382 D E T E R M I N A T I O N OF P H O S P H O R I C A C I D .

revolutions of the earth, this must be decided by geology : but

the remains exist.
Kow in nhich manner. do the organic substances of these
animals become decomposed ?
T h e animal substance consists essentially of iiitrogenated
ma.teria1 and fat. T h e former is easily decomposed, tlie latter
is very stable. a fact which has been very well known for a long
time and has been shomn again b y exact investigations. There-
fore we find tlie wax of cadavers i n old graves, therefore the
fat in the bones of mammals thousands of )’ears old, therefore
the fat or1 the bottoin of the ocean recently found.
Whether and how far the fat was decomposed i n this long
period by the water splitting up glycerol and forming the free
acid, for instance the fat iri the bones of mammals, cannot be
answered. Both, fat as well as the fatty acids form petroleum,
when distilled under pressure.
\Ye can imagine, that such remains wrapped in mud arid
transported by the currents i n the ocean, easily accuniulate and
later on, under the pressure of sedimentary layers or strata,
perhaps under the influence of heat too, are transformed into
petroleum. This is only one of the many possibilities by which
the mechanical process of the transniutation of fat into petroleum
may have happened.
Uiider any circumstances I think I have proven that from the
chemical standpoint, the formation of petroleum from animal
reniaiiis has the greatest probability, as we are able now to
transforni every aniiiial fat into petroleum.

THE DETERMINATION OF PHOSPHORIC ACID BY THE

TITRATION OF THE YELLOW PRECIPITATE
WITH STANDARD ALKALI.’
I l s H E S H Y I’EXUBKTOS,Jx.

I N the year 1882, I described a process for determining

phosphoric acid, volumetrically, by aninionium molybdate, on
the principle of Wildenstein’s sulphuric acid determination, or
of Gay Lussac’s silver method. An aqucous solution of am-
1 Read at the stated meeting of tlie Cheriiical Sectiou of the Fraiikliii Institute, held
September 19, 1893. Comniunicated by t h e author.
 DETERMIS.\TIOiS OF PHOSPHORIC ACID. 383
monium molybdate is run into the solution of the phosphate un-
til no futher precipitate is formed.
But it is not of this process that I now have to speak. I t is
referred to here, in order to draw attention to the concluding
paragraph of the paper, as follows :
“ I have obtained very sharp and accurate results by deter-

mining the amount of yellow precipitate (formed as above, after

thorough washing), by nieans of a standard solution of caustic
alkali, using litmus as an indicator; a description of which I
hope to present in a future paper. I mention it here simply to
place the fact on record.” ( / O U Y ? Z . F y a n k , h s f . , 113, 193;
Chem. ATcws, 46,7 . )
At that time I did a considerable amount of work upon the
last-ment:oned process, with very satisfactory results. X study
was made of the conditions most favorable to obtaining a phospho-
molybdate precipitate of constant composition, using solutions
of disodium hydrogen phosphate of known strength, and also a
solution of apatite, the determinations being checked by stand-
ard methods.
Before the process was in shape for publication, however, my
attention was called to nork of an entirely different nature, and
no description of the method was published other than that em-
bodied in the above-quoted paragraph.
Since that time several chemists have described processes
based upon the same principle.
E. Thilo, in the analysis of Thomas slag (Chcm. Z&., 11,
1g3), dissolves the yellow precipitate in standard ammonia, and
titrates back with acid, using litmus as an indicator.
Franz Hurideshagen (Zfschr.a z a l . C ~ L C ~28,V ZI.7, 1 ) uses stand-
ard sodium hydroxide in excess, and titrates back with nitric
acid, using phenolphthalein as an indicator.
C. E. Manby (1. A u a l . Apjl. Chcnz., 6 , 82) determines the
phosphorus in steel, iron, and iron ores, by dissolving the yellow
precipitate in ammonia, acidifying with nitric acid. evaporating
to dryness, and heating gently to expel nitric acid and ani-
inonini nitrate. H e then titrates, using the same solution and
indicator as Hundeshagen employs.
James 0. Handy (1. A n a l . A$$. Chcm., 6 , 204) avoids
384 I>I:TEKMISA’I’IOS 0 1 2 I’HOSPIIOKIC .-IC1 I).

the ei.aporatioii anti heating anti titrates directly a,< Thilo

and Huiitlc-Iiageii (lo. Lisiiig- staiidartl sotla and IjlieiiolIjlitlialeiii.
M . Rotliberg- a n d Il*.;I. Xucliiiivoiti ( -/, .-iv(i/, :Ipp/, Cht31u.
6, 2-1.3) a130 tlescril)? tiie saiiie pro(
. l l t h o u ~ hit is now eleven > ear; sirice i drew attention to this
process. its advantages arc so gi-eat that a n y itiformatioil toucli-
iiig it, i n addition to thxt furiiislieti the foregoing clieriiists.
m a y l)e of interest. I t fxr s u r p s s v s i i i quickiiess the process de-
scribed l)). nie ( o n the (.>a\- 1,iissnc or \\7iltletistt-iii priiiciple )
arid at the same time lacks ~iotliiiigi n accuracy. 111 iriost of the
p p e r s of the nl)ove-riieiitioiietl ciimiisth. tiie pi-oc‘ess is applied
t o the tleteriiii~iatioiioi pliospliortih i i i small quantities, a’; it oc-
curs, for itistaiicc, i n iron. steel. 01 ores. 1)uriiig the p h t !.ear,
I h i - c s 1 1 x 1 occasion to apply it to thy exaitiiiiatioii u i :Liiuiii1)er
of pliospliate rocks. as ell :is o l strong solutions oi 1)liosplioric
acid. coiitaiiiiiig over fift). per cent. Pic),, arid tlie method has
?,eel1 used coritiiiuoiisly during that tiriic. i alii iiitlei,tetl tc l l r .
I<dwiii Harris, n-ho had cliarge o l most of tlie lalm-:itor!. work
of the fartor!.. for iii:iny of thc. figures gi\.cii iieio\\.. I t was
seldoni that tn’o tests of tlie saiiic iii;iteri;il tliffcrtril i i ~ o r cthan
0 . I 1)er cciit, i n 1>:OF, even \vlicn tlic trit31 1’,20:,prneiit ;iiiiouiitetl
to as niucli as forty per cent. to fii’t!. iier criit. of tlie su?)staiict-
anal!~zctl.
T h e following solutions are used :
.-I i i ~ i ~ / o ~ ~,Z~~/3,bdnfc.-Niiict!-
i x i i ~ grams of tlie cr!.stals are dis-
s o l ~ w l( i n a large beaker) in somewhat less thaii one liter of
water. This is alloneti to settle, over night, arid tlie clear liquor
decanted into R liter flask. T h e siiiall quantity of insoluble
iiiolyhtlic acid. always present, is tlissolj-ed i n a little ammonia
and added to the main solution. Should the molybdate he
fouiitl to coiitairi traces of I’,-O!, a few decigrams of niagnesium
sulphate are atlckcl, aninioiiia being added to faint alkalinity.
The whole is then iiiade up to one liter. I t is this nqzlrous solu-
tion that is used, 110 nz’fricacid a*/zafrvev being emflojwi. Each
cc. precipitates three niilligranis of P,O,.
T h e n ? i i i i ~ o ? z i i s ?izitrufc
i~ solution is simply a saturated aqueous
solution of the salt. r)istilled water is poured into the bottle of
crystals in quantity insufficient to dissolve them all. Even i n
 DETERMINATION OF PHOSPHORIC ACID. 385
cold weather, ten cc. of this solution is amply suficient for each
test.
T h e nifric acid, used for acidifying the solution of the phos-
phate, has a specific gravity of 1 . 4 or thereabouts.
T h e sfandard#otassium hydroxide solution is of such strength
that one CC. = one mgm. P,O,. One hundred cc. of it will neu-
tralize 32.65 cc. of normal acid. It can be made from normal
potassium hydroxide (that has been freed from all carbonate
by barium hydroxide), by diluting 326.5 cc. to one liter. But
its strength is best determined empirically by a direct test
upon a phosphate solution of known strength, precipitating
with ammonium molybdate and making the analysis as de-
scribed below, all potassium carbonate having first been removed
by barium hydroxide.
T h e standard acid has the same strength, volume for volume,
as the potassium hydroxide and can be made by diluting 326.5
cc. of normal acid to one liter. I n testing it against the alkali,
phenolphthalein (and methyl orange) should be used.
T h e indicafor can be either litmus, rosolic acid, or phenol-
phthalein. I have used the latter almost exclusively, as it has
been shown, by J. H. Long ( A m . Chem. J . , II, 84) that titra-
tions with this indicator in the presence of ammonium salts
are perfectly reliable if the amount of the ammonium salt pre-
sent is not excessive, if the solution is cold, and if the phenol-
phthalein is used in sufficient quantity. One gram of the
phenolphthalein is, accordingly, dissolved in 100 cc. of sixty
per cent. alcohol, and at least 0 . 5 cc. of this solution is used for
each titration. T h e washing of the ammonium phospho-molyb-
date is done by water. (Isbert and Stutzer, Ztschr. anal. Chem. ,
26, 584.) I t ilay be of interest to quote from their results,
since it is by some chemists thought necessary to wash with a
neutral or acid solution of an ammonium salt. In all tests, as
made by them, twenty-five cc. of the sodium phosphate solution
were precipitated by ammonium molybdate and the phosphoric
acid determined therefrom, as magnesium pyrophosphate in the
usu a1 manner.
When the yellow precipitate was washed with ammonium
nitrate solutim, there were obtained :
(I). ..... ,... .... 0.1943 gram P,O, in 50 cc.
(2). ... .... . ... .. 0.1948 gram P,06 in 50 cc.
386 DETERMIXATION O F PHOSPHORIC ACID.

When washed with water, there was found :

( 3 ) .. . . . . . . . . . . . . 0.194; firtim I’?C):, i n jo cc.
( 4 ) . ............. 0.1942 gram I’,o~ i n 50 kc.
I n order to establish the fact more certainly, the precipitate
was washed with unusually large quantities of water :
c c . of water
usrd i n wnsliiiig. P 2 O 6 found.
( 5 ) . ............. Q”r, 0.1947 i n 50 cc.
( 6 ) . ............. 400 0.1944 in 50 cc.
( 7 ) . ............. 5 0 0 0.1948 i n 50 cc.
(8). ............. I ,m 0.1940 i n 50 ce.
There is, accordingly, no danger of loss in washing the yellow
precipitate with water.
T h e following is the method of performing the analysis :
One gram of the phosphate is dissolved in nitric acid, an ex-
cess of which can be used with impunity, and the solution fil-
tered into a 250 cc. flask and made up to the mark. T h e solu-
tion can eyen be poured into the flask without filtering, since
the presence of a little insoluble matter does not interfere in the
least with the titration. Moreover, since most phosphate rocks
seldom contain over ten per cent, of insoluble matter, and as this
has the sp. gr. of at least 2 , it occupies a volume of about
0.05 cc., an amount so small that it may be neglected. ( F o r
instance, even in the case of a phosphate rock containing forty
per cent. P,O,, the error is only 0.005 per cent. P,O,. )
After the clear solu-tion has been poured off, it is well to treat
the sand, etc., at the bottom of the beaker, with a cc. or so of
hydrochloric acid, in the warmth, to insure complete solution.
I t is not necessary to evaporate to dryness. Isbert and Stutzer
have shown, in their paper, that when the yellow precipitate is
washed with water, the soluble silica is removed, and that
evaporation (to render the silica insoluble) is superfluous.
Their results are corroborated by the test analysis that will be
given below. I n the event of its being desirable to remove
silica by evaporation, for any purpose, the evaporation should
be performed over a water bath, or, if on an iron plate, with
great care, since, otherwise, meta or pyro phosphates are formed
with results that are correspondingly low.
Twenty-five cc. of the solution (equal to 0.1gram) are now
 DETERMINATION OF PHOSPHORIC ACID. 387
taken for analysis. It may be thought, by some, that an analy-
sis made upon so small a quantity of material as one decigram,
and with a standard solution representing only one milligram
per cc., may be liable to errors that would not exist, when us-
ing larger quantities or stronger solutions. But it should be
borne in mind that the accuracy of measurement with a twenty-
five cc. pipette, is precisely the same, whether ten grams of the
original substance are taken or only one gram. Any error in
measuring with a pipette is, of course, entirely independent of
the quantity in solution. I n regard to the amount of material
to be manipulated (filtered, washed, etc.) it will be remembered
that the weight of the yellow precipitate is over twenty-eight
times the weight 6f the P,O, contained in it. Every milligram
of P,O, is accordingly represented by more than twenty-eight
milligrams of precipitate. T h e standard alkali, although repre-
senting only 0.001 gram per cc. is in reality more than three
times as strong as the decinormal solution generally employed.
Of course, in the case of materials containing only ten to fifteen
per cent. P,O,, as in fertilizers, two or three grams can be taken
for analysis, if desired, instead of one gram.
Rcturning, therefore, to the method of the analysis, twenty-
five cc. of the solution are measured out anddelivered into a beaker
holding not more than 100 to 125 cc. A large beaker requires
unnecessary washing to remove the free acid in washing the yel-
low precipitate. T h e solution is neutralized with ammonia-
until a precipitate just begins to form-and five cc. of nitric
acid of sp. gr. 1.4 added. Ten cc. of the ammonium nitrate
solution are added, and the entire bulk of the solution made up
to fifty to seventy-five cc. by adding water.
Heat is now applied and the solution brought to a full boil.
It is then removed from the lamp, no more heat being applied
and treated a t once, with five cc. of the aqueous solution of am-
monium molybdate, which is run into it from a five cc. volume
pipette, the solution being stirred as the precipitate is added,
T h e beaker is now allowed to rest quietly for about one minute,
during which time the precipitate settles almost completely.
T h e five cc. pipette is filled with the molybdate solution, and a
part of its coJtents allowed to drop in, holding the beaker up
388 D E T E K M I K A T I O S O F PIIOSI'IIORIC hCI1).

to the light. If a formation of a yellow clout1 takes place, it i.<

at once perceptible, i n which case the reniaiiider of the pipette
full is run in, the solution stirred and allowed to settle. X third
pipette full is now added as before.. Should it cause 110 furthcI
cloud, onIy about one-half of its conteiits is adtieti, the rmiain-
der being run into the beaker into which the filtrate a r i d \vasli-
ings from the yellow precipitate ;ire to g o . 111 tiit. test aiialyses
given below, it will lie slio\vn that c\-eii ivheii /(/?w>l o-.i j l t.t.(xs.Y
of the molybdate, were purposely ur;etl, over niitl a b o \ ~the cnl-
culated anloutit, the results 11 ere ;ic~urate---:i(~ ilioiy~)tiicacid
conling down with the yellow precipitate.
I t is seldom that more than fifteen cc. iii all (three five cc.
pipettes full) of the iiiolylxlate have to lie added. Since each
cc. precipitates three milligrams P,JOc>, fifteen cc. will precipitate
forty-five milligrams P,O,. This is equivalent to forty-five per
cent. on the 0 . 1 gram taken for analysis, aiid it is not often that
ariy material to be examined contains over this percentage. In
the analysis of materials rich in phosphoric acid, it is one of the
embarrassing features of the usual process, in \r.hicli the r i i t ~ i c
acid s o h f i o z of the niolybdate is used, that, i n the first place,
large quantities of the precipitant have to be used (frequently
several hundred c c . ) , and, in the second place, that the aiialyst is
never certain that enough has I~eeiiadded to throw tloirn all of the
phosphoric acid. This necessitates frequent testiiigs of small
portions of the phosphate solution, u r of the filtrate. There is
another difficulty peculiar to the process as usuall!. carried out
in all methods in which the determination is made directly u1mi
the phospho-molybdate itself, i n that much care niust be ob-
served to keep the solution at a certain teniperature, since other-
wise molybdic acid contaminates the precipitate 311d the anal!.-
sis is rendered worthless. 111the proc lierein described. us-
ing an aqueous solution of the mol!,bdate, the point :it wliicl~sui-
ficient of the precipitant has lieen added is easily seen. S o
molybdic acid separates, because, i n the first place, no great ex-
cess of molybdate is added ; and because, iii the second place,
the solution is filtered immediately, or as soon as it has settled,
which requires only a minute or two. The time required from
the first addition of the molybdate to the beginning of the filtra-
 DETERMINATION O F PHOSPHORIC ACID. 389
tion is never over ten minutes, and is generally less. T h e fil-
trate and washings from the precipitate when treated with ad-
ditional molybdate solution, give, on standing on a hot plate for
an hour or so, a snow-white precipitate of molybdic acid, show-
ing that all of the phosphoric acid has been precipitated. I
have observed t h i s hundreds of times.
A slight correction should be made to the statement made
above in regard to fifteen cc. of the molybdate precipitating
forty-five milligrams of P,O,. This is not strictly true, for the
reason that a small quantity (something over one cc.) of the
molybdate is required to neutralize the solvent action of the
nitric acid. Therefore, in very high grade phosphates a fourth
five cc. pipette full may be required.
T h e yellow precipitate is now filtered through a filter seven
centimeters in diameter, decanting the clear solution only.
This is repeated three or four times, washing down the sides of
the beaker, stirring up the precipitate, and washing the filter
and sides of the funnel above the filter each time. T h e precipi-
tate is then transferred to the filter and washed there. When
the precipitate is large it cannot be churned up by the wash
water and cannot be washed down to the apex of the filter.
This is generally the case when there is over ten or fifteen per
cent. phosphoric acid present in the substance analyzed. In
such an event, I am accustomed to wash the precipitate back
into the beaker, and to fill the funnel with water above the level
of the filter, doing this two or three times, then washing the
precipitate back into the filter. I t is not necessary to transfer
to the filter the precipitate adhering to the sides of the beaker.
It goes without saying that during the washing no ammonia
must be present in the atmosphere of the laboratory. Inasmuch
as the beaker, funnel, filter, and precipitate are small, the wash-
ing does not take long to perform. I t requires, in fact, from
ten to fifteen minutes, even when large precipitates (= thirty to
forty per cent. P20,) are handled. T h e precipitate and filter
are now transferred together to the beaker. By pressure with
the tip of the finger upon the double fold of the filter, it is
easily given a sideways motion and lifted out of the funnel with-
out any danger of breakage, the precipitate being still within it.
390 DETERMINATION O F PHOSPHORIC ACID.

T h e alkali solution is run in until the precipitate has dissolved,

at least twelve drops of the phenolphthalein ( I : 100) are then
added, and the acid run in without delay until the pearly color
disappears and the solution is colorless. T h e presence of the
filter paper does not interfere in the least. T h e reaction of the
indicator is not so sharp as when only acid and alkali are used,
but it is easy to tell with certainty the difference caused by one
drop of either acid or alkali. After deducting the volume of
acid used from that of the alkali, the remainder gives the per-
centage of P,O, directly, each cc. being equal to one per cent.
P,O,. Thus, if there are 28.3 cc. of alkali consumed, the
material contains 28.3 per cent. P,O, when one decigram is
taken for analysis. From the time the twenty-five cc. are
measured out until the result is obtained, from thirty to forty
minutes are required.
I have applied this process to deterniinatioris of phosphoric
acid in phosphates and fertilizers, and have had no experience
i n determining phosphorus in iron, steel, or iron ores. I am in-
clined to believe that in the presence of such large quantities of
iron salts when using the aqueous solution of the molybdate it
may be iiecessary to guard against contamination of the yellow
precipitate by ferric hydrate, perhaps by using larger quantities
of nitric acid than five cc., and perhaps by washing the precipi-
tate at first, with dilute nitric acid. It may also be the case
that the yellow precipitate will form more slowly.
T h e following analytical experiments were made in order to
test the process.
Several sodium phosphate solutions were used a t first, the
strength of which was only approximately known, as the aim
was to see how closely two readings would agree ; equal quanti-
ties of t?ie phosphate being taken for each pair of tests.
A.
cc.
( I ) KHO w i t h rosolic a c i d . . ...................... 15.0
(2) K H O with rosolic a c i d . . ...................... 15.15
B.
cc.
(I) K H O rosolic a c i d . . ........................... 30.5
(2) KHO rosolic a c i d . . ........................... 30.5
......................
( 3 ) KHO p h e n o l p h t h a l e i n . . 30.7
 DETERMINATION OF PHOSPHORIC ACID. 391

All titrations, after this, were made with phenolphthalein.

C.
cc.
(I) KHO ......................................... 31.7
(2) KHO.. ....................................... 31.8
A K H O solution was now made of such strength that one
cc. = o n e mgm. P,O,.
D.
mgm.
( I ) P,06 ......................................... 50.35
( 2 ) P,O, ......................................... 50.55
(3) p,o, ......................................... 50.40
E.
mgm.
( I ) P,06 ......................................... 51.6
( 2 ) P,05 ......................................... 51.55
F.
mgm.
(I) p,o, ......................................... 5IJ35
(2) P,Oj ......................................... 50.93
D, E, and F were solutions used in standardizing.
Two different samples of Florida phosphate rock were ex-
amined.
I.
Per cent.
( I ) P,O, ...................................... = 29.68
( 2 ) P,06 ....................................... - 29.84
11.
(I)p,O, ....................................... = 31.28
( 2 ) p,oj .......................................
= 31-34
T h e following were solutions of phosphoric acid:
1.
Per cent.
( I ) P,O, ....................................... = 46.78
( 2 ) P,O, ....................................... = 46.69
11.
(I) P,Oj ....................................... = 44.41
(2) P,O, ....................................... - 44.63
111.
(I) P,O, ....................................... = 48.95
(2) P,O, ....................................... = 48.80
T h e effect of an excess of ammonium molybdate was tried.
T h e soluble part of an acid phosphate was made up to a definite
volume and fifty cc. taken for analysis. I t was found to con-
tain 8.28 per cent. P,O,. T h e test was repeated, using fifteen
392 D E T E R M I N A T I O N O F PHOSPHORIC .4CII)

cc. of the molybdate more than was used in the first trial.
Result=8.36 per cent. P 2 0 , equal to a difference of 0.08 per
cent.
T h e following tests show that it is not necessary to remove
silicic acid by evaporating to dryness before precipitating with
the ammonium molybdate.
A sample of Florida phoshate rock contained :
Per cent.
( I ) P,Oj silica r e m o v e d . . ......................... 31.21
( 2 ) P,Oj silica not removed ....................... 31.21
Another sample of the same rock was also tried :
Silica removed. Per ceut.
(I) P,Oj ....................................... - 30.5
-
(2) p,03 ....................................... - 30.7
Average E’,Oj.. .................... = 30.6
Silica not removed. Per cent.
( I ) P,O, ......................................... 30.6
(2) p,o, ......................................... 30.7
__-
Average P,Oj ......................... 30.65
When silica is not removed the filtrate from the yellow pre-
cipitate has a yellow tinge.
T h e relation between the P,O, in the precipitate and the
potassium hydrate was established by determining the strength
of a solution of disodium hydrogen phosphate by precipita-
tion as the ammonium magnesium salt, and also by testing it
by this titration process. T h e phosphate of soda solution was
weighed (riot measured), and the magnesia precipitate, after
filtering, was dissolved and reprecipitated with ammonia.
(Gooch, Am. Clzem. I.,
I,405.) T h e results are given in
Table I , the last column giving the amounts of P,O, obtained
on a basis of ten grams of the solution:
TABLEI.
Weight of Weight of Equal Grams
Ka HPO m5?pzo, to P,O, in
sollution’ obtained. P,O,. IO grams
Grams. Grams. Grams. of solution.
I...... 75.824 1.2471 0.7956 0.10494
XI...... IOI .167 1.6637 I .&14 0.10492
111...... 101.622 1.6733 I ,0676 0.10gog

Average.. .......................... 0.10497

 DETERMINhTION O F PHOSPHORIC ACID. 393
Therefore, ten grams of the solution contain 0.10497 gram
p2°b'

Weighed portions of the same solution were now precipitated

with molybdate and the precipitate titrated with alkali. The
results are given in Table 11, the last column of which gives the
number of cc. of alkali equivalent to ten grams of the solution.
T h e indicator was phenolphthalein :
TABLE11.
Weight of Number of =Number of cc.
Na HPO cc. o f KHO sol.
solution.' KHO sol. required for
Grams. required. IO grams solution.

I ................... 74975 79.05 = 105.4

II................... 7.8255 82.40
Average.. .......... ...........
It follows, therefore, that 105.35 cc. alkali = 0.10497gram
P,O, ; therefore,
( I ) io0 cc. alkali = 99.64 milligrams P,05.
T h e HC1 solution was now titrated against the alkali, using
phenolphthalein as the indicator. 99.00 cc. alkali were found
to equal 99.05 cc. acid.
T h e HCI was then standardized against pure sodium car-
bonate.
( W i t h inethyl orange, cold :)
( a ) 1.1291grams Na,CO,=65.45 cc. HCl
(With phenolphthalein, boiling :)
( 6 ) 1.1934grams Na,CO,=69.2 cc. HCI
Therefore, 100 cc. =
( a ) 1725. milligrams Na,CO,
( 6 ) 1725. milligrams Na,CO,
Since 99.00alkali = 99.05 acid, it follows that :
alkali = 1726 milligrams NqCO,
(2) 100 cc.
It has already been shown by ( I ) that 100 cc. alkali = 99.64
milligrams P,O,. Therefore, combining (I) and (2) we obtain
99.64 milligrams P,O, = 1726. milligrams Na,CO,.
Dividing each by its molecular weight, we have :
for P,o,. .......................... .* =0.7014
141.06
for Na,CO,. ........................ .e
106.r
= 16.27
Therefore P,O, : Na,CO,= 0.7014 : 16.27= I : 23.2.
I*zbg3
394 DETER 31 I S .‘.TIO S OF P I I OS 1’11 O K I C .‘.C I 1 .
)

I n other words. 2 3 . 2 molecules of Sa,CO,% are required to

iieutralize the yellow precipitate coiltailling olie ~iiolecule P,O..
T h e above figures are based up011 tlir following atoiiiic
weights :
JI g ................................ - 24.3
0 ................. . . . . . . = iii
1’ .................................. __
- ;. I .o~;
S a ................................ - r,;.og
-
c.... .............................. -
- 12

Mg,P,O,=63.80 per cent. P,O,.

There is some uncertainty as to the correct atomic weight of
magnesium. If M g = 24 (instead of 24.3 a3 taken above) \\.it11
Mg, P,O, = 63.98 per cent. P,O, the ratio of Na,2C0,,to P,O, =
23.1 to I (instead of 23.2 to I ) . I t is difficult to obtain abso-
lutely pure Na,CO, ; any impurity in it will make the ratio of
Na,CO, to P,O, too high.
Practically, therefore, twenty-three molecules of Na,CO, are
required for one molecule P,O,. This agrees with Hundesha-
gen’s results.
I have never seen any explanation as to why twenty-three
molecules of alkali are required to neutralize one molecule of
the ammonium phospho-molybdate. A discussion of the sub-
ject, therefore, may be of interest. Hundeshagen has shown
(hoc.a?.)that (neglecting any water of crystallization) the yel-
low precipitate, after thorough washing with water, has the
following composition :
PO,
1 2 MOO,
3 ”,
Or, doubling the formula, for the sake of clearness:
PtO,
24 MOO,
6 NH,
K. T. Thomson has shown in his researches on indicators
(Chenz. News, 47, 127) that of the three hydrogen atoms in
H,PO, two must be saturated with alkali before the reaction
with phenolphthalein will be neutral, the next drop of alkali
after this causing the red color to appear. Writing the formula
 DETERMINATION OF PHOSPHORIC ACID. 395
differently, H,P,O, must become R,H,P,O, ( R being the radical
NH,, or any alkali metal). Therefore, when the yellow precipi-
tate is broken up by alkali, only four of the six molecules of N H ,
are required to form (with the P,O, of the precipitate) a phosphate
of ammonium that is neutral to the indicator. T h e remaining
two molecules of N H , unite with one molecule of MOO,, yield-
ing also a salt that is of neutral reaction. This leaves twenty-
three molecules of MOO,, representing the ' ' net available acidi-
ty " (if I may use the expression) of the ammonium phospho-
molybdate. These twenty-three molecules of MOO,, of course,
require twenty-three molecules of Na,CO, (or its equivalent of
K H O ) to form Na,MoO,. Q. E. D. T h e following is the
formula representing the reaction :
+
6NH,.P,0,.24Mo08 q N a , C O , + H,O =
(NH,),.H,P,O,+(NH,),MoO,+ qNa,MoO, f 23CO,
It may be well to give a short resume of this method. One
grain of phosphate rock, or two or three grams of fertilizer are
dissolved in nitric acid, and without evaporating to dryness
diluted to 250 cc. T h e solution need not be filtered. Twenty-
five cc. of the solution are delivered into a four-ounce beaker
and neutralized with ammonia-until a precipitate just begins
to form-and then treated with five cc. of HNO, of 1.4 specific
gravity. Ten cc. of a saturated solution of ammonium nitrate
are added and the solution diluted to a volume of fifty to seventy-
five cc. I t is then brought to a full boil, removed from the
lamp and five cc. of the aqueous solution of ammonium molyb-
date added. This is followed by a second and a third five cc.,
if necessary, the precipitate allowed to settle, and filtered at
once through a seven centimeter diameter filter. I t is washed
thoroughly with water by decantation and on the filter. T h e fil-
ter and precipitate are transferred bodily to the beaker. Stand-
ard alkali is then run in and at least 0 . 5 cc. of phenolphthalein
(one per cent. solution) added, and then standard acid, until
the color vanishes. Each cc. of alkali equals one milligram of
phosphorus pentoxide.