- Email: [email protected]
Effect of Structural Stress on the Intercalation Rate of Kaolinite
Journal of Colloid and Interface Science 250, 379–393 (2002) doi:10.1006/jcis.2001.8208
Effect of Structural Stress on the Intercalation Rate of Kaolinite Youjun Deng, G. Norman White, and Joe B. Dixon1 Department of Soil and Crop Sciences, Texas A&M University, College Station, Texas 77843-2474 Received July 16, 2001; accepted December 27, 2001; published online May 15, 2002
Particle size in kaolinite intercalation showed an inverse reactivity trend compared with most chemical reactions: finer particles had lower reactivity and some of the fine particles cannot be intercalated. Although this phenomenon was noted in the early 1960s and several hypotheses have been reported, there is no widely accepted theory about the unusual particle size response in the intercalation. We propose that structural stress is a controlling factor in the intercalation and the stress contributes to the higher reactivity of the coarser particles. In this study, we checked the structural deformation spectroscopically and indirectly proved the structural stress hypothesis. A Georgia kaolinite was separated into nine size fractions and their intercalations by hydrazine monohydrate and potassium acetate were investigated with X-ray diffraction (XRD) and Fourier-transform infrared (FTIR) analyses. The apical Si–O band of kaolinite at 1115 cm−1 shifted to 1124 cm−1 when the mineral was intercalated to 1.03 nm by hydrazine monohydrate, and its strong pleochroic properties became much weaker. Similar reduction in pleochroism was observed on the surface OH bands of kaolinite after intercalation. Both the bending vibrations of the inner OH group at 914 cm−1 and of the surface OH group at 937 cm−1 shifted to 903 cm−1 after intercalation by hydrazine. A new band for the inner OH group appeared at 3611 cm−1 during the deintercalation of the 1.03 nm hydrazine kaolinite complex. Pleochroism change in the apical Si–O band suggested the tetrahedra had increased tilt with respect to the (001) plane. The tilt of the Si–O apical bond could occur only if the octahedra had also undergone structural rearrangement during intercalation. These changes in the octahedral and tetrahedral sheets represent some change in the manner of compensation for the structural misfit of the tetrahedral sheet and octahedral sheet. As the lateral dimensions of a kaolinite particle increases, the cumulative degree of misfit increases. Intercalation breaks the hydrogen bonds between layers and allows for the structure to reduce the accumulated stress in some other manner. The reversed size effect on intercalation probably was not caused by crystallinity differences as reported in the literature, because the Hinckley and Lietard crystallinity indices of the four clay fractions were very close to each other. Impurities, such as dickiteor nacrite-like phases are not significant in the studied sample as suggested by the XRD and IR results, they are not the main reasons for the lower reactivity of the finer particles. C 2002 Elsevier Science (USA) Key Words: crystallinity; hydrazine; intercalation; kaolinite; particle size; pleochroism; potassium acetate; structural misfit; structural stress.
1 To whom correspondence should be addressed. Fax: (979)845-0456. E-mail: [email protected].
379
INTRODUCTION
Intercalation of kaolin minerals with inorganic and organic compounds has wide potential for scientific and industrial applications. Intercalation of kaolinite often induces structural perturbations. Therefore, the guest intercalating compounds can function as probes in the investigation of the crystallography and reactivity of kaolin minerals (1–4). Soil kaolinite and halloysite can be differentiated using the differences in intercalation rate by formamide (5, 6). Kaolinite has been modified to become hydrophobic with long-chain alkylamines (7, 8), fatty acid salts (9, 10), and several polymers (11–13). The modified kaolinite minerals have special thermal and mechanical properties that can be used in the polymer industry. The intercalation reaction has also been used to delaminate coarse kaolinite (14). The study on kaolinite intercalation began in the 1960s (15, 16), and numerous publications are available on this subject. It is commonly observed that particle size has a reversed reactivity trend in kaolinite’s intercalation compared with other chemical reactions: the finer kaolinite particles are intercalated more slowly than coarse particles. Some fine kaolinite particles cannot be intercalated at all. Several research groups have discussed the possible reasons, no agreement has been reached. In the intercalation study of kaolinite by potassium acetate, Wiewiora and Brindley (17) reported that the intercalation of the >1 µm kaolinite by potassium acetate (KOAc) was essentially complete within 1 week, less than 4% of the <0.5-µm fraction intercalated in a 1-week reaction period. Weiss (18) reported a similar reactivity trend in the intercalation of kaolinite by urea: when particle size decreased from 9.5 to 0.68 µm, the intercalation rate first increased to a maximum value in the 3.8- to 5.0-µm fraction and then decreased. It took 65 days for the 0.68- to 0.80-µm fraction to reach complete intercalation but 20 days for the 3.8to 5.0-µm fraction. Gomes (19) investigated the intercalation capacities of the <0.5-µm particles from five kaolinite samples by hydrazine and also showed that the finer particles were intercalated more slowly than the coarser ones. Uwins et al. (20) examined the intercalation rates of a wide range of kaolinites with different particle size, shape, and defect distributions. They first treated the samples with hydrazine hydrate and then investigated the reactivity of hydrazine-treated kaolinite toward N methylformamide (NMF). The >2-µm fractions reacted to near completion overnight, but less than 28% of the <0.3 µm could be intercalated by NMF even with longer treatments. The lower 0021-9797/02 $35.00
C 2002 Elsevier Science (USA)
All rights reserved.
380
DENG, WHITE, AND DIXON
reactivity of the finer kaolinite particles in the intercalation was also reported by Chen et al. (21, 22). The lower reactivity of the finer kaolinite particles was attributed to their higher crystallinity by Wiewiora and Brindley (17). High crystallinity kaolinites, however, have been reported as undergoing fast intercalation by potassium acetate in several papers (23–25). Well-ordered kaolinite also showed higher intercalation yield than the poorly ordered kaolinite in the intercalation by NMF as reported by Uwins (20). Furthermore, the finer fractions can have poorer crystallinity as well as better crystallinity than the coarser fractions depending on the origin of the mineral (26). The facts that there is no consistent trend in crystallinity associated with particle size and that high crystalline kaolin minerals are more responsive to intercalation suggest that the “better crystallinity” is not likely the reason why the finer kaolinite particles are intercalated more slowly or even cannot be intercalated at all. Weiss et al. (18) explained the reversed particle size effect using a “ring mechanism.” According to their hypothesis, the intercalating molecules act as wedges at kaolinite edges and cause the layers to be elastically deformed. The length of elastically deformed zone in the crystal is termed “cooperative action length,” a0 . The intercalating molecules penetrate faster in the deformed zone because of the weaker cohesion in the zone. One factor in determining the cooperative action length is the elastic properties of the layers. Particles with better crystallinity have longer a0 and intercalate faster. The intercalation on large crystals can start from all edges giving a “ring mechanism” and therefore have faster reaction rate. For small crystals, the intercalation cannot undergo from all edges because the elastic deformation induced by the intercalation from an edge are transmitted across the particle and cause contractions of the layers on the opposite edge. The intercalation on small particles proceeds at a slow rate via a “one-side mechanism.” By this mechanism, the high-crystallinity crystals should be more difficult to intercalate because it requires more energy to open the longer “cooperative action zone.” This mechanism cannot explain why certain fine kaolinite cannot be intercalated. Raussell-Colom and Serratosa (27) summarized the factors in determining the intercalation capacity of kaolinite. They concluded that the intercalation capacities depend on the cohesion energies of the kaolinite stackings. The stacking in kaolinite is strongly correlated with structural disorder or crystal imperfection. The reason why intercalation capacity depends on particle size is because crystal size is related to crystal imperfection. It has been found that the imperfections or defects in kaolinite are largely due to errors in the position of the vacant octahedral site and the errors constitute the introduction of dickite layers (28). The IR bands of dickite and nacrite have been identified in some fractionated kaolinite samples at low temperatures. The overall proportion of these dickite and nacrite stacking sequences increased with decreasing particle size in some kaolinite samples (29, 30). Dickite, nacrite, and kaolinite are 1 : 1 layer silicate minerals; they have identical chemical compositions but
differ structurally by rotations between layers, which causes changes in hydrogen bonding. The differences in hydrogen bonding may result in different reactivity to the intercalating agents. The appearance of these stacking faults in kaolinite may affect the overall intercalation rate. For example, dickite has been observed to have a slower intercalation rate than kaolinite when it is treated with hydrazine monohydrate; however, a high proportion of 5- to 10-µm dickite with low defects could be intercalated to 1.024 nm by hydrazine monohydrate and the intercalation was nearly complete within a 2-day reaction period (31). Similarly, nacrite can be intercalated to near complete by KOAc or N -methylacetatmide in a few days (32–35). The near completion of the intercalations of dickite and nacrite suggests that dickite- or nacrite-like phases in kaolinite should be intercalated too if a long reaction time, for example, a few days, is allowed. In other words, dickite, nacrite, or the stacking errors in kaolinite are not likely a main reason for the slow intercalation of the fine kaolinite particles. Gomes (19) reported another type of defect, that the presence of randomly interstratified single layers of smectite or illite are strongly correlated to the intercalation properties. These types of defects are more abundant in the smaller particle size fractions and are generally responsible for the asymmetry of the 001 X-ray reflections and the lower intercalation capacity. The interstratification separates the crystal to more small domains, and it cannot explain why the small kaolinite domains are less active in the intercalations. Furthermore, the asymmetry of the 001 XRD peak seemed more dependent on the origin of the sample. The asymmetry indices of the <0.1- and 0.1- to 0.2-µm fractions of an Ireland “bondclay” kaolinite are almost the same, but only about 13% of the <0.1-µm fraction was intercalated, whereas about 35% of the 0.1- to 0.2-µm fraction was intercalated (19). The same asymmetry indices but different intercalation capacities suggested the particle size is more important than the defects caused by the random interstratification of illite or smectite in determining of the intercalation rate or capacity. Uwins et al. (20) also similarly concluded that particle size appeared to be a more significant controlling factor than defect distribution in determining the relative yield of kaolinite–NMF intercalate. We propose that structural stress is a driving force in the intercalation. Kaolinite is a 1 : 1 layer silicate with a structure consisting of a tetrahedral Si sheet sharing apical O2− ions with an octahedral Al sheet (Fig. 1). The lateral size of the tetrahedral sheet is larger than the octahedral sheet. Structural stress is induced from the misfit of the tetrahedral sheet and the octahedral sheet in the 1 : 1-type linkage. Part of the structural stress can be relieved when kaolinite is intercalated. Coarser particles intercalate faster due to their larger structural stress. Because the stress is associated with the misfit of the tetrahedral and the octahedral sheets, it is expected that some deformation within the sheets will occur when the constraints resulting from hydrogen bonding between layers is reduced. Determining the changes in octahedral and tetrahedral conformation should provide some indirect evidence of the structural stress change.
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 1.
381
Structural models of kaolinite (according to Bish (36)).
The main objective of this study was to test the hypothesis that structural stress is a controlling factor in the intercalation of kaolinite by documenting structural changes in the octahedra and tetrahedra following intercalation with infrared data. Meanwhile, it is endeavored to discuss the possibilities of the hypotheses that impurities, stacking faults, or crystallinity are the reasons that the fine particles are intercalated more slowly. MATERIALS AND METHODS
Kaolinite and Intercalating Agents A Georgia kaolinite sample, provided by the J. M. Huber Corporation, was used in this study. The sample was treated with 1 M, pH 5, sodium acetate buffer solution to ensure that it was free from carbonate minerals and exchangeable high-valence cations (37). The sample was separated into nine fractions by sieving for the sand fraction (50–500 µm), settling for the silt fractions (2–5, 5–10, 10–20, and 20–50 µm), and centrifugation for the clay fractions (<0.2, 0.2–0.5, 0.5–1.0, and 1.0–2.0 µm). A solution containing 3 g Na2 CO3 per 20 L with a pH about 10 was used to disperse the sample. All of the fractions were oven dried at 60◦ C. The Hinckley (38) and Lietard (39) crystallinity indices were determined on the kaolinite in each fraction. It should be noted that determination of the Lietard crystallinity index is hindered by the presence of mica in the coarser fractions. Hydrazine monohydrate and saturated KOAc solutions were used as the intercalating agents in this study. Potassium acetate obtained from Aldrich Chemical Company, and hydrazine monohydrate (H2 NNH2 · H2 O, 99.8%) obtained from Fisher Scientific, were used without further purification. Monitoring Intercalation Rate Two methods were employed to study the intercalation rates of kaolinite by hydrazine. To study the initial reaction, dry kaolinite
powder was mounted on a quartz slide, and four to six drops of hydrazine monohydrate were added on the powder, immediately followed by 10 continuous XRD analyses in the 5–15◦ 2θ range (2). Typically, one run of the XRD took 18 min. For the <1-µm clay fractions, it was found that the intercalation was not complete within 3 h. To monitor the long-term intercalation rate, a 0.5-g sample from each fraction was immersed in hydrazine monohydrate in a 40-mL Nalgene centrifuge tube for various lengths of time. The tube was capped and shaken frequently. To measure the XRD pattern of the sample, the particles were concentrated by centrifugation, and a small amount of clay paste was smeared on a glass slide as a thin film. The film was analyzed with XRD immediately in the 5–15◦ 2θ range. The intercalation of kaolinite by saturated KOAc solution on quartz slide was similarly monitored by XRD as the first method used for the intercalation by hydrazine, except a broader 2θ range of 5 to 28◦ was monitored. Sample Preparation for Infrared Analysis Because the 1- to 2-µm size fraction contains no other minerals that are detectable by XRD or IR and has fast and near complete intercalation by hydrazine monohydrate, only the hydrazine monohydrate intercalated 1- to 2-µm kaolinite was examined by IR analysis. One gram of 1- to 2-µm kaolinite was mixed with 5 mL of hydrazine monohydrate in a Nalgene centrifuge tube. After more than 3 weeks’ shaking, a subsample of the suspension was diluted with hydrazine monohydrate. Six drops of the diluted suspension, containing about 1 mg of kaolinite, were pipetted onto a 32 × 2-mm ZnS Irtran window. The window was heated at about 45◦ C with a 250-W infrared heat lamp to accelerate the evaporation of excess hydrazine. When the evaporation of the free hydrazine monohydrate was nearly complete, the sample was checked with the X-ray diffractometer in an open atmosphere. After XRD examination, an additional
382
DENG, WHITE, AND DIXON
ZnS window and a rubber O-ring washer were used to seal the nearly dry film between the two windows. The two windows were mounted on an IR sample holder, and the film orientation was changed by rotating the windows to 0◦ and 45◦ with respect to the perpendicular direction of the IR beam. The film was alternately monitored with the X-ray diffractometer in open atmosphere, and with the infrared spectrometer in the sealed windows. The purpose of sealing the film between the two windows in the IR analysis was to reduce the evaporation while the incidence angle was changed. The alternate XRD and IR monitoring were repeated more than 10 times until the hydrazine IR bands had nearly disappeared from the IR spectrum. The potassium acetate–kaolinite intercalates were not investigated with IR spectroscopy because the intercalation of the clay fractions was not complete even after 1 month of reaction. It was also difficult to remove the excess salt without causing collapse of the intercalates. X-Ray Diffraction Analysis XRD analyses were performed with CuK α radiation on a Philips diffractometer. Operational conditions of 30 kV and
FIG. 2.
18 mA were used. The goniometer was equipped with a curved graphite beam monochromator. The measurements were conducted in the step-scanning mode, with a step size of 0.05◦ 2θ and a counting time of 5 s at each step. Fourier-Transform Infrared Spectroscopy IR spectra were recorded using a Perkin-Elmer 2000 system Fourier-transform IR spectrometer. For each spectrum, a total of 64 scans were obtained at a resolution of 1 cm−1 and a data collecting interval of 0.2 cm−1 . A triglycine sulfate (TGS) detector was used for the analyses. The optical bench of the instrument and the sample chamber were purged with dry air produced by a Balston 75-52 FTIR purge gas generator. RESULTS
Crystallinity of Different Size Fractions The XRD patterns of the nine size fractions used in this study (Fig. 2) demonstrated that the dominant mineral in all fractions was kaolinite. As indicated by the poorer resolution, broader peak width and weaker intensities of the prismatic peaks at
Powder XRD patterns of the different size fractions of a kaolinite sample.
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
TABLE 1 Particle Size Distribution and Crystallinity Indices of the Kaolinite Fractions Used in This Study Particle size (µm)
Percentage (%, w/w)
Hinckley indexa
Lietard R1 indexb
Lietard R2 indexb
<0.2 0.2–0.5 0.5–1.0 1.0–2.0 2.0–5.0 5.0–10 10–20 20–50 50–500
5.3 23.6 20.8 8.8 23.7 9.8 2.5 2.3 0.2
0.22 0.42 0.29 0.49 0.49 0.64 0.66 0.58 0.43
0.34 0.34 0.39 0.44 0.45 0.46 0.49 0.48 1.06
0.51 0.52 0.47 0.57 0.71 0.74 0.81 0.81 1.14
a
Calculated according to Hinckley (38). Calculated according to Lietard (39). The value for the Lietard crystallinity for the 50- to 500-µm fraction is the result of interference by mica peak overlaps. b
0.444, 0.434, 0.416, 0.256, 0.249, 0.238, and 0.229 nm, the finer fractions were generally less ordered than the coarser fractions. The small particle size also resulted in weak diffraction of the finer fractions. The crystallinity of kaolinite as characterized by the Hinckley (38) and Lietard (39) indices (Table 1). The
FIG. 3.
383
Hinckley indices of the fractions were less than 0.7, indicating the crystallinity of the samples were lower than some low defect kaolinites such as Keokuk kaolinite which has a Hinckley index of 1.54 (40). The clay particles (<2 µm) have lower crystallinity indices than the silt fractions (2–50 µm). The crystallinity indices among the four clay fractions (<0.2, 0.2–0.5, 0.5–1, and 1–2 µm) are very close to each other and there is no consistent trend associated with the particle size indicating their similarity in crystallinity. Infrared Analyses of Clay Size Fractions of the Kaolinite The four clay fractions showed almost identical IR characteristics in terms of band position, shape, intensity, and intensity response to the change of incidence angles (Fig. 3). The four IR bands at 3696, 3668, 3652, and 3619 cm−1 are the characteristic OH stretching vibration bands of kaolinite. The 3696 cm−1 band had strong pleochroism as shown by its increasing absorption when the incidence angle was changed from 0◦ (Fig. 3a) to 45◦ (Fig. 3b). The other bands at 1115 (apical Si–O), 1033, 1008 (Si–O–Si in-plane vibrations), 937, 914 (OH bending vibrations), 795, and 755 cm−1 (OH translational vibrations) are also typical bands of kaolinite. The strong IR pleochroism of
FTIR spectra of different kaolinite size fractions at 0 and 45◦ IR beam incidence angles.
384
DENG, WHITE, AND DIXON
FIG. 4.
Selected XRD patterns of 1- to 2-µm kaolinite during its intercalation (a) by hydrazine monohydrate and (b) by saturated KOAc solution.
the 1115 cm−1 band is demonstrated by its intensity increase when the incidence angle was changed from 0◦ (Fig. 3c) to 45◦ (Fig. 3d). The nearly identical IR characteristics of the four fractions imply that impurities, such as dickite- or nacrite-like phases, may exist in the sample, but they are not significant in these clay fractions because these impurities have different IR bands from kaolinite at room temperature. Particle Size Effects on Intercalation by Hydrazine Monohydrate All of the >1-µm fractions showed very similar intercalation characteristics. The 0.72 nm d(001) of kaolinite expanded to
1.03 nm by hydrazine monohydrate (H2 NNH2 · H2 O) as shown in the example XRD patterns of the 1- to 2-µm clay fraction during the intercalation by hydrazine monohydrate (Fig. 4a). No other diffraction peak or shoulder between the 0.72 and 1.03-nm peaks was observed, indicating no intermediate intercalates. The XRD 0.72-nm peak intensity decreased quickly in the initial 30 min and nearly disappeared within 2 h. The 1.03-nm d-spacing for the hydrazine-kaolinite intercalate agreed well with the 1.03- to 1.04-nm values reported in the literature (1–4). The >2-µm fractions showed very similar intercalation characteristics as the 1- to 2-µm clay: all of them underwent very quick initial reactions; there was no XRD-detectable transition state and the intercalations were nearly complete within
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
385
FIG. 5. XRD patterns of >2-µm fractions of kaolinite during their intercalation (a) by hydrazine monohydrate after 2 h reaction and (b) by saturated KOAc solution after 24 h reaction.
2 h (Fig. 5a). On the contrary, the <1-µm fractions exhibited different reactivities to hydrazine monohydrate. The ratio of the nonintercalated kaolinite (0.72 nm) to the intercalated phase (1.03 nm) increased as the particle size decreased. About half of the <0.2-µm kaolinite could not be intercalated by hydrazine monohydrate even after more than 1 month of reaction (Fig. 6a). Particle Size Effect on Intercalation by Saturated KOAc Solution The intercalation of kaolinite by saturated KOAc solution was slower than that by hydrazine. The 0.72-nm d(001) spacing of kaolinite was expanded to 1.40 nm by KOAc (Fig. 4b). The higher-order diffraction of the 1.40-nm peak also showed in the
patterns, i.e., the 002 at 0.70 nm, 003 at 0.469 nm, and 004 at 0.352 nm (Figs. 4b, 5b, and 6b). The position of the 004 diffraction of the intercalated kaolinite at 0.352 nm was very close to the 002 diffraction of the nonintercalated kaolinite at 0.357 nm, but they were well resolved in the 5-h intercalation pattern of the 1- to 2-µm kaolinite (Fig. 4b). The >2-µm fractions were also intercalated faster than the finer ones by saturated KOAc solution. The intercalation of all of the >2-µm fractions were nearly complete after 24 h of interaction (Fig. 5b). Broad, lowintensity peaks at around 0.7 nm remained in the XRD patterns, but were centered at 0.70 nm, representing the 002 of the 1.40-nm kaolinite–KOAc intercalates, rather than at the 0.72-nm spacing of the nonintercalated kaolinite.
386
FIG. 6.
DENG, WHITE, AND DIXON
XRD patterns of <2-µm fractions of kaolinite after one month of intercalation (a) by hydrazine monohydrate and (b) by saturated KOAc solution.
Different Responses of Particle Size to the Intercalations by Hydrazine Monohydrate and KOAc To obtain a more comprehensive image of the intercalation differences among the different fractions, the percentage of the 001 peak area of intercalated kaolinite to the total area of the 001 peaks of the nonintercalated and intercalated kaolinites were calculated and plotted in Fig. 7. The area ratio approximates the intercalation yield of kaolinite. For the hydrazine monohydrate intercalated kaolinite, the area of the peak at 1.03 nm was counted as the intercalated phase and that at 0.72 nm as the nonintercalated one. Because there are some mica in the >5-µm
fractions and the mineral had a diffraction peak at 1.0 nm, the area for the intercalated phase was slightly overestimated. The sand fraction was not plotted in Fig. 7 because the mica peak introduced too much uncertainty. For the KOAc intercalation, the area of the peak at 1.40 nm was attributed to the intercalated phase. The area of the 0.72-nm peak cannot be fully attributed to the nonintercalated kaolinite because the weak 002 peak of the intercalated kaolinite was located at 0.70 nm nearby the 0.72-nm peak. It was found that the 5- to 10-µm fraction showed the highest completion of intercalation with an area ratio of the 1.40-nm peak to 0.70 nm of 13.8. It was assumed that this ratio remained constant for all of the intercalated kaolinite regardless of the
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
387
FIG. 7. Peak area percentage of the intercalated kaolinite relative to total 001 peak area of the intercalated and nonintercalated kaolinites for (a) intercalation by hydrazine monohydrate and (b) intercalation by KOAc acetate solution.
particle size. During the calculation of the nonintercalated kaolinite, the area of the peak at 0.72 nm was corrected by subtracting 1/13.8 of the area of the 1.40-nm peak. In the intercalation of kaolinite by hydrazine and saturated KOAc solution, the proportions of nonintercalated kaolinite increased when the particle size decreased. The finer particles, however, showed different reactivity in the intercalations by the two intercalating agents. About half of the particles in the <0.2-µm fraction were intercalated by hydrazine, but the intercalation by saturated KOAc solution was negligible even at the end of the 35-day monitoring period. Most of the particles (>80%) in the 0.2- to 0.5- and 0.5- to 1.0-µm fractions were intercalated by hydrazine, but almost none of the mineral in the 0.2- to 0.5-µm fraction was intercalated by KOAc solution and only about 40% of the 0.5- to 1-µm fraction were intercalated by KOAc within the 35-day monitoring period. Nearly all of the kaolinite in the 1- to 2-µm fractions were intercalated by hydrazine within 2 h, whereas there was still about 20% was not intercalated by KOAc solution even after longer than 1 month of reaction.
Structural Adjustments after Intercalation from IR Evidences The apical Si–O and the surface OH groups of kaolinite are nearly perpendicular to the (001) plane of the mineral (Fig. 1). Their vibration bands are highly pleochroic. Structural deformations after intercalation may change their pleochroism properties. A successful IR pleochroism experiment requires that the mineral particles are well oriented. When 1- to 2-µm kaolinite– water and intercalated kaolinite–hydrazine suspensions were deposited and dried on ZnS windows, the prismatic peaks in the 20– 25◦ 2θ range disappeared from the XRD patterns (Fig. 8). The disappearance of the prismatic peaks indicated that the kaolinite platelets, both before and after intercalation, were highly oriented on the ZnS windows. The (001) planes of the particles were parallel to the window surfaces. As a result of the highly preferred orientation of the particles, some bonds in the structure become oriented with respect to the IR beam. As the orientation of the mineral relative to the IR beam remains the same before and after treatment, any changes in intensities of IR bands, or the pleochroism change, can be used as evidence for changes
388
DENG, WHITE, AND DIXON
FIG. 8.
XRD patterns of randomly oriented kaolinite powder and oriented films on ZnS windows.
in the geometry of the bonds relative to the basal plane of the kaolinite. IR Pleochroism Change of Si–O Band in the Hydrazine–Kaolinite Intercalate The hydrazine monohydrate–kaolinite (1–2 µm) intercalate was alternately monitored by XRD (Fig. 9) and IR. As revealed by the XRD patterns, the intercalate had well-defined sharp 1.03-nm 001 peaks in the first four XRD-IR monitoring cycles. The sharp XRD peaks imply that the intercalated kaolinite persisted in its highly ordered structure during the initial loss of the excess hydrazine molecules. The IR spectra of the hydrazine monohydrate–kaolinite intercalate taken during the third monitoring cycle (Fig. 10a) had a sharp band at 1124 cm−1 that arose from the apical Si–O stretching vidration of the mineral, similar to those observed for DMSO or formamide intercalated kaolinites (4, 41). Compared with the same vibration in the starting mineral (Fig. 3c), this band shifted 10 cm−1 from 1115 cm−1 and varied less in intensity than the starting kaolinite when the incidence angles were changed from 0◦ to 45◦ . The pleochroism change of the apical Si–O band was more obvious in the IR spectra during the fifth XRD-IR monitoring cycle (Fig. 10b). In the fifth monitoring cycle, the mineral was beginning to collapse as indicated by the reduced intensity and increased broadness and high angle asymmetry of the 1.03 nm XRD peak (Fig. 9). Some of the hydrazine molecules must have moved out of the interlayer and the mineral partly returned to the initial structural conditions. The new band at 1113 cm−1 results from the apical Si–O bond of the deinter-
calated kaolinite. The 1113 cm−1 band of the deintercalated kaolinite was highly pleochroic as shown by the almost doubled intensity when the incidence angle was rotated from 0◦ to 45◦ (Fig. 10b). Compared with the 1113 cm−1 band of the deintercalated kaolinite, the 1124 cm−1 band of the remaining intercalated kaolinite was much less pleochroic. At 0◦ of incidence angle, its intensity was higher than the 1113 cm−1 band; at 45◦ , however, its intensity was lower than the 1113 cm−1 band. The intensity changes of the 1113 and 1124 cm−1 bands relative to each other confirmed that the apical Si–O band became less pleochroic when kaolinite was intercalated. It also confirmed that kaolinite particles on the ZnS windows were highly oriented. Otherwise, no pleochroism would be observed for the deintercalated kaolinite. Two bands at 1038 and 1008 cm−1 arose from the Si–O–Si in-plane stretching vibrations of the kaolinite–hydrazine intercalate (Fig. 10a). The 1038 cm−1 band occurred at a slightly higher frequency than the 1033 cm−1 band for starting kaolinite (Fig. 3c). Similar bands shifts have been observed previously by Johnston et al. (42). These authors attributed these shifts to small structural changes induced by the presence of hydrazine and by dielectric effects. The blue shifts of the Si–O and Si–O–Si bands suggest that the Si–O bond strength was enhanced in hydrazine intercalated kaolinite. IR Pleochroism Changes of Surface OH Bands It is generally agreed that the 3695, 3668, and 3652 cm−1 bands in the IR spectrum (Fig. 3a) of kaolinite arise from the stretching vibrations of the surface OH groups, and that the
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 9.
389
XRD patterns of kaolinite–hydrazine monohydrate complex during deintercalation.
3619 cm−1 band arises from the stretching vibration of the inner OH group (43, 44). The 3695 cm−1 band was highly pleochroic (Fig. 3). The IR spectra of hydrazine-intercalated kaolinite in the 3250–3750 cm−1 range during the third monitoring cycle were plotted in Fig. 11. Relatively sharp bands occurred at 3618, 3365, and 3360 cm−1 and broad bands occurred at 3682, 3655, 3570, 3463, and 3310 cm−1 . The 3365, 3360, and 3310 cm−1 bands arise from the N–H stretching vibrations of intercalated hydrazine molecules and will be discussed in another article. The 3618 cm−1 band arose from the inner OH group of the kaolinite. The surface OH band at 3695 cm−1 disappeared in the kaolinite– hydrazine intercalate. The disappearance of the 3695 cm−1 band has been interpreted to result from the H-bonding of the surface OH groups with hydrazine molecules (1, 2, 45, 46). At least one of the bands at 3570 or 3463 cm−1 must belong to the perturbed
surface OH groups from kaolinite. When the film was rotated to different incidence angles (Fig. 11), however, the intensities of the 3570 and 3463 cm−1 bands did not change as distinctly as the 3695 cm−1 band of the starting kaolinite (Figs. 3a and 3b). The similarity of the intensities of these bands indicated that the Hbonded surface OH groups in the intercalate were less pleochroic than those in the starting kaolinite. In other words, the directions of the OH groups in the intercalate were biased more from the perpendicular direction of the (001) plane. Inner OH Group Perturbance The stretching vibration of the inner OH group of the intercalate had a frequency of 3618 cm−1 (Fig. 11), which was nearly identical to the position of the corresponding 3619 cm−1 band of the starting kaolinite. Since the inner OH group is located
390
DENG, WHITE, AND DIXON
between the tetrahedral and octahedral sheet and is not accessible to guest molecules, this OH group usually is not readily perturbed by the intercalation. Yet the deformation vibrations of the inner OH groups were shifted after intercalation. The 937 and 914 cm−1 bands (Fig. 3c) arise from the bending vibrations of the surface OH and inner OH groups, respectively (43). Russell and Fraser (47) gave reversed assignments for the two bands. Only one band at 903 cm−1 in the spectra of the intercalate could be found in the same range (Fig. 10a). The lack of two distinct bands suggested that the bending vibrations of the inner OH and surface OH groups had identical frequencies in the intercalate. Compared with the original frequencies at 937 and 914 cm−1 , both of the surface OH and inner OH group bending bands had red shifts in their frequencies. The red shifts of these two bands imply that not only the surface OH groups but also the inner OH groups were perturbed in the intercalation. As the kaolinite hydrazine intercalate lost interlayer hydrazine molecules, new bands were observed at 3624 and 3611 cm−1
FIG. 11. FTIR spectra (3800–3200 cm−1 ) of the hydrazine–kaolinite intercalate during the third XRD-IR monitoring cycle.
(Fig. 12). These two bands likely originated either from the surface OH groups due to a new arrangement of H bond or the inner OH groups perturbed during the deintercalation. As will be discussed in detail in another manuscript, the 3624 cm−1 band was the only surface OH band in the anhydrous hydrazine-kaolinite intercalate with a d(001) spacing of 0.96 nm. The 3624 cm−1 band during the deintercalation of the hydrazine hydrate kaolinite intercalate (Fig. 12) is likely due to the same kind of bonding when the 1.03-nm complex began to collapse. Namely, it was formed when water molecules were evaporated and the intercalate formed the same structure as the 0.96-nm intercalate. The 3611 cm−1 band was not observed in the anhydrous hydrazine– kaolinite intercalate and more likely originated from the inner OH groups. The Si–O band shifts and pleochroism changes have confirmed that the tetrahedra in the intercalate tilted more than in the starting kaolinite. The tilting of the tetrahedra should affect the inner OH groups too. As discussed earlier, the inner OH group in a fully intercalated kaolinite had the same stretching frequency (3618 cm−1 ) as the nonintercalated kaolinite, but the bending band of the inner OH group red-shifted to 903 cm−1 . This means the configurations of the tetrahedra in the starting and the fully intercalated kaolinite caused little change in the chemical environment of the inner OH groups. In the transition state of tetrahedra tilt from one configuration to the other one, the chemical environment of the inner OH might be slightly different from the equilibrium environment. It appears reasonable to assign the 3611 cm−1 band to the inner OH groups during the deintercalation. DISCUSSION
Crystallinity and Impurity Effects on Intercalation Rate FIG. 10. FTIR spectra (1200–850 cm−1 ) of (a) hydrazine-kaolinite intercalate during the third XRD-IR monitoring cycle and (b) partly collapsed hydrazine–kaolinite intercalate during the fifth XRD-IR monitoring cycle.
The crystallinity indices of the four clay fractions were very close, and there is no consistent trend associated with the particle size. High crystallinity kaolinite can be intercalated as mentioned early. Apparently the lower reactivity of the finer particles
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 12.
391
FTIR spectra of kaolinite–hydrazine intercalate during its deintercalation.
cannot be attributed to their better crystallinity as proposed by Wiewiora and Brindley (17). The different responses of the particle size to the hydrazine and KOAc solution treatments suggested that the KOAc was a less effective intercalating agent in the intercalation of kaolinite. Moreover, the different responses imply that the slower intercalation rate of the finer fractions cannot be attributed to other impurities, such as dickite- or nacritelike phases as reported in the literature (29, 30). Otherwise, the negligible intercalation of the <0.2 fraction will lead to the conclusion that most of the particles in this fraction are the impurities. The high similarity of XRD patterns and IR spectra of the four clay fractions, however, suggested that the presence of impurities is not significant in the sample and therefore they cannot be the fundamental reason for the lower reactivity of the finer particles. The impurity or stacking fault hypothesis cannot explain why the particle size has different responses to the two intercalating compounds. All of these data suggest that the intrinsic structural properties of kaolinite contributed to the differences in intercalation rate. Tilting of Tetrahedra in Kaolinite Intercalate The IR spectra of the fully intercalated kaolinite and the partly deintercalated kaolinite–hydrazine intercalate show that the apical Si–O bond became less pleochroic when the mineral was intercalated by hydrazine. The reduced pleochroism of the apical Si–O band in the intercalate suggested that the apical Si–O was biased more from the perpendicular direction of the (001) plane,
meaning that the tetrahedra tilted more. Apparently, the tilting was induced by the intercalation. When the guest molecules moved out the interlayer space, the tetrahedra tended to return to its original configuration as indicated by the recovered pleochroism and band position. The pleochroism changes of the apical Si– O band also suggested that the tilting of the tetrahedra during the intercalation-deintercalation is reversible. Similarly, the reduced pleochroism of the surface OH groups, and the shifts of the bending bands of the inner OH groups also imply that structural adjustments occurred in the octahedra. The appearance of the new band at 3611 cm−1 from the inner OH group also suggested structural changes in the tetrahedral and octahedral sheets. The reasons for the changes in the surface OH groups and tilt of the apical Si–O bonds are likely the result of releasing the structural stress accumulated from the misfit of the tetrahedral and octahedral sheets of the mineral. It is well established that an ideal Si tetrahedral sheet is larger than an Al octahedral sheet (48). This misfit induces structural stress on the mineral. The misfit is compensated by several types of structural distortion. One of them is the rotation of the tetrahedra from the ideal hexagonal arrangement to ditrigonal shapes. The second is the tilting of the tetrahedra such as that the basal oxygens do not occur in the same plane. The tilting of the tetrahedra produces a corrugated basal plane. The tilting of the tetrahedra also leads the apical Si–O bond to bias from the perpendicular direction of the (001) plane. Our calculations based on published crystallography data (36) indicated that the two apical Si–O bonds in a unit cell are tilted 4.8◦ and 5.7◦ from the perpendicular
392
DENG, WHITE, AND DIXON
direction of the (001) plane. The distortions of the tetrahedra and octahedra in kaolinite, however, cannot fully release the structural stress probably due to the effects of hydrogen bonding between layers. This accumulated stress is likely the reason why coarse kaolinite particles frequently exist as vermiforms in which the plates are not exactly parallel to each other (49). The structural stress might also be a reason why kaolinite cannot grow as large as other phyllosilicates such as mica and chlorite. When kaolinite was intercalated, the loss of the stabilizing effects of the hydrogen bonding caused the misfit of the tetrahedral and octahedral sheets to be compensated by different structural adjustments within the silicate layers. In this experiment, increased tilting of the apical Si–O bond and the surface OH groups appear to result from the structural adjustment. Aside from the spectroscopic evidences observed in this study, some morphology observation and crystallographic analyses on intercalates also indicated that the structural deformation may occur even though most of the literature on kaolinite intercalation assumed that the basic silicon–oxygen tetrahedra and aluminum–oxygen (hydroxy) octahedra did not deform during intercalation. For example, curling of kaolinite plates to a tubular structure was reported by Singh and Mackinnon (50) after 35 treatments with potassium acetate solution followed by washing and centrifugation. The transformation of the platy kaolinite to tubular structure implies more structural adjustment within the basic 1 : 1 sheet occurs during the intercalation–deintercalation process. Based on the principal components analysis of 15 structural parameters, Giese proposed that intercalation can distort the 1 : 1 layer structure, either by allowing the layer to relax or as the result of a strong interaction between the intercalating molecule and the atoms of the layers. The distortion caused increases in the sizes of the octahedra and tetrahedra, the thickness of the octahedral and tetrahedral sheet, the length of the shared octahedral edge, and the tilting of the tetrahedra (28). The structural adjustments of the tetrahedra and octahedra during the intercalation suggest that the structural stress is a driving force for the intercalation. When the stress is larger, the intercalation is faster and proceeds to a greater extent, and vice versa. Kaolinite particles appear platy in shape. To keep the plates flat, coarser particles will accumulate more stress induced by the misfit of the tetrahedral and octahedral sheets, and therefore, the coarser fractions are intercalated faster. For the same reason, the finer particles are intercalated more slowly due to their reduced structural stress. CONCLUSIONS
Finer kaolinite particles were intercalated more slowly than the coarser particles. Hydrazine monohydrate intercalated kaolinite faster than saturated KOAc solution. The KOAc intercalation was negligible in some fine fractions. The faster intercalation rates of the coarse fractions are likely dominated by the structural stress accumulated from the misfit of the tetrahedral and octahedral sheets of the mineral. The strong pleochroic properties of the apical Si–O band, and the surface OH bands
were reduced when kaolinite was intercalated. The pleochroism changes suggested that the tetrahedra tilted more and the octahedra may also have undergone structural adjustments during the intercalation. The red shift of the bending vibration of the inner OH group and the appearance of the 3611 cm−1 absorbance during the deintercalation of the hydrazine–kaolinite intercalate also supported the premise that structural deformation occurred during intercalation. The changes in the kaolinite structure are caused by releasing of stress from the structural misfit between the tetrahedral sheet and octahedral sheet. The IR pleochroism data support the hypothesis that the finer particles have slower intercalation rates due to the less structural stress on their layers. The unusual size effect on intercalation rate of kaolinite could not be explained by crystallinity differences as reported in the literature, because the particles in the clay fractions have lower crystallinity than the silt particles and there is no consistent crystallinity trend associated with the particle size among the four clay fractions. Impurities, such as dickite or nacrite-like phases, are not significant in the studied samples; they could not substantially contribute to lower reactivity of the small particles. ACKNOWLEDGMENT The authors thank Dr. Richard Loeppert for allowing access the FTIR-system 2000.
REFERENCES 1. Frost, R. L., Kloprogge, J. T., Kristof, J., and Horvath, E., Clays clay Miner. 47, 732 (1999). 2. Johnston, C. T., and Stone, D. A., Clays Clay Miner. 38, 121 (1990). 3. Inoue, K., and Okuda, S., Yogyo Kyokai Shi 81, 353 (1973). 4. Ledoux, R. L., and White, J. L., J. Colloid Interface Sci. 21, 127 (1966). 5. Theng, B. K. G., Churchman, G. J., Whitton, J. S., and Claridge, G. G. C., Clays Clay Miner. 32, 249 (1984). 6. Churchman, G. J., Whitton, J. S., Claridge, G. G. C., and Theng, B. K. G., Clays Clay Miner. 32, 241 (1984). 7. Komori, Y., Sugahara, Y., and Kuroda, K., Appl. Clay Sci. 15, 241 (1999). 8. Komori, Y., Sugahara, Y., and Kuroda, K., J. Mater. Res. 13, 930 (1998). 9. Sidheswaran, P., Mohan, S. V. R., Ganguli, P., and Bhat, A. N., Indian J. Chem. A 26A, 994 (1987). 10. Sidheswaran, P., Bhat, A. N., and Ganguli, P., Clays Clay Miner. 38, 29 (1990). 11. Sugahara, Y., Nagayama, T., Kuroda, K., Doi, A., and Kato, C., Clay Sci. 8, 69 (1991). 12. Sugahara, Y., Satokawa, S., Kuroda, K., and Kato, C., Clays Clay Miner. 36, 343 (1988). 13. Sugahara, Y., Satokawa, S., Kuroda, K., and Kato, C., Clays Clay Miner. 38, 137 (1990). 14. Barrios, J., Plancon, A., Cruz, M. I., and Tchoubar, C., Proc. Eur. Clay Conf. 3rd 6 (1977). 15. Wada, K., Am. Mineral. 46, 78 (1961). 16. Weiss, A., Thielepape, W., Goring, G., and Ritter, W., “Kaoliniteinlagerungs-verbindungen,” Proceedings of the International Clay Conference, 1963, Stockholm, Sweden. 17. Wiewiora, A., and Brindley, G. W., “Potassium Acetate Intercalation in Kaolinite and Its Removal: Effect of Material Characteristics,” Proceedings of the International Clay Conference, 1969, Tokyo, Japan. 18. Weiss, A., Becker, H. O., Orth, H., Mai, G., Lechner, H., and Range, K.-J., “Particle Size Effects and Reaction Mechanism of the Intercalation
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
29.
30. 31. 32. 33.
into Kaolinite,” Proceedings of the International Clay Conference, 1969, Tokyo, Japan. Gomes, C. S. F., Bol. Soc. Geol. Port. 23, 55 (1982). Uwins, P. J. R., Mackinnon, I. D. R., Thompson, J. G., and Yago, A. J. E., Clays Clay Miner. 41, 707 (1993). Chen, Z.-X., Yan, W., Wang, J., and Ji, C.-Y., Jianzhu Cailiao Xuebao 3, 240 (2000). [in Chinese] Chen, Z.-X., Yan, W., Wang, J., Ji, C.-Y., and Xian, Y.-M., Jianzhu Cailiao Xuebao 3, 151 (2000). [in Chinese] Frost, R. L., Kristof, J., Paroz, G. N., Tran, T. H., and Kloprogge, J. T., J. Colloid Interface Sci. 204, 227 (1998). Frost, R. L., Kristof, J., and Tran, T. H., Clay Miner. 33, 605 (1998). Frost, R. L., Kristof, J., Horvath, E., and Kloprogge, J. T., J. Colloid Interface Sci. 214, 109 (1999). Bish, D. L., and Chipera, S. J., “35th Annual Clay Mineral Society Meeting Abstract,” p. 90, Clay Mineral Society, Aurora, CO, 1998. Raussell-Colom, J. A., and Serratosa, J. M., in “Chemistry of Clays and Clay Minerals” (A. C. D. Newman, Ed.), Vol. 6, p. 480, Wiley, New York, 1987. Giese, R. F., Jr., in “Hydrous Phyllosilicates (Exclusive of Micas)” (S. W. Bailey, Ed.), Vol. 19, p. 725, Mineralogical Society of America, Chelsea, MI, 1988. Johnston, C. T., Elzea-Kogel, J., and Bish, D. L., “37th Annual (Wy-2-K+ ) Clay Mineral Society Meeting, p. 69, Clay Mineral Society, Aurora, CO, 2000. Prost, R., Dameme, A., Huard, E., Driard, J., and Leydecker, J. P., Clays Clay Miner. 37, 464 (1989). Cruz, M. D., and Franco, F., Clays Clay Miner. 48, 586 (2000). Ben Haj Amara, A., Ben Brahim, J., Besson, G., and Pons, E. C. H., Clay Miner. 30, 295 (1995). Ben Haj Amara, A., Clay Miner. 32, 463 (1997).
393
34. Ben Haj Amara, A., Ben Brahim, J., Plancon, A., and Ben, R. H., J. Appl. Crystallogr. 31, 654 (1998). 35. Ben Haj Amara, A., Ben Rhaiem, H., and Plancon, A., J. Appl. Crystallogr. 33, 1351 (2000). 36. Bish, D. L., Clays Clay Miner. 41, 738 (1993). 37. Dixon, J. B., and White, G. N., “Soil Mineralogy Laboratory Manual,” Texas A&M University, 2000. 38. Hinckley, D. N., Clays Clay Miner. 11, 229 (1964). 39. Lietard, O., “Contribution a L’etude des proprietes phsicochimiques cristallographiques et morphologiques des kaolins,” L’Institut National Polytechnique de Lorraine, 1977. 40. Brindley, G. W., Kao, C. C., Harrison, J. L., Lipsicas, M., and Raythatha, R., Clays Clay Miner. 34, 239 (1986). 41. Olejnik, S., Posner, A. M., and Quirk, J. P., Clays Clay Miner. 19, 83 (1971). 42. Johnston, C. T., Bish, D. L., Eckert, J., and Brown, L. A., J. Phys. Chem. B 104, 8080 (2000). 43. Farmer, V. C., “The Infrared Spectra of Minerals,” Mineralogical Society, London, 1974. 44. Farmer, V. C., Clay Miner. 33, 601 (1998). 45. Ledoux, R. L., and White, J. L., Silicates Ind. 32, 269 (1967). 46. Frost, R. L., Kristof, J., Paroz, G. N., and Kloprogge, J. T., J. Colloid Interface Sci. 208, 216 (1998). 47. Russell, J. D., and Fraser, A. R., in “Clay Mineralogy: Spectroscopic and Chemical Determinative Methods” (M. J. Wilson, Ed.), p. 367, Chapman and Hall, London, 1994. 48. Schulze, D. G., in “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), p. 1, Soil Science Society of America, Madison, WI, 1989. 49. Dixon, J. B., in “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), p. 467, Soil Science Society of America, Madison, WI, 1989. 50. Singh, B., and Mackinnon, I. D. R., Clays Clay Miner. 44, 825 (1996).
Effect of Structural Stress on the Intercalation Rate of Kaolinite Youjun Deng, G. Norman White, and Joe B. Dixon1 Department of Soil and Crop Sciences, Texas A&M University, College Station, Texas 77843-2474 Received July 16, 2001; accepted December 27, 2001; published online May 15, 2002
Particle size in kaolinite intercalation showed an inverse reactivity trend compared with most chemical reactions: finer particles had lower reactivity and some of the fine particles cannot be intercalated. Although this phenomenon was noted in the early 1960s and several hypotheses have been reported, there is no widely accepted theory about the unusual particle size response in the intercalation. We propose that structural stress is a controlling factor in the intercalation and the stress contributes to the higher reactivity of the coarser particles. In this study, we checked the structural deformation spectroscopically and indirectly proved the structural stress hypothesis. A Georgia kaolinite was separated into nine size fractions and their intercalations by hydrazine monohydrate and potassium acetate were investigated with X-ray diffraction (XRD) and Fourier-transform infrared (FTIR) analyses. The apical Si–O band of kaolinite at 1115 cm−1 shifted to 1124 cm−1 when the mineral was intercalated to 1.03 nm by hydrazine monohydrate, and its strong pleochroic properties became much weaker. Similar reduction in pleochroism was observed on the surface OH bands of kaolinite after intercalation. Both the bending vibrations of the inner OH group at 914 cm−1 and of the surface OH group at 937 cm−1 shifted to 903 cm−1 after intercalation by hydrazine. A new band for the inner OH group appeared at 3611 cm−1 during the deintercalation of the 1.03 nm hydrazine kaolinite complex. Pleochroism change in the apical Si–O band suggested the tetrahedra had increased tilt with respect to the (001) plane. The tilt of the Si–O apical bond could occur only if the octahedra had also undergone structural rearrangement during intercalation. These changes in the octahedral and tetrahedral sheets represent some change in the manner of compensation for the structural misfit of the tetrahedral sheet and octahedral sheet. As the lateral dimensions of a kaolinite particle increases, the cumulative degree of misfit increases. Intercalation breaks the hydrogen bonds between layers and allows for the structure to reduce the accumulated stress in some other manner. The reversed size effect on intercalation probably was not caused by crystallinity differences as reported in the literature, because the Hinckley and Lietard crystallinity indices of the four clay fractions were very close to each other. Impurities, such as dickiteor nacrite-like phases are not significant in the studied sample as suggested by the XRD and IR results, they are not the main reasons for the lower reactivity of the finer particles. C 2002 Elsevier Science (USA) Key Words: crystallinity; hydrazine; intercalation; kaolinite; particle size; pleochroism; potassium acetate; structural misfit; structural stress.
1 To whom correspondence should be addressed. Fax: (979)845-0456. E-mail: [email protected].
379
INTRODUCTION
Intercalation of kaolin minerals with inorganic and organic compounds has wide potential for scientific and industrial applications. Intercalation of kaolinite often induces structural perturbations. Therefore, the guest intercalating compounds can function as probes in the investigation of the crystallography and reactivity of kaolin minerals (1–4). Soil kaolinite and halloysite can be differentiated using the differences in intercalation rate by formamide (5, 6). Kaolinite has been modified to become hydrophobic with long-chain alkylamines (7, 8), fatty acid salts (9, 10), and several polymers (11–13). The modified kaolinite minerals have special thermal and mechanical properties that can be used in the polymer industry. The intercalation reaction has also been used to delaminate coarse kaolinite (14). The study on kaolinite intercalation began in the 1960s (15, 16), and numerous publications are available on this subject. It is commonly observed that particle size has a reversed reactivity trend in kaolinite’s intercalation compared with other chemical reactions: the finer kaolinite particles are intercalated more slowly than coarse particles. Some fine kaolinite particles cannot be intercalated at all. Several research groups have discussed the possible reasons, no agreement has been reached. In the intercalation study of kaolinite by potassium acetate, Wiewiora and Brindley (17) reported that the intercalation of the >1 µm kaolinite by potassium acetate (KOAc) was essentially complete within 1 week, less than 4% of the <0.5-µm fraction intercalated in a 1-week reaction period. Weiss (18) reported a similar reactivity trend in the intercalation of kaolinite by urea: when particle size decreased from 9.5 to 0.68 µm, the intercalation rate first increased to a maximum value in the 3.8- to 5.0-µm fraction and then decreased. It took 65 days for the 0.68- to 0.80-µm fraction to reach complete intercalation but 20 days for the 3.8to 5.0-µm fraction. Gomes (19) investigated the intercalation capacities of the <0.5-µm particles from five kaolinite samples by hydrazine and also showed that the finer particles were intercalated more slowly than the coarser ones. Uwins et al. (20) examined the intercalation rates of a wide range of kaolinites with different particle size, shape, and defect distributions. They first treated the samples with hydrazine hydrate and then investigated the reactivity of hydrazine-treated kaolinite toward N methylformamide (NMF). The >2-µm fractions reacted to near completion overnight, but less than 28% of the <0.3 µm could be intercalated by NMF even with longer treatments. The lower 0021-9797/02 $35.00
C 2002 Elsevier Science (USA)
All rights reserved.
380
DENG, WHITE, AND DIXON
reactivity of the finer kaolinite particles in the intercalation was also reported by Chen et al. (21, 22). The lower reactivity of the finer kaolinite particles was attributed to their higher crystallinity by Wiewiora and Brindley (17). High crystallinity kaolinites, however, have been reported as undergoing fast intercalation by potassium acetate in several papers (23–25). Well-ordered kaolinite also showed higher intercalation yield than the poorly ordered kaolinite in the intercalation by NMF as reported by Uwins (20). Furthermore, the finer fractions can have poorer crystallinity as well as better crystallinity than the coarser fractions depending on the origin of the mineral (26). The facts that there is no consistent trend in crystallinity associated with particle size and that high crystalline kaolin minerals are more responsive to intercalation suggest that the “better crystallinity” is not likely the reason why the finer kaolinite particles are intercalated more slowly or even cannot be intercalated at all. Weiss et al. (18) explained the reversed particle size effect using a “ring mechanism.” According to their hypothesis, the intercalating molecules act as wedges at kaolinite edges and cause the layers to be elastically deformed. The length of elastically deformed zone in the crystal is termed “cooperative action length,” a0 . The intercalating molecules penetrate faster in the deformed zone because of the weaker cohesion in the zone. One factor in determining the cooperative action length is the elastic properties of the layers. Particles with better crystallinity have longer a0 and intercalate faster. The intercalation on large crystals can start from all edges giving a “ring mechanism” and therefore have faster reaction rate. For small crystals, the intercalation cannot undergo from all edges because the elastic deformation induced by the intercalation from an edge are transmitted across the particle and cause contractions of the layers on the opposite edge. The intercalation on small particles proceeds at a slow rate via a “one-side mechanism.” By this mechanism, the high-crystallinity crystals should be more difficult to intercalate because it requires more energy to open the longer “cooperative action zone.” This mechanism cannot explain why certain fine kaolinite cannot be intercalated. Raussell-Colom and Serratosa (27) summarized the factors in determining the intercalation capacity of kaolinite. They concluded that the intercalation capacities depend on the cohesion energies of the kaolinite stackings. The stacking in kaolinite is strongly correlated with structural disorder or crystal imperfection. The reason why intercalation capacity depends on particle size is because crystal size is related to crystal imperfection. It has been found that the imperfections or defects in kaolinite are largely due to errors in the position of the vacant octahedral site and the errors constitute the introduction of dickite layers (28). The IR bands of dickite and nacrite have been identified in some fractionated kaolinite samples at low temperatures. The overall proportion of these dickite and nacrite stacking sequences increased with decreasing particle size in some kaolinite samples (29, 30). Dickite, nacrite, and kaolinite are 1 : 1 layer silicate minerals; they have identical chemical compositions but
differ structurally by rotations between layers, which causes changes in hydrogen bonding. The differences in hydrogen bonding may result in different reactivity to the intercalating agents. The appearance of these stacking faults in kaolinite may affect the overall intercalation rate. For example, dickite has been observed to have a slower intercalation rate than kaolinite when it is treated with hydrazine monohydrate; however, a high proportion of 5- to 10-µm dickite with low defects could be intercalated to 1.024 nm by hydrazine monohydrate and the intercalation was nearly complete within a 2-day reaction period (31). Similarly, nacrite can be intercalated to near complete by KOAc or N -methylacetatmide in a few days (32–35). The near completion of the intercalations of dickite and nacrite suggests that dickite- or nacrite-like phases in kaolinite should be intercalated too if a long reaction time, for example, a few days, is allowed. In other words, dickite, nacrite, or the stacking errors in kaolinite are not likely a main reason for the slow intercalation of the fine kaolinite particles. Gomes (19) reported another type of defect, that the presence of randomly interstratified single layers of smectite or illite are strongly correlated to the intercalation properties. These types of defects are more abundant in the smaller particle size fractions and are generally responsible for the asymmetry of the 001 X-ray reflections and the lower intercalation capacity. The interstratification separates the crystal to more small domains, and it cannot explain why the small kaolinite domains are less active in the intercalations. Furthermore, the asymmetry of the 001 XRD peak seemed more dependent on the origin of the sample. The asymmetry indices of the <0.1- and 0.1- to 0.2-µm fractions of an Ireland “bondclay” kaolinite are almost the same, but only about 13% of the <0.1-µm fraction was intercalated, whereas about 35% of the 0.1- to 0.2-µm fraction was intercalated (19). The same asymmetry indices but different intercalation capacities suggested the particle size is more important than the defects caused by the random interstratification of illite or smectite in determining of the intercalation rate or capacity. Uwins et al. (20) also similarly concluded that particle size appeared to be a more significant controlling factor than defect distribution in determining the relative yield of kaolinite–NMF intercalate. We propose that structural stress is a driving force in the intercalation. Kaolinite is a 1 : 1 layer silicate with a structure consisting of a tetrahedral Si sheet sharing apical O2− ions with an octahedral Al sheet (Fig. 1). The lateral size of the tetrahedral sheet is larger than the octahedral sheet. Structural stress is induced from the misfit of the tetrahedral sheet and the octahedral sheet in the 1 : 1-type linkage. Part of the structural stress can be relieved when kaolinite is intercalated. Coarser particles intercalate faster due to their larger structural stress. Because the stress is associated with the misfit of the tetrahedral and the octahedral sheets, it is expected that some deformation within the sheets will occur when the constraints resulting from hydrogen bonding between layers is reduced. Determining the changes in octahedral and tetrahedral conformation should provide some indirect evidence of the structural stress change.
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 1.
381
Structural models of kaolinite (according to Bish (36)).
The main objective of this study was to test the hypothesis that structural stress is a controlling factor in the intercalation of kaolinite by documenting structural changes in the octahedra and tetrahedra following intercalation with infrared data. Meanwhile, it is endeavored to discuss the possibilities of the hypotheses that impurities, stacking faults, or crystallinity are the reasons that the fine particles are intercalated more slowly. MATERIALS AND METHODS
Kaolinite and Intercalating Agents A Georgia kaolinite sample, provided by the J. M. Huber Corporation, was used in this study. The sample was treated with 1 M, pH 5, sodium acetate buffer solution to ensure that it was free from carbonate minerals and exchangeable high-valence cations (37). The sample was separated into nine fractions by sieving for the sand fraction (50–500 µm), settling for the silt fractions (2–5, 5–10, 10–20, and 20–50 µm), and centrifugation for the clay fractions (<0.2, 0.2–0.5, 0.5–1.0, and 1.0–2.0 µm). A solution containing 3 g Na2 CO3 per 20 L with a pH about 10 was used to disperse the sample. All of the fractions were oven dried at 60◦ C. The Hinckley (38) and Lietard (39) crystallinity indices were determined on the kaolinite in each fraction. It should be noted that determination of the Lietard crystallinity index is hindered by the presence of mica in the coarser fractions. Hydrazine monohydrate and saturated KOAc solutions were used as the intercalating agents in this study. Potassium acetate obtained from Aldrich Chemical Company, and hydrazine monohydrate (H2 NNH2 · H2 O, 99.8%) obtained from Fisher Scientific, were used without further purification. Monitoring Intercalation Rate Two methods were employed to study the intercalation rates of kaolinite by hydrazine. To study the initial reaction, dry kaolinite
powder was mounted on a quartz slide, and four to six drops of hydrazine monohydrate were added on the powder, immediately followed by 10 continuous XRD analyses in the 5–15◦ 2θ range (2). Typically, one run of the XRD took 18 min. For the <1-µm clay fractions, it was found that the intercalation was not complete within 3 h. To monitor the long-term intercalation rate, a 0.5-g sample from each fraction was immersed in hydrazine monohydrate in a 40-mL Nalgene centrifuge tube for various lengths of time. The tube was capped and shaken frequently. To measure the XRD pattern of the sample, the particles were concentrated by centrifugation, and a small amount of clay paste was smeared on a glass slide as a thin film. The film was analyzed with XRD immediately in the 5–15◦ 2θ range. The intercalation of kaolinite by saturated KOAc solution on quartz slide was similarly monitored by XRD as the first method used for the intercalation by hydrazine, except a broader 2θ range of 5 to 28◦ was monitored. Sample Preparation for Infrared Analysis Because the 1- to 2-µm size fraction contains no other minerals that are detectable by XRD or IR and has fast and near complete intercalation by hydrazine monohydrate, only the hydrazine monohydrate intercalated 1- to 2-µm kaolinite was examined by IR analysis. One gram of 1- to 2-µm kaolinite was mixed with 5 mL of hydrazine monohydrate in a Nalgene centrifuge tube. After more than 3 weeks’ shaking, a subsample of the suspension was diluted with hydrazine monohydrate. Six drops of the diluted suspension, containing about 1 mg of kaolinite, were pipetted onto a 32 × 2-mm ZnS Irtran window. The window was heated at about 45◦ C with a 250-W infrared heat lamp to accelerate the evaporation of excess hydrazine. When the evaporation of the free hydrazine monohydrate was nearly complete, the sample was checked with the X-ray diffractometer in an open atmosphere. After XRD examination, an additional
382
DENG, WHITE, AND DIXON
ZnS window and a rubber O-ring washer were used to seal the nearly dry film between the two windows. The two windows were mounted on an IR sample holder, and the film orientation was changed by rotating the windows to 0◦ and 45◦ with respect to the perpendicular direction of the IR beam. The film was alternately monitored with the X-ray diffractometer in open atmosphere, and with the infrared spectrometer in the sealed windows. The purpose of sealing the film between the two windows in the IR analysis was to reduce the evaporation while the incidence angle was changed. The alternate XRD and IR monitoring were repeated more than 10 times until the hydrazine IR bands had nearly disappeared from the IR spectrum. The potassium acetate–kaolinite intercalates were not investigated with IR spectroscopy because the intercalation of the clay fractions was not complete even after 1 month of reaction. It was also difficult to remove the excess salt without causing collapse of the intercalates. X-Ray Diffraction Analysis XRD analyses were performed with CuK α radiation on a Philips diffractometer. Operational conditions of 30 kV and
FIG. 2.
18 mA were used. The goniometer was equipped with a curved graphite beam monochromator. The measurements were conducted in the step-scanning mode, with a step size of 0.05◦ 2θ and a counting time of 5 s at each step. Fourier-Transform Infrared Spectroscopy IR spectra were recorded using a Perkin-Elmer 2000 system Fourier-transform IR spectrometer. For each spectrum, a total of 64 scans were obtained at a resolution of 1 cm−1 and a data collecting interval of 0.2 cm−1 . A triglycine sulfate (TGS) detector was used for the analyses. The optical bench of the instrument and the sample chamber were purged with dry air produced by a Balston 75-52 FTIR purge gas generator. RESULTS
Crystallinity of Different Size Fractions The XRD patterns of the nine size fractions used in this study (Fig. 2) demonstrated that the dominant mineral in all fractions was kaolinite. As indicated by the poorer resolution, broader peak width and weaker intensities of the prismatic peaks at
Powder XRD patterns of the different size fractions of a kaolinite sample.
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
TABLE 1 Particle Size Distribution and Crystallinity Indices of the Kaolinite Fractions Used in This Study Particle size (µm)
Percentage (%, w/w)
Hinckley indexa
Lietard R1 indexb
Lietard R2 indexb
<0.2 0.2–0.5 0.5–1.0 1.0–2.0 2.0–5.0 5.0–10 10–20 20–50 50–500
5.3 23.6 20.8 8.8 23.7 9.8 2.5 2.3 0.2
0.22 0.42 0.29 0.49 0.49 0.64 0.66 0.58 0.43
0.34 0.34 0.39 0.44 0.45 0.46 0.49 0.48 1.06
0.51 0.52 0.47 0.57 0.71 0.74 0.81 0.81 1.14
a
Calculated according to Hinckley (38). Calculated according to Lietard (39). The value for the Lietard crystallinity for the 50- to 500-µm fraction is the result of interference by mica peak overlaps. b
0.444, 0.434, 0.416, 0.256, 0.249, 0.238, and 0.229 nm, the finer fractions were generally less ordered than the coarser fractions. The small particle size also resulted in weak diffraction of the finer fractions. The crystallinity of kaolinite as characterized by the Hinckley (38) and Lietard (39) indices (Table 1). The
FIG. 3.
383
Hinckley indices of the fractions were less than 0.7, indicating the crystallinity of the samples were lower than some low defect kaolinites such as Keokuk kaolinite which has a Hinckley index of 1.54 (40). The clay particles (<2 µm) have lower crystallinity indices than the silt fractions (2–50 µm). The crystallinity indices among the four clay fractions (<0.2, 0.2–0.5, 0.5–1, and 1–2 µm) are very close to each other and there is no consistent trend associated with the particle size indicating their similarity in crystallinity. Infrared Analyses of Clay Size Fractions of the Kaolinite The four clay fractions showed almost identical IR characteristics in terms of band position, shape, intensity, and intensity response to the change of incidence angles (Fig. 3). The four IR bands at 3696, 3668, 3652, and 3619 cm−1 are the characteristic OH stretching vibration bands of kaolinite. The 3696 cm−1 band had strong pleochroism as shown by its increasing absorption when the incidence angle was changed from 0◦ (Fig. 3a) to 45◦ (Fig. 3b). The other bands at 1115 (apical Si–O), 1033, 1008 (Si–O–Si in-plane vibrations), 937, 914 (OH bending vibrations), 795, and 755 cm−1 (OH translational vibrations) are also typical bands of kaolinite. The strong IR pleochroism of
FTIR spectra of different kaolinite size fractions at 0 and 45◦ IR beam incidence angles.
384
DENG, WHITE, AND DIXON
FIG. 4.
Selected XRD patterns of 1- to 2-µm kaolinite during its intercalation (a) by hydrazine monohydrate and (b) by saturated KOAc solution.
the 1115 cm−1 band is demonstrated by its intensity increase when the incidence angle was changed from 0◦ (Fig. 3c) to 45◦ (Fig. 3d). The nearly identical IR characteristics of the four fractions imply that impurities, such as dickite- or nacrite-like phases, may exist in the sample, but they are not significant in these clay fractions because these impurities have different IR bands from kaolinite at room temperature. Particle Size Effects on Intercalation by Hydrazine Monohydrate All of the >1-µm fractions showed very similar intercalation characteristics. The 0.72 nm d(001) of kaolinite expanded to
1.03 nm by hydrazine monohydrate (H2 NNH2 · H2 O) as shown in the example XRD patterns of the 1- to 2-µm clay fraction during the intercalation by hydrazine monohydrate (Fig. 4a). No other diffraction peak or shoulder between the 0.72 and 1.03-nm peaks was observed, indicating no intermediate intercalates. The XRD 0.72-nm peak intensity decreased quickly in the initial 30 min and nearly disappeared within 2 h. The 1.03-nm d-spacing for the hydrazine-kaolinite intercalate agreed well with the 1.03- to 1.04-nm values reported in the literature (1–4). The >2-µm fractions showed very similar intercalation characteristics as the 1- to 2-µm clay: all of them underwent very quick initial reactions; there was no XRD-detectable transition state and the intercalations were nearly complete within
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
385
FIG. 5. XRD patterns of >2-µm fractions of kaolinite during their intercalation (a) by hydrazine monohydrate after 2 h reaction and (b) by saturated KOAc solution after 24 h reaction.
2 h (Fig. 5a). On the contrary, the <1-µm fractions exhibited different reactivities to hydrazine monohydrate. The ratio of the nonintercalated kaolinite (0.72 nm) to the intercalated phase (1.03 nm) increased as the particle size decreased. About half of the <0.2-µm kaolinite could not be intercalated by hydrazine monohydrate even after more than 1 month of reaction (Fig. 6a). Particle Size Effect on Intercalation by Saturated KOAc Solution The intercalation of kaolinite by saturated KOAc solution was slower than that by hydrazine. The 0.72-nm d(001) spacing of kaolinite was expanded to 1.40 nm by KOAc (Fig. 4b). The higher-order diffraction of the 1.40-nm peak also showed in the
patterns, i.e., the 002 at 0.70 nm, 003 at 0.469 nm, and 004 at 0.352 nm (Figs. 4b, 5b, and 6b). The position of the 004 diffraction of the intercalated kaolinite at 0.352 nm was very close to the 002 diffraction of the nonintercalated kaolinite at 0.357 nm, but they were well resolved in the 5-h intercalation pattern of the 1- to 2-µm kaolinite (Fig. 4b). The >2-µm fractions were also intercalated faster than the finer ones by saturated KOAc solution. The intercalation of all of the >2-µm fractions were nearly complete after 24 h of interaction (Fig. 5b). Broad, lowintensity peaks at around 0.7 nm remained in the XRD patterns, but were centered at 0.70 nm, representing the 002 of the 1.40-nm kaolinite–KOAc intercalates, rather than at the 0.72-nm spacing of the nonintercalated kaolinite.
386
FIG. 6.
DENG, WHITE, AND DIXON
XRD patterns of <2-µm fractions of kaolinite after one month of intercalation (a) by hydrazine monohydrate and (b) by saturated KOAc solution.
Different Responses of Particle Size to the Intercalations by Hydrazine Monohydrate and KOAc To obtain a more comprehensive image of the intercalation differences among the different fractions, the percentage of the 001 peak area of intercalated kaolinite to the total area of the 001 peaks of the nonintercalated and intercalated kaolinites were calculated and plotted in Fig. 7. The area ratio approximates the intercalation yield of kaolinite. For the hydrazine monohydrate intercalated kaolinite, the area of the peak at 1.03 nm was counted as the intercalated phase and that at 0.72 nm as the nonintercalated one. Because there are some mica in the >5-µm
fractions and the mineral had a diffraction peak at 1.0 nm, the area for the intercalated phase was slightly overestimated. The sand fraction was not plotted in Fig. 7 because the mica peak introduced too much uncertainty. For the KOAc intercalation, the area of the peak at 1.40 nm was attributed to the intercalated phase. The area of the 0.72-nm peak cannot be fully attributed to the nonintercalated kaolinite because the weak 002 peak of the intercalated kaolinite was located at 0.70 nm nearby the 0.72-nm peak. It was found that the 5- to 10-µm fraction showed the highest completion of intercalation with an area ratio of the 1.40-nm peak to 0.70 nm of 13.8. It was assumed that this ratio remained constant for all of the intercalated kaolinite regardless of the
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
387
FIG. 7. Peak area percentage of the intercalated kaolinite relative to total 001 peak area of the intercalated and nonintercalated kaolinites for (a) intercalation by hydrazine monohydrate and (b) intercalation by KOAc acetate solution.
particle size. During the calculation of the nonintercalated kaolinite, the area of the peak at 0.72 nm was corrected by subtracting 1/13.8 of the area of the 1.40-nm peak. In the intercalation of kaolinite by hydrazine and saturated KOAc solution, the proportions of nonintercalated kaolinite increased when the particle size decreased. The finer particles, however, showed different reactivity in the intercalations by the two intercalating agents. About half of the particles in the <0.2-µm fraction were intercalated by hydrazine, but the intercalation by saturated KOAc solution was negligible even at the end of the 35-day monitoring period. Most of the particles (>80%) in the 0.2- to 0.5- and 0.5- to 1.0-µm fractions were intercalated by hydrazine, but almost none of the mineral in the 0.2- to 0.5-µm fraction was intercalated by KOAc solution and only about 40% of the 0.5- to 1-µm fraction were intercalated by KOAc within the 35-day monitoring period. Nearly all of the kaolinite in the 1- to 2-µm fractions were intercalated by hydrazine within 2 h, whereas there was still about 20% was not intercalated by KOAc solution even after longer than 1 month of reaction.
Structural Adjustments after Intercalation from IR Evidences The apical Si–O and the surface OH groups of kaolinite are nearly perpendicular to the (001) plane of the mineral (Fig. 1). Their vibration bands are highly pleochroic. Structural deformations after intercalation may change their pleochroism properties. A successful IR pleochroism experiment requires that the mineral particles are well oriented. When 1- to 2-µm kaolinite– water and intercalated kaolinite–hydrazine suspensions were deposited and dried on ZnS windows, the prismatic peaks in the 20– 25◦ 2θ range disappeared from the XRD patterns (Fig. 8). The disappearance of the prismatic peaks indicated that the kaolinite platelets, both before and after intercalation, were highly oriented on the ZnS windows. The (001) planes of the particles were parallel to the window surfaces. As a result of the highly preferred orientation of the particles, some bonds in the structure become oriented with respect to the IR beam. As the orientation of the mineral relative to the IR beam remains the same before and after treatment, any changes in intensities of IR bands, or the pleochroism change, can be used as evidence for changes
388
DENG, WHITE, AND DIXON
FIG. 8.
XRD patterns of randomly oriented kaolinite powder and oriented films on ZnS windows.
in the geometry of the bonds relative to the basal plane of the kaolinite. IR Pleochroism Change of Si–O Band in the Hydrazine–Kaolinite Intercalate The hydrazine monohydrate–kaolinite (1–2 µm) intercalate was alternately monitored by XRD (Fig. 9) and IR. As revealed by the XRD patterns, the intercalate had well-defined sharp 1.03-nm 001 peaks in the first four XRD-IR monitoring cycles. The sharp XRD peaks imply that the intercalated kaolinite persisted in its highly ordered structure during the initial loss of the excess hydrazine molecules. The IR spectra of the hydrazine monohydrate–kaolinite intercalate taken during the third monitoring cycle (Fig. 10a) had a sharp band at 1124 cm−1 that arose from the apical Si–O stretching vidration of the mineral, similar to those observed for DMSO or formamide intercalated kaolinites (4, 41). Compared with the same vibration in the starting mineral (Fig. 3c), this band shifted 10 cm−1 from 1115 cm−1 and varied less in intensity than the starting kaolinite when the incidence angles were changed from 0◦ to 45◦ . The pleochroism change of the apical Si–O band was more obvious in the IR spectra during the fifth XRD-IR monitoring cycle (Fig. 10b). In the fifth monitoring cycle, the mineral was beginning to collapse as indicated by the reduced intensity and increased broadness and high angle asymmetry of the 1.03 nm XRD peak (Fig. 9). Some of the hydrazine molecules must have moved out of the interlayer and the mineral partly returned to the initial structural conditions. The new band at 1113 cm−1 results from the apical Si–O bond of the deinter-
calated kaolinite. The 1113 cm−1 band of the deintercalated kaolinite was highly pleochroic as shown by the almost doubled intensity when the incidence angle was rotated from 0◦ to 45◦ (Fig. 10b). Compared with the 1113 cm−1 band of the deintercalated kaolinite, the 1124 cm−1 band of the remaining intercalated kaolinite was much less pleochroic. At 0◦ of incidence angle, its intensity was higher than the 1113 cm−1 band; at 45◦ , however, its intensity was lower than the 1113 cm−1 band. The intensity changes of the 1113 and 1124 cm−1 bands relative to each other confirmed that the apical Si–O band became less pleochroic when kaolinite was intercalated. It also confirmed that kaolinite particles on the ZnS windows were highly oriented. Otherwise, no pleochroism would be observed for the deintercalated kaolinite. Two bands at 1038 and 1008 cm−1 arose from the Si–O–Si in-plane stretching vibrations of the kaolinite–hydrazine intercalate (Fig. 10a). The 1038 cm−1 band occurred at a slightly higher frequency than the 1033 cm−1 band for starting kaolinite (Fig. 3c). Similar bands shifts have been observed previously by Johnston et al. (42). These authors attributed these shifts to small structural changes induced by the presence of hydrazine and by dielectric effects. The blue shifts of the Si–O and Si–O–Si bands suggest that the Si–O bond strength was enhanced in hydrazine intercalated kaolinite. IR Pleochroism Changes of Surface OH Bands It is generally agreed that the 3695, 3668, and 3652 cm−1 bands in the IR spectrum (Fig. 3a) of kaolinite arise from the stretching vibrations of the surface OH groups, and that the
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 9.
389
XRD patterns of kaolinite–hydrazine monohydrate complex during deintercalation.
3619 cm−1 band arises from the stretching vibration of the inner OH group (43, 44). The 3695 cm−1 band was highly pleochroic (Fig. 3). The IR spectra of hydrazine-intercalated kaolinite in the 3250–3750 cm−1 range during the third monitoring cycle were plotted in Fig. 11. Relatively sharp bands occurred at 3618, 3365, and 3360 cm−1 and broad bands occurred at 3682, 3655, 3570, 3463, and 3310 cm−1 . The 3365, 3360, and 3310 cm−1 bands arise from the N–H stretching vibrations of intercalated hydrazine molecules and will be discussed in another article. The 3618 cm−1 band arose from the inner OH group of the kaolinite. The surface OH band at 3695 cm−1 disappeared in the kaolinite– hydrazine intercalate. The disappearance of the 3695 cm−1 band has been interpreted to result from the H-bonding of the surface OH groups with hydrazine molecules (1, 2, 45, 46). At least one of the bands at 3570 or 3463 cm−1 must belong to the perturbed
surface OH groups from kaolinite. When the film was rotated to different incidence angles (Fig. 11), however, the intensities of the 3570 and 3463 cm−1 bands did not change as distinctly as the 3695 cm−1 band of the starting kaolinite (Figs. 3a and 3b). The similarity of the intensities of these bands indicated that the Hbonded surface OH groups in the intercalate were less pleochroic than those in the starting kaolinite. In other words, the directions of the OH groups in the intercalate were biased more from the perpendicular direction of the (001) plane. Inner OH Group Perturbance The stretching vibration of the inner OH group of the intercalate had a frequency of 3618 cm−1 (Fig. 11), which was nearly identical to the position of the corresponding 3619 cm−1 band of the starting kaolinite. Since the inner OH group is located
390
DENG, WHITE, AND DIXON
between the tetrahedral and octahedral sheet and is not accessible to guest molecules, this OH group usually is not readily perturbed by the intercalation. Yet the deformation vibrations of the inner OH groups were shifted after intercalation. The 937 and 914 cm−1 bands (Fig. 3c) arise from the bending vibrations of the surface OH and inner OH groups, respectively (43). Russell and Fraser (47) gave reversed assignments for the two bands. Only one band at 903 cm−1 in the spectra of the intercalate could be found in the same range (Fig. 10a). The lack of two distinct bands suggested that the bending vibrations of the inner OH and surface OH groups had identical frequencies in the intercalate. Compared with the original frequencies at 937 and 914 cm−1 , both of the surface OH and inner OH group bending bands had red shifts in their frequencies. The red shifts of these two bands imply that not only the surface OH groups but also the inner OH groups were perturbed in the intercalation. As the kaolinite hydrazine intercalate lost interlayer hydrazine molecules, new bands were observed at 3624 and 3611 cm−1
FIG. 11. FTIR spectra (3800–3200 cm−1 ) of the hydrazine–kaolinite intercalate during the third XRD-IR monitoring cycle.
(Fig. 12). These two bands likely originated either from the surface OH groups due to a new arrangement of H bond or the inner OH groups perturbed during the deintercalation. As will be discussed in detail in another manuscript, the 3624 cm−1 band was the only surface OH band in the anhydrous hydrazine-kaolinite intercalate with a d(001) spacing of 0.96 nm. The 3624 cm−1 band during the deintercalation of the hydrazine hydrate kaolinite intercalate (Fig. 12) is likely due to the same kind of bonding when the 1.03-nm complex began to collapse. Namely, it was formed when water molecules were evaporated and the intercalate formed the same structure as the 0.96-nm intercalate. The 3611 cm−1 band was not observed in the anhydrous hydrazine– kaolinite intercalate and more likely originated from the inner OH groups. The Si–O band shifts and pleochroism changes have confirmed that the tetrahedra in the intercalate tilted more than in the starting kaolinite. The tilting of the tetrahedra should affect the inner OH groups too. As discussed earlier, the inner OH group in a fully intercalated kaolinite had the same stretching frequency (3618 cm−1 ) as the nonintercalated kaolinite, but the bending band of the inner OH group red-shifted to 903 cm−1 . This means the configurations of the tetrahedra in the starting and the fully intercalated kaolinite caused little change in the chemical environment of the inner OH groups. In the transition state of tetrahedra tilt from one configuration to the other one, the chemical environment of the inner OH might be slightly different from the equilibrium environment. It appears reasonable to assign the 3611 cm−1 band to the inner OH groups during the deintercalation. DISCUSSION
Crystallinity and Impurity Effects on Intercalation Rate FIG. 10. FTIR spectra (1200–850 cm−1 ) of (a) hydrazine-kaolinite intercalate during the third XRD-IR monitoring cycle and (b) partly collapsed hydrazine–kaolinite intercalate during the fifth XRD-IR monitoring cycle.
The crystallinity indices of the four clay fractions were very close, and there is no consistent trend associated with the particle size. High crystallinity kaolinite can be intercalated as mentioned early. Apparently the lower reactivity of the finer particles
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
FIG. 12.
391
FTIR spectra of kaolinite–hydrazine intercalate during its deintercalation.
cannot be attributed to their better crystallinity as proposed by Wiewiora and Brindley (17). The different responses of the particle size to the hydrazine and KOAc solution treatments suggested that the KOAc was a less effective intercalating agent in the intercalation of kaolinite. Moreover, the different responses imply that the slower intercalation rate of the finer fractions cannot be attributed to other impurities, such as dickite- or nacritelike phases as reported in the literature (29, 30). Otherwise, the negligible intercalation of the <0.2 fraction will lead to the conclusion that most of the particles in this fraction are the impurities. The high similarity of XRD patterns and IR spectra of the four clay fractions, however, suggested that the presence of impurities is not significant in the sample and therefore they cannot be the fundamental reason for the lower reactivity of the finer particles. The impurity or stacking fault hypothesis cannot explain why the particle size has different responses to the two intercalating compounds. All of these data suggest that the intrinsic structural properties of kaolinite contributed to the differences in intercalation rate. Tilting of Tetrahedra in Kaolinite Intercalate The IR spectra of the fully intercalated kaolinite and the partly deintercalated kaolinite–hydrazine intercalate show that the apical Si–O bond became less pleochroic when the mineral was intercalated by hydrazine. The reduced pleochroism of the apical Si–O band in the intercalate suggested that the apical Si–O was biased more from the perpendicular direction of the (001) plane,
meaning that the tetrahedra tilted more. Apparently, the tilting was induced by the intercalation. When the guest molecules moved out the interlayer space, the tetrahedra tended to return to its original configuration as indicated by the recovered pleochroism and band position. The pleochroism changes of the apical Si– O band also suggested that the tilting of the tetrahedra during the intercalation-deintercalation is reversible. Similarly, the reduced pleochroism of the surface OH groups, and the shifts of the bending bands of the inner OH groups also imply that structural adjustments occurred in the octahedra. The appearance of the new band at 3611 cm−1 from the inner OH group also suggested structural changes in the tetrahedral and octahedral sheets. The reasons for the changes in the surface OH groups and tilt of the apical Si–O bonds are likely the result of releasing the structural stress accumulated from the misfit of the tetrahedral and octahedral sheets of the mineral. It is well established that an ideal Si tetrahedral sheet is larger than an Al octahedral sheet (48). This misfit induces structural stress on the mineral. The misfit is compensated by several types of structural distortion. One of them is the rotation of the tetrahedra from the ideal hexagonal arrangement to ditrigonal shapes. The second is the tilting of the tetrahedra such as that the basal oxygens do not occur in the same plane. The tilting of the tetrahedra produces a corrugated basal plane. The tilting of the tetrahedra also leads the apical Si–O bond to bias from the perpendicular direction of the (001) plane. Our calculations based on published crystallography data (36) indicated that the two apical Si–O bonds in a unit cell are tilted 4.8◦ and 5.7◦ from the perpendicular
392
DENG, WHITE, AND DIXON
direction of the (001) plane. The distortions of the tetrahedra and octahedra in kaolinite, however, cannot fully release the structural stress probably due to the effects of hydrogen bonding between layers. This accumulated stress is likely the reason why coarse kaolinite particles frequently exist as vermiforms in which the plates are not exactly parallel to each other (49). The structural stress might also be a reason why kaolinite cannot grow as large as other phyllosilicates such as mica and chlorite. When kaolinite was intercalated, the loss of the stabilizing effects of the hydrogen bonding caused the misfit of the tetrahedral and octahedral sheets to be compensated by different structural adjustments within the silicate layers. In this experiment, increased tilting of the apical Si–O bond and the surface OH groups appear to result from the structural adjustment. Aside from the spectroscopic evidences observed in this study, some morphology observation and crystallographic analyses on intercalates also indicated that the structural deformation may occur even though most of the literature on kaolinite intercalation assumed that the basic silicon–oxygen tetrahedra and aluminum–oxygen (hydroxy) octahedra did not deform during intercalation. For example, curling of kaolinite plates to a tubular structure was reported by Singh and Mackinnon (50) after 35 treatments with potassium acetate solution followed by washing and centrifugation. The transformation of the platy kaolinite to tubular structure implies more structural adjustment within the basic 1 : 1 sheet occurs during the intercalation–deintercalation process. Based on the principal components analysis of 15 structural parameters, Giese proposed that intercalation can distort the 1 : 1 layer structure, either by allowing the layer to relax or as the result of a strong interaction between the intercalating molecule and the atoms of the layers. The distortion caused increases in the sizes of the octahedra and tetrahedra, the thickness of the octahedral and tetrahedral sheet, the length of the shared octahedral edge, and the tilting of the tetrahedra (28). The structural adjustments of the tetrahedra and octahedra during the intercalation suggest that the structural stress is a driving force for the intercalation. When the stress is larger, the intercalation is faster and proceeds to a greater extent, and vice versa. Kaolinite particles appear platy in shape. To keep the plates flat, coarser particles will accumulate more stress induced by the misfit of the tetrahedral and octahedral sheets, and therefore, the coarser fractions are intercalated faster. For the same reason, the finer particles are intercalated more slowly due to their reduced structural stress. CONCLUSIONS
Finer kaolinite particles were intercalated more slowly than the coarser particles. Hydrazine monohydrate intercalated kaolinite faster than saturated KOAc solution. The KOAc intercalation was negligible in some fine fractions. The faster intercalation rates of the coarse fractions are likely dominated by the structural stress accumulated from the misfit of the tetrahedral and octahedral sheets of the mineral. The strong pleochroic properties of the apical Si–O band, and the surface OH bands
were reduced when kaolinite was intercalated. The pleochroism changes suggested that the tetrahedra tilted more and the octahedra may also have undergone structural adjustments during the intercalation. The red shift of the bending vibration of the inner OH group and the appearance of the 3611 cm−1 absorbance during the deintercalation of the hydrazine–kaolinite intercalate also supported the premise that structural deformation occurred during intercalation. The changes in the kaolinite structure are caused by releasing of stress from the structural misfit between the tetrahedral sheet and octahedral sheet. The IR pleochroism data support the hypothesis that the finer particles have slower intercalation rates due to the less structural stress on their layers. The unusual size effect on intercalation rate of kaolinite could not be explained by crystallinity differences as reported in the literature, because the particles in the clay fractions have lower crystallinity than the silt particles and there is no consistent crystallinity trend associated with the particle size among the four clay fractions. Impurities, such as dickite or nacrite-like phases, are not significant in the studied samples; they could not substantially contribute to lower reactivity of the small particles. ACKNOWLEDGMENT The authors thank Dr. Richard Loeppert for allowing access the FTIR-system 2000.
REFERENCES 1. Frost, R. L., Kloprogge, J. T., Kristof, J., and Horvath, E., Clays clay Miner. 47, 732 (1999). 2. Johnston, C. T., and Stone, D. A., Clays Clay Miner. 38, 121 (1990). 3. Inoue, K., and Okuda, S., Yogyo Kyokai Shi 81, 353 (1973). 4. Ledoux, R. L., and White, J. L., J. Colloid Interface Sci. 21, 127 (1966). 5. Theng, B. K. G., Churchman, G. J., Whitton, J. S., and Claridge, G. G. C., Clays Clay Miner. 32, 249 (1984). 6. Churchman, G. J., Whitton, J. S., Claridge, G. G. C., and Theng, B. K. G., Clays Clay Miner. 32, 241 (1984). 7. Komori, Y., Sugahara, Y., and Kuroda, K., Appl. Clay Sci. 15, 241 (1999). 8. Komori, Y., Sugahara, Y., and Kuroda, K., J. Mater. Res. 13, 930 (1998). 9. Sidheswaran, P., Mohan, S. V. R., Ganguli, P., and Bhat, A. N., Indian J. Chem. A 26A, 994 (1987). 10. Sidheswaran, P., Bhat, A. N., and Ganguli, P., Clays Clay Miner. 38, 29 (1990). 11. Sugahara, Y., Nagayama, T., Kuroda, K., Doi, A., and Kato, C., Clay Sci. 8, 69 (1991). 12. Sugahara, Y., Satokawa, S., Kuroda, K., and Kato, C., Clays Clay Miner. 36, 343 (1988). 13. Sugahara, Y., Satokawa, S., Kuroda, K., and Kato, C., Clays Clay Miner. 38, 137 (1990). 14. Barrios, J., Plancon, A., Cruz, M. I., and Tchoubar, C., Proc. Eur. Clay Conf. 3rd 6 (1977). 15. Wada, K., Am. Mineral. 46, 78 (1961). 16. Weiss, A., Thielepape, W., Goring, G., and Ritter, W., “Kaoliniteinlagerungs-verbindungen,” Proceedings of the International Clay Conference, 1963, Stockholm, Sweden. 17. Wiewiora, A., and Brindley, G. W., “Potassium Acetate Intercalation in Kaolinite and Its Removal: Effect of Material Characteristics,” Proceedings of the International Clay Conference, 1969, Tokyo, Japan. 18. Weiss, A., Becker, H. O., Orth, H., Mai, G., Lechner, H., and Range, K.-J., “Particle Size Effects and Reaction Mechanism of the Intercalation
STRUCTURAL STRESS EFFECT ON KAOLINITE INTERCALATION
19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
29.
30. 31. 32. 33.
into Kaolinite,” Proceedings of the International Clay Conference, 1969, Tokyo, Japan. Gomes, C. S. F., Bol. Soc. Geol. Port. 23, 55 (1982). Uwins, P. J. R., Mackinnon, I. D. R., Thompson, J. G., and Yago, A. J. E., Clays Clay Miner. 41, 707 (1993). Chen, Z.-X., Yan, W., Wang, J., and Ji, C.-Y., Jianzhu Cailiao Xuebao 3, 240 (2000). [in Chinese] Chen, Z.-X., Yan, W., Wang, J., Ji, C.-Y., and Xian, Y.-M., Jianzhu Cailiao Xuebao 3, 151 (2000). [in Chinese] Frost, R. L., Kristof, J., Paroz, G. N., Tran, T. H., and Kloprogge, J. T., J. Colloid Interface Sci. 204, 227 (1998). Frost, R. L., Kristof, J., and Tran, T. H., Clay Miner. 33, 605 (1998). Frost, R. L., Kristof, J., Horvath, E., and Kloprogge, J. T., J. Colloid Interface Sci. 214, 109 (1999). Bish, D. L., and Chipera, S. J., “35th Annual Clay Mineral Society Meeting Abstract,” p. 90, Clay Mineral Society, Aurora, CO, 1998. Raussell-Colom, J. A., and Serratosa, J. M., in “Chemistry of Clays and Clay Minerals” (A. C. D. Newman, Ed.), Vol. 6, p. 480, Wiley, New York, 1987. Giese, R. F., Jr., in “Hydrous Phyllosilicates (Exclusive of Micas)” (S. W. Bailey, Ed.), Vol. 19, p. 725, Mineralogical Society of America, Chelsea, MI, 1988. Johnston, C. T., Elzea-Kogel, J., and Bish, D. L., “37th Annual (Wy-2-K+ ) Clay Mineral Society Meeting, p. 69, Clay Mineral Society, Aurora, CO, 2000. Prost, R., Dameme, A., Huard, E., Driard, J., and Leydecker, J. P., Clays Clay Miner. 37, 464 (1989). Cruz, M. D., and Franco, F., Clays Clay Miner. 48, 586 (2000). Ben Haj Amara, A., Ben Brahim, J., Besson, G., and Pons, E. C. H., Clay Miner. 30, 295 (1995). Ben Haj Amara, A., Clay Miner. 32, 463 (1997).
393
34. Ben Haj Amara, A., Ben Brahim, J., Plancon, A., and Ben, R. H., J. Appl. Crystallogr. 31, 654 (1998). 35. Ben Haj Amara, A., Ben Rhaiem, H., and Plancon, A., J. Appl. Crystallogr. 33, 1351 (2000). 36. Bish, D. L., Clays Clay Miner. 41, 738 (1993). 37. Dixon, J. B., and White, G. N., “Soil Mineralogy Laboratory Manual,” Texas A&M University, 2000. 38. Hinckley, D. N., Clays Clay Miner. 11, 229 (1964). 39. Lietard, O., “Contribution a L’etude des proprietes phsicochimiques cristallographiques et morphologiques des kaolins,” L’Institut National Polytechnique de Lorraine, 1977. 40. Brindley, G. W., Kao, C. C., Harrison, J. L., Lipsicas, M., and Raythatha, R., Clays Clay Miner. 34, 239 (1986). 41. Olejnik, S., Posner, A. M., and Quirk, J. P., Clays Clay Miner. 19, 83 (1971). 42. Johnston, C. T., Bish, D. L., Eckert, J., and Brown, L. A., J. Phys. Chem. B 104, 8080 (2000). 43. Farmer, V. C., “The Infrared Spectra of Minerals,” Mineralogical Society, London, 1974. 44. Farmer, V. C., Clay Miner. 33, 601 (1998). 45. Ledoux, R. L., and White, J. L., Silicates Ind. 32, 269 (1967). 46. Frost, R. L., Kristof, J., Paroz, G. N., and Kloprogge, J. T., J. Colloid Interface Sci. 208, 216 (1998). 47. Russell, J. D., and Fraser, A. R., in “Clay Mineralogy: Spectroscopic and Chemical Determinative Methods” (M. J. Wilson, Ed.), p. 367, Chapman and Hall, London, 1994. 48. Schulze, D. G., in “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), p. 1, Soil Science Society of America, Madison, WI, 1989. 49. Dixon, J. B., in “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), p. 467, Soil Science Society of America, Madison, WI, 1989. 50. Singh, B., and Mackinnon, I. D. R., Clays Clay Miner. 44, 825 (1996).