lodometric Determination of Cobalt

H. .4. LAITINEN kVD L. W. BURDETT', University of Illinois, lirbana, Ill.

The object of this investigation was to find a simple reacts with iodide in acid solution to yield cobalt(I1)
and convenient volumetric method for the deter- and iodine. The iodine is titrated with thiosulfate,
mination of cobalt in complex compounds. The using a visual starch or an amperometric end point.
sample is prepared by ignition to the oxide and In the absence of interfering metals, cobalt can be
fusion with pyrosulfate. The solution is treated determined accurately. The interference of moder-
with potassium bicarbonate and hydrogen peroxide ate amounts of iron and of small amounts of nickel
to form a carbonato complex of cobalt(Il1) which can be prevented by the addition of fluoride.

I N CON?;ECTION with the syiithesis of various cobalt(II1)

complexes,asimpleand rapid method for the determination of
c-obaltin these compounds was desired. Lack of interferences was
investigation of the nature of' the voniplex is under way in this
laboratory.
This same green complex qerved as a basis for a colorirnetric
of secondary importance because of the freedom from other procedure by Ayres ( 1).
metals.
One of the simplest of the existing procedures was first in- AMPEROMETRIC TITRATIOK APPARATUS
vestigated by Job ( 6 ) , who oxidized the cobalt with hydrogen A Fisher Elecdropode was used as a source of potential as
peroxide in the presence of potassium bicarbonate to give a solu- well as a means of current measurement.
ble green carbonato complex. This he reduced with ferrom pyro-
phosphate, the excess of which was titrated with potassium STANDARD SOLUTIONS
permanganate solution. Standard Thiosulfate Solution. This solution was prepared
Metal ( 7 ) also formed this soluble green complex as a step in from C.P. sodium thiosulfate and was standardized against a
his procedure. However, as a subsequent step he decomposed standard potassium iodate solution made from previously dried
the complex by boiling in sodium hydroxide solution to give cobalt analytical reagent grade potassium iodate which was used as a
primary standard.
(111) hydroxide. After the excess peroxide had been removed Standard Cobalt Solution. This solution was prepared from
hy boiling, the solution was made acid in the presence of potas- carbonatotetramminecobaltic nitrate, for which a method of
Yium iodide, which was oxidized as the precipitate dissolved. DreDaration is described bv Walton ( 8 ) . The comuound was
Engle and Gustavson (6) found that 0.5 to 2 hours were required twice recrystallized from aqueous solution by the addition of an
equal volume of ethyl alcohol, washed with alcohol, and then
to dissolve the cobaltic hydroxide completely, depending on the dried a t 90" C. for 1 hour. Its purity was verified by ignition t o
amount of cobalt in the sample, and therefore used a carbon CorOc as described by Brintzinger and Hesse (3). Thp purity,
dioxide atmosphere to avoid air oxidation of iodide. They calculated from five ignitions, was 99.95 f 0.07%.
oxidized the cobalt directly to the hydroxide, avoiding the forma- The standard solution was prepared by ignition of the complex
to Co804,followed by pyrosulfate fusion. The melts from sev-
tion of the carbonato complex. Difficulty in dissolving the last eral such fusions were dissolved, transferred to a volumetric
trares of cobaltic hydroxide was reported by Willard and Hall flask, and diluted to the mark.
(9 1. Other solutions were prepared by dissolving C.P. grade salts
Job's method has a number of objectionable features. The without further purification.
large volume employed (150 ml.) requires a large amount of
EXPERIMENTAL
potassium bicarbonate (30-gram excess) in order to give a stable
complex. This large excess must be carefully neutralized, to Details of Method Finally Adopted. To 20 to 25 ml. of solu-
prevent loss due to effervescence. The large volume would tion in a 250- to 300-ml. Erlenmeyer k k containing from 1.5
to 250 mg. of cobalt in sulfuric acid solution, add sufficient sodium
probably make complete removal of the excess peroxide difficult, or potassium bicarbonate t o neutralize the acid, then add 5
as the concentration of cobalt, which catalyzes the decomposition, grams in excess. Add 5 ml. of 30% hydrogen peroxide and after
would be decreased. An excess of bicarbonate before addition of effervescence has subsided somewhat, wash down the sides of the
the hydrogen peroxide is desirable, so that the carbonato complex flask with distilled water from a wash bottle t o ensure complete
oxidation of the cobalt. This also serves to wash down any
will form immediately, rather than cobaltic hydroxide which hydrogen peroxide which was splattered upon the walk of the
must be dissolved by addition of excess bicarbonate.
The use of a carbon dioxide atmosphere complicates the method
and should be avoided. Table I. Analysis of Standard Cobalt Solution by
An iodometric procedure would be simpler and more direct Iodometric Method
than the addition of ferrous iron followed by a titration of the Normality of Cobalt Cobalt Relative
excess. Only one stable standard solution would be involved, Cobalt Solution Present Found Error
rsther than two solutions of lesser stability. Mg. dug. 7%
An increased number of interferences would no doubt result -0.07
0.0
from the uae of iodide as a reducing agent. For the present 0.08
- 0 , 1.5
application, however, this is of no consequence. 0.0
The exact formula of the green carbonato coniplev in solution 0.01 14.72 ' 14.74 0.13
has not been determined. Job assigned it the formula KZCoO3, 14.73 0.07
14.75 0.20
while Durrant ( 3 ) considered it to be (KCO2-0)zCo-0- 14.73 0.07
C O ( O C O ~ K ) ~Duval
. (4),in a more recent investigation, iso- 0.001 1.473 1.485
1.491
0
1
8
1
lated from the green solution a green solid which he called cobalt l.503Q 2 9
(111)-triscarbonatocobaltiate, Co [Co(C08)a1, on the basis of 1.444a
1.49Ia
-2
1
0
2
determination of cobalt and carbon dioxide. In solution, how- 1.~503~ 2 0
1.497a 1 6
ever, it would appear plausible to assume all of the cobalt to be 1 .497a 1 8
in the same form, possibly C O ( C O ~ ) ~ - -or
- CO(CO3)2-. An a Amperometric titration of iodine
1 Present address, Standard Oil Co. (Indiana), Whiting, Ind
1268
V O L U M E 23, NO. 9, S E P T E M B E R 1 9 5 1 1269

flask by the effervescing solution, in order t o allow it to undcxrgo 1- is the original volume of the solution, X is the volunic of re-
decomposition. agent added, and i is the measured rurrent. This correction is
Heat the solution gently until effervescence due to the de- unnecessary if the concentration of the thiosulfate is ten or more
composition of the hydrogen peroxide ceases. (If the cobalt times the concentration of the iodine solution being t i t r a t d .
concentration is 0.05 ,Mor higher, standing a t room temperature
for 5 minutes will destroy the excess hydrogen peroxide. If the The amperometric method was employed only for ver>-dilute
holution is heated too long or too strongly, some decomposition solutions. Some results obtained are given in Table I. The
of the complex may occur.) Add several small portions (0.5 tendency to high results with 0.001 Jf cobalt solutions i!: caused
gram) of potassium (or sodium) bicarbonate a t intervals during
the heating period. (ThiP step has proved effective in keeping by the incomplete decomposition of the last trace of hyc11,ogen
the complex stable.) peroxide. The decomposition is complete with 0.005 11f cwbalt'
Cool the solution with a cold water (or ice) bath, dilute to solution. Some other results arv given in Table I1 in coniicvtion
100 ml., and add 5 grams of potassium iodide. Carefully neu- with application of the method in the presence of iron. .\ t>.I)ical
tralize the solution 1% ith 1 to 1 hydrochloric acid, adding the acid
dropwise while swirling the solution until the effervescence titration curve is shown in Figure 1.
(.eases, then add 10 ml. in excess and titrate the liberated iodine Interferences with Method. Any element which will oxidize
with standard thiosulfate solution. iodide to iodine after treatment with hydrogen peroxide and
Results obtained by the above procedure in the analysis of xidification under the conditio1i.Qof the cobalt determination will
htandard cobalt wlutionp are given in Table I. interfere with this method unless it has previously been rendered
inactive by complex formation or some other means. Among
these interfering elements are iron, chromium, manganese, copper,
antimony, molybdenum, tungsten, and vanadium.
Effect of Presence of Iron. It was found that the interfwence
o f iron could be largely prevented bj- the addition of 2 grams of
rodium or potassium fluoride aiid 2 grams of sodium acetate t o
the cobalt solution prior t o the addition of the bicarbonate.
17nder these conditions a light colored crystalline precipitate
formed, the exact nature of \vhic.h is still uncertain. As the addi-
tion of fluoride to an acidic ferric iron solution gave the same pre-
cipitate, it is assumed to be FvF3.41/2H,0,which is a yellow csrys-
talline compound only slightly soluhle in cold water. This
precipitate dissolved if the solution w:is heated longer than had
been done in the procedure previously outlined. More satis-
factory results were obtained by following the usual procedure
and allowing the precipitate to remain, but this made the starch
iodine end point somewhat more difficult to see when anslyzing
dilute solutions-e.g., 0.001 M . Somewhat high results were
obtained in the analysis of both 0.01 and 0.001 M solutions,
although in the case of 0.01 M solutions, the error was only of the
order of 0.4% or less in the presence of a 3-fold excess of iron
(Table 11). The error with the 0.001 M solution, in the presence
of a 20-f-!d excess of iron, was of the order of 4 to 5%. Smaller
amounts of iron gave correspondingly lower errors.

Table 11. Analysis of Cobalt Solutions in Presence of

Varying Amounts of Iron
Figure 1. Titration of Iodine w-ith Sodium Cdbalt Iron Cobalt Relative
Thiosulfate Present Present Found Error
Using rotating platinum electrode 111 ml. of 1.13 X 10-4 .M io-
MQ. ,MU. .MU. ?
dine with 9.74 X 10-4 M sodium thiosulfate in dilute hydro- 14.72 50 14.76 0.2i
chloric acid solution approximately 0.1 M in sodium chloride. 14.72 50 14.75 0.20
Applied e.m.f.. -0.2 volt DS. S.C.E. 14.72 30 14.77 0.33
14.72 30 14.77 0.33
1.472 10 1.426 -3.2
1.472 10 1.480 0.54
Amperometric Titration of Liberated Iodine. With 0,001 M 1.472 10 1.497 1.7
1,472 20 1.472 0.0
thiosulfate solutions, the end points were not so easily distinguish- 1.472 20 1.438 -2.4
able as with solutions of higher concentration. It was especially 1,472 30 1.538 4.3
1.472 30 1.544 4.9
difficult to distinguish the end point of a solution containing 1.472 50 1.47Za 0 0
1.472 50 1.47ga 0 34
various amounts of iron. Therefore, the amperometric titration
a Precipitate filtered off prios to heating of solution to r c m o v c cxcess
of the iodine with thiosulfate solution, using a rotating platinum peroxide.
electrode, was investigated and found to give very sharp end ~ . . ______. -
points when using an applied potential of -0.2 volt (its. the
baturated calomel electrode).
Good accuracy and precision were obtained by filtering off
Khen the amperometric titration was employed, the method the precipitate by suction, using a sintered-glass crucible of
was carried out exsctly as when the iodine-starch end point was medium porosity, prior to the heating of the solution. Under
used, up to the point of the actual titration. At that point the these conditions 1.5 mg. of cobalt were determined in the pres-
iodine solution was transferred from the 250-ml. Erlenmeyer
flask to a 250-ml. beaker, after which it was titrated. The ence of 50 mg. of iron with an accuracy and precision of 0.35%
total volume of the solution a t the end of the titration was or better. These results are also given in Table 11.
measured by means of a graduated cylinder, and this volume, Effect of Presence of Nickel. Sickel, unexpectedly, also
less the volume of thiosulfate added, was taken as the volume of interferes. When the solution is heated to remove excess per-
the iodine solution at the beginning of the titration. This value
~ 1 necessary
3 in order to make a correction for the volume oxide it becomes more basic anti a precipitate of what was ap-
parently nickel hydroxide forms. The presence of this precipitate
rhange according to the formula: i,,,,. = (' + where
ij
causes high results to be obtained possibly because of the fomla-
V
1270 ANALYTICAL CHEMISTRY

tion of some nickel dioxide from the nickel hydroxide in the pres- A preliminary separation of cobalt is necessary if large amounts
ence of hydrogen peroxide. The modified method employed in of iron, or significant amounts of nickel, are present. An ampero-
the presence of iron also gave improved results in the presence of metric titration of the liberated iodine with thiosulfate solution
nickel, as the nickel was not readily precipitated, because of provides a convenient method of determining the iodine content
complex formation with the fluoride. The values given in in dilute solutions.
Table I11 were obtained using this modified procedure. Results In the acidification step a relatively large excess of potassium
were still high, however; 1.5 mg. of cobalt were determined with iodide is employed to avoid loss of iodine through volatilization.
an accuracy of 3.7% in the presence of 50 mg. of nickel. Ampero- Hydrochloric acid is employed rather than sulfuric, as is specified
metric titration of the iodine with thiosulfate was made in each in many iodometric methods, on the theory that any chlorine
of the analyses listed. produced by oxidation of chloride ion by the trivalent cobalt
would react with iodide to give an equivalent amount of iodine
rather than being evolved, as would be the case Rith most of the
Table 111. Analysis of Dilute Cobalt Solutions in Presence oxygen liberated from a sulfuric acid solution.
of Varying Amounts of Nickel Table V shows the results of a number of experiments made on
Cobalt Kickel Cobalt Relative dilutions of a single cobalt sulfate solution to determine the re-
Present Present Found Error
quirements for, and extent of, hydrogen peroxide removal.
Me. MC7. Mu. %
Complete removal of hydrogen peroxide is attained in solutions
as little as 0.005 M in cobalt, as shown by the agreement between
the resultg obtained with dilute and concentrated solutions.
Solutions which were 0.01 M in cobalt gave high results when
titrated after standing for 10 minutes a t room temperature, but
gave the correct result after heating. Solutions which were more
I t is concluded that a preliminary separation of cobalt is than about 0.05 M in cobalt concentration did not even require
necessary if appreciable quantities of nickel are present. heating for peroxide removal, if they were allowed to stand a t
Determination of Cobalt in Cobalt Complexes. The applica- room temperature for 10 minutes before titration.
tion of the method to the determination of cobalt in complex
salts requires first the complete decomposition of the complex
salt, in order that the carbonato complex can be formed.
Table V. Effect of Various Conditions on Removal of
The only satisfactory procedure investigated involved the de- Hydrogen Peroxide from a Solution Containing the
struction of the complex by careful ignition, followed by fusion of Carbonatocobaltic Complex
the resulting oxide prior to analysis by the iodometric method. Concentration of
Cobalt Solution, Conditions Employed Cobalt Found,
Weigh out a sample large enough to give a suitable titration M for Peroxide Removal Me.
into a porcelain crucible and decompose it by careful heating 0.1 Sample heated till effervescence 2.472
with a gas burner. Heat the sample from the top by carefully ceased 2.477
pointing the burner toward the crucible until all the visible ma- 2.478
terial is decomposed. (This is to avoid loss of sample by being 0.1 Sample allowed to stand at room 2.482
blown out of the crucible as decomposition gases are evolved.) temperature for 5 minutes 2.478
Gently heat the bottom of the crucible to complete the decom- after complex formation 2.480
position. 0.1 Sample allowed to stahd at room 2.477
Fuse the resulting oxide with potassium pyrosulfate and temperature for 10 minutes 2.478
allow the melt to cool somewhat. Loosen the melt by adding a after complex formation 2,475
mall amount of water and gently heating with a gas burner with 0.1 Sample allowed to stand at room 2.477
a low flame. Transfer the contents of the crucible to a 250-ml. temperature for 30 minutes 2,475
Erlenmeyer flask and complete the dissolution of the sample. after complex formation 2.478
Complete the analysis as described above. 0.1 Sample allowed to stand a t room 2.478
temperature for 18 hours after 2.479
Results obtained with this method in the analysis of a, sample complex formation
of carbonatotetramminecobaltic nitrate are given in Table IF-. 0.01 Sample allowed to stand at room 0.2586
temperature for 10 minutes 0.2571
The maximum deviation from theoretical is O.l5%, while the after complex formation
average deviation is only 0.05%. 0.01 Sample heated till effervescence 0.2175
ceased 0.2473
0.2473
0.005 Sample heated till effervescence 0.1236
Table IV. Analysis of Carbonatotetramminecobaltic ceased 0.1236
Nitrate 0.1239
Purity from
Weight of Weight of Theoretical Iodometric
Sample Oxide Weight of Oxide Analysis
G. G. G. %
Table V also shows the stability of the complex, as satisfactory
0.8231 0.2684 0,2662 100.15
0.6769 0.2181 0.2181 loo. 00 results were obtained after the sample had stood for 18 hours after
0.7543 99,98 complex formation.
0.8239 100.08
0.7965 100,Ol
LITERATURE CITED
(1) Ayres, G. H . , Rept. S e w England Assoc. Chem. Teachers, 42,
Decomposition of cobalt complexes by wet oxidation with 143-7 (1941).
sulfuric and nitric acids or perchloric and nitric acids was un- (2) Brintainger, H., a n d Hesse, B . , Z . anal. Chem., 122, 241 (1941).
successful, because it was difficult or impossible to remove the (3) Durrant, C. H . , J. Chem. SOC.,87, 1781 (1905).
(4) D u v a l , C., Anal. Chim. Acta, 1, 201 (1947).
h a 1 traces of nitrogen oxides which interfere with the iodometric (5) Engle, J., a n d Gustavson, G . , J. Ind. Eng. Chem., 8, 901 (1916).
procedure. (6) Job, A., A n n . chim. phus., 20, 214 (1900).
DISCUSSION AND SUMMARY (7) Metal, G., Z . anal. Chem., 53,537 (1914).
(8) Walton, H . F., “Inorganic Preparations,” p . 91, N e w York,
The procedure as outlined was found to be rapid and simple, Prentice-Hall, 1948.
especially when applied to samples containing no metallic (9) Willard, H . H . , a n d Hall, O., J . A m . Chem. SOC.,44, 2237 (1922).
ions other than cobalt, such as the various complexes of cobalt. RECEWED
March 21, 1951.