September15,1933 INDUSTRIAL AND ENGINEERING CHEMISTRY 313

CONCLUSION (3) Heym, Ann. Phys., (9) 12, 443 (1919).

(4) Johnston and Adams, J. Am. Chem. Soc., 33, 829 (1911).
Calcium, strontium, and cadmium are all co-precipitated (5) Mellor, “Comprehensive Treatise on Inorganic and Theoretical
with barium sulfate. Chemistry,” Vol. 3, p. 766, Longmans, 1923.
Beryllium, magnesium, and zinc are not co-precipitated (6) Nitchie, (3. ( 3 . 3 I N D . EXQ.CHEM.,21, 1 (1929).
(7) Pollock, Sci. Proc. Roy. Dublin Soc., 11 (N.S.), 184 (1907).
with barium sulfate. (8) Popoff and Neuman, IXD.ENG.CHEW,Anal. Ed., 2, 45 (1930).
The data for cadmium indicate the preferential adsorption (9) Popoff, Waldbauer, and McCann., Ibid., 4 , 4 3 (1932).
of iodide ion or of Cd14--.
LITERATURE
CITED RECEIVED April 16, 1933. Presented before the Division of Physical and
Inorganic Chemistry at the 85th Meeting of the American Chemical Society,
(1) Cornog, J . Am. Chem. Soc., 43, 2573 (1921). Washington, D. C., March 26 to 31, 1933. From the thesis of E. St. Clair
(2) Fajans, “Radio Elements and Isotopes,” P. 95, McGraw-Hill, Gants for the M.8. degree, State University of Iowa.
1931.

Determination of Calcium in Lead-Calcium

Alloys of Low Calcium Content
BEVERLY A. WOOTEN,Bell Telephone Laboratories, Inc., New York, N. Y.
L. CLARKEAND LELAND

T H E purpose of this work was the development of a

rapid, high-precision method for determining calcium
in lead-calcium alloys containing from 0.02 to 0.06 per
cent calcium. Small amounts of several metallic impurities
were present in the alloy.
5 . Incomplete removal by washing of the excess ammo-
nium oxalate in the calcium oxalate precipitate.
The experimental basis for so characterizing the above
follows :
1. The Western Electric Company method specified a
After the removal of the lead as sulfate from the nitric cooling period of 5 minutes a t room temperature before
acid solution of the alloy, calcium and the impurities remain filtration of the calcium oxalate precipitate. Since the com-
as the nitrates and sulfates. It is generally agreed that the pleteness of the precipitation is a function of the final tempera-
best form in which to separate calcium for determination ture of the solution, results obtained in summer will show a
is the oxalate, and where speed is desired the volumetric negative error from this cause measurably higher than in the
evaluation of the calcium oxal’ate is obviously preferable to case of those obtained in winter. Experiment confirmed
the gravimetric. Our problem was therefore that of modi- this; microchemical examination of the filtrates from the
fying the classical method for determining calcium by the calcium oxalate precipitation showed that cooling in an ice
permanganate titration of the sulfuric acid solution of the bath reduced the error by approximately 50 per cent.
oxalate, so that our requirements of precision and rapidity 2. The presence of asbestos in the solution being titrated
would be met. It was desired to run two concurrent deter- was found definitely to obscure the end point, and to produce
minations in an hour, with a precision of +0.002 per cent erratic results in check determinations. A positive error was
calcium. introduced, caused either by a reducing action of the asbestos
Shaw, Whitternore, and Westby (2), of the Western or by the tendency to overtitrate in turbid media, or by both.
Electric Company, have published a method for this determi- A certain alloy gave successive values of 0.051, 0.056, 0.053,
nation. They dissolve a 20-gram sample of the alloy in fum- and 0.055 per cent calcium by the Western Electric Company
ing nitric acid, add sulfuric acid, filter off the lead sulfate, and method. When the same alloy was analyzed by the method
add successively ammonium hydroxide in excess, ethyl modified by the substitution of a Frittig glass filtering cru-
alcohol, and ammonium oxalate. The mixture is boiled for 2 cible for the asbestos Gooch, the values obtained by the same
minutes, and the precipitate allowed to settle for 5 minutes. analyst were 0.040, 0.041, 0.042, and 0.042 per cent.
The mixture is filtered through asbestos, the precipitate 3. It was originally suspected that a source of error
dissolved in sulfuric acid, and titrated hot with permanga- in the method lay in the failure to separate the Second and
nate in the presence of the asbestos. Third Group Metals as hydroxides before precipitating the
A careful trial of this method led us to suspect compensat- calcium as oxalate. Upon investigation, however, it was
ing errors. Repeated determinations on the same sample proved that no measurable error was produced from this
were generally consistent when performed on the same day cause, at least in the case of the alloy in which we were pri-
and by the same analyst; but were generally inconsistent marily interested. This alloy contained on the average the
when either of these conditions did not obtain. This ob- following impurities: copper, 0.07 per cent; tin, 0.002 per
servation pointed to the existence of uncontrolled variables. cent; bismuth, 0.001 per cent; antimony, 0.001 per cent;
A careful search for such variables led to the discovery of the arsenic, 0.002 per cent; and iron, 0.002 per cent. We found,
following sources of error and experimental difficulties: however, that when the Frittig crucible was substituted for
1. Incompleteness of the calcium precipitation under the the asbestos mat, the hydroxides of these metallic impurities,
prescribed conditions. if not previously removed, tended t o clog the Frittig crucible
2. Difficulty in selecting end point in presence of asbestos, and render filtration prohibitively slow. For this reason,
and possible reducing action of the asbestos. and also to guard against samples with abnormal amounts
3. Failure to make an ammonium hydroxide separation of impurities, it seemed desirable to filter off the ammonium
of the Second and Third Group Metals before precipitating hydroxide precipitate before precipitating the calcium.
calcium as oxalate. 4. Hahn and Weiler ( 1 ) state that calcium oxalate is appre-
4. Solubility of calcium oxalate in the unneutralized ciably soluble in hot water, but much less so in dilute am-
hot wash water. monia solution. I n corroboration of this statement, on a
314 ANALYTICAL EDITION Vol. 5, No. 5
synthetic lead-calcium solution containing 0.0396 per cent STANDARDIZATION OF SOLUTIONS. Calcium Chloride Solutions.
calcium, we obtained a value of 0.0344 per cent on washing Dry calcium carbonate was dissolved in c. P. hydrochloric acid
and the solution diluted to the desired volume. The calculated
with hot water, and 0.0381 per cent when the washing was calcium values of these solutions were found to agree very closely
done with cold 1 per cent ammonium hydroxide. The last with those determined by the standard oxalate-permanganate
washings were proved to be free from oxalate by testing method. Portions of the solutions (50 ml. or 100 ml.) were used
with permanganate. for the standardization. In precipitating the calcium oxalate,
the solution was allowed to digest overnight.
5. Tests with permanganate showed that washing a Standard Potassium Permanganate Solution. Merck's c. P.
calcium oxalate precipitate three times with hot water does potassium permanganate was used in the preparation of 0.05
not remove all the excess precipitant. To some extent this N solutions which were prepared and stored with the usual
error compensates that given in the preceding paragraph. precautions. The solution in use was frequently standardized
against Bureau of Standards sodium oxalate.
It is clear from the foregoing that the Western Electric
Company method, while doubtless satisfactory where speed The data are summarized in Table I and show that the
is essential and a moderate precision sufficient, involves un- method invariably gives results which are in error in the
controlled variables and compensating errors which make it sense of being too low, this error increasing somewhat as the
unsuitable for work requiring the highest precision. We calcium content decreases. The method is remarkably
attempted to devise an improved method which would be reproducible, however, as will be shown in the next section;
free from those objections. The procedure follows: and the existence of a constant error, if the source and magni-
tude of this error be certainly known, cannot be an objection.
PROCEDURE I n a search for the source of the constant error, the lead
To a 20-gram sample of the finely milled and thoroughly sulfate .and ammonium hydroxide precipitates were elimi-
mixed sample, contained in a 400-ml. beaker, add 100 ml. 1 t o 3 nated by spectroscopic and microchemical methods as con-
nitric acid. When the vigorous reaction has ceased, heat on a taining negligible quantities of calcium. The filtrate from the
hot plate to hasten solution. When solution is complete, wash calcium oxalate precipitate was found, however, to contain
the sides of the beaker with hot water and add 20 ml. 1 to 1 appreciable quantities of calcium. As a representative
sulfuric acid. Remove the lead sulfate by suction filtration
through an asbestos mat and wash thoroughly with hot water. example, a synthetic solution containing the equivalent of
Reject the precipitate. At this point the volume of the solution 0.0397 per cent calcium gave an apparent calcium content
should be between 150 and 175 ml. of 0.0381 calcium, an error of -0.0016 per cent calcium.
Neutralize the solution with ammonia and add 15 ml. in excess. The filtrate from this determination was found by micro-
Boil for 5 minutes and filter through a No. 42 Whatman filter
or an asbestos mat, washing the precipitate with hot water. chemical means to contain 0.34 mg. calcium, which is equiva-
Reject the precipitate. lent to 0.0017 per cent calcium on the 20-gram sample basis.
To the filtrate add 10 ml. concentrated ammonium hydroxide, The source of the constant error is therefore known, and
30 ml. of 95 per cent alcohol, and stir well. Add 2 rams of Kahl-
baum's "pro-analyse" ammonium oxalate. Boil for 3 minutes cannot be eliminated without sacrifice of speed. The con-
The volume of the solution should be about 300 ml. Remove stant error for any given calcium content within the range
from hot plate, rinse cover glass and sides of beaker with water, covered can be obtained by interpolation from the data in
and allow to stand in an ice bath for 5 minutes. Temperature of Table I.
solution after cooling should be 45" to 55" C. Filter the re
ci itated calcium oxalate on a clean Frittig glass crucible
1&4), washing beaker six times with cold 1 per cent ammonium
&Ti TABLEI. PER CENT CALCIUX ON 20-GRAM SAMPLE BASIS
hydroxide. Wash precipitate six times with the 1 per cent AVERAQE
DETERMI- DEVIATION
CONSTANT
ammonia solution, filling the crucible each time. Reject wash- NATIONS ADDRD FOUNDFROM MEAN ERROR
ings. % % % %
Remove crucible from holder, and wash holder and outside of 6 0.0080 0.0041 fO ,0004 -0.0039
crucible thoroughly with distilled water t o insure freedom from 6 0.0100 0.0061 =to. 0006 -0.0039
oxalate. Replace crucible in holder and dissolve the calcium 6 0.0149 0.0136 fO .0004 -0.0014
11 0.0198 0.0181 ;to. 0004 -0.0017
oxalate by pouring six 15-ml. portions of hot 10 per cent sulfuric 16 0.0296 0.0277 f0.0006 -0.0018
acid throu h the crucible. Wash once with hot water after each 16 0.0399 0.0381 f0. 0006 -0.0010
-0.0018
addition of acid by filling the crucible Add to the filtrate 20 7 0.0492 0.0482 f O ,0003
7a 0.0000 0 . 0005a fO.OOO1 +0.0006
ml. 1 to 1 sulfuric acid and dilute to 300 ml. Heat the solution a Blank determination on a 20-gram sample of commercial lead.
t o 80' to 90" C. and titrate t o a faint permanent pink with 0.05 N
potassium permanganate.
I n the last line in Table I are given the results of a series
NOTESON PROC~DURE. of blank determinations on 20-gram samples of the com-
1. New Frittig crucibles should be allowed to stand overnight in dilute mercial lead. The value obtained of 0.0005 per cent calcium
chromic acid mixture before being used. Crucibles should be kept in chromic (0.1 ml. 0.05 N potassium permanganate) corresponds closely
acid mixture when not in use. Accumulated lead sulfate may be removed
from the pores of the crucibles by washing with hot ammonium acetate to the indicator error to be expected on a volume of 300 ml.
solution. in which the titration is usually made.
2. All asbestos used should be treated with Concentrated hydrochloric
acid and then washed free of chlorides with distilled water. PRECISION OF METHOD

ERROR
THECONSTANT Eight determinations were made on each of four alloys
to determine the reproducibility of the method. The data
A series of experiments was performed to determine the are presented in Table 11. Since the correction applied for
accuracy of the method. Solutions for analysis were pre- the constant error was obtained by the same method, it is
pared by dissolving 20-gram samples of commercial lead, sukject to the same accidental error. Even taking this into
containing the impurities stated earlier in this paper, in consideration, it is obvious from an inspection of the data in
nitric acid and adding a measured amount of standardized Tables I and I1 that the precision of the method is better
calcium chloride solution. The lead was obtained from the than k0.002 per cent calcium. It should be noted that this
St. Joseph Smelting & Refining Company. For the prepara- high precision is obtainable only by skilled analysts, well prac-
tion of the standard calcium solution, Kahlbaum's purest ticed in the method. Such an analyst can run two concurrent
grade of calcium carbonate was used. A spectroscopic determinations by this method in an average time of 55
examination of this material showed traces of barium, minutes, which will agree within the stated precision. This
strontium, and magnesium only, the total impurities being time can be reduced by about 10 minutes if the use of fuming
estimated a t less than 0.03 per cent. nitric acid for the solution of the sample is not objectionable.
September 15,1933 INDUSTRIAL AND ENGINEERING CHEMISTRY 315
11. REPRODUCIBILITY
TABLE OF METHOD We are indebted to J. D. Struthers of this laboratory for
AVERAGE AVERAGE AVERAQE performing most of the experimental work reported.
DETERMI- APPARENT CORRECTED DEVIATION
ALLOY NATIONS CALCIUM ClLCIUM FROM MEAN

1 1 %
0.0177 %
0.0197 %
-0.0004 LITERATURE
CITED
1 1 0,0189 0.0209 $0.0008
1 1 0.0174 0.0194 -0.0007 (1) Hahn and Weiler, 2. anal. Chem., 70, 1 (1927).
1 1 0.0187 0.0207 $0.0006 (2) Shaw, Whitternore, and Westby, IND.ENQ.Cmix., Anal. Ed.,
1 1 0.0177 0.0197 -0.0004 2,401 (1930).
1 1 0.0177 0.0197 -0.0004
1 1 0.0193 0.0213 $0.0012
1 1 0.0177 0.0197 -0.0004 RECEIVED June 13, 1933. Presented before the Division of Physical and
2 8 0.0271 0.0291 1 0 .0003 Inorganic Chemistry at the 86th meeting of the American Chemical Society,
3 8 0.0406 0.0419 10.0004
4 8 0.0620 .... 10.0006 Washington, D. C.,March 26 to 31, 1933.

Silicon Tetrafluoride Volatilization

W. D. ARMSTRONG,
Laboratory of Physiological Chemistry, University of Minnesota, Minneapolis, Minn.

A highly simplijied apparatus, without gas-

T
H E volatilization of sili- and their connecting tubes. An
con tetrafluoride as a scrubbing devices, for the quantitative evolution all-glass apparatus with only a
step in the determina- and collection of silicon tetrafluoride gives good trap a t -35" c. b e t w e e n the
Of fluorine has reached its digestion flask and the receiver
h i g h e s t state of development
results since the apparatus can be thoroughly also to give complete re-
in the t h o r o u g h s t u d i e s o f dried and possesses no dead spaces* BY tempera- covery of fluorine under various
Reynolds, ROSS, and Jacob (6) lure control, the loss of fluorine caused by hy- conditions of p r o c e d u r e , A
and Shuey (6). These writers, drolysis of the tetrafluoride is usually prevented w h i t e , frost-like s u b l i m a t e
together with Adolph (1) and and, should hydrolysis occur, the Juorine-con- formed in the top of the trap and
Wagner and Ross (8)' ade- in the bends of the connecting
taining products can be driven into the receiver tubes when scrubbers were used,
quate reviews of the subject.
Treadwell and Koch (7), wagner by the application of heat to the delivery tube. in spite of all p r e c a u t i o n s to
and Ross, and Casares (4) have eliminate water from the appa-
reported the quantitative recovery of fluorine. However, ratus. This material may have been the same as the "strange
Reynolds, Ross, and Jacob recovered from calcium fluoride deposit" first mentioned by Treadwell and Koch (7) in con-
an average of only 93.5 per cent of the calculated fluorine. nection with their gasometric method for measuring silicon
Furthermore, the results of Shuey with sodium fluoride varied tetrafluoride. They, however, observed it only when working
from $9.77 to 100.36 per cent and averaged 95.72 per cent. below atmospheric pressure. It seems likely that this deposit
Willard and Winter (Q),using a modification of the Casares was silicic acid and hydrofluosilicic acid produced by the hy-
procedure, obtained good recovery with inorganic fluorides drolysis of silicon tetrafluoride by water retained on the glass
but found low results on the determination of fluorine added walls or formed in the reaction bulb. The rinsings of the
to feed ash. trap and tubes were acid in reaction, and after several deter-
Previous workers, except Casares and others who have minations a white solid accumulated in the apparatus a t the
employed a gasometric technic, have found it necessary to points where the original deposit collected. This substance
scrub the silicon tetra- was insoluble in water
fluoride before it en- but dissolved in boiling
tered the receiver, in alkali apd was therefore
order to remove sub- probably s i l i c i c acid.
stances, chiefly acids, I n an all-glass appara-
which interfered with tus designed to collect
the analysis of the re- s i l i c o n tetrafluoride
c e i v e r c o n t e n t s for quantitatively f r o m
fluorine. Various so d i u m silicofluoride
types of wash bottles decomposed by heat a
containing sulfuric acid similar deposit was ob-
or s u l f u r i c acid plus served, and the results
silver sulfate or of the titration of the
c h r o m i u m trioxide receiver contents f o r
h a v e been used t o hydrofluosilicic a c i d
a b s o r b hydrochloric were low. The wash-
acid, sulfur d i o x i d e , ings of the connecting
and oxides of nitrogen. tube of the apparatus
Some o b s e r v a t i o n s D required an additional
made by this writer in- a m o u n t of standard
dicate that these wash alkali sufficient to give
bottles cause a loss of a quantitative recovery
f l u o r i n e chiefly on of fluorine f r o m t h e
a c c o u n t of the diffi- d e c o m p o s e d silico-
culty of drying them FIGURE1. SILICON TETRAFLUORIDE EVOLUTION APPARATUS fluoride.