CEMENT md CONCRETE RESEARCH. Vol. 22, pp. 941-948, 1992. Printed in the USA.

0008-8846/92. $5.00-t00. Copyright 1992 Pergamon Press Ltd.

T HE R MODYNAMI C AND KI NETI C AP P ROACH TO T HE ALKALI -
SI LI CA REACTI ON. P ART 1: CONCEP TS
ABS T R AC T
Roger Dr on and Fr an~oi se Br i vot
Labor at oi r e Cent r al des Pont s et Chauss6es
58 Bd Lef ebvr e- 75732 Par i s cedex 15
(Communicated by M. Moranville-Regourd)
(Received Nov. 21, 1991)
The r eact i vi t y of si l i ca can be appr oached by usi ng t he t her mo-
dynami c concept of affi ni t y and t he ki net i c concept s of r eact i on
rate, ener gy of act i vat i on, and or der wi t h r espect to OH-.
The ki net i cs of di ssol ut i on of a gi ven var i et y of si l i ca mus t be
s t udi ed on agi t at ed s ys t e ms e ns ur i ng t he h o mo g e n e i t y of t he
sol ut i on at all t i mes. Thi s makes pos s i bl e de t e r mi nat i on of t he
ener gy of act i vat i on and of t he or der wi t h r espect to OH-.
On t he ot her hand, t he r educt i on in al kal i ni t y has no me ani ng
unl e s s it is d e t e r mi n e d on st at i c s ys t e ms not i nc l udi ng any
s u p e r n a t a n t .
Le pr obl ~me de la r 6act i vi t 6 de la si l i ce peut 6t re abor d6 en
f ai s ant appel au c o n c e p t t h e r mo d y n a mi q u e d' af f i ni t 6 et aux
c o n c e p t s ci n6t i que s de vi t e s s e de r 6act i on, d' 6ne r gi e d' act i -
vat i on et d' or dr e par r appor t h OH-.
La ci n6t i que de di s s ol ut i on d' une vari 6t 6 donn6e de si l i ce doi t
~tre 6t udi 6e sur des syst ~mes agi t 6s gar ant i s s ant l ' homog6n6i t 6
pe r ma ne nt e de la sol ut i on. El l e donne acc~s /~ la d6t er mi nat i on
de l ' 6ner gi e d' act i vat i on et de l ' ordre par r appor t h OH-.
Par cont r e, la r 6duct i on d' al cal i ni t 6 n' a de sens que si el l e est
d6t er mi n6e sur des syst ~mes ne c ompor t a nt pas de sur nageant .
lalx_a.a_ltr, lL0_a
Whi l e k n o wl e d g e of t he t h e r mo d y n a mi c and ki ne t i c l aws t hat gove r n
c h e mi c a l r e a c t i o n s is no s ubs t i t ut e f or an a p p r o p r i a t e e x p e r i me n t a l
941
942 R. Dron and F. Bfivot Vol. 22, No. 5
i n ve s t i g a t i o n , t he g u i d a n c e t h e y can p r o v i d e in c o n d u c t i n g e x p e r i me n t s
and i nt er pr et i ng t he r es ul t s is t oo of t en unde r e s t i ma t e d.
The al kal i - si l i ca r eact i on is fi r st of all a chemi cal r eact i on, and mor e par t i -
cul ar l y an a c i d- ba s e r eact i on. The aci d r eact ant is si l i ca in t he sol i d st at e.
The ba s i c r e a c t a n t s ar e p o t a s s i u m ( or s o d i um) h y d r o x i d e in t he p o r e
sol ut i on, and cal ci um h y d r o xi d e in t he sol i d st at e. The r eact i on me d i um is
wat er . The pr oduc t of t he r eact i on is a cal ci um p o t a s s i um si l i cat e hydr at e,
( C- K- S- H) or a cal ci um s odi um si l i cat e hydr at e ( C- Na- S- H) .
The not i on of af f i ni t y, whi ch f ol l ows f r om t he appl i cat i on of t he pr i nci pl es
of t h e r mo d y n a mi c s , is l i nked t o t hat of t he var i at i on of f r ee e ne r gy t hat
a c c o mp a n i e s any r e a c t i o n . It c o v e r s t wo f u n d a me n t a l a s p e c t s of t he
p r o bl e m t hat occur s in pr act i ce: t he r e act i vi t y, and t he me chani cal e ne r gy
t hat is t he c a u s e of t he d e g r a d a t i o n s . It is h o we v e r l i mi t e d t o one
de f i ni t i on of a pos s i bi l i t y of t r ans f or mat i on, wi t hout t aki ng account of i t s
dyna mi c . The br oa de r not i on of r e a c t i vi t y e n c o mp a s s e s t hi s ot her a s p e c t
and a c c o r d i n g l y c o n c e r n s t he c o n c e p t s of c h e mi c a l ki net i cs : ki net i c l aw,
or der , e ne r gy of act i vat i on, me c ha ni s ms , st eps, cont r ol i ng st eps.
The r at i onal a ppr oa c h to t he p h e n o me n a s houl d t he r e f or e make it pos s i bl e
to pos e s o me que s t i ons mor e cl ear l y. Al r e ady, a cer t ai n n umbe r of con-
ce pt s can be r e g a r d e d as f o u n d e d bot h e xp e r i me n t a l l y and t he or e t i cal l y.
Ot her s have run i nt o obs t a c l e s t hat wi l l be di f f i cul t to o v e r c o me and t hat
wi l l s o me t i me s ha ve t o be s i de s t e ppe d. Thi s is what we ar e goi ng t o
at t e mpt t o s how.
Conc e nt s
1. Th e r ol e of b a s e s in t he ASR
"It is t he h y d r o x y l i on t hat at t acks a ggr e ga t e , not s o d i um i on". Thi s
pr i nci pl e was cl ear l y st at ed by Dent Gl as s e r and Ka t a oka (1) as ear l y as
1981. It me a n s t hat t he al kal i ne c a t i ons pl ay s up p o r t i n g par t s i n t he
pr i mar y act of at t ack of si l i ca and t hat t hey ar e t her e onl y to a c c o mp a n y
t he OH- ani on t hat is t he act i ve agent . Expe r i me nt s of Chat t er j i et al. (2)
mi ght s ugge s t t hat this is not t he case, and t hat neut r al sal t s ar e s uf f i ci e nt
to pr oduc e t he r eact i on, si nce t hey l ead to l ar ge e xpans i ons whe n added t o
t he c e me n t - r e a c t i v e aggr e gat e s ys t e m, pr ovi de d h o we v e r t hat it is at hi gh
conce nt r at i ons ( 3N) . Thes e r esul t s ar e all t he mor e t r oubl i ng in t hat , at t he
s a me c o n c e n t r a t i o n , t he KOH and Na OH ba s e s c a u s e muc h s ma l l e r
e x p a n s i o n s .
The f ol l owi ng obs e r va t i ons may be made:
(a) The s odi um and pot a s s i um sal t s have ver y si mi l ar ef f ect s .
Vol. 22, No. 5 ALKALI-SILICA REACTION, CONCEPTS 943
(b) The expansi on process is char act er i zed by a r at her l ong i nduct i on time,
4 weeks for chl or i des and nitrites, 8 weeks for nitrates.
(c) When 3N solutions of KOH or of NaOH are used, the i nduct i on time is
zero, but the final expansi on is much less than with the salts.
(d) When an ani on preci pi t at i ng t he cal ci um ion is pr esent (for exampl e
t he phosphat e ion) the expansi on is zero.
These same aut hors gi ve the key to t he eni gma:
It is t hr ough a r eact i on wi t h the al umi nat es t hat t he neut r al salts cause
t he appear ance of t he act i ve agent OH- . Ther e agai n t he mai n r ol e is
pl ayed not by t he al kal i ne cat i ons, but by t he ani ons t hat accompany
t hem, namel y CI-, NO2-, and NO3-, whi ch have t he propert y of substituting
for the OH- ion of the al umi nat es to form homol ogues of Fri edel ' s salts.
The exi st ence of a l at ency t i me shows that this r eact i on is slow.
The Ca 2 ion has a much mor e f undament al role. Its pr esence is essent i al
for the f or mat i on of the fi nal del et er i ous compound. It bei ng admi t t ed that
t he por t l andi t e of the cement const i t ut es its source, the cal ci um ion must
be able to mi gr at e by di ffusi on to t he sites wher e this pr oduct appears.
We are goi ng to show that, for this reason, par adoxi cal l y, the concent r at ed
al kal i ne bases cause little expansi on:
In a me di um [OH-] = 3N the concent r at i on of cal ci um ions is ver y low,
because of t he const ancy of t he sol ubi l i t y pr oduct [Ca2+][OH - ]2 =10-4.5.
The di f f usi on fl uxes bei ng, in accor dance wi t h Fi ck' s law, pr opor t i onal to
t he gr adi ent s of conce nt r at i on, t hey are pr act i cal l y zer o unde r t hese
condi t i ons, so that t he di ssol ut i on of the cal ci um hydr oxi de is i nhi bi t ed.
This expl ai ns the ver y low rate of expansi on observed in such medi a.
A si mi l ar process takes pl ace in the pr esence of PO4 3- ions, whi ch preci pi -
tate the Ca 2 ion.
2. The probl em of si l i ca react i vi t y
The r eact i ons of t he chemi st r y of cement s i nvol ve t opochemi cal i nt er f ace
pr ocesses and i oni c r eact i ons in sol ut i on. Si nce each at om of si l i con is
i ni t i al l y connect ed to t he l at t i ce by four si l oxane bonds, it is obvi ous that
t hei r r upt ur e by OH- ions is of t he t opochemi cal type. These rupt ures lead
to t he r el ease of ani oni c entities that may i ncl ude several at oms of silicon,
but t hat evol ve in a basic medi um t owar ds the si mpl e monomer i c for ms,
by br eaki ng of t he resi dual si l oxane bri dges.
The Le Chat el i er' s post ul at e, on whi ch we shall base the devel opment s that
fol l ow, assumes t hat it is t hese i oni c for ms that, aft er a transit, possi bl y
ver y shor t , in t he sol ut i on, wi l l i nt er act wi t h t he cat i ons pr es ent to
pr oduce r el at i vel y i nsol ubl e compounds, (i.g. a cal ci um pot assi um si l i cat e
or a cal ci um sodi um silicate).
944 R. Dron and F. Brivot Vol. 22, No. 5
The pr i mar y act is t her ef or e t he di ssol ut i on of t he silica, whi ch we must
consi der f r om t he t wo vi ewpoi nt s of t her modynami cs and ki net i cs.
3. S ol u b i l i t y i n t h e t h e r mod y nami c sense
St ri ct l y speaki ng, t he sol ubi l i t y of a subst ance is expr essed by its concen-
t rat i on in t he sol vent phase when equi l i br i um is r eached. Thi s def i ni t i on
can not be appl i ed to silica, because t he speci es SiO2 does not exi st in
sol ut i on. We must consi der t he equi l i br i um bet ween a sur face hydr at e of
the type SIO2, H20 and the solvated species SiO4H4 in solution
SiO2, H20(surf.) + H20 . . . . . > SiO4H4(solution)
The const ant of this equi l i br i um is [SiO4H4] = k. It depends on the vari et y
of silica. Pascal (3) gi ves, for t he sol ubi l i t y of quart z and of amor phous
silica at 25C, the values of 0.03 g/1 and 0.15 g/l
or kquartz = 5 10 -4 and kam.sil" = 2.5 10 -3.
The react i on can also be wri t t en to i nvol ve the OH- i o n .
SiO2, H20 + OH- - . . . . > SiO4H3-.
Under t hese condi t i ons the const ant of equi l i br i um becomes
[SiO4H3-]/[OH-] = k'
k and k' are rel at i ve to the t ype of silica in question.
It can be seen that the sol ubi l i t y can be expr essed ei t her by the concen-
tration of the neut ral species [SiOnH4] or by the ratio [Si O4H3- ]/ [OH' ].
Since the pKA of the SiO4H4/SiO4H 3- coupl e is pK1=9.6, the t wo expressi ons
are l i nked by the rel at i on
[SiOnH3-]/[OH-] = 104.4[SIO4H4]
In a hi ghl y basic medi um the species SiO4H22- is also found, and it is easy
to est abl i sh t he pr opor t i onal i t y
[SiOnH22-]/[OH] 2 = 105.7[SIO4H4],
the pKA of the SiO4H3-/SiOaH22- coupl e being pK2 =12.6.
The "appar ent " sol ubi l i t y of si l i ca in a sol ut i on of sodi um hydr oxi de
depends on t he i ni t i al conce nt r at i on of t he sodi um hydr oxi de. In t he
absence of pol ymer i zed forms, it is expr essed as the sum of the concent r a-
tions of t he t hr ee speci es, the di st ri but i on of whi ch is a f unct i on of the
final pH: For pH bet ween pKI +I = 10.6 and pK2-1 = 11.6 t he pr eponder ant
form of silica is Si O4H3- ( t he Si O4H22-/ Si O4H3 - and Si O4H4/ Si O4H3- r at i os
bei ng less t han 1/10).
The el ect r oneut r al i t y of the sol ut i on is t her ef or e wri t t en
[SiO4H3"] + [OH-] = [Na +]
In the case of an amor phous silica, for whi ch k = 2.5 10 -3 , it can be seen
that [OH-] is also negligible with respect to [SiO4H3-]
Vol. 22, No. 5 ALKALI-SILICA REACTION, CONCEPTS 945
At pH 10.6
At pH 11.6
[SiO4H4] = 2.5 10 -3 what ever the pH
[SiO4H3-] = 2.5 10 -2 and [OH-] = 4 10 -4
[SiO4H3-] = 2,5 10 -1 and [OH-] = 4 10 -3
In t hese t wo e xt r e me cases, and thus for the i nt er medi at e pH, [OH-] is
negl i gi bl e with respect to [SiO4H3-].
Cons equent l y, [SiOaH3"] =[ Na +] and in consequence the apparent solubility
Sap p. = [ o n - ] i n i t i a l .
It may be concl uded f r om this t hat the appar ent sol ubi l i t y ( under st ood as
t he quant i t y of si l i ca that can be di ssol ved in a sol ut i on of sodi um hydr o-
xi de of gi ven concent r at i on and vol ume) can in no way lead to the t her mo-
dynami c sol ubi l i t y val ue, and t her ef or e can not be used to di st i ngui sh,
f r om t hi s poi nt of vi ew, di f f e r e nt t ypes of si l i ca ( f or e xa mpl e an
amor phous silica and a cr yst al l i zed silica).
On the ot her hand, this di st i nct i on can be made i f t he pH is fi xed, for
e xampl e by means of a buf f er sol ut i on. For a gi ven pH, t he appar ent
sol ubi l i t y is pr opor t i onal to t he t he r modynami c sol ubi l i t y, even t hough it
depends on the pH chosen. Thus, the buffer NH3/ NH4 + at pH 9.2 will give
a val ue 1.4 t i mes as l arge as a buffer at pH 7.
4. T h e d i s s ol u t i on rat e
The af f i ni t y of a r eact i on is def i ned at const ant pr essur e by t he rel at i on
A = - A G, in whi ch A G is the free ent hal py of reaction. It can be put into the
f or m A = RT Ln k/ a, wher e a is here the act i vi t y of Si O4H4 and k the
t h e r mo d y n a mi c sol ubi l i t y.
Ther ef or e A = RT Ln k/[SiOaH4] = RT Ln k' . [OH- ]/ [Si O4H3- ], whi ch shows
that t he affi ni t y of di ssol ut i on depends, via k, on the consi der ed silica, and,
for a gi ven silica, on t he under sat ur at i on fact or. It can r each high val ues i f
t he r at i o [ S i Oa H3 - ] / [ OH- ] is kept l ow, for exampl e by a pr eci pi t at i on
r e a c t i on.
It is cl ear that t he dynami c of di ssol ut i on is l i nked to the affi ni t y. In the
i deal case of a r ever si bl e r eact i on in a si ngl e step, t he l ogar i t hm of the
rat i o of t he rat e in di r ect i on 1 to di rect i on 2 is proport i onal to the affi ni t y.
Massar d and Gi r eaud (4) have not ed t hat , usi ng t he t he r modynami cs of
i r r ever si bl e pr ocesses as basis, it is possi bl e to est abl i sh a phe nome no-
l ogi cal l aw of di ssol ut i on but, in general , it is al ways necessar y to resort
to di r ect e xpe r i me nt , in or der to de t e r mi ne t he cons t ant act i ng in a
ki net i c law.
It is possi bl e for exampl e to fol l ow the concent r at i on of silica in sol ut i on
versus t i me dur i ng the at t ack of a sand by a sol ut i on of sodi um hydr oxi de.
The r at e of t hi s pr ocess depends on a number of f act or s, such as t he
cont act ar ea, t he conce nt r at i on of t he base, t he t e mpe r at ur e , and t he
condi t i ons of agi t at i on.
946 R. Dron and F. Brivot Vol. 22, No. 5
The s t a nda r d t es t me t h o d AS T M C 289- 71 f or p o t e n t i a l r e a c t i vi t y of
aggr e gat e s e e ms to be a good appr oach t o such an exper i ment :
It s at i s f i e s t he r e qui r e me n t of cons t ant ar ea of at t ack, s i nce t he gr adi ng
is s peci f i ed.
The t e mp e r a t ur e of 80C is p r o ba bl y ade quat e , be c a us e it is r e a s ona bl e
t o b e l i e v e t hat d i s s o l u t i o n t a ke s pl a c e by t he s a me me c h a n i s m as at
a mbi e nt t e mpe r a t ur e , and t hat it is s i mpl y acce l e r at e d.
The s t andar di zed ver s i on has a we a k poi nt t hat is eas i l y i mpr ove d:
Si nce t he at t ack is st at i c, t he s upe r nat ant s ol ut i on is not in e qui l i br i um
wi t h t he s ol ut i on i mpr e gnat i ng t he sand at all t i mes.
In f act t he pr oc e dur e cal l s f or agi t at i ng f r om t i me t o t i me, but does not say
how or how of t en. Si nce t he t est l ast s 1 day and 1 ni ght , it is l ogi cal to
t hi nk t hat in mos t cas es , manual agi t at i on, e ve n r egul ar , wi l l be appl i ed
onl y dur i ng t he day and wi l l be omi t t ed at ni ght .
The i deal wo u l d be t o use p e r ma n e n t me c h a n i c a l agi t at i on to ke e p t he
me d i u m h o mo g e n e o u s at all t i mes. Ho we v e r , e ne r ge t i c agi t at i on l i kel y t o
c a us e at t r i t i on of t he sand mus t be avoi ded, and so magne t i c st i r r i ng must
be r ul ed out . It woul d mor e ove r , be uns ui t abl e as a means of achi e vi ng a
c o mp l e t e s us p e n s i o n of t he sand, gi ven t he val ue of t he s and/ l i quor mas s
rat i o. Pr e f e r e nc e wi l l t he r e f or e be gi ven t o a s ys t e m of s l ow t ur ni ng over ,
say by mo un t i n g t he t est cont ai ne r s on a ver t i cal whe e l r e vo l vi n g in an
o v e n .
By var yi ng t he c ont a c t t i me f r om 0 t o 24 hour s ( or f r om 0 t o 8 hour s)
f ol l owe d by t he de t e r mi nat i on of t he si l i ca in sol ut i on, it is pos s i bl e to pl ot
t he c ur ve [Si O2] = f( t ) and obt ai n t he s ol ubi l i zat i on r at e by de t e r mi ni ng
t he s l ope at t he i nf l ect i on poi nt .
Ki net i c s t udy of t he si l i ca di s s ol ut i on can pr ovi de ot her i nf or mat i on, such
as t he e ne r gy of act i vat i on, whi ch can be de t e r mi ne d by var yi ng t he t est
t emper at ur e, and t he or der of t he r eact i on wi t h r e s pe ct t o OH- .
The not i on of or der coul d be usef ul as a mor e r i gor ous wa y of appr oachi ng
t he p r o bl e m of t he me c h a n i s m of di s s ol ut i on.
As st at ed above , t he at t ack of si l i ca is a c ompl e x pr oce s s i ncl udi ng at l east
f our s t eps , s i nce e a c h a t om of s i l i con is l i nked t o t he l at t i ce by f our
s i l oxane br i dges . It is t he s l owe s t of t hese st eps t hat cont r ol s t he r at e of
t he pr oces s as a whol e.
The t o p o c h e mi c a l r upt ur e of t he f i r st t hr ee s i l oxane br i dge s is p o s s i bl y o f
or der 2 wi t h r e s pe ct t o [OH- ]
2 OH- + Si - O- Si . . . . . > 2 Si O- + H2 0 (a)
But it ma y occur in t wo st eps:
Vol. 22, No. 5 ALKALI-SILICA REACTION, CONCEPTS 947
OH- + Si -O-Si . . . . . > Si OH + SiO- ( bl )
t hen Si OH + OH- - . . . . > Si O- (b2)
Whe n t hr ee br i dges ar e br oken, t hi s pr ocess l eads to - O- Si O3H2- , whi ch a
s i ngl e OH- suf f i ces to separ at e f r om t he subst r at e:
OH- + Si -O-Si O3H2- - . . . . > SiO- + SiO4H3- (c)
The sl ow st eps ar e a priori ( a ) o r ( bl ) . It wo ul d t hus be pos s i bl e in
pr i nci pl e to know whi ch of t he mode s of at t ack, (a) or (b), cor r e s ponds to
r eal i t y, accor di ng to whe t he r t he e xpe r i me nt al or der is 2 or 1.
Nor can t he e ve n t ua l i t y of an or de r zer o, whi ch woul d me a n t hat t he
hydr oxyl i c at t ack is not t he cont r ol i ng st ep, be r ul ed out.
The f i ne ki net i c st udy coul d pr ovi de s ome addi t i onal i nf or mat i on, in par t i -
cul ar about t he " r e duct i on in al kal i ni t y".
Thi s not i on pos es a di f f i cul t pr obl e m of pr i nci pl e. It pos t ul at e s t he exi s-
t ence of a mor e or l ess t hi ck l ayer f or me d of si l i ci c shr eds r esul t i ng f r om
i n c o mp l e t e r upt ur e of t he i ni t i al ne t wor k.
Thi s l ayer ma y be r e gar de d as bei ng f or me d by a "si l i ca gel " of whi ch t he
si l anol s ar e i oni zed. Be c a us e of its f r act al t ype, it is e xt r e me l y br i t t l e and
can not ke e p its i nt egr i t y i f t he s ys t em is agi t at ed. Un d e r t hese condi t i ons ,
it t ends to f r a g me n t and di s s ol ve qui ckl y in t he al kal i ne l i quor . Onl y t he
de e p zones havi ng a l ow l evel of r upt ur e of t he s i l oxane br i dges r e mai n
at t ached to t he gr ai n.
Th e r e f o r e t he r e d uc t i o n i n al kal i ni t y de pe nds on t he e n e r g y of agi t at i on
and is r a t he r poor l y d e f i n e d in t he s t andar d c ondi t i ons of AS TM t est
C 289- 71.
For s uch i n f o r ma t i o n to be me a n i n g f u l , it mus t be d e t e r mi n e d un d e r
c o n d i t i o n s t hat ar e st at i c and n e ve r t h e l e s s such t hat t he s ol ut i on is in
equi l i br i um wi t h t he s ur f ace of t he gr ai ns at all t i mes. It is t hen ne ce s s ar y
t o adj us t t he l i quor / s a nd r at i o so t hat t her e is no s upe r nat ant . It t hen
be c o me s pos s i bl e to f ol l ow t he r e duct i on of al kal i ni t y and so c ha r a c t e r i ze
t he ki ne t i c s of t he r e a c t i o n of t o p o c h e mi c a l br e a ki n g of t he s i l oxane
br i dges by t he OH- ion.
Concl usi on
Fr om t he poi nt of vi ew of t he conce pt s de ve l ope d above, t he mor e or l ess
gr eat s ens i bi l i t y of a gi ven si l i ceous aggr e gat e wi t h r e gar ds t o t he al kal i -
si l i ca r e act i on is r el at ed to its abi l i t y to di ssol ve i nt o i oni c f or ms in al kal i ne
medi a. Th e s e i oni c speci es ar e i nde e d abs ol ut e l y n e c e s s a r y to t he t hr ough
sol ut i on f or mat i on of t he de l e t e r i ous c o mp o un d CK- S- H.