Electroreduction of carbon dioxide

Electroreduction of carbon dioxide

J Elec1ru3nal Ekevxr Oem. Sequoia SA, 189 (1985) busanne ELECl-ROREDUCI’ION Pnnted V S BAGOTZKY. 26th November DIOXIDE REDUCI-ION AND N ...

817KB Sizes 2 Downloads 41 Views

J

Elec1ru3nal

Ekevxr

Oem.

Sequoia

SA,

189 (1985) busanne

ELECl-ROREDUCI’ION

Pnnted

V S BAGOTZKY.

26th November

DIOXIDE

REDUCI-ION

AND

N V OSETROVA

A N Frumkrn Irutttore of Electrochernutry. (Relved

m The Netherlands

OF CARBON

PART III. ADSORPTION METALS

Yu B VASSILIEV,

311

311-324 -

Academy

OF CO, ON PLATlNUM

and A A

MIKHAILOVA

of Screncer of the USSR

Moscow

(US

S R)

1984)

ABSTEUCT

The mam prcperttes of reduc:lonal adsorpllon of CO, on the plalmum n~e~als are stud& Chenusorbed parhcles are found to be produced only on plalmum and rhodnm-, Ueclroreducllon ol CO, on these merals prcwx-eds as a rcsulr 01 the m~eractmn

of Cq

molgulcs

actwated

III the cow

of adsorptmn

on

the metal surface wxth chenusorbed

hydrogen As a result, strongly the-rbed partxles are obtamed on the surface whxh are the products 01 more profound reducuon of CO, than LO rorrmc acid The further reduction of these chemusorbed parucles, accompamed by &em desorpbon mto the soluuon. IS very slow due to very strong couplmg of sorbed pticles wrth the sufaw and 10 very fasl backward adsorption of the reduchoo products Neither reduc?lon,al cheuusorphon of CO, nor mterachons of CO2 WILII adsorbed hydrogen were observed for mdlum, palladmm. osmnun or ruthemum

INT’RODUCTION Ln Parts I and II of the present work [l] the mechanism and kmetics of CO, electroreduchon in aqueous and aprotic solutions on metals wth lugh and moderate hydrogen overvoltages were Investigated. Gmer [2] was the frost to drscover that COz mteracts ant_h hydrogen adsorbed on a platmum electrode, producmg chemisorbed particles which are oxuhzed at the same potentials as the particles formed on platinum durmg adsorption of some organic compounds (such as methanol, formaldehyde, formic acid). Later on, the adsorption of CO, on platmum electrodes was studied in many works [3-121 and chernisorbed partxles were shown to be ttghtly bound to the surface and not to be desorbed by nnsing Using analytrcal methods, Johnson and Kuhn [12] detected no desorbed products such as CO, formaldehyde or forrmc acid m the soluuon. The chernisorbed particles remam stabie at any negative polanzatron [&IO], Sting removed only at posrtrve polanzatton Although tlus problem has been mvestigated u-r many works, the mechanism of adsorption and cathodic ieduchon of CO, and the composthon of the particle adsorbed on plattnum are not clear up to now. 0022-0728/85/$03

3iI

Q 1985 Elsener

Sequoia S A

312

The present work JS devoted to studies platmum metsls in aqueous solutions

of CO1 electroreductlon

on a :enes

of

EXPERIMENTAL

In order to study the mechamsm of CO, electroreducuon on platmum metals n-t aqueous solutrons the followmg methods were apphed. polanzation measurements [l], the method of pulse voltammetry of non-stationary current [13,14], the method of potentral shrfts at open circurt [13,14], and the method of radloactive mdtcators [15] Smce a lot of expenmental data on the mteractron of CO, with hydrogen adsorbed on platmum are avaJable now, attentron was pad mamly to the poorly rnvesugated aspects, VIZ to effects of pH, temperature and of the nature of the electrode on thrs reactton. Adsorptron measurements were done wrth the use of an oscillograpluc potenttostat PO-5122. In these measurements the decrease of the amount of adsorbed hydrogen dunng the apphcatton of a fast negative pulse and the charge spared for the oxldatlon of cherrusorbed particles were regtstered [13,14] The studies were can-ted out with smooth electrodes (platmum, u-r&urn, rhodtum, palladmm, ruthemum, osnuum) as well as wrth electrodes with htgh surface areas (platimzed platmum, rhodlzed rhodtum, uidrzed u-tdmm, palladlzed palladmm). The programs of prehnunary and worhng pulses for smooth electrodes were prepared using the data of refs 16-21. For all electrodes except platmum, counterelectrodes were made of gold fotl In the maJor part of the expenments the buffer solutions had the same loruc strength wth composittcn 0 1 M H,PO, + 0.35 M H,SO, and 1 M KOH The pH of the worhng solutrons was fmally estabhshed after prehmmary bubblmg and scturatron wtth carbon dioxrde The CO2 bubbhng dunng measurements &d not change the pH Srnce the pK of carboruc acrd IS 6 35 and CO2 Interacts chermcally wrth KOH at pH 3 6.3, mstead of CO, bubbhng K&O, was mtroduced m alkali soluttons m amounts equivalent to the solublhty of CO, m acrd soluhons, so that the total concentrauon of CO,, HCO,, and CO,‘was the same for every soluhon (2x10-’ M) Wh en the activauon energy was measured, the changes m CO, solublhty w-tth temperature were taken mto account The mvesugatrons worth labeled atoms were done by the method of ref 15, usmg Nn$O, labeled worth 14C, the total concentratron of CO, m the soluuon (10s4 to 2 X 10m3 M) bemg much less than m a CO,-saturated solutton at 1 atm After the surface coverage by chemisorbed particles achieved a statronary level, the cell was bubbled wtth an inert gas to remove Ca_ at a controlled value of potentml mthout dtsconnecting the circuit The number (n) of electrons used for the oxidatron of carbon-contaming parucles was defined from the adsorption of a r4C-labeled orgamc compound on the electrode_ During electroreduction of chernisorbed particles the number N of carbon-containtng parttcles resident on the electrode was measured versus the charge Q passed via the arcutt The slope dQ/dN corresponds to the value of n The

313

absolute amount of carbon-containmg particles was defmed by calrbr&on against a standard [22,23] The interaction of CO, with hydrogen prevtously adsorbed on an electrode was studied using the curve of potential decay at open ctrcuit while stmultaneously measuring the surface coverage by adsorbed hydrogen and by chen-usorbed carbon products (thrs tecluuque IS described m ref 14). Pt/Pt electrodes were prepared by electrodeposrtron at E,dq = 0 2 V as described m ref. 24 After agmg the electrode was subJected to cathodic-anodic activation at potentials of O-l 2 V. Palladium was deposrtea from a 3% solutron of PdCl, at Ed- = 0.2 V. After bemg cleaned of chlorine ions and aged and before 0 5 V are these particles oxidized ~th formauon of CO, In Fig 1 the results of measurements by the method of labeled atoms are presented Noticeable adsorption of water appears at potentlals m the range of hydrogen adsorption on platmum, at 0.1 V the adsorptton aclueves a stationary level and rt stays constant at more negative potentials, i e tile amount of chenusorbed par‘icles neither grows due to mcreasmg surface coverage by hydrogen nor decreases by further hydrogenation of chermsorbed particles mto products released mto solution The dependence of the Pt/Pt electrode surface coverage by chemrsorbed particles (curve 1) on potential is stationary at a given value of CO2 concentration, smce the values of coverage may be defined independently from the kmettcs of adsorpuon on a pure surface at every value of potential

-2

01

02

a3

04

Fig 1 The dependence of adsorpuon of chermsorbed par~cles on a plaImuxl platinum electrode m Cl5 M solution of HISO_, In the presence of 2X lo-’ M CO2 (l, 2) and after CO; is blown off (3) on potentA vaned as shown by arrows

When changmg the potentral from E, = 0 V to E, = 0.4 V in an electrolyte with dissolved CO, (curve 2), the surface coverage stays constant; rt decreases drastically only at E, > 0 4 V due to electrooxldatton of cherrusorbed partrcles So, a hysteresis exrsts m the I’-E dependence If CO, IS blown off the solution by an inert gas at E, = O-O.1 V and after that the potenttal IS increased, the value cf adsorptton is also constant unttl E, = 0.4 V and then rapidly decreases due to oxidation of chenusorbed parhcles (curve 3) [26] The chemisorbed part1c1e.s cannot be removed from the surface rf the electrode IS subjected to long term cathodic polarization (1 h at E, = -0.05 v) after achieving stationary coverage at E, = 0 -0.1 V and after removmg CO, from the solutron These parhcles desorb only by anodic oxtdation at E, > 0.4 V A smular #k-E, dependence was observed for a smooth platmum electrode [27,28] At a temperature of 83°C the dependence of adsorption on potentml has a different shape wxth a maxrmum at E, = 0 1 V and mth a drop near E, = 0 V (Fig. 2) Tlus decrease could not be connected wrth hydrogenation of chenusorbed

I 02

04

O6 &JY

Rg 2 The dependence of stahonaq surface coverage of a smooth platrnum elecuode by chemrsorbed o;gamc particles m CO+aturakd G 5 M soluuon of H,SO, at p =I aim ou potentlal for &ffercnt temperatures (1) 25, (2) 40, (3) 83T

315

Kg 3. The d-endence of stauonarj surface coverage of a smcuxh pla~~~rn electrode orgaruc parkles al E, = 0 1 V m 0 5 M solution of H,SO, on CO2 partA pressure

by cherrusorbed

particles to products released mto solution According to the experimental data, however, no hyd-ogenation of chermsorbed parucles was observed at this temperature following long-term cathodic polarizatron after adsorption at E, = 0.1 V both in the presence of CO, aid after blowmg CO2 off the solution. At the potential OF maxunum adsorptron (Ftg 2) the surface coverage of platmum by tightly chemisorbed particles mcreases with the growth of the CO2 volume concentration @artial pressure), this dependence betng described by the Temkm isotherm (Fig 3)

&=u+(l/f)

0)

lnc

The Roginskr-Zeldovtch

equatron

v,, = aek/aT= k&o,

exp( -

(Fig

4)

af9;)

describes the adsorption kmetics for smooth and platinized platmum electrodes any values of croncentrauon, pH, potential and temperature (see Frg 5)

(2) at

316

83 Fig 5 K~ncuc ~cnhcrrns of CO2 adsovtmn m 0 5 M soluuon of H,SO, ( ,J,,,~ = 1 atm) at 25°C (curves 1.3) and 85°C (curves 2,4) (1,2) Rho&urn elecrrcxk, E, = 0 02 V. (3.4) platmum electrode, E, = 0 1 V. hg 6 The dependence of Lhelogan~hm of the ~rut~alrate of CO1 adsorptIon on plaunum on potential for dIfferen solutions (1) 0 5 M H2S0, at 20°C, (2) buffer solution wth pH = 5 1. at 25°C. (3) 05 M H,SO, at 40°C. data of ref 10

The adsorption rate-constant k.Ed, IS a complicated potenttal in 0.5 M HISO, it has a m-mum at E, = 0 rate-constant for a smooth platmum electrode IS two than that for Pt/Pt The actrvauon energy of CO, electrode grows approximately hnearly with the surface G,,, = Go +

af

function of the electrode 1 V (Frg 6) The adsorption orders of magmtude lugher adsorption on a platinum coverage, i e.

‘RTek

where G,( E, = 0 1 V) = 31 kl/mol and of’ = 8 3. Thus, the reductional adsorption of CO2 on platmum electrodes IS very sirrular to that of other compounds [13,14]. THE EFFECf OF pH ON THE PROCESSOF REDUCTIONAL ADSORl’TION As seen m Fig 7, the @k-E, dependence changes very little when the soluhon pH IS vartecl, i e. in the scale of potentral E (measured wrth respect to the reference electrode) this dependence shifts in almost the same way as the range of hydrogen adsorptton does At E, = 0.05-O2 V the value of stattonary coverage decreases very weakly when the solution pH changes from 0.4 to 6, but at pH > 6.0 the coverage drops down to zero In buffer solutions wrth pH > 8, the particles dunimshing hydrogen adsorption and capable of oxrdahon under application of a positive pulse were found not to be adsorbed LI-I the presence of either CO, and KzCO, or carbon-labeled K &OJ The expenments on a Pt/Pt electrode with labeled CO2 mdicated that the surface coverage by chenusorbed particles in phosphate buffer solution wrth pH = 5.4 in the presence of CO, also grows (as m acid) as the potential IS changd from 0.3 to 0.1 V but the maxrmurc value is 30% smaller than that for an actd solution However, in

317

,-.-G

.7---W 3-0-13 ,-.-Al 5-o

-16

05 -Ek Fig 7 The dependence of stauonaq surface coverage of a smmth platinum elccwode bj chenusorbed orgamc parucles on potentA tn phosphate buffer soluuons contarrung2X10-’ M CO, at drfferent valuesofpH 04.10.13.30.56

contrast to acid solutrons, when the potenttal is decreased down fro-n E, = 0.05 V the surface coverage increases by about 20% of p at E,“& = -0-l V (albett tis mcrease is not eastly reproductble). In thts case the adsorbed partrcles are not tightly

stuck to the surface and can be removed by biowmg off CO, The expenments w=tth 0 1 M KOH soluuons 111 the presence of Na214C03 m concentratrons up to lo-’ M demonstrated that no carbon-contaniug chernisorbed partxles

are accumulated

on a Pt/Pt

V. To clarify the cause solution at E, = 0 V Then CO, was blown off the solution, the electrode was rinsed tn doubly dlstJled water wA hydrogen flow, and the solutron was replaced by 0.1 M

of tlus effect, COz was chenusorbed

electrode at E = O-O.8

on a Pt/Pt

electrode in 0.5 M H,SO,

KOH soluhon The chemtsorbed parttcles proved ED stay on the electrode surface. Vanatron of potentral and a long-term stay at E, == -0 1 to - 0.3 V also drd not decrease the adsorpuon. The chen-usorbed partrcles can be removed from the electrode only by vu-tue of electrooxldation at E, > 0 3 V Thus, only CO, produces chemisorbed organic particles whrle HCO, and CO:anions exhlblt no reductronal chenusorptron on a plaumun electrode However, the particles adsorbed in an acid solution can also extst tn an alkali solutton tn a wide range of potentrals. thetr electrochemrcal behavior being sinnlar to that in actd solutrons The kinetics of CO, adsorptron was measured on a smooth electrode m solutions

with vanous pH As seen tn Frg 4. the kinetrc isotherms of CO, adsorption at E, = 0.05 V in buffer solutions at pH from 0.4 to 6.3 almost comcrde These results show that the process of CO, adsorption is affected mainly by the surface coverage of adsorbed hydrogen, which is approxunately constant at E, = constant and at different pH, rather than by the potenhal E, ill the curves being different m the latter case. Therefore, the rate-determuting step of CO, adsorption on platinum 1s the interaction 01’ Cq with adsorbed Lydrogen, this being m agreement ~th the results of [2,3-5,8].

318 CATALYTIC PLATINUM

INTERACIION

OF

CO,

WITH

HYDROGEN

PREVIOUSLY

ADSORBED

ON

The catalytrc interaction of CO2 vnth hydrogen previously adsorbed on the surface of a Pt/Pt electrode was studred by the method measunng of the shaft of potential curves III open circurt urlth simultaneous measurements of changes m surface coverage by adsorbed hydrogen and chenusorbed parucles [14] Since the value of the plattnum electrode potential III an open crrcult is defined by adsorbed hydrogen (even at consrderable coverage by orgamc compounds), the decrease of the surface coverage by hydrogen resultmg from tnteraction with CO2 would give nse to a shrft of potential Hence, the rate of vanatton of the potenttal g.tvcs information on the rate of removal of adsorbed hydrogen, i e on the rate of us mtelactton with CO1 Expenments have shown that the rate of mteraction decreases exponentially versus accumulatron of ttghtly chernisorbed particles on the surface (see Frg 8) The nutral rate of CO, adsorpuon on a Pt/Pt catalyst m the open clrcurt is a cornphcated funchon of the nuual hydrogen coverage (or of mttral potentral) wrth a m-mum at t9h = 0 5-O 7 (or at E, = 0 1 V) (see Fig 9). The uuttal hydrogen coverage and the inittal potenual both being constant, the rate of CO, chenusorptton does not depend on pH at 0 4 -Z pH < 6 (see Frg 10). A certain decrease of the adsorptton rate at pH = 3-5 IS probably due to the effect of adsorptron of phosphone acrd amons on CO, adsorption So, it may be concluded that the rate of CO,

Fig 8 Vanauons of surface coverage by chenusorbed versus the clme of antact of tie electrode wlh CO, fi2,2’) 0 1 v, (3.3’) 0 15 v. (4.4’) 0 25 v

parkles (l-4) and by adsorbed hydrogen (1’-4’) at dlfferent values of nut~al potential (1.1’) 0 0 V.

FIN 9 The dependence of Lhe rate of polenkxl vanatlon (mV/mm). (1) AE ( = E. - E,,, ) = 0 0 V. (2) AE = 0 025 V. and of the CO2 adsorptIon rate [email protected],/Al (m e/s)- (3) 0& = 0 0. (4) &Jk= 025 on Lhe wuual value of lhe adsorpt on polegLal (E,,, ) or on the ~rut~al sufiace coverage by adsorbed hydrogen (OF’ )

319

FIN 10 (A) The curves of the sluf! of the potenrlal of a platmum electrode at E,,,, = 0 15 V aIter II IS pul mto con~ct wlh CO, in phosphate buffer solunons wLh pH = 0 4 (curve 1). 0 6 (2). 2 0 (3). 3 0 (4). 4 0 (5). 6 0 (6) (B) The dependence of the marumum shift of polenllals AE,, = E - E,,, on the soluuon PH

adsorption on platinum IS not directly affected by the electrode potential but IS affected by the hydrogen surface coverage, wl-uch 1s connected both with potential acd with pH THE

PRODUCTS

OF REDUCTION

ON PLATINUM

In expenments on CO2 electroreductlon on platinum, Johnson and Kuhn [12] have detected no products hke CO, formaldehyde or formic acid m the solution. According to our expenmental data the number of chermsorbed particles produced a’lways corresponds to the amount of adsorbed hydrogen involved m the reaction, I e no reaction products were detected whch are released mto the solution. This effect can be shown by the equahty:

AQH + Qeds= Pa where AQH 1s the decrease in the amount of hydrogen due to CO, adsorption at the gven potential, Qaas 1s the charge passed through the electrode during adsorption and Q. 1s the charge released dunng total oxldatlon of all chemisorbed particles to CO, Therefore, the CO, adsorption on platmum occurs through the formatton of tightly chenusorbed particles, no product berg released mto the solution The chermsorbed part~les remam stable at a strong negative polanzation (cf. refs. 8 and lo), they can only be removed by anodic oxldatlon. The study of the stolcluometnc composition of the chermsorbed particles by various electrochermcal methods [13,14] and by the method of labeled atoms leads to the conclusion that the nature of those partuzles stays the same under all condltlons utvolved. that it takes about 3 electrons to oxidize such particles m CO2 U-I0 5 M H,SO, solution and that such a particle occupies about 3 adsorption sites, therefore its composltron is ICOH (se-e Table 1).

TABLE

1

The data on the naiure of the parldes CO, III 0 5 M H,SO, E,/V 103 c/M G./mm ok= AQ,/Q;

01 1 67 [email protected] 0 36 0 36 1 16

Pm/Q: n ,(electr /sue) lo-l4 N/parhclc 10’9 q ( =

10”

10’ Q/C cm-’ “3( =

131x10’~

n2/eleclr

cm-’ QJC

0 15 23 150 0 52 0 51 0 98 21 5 17

)

.N.

chermxtrbed on a

75

10 8

plat~ruzedplaunum eicctrode from a solul~on of 0 15 1 67 70 0 31 0 27 0 88 12

2 93

3 23

0 25 37 75 0 2-o 22 0 18 0 83 08

48 5 75

8’ R,

N (parude)-’

02 1 67 15 02-022 0 17 0 117-o 91 0.76

30

37

3 85

3 42-3 56

2 27-3 6

31

30

Thus, the products of CO2 adsorphon on platmum are srmrlar to those of methanol adsorption Our results also Indicate that these products doffer from those produced dunng CO adsorptron on platinum THE

EFFECT

OF THE CATALYST

MATERIAL

When the CO2 adsorptron on rndrum and palladium [29] (the metals most smular m theu physIcal and cherrucal properties and m adsorptton of hydrogen) was studred, farrly unexpected results were obtamed, VIZ the cherrusorption was detected nerther at room temperature nor at lugher temperatures The measurements w-tth labeled atoms on a paIladmm electrode [30] mdrcated that at potentrals from 0 to 0 8 V no accumulation of carbon-containing particles ouzurs on the electrode surface. Hence, no mteractton of CO1 wtth adsorbed and absorbed hydrogen takes place on palladium and u-rdium, and no ttgbtly chenusorbed carbon-contaming partrcles are produced in this case. But accordmg to the data on me~thanol [19,31] and formaIdehyde [32] on palladmm and n-rdtum, such particles can exist on palladium in a wde range of potenttals, provrded they were produced during adsorption under dehydrogenation of organic molecules The absence of chemisorbed particles on the surface 1s probably caused by CO, not mteractmg with adsorbed hydrogen If, on the other hand, such an mteractron does take place, Its products are not accumulated on the surface but are released into the solutton In measurements with Pd/Pd and Ir/Ir electrodes rt was shown that rf CO, IS mtroduced into the solutron at Eim'= 0.05V m the case of an open ctrcurt, none of the following effects take place potential shifts, adsorbed hydrogen expenditure, accumulation of carbon-contammg chenusorbed particles. If the cucuit IS connected, no stationary current of electroreduction 1s observed until hydrogen starts evolvmg Thus, there 1s no interaction of CO, ~th hydrogen adsorbed on

321

palladmm and indmm which could lead to the formation of products either adsorbed at the surkace or released mto the solution. On the other hand, rhodmm, whose hydrogen adsorptron spectrum differs strongly from that of platmum, exhrbits CO? chenusorphon at room and higher temperatures at potentrals in the range of hydrogen adsorptron (Frg 11) It should be noted thdt the hysteresis IS more pronounced for rhodrum than for platmum The statronary T-E, dependence for rhodrzed rhodmm agrees well worth the data for a smooth electrode [29]. The disagreement with ref 33 IS mamly connected wrth the different method of Q”, calculatrons used therem. The kmetics of CO? adsorptron on smooth and rhodrzed rhodium electrodes can, u-r the first approxrmatron, be described well by the Rogmsku-Zeldovrch equation (see Frg. 5) but the rate of CO2 adsorption on rhodmm IS much smaller than on platinum. In addrtron, the results for rhodium can hardly be quantrtatrvely treated since the CO? adsorptron is weak and the range of potentials of hydrogen adsorptron LS much narrou-er Only the average composrtion of chenusorbed products on rhodrum can be defined, since the precrsron and reproducrbrhtj of the measurements are rather poor because the ranges of electrooxidatlon of chenusorbed partrcles and of oxygen adsorption overlap Our results Indicate that on rhodrum a mixture of particles IS chermsorbed wluch IS on average more reduced than that on platinum, Its composition dependmg on expenmental condrtions. Smooth ruthenium and osrmum electrodes which adsorb hydrogen actrvely both at room and at higher temperatures do not exlubrt CO, chermsorptron. It should be noted that methanol and other orgamc compounds are also poorly adsorbed by these metals. When Raney-nickel and porous Iron electrodes are put mto contact wrth CO, at open cucurt, a certam sluft of potentral IS detec:c:d that tndrcates rnteractron of CO2 and adsorbed hydrogen to be possrble According to the results obtamed, it may be concluded that CO, adsorption does not correlate umquely wrth the presence of adsorbed hydrogen and wth the energy of hydrogen binding CO, IS neither adsorbed nor reduced on a palladmm eiectrode though the latter has a wide energy spectrum of both adsorptron and absorption of

(I’- Rh 22’-PI

FIN 11 The dependence

of the surface coverage of rhodmzd (1 1’) and platmmd (2.2’) electrodes by partxles UI Lhe solution 0 5 M H,SO, + 3 8 x 10m4 M CO1 on potentml

chermsorbed carboncontammg

hydrogen However, the chermsorbed parhcles, su-nilar to those produced during CO2 adsorptton on platinum, are well known to be produced dunng adsorption 01 methanol on palladium and n-rdium. These parhcles cannot be formed from CO, due to dtfftculues of CO2 mteractton wth adsorbed hydrogen on these metals, 1-e CO, molecules are not, seenungly, sufficiently activated to react with hydrogen, though the amounts of hydrogen adsorbed on the metal surface are always sufficient. Therefore, the presence on the surface of adsorbed hydrogen w-~th a deftmte energy is necessary but not sufftctent for CO2 chenusorption on a catalyst-electrode The activauon of CO, molecules is also of unportance, 1 e the character of the mteractton of CO, with the metal is essenual. THE

MECHANISM

O= THE

PROCESS

The expenme:ital data show that the mechanisms of CO2 electroreductton m aqueous solutrons on metals wtth lugh and low hydrogen overvoltages dtffer. Reduction and reductional chermsorptlon of CO, occur by virtue of the interaction of molecular carbon dwxide wth adsorbed hydrogen which was produced dunng the prev-rous step of fast hydrogen adsorption (lust as in the case of metals with lugh overvoltages, HCO; and CO:ions are not reduced and do not form tightly chermsorbed oxlduing parhcles) CO1 molecules adsorbed on the electrode take part m the process Oust as m the case of metals with high overvoltagcs) smce some platinum metals exlnbit no reductional chermsorptron. The full reachon For platinum, wl-nch IS the most representative member of the platinum group, IS O=C=O+ P!

A

3 Pt-H

+ 3 Pt + Hz0

(III)

FL

Th_ts reachon Pt,(CO,),, Pt-COOH Pt’C=O Pt’

Pt -I”>C-OH

can proceed

+ Pt-H,,, + Pt-H,,, + Pt-H,,,

-

through a series of mterrnedtate

Pt-COOH ;:>C=O

+ H,O

steps [34] (IV) (V)

Pt -PtlC-OH Pt’

In the range of hydrogen adsorptton, reactton (IV) detemunes the total rate of CO, adsorption and reductton The rate of thts step depends on the surface coverage by either chenusorbed hydrogen or by weakly adsorbed CO, The standard Gibbs energy of CO, adsorptton IS a quadratrc function of potenhal [35]. The CO, adsorption IS too weak to compete ~t.b hydrogen adsorptton, hence it can take place only on a hydrogen-free surface, 1 e. on (1 - 8,). Therefore, the maxunum rate of reductional adsorption (at ek = 0) is

(5)

323

on Tlus equation explams the extreme dependence ci Lhe rate of CO, chemisorpUon platinum on surface coverage by hydrogen or on pote;ltial It must be taken into account that the reaction occurs only if the centers with adsorbed CO, and H are adJacent (steric factor), hence the rate-constant changes lvlth the number of such adJacen1 centers which m turn changes wth d, and Oco,_ The number of such centers can be calculated using the theory of “order-disorder cf A + B alloys” [36,37] The= calculations also give an extreme dependence of the rate of reductional chenusorption on 8, (or on [email protected] with a m-urn at 8, = 0 5. Therefore, the process of CO, electroreduchon on platmum metals IS a rather fast irreversible interachon of a CO, molecule activated as a result of the weak adsorption on the metal surface Hrlth hydrogen chemisorbed on the metal Ths process results m the formation of tightly chermsorbed particles which are the products of more profound reducbon of CO, than that mto formic acid. The step of further reduction of chenusorbed particles accompamed by desorption of the products into the solution (under normal tond:tlons) 1s extremely slow due to the very strong bond of the chenusorbed parbcles WLLIIthe surface. REFERENCES 1 Yu B Vzsd~ev, V S Bqotiy. N V Osetrova 0 A Khazova and N A Mayorova J ElecIroanzl Chem. 199 (1985) 271. Yu B Va_whev, V S aagoLz!cy. 0 A Khazova and N A Mayoro\q IbId. 189

2 3 4 5 6 7 8 9 i0 11 12 13 14 15 16 17 18 19 20 21 2.2 23 24

(1985) 295 J Gmer. Electrochlm [email protected] 8 (1963) 857, 9 (1564) 63 2 W Nxdrach. S Gllman and I Wernstock, J Ele-clrochem Sot , 112 (1965) 1161 B Plersma. T B Warner and S Schuldmer J Uectrochem Sot, 113 (1966) 841 M W Brelter, Electroclum [email protected] 12 (1967) 1213 S B Brummer and K ColuII, J EIeclroanal Chem. 21 (1969) 463 V N Kamalh and Hera Lal J EIectroanal Chem 19 (196R) 137 V E Kazannov, V N Andreev and G Ya Tysyachnaya. EILklrokhmuya. 8 (1972) 927 H B Urbach, L G Adams and R E Srruth, J Ekuochem Sot , 121 (1974) 233 J Sobkowslu and A Czenvms~. J EIeclroanal Chem , 55 (1974) 391. 59 (1975) 41 B I Pwllovchenko. V F. Slenm and AA_ -haeva EIeh,trokhlrmya. 4 (1968) 1004 P-R Johnson and A T Kuhn, J Elecwocnem Sac , 112 (lS65) 599 V S Bogotzky and Yu B Vasnhev Elecuocb~m AC11 (1569) 1439 Yu B Vasslhev. V S Bagotiy, 0 A Khazovq V V Chemy and A M Merelsky, J Electroanal C hem, 98 (1979) 253. 273. \ E. Kazannov. EIeklrotimlya. 2 (1966) 1170 N B Gorlna and V S Bagotzky. V-V Chemy, V S Zwkova, Yu B Vasslhev, V M Gryamov, Eleklrohhmuya, 8 (1972) 1341 V S Bagotiy, A M Skundm and EK Tuseeva_ Eleztrochun [email protected] 20 (1975) 834 hf R Tsrasevlch. KA Radyushkma and RKh Burshtem. IllekwoNunuy& 3 (1967) 455 V S Bagotzky, Yu B VasslLev, 0 A ‘Khazova and S S Sedovs Electroctum Act& 16 (1971) 2141 b/.A Gromyko. EIektrokhmuya. 8 (1972) 707 N B Goruxx V-S Kondrashevq V P Polyahov& EM Savllskq A M Skundrn and V V. Yahushev. EIektroLhmuya, 12 (1976) 1756 VELK azarmov and G Ya Tysyachnaya. EIektrotiya. 6 (1970) 892 V E Kazannov, V S Bagotiy. Yu B Vaulhev. V N Andree\- and S A Kuhev, Elektrokhlmlya, 17 (1981) 205 A A Mkhdov~ N V Osrtrova and Yu B Vasstiev, EIeklrokhlrmya, 21 (:985) 131

324 25 26 27 28 29 30 31 32 33 34 35 36 37

B I Podlovchenko and L Ah= Elektrokhmuya, 7 (1971) 1031. V N. Andreev. Yu B Vasnhev. N V. Os-etrova and T-N. Yassrrebova. Elektrokhmuya, 19 (1983) 381 A V Zakharyy1. N V. Osetrova and YwB Vawhw. Elektrokhuniya. 13 (1977) 1011 N V Oselrova. Yu B Vass~hev and V S Bagotiy. Uektrokhumyq 13 (1977) 5:2 A V Zakharyan, N V Oselrova and Yu B. Varslhev. Elektrokhmuya. 12 (1976) 1854 Yu B Vass~hev. V-N Andreev. N V Oselrova and TN Yastrebova. Elektroklurmya, 19 (1983) 414 VA Gromyko. Elektroklumlyn, 7 (1971) 885 S A Kubev, V N Andrew. NV Osetrova. V S Bagotzky and Yu B Vassrhev. Elekrrokhmuya 18 (1982) 787 J Sobkowskr, A W~eckowsh. P. Zelenay and A_ Czerwmskr, J Elecrroanal Chem , 100 (1979) 781 Yu B Vasslhev. V S Bagotiy and 0 A Khazcw~ J Electroanal Chem, 81 (1977) 229 B B Darnaskm. OA Petru and V V Batrakov. Adsorptron of Orgamc Gmpunds on Wec~rodes, Nauka, Moscow, 1968, p 61 (m Russian) W A Bragg and FJ Wdharns. Proc R Sot London Ser. A. 145 (1934) 699 R H Fowler and F A Guggenheim. Staatlsl~cal Thermodynanua, Cambndge Unwerslly Press, London. 1939