RESOLUTION OF THE POLYTYPE STRUCTURE OF SOME ILLITIC CLAY MINERALS THAT APPEAR TO BE 1Md

Transcription

1 Clays and Clay Minerals, Vol. 37, No. 2, , RESOLUTON OF THE POLYTYPE STRUCTURE OF SOME LLTC CLAY MNERALS THAT APPEAR TO BE 1Md G. S. AUSTN, 1 H. D. GLASS, 2 AND R. E. HUGHES 2 New Mexico Bureau of Mines and Mineral Resources, Campus Station, Socorro, New Mexico llinois State Geological Survey, 615 East Peabody Drive Champaign, llinois Abstract--llite/smectite (/S) and illite samples that appear to be 1Md polytypes on the basis of the lack of diagnostic reflections can be resolved as partly either 1M or 2M~ or both by either vapor solvation in ethylene glycol for 40 to 80 hr at ambient temperatures or for 4 hr at 80~ or by heating to 375"C for 2 hr. f samples are so treated, a distinct narrowing of the /S 003/005 peak occurs and diagnostic hkl reflections are revealed. These treatments, however, do not seem to affect diagnostic polytypic peak intensities, but rather the adjacent background. We propose that the 1Md term be restricted to its original definition--the disordered polytype associated with only the 1M ordered polytype. For illites in which disorder exists with a mixture of polytypes, the generic term Ad (disordered group A micas) should be used. Correct polytype determination of illitic material is significant because it reveals thermal history of the sediments and allows greater understanding of the depositional and diagenetic history of sedimentary basins. Key Words--Ethylene glycol solvation, Uite, Uite/smectite, Polytype, Order-disorder, X-ray powder diffraction. NTRODUCTON Confusion exists concerning the determination of polytypes for illite and the definition of 1Md illite (disordered, one-layer monoclinic). Currently five polytypes are recognized for dioctahedral micas, 1Md, 1M, 2M1, 3T, and rarely, 2M2 (Bailey, 1980). These possibilities presumably also exist for illitic material, and possibly smectite material (Drits et al., 1984). Yoder and Eugster (1955) and Levinson (1955) identified 1Md illites by the absence or low intensity of diagnostic polytype reflections, in particular the 112 and 112 reflections, characteristic of the 1M structure. Velde and Hower ( 963) used a ratio of peak intensities as a basis for their claim that the 1Mcl structure was the most common polytype, particularly in the < 1-ttm size fraction of illiles. Reynolds (1980) concurred and used the designation 1Md, but he questioned its appropriateness. Brindley (1980) suggested that a 1Md polytype may exist only ifa small amount ofinterstratification exists, usually with montmorillonite (sic). He also defined the presence of a l Md polytype on the basis of missing or diffuse reflections; however, he saw no obvious general rule that the missing or diffuse reflections corresponded to particular hkl indices, and, therefore, that no particular pattern of disorder was discernible. Srodofi and Eberl (1984) stated that the 1Md polytype can exist only if smectite interstratification is present, and furthermore, that the degree to which X-ray powder diffraction (XRD) effects of the polytype are observed is directly related to the smectite content of the sample. Copyright , The Clay Minerals Society 128 According to Bailey (1987 and personal communication, S. W. Bailey, Department of Geology, University of Wisconsin, Madison, Wisconsin, 1988), for hkl reflections, the 1M polytype shows both discrete k = 3n reflections and k ~ 3n reflections, whereas 1Md illites exhibit only k = 3n reflections. Correct determination of polytypes, especially 1M and 2M~, is important because they are environmental indicators (Hoffman and Hower, 1979; Bruce, 1984; Weaver and Brockstra, 1984). n sedimentary basins, the 1Md and M polytypes are commonly diagenetic and the 2M~ polytype is typically detrital. Yoder and Eugster (1955) concluded that the 1Md structure was the initial or metastable low-temperature form, that the 1M structure forms from the 1Md structure, and that transition to the 2M~ polytype occurs between 200 ~ and 350~ at 15,000 psi water pressure. Bailey's criteria for the identification of disordered polytypes are fully acceptable. The elevated background in the XRD pattern of illitic material between 20* and 33~ CuKa radiation (see stippled area in Figure 2a) has been attributed to disorder in the 1Md polytype (Drits et al., 1984; Eberl et al., 1987; and Bailey, 1988). We have observed that a large part of the elevated background is due to diffuse 00l reflections and at least three other factors: (1) broadening diagnostic reflections for the three common ordered polytypes in the 20 ~ to 33*20 range due to fine particle size, (2) noncrystalline inorganic material, and (3) noncrystalline organic material. Part of the confusion in identifying mixtures ofpolytypes in illitic material has been

2 Vol. 37, No. 2, 1989 Polytype structure of 1Md illite 129 using only diagnostic hkl reflections, or their absence, or the elevated background described above. f it can be demonstrated that most illitic material is 2M~ or a mixture ofpolytypes, then 1Md is no longer a generally useful term. Disorder is possible for all the mica polytypes. The term 1Md carries a genetic implication (Yoder and Eugster, 1955), and it has also been widely misused for samples containing polytypes other than 1M and for samples including smectite interstratifications. Furthermore, we find little evidence that suggests smectite interstratification to be a necessary condition for a disordered polytype. For all these reasons, the 1Md term should probably be restricted to its original definition (Yoder and Eugster, 1955)- the disordered polytype associated with only the 1M ordered polytype. For samples in which disorder exists, but for which the ordered polytype is uncertain, or for a mixture of ordered polytypes, S. W. Bailey (personal communication, 1988) proposed the term Ad (disordered group A micas). Samples having only one ordered polytype (e.g., 2M0 and evidence of a disordered polytype might be described as a mixture of 2M1 and "2Mtd"; however, this usage should be decided by appropriate clay minerals nomenclature committees. For this report, the terms 1Md and Ad, as defined above, will be used. The purpose of this paper is to illustrate how a few simple treatments, such as K + exchange, ethylene glycol solvation for extended times, and heating to 375~ can resolve at least part of the elevated background associated with the Ad polytype and reveal diagnostic polytype reflections that might otherwise be obscured. Materials EXPERMENTAL The present study involved the determination of the mica polytypes in K-bentonites (Table l) using X-ray powder diffraction (XRD) techniques. K-bentonites were chosen for this study because the illitic fraction of K-bentonites is entirely diagenetic (Hoffman and Hower, 1979; Bruce, 1984; Weaver and Brockstra, 1984). f 1M illite is evidence of diagenesis, then by definition, any disordered illitic structure must be 1Md. Polytype reflections are present in XRD patterns of materials that contain as much as 35% smectite layers in the illite/smectite (/S). The samples examined typically consisted of mixtures of /S, authigenic K-feldspar, and minor quartz. K-bentonites were used also because Srodofi and Eberl (1984) reported the same kinds of polytypes regardless of grain size or "bentonitic /S." Methods XRD analyses were performed on randomly oriented powder packed into 2-mm-deep, side-loading sample holders. The samples were examined using a No- Table 1. Location, percentage of smectite interstratification, and /S ordering of samples ~ examined in this study. % Smectite Type of Sample name Locality layers ordering Bravaisite Noyant, France 24 R = 1 DRK-C Jo Daviess County, 30 R = 1 llinois DRK-35 Grant County, 22 R = 1 Wisconsin From llinois State Geological Survey collections. relco XRD unit with CuKa at 50 kv and 28 ma. They were scanned from 4 ~ to 65~ using a vertical goniometer, a graphite crystal monochromator, and receiving and scatter slits of 1 ~ and 4 ~ respectively. One /S sample, "bravaisite" from Noyanl, France, contained about 24% smectite layers. To determine the illite polytype in this sample, a <2-~tm size fraction (to eliminate or reduce non-clay minerals) was isolated using standard sedimentation techniques. Diagnostic 1M peaks of some K-bentonites, particularly those containing an appreciable amount of K-feldspar, were not resolved with ethylene glycol solvation or by heating, unless the non-clay minerals content had been reduced to a minimum by using finer size fractions. For all samples, the < 1-#m size fraction was sufficiently fine grained for polytype determinations. RESULTS The XRD trace of the untreated sample (Figure 1 a) showed the diagnostic 112 (3.66/~, 24.3~ and 112 (3.07 Zk, 29.1~ 1M reflections (see Bailey, 1980, p. 58), but they are poorly defined. Without disturbing the clay materials in the sample holder, the sample was solvated in an ethylene glycol vapor at ambient temperature for about 24 hr and analyzed as above. The second trace (Figure 1 b) showed a better definition of the 112 and 112 reflections. The peak height of the /S 003/0053 reflection had increased and that of the /S reflection between 5 ~ and 10~ had decreased as a result of the Mrring effect. Mrring (1949) demonstrated that if the 00l diffraction from each layer in a mixedlayer mineral diffracts at nearly the same d-value, a strong, sharp diffraction peak results. f the 00l diffraction effects are at significantly different d-values, the resulting XRD peak is broad, has a relatively low peak height, and may form a doublet. A third XRD trace (Figure lc) was made after the sample had been re-exposed for an additional 24 hr to the ethylene glycol atmosphere. The 112 and 112 re- 3 The XRD peak for air-dried (partially hydrated) samples and for the heated, collapsed /S at ~26.5~ is referred to here as the 003/005 peak, even though the peak is only an 003/005 reflection if the sample is solvated with ethylene glycol vapor.

3 130 Austin, Glass, and Hughes Clays and Clay Minerals 1/S 003f005 /S 001/ /S /S Glycolated, 408 hi --f hr in air hr 90~ --e, hr in air GlyCOlat eel, 128 hr +30 rain 110~ Gycolatecl, 48 hr V Glycolat ed, 24 hr J\ k/ --a. 0 hr in air Ungt~coated Degrees 20 (Cu Kc0 i p, Degrees 20 (Cu KS) Figure 1, X-ray powder diffraction traces of the <2-tzm size fraction of randomly oriented bravaisite (about 24% expandable layers) showing increasing definition of the diagnostic 11 ~ and 112 1M reflections with increasing time of glycolation. The reflection consists principally of 131, but also of the 130, 200, and other tess intense reflections. /S 003/ 005 is composed of the illite 003 and smectite 005 reflections. The same is true of /S 002/003, /S 001/002, and US 001/ 001 reflections. Figure 2, X-ray powder diffraction traces of bravaisite after 430 hr of ethylene-glycol vapor solvation and exposure to air at ambient temperature. Little change in the 112 and 112 reflections can be seen for exposure as long as hr in air; however, minor changes in intensity of those peaks and recombination of the /S 001/001 and /S 001/002 reflections were observed when the sample was treated at 110~ for min. Figure 2a includes stippling of the radiation band caused by smearing ofpolytypic hkls (Bailey, 1980), rotational stacking disorder (Drits et al, 1984), line broadening ofpolytypic hkls due to small crystallite size, and material amorphous to X-radiation. flections were better defined, and the /S 003/005 reflection had further increased in peak height. The 001/ 001 and 001/002 reflections were resolved, and the 002/003 reflection at about 17~ shifted toward larger d-values, The sample was returned again to the ethylene glycol atmosphere, and, at the end of three additional days (five days total), a fourth trace was made (Figure l d). The 112 and 112 reflections were only slightly sharpened. The fifth trace (Figure 1 e) showed little change over the fourth trace, and 45 to 60 hr in an ethylene glycol atmosphere at ambient temperatures was apparently adequate to resolve the diagnostic peaks clearly. Note the relative variation in peak height of the 12 and 112 reflections with increased glycolation. A similar modification of the 003/005 peak was noted after as little as 1 hr of ethylene glycol solvation if the temperature was raised to 110~ however, 4 hr at 80~ was adequate to produce the response.

4 Vol. 37, No. 2, 1989 Polytype structure of 1Md illite A, 3,07A 112' /S 003/005 /S 001/001 --c. Glycolated, 240 hr 3.07A 1326 ~ -- f. Glycolated, hr --b. Glycolated, 48 hr ~ e. Glycolated, hr F r fly/ Jr/ ~ -- d. Glycolated, -/2 hr. J-- r Glycolated, 49.5 hr Degrees 2e (Cu Kt~) Figure 3. X-ray powder diffraction traces of the < 2-/~m size fraction of randomly oriented K-bentonite sample DRK-C (about 30% expandable layers) after solvation with ethylene glycol for 48 hr at ambient temperature. Note increased definition of the 112 and 112 reflections, but relatively minor changes in these reflections between 48 and 240 hr. ~ b. Glycolatc~l, 25 hr The possibility of ethylene glycol loss in air after glycolation was also examined by repeating the XRD traces of bravaisite at different time intervals (Figure 2). No significant change was noted for the 112 and 112 reflections, even after 48.5 hr (Figure 2c). The resolution of the 1Mreflections decreased significantly, however, after heating the same sample (Figure 2c) at 110~ for 30 min (Figure 2d), but after a few hours exposure in air, the 112 and 112 reflections reintensifted. The 112 and 112 peaks were always better defined, however, than before glycolation (Figure l a), even after 10 days in air (Figure 2e). Furthermore, mixed-layered /S peaks in the vicinity of the 001,002, and 003 illite positions broadened, and the 001/001 and 001/002 reflections nearly recombined. The same technique was applied to two other K-bentonites to test the glycolation procedure of randomly oriented samples further (Figures 3 and 4). The diagnostic reflections of 1M illite without ethylene glycol treatment were less apparent for sample DRK-C (Fig- Degrees 20 (Cu Kcx) Unglycolated Figure 4. X-ray powder diffraction traces of randomly oriented <2-#m size fraction of K-bentonite sample DRK-35 (about 22% expandable layers) showing increased definition of the diagnostic tm reflections with increasing time of glycolation. ure 2a) than for the bravaisite sample (Figure 1). After 48 hr ofglycolation at room temperature, however, the 112 and t 12 reflections were well defined (Figure 3b). The /S 001/001 and 001/002 reflections were resolved after 240 hr (Figure 3c), but no further change to the diagnostic 1M reflections was noted, One split of sample DRK-35 was treated with eth-

5 132 Austin, Glass, and Hughes Clays and Clay Minerals Degrees 20 (Cu Kc 0 -- e. Heated 2 hr at 600~ -- d. Heated 2 hr at 520~ -- e. Heated 2 hr at 375~ -- b. Heated 2 hr at 100~ -- a. Unheated Figure 5. X-ray powder diffraction traces of the same sample as shown in Figure 4 showing increased definition of the 11 and 112 reflections with increasing temperature. Note shift in d-values of reflections after the sample was heated for 2 hr at 600"C. ylene glycol; the other was heated for 2 hr at 1000, 375 ~ 520 ~ and 600~ (Figure 5). XRD traces were obtained after each heat treatment. Based on the effects of heating mixed-layer /S (Schultz, 1960; Austin and Leininger, 1976), temperatures of 350 ~ to 400~ for about 2 hr were adequate to cause narrowing of the /S 003/ 005 peak and to reveal the diagnostic polytype reflec- tions. XRD traces of a sample exposed for as long as hr to ethylene glycol vapor at ambient temperatures (Figure 4) showed maximum definition of the 112 and 112 reflections after 72 hr (Figure 4d). Note that the base of the /S reflection at 26.6~ narrowed and the /S 003/005 reflection sharpened by about the same amount after glycolation (Figure 4) or heating (Figure 5). This similarity suggests that glycolation and heating were equally effective in resolving polytype compositions. After heating the sample for 2 hr at 100~ (Figure 5b), the 112 reflection was distinctly resolved from the peak at about 26.6~ for the air-dried sample; however, the 112 reflection was represented only by a shoulder on the peak. Although the 26.6~ peak did not intensify, it narrowed due to collapse of expandable material. The sample was then heated for an additional 2 hr at 375~ and a third XRD trace was made (Figure 5c). The 12 reflection was deafly resolved, and a weak 112 reflection was noted. The 003/005 peak was quite narrow and had increased in peak height by ~ 10%. The sample was heated for an additional 2 hr at 520~ and a fourth XRD trace was made '(Figure 5d). Both the diagnostic 1M reflections (112 and 112) and the 003/005 peak showed only minor changes from the previous trace. A fifth XRD trace (Figure 5e), made after heating the sample to 600~ for an additional 2 hr, showed intensified 112 and 112 reflections. Both reflections, however, had shifted to larger d-values (by about 0.6~ suggesting structural modification. This modification was further indicated by the better resolution of the 112 vs. the 112 reflection. The 1M structure may be unstable even at 520~ at ambient pressures. The diagnostic 1M reflections of some samples were completely destroyed after 2 hr at that temperature. DSCUSSON The 1Md band, stippled from 20~176 in Figure 2a (Eberl et al., 1987), is probably due in part to 00l peaks of partially hydrated (0, 1, and, perhaps, 2 water layers) /S. As seen in Figures 1 and 3, the sharpening of the /S 003/005 reflection was accompanied by resolution of the 112 and 112 diagnostic 1M reflections. The remainder of the elevated background near the / S 003/005 (~ 26.6*20) peak may have been due to structural disorder, particles small enough that line broadening of hkl polytypic reflections may make them appear noncrystalline to the X-ray beam, or to truly noncrystalline material. The treatments used in the present investigation probably did not affect the resolution of the 112 and 112 peaks (or other diagnostic hkl reflections), if the weaker, broader, and less distinct diagnostic reflections of air-dried samples were due only to rotational stacking faults (Drits et al, 1984) or noncrystalline material. Any remaining lack of distinctiveness after these treatments, however, most

6 Vol. 37, No. 2, 1989 Polytype structure of 1Mdillite 133 probably was due to that type of disorder and/or crystallites that appear to be noncrystalline to X-rays. Thus, the treatments used merely "uncovered" the diagnostic peaks. Careful investigation of the lattice nodes for smectite containing various cations or intercalates and for collapse on heating might indicate other treatments, such as K + exchange, that would sharpen the /S 003/005 peak even further and that would reduce sample-preparation time. A comparison of the XRD traces resulting from various ions, intercalations, and heat treatments would increase the likelihood that a refined model of the actual nature of the mixed-layered mineral could be described. SUMMARY This study suggests that truly Md illitic clay minerals are far less common than previously thought and that mixtures of 1M, 2Ml and Ad illite structures may be much more prevalent. Earlier workers defined the 1Md polytype on the basis of the lack of or the low intensity of diagnostic 1M or 2M reflections. Recent workers have defined 1Md by the absence of discrete hkl reflections, with k 4= 3n as the criterion, and by the presence of an elevated background between about 20 ~ and 33~ (CuKa). Expansion of hydratable layers in illite by ethylene glycol for 40 to 80 hr at ambient temperatures, or 4 hr at 80~ or collapse of the hydratable layers on heating the sample to 3750(2 for 2 hr, may resolve diagnostic 1M and 2M~ XRD reflections. The study also suggests that the resolution of diagnostic hkl reflections by glycolation or heating is relatively long-lived, because those reflections remained resolved, even after days of exposure to ambient conditions. n addition to the diffuse 00l diffraction effects discussed in this report, most illitic samples contain a diffraction band or elevated background between about 20 ~ and 33~ due to structural disorder, fine particle size, or the presence of noncrystalline material, or some combination of these factors. Further research is necessary to determine whether 2- and 3-layer disordered structures can be distinguished. For now, it seems appropriate to restrict the use of 1Md to samples in which a disordered polytype is present, i.e., hkl reflections in which k v~ 3n reflections are diminished and an appropriate amount of elevated background between 20 ~ and 33~ is present, and 1M is the only ordered polytype. Mixtures of ordered polytypes and a disordered structure, as well as disordered illites in which no ordered polytype can be identified, should be labeled as partly or wholly Ad (disordered group A micas). Because of the importance of correct illite polytype determinations for the understanding of the weathering, depositional, thermal, and diagenetic history of illitic sedimentary rocks, we recommend that all samples be treated as described in this report before being labeled as 1Md or being analyzed for proportions of 1M, 2M1, and Ad illite. ACKNOWLEDGMENTS We thank the New Mexico Bureau of Mines and Mineral Resources for supporting a sabbatical at the llinois State Geological Survey and the llinois Survey for some expenses and space for the senior author. We also thank our colleagues, Jacques Renault, D. D. Eberl, F. A. Mumpton, R. C. Reynolds, Jr., and Jan Srodofl who read earlier versions of the manuscript. Special thanks, too, go to S. W. Bailey, D. M. Moore, and R. C. Reynolds for their extra help and advice on nomenclature problems and structural interpretations. REFERENCES Austin, G. S. and Leininger, R. K. (1976) Effects of heattreating mixed-layer illite-smectite as related to quantitative clay mineral determinations: J. Sed. Petrol 46, Bailey, S. W. (1980) Structure of layer silicates: in Crystal Structures of Clay Minerals and their X-ray dentification, G. W. Brindley and G. Brown, eds., Mineralogical Society, London, Bailey, S.W. (1988) X-ray identification of the polymorphs of mica, serpentine, and chlorite: Clays & Clay Minerals 36, Brindley, G.W. (1980) Order-disorder in clay mineral structures: in Crystal Structures of Clay Minerals and their X- ray dentification, G. W. Brindley and G. Brown, eds., Mineralogical Society, London, Bruce, C. H. (1984) Smectite dehydration--ts relation to structural development and hydrocarbon accumulation in northern Gulf of Mexico Basin: Amer. Assoc. Petroleum Geol. Bull. 68, Drits, K. A., Plan~on, A., Sakharov, B. A., Besson, G., Tsipursky, S.., and Tchoubar, C. (1984) Diffraction effects calculated from structural models of K-saturated montmorillonite containing different types of defects: Clay Miner. 19, Eberl, D. D., Srodofi, J., Lee, M., Nadeau, P. H., and Northrop, H. R. (1987) Sericite from the Silverton Caldera, Colorado: Correlation among structure, composition, origin and particle thickness: Amer. Mineral 72, Hoffman, J. and Hower, J. (1979) Clay mineral assemblages as low-grade metamorphic geothermometers: Application to the thrust faulted disturbed belt of Montana, U.S.A.; in Aspects ofdiagenesis, P. A. Scholle and P. R. Schluger, eds., Soc. Econ. Paleontol. Mineral. Spec. Publ. 26, Levinson, A. A. (1955) Studies in the mica group: Polymorphism among illites and hydrous micas: Amer. Mineral. 40, Mrring, J. (1949) L'interfrrence des rayons-x dans les syst~mes ~ stratification drsordonnre: Acta Crystallogr. 2, Reynolds, R.C. (1980) nterstratified clay minerals: in Crystal Structures of Clay Minerals and their X-ray dentification, G. W. Brindley and G. W. Brown, eds., Mineralogical Society, London, Schultz, L.G. (1960) Quantitative X-ray determinations of some minerals in rocks: in Clays and Clay Minerals, Proc. 7th Natl. Conf., Washington, D.C., 1958, Ada Swineford, ed., Pergamon Press, New York, Srodofi, J. and Eberl, D. D. (1984) llite: in Micas, Reviews in Mineralogy 13, S. W. Bailey, ed., Mineralogical Society of America, Washington, D.C.,

7 134 Austin, Glass, and Hughes Clays and Clay Minerals Velde, B. and Hower, J. (1963) Petrological significance of illite polymorphism in Paleozoic sedimentary rocks: Amer. Mineral 48, Weaver, C. E. and Brockstra, B. R. (1984) llite-mica: in Shale Slate Metamorphism in the Southern Appalachians, C. E. Weaver and Associates, eds., Elsevier, New York, Yoder, H. S. and Eugster, H.P. (1955) Synthetic and natural muscovites: Geochim. Cosmochim. Acta 8, (Received 6 August 1987; accepted 28 September 1988; Ms. 1705}