MINERALOGICAL MAGAZINE, SEPTEMBER 1986, VOL. 50, PP. 521-6

Thermal decomposition reactions of and their products

D. J. MORGAN British Geological Survey, 64 78 Gray~s Inn Road, London WC1X 8NG

S. ST. J. WARNE Department of Geology, University of Newcastle, New South Wales 2308, Australia

S. B. WARRINGTON Stanton Redcroft Ltd., Copper Mill Lane, London SW17 0BN

AND

P. H. A. NANCARROW British Geological Survey, 64-78 Gray's Inn Road, London WC1X 8NG

ABSTRACT. The thermal decomposition of caledonite Description and initial characterization of has been examined by simultaneous differential thermal sample analysis, thermogravimetry and mass spectrometry. Structural H20 and CO 2 are liberated endothermically The caledonite occurred in the centre of a geode between 300 and 400 ~ leaving a residue of sulphate, approximately 4 cm in diameter, from the oxysulphate, and Cu(0 and CuO0 oxides. A series of sharp district, Strathclyde Region, Scotland (British Geo- endothermic peaks between 850 and 950 ~ correspond to logical Survey, inventory no. 7443). It phase transition and melting reactions of the PbO- was present as prismatic bluish-green crystals, PbSO, mixture. The sulphate anion breaks down above elongated along (001), intergrown with 880 ~ Mass spectra of the gaseous decomposition and minor . The caledonite crystals were products show SO2, SO, and 02, although SO is an artefact arising from ion fragmentation of the SO2 within carefully hand-separated under a binocular micro- the mass spectrometer. The residue at 1060 ~ is com- scope and ground to pass a 120 mesh (125 pm) posed predominantly of 2PbO-PbSO, and Cu(I) and screen. The ground material gave an X-ray powder Cu(n) oxides. pattern almost identical with that recorded KEYWORDS: caledonite, differential thermal analysis, on JCPDS card 29-565A (see also Giacovazzo thermogravimetry, mass spectrometry. et al., 1973). The infra-red spectrum of the material was recorded by Dr J. D. Russell A N U M~ ER of papers have recently been devoted to who reported (personal communication, March the characterization of and distinction between the 1985) that it 'is identical to that over the range polymorphs of Pb4SO4(COa)2(OH) 2 (Russell et al., 1800-400 cm -1 shown by Moenke (1974). The 1983, 1984; Livingstone and Sarp, 1984) and principal OH stretching band occurs at 3408 the thermal decomposition of the most common cm 1 with weak inflexions at 3434, 3377, and 3300 of these----has been comprehensively cm 1, compared to the positions listed by Moenke described by Milodowski and Morgan (1984). The at 3625, 3585, and 3410 cm-1 but not illustrated. present paper describes the reactions undergone by There was little or no interaction between the the related mineral caledonite (PbsCu2(SO,) 3 mineral and the alkali halide, the spectrum (fig. 1) (CO3)(OH)6) on heating, using results from simul- being virtually identical to those obtained in inert taneous DTA-TG equipment linked to a quadru- mulling agents, the latter showing only slight pole mass spectrometer for evolved volatiles weakening of SO4 absorption bands at 1120 and analysis. 620 cm - 1., ~) Copyriffht the Mineraloffical Society 522 D.J. MORGAN ET AL.

I I I I I I I I I I

c

621 606 c~ -3300

1- \ 3408 1397 1059 1122

I I I I I I I I 4000 3000 2000 1600 1200 800 400

/era -1 FIG. 1. Infra-red spectrum of caledonite. 0.8 mg in 13 mm diameter Csl disc, recorded on Perkin Elmer 580B instrument.

930~/A 863

A m m 10

0 --

0

J< t. 10 E

346 j \4.___

H20 S02

CO 2

02 cooling

~ 200 300 400 500 600 700 800 900 1000 500 Temperature ~ 1060 FIG. 2. DTA, TG, and volatile evolution profiles for caledonite. 30 mg sample in flowing nitrogen. Heating rate 15 ~ to 1060 ~ natural cooling rate. THERMAL REACTIONS OF CALEDONITE 523

TABLE I. Phases identified in products cooled from different

temperatures on the caledonite DTA curve (see Fig. 2).

~ CuO Cu20 PbSO 4 PbO.PbSO 4 ~-2PbO.PbSO 4 4PbO.PbSO 4

550 o o o*

800 o o ,*

950 o o

1060 o o o

major (* dominant)

o minor

Methods This initial decomposition can be represented by: Simultaneous DTA-TG curves and volatile Pb5 Cu2(SO4)3(C0 3)(OH)6 -+ evolution profiles were obtained in flowing nitro- 2[PbO PbSO~] + PbSO4 + gen and at a heating rate of 15 ~ using 2CuO + CO 2 -]- 3H20. (1) a Stanton Redcroft STA 781 thermal analyser This agrees with the above observations except that linked to a quadrupote mass spectrometer. Sample it does not take account of the identification of weights varied between 20 and 50 mg. Interfacing Cu20 in the decomposition products. Copper is between the thermal analyser and the mass spectro- present as Cu(II) in the caledonite structure, each meter was by an inert glass-lined capillary which Cu atom occurring in pseudo-octahedral sites was heated to avoid condensation. surrounded by 4OH and 20 (Giacovazzo et aL, X-ray studies on intermediate decomposition 1973). The occurrence of both CuO and Cu20 in products were made on spindle mounts in a 114.6 the decomposition residue is therefore anomalous mm diameter Debye Scherrer powder camera as there does not appear to be any scope for using Cu-Kct radiation. reduction of Cu(n) during decomposition by any other products of the reaction. In addition, no Results and discussion was liberated during this reaction (see 02 trace on fig. 2). However, in order to accommodate Simultaneous DTA-TG curves are given in fig. 2, the X-ray identification of both CuO and Cu20 together with evolution profiles for H20, CO2, O2, as reaction products, the decomposition may be SO2, and SO. Phases identified in products cooled represented by from intermediate temperatures indicated on the DTA curve are recorded in Table I. Pb5 Cu2(SO4)3(COa)(OH)6 --> 2l-PbO - PbSO4] + PbSO, + CuO + Reactions 20 800 ~ Water and CO2 are ex- pelled in a single sharp endothermic reaction with a 189 + 3H2OT. (2) peak at 346 ~ on the DTA curve. The temperature In this equation the assumption is made that any of this peak was unaffected by Pco2 up to 1 atmos- oxygen liberated is retained within the reaction phere (cf. the behaviour of basic copper products; as will be seen from the next section, there described by Henmi et al., 1985). The weight loss is some experimental evidence for this. corresponding to this endothermic reaction is 6.2 ~o Reference to the phase diagram for the system (theoretical H20+CO 2 content of caledonite = PbO PbSO4 shows that the lead sulphate/ 6.08 ~). In contrast to that of H20, the CO2 profile oxysulphate decomposition product plots in a shows a marked 'tailing off' towards higher tem- position equivalent to point B in fig. 3. peratures and the TG curve does not plateau until Reactions 800 1060 ~ The DTA curve in fig. 2 640 ~ X-ray powder photographs of products shows three sharp endothermic peaks at 883, 928, heated to 550 and 800 ~ were virtually identical and 936 ~ These three temperatures correspond (Table I), these indicating major PbO' PbSO4, respectively to the ct-fl PbSO 4 transition (point C, significant PbSO4, and minor CuO and Cu20. fig. 3), the start of incongruent melting of the 2: 1 Both CuO and Cu20 lines were slightly stronger PbO'PbSO4 decomposition mixture (point D), in the product from 800 ~ but no copper and the formation of a true liquid (point E). sulphate lines were detected in either powder Breakdown of the sulphate anion begins at ap- photograph. proximately the same temperature as that of the 524 D. J. MORGAN ET AL. 1200

1100 Liquid.

1000

900

,,_2PbO. l=-2PbO. C 800 PbS04 J PbS04

2 700 PbS04 I -4Pbo.Pbso, 11 ,

P 600 - IT |' Q) (metastable PbO.PbS04 {3. a-2PbO.PbSO 4) E + F- 500 - I oc-PbSO, oc-PbO 400 - + 4PbO.PbS04 metastable f3-2PbO.PbSO 4) B 300 - 1,I I 200 - I o" (./3 .~ ..Q AS) 100 - t3.. m o (5 O o" 0 ..~ x2t .s 03 Q_ 13_ IX. IX. 0 a_ I ~ 1 TM I I I I I IO0 90 80 70 60 50 40 30 20 10 0 mole% PbO FIG. 3. Phase diagram for the system PbO-PbSO 4 (after Billhardt, 1970) showing decomposition route of caledonite. Points B J are discussed in the text.

~-fl PbSO 4 transition and appears to be accom- and a smaller, less well-defined one at 930 ~ The panied by evolution of 502, SO, and 0 2. The coincidence of these peaks with those on the DTA evolution profiles of all three volatiles have similar curve corresponding to the ~-/~ PbSO 4 transition shapes and peak at approximately the same tem- and subsequent melting reactions suggests that peratures but the oxygen profile contains two they are due to physical release of oxygen which additional peaks, a relatively sharp one at 890 ~ was trapped within the products of the first THERMAL REACTIONS OF CALEDONITE 525 decomposition. Isolation of decomposition pro- Lombardi (1984) figured mass spectra of the ducts of other lead by oxysulphate gaseous decomposition products of three alunites formation has been postulated previously by Gray which showed SO3, SO2, and SO, but in widely et al. (1967) and Bugajska and Karwan (1979). differing ratios. SO is a thermodynamically un- The heating programme was stopped at 1050 ~ stable, fugitive species with a lifetime of a fraction of following a weight loss of ~ 10~o for this second a second, and is unlikely to survive the journey from decomposition reaction. A product cooled from sample to mass spectrometer (see e.g. Greenwood 950 ~ i.e. half-way through the reaction, showed and Earnshaw, 1984). Therefore, the most likely major PbO-PbSO4 and minor CuO and Cu20 explanation for its presence on mass spectra of (Table I). The residue from this reaction (1060 ~ sulphate decomposition products is that it is a showed major 2PbO'PbSO4 and minor 4PbO. fragment ion of the decomposition of SO2 within PbSO4, CuO, and CH20, The reaction to 1060 ~ the mass spectrometer. This would explain the could thus be represented in general terms by the similarity between SO and SOz profiles in fig. 2 (see equation: also Lombardi, 1984, fig. 5). Equation (3) can thus be modified as follows: 6[PbO- PbSO4] + 3PbSO4 -* 5[2PbO.PbSO,] + 3SOz~ +-~O2T + SOT. (3) 6[PbO PbSO4] + 3PbSO4 -* 512PbO. PbSO4] + 4SO 3 This takes account of the relative ratios of SO 2 to SO evolved and also the fact that 2PbO.PbSO 4 dissociation was the dominant lead oxysulphate phase identified > 800 ~ in the residue; the theoretical weight loss for this 4SO 2 + 202 reaction is 10.1 ~o. However, equation (3) can only partial dissociation in I be an approximation of the process as (i) 4PbO. mass spectrometer ~. PbSO4 was identified as a minor phase in the 4SO § . (4) residue; (ii) although primarily due to breakdown of the sulphate anion, this second weight loss must The result of this second decomposition to have involved some evaporation of PbO also, as 1060 ~ was to increase the amount of PbO relative yellow deposits were noted in the furnace and gas to PbSO4 in the decomposition residue, as shown line at the end of the run; (iii) a thermal reaction by the line E-F in fig. 3. The position of point F is yielding SO2 + SO + 02 simultaneously seems only approximate, but can be justified by reactions thermodynamically unlikely. undergone by the residue on cooling (see next With regard to point (iii) above, there is some section). controversy over the exact species evolved during Reactions on cooling from 1060 ~ Major 2PbO" thermal decomposition of the sulphate anion, PbSO4 and minor 4PbO. PbSO4 were identified in which has arisen both because of the temperature- the residue cooled from 1060 ~ On the phase dependent dissociation SOa -* S02 + (Stern diagram in fig. 3, this would place its composition and Weise, 1966) and differences in experimental just to the left of the vertical composition line conditions under which the thermal decomposition 2PbO.PbSO4, i.e. point F (at a temperature of has been investigated. Truex et al. (1977) studied the 1060 ~ All the peaks on the DTA cooling curve decomposition of aluminium sulphate in a flow agree with this positioning. The peak at 930 ~ reactor system in the range 500-700 ~ They found represents the transition from true to incongruent that SO 3 comprised 97~ of the emitted sulphur melt (point G) and that at 863 ~ the transition oxides; the small amount of SO2 identified was from incongruent melt to solid (point H). A very attributed to the extremely short residence time of small peak at 662 ~ could be due to minor the SO3 in the system (1 see) before collection, decomposition of ct-2PbO.PbSO4 into PbO' which thus minimized the SO3-*SO2+ 2 PbSO4 and 4PbO-PbSO4 (point I; see also Bill- dissociation. Collins et al. (1974), using mass hardt, 1970), but the bulk of the ~-2PbO-PbSO4 spectrometry, showed that for anhydrous CuSO4 changes to the fl-form, as represented by the peak and A12(SO~)3, SO3 was the initial gaseous de- at 447 ~ (point J). composition product. The SO 3 dissociated to SO 2 and O2; SO also appeared on the mass spectra due Summary and conclusions to the dissociation SOz-*SO + +O- within the ionization chamber of the mass spectrometer. No On heating, caledonite decomposes between 300 SO3 was detected in mass spectra of gases evolved and 400 ~ to give a mixture of PbO-PhSO4, from alunite, however, and it was suggested by PbSO4, and Cu(I D and Cu(I) oxides. CO2 and Collins et aI. that SO2 and SO were primary hydroxyl water are liberated during this reaction; products of the decomposition of this mineral. some oxygen also appears to be generated but is not 526 D. J. MORGAN ETAL. physically released from the decomposition residue Collins, L. W., Gibson, E. K., and Wendlandt, W. W. until higher temperatures are attained. (1974) The composition of evolved gases from the Subsequent reactions undergone by the de- thermal decomposition of certain metal sulfates. Ibid. 9, composition residue are governed only by the 15-21. nature of the lead sulphate and oxysulphate com- Giacovazzo, C., Menchetti, S., and Scordari, F. (1973) The of caledonite, CuzPbs(SO4)3CO 3 ponents, with temperatures for phase transition (OH)6. Acta Crystallogr. B29, 1989-90. and melting reactions agreeing well with those Gray, N. B., Stump, N. W., Boundy, W. S., and Culver, predicted from the PbO-PbSO4 phase diagram. R. V. (1967) The sulfation of lead sulfide. Trans. Breakdown of the sulphate anion commences Metallurgical Soc. AIME, 239, 1835 40. at 880 ~ with SO2, SO, and O2 appearing on Greenwood, N. N., and Earnshaw, A. (1984) Chemistry of the mass spectra of the decomposition products. the Elements, p. 824. Oxford, Pergamon Press. The initial gaseous decomposition product was Henmi, H., Hirayama, T., Mizutani, N., and Kato, probably SO3, which dissociated completely to M. (1985) Thermal decomposition of basic copper SO 2 and Oz at these high temperatures. The SO is carbonate, CuCO 3. Cu(OH)~ - HzO, in carbon dioxide atmosphere (0 50 atm.). Thermochim. Acta, 96, 145-53. an artefact arising from partial fragmentation of Livingstone, A., and Sarp, H. (1984) , a SO 2 within the mass spectrometer. Above 1000 ~ new mineral from Leadhills, Scotland, and Saint-Prix, some evaporation of PbO occurs but CuO and France--a polymorph of leadhillite and . Cu20 are still present in products cooled from Mineral. Mag. 48, 277-82. 1060 ~ Lombardi, G. (1984) Thermal analysis in the investigation Whilst linking evolved gas detectors to DTA- of zeolitized and altered volcanics of Latium, Italy. Clay TG equipment is by no means a new technique in Minerals, 19, 789 801. mineral thermal analysis, the present investigation Milodowski, A. E., and Morgan, D. J. (1984) Thermal has confirmed both the need for this approach reactions of leadhillite Pb4SO4(CO3)z(OH)2. Ibid. pp. 825 41. when dealing with minerals showing complex Moenke, H. (1974) Mineral Spektren, 1. Akademie-Verlag, volatile evolution behaviour and also the flexibility Berlin. afforded by the quadrupole mass spectrometer Morgan, D. J. (1977) Simultaneous DTA-EGA of compared to a system employing separate detectors minerals and natural mineral mixtures. J. Thermal for each volatile (cf. Morgan, 1977; Milodowski and Anal. 12, 245-63. Morgan, 1984). Russell, J. D., Milodowski, A. E., Fraser, A. R., and D. R. (1983) New IR and XRD data for leadhillite of ideal Acknowledgements. Dr J. D. Russell, Macaulay Institute, composition. Mineral. Mag. 47, 371 5. is thanked for providing IR data. S.St.J.W. acknowledges -Fraser, A. R., and Livingstone, A. (1984) The infra- the opportunity to work in the laboratories of the red spectra of the three polymorphs of Pb4SO4(CO3)2 Geochemistry Directorate (BGS) and Stanton Redcroft (OH)/(leadhillite, susannite, and macphersonite). Ibid. during sabbatical leave in 1984. D.J.M. and P.H.A.N. 48, 295 7. publish with the permission of the Director, BGS (NERC). Stern, K. H., and Weise, E. L. (1966) High-temperature properties and decomposition of inorganic salts. Part 1. Sulfates. NSRDS-NBS 7. US Dept. of Commerce, REFERENCES Washington, USA, 38 pp. Billhardl, H. W. (1970) New data on basic lead sulfates. Truex, T. J., Hammerle, R. H., and Armstrong, R. A. (1977) J. Electrochem. Soc.: Solid State Sci. 117, 690-2. The thermal decomposition of aluminum sulfate. Bugajska, M., and Karwan, T. (1979) Characteristics Thermoehim. Acta, 19, 301-4. of the oxidation products of spherical samples of lead sulphide in the temperature range 773-1023 K. [Manuscript received 11 December 1985; Thermochim. Acta, 33, 41 50. revised 28 January 1986]