Hawkins, A. B., Lawrence, M. S. & Privett, K. D. (1988). zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK Giotechnique 38, No. 4, 517-532 Implications of weathering on the engineering Fuller’s Earth formation A. B. HAWKINS,* M. S. LAWRENCE? properties and K. D. PRIVETT of the $ La formation de terre a foulon est un limon The Fuller’s Earth formation is an overconcompact calcaire surconsolide jointe et fissure au solidated, jointed, randomly fissured, calcareous hasard qui contient des bandes de chaux argillacee mudstone with some thin to medium, jointed argiljoint& d’epaisseur faible ou modtree. Au sud de laceous limestone hands. South of Bath, the formaBath la formation comprend la couche de terre a tion contains the commercial (montmorillonite foulon commerciale qui est riche en montmorillonrich) Fuller’s Earth bed. This Paper reviews the ite. L’article examine I’influence de la proportion effect of the proportion of calcite present on the de calcite prbente sur la teneur en eau, les limites moisture content, Atterberg limits, particle size, d’Atterberg, la grandeur des particules et la rbisand residual shear strength. Attention is drawn to tance au cisaillement rbiduelle. On attire the importance of appreciating that the calcite perL’attention sur la nbessite de comprendre qu’il centage and the clay mineralogy are both likely to est probable que le pourcentage de calcite et la change with time as a result of weathering promineralogie argilleuse changeront au tours du cesses. This will affect the stability of natural temps ce qui entraine un processus d’effritement slopes and man-made cuttings, excavations, etc. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH cela influencera la stabilite des pentes naturelles, KEYWORDS: calcite; clays; index properties; montmodes deblais, des excavations, etc. rillonite; mudstones; weathering in any mudstone sequence varies both as a result of the natural lithological controls (the accumulated sedimentary material and diagenetic cements) and as a result of the inherent solubility of calcite. With time, ground water level fluctuations in natural slopes and alterations of the ground water regime as a result of engineering cuttings create significant changes in the proportion and nature of the calcite. This has consequential effects on the engineering properties of the strata. This Paper reviews the effect of the proportion of calcite present on the moisture content, Atterberg limits, particle size and residual shear strength in the Fuller’s Earth formation of the Bath area of the south Cotswolds (see Fig. 1.). Emphasis is placed on the importance of determining the percentage of calcite present and appreciating the likelihood that this may alter with time. The stratigraphy of the Fuller’s Earth formation is described and the difficulty of distinguishing calcareous mudstones from argillaceous limestones, particularly in moist samples, is emphasized. The mineralogical variations of the sub-divisions are considered. Attention is drawn to the change in clay mineralogy and effects of decalcification which occur as a result of weathering, and how these will affect the commonly measured engineering parameters of a calcareous mudstone. It is now well established practice to correlate various engineering parameters with intrinsic soil or index classification properties, primarily particle size and Atterberg limits. The Casagrande plasticity index provides a valuable indicator of the likely engineering behaviour for most clays and silts, mainly because the index properties are related to the soil mineralogy. Analysis of the correlations, including the bulk material properties such as particle mineralogy and bulk soil chemistry, has highlighted those geological factors which exert the greatest influence not only on the index test results, but also on the measured engineering properties of the overconsolidated, fissured, calcareous mudstones which form the dominant lithology of the Fuller’s Earth formation in the south Cotswolds. Research has confirmed a number of mineralogical changes that occur within the weathering profile. In addition, it has been shown that the variable proportion of calcite present has a profound effect on both the index and other geotechnical properties. The calcium carbonate content Discussion on this Paper closes on 1 April further details, see p. ii. * University of Bristol. t Johnson Poole & Bloomer. 1 Sir Robert McAlpine & Sons Ltd. 1989. For 517 518 HAWKINS, LAWRENCE STRATIGRAPHY The Fuller’s Earth formation of Middle Jurassic age consists mainly of calcareous mudstones/ clays, with the oolitic limestones of the Inferior Oolite below and those of the Great Oolite above (see Fig. 2). The full stratigraphic sequence in the Bath area according to the work of Kellaway & Welch (1948) and Arkell & Donovan (1952) is Upper Fuller’s Earth Claystone with argillaceous, limestone bands (c. 25 m) shelly and oolitic Fuller’s Earth bed Claystone rich in montmorillonite 7 m from top of sequence (c. 2-3 m) occurs about Fuller’s Earth rock Rubbly, argillaceous, stone shelly and (c. 4 m) oolitic lime- Lower Fuller’s Earth Silty claystones with limestone bands argillaceous (c. 10 m) and oolitic Various authors have shown that this general succession is applicable over an area extending from the Mendips to just north of Bath, although AND PRIVE’IT zyxwvutsrqponmlkjihgfedcbaZYXWV the Fuller’s Earth bed is of only limited area1 extent. North of Bath the formation becomes increasingly calcareous and several of the argillaceous limestone bands seen in the Bath area coalesce to develop thick, individually named units. SAMPLING Samples have been collected from boreholes, mine workings and trial pits at the locations shown in Fig. 1. In view of the extensive landslipping to the west of Soper’s Wood, in an area where the Fuller’s Earth bed was not known, a specially drilled, cored, borehole was sunk at ST 746677 in order to obtain samples through the complete Fuller’s Earth formation. Although the succession includes argillaceous limestone bands, (Fig. 2) there was nothing resembling the Fuller’s Earth bed which is present in the Combe Hay area south of Bath. Visual examination of the cores indicated the Upper Fuller’s Earth clays to be weathered and significantly discoloured to a depth of 4-5 m. Below this they were grey in colour and effectively fresh in weathering grade. zyxwvuts Fig. 1. Outcrops of the Fuller’s Earth in the Bath region and sample locations: DFB = Down Farm borebole; HB = Horsehouse Brake boreholes; SWB = Soper’s Wood borehole; WB = Wellsway borehole; CHM = Combe Hay mine WEATHERING OF FULLER’S EARTH FORMATION 519 zyxwvutsrqponmlkjihgfedcbaZYXWV MINERALOGY Samples of the Fuller’s Earth formation from the Soper’s Wood and Wellsway borehole cores were taken at 1 m intervals, and those from the face of the mine within the Fuller’s Earth bed at 200 mm intervals. Ten samples were also taken from the Down Farm borehole, at levels permitted by the UlOO sampling. Additional samples at selected depths were obtained from a series of (Approximate stratlgraphlc level of the Fuller’s Earth bed south of Bath) trial pits to the south of Soper’s Wood. The mineralogy of the Fuller’s Earth formation was described recently by the Authors (Hawkins, Upper Fuller’s Earth clay - blue grey. Lawrence & Privett, 1986) and some general j thinly laminated, calcareous. comments were made by Robertson (1986). 1overconsolidated, jomed, randomly fwured silty claystone wth some Analysis by X-ray diffractometry (XRD) shows jonted. very thin to medium beds of $ clearly that, throughout the succession, illite is the argillaceous ilmestone 2 dominant mineral present, with kaolinite as a 2 common subordinate constituent. Generally the + glycolated XRD samples show only a very weak I2 25 v1 to weak expanded montmorillonite peak, except in the Soper’s Wood borehole where, within the top 4-5 m of the Upper Fuller’s Earth, reasonably strong diffraction peaks are evident. It is j Fuller’s Earth rock - thinly to medium5bedded shelly and oolltlc llmestone likely that this increased montmorillonite content is related to weathering rather than initial lithoLower Fuller’s Earth clay - blue grey. logical variation in the clay mineralogy. The calcareous. lolnted, randomly fissured, weathering process envisaged is that postulated ~overconsolldated. silty claystone wth i some poorly developed thin beddlng by MacEwan (1949) shown in Fig. 3. In this, illite and occasional medfum, Jointed. is transformed to a mixed layer, interstratified, aroillaceous limestone bands th;oughout illite-smectite (montmorillonite) assemblage by the release of potassium and the subsequent i lnferlor Oollte - thinly to mediumhydration and adsorption of calcium or sodium bedded shelly and oolltic lImaStOne zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFE into the preferentially weathered planes of the interlayer spacings (see Jackson, Hseung, Corey, Fig. 2. Litho/stratigraphic log of the Soper’s Wood boreEvans & Vanden Heuvel, 1952). hole Mineralogical and chemical determinations on samples from the Down Farm borehole and Soper’s Wood trial pits suggest a change in the clay-mineral assemblage in the more weathered Additional fresh material from the area south near surface zones, possibly exacerbated by postof Bath was sampled from the cores of a site glacial acid weathering transforming some of the investigation borehole undertaken immediately illite to a mixed layer clay in the near surface west of the Wellsway Road (ST 742625). This zones. The expandable smectite component of borehole penetrated the Upper Fuller’s Earth but this mixed layer mineral has been measured at again did not encounter the commercial Fuller’s 23-39% (Lawrence, 1985; Hawkins et al., 1986). Earth bed. Samples of this bed were obtained At none of the localities reported, however, has from the face of the working Fuller’s Earth mine the montmorillonite content approached that at Combe Hay (ST 730612) before it ceased active observed in the unweathered Fuller’s Earth bed. production in 1980. At Combe Hay, the Fuller’s Earth bed is about Samples of weathered Fuller’s Earth clay were 2.4 m thick and contains high proportions of the obtained from hillslope sites within the Swainssmectite mineral Ca-montmorillonite (Jeans, Merwick Valley. At Down Farm (ST 755695) a boreriman & Mitchell, 1977). In the Top Seam (Fig. 4) hole located on the line of the A46 realignment large calcareous concretions and a significant (Privett, 1980) was drilled to obtain continuous fossil content contribute to a calcite percentage UlOO samples. Across the valley, adjacent to between 33-52%. This reduces the proportion of Soper’s Wood, slips reactivated between 1980 and Ca-montmorillonite present, and thus the com1984 involving the weathered Upper Fuller’s mercial value of this part of the bed. In the Whole Earth have been examined by Lawrence (1985), Seam, however, the calcite content has been meawho tested samples taken from shallow trial excasured as less than 15% for seven of the nine vations. Depth: m Tonso~l and hlllwash Great Oollte limestone - thinly to medium-bedded coarse. shelly and OOII~IC limestone 520 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA AND PRIVETT HAWKINS, LAWRENCE zyxwvutsrqponmlkjihgfedcbaZYXWVUTS Loss of poiasslum with subsequent hydration and adsorption of Ca’and/or Na’ Into Interlayer spaces along preferential weathering planes Unaltered crystallite core Frayed crystallite edges with mcreaslng hydration I K’ llllte (dioctahedral) Montmortilonlte (dloctahedral) Fig. 3. Representation of the weathering of illite to montmorillonite samples. The high clay-mineral assemblage of the whole seam is dominated by interstratified illitesmectite containing between 92% and 100% smectite. According to Penn, Merriman & Wyatt (1979) this is equivalent to a bulk Camontmorillonite content of 60-80%. Visual field observations and X-ray diffractogram analysis also confirmed the presence of gypsum within the weathered and landslipped horizons at the Soper’s Wood site. The trial pits indicated gypsum growths notably between 1.0 and 2.5 m below ground level within the weathering zones III and IV, a depth similar to that described by Chandler (1972) for the Lias clay. After about ten years of storage, during which the fresh Fuller’s Earth clays were allowed to oxidize and desiccate, the Soper’s Wood cores showed a prolific growth of gypsum on the exposed core surfaces. Such gypsum growth was also evident on Westbury Beds bag samples after only a few months (see Hawkins & Pinches, 1986; 1987a; 1987b). Mineralogical analysis has indicated a trace presence of quartz, feldspar (of variable composition) and pyrite. For a more detailed, comprehensive, resistate mineral analysis for the Fuller’s Earth, reference should be made to Hallam & Sellwood (1968) and Jeans et al. (1977). GEOTECHNICAL PROPERTIES To date, the only published information on the index properties of the Fuller’s Earth formation in the Bath area is that presented by Chandler, Kellaway, Skempton & Wyatt (1976). They indicated a range of liquid limit values for samples from the Swainswick Valley, varying from 35 to 55% (mean 44%) for the clays of the Lower Fuller’s Earth and from 40 to 65% (mean 51%) for the Upper Fuller’s Earth. The average plastic limit for the material in the first 10 m below the surface is reported as between 17% for the Lower and 20% for the Upper Fuller’s Earth. With mean plasticity indices of 27% and 31%, the results suggest there is no significant difference between the classification test results for the Lower and Upper Fuller’s Earth at this locality. Moisture content and Atterberg limit profiles Moisture contents and Atterberg limits were obtained on samples at 1 m intervals on the core of the Soper’s Wood borehole (Fig. 5); on WEATHERING OF FULLER’S EARTH FORMATION 521 zyxwvutsrqponmlkjihgfedcbaZYX Pale grey fme gramed argillaceous llmestone Sample levels CH 1 0.2 CH 2 0.4 CH 3 0.6 0.8 CH 5 1.0 CH 6 1.2 - - - 5 $ Grey, shelly, highly calcareous silty claystone with occaslonal, boulder sized. grey calcareous nodules. ------- ---- -------- E t-- Ad0 i CH 4 Roof of adlt I Grey, slightly calcareous claystone zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA with vague lamlnatlons and bloturbations. B OccasIonal grey calcareous nodules and patches of coarse. shelly material. I w” Cl 1 CH 7 0 1.4 Pewstent, very thm band of gravel to cobble sized grey calcareous nodules 000000 1 CH 8 1.6 CH 9 1 .E 2.0 CH 11 2.2 ; zyxwvutsrqponmlkjihgfedcbaZ 3 0 Grey, slightly calcareous claystone with occasional grey calcareous nodules. 1 CH 10 s cz 0 -- 2.4 1 Floor of adtt Fig. 4. Lithological sequence and sampling levels in the Fuller’s Earth bed at Combe Hay mine, Bath (modified from Jeans et al., 1977) 0 I 10 ’ 20 I Molsture content:% 30 40 50 I I I % calcite 60 I 70 I 5- E g 20$ 25 - 3 Fig. 5. Natural moisture content, Atterberg limits and calcite variation through the Fuller’s Earth formation at Soper’s Wood, Bath (see Fig. 2 for terminology) 522 HAWKINS. 7 2: 4: LAWRENCE AND Moisture content: % 6; 8: PRIVET? % calclte zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR I:0 $0 3: 5: 7: CH1 CH2 CH3 - D B = T CH4 - ; CH5 - 5 g CH6 - -F CH7 - e : CH8 - LL CH9 CHlOCH1 1 - Fig. 6. Natural moisture content, Atterberg limit and calcite variation for the Fuller’s Earth bed from the Combe Hay mine, Bath material from the Fuller’s Earth mine at 200 mm intervals (Fig. 6) and from UlOO samples from the Down Farm borehole (Fig. 7). Although Chandler et al. (1976) suggest ‘the natural moisture contents are, as usual in weathered overconsolidated clays, approximately equal to the plastic limit’, the evidence indicates they are generally below the plastic limit in the less weathered/unweathered material but above it in the more weathered samples. In the unweathered Fuller’s Earth bed, the moisture content profile is parallel to that of the plastic limit (Fig. 6) but consistently 5-10% lower. The data from the typical succession at Soper’s Wood (Fig. 5) show that, in unweathered Fuller’s Earth, the moisture content approaches the plastic limit upwards through the sequence. The two profiles cross near the top of the borehole at 01 E E 0 23- 4- 5 5- Molsture content: % 4: 6; 2: the same level where weathering has penetrated sufficiently to increase the montmorillonite content. The effect of weathering on the moisture content-plastic limit change-over is also evident for the samples from Down Farm (Fig. 7) which afford a more detailed profile of the weathered portion of the sequence. It can be seen that, in the upper part of zone I and in the more weathered material above, the moisture content increases progressively, relative to the plastic limit. A similar trend is indicated by the test data from the Horsehouse Brake Cutting on the A46 realignment (MRM Partnership, 1985) although in this case the colluvial nature of the near surface soils produced values in the upper 2 m which are not indicative of in situ Fuller’s Earth clay. Table 1 summarizes data from the Soper’s Wood area and indicates a clear, progressive % calclte 6; l:O 1: 3: 5: 7c1 v-5-J WP W WL Fig. 7. Natural moisture content, Atterberg limit and calcite variation for the Fuller’s Earth from Down Farm borehole, Bath: I-IV represent weathering zones WEATHERING OF FULLER’S Table 1. Moisture and index values from trial pit samples at Soper’s Wood (SWTP) for weathering zones II and IV of the Upper Fuller’s Earth clays and for zone I from borebole samples (SWBH). EARTH FORMATION 523 zyxwvutsrqp values for the zone IV material. It is suggested that the different values reflect the change in clay mineralogy and calcite content as weathering progresses in the illite-rich calcareous mudrock. The index limits of the unweathered Fuller’s Earth bed are similar to those of the zone IV Fuller’s Earth clay, suggesting that the high limits in the former reflect the greater proportion of Camontmorillonite in the clay mineralogy of the Fuller’s Earth bed. increase in moisture content and Atterberg limits WEATHERING OF THE FULLER’S EARTH in the more weathered material. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDC The boreholes show a weathering profile in the Fuller’s Earth clays, typically within 4-5 m of the ground surface (note about 6 m depth in the Index properties and calcite content Soper’s Wood borehole, below the 2.4 m of Great X-ray diffraction demonstrated the presence Oolite.) With the data from a large number of of calcite, the percentage of which was determined trial pits on the Soper’s Wood site and informaby acid digestion. The results are shown in Figs tion from the Down Farm borehole, a visual 5-7. In Fig. 5 the steady, upward increase of the weathering classification scheme has been estabmoisture content and plastic limit and the more lished (see Table 2, after Lawrence, 1985). The sensitive increase in liquid limit, is accompanied progressive alteration in material fabric and the by an overall decreasing trend in the percentage state of oxidation forms the basis of the classiof calcite. A similar trend is observed when comfication. The zonal scheme has been adapted from paring the moisture content and Atterberg limits that presented by Chandler (1969 and 1972); with the calcite percentages from the more EGGSWP (1972); and BSI (1981). weathered samples of the Down Farm borehole The main trends in index properties which can (Fig. 7). be associated with weathering of the Fuller’s There is an observable general trend in the Earth clay have been represented in a series of calcite content, although in detail the profiles are bar graphs (Fig. 8). The figure includes all the highly irregular, reflecting the lithological variaavailable moisture contents and Atterberg limits tions within the Fuller’s Earth clays. The Atterbdata plotted for each of the weathering zones, and erg limits also have very irregular profiles, yet a demonstrates clear trends of increasing moisture comparison of the liquid limit and calcite percentcontent and Atterberg limits with weathering. As age indicates a strong inverse correlation which would be anticipated, the most obvious trend is confirms both the general trend and the variation displayed by the variation in liquid limit; the two due to lithology. A similar inverse relationship is zone V/IV results came from disturbed material. indicated for the Fuller’s Earth bed (Fig. 6) The results from the Fuller’s Earth bed are although the trend is in the opposite direction. In included below the weathering zones for comparithis case the values are influenced by the high son. calcite percentages associated with the impure The bar graph plot for measured calcite nature of the Top Seam (Fig. 4) concurrent with a content (Fig. 9) indicates a general decreasing clear lithological gradation into the overlying trend with weathering grade. The very wide ‘roof-bed’ limestones. spread of plots, especially for the zone I material, Comparison of the weathered Fuller’s Earth is clearly related to lithological variations; it sequence at Down Farm (Fig. 7) and the being very difficult to separate visually the highly unweathered material from the Soper’s Wood calcareous mudstones from the argillaceous limeborehole (Fig. 5) indicates the liquid limit values stones. This decrease in calcite with weathering is of the former (6&100%) to be considerably more obvious than is indicated by the limited higher than those of the latter (21-67%). In addidata reported by Chandler & Apted (1988). tion, trial pit samples from the Soper’s Wood site gave higher liquid and plastic limits for the zone II and IV material, compared with results from CALCITE CONTENT AND ITS CONTROL ON the fresh borehole samples at the same altitude, GEOTECHNICAL PROPERTIES and hence probably the same stratigraphic level In order to examine the inverse relationship (Table 1). It is of note that although there is little between the calcite content and the index properdifference in the upper values in the ranges ties, the information has been plotted in graphical between the zone I and II material, there is quite form in Figs l&14. In these figures, the test data a marked increase between these and the upper 524 HAWKINS, LAWRENCE AND PRIVETT zyxwvutsrqponmlkjihgfedcbaZYXW Table 2. Weathering scheme for the Upper Fuller’s Earth clay based on the scheme for Soper’s Wood (Lawrence, 1985) Residual soil (hillwash/solifluction) VI Soft to firm In the superficial material, claystone is weathered to a soil in which the original rock fabric is completely destroyed. Mottled yellow-brown oxidation above the ground water level is extensive to complete. Some gleying occurs at shallow depths. Frequently contains many gravel and cobble fragments of weathered Great Oolite limestone derived from higher slopes. Seasonal desiccation features are common. Completely weathered V Soft to firm Claystone is discoloured and oxidized and completely weathered to a yellow-brown clay soil. Occasional small corestones may preserve relic laminations. Frequent gleyed fissures and root channels. Calcite content variable. Highly weathered IV Firm to stiff Fissured clay with extensive oxidation discoloration. Fissures are open and may be gleyed on surfaces. Oxidation alteration penetrates deeply from fissures. Lithorelicts (1040%) indicate some remaining structure. Calcite content variable. Some gypsum crystals on discontinuity surfaces. Many minor striated shear surfaces. Moderately weathered Firm to stiff Claystone is partially oxidized and altered to clay. Open fissures and discontinuities are discoloured orange-brown on surfaces. Alteration penetrates into soil mass, becoming yellow and greybrown along 15 mm wide zones parallel to fissures. Lithorelicts comprise 4G70% of bulk soil. Original soil structure very evident. Calcite content variable. Many large gypsum crystals along fissure and bedding surfaces. Minor shear surfaces present with small displacements, resulting from overburden stress release and swelling. Slightly weathered Stiff to very stiff Slightly discoloured claystone, confined to narrow oxidation zones adjacent to discontinuities which may be open and dark brown. Intact unweathered lithorelicts predominate (7&95%) and original structure is largely unaltered. Fissures are very closely to closely spaced. Calcite content variable. Some large selenite crystals, often along major discontinuities. Fresh Weak to very weak Parent blue-grey silty claystone showing no oxidation discoloration or any other weathering effects apart from closed stress release fissuring. Occasional open and extensive joints are iron stained. Occasional iron-rich fossil nodules with local brown discoloration zones. Calcite content variable. - are presented in full to illustrate the range and variation of typical results from the Fuller’s Earth. Such plots are clearer than data in table form, and allow the information to be more easily assimilated, particularly the relationship between the three groups of samples representing weathered Fuller’s Earth clay, unweathered Fuller’s Earth clay and the unweathered, but montmorillonite rich, Fuller’s Earth bed. With some parameters, e.g. liquid limit, the effect of weathering is similar to that produced by an increase in the authigenic expanding lattice clay mineral content. Other parameters, such as the liquidity index and residual shear strength, show a dissimilar behaviour. These relationships will be examined in more detail in the following sections. WEATHERING OF FULLER’S Zones EARTH 525 zyxwvutsr FORMATION Llquld limit: % I I V/IV IV III II I IIIIIII II FEB I 20 zones 3‘0 40 Moisture I 5; content: $0 I II 70 80 90 zones % I III1 I 100 Plastic hmlt: % V/IV V/IV IV IV Ill Ill II II I , FEB 0 g, , ,, ,I,, lb 2b I IE jo 40 50 FEB ill lb One determination io 30 J zyxwvutsrqponmlkji 40 Fig. 8. Variation of natural moisture content and Atterberg limit with weathering grade for the Fuller’s Earth. Fresh Fuller’s Earth bed (FEB) samples are included for comparison zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Index properties and calcite content below 30%, with some of the more calcite-rich argillaceous samples (technically weathered limestones) having low moisture contents. Although the moisture contents separate into two A distinct relationship between the calcite content, the moisture content and the Atterberg limits can be seen. In Fig. 10 the less weathered material is shown to possess moisture contents % calclte IV Ill II I I FEB b lb I I II ;o 30 4b kio $0 70 80 9b Fig. 9. Variation of calcite with weathering grade for the Fuller’s Earth clay; fresh Fuller’s Earth bed (FEB) samples are included for comparison HAWKINS, LAWRENCE AND PRIVETT 526 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 50Fuller’s Earth bed Weathering zones I-II Weathering zones Ill-V 40 - I I 30 60 I 90 zyxwvutsrqponmlkji % calcite Fig. 10. Relationship between calcite and moisture content for the two groups of weathering zones and the Fuller’s Earth bed distinct groupings with only limited overlap, the data indicate no pronounced correlation. The clearly aligned data points for the Fuller’s Earth bed demonstrate a very strong negative correlation of moisture content with calcite content. These results from unweathered material obtained from a limited stratigraphic horizon show less scatter than the others. The negative correlation is slightly more evident when comparing the calcite content to the plastic limit (Fig. 11). The Fuller’s Earth bed is characterized by generally high plastic limits (nearly 40%) where the calcite contents are low, and again a near straight-line relationship is exhibited. There is considerable overlap between the results from the weathered and unweathered samples. The broader zone indicates a lower degree of correlation than for the Fuller’s Earth bed. The weathered samples generally plot slightly below those of the Fuller’s Earth bed, in contrast to the moisture content results seen in Fig. 10. The relationship flattens at low plasticities, such that samples with plastic limits of between 15-20% have calcite contents varying from 40% to nearly 90% (i.e. technically argillaceous limestones). The results for the weathered material and the Fuller’s Earth bed lie in the same zone when the liquid limit is plotted against calcite content (Fig. 12). The scatter is greater than for the plastic limit, and samples with a difference in liquid limits of only 5% can have a calcite content varying by up to 50% ; note the reduced vertical scale of Fig. 12 compared with Figs 10 and 11. A clearer distinction between weathered and unweathered Fuller’s Earth is evident when the plasticity and liquidity indices are plotted against calcite content, than when the actual individual Atterberg limits are plotted (compare Figs 13 and 14 with Figs l&12. There is a clear division between the weathered and unweathered material at a plasticity index of WEATHERING OF FULLER’S V Fuller’s Earth bed 0 Weathering zones I-II l Weathering zones Ill-V 10 0 EARTH 527 FORMATION I 60 I 30 % calcite I 90 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON Fig. 11. Relationship between calcite and plastic limit for the two groups of weathering zones and the Fuller’s Earth bed V Fuller’s Earth bed 0 Weathermg zones I-II 0 Weathering zones Ill-V -. - I OO I I 30 60 --_ .,““& ./ J 90 % calcite Fig. 12. Relationship between calcite and liquid limit for the two groups of weatbering zones and the Fuller’s Earth bed 528 HAWKINS, LAWRENCE AND PRIVET-T l30- v Fuller‘s 0 Weathermg Earth zones bed I-II l Weathermg zones Ill-V Fig. 13. Relationship between calcite and plasticity index for the two groups of weathering zones and the Fuller’s Earth bed approximately 42%, and there is more scatter in the data from the weathered than from the unweathered. As with the plastic limit and liquid limit, the Fuller’s Earth bed results fall in the area of the weathered material. The plots of calcite content against liquidity index (Fig. 14) indicate a marked separation between the two weathering groups. The weathered Fuller’s Earth clay (zones III-V) is characterized by positive liquidity indices and the unweathered clays by negative liquidities, which become even more negative with higher calcite content. The Fuller’s Earth bed samples fall in a narrow range between -0.1 and -0.25, i.e. within that of the less weathered clays. Around the zero liquidity index (f0.25) the plot flattens out, representing calcite contents of approximately 50% and below. Summarizing the relationships between the various plasticity parameters and their negative correlations with calcite content for the weathered and unweathered Fuller’s Earth clay montmorillonite-rich and the unweathered, Fuller’s Earth bed, the following points emerge (4 weathered Fuller’s Earth is characterized by positive I,, I, greater than 40% and w generally above 30% @I unweathered Fuller’s Earth is characterized by negative I,, I, less than 42% and w below 30% (4 high calcite contents produce low I, and large negative values of I, because of low w%, wL and wP (4 the Fuller’s Earth bed has slightly higher wP, similar wL , similar I,, but lower w, than the weathered Fuller’s Earth. The lower I, is similar to the low calcite unweathered Fuller’s Earth clay. Particle size and calcite content Comparison of the particle size distribution data with the calcite content reveals two trends, one for the fraction finer than 2 nm (clay grade) and the other for the fraction coarser than 20 urn (coarse silt and above). In the case of the Fuller’s Earth bed, the calcite content correlates well with both particle size fractions, but the relationship WEATHERING OF FULLER’S 529 EARTH FORMATION v Fuller’s 0 Weathermg Earth zones l-11 . Weathermg zones III-V \ bed ‘\ q\ \ \ zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP 0 \ \ 4 \ ‘\O\ I \ \ I I 30 \ I I I bl 60 , 90 % calate Fig. 14. Relationship between calcite and liquidity weathering zones and the Fuller’s Earth bed index for the two groups of shear strength is an important parameter in trends are opposite; being negative for the finer engineering designs in this area. The residual than 2 urn but positive for the coarser than shear strength parameter &_ was measured for a 20 urn material, demonstrating that calcite is number of effective normal stress values using the present in the coarser fractions. This is evident for Bromhead ring shear apparatus. The values all types of Fuller’s Earth, although the obtained for a 25 kPa normal effective load, weathered material shows a greater scatter. The equivalent to the shallow depth slide, have been scatter represents, to some degree, the difficulty of plotted against the calcite content in Fig. 15. disaggregating calcareous clays/mudstones for Although the data shows two clear zones for particle size determination. the weathering groups of the Fuller’s Earth clays, It is observed that when sample CH1 (see Fig. there is no pronounced linear trend for either of 4) was decalcified under laboratory conditions, these. It can be seen that, at an effective normal there was an increase of almost 40% in the relastress of 25 kPa, the more weathered material has tive proportion of the clay fraction. The effect of a & of 11-16” ; although this is only 5-8” for decalcification on the coarser fraction is repplots at 400 kPa. A similar decrease is noted for resented by a marked reduction, from 33% to the samples from the Fuller’s Earth bed, the 5%. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA natural samples having a 4; of 13-27” at 25 kPa and 7-21” at 400 kPa. In both cases the totally decalcified sample gives a slightly lower figure. Residual strength and calcite content The difference between the 4; of weathered Many of the hill slopes in the south Cotswolds samples and the Fuller’s Earth bed at the same have suffered from both deep seated and near calcite content may be explained by the mode of surface slope movements and hence the residual 530 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA HA W KINS. LA W REN CE A ND PRIV ETT (rolling) mode of failure compared with the smaller effect of the partially dissolved finer grains in the weathered material. It is important v CHI natural to appreciate this, as it is contrary to what would V be expected from the presence of smectite l V (montmorillonite) in the Fuller’s Earth bed and V demonstrates the significance of the calcite. Figure 16 shows the residual strength ,A9 / envelopes for samples from the 2.4 m-thick / Fuller’s Earth bed. The figure demonstrates the / d’ curvature of the residual failure envelopes (see I Hawkins & Privett, 1985) which fall into three zonesI-11 groups, depending on calcite content. Samples with greater than 30% calcite have steeper envelopes (and hence higher 4;) than the others; CH1 is the steepest (52% calcite). The steeper envelopes are also those with the less pronounced curvature at low normal stresses, i.e. the lower the calcite content, the greater the difference in #s between high and low normal stress conditions. In order to confirm the influence of calcite on residual shear strength, sample CH1 (52% calcite) was decalcified in the laboratory by acetic acid digestion. A dramatic reduction in 4; occurred with decalcification, from 27” to 11” at 25 kPa 30 60 and from 21” to 6” at 400 kPa. % calcite zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 25 w V V FEE ‘CH1 decalcified 0 Fig. 15. Relationship between calcite and residual shear strength parameter 4, at 25 kPa for two weathering groups and the Fuller’s Earth bed CONCLUSIONS Visual observations of the weathering profile within the Fuller’s Earth clays confirm the physical changes typically associated with oxidation and softening. A general colour change from grey occurrence of the calcite; the larger grains in the fresher Fuller’s Earth bed resulting in a turbulent Residual ------- --- m 200 1 friction ratlo - normal Residual stress curves _-_________ 130 g ___________CH5_I Decalafed CH1 shear stress - normal Stress envelopes B Calcite contents c ,_ 2 ; 100 : 0) 0 CH1 1 CH6 CH9 CH6 CH10 0 100 200 300 Effective normal 400 500 600 stress, On’: kPa Fig. 16. Residual strength envelopes and residual friction ratio carves for tkc Combc Hay Fuller’s Earth bed WEATHERING OF FULLER’S to brown is commonly noted at the levels where stress release fissuring becomes prominent, with the consequent loss of inherent strength and the overprogressive degradation from an consolidated claystone towards a firm, structureless clay. X-ray diffractometry analyses on samples from the Fuller’s Earth formation indicate the mineralogical composition to be dominated by illite with subordinate kaolinite. The Fuller’s Earth bed is exceptional, however, being composed mainly of Ca-montmorillonite. Whilst the formation is calcareous throughout, it is the lithological variations due to authigenic cements, fossil content and calcite crystals that give rise to the significant differences in the proportions of calcite content. Analysis of fresh material from borehole cores and of samples taken from the trial sections through the surface weathered layers has been undertaken. This revealed a chemical weathering sequence involving the following processes. (4 Stress relief allows the development of open fissures, facilitating the percolation of surface water into an otherwise relatively impermeable mudstone/clay mass. (b) Decalcification by weak acids, including the carbonic acids in rainwater and humic acids from the vegetation layers. (4 Decomposition of pyrite by oxygenated water, with the production of an acidic ground water (sulphuric acid). (4 Subsequent sulphuric acid attack on calcite (calcium carbonate) liberating calcium which facilitates the formation of gypsum (calcium sulphate), see Hawkins & Pinches (1986, 1987a and 1987b). (4 Ferrous ions converted to the ferric state produce limonite and haematite, thereby effecting the colour change from blue-grey to yellow-brown. (f) Acidic ground water reacts with the clay minerals, depotassification and hydration transforming illite to interstratified illite-smectite (montmorillonite). Within a sequence of calcareous mudstones, calcite generally comprises the coarser, more angular particles, such as secondary concretions, fossil shell fragments, broken calcitic veins, and cementitious material. euhedral grains, Although X-ray diffractometry on the separated ‘less than 2 urn’ clay smears suggests the persistance of calcite to at least fine silt grades, this mineral (in addition to quartz, pyrite and feldspars) comprises the bulk of the coarse silt and sand fractions. Calcite is characterized by a relatively small surface area to particle volume ratio which, together with very low sorptive properties, EARTH FORMATION 531 zyxwvutsrqponmlkjihgfedcbaZ accounts for the negative correlations between calcite content and the Atterberg limits. However, the predominance of clay minerals throughout the formation also explains the positive correlations of clay grade material with the index properties. Within the Fuller’s Earth bed and the near-surface weathered horizons, the presence of expanding Ca-montmorillonite, with its superior sorptive properties and higher surface area to particle volume ratios, can be seen to correspond to the increased moisture content and plasticity values. It has been demonstrated that, within a calcareous mudstone sequence such as the Fuller’s Earth of the Cotswolds, weathering modifies the engineering behaviour in two major transformations. (4 A decrease in calcite content and the removal (4 of coarser particles leads to a relative increase in clay grade minerals. The reduction in a content leads to mineral non-sorptive increased plasticity. The transition of illite to interstratified illite(montmorillonite) increases the smectite expanding lattice clay mineral proportion of the soil and Atterberg limit values increase accordingly. Laboratory results suggest that for the Fuller’s Earth formation, calcite content is correlated to the residual shear strength parameter 4;. This is probably because calcite occurs predominantly as equigranular, often euhedral, coarse particles (along with other common primary sedimentary minerals, such as quartz, feldspar, etc.). Such grains contribute to a ‘rolling’ frictional shear resistance (Lupini, Skinner & Vaughan, 1981). If decalcification is the prevalent weathering process in a calcareous mudstone under the typical United Kingdom climate, then an effective reduction in the calcite grain content will produce a relative increase in clay mineral platelets, hence promoting the weaker ‘sliding’ frictional shear resistance. It is suggested that this process is operative within the Fuller’s Earth sampled from the Bath area. The implications of this decalcification process and the associated potential reduction in shear strength must be considered when assessing the long term stability of natural hill slopes, and in the design of civil engineering cutting and embankment slopes. Decalcification by the initial removal of the authigenic cements results in a progressive reduction in the peak strength values (which are not discussed in this Paper). For most weathered calcareous mudstones, softening will already have begun to reduce the shear strength. In slopes in which landslipping has occurred, and hence appropriate residual shear strengths are 532 HAWKINS, LAWRENCE operative, continued decalcification will accelerate this decrease. The implications of clay mineral transformation along the relatively permeable shear surfaces must also be considered, as it will have an important influence on the long term stability of such slopes. A study of selected Fuller’s Earth clay localities in the Cotswolds has demonstrated that, in a weathering profile through a calcareous mudstone sequence, decalcification and depotassification results in significant changes in fundamental engineering properties. Preliminary laboratory tests and analyses have confirmed a relationship between the calcite content and the shear strength values. Further work is required on the process of decalcification within the time-scale of particular man-made slopes and designed earth structures, as well as the long term behaviour of natural slope morphologies. The evaluation of the changes in shear strength with progressive reduction in the calcite content, and the likely long-term effect on man-made slopes, is continuing. ACKNOWLEDGEMENTS Laporte Industries are to be thanked for allowing access to the Fuller’s Earth mine. The work was partially supported by the National Engineering Research Council through studentship grants. The Authors would like to thank Marian Trott and H. Gudge for their assistance in the preparation of this Paper. REFERENCES AND PRIVETT EGGSWP (1972). The preparation of maps and plans in terms of engineering geology-Engineering Group of the Geological Society Working Party Report. Q. J. Engng Geol. 5, Part 4,293-381. Hallam, A. & Sellwood, B. W. (1968). Origin of Fuller’s Earth in the Mesozoic of southern England. Nature 220, Dee 21, 1193-l 195. Hawkins, A. B., Lawrence, M. S. & Privett, K. D. (1986). Clay mineralogy and plasticity of the Fuller’s Earth formation, Bath, UK. Clay M inerals 21, Part 3,293-310. Hawkins, A. B. & Pinches, G. M. (1986). Timing and correct chemical testing of soils/weak rock. Site Investigation Practice: _ Assessing ES 5930 (ed. Hawkins, A. B.). London: Geological Societv. Hawkins, A. B. & Pinches, G. M.jl987a). Cause and significance of heave at Llandough Hospital, Cardiff-a case history of ground floor heave due to gypsum growth. Q. J. kngn> Geol. 20, Part 1,41-57. Hawkins. A. B. & Pinches. G. M. (1987b). Sulnhate analysis on black muds&es. Gto;echni&e 37; No. 2, 191-196. Hawkins, A. B. & Privett, K. D. (1985). Measurement and use of residual shear strength of cohesive soils. Ground Engng zyxwvutsrqponmlkjihgfedcbaZYXW 18, No. 8,22-28. Jackson, M. L., Hseung, Y., Corey, R. B., Evans, E. J. & Vanden Heuvel, R. C. (1952). Weathering sequence of clay-size minerals in soils and sediments II: chemical weathering of layer silicates. Proc. Soil Sci. Sot. of America 16,34. Jeans, C. V., Merriman, R. J. & Mitchell, J. G. (1977). Origin of middle Jurassic and lower Cretaceous Fuller’s Earths in England. Clay M inerals 12, Part 1, 1144. Kellaway, G. A. & Welch, F. B. A. (1948). British regional geology : Bristol and Gloucester District. London: Institute of Geological Sciences. Lawrence, M. S. (1985). The engineering geology of selected slopes on the Jurassic strata of the south Cotswolds. PhD thesis, University of Bristol. Lupini, J. F., Skinner, A. E. & Vaughan, P. R. (1981). The drained residual strength of cohesive soils. Gko- Arkell, W. J. & Donovan, D. T. (1952). The Fuller’s technique 31, No. 2, 181- 213. Earth of the Cotswolds and its relation to the great MacEwan, D. M. C. (1949). Some notes on the recordJ. Geol. Sot., London 107, Part 3,227-253. oolite. Q. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA ing and interpretation of X-ray diagrams of soil BSI (1981). Code of practice for site investigations: BS clay s. J. Soil Sci. 1, No. 1,9&1Oj. 5930. London: British Standards Institution. MRM Partnershin (1985). A46 UoDer Swainswick to Chandler, R. J. (1969). The effect of weathering on the A420 Site i&sti&tion interprektive report for the shear strength properties of Keuper Marl. GhotechDepartments of the Environment and Transport. nique, 19, No. 3, 321-334. London: HMSO. Chandler, R. J. (1972). Lias clay: weathering processes Penn, I. E., Merriman, R. J. & Wyatt, R. J. (1979). The and their effect on shear strength. GCotechnique 22, Bathonian strata of the Bath-Frome area. Report of No. 3,403431. the Institute of Geological Sciences 78122. London: Chandler, R. J. & Apted, J. P. (1988). The effect of HMSO. weathering on the strength of London clay. Q. J. Privett, K. D. (1980). The engineering geology of slopes Engng Geol. 21, Part 2,59-68. of in the south Cotswolds. PhD thesis, University Chandler, R. J., Kellaway, G. A., Skempton, A. W. & Bristol. Wyatt, R. J. (1976). Valley slope sections in Jurassic Robertson, R. H. S. (1986). Fuller’s Earth: a history of strata near Bath, Somerset. Phil. Trans. Royal Sot. Society calcium montnwrillonite. (Mineralogical A, 283,.527-556. occasional publication). Hyde: Volturna Press.