Langmuir 1991, 7, 2636-2641
2636
Hydrolysis of (y -Aminopropyl)triethoxysilane-Silylated Imogolite and Formation of a Silylated Tubular Silicate-Layered Silicate Nanocomposite Leighta M. Johnson and Thomas J. Pinnavaia' Department of Chemistry and Center for Fundamental Materials Research, Michigan State University, East Lansing, Michigan 48824 Received March 28, 1991. In Final Form: June 17, 1991
The external AlOH and internal SiOH surfaces of the tubular aluminosilicate imogolite have been silylated by reaction with hydrolyzed (y-aminopropy1)triethoxysilane(APS)in acidic aqueous solution. By monitoring the depletion of APS from the imogolite surfaces upon dialysis against distilled water at 25 "C,we observed two pseudo-first-orderhydrolysis processes. The faster depletion process with hob = 6.7f 0.9 X 10-2 h-l was attributed to the hydrolysis of APS from the external AlOH surfaces of the tubes. This assignment was consistent with the depletion behavior of silylated y-alumina at comparable APS loadings in the range 46-78 wt 7' 6 APS. Also, proton-decoupled BSi MAS NMR spectroscopy indicated that APS-silylated imogolite and y-alumina possess similar siloxane polymer structures. The slower APS depletion process, for which hob = 1.4 f 0.1 X 10-2 h-l, was assigned to the hydrolysis of APS from the internal SiOH surfaces of the imogolite tubes. The reaction of the fully silylated tubes with Na+-montmorillonite afforded a tubular silicate-layered silicatenanocomposite in which the imogolite tubes and clay plates were irregularly aggregated, as judged by X-ray diffraction. Selective hydrolysis of APS from the external imogolite surfaces occurs upon washing of the irregularly aggregated products. This resulted in the formation of a well-ordered tubular silicate-layered silicate intercalation compound in which only the internal surfaces of imogolite were silylated. Introduction
The modification of oxide surfaces by coupling with (y-aminopropy1)triethoxysilane (APS)and other function-
alized organosilane coupling agents has found useful applications in the areas of electrochemistry, chromatography, catalysis, and composite materials.'* In some systems, however, such as those involvingFez03 and A1203 surfaces, the performance characteristics of the organosilane-oxide interface have been limited by the lability of the surface bonds in humid or aqueous environments.7-9 We recently have been concerned with the surface chemical properties of the novel tubular aluminosilicate imogolite.lO This compound has a molecularly regular outer diameter of -25 A and an inner channel diameter of -8.7 A.11 The outer surface consists of a gibbsite-like array of AlOH groups, whereas the inner surface is composed of a regular distribution of SiOH groups. Our earlier FTIR and 29Si MAS NMR studies provided evidence for the binding of APS polymer to the external alumina-like surfaces of imogolite,1°but it was not possible to determine whether the internal surfaces were silylated. Organosilane derivatives of imogolite are potentially useful pillaring reagents for the synthesis of a new class (1)Moses, P. R.; Wier, L. M.; Lennox, J. C.; Finklea, H. 0.;Murray, R. W. Anal. Chem. 1978,50, 576. (2) Silylated Surfaces; Leyden, D. E., Collins, W. T., Eds.; Gordon
and Breach New York, 1980. (3) Silanes, Surfaces and Interfaces; Leyden, D. E., Ed.; Gordon and Breach New York, 1985. (4) Chemically Modified Oxide Surfaces;Leyden, D. E., Collins, W. T.,Lochmuller, C. H., Eds.; Gordon and Breach New York, 1990. (5) Sung, C. S. P.; Lee, S. H.; Sung, N. H. Polym. Sci. Technol. 1980, 12B, 757. (6) Garbassi, F.; Occhiello, E.; Bastioli, C.; Romano, G.; J. Colloid Interface Sci. 1987, 117, 258. (7) Plueddemann, E. P. Silane Coupling Agents; Plenum: New York, nor\ 1
IJOL.
(8) Kulkarni, R. D.;Goddard, E. D.Int. J. Adhes. Adhes. 1980,1,73. (9) Furukawa, T.; Eib, N. K.; Mittal, K. L.; Anderson, H. R., Jr. J. Colloid Interface Sci. 1983, 96,322. (10) Johnson, L. M.; Pinnavaia, T. J. Langmuir 1990,6,307. (11)Cradwick, P. D. G.; Farmer, V. C.; Russell, J. D.; M u o n , C. R.; Wada, K.; Yoehinaga, N. Nature (London) Phys. Sci. 1972,240, 187.
0743-7463/91/2407-2636$02.50/0
of microporous tubular silicate-layered silicate (TSLS) intercalation compounds. Pristine imogolite itself functions as a pillaring reagent by intercalating into smectite clays at the monolayer level and forming regular intercalates that exhibit severalorders of 001X-rayreflections.12 However, the intercalated tubes are laterally in van der Waals contact; consequently, the microporosity of the TSLS complex is restricted primarily to the internal channel of the tubes. By increasing the size of the tubes through the coupling of organo groups, one might expect to increase the lateral separation between tubes and thereby generate hydrophobic micropores suitable for the adsorption of organic molecules from aqueous solution. Also, permanent intertube microporosity might be formed upon removal of the organo groups through calcination. The results of the present study demonstrate in part that APS-modified imogolite does indeed form nanocomposite materials when allowed to react with the smectite clay Na+-montmorillonite. In an effort to better understand the surface chemistry of APS-modified imogolite in the formation of these nanocomposites, we have undertaken a study of the hydrolytic lability of this material. Also, we have examined the hydrolysis of APS-modified y-alumina in order to facilitate a comparison between the hydrolytic properties of the APS-modified external AlOH and internal SiOH surfaces of imogolite. Experimental Section
Materials. Imogolitewas synthesizedaccordingto the method described by Farmer and FraseP and purified by dialysis againat deionizedwater. y-Alumina (CatapalB)was obtained from Vista Na+-montmorillonite(SWy-1,Source Clay MineralsRepository) was purified by sedimentation. APS was purchased from Petrarch Systems and used without further purification. Silylation Reactions. Silylation reactions with APS in aqueous solution were carried out using 0.17 and 0.08 wt 76 (12) Johnson, I. D.;Werpy, T. A.; Pinnavaia, T. J. J . Am. Chem. SOC.
1988,110,8545.
(13)Farmer, V. C.; Fraser, A. R. Deu. SedimentoL 1978,27,547.
0 1991 American Chemical Society
Hydrolysis of APS-Silylated Imogolite suspensions of y-alumina and imogolite, respectively. The mineral suepensions were adjusted to pH 4.2 with acetic acid before the addition of APS. A 2.0 wt % solution of APS was adjusted to pH 4.2 with acetic acid and allowed to age at room temperature for 0.5 h. The desired amounts of the APS solution were then added to the mineral suspensions, and the mixtures were allowed to stir overnight. The initial amount of organosilane added to each suspension equaled 78 and 85 wt % of the air-dried solid imogolite and y-alumina, respectively. These quantities of APS exceeded the amounts needed for monolayer coverage of the substrates. A 0.5 wt % solution of pure APS in deionized water also was prepared, and the pH was adjusted to 3.7 with 1.0 M acetic acid. Hydrolysis of Silylated Imogolite and 7-Alumina. The silylated imogolite and y-alumina suspensions were divided into 25-mL aliquots and placed into tubular cellulose dialysis membranes having a molecular weight cutoff of 12 000-14 000 (Spectrapor-Spectrum Medical Industries). The membrane tubes were subsequently placed into 5 L of deionized water. A membrane tube was removed every 2 h for the first 18 h, and at longer intervals thereafter. Each time a sample was removed, the deionized water was replaced. With the exception of the hydrolyzed APS, the dialyzed suspensions were air-dried, and the amounts of bound APS were analyzed by Fourier transform infrared (FTIR) spectroscopy. In the case of pristine APS, the depletion of hydrolyzed silane from the dialysis membrane was determined by allowing the solution to evaporate in air and weighing the residue recovered. (y-Aminopropy1)triethoxysilane-Montmorillonite Complex. A 1.33 wt % Na+-montmorillonite suspension was mixed with a known volume of a 1 wt % APS solution that had been adjusted to pH 3.7 by the addition of 1.0 M acetic acid. A portion of the mixture was allowed to dry in air to afford a composite product containing 57 wt % APS. A portion of the product was also washed with water and air-dried. Silylated Imogolite-Montmorillonite IntercalationComplex. A 0.1 wt % suspension of imogolite was combined with a known volume of a 1.0 wt % APS solution that had been adjusted to pH 3.7by the addition of 1.0 M acetic acid. The concentration of APS was made equal to 7 wt % of the components in suspension (imogolite and silane). Na+-montmorillonite in 1.33 wt % aqueous suspension was added to the APS-treated imogolite suspensions in a weight ratio of 2.5:l imogolite-montmorillonite. The mixture was stirred overnight, collected by centrifugation, resuspended in water, and air-dried. Physical Measurements. X-ray diffraction patterns were obtained for oriented film samples formed by air-drying the suspensions on the surface of a glass slide. A Rigaku Rotaflex Model RU-2OOBH diffractometer equipped with a copper target was used to measure basal spacings. FTIR vibrational spectra for the air-dried products were recorded over the frequency range from 4000 to 400 cm-1 on an IBM Model IR 40s instrument equipped with an IR44 workstation. The samples (40 mg) were weighed on an analytical balance and pressed into 2 wt % KBr pellets. The concentration of APS in each sample was determined from the absorbance ratios of selected APS and substrate stretching frequencies. The absorbance of the u,(NHs+) vibration at 1570 cm-I was used to determine APS concentrations for the APS-modified imogolite and y-alumina samples. The intensities of the silane vibrations were normalized with respect to the u(SiOA1) stretching band centered near 950 cm-I for imogolite and the 1070-cm-' u(AlOA1) vibration in y-alumina. APS concentrations determined in this manner were confirmed by elemental analysis. The lH-decoupled MAS 29% NMR spectra were obtained on a Varian VXR 400 spectrometer equipped with a Doty probe. The W i spectra were generated at a frequency of 79.459 MHz. A 30' pulse of 3 ps, a delay time of 20 s, and a sample spinning rate of 5 KHz were used. All chemical shifts were reported with reference to TMS. Carbon and hydrogen elemental analyses were performed by Galbraith Laboratories in Nashville, TN. Silicon and aluminum analyses were provided by the inorganic chemistry laboratory of the Department of Toxicology at Michigan State University. Three-point nitrogen BET surface areas for the materials were measured on a Quantachrome Quantasorb Jr. sorptometer.
Langmuir, Vol. 7, No. 11,1991 2637
4000 3500 3000 2500 2000 1500 1000 500
Wavenumber, cm-I Figure I. FTIR absorbance spectra (KBr pellets) of the substrates (A) imogolite and (B) y-alumina.
Results (y-Aminopropyl)triethoxysilane is known to hydrolyze rapidly in dilute aqueous solution to form the organotrisilanol and an organosiloxane dimer.7 Allowing the solution to evaporate in the open atmosphere leads to the protonation of the amino group by carbonic acid and to the condensation of APS units to form a (y-aminopropy1)siloxane polymer.1° When an oxide surface is exposed to a hydrolyzed APS solution, the APS initially becomes hydrogen bonded to the surface hydroxyl^.^ Dehydration normally results in the formation of a surface-coupled organosiloxanepolymer. The characterizationof the surfacebound polymer can be accomplished using FTIR and %Si MAS NMR spectroscopy. FTIR and %SiMAS NMR measurements from previous worklohave shown that a surface-bound polysiloxane forms on imogolite upon evaporation of a solutionof the substrate in 2.4 w t 96 APS at pH 3.6 in air at ambient temperature. This siloxane polymer, however, was readily dissociated from the imogolite surface upon dialysis against distilled water. In the present work we have monitored by FTIR spectroscopy the APS concentration throughout the dialysis of the APS-treated imogolite suspension in order to determine the hydrolytic lability of the silylated AlOH and SiOH surfaces. We also have examined the change in APS concentration with time for the silylated surface of y-alumina, a poorly ordered oxide having surface AlOH units. A comparison of the hydrolysis behavior could then be made for APS bound to the distinguishable internal SiOH and external AlOH surfaces of imogolite and the monofunctional surfaces of alumina having only AlOH sites. FTIR spectra for the unreacted imogolite and y-alumina substrates are shown in Figure 1. Imogolite exhibits two bands due to the presence of water, namely, the v(OH) stretching frequency near 3500 cm-l and the 6(HOH) deformation band at 1630cm-I. There also are two bands at 995 and 940 cm-' due to v(SiOA1) stretching modes that are indicative of the isolated orthosilicate units present in the imogolite structure.ll The y-alumina FTIR spectrum
Johnson and Pinnauaia
2638 Langmuir, Vol. 7, No. 11, 1991
I C ' 1 ' I . I ' 4000 3300 3000 2500 2000 1 do0 1do0 560 1
'
1
I
'
I
-
I
'
I
'
I
'
4000 3500 3000 2500 2000 1500 1000 500
Wavenu mbe r , cm-' Figure 2. FTIR absorbance spectra (KBr pellets) of APSmodified imogolite samples isolated after dialysis for the time intervals indicated. contains prominent hydroxyl stretches in the 30003700-cm-I range and a sharp v(AlOA1) band at 1070 cm-1. The FTIR absorbances characteristic of pristine APS polymer, formed by hydrolysis in the presence of acetic acid a t pH 3.7 and evaporation of the solution in air at room temperature, were analyzed in detail in our previous work.lO Two v,(SiOSi) stretching modes were assigned to absorbances at 1130 and 1030 cm-l.ll The v,(CHz) frequency of the propyl group and the alkylammonium ion stretching frequency, v,(NH3+), were assignedto bands a t 2926 and 1573 cm-l, respectively. Acetic acid, along with carbonic acid formed from atmospheric COz, caused protonation of the amino group in the polymer. Strong absorbances due to the counterions occurred a t 1400 cm-l for the acetate u,(RCOz-) stretch and at 1638 and 1336 cm-1 for the bicarbonate v,(COz-) and vS(COz-) absorbances, respectively. We consider next the FTIR properties of APS bound to imogolite and y-alumina a t initial loadings corresponding to 78 and 85 w t % APS, respectively. These loadings are in excess of the amounts necessary for monolayer coverage. On the basis of the Nz BET surface areas for imogolite (500 m2/g) and y-alumina (290 m2/g), a hydroxyl group surface concentration of 12-14/100 A2(as estimated from the 7-8-A2area occupied by the triangular Al(OH)3groups of an edge-shared A106 octahedral sheet), and the coupling of one RSi03 moiety to three surface OH groups, complete monolayer coverage could occur at 46 and 31 wt 5% APS, respectively. Of course, these values represent upper loading limits for ideal monolayer coverage because they are based on estimates that disregard the surface occupied by the organo groups and the likely formation of siloxane polymer. Vibrational frequencies attributable to bound polysiloxane were readily observed in the FTIR spectra of the initial APS-modified samples. Spectra for air-dried samples obtained after dialysis periods of 0,4,12, and 50 hare shown for the APS-modified imogolite and y-alumina systems in Figures 2 and 3, respectively. The polysiloxane in the fully loaded samples was identified by a band
-
Wavenumber, cm-' Figure 3. FTIR absorbance spectra (KBr pellets) of APSmodified y-alumina samples isolated after dialysis for the time intervals indicated. in the 1120-1140-~m-~ region, which was assigned to the asymmetric stretching frequency of SiOSi linkages." An alkylammonium ion stretching frequency, v,(RNH3+), similar to that observed in the pure polymer, was found near 1570 cm-'. The acetic acid responsible for the initial protonation of the amine gave rise to the strong acetate v,(RCOz-) band near 1400 cm-l. The amine also reacted with carbonic acid formed from atmospheric COz. Vibrations due to the presence of bicarbonate ion were visible at 1638 cm-l due to v,(COz-), and 1336 cm-l due to v,(COz-). The sharp band near 2930 cm-' has been assigned to the v,(CH2) frequency of the propyl group. As the dialysis time was increased, a reduction in siloxane absorbances relative to the substrate absorbances was observed for both APS-modified substrates. A comparison of the FTIR spectra for the pristine substrates in Figure 1 with the APS-modified substrates in Figures 2 and 3 provides verification for the retention of significant amounts of siloxane polymer by both substrates after 12 h of dialysis. The presence of the y-aminopropyl group is especially visible in the v,(CHz) stretching region near 2930 cm-l and in the v,(NH3+) region near 1570 cm-l. The latter absorbance was utilized for the determination of the APS concentration. After 12 h of dialysis, the concentration corresponded to 40 and 24 wt 5% APS for the imogolite and y-alumina substrates, respectively. The latter value represents the limit a t which reliable estimates of APS concentration could be made from the FTIR spectra. Plots of the log (wt ?6 APS) versus dialysis time for the pristine APS, APS-modified imogolite,and APS-modified y-alumina systems are shown in Figure 4. It is especially significant that the initial rate of depletion of pristine APS from the dialysis membrane is at least 5 tines faster than the depletion of bound APS from the surfaces of imogolite and y-alumina. That is, the monomeric and dimeric units that form upon APS hydrolysis diffuse relatively rapidly through the pores of the dialysis membrane. Thus, the depletion behavior for APS-modified (14)Chiang, C. H.; Ishida, H.; Koenig, J. L.J . Colloid Interface Sci. 1980, 74, 396.
Langmuir, Vol. 7,No. 11,1991 2639
Hydrolysis of APS-Silylated Zmogolite A APS
o APS-lmogolita APS-Alumina
n
*
v,
a
a
'
\
1.2
K
W
0 -m
*
A A
A
*
0.4
-67 -59 n
0.0
0
10
20
30
40
50
60
Time (hours)
Figure 4. "Plotsof log (wt % APS) versus dialysis time for hydrolyzed APS, APS-modified imogolite, and APS-modified y-alumina. imogolite and y-alumina is determined by the rate of hydrolytic displacement of the siloxane from the substrate surfaces. Although the results in Figure 4 for the depletion of APS from the surfaces of imogolite are limited to a rather narrow composition range, two pseudo-first-order depletion processes are evident. A best least-squares fit of the data at coverages corresponding to 46-78 wt % APS afforded an initial pseudo-first-order rate constant, hob = 6.7 f 0.0X 1W2h-l. The depletion rate for APS-modified y-alumina over a comparable composition range was h-l. This suggests that the similar, hob = 7.6f 0.9 X initial APS depletion process is similar for the two substrates. At coverages of 146 wt ?6 APS, the rate of siloxane depletion from imogolite decreased substantially, corresponding to a pseudo-first-order rate constant kobs = 1.4 f 0.1 X h-l. The 5-fold decrease in rate suggests that bimodal mechanisms operate in the depletion of siloxane from the surfaces of APS-modified imogolite. In contrast to the imogolite system, the depletion of siloxane from y-alumina markedly accelerated with decreasing APS surface coverage (cf. Figure 4). In order to better characterize the APS polymer bound to imogolite and y-alumina, the air-dried silylated oxides were investigated by 'H-decoupled 29SiMAS NMR. The spectrum for APS-modified imogolite at a 20 wt % concentration is shown in Figure 5A. This sample was prepared by direct reaction with APS at pH 3.7 and was not subjected to dialysis. However, dialyzed samples showed the same characteristic NMR resonances as nondialyzed samples at comparable APS loading. The spectrum in Figure 5A contains a very intense resonance a t -79 ppm due to the HOSi(0Al)senvironment of the unmodified imogolite orthosilicate units.l5J6 An envelope of resonances in the -45 to -70 ppm region contains a shoulder at -48 ppm and two resolved resonances at -58 and -68 ppm. These three resonances are assigned to the following silicon sites of a surface-bound siloxane polymer: RSi(OH)?(OM),RSi(0H)(OM)z, and RSi(OM)3, where M is an adjacent Si site in the surfacebound polymer, or an A1 or Si site at the imogolite surface. These assignments are in accord with the chemical shifts (15) Barron, P. F.; Wilson, M. A.; Campbell, A. S.;Frost, R. L. Nature 1982,299,616.
(l.6) Goodman, B. A.; Russell, J. D.; Montez, B.; Oldfield, E.; Kirkpatrick, R. J . Phys. Chem. Miner. 1985, 12, 342.
I
-501
1
1
1
1
1
-10 -30 -50
I
I
I
I
-70 -90
I
I
I
-120
I
I
I
i
ppm
Figure 5. 'H-decoupled %i MAS NMR spectra (79.459 MHz) of (A) 20 w t % APS modified imogolite and (B) 24 wt % APS modified y-alumina. observed for APS-modified silica surface^.^^-^^ A fourth broad resonance a t -90 ppm is assigned to silylated silicon sites in the imogolite structure. The 11 ppm upfield shift is consistent with the conversion of some imogolite Si(OH)(OAl)s sites at the inner surface of the tubes to Si(OSi)(OA1)3sites by reaction with APS. The 'H-decoupled %i MAS NMR spectrum of APSmodified y-A1203at a 24 wt % concentration is shown in Figure 5B. This sample also was prepared by the direct reaction with APS at pH 3.7 without undergoing dialysis. Resonances due to the surface-bound polymer at -50,-59, and -67 ppm were assigned to RSi(OH)2(0M),RSi(0H)(OM)z, and RSi(OM)3 (M = Si, Al) sites, respectively. The resonances here are sharper than the corresponding lines for APS-modified imogolite,suggestingthat the APS siting is less heterogeneous on the monofunctional alumina surface. We next investigated whether APS-modified imogolite could be intercalated into the galleries of a Na+-montmorillonite clay to form a silylated tubular silicate-layered silicate intercalation complex. The intercalation of APSmodified imogolite into Na+-montmorillonite was anticipated, because the positive charge provided by the APS polymer should allow for electrostatic interaction between the negatively charged clay layers. However, hydrolyzed APS itself may compete with APS-modified imogolite for intercalation in montmorillonite. Therefore, prior to investigating the reaction of APS-modified imogolite with Na+-montmorillonite, we examined the reaction of the clay with hydrolyzed APS at pH 3.7. Parts A and B of Figure 6 compare the X-ray diffraction (17) Caravajal, G. S.; Leyden, D. E.; Maciel, G. E. In Silanes, Surfaces and Interfaces; Leyden, D. E., Ed.; Gordon and Breach: New York, 1985;
pp 283-303. (18) Sudhoter, E. J. R.; Hub, R.; Hays, G. R.; Alma, N. C. M. J. Colloid Interface Sei. 1985, 103, 554. (19) De Haan, J. W.; Van Den Ven, L. J. M. J. Colloid Interface Sci. 1986, 110, 591. (20) Rudzinski, W. E.; Montgomery, T. L.; Frye, J. S.;Hawkins, B. L.; Maciel, G. E. J. Chromatogr. 1985, 323, 281.
Johnson and Pinnavah
2640 Langmuir, Vol. 7, No.11, 1991
'v(
-
APS-mont. unwashed
2 I-
