Session 1/14 The Colloidal “Activity” of Clays L’Activité colloidale des argiles by A.W.SkEMPTON, D.Se., AM.LCE,, F.GS., University Reader in Soil Mechanics and Assistant Professor at Imperial College, University of London, England Summary In any particular clay stratum the ratio of the plasticity index to the clay fraction content is approximately constani, and may be defined as the “activity” of the clay. Values of activity are given for many clays and also for the more common minerals. Tis shown, that activity is related to the mineralogy and geological history of clays, and to the proportion oftheir shear strength contsibuted by true Cohesion. Field data is presented which indicate tat the diffe culties of taking satisfactory undisturbed samples in deep beds of sensitive clay are reitrited to those clays with an activity of less than 0.75, Introduction The properties of a clay are determined fundamentally by the physico-chemical characteristics of the various constituent minerals and by the relative proportions in which the minerals are present. ‘The determination of these characteristics is a lengthy and difficult process requiring the use of an X-ray spectrometer, thermal analysis, etc.,and itis evident that such techniques can never become part of the normal laboratory procedure in soil mechanics. Some simple tests aro therefore required that give a quantitative measure of the composite effects ofall the basic properties of a clay and, as is well known, the Atterberg limits full this funcifon in large measure. But they are not wholly sufficient, and in the present paper evidence is given which shows that valuable additional information is ied by an index property combining the Auterberg limits and the particle size distribution of a clay; yet requiring for its determination only the results of these routine test, The Ratio: Pi/Clay Fraction If a number of samples are taken from a particular clay stratum and the clay fraction content (percentage by weieht of particles finer than 2 miccons) and the Atterberg plasticity index (P1) are determined for each sample, then there is gene- ‘Sommaire Dans une couche dargile fe rapport de indie de plastics 81a proportion d'argle (moins de 2 microns) est presque constant ct ‘peut Eire défin) comune «T'actvitéy de la couche, Les valeurs de ‘cette activité sont donntes pour diverse angles et aus pour Tes minéraux les plus sépandus. I! est démontré que Fact dépend ide Ta minéralogie et de la péoiopie des aries et de leur résistance fu eisillement die & In cohéeion vmaiewLes observations sur le terrain montrent que extraction <'chantillons satsaisants, dane e profondes couches dargile sensitive, ne présentent des difcultés ‘que pour les artes dont Pacivits est aféreure & 075 rally a quite wide range in the numerical values for both pro- perlies. Yet if the plasticity index is plotied against clay frac- tion it wil be found that the points Tie about a straight line which extrapoletes back to the origin. Typical ses of results obiained from such tests on four clays are given in Fig. 1. ‘The degrce of scatter about the mean line is presumably a measure of the variations in composition within the stratum, Now it is widely recognised that the higher the plasticity index the more pronounced are the colloidal properties of a clay. Moreover the colloidal properties are contributed largely, by the finest particles and, in particular, by the “clay fraction”. But reference to Fig. 1 will at once show that two clays wl may have the same content of clzy fraction can have widely different plastic indexes, and it would seem logical to assume that the clay with the higher Pf, for a given clay fraction con- tent, is more colloidally active than the clay with a lower PJ, fog the same given clay fraction content. "The direct linear relationship between PJ and clay fraction content for any particular clay enables this degree of colloidal, activity to be expressed very simply by the rat plasticity index activity = Y= "day fraction cyRRR nw “This ratio is, in fact, the slope of the fines such as those in Tie: {and it provides a convenient single-alued paramcter for any particular clay ‘rhe above definition of activity as given by the author in 1950 and isa denelopment of an earlier conception (Skemp/on, 1948) in which liquid limit was plotted against clay fraction Fiat Relation Between Plasticity Index and Clay Fraction Ron er Tne de plist le powrcntage tale In the 1948 paper three classes of clay were revosnised, from {hs point of view, namely “inactive”, “normal” and “active” wee Sptsined subsequently has not lead to any essential erate in this classification which, in terms ofthe ratio Pielay fraction, may be stated as follows: inactive clays — activity <0.75 normal clays — activity 0.75 to 1.25 fetive clay — activity >1.25 “The relation between liquid limit and lay fraction, although tinear, is not one of direet proportion and is therefore Jess cOn- TSent than the ratio Pifclay fraction. For the idea of plotting Say fraction against plasticity index rather than Tiguid Umit Ro author is indebted to 2 graph in a paper by 4. Casagrande sand Shannon (1948). ‘Activity of Various Minerals In examining more fully the significance of activity itis fa athe fst place of interest to discover the values of PUstay frac~ tio for the commonly occurring minerals in clays. The prin- ‘Gpal data are assembled in Table 1. “Table 1 Values of PUClay Fraction for some Clay-Mineral, ‘Mineral ‘criiey steers | Quartz 00 von Moos (1938) | Caleite os. ton Boos (1939) Mica (muscovite) 023 tom Moos (1938) ean (033 TNorthey (950) 1046 Sernele (1950) ite 090 Northey (1950) Camontmosillonite | 15S Samels (1980) Nemontmorionie | 7. Semmets (1950) 38 “The three minerals quartz, calcite and mica, tested bY Wom “Moos, were ground to 2 very small pasticle size and the PZ Meee termined on the fraction finer than 2 microns. ‘The array of these minerals is ow, as might be expected from seat uatvely simple erystal structure. Of the true clay mine eee far examined Kaolinite has the Lowest activity. ite is Trapably the most widespread ofall clay minerals but usually ares ia conjunction with other minerals. Fortunately, how: Rees ay shale exists in Hinois, the clay fraction of which ever yate mast endrcl of illite, A large sample if this material sorfindly sent by Professor Grin and the average result of wee ttded out by Dr. Norther, in the author's laboratory, given in Table 1. The clay known as bentonite consists ah is erenclusively ofthe mineral montmorillonite, In is natural ‘Mate bentonite is usually a sodium clay ang, in this state, it cate terry exceptionally high activity, By effecting a base ex Tass from sodium (monavalent) to calcium (bivalend SSumucis (1980) has shown by repeated tests that the activity ie considerably lowered; although even the Ca-bentonite bas A gi netiity, He has aso shown that bentonite carrying & 2 lent base Al has an activity of about 1.3 In contrast, ‘Sevnels (1980) found that base exchange fas only a minor Setecnee on kaolinite, No base exchange tests appear to have tioer made on lite but, since this mineral shows moderate petty, the effect would probably be appreciable i clear from the above results that activity i, broadly speaking, related to the structural complenity of the minerals seoging from quartz through Kaolin up to montmorillonite “Activity and Geological History Information concerning 27 clays is given in Table 2, from ‘uthch it may be deduced that there s some degree of correla~ “Naw between aetivity and the mineralogy and geological bis- tory of a clay. = jnaetive” clays (activity ess than 0.75) seem to possess cone of more of the following characteristics: Go) cay fraction either consists predominantly of Kaolinite, ‘Contains litle true clay mineral: (deposition in fresh waters {Gy deposition in salt water, but subsequently Ieehed by pst colation of fresh water Clays combining the characteristics (@) and (8) 0 (@) and © ae he Teant active group [ (activity les than 0.5) Apart ‘Rem aol the tpicel members of this group ave Tatesiacia) tans derived largely by mechanical erosion of non-arelaceous ys Soy loeahests, and deposited in ie-demmed lakes; and (Setalcial marine or estuarine clays which have subseauensty Parr senched by fresh water, usually following isostatic voli aeerre's evidence, both from the fied and the laboratory. that seer of the extra sensitive clays belong to this category of hed post-slacial marine deposits (Rosenauist, 1946; Skerp- reer Northey, 1952). Clays formed by normal weathering or posted in fresh water seem 0 fall into the group 2 with activities between about 0.5 and 0.75. ‘tye Tergest_group is that with activities between 0.75 and t2svand it inludes the marine and estearine clays with ite athe predominant clay mineral. Only 8 examples are gives 3 Table 2, but many more clays could be included ranging in iological age from the Jurassic tothe postslacial periods Gotu 4 the members of svhichmay be described as "active lavsccunsiss of deposits Which contain an appreciable amount cree coftods, although in other respects they woulé be clases as““normnal™. Te may be expected that cays containing‘Table 2 Correlation between Activity and the Mineralogy and Geology of Some Clays a mae Group | of ‘Location Geology Clay Fraction \ activity Authority ac te | aa ‘rcve) tm [6 Tar or Gober | ot Gal maroon vy =< | m | ass | peta cin ny | — | 03» | sete cabo, UA a rte jee apes | | feo Nin | reine ess | Bens | 88 | Rieu | inact 05t0| Wren, Ws LateGica! rombrnanene |---| --- | 03 | am | Fe te naan Weneae Conese | Tk oe See ono. ‘Seagrove Bay, 1.0.W., Engl. | Oligocene, fresh-water 0.73 | Skempton | Namal] 075] Pemtoroah Ended | Overt cy tras nanre | cos [ans | sm) Rare a Poetic ares SEE | Been Nepte Sree. usa. | Gove Cavomres iy Ly E | 88s | Remrrene pees, | (Someta | 93) | Eine stein ee. jae | Soe | Seto Aociritec emacs | soto marae and auatae tee | BRS" ae We as Maat noe RRS, Weta | Re bette Bree mee 188 | eras ce | La eine yal Ree ait Spas te 133 | Rene Raewies, « |feauuemetatas | 22 2 | BRS eae ee | cena, copes 18 | BRE Seem eee ica | React cram mee ca 13s | Brat | sete | more | Meo cy enone Cay mo | ee | a3 | aaratet a | AS | Neca. ime = 3 | May 3 | Sa > Sa tees oy — ontte SBS Elm fens = So 2 Meta Ccusmontmorlonte woul also fl into this group, but the t= 6 +6) 10 4, author does not know of any data on such materials. Group 5 includes only bentonitc clays, no others are known with such high activity values; and the reason is immediately apparent from the fact that they consist predominantly of Ne-mont- morillonite (Gee Tabie 1) Boulder clays have not been given in Table 2 since they ean vary between Groups I to 4, depending upon the nature of the ground from which the glacier or ice-sheet derived the material. Thus the boulder clays of East Anglia, being derived from the Jurassie and Cretacsous clays of the southern and caster Midlands, fll into group 3. But some of the boulder clays of northern England and Scotland fall into groups 1 or 2 as they consist largely of finely ground rock minerals with litle iF any true clay minerals incorporated in the matrix. Acitivity and True Cohesion ‘The shear strength of a clay is made up of two parts, the cohesion c, and the coeficient of internal friction tan 9, according fo the expression (Hvorsler, 1937) where o;! is the effective pressure normal to the sheer plane. fa clay is normaly-consoligated from a slurry under e pres- ssure 0," and is then sheared sufficiently slowly forall the pore ‘water pressure 10 be fully dissipated (a “drained” shear test), then: ry = 9 tn ee where gis the angle of shearing resistance in the “drained” state. If, moreover, ¢, is the cohesion of the clay atthe water ‘content at failure in the drained shear test, then the proportion, of the shear strength due to cohesion is 3. tan pe ‘and the proportion due to internal ftietion is tmor_,_[_@ | van gu Loren ge In Fig, 2 the components of shear strength in 8 normally= consolidated materials are plotted agsinst their activity. Tis 8