"dealumination" and aluminum intercalation of vermiculite

0 downloads 0 Views 975KB Size Report
was then ground to
Clays and Clay Minerals, Vol. 39, No. 3, pp. 270-280, 1991.

" D E A L U M I N A T I O N " A N D A L U M I N U M I N T E R C A L A T I O N OF VERMICULITE JEAN-BAPTISTE D'EsPINOSE DE LA CAILLERIEAND JOSI~J. FRIPIAT Department of Chemistry and Laboratory for Surface Studies, University of Wisconsin-Milwaukee P.O. Box 4t 3, Milwaukee, Wisconsin 53201 Abstract--Dealuminationof vermiculite was carried out using (NH4)2SiF6 solutions. The dealuminated products were studied by high-resolution solid state zgsi and 27A1nuclear magnetic resonance. A decrease in the cation-exchange capacity (CEC) resulted from the partial removal of A1 from the tetrahedral layer, which decreased the framework negative charge, and from the partial replacement of Mg by A1 in the octahedral layer, which increased its positive charge contribution. The lowest CEC was obtained by swelling the structure with butyl-ammonium prior to the reaction with (NH4)2SiF6. Thus, CECs in the range observed for beidellite were measured; however, the lowest (AI/Si)TMratio was still more than twice as high as in beidellite. In addition, the dealumination reaction yielded noncrystalline silica as a byproduct. In contact with a solution of A1 hydroxypolymer (Altz), the dealuminated vermiculite showed no 18-/~. reflection characteristic of Al~3-intercalated smectite; instead it showed an ill-defined interstratification. For some samples, however, a significant increase in the specific surface area (as much as 230 m2/g) was observed, suggesting that an intercalation of A1 moieties did occur. The 27A1resonance spectra of the intercalated structure showed at least two components in octahedral coordination. On thermal activation, a resonance line attributable to pentacoordinated A1 was observed. Key Words--Aluminum,Cation-exchange capacity, Dealumination, Intercalation, Nuclear magnetic resonance, Vermiculite.

INTRODUCTION Vermiculites are swelling trioctahedral micas conraining Al-for-Si substitutions in tetrahedral layers and AI-, Fe-, and Ti-for-Mg substitutions in octahedral layers. They have been extensively studied, and an excellent review of their hydrated and cation-exchanged structures was recently published by de la CaUe and Suquet (1988). Because of the tetrahedral and octahedral substitutions, the overall negative charge of the structure results from an imbalance between the negative charge of the tetrahedral layer a n d the excess positive charge of the octahedral layer. The present work was undertaken to investigate (1) the possibility of lowering the net negative charge by substituting A1 by Si in the tetrahedral layer, and (2) the possibility of pillaring the negative-charge-depleted vermiculite with the A1 hydroxypolymer Keggin-like cation A11304(OH)E4(H20)~27+(abbreviated here as Al~3). The procedure for substituting Si for A1 in the tetrahedral layer has been used successfully on zeolites (Breck et al., 1985). It consists of a mild treatment with (NHa)2SiF6 in aqueous solution. The use of this reagent for modifying layered silicates was suggested earlier by Miyake et al. (1987). As is shown below, the procedure suggested by Breck et al. (1985) resulted in a decrease of the cation-exchange capacity of vermiculite, the tetrahedral and octahedral compositions both being affected. Although Al species were introduced in the interlayer, as evidenced by a noticeable increase in specific surface area, Copyright 9 1991, The Clay Minerals Society

pillaring in the classical was not achieved, inasmuch as no rationale d(00/) reflections with d(00/) ~ 18/~ were observed. Instead, an ill-defined interstratified structure was obtained. This result is rich in information concerning the cation-exchange mechanism involving Al13. MATERIALS Vermiculite

Two different samples of vermiculite were studied. Low-charge vermiculite from Benahavis having a halfunit cell composition (L6pez Gonzalez and Barrales Rienda, 1972) of Mgo.24 Ca0.03 (Si2.sl AllAo Fe3+o.o9)TM (Mg2.46 Tio.,1 Fe3+0.43)vI O1o(OH)2 was ground to < 2 /,m (Stoke's radius). This material was converted into the sodium form by repeated contact with a 1 M NaC1 solution and stirring at room temperature. The excess NaC1 was washed out with distilled water, and the slurry was separated by centrifugation until C1- free. High charge Llano vermiculite having a half-unit cell composition (Raussel-Colom et aL, 1980) of Mgo.4s (Si2.Ts All.22) IV (Mga.94 AloA Tio.o2 FeS+0.0~)VlOlo(OH)2 was treated with dilute H2SO 4 until no magnesite could be found in the X-ray powder diffraction diagram. It was then ground to

0.6

o 0.4

0.2

,,,-

......................... 2

1

3

~/time

4

5

(hr)

Figure 1. Decrease of observed relative cation-exchange capacity (CEC) vs. square root of reaction time. R = [Si added as (NH4)2SiF6/(AI) nv] = 1; temperature = 60~ From left to right: (rq) Llano, Llanoa 12, Llanoa24, Llanoa36; ((3) Benahavis, Benahavisa; (A) Llanob24. Solid line = linear regression obtained for Llanoa a ~/t(hr); arrows represent errors in measured relative CEC. T o this point, the d e a l u m i n a t i o n process b e h a v e d as expected. C o n s e q u e n t l y the m a t e r i a l o b t a i n e d after 36 hr o f t r e a t m e n t " a " o f the coarse L l a n o sample (Llanoa36) was carefully inspected by 29Si a n d 27A1 M A S N M R . T h e relative C E C o f this sample was 0.33. T h e d e c o n v o l u t i o n o f the 295i and 27A1 signals are displayed in T a b l e 1. W h a t is not s h o w n in this table is the noticeable c o n t r i b u t i o n o f a b r o a d 29Si line centered at - 111 p p m , c o r r e s p o n d i n g to a 4 Q e n v i r o n m e n t . T h e initial sample had a n a r r o w and w e a k c o n t r i b u t i o n at a b o u t this frequency, w h i c h can be attributed to residual quartz or cristobalite. T h e ratio (AI/Si) TM in the v e r m i c u l i t e structure was o b t a i n e d quite accurately f r o m the e q u a t i o n

(Al

n

TM

~]

= n=0 ~ ~I(Si, nA1),

(1)

where I(Si,nAl) is the relative intensity o f the 3Q(Si,nAI) c o n t r i b u t i o n to the overall intensity o f the f r a m e w o r k 29Si. T h e results in T a b l e 1 were d u p l i c a t e d for o t h e r samples (vide infra).

Table 1. Magic-angle spinning nuclear magnetic resonance results obtained by deconvolution of the 29Si and 27A1spectra. Integrated intensities Vermiculite sample'

3Q (Si,2A1)

3Q (Si, IAI)

3Q (Si,0A1)

Na-Llano Na-Llanoa 36 z 6(TMS) (ppm) ~(AP +) (ppm)

41 28 -84

53 49 -88

6 22 -93

(A1/Si) Iv3

45% 35% 68

(AlVl/Altot~)

10% 16.5% 3

i See text for descriptions. 2 This is the coarse Llano submitted to dealumination procedure "a" during 36 hr (cation-exchange capacity = 70 meq/ 100 g). 3 Obtained using the Lowenstein rule (Eq. 1).

C Figure 2. Schematic representation of effect of ApV-by-SiTM substitution mimicking trend (described in Table 1) in variation of silicon environment. [3, A1; e , Si; O, O. (a) A1TMhas no a l TM as second neighbor. Its substitution by SiIv leads to the creation of four [Si,3Si] sites and the disappearance of three [Si, 1Si,2A1] sites. (b) A1TMhas one A1TMas second neighbor. Its substitution by Siw leads to the creation of three [Si,3 Si] sites and the disappearance of one [Si, 1Si,2A1] and of one [Si,2Si, IA1] sites. (c) the A1TMhas two A1TMas second neighbor. Its substitution by SiIv leads to the creation of one [Si,2AI, ISi] and of two [Si,3Si] sites and the disappearance of two [Si,2Si, 1A1] sites. F r o m the data r e p o r t e d to this point, d e a l u m i n a t i o n seems to h a v e affected significantly the site [Si, 1Si,2AI], i.e., a silicon site s u r r o u n d e d by one Si and two A1 atoms, as well as the site [Si,3Si]. T h e relative intensity o f the latter increased at the expense o f the former. T h i s o b s e r v a t i o n is schematically a c c o u n t e d for in Figure 2, in w h i c h the [Si,3AI] site has b e e n neglected because its c o n t r i b u t i o n was t o o w e a k to be clearly detected. I f a m a x i m u m o f two AI per p s e u d o - h e x a g o n a l ring existed for reasons o f electrostatic stability described by H e r r e r o (1985), the substitution o f one o f t h e m by Si increased the [Si,2Si, 1A1] contribution, whereas the substitution o f A1 by Si in this e n v i r o n m e n t increased the [Si,3Si] contribution. Hence, i f substitution occurred at r a n d o m , the relative intensity o f the (Si, 1A1) e n v i r o n m e n t did n o t change m u c h , as o b s e r v e d in T a ble 1. T h e overall (A1/Si) TM ratio decreased f r o m 45% to 35%. I f the d e a l u m i n a t i o n reaction affected only the tetrahedral layer, the relative CEC should h a v e decreased f r o m 1 to 0.84 instead o f f r o m 1 to 0.33, as was observed. O n the o t h e r hand, the ratio o f the sixfold c o o r d i n a t e d A1 signal to the overall 27A1 signal increased significantly f r o m 10% to 16.5% (Table 1). T h i s i n f o r m a t i o n is only q u a l i t a t i v e because the two 27A1 resonances at a b o u t - 6 8 and a b o u t 3 p p m do n o t b e h a v e similarly u n d e r the r a d i o - f r e q u e n c y excitation. A c c o r d i n g to the c h e m i c a l formula, in the initial ver-

Vol. 39, No. 3, 1991

Vermiculite dealumination and pillaring

miculite this ratio was about 9%. Thus, some of the A1 extracted from the tetrahedral layer was probably not washed out a s ( N H a ) 3 A 1 F 6 . The solubility of this salt was low, as indicated above, but the dilution of the reacting solution and the subsequent washings should have removed it. I n addition to observing an increase in the (AlVI/total A1) ratio, the very significant increase o f a 4Q component in the 29Si spectrum, which indicates the formation of extra-framework silica, cannot be neglected. If some kind of silica gel was formed at the expense of the reagent, it could have combined with the extracted A1, resulting in the formation of a noncrystalline silica-alumina whose 29Si resonance would have been split into a 4Q(Si,4Si) signal at about 110 p p m and a weaker 4Q(Si,3Si, 1A1) signal between - 1 0 0 and - 1 0 5 p p m (Man et al., 1990). If this gel had a low CEC, the overall observed CEC would have been less than what could have been predicted from the (A1/Si) TM ratio in the vermiculite framework. From these conditions, the lowering of the CEC may have resulted from three factors; (1) the removal of AITM from the tetrahedral layer, (2) the translocation of extracted A1 from the tetrahedral layer into the octahedral layer (AITM ~ A1v~) and extraction of Mg from the octahedral layer, and, (3) formation of extra-framework silica. The A1TM-~ A1vI translocation would have increased the positive charge in the octahedral layer and lowered the overall negative charge of the structure, already depleted by the AFV-by-SiTM substitution in the tetrahedral layer. The following scheme accounts for different reactions that could have occurred. The octahedral substitutions by other elements than A1 have been neglected for sake of clarity. -

(1) cation substitution in the tetrahedral layer: M+(x_y)[Si(4_x), Alx]IV[Mg(3-y), Aly]VlOlo(OS)2 + (x - x')(NH4)2SiF6 (NH4+)(x._y)[Si(4_x,), Alx,]tV[Mg(a_y), mly]VIOlo(OH)2 + [M(x_y)fNH4)(2x 3x,+y)][A1F6](x-x,).

(2)

If M ยง is N H Z (procedure "b"), the last right-hand m e m b e r is (x - x ' ) ( N H 4 ) 3 A 1 F 6 . (2) cation substitution in the octahedral layers, the last right-hand m e m b e r in (2) being abbreviated (x - x')M3A1F6. (x -- x')M3A1F6 + (Na4+)(x,_y)[Si(4_x,), Alx,]TM 9[Mg(3-r), Aly]VlO,o(OH)2 --. (NH4+)(x,_y,)[Si(a_x,), Alx,]IV[Mg(3_y,),Aly,]WOlo(OH)2 + [(x -- x') -- (y -- y')]NH4MgAIF6 + [(x - x') - (y - y')]M3A1F6, where y - y' -