John Elihu Hall.

The American journal of science online

. (page 20 of 61)
Online LibraryJohn Elihu HallThe American journal of science → online text (page 20 of 61)
Font size
QR-code for this ebook

525®, a sample of marcasite heated 4^^ hours in hydrogen
sulphide gave the oxidation number 55'8. Under similar
conditions at 450®, a sample of marcasite heated 4 hours gave
the oxidation number 15*7, which corresponds to 5*5 per cent
of pyrite. The sample was returned to tne furnace and heated
again 5 hours. This time the top layer in the crucible gave,
after purification, the oxidation number 27, corresponding to
53 per cent pyrite, while a deeper layer in the same crucible
gave 31, which indicates about 61 per cent of pyrite. At 450®,
therefore, dry heating in H,S changes marcasite, to pyrite
rather slowly — 50 per cent-60 per cent in 9 hours. Heated
to 410® for 4 hours, the oxidation number was 13'5, showing
that no measurable change had occurred.

J. Konigsberger and O. ReichenheimJ found a marked
decrease in the electrical resistance of marcasite in the neigh-
borhood of 520®. They noted that the sulphide then possessed
a specific resistance of the same order as pyrite, and rightly in-
terpreted their results to mean that marcasite had changed into
pyrite and the change is irreversible. It is difilcult to see in
their results, however, any support of their statement that the
change appears to begin between 250® and 300®, while our
results contradict it.

An effort was made to effect the transformation at a lower
temperature in the wet way. At 350® marcasite heated in a
sealed tube with a small amount of dilute sulphuric acid partly
changed to ferrous sulphate and sulphur dioxide, but the solid

•Ann. Chem. Pharm., 90, 256, 1854.

t PTiite is naturally a more lastrons mineral than marcasite. The daUer
color of heated marcasite is to be ascribed to the very large number of
minute crystals in the product, and the lack of continuous surfaces.

{Neues Jahrb., ii, 86, 1906.

Digitized by VjOOQ IC

188 AUeUj OrenshaWy Johnston ^ and Lar%en —

portion gave no sign of change, and at 300° a powdered sample
which was repeatedly heated for several days' time with
sodium sulphide and polysulphide solutions^ was equally

Density, — The density of the marcasite heated to 610° rose
from 4'887 to 4*911. The density of pure pyrite is 5*02+
The change in color, and, more convincingly,*^ the oxidation
number, show that the substance is pure pyrite after heating,
yet its density is too low. The explanation is probably to he
sought in the porosity of the product.

Influence ofpresaure on the change marcasite -^pyrite. — Dr.
A. Ludwig, at our request, kindly undertook some experiments
on the influence of pressure in transforming marcasite to

Eyrite. A few grams of marcasite were compressed for five
ours at a pressure of about 10,000 atmospheres. At the end
of the penod, the oxidation by Stokes' method showed no
change m the substance. Later Johnston and Adams devised
an apparatus in which the marcasite could be heated by a
resistance coil while subjected to hydrostatic pressure of about
2000 atmospheres in petroleum oil. A number of experiments
were tried between 300° and 400°, but Stokes' reaction showed
no pyrite formation. The oxidation method was perhaps not
quite so certain here on account of the fact that the oil was
partially decomposed at the higher temperatures and the
product may have contained some reducing matter which could
not be removed by petroleum "ether." A mixture of pyrite
and marcasite containing any such reducing impurity, as we
have seen (p. 184), would give too low an oxidation number.
Thus it might happen that a little pyrite could be overlooked.
These experiments are of considerable interest because there
are very few data on the effect of pressure in irreversible
changes.* We do not know whether a difference in densitv in
the right direction would favor the change or not, since
Le Chatelier's law applies only to reversible changes. If the
speed of the chan^ is influenced by pressure quite apart from
the volume relation, it may perhaps be retarded rather than
accelerated. Until apparatus is developed which will give
higher temperatures and at the same time high pressures, this
problem must wait. At present the a8sumption,f which has
been made in geology, that pressure favors all changes which
are accompanied by reduction in volume, irrespective of their
reversibility, is unwarranted.

Monotropic relation of marcasite to pyrite, — A crucible con-
taining 50 g. pyrite was rapidly heated (20° per minute) in H,S

* Van't Hoff, Vorlestmgen, 2nd. £d., Bratmschweig, 1901, vol. I, p. 286.
f See Van Hise, A Treatise on Metamorphism, Monograph No. 47, U. S.
Geol. Survey, 1904, pp. 215, 363.

Digitized by Vj005 IC

Mineral Svlphides of Ito7i,


over a range from 400^ to 600°. The curve was perfectly
smooth. When a similar charge of marcasite was heated in
the same way, an acceleration of the temperature was plainly
seen on the curve between 500° and 600°. (See fig. 5.) The

520 *•
500 **

S 440



Fig. 5.

-i zr .

Z 7 X -

j.j.J'^Jl _

jju/f _

Tti. w


jti/ilT :








^ //' / i it







- ^? ^-j-




Jl tj.





Jl tl t


111 tl H-

tiX-tt ^




Fio. 5. Thermal curves showing eyolntion of heat when marcasite is
changed into pyrite. (Carves 3^ 4 and 5.) Carves 1 and 2 are thermal
curves of pyrite.

Digitized by VjOOQ IC

190 AUe7iy Crenshaw^ John%tcni^ and Laraen —

experiment was repeated several times with similar results.
Under these conditions, there is a plain evolution of heat
accompanying the change of marcasite into pyrite. This
shows, of course, that marcasite possesses the more energy* of
the two and is a monotropic form. This condition of instability
is in accord with the more rapid oxidation of marcasite in
nature, and in it is probably to be found the reason for the dif-
ference in behavior between marcasite and pyrite toward other
oxidizing agents. Monotropic forms often crystallize from
some particular solvent or within a limited temperature range.
The formation of marcasite from acid solutions is in accord
with this, though as yet we do not understand the reason for
it. A rise in temperature doubtless increases the velocity of
the change marcasite -> pyrite. At low temperatures this
becomes mfinitesimal or zero ; above 450° it becomes measur-
able. This irreversible relation has a bearing on the question
of paramorphs of iron disulphide, for it is impossible to see
how pyrite could change to marcasite without first passing into
solution, while the opposite change is experimentally estab-
lished. Pai'amorphs of pyrite after marcasite are certainly
possible, but paramorphs of marcasite after pyrite are evidently

The agency of organic matter in the formation of natural
pyrite and marcasite. — The fact that pyrite is sometimes found
in nature replacing wood has been alluded to. Liversidgef gives
an example of recent pyrite which is found on twigs in a hot
spring at Tampo, N. Z. The sulphides of southwestern Mis-
souri, including pyrite and marcasite, are frequently associated
with asphaltic matter, and in Oklahoma this is sometimes so
great in quantity as to interfere with the concentration of the
ores. (Lindgren.) It is pretty generally believed by geolog-
ists that the organic matter of certain shales acted as a precip-
itant of the pyrite they contain. Such a shale underlies the
sulphide deposits of Wisconsin. We learn from Mr. W. H.
Emmons of the U. S. Geological Survey that this shale con-
tains a small quantity of hydrogen sulphide, which naturally
may have been the precipitating agent. Coals, also, in which

fyrite is commonly found, are frequently permeated within
ydrogen sulphide.
The '^ole m micro-organisms in the formation of iron di-
sulphide. — Tjie question naturally arises whether there is any
connection between organic matter and the formation of

*Cayazzi (Rend. Accad., BologDa, N. S., il, 205, 1898) states that the
heats of combustion of pyrite and marcasite are identical (1550 cal.). This is
certainly incorrect.

t J. Royal Soc. N. S. Wales, xi, 262, 1877.

Digitized by VjOOQ IC

Mineral Sulphides of Iron. 191

hydrogen sulphide. In the putrefaction of organic matter
hydrogen sulphide is one of the products, and Gautier* has
surmised that the pyrite which sometimes forms the substance
of fossil bones and shells is precipitated by hydrogen sulphide
which is given off slowly by the organic matter during its de-
composition. The formation of pyrite in this case would dif-
fer from that described in the previous pages only in the source
of the hydrogen sulphide, which is here a product of micro-

There is another way in which micro-organisms produce
hydrogen sulphide, and that is by the reduction of sulphates.
According to Beyerinck,t a considerable number of bacteria,
algae, flagellata and infusoria show this kind of activity.
Spirillum desulf uricans is one of the most important. As these
organisms are active only in neutral or alkaline solutions, fer-
rous sulphide is precipitated whenever ferrous salts as well as
sulphates are present. The black mud of many swamps, pools,
and even seas (e.g., the Black Sea), as well as sea coasts,:|:
which are intermittently overflowed, contains ferrous sulphide.
Mr. C. A. Davis of the tJ. S. Bureau of Mines, who has had
large experience on this subject, informs us that he has always
found hydrogen sulphide in peat-bogs into which tide-water
finds its way. Apparently, the formation of pyrite or mar-
casite through the agency of micro-organisms has not been
observed, but only an influx of air with excess of hydrogen
sulphide would be needed to change the ferrous sulphide into
disulphide. That micro-organisms are directly responsible for
any great quantity of the pyrite or marcasite of nature seems
unlikely because in the first place they are probably not active
far from the surface of the ground. They have been dis-
covered at depths of only four or five raeters.§ A fraction of
a per cent or free acid usually inhibits the growth of these
organisms ; therefore they could not live in the solutions from
which marcasite appears to have formed. Pyrite and marca-
site are not infrequently associated with minerals like chalco-
pyrite, which proves the presence of copper in the original
solutions, and copper is exceedingly poisonous to practically
all micro-organisms. It is possible, however, that the reduc-
tion of sulphates like gypsum and sodium sulphate by micro-
organisms may be an important source of hydrogen sulphide
in nature.

Distinct conditions leading to pyrite or marcasite in nature,
— The geological relations of marcasite indicate that it is a

♦C.R., cxvi, 1494, 1893.

f Centr. Bakter. u. ParaBitenknnde, i, pp. 1, 49, 104, 1895.

i HiS was found in sea-water by B. Leroy, Ann. Ch. Ph., Iviii, 382, 1846.

^ Hygiene des Bodens, Josef yon Fodor, Jena, vol. i, Pt. I, p. 187, 1898.

Digitized by VjOOQ IC

192 Alleuy CrenshaWj Johnston^ and Larsen —

product of surface conditions. The oxidation of either pyrite
or raarcasite gives first a mixture of sulphuric acid and ferrous
sulphate which by further oxidation easily gives ferric sul-
phate. The action of hydrogen sulphide and atmospheric
oxygen simultaneously on the acid ferrous solution would
lead to the same goal. We have found how hydrogen sulphide
acting on acid solutions, especially in the cold, gives rise to
marcasite. We have also found that above 450° marcasite
could not form, thus further confirming geological deductions.

Pyrite, being a stable form, probably crystallizes under a
considerably wider range of conditions than marcasite. The
evidence oi synthetic study is that the formation of pyrite
is favored by high temperature and by solutions which con-
tain little or no tree acid. In accord with these, we have the
following geological deductions. First, pyrite is a product of
hot springs. In the springs of Carlsbad, which have a tem-
pe^-ature of about 55° C, recent pyrite is observed.* The
waters contain sulphates and a trace of hydrogen sulphide,
and are slightly alkaline. The lagoons of Tuscany are deposit-
ing pyrite from their hot waters. Bunsenf found that the
hot vapors of the fumaroles of Iceland were gradually chang-
ing the ferrous silicate of the basalts into pyrite.

More important geologically is the fact that the product of
deep veins by ascending waters is always pvrite, never marca-
site. Such waters are naturally hot, and commonly if not
always alkaline.:|: We can now see that a separation of pyrite
from a magma is entirely possible, while the temperature of
any magma would doubtless be incompatible with the existence
of marcasite.

27i6 occurrence of pyHte and marcasite together, — Hintze^
mentions thirty-one instances where pyrite and marcasite are
found intergrown or precipitated one upon the other. Stokes ||
also tested a number of specimens which proved to be mix-
tures of pyrite and marcasite, some of them intergrown in con-
centric layers. In other places, e.g., in Joplin, Missouri, the
two minerals have been ooserved by us in the same hand spe-
cimen .^^ According to F. L. Ransome,** the two minerals
occur together, though perhaps not intergrown, in Goldtield,
Nevada. These facts show veiy strikingly not only the small
influence of nuclei tn directing the form of the disulphide

* Daubr^y Q^ologie exp^rimentale, Paris, 1879, p. 98.

+ Pogg. Ann., Ixxxiii, 259, 1851.

XA. hot acid solntion in contact with carbonate or most silicate rocks
wonld first be neutralized and then become alkaline.

SLehrbnch der Mineralogie, vol. i, pp. 724-778, 820-882.

I Loc. cit.

^See W. S. T. Smith and Siebenthal, U. S. G. S. Folio 148.

** Private commanication. Specimens were also submitted by Mr.

Digitized by VjOOQ IC

Mineral Sulphides of Iron, 193

which separates from solution, but also that comparatively
slight differences in conditions may give rise to one or the
other. Further, that the two minerals may have formed
at the same time in some instances. The synthetic experi-
ments which have been described proved that the minerals
very commonly formed together,* as polymorphic forms which
are monotropic are apt to do. Cold solutions which were suffi-
ciently acid gave marcasite ; warm or hot solutions, either neu-
tral or alkaline, gave pyrite, and intermediate conditions gave

IL Pyrrhotite,

Composition. — Special interest attaches to the composition
of pyrniotite, which, despite much discussion, is still an unset-
tled question. The various formulae,! Fe^Sg, FCj^Sja, FcnSn + ,
and FeS which have been assigned to it by various authors
rest on widely varying analytical data. The analytical
methods have no doubt l>een at fault, but the more important
question concerns the physical homogeneity of the substance.
This has been unusually troublesome. Pyrrhotite almost
always occurs in the massive condition, a circumstance which
naturally arouses suspicions of its purity, while its opacity
makes it impossible to put the question to an optical test.

Many years ago Lindstrom;]: subjected all the Known analyses
of this mineral to a careful critique. Those which for anj'
reason, such as defective analytical methods or impure mate-
rial, appeared unconvincing were reiected.§ In the remainder,
the ratio! of iron to sulphur was calculated and found to vary
from 1:1-06 up to 1:1*19. Some yeara later HabermehlTi inves-
tigated the same question. He crushed pyrrhotite to a fine
powder, covered it with water, and endeavored, by means of a
strong horseshoe magnet, to separate it into fractions varying
in magnetic intensity. Such fractions as he obtained in thjs
wav showed no systematic difference in composition. Haber-
mehl used in his experiments the pyrrhotite from Bodenmais.

A very satisfactory general discussion of the question of
admixtures in pyrrhotite is also given in Habermelil's paper.
He decided that pvrrhotite could not contain free sulphur
because carbon disulphide removes none from it, neither could

* It may be, howeyer, that pjrite was formed first and was succeeded by
marcasite as the acidity of the solation increased.

t Sidot jndged from experimental work with FcsO* and HjS that pyrrho-
tite should have the formula FeaS*, C. R. Ixvi, 1257, 1868.
t Ofv. Ak. Stockh., xxxii, No. 2, 25, 1875.
S 18 analyses out of 43 were thus rejected.
I I In the calculation of the ratios Lindstrom took, in place of the small

I percentage of nickel found in many of these analyses, the equivalent of iron.

^ Ber. Oberhess. Ges. far Natur- und Heilkunde, xviii, 83, Oiessen, 1879.

Digitized by VjOOQ IC

194 AUeUj CrenshaWy Johnston^ and Larsen —

it contain any disulphide of iron because this is insoluble in
hydrochloric acid, wnile only sulphur remains after pyrrhotite
has been boiled with this reagent. Judging from the proper-
ties ascribed to Fe,S,, he concluded that this also could not be
present. Habermehl was thus forced to conclude, like Lind-
strom, though on the basis of further evidence, that pyrrhotite
was variable in composition, and it may here be said that no
evidence of later date has ever disproved this conclusion. At
that time a homogeneous solid of variable composition was an
anomaly. To-day such substances are quite generally recog-
nized under the category of solid solutions.

Hyvoihesia of solid solution, — We proposed to test the
hypotnesis of solid solution by preparing a series of synthetic
pyrrhotites and measuring some property of them. Pyrrhotite
was prepared by Berzelius, Rammelsberg and others, generally
by heating pyrite. One can also begin with marcasite, which,
as we have seen, is first changed to pyrite between 450® and
600°, or with sulphur and iron. We have tried all three
methods, though most of the work has been done with pyrite
from Elba, lliis very pure mineral, an analysis of whicli was
given on p. 177, Part I, was kept in a vacuum desiccator,
from which portions were taken from time to time as required.

Apparatus. — The apparatus used in the synthesis of pyrrho-
tite appears in fig. 6. The crucible, C, containing the pyrite
is of unglazed porcelain, 48""* high X 37"°" outside diameter at
the top and 22""° at the bottom. It has a doubly perforated
graphite cover, E^ through the central orifice of which passed
the glazed Marquardt jacket, -4, which shields the thermo-
element. Through the second orifice passes a similar tube, B^
open at the lower end, which is traversed by a current of
hydrogen sulphide.*

The crucible is inclosed in a large porcelain tube, /?, 40—45°'™
inside and 50™°* outside diameter, and 50*^" in length. In some
of the experiments the crucible was supported in the hot zone
by a strong graphite rod, (r, which was fastened to the cover,
-c, and clamped outside the large tube, 2?, while the cover itself
was fastened to the crucible by three small pegs. By means of
this device, the crucible could be quickly lowered at any time
to the bottom of D and thus rapidly cooled. In other experi-
ments, a much shorter porcelain tube was substituted for J9,
in which instances the crucible was supported by a fire-clay
pedestal which rested on the bottom of the tube. The upper

* Since the ferrous snlphide from which hydrogen snlphide is generated,
contains free iron, the gas was passed through boiling snlphur to remove
hydrogen, before reaching the furnace.

Digitized by VjOOQ IC

Mineral Sulphides of Iron.


Fig. 6.


Fig. 6. Apparatne for the preparation of pyrrhotite.

end of the tube in all cases is closed by the graphite cover, J7,
through which pass A and B. The crucible and its contents
are heated by a platinum-resistance furnace, F^ as shown in
the figure.

Digitized by VjOOQ IC

196 Allen^ Crenshaw^ Johnston^ and La/raen —

Synthesis. — When heated in hydrogen sulphide, pyrite de-
composes gradually into sulphur and pyrrhotite. The decom-
position may be first detected at about 575° (see p. 205). At
about 665° it proceeds rapidly ; still, even after some hours at
750°, the dissociation is never quite complete, and when the
pyrite used is rather coarse (sized between screens of 8-40
meshes per cm.), several per cent of it persist in the product.
The latter is therefore melted and cooled, and the resulting
material, which is now free from pyrite,* serves as a starting
point in the preparation of pure pyrrhotite.

When this is reheated it loses or gains sulphur according to
conditions. A series of products was made by heating the sul-
phide in hydrogen sulphide for some hours at different meas-
ured temperatures, and then cooling it in nitrogen.f To
facilitate the formation of homogeneous products the sulphide
was carefully sized between screens of 16 and 40 meshes per
cm. These products were all similar in appearance to natural
pyrrhotite. They were all dense, opaque, metallic, more
brownish than pyrite, and only very slightly tarnished. Some
tests were made to prove that the quantity of oxide on the
surface was negligible.

Weighed samples were heated in dry hydrogen to a red heat
and the water formed was absorbed by passing through a
calcium chloride tube. The surface of the grains became
bright in a few minutes. The tube was cooled in hydrogen,
which was then displaced from the apparatus by dry air. The
water thus collected corresponded to less than 0*1 per cent of
oxygen in two different tests. Furthermore it will be noted
later that the preparations which were cooled in nitrogen were
comparable in density with those which were prepared in
another way and were not tarnished in the slightest degree
(seep. 199).

Comj^ositioti of synthetic pyrrhotite. — The sulphur was
determmed in each of the synthetic pyrrhotites by a method:}:
worked out in this laboratory, which was proved accurate
within at least 0*2 per cent of the sulphur present.

Treitschke and Tammann§ state that the fused sulphide of

* Except for a slight Buperficial layer which decompoBes in the next

f The nitrogen was prepared by Enorre's method (Die Chem. Ind., xxv.,
581, 550, 1902 ; Chem. Centralb., i, 125, 1903), i. e., by dropping a saturated
solution of sodium nitrite from a dropping funnel into a solution of ammo-
nium sulphate and potassium chromate. The gas was passed through dilute
sulphuric acid to remove ammonia, and then successively through long col-
umns of chromic acid to remove oxides of nitrogen ; sulphuric acid to remove
moisture, and flnaUy over-heated copper to remove oxygen, or any traces of
oxide of nitrogen which may have escaped.

t Allen and Johnston, Jour. Ind. & Eng. Chem., ii^ 196, 1910; Zs.
anorg. Chem., Ixix, 102, 1911.

§ Z. anorg. Chem., xlix, ^20, 1906.

Digitized by VjOOQ IC

Mineral Sulphidea of Iron.


iron corrodes and dissolves porcelain, bnt it should be noted
that they made their fusions in carbon-resistance furnaces
without further protection from the air.

Our unglazed crucibles appeared quite unattacked in an
atmosphere of hydrogen sulphide. Analyses revealed the
presence of about 0*25 per cent of silica in our preparations,
though we believe that most of this was derived from minute
fragments of the crucible, which are difficult to exclude entirely
when the cake of sulphide is broken out of the crucible. Thus
No. 7 gave 0'33 per cent. No. 10, 0*26 per cent, and No. 3
gave 0*22 per cent and 0*24 per cent of silica in duplicate deter-
minations. This of course includes the silica in the original
pyrite, which, however, was negligible, — '04 per cent.

JSdation of the specific volume to the composition. — The
specific gravity* of each preparation was determined and from

Table IV.f
CompoBition, deneitj and specific YolnmeB of pyrrhotites.


Cal. FeS





Cal. dene,
at 4''

Sp. V.

























































Online LibraryJohn Elihu HallThe American journal of science → online text (page 20 of 61)