Indi an Journal o f Chemistry Vol. 43A, May 2004, pp. 1033-1038

Microkinetic model studies of alloying effects on the CO+NO+02 reactions over Pt-Rh nanocatalysts

~bir De Sarkar' & Badal C Khanra

Theoretical Conde nsed Matter Physics Divi sion, Salia Institute o f Nuclear Physics, l/AF , Bidhan Nagar, Ko I kat a 7mnr64~ maia -

Email: abir @cmp.sal;a-:e;:n;i. in

Received 21 October 2003; revised 25 February 2004

Microkinetic mode l has been used to study the kinetics o f CO+NO+02 reactio ns on Pt-Rh nanocatalysts. The a lloy particles show an intermediate behavior for the CO+02 reactio n. For the CO+NO reacti on, the kinetic behavior o r alloy particles are elose to that of Rh, while for the CO+NO+02 reac ti on, it is very close to that of Pt. The rate o f CO+02 and CO+NO+02 reac tions is hi gher than that o f CO+NO reactio n o n Pt, Rh and Pt-Rh nanocatalysts . The li ght-off temperature in case o f CO+NO reac ti on is de layed in the absence of oxygen. The overall kine tics o f the CO+NO+02 reaction on Rh is dominated by CO+NO reactio n kinet ics rather than by CO+02 reaction kine ti cs which has a hi gher inherent rate . The scenari o reverses itse lf o n Pt nanopartic les. The mag nitude o f the NO inhibiti on effect on the CO ox idati on by O 2 is considerabl y smaller o n Pt than on Rh . The simulated results are in qualitative agreement with the experimental findings.

PACS : 68.43. Mn ; 82.20.Pm ; 82.65.+r ; 82.45 .1n ; 82.20. Wt IPC Code: Int. C I. 7 BOil 23/42, BOil 23/36

The catalytic cleaning of vehicle exhaust gas is based on the so-ca lled three-way catalysts (TWC), converting simultaneously CO, the uncombusted hydrocarbons (HC) and nitrogen oxides (NO,) into non-toxic products. In the li ght of the increasing stringency of vehicular emission norms, the physico­chemical properties of the currently used Pt-Rh bimetallic nanocatalysts require a thorough understanding for optimum performance. The reactions of CO with O2 and NO over Pt, Pd, Rh catalysts as well as on Pt-Rh nanocatalysts are important steps in the catalytic control of automobile exhaust emissions. ]n our earlier work 1

, we studied the effect of alloying on the kinetics of CO+02 and CO+NO reaction on the Pt-Rh nanocatalysts. rn this paper we examine how the rates of the CO+NO+02

reactions compare with those of CO+02 and CO+NO reactions. We have used the microkinetic model which takes into account the rates of each elementary reaction step before arriving at the overall reaction rate of CO2 form ation .

Theoretical model We consider the microkinetic model2

-4 in o rder to study the kinetics of CO+NO+02 reaction s on Pt, Rh and the PI - Rh nanocatalysts. The model is described below:

Kinetics of CO+NO+02

reaction

For the CO+NO+02 reaction the followin g elementary steps were considered3

.4 111 the microkinetic scheme: (i) Adsorption of CO and NO molecules onto the surface of the metal or the alloy , (ii) desorption of adsorbed CO and NO molecules from the surface, (iii) dissociation of O2 molecule on the surface followed by adsorption of atomic oxygen, (iv) dissociation of molecularly adsorbed NO into adsorbed atomic nitrogen and oxygen atom, (v) formation of N2 molecules through two possible steps, namely, ~-N2 and 8-N2 steps, and finally , (vi) formation of CO2 molecules through the combination of adsorbed CO molecule and adsorbed oxygen atom.

CO HCO" NOHNOa

0z--720" NOa-7Na+ 0 " Na+N" -7N2(~- N2) NO,,+ N" -7 N2+ 0 " (8 - N2)

CO" + 0 " -7COz

( I ) (2)

(3)

(4) (5)

(6) (7)

The suffix 'a' denotes the component in the adsorbed phase. Adsorption is the first step in the overall reaction scheme. The rate of adsorption of

1034 INDIAN J CHEM, SEC A, MA Y 2004

species i o n metal m is given by molecular colli sion h 2.4. (.) 1/2 t eory . ' a. i m=(RTl2nMi ) (Jill (S i) m C; 8" where Mi

is the mo lecular weight of the species i; (Jm is the area occupi ed by 1 mo le of surface a to ms (cm2/mo le) of

the metal m ; (Si)m is the initial sticking coeffic ient of the species i on the me ta l m; C; is the conce ntration of the component i in the gas phase and is calculated from the gas pressure through the re lation Ci =P;lRT(moles/c.c) ; and Sv is the fraction of vacant surface s ites availab le fo r adsorption a nd is g iven by

Sv = I-Sco- SNO -SwSo; R is the uni versa l gas constant

and T is the temperature in Ke lvin. SI is the s urface coverage of adsorbed species i.

The rate of desorption of CO and N02.3 are g iven

by rd.eo=A exp[- (E - aco. Seo -aN ,SN)IR71 .Seo and

rd.NO = A ex p[- EIRT]. SNO respecti vely; a eo and aN are the inte raction parame ters (kcal/mo le) which account for the repul sive inte raction between COa-COa and C O il-N il' Here A a nd E denote respective ly the pre­expone nti a l facto r (S· I) and the activation energy (kcal/mo le) for a particular rate of the e le me ntary reaction step occurring on the surface. The rate of NO

dissociation is g ive n by rdiss .NO=A exp (- ElK!) SNO. Sv'

and the ra tes o f 2 formation are g ive n by rN2 .r)= A exp[ - (E - aN,SN IRT) lSN2 and rN2.0 = A ex p(-EIRT) SNO.SN. The rate o f fo rmation o f CO2 mo lecules is

g ive n by rem = A exp(-EIRT) . Seo. So. 8-N2 formation step is no t found to occur on PtS

.7

.

The s teady state continuity equ ation s fo r the

adso rbed species CO"' NO", N" and 0 " are . I . b 14 respective y g iven y'

r".eo- rd.eo - re02 = 0

r" .NO- rcl. NO- rdiss.NO -rN2.8= 0

rdiss.NO -rN2.8 -2 rN2.IF 0

rdiss.NO +2r,,,02+rN2.1i - re02= 0

. .. (8)

... (9)

(10)

... ( II )

Fo r study ing the kinetics of the CO+NO+02 reaction , o ne has to so lve the set o f the above four

equatio ns self-consistently for Seo, SNO, SN and So. Thi s he lps us to estimate re02 which is a yards ti ck for

comparing the rates o f differe nt reacti o ns, namely, CO+02, CO+NO and CO+NO+02.

The act iv iti es of Pt and Rh for the CO+02 and CO+NO reactions are known and we resort to the first o rde r approximati on fo r the supported PtsoRhso alloy parti c les by whi ch the rate o f adsorption of CO (say) for the a lloy may be ex pressed as

... ( 12)

Here, (xs)Pt is the surface concentrati on of Pt atoms on the surface of the supported Pt-Rh nanoparticles . The other elementary reaction steps were al so treated on the same footing . The values of the surface concentration of Pt atoms, (xs)Pt were taken from our earlier Monte Carlo calcu lations8

.

Kinetics of CO+02 reaction For the CO+02 reacti on, steps represented by Eqs

I , 3 and 7 are the on ly re levant e lementary react ion ste ps and the rate Eqs 8 and II wou ld on ly require attention . One has to consider on ly the second and the last term of Eq . II which represent the ra te of adsorptio n of O2 and the rate of C O2 formation respectively . The rates re lated to NO di ssociatio n a nd

N2 form ation do not exist for this case. Here, Sv = 1-

Sco and aN = 0 for the CO desorption s tep .

Kinetics of CO+NO reaction For the CO+NO reaction , a ll the e le me ntary

reacti on ste ps barring the step represented by Eq. 3 are releva nt. Therefore, a ll the ra te eq uations need to be considered here. However, the second term in the rate Eq. II (i.e . the one for the rate of adsorption of O2) would no t appear he re for obvi ous reasons. He re,

Sv = 1 - Sco- SNO -8N . The N20 formation is s ig nifica nt only in Ultra Hi gh

Vacuum (UHV) conditions9. Therefore, N20

formation ste p has been ig no red in our model studi es o n higher pressures.

Results and Discussion The paramete rs used in our studi es are g iven in

Table I . The Mo nte Carlo s imulated va lues of surface concentra tio n of Pt, (xs)Pt va lues are 0.548, 0 .551 , 0.559, 0.553 , 0.538, 0.564, 0 .528, 0.5 54, 0 .572 and 0.523 at 450, 475 , 500, 525 , 550, 575, 600, 625, 650 and 675 K respec tive ly. The kine ti cs o f CO+02 and CO+NO reactio ns were in ves tigated separate ly as a function of both tempe rature :.ll1d pressure for obta in ing a deeper insight into the reactivity of both NO and O2 towards CO. A compara tive study of the ki ne ti cs of the CO+02 reaction o n three d i ffe ren t sys te ms is de picted in Fig. I. T he tu rnover numbe r is no thing but the number of CO2 mo lecules th at 'are released pe r s ite per second . The rate o f formation of C O 2 mo lecules is found to be highe r on Pt than on Rh which is quite obv io us. It is q uite evide nt from the

SARKAR el al.: ALLOY IN G EFFECTS ON CO+NO+02 REACTIONS OVER Pt -Rh NANOC ATAL YSTS 1035

Table I- Parameter va lues used III Illode l calculations for CO+NO+02 reacti o n

CO adsorpti on

o m(cm2 Imole )

Seo CO desorpti on A(S-I)

£ (kcal /mole)

Clco(kcal/mo le) uN(kca llmole) CO, formati on A(S:I)

£ (kcal /mole) NO adsorpti on

S NO NO desorption A(S·I)

£ (kGli/mo le) NO di ssociati on A(S-I)

£ (kcallmo le) O2 adsorption

So o-N , formation A(s·i )

£ (kcallmole)

~- 2 format ion A(S-I )

£ (kcallmole)

u N(kcallmo le)

4.03 x l OR

0 .5

1.0 x 10 IJ

29 .63 6.5

1.0 X 10 1:1

13.5

0.5

5 X 10 1:1

26

3.0 X 10 10

29

0.01 2

1.3 )( 10 " 20 .2

3.75 X 108

O. 5

1.6 X 10 14

3 1.6 4 .5 10

1.0 X 10 12

14.3

0 .5

3.0 X 10 10

6x l0 1J [(III) plane] 19

0 .01

2 X 109

2 1

3.0 X 1010

3 1 4

rate constants (i.e. pre-ex ponential factor and activation energy) and the expressio n for the rate of formation o f CO2 molecul es (rC02) that Pt is a bette r

catalyst fo r CO ox idati on than Rh. The PtsoRh so alloy particles show an inte rmed iate behavior.

The variation of the kinetics of the CO+02 reacti on with the partial pressure of the adsorbates (Peo and P02) at constant temperature is depicted in Figs 2 and

3. Figures 2 and 3 reveal an intermediate kineti c behavio r for the supported alloy particles. Figure 2 shows that the turnover number (TON) decreases with the increase in the partial pressure of CO at constant partial pressure of O2. The sudden drop in TON by 2 orders of magnitude at low CO pressures as shown in Fig. 2 is a direct consequence of the inhibiti on of oxygen adso rption caused by a high concentration of COa. This is also corroborated by the findin gs o f Oh3

on Rh . In the limit of hi gh Beo, the CO2 formati on step, despite its low activation energy, would be

kinetically hindered due to very low 80. Again , the pathway for CO desorption is intrinsica lly constricted

n:·· .. ......... -''U>' .... " "".

Pt - 0 -

Rh

supported PtsoRhso

". '0.,

.......... " ...... . .....

-' '0,

Pco = 0.01 atm = P0 2 ............

1.6 1.7 1.8 1.9 2 2.1 2.2

1000fT(K- I)

Fig. I- A compari so n o f the spec ific rates of CO oxidation on three different systems at partial pressure o f CO, Peo = 0.0 I atm = part ial pressure o f O2, POz .

V)

2 '0; Ul Q)

"S u Q)

0 E N

0 2-Q; .D E :::> z Q; > 0 EO ::J t-

102

101

10°

10-1

••••• n.

'''.

Pt -u-­

Rh .... 0 .. • ·

supported Pt50Rh50

P02 = 0.01 atm . T = 500 K

o 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

Partial Pressure of CO (Pea) (in atm)

Fi g. 2-A co mpari son of the va ri ation o f the rate o f CO+02

reac ti o ll with the partial pressure of CO, Peo on 3 d ifferent sys te ms a t temperature. 7 = 500 K and parti al pressure o f O2 .Pm = 0.01 atm .

due to its high activation energy. Therefore, we observe that in the regime of h igh CO pressure, the CO+02 reaction gets poisoned by the accumulation of CO". On the contrary , in Fig. 3, we observe that the rate of CO+02 reaction increases with the inc rease in partia l pressure of O2. This is because of the fact that

the rate of CO adsorption and consequently Beo increases s ignificantl y with the increase in Peo due to the high sticking co-effi c ient of CO. Again , the st icki ng co-efficient of O 2 being sufficiently small er than that of CO, the rate of O 2 adsorp tion and

conseq uently, 80 does not increase substanti a ll y with the increase in POz so as to inhibit the adsorption of

CO. However, an increase in Bo with the in crease in

1036 INDIi\N J CHE 1. SEC A, MA Y 2004

~02 ,----T�--,--,---.-�p-t~1 -_~I-j

Rh j supported PIsoRhso _

.. u'-

• 1 __ ~.-- , ___ 0 ~

: __ : i ._.- 0'- i

1 1

T ~ 500 K • Pco = 0.01 tm L_----L- ...L .~ _l.-....-J~ ~_

o 0.01 0020.030.040 050.06 0.07 O.OS 0.09 0. 1 Partial °ressure of 0;> (P O

2) (in aIm I

rig. :l-A compari soll "rl re variati oll of the rate: of CO+O~

n:actio ll wi th the partia l pre:ssure of 0 :. I'm on 3 different systellls al lemperature: . T = 50() K a,ld panial pressure of CO. Pea = 0.0 I at lll.

-',

.. -1'\ ... 7

---"'--

supported PI

3upported Rh

Rh(111)

supported PIsoRhso "li- ••

--., .. "-- ..

Peo = 0.01 aIm = PNO

1 A i " ~

~ i ~ J

.., I ~

'-~--L-~-'--~~ 1.7 1.S 1.9 2

1 OOOff(K -')

Fig. 4-/\ compara l iv .~ sludy of the rale of CO+NO reacl io ll on three differenl syslcrm al p~lllia l pressure of CO, PC( , = part ial pressure of NO, P",o = 0.0 I atm.

P02 promotes the CO oxidation by O2, as the

expression for reo~ shows a first order dependence on

00 .

Figure 4 compares thc rate of the CO+NO reaction on three different systems. It may be noticed that the tu rnover number for the fo rmation of CO2 molecules is appreciabl y higher on Rh. However, the kinetic behav ior o f the supported all oy parti cles is an intc:rmediate between the 2 metals lO but is closer to tha t of Rh . The low tu rnover number fo r format ion of CO2 mo lecules on supported Pt can be attributed to its low NO dissociation capabi lity. The NO dissociation

CO+NO

f I

supported PI

supported Rh

Rh(111 ) .

supported PIsoRhso

• I

~ ·--~--~ --·~·:e-- ;;.--~ - ~--~ __ ~ __ ~ __ ~ •. " -9-'1/ --Q--9- <>---9--0

T = 500 K. PNO = 0.0 1 aIm

0.020.040.060.0S 0.1 0.120.1 4 0.16 0.1S 0.2

Part ial Pressure of CO (PCO) (in aim)

Fig . 5-A comparison of Ihe varialion o f the rate of CO+NO reaclio n with Ihe part ial pressure of CO. Peo 011 3 difrcrelll sys lems al temperature. T = 500 K ;lI1d panial pressu re of NO. p~() = 0.0 I alill .

step is the most impo rtant step in CO+ 0 reaction because it furni shes the adsorbed oxygen atom with wh ich molecula rl y adsorbed CO can un ite to form CO2 molecule. T hi s means that for the CO+NO react ion the rate of fo rmation of CO2 mo lecules is contro ll ed by the ra te of NO dissoc iation. T hc acti vation energy for 0 dissociation is much higher on Pr than on Rh. In other words, the rare of 0 dissociation on Pt is very small. Thus , the rate of format ion of CO2 molecuies suffers a sharp decline on Pt.

Figures 5 and 6 delineate how the kinetics of the CO+NO reaction vary with the part ial pressure of the adsorbates, Peo and PNO, at a constant temperature . It is ev ident from these Figs 5 anc! 6 that the kinetics o f the CO+NO reaction have a we:lk dependence on the partial pressure of the adsorbates and also that the su pported a lloy partic les closely resemble the kinetic behavior of supported Rh. The kinetics of the CO+NO reaction have much stronger dependence o n temperature as may be noticed in Figs 4, 5 and 6. The reaction pathways for the CO+NO reaction are much more complicated than that for the CO+0 2 reaction. The number of adsorbed species is higher for the CO+NO reaction than for the CO+02 reaction. Therefore, after the CO+NO reaction has progressed sufficiently, the adsorption of CO or NO would always be inhib ited to some ex tent by the presence of other adsorbed species. Hence, the rate of CO+NO reacti o n shows a weak dependence o n the partial pressure o f the adsorbates. Fro m Fig . 5 it is eviden t

SARKAR 1'/ al.: ALLO Yl G EFFECTS ON CO+NO+O~ REACTlO S OVER Pt-Rh NA OCAT LYSTS 1037

I I~:: '" ~ I supported PtsoRhso

I • I I

supported Pt -e--

supported Rh _ ... (> ••••

Rh(111 )

E 100 .. . _.A ._ l!> .... ... .. -A _J'>

.... .. a __ _ ... ... .tI. ••• A ... ... ... -

T = 500 K • PCO = 0.01 atm

0.020.040.060.08 0.1 0.120.14 0.16 0.1 8 0.2

Partial Pressure of Nit ric Oxide (PNO) (in atm)

Fig. 6--A comparative study of the variation of the rate of CO+NO reacti on with the pJrlia l pressure of nit ri c ox ide. PNO at temperature. T = 500 K and partial pressure of CO. Peo= 0.0 I atm on 3 di ffe rc III systems.

that the ki netics of the CO+NO reaction varies weakl y with the partial pressure o f CO. So, in comparison with the CO+O'2 react ion , the inhibiting ro le of CO in the rate of CO+NO reacti o n is almost absent fo r Rh ( II I) and is much less pro nounced for the rest of the system. In Fi g. 3 we have already found that th e rate of CO+02 reaction increases w ith the partial pressure of O2. In Fi g . 6 it may be no ti ced that the rate of CO+ 0 reacti o n on Rh ( I 11 ) increases very sl uggishly with the pa rt ia l press ure of NO, whereas it diminishes slightly wi th the partial pressure of NO for the rest o f the three systems. Thus, a comparison of Figs 3 and 6 would sho\.\' that the increase in pressu re of 0 marginally affects the rate o f the CO+ 0 reaction unlike O2 which appreciab ly promotes the rate of CO+02 react ion.

The inherent rate of CO+O'2 reaction is much hi gher than that o f CO+NO reaction on Rh , Pt al d Pt­Rh nanocatalysts , We have included the results for Rh( 11 1) surface (Fig . 7) as it is a standard sys tem on which immense work has been done3

.9

. It may be noticed that the o nset of the CO+02 reaction takes place at a much lower temperature than that o f the CO+ 0 react ion . In o ther words, the light-off temperature for the CO+NO reaction is hi gher than that of CO+02 reaction . It may be recalled here that the light-off temperature fo r a particul ar reacti on is the min im um temperature at which it is fired ). T his observation is in agreemen t w ith ex perimental findin gstl fo r Rhl A120 3. It is al so evident from Fig. 7 that the kineti cs of CO+NO+0 2 reaction on Rh(l ll )

Ul 2 'iii Ui ~ ::J () Q)

(5 E

N

0 ~ CD .D E ::J Z

ID > 0 E ::J f-

103

102

101

100

10-1

0 ..

o

"0

CO+NO+0 2 -e-­

CO+NO CO+0 2

Rh(111)

1.5 1.6 1.7 1.8 1.9 2 2.1 2.2

1000rr (K I)

Fig. 7-A comparati ve slUdy of the rates or 3 diffe rent reactions on Rh ( III ) at partial pressure of CO. Peo = parti al pressure of NO. PNO = partial pressure or 0 1• P02 = (J.G I at l11.

is closer to that of CO+NO reaction rather than th at of CO+02, Thus, the kinetic behav io r of the CO+NO+0 2

syste m o n Rh ( II I) is domi nated by the CO+NO reac ti on kinetics ra ther than by the CO+02 reacti on kinetics, Here, the overall kine tics of the CO+NO+0 2

s stem is foun d to be controlled by the CO+NO reaction which has a lower inherent rate. F rom thi s observation one can at once in fer that the presence of NO in the gas mixture in hibit. the fo rmat ion of CO~ on Rh, In other words. the rate of 0+02 reacti o n on Rh( III ) is strong ly suppressed when 0 is present in the reacting gas , whereas the CO+' 0 reacti o n is only mildl y affected by the presence o f oxygen. Th is findin g is al so corroborated by e peri mental resu lt II

for Rh/AI 20 J ,

The o verall kirwtics of the CO+ 0 +0 2 reaction on supported Pt is governe<i by the CO+02 reaction ki netics. Therefore, the ki net ic of the CO+NO+0 2

reac ti on on supported Pt has more of CO+02

character and less of CO+NO character. It is quite o' vious th at the rate o f CO oxidation by O2 is much iess impeded by the presence of NO. T hi s o bservation agrees wei ! with the experimental findi ngs ll.

ow, Pt-N bond strength being much less than Rh-bond strength J(), 2 is instantaneous ly re leased from

the surface of Pt through the co mbinati on of 2 Na's as soon as Na is formed from the di ssoci ati o n o f NO/-7

.

The ro le o f inhibiti on of 0 in the o xidation of CO by O 2 on Rh can thus be exp lained in terms of blocking o f sites by N" as Na binds to Rh-sites more strong ly . The kinetics of the CO+NO+02 on

1038 IND IAN J CHEM. SEC A, MA Y 2004

supported Pt is very close to that of supported alloy parti cles .

The rate of CO+NO+O~ reaction o n Rh , Pt and Pt­Rh systems is higher than that of CO+NO reaction and the light-off temperature fo r the CO+NO reaction is higher than that of CO+NO+Oz reacti on o n a ll these 3 ~ystems . Therefore, it can be infe rred here that the lig ht-o ff is delayed in the absence o f 0 2. From th is observatio n o n the ligh t-o ff behavi or o ne may arrive at the conclu sion that it is the presence o f oxygen that is of c rucial impo rtance in three-way catalys ts.

Conclusion Microk inet ic model has been used to study the

effect of alloying on the kinetics of CO+NO+02

reacti ons on Pt-Rh nanocatalysts. T he fo ll owing res ul ts have been fo und.

(i) T he suppo rted alloy particles show an inte r­medi ate kine tic behavior fo r the CO+Oz reaction and the kinetic behav io r of the supported alloy part icl es is close to that of Rh fo r the CO+NO reactio n, whil e for the CO+NO+02 reacti on it is very c lose to that o f Pt ; (ii) the inherent rate of CO+Oz and CO+NO+Oz reacti on is hi gher than that of the CO+NO reacti on on a ll the systems studied in this work; (iii ) the light-off temperature o f the CO+Oz and CO+NO+Oz reaction

is higher than that of CO+NO reacti on on the sys tems studied; and (iv) the kinetics of the CO+NO+02

reacti on on Rh show more of CO+ NO character and less o f CO+Oz character. T he situation is j ust the o pposite fo r Pt. Though the model used here is ve ry simple, it can account fo r many experimentally observed facts. The present find ings may be gainfully ex plo ited in three-way catalysis.

References I Sark aI' A De & Kh anra B C. lilt/ian J Physics. 77 A

(2003)603. 2 Herz R K & Marin S P, J Cala/, 65 ( 1980) 28 1 . 3 Oh S 1-1, Fisher G B, Carpenter J E & Goodman 0 W, J

Caw/, 100 ( 1986) 360. I[ I-Ioebin k J H B J, van Ge illeil R A. van de n Tillaan J A A &

Marin G B, Chell/ Ellg Sci, 55 (2000) 1573. 5 Kon luke 0 & von Ni essen W, Sill! Sci, 40 I ( 1998) 185. 6 Fink Th , Oath J -P, Basse tt M R, Illlb ihl R & Enl G. Sil l!

Sc i, 245 ( 1991 ) 96. 7 Fink Th , Oath J-P, Illlbihl R & En! G, J Chell/ Phys, 95

( 1991 ) 2 109. 8 SarkaI' A De & Kh anra B C, Illdiall J Chelll A , 41 A (2002)

1784. 9 Zhdanov V P & Kaseillo B, Sill! Sci Rep, 29 (1997) 3 1. 10 Heezen L, Kilian V N, van Slooten R F, Wolf R M &

Ni euwenhuys B E, Caw/ysis alld A lllolI/olive POlllllioll COlllro/lf , (Elsevier, Amsterdam) ( 199 1) 38 1.

II Oh S H & Carpenter J E, J CaW/, 101 ( 1986) 114.