Title

Studies of colloidal catalysts by thermal analysis of reactionvelocity VII : on the permanent poisoning of catalysts(Commemoration volume dedicated to Prof. Shinkichi Horibain celebration of his sixtieth birthday)

Author(s) Suito, Eiji

Citation The Review of Physical Chemistry of Japan (1947), 20: 35-41

Issue Date 1947-09-15

URL http://hdl.handle.net/2433/46649

Right

Type Departmental Bulletin Paper

Textversion publisher

Kyoto University

~ t/ __,

' ~ The Review of Physical Chemistoy+of Japan Vol. 20 (1946)

STUDIES OF COLLOIDAL CATALYSTS BY THERMAL

ANA~i,YSIS OF REACTION VELOCITY. VII.

On the Permanent Poisoning of Catalysts.

• 13y Elp Sutxo. .

Generally, in the poisaling of catalysts hvo types arc recognized, "temporary

and permanent ". As ,ate .of the' former cases, the poisoning of the catalytic decomposition of lrydrogen peroxide on colloidal platinum by carbon monoxide was

reported in the previous report of this series'/. Now, as the latter case, the effect

of mercuric chloride, mercuric cyanide and potassium cyanide (and; as the comparing , common salt, potassium chloride) on that reaction has been investigated by the

method of thermal analysis of reaction veloci[ypt. 'then the behaviour of the poison,

.the cation Hg~' and the anion CN- (Ii+ aad CI ions male no poisoning), examined

and the ionic adsorption discussed.

Experimental.

The apparatus and the pratedure-of the thermal analysis were similar to

those in the previous report"t. Namely, ; 5o cc, of hydrogen peroxide solution and 5o cc. of platinum sol, to either of which a given amount of ffie poison was

added, were mixed together in a glass calorimeter. The temperature rise, pT, in

the reaction system caused by the decomposition of hydrogen peroxide is measured with time, t. From the thermoanalytical equation,

h': the cooling constant of the reaction system, 0.013

lip: the water equivalenh of the reaction system, t 19.5 cal.

Q : the decomposition heat of I-L_O„ 2g.q Kcal/mol. .

the reaction velocit} ~[ , is obtained. The relation pT•-t (curve' (a) in the • figures) and log ~ -t (curve (b)) show the type of the reaction. I( the reac-

tion is first order, the latter relation is expressed by a strcight line and its slope

i) Eizi Suito, Thir founzul, I8, 96, tog (i9gq). ~ z)• Eizi Suito, rdlJ., 13, 76 (I939)• IS, t, t55 (19gt). 16, r (rgyz).

i I

,. .::~x~.

1The Review of Physical Chemistry of Japan Val.

1

36 F.. SU1T0 ~ ,

indicatcsthe velocity constant. The activities of catalysts are compared by using

the first order velocity constant, k, and the initial reaction velocity, C ~ /o The pln[iuutn sol used was prepared electricdly by high frequency alternating

current ($vedberg method)in air and treated -with hydrogen gas; the sol obtained

was accordingly very stable hydrogen sol which had-the surface of platinum'I and

its eoncentratio» u•as 3-6 x to ° g-atom/1 at the reaction. The amount of the

poisons .added leas so small that .the sol did riot coagulate though all the experi-ments. (hut in the case of a large amowtt of KCl it coagulated.) The concent-ratiat of leydi-ogeu_ peroxiFfe tvas'abvays o.05 mol/1 at-the reaction. The tnuperatnre of the ctpcrimcnts was always 25°C.

Experimental Results.

Series A: The reaction hetwcen the HoO,-

solution ;md the Pt-sols containing I~ various amowtts of poisons.

Series B: The reaction between the HoO.,-1

solutions containing various amount • of poisons and the pt-sot. d

~5 ~ •t:d fY. ~ (1) Mercuric chloride, HgCI_. v °vvrovt of oo v's .Series fl. (];Xp. -INTO. I~g ShOWR 111 PIg. L) a °ovvv~ cost svs

0 oavwas °war fy

i °°°o~a°o~°O,'+ ugObPO u The reaction type is .not different from the e Doer aoaer u

+•«5 stuldard , reaction in the absence of poison

i° , P (Lxp. No. t.) and is mailily first order, except

at the initial st<1ge which is caused particularly

m o by the hydrogen sot. As the amount of the ,r P.. 5 ,

rts. I poison is increased, the activity, k or C ̀~) , tGt o

«r m'.,. 4~~ F /y9..w'~ ~ decrease. ^ ! O aPPd 1095 s ovovoot orsv v2a Series ~ (I'.xp. NO. 9^rII In Pig. 2.)

/O 5 °W95 f_S vevam a°o!° 6Y 'Pl.e ..~H,... •....e ....A N.,. .1.... _......L... _!

~ an ~ n.rd

. .,

F'ig. ~

c utii~is~i

Series B (Isxp. No. 9,-ii in Pig. 2.)

The reaction type and the depression the. activity is much the same as in Series

(2) Potassium cyanide, ACN. Serres A (i:p. No. ~2~i9 in Fig. 3.)

The relation between ]og t~ -and t is

~.. .,

~ :. ,

of

A.

net

(1946)20

=~ -

- -~ The Review of Physical Chemistry of Japan Vol

STUDIES OF COLLOIDAL CATALYSTS RY TIIL'RMAL ANALYSIS OF REACTION ;y 1

linear but " S "-shaped ; 'viz. the reaction is not first order. Although k is not

determined, 4he activity ilidicated by C r~ /u is remarkably decreased ivith the increase of concentration of KCN. 'Series B (I•:xp, No. zo-~-26 in Fig. 4.) I

The reaction is of ." S " type, being the same as in Series ~, but it differs

from Series A that all the log (~~t-)--t curves caiverge to the ppint (ft )aa of standard reaction when t-o.

_ -,-

d.~ , v

f _

1, i 9.

.,

I/f U 0:/4 /990 /3 O y' ~

!f~ + S 40000/ 0010 3d/ N~ p /0 00000/ PO/d 'fLY' A O 9v 3/ I s J dd f90

/L 93 1: ] /IS •f L 3d 1J t _ Sd 9A3 /f OWO/ 33 S UOOGId 4J3 /:~ ? OS ~ So 2 pyp

h

»

\ V6

K .. A b a• .r~o~iu T.2~T~

, L..

Fig. 3

(3) Mercuric cyanide, Hg(CI~,..

Series A (I's.p. No: 2733 in I~ig. Series B (I?xp. No. 3'4^-35 in Fig.

In both Series, the reaction is of

affected by tlic concentration of the

(4) Potassium chloride, KCI..

Series A (ISxp. No. 36--41 in Fig. Series B (Isxp. No. 42^-45 in Fig.

It is the first order reaction.

I

5.) ~. 6.)

" S "type and

poison.

7~) 8.)

In comparison

(1946)'20

N

~t (~:..

r~ a

the activity,(~)n is si~nilary

with the case of poison, the

- :sue : e-. ... ~

38 E. SUITO

The Review of Physical Cfiemistry of Japan Vol. 20 (1946)

d

41 ~

a

S` ̀' 1

a

(O

ly

anti. c..~.4r_.,, t ..L

a e,mnca3 e:••• ~s

)o v 33 a9~~0G/5

]9 J'\ •,

1~~

~~

l

1

l

.

I

.,

Y

R1

W. S

H O60/ Q 96 bfM1 JS s ma. 360

~] aw .oej af) 0 3 ~JOJ] /9J y

b/ 66]b 99

a - .

~•.: -

m _. r a., , .; nr w ~e

Pig. 5

1Y a ~ a!Y! 1990 ,t) J! a9a00aT ]9a

Js a9om9 _ )zs 4.~lOY6.O l~

~• ~ .

n ! >• ' _

depression of the

salt, that sol coa

ReaMion type.

In the case

first order. This

does not react wi

the cases of I:G

rather indicate th.

Proceeds accordin

-/ ,.n,

rlg. ~

WW ^y / J6 a `~91J9 Isla ~ O) UW6 C%Y 6n0 u aw 9a+ us 14 n06 0919 )!i .y •6 a/ N1/ I3l ,..

w .p -~i ,..H ~t M1.. , -

r•~g. v rag. s

activity is attended by the addition of so large amounts of the

gulates in6the concentration of KCl above o.oj nto]/l.

Considerations. ,

of I-TgCI, (and KCI), the reactia~ type has not changed, being the

fact demonstrated that the I-Ig*+ ion once adsorbed on the catalyst

i th the react.vtt~ H:O.,'and Is a completely permanent poison. In V and Hg (CN)D, at the other hand, the " S " type reaction may 2t the activity of the catalyst diminishes slowly as the reaction g to the influence of'thc reactant on the poison CN' ion; than

1 ; ,u .y~.

~ . i

The Review of Physical Chemistry of Japan Vol.

STUIIIRS OF COLLOIDAL CATALYSTS BY TFIRR~IAL ANALYSIS OF RfiACTION 3g

that hvo kinds of first order reaction with the constant k, and k., (k;> kz) consecu-

tively proceed. `

Toxicity of various poisons.

The ratio of the first order reaction- constant at various concentration of a

poison to that in absence of the poison, k~/ko, conincides with the same ratio of

the initial ruction velocity, I di l I/ ̀fz l t, in the case of HgCI~ and KC1. As dt a,c ̀ dt 0,0

the concentration of the catalyst, C,rin a series is slightly different from each other,

the ratio of the concentrations of poison, to that of I't-sol in each series, c'-c/C, indicates the relative amounts of lJbison th-ough all series. The relation between the

i relative activity, k°/ko or ~ dx ~ ~ ~~ dx

dt n,c dt o.o

,,, and the relative amount of poison, 1, is °,

i : ~ _ shown Fig. 9. Lr the case of HgCh,

a` '+"' ~ ~',"" IIg(CN)$ and KCI, the eaves of Series A '

~, and B fall on each other. The order of

i ~ ;;t ,,, `,~ toxicity is [-Ig(CN).,> I-IgClp> KCN (KCI) a jt '` as indicated by the relative amounts of

~• ~ '~~ poison to reduce the relative activity to I/z - •v-5•° ~• ~-~-°-~ and I/IO . (see Table I). Obviously KCI

F'g' 9 does nol act as a poison,. and 1:+ and Cl-

ions, therefore, -can be neglected in the discussion of poisoning effect.

TaUlt i. - '

poison

taxisityReaction lype

ttfs hna

t[sclr

1 [g(CN).

KCNKL'I

o.ao8

0.0008

to

o.o:

0.008

(0.03)s5a

A=D

A=D

A~D

A=1i

6rst order"5" type

~~ y ~~ type

first order

20 (1946) .

The curves of HgCh anr~ f Ig(CN}' in Fig. 9. are linear, bending at c~=o.5 mol. For the cause of this point, i\Iaxted'~ attributed it to two kinds of• active

centre on the catalyst. But the author thinks that it is equivalent to the saturation

value of adsorption that all parts of the catalyst sarface is covered by the poison)

Hg++ Now; let us calculate the .area of the catalyst sarface. If a particle of the

3) E. 13. Tfaxted, J. Chriu. Sor., 119, zx5 ti9z6), 121, r76o Q9zz), 127, 73 (i9z5)• E. ii. Mazted ~ G. J. Lex~is, ibid.; goz -Q 9bg).

y w

- , _

The Review of Physical Chemistry of Japan Vol.

, ao i. sumo

~'

platinum-colloid is a sphere of z ntp radius (sp. wt. of Pt is zr4); the number of colloid particle is 3 x tom and its total surface area is r.5 x to' cm- ler r g-atom.

As the radius of Hg*~' ion (fIg atom), is c.r'(t.j) ~, its number to cover

completely the total surface of the catalyst as monovtolecular film is 3 X_io' ([.y x ro=') and naznely o.5 (0.3) mol. "llus value coincides with the above value of

~. Of course, it is hssumed that the surface of the catalyst is smooth, because its

colloidal particles have primary stlbility.

On the ionic adsorption.

It is sure that the falling of activity is due to strong adsorption of the poisons

on the ,surfaces of colloidal platinum. Generally, there arc two tylxs of adsorption t in solution, -

(a) Molecidar adsorption-non-polar adsorption (non-electrolyte) (b) Ionic adsorption.-polar adsorption (electro]}4e).

And the latter can be divided into 4 cases as follow:

(t) Tme adsorption of electrolyte (Gibb's adsorption) (') Adsorption of potential dcterminiiig ion.

(3) Exchange adsorption of counter ion. • (4) F.uhangc adsorption of lattice ion.

It is recognized that the adsorption isothemt is expressed by Freundlich equatidn log x=alogc iu the case (t) and x=r<}blogc, which is led from the IQentst-

equation, iit, the (z).

As seen in Fig. q. in the case of HgCI= and I ig(CN). the activity of catalyst

is in proportion to the logaritlun of the poison concentration, via. ~°. =iz-b IogC, ° below its saturation, value, ht the case of KCI, on the othcrhaitc3, the relation

between k°/k~ and log d is net linear, that between log k°/k° and log c' being linear. Accordingly in the case of IiCI which does not -act as a poison, it seems

-that neither Iif nor CI- ion has special affinity and can'be selectively adsorbed on

platinum surface, but the accunudation of the electrolyte-both cation and anion-retards the reaction velocity. (case (t)). It is quite naturaLthaY the infltiencc comes out only when the concentration of the electrolyte is very high as compared with the other cases. On the otherhand, owvlg to the special.a(Iinity ofinercury to

platinum, Hg++ ion will be strongly adsorbed on the surface of the catalyst and never desorbed by any cause (case (2)). Thus, this ionic adsorption curses

phenomena of permanent poisoning,' Lt this case, both NgCI. and Hg(CN). have very small ionisation values (IIgCI.. is regarded as a non-electrolyte), but the

J.. u

b

20 (1946)

~ .:' . ~..i.1J...~ ~..,-

v'-:4q: -- ~, '

The Review of Physical Chemistry of Japan VoY. 20 (1946)

, '

STUDIES OF COLLOIDAL CAT.ll,YSTS BY' TIiERMAL ANALYSIS OF REACTION qt.

dissociation will be increased by the strong ionic adsorption. A(tcr the saturation concentration of poison to cover the total surface of catalyst, the hydrogen peroxide , is not to decompose, but practically it decomposes a little. This fact seems to show tliat the decomposition occurs on the surface of mercury which is adsorbed

and deposited out on p]atinum, Hourly, by the I-Ig-sol. (It was tested that H_0=

tvas never decomposed by IIgCh alone).

In the case of CN poison, it may be attributed' to the molecular adsorption of- HCN rntlter t1lan to the ionic adsorption of CN- ion, became IiCN is easily

formed by hydrolysis of KCN in .solution. And, as~ HCN is a very weak acid

(ionization const. is y.z x Yo '~, its dissociation degree tivill be changed by H_Q (an acid), and then this seems to cause " S " type reaction. .

y'That the initial reaction velocities in Series A and R arc equal to each other ,. shows that dIe ionic adsorption just discussed need very short time to reach its

equilibrium state.

In closing, it is the author's pleasant duty to acknowledge his sincere thanks to Professor S. I Ioriba for his valuable guidance during the course of this research,

t The cost of this research has been defrayed from the Scientific Research

Expenditure. of the Educational Department to Ivhich the author's thanks are due.

Iwsti9de far Chemiasl Resereb. (Takatveki),

_ Kyoto Inryerial Uiaiuer+n.ty.

i

~ ,