Electrochemical oxidation of cyanide ion at platinum electrodes

JOURNAL

OF

ELECTROCHEMICAL

ELECTROANALYTICAL

OXIDATION

PLATINUM DONALD Department

of Chenzistry,

CFJEMISTRE-

T. SAWYER U+riversiZy

(Received

OF

=95

CYANIDE

ION

AT

ELECTRODES AJSD

ROBERT

of CaZifonria. July

9th.

J_ D-XY

Riverside.

Calif.

( i7.S.A

_)

1961)

Two recent studiesi. 2 have shown that cyanide ion has a pronounced effect on the electrochemical reduction of o,vygen at platinum. To understand this phenomenon more fully, as well as the general problem of electrode “poisoning”, a detailed investigation of the electrochemical oxidation of cyanide ion at platinum electrodes has been undertaken. This seemed particularly desirable because the mechanisms proposed for the electrochemistry of osygen and hydrogen at platinum have involved a platinum oxide coatingi.a_ Although cyanide ion has been shown to be oxidized by numerous chemical reaction+, recent reports of electrochemical oxidations are limited to two studies5.6. In these the investigators conciuded that cyanide ion is oxidized to cyanogen, which undergoes further reactions to form a brown polvmer. SCH~KDT AND MEINERT” also concluded that in sufficiently basic solutions the formation of the polymer is inhibited, and cyanate ion is the product of the cyanogen decomposition_ They also give references to earlier works in which cyanate ion as well as ammonia and carbon diotide were found to be the ultimate products of the electrolysis of cyanide solutionsThe present study has been concerned with the effect of solution pH and electrode pre-conditioning upon the kinetics and mechannsm for the oxidation of cyanide ion at a platinum surface. The results of such variations have been investigated by voltammetry, chronopotentiometry and the galvanostatic method for electrode kinetics7_ EXPERniEKTAL

The electrochemical has been previously

studies were made with a versatile electronic instrument which described by DEFORDB_ This instrument, which is based on the operational amplifiers, permits voltamm etric, chronopotentiometric

use of Philbrick and galvanostatic studies by appropriate interconnection of the operational modules. Potentials were measured to an accuracy of + IO mV and currents to an accuracy of etric measurements were made in a modified H-cell to prevent o.I~/~. The voltamm Zk attack of the agar salt bridge by alkaline solutions as well as contamination of the calomel electrodeg. The cell for chronopotentiometric measurements was of conventional design and utilized a platinum foil working electrode and a platinum-gauze auxiliary electrode_ All potentials were measured relative to a saturated calemel electrode, unless otherwise indicated. J_ Ekctroanal.

Cirem.,

5 (1963) Ig5-zq

D_ T_ SAWER

196

AND

R.

Theelectrodesforvoltanunetricmeasurements platinum

wire into

soft glass tubing;

planar surface. Chronopotentiometric prepared from reagent grade platinum

J_ D_4Y

wereprepared.by sealing

the exposed

portion

was

groundflatto

IS gauge give a

measurements were made with electrodes foil, I cm2 on each side, welded to zz gauge

platinum wire sealed into soft glass tubing. The effective area of this electrode was determinedto bez_39 cm"byusingthechronopotentiometricreduction of potassium ferricyanide together with the known diffusion coefficient for fenicyanide ion, 0-77-10-5 cm" set-i (0-004 F KaFe(CN)a, 0.5 F KCl)iO,inthe Sand equationllpH measurements were made with a Beckman ModelG or a line operated Leeds andNorthrup pH meter equippedwith high range glass electrodes; the meters were standardized using N.BS_ buffers All measurements were made at ~5-0 + 0-1~ in a thermostattedbath. The chemicalswere reagent grade andwereusedwithoutfurtherpurification.The dis~ledwaterusedintheespe~mentswasredisti~edfromanall-glassstillcontaining abasicperrnanganatesolutionin distilledwater The presence of buffering agents eliminated the chronopotentiometric waves for cyanideion; thus,-thesemeasurementswere made in the absence ofbuffers with the pH adjusted by KOH or H&04_ The supporting electrolyte for all electrochemical measurements was o_r-o.5 F K&04_ The~neticparametersfortheoxldationofcya~deionweredete~nedusiTlgthe galvanostatic

method

as discussed

~~DEL~AY~_

RESULTS

AND

DISCUSSION

Voltamwzetry Cyanide ion between S and

gives a well defined II (Fig_ I)_ Although

anodic wave at the platinum electrode the diffusion current is proportionalto

for pH's cyanide

concentration at a given pH, the diffusion current changes with pH for a given total cyanide concentration_ However,ifthe cyanide ion concentrationis computed from the curve shown in Fig. z the dissociation constant for HCNi". PKHCN = 9.216,

J

Q-2

0.0

0.2

0.4

0.6

E vs. SCE, Fig. I_ Voltan&etic 5 -10--j F KCN and

0.1

o_xidation FX&Qa,

of cyanide ion-at a&l was adjusted

08

10

L2


a platinum electrode. The solution contained to pH xo_o,The szan rate was 0.3 V-per minute. _T_ Elecbroanab.

Ghem.,-5

-(x963)

_Igj-?o3

OXIDATION

OF CYANIDE

197

ION

results_ This behavior indicates that the rate of oxidation is significantly greater than the rate of dissociation for HCN; TANAKA AND MURAYAMAV have reported that the rate constant for HCN dissociation is I - 10-7 set-1 at 2~~. Thus,-these data lead t0 the conclusion that the electrochemical o,xidation of dissolved hydrogen cyanide only occurs for the free cyanide ion and that any kinetic expression should be concerned

with the free cyanide ion concentration

only-

5

pH

-

p H

9.5

*

pti

9.0

Fig. 2. The voltammetric &iffusion cm-rent as a function of cyanide different pR's. The cyanide ion concentration was calculated from which RCN in the solution and the dissociation constant for NC=.

10.0

ion

concentration

has

a p&Z,

for three

the formal concenWatioa of 9x16

of

at zs"-

-

0

5

I5

10

TIME,

Fig- 3- Chronopotentiomebic containedo.02 FKCN ando-

oxidation FK&O

i

20

bed

of cyanide ion ataplatinumfoil electrode.-The *,andwasadjustedtopH ro.o.Thecurrentw~~

solution 3-00 rnA_

Chronopotenticwnetry

The -high

chronopdtentiometric degree-

of irreversibility

chronopotentiogram

oxidation is shown

is for an untreated

wave by

for cyanide the

platinum

large

ion is shown

overpot&kial.

electrode. J- Electraanal.

However, C&em_.

in Fig.

.Tbis

3;

the

particular

pre-oxidation _s [x963)

~9~5~203

D_

198

T_ SAFVYER

AND

R.

J_ DAY

curve; a pre-reduced electrode gives a less well-definedwaveuntilithasbeenoxidizeddu~ngthecourseoftheanodicpulse_All

of the electrode gives virtually the same

of the remaining studies have been carried out using an untreated or pre-oxidized electrode. To establish the current Limiting factors for the oxidation of cyanide, a series of chronopotentiometric tionshavebeenmade.

oxidations at different currents, pH's For a diffusion controlledprocessthe

and cyanide Sandequation

concentra-

YL*?I FAD*

‘k*

-= C

(I)

2

indicatesthatiz&/C is a constant for a given electrode-cell combination at a given temperature_ Table I summarizes the values obtained for the cyanide oxidation; cyanideionconcentrationshavebeencalculatedfromtheHCNdissociationconstantl"_ Thegeneralconstancyofthe valuesindicatesthatthe current is diffusion controlled andthat

adsorptionis

notarate-controlfingfactorTABLE OF

V-4LUES

A_

0-02

F

ICCN.

0.5

F

ir4/C

FOR

THE

I

OXIDXTION

OF

IO.,

9-7

9-5 9-3 90 0.002

to

o.oz=j

F

ION

&SO4 IO.0

B_

CYXSIDE

KCN.

0.5

F

r-57-

10-I

0.84 0.82

I-38-10-I-24- 10-Y I-IO- 10-Z o-57 - 10-Z 0.62 - IO-?-

o-s3 ox,-

0.81 0-Q

&SO4 9-5

0.81

-+o.g1

Use of the calculated cyanideion concentrationsin obtaining the constant values further supports the conclusion that only the free cyanide ion is oxidized and that therateofHCNdissociationisslowrelativetotherateofoxidation_ The chronopotentiometric data for cyanide ion can also be used to establish the numberofelectronsreleasedinthe oxidation reaction_ By using an average value of 0_83foritrf/C and a value of+g4-ro-3forD?t (for CN-in 0.25 F NaCl and o-01 F in eqn. (I), the value of n is calculated to be 0.83. Thus the oxidation NaOHj14.15 reactionc;LI1beconcludedtobeaoneelectronstepandc~be~tten CN-

+

CN-

+

e-

(2)

withcyanideradicalscombmingtogivecyanogen. With this preliminary understanding of the system, a series of galvanostatic studies oftherate of cyanideion oxidation has been made. The potential of the working electrode vs_ the normal hydrogen electrode fdr a-constant current source hasbeenshownby DELAHAY~ tobeexpressedbytherelation

E=

-RT (I

-m)n,F

ln

nFCo

Mf,a io

-

(1 _:;nmFln

(3)

r?]

J_ ElecfroanaZ.

Chem,.

5 (1963)

Ig5--203

OXIDATION

OF

CYANIDE

ION

=99

where the terms have their usual electrochemical significance_ By of known current and following the potential as a function-of time

appQi.ug a pulse with a recorder it

is. possible, using an extrapolation to time equal to zero, to cause the last term in eqn. (3) to go to zero With this operation applied at several current densities the anodic transfer coefficient, (I: - DC),can be evaluated from the slope of a plot of E vs.

log io_ The

heterogeneous forward rate constant (relative to the normal hydrogen electrode), ko~m~,can then be evaluated by applying eqn. (3)_ To evaluate (I - OL), the value of ?ta must be known; it is assumed to be one from the arguments given for eqn.

(2). Figure

4 shows

the extrapolated

values

of E US. log io for a series of different

,LDPE.

--

..w-bg

A-0025F e - 0.010F

330 330

c-

310

O.ooBF

0- O.OC4F E-O.M33F F - 0.001 F

r

200 230 170

Fig.+Theextiapolatedpoteniials, El-0 as a function ofcurrentforthegalvanostatico_tidation of cyanide ion at a platinum foil electrode_ The several ICCN concentrations studied are indicated as wella~the slopes for their respective curves_ AI1 of the solutions contained 0.5 F XC&OS and were adjusted to pH g-5_

concentrations of KCN at pH g-5_ The slopes for the curves, in mV/log Z, are given on the figure. Normally the slopes for all of the curves would be expected to be the same; however, for this system there is a trend of increasing slope with increasing KCN concentration_ Table II tabulates values of (I - a~) and k0f.h for various pH”s and KCN concentrations. The free cyanide ion concentration calculated from the dissociation constant for HCN, has been used in eqn_ (3) to determine the values of R-O~,h_ TABLE SU34ICLaRY

OF CYANIDE

IiINETIC IOh’

P_GxA?.~ETERS AT

A

II EV-AI_UATED

PRI-OXIDIZED

FOR

PLATINURZ

OSIDATIOX

THE

o-35

0-g.

9-5

O-26

1.6-10-m

O-30 0.19

0.1.10-a

O-18

=j_o.10-8 2-0 -10-B I-6 - 10-a 0.8 - 10-s 2.0 - 10-a 2.x - IO-B

9-5

5-o.IO-3

0.22 0.X8

IO-5

o-26

TO.5

O-20

-

10-B

95

9-5 9-5 9-5 9-5

OF

ELECTRODE

Average

J_ Electroanal.

Cirern..

5 (1963)

xgy-203

3-00

The comes

D--II_

SAmRAND

R. J_ DAY

potential required to o_xidize cyanide ion beconcentrationincreases,whichis asurprising

datain Fig. 4indicatethatthe more positive asthecyanideion

and unexpected phenomenon. A number theseobservationstothe extentthatthey mentaJerror.Figure5 matelylinearcurve.

shows The

aplotof

of additional experiments have arebelievedvalidandnotdueto

(1 --,)zIs_ -log(CN-)whichgives

equationforthis (I - cx)=

confirmed esperian approxi-

curveis

-0.06

- 0.123 log(CXp)

(4)

1

c

-0.21 0

1

2 -

log

3

4

[CN-]

Fig. 5_ The anodic transfer coefficient, (I -a), as a function of cyanide ion concentration-The values of (I -n) were determined b-4_ the galvanostatic method and the cyanide ion concentrations were determined from the formal concentrations of ECCIS and the dissociation constantfor HCN. The slope of the curve is -0_1?3 andithas aninterceptof -0.06.

andindicatesthat (I -a) would be negative for a I M cyanide ion concentration_ Equation (4)indicates that the higher the cyanide ion concentration is, the smaller is the fraction of the applied potential that is effective in accelerating the rate of electrochemical oxidation- Such behavior can be thought of as a poisoning of the platinum electrode surface bq- cyanideion itself,e.g., Pt(OH)p

+

zCN-

+Pt(CN)z+zOH-

(5)

Thus if Pt(CN)z (or Pt(CN)az- ) is catalytically much poorer than Pt(OH)z for the oxidation of cyanide ion, increasing concentrations of cyanide ion would cause the reactionineqn.(5)togototheIlghtandbringaboutadecreasein(I -~~)asindicated byeqn (4)_Theobservedchangesof(1 - a)alsomightbeexplainedbyconsiderations similar to those presented by TONDE~R, DOMBE-T AND GIERST on the effects of changes in the double layer r6_ Thus,increasedconcentrations of cyanideion would bring about increased adsorption of CNor Pt(CN)a"-- on the electrode and cause attendantchangesinthedoublelayer. Thiswouldthenaffectthekineticsofoxidation for cyanideionin amannersuch as has been observed. These phenomena are sufficientlyinterestingto warrantadditionalstudyintheiirt-me to estabhshjustwhatis bringing

about the apparent

change in-(I -a)

with

changes

of cyanide

ion concen-

tration_ Equation(3) indicatesthataplotof E ZK_log (d-t%) shouldgive alinearcurve from-whichtheslope also shouldpermitevaluationof (1 -cx)_~Anmnberofchronopotentiograrns similar to Fig. 3 have been plotted andin all cases curves similar to J_ EZectroanaZ-

Chem..

5 (1963)

kg5--203

OXTDATIO?Z

OF

CY_mIDE

ION

201

the one shown in Fig_ 6 have resulted_ The upper portion of this has a slope approaching mV per log unit, but the lower straight-line portion only has a slope of 73 mV/log (& - &) unit. This behavior indicates that initially (I: - a) is small, but during the

300

course of the chronopotentiogram its value becomes more effective catalytically_

increases

markedly;

that

is, the electrode

1.4

E i

x

IfI

‘1:\‘,:

e w

1.0 a01

002

004

QOEI

a406

1

E

4

6

8 -10

Fig. 6. Voltage-time analq-sis of the curve for the chronopotentiometic oxidation at a platinum foil electrode_ The upper portion of the curve has a slope appI’oximating per log unit, while the contained 0.02 F KCN

lower. straight-line and 0.5 F K&O*

portion and was

has a slope of 73 mV adjusted to pH 10.5~

per The

of cyanide ion

300-400 rnv log unit- The solution current was 3.00 m-4.

CONCLUSIONS

Although (I - n) changes with cyanide ion concentration, k,*f,ti is independent of both pH and cyanide ion concentration_ This observation plus the preceding experimental data can be accounted for by the foilowing proposed reaction mechanisms. l?t(0H)~

+

Pt(OH)&X)-

CN-

--+ l?t(OH)r(CN)-

Pt(OH)z

2 (CN)-

-+

+

(CN)-

+

c-

(CN)?

As the cyanide concentration increases the Pt(OH) 2 is metathesized Pt(CN)a”-) as indicated by eqn. (5)_ The Pt(CNj2 surface is postulated mechanistically and to inhibit the oxidation of cyanide ion. An alternative mechanism involves the formation of PtOz Pt(OH)a Pt03

f

?- CN-

f-

?- OH-

+

3 H-0

4 +

PtOE

+

Pt(OH)z

z (CN) - +

3- Hz0 +

f

3 OH-

(6)

slow

(7)

fast

w

to Pt(CN)B (or to be ineffective

fast

z e+

fast

2 (CX)-

(CN)s

(9)

slow

IlO)

fast

(+I)

Metathesis of Pt(OH)z by CNmay lead to a compound, Pt(CN)z, that cannot be oxidized to PtO,_ However, this would tend to decrease the amount of PtOz available for the reaction shown by eqn. (IO) _ Because i&/C is a constant this seems unlikely; the mechanism expressed by eqns. (6), (71, (8) is thought to be more consistent with the data. Efforts to isolate oxidation products for the oxidation of cyanide ion have resulted in the detection of ammonia. This product can arise from the basic hydrolysis of and to ammonia cyanogen, first to cyanate and cyanide 17 and then of the cyanate carbon dioxidel8_ Detection of oxidation products by reverse chronopotentiometry has been totally unsuccessful. Electrochemical reduction of dissolved cyanogen gas and of cvanate ion has been unsuccessful also. J. EEectvoamzl.

Chem_,

5 (1963)

xg5-zo3

De 1.

202

SAIWY-ER

AND

R.

J_ DAY

The need for a platinum oxide film to catalyze the-oxidation of cyanide ion is somewhatslmilartotheconditionsnecessaryforthereduction ofdissolvedoxygenl. As shown in the latter study,a platinum oxide fihn is necessaryto get a rapid and efficient reduction of dissolved oxygen Cyanide ion causes a virtual elimination of the oxygen reductionwave,which can be accounted forifthereis conversion of the platinumoxideto platinum cyanide-The presentstudyindicatesthatsuch aphenomenon probably takes place andleadsto an electrode thatalsois ineffective for the o~dationofcyarudeioIl_Thusplatinumoxidea~pearstobeahighlyimportantcatalytic material in electrochemical reactions_ This has been seen to be the case for the electrolyticreduction of dissolved osygenr-2, the electrolyticoxidation of dissolved hydrogen3 and now, the electrolytico_xidation of cyanide ion_ However, for the last system,theveryion *hat is beingosidized also is poisoning the eIectrode forits own oxidation_ The decrease of the transfer coefficientwith increasing concentrations of the active species, as observedhere,is believedto be somewhat unique. ’ Although the proposed mechanisms are quite tentative, and certainly do not and do preclude other possibilities,they do account for the observed phenomena reasonably explain the rather surprising phenomena of having a higher oxidation potential with increased cyanide concentrations_ Without the change in slope of the curves in Fig_ 4 and the attendant change in the values for the transfer coefficient, extremely diverse results would havebeen obtained for the rate constant. ACKKOWLEDGEMENT

This work was supported by the United States Air Force, Geophysics Research Directorate, Air Force Carnbridge Research Laboratories, under Contract No. AF rg@o&8347Voltammetric, c~o~opoten~omet~c and galvanostatic studies have been used to investigate the electrochemical oxidation of cyanideion at platinum electrodes-The reaction is a diffusion controlled process involving a one-electron oxidation of free cyanideionto cyanogen.The anodic transfer coefficient,(I -a), for apre-oxidized electrodein apHg_5solutionvariesfrom0_35 forro-3 F KCNto o.rBfor0.25~FKCNThe average value for the forward, heterogeneous rate constant, k"f,n,is2.1-10-~ cm set-l (using potentials relative to the normal hydrogen electrode). The proposed mechanism for the o.xidation reaction involves a platinum oxide film. Increasing cy~ideionconcentrationsasepostulatedtocausemetathesisoftheoxidetoplatinum cyanide, which brings about a decrease in (I -5~) and inhibition of the oxidation reactionREFERENCES 1 D. 2 I).

“I-. SAWTFER

INTTERRXX~E,J_ Ef~~~roanaE. Chem_. 2 (1961) 310. AKD E_T.SEO. ibid_, 3 (1962) 410.. AXD E.T..SEO. +bid_,5 (1963) 23_ F_ PEEUONE. Anti_ Chim_ _dppiicata. 18 (191s) 550~ AXD

L. V_

T_ SAVWER 3 D-T. SAWYER .i Is.. R_Icci AND 5 J_ S. FITSGERU, Chem. 6 1~~2.; (1955) ~7. e-H_ SCEKBIIDT AND H_ MEINERT,Z_ _4norg_AIZgenr_ Chem., 293 (x957) z14_ ~7 P. DELMXAY, LV~W Inslirumental &l&hods Ga EZeedrocllemisiry. Interscience Publishers, Inc.. NewYork,xg54.p_ 186. 8 D. D_ DEFORD. Private Cornrnunication. presented at the r33rd -American Chemic&Society Meeiing~ San Francisco,Calif-,ApriZ, xg5S

OXIDATION

OF

CYANIDE

TON

203

B R. L_ PECSOK AND R. S_ JWET, JR., Anab. Chent., ~7 (1955) 16.5~ lo M_ voh- STACKELEERG, M. PILGRAM. AV_ TOOME, Z_ EZek~rocliem.. 57 (x953) -34~~ 11 H. J_ S_ SAND, P&Z_ Mug.. I (xgor) 45_ 11 K. P_ ANG, J_ Chews. Sot_, (xq5q) 3Sz2_ 1s N. TANAKA AXD T. MuRAYAMA,Z. Pky.srk.~Chet?z. (Frankfurt), 15 (1959) 146. 14 L. MEITE~S, PoZarogruphiC Techniques. Interscience Publishers Inc., New York, 19%. 15 W. H. JURA, Anail. Chenz.,-26 (1954) 1121. 16 J_ J. TONDEUR, -4. DOXSFZET, AND I.. GIEFCST.,J_ EZectroanaZ. Ckem., 3 (1961) =5Publishers 17 I_ M. KOLTOFF AMD J_ J_ LINGKXE. Polarogvaphy, znd ed.. Interscience York, rqsz, p- 5+xX8 M. W. LISTER, Cam_ J_ CRem_, 33 (rqg5) 426J_

EZecfroanaZ.

CkemL,

5 (x963)

p. 256. Inc.,

New

195-203