Transition of Sulphide Self-Heating from Stage A to Stage B †

: Previous work has shown that sulphide self-heating occurs in three distinct stages, referred to as Stage A, Stage B and Stage C. In this publication, the focus is the transition from Stage A to Stage B which occurs at ca. 100 ◦ C. Background literature hints that the transition corresponds to the transformation of the rhombic form of elemental sulphur to the more reactive monoclinic form that occurs at 96 ◦ C. A test apparatus is modiﬁed for adiabatic heating to track the transition. The results support this transformation of sulphur as being key to the transition, and the transition temperature is thus modiﬁed to 96 ◦ C. Variations in a sample’s response under adiabatic conditions are observed and possible reasons are discussed. Testing in adiabatic mode provides new insights into the sulphide self-heating process that complements the test designed to identify propensity to self-heat.


Introduction
Some materials, when exposed to ambient conditions, can exhibit a rise in temperature without requiring an external heat source. Generally referred to as pyrophoric or self-heating substances, they pose a challenge to the safe handling of materials in storage and transportation. Numerous substances self-heat, ranging from wood chips and powdered milk to coal and sulphide minerals [1,2]. Combinations of sulphide minerals, most notably those containing the iron sulphide pyrrhotite (Fe (1−x) S), commonly encountered in the extraction of base metals (e.g., copper, zinc, lead, and nickel), are particularly prone to self-heating [3,4]. If not controlled, sulphide self-heating can lead to work disruptions, loss of infrastructure, delays in operations and in some severe cases, loss of life [5]. Dramatic examples include in-situ smelting of ores and resultant fires [5,6] and loss of a ship transporting Cu-concentrate [7].
Testing for sulphide self-heating risk is required by the International Maritime Solid Bulk Cargoes Code [8] and tests continue to be proposed [9][10][11][12]. Recent work [13] has shown that the U.N.-recommended test can give false negatives and provides an alternative.
Studies on sulphide self-heating mechanisms identify moisture as key to exothermic reactions with elemental sulphur as one the products [14,15], the critical range being about 3-8% w/w [16]. Therein lies a conundrum: as moisture is consumed in the reactions and evaporated as temperature rises to ca. 100 • C, what causes the further temperature rise? Rosenblum and co-workers [3,[17][18][19][20] have argued that sulphide self-heating follows a three stage process: Stage A, up to ca. 100 • C, where moisture is present and elemental sulphur is produced; Stage B, ca. 100 • C to 350 • C, where the sulphur oxidizes; and Stage C, above ca. 350 • C, when direct oxidation of the sulphides occurs sulphur oxidizes; and Stage C, above ca. 350 °C, when direct oxidation of the sulphides occurs (i.e., smelting). Studies have largely concentrated on the properties of the individual stages, with no studies reported on the transitions between the stages. Without transition from Stage A to Stage B, for example, the runaway condition Stage C, which causes most of the problems, does not occur.
The present paper focuses on the factors associated with the Stage A to Stage B transition. We start with a review of the literature related to the transition.

A Brief Review of Self-Heating Literature Related to the Transition
In Rosenblum and Spira [18] the notional temperatures at which the transitions occurred were identified. The paper also described the test apparatus and protocol designed to assess Stages A and B (now referred to as the FR-2 apparatus and test). Stage A (i.e., below ca. 100 °C) conditioning (or "weathering") was simulated using a thermally active waste rock sample from Brunswick Mining and Smelting. The sample was dry ground then moistened to 8% moisture. The sample was subjected to cycles of 15 min air injection every 5 h. Weathering at 50 °C showed that elemental sulphur was generated ( Figure 1) along with other oxidation products (in this example, goethite and sulphates) and that the pyrrhotite, which was the source of the oxidation products, diminished with weathering time. Figure 1. Effect of weathering in Stage A at 50 °C on reduction in pyrrhotite content and increase in oxidation products (adapted from Rosenblum and Spira [18]).
Using differential scanning calorimetry (DSC), the heat-generating potential of a sulphide ore sample was assessed ( Figure 2). The fresh (un-weathered) sample exhibited practically zero heating and with increased weathering (at 40 °C) the exothermic peaks increased showing a direct correlation between the extent of weathering and heat generation. In terms of the stages, the weathering was in Stage A with resultant heat flow occurring in Stage B (DSC peaks are >100 °C). Using differential scanning calorimetry (DSC), the heat-generating potential of a sulphide ore sample was assessed ( Figure 2). The fresh (un-weathered) sample exhibited practically zero heating and with increased weathering (at 40 • C) the exothermic peaks increased showing a direct correlation between the extent of weathering and heat generation. In terms of the stages, the weathering was in Stage A with resultant heat flow occurring in Stage B (DSC peaks are >100 • C). In an extension to the results in Figure 2, which represents about 4 days, weathering was conducted at 50 °C for 45 days followed by DSC measurements (Figure 3). The heat flow continued to increase with increased weathering, reaching a maximum at about 20 days then dropping, reaching a plateau, followed by a further drop. Based on the prior results, the following was argued to be the sequence of events giving rise to these swings in reactivity. In the initial stage of weathering elemental sulphur is formed which becomes the fuel that drives the observed exothermic reaction in the DSC test (i.e., Stage B reaction/s). With increased weathering, the other oxidation products (oxides, etc.) that are formed could interfere with the oxidation of the elemental sulphur thus reducing the heat flow. Note that after washing, filtering and drying the 45-day weathered sample, reactivity increased significantly. A suggestion is that washing either dissolved or dislodged the oxides, etc., which are also formed, thus exposing more of the elemental sulphur resulting in increased heating. In an extension to the results in Figure 2, which represents about 4 days, weathering was conducted at 50 • C for 45 days followed by DSC measurements (Figure 3). The heat flow continued to increase with increased weathering, reaching a maximum at about 20 days then dropping, reaching a plateau, followed by a further drop. Based on the prior results, the following was argued to be the sequence of events giving rise to these swings in reactivity. In the initial stage of weathering elemental sulphur is formed which becomes the fuel that drives the observed exothermic reaction in the DSC test (i.e., Stage B reaction/s). With increased weathering, the other oxidation products (oxides, etc.) that are formed could interfere with the oxidation of the elemental sulphur thus reducing the heat flow. Note that after washing, filtering and drying the 45-day weathered sample, reactivity increased significantly. A suggestion is that washing either dissolved or dislodged the oxides, etc., which are also formed, thus exposing more of the elemental sulphur resulting in increased heating.
In the 1995 publication [9] SO 2 off-gas was measured as a function of temperature. Figure 4 (open symbols, % scale) shows SO 2 began to appear at about 100 • C and with increasing temperature substantially more SO 2 was generated. From these results, it appears that SO 2 is emitted in Stage B only.
Included in the figure is experience at one mine site [21] where high levels of SO 2 were detected which prompted measurements to correlate SO 2 emissions with the corresponding in situ temperature. Results (closed symbols, ppm scale) showed that most data fell above 100 • C, that is, corresponding to a Stage B event. When plotting sulphur vapour pressure as a function of temperature [22], it is noted that vapour first appears at a temperature above ca. 100 • C (dashed curve).
These results support the interpretation that self-heating of sulphides proceeds in at least two distinct stages: Stage A, where the presence of moisture causes exothermic reactions producing elemental sulphur, followed by Stage B, where the sulphur oxidizes to generate heat and sulphur dioxide. There is evidence [23] that the elemental sulphur formed in Stage A is the rhombic form (α-S8) which is stable below 96.5 • C, and above this temperature it transitions to the less stable monoclinic form (β-S8) that readily oxidizes to generate heat and SO 2 . As an average we adopt a sulphur-form transition temperature of 96 • C, the range being from 95 • C [23-25] to 96.5 • C [26]. In the 1995 publication [9] SO2 off-gas was measured as a function of temperature. Figure 4 (open symbols, % scale) shows SO2 began to appear at about 100 °C and with increasing temperature substantially more SO2 was generated. From these results, it appears that SO2 is emitted in Stage B only.
Included in the figure is experience at one mine site [21] where high levels of SO2 were detected which prompted measurements to correlate SO2 emissions with the corresponding in situ temperature. Results (closed symbols, ppm scale) showed that most data fell above 100 °C, that is, corresponding to a Stage B event. When plotting sulphur vapour pressure as a function of temperature [22], it is noted that vapour first appears at a temperature above ca. 100 °C (dashed curve).
These results support the interpretation that self-heating of sulphides proceeds in at least two distinct stages: Stage A, where the presence of moisture causes exothermic reactions producing elemental sulphur, followed by Stage B, where the sulphur oxidizes to generate heat and sulphur dioxide. There is evidence [23] that the elemental sulphur formed in Stage A is the rhombic form (α-S8) which is stable below 96.5 °C, and above this temperature it transitions to the less stable monoclinic form (β-S8) that readily oxidizes to generate heat and SO2. As an average we adopt a sulphur-form transition temperature of 96 °C, the range being from 95 °C [23][24][25] to 96.5 °C [26].
With this background, experiments were designed to identify factors affecting the Stage A to Stage B transition.  . Compilation of literature data as a function of temperature: emission of SO2 from lab test [9] and mine site [21]; sulphur vapour pressure; phase transition.

Samples
Samples of rod mill discharge (RMD) from Glencore's Brunswick Mine and nickel concentrate (Ni Con) from Glencore's Raglan Mine were used in this test series. The RMD sample was ground to 80% passing 75 µm, split in a rotary splitter into 500-g samples then bagged and sealed. The Ni Con sample was shipped under a water cover. On receipt, the sample was filtered, spread and dried in a convection oven at 40 °C for 5 h. The sample was then split using the rotary splitter into 500 g samples, . Compilation of literature data as a function of temperature: emission of SO 2 from lab test [9] and mine site [21]; sulphur vapour pressure; phase transition.
With this background, experiments were designed to identify factors affecting the Stage A to Stage B transition.

Samples
Samples of rod mill discharge (RMD) from Glencore's Brunswick Mine and nickel concentrate (Ni Con) from Glencore's Raglan Mine were used in this test series. The RMD sample was ground to 80% passing 75 µm, split in a rotary splitter into 500-g samples then bagged and sealed. The Ni Con sample was shipped under a water cover. On receipt, the sample was filtered, spread and dried in a convection oven at 40 • C for 5 h. The sample was then split using the rotary splitter into 500 g samples, bagged, sealed and stored at room temperature.

Self-Heating Test Apparatus
The FR-2 apparatus ( Figure 5), described in detail elsewhere [3,27], was used. In the standard test, Stage A is conducted on a 500-g dry sample moistened to 6% w/w, the temperature is set at 70 • C with air injection (via a water bath below the sample) at 100 mL/min for 15 min every 5 h for ten such cycles. A reactive sample is identified by peaks in the temperature-time plot (thermogram) accompanying the air injections. Stage B is conducted after raising the temperature to 140 • C with air injection at 200 mL/min for 15 min every 5 h repeated until no more heating is observed (i.e., there are no further peaks on the thermogram).

Weathering
For the weathering tests, moist (6% w/w) samples were exposed at 70 °C for a set number of air injection cycles ranging from 0 to 20. The sample was then dried, re-moistened to 6% w/w and tested at the standard Stage A condition (70 °C; 10 cycles). Some of these samples were also analyzed for elemental sulphur. There were two objectives: (1) to correlate weathering with the generation of elemental sulphur; and (2) correlate the effect of weathering (number of cycles) with self-heating capacity of Stage A

Transition Stage A to Stage B
Both the as received and weathered samples were tested. The apparatus was first set to CTC mode (70 °C) with no air injection. After the sample reached the set temperature, control was switched to ATC mode. At that point, air injection (100 mL/min) commenced. To prevent runaway temperatures (that can result in a fire), the ATC mode was limited to 140 °C, at which point it was switched back to CTC mode. When the sample temperature equilibrated with the oven temperature,

Adiabatic Temperature Control (ATC)
To follow the transition, the apparatus was modified to operate in adiabatic mode where oven temperature follows the sample temperature; that is, little or no heat flows in or out of the sample.

Weathering
For the weathering tests, moist (6% w/w) samples were exposed at 70 • C for a set number of air injection cycles ranging from 0 to 20. The sample was then dried, re-moistened to 6% w/w and Minerals 2020, 10, 1133 6 of 12 tested at the standard Stage A condition (70 • C; 10 cycles). Some of these samples were also analyzed for elemental sulphur. There were two objectives: (1) to correlate weathering with the generation of elemental sulphur; and (2) correlate the effect of weathering (number of cycles) with self-heating capacity of Stage A.

Transition Stage A to Stage B
Both the as received and weathered samples were tested. The apparatus was first set to CTC mode (70 • C) with no air injection. After the sample reached the set temperature, control was switched to ATC mode. At that point, air injection (100 mL/min) commenced. To prevent runaway temperatures (that can result in a fire), the ATC mode was limited to 140 • C, at which point it was switched back to CTC mode. When the sample temperature equilibrated with the oven temperature, the oven was shut-off.

Sulphur Determination
On selected samples, elemental sulphur was determined using the carbon disulphide method which brings elemental sulphur directly from the sample into solution [28]. Five-gram samples were contacted with 15 mL carbon disulphide (Sigma-Aldrich, 99.9% A.C.S reagent grade) in a 50 mL beaker, stirred for 10 min then filtered into a 50 mL flask. To ensure that all the sulphur had been recovered, an additional 5 mL carbon disulphide was added, the contents stirred and filtered into the same 50 mL flask. The filtrate was evaporated using an overhead fan in a fume hood and weighed to determine the elemental sulphur content (%). Figure 6 shows that increased weathering increased the formation of elemental sulphur. The as-received RMD sample (i.e., zero weathering) had less than 0.1% sulphur, which after 15 cycles rose to 1.8%. In contrast, the as-received Ni Con sample contained almost 1% elemental sulphur (suggestive of prior weathering) and weathering only marginally increased the sulphur content.  Figure 6 shows that increased weathering increased the formation of elemental sulphur. The asreceived RMD sample (i.e., zero weathering) had less than 0.1% sulphur, which after 15 cycles rose to 1.8%. In contrast, the as-received Ni Con sample contained almost 1% elemental sulphur (suggestive of prior weathering) and weathering only marginally increased the sulphur content.

Effect on Self-Heating
Starting at ten weathering cycles, Figure 7 shows that increased weathering cycles resulted in increased Stage A self-heating capacity for both the RMD and Ni Con. The rate of increase for both is similar (lines almost parallel). The higher values exhibited by RMD are attributed to the fact that after the initial 10 weathering cycles there was more elemental sulphur present compared with the Ni Con sample (Figure 6).

Effect on Self-Heating
Starting at ten weathering cycles, Figure 7 shows that increased weathering cycles resulted in increased Stage A self-heating capacity for both the RMD and Ni Con. The rate of increase for both is similar (lines almost parallel). The higher values exhibited by RMD are attributed to the fact that after the initial 10 weathering cycles there was more elemental sulphur present compared with the Ni Con sample ( Figure 6).
Minerals 2020, 10, x FOR PEER REVIEW 8 of 13 Figure 7. Effect of weathering on Stage A self-heating capacity (for calculation for self-heating capacity see [3]).

Example of a Successful Transition
The sample temperature was first raised to 70 °C under CTC mode with no airflow. When stabilized, control was changed to ATC mode and air injection started (Figure 8). The sample temperature increased, plateaued near the transition temperature for about 3 h (10 h to 13 h) then continued to rise. At 140 °C oven temperature control was changed back to CTC mode, with sample temperature continuing to rise reaching a maximum of 166 °C then dropping back. The temperature rise from 70 °C to ca. 100 °C represents Stage A self-heating while the further temperature rise to 166 °C represents Stage B self-heating. Stage B was sustained for about 8 h (13 h to 21 h) when the sample temperature began to drop, which is consistent with the fuel, elemental sulphur, having been consumed. Figure 7. Effect of weathering on Stage A self-heating capacity (for calculation for self-heating capacity see [3]).

Example of a Successful Transition
The sample temperature was first raised to 70 • C under CTC mode with no airflow. When stabilized, control was changed to ATC mode and air injection started (Figure 8). The sample temperature increased, plateaued near the transition temperature for about 3 h (10 h to 13 h) then continued to rise. At 140 • C oven temperature control was changed back to CTC mode, with sample temperature continuing to rise reaching a maximum of 166 • C then dropping back. The temperature rise from 70 • C to ca. 100 • C represents Stage A self-heating while the further temperature rise to 166 • C represents Stage B self-heating. Stage B was sustained for about 8 h (13 h to 21 h) when the sample temperature began to drop, which is consistent with the fuel, elemental sulphur, having been consumed. continued to rise. At 140 °C oven temperature control was changed back to CTC mode, with sample temperature continuing to rise reaching a maximum of 166 °C then dropping back. The temperature rise from 70 °C to ca. 100 °C represents Stage A self-heating while the further temperature rise to 166 °C represents Stage B self-heating. Stage B was sustained for about 8 h (13 h to 21 h) when the sample temperature began to drop, which is consistent with the fuel, elemental sulphur, having been consumed.

Example of Failure to Transition
In an expanded first part of the temperature response, Figure 9 shows that in this case while the sample temperature rose when the system was switched to ATC mode, it did not pass the transition temperature.

Example of Failure to Transition
In an expanded first part of the temperature response, Figure 9 shows that in this case while the sample temperature rose when the system was switched to ATC mode, it did not pass the transition temperature. Figure 9. Thermograph of a test sample that did not pass the transition temperature.

Transition and Weathering
The RMD samples weathered for various durations and analyzed for elemental sulphur were remoistened to 6% and then evaluated for the transition from Stage A to Stage B. Using the maximum sample temperature recorded (see Figure 8) to characterize the sample's self-heating propensity, Figure 10 shows a bell-shaped response: the sample with less than 0.5% sulphur did not pass the transition (in fact did not self-heat as the temperature remained at 70 °C), samples from ca. 0.6% to 1.5% sulphur showed increasing maximum temperatures, and samples with greater than 1.5% sulphur showed a decrease in maximum temperature. The decrease is attributed to the growing presence of other oxidation products (sulphates, oxides/hydroxides), which appear to restrict access of oxygen to the elemental sulphur.

Transition and Weathering
The RMD samples weathered for various durations and analyzed for elemental sulphur were remoistened to 6% and then evaluated for the transition from Stage A to Stage B. Using the maximum sample temperature recorded (see Figure 8) to characterize the sample's self-heating propensity, Figure 10 shows a bell-shaped response: the sample with less than 0.5% sulphur did not pass the transition (in fact did not self-heat as the temperature remained at 70 • C), samples from ca. 0.6% to 1.5% sulphur showed increasing maximum temperatures, and samples with greater than 1.5% sulphur showed a decrease in maximum temperature. The decrease is attributed to the growing presence of other oxidation products (sulphates, oxides/hydroxides), which appear to restrict access of oxygen to the elemental sulphur. A second test series used the Ni Con, which entailed weathering of two samples in duplicate (at 5 and 20 cycles) and one sample in triplicate (at 10 cycles) with testing in ATC mode. Figure 11 shows that for the duplicate at 5 cycles both samples transitioned to Stage B and reached a maximum temperature of ca. 130 °C. The other duplicate (at 20 cycles) and the triplicate sample (10 cycles), however, produced a range of results from just reaching the transition temperature to rising above 140 °C. A second test series used the Ni Con, which entailed weathering of two samples in duplicate (at 5 and 20 cycles) and one sample in triplicate (at 10 cycles) with testing in ATC mode. Figure 11 shows that for the duplicate at 5 cycles both samples transitioned to Stage B and reached a maximum temperature of ca. 130 • C. The other duplicate (at 20 cycles) and the triplicate sample (10 cycles), however, produced a range of results from just reaching the transition temperature to rising above 140 • C. A second test series used the Ni Con, which entailed weathering of two samples in duplicate (at 5 and 20 cycles) and one sample in triplicate (at 10 cycles) with testing in ATC mode. Figure 11 shows that for the duplicate at 5 cycles both samples transitioned to Stage B and reached a maximum temperature of ca. 130 °C. The other duplicate (at 20 cycles) and the triplicate sample (10 cycles), however, produced a range of results from just reaching the transition temperature to rising above 140 °C.

Discussion
The original hypothesis was that transition from Stage A to Stage B occurs at 100 • C; that Stage A involved the formation of elemental sulphur which was the fuel for Stage B. Based on the literature and the present findings, the current interpretation refines Stage A as up to 96 • C, corresponding to the transition from rhombic sulphur (α-S8) to the more reactive monoclinic sulphur (β-S8).
In this interpretation, the self-heating in Stage A is the result of exothermic oxidation of sulphides in the presence of moisture that generates rhombic elemental sulphur. Under favorable conditions, the material could reach 96 • C and the sulphur start to transition to the monoclinic form which requires latent heat. Figure 8 appears to show this latent heat, the temperature remaining roughly constant around 96 • C for a period before continuing to rise. This transition time could range from minutes to days and may not always occur.
Failure to pass the transition temperature can have several causes. Some possibilities include: (1), moisture level may decrease below that to sustain Stage A reactions (moisture is consumed in the reactions and is lost to evaporation and replenishment via humidity may not be sufficient); (2), while transition probability increases with sulphur content (i.e., weathering), the increasing level of other oxidation products, sulphates, oxides, etc., may act to coat surfaces and prevent oxygen from reaching the elemental sulphur, as suggested in Figures 3 and 10; and (3), there may be differences in sample packing that affect access to oxygen and moisture.
These possibilities help explain the finding that the same samples can vary in response to the degree of weathering ( Figure 11). This variation is mirrored in the field where self-heating typically occurs in isolated "hot spots" and not generally across the entire pile [18].
The standard Stage A/Stage B test tends not to encounter variations in response on the same sample. In these tests the temperature in Stage B is set at 140 • C consequently whether transition occurs is not at issue. The test, designed to identify a material's propensity to self-heat and progress from Stage A to Stage B (and by extension Stage C), is independent of whether, on occasions, conditions may suppress transition. Testing in adiabatic mode, as introduced here, does offer further insights into the self-heating process and complements the standard FR-2 test.

Conclusions
This study focused on the transition from Stage A to Stage B in the self-heating of sulphides. A new test using adiabatic conditions is introduced. Evidence is presented that the transition occurs at 96 • C, corresponding to the transition from the stable rhombic form of elemental sulphur to the more reactive monoclinic form. The transition temperature, formerly designated as 100 • C, is duly revised to 96 • C. Variations in sample response to transitioning are detected and possible reasons are discussed.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the result.