Gilbert Hayes Whiteford.

A study of the decomposition of silicates by barium salts online

. (page 1 of 2)
Online LibraryGilbert Hayes WhitefordA study of the decomposition of silicates by barium salts → online text (page 1 of 2)
Font size
QR-code for this ebook

3 1151 02564 2475





''''''' ^^m










Gilbert Eayee 'iVhiteford,

Baltimore, ILd.,

June, 1917.

Table of Contents. Page

Acicnov;leclgjr:ent 3

Introduction 4

Pxirif ication of Reagents 6

Eutectic .fixture of Barium Carbonate and Chloride.... 7

Decomposition 9

Determination of Silica 11

Residue from Silica 12

Leterrjination of Iron and Alumina 14

Determination of Calcium 17

;aethod for Separation of Calcium from Barium 21

Determination of Alkalis 23

Determination of IJagnesium 29

Summation of Analyses 30

Conclusions 28

Biography 40

The investigation described in the following pages was

suggested by Prof. J. C. 7. Frazer and carried out undor his
direction. T'hatever of merit therfinay be in the work is to
be ascribed to his constant attention and suggestion. To him
the author is deeply indebted. -rhe author also wishes to
thank Prof, ^dv.ard Z-.. '..lathews and Dr, J. T. Singcwald for ma-
terial for analysis and helpful suggestions; also Professors
:iOT8e, 3eid, Lovelace and Gilpin and Dr, Davis for kindly
interest and helpful instruction throughout his course.

A Study of the DecornpoBition of Silicates
by Barium Salts.

Among the numerous reagents for the decomnosit ion

of silicate rocks both barium carbonate and uariu/ri chloride have

been used separately. Accordine- to J. Lawrence Smith ''The ci;lo-

rid of barium has been lately proposed; but its decomposing prop-
erties are very feeble, as the chlorine in combination with the
baryum is not liberated at a white heat, and few silicates are
able to produce the decomposition. It miay succeed with some
of the feldspars, bat decomposes very imperfectly even the micas.
So it is r.-^ther a risk to employ it with an unknowx substance.

'The carbonate of baryta is the compound of bary-
ta most generally employed for silica decompositions; still this
is attended with much difficulty, owing to the in'^usibility of
this salt, and the impossibility of driving off the carbonic
acid by heat alone; and even if this latter were possible, the
objection pertaining to caustic baryta v/ould then arise,

'The following extract from Rose's Analvtical
Chemistry, ( transle tioij by I'iormandy, i; a note by the trans-
lator,) presents fairly the difficulties attending the nethod
of decomposing the silicates:

1. J. Lawrence Smith, Am. Tour. Sci., ?. ' Series. 15, ?25, (1853),

'The heat applied is so Intense, th^ut some precautions

roust be taken. The platinum cracible containir.R' th'i mixture

should be exposed first to the heat of an ar5:and-lamp, ar.d when

the iruiss begins to agglutinate, ' 'c'i crucible should bo closed,

and its cover tied down with platinum wire, tiien placed in a

liessian crucible, closed up also; the whole x^ pi- ced upon an

inverted crucible, and i-ubmitted to th'.- action of the bl' st of

a wind furnace, b^ffinninp: first ?-radually vith a red heat, pilinp-

on more coice, so as to fill up the furrrrce.and increasing the

heat to the highest possible pitch, until the Hessian crucible

begins to soften. It is absolutely necessary to the success

of the operation, ';h;:t the Heseian crucible thould be closed as

well as possible, vhich is best done b:; luting tho cover /.ith

fire-clay; the r^eseien crucible and its cover having fused to-

^■elher, cannot be separated, except by breaking, v-c. , ^c. , '

It will be seen in reading- this extract, that the

heat req^uired is not or inarilj' et the coraiLf-nd of Kost chemists,

in fact no other variety of furnace than a Tefstroem can be

depended on for s coranlete decomposition."

In a recent investigation ilempol decomposed the

most refractory silicates 07/ heating in a specially constructed

oven and blast furnj^ce et 1360°, Hempel, however, records no

analyses of the silicates he decomposed in this way.

The purpose of this investigation to effect a

complete decomposition at a lower temperature by using a rr.ixture

of barium carbonate and chloride end to compare the analyt-

1. "empel Zeit. f. Anal. Chem. 5.", ^6, (1913).

ical advar.ta£;es of tVie method witli the methods in genoral use
at preser.t.

Darlum Chloride . The purest comrccrcial barium chlo-
ride was dissolved in ditftilled water, filtered and reprecipitated
by a current of hydrochloric acid g*as which made to pass
through two wash bottles. It was again dissolved and reprecip-
itated in the same way. After washing with a very small quantity
of water, it was filtered on a Titte filter and dried to constant
weight. To obtain the anhydrous chloride it was heated in the
muffle furnace to constant weight.

Barluir. Carbonate . The hydrous chloride prepared as
above was dissolved in water, the solution made amjnonical and the
carbonate -irecipitated by a current of washed carbon dioxide ?-as.
The precipitate was then washed by decaitation until practically
free from chlorides, collected on a ^itte filter and 7;ashed until
the washings rave no -^precipitate, with silver nitrate. It whs
then dried in t -.e electric oven at 150° until a constant weight
was obtained.

Sodiuri! Carbonate. "Chemiaally pure" crystals were
dissolved in distilled water, the solution boiled and filtered.
It was then put in a thick-walled ilask and precipitated under
pressure by a current of washed carbon dioxide ^^-as. Tr.e precio-
i;ated bicarbonate was then dried as mucji as possible on a suction
filter and finally heated in an air bath to constant weight.

CALCIU:,! Carbonate , the purest commercial chloride
was dissolved in distilled water, a little ammonium dydroxide

added, the solution boiled and filtered on cooling. It was
than precipitated by anrnonium carbonate ..hich had been made by
passing v^ashad carbon lioxide gas into a/nonia. The precipita-
ted calcium carbonate tnen washed a number of tiroes by decani
tation, collected on a porcelain filter, washed until free from
chlorides and dried to constant weight.

Ainmonlum chloride . This was made by dissolving washed
hydrochloric acid gas in distilled water and neutralizing with
ammonia gas ii.ade by boiling concentrated ammonium h:7dr6xide.
After filtering it was dried to constant weight.

^-ydrofluoric acid . This was tested by evaporation
in a platinum crucible. lo non-volatile residue was found.

Sulphuric acid . The method of 7resenius was used.
3 graTEB of ammoniiiin sulphate were added to a liter of sulphuric
acid and the solution heated in an evaporating dish until copious
fumes were evolved. After cooling 5 grams of manganese dioxide
were added and the acid again heated to boiling. On cooling it
was placed in a retort ar.d distilled. The first portion of the
distillate was discarded and the remainder redistilled until no
non-volatile residue remained after the evaporation of 3 to 5 cc.


The meltine- ooint of anhydrous barium chloride is given
by different investigators as follows: -

1. Landolt-"ornstein PhysiK. -chem. Tarbellen, 270

922° Llyer, Ridille and Lamb, (1S93).

347 Le Chatelier, (1887).

915.6 Iwc Crae, (1895), by Pt-i-tlr.

344 Ramsay and EumorfoooulOB, (1896).

960 Rufl' and Plato, (1903).

The melting point of barium carbonate according to
Le Chatelier, (1887), is 795°.

Ir order to secure decomposition at the lowest temperature
possible, the composition of the eutectic mixture was determined.
The melting points obtained, however, are only approximate since
the two salts when rrixed appeared to have no definite melting
point. A pi at intra- iridium junction protected by a thin platinum
sleeve was placed in the centre of the mixture and the crucible
heated in tr.e Korea electric furnace. An attempt to plot the
melting and cooling curves was made, but this had to be ab- ndoned
since no definite breaks in t ne curve either on melting or on cool-
ing Y/ere noted. The covers were then removed from tr.e furnace
and the temperatures at wnich the mixture egan to melt and to
solidify were noted. In this way the eutectic point was deter-
mined with a reasonable deg-ree of accuracy, though, as stated
the temperatures giver, below are only approximate.

BaCOc moles SaClg moles ...elting Point

1.0 .67 740°

1.0 1.0 700

1.0 1.1 720

1.0 1.5 725

1. Le Chatelier, Bull, Soc, Chem. 47, 301, (1387)

The eutectic mixture not used ^or tho (lecomposition
because socie of the carbonate ions are used up during- the jecom-
position, thus charginc- the composition of the mixture. It is
therefore advantatieouB to start with a consioerable oiceee of the
carbonate ;.iA allow the eutectic point to be approsched durir.g the
decomposition. I'he following tes's were Kiade with one gram sam-
plee of ":ebsterite in order to determine the best ratio tocix the
carbonate and chloride, he amount of the mixture necessary to
effect a complete decomposition ?ind the length of time neces-
sary to heat in tne flame of the ordinary Bunsen burner.

Amt. of r;:ixture P.atio in moles Time heated State of


=2-ritty residue











































. "
n ti

The rock powder was ground in an ag&te mortar ur.ti]
all particles passed through a sieve of the finest silk bolting
cloth. In all tests the finely ground rock oowder was inti.irite-
ly mixed with the decomposition mixture by thorou^^-L ^-rinding
in the mortar. A sm^^ll amoimt of the mixture of barium salts
being used as a pad in the bottom of the crucible, then the
rock powder v.ith the mixture and finally on top of this a

little of the mixture whio'n waa reserved to rinse off the
watch glass and tiie irortar. i'he heat of an ordi ary Bunsen
burner, producing- a dull redness as far un the crucible as tho
mixture extends is sufficient to sinter the mass together and
produce complete decomposition. A hie-her heat, as will be
shown later, is advantag-eous since it decreases the adsorption
power of the silica. The mass neel not fuse but by simply
sintering together, produces a cake which is very easily removed
by the gentlest tapping on the bottom of the crucible. This
has a distinct advantage over tne usual method of fusing v.ith
sodium carbonate, since all loss by spattering out of ths cru-
cible during the fusion is avoided. If too high a heat is
used in the decomposition, a fusion takes place which in some
cases requires considerable time to remove from the crucible.

Anhydrous barium chloride must be used, otherwise,
the barium chloride v/ill hydrolize at the high temperature and
the crucible may be badly attacked by the hydrochloric acid, in
conta'ct with air. In one case after the crucible was cleaned
a loss of over two-tenths 'f ">- gram in weight was noted; while,
due to the dissolved platinum^ the fused mass was badly discol-
roed ai.d unfit for analysis. V/ith t;ie arhydrous salt the plat-
inum crucibles were never attacked, and even v.ith two to three
hours heating to a bright redness no loss in weight and no dis-
coloration could be detected. In two cases the fusion with
sodium carbonate slightly attacked the crucible, while in the
decomposition with the barium salts the crucible was not attacked


One gram sj^mpleB of the rock powder were used ir: oach
case. This was intimately mixed with five graiT^s of the barium
salts in tiie ratio of four moles of bariurr. carbonate to one mole
of anhydrous barium chloride and deco.r.posed in the heat of an
ordinary junsen burner. The sintered mass v/as dissolved in
dilute hydrochloric acid and evaporated to dryness twice on
the water bath. Transparent quartz dishes were used for this
evaporation, 10° hydrochloric acid was used to moisten the sil-
ica. This is believed to be advanta^-eous since it in nib case
renders the silica gelatinous as is frequently the case when
too great an excess of the concentrated acid is used. After

filtering, washing, igniting to constant weight and evaporating

with hydrofluoric acid in the usual way, Lenher's method of treat-

ingthe residue was used. That is, instead of putting the pre-
cipitate of iron, alumina &c. into the unwashed crucible the
residue from the evaporation of the hydrofluoric acid and assum-
ing that all of this resi ue consisted of oxides of the aluminum
group, the residue was fused with a small amount of sodium car-
bonate, th . fusion taKen up with a little dilute hyarochloric
acid andadded to the filtrate from the silica. This is more
accurate since all the matter adsorbed on the silica is not
aluminum and ferric oxides, and in case a large amount of iron
be present it may not all be in ti.e ferric state. A compar-
ison of the results obtained by this method and by the ordinary

1. Lenher and Truog. C. Am. Chem. Foe. 38, 1050 (1916)


method of decomposing with sodiuin carbonate is p-ivon in the
following table.

fample % SiOg

KaeCOs 3a salts.

fl rourraaline 39.06 39.16

f2 Websterite 62.18 52.18

fZ ileta-rhyolite 76.61 76.51

r4 Gneissic Inclusion 78.05 77.99

*5 Light band in gneiss 69.70 69.67

*6 Queleuzite 9.39 9.38

f7 Garnet rook 37.01 37.07

RESIDUES. Silica always adsorbs iron, aluminum and
other salts and no amount of wasliing will completely reiLOve
these from its surface. It was noticed that while decomposi-
tion was complete at a dull redness, the aniount of adsorbed
matter could be greatly reduced in some cases by heating to a
higher temperature. To study the relation between the temper-
ature and the amount of adsorbed matter another series was run
at the full heat of the l-lorse electric furnace, -about 900°.
The following table shows the results.


Sample KagCOg decomp. Dull red. Ba salts decomp.

Elect, furnaoe

#1 .0070 .0119 .0101

#2 .0045 .0Z42 .0072

#3 .0068 .0111 .0083

#4 .0111 .0203 .0135

#5 .0065 .0243 .0128

#6 .0186 .1603 .01£3

#7 .0023 .0215 .0185

Ko real advantage seems t4 be gained by the use of
high temperatures in the decomposition other than to decrease
this residue, except in the case of #6,- a garnet rock very rich
in manganese. In the case of this specimen accurate checks of
SiOg could not be obtained with so large a residue, while when
decomposed at a high temperature it was quite easy to check the
results within the limits of experimental error. With sodi\un
carbonate this fusion strongly attached the crucible, forming
a dark coating which only could be removed by long and repeated
heatings with concentrated hydrochloric acid. The fusion with
barium salts did not attack the crucible at all, not even dis-
coloriTjg it. In all cases the mass decomposed at higher temper-
atures formed a loosely sintered mass which was easily removed
from the crucible, could easily be broken up with a stirring
rod and readily went into solution,- in some case without any
evolution of carbon dioxide -Ahen treated with hydrochloric acid.
In order to determine if the amount of bariu salts


used in the clecon.position iiad any relation to the amount of the
residue, one decomposition was made using eiplit grams o' barium
salts instead of five, and heating to the full heat of the elec-
tric furnace. A difference of only .0001 p-ram v-as found in the
amount of the residue.


In precipitating iron and aluminum as hydroxides the

use of i!.ethyl red to show the end point as sug ested by rslum was

tried. "^en, however, the amount of ferric oxide is large, un-
less the amount of the solution is correspondingly large consid-
erable time is lost in waiting for the precipitate to settle in
order to be able to see the color of the solution. In samples
#6 and #7, silicates rich in manganese, the color of tho indica-
tor was destroyed as soon as it touched the solution, thus render-
ing methyl red useless in these cases. In these samples eveii a
cubic centimeter of the methyl red did not impart the slightest
color to the solution, in all other cases one or two drops gave
the solution a distinct colot.

I-norder to determine the end point a small piece of
red litmus paper was ir.oistened and placed on the convex side of
the watch glass. When the precipitation was apparently com-
pleted by slowly dropning amDmonium hydroxide from a burette into
the not solution, the supernatant vapor was blown off at two or
tliree times and the watch glass placed over the beaker. In this
wa" the iron and aluminum hydroxides v.ere completely precipitated,

1. Blum. J. Am. Ghem. Soc. 38, 1282 (1916), Tciectific laper 7 296
U. £. Bureau of Standards.


Tests made on the filtrate showed that none of the aluminum
hydroxide was redissolved. As soon as sufficient armonium
hydroxide was added to turn the color of the litmus paper,
the contents of the beaker were boiled for one minute, the
precipitate allowed to settle, then filtered and washed,- first
with ammonium chloride solution, (1%) , then with water. The fil-
ter paper with the precipitate was then put back in tiie orie-inal
beaker, dissolved in nitric acid and reprecipitated in the same

Instead of ie-niting the precipitate of ferric and alum-
inum hydroxides in the unwashed crucible containing the residue

left from the evaT)oration of the silica by hydrofluoric acid,

• ' 1
(the usual method), Lenher's method was used, as described under

silica. If the residue consisted entirely of the oxides of this
group this would be unneaessary. But in the cases of some sil-
icates the amount of other substances in this residue is consid-
erahle. Ihis was found to be especially true in the cases of
samples t6 and .'''7, iLan^arr.ese silicates. In the case of these
a large part of the residue from tr.e silica consisted of rr.anga-
nese salts. Lenher's modification of the usual metiiod is un-
douotedly more accurate than the old way.

In the case of .-iianganese silicates no trouble at all
was experienced in separating the iron alumina *c. from the man-
ganese. 5,ualitative tests were rijade and it was found that when
the end of the precipitation of tr.e iron and alumina by the above
m.ethod y;as observed an excess of a number of 'Irops of the amx^onium

1. Lenher and Truog. J. Am. Cherc. ""00. 38, 1C5D (1916)


hydroxide could be added witiiout i.rinpiint- aowii any ii\HDre.neee
hydroxide. Soao of the reprecipitated iron and aliiiLina was
dissolved was tested and found to be free from .-nanganesea
Some manganese could be found ir the first precipitation, but
this is, no doubt, due to the fact that the precipitate cannot
be entirely freed froia soluble salts by washing alone. Some
manganese remains in the jrecipitvte j-st af soinc .alci-i;ja h.:.1
ijiagnesiuEi remain, and the manganese is coi.'.nlctely reiLOved in the
reprecipitation just as the otrior soluble salts are. This frac-
tional precipitation has the advantage of saving consioerable time

in the analysis.


TitaniTiB-. oxide was determined by T^eller's coloronetric

method. The other elements which come down with tnis group were
determined by the usual raeti.ods.

The difference in the results obtained from the two

decompositions was very slight as is shown by the following table.

Total Fe

as FegOg





3a "alts













































. 0029









In the above table the ferric oxide represents the
total iron in tin sample calculated as ferric oxide. The ^ reater
1. 7/eller. 5er. ?595, (1882).


part of this iron in all the samples existed in the ferrous state,
but es this had to bo determined hy decomposition with hydroflu-
oric acid this need only be shoiwn in tne sunn ations of the sot)-
arate samples.

ri\imerous methods have been proposed for the separation
of bariu:; and calcium. iio^ever i.o one seems to have viorked
with so sanll an amount of calcium in the presence of so large
an aniount of barium. In order to get a method v/hich is at the

sarre time accurate and raT>id a number of experiments v/ere tried.
Hose states that both the barium and cslcium may be

precipitated from the solution as sulphates, the mixed sulphates
then treated with sodium, potassium or ammonium carbonate and
the calcium sulphate will be converted quar.titatively into cal-
cium carbonate while the barium sulphate will be unacted i^pon.
Although Rose states that the reaction will be complete at the
end of about six hours, it was found to give low results, even
though allowed to stand 30 days with stirring at frequent inter-
vals. Equilibrium betv/een the sulphate and carbonate is reached
so very slowly as is shown by the fallowing table that this n:eth-
od is wort 1 J. ess.

CaO found ..lixture stood.
.0807 over night
.0877 5 days

.0924 50 days
The method of simultaneous precipitation was
next tried. An an.ount of sulphuric acid about 10^! in excess of

i. Tioae, i-.nalyse Quantitative 37-40.



CaO taken

4.0 grams






that which is neceseary to precipitate all of the barium was
neutralized with sodiu:: carbonate and then about 50^ excess of
sodiuTi carbonate added to the dilute solution of the sodium
sulphate and the solution boiled. The precipitant while still
hot was added droi- by drop to the boiling solution of the barium
and calcium chlorides, which was stirred vigorously all the while,
After cooling and allowing to stand long enough for tho barium
sulphate to become granular, the mixture of barium sulphate
and calciun: carbonate was filtered and v.'ashed. The filter paper
containir.g the orecipitate was then spread out ap-ainst the back
of the beaker and the precipitate completely washed 'back ir.t
the beaker by a jet of water from the wash bottle. This was
then treated from one to two hours with dilute hydrochloric acid,
refiltered and thoroughly washed. The calcium, now all in the
filtrate, was precipitated as oxalates In all cases except one
low results were obtained.
BaG03 used CaCOg used equal to CaO

4.0 grams























CaO found














. 0492




. 1430






^he irethod of Sidorsky, auustituting anr onimn oxalate
for sodium carbonate v;as then tried, the nrocese being carried
out in tlie san.e nar.ner. RgBultB were as follows:

BaCOg used CaCOg used equal to CaO

5 frams


CaO found







It is noticeable thst the ftirst part of tlie
series Ls regular, though low. The lest one of the series ran
high. This may be due to the fact that tha amount of hydro-
chloric acid used to dissolve the calcium xalate was roughly
estimated, and too allrge an excess being used some of t'ne barium
sulphate was dissolved as it is a well known fact that some
barium sulphate will dissolve in strong hydrochloric acid.
It is more likely however tho-t the high results are due to the
fact that some of the barium sulphate ran through the filter,
since in almost all the barium sulphate after the treatment with
hydrochloric acid was exceedingly hard to filter.

A series v;as now tried by precipitating tne bariu;;
and calcium as oy.aletes and then converting the barium oxalate
to barium sulphate by a solution of ennionium sulphate contain-
ing an amount of su|}phate ions Just in ejcese of the theoretical


Online LibraryGilbert Hayes WhitefordA study of the decomposition of silicates by barium salts → online text (page 1 of 2)