THE AMERICAN MINERAI,OGIST, VOL, 46, NOVEMBER_DECEMBER, 1961
IONIC COORDINATION IN ALUMINO-SIICIC GELSIN RELATION TO CLAY MINERAL FORMATION
C. DB Krurr, M. C. G,q,srucsB* ano G. W. Bnrrvor,evt
Assrnacr
Syntheses of aluminum and magnesium silicates have been carried out at low tempera-tures and normal pressures, with the production of various proportions of gels and crystal-Iine phases. with aluminum, the gel phase is the more abundant and identification of thecrystals is possible only by electron diffraction; with magnesium, the yield in crystals ismuch higher and r-ray identification is possible. It is shown that the properties of the gelsinfluence the kind of crystals synthesized. The main factors are pH, salt concentrationand the ratio of aluminum or magnesium content to the silica content. For aluminum, thechange from six-fold to four-fold coordination increases with pH. Kaolinite has beenidentified at low pH and mica-like structures at higher pH. Serpentine minerals havebeen obtained in an intermediate pH range. The better yield of magnesium-bearingminerals may be attributed to the six-fold coordination of this cation.
INrnonucrroN
According to H6nin, Cail lbre and their collaborators (1953), the forma-tion of clay minerals under ordinary conditions of temperature and pres-sure appears to be determined by the existence of a brucite-type hy-droxide layer which, even in solution, induces the Sioa tetrahedra to de-velop a layer lattice. Various clay minerals are formed depending on thetype of cation involved and on the pH of the solution, with high pH,s(8-9) favoring formation of 2;1 lattices and lower pH's (6-7) favoring1: 1 lattices. fn the pH range 6-7, precipitation of aluminum hydroxideoccurs but, according to these authors, its rapid recrystallization toboehmite prevents the formation of kaolinite under these conditions.
The present experiments were undertaken with a view to studying theinfluence of simultaneous additions of Al and Si in the pH range ofaluminum hydroxide precipitation with the hope that the recrystalliza-tion of the hydroxide might be retarded and the formation of a layersilicate structure facilitated. Experiments were also carried out withmagnesium with a view to obtaining additional information on claymineral formation.
Part of the present study carried out in the University of Louvain hasbeen described by Gastuche and De Kimpe (1959) but subsequent work
x Laboratoire des Colloides des Sols Tropicaux, I.N.B.A.C., Institut Agronomique,H6verl6-Louvain, Belgium.
t Department of ceramic Technology, The Pennsylvania State university, universityPark, Pa. (U.S.A.).
contribution No. 60-50 from the college of Mineral rndustries, The pennsylvaniaState University, University Park, pa., U.S.A.
1370
IONIC COORDINATION IN ALUMINO-SILICIC GELS r37 |
at the Pennsylvania State University, utilizing additional techniques,now enables an integrated account to be given of these investigations.
Pnocr,ounB
The influence of the following factors on clay mineral formation wasstudied (see also Table 1) :
a) pH control,: Daily additions were made of normal solutions of either NaOH orHCI to maintain constancy of pH in a desired range between pH 4 and 9.
b) Salt concmtration: 'lwo series of experiments (experiments IIIa and IIIb) were
run with and without a soiution saturated with NaCl. High salt concentration favorsgel flocculation but no evidence of preferential orientation is found.
c) Alurninum in solution: Initially (experiments I and IIa) aluminum was brought intosolution by slow dissolutidn of a metallic plate, but subsequently (IIo, III) bydaily additions of an ionic form, AlCh 6HzO.
The experiments were made in a constant volume of 2 I which was maintainedconstant by evaporation of excess water; concentration increased, therefore, as thethe experiments progressed.
d) Sil,ica soraces: Either the sol, 'Ludox SM,' (exp. I) or ethyl silicate (expts. II & IIDwas added daily in small amounts. In one experiment (exp. I) no silica was addedapart from that extracted from the Pyrex flask.
e) The ratio AIzOt/ (AIzOz*SlO) : In the first experiment, where the amount of alumi-num dissolved from the plate depends on the pH, the relative content of this ele-ment is always low. Subsequently in order to keep this ratio as nearly as possible5070, Al ions were used.
Each experiment lasted for at least two months. Afterwards, the sam-ples were removed, washed and oven-dried at 105' C.
Srury oF THE GELSAlumina-silica gels
The slow addition of both Si and AI ions in the pH range studied wasthought to be favorable to the fixation of SiOa tetrahedra at the time offormation of the aluminum hydroxide framework. X-ray studies of suchgels give no indication of crystallized aluminum hydroxide, while in asimilar experiment carried out at pH 4.2 with the Pyrex glass as the onlysource of sil ica, boehmite was found as the final result, as proved by bothelectron and x-ray diffraction techniques.
To explain the iack of $-ray reflections in most experiments, it can besupposed that sil ica controls the development of the sample, imposingmainly its three-dimensional framework.
A closer study provides further information. It has been concluded byTamele (1950) and by Mill iken, et al. (1950), that in fresh alumina-silica gels, obtained by coprecipitation, the aluminum is involved in thefour-fold coordinated silica framework, provided the (AlrOa/AlrOi
R d
B !
E
o+
o
l a
t =
t ft Il 9
t -
I E
l "
II
tr
+
I
nif
o
lI
Ill
l o
t +t fl 9
l ( ,I
l3
l;
I Fi o
l "t ?l oI
l "Il rI
. = l6 l
T If l9 l
" lI
" lI
Eli
l e
t ?t . b
t -
I€
l 3l ' 6i ?
I
:9 Il_:
,x
ol - l -l " ; l *
l=l=9E
i l !
*l-; i . ;- j -
d l <s t +d t ^
t , hI
E
(J
za
O
7
E
O
z@ E
C)
zE
E ^ t ,^; l;
i
3
+
a
o, a p
t h "
il
H
rJ]
E
T372 C. DE KIMPE, M. C. GASTUCHE AND G. W. BRINDLEY
6g
z
z
E
X
t{s
a
trF
IONIC COORDINATION IN ALUMINO-SILICIC GELS 1373
+SiO, ratio remains less than 30-4070. As the ratio increases, the excess
aluminum may take six-fold coordination. According to Iler (1955), the
the following equil ibria can be considered:
IoI
o si-o oH
r;iil
For each four-fold coordinated aluminum ion in formula (i), one nega-
tive charge arises, which is balanced by a cation. Gels of this form are
stabilized under conditions of high pH and high salt concentration. With
decreasing pH, at about pH 4.5, the unstable acid form of the gel, for-
mula (ii), transforms quickly into an arrangement having uncharged six-
fold coordinated aluminum, formula (i i i).o) The change in the base exchange capacity (B.E.C.) with respect to
the relative content in AIzOr (Fig. 1), shows a maximum as already proved
by several authors, Tamele (1950), Mil l iken, et al. (1950), Bosmans and
Fripiat (1958). It can be interpreted as follows: As long as the (AlzOt/
AhOa*SiOr) ratio is small enough to allow the formation of tetrahedralaluminum, the charge will increase with aluminum contentl when the
ratio reaches the upper limit above which there are no further steric pos-
sibilities of Al-tetrahedra sharing corners with Si-tetrahedra, the charge
decreases, and the excess aluminum adopts octahedral coordination.
1374 C. DE KIMPE, M. C. GASTUCHE AND G. W. BRINDLEY
B E CPeq/9
500
400
300
200
t00
0
Frc. 1. Change in B.E.C. following rel-ative aluminum content for experimentL.
2 gta2C00
tat59
la258o
1atsro
r11160
aE.c.yeq./g
2000
4 5 6 7 8 9 p H
Frc. 2. Variation of B.E.C. accordingto pH of gel formation for the same rela-tive aluminum content. A. ExperimentIII' (in solution saturated in NaCl. B. Ex-periments II" and III1, (in dilute solution).
g t o %
Kaolinilc
8060t020
II
l
\
2
I /
7
pH!.N.cr (UrJ,
\ e l xe . r i . o r s - r
pH 5 - l { . ( M r
ex s. rrcr (!l! c)
Frc. 3. Change in the difiraction angle 20 with aluminum coordination.
IONIC COORDINATION IN ALUMINO-SILICIC GELS 1375
6) In experiments II and III, where the aluminum content was kept
constant, one observes the influence of pH and salt concentration on the
transformation of aluminum coordination. In Fig. 2, it is seen that the
aluminum coordination number was determined by o-ray fluorescence,
following the method described by white, et, al. (1958). In effect, one
measures the emission wavelength of AlKa, using a reflection from an
EDT crystal in the region oI 20:142.5". A General Electric XRD-5 *-
ray unit with a flow-proportional counter and helium path was em-
pllyed. Small but significant shifts in the angular position of the AlKa
lin" u.. obtained for difierent corrdination states. The results are cali-
brated by reference to AFvPOa and AIzvrSizOo(OH)e (kaolinite). A cali-
bration line is drawn between the 20-values obtained with AlPoa and
So m2l
100
OH(in arbilrary'
units )
50
15678gpHFrc. 4. Evolution of surface area and oH content following thepH of gel formaticn. A. Gels
prepared in concentrated Nacl solution. B. Gels prepared in dilute solution.
1376 C. DE KIMPE, M. C. GASTUCHE AND G. W. BRINDLEY
with kaolinite, as in Fig. 3, from which an estimate is obtained of theproportions of Al ions in the two coordination states in a mixture. Al-*gysh high accuracy cannnot be claimed for these estimates of Alrv andAlYr, nevertheless when the results for the gels are inserted in the dia-gram a clear trend is seen towards six-fold coordination as the pH con-ditions become more acidic.
c) valuable data also are provided by surface area (56) measurements,which are compared with the OH content of the gels (Fig. a). The formerdeterminations were made folrowing the Brunauer, Emmett and Teller!B_:E
f ) method, by adsorption of nitrogen at low temperature. TheOH.contents of the gels were obtained by infra_red spectrophotometry;the intensity of the OH stretching vibration band at Z.SSp *u, comparedwith the intensity of the HrO deformation band at 6p, and this ratio, whenmultiplied by the total water content of the gel determined by chemicalanalysis, gives a value related to the totar oH content of the sample.
The curve for the specific surfaces of gels obtained under condiiions ofgreat dilution (see lower part of Fig. 4), shows a maximum around pH5'S.which, according to rler (1955), is the pH of maximum instabil ity fora silica gel. rt is noticeable that this disorganization appears also in theformation of mixed gels.
Frc' 5. X-ray patterns of the rnagnesium products. A. Serpentine mineral, plus brucite.B' serpentine mineral; in the presence of Nacl, the reaction is complete and brucite iswholly transformed.
IONIC COORDINATION IN ALUMINO-SILICIC GELS
Frc. 6. Example of a fiber formed in
the experiment with magnesium and ethyl
silicate, in dilute solution.
Frc. 7. Formation of kaolinite crystals.
Notice the variable thickness of the
platy crystals. Experiment IIo; ionic alu-
minum and ethyl silicate at PI{ 4, 5.
The decrease in surface area observed at high pH can be explained by
a strain due to the water held by the cations. I ler (1955) admits this
possibil i ty for alumina-sil ica gels. Bosmans and Fripiat (1958) observe
also, for such freshly coprecipitated gels, that the maximum in B.E.C. at
about 30 to 40/o of relative aluminum content is connnected with the
minimum in surface area, measured by the B.E.T' method.The slow increase in OH groups along with increasing B.E.C. for
these samples (compare data in upper part of Fig. 4 with data in Fig. 2),
comes from the stabilization of the tetrahedral aluminum form. The low
surface area measured for the gel prepared at pH 4'5 comes along with a
sharp increase in OH groups and a very low B.E.C.; in this case, the six-
fold coordinated aluminum is stabilized.
Mognesia-sili,ca gels
Experiment 116 was performed in order to investigate the formation of
gels starting from a cation which takes only the six-fold coordination.
Two kinds of magnesium sources have been used. The ionic form,
MgCh'6HzO, was added slowly along with ethyl sil icate at pH 5'5' Two
1378 C. DE KIMPE, M. C. GASTUCEE AND G. W. BRINDLEY
other experiments were performed with meta[ic magnesium; one in asolution saturated in Nacl, the other in dilute solution. The metalquickly transformed into hydroxide; the reaction was more rapid inpresence of NaCl.
Though the pH had to be kept constant in the acid pH range, therapid dissolution of magnesium into hydroxide increased the pH upto 7.5.
The products obtained in these experiments are characterized by veryhigh specific surface, high water content and very low B.E.C.
Tnn Cnysrar,ltNr PrresrMagnesium proilucts
For the reasons given above, the yield in crystaline phase is importantand good cc-ray patterns of a 1: 1 serpentine mineral were obtained.
The presence of Nacl promoted considerabry the result; in the case ofdilute solutions, the brucite pattern is sti l l present (Fig. 5).
Frc. 8. Formation of mica-like plates. Crystal coming from experiment Io,with Ludox and metallic aluminum at oH 9.
IONIC COORDINATION IN ALUMINO-SILICIC GELS 1379
Frc. 9. Example of diffraction by pseudo-hexagonal crystal found at low pH, kaolinite.
Ring pattern formed by aluminum metal used as internal standard.
Under the electron microscope, two difierent crystalline forms wereobserved. A platy form was obtained which yielded an electron difirac-tion pattern showing a hexagonal distribution of spots with no evidencefor the long o-parameter of antigorite found by Zussman, et al. (1957).This form is therefore a platy serpentine, possibly lizardite, but completeidentification is not possible. At least it can be said that the platy form isnot antigorite in the strict sense of this term. Fibrous forms are also ob-served in the electron microscope and these appear to be chrysoti le(Fig. 6) .
Aluminum products
As said previously, the yield in crystalline phase was never sufficientto give a useful tc-ray pattern. Therefore, all the determinations on thesecrystals were made using electron microscope and electron difiractiontechniques.
1380 C. DE KIMPE, M. C. GASTUCHE AND G. W. BRINDLEY-
In most cases, the first step in the crystalline process appears to be theformation of plates with a widely variable thickness. Their developmentin the 001 plane depends on the pH. In the low pH range, several pseudo-hexagonal crystals were found (Fig. 7) some showin g 120" angles, verysimilar to those of a well crystallized kaolinite. In the high pH range, thecrystals are better developed in the 001 plane, but their morphology isless well defined. They do not show any obvious geometricar form exceptthat occasionally they are platy in appearance (Fig. 8).
The difiraction unit of the R.C.A. microscope E.M.U. 2D was used,equipped with an aperture specially designed by Charteron and Oberlin(1956) which permitted selected area diffraction. Single-crystar patternswere obtained showing hkj rcflections. rdentification on the basis ofthese reflections is difficult and has been discussed fully by Brindley andDe Kimpe (1961). They find that a clear disrinction between the hkTdiagrams of different layer lattices is possible only by accurate measure-ment of the lattice parameters, using metallic aluminum shadowing as aninternal standard. By this method it is possible to measure the D-param-eter with an accuracy better than +0.270, which is sufficient to dis-tinguish between the principal clay lattices (see Fig. 9).
Using this method, D-parameters were measured for crystals among thegel phase: At pH 4.5, the b-parameter is 8.93+0.03 A. ft.un be corre-Iated with kaolinite. At higher pH, a value of D equal to 9.02+0.02 Acorresponds to a mica or micaJike mineral. The sodic mica, paragonite,is a possible explanation.
CoNcrusroNs
Though the yield of crystalline phase is very poor in the experimentscarried out in the presence of aluminum, a detailed study of the gelphase shows its tendency to organization, confirmed by the crystallinephase study. Kaolinite appears at low pH, where the six-fold coordinatedstructure of aluminum is stabilized. At higher pH, a mica-like claymineral appears, while the four-fold coordinated aluminum increases inthe gel structure.
The low yield in crystalline phase can be attributed to:
(i) The insolubility of silica at the pH studied induced the polymeriza-tion of an alumina-silica gel.
(ii) The aluminum ion easily gives an isomorphous substitution withsilicon, the hexacoordinated form being stable only at low pH.
The latter difficulty does not arise with magnesium ions and in con-sequence magnesian clays are developed more easily.
IONIC COORDINATION IN ALUMINO-SILICIC GELS 1381
AcrNowr-BncMENTS
Thanks are due to Professor J. J. Fripiat for his interest and encourage-
ment dur ing th is work.Some of the experimental measurements were performed at the De-
partment of Ceramic Technology, The Pennsylvania State University,
and form part of the program of Project 55, sponsored by the American
Petroleum Institute to whom our thanks are due.
RETERENCES
Bosua.xs, H., auo Fnrrtar, J. (1958), Ionenuitwisselingsvermogen en oppervlakte ladings-
dichtheid van SiOTAI:Oa gelen. Pedologi'e,8, 184-198'
BrrNol-ov, G. W., arn De KrlIPr, c. (196-), Identification of clay minerals by single
crystal electron diffraction. Am. M ineral,. 46' 1005-1016'
carr_r,nn.E, S., eNo HnNrx, s. (1948), Essais de synthbse des min6raux argileux. verres et
Silicales Ind., 12, 63-64.
Ceu-r-rnn, S., Her.rrN, S., eNo EsqurvrN, J. (1953), Recherches sur la synthbse des min-
eraux argileux. Bul,l,. Soe. Fr. Min',761 300-314.
cnanrenoN, R., ewo OnnnuN, A. (1956), Description d'un selecteur pour microdiffraction
6lectronique. Bul,l'. Soc. Fr. Mi'n. Crist.,79,479-490.
Gastucun, M. C., lNo Dn KruPE, C. (1959), Tentative de synthbse des argiles du groupe
du kaolin dans des conditions normales de temp6rature et de pression. Acail. Royol,e
ite Betrgi'qwe, Bul,l,. Cl. Sc.,1th series,45' 1087-1104.
I{nNru, S., eNo RosrcHnr, O. (1954), Resultats obtenus du cours de nouveaux essais de
synthbse des min6raux argileux. c. R. il.es R,uni,ons d.u Groupe Fr. d,es Argi'les,w,
Nouvelle serie no. | (1954) 19-22.
h,an, R. K. (1955), The colloid chemistry of silica and silicates. corneil univ. Press,
Ithaca, N. Y.
Mrlltrnlt, J. H., Mrlr,s, G. A , ,r'No Onr-ao, A. G. (1950), The chemical characteristics
and structure of cracking catalysts. Disc. Farad S oc., 8, 27 9'
Tlurr.e , M. W. (1950), Chemistry of the surface and the activity of alumino-silica cracking
catalysts. Disc. Farad. Soc., 8, 270-279.
wnrn, E., McKrNsrny, H., enn Blrns, T. F. (1958), Crystal chemical studies by x-ray
fluorescence. Seventh Ann. Conf. Ind' App. X-Ray Analysis, Univ' of Denver, 239-
1 A <
Zussuan, J., Bnrmr,nv, G. W , aNo Conrn, J. J. (1957), Electron difiraction studies of
serpentine minerals. A m. M i.ner al'., 42, 133-1 53.
Manuseript receiaed' January 18, 1961.