John Elihu Hall.

The American journal of science online

. (page 19 of 61)
Online LibraryJohn Elihu HallThe American journal of science → online text (page 19 of 61)
Font size
QR-code for this ebook

IB not dry but it does not contain enough water to affect the
" result essentially. In the determination of the ferrous iron,
- . W€i,^ have employed a weight burette instead of a volume
. Hesults with natxcral pyrite and marcasite. — The minerals
^taployed were pyrite from Elba* and marcasite from Joplin,
Ma They were purified with great care and then analyzed.f
jxhe only impurities found were small quantities of silica and
^a HQinute trace of copper in the marcasite.

4 Marcasite Pyrite Gal. for FeS«

1 Fe 46-53 46-49 46-56

. S 53-30 63-49 53-44

? ■ ' SiO, -20 -04


^ ... 100-03 100-02 10000

iThd oxidation coefficients obtained by us were 56 for pyrite
.^RHt 14 for marcasite, while Stokes found 60 and 18 respect-
ively^. The differences have not been entirely accounted for,
illHi^gh we have taken somewhat greater precautions in our
%roilc. However, and this is the point to be emphasized, the
millts of each are probably consistent among themselves.

Metermination of the relative quantities of pyrite and
oasite in mixtures, — By grinding together the two min-
, in different proportions and then determining the oxida-
tit^ number for tne mixture, Stokes constructed a curve rep re-
some of the later experiments, a pjrrite from LeadviUe, Col., con-
ig 0*1 per cent copper, was nsed.

pr the method of analysis, see Allen and Johnston, Zs. anorg. Chem.,

Digitized by VjOOQ IC

178 Alleuj Crenshaw^ Johnston, and Larsen —

senting the oxidation coefficients for all possible mixtures.
His curve had the form of a eutectic curve with the lowest
point at ten per cent, pyrite having an oxidation coefficient of
about 15. Fig. 4 shows the curve which we obtained for mix-
tures of pyrite and marcasite. Careful experiment failed to
reveal any mixture which gave a lower oxidation coefficient
than pure marcasite. The difficulty of determining accurately
the composition of an unknown mixture, by a measurement of

Fio. 4.


I 50











Digitized by VjOOQ IC

Mineral Sulphides of Iron.


site. Of course, the method only applies where the substance
contains no other reducing agent than pyrite and marcasite.
If, therefore, we had present another crystal form of FeS„ the
results would be unreliable. Although both pyrite and mar-
casite have been repeatedly detected by the microscope in the
synthetic products, no evidence of another form has come to
light. It is possible that in some instances (which will be

Eointed out in the proper place) amorphous FeS, may
ave been present. In most cases, however, the crystalline
structure was so apparent even to the unaided eye that the
presence of any amorphous material was quite unlikely, and
this conclusion was only confirmed bv the microscopic study.

Marcasite, — Marcasite is the principal product when hydro-
gen sulphide acts directly on ferric sulphate at 200^.

Exp. 1-5 g. NH,Fe(S0,),.12H,0 m 100" water saturated
at room temperature with hydrogen sulphide, was sealed up and
heated in a bomb for several days at 200°. To insure a suflS-
cient quantity of the product for experiment, three tubes were
thus heated under the same conditions. The product was
removed, ground fine, purified, and the oxidation number
determined. It was found to be 23*6, corresponding to about
48 per cent pyrite.

Infiuence of free acid on the proportion of marcasite. — The
equation H,8 -f Fe,(SO,). = S + 2FeS0, + H,SO, shows that
sulphuric acid is a product of the reaction in which the marca-
site forms, and its concentration evidently increases as the
reaction proceeds. It was, therefore, a plausible hypothesis
that the concentration of acid influences the crystal form ; it
seemed possible that the pyrite might have formed in the
earlier stages of the process, when the acid was weaker. If
this view were correct, a greater initial concentration of acid
should result in more marcasite. The hypothesis is proved
correct as shown by the results collated in the adjoining table.

Table I.
The effect of free HaSOi on the formation of marcasite at 200°.


! Foand

NH4Fe(S04),.12 H,0

Water saturated
with H,S

Vr^^ w QO • Oxidation
FreeH,S04» number

Per cent



0-60 g. 1 23-6
0-57 g. 18-9
0-78 g. j 17-0
1-18 g. 1 16-5


*Thi8 includes the acid formed in reduction of the ferric iron by hydro-
gen sulphide.

Digitized by VjOOQ IC

180 AUen^ Crenshaw^ Johnston^ and Larsen —

Tablb IL
EMeet oi tempenitiire on the fomiAtioii of maieante.





Sulphnr Free HjSO*

Water sata-
rated with

Oxidation Percent


H.S at 0*





0-6 g.

017 g.

100 «=


300"" ■


0-6 g.

017 g.

100 «





0-6 g.

0-17 g.

• 100"





0-6 g.

017 g.

100 '^^^





0-5 g.

017 g.




Table II shows the influence of temperature on the reaction.
Here ferrous sulphate and sulphur, the products of the action
of liydrogen sulphide on ferric sulphate, were taken. The
quantities of acid used, by inadvertence, were considerably less
than intended, but the evidence shows the influence of tempera-
ture very plainly. The higher the temperature the greater is the
quantity of pyrite form^. A word is here needed on the
question of the formation temperature. The furnace was first
heated to the required temperature, then the cold bomb con-
taining the tube was introduced. Obviously, the reaction could
not take place entirely at one temperature. At the lower tem-
peratures, however, the reaction is very slow, so that the time
required for the bomb to reach the temperature of the furnace
is not important. When the bombs were heated to 300°, many
hours were required to reach the maximum temperature, it
would not be worth while, after having shown that the two
variables, temperature and acid concentration, both influence
the product of this reaction, to study the problem in great
detail, but it is interesting to note that the reaction will proceed
at ordinary temperatures, and also that pure marcasite may be
obtained by a proper combination of the two variables. Thus
at 100° the precipitate formed from a solution containing
1'18 per cent of free sulphuric acid gives the oxidation coetB-
cient 14'5 and is therefore pure marcasite. From 2 liters of a
solution which contained 3 per cent of hydrous ferrous sulphate
and 0*15 percent of free sulphuric acid, 1 gr. of precipitate was
obtained at room temperature in about three weeks. Unfortu-
ately, there was an accident in the determination ^f the oxida-
tion coefficient of this product, but we can state that it contained
less than 10 per cent of pyrite. For every temperature there
appears to be a quantity of acid which inhibits the reaction:

Digitized by VjOOQ IC

Mineral Sulphides of Iron, 181

I eSO,+ H,S + S = FeS,+ H.SO,. The quantity is smaller the
lower the temperature. It appears to bear no relation to the
solubility of the sulphide, for at room temperature this quantity
is only a small fraction of 1 per cent. At 200° it lies between
3*5 per cent and 5 per cent ; at least this is true for periods of
a few weeks.

Crystals of mareasite, — Measurable crystals of marcasite
were obtained by the slow action of hydrogen sulphide on fer-
ric sulphate or chloride at several temperatures up to 300®
(see p. 174). The general problem of malcing measurable crys-
tals is one of the most troublesome in synthetic mineralogy.
As yet we have no light on the means of controlling the
number of nuclei which form in the process of crystallization.
In general, the more soluble minerals are obtained in larger
crystals. Likewise, larger crystals are obtained from a medium
in which they are more soluble. In preparing well-formed
marcasite crystals, some unaccountable failures were met with,
though generally the products obtained by the method previ-
ously described contained a number of crystals which were
measurable. The marcasite crystals were like the natural
mineral in color and luster and the axial ratio deduced from
the angular measurements was a\h\c ^ 0*7646 : 1 : 1-2176 as
compared with 0*7680 : 1 : 1*2122 for natural marcasite (Gold-
schmidt). The striations which marked the crystals agreed
with orthorhombic and not with regular symmetry (see III,
Crystallographic Study).

Formation of pyrite* — While the product of the action of
sulphur and hydrogen sulphide on ferrous salts is largely mar-
casite, the percentage increasing with the quantity of free acid
present, pyrite is the principal product where the solution
remains neutral or but slightly acid.

Action of hydrogen sulphide on ferric hydroxide. — Freshly
precipitated ferric hydroxide is instantly blackened by hydrogen
sulphide. The product is a mixture of ferrous sulphide and
sulphur, as shown by the following. Freshly precipitated fer-
ric hydroxide was washed free of soluble salts, suspended in
water and treated for some time with hydrogen sulphide. A
portion of the black amorphous precipitate dissolved in cold
dilute acid with evolution of hydrogen sulphide, leaving a
residue of amorphous sulphur. Another portion was hrst
digested with ammonium sulphide solution. After filtering

*For former syDtheses of pyrite, see Wohler (Ann., xvii, 260, 1836).
Wohler heated an intimate mixture of FeaOs, S, and NH4CI till the NH^Cl
was sublimed. He obtained some small brass yellow tertrahedra and oota-
hedra. Senarmont (loc. cit.) obtained FeSa by heating ferrous salt solutions
with alkaline poly sulphides. Geitner (Ann. 129, 350, 1864) heated metallic
iron with a solution of sulphur dioxide to about 200°. His product may
have been marcasite. See also Doelter (Zs. Kryst., zi, 80, 1886).

Digitized by VjOOQ IC

182 AUen^ Ot'enshaw^ Johnston^ and JLarsen —

and washing out the excess of reagent, the black residue dis-
solved without leaving any sulphur behind. The product
therefore must have contained free sulphur and could not have
been ferric sulphide, Pe,S„ though the latter, according to
Gedel,* decomposes with dilute acid, giving the same products
as the above mixture. A product made bv the action of
hydrogen sulphide on ferric hydroxide was washed into a glass
tube with about 100^ water, saturated with hydrogen sulphide
at room temperature, sealed and heated at 140° for seven days.
The solution when cooled and opened still smelled strongly of
hydrogen sulphide. The product had become quite dense and
had a yellowish grey color. It was boiled in hydrochloric acid
for some time to dissolve any unchanged ferric hydroxide, or
ferrous sulphide, and further purified as usual. The oxidation
number was 49, corresj^onding to 87 per cent of pyrite. The
work was repeated with ferric hydroxidef which had been
dried at 100*^ to make it easier to handle. It proved, however,
much less susceptible to hydrogen sulphide. It had to be heated
repeatedly at 140° with saturated hydrogen sulphide water
before the color of the oxide of iron had disappeared entirely.
After purification, the product gave the oxidation number 50*4,
corresponding to 90 per cent pyrite.

Action of sulphur on pyrrhotite in the presence of a sol-
vent. — The formation of pyrite, just described, is evidently a
result of the direct union of sulphur and ferrous sulphide, the
first product of the reaction. The hydrogen sulphide water
probably acts as a weak solvent. Similarly, the marcasite may
be re^rded as a product of the addition of sulphur to ferrous
sulphide, which forms gradually from solution. The formation
of pyrite by the action of sulphur on crystalline pyrrhotite,
the relation of which to ferrous sulphide will be shown farther
on, proves conclusively that, at a given temperatm'e, it is not
the exact nature of the solid phase which reacts with the sul-
phur, but the composition of the solution in which it forms,
that determines whether the product shall be pyrite or marca-
site. 2"2 g. pyrrhotite prepared in the laboratory, and 0*8 g.
of sulphur were put into a glass tube, to which was added a
solution of 0*1 g. of sodium bicarbonate in 100^° water. Before
sealing the tube, the solution was partially saturated with hydro-
gen sulphide. In composition this solution was similar to that
of a warm "sulphur" spring, and it served as a solvent for the
sulphur, which was graaually absorbed by the pyrrhotite. The
tube and its contents were heated for two months at 70°.
The product at the end of that time still contained sulphur and
undecomposed pyrrhotite. To remove the latter it was boiled

♦ Jonr. fOr Gasbel., xlvili., pp. 400 and 428, 1905.

f According to Gedel (loc. cit.), Fe«Os.HsO is thus obtained.

Digitized by VjOOQ IC

Mineral Sulphides of Iron,


for a long time with 20 per cent hydrochloric acid. The res-
idue was dense and brassy-yellow. It was finely ground, puri-
fied as usual, and tested by Stokes's method. It gave an oxi-
dation number of 55'1, which corresponds to pure pyrite within
the limits of error of the method. To make sure that no mis-
take had been made in this test, some pure natural pyrite was
compared the next day with the same ferric sulphate solution.
lOO*^*^ of the sulphate solution, after it had been reduced with
the synthetic pyrite, required 42'51 g. of permanganate solution.
100^^ which had been reduced by natural pyrite took 42*56 g.
of the permanganate.

The action of sulphur on pyrrhotite was tried again at 300°,
where the reaction was of course much more rapid than at 70°.

Into the tube were put 5 g. powdered pyrrhotite, 1"75 g. sul-
phur, 0-2 g. KaHOOj, and lOO''*' water partially saturated with
H,S. The tube was heated four days at 300°. The oxidation
number of the purified product was 52-0, corresponding to
about 95 per cent pyrite. It is possible that these products
contained a very little undecomposed pyrrhotite or perhaps
amorphous FeS„ both of which would nave undoubtedly had
a similar effect as marcasite in lowering the oxidation number.

Table III.
The oxidation numbers of FeSs formed from alkaU polysnlphide solution.









3 days

100 '^^^



0-76 g.




5 days




•75 g.




7 days








The action of alkali polysvlphides mi ferrous salts. —
Senarmont* in 1851 showed by analysis that the product of
the action of alkali polysulphides on ferrous salts at 180° is
FeS,. The question of the crystal form was not investigated.
The black amorphous precipitatef which is obtained at room
temperature by tne above reaction appears to be a mixture of
sulphur and ferrous sulphide, at least it decomposes with dilute
acids, giving a residue of amorphous sulphur, while hydrogen
sulphide escapes. On heating, disulphide of iron gradually

• Loc. cit.

f Gedel (loc. cit.) claims that this precipitate is FeaSi.

Digitized by VjOOQ IC

184 AUen^ Crenshaw^ Johnston^ and Lar%en —

forms, though some of the black precipitate is still unchanged
after it has been heated several days at 100° with excess of the
polysulphide. The oxidation coefficients of several such prod-
ucts formed at different temperatures, and carefully purified as
usual, are given in Table III. Evidently they are not pure
pyrite, a result somewhat surprising in view of the previous
work ; for if we obtain marcasite from the more acid solutions,
marcasite with pyrite from those which contain less acid, and
pure pyrite from practically neutral solutions, we should natu-
rally expect pure pyrite from alkaline solutions. Further inves-
tigation has led us to believe that the products of the alkaline
solutions do not contain marcasite, but are mixed with amor-
phous disulphide. Stokes explained, very plausibly, that the
reason why marcasite gave more free sulphur than pyrite when
it reacted with ferric sulphate was because it was more soluble ;
a fortiori^ amorphous disulphide would give, under similar
conditions, still more free sulphur because it is the most solu-
ble of the three. The evidence for the existence of amorphous
disulphide in the products of alkaline solutions is as follows :
While the products of acid solutions which contain the most
marcasite are the best crystallized, those from alkaline solu-
tions which, judging by their oxidation coefficients, contain
the most, are almost black, dull and lusterless at the lo.wer
temperatures. The quantity of pyrite is increased by raising
the temperature or prolonging the time of reaction, — both con-
ditions which are favorable to the crystallization of an amorph-
ous substance. Moreover, marcasite is not changed by heating
in alkali polysulphide solutions, as we found by heating some
of the natural mineral in powdered form for several days at
300° with polysulphide of sodium. The oxidation coefficient
remained 14'5. The influence of temperature on the formation
of pyrite from ferrous salts and alkali polysulphides is shown
in Table III. The influence of time is proved by the two fol-
lowing experiments : 3 g. FeS0,.7H,0, 2*5 g. Na,8, and 0*75 g.
sulphur, and 100°*^ water, were heated 2 days at 100°. The pro-
duct contained about 75 per cent pyrite.* A similar system
heated for 7 days at the same temperature gave a product contain-
ing about 95 per cent pyrite. The results are calculated on the
supposition that they contain marcasite; if they contain amor-
phous disulphide instead, the true percentage of pyrite should
of course be higher, since a given quantity of amorphous disul-
phide would be equivalent to a greater quantity of the less sol-
uble marcasite, but the order of the results would of course
remain the same. The products obtained at 300° were yellower,

* There is an apparent discrepancy between this last result and the one
quoted in Table III under the caption ** IOC.'* In the latter case the excess
of polysulphide was much smaller.

Digitized by Google |

Mineral Sulphides of Iron, 185

denser, and in direct sunlight showed a decided metallic laster,
while their oxidation coefficients approach those of pure pyrite.
It may therefore be safely stated that the product of the union
of ferrous sulphide and sulphur from an alkali polysulphide
solution is at first amorphous disulphide of iron which gradu-
ally crystallizes to pyrite.*

The formation of iron disuk>hide hy the action of sodium
thiosvlphate on ferrous salts. — In the endeavor to explain the
formation of pyrite and marcasite in nature, the following
hypothesis presented itself. Iron disulphide of either form
may oxidize under surface conditions to ferrous thiosulphate
by direct addition of oxygen ; this is transported by circulating
waters to some point wnere it is reduced again to its former
condition. In a partial study of the oxidation of marcasite,
no trace of thiosulphate was discovered. At the same time
the effort to obtain the disulphide of iron by reduction of
the thiosulphate was successful. When water solutions of
ferrous sulphate and sodium thiosulphate are heated in sealed
tubes, even to temperatures under 100®, the iron disulphide is
precipitated with sulphur. By quantitative experiments which
loUow, the reaction is proved to be :

4Na,S,0, + FeSO, = FeS, + 3S + 4Na,S0,.

Exp. 1. 5 g. FeS0,.7H,0, 18 g. Na,S,0,.5fl,0, and 2h^
water were heated in a sealed tube for 9 days at 90®. All but a
trace of the iron was precipitated. The precipitate of FeSj+ S
was washed in air-free water and dried in vacuo. The sulphur
was extracted by carbon disulphide and the residue of FeS, was

Cal. from the
Found above eqnation

FeS, + S 3-82 3-86

FeS, 2-l'7 2-16

Exp. 2. 2 g. FeS0,.7H,0, 18 g. Na,S,0,.5H,0 and Zf>'^ water
were sealed in CO, and heated for one day at about 150®. The
precipitate was filtered and washed with air-free water. The
solution was boiled in a current of carbon dioxide to remove a
trace of hydrogen sulphide, and an aliquot part was titrated
with standard iodine solution.

No. 1. 1/5 of the solution required 16*135 g. iodine solution.
Cal. for the whole, 80*675 g. iodine solution.

No. 2. 2/5 of the solution required 32-332 g. iodine solution.

* In a recent paper (Zs. angewandte Chem., xxiv, 97, 1911), ** Die Bildnng
von Eisen Bisulphide in LSsungen tmd die Enstehnngen der nattlrlichen Pyrit-
lagem," W. Feld states that whenever sulphur and ferrous sulphide are boUed
in neutral or weakly acid solutions, pyrite forms.

Am. Jour. Sci.— Fourth Series, Vol. XXXIII, No. 195.— March, 1912.

Digitized by VjOOQ IC

186 AUeUj Crenshaw^ Johnston^ wad Larsen —

Cal. for the whole, 80*830 g. iodine solution. The iodine solu-
tion contained '005445 g. iodine per g.

80*675 and 80*830 g. iodine solution are respectively equivj^ i-
lent to 0*858 and 0*860 g. Na,S,0,-Sil,0. Therefore 7*142 g.
and 7*140 g. NaaS,0,.5H,0 were consumed in the reaction.
The equation demands 7*137 g. for 5 g. FeS0,.7H,0. \

Foimd Cal.

FeS, + S 1*547 1-556

FeS, 0*872 0-863

Sodium thiosulphate when heated with water in sealed tubes
forms hydrogen sulphide and sodium sulphate. Na,S,0,+H,0=
H,S+l5ia,S0^. The reaction at 200^ is quite incomplete,
though no thiosulphate was obtained when a solution of sodium
sulphate saturated with hydrogen sulphide was heated under
the same conditions. When 1 g. Na,S,0,.5H,0 and 20*^*^ water
were heated 4 days at 200^, the thiosulphate undecomposed, as
determined b^ standard iodine solution, was 0*753 g. 0*247 g.
decomposed is equivalent to 0*141 g. Na^SO^. The solution
after titrating with iodine was precipitated with barium
chloride. BaSO, found 0*223 g. equivalent to 0*136Na,SO,.

At first it was thought that the reaction between ferrous sul-
phate and sodium thiosulphate was to be explained as follows :
(1) Na,SA+H.O = Na.S0,+H,8, (2) FeS0,+4H,S = FeS,+
3S+4H,0. Later it was found that from ferrous chloride the
same mixture of sulphur and FeS, was precipitated. Of
course, the reaction represented by equation (2) could not go
on with ferrous chloride. Therefore, the following is probably
the true explanation of the reaction.

(1 .) FeSO, + Na S O, = Na,SO, + FeS.O,
(2.) FeS,03 + 3Na,S,0, = FeS, + 3S + 3Na,SO,

Form of FeS^ obtained by heating ferrous salts with sodium
thiosulphate. — Neither marcasite nor pyrite is obtained pure in
this reaction. The product shows in crusts the color of pyrite,
but it is poorly crystallized and may contain amorphous disul-
phide. A product prepared at 90°, tested by Stokes' method,
behaved like a mixture of 70 per cent pyrite and 30 per cent
marcasite. Another product formed at 300°* tested in the
same way acted like a mixture of 72 per cent pyrite and 28
per cent marcasite. Though this reaction — the reduction of
ferrous thiosulphate by sodium thiosulphate — doubtless has no
significance as applied to geology, it is probable that the fer-
rous thiosulphate might be reduced by other reagents, and it is
possible that ferrous thiosulphate may be formed in nature
under some conditions, but of this we have no evidence.

* This was the maximum temperature. The reaction may have been com-
plete before this temperature was reached. /

Digitized by


Mineral Sulphides of Iron. 187

The transformation of marcasite into pymte. — More than
fifty yeai-B ago, Wdhler* tried the experimeDt of heating both
minemls for four hours at the temperature of boiling sulphur
(about 445®) without observing any change in either of them.
Our own results indicate that marcasite undoubtedly changes
here, but very slowly. When marcasite was heated at 610® in
hydrogen sulphide gas for 3 hours, it lost about 2'5 per cent
sulphur and became strongly magnetic, owing, of course, to the
formation of some pyrrhotite. A finely ground sample, after
being thoroughly boiled out with hydrochloric acid, appeared
decidedly yellower and less lustrous than marcasite.f

The comparison is best made by placing the sample to be
tested alongside of a fragment of marcasite which has had all
tarnish removed by recent boiling in hydrochloric acid (Stokes).
A finely ground and purified sample of the heated marcasite
gave the oxidation number 56 instead of 14 as previously. At

Online LibraryJohn Elihu HallThe American journal of science → online text (page 19 of 61)