Arnold Bennett.

Chemical news and journal of industrial science online

. (page 38 of 88)
Online LibraryArnold BennettChemical news and journal of industrial science → online text (page 38 of 88)
Font size
QR-code for this ebook

J. D£WAK, M.A , F.R.8., lacksonian Professor,

University of Cambridge.

Wb have already described (Phil. Trans,, A, z888) the
remarkable spedrum of the oxy-hydrogcn flame burning
at the ordinary atmospheric pressure. Recently we have
examined the f pedlrum of the same flame at various pres-
sures : hydrOj^en burning in excess of oxygen up to a
pressure of 4) atmospheres, and oxygen in excess of hy-
drogen up to a pressure of 25 atmospHeres, also that of
the mixed gases buming in carbonic acid gas.

The apparatus employed was an adaptation of one of
the tubcfs used in our experiments on the absorption
spcdlra of compressed gases {Phil, Ma^., September, 1888,

* AbstraA of a Paper read before the Royal Society, Feb. 19, xbgu

Digitized by



Influence of Pressure on the Spectra of Flatnes.

r CMiyiCAL llBWt,

\ March 26, 1891.

and Roy. Soc, Proc, vol. xlvi., p. 222). It consisted of a
steel cylinder, about 50 m.m. in internal diameter, and
225 m.m. long, fitted at one end with a quartz stopper, a,
in the annexed figure, and with a jet, 6, for burning the
gas, adapted by a properly fitting union joint to the oppo-
site end. There were two tubes, c and d^ conne^ed to the
cylinder at the sides, of which one, c, served for the intro-
duAion of gas, while the other, d, was fitted with a stop-
cock and was used to draw off the water formed, or to re-
duce the pressure of the gas in the cylinder if that was
deiired. The flame was observed, nearly end on, through
the quartz stopper. The whole apparatus was kept cool
by a stream of cold water running on to a sponge cloth
wrapped round the cylinder. In the course of the tube
conveying gas to the jet b was interposed a small cylinder,
e, in which sodium was placed, and by heating this, the
gat entering could be charged with sodium vapour.

The gases were supplied from steel cylinders into which
they had been cot: pressed, and the pressure was
registered by a gauge attached to the tube by which the
gas entered the experimental cylinder. Commercial com-
pressed gases were used, containing a sensible percentage
of air.

When hydrogen was the gas forming the burning jet, it
was lighted at the end of the tube b before introducing it
into the experimental cylinder. When it was desired to
have a jet of oxygen burning in hydiogen, this could be
managed by introducing oxygen through the second tube
and increasing the supply of hydrogen until the flame

passed over to the oxygen jet. The same result was
sometimes attained by first filling the experimental
cylinder by a gentle stream of hydrogen through the side
tube e before the end with the tube b was screwed on ; the
hydrogen as it issued was then lighted, aud the jet, with ,
a gentle stream of oxygen issuing, inserted and screwed
down. The stopcock, ;, was kept open until this was
done, and then by closing s, and admitting more gas from
the reservoirs, the pressure in the experimental cylinder
could be increased at pleasure.

Hydrogtn Burning in Oxygen,
The first observations were made with a jet of hydrogen
burning in oxygen. As the pressure rose, the luminosity
of the flame increased, as long ago described by Frankland
(•• Experimental Researches," p. 905). The colour of the
flame, \iewed end on, was yellow, as if it contained
sodium ; but on examining it with a spedroscope, it was
found to give a ccntinuous spedrum interse^ed by many
shaded bands, and the D lines of sodium were only
faintly present. The shaded bards were faint at a pres-
sure of 5 atmospheres, tut at pressures of 20 atmospheres
and upw aids they C£ me cut sirorgly. They were evidently
the absorption bands of NO2, derived from the residue of
atmospheric air mixed >Kiih tl:e condensed gases. We
tcok a photograph of them, and on compaiing this with a
photograph of the KO3 bands, we found the two to
be identical. Exceft for the bands, and the bright lines
of sodium, the speiflium appealed to be continuous, and
to extpnd frcpi fik?put ^ 02pp to ^ 4150, with the brightest

part about A 5150. It increased in brilliance as the pres*
sure increased, as well as in extent, being visible at 3
atmospheres pressure from about \ 6720 to X ^040. The
greater distindness of the NO2 bands at the higher pres-
sures was due both to the greater brightness of the con-
tinuous spedrum and to the greater quantity of NOa
formed. A large quantity of water accumulated in the
experimental tube, and when this was drawn off by the
stopcock, s, it effervesced with escape of NO, and was
found to be strongly acid. A specimen titrated was found
to contain very nearly 3 per cent of nitric acid. Tne ob-
servations were continued up to a pressure of 40 atmo-
spheres. There was no indication that the continuous
spedrum had any connexion with the line spedrom of
hydrogen. There was no increase of brilliance in the
neighbourhood of the C, F, or G lines of hydrogen. The
charaders of the spedrum were, however, better seen ia
the absence of NOai and will be described in the next

Oxygen Burning in Hydrogen.

In this case the colour of the flame was very different
from that of hydrogen burning in oxygen. Instead of
being yellow, it appeared, to the unaided eye, to have a
lavender hue. In the spedroscope it showed a perfedly
continuous spedrum, brightest in the green, about the
region of the Fraimhofer line 6, and very gradually fading
away on either sic'e. On the red side it could be just
traced up to about A 6150, and on the violet side to about
A 4285, at ordinary pressures. The sodium lines were
absent. With incr.ase of pressure it increased very much
in brightness, and at 8 atmospheres pressure it could be
traced as low as A 6630 and as high as A 3990.

The dispersion used was that of a dired-vision spedro-
scope (such as was described by us, Roy. Soc. Proc., vol.
xli., p. 449), equivalent to three prisms of white flint
glass, but the collimator and telescope veiy short, so as to
obtain plenty of light. With less dispersion, perhaps, the
continuous spedrum might have been traced further.
Photographs, however, showed that it scarcely extended
into the ultra-violet. There was no indication that this
spedrum was due to an expansion of the lines of either
the first or second spedrum of hydrogen. It is true that
the maximum brightness (which could not be determined
with any preat accuracy) was not very far from F, bu^
no indication of any second maximum in the neighbour^
hood of either C or G, or anywhere else, could be deteded."
The pressure was carried up to 12 atmospheres, and at
this pressure the visible spedrum was brilliant, but in the
ultra-violet, photographs showed that the spedrum con-
sisted only of what we have called the " water-spedrum,**
very strong and sharp. The lines of this spedrum
showed no signs of expansion even at a pressure of Z2
atmospheres, and, though much more intense than at
ordinary pressures, remained clearly defined.

Observations were continued with the eye up to 25
atmospheres pressure, but no trace of emission, or absorp-
tion, corresponding to either spedrum of hydrogen, could
be deteded, and it is doubtful if either spedrum can be
produced in such a flame. Since the formation of steam
from its component gases is attended with a diminution
of volume, increased pressure will increase the stability of
the compound, and the flame will contain a larger pro-
proportion of steam, as well as have a higher temperature,
than at ordinary pressures.

The water formed when the flame was a jet of oxygen
burning in hydrogen was found to be alkaline, and to con-
tain ammonia. But the proportion of ammonia was
much less than the proportion of nitiic acid formed when
the jet was hydrogen burning in oxygen ; a specimen
titrated contained 0-0004 P*^'' cent of ammonia.

Effects of Pressure on the Sodium Spectrum.
In order to see what eficd would be produced by in-
creased pressure on the spedrum of ether substances in
the flame, we charged the hydrogen with sodium vapour
by making it pass, before entering the experimental

Digitized by


CttKUlCA^ NEWft,)

March 26, 1891. 1

Potassium Cyanide Assay of Lead Ores.

cylinder, through a small iron cylinder, *, in the figure,
containing metallic sodium, heated by a lamp. As the
D lines of sodium are very easily expanded and self-
reversed in a flame at ordinary pressure, some care was
needed to discriminate the effects which were really to be
ascribed to pressure. The gas was easily charged with
sodium vapour, and when burning in oxygen, not only
the D lines, but the citron and green pairs, and sometimes
the blue pair (X 467), and the orange pair (A 616), were
well seen ; but we could not find that they were expanded
by increase of pressure. A sudden change of pressure
generally produced an expansion, but it did not last; the
lines fined down again when the pressure was sieadyi
whether that pressure was high or low. These experi-
ments were continued up to a pressure of 40 atmospheres
without any definite effed on the width of the lines which
could be ascribed to the pressure.

It may be said that at the higher pressure the evapora-
tion of the sodium would be slower, and so the proportion
of sodium vapour to hydrogen be diminished ; also, when
the lines are diffuse at the edges to begin with, it is
extremely difficult to judge whether there is any expan-
sion. At all events, we may say that there is no expan-
sion produced by pressure at all comparable with that
produced in a flame at ordinary pressure by increasing
the quantity of sodium in the flame. We noticed, how-
ever, that the presence of sodium, which produces a feeble
continuous spedrum in a flame at an ordinary pressure,
teemed to increase the continuous spe^rum of the flame
under pressure, especially in the orange and green.
(To be continned).


AiSAjer at the Phcenix Test Works, Battertea.

(Concluded from p. 30).

OwiNO to press of other work, I have been unable to
finish the experiments upon the cyanide assay of lead
quite as soon as I expeded. The experiments have been
continued upon the remaining ores of commercial value,
viz., anglesite, pyromorphite, and mimetisite, and with
the two former minerals results have been obtained ex-
ceeding all expedations. As was to be expeded, with the
last named mmeral, the result was vitiated owing to the
redudion of arsenic, giving too high a result. The lead-
antimony minerals such as jamesonite, &c., were not
treated, since the lead and antimony would be reduced
together. Of course, the lead and antimony might be
estimated in the button by the usual wet process, but
such a mode of procedure offers no advantage over the
usual chemical analysis.

Anglesite, — The anglesite employed for the following
experiments contained a small quantity of cerussite, and
here and there a minute crystal of galena was seen
glistening. It contained 59'62 per cent of lead as shown
by chemical analysis. The mineral was crushed and
passed through a 60 sieve, samples being taken when
required by the method detailed in my first paper
(Chbm . Nbws, vol. Ixiii., p. 30). 100 grains of the ore
were mixed with 600 grains of ** stick '"KCy, and placed
in a small Battersea crucible with the precautions
observed in the galena assay. This mixture was heated
for twenty minutes, using a low red heat rising gradually
to a bright red. A good finishing temperature was
required, otherwise the slag would have been too viscid.
The button of lead was obtained in the usual way by
breaking the crucible.
Exp. Vo. Ondni.


Button of lead obtained 59-3 1 Mean 59-35 p.c.
f» I If 59 4)


The slag was of a light pink with a white crystalline
core. This result is as good as it is possible to get with
any furnace assay, since it is only 0*27 per cent below
that given by chemical analysis.

Pyromorphite containing 74'oz per cent of lead was the
mineral next worked upon. A sample was made as above,
and xoo grains of the ore treated, in duplicate, as in
experiments 21 and 23, finishing, however, with a mode-
rate red heat. The crucible showed only a spongy mass
at the bottom instead of the expeded button. Further
experiments were made using a higher finishing tempera-
ture as well as modifying the starting temperature. In
experiments 23 and 24 the temperature at the commence-
ment was a low red.

Exp. No. Oraini.

23. Button of lead obiained 72-85 1 ^ean 7277 P-c

24. „ 1, M 7270 j ' ''*^

In experiments 25 and 26 a moderate red heat was
used at the start, and finishing at a bright red as in
experiments 23 and 24.

Exp. No. Grains.

25. Button of lead obtained 73-0) ^^^^^ „

20. „ „ ,t 73'i )

The mean of five experiments gave a result of 73*1 per
cent. From these experiments it is obvious that a higher
temperature is required to complete the readion with
pyromorphite than with any of the preceding minerals.
The result is rather astonishing considering the high
temperature used, and certainly points to the loss in the
galena assay being a chemical one, and not so much to
loss by volatilisation. This assay yields 0*91 per cent
lower than the chemical process. For comparison, two
assays were made, using the mixture of

Ore . • • • . • • • 200 grs*
Carbonate of soda . . 250 „

Argol fio „

Covered with zoo grs. borax.

This mixture was fused for twenty^five minutes, using a
bright red heat, and the fused charged poured into a warm
mould. The following results were obtained :—

Exp. No. Grains.

27. Button of lead obtained 142-2 ) ^

The slag thus obtained is viscid, and there is great
danger of it encloiing small globules of metal unless the
crucible is handled rapidly. This result is 2*06 per cent
below the cyanide method.

A sample of impure mimetisite was next experimented
upon, with, however, very varying results. The button of
lead obtained was brittle and of a very dark colour, due to
the presence of arsenic. Under these circumstances, after
making four experiments with similar results, no more
work was done upon this mineral.

The fad that one flux gives good results with all the
commercial ores of lead is of great value. As galena,
cerussite, and anglesite frequently occur together, it is of
great advantage to be able to assay all these minerals
with the same flux. Of course this method does not give
such accurate results as the sulphate chemical method,
but as the furnace method is still largely used (and in all
probability will continue to be used) in most lead mines,
any method which will give higher results than those now
in use is well worth considering.

It is true, as Mr. Cooper states (Chem. Nbws, vol.
Ixiii., p. 73), that with lead ores containing other reducible
metals the lead button obtained contains a notable quan-
tity of those metals. In those cases where the impurity
is large, the furnace method is inapplicable, and the
chemical method must perforce be employed. The galena

Digitized by



Determination of Ferric Oxide and Alumina.

f CHimeAL Nswt,
I March 26, 1891.

used in experiments 15 and 16* contained nearly 2 pe^ cent
of copper pyrites, and most of the copper was in all pro-
bability carried off in the slag« as the lead button obtained
was quite clean and soft. Percy made a number of
experiments some years ago in the Metallurgical Labora-
tory of the School of Mines upon cupriferous galena. A
galena containing 10 per cent of copper pyrites was
reduced, using a carbonate of soda and argol (lux, and the
amount of copper deei mined in the button of metal
obtained. It was found that tlie button of lead contained
98-9 per cent of lead and I'l per cent of copper ; with a
galena containing 25 per cent of pyrites ; the button of
metal contained 96*5 per cent of lead and 3*5 per cent of
copper, the rest of the copper being contained in the sla^.
From these figures it is evident that with a galena having
less than 5 per cent of copper pyrites the error introduced
is negligible, since the button would contain less than 0*4
per cent of copper. I may point out, however, that the
quartz worked on by Mr. Cooper could hardly be called a
Uad ore, inasmuch as it only contained 3 per cent of
galena and about 10*5 per cent of copper pyrites ; surejy
this is more like a copper ore. The results obtained from
a mineral of this charaAer do not warrant the condemna-
tion of the cyanide assay of lead ores.





By Dr. von GRUBER.

The determination of ferric oxide and alumina in phos-
phates is of the greatest importance for the manufa^ure
of superphosphates, as it is the only method of foreseeing
the results of the process and the value of the finished
article. Even the phosphoric acid which has been
rendered soluble goes back and becomes insoluble in pro*
portion to the proportion of ferric oxide and ahraiina in
the manure.

This fad is admitted, although the process of reversion
18 still unexplained. The author suggests that the strong
attradion of water by the gypsum as it passes from the
precipitated to the crystalline state deprives the existing
phosphates of sesqu'oxides of their constitutional water,
and thus converts them into an insoluble modification ;
at least certain experiments made point in this diredion.
The objed of the present communication is to discuss the
methods of determining the sesquioxides which have only
just been perfeded, which, considering the pradical im-
portance of the question, is truly surprising.

There has been hitherto employed a so-called " con-
ventional method,'* the result of which showed the most
striking discrepancies. It was customary to dissolve in
acetic acid the precipitate obtained by adding ammonia
to the hydrochloiic (or nitro-hydrochloric) solution of the
phosphate, containing the phosphates of the alkaline
earths, of ferric oxide, and alumina. Calcium and mag-
nesium phosphates were supposed to be completely dis-
solved, whilst ferric and aluminium phosphate remained
ondistolved. But as a considerable quantity of calcium
phosphate always remained in the precipitate the precipi-
tate was re-dissolved in hydrochloric acid, re-precipitated
with ammonia, and the precipitate dissolved a second
time with acetic acid. The precipitate was thus rendered
purer, and was, indeed, generally free from lime; but
whilst the first result was too high in consequence of the
presence of calcium phosphate, the second precipitate
gave too low a result, as the cold acetic acid has a sol-
vent action, especially on the aluminium phosphate.

Still this conventional method was retained, though it
could not claim the rank of a scientific process, and must

* CucuicAL News, vol. Ixiii , p. 30,

even be regarded as insufficient for the demands of prac-
tice. There was no other method of determination.

From this point of view we can understand the discre-
pancies announced from the laboratories of leading
analysts : —

A. B. 0. D. E.

Per cent.


















The letters A — E stand for the various materials
examined, and the numbers I.— IV. stand in place of the
names of the analysts. It is seen that there can scarcely
be any idea of an agreement. It seems as if one analyst
had precipitated only once and others twice.

WniUt making known these conflicting results the
author endeavoured about six years ago to draw attention
to the unsatisfadory nature of this method of determina-
tion, and expressed the wish that as long as no better
process was known this ** conventional method'* might be
developed by fixing exadly the quantity of acid to be
used' for the crude phosphate, the quantity of ammonia
to be used for neutralising, and the quantity of acetic
acid, so that the results obtained might at least be con-
cordant if not stridly corred. He also made the request
that the experience gained in seeking for a better method
might be mutually communicated.

Not one reply was obtained to this circular letter sent
out by the author and his collaborator, Herr Otto
Schonherr, though he does not venture to infer that the
subjed was ignored. There was never found an abso-
lutely corred result even when working with pure mate-
rials. Among other expedients he tried the precipitation
of the phosphoric acid of the precipitate with molydenum
as proposed by Stutzer. The worst point was that no
explanations for the deviations — often considerable —
could be given.

The method of Eugen Olaser (published in November,
1889) solved the question in a manner sufficiently accu-
rate for technical purposes, and avoided the errors of the
conventional method. This so called ** alcohol method*'
is very simple: —

5 grms. phosphate are dissolved in the usual manner 10
25 c.c. nitric acid of sp. gr. 1*2, and about 12-5 c.c. hydro-
chloric acid of the same sp. gr., and made up to 500 c.c, ;
100 c.c. of the filtrate (a one-fifth of the sample) are
placed in a (-litre flask, with the addition of 25 c.c. sul-
phuric acid of 1*84 sp. gr. The flask is let stand for five
minutes, shakini; occasionally; about 100 c.c. alcohol
(95 per cent) are added ; the flask is cooled, filled up to
the mark with alcohol, and well shaken up. After stand-
ing for half an hour the liquid is filtered. 100 c.c. of the
filtrate (^ 0*4 grm. of the sample) are evaporated down
in a platinum capsule until the alcohol is expelled. The
solution is then mixed with 50 c.c. water in a beaker and
heated to a boil. Ammonia is added to the solution until
the readion is alkaline, but not whilst boiling, to avoid a
violent effervescence.

The excess of ammonia is boiled away ; it is let cool,
filtered, the precipitate washed with hot water, ignited,
and weighed as Tetric and aluminium phosphate. On
trying this method the author obtained results which
agree with those of calculation. The precipitate
obtained was pure, consisting entirely of ferric and alu-
minium phosphates.

In January, 1890, Stutzer published another method for
the determination of ferric oxide and alumina in phos-
phates. The first (calciferous) precipitate which remained
on precipitation with ammonia and treatment with acetic
acid, is colleded and treated with molybdic solution in
the usual manner to remove the phosphoric acid. In the
filtrate the ferric oxide and alumina are precipitated with
ammonia in the usual manner.

This method seems, theoretically speaking, to ensure

Digitized by



26,1891. f

Theory of Solution.


absolutely corredt results. In the menntime the author
has tried it years ago, and not obtained satisfadory
results. On repeated trial it was the same. Sometimes
he obtained several satisfadory analyses in fuccession,
but then all at once the method failed without any per-
ceptible reason. Or. Stutzer himself has had similar
experiences ; whence the method cannot be pronounced
satisfadory until the necessary precautions have been
discovered. However, a combination of Glaser^s method
with that of Stutzer meets the objedlion that half the
quantity of ferric and aluminium phosphate found is cal-
culated as sesquioxides. This objection is justified in
view of the different molecular weights of ferric oxide and
alumina if scientific accuracy is required, but for techni-
cal purposes no scruples need be entertained. If it is
necessary to remove the error the precipitate obtained,
according to Glaser's method, which is free from lime
and magnesia, is re-dissolved and treated with molybdic
solution, to that the pure sesquioxides can be precipi-
tated from the filtrate with ammonia and separated in
the usual manner.

The ** alcohol method" was formally accepted at the
Congress of German Experimentr.l Stations held on Sep-
tember i8ih. — Zeitschrifl Anal, Chtmie.^ xxx., 'p. 9.


The following discussion ensued on the reading of the
paper on •* The Present Position of the Hydrate Theory
of Solution," by Mr. S. U. Pickering. F.R.S.. which
appeared in the Chemical News, vol. Ixii., pp. 185 and

Df. Gladstone made a communication on *' The Mo-
lecular Refradion of Substances in Solution," in which he
reconsidered the five reasons given in 1863 and i86g for
believing that ** the specific refradive energy of a solution
is the mean of the specific refradive energies of the sol-

Online LibraryArnold BennettChemical news and journal of industrial science → online text (page 38 of 88)