The accurate determination of cobalt

ANALYTICA

20

THE II.

ACCU

LiATE

Gl~AVIMIYl’IC1C THE

CHIMICA

ACTA

I~ETEI~MINATION MIII‘IIODS I’HOSPHA’I’JS

OF

COBALT

AND IN PARTICULATC METHOD

A. G. IWSTI’:I< ANDLV. J. wIr~r,IAikls+ IZoycilHoIlowcry

Collep,

Univevsily

(Rcceivcd

January

of Lo?tJo~r. (Gvcal

14th.

Bvilnirr)

~cjGo)

I NT-ROD UCTION

Gravimctric methods used in the determination of cobalt have been critically discusscdl. The precipitation of cobalt by a-nitroso-P-naphthol introcluccd by TLIXSI~I AND VON KNORRIJ~ in 1885 was one of the first methods used for the determination of a metal with the aid of a sclcctivc organic rcafq!nt. In spite of its many dcfccts, and the large number of other methods proposed, it remains, in modified form, one of the most frequently recommended methods for the dctcrmination of cobalt. No matter which modification is usccl, the method is open to much criticism. The precipitate itself is impure, due to the cobalt being in two valcncy states; the reagent also acts as an oxiclant, itself becoming reduced to products which further contaminate the prccipitate. Poisonous fumes arc given off on heating, and the bulky nature of the prccipitatc ancl its tcndcncy to stick to glassware make filtration difficult. Ipition of the complex to the ‘oxiclc’ leads to further difficulty owing to its uncertain composition. Ticduction of the ‘oxide’ by hydrogen to metallic cobalt for final weighing is the form of the method claimed to be most accurate. The important question here is whether the metallic rcsiclue contains tracts of unreduced oxide. Also, in this form, weighing errors hnvc the masimum cffcct. This, no doubt, is a reason why conversion of metallic cobalt to Co!304 for final weighing has also hccn recommended. Recluction by hydrogen to the metal via “C0304” of uncertain composition has been rccommcndcd as the final proccclurc in many mcthocls of cobalt analysis. Variation in composition of the prccipitatc or final wcighini form is the major source of error in a number of cobalt gravimctric mcthocls e.g. the phcnylthiohyclantoic acid and cobaltinitritc methods. Its cffcct on the sulphatc method will be apparent from the esaminntion of CoSO.4 as a possible rcfcrence standardl. Slight solubility of the precipitate is the major source of error in other cobalt methods, particularly the phosphate method. The aim of the present work was to csaminc csperimcntally a number of cobalt methods using a cobalt rcfcrcncc standard previously shown to be reliablcl. In selccting methods for clctailccl examination, consideration was given to rclcvant data already available e.g. thcrmogravimetric bchaviour of prccipitatcs involved. Those *

Prcscnt aclclrcsv: Collcgc of l’ccjinology,

Wistol

(Great

Britain). Awd.

C/rim.

Acln,

24 (196x)

20-32

ACCURATE DETERMINATION OF CO. II.

21

finally chosen for critical examination were the anthranilate, phosphate, and electrolytic methods, the last being included for convenience with gravimetric methods. Methods depending on final reduction to metallic cobalt, or ignition as CoSO4 were rejected for reasons given above. I
of the ammonium phosphate yields the pyrophosphatc, form. 2 CoNI-141'04~1-120+ Co2zJ207 + 2 NI-ia _t-3 1130

an altcrnxtivc

The corresponcling mcthocl for magnesium has an extensive litcraturc; ;L good account of the difficulties involved is given by KOLTIIOFF ANI:, SANDELL~. Earlier work on the cobalt method took no account of residual cobalt in the filtmtc which DuI:T\*~ and later SCIIOISLLIX~~ clctermincd gravimetrically following precipitation as sulphidc. I>uI:TY’s results, which indicated that there was a significant amount of cobalt in the filtrate, wcrc ignored by some later workers. [n assessing the phosphate method \VII,LARD AND HALL 10 invariably obtained higll results. Marc recently MATSUO 11 has investigated the effect of plr and ammonium ion concentration on the precipitation of cobalt ammonium phosphate. I+xipitates producccl under various conditions, wcrc in each case ignitccl to the pyrophosphatc and correctecl for cobalt in solution, which was dctcrminccl using a-nitroso+naphthol. Low results were obtained under all conditions, and the method Was rejected. The pyrolysis curve 12 for CoNHJ?O4.H20 indicates that the pyrophosphate, which appears at ~SO”, is a good weighing form. Much of the work done on the pl~osph.‘tc method has been concerned with sepqrating nickel from cobalt as well as a means of determining the latter. Little attention has been given to the best method of producing a precipitate of satisfactory composition.The conflicting results of WILLARD RNDHALL~O~~C~ MATSUO~~ suggest v‘ariation in composition of the precipitate, although use of faulty rcfcrcncc standards may have been a contributory factor. In its existing form, with the determination of cobalt’,~, in the filtrate by sulphide Amal. Ghim. Ada, 24 (x96x) 20-32

A.

22

G. FOSTER,

W.

J. WILLIAMS

precipitation followed by ignit,@ to ‘Co304’, two gravimetric procedures are involved which detract considerably from tile value p_f_Fthe method. The electrolytic method has been rccommen%zd as the best for determinations involving large amounts of cobalt, and where high accuracy is desired. The method has been used to stanclardisc cobalt stock solutions for USCin assessing other methods of cobalt analysis. Electrolytic cobalt has been used by many workers as a reference standard. Some of the unsatisfactory features are said to bc: (a) incomplete deposition; (b) an occasional slight anode deposit in the abscncc of a reducing agefit; (c) a contaminated deposit in the presence of a reducing agent ; (~2)CLtcndcncy for the platinum anoclc to dissolve during prolonged electrolysis. l’hc electrolytic method has been carried out with a greater diversity of procedure than any other method. Chemical composition of the elcctrolytc, time of electrolysis, current density, temperature, type of clectrocle, rate of stirring, and prr of the solution have been varied. Various procedures have been cvolvecl in an attempt to overcome the unsatisfactory fcaturcs of the method. l’rcatmcnt of the spent electrolyte with H& to determine dctermincd sulphur residual cobalt via ‘Co304’ has been rccommendcd 1.3.KALLMAN~~ in the clcctrodeposited cobalt (arising from the use of sodium bisulphitc), as BaS04, and corrected the weight of deposit accordingly. EXAMINATION

OF METIIODS

Alicluots of a standard solution of cobalt were analyscd using this method. Previous examination of filtrates from the anthranilate method had indicated the presence of small amounts of cobalt, All filtrates and washings, therefore, were made up to 5oo ml and analysccl for cobalt using a Nitroso-R-salt mcthocl. Generally, the filtrates were analysecl as soon as possible because light turns sodium anthranilate brown. Otherwise the filtrates were stored overnight in a dark place. Rccovcrics of cobalt in the Nitroso-I2-salt method were shown to be unaffected by the amount of sodium anthranilate present. Some representative results are given in Table I. In carrying out the above procedure, a drawback was encountered which previous TARLE ANTHRANILATIS

Normal

I 2

3 4 2 M&l

120.3 120.3 120.3 120.3 120.3 120.3 crrorr”t”b331

10

15 20

30 30 g

rng

I WETIIOD

proccdurc

120.6

0.

120.5

0.2

120.2

0.5 0.2 0.2 0.1

120.7 120.7 120.7 120.5 120.5 12o.G

0.1

120.6

120.3 I 20.3 120.5 120.5

I

i-o.4 -t-o.4 -to.4 j-O.2

j-o.2 -l-o.3 i-o.3

-kO.ZGO/”

A-al.

Ckim.

Acta,

24 (x961)

20-32

ACCURATE

DETERMINATION

*

OF CO.

II.

23

workers do not appear to have mentioned. The precipitate is very light, and even on the most gentle boiling of the test solution spraying occurs, which results in particles of the precipitate being deposited on the walls of the beaker and the underside of the covering glass. Much care is required to remove this quantitatively. In an attempt to overcome this difficulty, experiments were carried out to find whether heating the solution on a water bath could replace the five minutes’ boiling. The results are shown in Table II. ANTIiHA.SlLATE

hllfTiiOI>

Effect of heating on water bath rather than boiling

--._.__---.___--.--_-_____--__-_--_ uscll %I

tnitr

120.3 I 20.3 120.3 120.3

IO ‘5 20 30

30 30 30 30

I IQ.0 120.1 120.5 120.5

2 _I 1.1 0.5

121.X

-+-0.8

121.2 121.0

-i-O*9 -l-o.7

(J-3

120.8

120.3 120.3 120.3

30 40

45 4.5 45

120,s 121.2

0.2 a

121.2 121.0

121.0

0.1

121.1

laker

No.

-

,,,g

-__1 2

3 4 2

7

>IcsLI~

CrrOr:

-.-__-_-_--I___ Coball

Tl,?W hcalrd on bath

CObldL

Exccsr

J yb

at:lhmnitalc

- ___. .___-__-_-

-f-O.76

30 lng +0.63~~~

TOfd cnbult

Fillrnlc

/ortrrJ a.9 aulhrudatc

c&al t mif

W,K

---_------_--.---

I:‘rro,

fowi

--_--_-_-

wz

tw

In his determination of the per range for complete precipitation, Goro4 used an acetate buffer. He did not detect the increased solubility of the precipitate in its presence, as SIXNNAN, Sh1I-rI-Iand WARD” did in later work. Owing to the ease with which cobalt in solution mxy bc determined using NitrosoIi-salt, some determinations were made in acctatc and tartrate buffers. The y&hod was similar to the first method described above. ‘i’he results arc given in (Table III. Nitrate ions arc allcgccl to interfere with some cobalt methods. The anthranilatc method was therefore carried out in the presence of I g potassium nitrate, but the filtrate cobalt was normal and the total cobalt unaffected.

ANTIIHANILATIS

LSETHOD

Effect of acctato and tartr;rtc

120.3

30

120.3

30

120.3

30

3 g sodium

:Kxtatc! 3 6 sodium Clcctatc 3 g sodium tartratc

118.4

0.2

118.6

-1.7

118.4

0.3

I IS.7

-1.G

x13.0

6.4

x IS)*4

-0-g

Ex+crint.ental* lieferevtca standard: [Co(NH3)tCl]C12 was used. Stock described in Part I. I_-_-* General tcchniquc described in Part I (rcf.1).

solutions

were prepared

Anal. Chins. Ada,

as

24 (xgGr) 20-32

24

A. G.

FOSTER,

W. J.

WILLIAMS

So(Eiz~z nntltratitilale solaltion : Ratches of 5-8 g ordinary grade anthranilic acid were boiled with water in the presence of animal charcoal, filtered, and left to re-crystallise. l’hc product after three treatments was very faintly yellowi.sh, m.p. 14.4”. It was dried over 1-‘208. Sodium anthranilatc solution 3% was prepared in the following way. Slightly less than 7.5 g anthranilic acid was shaken up with 55 ml N NaOH diluted to zoo ml with water. One drop of phcnolpl~tl~alcin was added, and the solution made slightly acid by adding a saturatccl ac~ucous solution of anthranilic acid. ‘The total volume was’ mnclc up to 250 ml with water ancl the reagent stored in a dark bottle in a cool place. Nilroso-Z\‘-salt solzflion: A 0.2.~/~aqueous solution was prepared. It was found to be stable for several months if kept in a dark bottle and stored in a dark place.

l’hc volume of the cobalt solution was adjusted to zoo ml, then 1.o Nsodium carbonate solution added until a slight clouclincss appcarcd. This was just clcarccl by acldition of I : zoo (v/v) HCl. ‘I’he solution was heatccl until it boiled, further additions of HCl being made to maintain a clear solution, An excess of 3% sodium anthranilate was slowly adclecl, the solution being stirred. After five minutes of gcntlc boiling it was left to stand for bnc hour, filtered (X4 sintcred glass filter), and washed with cold u.TS~/~ solution followed by 5 ml alcohol. After drying for one hour at 115~ it was coolccl ancl weighed.

The nnthranilnte mctllod is clifficult to apply bccausc there is a tcnclency for the prccipitatc to splash. The results arc all sli@tly high, probably clue to contamination of the precipitate by the reagent itself. When heating is carriccl out on a water bath, the error is doubled, thus making this proccclurc inadmissible. If a factor were employed, it would have to bc based on the total cobalt, inclucling cobalt rccoverccl from the filtrate, on account of the latter not being constant. From the results shown in Table I, such a factor woulcl have the value o.gg7. ‘I’llc dcterminntions maclc in the prcscncc of sodium acctnte ancl sodium tartratc indicate that the bchaviour is not simply a question of solubility of the precipitate as previously sue;gcstcd. In the prcscncc of acetate ions, cobalt in the filtrate is normal, but the total cobalt is low. Tn contrast, the presence of tartrate ions leads to an abnormally high amount of cobalt in the filtrate, but a lower negative error for total cobalt. The method cannot be used in the presence of thcsc ions cvcn when residual cobalt is determined in the filtrate. ‘flit anthrnnilatc method is not entirely satisfactory. Its only aclvantagc appears to be the non-interfcrcncc by nitrate ion.

first proccclurc .triccl was basccl on the standard zinc methodlb. This consisted of adding an csccss of xolyb cli-ammonium phosphate solution to a hot neutral cobalt solution buffered with a misturc of ammonium chloriclc and sodium acetate. The prccipitatc of cobalt ammonium phosphate was subsequently filtered, washed with r% cli-ammonium phosphate solution followed by alcohol, and ignited at 730” to the pyrophosphate for weighing. It was the phosphate method I usecl in Part I (ref. l).

The

Anal. Ghim. Ada, 24 (196x) 20-32

ACCURATE DETERMINATXON OF CO. II.

25

Using [Co(NH3)&l]Cl2 as reference standard, results were consistent but invariably high. The precipitate was rather fine, and easy to transfer, but had a tendency to make washing and filtering difficult by becoming tightly packed under suction. In order to determine whether under different conditions of PI-I (n) the residual cobalt might be less or (b) the composition of the precipitate might be slightly different and give more accurate results, the PII values were adjusted before precipitation using r.o LV NaOH or 1.0 1V HCl, so that for the series the values were distributed between PH 6.o and p~r 8.5. The remaining procedure followed that of Phosphate method I. After precipitation, the precipitate undergoes a series of colour changc~, passing through a royal blue phase and ending in the violet form. It is essential fdr this scin any of the intermediate forms the prcquence to be completed before filtration; cipitate passes through the filter. In this series, the analyses at the higher pr-r values took about 3 h on the water bath to turn violet. The resulting loss of ammonia led to a large drop in ~EI, so that the final per range for the series was narrow. The results indicated that the relative error was independent of per for the proccdurc used. MATSUO~~ in investigating the effect of prr on the precipitation of CoNH.#04. Ha0 adjusted the p~r after precipitation. This work, with slight modification, was repented. The results arc shown in Tnblc IV.

t?FIrECT

OF

.

.C&d1 /okr,i

_

I>11 ON

..

wz

.._ .._. 99.5

99.5 99.5

‘I.1115 I’RISCII’I’TA’t’ION _ _

. Co6ull /mud ((5 CotI’. m#

._. _ _.. __.

lO0.G

100.3 100.0

99.5

99.5 99.5 09.5 99.5

. -._--.._

____. ____...___.____ _.__..^

c)c).x

99.6 99.3 99.2 g8.H

OF

_

_ I:ilfrufe cohdl w

^_..

COIiAL’I’

^.

,_

hJI>lONtUhl

I’tfOSl’tlh’TiS

_ Tofal cobnll vrg

_ .__. .

_.

w fillrtrlr

_._ ~.__ . . .

0.5

1oo.H

8.36 8.04

0.0

ro0.G

7.c)’

0.8

100.6

0.8

loo.‘}

7.03 6.74 640 G.L!H 5.63

0.0

I .o

0 *c.,

101.2

100.3 100.1

I .2 100.0 . ..I_ ^_ ..._-_..___,_._._.._. ._.._. .. _..,. .._..._. _ __. __.

The alternative method of precipitating CoNH4P04.HzO was then investigated. This consists of adding ammonia to an acid cobaltous solution containing di-ammonium phosphate. The method has been rccommendecl by SCIIOELLEHQ who dctermincd soluble cobalt gravimetrically as sulphide and COrreCtCd the major fraCtiOn as coZI-‘207 accordingly. SCHOELLER’S method, with filtrate cobalt determined by Nitros&R-salt was found to bc very reproducible. A modification of it was used (Phosphate method II) in Part I. Results on the basis of [Co(NH+Cl]C12 arc included in Series II, Part I. The precipitate in this form of the method is The mean relative error was o.q%. more difficult to handle. It has a tendency to stick to glassware. It was found that if an equivalent amount of HaSO4 was used instead of HCl, the final results were lower by o.r-o.2y0. As the mean error using HCl was +o.xg%, use of H2S04 would produce almost exact results. A large number of determinations has confirmed this. The only disadvantage found in using H&04 rather than HCl was that the change Anal.

Cltirn.

Ada,

24

(IgGI)

20-32

26

A. G. FOSTER,

W.

J. WILLIAMS

from the blue amorphous precipitate formed initially, to the final violet form took twice as long (30-40 rnin). Further investigation of this form of the phosphate method was carried out, the influence of the following factors being determined: (e) excess of di-ammonium phosphate used; (b) excess of T : I ammonia added to effect precipitation; (c) time interval hcfore filtration. A series of analyses using Phosphate method II W:LScarried out, the precipitates being filtered TO min, 20 min, I h,4 h and G h after changing into the violet form ; (d) effect of washing the precipitate with water rather than r”k, di-ammonium phosphate solution; (e) effect of weighing as CoNI-I,#O~.HZO rather than coZho7. A series of determinations was carried out using Phosphate method II, in which the ammonium phosphate was weighed after drying at moo and after ignition at 730”. The results arc given in Table V. TABLE

Effect

V

of weighing as CoNI-1~lJO~~H~O and Cozl’zO~ ..._._.__..___._.. __._- .-.._ - .__...._.._.. __ ..__ -_ ._ . . _.. Cobdl /ouml as CoNMdTh~lJrV (mg) _ .._ . ..-... _.- _--..____._. -... -_.--_

Cobalt fow~d ns Co;J’n0~ (mg) .- _. __. ____._._.__.-._

[email protected]

x47.8 77.6 77.5 77.4 77.2 77.3

77.3 77.2 77.2 77.1 77.0 106.0 106.1

102.5) lOG.2

_._ -_-_--_.

___

147.4 r47.6

___--..- __--.-_-_--.-.--

._.._,__.__ -.--. -

Samples of CoNHJJO~*H20 produced by l’hosphatc method II were heated for intcrvnls of 5” above m#, cooled and rc-weighed. A scrics of determinations using Phosphate mcthocl II was carried out in which the CONH~POII~H~O was ignited at various temperatures bctwccn 650~ and goo’.

I h at

Expcrirnental details of the two principal phosphate mcthocls (Phosphate m&hods T and IT) have been given in Part I. I~efe~e~c stn&ar~: [Co(NET&Cl]Cla * Effect of t)H 01s the firecifiitntion of CoNH@04 -HaO. Sufficient NH&l was added to the cobalt solution to give, together with 25 ml of I~O/~(w/v) di-ammonium phosphate solution aclcled later, a I: : 50 Co+” : N&+ molar ratio. The total volume was adjusted to roe ml. After heating on a water bath for a few minutes, 2.5 ml di-ammonium phosphate solution, roO/, (w/v), was added, and the pH acljustccl using 1.0 N NHdOH. The precipitate was left to stancl overnight. It was filtered, washed with water followed by a little alcohol, dried at roo” and ignited to the pyrophosphate. The PM of each filtrate was measurecl and the rcsiclual cobalt determined using Nitroso-R-salt. Effect of excess di-atmm&wt ~lzos#hzte. The procedure was identical with Phosphate methocl II except that ro ml I : 2 H&04 (v/v) was used instead of HCl, the precipitate Aml.

Chim.

A&z,

24

(1961)

20-32

ACCURATE

DETERMINATION

OF CO.

27

II.

was washed with water instead of 1% (w/v) di-ammonium phosphate, and the amount of di-ammonium phosphate used (as TOO/~ w/v solution) was that indicated in Table VI which contains the results. TAULE VI PHOSPHATIS

METHOD

Effect of excess di-ammonium

99.4 99*4 99.4 99.4 99.4

2.5 4.3 10.0 25.0

50.0

phosphate

99*4 99-4 99.6 99a9

100.3

E/fed of nddiq differed aw~owtls of I : I nmmodz. In Phosphate method II additional I : I ammonia was added so that the precipitate formed, just dissolved. The solution was stirred vigorously; the prcdipitatc slowly appeared and the analysis continued.

The results given for Phosphate methods I and II in Part I of this scrics indicate that variation in procedure leads to different results. ‘This must be due to slight variations in the composition of the precipitates. Some of the published results in the literature using methods similar to Phosphate method I have claimed good results in contrast to the high results found in the present work. ‘This may be due to a compensation of errors. For example, DIclcla and KHAUS~~ use similar methods, but both ncglcct residual cobalt. In all the phosphate determinations, some residual cobalt was found in the filtrate, the amount varying from 0.4 mg to 3.0 mg for a total cobalt of TOO mg. Procedures based on aciclimctric titration of CoNH4P04 - HzOl* would therefore lead to erroneous results. As the rcsiclual cobalt is not proportional to the total cobalt the USC of a factor is not possible. The results shown in Table IV support the trends shown in MATSUO’S workll, namely that the total cobalt is greater as the p11increases. However the actual values for total cobalt found, in relation to the amount taken, differs considerably, all the present results being high in contrast to those of MATSUO. . For example, at PH 8.0, for a NH4+: Co+2 molar ratio of 50 : T the result is 0.4% low whereas the prcscnt work indicates the total cobalt to be 1.4% high. Washing the prccipitatc with water instead of IO/, di-ammonium phosphate resulted in the amount of cobalt in the filtrate being increased by so%. As the filtrate cobalt is determined in any case, there appears to be no advantage in using the ammonium phosphate. Results on the basis of CONH~PO~*H~O dried at roo” and then ignited at 730” are shown in Table V. In all cases except one weighing as CONH~PO~*H~O leads to high results. Heating the cobalt ammonium phosphate at increments of 5” above moo produced erratic results. In the present work, 730” was used for ignition to the pyrophosphate. DUVAL’S workla indicates that similar results could be expected over a wide range of temperAiraZ.~Gh&

Acta,

24’(‘1&)

n&32

A. C. FOSTER,

28

W.

J. WILLIAMS

~~turc.l’l~iswasvcrificdinaseriesignitcd at various temperatures between 65o”and gi>o”. ‘I’ablc VI indicates that the final result in the Phosphate method is influenced by the quantity of di-ammonium phosphate added. Use of IO ml of 10% di-ammonium phosphate represents a phosphate : cobalt molar ratio of 4.3. The results indicate that no npprcxzial~lc error woulcl result from a relatively large change in this ratio. In determinations whcrc just sufficient I: : T ammonia had been qddecl to dissolve the precipitate complctcly, the rcsiclual cobalt was much higher. As the Nitroso-R-salt method is only about IO/~accurate, this procedure is liable to introduce a slight error, and could account for slightly less reproducible results hcing obtained. A similar error is possible if the filtratiqn is made too soon e.g. within ro xnin of the precipitate having turned into the violet form. It was found that by filtering the hot solution over TO mg cobalt was found in the filtrate Furthermore, on cooling bcforc making up to 5oo ml for application of the Nitroso-R-salt proceclurc, there was a slight tendency for the CoNMJDOz~~H~O to rcprecipitatc. On the basis of the above work, tllc following procedure is rccommcnliccl for the accurate tlctcrminntion of cobalt.

‘I’0 the cobalt solution, which should be about neutral add x0 ml T : 2 H&O.~ (v/v), and IO ml 70’%, (w/v) cli-ammonium phosphate solution. Acljust the volume to rgo ml. Heat until boiling and then transfer to a water bath; add sufficient T : f (v/v) ammonia to cause the prccipitatc formed to turn slightly purple and partly dissolve (approx. 25 ml recluirccl). Remove the beaker from the bath when the precipitate is violet ancl stand it in a cool place for at lcast one hour. .L;iltcr through a X4 sinterecl crucible, and wash with cold water followed by 5 ml ~CJ% (v/v) alcohol. Dry the precipitate for Ts min at xoo” and then ignite for one hour at 7~~o”-8000. Cool and weigh as Co8P207. Dilute the filtrate ancl washings to 500 ml in a volumetric flask and transfer cithcr :L To- or zs-ml alicluot to a beaker, Dilute to 40 ml, add one drop of phenolphthalein, ;ux.l neutralist with I : 50 (v/v) HCl. Add a few ml cxccss of the acid, and boil gently on ;L hot plate. Adcl 4 ml so”/0 (w/v) sodium acetate solution, continueboiling, &nd then adcl 4 ml 0.2”/~ (w/v) Nitroso-K-salt. Allow the colour to develop fully (about 2 min), add 2 ml cont. HNOo and boil for a further minute. Cool rapidly and make up to TOC)ml. ‘I’rcat a blank similarly. Measure the optical density using a 4-cm cell and IL 545-rnp filter. From II previously constructed calibration curve, cletermine the residual cobalt and aclcl this to the cobalt found as pyrophosphatc. Tnblc VII contains a typical series of results using the methocl.

Cobult tokll IJIg ---._‘L‘(l.0 2‘1 I .o

-------

12o.g @.2 48.2 24.z 24.2

Cobaltjotruci w1:

-_

Kclctliv~~ ~rrcw %I

2,*0.g 2‘1 I .3

,:::

120.5 48.2 48.1 24.2 24.2

0.0 0.0 -0.2 0.0 0.0

A~zd.

Clriwz. Actn,

24

(1961)

20-32

ACCURATE

DETERMINATION

The Electrolytic

OF CO. II.

29

Method

The present work was mainly restricted to an examination and comparison of some existing procedures, and an assessment of electrolytic cobalt 3s a reference standard in the determination of cobalt. After each test, the spent electrolyte and washings were examined for cobalt using Nitroso-R-salt. The usual assumption was made that all platinum lost from the anode would have been deposited on the cathode. Some representative results arc given in Table VIII. TABLE ELECTROLVTIC

____--.

-8 For this clectrolyscs

series, cotxrlt a high rate

__--_--_

PROCIZIJURE

VIII UNDER

Zt\

99.0 g8.6

45 30

2t\ 3.25v 2i\ 3.25V

roonl room

y8.G 98.7

30 45

2A 3A

roe nl room

98.7 98.7 98.7

4.5 45

--

2

3.5ov 2.25v

3h 3A

.--

--OS5

-0.8 -0.9

-0.5 --0.3

room 2.25V 2.2gV

_.._

0

Go

-1.G

75

----Lao

99.9

i-l.3

100.5 98.9

t-1.5 i-o.3

9g* I 100.1

-l-o.5 -t-J.4

100.3 100.1 100.3

fX.G -t-1.4 fl.3

~_I~----.~.-.---~----

were standnrdiscd using tlw maintainccl rend the N;rHSOn

X~RObl

Nos.

3.25v

1,5A

120

stock solutions of stirring was

solution;

room

30

TA13L15

I was a sulphste

CONDITIONS

g&G

ISLECTRODEI’OSITION

8 No.

VARIAULR

phosph;rto method usud WLS ordinary

; in all grade.

IS

UIC,\RIjONATE--PXIHOS1I)F3

and 3 wore: chloride

COMI’LB,S

solutions.

Attempts were made to deposit cobalt quantitatively from the bicarbonate-hydrogen peroxide complex. The results are given in Table IX. The evidence suggesting that electrodeposited cobalt is contaminated is indirect. A typical electrolytic deposit was analysed by the phosphate method and shown to be impure. In all the above work, the cathode deposit!wa& weighed.after.drying at 100”. The effect of drying at’ higher’ temperttures iklicated nb bhanke .in’ weight up to 140~. AnaL. C++

Acta, 24 (rg6r)

20-32

A. G. FOSTER,

30

W.

J. WILLIAMS

Electrodeposition apparatus : A Griffin and Tatlock electrochemical analysis apparatus with provision for heating and stirring the electrolyte was used. Platinum gauze cylindrical elcctrocles were employccl. Z?cfertince slamlard : [Co (NH3)eClJClz Gcnerul ~roced~~~re. l’hc solution, containing about TOO mg cobalt was elcctrolysed for a known time, a high rate of stirring being maintained. The average current passing and the potential difference bctwccn the elcctrodcs was notccl. At the end of the electrolysis, the beaker was lowered from the electrode with the current still passing, tllc clcctrodcs being washed with water at the same time, l’he spent clectrolytc and washings wcrc made up to fjoo ml ancl cobalt was determined in an aliquot by means of Nitroso-1X-salt. Blank tests carried out on the reagents other than cobalt indicated no intcrfcrcncc with the methocl. The washed clcctroclcs wcrc dipped into ethyl alcohol, dried at IOO~ for 15 min, coolccl [Lncl weighed, De@riitiovt from the cobdl bicarlov~~tc-llcvoxi(~e conr.+lcx. Sodium bicarbonate, ro g and TO ml ~cJ”/~ H& were adclcd to an alicluot of the! cobalt solution. When cffcrvcsccnce had ccascxl, the siclcs of the beaker were washed clown with water ant1 the solution clcctrolysccl in the mnnncr describccl above. Analysis of cZect~oZytic cobalt. 5~) ml of the cobalt solution was analysccl using the Phosphate method IT but nitric acicl was usccl instead of sulphuric acid (because the subsequent cobalt &posit was dissolved in nitric acid, and nitrate ion affects the phosphntc method giving low results). Another 50 ml of the cobalt solution was clcctrolysccl in the prcsencc of ammonia, ammonium sulphatc ancl hydrazinc sulphnte. The weighed deposit was dissolved in nitric acid, and the solution annlysccl for cobalt using the same phosphate mcthocl with a correction for the amount of cobalt in the filtmtc. The quantity of nitric acid usccl to clissolvc the deposit was similar to that used in the first annlysis.

‘Table VIII contains representative results using slight variations of some rccommended procedures for the cletcrmination of cobalt by clcctroclcposition. Results of clcctrolyscs of the bicarbonntc-pcroxiclc complex arc given in Table IS. In all casts cobalt rcmaincd in the spent electrolyte at the end of the clcterminations and it rangcc-1from 0.2 to 5.3 mg. The total cobalt in every cast was invariably high, indicating that the &posit was contaminatccl. Excluding those containing no reducing agent, tlic results in Table VIII arc r.q.lyOhigh. Analysis of a typical deposit using the phosphate method gave the following results: ____ -__---._-_-

..-..-.

-______-_______~*----~._--_-____-_.

-. -.__.__..-.._._.-. - .- --.-.--_ .----_-“---. _-__-_-__-_

. ..___

-.-.--

.--......-.--._.._ ___--_._.____.--_._-_--_ Cobult .-.-_

-.-

.. . .

.._ ___.__.-__---_--_____

of 50 ml col~ht ~toclc solution by pliosl~lmtc nwtlwcl . 4 . . , . . Weight clcposit from 50 ml ?ianw solution . . . . . . . . . . . , . . . . Rcaidw.1 cobalt in spent clcctrotlc . , , . . . . . . , . . . . . . , . . Deposit should thcrcIorc!contain 9X.4 rng cobnlt (95.7 0.3) Cobalt iLCtllik1ly found in tlcposit by plloSpll~Ltc ;UXllySiS . . . a . . . . . . . -- .-._---_.--__---.-__ __-.-~.__-.-_--_._____.-~I_-___ --I----._.Amlysis

In this particular

case the deposit contained

about I.C,(g impurity.

fotrml

. . .

cJ8.7 tng . loo.1 mg . 0.3 mg

.

.

. , , . .

.

,

.

.

g8.G mg

ACCURATE

DETERMINATION

OFCO.

II.

31

As to the nature of the contamination, sulphur is almost certainly one substance present; in the absence of sulphur the positive en-or is lower. Some previous investigations have suggested that carbon is one of the contaminants. Some electrodepositions were made in the absence of both sulphur and carbon, the electrolyte containing NH&l, NH40H, and a little hydrazine hydrochloride. The results were still o.7% high. It is difficult to avoid the conclusions that the deposit is contaminated with oside or hyclroside. A rcccnt study of the electrodeposition of cobalt has been made by S~UYEH AND Swmxl~, who suggest, on the basis of another investigation 20, that the contamination might bc due to the osicle, hydroxide, or water. The hypothesis is considered that below a cobalt concentration of about o.005 N such compounds as CoS04.3Co(OH):!Hz0 are deposited, although their cspcrimental results do not support this theory in its simplest form. De~roxlsis AxI) P~~RBAI.~~~iii r’l recent stucly of the equilibrium-pi-1 study of the system Co-I&O have shown that in the absc~~cc of complexing substanccs, a deposit contaminated with oxide is possible above per 6.3. Knowledge of the system Co-NH a-H&I would bc required to decide whether such contamination was possible in ammoniacal solution. From the analytical standpoint, the results given by SALYISR ANL, SWEET agree with the present work, the average error in their case being -+~.a(;/,. In view of the contaminated cleposits obtained using the various procedures, the electrolytic method is rejected for the nccuratc determination of cobalt.

Of the gravimetric mcthocls cxaminccl for the accurate dctcrminntion of cobillt, only a modification of the phosphate method is capable of giving nccuratc results. ‘The electrolytic method gives very high results owing to a contaminated deposit.

WC arc indebted to the IMinistry of Education for a maintcnancc award (W. J.W.) and to the Central Research Fund of the University of London for the purchase of a Spckkcr nbsorptiometcr.

Some gravirnctric mcthocls for tlctcrmining cobalt hsvc been cxarnincd in orclcr to assess their value for the accurrrtc dctermiuation of the rnctzrl. ‘l’hc clcctrolvtic rncthod wus founcl to givr: high results (about I .4%). ;rnd the anthrariikrtc method slightly high results (0.2 - o.3’;/0). A rnodification of the phosphtrtc mcthotl has been dcvclopccl in which a former drawlxrck, the solubility of CoNl-I~P0~~l-l&, has been ovcrcomc using a rapid spcctrophotomctric dctcrmination of residual cobalt. The final proccduru dcvcloped has been found lo give rrccurato results (&o. I - 0.2%) and its use is rccommcndcd whcrc this dcgrcc of accumcy is rccluircd.

Quclqucs m&ho&s dc dosage gravim&riquc phate ;L dotim Its rdsultats lcs plus prdcis.

du cobalt

ont

&b cxaminbcs.

IAL mdthodc:

uu phos-

ZlJSAMlLIENJ~ASSUNG Ueschrcibung cincr kritischen Untcrsuchung libcr cinige gravimetrischc Mcthodcn zur Uestimmung van Iiobalt. Die bestcn Ergcbnisse wurdcn mit ciucr rnoclifizicrtcn Phosphatmcthoclc in Verbindung mit ciner spcktrophotomctrischen 13estimmun.g erziclt. Anal.

Clrivn. A&a,

24 (1,961)

20-32

32

1

A. G, POSTER,

1\.

G.

1’OSTICR

* 31.

ILINSKI

a

PUXK

tt.

AND AND

AND

J.

Iv.

G.

VON

i\-t.

\vJJ,LlAMS,

fi

1:.

0

I<.

7

I.

liOLMJ?S,

Js

t<.

I)ITT,

%.

M.

8 I,. l_)uxvy, w.

lo

11.

11

1’.

I(.

UJJfd.

1-t.

M’.

1:.

13.

I<.

S/ccl

ASIJ

Chw., !%.

“

t-t.

*n

1;.

Suwos,

‘”

13).

SALYliI<

2”

J. L.

2’

(1954) 353. 13. t’)rx.TOMUIc UJ/d

%.

IiHAlJS,

UJld.

AND Proc.

Z’ll.

538.

(IC#O)

AJJul. \\‘AI
CjliJn.

of

312.

15 (1956)

Acta,

Artalysf,

GI

(1936)

Qzrurlfit~ifive

395.

Inorgctriic

A

3rd

Judysis,

Ed.,

/tJrJ.

ChWJ.

IS.

t:.

I-.

44

Srcl.,

(1922)

54

(rgsl)

ISlsovicr,

.*I Jlalysis,

;

Applird

t.~:sI~lsI.L.

4”).

( I.,J’J~

2223.

707. i\nlstcrclanl.

I JJorguJJic

A

1953.

Jiutysis.

ant1 SUIIH,

\\‘ilcy

12.’

sWI’.I’-’

C’h’tJ/.

.%,C.,

S,?

(IC,jO)

120.

\.n” ( lCJ.45)

Z.tI.

63th , .*t Jld.

~~ACZKVWSKI’. t<.

52.

.%C.,

/JM/. C&m.

.I~lulysis. t<.

(1914)

8.

LL (lq_=g.: l510.

c’:.C’Jh

‘I‘. t<.

<;.

x.!

I’utlrJ~Jrl~i ASIJ

(1933)

22

I_+.

go

(rq+t) J.

:\f;NtlSS

ChJJl..

%.

ttAGl:IC,

fr’ijrelics,

31.

NJl
n&d. Ggg.

31.

~l’lfcrJJ~o~rufriJ~,rc~lric:

ASIJ

C/Ji?Jt. (1885)

Y’c.~fboo/c

JU~UJI.,

Ifi ‘1’. ti. thCK,

I\.

(I.UJI~OJJ),

ttALL,

New ScJrlc, trlq. 14 S. IirlLLnl.~w, .4Jrd. J.

ASD

Gg

S’uc.

t~tILLI’I3I~AsIl

In

C~JL’JJZ., 93 55 (1)3,1) I \V. I<. C:nInre!rN,

SASI>IxL,

IJM.

t).

C/IOJI.

IJror~uJJic~

t3ALI.

18

1~)s~.

ANIl

./.

Uev.,

SsII.r&I

/1ndysl,

\\‘ILI.AI
Al~‘rsuv,

AND

I:.

AND

I.ontlon, J, /mJl

ScIioIiI.LIsrc,

12 C. l>UVAL, 13

J. Il.

tc01,71101~1’

MacBlillsr~, ”

I’
G.

SHJINNAN,

Atld.

I
4 1.1. GOTO, J. Chcnz. Sot. Jafiun,

W. J. WILLIAMS

t’OUHUAIX,

of 6th hkcting.

hi., C‘kJJJ..

ASIJ

t I.

tJl!CJ’JlUfilJJJ~tl

I’oiticrs.

I.~mthm. 3”

(1958)

A.

t1HI(iIlr, ~OJilJJfi~tcT

195s.

1b32.

J.

Nuti.

iCrscwrcj1

(If

i~~C~CfVOC~ll~JJliCcrl.

Ilrrr.

Stajrdurds,

53

?‘~Jl~rJJlfldyJJc~Jllif.~

1cjfi.1. Atrcil.

C’hirrr.

Ackc,

‘r.+

(rgGl)

20-32