FSE wiki: SI values recomputed from the US customary originals (Table 9.5.1 pressures and temperature, travel distance, container volume, air gap, powder layer) (written by Claude Opus 5.5) |
Remove bibliographic source lines from image captions and descriptions |
||
| (5 intermediate revisions by 3 users not shown) | |||
| Line 1: | Line 1: | ||
A '''combustible metal''' is a metal that is either combustible or explosive, because of its inherent properties or because of its physical form.<ref name="CH-153">Christman (2008), p. 6-153.</ref><ref name="CH-163">Christman (2008), p. 6-163.</ref> Nearly all metals burn in air under some conditions, but magnesium, titanium, zirconium, the alkali metals and several others ignite easily as thin sections, fine particles or molten metal. As massive pieces the same metals are very difficult to ignite, except for the alkali metals, which react with water and release hydrogen.<ref name="CH-153" /> | A '''combustible metal''' is a metal that is either combustible or explosive, because of its inherent properties or because of its physical form.<ref name="CH-153">Christman (2008), p. 6-153.</ref><ref name="CH-163">Christman (2008), p. 6-163.</ref> Nearly all metals burn in air under some conditions, but magnesium, titanium, zirconium, the alkali metals and several others ignite easily as thin sections, fine particles or molten metal. As massive pieces the same metals are very difficult to ignite, except for the alkali metals, which react with water and release hydrogen.<ref name="CH-153" /> | ||
Fires in combustible metals form Class D of the [[fire classes]] used in the United States.<ref name="CO-71">Conroy (2008), p. 17-71.</ref> They burn much hotter than flammable liquids, some keep burning in carbon dioxide, nitrogen or steam, and water can make them react violently.<ref name="CH-154">Christman (2008), p. 6-154.</ref><ref name="CH-155">Christman (2008), p. 6-155.</ref> Most ordinary [[extinguishing agent]]s fail on | Fires in combustible metals form Class D of the [[fire classes]] used in the United States.<ref name="CO-71">Conroy (2008), p. 17-71.</ref> They burn much hotter than flammable liquids, some keep burning in carbon dioxide, nitrogen or steam, and water can make them react violently.<ref name="CH-154">Christman (2008), p. 6-154.</ref><ref name="CH-155">Christman (2008), p. 6-155.</ref> Most ordinary [[extinguishing agent]]s fail on Class D fires; the agent must match the metal. Suitable options include certain dry powders, sand and salts. Argon or helium can blanket some fires in enclosed spaces.<ref name="YN-91">Yu and Newman (2008), p. 2-91.</ref><ref name="CH-161" /> Finely divided metals add a second hazard, the [[dust explosion]].<ref name="CH-154" /> In the US practice described by the 2008 handbook, NFPA 484, ''Standard for Combustible Metals'', addresses these hazards.<ref name="CH-153" /> | ||
== Definition and scope == | == Definition and scope == | ||
| Line 18: | Line 18: | ||
The explosivity assessment in the 2006 edition of NFPA 484, as reproduced in the handbook, runs in four steps.<ref name="CH-154" /> A screening test first classifies the explosibility of the dust. The [[minimum ignition energy]] and the minimum ignition temperature of a dust cloud follow. The third step measures the explosion severity in a 20-litre sphere: the maximum pressure, the maximum rate of pressure rise and the [[deflagration index]] ''K''<sub>St</sub>. The last step gives the limiting oxygen concentration and the minimum explosible concentration.<ref name="CH-154" /> | The explosivity assessment in the 2006 edition of NFPA 484, as reproduced in the handbook, runs in four steps.<ref name="CH-154" /> A screening test first classifies the explosibility of the dust. The [[minimum ignition energy]] and the minimum ignition temperature of a dust cloud follow. The third step measures the explosion severity in a 20-litre sphere: the maximum pressure, the maximum rate of pressure rise and the [[deflagration index]] ''K''<sub>St</sub>. The last step gives the limiting oxygen concentration and the minimum explosible concentration.<ref name="CH-154" /> | ||
[[File:fse-combustible-metal-v1.svg|thumb|upright=1.5|alt=Flowchart: combustibility test and four-step explosivity determination leading to the decision whether a material is a combustible metal.|Assessment of a material under NFPA 484 (2006 edition): failing either the combustibility test or the explosivity determination makes it a combustible metal.]] | |||
Other US documents refer to NFPA 484 for particular situations, such as the use of sprinklers where magnesium is present and the dust collectors for aluminium.<ref name="FR-41">Frank (2008), p. 17-41.</ref><ref name="CH-161" /> Under NFPA 101 and NFPA 5000, as described by Demers (2008a), industrial facilities that produce explosive dusts of aluminium or magnesium are high-hazard industrial occupancies.<ref name="DE-208">Demers (2008a), p. 20-208.</ref> In the international transport hazard classes described by C. J. Wright (2008a), Division 4.1 (flammable solids) includes any metal powder that can be ignited.<ref name="WR-120">C. J. Wright (2008a), p. 13-120.</ref> Division 4.3 covers materials that, when wet, can become spontaneously flammable or give off flammable or toxic gas at more than 1 litre per kilogram per hour. Potassium is one of its examples.<ref name="WR-120" /><ref name="WR-119">C. J. Wright (2008a), p. 13-119.</ref> | Other US documents refer to NFPA 484 for particular situations, such as the use of sprinklers where magnesium is present and the dust collectors for aluminium.<ref name="FR-41">Frank (2008), p. 17-41.</ref><ref name="CH-161" /> Under NFPA 101 and NFPA 5000, as described by Demers (2008a), industrial facilities that produce explosive dusts of aluminium or magnesium are high-hazard industrial occupancies.<ref name="DE-208">Demers (2008a), p. 20-208.</ref> In the international transport hazard classes described by C. J. Wright (2008a), Division 4.1 (flammable solids) includes any metal powder that can be ignited.<ref name="WR-120">C. J. Wright (2008a), p. 13-120.</ref> Division 4.3 covers materials that, when wet, can become spontaneously flammable or give off flammable or toxic gas at more than 1 litre per kilogram per hour. Potassium is one of its examples.<ref name="WR-120" /><ref name="WR-119">C. J. Wright (2008a), p. 13-119.</ref> | ||
== Combustion behaviour == | == Combustion behaviour == | ||
Machining turns solid stock into chips and fines with a much larger surface-to-mass ratio, and some of these can ignite spontaneously or from outside sources.<ref name="DO-68">Dobbs (2008), p. 9-68.</ref> The thinner the cutting and the smaller the particles, the greater the likelihood of fire, and heavy chips can hide a significant amount of fines.<ref name="DO-70">Dobbs (2008), p. 9-70.</ref> According to Christman (2008), reactivity is somewhat reduced for sections thicker than 381 µm (0.015 in.). Finer particles and thinner sections are an increasing hazard.<ref name="CH-154" /> | |||
Alloys can differ widely in combustibility from their constituent elements.<ref name="CH-154" /> A special case is the eutectic, an alloy with a melting point below that of each of its constituents, which can react vigorously and exothermically at lower temperatures. A titanium plate heated in contact with steel scale is one recorded example. The eutectic liquefies below the melting point of either material and the hot liquid spreads by flow.<ref name="CH-154" /> | Alloys can differ widely in combustibility from their constituent elements.<ref name="CH-154" /> A special case is the eutectic, an alloy with a melting point below that of each of its constituents, which can react vigorously and exothermically at lower temperatures. A titanium plate heated in contact with steel scale is one recorded example. The eutectic liquefies below the melting point of either material and the hot liquid spreads by flow.<ref name="CH-154" /> | ||
| Line 53: | Line 54: | ||
'''Worked example.''' Magnesium burns to magnesium oxide, MgO, so ''x'' = ''y'' = 1. With atomic weights of 24.31 for magnesium and 16.00 for oxygen, ''r'' = 16.00/24.31 = 0.658 kg of oxygen per kilogram.<ref name="DR-14" /> Dividing 24.73 MJ/kg by this ratio gives 37.6 MJ per kilogram of oxygen, about 2.9 times the value for organic fuels.<ref name="BA-281" /> A hydrocarbon fuel needs about 3.46 kg of oxygen per kilogram, so magnesium uses about one fifth as much.<ref name="VR-3.3" /> The result assumes complete oxidation and says nothing about how easily a given piece ignites, which depends on its surface-to-mass ratio.<ref name="DO-70" /> | '''Worked example.''' Magnesium burns to magnesium oxide, MgO, so ''x'' = ''y'' = 1. With atomic weights of 24.31 for magnesium and 16.00 for oxygen, ''r'' = 16.00/24.31 = 0.658 kg of oxygen per kilogram.<ref name="DR-14" /> Dividing 24.73 MJ/kg by this ratio gives 37.6 MJ per kilogram of oxygen, about 2.9 times the value for organic fuels.<ref name="BA-281" /> A hydrocarbon fuel needs about 3.46 kg of oxygen per kilogram, so magnesium uses about one fifth as much.<ref name="VR-3.3" /> The result assumes complete oxidation and says nothing about how easily a given piece ignites, which depends on its surface-to-mass ratio.<ref name="DO-70" /> | ||
[[File:fse-combustible-metal-v2.svg|thumb|upright=1.5|alt=Bar chart of heat released per kilogram of oxygen for nine metals, 17 to 40 MJ/kg, against a line at 13.1 MJ/kg for organic fuels.|Heat released per kilogram of oxygen consumed, calculated for nine metals burning to their oxides, compared with the near-constant 13.1 MJ/kg of most organic fuels.]] | |||
{| class="wikitable" | {| class="wikitable" | ||
| Line 127: | Line 129: | ||
|} | |} | ||
Schwab (2008c) defines the deflagration index as the maximum rate of pressure rise in a closed, non-vented vessel multiplied by the cube root of the vessel volume. It is measured in a 1 m³ vessel or in a 20-litre sphere calibrated to give the same results.<ref name="SC-147" | The zirconium and uranium ignition-energy entries are from 1 g test charges; larger quantities ignited spontaneously. The 1964 ignition energies are method-specific, and later test methods can yield lower values.<ref name="DO-69" /><ref name="SC-147">Schwab (2008c), p. 6-147.</ref> | ||
Schwab (2008c) defines the deflagration index as the maximum rate of pressure rise in a closed, non-vented vessel multiplied by the cube root of the vessel volume. It is measured in a 1 m³ vessel or in a 20-litre sphere calibrated to give the same results; Schwab also notes that the smaller sphere often yields a lower value.<ref name="SC-147" /> He lists 508 bar·m/s for a magnesium dust of 28 µm, with a maximum pressure of 17.5 bar (gauge).<ref name="SC-145">Schwab (2008c), p. 6-145.</ref> An aluminium powder finer than 10 µm reaches 515 bar·m/s and 11.2 bar.<ref name="SC-145" /> Zalosh (2002) gives the same 515 bar·m/s for aluminium, and no other dust in his table exceeds 209 bar·m/s.<ref name="ZA-404">Zalosh (2002), p. 3-404.</ref><ref name="ZA-409">Zalosh (2002), p. 3-409.</ref> | |||
[[File:fse-combustible-metal-v3.svg|thumb|upright=1.5|alt=Bar chart of deflagration index: aluminium 515 and magnesium 508 bar·m/s, all other listed dusts 42 to 209 bar·m/s.|Deflagration index KSt of metal dusts compared with other industrial dusts. The data are a qualitative indication only and not for design.]] | |||
<math display="block">K_\mathrm{St} = \left(\frac{\mathrm{d}P}{\mathrm{d}t}\right)_\mathrm{max} V^{1/3}</math> | <math display="block">K_\mathrm{St} = \left(\frac{\mathrm{d}P}{\mathrm{d}t}\right)_\mathrm{max} V^{1/3}</math> | ||
| Line 142: | Line 147: | ||
|} | |} | ||
'''Worked example.''' For the aluminium powder with ''K''<sub>St</sub> = 515 bar·m/s, the maximum rate of pressure rise in a 1 m³ vessel is 515 bar/s. | '''Worked example.''' For the aluminium powder with ''K''<sub>St</sub> = 515 bar·m/s, the maximum rate of pressure rise in a 1 m³ vessel is 515 bar/s. If that same ''K''<sub>St</sub> is assumed in a 20-litre (0.020 m³) sphere, its cube root is 0.271 m and the formula gives about 515/0.271 ≈ 1,900 bar/s. A dust at 209 bar·m/s would similarly give about 770 bar/s under this assumption. These are mathematical illustrations, not predictions for a different vessel or design values: the measured index can differ with vessel and test conditions, particle size, shape, moisture and solvent content.<ref name="SC-147" /> | ||
Metal dusts can also burn at lower oxygen concentrations than organic dusts. With carbon dioxide inerting, organic dusts have limiting oxygen concentrations of about 11–15 % by volume, but the [[limiting oxidant concentration]] of some metals is much lower.<ref name="ZA-404" /> Zalosh (2002) gives 2 % with carbon dioxide and 5 % with nitrogen for aluminium, and zero for zirconium, which indicates that zirconium powder is pyrophoric.<ref name="ZA-403">Zalosh (2002), p. 3-403.</ref><ref name="ZA-404" /> Carbon dioxide should not be used for [[inerting]] magnesium, silicon or zirconium, for which a noble gas such as argon may be needed.<ref name="UG-149">Ural and Garzia (2008), p. 17-149.</ref> | Metal dusts can also burn at lower oxygen concentrations than organic dusts. With carbon dioxide inerting, organic dusts have limiting oxygen concentrations of about 11–15 % by volume, but the [[limiting oxidant concentration]] of some metals is much lower.<ref name="ZA-404" /> Zalosh (2002) gives 2 % with carbon dioxide and 5 % with nitrogen for aluminium, and zero for zirconium, which indicates that zirconium powder is pyrophoric.<ref name="ZA-403">Zalosh (2002), p. 3-403.</ref><ref name="ZA-404" /> Carbon dioxide should not be used for [[inerting]] magnesium, silicon or zirconium, for which a noble gas such as argon may be needed.<ref name="UG-149">Ural and Garzia (2008), p. 17-149.</ref> | ||
| Line 149: | Line 154: | ||
== Reaction with water == | == Reaction with water == | ||
Water is usually the wrong agent for metal fires. A number of metals react exothermically with water to form hydrogen, which burns rapidly, and violent steam explosions can occur if water enters molten metal.<ref name="YN-91" /> Christman (2008) compares applying water to a combustible metal fire with pouring gasoline on a house fire. | Water is usually the wrong agent for metal fires. A number of metals react exothermically with water to form hydrogen, which burns rapidly, and violent steam explosions can occur if water enters molten metal.<ref name="YN-91" /> Christman (2008) compares applying water to a combustible metal fire with pouring gasoline on a house fire.<ref name="CH-155" /> In the magnesium example he explains that oxygen from the water combines with magnesium while the released hydrogen may burn.<ref name="CH-157" /> Frank (2008) notes that water otherwise remains stable in a fire and breaks down only in very special circumstances.<ref name="FR-31">Frank (2008), p. 17-31.</ref> | ||
The reaction ranges from minor to explosive.<ref name="FR-41" /> As a rule, water is not used on fires in magnesium, titanium, sodium or hafnium, or in calcium, zinc and aluminium where these burn, although it can still cool exposures.<ref name="FR-41" /> Large quantities of water have extinguished small amounts of burning magnesium without molten pools.<ref name="YN-91" /> For zirconium, a water spray has no adverse effect on burning turnings, and a straight stream into a drum of burning scrap extinguished the fire.<ref name="CH-161" /> | The reaction ranges from minor to explosive.<ref name="FR-41" /> As a rule, water is not used on fires in magnesium, titanium, sodium or hafnium, or in calcium, zinc and aluminium where these burn, although it can still cool exposures.<ref name="FR-41" /> Large quantities of water have extinguished small amounts of burning magnesium without molten pools.<ref name="YN-91" /> For zirconium, a water spray has no adverse effect on burning turnings, and a straight stream into a drum of burning scrap extinguished the fire.<ref name="CH-161" /> | ||
| Line 159: | Line 164: | ||
The agents for Class D fires are called dry powders, a term that must not be confused with dry chemical.<ref name="CO-72">Conroy (2008), p. 17-72.</ref> A [[dry chemical extinguishing agent]] is not effective on Class D fires, because the heat of the burning metal consumes it, and multipurpose dry chemical may react violently with some metals.<ref name="LA-17">Lake (2008), p. 17-17.</ref> A [[dry powder extinguishing agent]] is tested and listed for specific combustible metals.<ref name="CO-76">Conroy (2008), p. 17-76.</ref> By 2008, one portable dry powder extinguisher of 13.6 kg was available, with wheeled models of 68 and 159 kg. Its agent is sodium chloride with a thermoplastic that binds the particles into a crust on the fire.<ref name="CO-81" /> | The agents for Class D fires are called dry powders, a term that must not be confused with dry chemical.<ref name="CO-72">Conroy (2008), p. 17-72.</ref> A [[dry chemical extinguishing agent]] is not effective on Class D fires, because the heat of the burning metal consumes it, and multipurpose dry chemical may react violently with some metals.<ref name="LA-17">Lake (2008), p. 17-17.</ref> A [[dry powder extinguishing agent]] is tested and listed for specific combustible metals.<ref name="CO-76">Conroy (2008), p. 17-76.</ref> By 2008, one portable dry powder extinguisher of 13.6 kg was available, with wheeled models of 68 and 159 kg. Its agent is sodium chloride with a thermoplastic that binds the particles into a crust on the fire.<ref name="CO-81" /> | ||
[[File:fse-combustible-metal-v5.jpg|thumb|upright=1.5|alt=a portable Class D dry powder fire extinguisher with its applicator nozzle and Class D marking.|A portable dry powder extinguisher for Class D fires, with a low-velocity applicator for gentle discharge.]] | |||
In bulk, the two most common agents are sodium chloride and G-1 powder, a graded granular graphite with phosphorus compounds that is applied by hand. The heat makes the phosphorus compounds release vapours that keep air away, while the graphite conducts heat from the metal.<ref name="CO-81" /><ref name="CO-82">Conroy (2008), p. 17-82.</ref> Burning metal on a combustible surface is first covered with powder and then shovelled into a 25.4–50.8 mm (1–2 in.) layer of powder spread nearby.<ref name="CO-81" /> | In bulk, the two most common agents are sodium chloride and G-1 powder, a graded granular graphite with phosphorus compounds that is applied by hand. The heat makes the phosphorus compounds release vapours that keep air away, while the graphite conducts heat from the metal.<ref name="CO-81" /><ref name="CO-82">Conroy (2008), p. 17-82.</ref> Burning metal on a combustible surface is first covered with powder and then shovelled into a 25.4–50.8 mm (1–2 in.) layer of powder spread nearby.<ref name="CO-81" /> | ||
[[File:fse-combustible-metal-v4.svg|thumb|upright=1.5|alt=Two cross-sections: left, burning metal chips covered by a dry powder layer; right, a water jet scattering chips while a metal-water reaction releases hydrogen.|Left: a compatible dry powder applied gently covers the burning metal and excludes air. Right: a water jet can scatter burning particles, while the metal-water reaction can release hydrogen.]] | |||
{| class="wikitable" | {| class="wikitable" | ||
| Line 201: | Line 209: | ||
== Limitations and misconceptions == | == Limitations and misconceptions == | ||
Ignition is also inconsistent: fines that seem impossible to ignite can be ignited by static from a polyethylene bag. A long period of "being lucky" can give a false sense of security.<ref name="CH-155" /> | |||
The terms dry chemical and dry powder are easily confused.<ref name="LA-17" /> NUREG-1805, for example, lists "dry chemical agents" for Class D fires in one table, whereas the handbook states that dry chemical is not effective on them.<ref name="IS-B16">Iqbal and Salley (2004), p. B-16.</ref><ref name="LA-17" /> Even a genuine Class D agent may be hazardous on the wrong metal, and tables of agents differ: Yu and Newman (2008) list Met-L-X for sodium only, while NUREG-1805 lists it for seven metals.<ref name="CO-87" /><ref name="YN-91" /><ref name="IS-B31">Iqbal and Salley (2004), p. B-31.</ref> | The terms dry chemical and dry powder are easily confused.<ref name="LA-17" /> NUREG-1805, for example, lists "dry chemical agents" for Class D fires in one table, whereas the handbook states that dry chemical is not effective on them.<ref name="IS-B16">Iqbal and Salley (2004), p. B-16.</ref><ref name="LA-17" /> Even a genuine Class D agent may be hazardous on the wrong metal, and tables of agents differ: Yu and Newman (2008) list Met-L-X for sodium only, while NUREG-1805 lists it for seven metals.<ref name="CO-87" /><ref name="YN-91" /><ref name="IS-B31">Iqbal and Salley (2004), p. B-31.</ref> | ||
Latest revision as of 08:27, 8 November 2026
A combustible metal is a metal that is either combustible or explosive, because of its inherent properties or because of its physical form.[1][2] Nearly all metals burn in air under some conditions, but magnesium, titanium, zirconium, the alkali metals and several others ignite easily as thin sections, fine particles or molten metal. As massive pieces the same metals are very difficult to ignite, except for the alkali metals, which react with water and release hydrogen.[1]
Fires in combustible metals form Class D of the fire classes used in the United States.[3] They burn much hotter than flammable liquids, some keep burning in carbon dioxide, nitrogen or steam, and water can make them react violently.[4][5] Most ordinary extinguishing agents fail on Class D fires; the agent must match the metal. Suitable options include certain dry powders, sand and salts. Argon or helium can blanket some fires in enclosed spaces.[6][7] Finely divided metals add a second hazard, the dust explosion.[4] In the US practice described by the 2008 handbook, NFPA 484, Standard for Combustible Metals, addresses these hazards.[1]
Definition and scope
Combustion is a set of mainly exothermic chemical reactions between a fuel and an oxidiser, usually oxygen, that releases heat and light as a flame or a glow.[8] Most fuels contain a high proportion of carbon or hydrogen, as wood, coal, oil and gas do.[9] Pure metals are elements, made of only one type of atom, and contain neither.[10] Yet nearly all of them burn in air under certain conditions. Some oxidise so quickly in air or moisture that they heat themselves to their ignition temperature, while others oxidise so slowly that the heat escapes first.[1]
The handbook chapter on metals, revised by Christman (2008), defines combustible metals as metals that are either combustible or explosive.[1] It names calcium, hafnium, lithium, magnesium, niobium, plutonium, potassium, sodium, tantalum, thorium, titanium, uranium, zinc and zirconium, because thin sections, fine particles or molten metal of these elements ignite easily.[1] NUREG-1805 links the ease of ignition to a high specific surface area.[11] The hazards arise in production, processing, finishing, handling and storage, and in related systems such as process ventilation.[1]
Metals that are not normally regarded as combustible can still burn. Iron and steel ignite when finely divided or in contact with other burning materials, and clean, fine steel wool ignites fairly easily.[4] The usual forms of aluminium are not a factor in most fires, yet aluminium can burn vigorously when other materials burn around it.[7] Particle size, shape, quantity and alloy are therefore all factors in judging whether a metal is combustible.[4]
The alkali metals caesium, francium, lithium, potassium, rubidium and sodium, and alloys such as the sodium-potassium alloy NaK, react with water to generate hydrogen.[1] For them neither the quantity nor the particle size affects the risk of ignition. Lithium, sodium, potassium and NaK react with water at room temperature, while the others may react only when heated or in a fire.[1] They are typical examples of a water-reactive material.
In NFPA 10, Standard for Portable Fire Extinguishers, as described by Conroy (2008), Class D covers fires in combustible metals. The other classes are A (ordinary combustibles), B (flammable liquids and gases), C (energised electrical equipment) and K (cooking oils and fats).[3] NUREG-1805 gives the same classes from the 2002 edition of NFPA 10, and Grant (2002) calls Class D the class of reactive metals.[12][13]
Requirements and classification
In the United States, NFPA 484, Standard for Combustible Metals, as described by Christman (2008), has a chapter for each significant combustible metal. A chapter was added to assess whether a material should be treated as a combustible metal at all. It uses one test for combustibility and one for explosivity, and a material that fails either is a combustible metal to which NFPA 484 applies.[1]
The explosivity assessment in the 2006 edition of NFPA 484, as reproduced in the handbook, runs in four steps.[4] A screening test first classifies the explosibility of the dust. The minimum ignition energy and the minimum ignition temperature of a dust cloud follow. The third step measures the explosion severity in a 20-litre sphere: the maximum pressure, the maximum rate of pressure rise and the deflagration index KSt. The last step gives the limiting oxygen concentration and the minimum explosible concentration.[4]
Other US documents refer to NFPA 484 for particular situations, such as the use of sprinklers where magnesium is present and the dust collectors for aluminium.[14][7] Under NFPA 101 and NFPA 5000, as described by Demers (2008a), industrial facilities that produce explosive dusts of aluminium or magnesium are high-hazard industrial occupancies.[15] In the international transport hazard classes described by C. J. Wright (2008a), Division 4.1 (flammable solids) includes any metal powder that can be ignited.[16] Division 4.3 covers materials that, when wet, can become spontaneously flammable or give off flammable or toxic gas at more than 1 litre per kilogram per hour. Potassium is one of its examples.[16][17]
Combustion behaviour
Machining turns solid stock into chips and fines with a much larger surface-to-mass ratio, and some of these can ignite spontaneously or from outside sources.[18] The thinner the cutting and the smaller the particles, the greater the likelihood of fire, and heavy chips can hide a significant amount of fines.[19] According to Christman (2008), reactivity is somewhat reduced for sections thicker than 381 µm (0.015 in.). Finer particles and thinner sections are an increasing hazard.[4]
Alloys can differ widely in combustibility from their constituent elements.[4] A special case is the eutectic, an alloy with a melting point below that of each of its constituents, which can react vigorously and exothermically at lower temperatures. A titanium plate heated in contact with steel scale is one recorded example. The eutectic liquefies below the melting point of either material and the hot liquid spreads by flow.[4]
Burning metals generally reach much higher temperatures than burning flammable liquids. Some hot metals keep burning in carbon dioxide, nitrogen or steam, where ordinary combustibles would go out.[4] Sodium, potassium, magnesium, titanium and zirconium decompose carbon dioxide, so it cannot extinguish them.[20] According to Grant (2002), the same metals burn so intensely that they overpower Halon 1301.[21] Halogen-containing agents, including the halons, react violently with burning magnesium, and sodium reacts violently with halogenated hydrocarbons.[22][23]
Burning titanium produces little smoke, while burning lithium gives dense, profuse smoke. Zirconium powder moistened with water burns with near-explosive violence, but the same powder wet with oil burns quietly. Sodium melts and flows while burning; calcium does not.[4] Thorium, uranium and plutonium also emit ionising radiation, which complicates fire fighting and introduces contamination.[4]
Heat of combustion and oxygen demand
The heat of combustion of several metals is high. Babrauskas (2008a) lists 31.06 MJ/kg for aluminium, 24.73 MJ/kg for magnesium and 43.08 MJ/kg for lithium, each burning to its oxide.[24] Dry wood has a gross heat of combustion of about 18–21 MJ/kg. For most organic fuels the heat released per kilogram of oxygen is nearly constant at 13.1 MJ/kg, an assumption used in heat release rate measurements.[24]
The oxygen demand of a metal follows from stoichiometry, the proportions in which substances react.[25] A metal that burns to an oxide with x metal atoms and y oxygen atoms needs a mass of oxygen r per kilogram of metal. The heat released per kilogram of oxygen is its heat of combustion divided by r:
| Symbol | Description | Unit |
|---|---|---|
| mass of oxygen needed to burn one kilogram of metal to its oxide | kg/kg | |
| , | number of metal atoms and oxygen atoms in the oxide formula | – |
| , | atomic weight of the metal and of oxygen | g/mol |
| heat of combustion of the metal | MJ/kg | |
| heat released per kilogram of oxygen consumed | MJ/kg |
Worked example. Magnesium burns to magnesium oxide, MgO, so x = y = 1. With atomic weights of 24.31 for magnesium and 16.00 for oxygen, r = 16.00/24.31 = 0.658 kg of oxygen per kilogram.[10] Dividing 24.73 MJ/kg by this ratio gives 37.6 MJ per kilogram of oxygen, about 2.9 times the value for organic fuels.[24] A hydrocarbon fuel needs about 3.46 kg of oxygen per kilogram, so magnesium uses about one fifth as much.[25] The result assumes complete oxidation and says nothing about how easily a given piece ignites, which depends on its surface-to-mass ratio.[19]
| Metal | Oxide formed | Heat of combustion (MJ/kg) | Oxygen demand r (kg/kg) | Heat per kg of oxygen (MJ/kg) |
|---|---|---|---|---|
| Aluminium | Al2O3 | 31.06 | 0.890 | 34.9 |
| Calcium | CaO | 15.85 | 0.399 | 39.7 |
| Iron | FeO | 4.87 | 0.286 | 17.0 |
| Lithium | Li2O | 43.08 | 1.153 | 37.4 |
| Magnesium | MgO | 24.73 | 0.658 | 37.6 |
| Sodium | Na2O | 9.00 | 0.348 | 25.9 |
| Titanium | TiO2 | 19.6 | 0.668 | 29.3 |
| Zinc | ZnO | 5.37 | 0.245 | 21.9 |
| Zirconium | ZrO2 | 12.03 | 0.351 | 34.3 |
Ignition temperatures
Ignition temperatures of the pure metals in solid form range widely, from 69 °C for potassium in oxygen to 1,593 °C for titanium.[5] Many tabulated values refer to ignition in oxygen rather than air, and some are only bounds. Finer forms ignite much lower: magnesium ribbons and shavings at about 510 °C and fine magnesium powder below 482 °C.[5] NUREG-1805 reproduces the table from an earlier edition of the handbook with some different values, such as 555 °C instead of 1,000 °C for aluminium.[26]
| Metal | Melting point (°C) | Boiling point (°C) | Ignition temperature of solid metal (°C) |
|---|---|---|---|
| Aluminium | 660 | 2,452 | 1,000 (in oxygen; above this value) |
| Calcium | 842 | 1,441 | 704 |
| Iron | 1,535 | 3,000 | 930 (in oxygen) |
| Lithium | 186 | 1,336 | 180 |
| Magnesium | 650 | 1,110 | 623 |
| Potassium | 62 | 760 | 69 (in oxygen; spontaneous in moist air) |
| Sodium | 98 | 880 | 115 (above this value) |
| Titanium | 1,727 | 3,260 | 1,593 |
| Uranium | 1,132 | 3,816 | 3,816 (in oxygen; below this value) |
| Zinc | 419 | 907 | 900 (in oxygen) |
| Zirconium | 1,830 | 3,577 | 1,400 (in oxygen) |
Dust explosion hazard
Clouds of most metal dusts in air are explosive, so fugitive dust must be controlled.[4] Christman (2008) gives a rule of thumb: a dust layer thick enough to hide the colour of the surface below, or to write a name in, is an explosion hazard.[4] Tests by the US Bureau of Mines, published in 1964 and reproduced by Dobbs (2008), show how little energy and dust are needed. Their explosibility index runs from none (0) through weak (0.1), moderate and strong to severe (10 or more).[27]
| Metal | Dust ignition temperature (°C) | Minimum explosive concentration (g/m³) | Minimum ignition energy of dust cloud (mJ) | Maximum pressure (kPa, gauge) | Explosibility index |
|---|---|---|---|---|---|
| Aluminium | 343 | 45 | 50 | 503 | >10 |
| Magnesium | 327 | 40 | 40 | 621 | >10 |
| Titanium | 166 | 45 | 25 | 483 | >10 |
| Zirconium | −7 | 45 | 15 | 379 | >10 |
| Uranium | −7 | 60 | 45 | 365 | >10 |
| Zinc | 360 | 501 | 960 | 331 | <0.1 |
The zirconium and uranium ignition-energy entries are from 1 g test charges; larger quantities ignited spontaneously. The 1964 ignition energies are method-specific, and later test methods can yield lower values.[27][28]
Schwab (2008c) defines the deflagration index as the maximum rate of pressure rise in a closed, non-vented vessel multiplied by the cube root of the vessel volume. It is measured in a 1 m³ vessel or in a 20-litre sphere calibrated to give the same results; Schwab also notes that the smaller sphere often yields a lower value.[28] He lists 508 bar·m/s for a magnesium dust of 28 µm, with a maximum pressure of 17.5 bar (gauge).[29] An aluminium powder finer than 10 µm reaches 515 bar·m/s and 11.2 bar.[29] Zalosh (2002) gives the same 515 bar·m/s for aluminium, and no other dust in his table exceeds 209 bar·m/s.[30][31]
| Symbol | Description | Unit |
|---|---|---|
| deflagration index of a dust | bar·m/s | |
| maximum rate of pressure rise in a closed, non-vented test vessel | bar/s | |
| volume of the test vessel | m³ |
Worked example. For the aluminium powder with KSt = 515 bar·m/s, the maximum rate of pressure rise in a 1 m³ vessel is 515 bar/s. If that same KSt is assumed in a 20-litre (0.020 m³) sphere, its cube root is 0.271 m and the formula gives about 515/0.271 ≈ 1,900 bar/s. A dust at 209 bar·m/s would similarly give about 770 bar/s under this assumption. These are mathematical illustrations, not predictions for a different vessel or design values: the measured index can differ with vessel and test conditions, particle size, shape, moisture and solvent content.[28]
Metal dusts can also burn at lower oxygen concentrations than organic dusts. With carbon dioxide inerting, organic dusts have limiting oxygen concentrations of about 11–15 % by volume, but the limiting oxidant concentration of some metals is much lower.[30] Zalosh (2002) gives 2 % with carbon dioxide and 5 % with nitrogen for aluminium, and zero for zirconium, which indicates that zirconium powder is pyrophoric.[32][30] Carbon dioxide should not be used for inerting magnesium, silicon or zirconium, for which a noble gas such as argon may be needed.[33]
Bag-type dust collectors are usually avoided for fine metal dust, because the filter holds the particles in an envelope of oxygen where a small spark can ignite them.[19] Many light-metal dusts react with water to form hydrogen, and their wet sludge is highly explosive unless kept submerged.[34] Collectors for aluminium must be dedicated to aluminium, because iron or steel in the system can lead to a thermite-type reaction.[7] Aluminium dust is also the most challenging material in explosion suppression tests. Its total suppressed pressure can be held to 10–20 % of the maximum explosion pressure, against less than 5 % for organic fuels.[35]
Reaction with water
Water is usually the wrong agent for metal fires. A number of metals react exothermically with water to form hydrogen, which burns rapidly, and violent steam explosions can occur if water enters molten metal.[6] Christman (2008) compares applying water to a combustible metal fire with pouring gasoline on a house fire.[5] In the magnesium example he explains that oxygen from the water combines with magnesium while the released hydrogen may burn.[22] Frank (2008) notes that water otherwise remains stable in a fire and breaks down only in very special circumstances.[36]
The reaction ranges from minor to explosive.[14] As a rule, water is not used on fires in magnesium, titanium, sodium or hafnium, or in calcium, zinc and aluminium where these burn, although it can still cool exposures.[14] Large quantities of water have extinguished small amounts of burning magnesium without molten pools.[6] For zirconium, a water spray has no adverse effect on burning turnings, and a straight stream into a drum of burning scrap extinguished the fire.[7]
Wetted fines are particularly dangerous. Finely divided metal scrap that is wet, or wetted with water-soluble lubricant, may burn so rapidly and violently that nobody can get close enough to apply an agent.[37] Where metals are machined with combustible liquids, Class B and Class D materials burn together, and the fire will probably continue unless both are extinguished together.[5]
Extinguishing Class D fires
Portable Class D extinguishers, or agents developed for the metal involved, may control small fires.[5] The burning metal must be completely covered with a compatible agent, and its quantity must be well below what the agent on hand can handle.[5] Metal fires are hard to extinguish because of their very high temperatures and long cooling times.[6]
The agents for Class D fires are called dry powders, a term that must not be confused with dry chemical.[38] A dry chemical extinguishing agent is not effective on Class D fires, because the heat of the burning metal consumes it, and multipurpose dry chemical may react violently with some metals.[39] A dry powder extinguishing agent is tested and listed for specific combustible metals.[40] By 2008, one portable dry powder extinguisher of 13.6 kg was available, with wheeled models of 68 and 159 kg. Its agent is sodium chloride with a thermoplastic that binds the particles into a crust on the fire.[37]

In bulk, the two most common agents are sodium chloride and G-1 powder, a graded granular graphite with phosphorus compounds that is applied by hand. The heat makes the phosphorus compounds release vapours that keep air away, while the graphite conducts heat from the metal.[37][41] Burning metal on a combustible surface is first covered with powder and then shovelled into a 25.4–50.8 mm (1–2 in.) layer of powder spread nearby.[37]
| Agent | Main ingredients | Used on |
|---|---|---|
| G-1 or MetalGuard powder | graphitised coke and organic phosphate | magnesium, aluminium, uranium, sodium, potassium |
| Lith-X powder | graphite and additives | lithium, magnesium, zirconium, sodium |
| Foundry flux | mixed chlorides and fluorides | magnesium |
| Sodium chloride | NaCl | sodium, potassium |
| Soda ash | Na2CO3 | sodium, potassium |
| Lithium chloride | LiCl | lithium |
| Dry sand | SiO2 | various metals |
| Boron trifluoride gas | BF3 | magnesium |
| Argon or helium | Ar, He | any metal |
| Nitrogen | N2 | sodium, potassium |
A fire in an enclosed space can be smothered with an inert gas, such as argon or helium for zirconium.[7] Because certain metals react exothermically with nitrogen, helium and argon are the only acceptable inert gases for them, and halocarbons should not be used on metal fires.[6] Copper powder is very effective on lithium fires, and an alkali metal burning in an apparatus can often be extinguished by closing all openings.[42]
A forceful jet can scatter light turnings or powders and raise an explosive dust cloud.[4][43] The discharge of a dry powder extinguisher is therefore started slowly from a safe distance.[40] For small quantities of burning titanium or zirconium powder, the safest procedure is to ring the fire with a Class D agent and let it burn out.[44][7] In the US practice described by Conroy (2008), the amount of agent depends on the surface area and form of the metal. The maximum travel distance to Class D extinguishers is 22.9 m (75 ft).[45]
Prevention, storage and handling
Christman (2008) advises assuming that a fire will occur at some time when thin-section products of these metals are handled, so that good housekeeping and careful logistics limit the damage.[4] Turnings, chips and fines must be removed and fugitive dust collected. Failure to follow these principles is likely to cause significant fire loss and injury.[2]
Particulate forms are best stored in individual sealed metal containers rather than in bulk, in packages sized to the quantities used.[4] Some powders are stored under water, but spontaneous combustion can occur as they dry to a critical moisture level.[4] Zirconium powder is stored and shipped in 3.79-litre (1 gal) containers with at least 25 % water by volume.[42]
Radioactive metals add a contamination problem, because radioactivity is not altered by fire and continues wherever the metal is spread.[7] Collecting uranium, thorium or plutonium fines in containers of less than 3.79 litres (1 gal) has prevented spontaneous ignition. Fire tests showed that an air gap of about 102 mm (4 in.) between chip containers can stop a fire spreading from one to the next.[46]
Severson (2008) lists combustible metal waste, such as magnesium shavings, as solid waste.[47] Fine magnesium scrap can be burned in a safe outdoor area in a layer about 102 mm thick, and titanium fines coated with water-soluble oil have ignited spontaneously.[48] Christman (2008) also recommends practice in an isolated outdoor location before a real fire, and prefire planning by fire departments that protect businesses using these metals.[5]
Limitations and misconceptions
Ignition is also inconsistent: fines that seem impossible to ignite can be ignited by static from a polyethylene bag. A long period of "being lucky" can give a false sense of security.[5]
The terms dry chemical and dry powder are easily confused.[39] NUREG-1805, for example, lists "dry chemical agents" for Class D fires in one table, whereas the handbook states that dry chemical is not effective on them.[49][39] Even a genuine Class D agent may be hazardous on the wrong metal, and tables of agents differ: Yu and Newman (2008) list Met-L-X for sodium only, while NUREG-1805 lists it for seven metals.[45][6][50]
Published property data are uncertain. NUREG-1805 and the handbook print different ignition temperatures for the same metals.[26] Schwab (2008c) notes that the 20-litre sphere often gives values a factor of two lower than the 1 m³ vessel.[28] The sources also conflict on heat. Christman (2008) states that burning sodium releases about as much heat as the same mass of wood. Babrauskas (2008a), however, lists 9.00 MJ/kg for sodium against 18–21 MJ/kg for dry wood.[23][24]
Finally, the rule against water is not absolute. Large quantities of water have extinguished small magnesium fires, and water extinguishes burning zirconium scrap.[6][7] In the United States, NFPA 484, as described by Christman (2008), recommends automatic sprinklers for magnesium in combustible buildings, so that a fire is controlled before the metal becomes involved. That case is described in magnesium fire.[51]
See also
Notes
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Christman (2008), p. 6-153.
- ↑ 2.0 2.1 Christman (2008), p. 6-163.
- ↑ 3.0 3.1 Conroy (2008), p. 17-71.
- ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 Christman (2008), p. 6-154.
- ↑ 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 Christman (2008), p. 6-155.
- ↑ 6.0 6.1 6.2 6.3 6.4 6.5 6.6 Yu and Newman (2008), p. 2-91.
- ↑ 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 Christman (2008), p. 6-161.
- ↑ Van Rossum (2027), §4.1.
- ↑ Van Rossum (2027), §4.2.
- ↑ 10.0 10.1 Drysdale (2008), p. 2-14.
- ↑ Iqbal and Salley (2004), p. B-29.
- ↑ Iqbal and Salley (2004), p. B-13.
- ↑ Grant (2002), p. 4-153.
- ↑ 14.0 14.1 14.2 Frank (2008), p. 17-41.
- ↑ Demers (2008a), p. 20-208.
- ↑ 16.0 16.1 C. J. Wright (2008a), p. 13-120.
- ↑ C. J. Wright (2008a), p. 13-119.
- ↑ Dobbs (2008), p. 9-68.
- ↑ 19.0 19.1 19.2 Dobbs (2008), p. 9-70.
- ↑ Wysocki (2008), p. 17-6.
- ↑ Grant (2002), p. 4-152.
- ↑ 22.0 22.1 Christman (2008), p. 6-157.
- ↑ 23.0 23.1 Christman (2008), p. 6-159.
- ↑ 24.0 24.1 24.2 24.3 Babrauskas (2008a), p. 6-281.
- ↑ 25.0 25.1 Van Rossum (2027), §3.3.
- ↑ 26.0 26.1 Iqbal and Salley (2004), p. B-30.
- ↑ 27.0 27.1 Dobbs (2008), p. 9-69.
- ↑ 28.0 28.1 28.2 28.3 Schwab (2008c), p. 6-147.
- ↑ 29.0 29.1 Schwab (2008c), p. 6-145.
- ↑ 30.0 30.1 30.2 Zalosh (2002), p. 3-404.
- ↑ Zalosh (2002), p. 3-409.
- ↑ Zalosh (2002), p. 3-403.
- ↑ Ural and Garzia (2008), p. 17-149.
- ↑ Dobbs (2008), p. 9-71.
- ↑ Zalosh (2002), p. 3-418.
- ↑ Frank (2008), p. 17-31.
- ↑ 37.0 37.1 37.2 37.3 Conroy (2008), p. 17-81.
- ↑ Conroy (2008), p. 17-72.
- ↑ 39.0 39.1 39.2 Lake (2008), p. 17-17.
- ↑ 40.0 40.1 Conroy (2008), p. 17-76.
- ↑ Conroy (2008), p. 17-82.
- ↑ 42.0 42.1 Christman (2008), p. 6-160.
- ↑ Dobbs (2008), p. 9-72.
- ↑ Christman (2008), p. 6-158.
- ↑ 45.0 45.1 Conroy (2008), p. 17-87.
- ↑ Christman (2008), p. 6-162.
- ↑ Severson (2008), p. 7-60.
- ↑ Severson (2008), p. 7-66.
- ↑ Iqbal and Salley (2004), p. B-16.
- ↑ Iqbal and Salley (2004), p. B-31.
- ↑ Christman (2008), p. 6-156.
References
- Babrauskas, V. (2008a). Tables and charts. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-269–6-294). National Fire Protection Association.
- Christman, T. (2008). Metals. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-153–6-164). National Fire Protection Association.
- Conroy, M. T. (2008). Fire extinguisher use and maintenance. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-71–17-92). National Fire Protection Association.
- Demers, D. P. (2008a). Industrial occupancies. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 20-203–20-210). National Fire Protection Association.
- Dobbs, P. G. (2008). Metalworking processes. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 9-67–9-72). National Fire Protection Association.
- Drysdale, D. (2008). Physics and chemistry of fire. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-3–2-20). National Fire Protection Association.
- Frank, J. A. (2008). Characteristics and hazards of water and water additives for fire suppression. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-31–17-43). National Fire Protection Association.
- Grant, C. C. (2002). Halon design calculations. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 4-149–4-172). National Fire Protection Association; Society of Fire Protection Engineers.
- Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
- Lake, J. D. (2008). Chemical extinguishing agents and application systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-17–17-29). National Fire Protection Association.
- Schwab, R. F. (2008c). Dusts. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-141–6-152). National Fire Protection Association.
- Severson, P. R. (2008). Hazardous waste control. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 7-57–7-69). National Fire Protection Association.
- Ural, E. A., & Garzia, H. W. (2008). Explosion prevention and protection. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-141–17-165). National Fire Protection Association.
- Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
- Wright, C. J. (2008a). Managing the response to hazardous material incidents. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 13-117–13-144). National Fire Protection Association.
- Wysocki, T. J. (2008). Carbon dioxide and application systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-3–17-15). National Fire Protection Association.
- Yu, H.-Z., & Newman, J. S. (2008). Theory of fire extinguishment. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-79–2-92). National Fire Protection Association.
- Zalosh, R. (2002). Explosion protection. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-402–3-421). National Fire Protection Association; Society of Fire Protection Engineers.