be, Cb, Cd, Co, Cr, Dy, Er, Eu, Ga, Gd, Ge, Hf, Hg, La, Li, Lu, Mo, Nd, Ni, P, Pb, Pt, Ra, Sb, Sc, Sn, Sm, Ti, U, V, Zn, Zr. Specimens: LB 1 (NMNH* 87183): Amherstburg, Ontario, Canada. Thick tabular crystals. Pale blue. LB 3 (NMNH 87469): Mt. Bonnell, near Austin, Texas. Cleavage fragments. Pale blue. LB 4 (NMNH 119651): Ottawa County, Ohio. Thick tabular crystals. Pale blue. LB 5 (NMNH 104583): Brown County, Kansas. Intergrown tabular crystals. Brownish orange. LB 6 (NMNH 94275): Girgenti, Sicily. Subhedral crystals in parallel growth. Blue. LB 7 (NMNH R16219): El Rayo Mine, Durango, Mexico. Cleavage fragments. Pale blue. LB 8 (NMNH 16758): Jena, Thuringia, Germany. Cross fiber vein 1.5 cm thick. Blue. LB10 (NMNH 116662): Caledon Township, Ontario, Canada. Tabular crystals. Orange. LB11 (HMM 101020): Clay Center, Ohio. Elongated 100, thin tabular {001) crystals. Blue and colorless zones are from the same crystal. *NMNH = National Museum of Natural History, Washington, D.C. HMM = Harvard Mineralogical Museum, Cambridge, Massachusetts.
Table 2. Thermal bleaching data for blue celestite
Specimens of vivianite (a transparent green fragment from Cameroun and a blue fibrous specimen from Nelson Island, Alaska) were also heated. Results The blue celestite crystals all began to slowly bleach at around 170°C. The color in all specimens simply became paler, without any detectable change in hue. With increasing temperature, a point was reached at which bleaching became rapid, around 200°C. This temperature, the minimum at which bleaching is rapid, is generally defined as the bleaching temperature. The irradiated crystals showed similar behavior to the natural crystals. The bleaching times at 190°C varied widely, from one minute to about six days (Table 2). The two orange specimens (LB 5 and LB 10) behaved entirely differently, the orange color remaining stable through 850°C. A slight brown tinge to the color was present above 350°C, but this faded upon cooling to room temperature. The fibrous blue vivianite became dark brown at 180°C, and the green vivianite became opaque green at 180°C and dark opaque brown at 250°C. Discussion The very low bleaching temperatures of the blue celestite, not accompanied by changes in color, strongly suggest the presence of color centers. The variable bleaching times and temperatures could indicate the presence of several color centers that have different "trap depths" (the thermal energy needed to destroy the color center). However, since the figures for bleaching times (which are probably the most accurate measure of the thermal stability) do not show a modal distribution, this mechanism does not appear to be likely. It is more likely that the rate of color center destruction is being affected by a variable quantity, such as trace element content. The thermal behavior of the orange celestite virtually precludes the possibility that the orange color is caused by color centers. The thermal behavior of the vivianite is not consistent with the theory that vivianite inclusions cause the blue color of celestite. Irradiation studies Experimental methods Fragments used for irradiation were about 5 x 5 x 1.5 mm. The blue specimens were first bleached at 300°C for 30 minutes, and all became colorless. Several specimens were chosen for particular properties, specifically: (1) colorless and blue sections from a single crystal of LB 11 were both irradiated, and (2) a cleavage fragment of LB 6 that contained a well-defined colorless zone 1 mm thick was irradiated after bleaching. Specimens of LB 1, LB 3, LB 4, LB 7, LB 8, and LB 10 (orange) were also irradiated. The specimens were each taped in front of an unfiltered port on a Philips X-ray generator, with an Fe tube operated at 45 kV and 10 mA. The specimens were observed (with the X-ray generator off) after exposure for 15 min, 30 min, 45 min, 60 min, 90 min, 3 hrs, and at longer intervals if warranted. Several of the specimens were also used in the optical absorption spectrometer, both before and after irradiation, as described in the following section. Results All the originally blue crystals became noticeably blue after 15 minutes of irradiation, and reached the maximum depth of color within 45 minutes. The final depth of color was roughly proportional to the depth of the original color. This latter quantity was somewhat difficult to judge, however, since the color developed most intensely in a paper-thin layer facing the X-ray tube, and faded very rapidly going into the crystals. The orange crystal, LB 10, showed no change after 6 hours. The originally colorless section of LB 11 became only extremely pale blue after 24 hours exposure to the radiation, though the originally blue section quickly became blue. The zoned crystal fragment, LB 6, which was uniformly colorless after bleaching, faithfully reproduced its zoning after irradiation (Fig. 1 ).
Discussion The fact that the blue crystals rather quickly reached a maximum depth of color under radiation indicates an apparent approach to saturation of color-producing sites. The limited number of sites is substantiated by the production of color roughly in proportion to the depth of color originally present, and the failure of originally colorless zones to become significantly colored. The factor controlling the number of color-producing sites could be trace elements, which could either sensitize the blue zones to radiation or quench the color-producing process in the uncolored zones.
Optical absorption spectroscopy Experimental methods The absorption spectra of blue and orange celestite were investigated in the visible and near-infrared regions. The effects of thermal bleaching and of irradiation on the spectra were also studied. A spectrum of blue barite, from Cumberland, England, was observed for comparison. A Cary-17 spectrophotometer in the transmission mode was used. The illumination was non-polarized, in the range 400 to 1500 rim. Specimens LB 1, LB 6, LB 10, and LB 11 were studied. The specimens consisted of plates about 3 mm thick cut roughly parallel to {001} and polished on both sides, with the exception of LB 11, which was an unpolished tabular {001} crystal about 3 mm thick. Results The absorption spectra of representative specimens are reproduced in Figures 2 and 3. The major absorption peaks in all the blue celestite are at 620±5 nm, 575±5 nm, and <400 nm. The blue color results from absorption at both the red and violet ends of the visible spectrum, with blue light preferentially transmitted. Minor absorption peaks may occur at 540 and 710 nm. There is virtually no difference between the absorption spectra of the natural and irradiated blue celestite specimens.
Absorption in the blue barite occurs primarily around 640, 595, and <400 nm. The orange celestite has a multiple peak or absorption edge beginning at 600 nm and extending beyond 400 nm. There is also a small "shoulder" at about 540 nm. No absorption was recorded in the near infrared for any of the specimens. The small peak at about 1385 nm is an instrumental artifact. Discussion Since the absorption spectra of natural and irradiated blue celestite are virtually identical, the sane mechanism is probably responsible for the color in both, and the color in the natural samples may be radiation-induced. The absence of any of the characteristic Fe2+ and Fe3+ absorption bands (Burns, 1970; Loeffler and Burns, 1976) virtually rules out the possibility of these ions contributing to the color, either by themselves or through charge transfer. The spectra of the blue and orange celestite have almost nothing in common. The two colors are almost certainly caused by different mechanisms, as is also revealed by their thermal behaviors.
The spectra of blue barite and blue celestite have a very similar shape, but the peaks of barite are displaced about 20 nm to longer wavelengths (less energy). This shift may be related to the larger unit-cell size of barite (isostructural with celestite), which results in a slightly lower crystal field energy acting on the coloring mechanism. Electron spin resonance spectroscopy Introduction
ESR spectroscopy measures the transitions of unpaired electrons between the spin states ms = ±
½. The technique can detect paramagnetic ions and radicals, and also free unpaired electrons. The transition is generally expressed by the g-factor, where hp =
gµBH (h = Planck's constant, v = frequency of applied alternating magnetic field, µB = Bohr magneton, H = static magnetic field strength). The g-factor in anisotropic crystals is directional and is expressed as the g-tensors
gx, gy, and gz (or, sometimes g||, and
g Table 3. ESR data for celestite and for some related substances
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
for centers with axial symmetry). The g-factor is further split by electron interactions with nuclei having non-zero spin (I ≠ 0), and with neighboring paramagnetic centers. The measurement of these influences, together with several other quantities measured by
ESR, can give a highly distinctive and unambiguous identification of the composition, symmetry, and location of the paramagnetic centers. In general, centers with an average g-factor less than
g0 (that of a free electron, 2.0023) are electron-excess centers, and those with an average value of g greater than
g0 are hole centers.