PLEASE 00 Nor REMOVE FRCM LIBRARY

Bureau of Mines Report of Investigations/1987

Analysis of the Oxidation of , , , , , and

By G. W. Reimers and K. E. Hjelmstad

UBRARY ""OKANERESEARC H CENTER .... RECEIVED

OCT 021 1981

US BUREAU Of MINES t 315 MONTGOMBlY ",'IE. SPOICANl,. WI>. 99201

UNITED STATES DEPARTMENT OF THE INTERIOR Report of Investigations 9118

Analysis of the Oxidation of Chalcopyrite, Chalcocite, Galena!' Pyrrhotite, Marcasite, and Arsenopyrite

By G. W. Reimers and K. E. Hjelmstad

UNITED STATES DEPARTMENT OF THE INTERIOR Donald Paul Hodel, Secretary

BUREAU OF MINES David S. Brown, Acting Director Library of Congress Cataloging in Publication Data :

Reimers, G. W. (George W.) Analysis of the oxidation of chalcopyrite, chalcocite, galena, pyr­ rhotite, marcasite, and arsenopyrite.

(Bureau of Mines report of investigations; 9118)

Bibliography: p. 16 .

Supt. of Docs. no.: 128.23: 9118.

1. . 2. Oxidation. I. Hjelmstad, Kenneth E. II . Title. III. Series: Reportofinvestiga­ tions (United States. Bureau of Mines) ; 9118.

TN23.U43 [QE389.2] 622 s [622'.8] 87-600136 CONTENTS

Abst 1 Introduction ••••••••• 2 Equipment and test • .,.,.,.,.,.,., .•. .,., ...... ,.,.,.,., ... ., . ., ...... , .... ., ...... , .. ., . .,. 3 Samp Ie ., .. ., ., ., ., ., ., ., ., ., ., ., .. ., .• ., .. ., ...... , •.••• ., . ., ., ..•. ., ., ., . ., ., ., ., • ., ., ., ., • 4 Results and discussion .. ., . .,.,.,...... ,., ...... , ., .... " . ., . .,.,., . .,., . .,., ,.,.., ., . ., 5 -type ., . .,.,.,.,.,., ...... ,., .... ., ...... ,., ... .,., ...... ,.,.,.,.,.,.,.,.,., ... .,. 5 Thermal ric analyses., • ., ...... , .... ., . .,.,.,., . .,.,., .. .,.,., . .,.,., . .,.,.,.,.,.,.,.,., ... .. 5 Chalcopyrite ... ., .. .,.,.,., . .,., .. ., . .,., It.,., • ., .. .,.,.,.,.,., •• .,.,.,.,.,.,.,., • .,., .. ., .. ., •• ., • .,., • .,.,.,., 5

Chalcoci tee ••• ., .•. ., .•. .,., ., .. ., ., ., •• ., ., ., .. ., ., ., • ., ., ., ., .,., ., ., ., ., .. ., ., .• ., ., . ., ., . ., . ., ., It •• ., .... 9 Galena .•.•••••...••...... ••.•.•..••••...••..•.•...•.••.....•...•...... 9 Differential thermal 10

Di s cus s ion . ., • ., ., ., ., .. ., .. ., ...... II • II ...... 10 11

ric analyses ...... II •••• II • •• II 11 Pyrrhoti tea ...... ,. .. 11

Marcasi tea ...... II ...... 11 Arsenopyrite ••••••••••••• 12 Differential thermal analyses ..•.. " ...... " ...... ,. 13 Di s cus s ion ...... 14 Summa ry and conclusions .••••••..•...••••.•••••...•...... ••.•...... 14 Re f e rences •...... •...... •••.•...... •.....••.•...•...... •...... 16

ILLUSTRATIONS

L Thermal gravimetric is illustrating effect of milling time on oxi- dation of chalcopyrite A in air ...... ,. ... . 7 2. Isothermal oxidation of chalcopyrite A at 290 0 C, with air and air contain- 40 pct water vapor and with oxygen and oxygen 40 pct water vapor. • • • ..... • • •• •• •• •• •••• •• ••• ••••• •••• •• •• ••• .... •••. •• •••• •• •••...... 7 3. Thermal tric of B oxidized with oxygen con- taining pct water vapor, chalcopyrite C oxidized with air, chalcocite A oxidized with air, and oxidized with oxygen containing 40 water vapor ••••••• ., •• .,,,,, ...... " • • •• .. • •• .. ... •• •• ...... •• • .... • ••• •••• ••• 8 4. Isothermal oxidation cna~.copy'rlte B with air, chalcocite B with air, and with oxygen pet water vapor ••••••••••••••••••••••••.• 9 5. Differential thermal chalcopyrite B, chalcocite B, and milled for 120 min and oxidized with air •••••••••••••••••••••••••••••. 10 6. Thermal ric analyses of pyrrhotite A milled for 15 min, A milled for 120 ite B milled for 120 marcasite milled for 120 and milled for 120 min •••••••••••••••••••••••• 12 7. Isothermal test of marcasite milled for 120 min and oxidized with air 40 pct water vapor •••••••• 13 8. Differential thermal analyses of A, marcasite, and te milled for 120 min and oxidized with air 4.5 pct water vapor ••• 13

TABLES

1. Standard heats of reaction (8H~98) ••••••••••••••••••••••••••••••••••••••••• 3 7.. of s ulf ide ...... , ...... " ...... , .. . 4 3. Particle size and 6 -----

UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT cm centimeter kcal/mol kilocalorie per mol cm ~ cubic centimeter ).lm micrometer cm 3 /min cubic centimeter per minute mg milligram

°c degree Celsius min minute

°C/min degree Celsius per minute mm millimeter g gram pct percent h hour wt pct weight percent AN ALYSIS OF THE OXIDATION OF CHALCOPYRIT E, CHALCOCITE, GALENA, PYRRHOTITE, MARCASITE, AND ARSENOPYRITE

By G. W. Reimers 1 and K. E. Hjelmstad2

ABSTRACT

Conditions in the underground mine environment can cause self-heating of as a result of exothermic oxidation reactions, which may result in mine fires. This Bureau of Mines report describes thermal analyses of finely ground chalcopyrite, chalcocite, galena, pyrrhotite, marcasite, and arsenopyrite, to characterize their responses under oxidizing conditions. Thermal gravimetric analysis (TGA) and differ­ ential thermal analysis (DTA) were used, in the temperature range 100 0 to 500 0 C. Chalcopyrite showed an initial weight gain, due to formation of sul­ fates, a weight loss, caused by ignition of sulfide to iron , and a final weight gain, indicating the conversion of sulfide to sulfate. Chalcocite did not ignite in the temperature range studied but exhibited a rapid weight gain, because of the rapid conversion of to sulfate. Likewise, g a lena failed to ignite but underwent a steady weight gain when oxidized in a moist atmosphere, showing conver­ sion of sulfide to sulfate below the ignition point. Marcasite underwent a rapid weight loss, indicating ignition, at 200 0 C. Moisture in the oxidizing atmosphere lowered the ignition point of marcasite and arsenopyrite but not that of pyrrhotite. All the sulfides exhibited exothermic behavior at temperatures below the ignition point. l p hysical scientist. 2Geophysicist. Twin Cities Research Center, Bureau of Mines, Minneapolis, MN. INTRODUCTION

This report describes the results of spontaneous combustion in sulfide depos­ laboratory thermal analysis experiments its are not well understood. Thus, cur­ conducted to follow the oxidation of rent efforts by mine personnel to control chalcopyrite, chalcocite, galena, pyr­ spontaneous combustion problems, regard­ rhotite, marcasite, and arsenopyrite. less of how well intended, are not always These tests are part of a broader Bureau based on sound engineering and technical of Mines research program to develop principles. The intent of this research methods of controlling the oxidation of is to delineate the thermal behavior of sulfide ores in underground mines, thus sulfide exposed to oxidizing preventing spontaneous combustion mine conditions, thereby providing a back­ fires. ground of data that will enable mine per­ Previously, Ninteman (~)3 completed an sonnel to undertake fire prevention extensive literature review of the nature efforts with maximum effectiveness. and problems associated with the oxida­ In his monograph on chalcopyrite, tion and combustion of sulfide minerals Habashi (2) refers to work by Tafel and in underground metal mines. Sulfide oxi­ Greulich, who followed the oxidation of dation is an exothermic process that pro­ powdered chalcocite in air. They found duces heat, which, if not dissipated, that both cupric sulfate and ferrous sul­ causes a temperature rise within the min­ fate were formed slowly at about 350 0 C. eral mass. If the temperature begins to Ferrous sulfate reached its maximum form­ rise, a thermal runaway situation may ation at 400 0 C; at 500 0 C, its rate of develop that will cause the sulfide and formation equaled its rate of decomposi­ other combustibles to burn. tion. Cupric sulfate reached its maximum Reactions similar to those that cause formation at 550 0 C. Above 550 0 C, it spontaneous combustion in underground gradually decomposed to copper oxysul­ mines can also be responsible for the fate. Habashl also ci ted the work of self-heating of sulfides during shipment. Meunier and Vanderpoorten in studying the Kirshenbaum (~) described many such inci­ oxidation of chalcopyrite cubes (2 to 3 dents, involving both rail and ship car­ cm 3 ) in the temperature range of 300 0 to riers of sulfide concentrates. His 900 0 C. At 330 0 C, they noted oxidation teview provided details of the conditions on the surface of the cube. At 500 0 C, that led to the oxidation and self­ the surface film was thick enough that heating of sulfide cargoes and, ultimate­ when the cube was sectioned the red outer ly, to combustion. Mention was also made layer could be identified as ferric of the high levels of toxic S02 generated oxide. Work by Razouk and others to during such events, which impeded fire study the oxidation of chalcopyrite by control efforts. TGA was also reviewed (2). Razouk pro­ The spontaneity and prolonged incipient posed a mechanism for the temperature stage (weeks or months) of these fires range of 350 0 to 400 0 C, suggesting that make their occurrence difficult to pre­ chalcopyrite oxidized to form ferric sul­ dict and to detect. Furthermore, these fate and cuprous sulfide. Between 450 0 fires often start in abandoned, backfill­ and 600 0 C, the cupric oxide and ferric ed, and/or caved areas where abundant sulfate reacted to form cupric sulfate fuel is present but where access for fire and ferric oxide. This interpretation fighting is difficult or in some cases seems to contradict that of other re­ impossible. Since detection and suppres­ searchers, but it does propose the forma­ sion are so difficult, a high priority tion of sulfates during oxidation. is placed on fire prevention. However, Larger scale Bureau research experi­ the chemical reactions giving rise to ments in progress on the low-temperature (90 0 to 200 0 C) oxidation of have 3Undcrlined numbers in parentheses re­ shown that ferrous sulfate was formed on fer to items in the list of references at the surface of pyrite in the presence the end of this report. of moist air. In these same tests, the 3

temperature of the te increased sub­ TABLE 1. Standard heats of reaction stantia during oxidation. However, in (6H;98) similar tests conducted in the absence of water vapor, only a slight temperature Oxidation rise was noted and the presence of sul­ fate on the surface was not Chalcocite •••••• CuO,CUS04 •••••• detected X-ray diffration sis. The role that moisture has in pro­ the oxidation of pyrite to sulfate is not understood. However, the vapor may assist the oxygen in Galena •••••••••• PbS04...... -197.0 the sulfate formed on surface. and Herbert Marcasite ••••••• Fe ,S02 •••••• -200.4 4, among others, have also suggested FeS04. , ••••• -253.7 the formation of sulfates contri­ butes to the of sulfides. Fe ,S02 •••••• -132.1 Table 1 lists the molar heats of re- FeS04, -178.9 action of a number of common sulfides to form and sulfates Only the Pyrite •••••••••• Fe ,S02...... -199.4 anhydrous forms of the sulfates are re­ -252.7 , as the degree of hydration would be influenced the level and ite, marcasite, and the temperature at which the sulfates were measured in the temperature range were formed. It should be noted that the 100 0 to 500 0 C. This weight in­ reactions responsible for forming the formation, the XRD analysis of the oxi­ sulfates more energy ) dized products. and the results obtained do the reactions that precede from DTA were used to characterize the and form oxides. However, oxidation of the sulfide min­ sulfide self-heat is only caused erals. These data, together with pre­ through the formation of sulfates is vious published data on the oxidation probably oversimpl a complex pro­ of pyrite will be used to develop cess. Previous research indicates that a models that the kinetics and number of other factors and reactions, mechanisms of sulfide oxidation reac­ as the presence of free , can tions. A detailed unders of these contribute to sulfide self (~-2). reaction and mechanisms is In this inves the weight optimized oxidation spontaneous combustion chalcocite, fires.

EQUIPMENT AND TEST PROCEDURE

The thermal used in plotter. Modifications to the TGS-2 ana­ this consisted of the included the installation of a Perkin Elmer 4 components~ a TGS-2 ther­ tape around the tube shell ric analyzer and a DTA 1700 that surrounds the furnace assembly, in differential thermal , in con- order to reduce moisture condensation on junction with a 7/4 thermal the balance hangdown wire The sis controller. Data from the analyzers balance was also continually with were continuously fed to a thermal ana a few cubic centimeters per minute of sis data station (TADS), and the test re­ helium to prevent corrosion from moisture sults were on paper by a TADS-l gas were used to products does supply the nitrogen, air, and oxygen used the Bureau of in the tests. In the TGA tests, a flow Mines. rate of 500 cm 3 /min was used for purging 4

the system with nitrogen and for supply­ the nitrogen was replaced with the oxi­ ing air or oxygen during the oxidation dizing gas, and the sample was heated at tests. To add moisture to the system, 25 0 Cimino A starting temperature of the oxidizing gas (air or oxygen) was 100 0 C was used to prevent condensation bubbled into water contained in a heated of moisture on the sample when water flask and routed to the TGA furnace vapor was added to the oxidizing-gas through an insulated tube to prevent con­ stream. Selected sulfide samples were densation. The water content of moist also tested isothermally under oxidizing gas streams was verified over measured atmospheres at fixed furnace tempera­ time intervals by weighing the water col­ tures, using the TGA system. These tests lected in drying tubes filled with a dry­ were run at temperatures near the igni­ ing agent. When moisture was added to tion point or near the point where rapid the gas stream, the flow rate of the oxi­ weight change occurred. The same system dizing gas was reduced to compensate for loading procedure was used, and the sul­ the volume of the added water vapor, fides were heated in nitrogen until the keeping the total flow rate of 500 desired test temperature was reached. cm 3 /min. A flow rate of 40 cm 3 /min was DTA tests used a sample weighing about used in tests with the DTA equipment. 30 mg. After the sample pan was filled, For the TGA tests, approximately 40 mg the furnace assembly was locked into po­ of sulfide was used. The sample was sition, and the assembly was purged with loaded into the balance pan, and the sys­ nitrogen while the sample was heated to tem was closed by raising the furnace as­ 100 0 C. At 100 0 C, the oxidizing gas was sembly into position. The system was allowed to flow into the assembly, and then purged with nitrogen while the sam­ the thermal behavior of the sample was ple was heated to 100 0 C. At this point, monitored at a heating rate of 25 0 Cimino

SAMPLE PREPARATION

The sulfide minerals used in this in­ further size reduction in a Bleuler-brand vestigation were specimens purchased from mill. Table 2 lists the partial chemical a commercial supplier. The primary analysis of the various samples. XRD crushing of the minerals involved break­ analysis was used to confirm their com- ing with a sledge hammer, followed by positions. Pyrite was present as a minor roll-crushing in several stages, along constituent in the samples designated with sorting to remove visible gangue. "chalcopyrite A" and "pyrrhotite A" and After being crushed to minus 6 mm, the in the marcasite and arsenopyrite. sulfides were split into 20-g lots for

TABLE 2. - Partial chemical analysis of sulfide samples, weight percent

Element Chalcopyri te Chalcocite Galena Pyrrhotite Marca- Arseno- A B C A B A B site pyrite Fe ••••••••• 37.4 33.3 30.8 5.3 5.6 0.26 45.6 44.0 45.6 28.5 Cu ••••••••• 21. 2 30.7 33.3 71.4 68.0 .24 .10 • 14 ND ND P b ••••••••• ND ND ND ND ND 82.2 ND ND ND ND S •••••••••• 31.8 33.0 31.6 20.7 19.0 12.1 30.2 24.2 50.7 13.4 AI ••••••••• .68 < .20 < .20 .31 .37 <.20 <.20 2.4 ND .31 As ••••••••• ND ND ND ND ND ND ND ND ND 27.6 Ca ••••••••• <.50 < .50 <.50 <.50 <.50 < .50 1.8 1.9 ND <.30 Mg ••••••••• .42 < .01 < .1 < • 1 < • 1 < • 1 .16 2.0 ND < .10 Ni ••••••••• ND ND ND ND ND ND .08 3.8 ND ND S i ••••••••• 1.7 .48 1.1 .34 2.2 < .10 .77 8.5 ND 9.7 Zn ••••••••• ND ND "ND ND ND < _ 01 ND ND ND ND ND Not determlned. 5

In previous tests, it was found that galena samples that were ground for 15 the reactivity of sulfides increased as and 120 min. The mean and median parti­ the particle size decreased; therefore, cle sizes for the various samples are in­ the samples in this study were ground in dicated in the table footnotes. A Leeds the Bleuler mill to fine sizes, to reduce and Northrup Microtrac particle size mon­ the time required to conduct the experi­ itor was used to collect the size inform­ ments. This was achieved by milling 20-g ation on the ground sulfides. Heptane batches for either 15, 30, 60, or 120 was evaporated from the samples before min. To decrease tne possibility of oxi­ they were added to the water carrier used dation during this final size-reduction with the monitor. The average particle step, the samples were milled and stored size of these sulfides was in the range in heptane. Just prior to thermal analy­ of 5.8 to 8.7 ~m and 3.4 to 4.8 ~m for sis, a portion of the sample slurry was the 15- and 120-min grinding times, re­ withdrawn from the storage bottle, dried spectively. Under the assumption that in air at room temperature, and then grinding would produce similar sizes in loaded into the analysis units. all samples, the remaining milled samples Table 3 illustrates the size distribu­ were not measured. tion of chalcopyrite, chalcocite, and

RESULTS AND DISCUSSION

ORE-TYPE SULFIDES increase the sample weight gain prior to ignition. Several sulfides that are commonly Figures 2A and 2B illustrate the test mined for their metal content were exam­ results of oxidizing the chalcopyrite A ined first: two copper sulfides, chalco­ sample with the finest particles at pyrite and chalcocite, and the lead 290 0 C. An increase in weight was ob­ sulfide galena. served when water vapor was added to either air or oxygen. Figure 2A also il­ Thermal Gravimetric Analyses lustrates that an increased rate of weight gain can be initiated when mois­ Chalcopyrite ture is added to the atmosphere even after exposing the sample to dry air for Chalcopyrite was the first ore-type 60 min. When this test was repeated at mineral selected for examination in the 300 0 C, the sample immediately ignited series. It is a sulfide of copper and when exposed to the moist-air oxidizing iron and has a typical formula of CuFeS2 atmosphere. Ignition was verified by the and a tetragonal (9). rapid weight loss and the glow observed Figure 1 shows the results of TGA tests through the glass wall of the furnace. conducted on chalcopyrite sample A, mill­ After cooling, the sample was noted to ed for 15, 30, 60, and 120 min, in the have the red color characteristic of he­ temperature range of 100 0 to 400 0 C in matite. XRD analysis of the sample after dry air. In these tests, the ignition oxidation confirmed that ferric oxide temperature, shown as the point where a () had been formed. rapid weight loss occurred due to the The TGA results for chalcopyrite B (15- oxidation of iron sulfide to the oxide, and 120-min grinds) oxidized in oxygen decreased with increasing milling time containing 40 pct water vapor are shown (decreasing particle size). Replacing in figure 3A. This figure illustrates air with oxygen had little effect on the that over the temperature range of 100 0 ignition point. This indicates that the to 400 0 C the finer material ignited at a oxygen content of the air atmosphere was lower temperature. Similar effects were not limiting the results illustrated in noted when these samples were oxidized in figure 1. The addition of moisture to air and in air containing moisture. The the oxidizing atmosphere was found to TGA data also indicate a weight gain in 6

TABLE 3. - Particle size analyses of chalcopyrite A, chalcocite A, and galena

Sample cumulative Sample volume, Particle size volume, pct undersize pct fraction, ]..1m 151I1in 1201Ili n 15-min 120-min grind I grind grind grind CHALCOPYRITE A' Minus 42.21 plus 29.85 •••••••••••.••••• 100.0 100.0 4.2 0.0 Minus 29.85 plus 21.10 ••.•••••••••••••• 95.8 100.0 7. 1 .0 Minus 21. 10 plus 14.92 •••••••••.••••••• 88.7 100.0 5.3 .0 Minus 14.92 plus 10.55 •••••••••••••••.• 83.3 100.0 7.5 .0 Minus 10.55 plus 7. 46 ••••••••••••••••• 75. 7 100.0 4.8 .0 Minus 7.46 plus 5. 27 ••••••••••••••••• 70.9 100.0 20.6 15. 1 Minus 5.27 plus 3. 73 ••••••••••••••••• 50.2 84.8 25.0 30.2 Minus 3.73 plus 2. 63 ••••••••••••••••• 25.2 54-.6 7.9 18.9 Minus 2.63 plus 1. 69 ••••••••••••••••• 17.2 35.6 9.0 11. 3 Minus 1. 69 plus 1. 01 ••••••••••••••••• 8.2 24.2 6. 1 13.8 Minus 1. 01 plus O. 66 ••••••••••••••••• 2. 1 10.4 1.5 7.4 Minus 0.66 plus O. 43 ••••• • •• ••• • • •••• .5 3.0 .5 2.2 Minus 0.43 plus O. 34 ••••••••••••••••• .0 .8 .0 .8 Minus 0.34 plus O. 24 •••••••••••••.••• .0 .0 .0 .0 CHALCOCITE A2 Minus 42.21 plus 29.85 ••••••••••••••••• 100.0 100.0 0.0 0.0 Minus 29.85 plus 21. 10 ...... 100.0 100.0 .0 .0 Minus 21. 10 plus 14.92 ••••••••••••••••• 100.0 100.0 .0 .0 Minus 14.92 plus 10.55 ••••••••••••••••• 100.0 100.0 7.8 .0 Minus 10.55 plus 7. 46 ••••••••••••••••• 92.1 100.0 13.5 3.4 Minus 7.46 plus 5. 27 ••••••••••••••••• 78.6 96.5 30.8 31.3 Minus 5.27 plus 3. 73 ••••••••••••••••• 47.8 65.2 31.9 39.2 Minus 3.73 plus 2. 63 ••••••••••••••••• 15.8 26.0 11.6 16.7 Minus 2.63 plus 1. 69 ...... 4.2 9.2 1.9 5.4 Minus 1. 69 plus 1. 01 ...... •...... 2.2 3.8 1.6 2.3 Minus 1. 01 plus O. 66 ••••••••••••••••• .5 1.4 .5 1.4 Minus 0.66 plus O. 43 ••••••••••••••••• .0 .0 .0 .0 GALENA3 Minus 42.21 plus 29.85 ••••••••••••••••• 100.0 100.0 0.0 0.0 Minus 29.85 plus 21.10 ••••••••••••••••• 100.0 100.0 5.5 .0 Minus 21.10 plus 14.92 ••••••••••••••••• 94.5 100.0 6.9 1.8 Minus 14.92 plus 10.55 ••••••••••••••••• 87.6 98.2 10.5 1.9 Minus 10.55 plus 7. 46 ••••••••••••••••• 77.1 96.3 14.5 5.3 Minus 7.46 plus 5. 27 ••••••••••••••••• 62.6 91.0 27.8 28.3 Minus 5.27 plus 3. 73 ••••••••••••••••• 34.9 62. 7 24.3 32.4 Minus 3.73 plus 2. 63 ••••••••••••••••• 10.5 30.2 5.3 14.2 Minus 2.63 plus 1. 69 ••••••••••••••••• 5.2 16.0 2.9 9.9 Minus 1. 69 plus 1. 0 1 ••••••••••••••••• 2.3 6.0 2.3 3. 7 Minus 1.01 plus O. 66 ••••••••••••..••• .0 2.3 .0 1.2 Minus 0.66 plus O. 43 ••••••••••••••••• .0 1.1 .0 1.1 Minus 0.43 Jllus O. 34 •.•••.•..•.•.•..• .0 .0 .0 .1 'Mean and median particle sizes: for the 15-min grind, 8.68 and 5.31 ]..1m, respec- t i ve ly; for the 120-min grind, 3.40 and 3.42 ]..1m, respectively. 2Mean and median particle sizes: for the 15-min grind, 5.78 and 5.55 ]..1m, respec- tively; for the 120-min grind, 4.72 and 4.68 ]..1m, respectively. 3Mean and median particle sizes: for the 15-min grind, 8.44 and 6.27 ]..1m, respec- tively; for the 120-min grind, 4.83 and 4.43 ]..1m, respectively. 7

105r------.------.------.------.------,,------.

0

0 0- 100 -~ W I- cr >- 0... 0 KEY u --l A 15 min

90 ~ ______-L ______L______L ______L______L ______~ 100 150 200 250 300 350 400 TEMPERATURE, °C

FIGURE 1.- Thermal gravimetric analysis illustrating effect of milling time on oxidation of chalcopyrite A in air.

the sample prior to the rapid weight 10s8 104,----,-----,----,-----,----.-----,----, associated with ignition of the sulfide. A From these curves, the occurrence of 102 B several reactions can be postulated. The Moisture added""V first reaction, as evidenced by the A slight gain in sample weight prior to the IOO~ KEY rapid weight loss, was probably the par­ A Air tial oxidation of the iron sulfide frac­ B Air containing 40 pet water vopor U tion of the chalcopyrite sample to the 0- 98 i sulfate. Although iron sulfate was not detected by XRD analysis of the partially ~ 108 oxidized samples, its presence was con­ 0:: B >- firmed by dissolving the soluble sulfate g, 106 u from partially oxidized samples and then ...J

110 I I I I 115 A Cha Icopyrl te B 8 Chalcopyrite C

+- u KEY u Ar: a. -Q. 105 - 110 - A 15-min grind ~ -~ - 8 120-min grind w 8 ·w I- ;.- .-f l- 100 105 e::: - e::: ~ >- Cl.. Cl.. 0 0 U U -.J -.J r- <{ 100 <{ 95 V - ::x: I u U V, 90 I I I I 95

120

u C Chalcocite A -Q. 104 D Galena ~ 110 u -Q. W 102 I- -~ u 0

FIGURE 3.-Thermal gravimetric analyses of (A) chalcopyrite B oxidized with oxygen containing 40 pct water vapor, (8) chalcopyrite C oxidized with air, (C) chalcocite A oxidized with air, and (D) galena oxidized with oxygen containing 40 pct water vapor. of iron oxide, copper sulfate, and cupric moisture was added to the oxidizing atmo­ oxide. sphere, an increase in the rate of weight Samples of chalcopyrite B that were gain was noted at temperatures below the ground to the finest paticle size and ignition point. then oxidized in air for 60 min at con­ TGA results for sample chalcopyrite C, stant furnace temperatures of 280°, 310°, ground for 15 and 120 min and oxidized in and 320° C are shown in figure 4A. The dry air, are shown in figure 38. The rate of sample weight gain increased as finest fraction of chalcopyrite C ignited the test temperature increased. At at a lower temperature than was noted for 320° C, the sulfide was found to ignite, samples A and B. It also underwent a as indicated by the rapid initial weight smaller weight loss at ignition, which loss. Following the weight loss, the was probably a result of its lower iron chalcopyrite gained weight, indicating content. Sample C also exhibited a ten­ the conversion of copper sulfide to dency to gain weight prior to the rapid copper sulfate. In similar tests where weight loss that occurred at ignition. 9

120 weight gain associated with copper sul­ A Cha Icopyr i Ie 8 fate formation . XRD analysis of the ~ e a. 115 chalcocite after ox idation confirmed the i r presence of copper sulfate. The addition w 110 KEY of moisture to the oxidizing gases had t:: A 280° C only a small effect on the behavior of 0:: >- B 310°C the samples during TCA. The effect of D- 105 0 f0- e 320°C B u moisture in the oxidizing atmosphere was -.I « I A also studied at a constant furnace tem­ I 100 perature of 300° C. When the chalcocite U -- V ground to the smallest particle size was 95 I oxidized for 60 min in air containing 40 pct water vapor , it gained about 4 wt 120 pet. This c ompares with a gain of 2 wt BChalcocite 8 pct for samples oxidized in dry air. KEY ~ 115 E TCA tests were also conducted on a a. A 300°C B 320°C o second sample of chalcocite (sample B). e 340°C The results of ~he analysis were very 110 o 350°C e W'" similar to those obtained for sample An t: E 360° C u At test t e mperatures ranging from 300° to 0 105 B u 360° C, weight gains were noted. Results -.I « A of oxidizing sample B, ground to the I u 100 smallest particle size, in dry air at fixed furnace temperatures ranging from 95 300° to 360° C, are shown in figure 4B. This figure illustrates the weight gain due to the formation of copper sulfate as 106 a function of temperature . It also indi­ e Galena cates that if chalcocite underwent slow KEY ~ 104 A 250°C heating in an ox idizing atmosphere, the a. B 290°C. sulfide would tend to be converted to the e 300°C ~ B sulfate rather than igniting to form an

Calena (PbS) , a lead sulfide , is the 98 most common ore of lead and occurs as 0 10 20 30 40 50 60 cubic crystals (1). TCA results for the TIME. min galena sample oxidized in oxygen contain­ FIGURE 4.-lsothermal oxidation of (A) chalcopyrite B with ing 40 pct water vapor are shown in air, (8) chalcocite e with air, and (C) galena witn oxygen con· figure This figure illustrates the taining 40 pet water vapor. 3D. higher r eac tivity of the fine r samples . As the galena was heated to 400° C, a Chalcocite gradual gain in weight was seen. The abrupt wei~ht change common to the copper Chalcocite is coppe~ sulfide having an sulfide samples was absent. Substituting orthorhombic crystal structure and the ail' for oxygen had no measurable effect formula CU2S (9). TCA results for chal­ on the weight change results; however, cocite A ground for 15 and 120 min are the presence of moisture promoted the shown in figure 3C. In this test, the weight gain. Results of tests conducted sulfide was oxidized in dry air, and the on the finer sample of galena at constant results indicate that in the range of f urnace temperatures of 250°, 290°, and 340 0 to 380 0 C the sample underwent a 300 0 C, with oxygen containing 40 pct rapid weight gain, consistent with the moisture , are shown in figure 4C . XRD 10

analysis of the sample oxidized isotherm­ Examination of the DTA results (fig. 5) ally at 300 0 C confirmed that lead sul­ indicates that exothermic behavior for fate was formed. both of the copper sulfides was initiated at about 1500 C. The exothermic peak for Differential Thermal Analyses the chalcopyrite is at about 310 0 C; the exothermic peak for the chalcocite is at DTA was also conducted on the various about 370 0 C. These values are fairly types of ore sulfides. These tests were close to the temperatures at which rapid conducted in an atmosphere of dry air and weight change was first noted during TGA. spanned a temperature range from 1000 The behavior of galena indicated a steady to 500 0 C. All of the sulfides were increase in exothermic behavior until the ground for 120 min prior to testing. peak occurred at about 440 0 C. This is consistent with the steady gain in weight

5 r----,----,---,---,---r---r---,---, measured by TGA. A Chalcopyri te B Discussion

~ The results of the thermal analysis ~ n:: experiments conducted on chalcopyrite, ~ 0 chalcocite, and galena samples show that I­o increasing the grinding time, which X w should be synonymous with generating finer particles, increased the reactivity of the sulfides and caused them to ignite -5 or undergo weight change at lower temper­ u atures. Because the chalcopyrite samples 0 W 5 contained about 50 pct iron sulfide, the n:: ::;) B Chalcocite B evidence of ignition was based on the «I- rapid weight loss during TGA and the red n:: w color of the sample after oxidation. In u. ~ addition, the presence of ferric oxide in W I- the oxidized samples was verified by XRD 0 --l analysis. In the case of the chalcocite

-5 L-__~ ____L- __~ __~ __L- __L- __~ __~ scanned over the temperature range of 100 150 200 250 300 350 400 450 500 the experiment and in isothermal tests. TEMPERATURE,oC This weight gain of the chalcopyrite sam­ ple prior to ignition was assumed to be FIGURE 5.-Differential thermal analyses of (A) chalcopyrite B, (8) chalcocite B, and (C) galena milled for 120 due to the formation of sulfates. How­ min and oxidized with air. ever, as the weight gain was small, the 11

presence of sulfates was not detected by oxide. This sulfide was also tested iso­ XRD analysis. Further verification of thermally at 330 0 C, and XRD analysis of sulfate formation by other methods and the sample after oxidation for 2 h in­ the role water vapor has on promoting its dicated that partial conversion to fer­ formation are areas for further research. rous sulfate had occurred. Results of TGA of pyrrhotite B milled NON-ORE-TYPE SULFIDES for 120 min are shown in figure 6C. As observed for the previous pyrrhotite sam­ Sulfides that are considered gangue or ple, the presence of water vapor had lit­ byproducts were examined next: pyrrho­ tle effect on the ignition point of this tite, marcasite, and arsenopyrite. sample. Water vapor did, however, pro­ mote the weight gain of this sample at Thermal Gravimetric Analyses lower temperatures. Ferrous sulfate was also detected in this sample after iso'" Pyrrhotite thermal oxidation in air containing water vapor. Pyrrhotite was the first non-ore-type mineral examined in this test series. Marcasite Pyrrhotite is expressed as Fen-ISn, with n ranging from about 5 to 16. This iron TGA conducted on the marcasite sample crystallizes in the hex­ indicated that this was the most reactive agonal system (~). Results of tests con­ sulfide tested to date. Marcasite is ducted on pyrrhotite A, milled for 15 often called white iron pyrite and has min, in the temperature range of 100 0 to the same chemical formula, FeS2, as py­ 450 0 C, are shown in figure 6A. The sam­ rite. It is the orthorhombic dimorph of ple weight change was monitored in nitro­ pyrite and is considered less stable (~). gen, air, and air containing 40 pct water Figure 6D illustrates the TGA results for vapor. For this pyrrhotite sample, the a sample ground for 120 min. It shows presence of water vapor failed to have an the weight changes during oxidation of effect on the ignition point of the sam­ the sulfide in air and in air containing ple or the sample weight change at tem­ 40 pct water vapor, as well as under in­ peratures below 300 0 C. . The sample ert conditions. The presence of water weight loss while heating in nitrogen was vapor in the oxidizing atmosphere had a likely due to the volatilization of sul­ strong influence on the ignition point of fur. (The information on the sample this sulfide. It was found to lower the weight change in a nitrogen atmosphere ignition point to 257 0 C, which was 65 0 C was included, because this group of min­ below the value obtained in dry air. XRD erals tended to exhibit a more pronounced analysis of the samples at the completion weight loss during heating than was ob­ of testing indicated that the sulfide had served for the copper and lead sulfides. been converted to ferric oxide. A similar observation at higher tempera­ The results of isothermal tests con­ tures was reported by Chaubal and Sohn ducted on marcasite milled to the finest (~). ) particle size and oxidized with air con­ Figure 6B illustrates the results of taining 40 p~t water vapor are shown in TGA on pyrrhotite A ground to a finer figure 7. This figure illustrates the particle size by increasing the milling curves that were obtained for tests run time to 120 min. Decreasing the sulfide at 190 0 and 200 0 C. At 190 0 C, the sam­ particle size lowered the ignition point ple was found to slowly gain weight, and by about 50 0 C. Water vapor had little XRD analysis of the sample after oxida­ effect on the ignition point, and no ef­ tion for several hours indicated that the fect was observed on oxidation in the formation of ferrous sulfate had occur­ lower temperature regions. At the com­ red. At 200 0 C, the sample immediately pletion of the oxidation tests, the sam­ ignited. During thermal analysis of var­ ples were red, and XRD analysis confirmed ious sulfides, it has been observed, in the conversion of the sulfide to ferric the isothermal test, that the temperature 12

105 -g 105~~---.---.---.---.---.--~ A Pyrr hotite A fJ Pyrrhotite A ~=:;;;'\ A -~ 100 • 100' f---~--- u w a. C B ~

~ 95 b 95 . I w 0:: 0:: ~ >-90 L.....----L----'----'--...... - ...... -"""------' ~ 0... 0 105 I I I I C Pyrr hoti te B 0:: 0:: ) .- 100 0... o Marcasite

95 gI00 r-~~~~:;~~ ____j ~ w 90 90 ~ (f) 3 80 105 0:: - 90 A Air 0... o B Ai r containing 40 pet Z '.... ate r w 85 (f) C Nitrogen 0:: <:( 80 100 150 200 250 300 350400 ·450 500 TEMPERATURE,oC

FIGURE 6.- Thermal gravimetric analyses of (A) pyrrhotite A milled for 15 min, (8) pyrrhotite A milled for 120 min, (C) pyrrhotite B milled for 120 min, (D) marcasite milled for 120 min, and (E) arsenopyrite milled for 120 min.

rise produced when oxygen reacts with the pyrite, making it a sulfarsenide of iron, fresh sulfide surface induces ignition at FeAsS. This mineral crystallizes in the lower temperatures than are obtained when monoclinic system, with the crystals the sulfide is being oxidized as it is having a pseudo-orthorhombic symmetry be­ heated during the TGA run. cause of twinning (~). Figure 6E illus­ trates the TGA results for a sample mill­ Arsenopyrite ed for 120 min. This figure shows that water vapor had the effect of lowering The final sulfide tested in this series the ignition point of the sulfide and was a sample of arsenopyrite. In this causing the sample to gain we ight more mineral, arsenic replaces a sulfur in rapidly at lower temperatures. For this 13

6 ,---,----,---,---,----,---,----,---, 105 I A Pyrrhotite A A 100 U u a. ::!: +- 0:: W 3 95 :r: o t­ W- o I- x KEY w (f) « 90 B A 190°C u 0:: « B 200 °C ~ -6 85 f? 6 W 0:: B Marcasite ::l 80~--~-----L----~----L---~----~ t- o 5 10 15 20 25 30 <:( 0:: W TIME, min a. ::!: FIGURE 7.-lsothermal testing of marcasite milled for 120 w t- min and oxidized with air containing 40 pct water vapor. O ...J <:( t- Z W 0:: W sample, water vapor lowered the initial LL 0 0 LL ignition point from 372 to 337 C. The 0 final temperature in these TGA tests was -6 0 raised to slightly above 450 C to better 6 illustrate the unusual double weight loss C Arsenopyri te that occurred when the arsenopyrite was oxidized to the end of the first step, which was about a lO-pct sample weight u ::!: loss. XRD analysis indicated the pres­ 0:: W :r: 0 ence of hematite, but it failed to indi­ t- o cate if the iron sulfide or iron arsenide 0 z component of the mineral had oxidized. w

Differential Thermal Analyses

-6 L-__~ __-L __ -L __ ~ ____L- __~ __~ __~ The results of DTA on pyrrhotite A 100 150 200 250 300 350 400 450 500 milled for 120 min are shown in figure TEMPERATURE, DC 8A. This sample was oxidized in air con­ FIGURE B.-Differential thermal analyses of (A) pyrrhotite A, taining 4.5 pct water vapor, and the DTA (8) marcasite, and (C) arsenopyrite milled for 120 min and ox· trace shows two prominent peaks in prox­ idized with air containing 4.5 pct water vapor. imity to the areas of most rapid weight gain and weight loss shown in figure 6B. Based on the TGA results and XRD analy­ broad peak reaches its maximum value at sis, the exothermic peaks shown in figure about 300 0 C. A shift of this peak to 8A correspond to the formation of iron the left would be expected if additional sulfate and then iron oxide. The DTA water vapor could be added to the furnace trace of pyrrhotite B was similar to that assembly. Figure 8e is the DTA plot of for sample A and included two exothermic arsenopyrite during oxidation in air con­ peaks at 305 0 and 400 0 C. The DTA anal­ taining 4.5 pct water vapor. A large ysis of marcasite oxidized in an atmos­ peak was found at about 340 0 C, along phere of air containing 4.5 pct water with alllinor peak that initiated at vapor is shown in figure 8B. The single 480 0 C. 14

Discussion if the iron sulfide or iron arsenide com­ ponent of the mineral underwent a single The results of thermal analysis con­ weight loss on ignition. ducted on fine pyrrhotite, marcasite, and DTA indicated that marcasite and arse­ arsenopyrite showed that decreasing the nopyrite reacted in an exothermic manner particle size of the sulfide lowered the after oxidation was initiated at 100 0 C. ignition point. Decreasing the particle However, pyrrhotite required heating to size also increased the volatility of the 1400 to 1500 C before exothermic behavior sulfur, shown as sample weight loss dur­ was initiated. This result suggests that ing heating in nitrogen. Adding water the self- heating of pyrrhotite in a mine vapor to the oxidizing atmosphere had may be initiated by a mechanism other little effect on the ignition point of than sulfate formation from oxidation by the pyrrhotites tested in this investiga­ air. tion. The presence of water vapor did In this investigation, the sulfides not enhance the weight gain or sulfate were milled to very fine particle size to formation for pyrrhotite sample A, but it increase their reactivity. This enhanceo accelerated the weight gain prior to ig­ the degree of weight change and allowed nition for sample B. Ferrous sulfate was the completion of isothermal testing in a found in pyrrhotite samples following ox­ reasonable time span. From the thermo­ idation for several hours at temperatures dynamic data listed in table 1, it is ap­ below the sulfide ignition point. parent that the reactions that form sul­ The ignition point of marcasite was fate could be a possible source of heat significantly lowered when water vapor generation in a sulfide mine. The ther­ was present during oxidation. Partial mal analysis tests were conducted with conversion of the sulfide to the sulfate equipment that has a high thermal loss, below 200 0 C was verified. but future experiments are planned that Arsenopyrite was found to undergo an will use larger samples in tests that unusual double-step weight loss during will retain the heat from exothermic re­ oxidation in the presence of water vapor. actions. These tests are expected to XRD analysis was used to compare the oxi­ verify if the sulfates formed by air oxi­ dation products following the first dation in a warm, moist mine environment weight loss and the second weight loss; are part of the mechanism that to however, the analysis failed to indicate spontaneous combustion.

SUMMARY AND CONCLUSIONS

Common sulfide minerals, including air, or oxygen with or without the addi­ chalcopyrite, chalcocite, galena, pyrrho­ tion of water vapor. Using the weight tite, marcasite, and arsenopyrite, were change of the samples under the various subjected to thermal analysis using TGA test conditions and the XRD analysis of and DTA methods. One condition of the the oxidized samples, reactions were pro­ thermal analyses required grinding of the posed that could promote the heating of sulfides to very fine sizes to promote sulfides and eventually lead to a mine their reactivity. This allowed the ther­ fire. mal analysis to be conducted at relative­ As expected, the sulfide samples that ly low temperatures, and in experiments had the finest particl~ size were the with fairly short timeframes. Samples most reactive. This increased reactivity were tested in atmospheres of nitrogen, provided test results indicating that IS chalcopyrite, a common copper ore, under­ For all three of the non-ore-type min­ went a weight gain prior to ignition. erals, an exothermic weight gain was ob­ Evidence was found that this weight gain served prior to the ignition point, is related to the formation of ferrous caused by the formation of iron sulfate. sulfate, which is similar to the reac­ The TGA results show that the ignition tions that precede the ignition of py­ point of both marcasite and arsenopyrite rite. Upon ignition, or weight loss, of was decreased when water vapor was in­ the chalcopyrite, some of the iron pres­ cluded in the oxidizing atmosphere. ent in the sulfide is oxidized to ferric Water vapor appeared to have little ef­ oxide. Following the weight loss, the fect on the ignition point of pyrrhotite, chalcopyrite rapidly gained weight dur­ and it had a mixed effect on promoting ing the formation of copper sulfate. the formation of iron sulfate at tempera­ Chalcocite and galena did not exhibit the tures below the ignition point of the weight loss indicative of oxide formation sulfide. This suggests that water vapor in tests conducted below 450 0 C. How­ has a lesser role in the exothermic be­ ever, chalcocite exhibited a rapid weight havior of this mineral. The self-heating gain at a temperature just slightly above of pyrrhotite is often suggested as the the ignition point of chalcopyrite. This principal cause of spontaneous combustion weight gain was due to the formation of fires in sulfide ore mines. However, of copper sulfate, and if an initial source the minerals examined in this report, of heat were present, the chalcocite marcasite was found to be the most re­ could display a strong exothermic re­ active in that it exhibited a self­ sponse. Galena underwent a steady weight heating tendency at the lowest tempera­ gain as determined by the TGA method, and ture. As marcasite may be associated no rapid gain in weight was observed dur­ with pyrrhotite, but in much smaller ing the conversion of lead sulfide to quantities, its role in the spontaneous lead sulfate. Based on the thermal anal­ heating event may incorrectly be attri­ ysis results, it appears that neither buted to the pyrrhotite. The apparent chalcocite nor galena would tend to ig­ higher reactivity of the marcasite could nite during the long periods of time that also be a result of specific parameters are usually associated with self-heating employed in the test program that niffer of sulfides in the mine environment. from in-mine conditions. Further study However, because both chalcocite and is thus indicated to define the mecha­ galena appear to form sulfates, they nism responsible for these self-heating could contribute to the problem of self­ tendencies. heating. 16

REFERENCES

1. Ninteman, D. J. Oxi- Mineral BuMines B 677, 1984, dation and Combustion of Sulfide Ores in 355 pp. Underground Mines. A Literature Survey. 6. B. H. The Oxidation of Sul­ BuMines IC 8775, 1978,36 pp. Minerals in the Sullivan Mine. 2. Kirshenbaum, N. W. and CIM Bull., v. 70, No. June 1977, of Sulfide Concentrates. Ph.D. pp. 83-88. Thesis, Stanford Univ., Palo Alto, CA, 7. Farnsworth, D. J. M. Introduction 1967, 199 pp. to and Background of Fires in 3. Rabashi, Pillar Mining at the Sullivan Mine. and Me CIM Bull., v. 70, No. 782, June 1977. 1978,165 pp. pp. 65-71. 4. Simpson, R. N. F., and 1. C. 8. Pahlman, J.E., and G. W. Reimers. Herbert. and Handling Pro­ Thermal Gravimetric Ana is of te Sulphide Concen- Oxidation at Low rature. BuMines trate Spontaneous Combus- RI 9059, 1986, 12 pp. tion. First International 9. Thrush, P. W. A Dictionary of Symposium on Transport and Handl of Mining, Mineral, and Related Terms. Bu­ Minerals ish Columbia, , Mines Spec. Publ., 1968, 1269 pp. Oct. 20-23, 1971). Miller Freeman Publ. 10. Chaubal, P. C., and H. Y. Sohn. Inc. , San Francisco, CA, v. 1, 1 Intrinsic Kinetics of the Oxidation of pp. 138-153. Chalcopyrite Particles Under Isothermal 5. Pankratz, L. B., J. M. Stuve, and and Nonisothermal Conditions. Metall. N. A. Gokcen. Thermodynamic Data for Trans. B, v. 17B, Mar. 1986, pp. 51-60.

u.s. GOVERNMENT PRINTING OFFICE: 1987 605-0171600a6 INT.-BU.OF MINES,PGH.,PA. 28548