sign in sign up
<<
Home , Blue John (mineral), Fluorite, Treak Cliff Cavern

MINERALOGICAL MAGAZINE, JUNE I97 9, VOL. 43, PP. 243-50

The Blue coloration in banded () from Castleton, ,

A. K. GALWEY, 1 K. A. JONES, 2 R. REED 3 1 Department of Chemistry, 2 Department of Geology, 3 The Electron Microscopy Laboratory, The Queen's University, Belfast BT7 INN, Northern Ireland

AND

D. DOLLIMORE Department of Chemistry and Applied Chemistry, University of Salford, Salford, M5 4WT, Lancashire, England s u MM A R Y. From the microscopic examination of lightly of chromophore that have been proposed to etched, cleaved (I I I) surfaces of banded blue fluorite explain the blue coloration. These are incorporated (Blue John, from Castleton, Derbyshire) it is concluded , inorganic impurities, and lattice that the coloured lamellae are not necessarily or exclu- sively associated with lattice-line imperfections. Ele- imperfections or damage produced mechanically mental analyses, by electron dispersive methods, of fresh or by radiation. These workers applied the usual (I I I) surfaces in the immediate vicinity of small complementary chemical and physical techniques inclusions (c. IO #m diameter) possessing associated and conclude that the chief colouring material is coloured haloes detected no appreciable concentrations colloidal Ca, banded as a consequence of changing of impurities. From the evidence available, it is suggested conditions during mineralization. Braithwaite et al. that the zones of blue colour consist of colloidal (I973) find no correlations between the distribu- resulting from radiation damage caused by the inter- tions of colour and of either inorganic or organic mittent deposition of radioactive material on the surfaces impurities in Blue John and related . They of fluorite during development. The of colloidal calcium produced is particularly stable as a conclude that physical rather than chemical factors consequence of the close correspondence between lattice are responsible for the coloration and suggest that spacings in calcium and in calcium metal. colloidal particles of Ca have been formed by the aggregation of Ca atoms preferentially within the more defective regions of the crystal. These atoms THE unusual and aesthetically pleasing banded are believed to have been released from the lattice fluorite, often referred to as 'Blue John', occurring during long-term irradiation by radioactive ele- near Castleton in Derbyshire has been the subject ments, which are known to have been available by of long and continuing investigations of the their continued presence in associated uraniferous reasons for the striking blue coloration. Compari- deposits (collophane). sons have been made with the properties of colour- Calas (I972) concluded that colloidal Ca is not less material from the same location and, or, necessarily the colouring matter in all banded banded or colourless from other widely fluorites and discusses the possible existence of scattered sources. However, despite many alternative light-absorbing centres containing in- investigations no general agreement concerning the organic impurities, notably . Holgate identity of the chromophore has yet been reached (I973), from consideration of the dichroism of perhaps partly because much of the experimental coloured fluorites, believes that the most probable evidence has been negative in character. More chromophore in Blue John is a hydrocarbon, possible explanations have been eliminated than though the particular compound or molecular proposals positively proved. grouping is not identified. Thus, from these and MacKenzie and Green (I97I) summarize the other recent reports, we must conclude that the earlier literature concerned with Blue John and light-absorbing centres in coloured fluorite have related fluorites and identify the three general types not been positively and finally characterized and Copyright the Mineralogical Society 244 A. K. GALWEY, ET AL. that the nature of the chromophores present dark blue. In addition to these readily recognizable remains a matter for active discussion. patterns of colour distribution, with which we are The present article reports a study of two aspects particularly concerned here, there were also zones of the structure and composition of Blue John. of less regular tinting, ranging from what appeared Firstly, we have investigated the relationship be- to be lamellae of restricted extent to regions of ill- tween the distribution of lattice-line imperfections defined outline containing no recognizable struc- (dislocations) now present, as revealed by etching of ture. It should be remembered that any explanation cleaved surfaces, and the distribution of the blue of the coloration, based on consideration of the coloration. This was intended to determine whether properties of representative lamellae and the most direct support could be obtained for the theory that concentrated groups of point features with asso- dispersed Ca in the bulk crystal was precipitated or ciated haloes, must also be capable of accounting aggregated in the form of a stable colloid prefer- for the existence of the ill-defined and more diffuse entially within defective zones. Although deforma- regions of tinting. tion and annealing under natural conditions could Lamellae of coloration. Planar lamellae of colora- be expected to cause continuing variations in lattice tion, many being c. o.I mm thickness, obliquely imperfection distribution, it was thought probable traverse the thin plates (o.2-1.o mm) of fluorite that extensive precipitation of colloidal material prepared for examination by parallel cleavages at within dislocations would stabilize them against the (i I I) plane. This appearance is consistent with further movement. It seems inherently more likely the production of the thin coloured lamellae at that progressive dispersal of single Ca atoms would crystal planes parallel to cubic growth faces, during occur more readily than the annihilation of the or following addition of CaF2 in mineralization: dislocation networks within which such atoms were natural deposits of fluorite are often produced in preferentially accommodated in the first place. the form of assemblages of intergrown cubes, for Secondly, we have attempted to establish the which the (ioo) faces predominate. Growth direc- composition of features having the appearance of tions, and their relict structures, therefore, appear inclusions and which characteristically occurred in as planes orientated at 54 ~ 44' to the strongly irregular groups comprised of numerous individual preferred (Ill) direction of the fluorite members, each possessing an associated zone of crystal. blue coloration. We have been unable to identify Assemblages of inclusions. Slivers of fluorite, any foreign constituent and conclude that there are obtained on cleavage, occasionally contained zones no appreciable local concentrations of any of irregular dispositions of large numbers of small included non-volatile impurity. features provisionally identified as inclusions. The In this article we also make brief reference to individual members of these assemblages were too certain other of our measurements. These report no small (diameters c. Io/zm) to permit the resolution new information but are entirely consistent with of detail by optical microscopy. In some places each and serve to confirm earlier observations by others. opaque zone was surrounded by an indistinct blue The nature of the colour centres resulting in the halo. blue banding and the manner of its production Line dislocations in Blue John under natural conditions are discussed. The preferential dissolution of crystal surfaces initiated in the vicinity of imperfections has been Distribution of coloration in Blue John widely used as a method of locating points of The present study was exclusively concerned intersection of line dislocations with surfaces. Pre- with samples of natural banded fluorite (Blue John) liminary work with the present showed collected from a surface exposure directly above that triangular etch pits were developed after the entrance to , near Castleton, c. 2 min dissolution in IO ~o aqueous hydrochloric Derbyshire (Nat. Grid Ref. SK 135 83I). From acid at c. 35 ~ K. An unidentified deposit produced relatively perfect single (c. Io-3o mm linear in this treatment was removed from etched sur- dimensions) selected specimens were cleaved, by a faces by washing with hot distilled water before sharp blow with a knife edge, into thin plates microscopic examination or analysis. suitable for microscopic examination. Coloration About twenty different thin cleaved (i i i) plates in banded fluorite occurred in two characteristic of Blue John, each specimen having been selected to distributions: in the form of planar lamellae and in include at least one strongly coloured lamella, were the immediate vicinity of individual members of individually etched under an appropriate range of features having the appearance of irregular groups conditions (of severities greater than and also less of inclusions. There were significant variations of than the optimum) and the treated surfaces hue, even within a single crystal fragment; examined in detail in a scanning electron micro- perceived colours varied from pink-magenta to scope. This technique has the disadvantage that the BLUE COLORATION IN BANDED FLUORITE 245 precise position of the coloured zone does not located in a previous optical examination. A appear on the image, but it was possible to prepare representative composite photograph of an overlapping sequences of photographs of areas of etched surface is shown in fig. I, for which it interest for comparison with features identified and was independently known that a blue lamella

FIGS. I and 2: FIe;. I (above). Texture of(III) cleavage surtace of Blue John etched briefly with Io o; aqueous at c. 35o K. A blue lamella, aligned in the direction shown by the arrow, intersected the surface somewhere between the limits given by the horizontal white lines. None of the several aligned etch features could be positively ascribed to the presence of the coloured zone. SEM photograph. FIG. 2 (right). Typical inclusions found on (I ~I) cleavage surfaces of Blue John: both examples occurred on faces exposed by the vigorous disruption of the mineral on heating at 750 K. SEM photogra)h. 246 A. K. GALWEY, ET AL. intersected the surface as a horizontal band (of Assemblages of inclusions in Blue John unknown width) somewhere between the limits shown (as horizontal lines). No feature of the Representative small opaque zones, each with an etched surface could be correlated with the zone associated blue halo, were examined in an attempt of blue coloration. to identify any included material, since this could Etching invariably revealed the presence of give evidence of the composition of substances alignments of crystallographic imperfections, present during hydrothermal deposition of the which appeared as linear troughs apparently com- mineral. Optical microscopy yielded no useful posed of closely spaced, overlapping triangular information and it was difficult to ascertain etch pits. Dimensions of such troughs were signifi- whether the very small opaque spots (< IO pm) of cantly less than the thicknesses of the coloured interest were exposed at the surfaces of samples lamellae. While, in some instances, etch pit align- cleaved for stereoscan examination, even for ments coincided with known positions of blue crystals containing locally dense concentrations of zones, in other examples no such correlation was these features. Accordingly, numerous compari- evident and elsewhere, again, linear pattems of etch sons were made between stereoscan pictures of pits were present in places remote from any colora- cleavage surfaces traversing inclusion-rich regions tion. Thus, blue lamellae cannot be identified with and those from clear fluorite. The only general the presence of line imperfections, lattice disorder, conclusion to emerge from this comparison was or increased surface reactivity resulting in any that there was an enhanced incidence of pitting in significant enhancement of the local rate of salt areas from the inclusion-rich material. Some repre- dissolution. sentative pits are shown in fig. 2. We believe, While it is possible that extensive redistribution therefore, that the opaque spots were holes within and perhaps removal of dislocations may have the crystal filled with a or volatile liquid. The occurred during the post-depositional interval, it interiors, edges, and immediate vicinities of up- was none the less unexpected that there was no wards of IOO of these holes, and, in addition, such evidence of an enhanced, or even changed reactivity rare atypical surface features as were noted, were of the crystal in the vicinity of blue zones. This analysed by the EDAX method (Energy Dispersive would appear to suggest that the crystal lattice Analysis of Xrays: Edax International). The only within the lamellae does not now contain high element (of atomic number > I5) present was Ca, concentrations of line imperfections or impurities. though one or two instances of the presence of Fe in If it is believed (Braithwaite et al., 1973) that the trace amounts and at the lower limits of detection colour banding arose as a consequence of the were indicated. There was no evidence, therefore, dislocation net providing sites for the preferential that elements, held in any solution that was con- aggregation of Ca particles of colloidal dimensions tained within these inclusions, were deposited on from atoms released by radiation damage within solvent evaporation. We cannot be certain from the crystal bulk, it is difficult to explain how such our observations that included minerals containing an impurity-laden and locked network could have A1, O, Si (etc.) were entirely absent. We can be sure, been reorganized without achieving conditions of however, from detailed consideration of data for sufficient severity to either redisperse the Ca as both typical and atypical areas of representative atoms or permit its continued aggregation within cleaved surfaces that all analytical results were the imperfect zones. It is also necessary, on this entirely consistent with the view that the mineral model, to make the additional assumption that the was pure . Again, within the very trapped colloid has particular stability, since few zones of exceptional texture found, it was coloured fluorites from several distinct sources give apparent from EDAX analyses that calcium was (MacKenzie and Green, I97I) rather similar the preponderant, possibly exclusive, cation. spectra, presumably attributable to the presence of Crystals heated to c. 70o-8oo K (below the particles within the same size range. temperature at which colour is completely lost: From this evidence, therefore, we conclude that, some aspects of thermal bleaching are discussed by since the imperfection structure assumed to have Braithwaite et al., I973) disrupted violently with been present during colloidal particle development preferred cleavage at (I I I) planes. Stereoscan is not now detectable, the maintained colour examination showed that the surfaces exposed distribution in bands cannot be ascribed to the were rich in pits and EDAX analysis again con- formation of a colloidal dispersion of Ca metal in a firmed that calcium was apparently the exclusive preferred size range resulting from Ca atom surface cation. Surface pits were not invariably aggregation exclusively within a pre-existing and, associated with colour and the exposure of such in places, highly inhomogeneous, distribution of features on heating suggests that disruption of the lattice imperfections. crystal may be a consequence of pressure developed BLUE COLORATION IN BANDED FLUORITE 247 by fluid within the inclusion at elevated tem- electron spin resonance and proton spin resonance perature. spectra. A more precise application of the same general X-ray spectral analysis of powdered approach to the chemical identification of included samples revealed no detectable difference in im- material was made by selecting for examination a purity concentrations between coloured and single relatively large, separate, and distinct inclu- colourless Blue John; approximate amounts de- sion, e. 1o pm diameter, surrounded by a coloured tected were Sr 25, Fe I5, Y Io, and Mn 25 ppm. halo, c. 8/~m thickness. The fluorite overlaying this Combustion analyses gave H o.15 and C o.35~. structure was gently removed by systematically Mass spectral analysis, using an AEI MS9o2 instru- finer polishing until the dark spot was adjudged by ment, for the immediate analyses of the volatile optical microscopic examination to be at the products obtained on vacuum heating of Blue John mineral surface. This crystal was then mounted in to 700-50 ~ identified only H20, CO, and CO2. an electron microprobe and the beam (accurately Although the background response of the instru- located by the intentional production of a blue spot ment within the mass range Ioo-4o was generally on the surface at a point somewhat removed from increased during sample heating, no significant the feature of interest) was focused on the centre of quantity of any identifiable product was detected the selected inclusion prepared for examination. within the mass range Ioo-25o. , a possible Analytical scans were made extending across the product of radioactive decay, was sought, but not complete spectrum available (3o-i3 o~ for three found, in the released on heating the mineral crystal analysers: , , and to 800 ~ There was no indication of the forma- KAP. In addition, specific impurities were sought. tion of any detectable (i.e. < Io -9 mol g-~ or No perceptable concentration of any element other o.ooooI ~) quantity of organic products in the than Ca was detected, from which it is concluded C4-C~ o hydrocarbon boiling range by the thermal that the local concentrations of impurities (atomic desorption method (Galwey, I965), on heating number > IO) were all < o.I ~. The same result various samples of Blue John to the temperature at was obtained for two further selected inclusions, which the colour was lost, c. 77 ~ ~ similarly exposed on the surface and analysed. Electron microprobe analyses at each of a linear DISCUSSION sequence of surface points extending across the zone of intersection of a blue band with a cleavage Previous studies, together with limited observa- face gave no evidence of the presence of any tions within the present work, have shown that the detectable quantities of the following possible bulk composition of the coloured fluorite zones is impurities: Fe, K, Mn, P, S, Si, Sr, U, or Y. The only indistinguishable from that of the colourless major response was (again) for Ca and, indeed, the mineral. Our attempts here to identify any highly indications were that the concentration of this localized compositional changes that might be element was perhaps marginally greater (by associated with microstructural anomalies or inclu- < 0.5 ~o) within the blue band. While this apparent sions (dimensions c. 5 /~m upwards) have been local enhancement of concentration is too small to unable to detect the existence of any variations in be accepted as reliable, it does confirm that there is composition, or impurity aggregation, capable of not a diminished concentration of Ca a+ within the accounting for the coloration. Ca was virtually the coloured zone; this is, perhaps, the most probable only element detected by the techniques employed. consequence of the presence of cationic impurities. These analytical investigations were conducted in conjunction with etching studies of cleaved Other measurements surfaces, which revealed, from the pattern of etch pit development, that line imperfections were dis- Other of our measurements agree with, and serve tributed throughout the crystal and dislocations to confirm, conclusions published elsewhere; rele- were not associated with either lamellae of colora- vant experimental data which we have obtained tion or assemblages of inclusions. From this we may be briefly mentioned: conclude that structures within tinted zones were The coloration was not removed on boiling generally similar to those elsewhere and that the crushed Blue John with organic or inorganic chromophores are accommodated within a zone of (aqueous NaOH, HC1, or H2SO4) solvents. Lattice relatively perfect crystal, containing no recogniz- spacings, measured by X-ray diffraction, for blue able lattice distortion features. and for colourless material were indistinguishable. The available evidence, therefore, indicates that The absorption spectrum (MacKenzie and Green, Blue John is a remarkably pure and homogeneous I97I) included a broad band between 55o-6oo nm, fluorite. Bulk and local concentrations of impuri- but no useful information could be obtained from ties are probably too low to account for the 248 A. K. GALWEY, ET AL. observed absorption spectrum. No compositional as large organic molecules. The required concentra- feature, local reactivity variation (etching), textural tions of such impurities, disposed on original anomaly, lattice imperfection or characteristic growth (ioo) surfaces, would result in an appreci- microstructure capable of accounting for the able reduction in hardness and so the crystal would coloration was recognized. We, therefore, con- more easily break in this direction, unless the clude, with other authors (MacKenzie and Green, impurity forms bonds with the host lattice. I971; Braithwaite et al., 1973), that the blue colora- The indistinct local zones of blue, apparently tion is due to the presence of colloidal Ca, though associated with small inclusions, are not readily we propose below explanations for the develop- accounted for on this, or, indeed, most other simple ment of the banding and for the structure of this explanations. It is difficult to provide a realistic colloid that are somewhat different from those model for such inclusions, since detailed structures previously advanced. of the features could not be resolved by optical microscopy and no evidence of impurity retention Development of banded coloration in Blue John could be obtained. The most acceptable representa- tion that we can propose at present is that the Our observation that the blue coloration is not temporary residence of particles of sediment inter- consistently correlated with the existing distribu- rupted crystal growth locally, radioactive con- tion of line dislocations within the mineral us stituents resulting in damage to the fluorite in the to reject (see above) the suggested explanation immediate vicinity of the resting fragment. Such (Braithwaite et al., I973) that banding arises be- deposited material was not, however, itself included cause Ca atoms released in the crystal bulk (during within the growing mineral in analytically detect- irradiation by a radioactive material, e.g. U) segre- able amounts, though the temporary residence of gate by aggregation as colloidal particles prefer- such particles may have inhibited (or poisoned, by entially located within zones of imperfection, monolayer deposition) local surface advance suffi- banded and oriented in the (ioo) direction. We ciently to result in the formation of some (solvent- prefer an explanation invoking intermittent local filled?) inclusions. superficial radiation damage to the fluorite crystals This general model is consistent with the view during interruptions to their hydrothermal growth (Stevenson and Gaunt, I971) that Blue John was (Pering, I973). Mineral development through developed in its present position earlier than the advance of the cubic faces of the growing crystal is main mineralization phase of the region since it envisaged as proceeding by acquisition of CaF2 tends to occur in cavities and pipe vents rather than from a cooling solution, also containing a soluble in the veins proper. radioactive substance (U), which is later deposited. During this period a thin layer of the external Structure of colloidal Ca in Blue John fluorite surface suffered some radiation damage before the arrival of a further pulse of hot minera- In the observations reported above, we have lizing solution. This rapidly redissolved the radio- obtained no direct information concerning the active material, perhaps also some fluorite, before sizes, shapes, relative dispositions, concentrations, cooling to the point at which CaF2 deposition, or structures oftbe colloidal particles of Ca that we crystal growth, was resumed, covering and protect- believe to be responsible for the coloration of Blue ing the damaged (IOO) planar zone. Intermittent John, except that the dimensions were below the repetition of such cycles accounts for the irregular limits of resolution of the techniques available to patterns of blue planar zones, parallel, but of us. The only relevant information would appear to varying intensity, the presence of incomplete bands be the absence of a Tyndall effect (Braithwaite et al., (as a consequence of uneven dissolution) and the I973), indicating that the average diameter of the absence of retained radioactive material. On this colloidal particles is < 3o nm. model radiation damage is of limited thickness and The possible structure of colloidal Ca in fluorite crystal annealing may occur by local and limited is appropriately discussed with reference to recent ionic movements, which largely restore the perfec- observations and conclusions concerning radiation tion of the lattice, enveloping and protecting the damage in CaF 2 by Chadderton et al. (I976a, b). incorporated crystallites of the colloidal dispersion. These workers have subjected calcium fluoride There is little reduction of crystal coherence along crystals to a very high electron bombardment planes of blue coloration since disintegration does and found that instead of complete disintegration, not preferentially propagate within the chromo- as might have been anticipated, complicated lattice phore-rich laminae. This last important property of reorganizations occurred, leading to new structures Blue John does not seem to have been adequately of considerable stability. There was fluoride considered by those who identify the chromophore migration in the anionic sub-lattice and interstitials BLUE COLORATION IN BANDED FLUORITE 249 became trapped at bubbles of gas. Con- penetration for the complicated physical and currently cations were transformed into Ca crystal- chemical processes involved in the production of lites, coherent with the matrix and providing the stable colloid. It seems to us probable that nucleation centres for the anionic voids. The signifi- crystal damage within the coloured lamellae at cant and perhaps unique feature of this metal distances furthest from the source must arise in an production process (presumably also relevant to asymmetric radiation field and, if this results in the the production of chromophores, colloidal Ca, in generation of asymmetric colloidal particles, an the coloured natural fluorite) was that the cubic explanation of the existence of the dichroic (Hol- lattice of Ca metal could be accommodated within gate, I973) effect is provided. the (cubic) fluorite lattice with a minimum of The model proposed above, therefore, provides strain, since the unit cell edge of Ca metal is but an acceptable explanation of the chromophore, 2. I ~/o greater than that of calcium fluoride, consistent with available information and not in We do not intend to suggest that the radiation conflict with previous observations reported in the damage that resulted in colour development in Blue literature. It is suggested that lattice reorganization John, or any other coloured natural fluorite, following radiation damage, by annealing through approached the severity of conditions that relatively short-distance migrations, removes most occurred during the electron-microscope experi- large imperfections from the lattice and no excep- ments referred to above. The point relevant to the tional concentration of dislocations remains at the present discussion, however, is that, unlike most blue zones. other substances, lattice reorganization following irradiation can generate a Ca metal dispersion of particular stability, coherently accommodated CONCLUSIONS within the host CaF2 lattice with a minimum of We conclude, therefore, that the banded blue strain. Metal enveloped within the stable, well- coloration in Blue John represents zones of radia- crystallized, and insoluble mineral may be expected tion damage, produced during the intermittent but to be unreactive and protected from chemical temporary deposition of radioactive material at the change during geological periods of time. It is growing cubic faces of fluorite mineral. Local reasonable to suppose that in such a dispersion damage in a thin surface layer, during and follow- there would be a particle size range of preferred ing passage of this radiation, resulted in the aggre- stability influenced (inter alia) by strain energy due gation of fluorine as gas bubbles and the generation to molar volume differences (lattice spacings do not of small accumulations of metal atoms coherently constitute an exact match) and strain at the Ca- accommodated within reorganized crystals of CaF2 interface, which is insufficient to cause a fluorite. During this recrystallization, local con- marked reduction in crystal coherence within centrations of lattice imperfections were reduced. colloid-containing (coloured) zones. Properties of The colloidal particles of metal were of a preferred the stable colloid, including metal particle size, are, size range, of particular stability, and protected therefore, controlled by spacings and elastic con- within a relatively unreactive matrix. This explana- stants of the constituent phases. This explains the tion is based on the properties of the fluorite lattice occurrence of broadly similar tinting of fluorite (since the conditions that obtained during deposi- obtained from widely differing localities, while tion of the mineral cannot now be verified directly) minor differences in colloid dimensions account for and possesses the important feature that it does variations in spectra and colours. In contrast, it is not require the participation of impurities and, not evident that the growth of metal at zones of or, structures, the presence of which it has proved imperfection from atoms released in the bulk under so difficult to demonstrate. the influence of radiation would yield particles within a preferred size interval. Although it might be expected that coloured Acknowledgements. The authors thank Mr Ian Meighan bands, radiation-induced by this mechanism, for assistance in the determination of trace impurities in the bulk sample. would reduce in intensity with distance from the source (and we are indebted to a referee for this suggestion), this possibility could not be investi- REFERENCES gated by direct measurement since the blue lamel- Braithwaite (R. S. W.), Flowers (W. T.), Haszeldine lae were too thin. It was our qualitative impression, (R. N.), and Russell (M.), I973. Mineral. Mag. 39, however, that both bounding surfaces of coloured 40I-I I. zones were relatively sharp. Moreover, we are Calas (G.), I972. Ibid. 38, 977-9. aware of no reliable method of predicting the Chadderton (L. T.), Johnson (E.), and Wohlenberg (T.), variation in optical density with depth of radiation I976a. Radiation Effects, 28, I I I-I2. z5o A. K. GALWEY, ET AL. Chadderton (L. T.), Johnson (E), and Wohlenberg (T.), Holgate (N.), I973. Mineral. Mag. 39, 363-6. i976b. Physica Scripta, 13, I27-8; Monographs 76-o8, MacKenzie (K. J. D.) and Green (J. M.), I97I. Ibid. 38, 76-IZ of Fysisk Laboratorium l I, Orsted Institute, 459-7o. KCbenhavns Universitet. Pering (K. L.), I973. Geochim. Cosmochim. Acta, 37, 4ox- Galwey (A. K.), I965. Proceedings 3rd International 17. Congress on Catalysis, Amsterdam, North-Holland, Stevenson (I. P.) and Gaunt (G. D.), I97I. Geology of the Amsterdam, 79I-8O3 . country around Chapel-en-le-Frith, HMSO, London, --I965. J. Catalysis, 4, 34-42. p. 3o7.