THE COLLIERY GUARDIAN

AND

JOURNAL OF THE COAL AND IRON TRADES.

Vol. CXIII.	FRIDAY, JUNE 22, 1917.	. No. 2947.

THE SPONTANEOUS FIRING OF COAL.*

By J. S. HALDANE, M.D., F.R.S.

During the last four years, since the establishment
of the Doncaster Coal Owners’ Research Laboratory
in 1913, 13 papers (of which 12 were by Mr. T. F.
Winmill and Mr. J. Ivon Graham) on the conditions
under which spontaneous fires in coal occur have
appeared in the Transactions of the Institution of
Mining Engineers as reports to the Doncaster Coal
Owners’ Committee, f Each of these papers deals in
detail with some aspect of the central subject, and it
is perhaps somewhat difficult to follow out their full
bearing. It has therefore been suggested to the writer
that a paper very shortly reviewing the general results
so far reached might be of service to mining engineers.

It is unnecessary to refer to the great losses of coal,
dangers to life, particularly in gassy seams, and
expense caused by spontaneous firing or heating of
coal. It is only with the causes of this process that
the writer proposes to deal. A large number of facts
with regard to spontaneous heating have been familiar
for many years. Thus it is well known that only cer-
tain varieties of coal are readily liable to spontaneous
heating. It is equally well known that in practice
coal does not spontaneously fire in the absence of
oxygen, or if the coal is in large lumps, or is freely
exposed to the air in a layer not exceeding a certain
depth depending on the nature of the coal and the
degree to which it is broken up. For many years
there has been no practical doubt that the spontaneous
heating of coal is due to an oxidation process occur-
ring under such conditions that the heat formed by
the oxidation cannot escape as rapidly as it is
liberated. A very valuable review of work on this
subject was recently published by Prof. J. B. Porter,
of McGill University, Montreal.+ Very little was
known, however, as to the nature of the oxidation pro-
cess, the laws determining it, and the amount of heat
produced when a given volume of oxygen combines
with coal at the ordinary temperature of a mine.

The first experiments completed for the Doncaster
Committee were on this latter point, and were carried
out at Cambridge by Mr. Lamplough and Miss Hill§
by a delicate calorimetric method which had already
been used at Cambridge in a different form for physio-
logical work. These experiments showed that, for a
given consumption of oxygen, less heat is produced
than during ordinary combustion, at a high tempera-
ture. Mr. Winmill found later that the heat actu-
ally produced had been somewhat over-estimated, and
that it is about 2T calories per litre z of oxygen
absorbed. This is scarcely half the' heazt produced
when one litre of oxygen combines with coal during
combustion at a high temperature. Nevertheless, Mr.
Winmill was able to show that when a small sample of
suitable coal dust placed at ordinary temperature on
the laboratory • table was supplied with sufficient
oxygen, and at the same time prevented from losing
any heat, it burst into flame within about 24 hours.
Members of the Institution of Mining Engineers who
visited the Doncaster laboratory in 1915 will doubtless
remember the demonstration of the rapid rise in tem-
perature of a sample of powdered coal in his beautiful
apparatus.

One of the difficulties in obtaining accurate measure-
ments arose from the fact, of which we were at first
unaware, that coal takes up oxygen in two different
ways. In the first place, it combines chemically with
oxygen, evolving much heat in doing so. The oxygen
so combined cannot be recovered from the coal by the
vacuum pump. But coal also takes up oxygen in
another way, apparently analogous to simple solution
of gases in liquids, and without sensible evolution of
heat. The whole of this oxygen can be recovered by
pumping, as was quite recently shown in a paper by
Mr. Graham. In the case of oxygen, nitrogen, and
hydrogen, the amount of gas which goes into solution
in the coal is far greater than the amount dissolved by
an equal weight of.water; but, just as in the case of
water, the amount dissolved varies directly as the
pressure of the gas, and thus follows Henry’s law for
the solution of gases in liquids. The extraordinary
solubility of gases, and liquids in coal and other solids
has turned out to be a fact of very great scientific and
practical interest, and Mr. Graham’s experiments
have thrown a flood of n* * * §ew light on the presence of
firedamp and other gases in coal; but the writer must
not do more at present than point out the sharp dis-
tinction between the solution or “adsorption” of
oxygen in coal and the true “ chemical ” combination
which causes spontaneous fires.	1

Rate of Oxidation of Coal.

The next point investigated was the manner in which
the oxidation of freshly-powdered coal diminishes with
time. At first the oxidation is very rapid, but within
a few hours the rate diminishes greatly, and then

* Paper read before the Institution of Mining Engineers,

t Trans. Inst. M. E., 1913 to 1917, vols. xlvi. to lii.

t“ Weathering of Coal.” 1915; Report to the Depart-
ment of Mines, Canada. Government Printing Bureau,
Ottawa.

§ Trans. Inst. M. E., 1913, vol. xlv., p. 629,

becomes daily less and less, until after a few weeks
the rate is so small that it can hardly be measured.
Mr. Winmill, who investigated this matter, has pub-
lished a number of curves showing quantitatively this
characteristic behaviour of coal dust at various tem-
peratures up to 160 degs. Cent. It appears that it is
some particular kind of material in the coal that is
attacked at ordinary temperatures by the oxygen, and
that only a limited quantity of this material is pre-
sent. When it has all been oxidised, the coal is
incapable of any further spontaneous heating, unless
its temperature is first artificially raised. There seem,
however, to be different varieties of easily oxidisable
material; for with rise of temperature not only is the
rate of oxidation increased, but the total amount of
oxygen entering into combination may also be greatly
increased. This indicates that the greater part of the
easily oxidisable substances are so constituted as not
to be attacked by oxygen at a lower temperature, but
begin to be oxidised when the temperature rises, and
then accelerate the rate of heating.

The rate at which coal combines with oxygen from
air passed through it was found by Mr. Winmill to be
dependent on the fineness of subdivision of the coal.
Fine coal dust oxidises rapidly, and the oxidisable
material contained in it is correspondingly rapidly
used up. With increasing coarseness of subdivision,
the rate at which the oxidation occurs becomes
increasingly slower, and the time required for comple-
tion of the oxidation correspondingly longer. This
behaviour is explained by Mr. Graham’s more recent
experiments on the penetrability of coal by gases.
When thin slices of coal were tested as regards the
rate at which various gases pass through them, it was
found that the rate of passage is excessively slow.
Intact, coal is, in fact, a highly gastight material.
The very slow rate at which a lump of gassy coal gives
off the firedamp dissolved in it is thus easily intelli-
gible ; and the correspondingly slow rate at which
oxygen penetrates into a lump of coal explains the
slow and prolonged character of the oxidation process
in coarse coal. When solid coal is broken up or
fissured, the penetration of oxygen is, of course,
greatly accelerated. A solid pillar or side of coal
cannot oxidise and heat, because the oxygen cannot
penetrate. As soon, however, as the pillar or side
begins to disintegrate under the influence of pressure,
or from any other cause, the oxygen gets in, and heat-
ing may result. Mr. Graham’s experiments explain
completely the fact that heating started by fresh
breaks in the coal may occur along the sides of a road
in solid coal many years after the road has been made.
The oxygen has never got to this coal till the fresh
breaks opened a passage for it. The writer at one
time thought that coal is almost as porous as certain
other notoriously porous strata, so that with every
change of atmospheric pressure air passes in and out
of it. This idea is entirely erroneous: changes of
atmospheric pressure can have no appreciable influence,
on the giving off or taking up of gas by solid coal.
Air will pass quite freely through bricks or ordinary
building stone, but not through solid coal.

An examination by Mr. Winmill of fresh samples of
coal from various seams and districts disclosed great
variations in the amount and nature of the easily
oxidisable material which causes spontaneous heating.
Every variety of coal examined, from anthracite to
lignite, contained some of the oxidisable material, and
was thus liable to liberate some heat in the presence
of air. But in the case of some coals—for instance,
the Welsh steam coals examined—the total quantity
of material oxidised by air at temperatures up to 140
degs. Fahr, was so small that spontaneous firing could
only occur in very exceptional circumstances. In the
case of the Welsh steam coal, it was specially remark-
able that nearly the whole of the easily oxidisable
material was oxidised at as flow a temperature as
85 degs. Fahr. On raising the temperature to 140
degs. Fahr., the oxidation was more rapid at first,
but also more rapidly over, so that the total amount
of oxygen which entered into combination, and conse-
quently the total heat production, was only very
slightly increased. Coal from the seams in question
never gives rise in practice to gob-fires or spontaneous
fires in cargo ships ; and the- results of the laboratory
experiments made the reason of this perfectly dear.

In most of the other varieties of coal the total
amount of oxidation increased rapidly with rise .of
temperature, and spontaneous firing would occur
easily if loss of heat was prevented. The actual
liability of a mine to spontaneous fires could not,
however, be predicted, except in a limited sense, from
the laboratory examination of the coal. The reasons
for this seem fairly evident. In order that spon-
taneous fives in coal may occur, there must not only
be coal which is capable of firing:, but it piust be finely
broken in sufficient amount, with sufficient heat insu-
lation and a sufficient supply of oxygen. In the actual
working of a mine one or other of these necessary con-
ditions may not be present. If, for instance, the
whole of a seam, is removed in working, or little or no
Crushing of coal occurs, there will be no spontaneous

fires; and even when there is much finely-divided coal
in the goaf, no gob-fires can occur if there is no access
of air to the goaf. It is quite clear, therefore, that
other factors beside the liability of the coal to oxida-
tion are of the utmost importance in determining the
actual occurrence of spontaneous firest The pecu-
liarities of chemical composition which render coal
liable to oxidation are not under human control; but
the other factors which are essential to the actual pro-
duction of spontaneous fires are under engineering
control at one point or another, although sometimes
the cost of an effective control may be prohibitive.

Spontaneous Firing of Coal.

The writer will now consider more closely the manner
in which a spontaneous fire in coal is produced. The
simplest case is where a quite limited amount of coal
is concerned, as in Mr. Winmill’s experimental fires
with about Jib. of freshly-powdered coal placed in a
Dewar flask in such a way as to prevent loss of heat.
When an air current is passed through the coal, there
is, of course, a rise of temperature. If the current is
very brisk, so that the oxygen percentage of the air is
only very slightly reduced, the rise of temperature
will be very slight, and will soon come to an end, since
with a very slight rise of temperature the rate at
which the air carries off heat becomes equal to the
rate at which heat is formed. If, however, the air
current is slower, equilibrium of heat production and
heat loss does not occur till the temperature has risen
considerably. At this higher temperature, the rate of
oxidation in the coal will have considerably increased,
and material which was not oxidisable at all at the
lower temperature may have begun to oxidise. For
this reason, the equilibrium temperature will be much
higher than it would otherwise be. It is easy to see
that with a still slower air current the equilibrium
point may never be reached till the coal is actually on
fire. It is equally easy to see that the amount of
oxidation which occurs may be insufficient to raise the
temperature of the coal more than a little, even if the.
air current is cut down to a minimum, so that practi-
cally no heat is carried away in the air. With the
Welsh steam coal and anthracite this was the case.

The Welsh anthracite and steam coals which Mr.
Winmill examined were incapable of firing in a small
heat-insulated mass; but various other coals were quite
capable of firing. From the data as to oxygen absorp-
tion, it is now easy to .calculate whether a small mass
of coal is capable of spontaneous firing, since the heat
resulting from the oxygen absorption is now known,
as well as the specific heat of the coal and the heat
required to evaporate the moisture contained in it.

‘ When a large mass of coal is concerned, the condi-
tions are more complicated. In the case of a large
mass of coal, with air passing through it constantly
in a certain direction, the heat produced by oxidation
in each layer of the coal is carried onwards to the next
layer; and in each successive layer there is a further
rise of temperature. As the course of the air current
is followed, the coal in time becomes hotter and hotter,
until at last it becomes incandescent. This cumula-
tive rise of temperature cannot occur in a small mass
of coal, however perfect the heat insulation, and how-
ever favourably adjusted the air current may be; for
in a small mass of coal the temperature will be more
or less equalised by conduction. The mass must be
large enough to permit of very unequal temperatures
at different parts. When this condition is fulfilled,
and there is also practically complete external heat
insulation, it seems to the writer that any sort of
broken coal exposed to air may fire, even though this
coal would be quite incapable of spontaneous firing in
a small mass with complete external heat insulation.
The fires which occur in pit heaps are probably due
almost entirely to the cumulative heating effect just
discussed. Their occurrence does not seem to depend
merely on the external heat insulation resulting from
the size of the heap. Considering the large propor-
tion of stone usually present in pit heaps, it is difficult
to see how the mass of material could spontaneously
heat up evenly and as a whole to any considerable
extent.

The conditions underground are often peculiarly
favourable to a cumulative heating effect; for, owing
to the ventilation, it is very apt to happen that a
continuous slow current of air passes horizontally
along a layer of broken coal. The heat is thus carried
on, and the tepiperature rises higher and higher along
the course of the air current, until at a certain dis-
tance a dangerous temperature is reached. Ordinary
gob-fires seem to be always produced in this way. ' A
layer of'small'coal and stone 2 or 3 ft. deep, and with
free air above it, will seldom heat appreciably, since
the convection currents in the coal are in an upwards
and downwards direction, so that any heat formed
escapes rapidly to the air. But when the roof has
settled down, and there are still conditions which cause
a slow flow, of air along the layer of small coal, cumu-
lative heating will at once begin to tell.

It must be remembered that the heat carried for-
ward by an air current is not merely heat measured
by simple rise in temperature in the air, but all the
latent heat due to evaporation of moisture. This
latent heat, deposited again further on, becomes of
more and more preponderating importance as the
temperature approaches, boiling point. Of course, the
moisture contained naturally in coal, or deposited in