THE ALTERATION PRODUCTS OF POTASSIUM DEPLETED OXYBIOTITE

Transcription

1 Clays and Clay Minerals, 1973, Vol. 21, pp Pergamon Press. Printed in Great Britain THE ALTERATION PRODUCTS OF POTASSIUM DEPLETED OXYBIOTITE R. J. GILKES Department of Soft Science and Plant Nutrition, Institute of Agriculture, University of Western Australia, Nedlands, W.A (Received 19 February 1973) Abstract--Artificial weathering of biotites, which contain various levels of structural ferric iron, by NaCI and NaBPh4 solutions produces minerals and structures similar to those described for naturally weathered biotites. Oxidation of structural iron leads to K removal from alternate layers and development of hydrobiotite. The growth of order with increasing ferric iron content has been assessed by comparison with theoretical calculations for random and most ordered interstratified structures. There is evidence for the existence of two layer types in biotite prior to oxidation. The depression in rates of K release due to oxidation has been confirmed. INTRODUCTION DESPITE numerous investigations of rates of K release from micas and the structures of the alteration products, there still remains a large degree of uncertainty as to the influence of mica structure and chemistry (mica polymorph, hydroxyl configuration, fluorine content, oxidation state, etc.) on these aspects of mica weathering (Rausell-Colom et al., 1965; Newman, 1969; Scott and Reed, 1962). The weathering of biotite is of special interest to soil scientists since this mineral is a common component of a variety of igneous and metamorphic rock types from which many soils have developed. Biotite may also contain a significant proportion of the soils' store of essential micro- and macronutrient elements so that the availability of these nutrients will depend on the rate and mode of alteration of this mineral. The work of Gilkes et al. (1973b) has shown that the availability to plants of Mg and Zn from biotite as measured by 0.1 M HC1 extraction is much depressed by oxidation of structural iron. This investigation describes the influence of oxidation of structural iron and type of extractant on the structure of K depleted biotites. Several workers have described the structures present in synthetically (Gilkes et al., 1973a; Scott and Reed, 1962; De Mumbrum, 1959; Mortland, 1958) and naturally (Walker, 1949; Wilson, 1970) weathered biotite but their results are often fragmentary and conflicting. In particular, the structures found in naturally weathered biotites are not necessarily those predicted by laboratory weathering experiments. In no case has a systematic study of the effects of progressive oxidation and K depletion been undertaken. MATERIALS AND METHODS Four oxidized specimens of the less than 10/xm fraction of Bancroft biotite were prepared by the bromine oxidation technique described by Gilkes et al. (1972). The ferrous iron contents of the biotites as Fe were 12.0, 10.1, 7.9 and 4.7 per cent with total iron as Fe remaining constant at 14.6 per cent. Structural formulae for similar specimens are given by Gilkes et al. (1972). The method of potassium extraction from these specimens by sodium tetraphenylboron (NaBPI'u) and NaCI solutions was similar to that described by Gilkes et al. (1973a) with some modification so as to produce a greater range of K levels in the biotite specimens. The 50 mg samples of each biotite were shaken with 25 ml of I M NaCI in 30 ml polypropylene centrifuge tubes for 24 hr at 20~ The suspension was centrifuged at 10,000 r.p.m, for 10 min. A 20 ml aliquot was removed and replaced by 20 ml I M NaC1. All the extracting solution was not removed since this invariably leads to loss of biotite. This procedure was repeated up to 20 times with samples of biotite being taken for X.R.D. analysis at approximately equal intervals of K depletion. Potassium was determined by atomic absorption spectrometry (A.A.S.). A further ten 50 mg samples of each biotite were extracted with a I M NaC1/0.02 M NaBPh4 solution; depleted biotites produced by this treatment will be referred to as NaBPI'u extracted biotites. The extractions for different biotites ran for various times 303

2 304 R.J. GILKES as shown in Fig. 1 to compensate for the decreased rates of K release that occur in oxidized biotite. Samples were shaken in 12 ml polypropylene centrifuge tubes at 20~ On completion of shaking, 5 ml of A.R. acetone were added, the samples shaken immediately to dissolve precipitated KBPh4 and centrifuged at 5000 r.p.m, for 5 rain. The resulting supernatant was transferred by micropipette to a 25 ml volumetric flask. The samples were then washed with 10 ml 1 : 1 acetone/water, shaken and re-centrifuged at 5000 r.p.m, for 5 rain. The two supernatants were combined and made up to volume with 1:1 acetone/water: 5 rnl of extractant solution were added, and the process repeated 10 times. Potassium was analysed by A.A.S. The total potassium content of a given biotite was determined by A.A.S. after dissolution in 2 ml of 40 per cent hydrofluoric acid and 2 ml of 70 per cent perchloric acid. The residues were taken up in concentrated HC1. Samples were made up to volume in 3000 ppm strontium chloride. Ferrous iron was determined by dichromate titration after dissolution in 3 ml hydrofluoric acid and 20 ml conc. HC1. Prior to X.R.D. examination the K depleted biotite samples were shaken for several hours in 1 M magnesium chloride to replace the Na on exchange sites by Mg. The biotites were then sucked onto porous ceramic plates and given several washes with distilled water. After 24hr drying at room temperature and humidity, X-ray difffraction patterns were obtained using a Philips vertical goniometer with copper radiation. O ~ 120% Fe 2. //~,/A~l~ RESULTS The release of K from biotites as a function of cumulative time of the successive extractions by NaBPI~ and NaC1 solutions are shown in Figs. 1 and 2. These results demonstrate that the depressed rates of K release due to oxidation of octahedrally coordinated iron continue to higher values of total K lost than those measured by Gilkes et al. (1973a) for similar oxybiotites. Specimens for X-ray diffraction examination were taken at 10 successive increments of K depletion from both NaC1 and NaBPI~ extractions. With the exception of the NaCl-extracted biotite which contained 4-7 per cent Fe 2+, more than 50 per cent of initial K was removed by these treatments. In all cases the alteration products were sufficiently well developed to give strong diffraction patterns. Gilkes et al. (1973a) found that some replicate NaBPh4 extracted biotites gave different diffraction patterns. It is now believed that these differences were due to incomplete exchange of interlayer Na by Mg prior to X-ray diffraction. The biotites had been leached with 1 M magnesium chloride after preparation on ceramic plates, a technique that is successful with clay size materials but does not appear to allow for complete exchange for large particle sizes. In this study complete exchange was achieved by shaking the K depleted biotites with 1 M magnesium chloride solution for several hours prior to preparation of the specimen on porous ceramic plates, Replicate samples give identical diffraction patterns using this technique. Examples of X-ray diffraction patterns for each of the biotites are given in Figs. 3 and 4. Patterns for the unextracted biotites are not shown but in 100 8(: ~ 4C _e /o,,/ /~ />~...,,.~,,, I~0 O,,.O.O,,.ooO ~ 120"/* F'~'* 0 f /o y 60 0/0/0 ~...~A---'~ ~A''~-Z~" ~01 ~176 9 ^.-~ ~40 /~ ~" I~.~E7 x x x_x--x.--x--x--x--47% -- 2C.~IxT~1~ x'~x i ~0 ~0 CUMULI~i'IVE EXTRACTION TIME (HRS) Fig. 1. Percentage of initial K removed by successive NaBPh, extractions of oxybiotites, i i i NUMBER OF EXTRACTIONS Fig. 2. Percentage of initial K removed by successive NaC1 extractions of oxybiotites.

3 Potassium depleted oxybiotite 305 IO.~A I "~,2A 47A 120Wo Fe 2' I / S l/ 357g I V~SA 47W.Ve 2. / J I~t 24A ~ 40WOK,i, 24~l 4 7~ I Il~l \ I 40WOK 4.84A / ~1 kj \ 4 7"l're2" 510/,K / I x 87.,2A 70 ~176 Fig. 3. Diffraction patterns for basally oriented, Mg ~*- saturated oxybiofites after K extraction by NaBPI~. Initial ferrous iron content and percentage of initial K remaining after extraction are shown for each pattern. each case they gave a sharp integral 00( series of reflections with a basal spacing of 10,h,. (i) NaBPh4 extraction (Fig. 3). The 12 per cent Fe z* biotite altered to vermiculite with minor amounts of a mixed layer phase and no evidence of a regularly interstratified hydrobiotite phase. Increasing oxidation led to a gradual increase in both the intensity and sharpness of hydrobiotite reflections in the potassium depleted specimens. This change was accompanied by the disappearance of a discrete biotite phase, its place being taken by an interstratifled phase. This new phase increased in basal spacing as the percentage K extracted increased, up to a value of 12.2A at which stage a regular series of hydrobiotite reflections had developed. For all oxidation states the amount of vermiculite present increased with amount of K extracted. Fig. 4. Diffraction pattern~ for basally oriented, Mg 2*- saturated oxybiotites after K depletion by NaC1 solution. Initial ferrous iron content and percentage of initial K remaining after extraction are shown for each pattern. (ii) NaCl extraction (Fig. 4). The least oxidized specimen (12-0 per cent Fe 2+) shows a different behavior to that found for NaBPh4 extraction. Interstratification of the whole biotite crystal occurred as K was removed, evidenced initially by a broadening of the (001) and (003) reflections and subsequently by displacement of these reflections to higher basal spacings. Vermiculite and weak, diffuse hydrobiotite reflections appeared as the amount of K removed increased. Increased oxidation resulted in sharper hydrobiotite reflections, although this mineral never became as abundant as in the NaBPh4 treated specimens. Mixed layer phases, which increased in basal spacing as the percentage K removed increased, are present in all specimens.

4 306 R. J. GILKES INTERPRETATION Alteration products The diffraction patterns shown in Figs. 3 and 4 represent a small proportion of the total number of patterns obtained in this study. A simple summary of the alteration products for all combinations of ferrous iron content and K loss is difficult to represent since several phases may exist in each specimen. These phases may also vary in degree of structural regularity; i.e. biotite, vermiculite, hydrobiotite of variable stacking perfection, mixed layer minerals of variable spacing. Representation is considerably simplified if one considers that only two types of layers are involved in all these structures. These types are biotite-like layers (called B layers hereafter, d(001)= 10~), and vermiculitelike layers (V, d(001)= 14.2A) formed by the replacement of interlayer K by hydrated Mg ions. It is possible to plot the distribution of these two types of layers in the different mineral species as a function of K depletion for the series of oxybiotites. This representation is at best only semiquantitative since it requires an estimation of the relative abundances of the various minerals from their X-ray diffraction patterns, a procedure that is notoriously unreliable for such minerals (Brindley, 1961). Estimation of the abundance of discrete biotite and vermiculite phases can be done with reasonable accuracy since diffraction patterns of pure end members can be used as standards. Errors due to differences in the structure and chemistry of standards and unknowns are thus avoided. Assessment of the abundance of hydrobiotite by the same technique can also be used since some K depleted specimens contain only hydrobiotite and vermiculite. This situation allows the intensity of basal reflections from pure hydrobiotite to be computed after correction for the amount of vermiculite determined to be present in the specimen. Unfortunately, the stacking perfection of hydrobiotite in different samples is variable so that errors in quantitative analysis due to differences in hydrobiotite structure will arise. The influence of stacking order on the spacing and intensity of hydrobiotite basal reflections is discussed in the next section. Despite these departures from perfectly regular stacking in hydrobiotite, the mineral was assumed to be composed of equal numbers of B and V layers. Quantitative estimation of the abundance of interstratified phases of variable basal spacing is even more prone to error, particularly when hydrobiotite and interstratified mineral reflections occur at the same angle. A value for the abundance of these phases was determined by difference in this study and checked by inter-comparison of samples containing similar minerals to see that equal peak intensities gave similar concentrations. The abundances of the different minerals so determined were then normalised to give a total mineral content of I00 per cent. The proportion of B and V layers in the interstratified phases was determined from MacEwan et al. (1961) curves relating basal spacing to the proportion of each phase present. Reflections were used which are not sensitive to the degree of ordering of the component layers. The identification of this type of reflection is discussed in the next section. The proportion of B and V layers for each of the mineral species present and the total proportion of these layers in each sample were plotted against percentage K replaced. The plots of total numbers of B and V layers present against K replaced Should each show an approximately linear relationship. Figure 5 is such a plot for NaBPhrtreated biotite with 4.7 per cent Fe 2+, each of the 10 experimental points corresponding to a diffraction pattern. The major features of the alteration of this mineral are clearly seen using this form of representation. An interstratified phase containing a preponderance of B layers formed during the initial stages of K depletion along with lesser vermiculite. No unaltered biotite remained. The proportion of V layers in the interstratifled mineral and the abundance of vermiculite increased as successively more K was removed. As the proportion of V layers in the interstratified mineral passed 40 per cent, regular alternation of the B and V layer types became apparent and hydrobiotite reflections became more intense. Thus, the interstratified mineral in this sample was a precursor of hydrobiotite. With further losses of K some B layers of the hydrobiotite phase were converted to V layers and the discrete vermiculite phase grew in abundance. There is no evidence for interstratification of hydrobiotite with excess V layers, which suggests that this stage of alteration was accomplished by a pure V front moving through the crystal rather than the replacement of complete B layers by V layers throughout each crystal. The total V layer content of the three minerals shows an approximately linear relation with percentage K removed indicating that this representation gives a reasonable summary of the alteration process. This particular example shows a larger than normal discrepancy between the experimental and theoretical lines relating percentage K removed to percentage V layers. Most plots do indicate that the percentage V tayers has

5 Potassium depleted oxybiotite 307 TOTALV L,! 8O- // eo- / -," ~.:-~ ', ~. K J Ix'"'~ l ~\Q I i J \ O 60 8O ~ OF INfTIAL K REPt.ACED Fig. 5. The distribution of V and B layers in the different mineral phases present in NaBPIh treated oxybiotite (initially 4.7 per cent Fe:+): [] [] Total V and B layers; ( Interstratified mineral; x... x hydrobiotite; I vermiculite. been over-estimated and therefore the percentage B layers under-estimated. Similar curves at a reduced scale and omitting experimental points for clarity are shown for all specimens in Figs. 6 and 7. The following trends in alteration behavior are demonstrated by these curves: (a) Hydrobiotite increased in abundance as oxidation increased for both the NaBPh4 and NaC1 extracted biotites. (b) Interstratified phases showing an increase in percentage V layers with K removed were the first alteration product for all NaCI extracted biotites and for the most highly oxidized (4-7 per cent Fe 2 NaBPI~ extracted biotite. (c) No residual biotite remained in the NaC1 extracted biotites or in the 4.7 and 7.9 per cent Fe 2 NaBPI~ extracted biotites. (d) Vermiculite formed at a lower value of percentage K removed for oxidized biotites. This complex interaction between potassium level, oxidation status and the type of extraction helps explain why different authors have found a variety of mineral species in both naturally and artificially weathered biotites. Walker (1949) in his study of naturally occurring weathered biotites in Scottish soils reported vermiculite formation with no intermediate regularly interstratified phase. He did detect a shift of the biotite basal spacing to 10.4 A. Wilson (1970) has followed the conversion of biotite to vermiculite via a hydrobiotite phase in a Scottish soil profile. In both of the above exam- pies alteration was accompanied by oxidation of most of the structural iron. Scott and Reid (1961) found that biotite altered directly to vermiculite during NaBPI~ extraction with no evidence of interstratified phases. Mortland (1958) continuously leached biotite with 0.1 M NaC1 and also reported direct conversion to vermiculite. Mortland (1958) suggested that these results may have been due to the preservation of low levels of K in the extracting solution which allowed K to diffuse rapidly from within the mineral in the region behind the alteration front. The NaC1 extraction technique used in this work resulted in a gradual increase in the K level in solution with duration of extraction. This increase may have resulted in the termination of exchange in different layers at different solution K levels and led to the development of an appreciable quantity of interstratified phase. De Mumbrum (1959) used NaBPI~ to extract K from biotites and, in contrast to Scott and Reid (1962), found that large amounts of an interstratified phase developed. This difference may have been due to oxidation of structural iron in biotite which occurred during NaBPh4 extraction (Newman and Brown, 1966). As described above, the structure of K depleted biotites is very sensitive to oxidation state. Stacking order in hydrobiotite The interpretation given in this section makes use of the nomenclature and theory used by Ruiz Amil et al. (1966). Since these authors calculated diffraction effects for mixtures of 10.0 ]k (layer type A) and 14.0 ~ (layer type B) components instead of

6 308 R.J. GILKES,. "l.v 12.0,/.Fe2+ //80- 'I.B 12.0.i, Fe 2, //~ov!,, /./ 60- j./ 40- /" 20-,.-/' "/, Fe 2..,K. o]ov ~" // 80- /./ 60- / /// " 40" P / s \ / */.V 10'1"f* Fe~i~ ' *I.B \ \\ \\ \\\...?.%,s,.. " % Fe 2. / */.V "- ~/.z / 80- L / f. 9 fi "". \,'~ 7,9*l.Fe 2o % B ~30- ~ v s' t," 60" 40-60" J // 40- " "/,~ *1.V *lab..j 40' I "'9 "'.. "9... \,,s /... \ 4.-/./.Fe *I.V 90' 60" 40 " *I,B X \ "" I I "'. I 9,o 6o 8o 2b 9 1o OF INITIAL K REPLACED 9 "1 I' {Y3 Fig. 6. The distribution of V and B layers in the different mineral phases present in NaBPh, treated oxybiotites. The initial ferrous iron contents of each specimen are shown in the figures. (- total V and B layers, -... vermiculite, interstratified minerals, 9... hydrobiotite, biotite). the 10.0 A and 14.2-& components which occur in interstratified mica-vermiculite minerals, there are minor discrepancies between observed and calculated basal spacings. The values of basal spacing for the various stacking parameters p(a), p(aa), etc. used to describe interstratified structures can be derived from curves published by Ruiz Amil et al. (1966). Use was made in the previous section of the observation that basal spacings of some reflections depend solely on p (A) and are not influenced I 20 4O % OF" INITIAL K REPLACED Fig. 7. The distribution of V and B layers in the different mineral phases present in NaC1 treated oxybiotites. The initial ferrous iron contents of each specimen as Fe are shown in the figures. Key as for Fig. 6. by the value of p(aa). Other reflections vary in spacing or may be weak or absent depending on the value of p(aa). The behavior of several reflections is shown in Fig. 8, where basal spacing is plotted against p(a) for random (p(a)= p(aa)) and most regular (p(aa) = minimum) arrangements of layers. The concept of a most regular arrangement has been discussed by Gilkes and Hodson (1970), who show that p (AA) takes the following minimum values: p(aa)=o, p(a)<-0.5 ''

7 Potassium depleted oxybiotite 309 _p(a)-p(b) p(aa)-- ~ -, p(a)>0.5. This is simply a statement that minority components are never adjacent. The (001/001), (002/003) and (003/004) curves for random and most regular arrangements coincide as do those for intermediate values of p(aa). This allows these curves to be used for the determination of p(a) irrespective of the value of p(aa) as was done in the quantitative estimation discussed in the previous section. Rhodes and Coleman (1967) have also used this result, derived from a comparison of spacings for random and perfectly regular arrangements, to interpret diffraction patterns for interstratified mica-vermiculite minerals. They did not discuss extension to the case of most regular arrangements which occur when p(a) is other than 0.5. A tendency towards alternation of layer types (i.e. a most regular arrangement) should be revealed by the appearance of reflections at approximately 2r BEEN]" ~ MIN oo,,.,,,2[-. ~,6v,j,,, o,;o /002 / 8 ~-H X~O 'x 7 i, 2.,3 / X 5~ / 47 L~'~'~ Ox'~ '- ' I I t~, "#" 3.3 ~ J a. L I 31 t 0O31O05 / 3o[. / [, ~,,,, 26[" i i i x, ~' o12 c~ ~ ds r o2 o4 o6 oe ~!o p(a) p (a) Fig. 8. Basal spacings for interstratified 10-& (layer type A) and 14 ~k (layer type B) components as a function of p(a) taken from Ruiz Amil et al. (1966). The two sets of curves correspond to completely random (O O) and most regular (. x ) arrangements. CCM--Vol. 21 No. 5--D 24_~ ( ), 6"0/~ (002/002) and 3-0/~ (003/005). The reflections at approximately 8A (001/002) and 2.67 fk (004/005) will move towards values predicted by the curves for most regular interstratification in Fig. 8. Inspection of the diffraction pattern of well crystalline hydrobiotite shown in Fig. 3 indicates that the intensities of the diagnostic 8 and 6,~ reflections are low and may not be measurable if little hydrobiotite is present or if a significant deviation from perfect regularity occurs. The reflections at 24 and 3 A are relatively stronger and are the best indication that the structure has a tendency towards regular interstratification. Accurate spacing measurements for the 3.0 and 2.67/~ reflections will confirm that the structure is the most regular one for a particular p(a). Applying these results to the poorly ordered hydrobiotites formed by K removal, we find that these two reflections are usually too diffuse to permit accurate measurement of basal spacing. Since the curve relating basal spacing to p(a) for the 003/005 reflection is fairly flat in the vicinity of p(a)=0.5, this uncertainty leads to a large error in p(a). Ruiz Amil et al. (1966) have shown that the relative intensities of some reflections from interstratified minerals are very sensitive to the degree of order (i.e. p(aa)) of the component layers. This sensitivity allows the relative intensities of particular pairs of reflections to be used as a measure of the degree of order of hydrobiotite formed by K depletion of biotite. Figure 9 shows the ratio of amplitudes of the mixing for the 24 and 12/~ reflections of hydrobiotite as a function of p(aa) taken from Ruiz Amil et al. (1966). We see that for p(a)= 0.4 to 0-6 a moderately strong 24 reflection may occur but its intensity decreases rapidly as p(aa) moves away from its minimum value. Thus, a plot of relative experimental intensifies for this and other sensitive pairs of reflections against ferrous iron content can be used to show the influence of oxidation on hydrobiotite ordering (Fig. I0). These experimental curves show that the stacking sequence of hydrobioute formed by K depletion of oxybiotites using NaBPh4 and NaCI solutions increases in regularity with increasing oxidation status. An increase in p(a) at constant p(aa) may also increase these ratios (Fig. 9) but the slight shift in basal spacing that occurs (Figs. 3 and 4) shows that these differences are occurring at approximately constant p(a). Figures 9 and 10 cannot be compared directly since the ratio of mixing functions is not proportional to the intensity ratio. These two quantities

8 310 R.J. GILKES O.O2 0 )=05 0 oi o p(aa) Fig. 9. Ratio of amplitudes of the mixing function (b for the 24 and 12 t%, reflections of hydrobiotite as a function of p(aa). Values are taken from the theoretical curves of Ruiz Amil et al. (1966). ~ Io I (347A) /,,"-~ 07 F~ = layer structure factor qb = mixing function. Unfortunately, many sources of error prevent a comparison of the theoretical ratio with experimental results. The angular factor for the oriented powder specimens used in this investigation lies between those for a single crystal and for a random powder and may vary between specimens. The layer structure factor F~ cannot be calculated for non-regular mixed layer structures and is usually approximated by use of the structure factor for a 10/~ thick mica sheet. This approximation may lead to a considerable error for highly ordered interstratified structures. Despite these limitations the ratios of intensities of sensitive reflections to insensitive reflections shown in Fig. 10 do depict in an easily visualised form the increased order that occurs with increasing oxidation of structural iron. O2 ~5 ~ I I J I I I I I 1 i I I i i ~d I:~I~CENT Fe ~' IN OXYBIO'mE Fig. 10. Ratios of intensities of pairs of hydrobiotite reflections as functions of initial ferrous iron content of oxybiotite. Measurements were made on patterns corresponding to the maximum development of hydrobiotite. O NaBPh,, Z~ NaC1. are related by the equation (Ruiz Amil et al., 1966) I= OIFtl2(I ) I = intensity O = angular factor Model o.f the weathering process In the preceding sections the types of minerals and structures present in K depleted oxybiotites and their relative abundances were determined. It is necessary for an understanding of the weathering mechanism to visualize the distribution and development of these components in a single crystal. This has been done schematically in Fig. 11 which assumes a symmetrical distribution of phases in each crystal with the latest developing phase starting at the crystal edge. Observations of the alteration of large mica crystals confirm that this is the most important mode of alteration (Rausell-Colom et al., 1965). A different nomenclature to that used in the previous sections has been adopted for Fig. 11. The major difference lies in the use of the terms I* and I** to describe interstratified minerals containing approximately equal numbers of B and V layers but with different degrees of order (p(aa)), a double asterisk signifying a higher degree of order. In the semiquantitative analysis used to derive Figs. 6 and 7, the intensities of specific hydrobiotite reflections were compared with those of perfect hydrobiotite in the sample. As already discussed, these reflections are still present for minerals showing considerable departures from regular interstratification. Also, an approximate measure of the degree of order may be obtained from the intensity and sharpness of these reflections. Thus, three minerals showing specific hydrobiotite reflections but with an increasing degree of regularity are now recognised as I*, I** and H.B., respectively. The min-

9 Potassium depleted oxybiotite 311 ~1:05 INCREASING OXIE~nON f~)ne "rpb BIOTITES B I ~ p~b)t 0.5 INCREASING OXIDATION (b) NeO BIOTITE~CJ I ~ V I ml V ata~5 V ~,O5 m lib I V V i LA ~ v, ro{b)'-(>9 i Hl~ v 0-45 ~)~OB Fig. 11. Schematic representation of the development of different alteration :products from oxybiotite treated with NaBPh, and NaC1 solutions. A mirror plane exists to the L.H.S. of each crystal. The abbreviations used in this figure are identified in the text. eral species present in the alteration products have been described earlier and are now shown schematically in Fig. 11. By reference to this figure it becomes easy to visualize the successive stages of alteration. Oxidation of octahedral iron clearly results in the presence of two types of layer which alternate in sheets throughout the crystal and which show different rates of K release to extracting solutions. In addition to the two classes of interlayer K produced by oxidation, the interlayer K sheets in unoxidized biotite also show a range of K releasing capacities. There is a wide variation in the magnitude of this property between layers since complete layers of K are removed by NaCI treatment while adjacent layers are unaffected. The distribution of these layers is random in this case since only a weak ten- dency towards ordering develops in the interstratified phases present in NaCI- and NaBPh4- extracted biotite containing 12.0 per cent Fe 2+. The small amount of ordering that is present may relate to the 2-6 per cent Fe 3 present in this specimen. Once half the B layers have altered to V in the 12.0 per cent Fe 2+ biotite, little further exchange occurs throughout whole layers of the crystal but a V front moves in from the edge of the crystal replacing all remaining B layers at the same rate. This suggests that two classes of interlayer K sheets may also exist in roughly equal amounts in unoxidized biotite. In this case they are distributed in an approximately random arrangement and do not regularly alternate as in oxidized biotite. Thus oxidation may simply lead to an ordered arrangement of these two types of layer. This interpretation is supported by evidence of the structure adopted by both NaBPh4 and NaC1 extracted, oxidized biotite (Fig. 11). The initial interstratified phase I contains V layers extending throughout the crystal which are probably distributed randomly throughout the alternate layers that ultimately become V layers in perfectly ordered hydrobiotite. Diffraction patterns show that no ordered occupation of these alternate sites occurs since reflections corresponding to regular arrangements such as BBBVBBBVBBBV, etc. are not present. After further K depletion a strong hydrobiotite pattern develops, this mineral being more abundant in NaBPI~ extracted oxybiotite. The next stage consists of a sharply defined alteration front moving through the crystal at which all the B layers in hydrobiotite are converted to V layers at the same rate. Rhoades and Coleman (1967) observed a similar effect in their study of K fixation by Na and Mg 2+ saturated vermiculites and K depleted Bancroft biotite which was probably partially oxidized by NaBPh4 treatment (Newman and Brown, 1966). Hydrobiotite, rather than a randomly interstratified mineral with excess V layers, was formed during the early stages of K absorption up to the point at which all vermiculite disappeared. As further K was absorbed a randomly interstratified mineral with an increasing content of B layers developed. This sequence, which is also reported by Sawhney (1972), is exactly the reverse of that observed in this study for K removal from oxybiotite. Differences in the diffusion constant of K from within the two types of layer to NaBPI~ and NaCI solutions may control the relative amounts of hydrobiotite and vermiculite present in K depleted oxybiotite. Mortland (1958) has suggested that the

10 312 R. J. GILKES extent of development of interstratified phases during K release depends on the K level in extracting solutions. Thus in Mortland's (1958) work which used continuous leaching with 1 M NaC1 or Scott and Reeds' (1962) experiments with NaBPh4 solutions, the concentration of K in solution remained very low so that little or no interstratified phase developed during the conversion of biotite to vermiculite. In this study little interstratified phase formed during NaBPh~ extraction of 12.0 per cent Fe 2+ biotite. Major interstratification throughout the entire volume of each crystal occurred when successive extractions with 1 M NaC1 containing variable levels of exchanged K were used. The level of K in solution controls which interlayers will lose K, the complete K content of an interlayer being lost once exchange has been initiated. CONCLUSIONS This investigation has helped clarify relationships between the structure of weathered biotite and mode of leaching and oxidation status. The extent of random interstratification in slightly oxidized biotites is sensitive to K level in the extracting solution. A tendency towards ordered interstratification develops as the proportion of oxidized structural iron increases. Perfectly ordered hydrobiotite is produced by K depletion of highly oxidized biotite. Many of the structures or mineral assemblages reported by various authors (Walker, 1948; Wilson, 1970; Scott and Reed, 1962; De Mumbrum, 1959; Mortland, 1958; Rausell-Colom et al., 1965) for both naturally and artificially weathered biotites have been formed during this study. It may be possible to relate the structures adopted by naturally weathered biotites in soils to oxidation state, ease of drainage and concentration of K in soil solution. The results of these experiments also indicate that biotite oxidation may result in the ordering and modification of two types of layer which are initially present in equal amounts and in random arrangement in unoxidized biotite. The different properties of the two layer types have not been identified in this study. The presence of octahedrally coordinated ferric iron (Gilkes et al., 1972) is unlikely to be directly responsible since Tomita and Dozono (1972) have shown that K depleted sericite, low in iron, when heated to 800~ also adopts a regular 24 -A structure. Rectorite (Brown and Weir, 1963) also has a regularly alternating structure of paragonite--like sheets containing no structural iron. Thus oxidation of biotite must induce some structural change such as altered hydroxyl orientation, distorted or contracted tetrahedral sheets (Gilkes et al., 1972) or loss of hydroxyl (Farmer et al., 1971) which then leads to differences in exchangeability of K from between adjacent sheets. Acknowledgements--This study was supported by a University of Western Australia Research Grant. I am grateful to Mr. R. C. Young, Professor J. P. Quirk and Mr. Alan C. Wright for their assistance. REFERENCES Brindley, G. W. (1961) Quantitative analysis of clay mixtures. In The X-ray Identification and Crystal Structures of Clay minerals (Edited by Brown, G.) Mineralogical Society, London. Brown, G. and Weir, A. H. (1963) The identity of rectorite and allevardite: Proc. Int. Clay Conf., Stockholm, 1, De Mumbrum, L. E. (1959) Exchangeable potassium levels in vermiculite and K-depleted micas, and implications relative to potassium levels in soils: Soil Sci. Soc. Am. Proc. 23, Farmer, V. C., Russell, J. D., Mchardy, W. J., Newman, A. C. D., Ahlrichs, J. L. and Rimsaite, J. Y. H. (1971). Evidence for loss of protons and octahedral iron from oxidised biotites and vermiculites: Miner. Mag. 38, Gilkes, R. J. and Hodson, F. (1971) Two mixed-layer mica-montmorillonite minerals from sedimentary rocks: Clay Miner. 9, Gilkes, R. J., Young, R. C. and Quirk, J. P. (1972) The oxidation of octahedral iron in biotite: Clays and Clay Minerals 20, Gilkes, R. J., Young, R. C. and Quirk, J. P. (1973a) Artificial weathering of oxidized biotite--i. Potassium removal by sodium chloride and sodium tetraphenylboron solutions: Soil Sci. Soc. Am. Proc. 37, Gilkes, R. J., Young, R. C. and Quirk, J. P. (1973b) Artificial weathering of oxidized biotites--ii. Rates of dissolution in 0.1, 0.01, M HCI: Soil Sci. Soc. Am. Proc. 37, MacEwan, D. M. C., Ruiz Amil, A. and Brown, G. (1961) Interstratified Clay minerals In The X-ray identification and Crystal Structures of Clay minerals. (Edited by Brown, G.) Mineralogical Society, London. Mortland, M. M. (1958) Kinetics of potassium release from biotite: Soil Sci. Soc. Am. Proc. 22, Newman, A. C. D. (1969) Cation exchange properties of micas: J. Soil Sci. 20, Newman, A. C. D. and Brown, G. (1966) Chemical changes during the alteration of micas: Clay Miner. 6, Rausell-Colom, J. A., Sweatman, T. R., Wells, C. B. and Norrish, K. (1965) Studies in the artificial weathering of mica. Experimental pedology, Proc. Univ. Nottingham llth Easter Sch. Agric. Sci., pp Rhoades, J. D. and Coleman, N. T. (1967) Interstratification in vermiculite and biotite produced by potassium sorption--i. Evaluation by simple X-ray diffraction pattern inspection: Soil Sci. Soc. Am. Proc. Am. Proc. 31, Ruiz Amil, A., Garcia, A. R. and MacEwan, D. M. C. (1966) X-ray Diffraction Curves for the Analysis of interstratified Structures. Volturna Press, Edinburgh. Scott, A. D. and Reed, M. G. (1962) Chemical Extraction of potassium from softs and micaceous minerals with

11 Potassium depleted oxybiotite 313 solutions containing sodium tetraphenylboron: II Biotite: Soil ScL Soc. Am. Proc. 26, Tomita, K. and Dozono, M. (1972) Formation of an interstratified mineral by extraction of potassium from mica with sodium tetraphenylboron: Clays and Clay Minerals. 20, Walker, G. F. (1949) The decomposition of biotite in the soil: Miner. Mag. 28, Wilson, M.J. (1970) Astudyof weathering in a soil derived from a biotite-hornblende rock--i. Weathering of biotite: Clay Miner 8, 291~303. R6sumr-L'altrration artificielle par des solutions Na C1 et Na B Ph4 de biotites de teneurs varires en fer ferrique de constitution, produit des minrraux et des structures semblabtes h ceux que/'on drcrit pour les biotites altrrres naturellement. L'oxydation du fer de constitution entraiiae l'extraction de K ~ partir de feuillets alternrs et le drveloppement dune hydrobiotite. Le drveloppement de l'ordre avec les teneurs en fer ferrique croissantes a fit6 6tabli par comparaison avec les calculs tbroriques concernant des structures interstratifires au hasard ou d'une manibre plus ordonn6e. On apporte une preuve de l'existence de deux types de feuillets existant dans la biotite avant son oxydation. La diminution de la vitesse de librration de K due h l'oxydation a 6t6 confirmre. Kurzreferat-Die k~instliche Verw/tterung von Biotiten, die uaterschiedliche Gehatte an Gitter-Fe ~+ aufweisen, mit NaCI- und Natriumtetraphenylborat-L6sung ergibt/ihnliche Minerale und Strukturen, wie sic ffir natiirlich verwitterte Biotite beschrieben worden sind. Die Oxidation von Gittereisen ffihrt zur K-Freisetzung aus alternierenden Schichten und zur Bildung yon Hydrobiotit. Die Zunahme der Schichtordnung mit steigendem Fe~+-Gehalt wurde durch einen Vergleich mit theoretischen Berechnungen fiir zuf~illige und vollst~indig geordnete Wechsellagerungsstrukturen abgesch~itzt. Es besteben Hinweise auf das Vorliegen von zwei Schichttypen in Biotit vor Eintreten der Oxidation. Die Herabsetzung der Freisettungsrate yon K als Folge der Oxidation wurde best/itigt. Pe310Me -- ]IpH HCCKyCTBeHHOM BblBeTpI4BaHHH 6/4OTHTOB, co~epxamr~x pa3~uqhble ypoahl'l ctpyk- TypHoro xesle3a, nocpeactaom pactaopa NaCI ~i NaBPb4, noy~yqrl~t MnHepas;br ctpyktyp CXO~HblX CO ctpyktypamn ectecraeuuo abmetpehh~,~x 6HOTHTOB. OKHczeHr~e crpyktyphoro xe~e3a aeaet K or~enaermk~ K n3 nepemexaromhxc~ caoea n K 06pa30BaHHrO rn~porhotr~ta. Ko~H'~ecraeHno otteaaaaac~ uop~)ior pocta npn nob~marotttemca co~ep>kar~n ~e~e3a cpabhehhem C TeopeTr~tleCKIIM ahanlt30m npoh3bo.ffbhofo 14 camoro ynopfi.roqehhofo r~epeme~a}otueroc~ nau~actobann~. CyRtecrny~oT 3KcrleprlMeHTahbnble ~aunb~e, no~aep~baoa~ne cy~ectaoaauue ~Byx znuob caoeb 6nOTaTe ~10 okncaennz. B~,~o uo~at~ep~r~eno nonn~enue cter~eun c~opocta ~b~eaenua K 6~aro~apa OKI,ICaeHrllO.