IS! Hollingsn-orth, R. P.. -Lx.\L. CFEU. 29, 11.30 ( 1%; J .
\V. X.,Ferris, C. -1.ISD. ESG. &EX, -\SAL. ED. 8, 6 (1936;. (10; Jacob, K. D., Hill, K. L., Marshall, H. L.. Reynoidc, D. S..U. S. Dept. Agr.. Tech. Buli. 364 (1933). ( 1 1 ) Klemm, K., Voss, E., Z . anorg. u ( 9 ) Hoskins,
allgem. Chem. 251, 233 (1943). (is) IIcConnell, D., A m . Jlineralogist 23, 1 (1938). (13) McIienna, F. E., Xuclwnics 8, S o . 6, 24-33; 9, KO.1, 40-9; S o . 2, 51-8 (1961). (14) Slunter, P. h.,Aepli, 0. T., Boesatz, R . .I.,I d . Eng. Chem. 39,427 (1947 I .
(16) Sichols, \I.L., Kindt, B. E.. .%SAL. CHflI. 22, 'is5 (1950). (16~~ Reynolds, D. S.,J. Assoc. O@. Agr. Chemists 17, 323 (1931j. (1;' Reynolds, D. 5.. Hill, K. L., ISD. ESG.CHEU.,- 1 s ~ED. ~ .11, 21 (1939). (1s: Reynold-., E. S., Kershan, J. B., Jacob, B. D., J. -4sscrc. O B . I g r . Chevzists 19, 1% (1936). (19) Rochester, G. D., Phys. Rec. 56, 305 (1939). (20) Scarseth? S. D.. Better Crops W i t h Plan: Food, 24, 10 (1940); T h o & phates in Agriculture," rev. ea., p. 89, Darison Chemical Corp., Baltimore, lid., 1951.
R., Crsig, E:. L.! .\s.*L. Cmx. 26,996 (19%). ( 2 2 ) Wagner, 6.R., Ross. K. E.: 3. It!i. (21) Sneli, E.
Eng. chem. 9,1116 (1917;. (23) Khittaker7 (2. Vi'., Fox,
E.S., Zbid., (21)Wh-ynes, A. L., Dee,T. P., Specis! 19, 46i (19'27:.
Rept. Tech. Meeting, Intern. Superphosphate Slanufacturers , . k o c . , London, September 1953; Inst: Chem. Eng., Graduates & Students Section, Februarv 19%. ( 2 3 ) Kilia&, E. H., Winter, 0. E.. Iso. ESG.CHEU.,.LuL. ED.5, i (1933 j. RECEIVEDfor review January 14, 1850. Accepted June 26, 1959.
Colorimetric Determination of Chloride in Concentrated Hydrogen Peroxide ISIDORE GELD and IRVING STERNMAN Metal Chemistry Section,
U. S. Naval
Material laboratory, New York,
b Aluminum containers, commonly employed for storage of concentrated hydrogen peroxide, exhibit accelerated corrosion in the presence of traces of dissolved chloride. To study this effect, development of a sensitive method was necessary for determining chloride in concentrated hydrogen peroxide. A colorimetric method, based on the reaction of mercuric thiocyanate with chloride ion to release thiocyanate, which forms a reddish orange complex with iron(lll), is capable of determining chloride in 90% hydrogen peroxide, up to 4 mg. per liter with a sensitivity to 0.01 mg. per liter and an accuracy equal to or better than 0.1 mg. per liter. Variables are studied with respect to hydrogen peroxid e decomposition, rea gents, temperature, and interferences
D
of a method for determination of trace quantities of chloride in concentrated hydrogen peroxide was undertaken in conjunction with studies of the corrosive effect ( 6 . 7 ) of such chioride on aluminum containers. The government specificatioxi for hydrogen peroxid? (JIIL-H16005C) pernlits a maximum chloride content of 1 mg. per liter (1 p.pm. on a n.eight-\-olume basisj. The method described is a niodfication and extension of that proiiond by Zall? Fisher! and Garner is:. anri is intended for the determination of 0.01 to 4 mg. per liter cf chloride in co'icentrat,ed 1907G)hydrogen peroxide. EVELOPMEST
REAGENTS AND STANDARD SOLUTIONS
Mercuric thiocyanate, saturated solu1662 *
ANALYTICAL CHEMISTRY
N. Y.
tion in deionized water ( O . O i 5 a t C.). Filter to remoi-e excess. Ferric Perchlorate. Dissolve 14.0 grams of pure iron mire in dilute nitric acid. Add 120 mi. of perchloric acid (70%) and heat to fumes of perchloric acid. Continue fuming for 30 minutes. Cool, add 100 d.of hot deionized water, and boil for 5 to 10 minutes to remove chlorine. Cool and dilute to 1 liter with deionized water. Composite Reagent. Prepare as needed by mixing 2.00 parts by volume of ferric perchlorate solution n-ith 1.00 part by volume of mercuric thiocyanate solution. The rea-gent is stable for approximately 8 hours. Standard Solutions. By suitable dilution. of a sodium chloride master solution (0.1 nip. of C1- per ml.) prepare standard solutions ranging from 0 to 40 y of C1- per mi.. in increments of 2 y per d. These will be employed in preparation of thr calibration curve. 23'
CALIBRATION , Pipet 5.00 m!. of each standard solution into 10-ml. volumetric flasks and add 1.0 nil. of 6-j- sodium hydrosjde and 0.6 ml. of perchloric acid (iOYc). Add 3.00 ml. of the composite reagent and fill to the 10-ml. mark with deionized water. Mix. transfer to a testtube colorimeter cell, and compare the yellow-orange color against Rater a t approximately 460-mp wave length (KS-47 filter is satisfactory). The color is stable for approximately 1 hour, Kith gradual increasing absorbance thereafter. Plot the colorimeter readings, or absorbances. against the corresponding micrograms of chloride present. Make a separate calibration for each nen- solution of mercuric thiocyanate reagent prepared. Conduct the calibration a t approsimately the ambient temperature expected during the actual analysis. -4 representative
curve [Figure 1) indicates that Beer's law is not obeyed. The curve shows an approximate sensitivity of 0.5 y in the l o r e r portion of the curre-beloK 30 7 of chloride. The sensitivity decressec: a t higher chloride concentrations. PROCEDURE
Transier 50 mi. of the hydrogen peroxide sample into a covered 500-nii. tall-form borosilicate glass beaker. Carry along duplicate reagent blanks. Add 10 ml. of deionized water and 1.0 d.of 6-\- sodium hydroxide. Place on a steam bath to initiate a rigorous reaction. Maintain the temperature of the solution between 75' and 90" C. until decomposition is almost complete, then evaporate to dryness on the eteani bath. Dissolve salts and r i w . inside of beaker n-ith 1 to 2 mi. of tipionized water. Add 0.6 ml. of pcrchlorir acid (707,) and 3.00 ml. of the composite reagent. Transfcr the s o h tion to 3 10-ml. volumetric flask, rincing the beaker at least tlvice with c i t ionized water. Dilute to the 10-mi. mark, mix, and compare against n-nter Kith the photoelectric colorimeter. a: described under calibration. Obtair the quantit- of chloride from the calibration curve and deduct the aver'!g;c of the blanks. Several samples may be run simultaneously, escept for peroxide drconiposition, which should preferably be conduct,ed in pairs. Suitable safety precautions should be talien when handling concentrated hJ-droge:> peroxide, especialiy.during decomposition. PRECAUTIONS AGAINST CONTAMINATION
Decomposition and evaporation should be conducted in a hood with suction off to minimize motion of at-
~
22c
Table I.
0.400
2co
Figure 1. Calibration curve (K. S. No. 47 filter)
0 10 0 '70 0.30
0320 160
._
__
B
-
120
-
100
s f
E
4
-
80
I
+
I 0000. 0
0.44
..
0 40
-.
20
40
60
80
r
00
120
160
140
180
-
60
-
40
-
20
*
I,
200
CKORIOE
De\. . irg.,L. +O 03 1 0 01
0 13 0 21
0.40
140 0240
Proposed
Chloride, N g . Liter Addedo Fouudb
80
0280
Accuracy of Procedure
1oor c1-
0.50 0.60 0.70 0.80 0.90
1.00 L20 1.40 1.60 1.60 '00 2 20 2.40 2.60 "SO
2 133 2 . s
0:; -0.04 0 CK)
-(I
3.00
3.00 .ldded to ,WYc H,02 contnining 0.34 mg./1. of chloride. After deduction of 0.34 mg. 1. C1-. Table 11.
0 36 _.
.
032
..
028
..
024
'.
e 0
j *
Inipurit iee or
0 20 '. 016
012
._ -1
06
0.7
0.8
09 Hi
Figure 2.
LO
LI
I 2
L3
L4
L5
1.6
I?
I8
I9
2.0
YERCURIC THIOCYANATE R E A G E N T l S A T U R A T E 0 )
RESULTS
The accuracy of the proccdure was
20 Temperoture
0 2 0. 9
2 J
* . h i I y s k .~iippIiedby ninnufncturci..
Effect of mercuric thiocyanate on absorbance
siderable c o n t a l ~ l l a t i o n ~~Iltroduction of traces of sl;in ,jurillg handling should be avoided.
10
.ldditives Chloride Sitrnte Phcisphnte Suliate Tin .lluniinum .lniniuniuni
30
40
'C.
Figure 3. Effect of temperature on absorbance
nospheric contaminants Lpriucipally i-olat.ilechlorides and sulfides). During evaporation, hood doors should be kept t,ightly closed. Dripping of condensed steam can he avoided by ninint.:iining a stream of cold n-atcr in the hood and avoiding escessive steam. special isoIat.ed room specifically designcd for the detcrnlination of chloride would help minimize atmospheric contanlination. Tobacco smoke will t'nuo? con-
quantitative comparison oi chloride added with chloride recovered. up to 3.00 mg. per liter. -4ccuracy is either equal to or better than 0.1 mg. per liter, with s n average deviation of 0.03 mg. per titer. DISCUSSION
Preliminaxy Treatment. The colorimetric determination of chloride in hydrogen perovide by the mercuric thiocyanate method prcsrnted difficulties not previously cncountered n ith similar methodj for determining chloride in m t e r s (8). L-nder the test conditions the prese w e of more than 25 y of hydrogen
peroside cnusetl low rcsiilte, prot):ibl>. becaLls? it tfCcOIllpOsCs t!liOc\~:lIl:ltC' to sulfuric acid :ind hytlrocyanic :icid ( 3 . as follows: 6H:O.
+ Hg (CSP:b + 2HCS
-+
H;SOI
4- H@O; - 4H:CI
sary. It \vas necessary to conduct the peroside clecomposition under alkaline conditions, to avoid loss of chloride by volatilization. One milliliter of 6-V sodium hydroxirlc was found optimum as an additive, resuking in a n average decomposition time of 30 minutes per snmple. )fore than 1.3 nd. causc't1 ~, a siliceous attack on the g l : ~ producing precipitatc. Sniallrr nniounts incre:iwd decomposition time :ind c:iused loss oi chloride-for e\;anil)lc, 0.1 1111. of G.\sodium hytlroxide incrc:iscd dt.conipoaitioil tinir t o -1 hours n n i l rmilted i n complete loss of chloride, originally prewnt, as 1.3 nig. per liter. .%lthough deconiposition of the a k a VOL. 31, NO. 10, OCTOBER 1959
* 1663
h e hydrogen peraside was \igorous, n i t h rapid e\.olution of copious amounts of \rapor, no significant mechanical loss of liquid occurred, as tests of the va or with n-et pH paper revealed no a k a 'ne spr3y. However, some hydrogen peroside vapor is evolved n i t h the oxygen, and ma!- cause iow results if allowed to condense on glassware used for subsequent. operations. Addition of 10 ml. of water to the hydrogen peroxide sample helped maintain a controllable decomposition rate. Color Reaction. REAGESTS. Use of the perchlorate ion ( 6 ) in t h e composite reagent and for acidifying t h e hydrogen peroxide decomposition residue results in a considerably lower blank than n i t h nitrate or sulfate anion. The composite reagent, n-hich is stable for S hours, reduces the number of handling steps and possibilities of contamination. Another advan?age is elimination of possible error from ammonia in the laboratory atmosphere during the colorimetric procedure. Ammonia, if in contact with the neutral mercuric thiocyanate, will precipitate mercuric hydroside, causing high results. The composite reagent, which is acidic, will absorb the traws of ammonia n-ithout precipitation of mercur!.. Tile proposed procedure specifies a saturated solution of mercuric t h i e cyanate (8),which is approximately &Oyyo at 25' C. ( 3 ) . This concentratioii c m vary (lj, depending on the ambient temperature during preparatiov:. Figure 2 shows the effect. of mercuric thiocyanate on the absorbance of 20,50, and 100 y of chloride. In place
e
of the composite reagent, 2.00 ml. of ferric perchlorate and varying amounts of mercuric thiocyanate solutions were used. It is evident that slight changes in the concentration of t h e mercuric thiocyanate may have a significant effect on the results, especially n i t h higher concentratiocs of chloride, necessitating a new calibration cume for each nen mercuric thiocyanate solution prepared. This phenomenon may be the result of the mass-action effect of mercuric thiocyanate in the reaction:
peroxide, are inciuded for comparison. The data indicate that such contaminants and additives ~ i l not l interfere, even in concentrations sign.nificnntiy higher than those normally encounterea. Interferences from hydrogen peroxide, sulfides ($1, tobacco smoke, and ammonia have been cited. ACKNOWLEDGMEN:
The authors thank A. R. .4liison a n t W. L. llilier of the S a v y Naterial Laboratory for their inkrest and helpful suggestions. LITERATURE CITED
This equation may also explain the apparent lack of conformity t o Beer's law, because increasing concentrations of chloride ill also increase the concentration of mercuric chloride and partiall). shift the reaction t o the left. Addition of concentrated mercuric chloride to the developed color will cause complete bleaching ( 2 ) . TEXPERATCRE. Large variations of temperature have a significant effect on the absorbance (Figure 3). Chloride determinations should therefore be conducted at approsimately the ambient teniperature encountered in calibratiori. ISTERFERESCES. By the proposed method, the contaminants and additives generally present in commercial 9OYc hydrogen peroside produce a n error of less than 0.01 mg. of chloride per liter, in the concentrations shon-n in Table 11. The maximum allon-able limits specified in lIIL-H-l65005C, and a n analysis of a sample of commercial 90% hydrogen
M.,Hahn tionary of Chemical Solu llacmillan, >-e&-'1-ork !.
