Interaction between mineral composition, water content and mechanical properties of saturated cohesive soils Bojana Dolinar 1, Miha Mišič 2 1 Faculty of Civil Engineering, University of Maribor, Smetanova ul. 17, Maribor, Slovenia; e-mail: bojana.dolinar@uni-mb.si 2 Geological Survey of Slovenia, Dimiceva ul. 14, Ljubljana, Slovenia; e-mail: miha.misic@geo-zs.si Received: April 12, 200S Accepted: September IS, 2005 Abstract: It is known that mechanical properties of cohesive soils, which are determined within the scope of geotechnical investigations, depend on the water content, while the latter mostly on the mineral composition of soils. These relationships are given as empirical expressions that are based on the results of laboratory tests performed on artificial monomineral clay mixtures. The aim of the investigations described in the article was to verify practical applications for the above-mentioned findings. Five samples of heterogeneous soils were tested for this purpose. The values of the geomechanical properties, which were calculated from selected mineralogical properties, are compared to experimentally tested values. The testing results show good correlation. Key words: clays, mineral composition, undrained shear strength, compressibility Introduction It is known that mechanical properties of saturated cohesive soils, which are determined within the scope of geotechnical investigations, depend on the water content, while the latter mostly on the mineral composition of soils. Several researchers have tried to determine these relationships by comparing the quantity of water in soils to the sizes of grains, the specific surface, the quantity of clay fraction or the cationic exchange capacity. The results of their investigations, however, vary considerably, being valid for investigated soils only and not generally applicable. The reasons for the different conclusions reached by researchers in previous studies have been explained by Dolinar & Trauner (2003, 2004) as follows. Clay and non-clay minerals are present in cohesive soils. To understand the influence of soils' composition on the water content it is necessary to know that clay minerals, as well as water, are not chemically inert; therefore they are subject to interaction. In non-expanding clay minerals, water is strongly adsorbed on the external surfaces of the grains (w ) whilst in expanding clays it is bonded to the external surfaces (w ) and internal surfaces (w.) of the grains. In addition to strongly adsorbed water, the free pore water we, occurs in saturated clays (Fig. I). Using a standard method for measuring the water content w, the total quantity of intergrain water we (we = wef + w—) and interlayer water w. can always be determined by drying at a temperature higher than 100 °C. The total quantity of water w is, therefore, equal to the intergrain water content in non-swelling clays (Eq. (I)), whilst it is higher for the portion of interlayer water in swelling clays (Eq. (2)). w = w = w , + w (I) e ej ea v ' w= w + w = w + w + w (2) e i ej ea i v ' Interaction forces between the external surfaces of the grains and the adsorbed water are the same for all clays due to equal surface structure. It leads to the conclusion that at an equal quantity of free pore water wej, the ej thickness of the adsorbed water film t is the a same for different clay grains (Mitchell, 1993). The quantity of water, adsorbed on the external surfaces of the clay grains wea, depends, in this case, on the surface of the clay grains expressed on the unit weight of clay grains Asc (Eq. (3)). Whilst the quantity of intergrain water we depends on the sizes of the clay grains, the interlayer water quantity wi mostly depends on the types and quantities of interlayer cations (Grim, 1962). This means that it is impossible to express the total quantity of water w in expanding and non-expanding clays by considering the same mineralogical properties. The interlayer water w. is strongly adsorbed between the layers, therefore, it cannot influence the strength and compressibility of the clays. Any dependence between the water content and selected mechanical properties can be determined on the basis of the intergrain water content we only. Taking into consideration the assumptions mentioned above the intergrain water content we depends on the free pore water quantity w and on the sizes of the clay grains. It can be expected that the quantity of free pore water thickness of the firmly adsorbed water t„ Figure I. Graphical presentation of water distribution in fully saturated clays; the intergrain W = W , + W . e ef m Slika I. Grafična predstavitev razporeditve vode v zasičenih glinah; medzrnska voda w = w , + w . only depends on stress state, which means that it must be the same for different clays at equal effective stress a' and equal undrained shear strength cu and also the thickness ta of the adsorbed water film (Eq. (4)). w,, =w,,, +t i , • A-r (4) e\ct',c„ ef\a ',c„ a\a',cu AC V^V Equation (4) is valid for soils that contain clay minerals only. In cases of clay- and non-clay minerals in soils, the findings of Mitchell (1993) and Seed et al. (1964) must be considered, i.e. water in soils is mainly bonded to clay grains. This means that the intergrain water quantity also depends on the weight portionp of the clay grains in the soil (0 < p < 1). In this case, the Eq. (4) is rewritten as follows: w, , =p-(wfi , +t i , -A--) (5) e\a ,c„ r V ef\a ,cu a\a ,cu AC/ V-V where p ■ Asc = As; As (m2/g) is the grain surface per unit weight of the dry soil. In determining the relationship between the given mineralogical properties of soils, the intergrain water content w, and the undrained shear strength c , the researchers Dolinar et al. (2004) proceeded from the known experimentally defined relationship (Koumoto & Houlsby, 2001; Dolinar, 2004), where ae and be are soil-dependent parameters. ae (%) is the intergrain water content at the undrained shear strength cu = I kPa, and be is the slope of the linear function which represents the ratio between the intergrain water content we (%) logarithm and the undrained shear strength cu (kPa) logarithm. It follows from equations (5) and (6) that the intergrain water content at the undrained shear strength cu = I kPa can be expressed as We\cu=\ =£le=P' 1 + ' AS (U) Equation (7) shows that the parameter ae depends on the specific surface Asc and the weight portionp of the clay grains in the soil, because wef\c„=\ and ^a|c„=i are constants. It follows from equations (5), (6) and (7) that the parameter be depends on the specific surface Asc only. The relationships between the specific surfaces of grains and both parameters are given by Eqs. (8) and (9), where a = 33.70, p = 0.99, y = 0.05 and X = 0.27. The coefficients in Eqs. (8), (9), (II) and (12) were determined by testing samples which had clearly defined chemical and mineral composition, a known structural formula and a specific surface. ae=p-a + $-As (8) be=yAscx (9) Equations (7) and (8) have the same structure, therefore, the first and the second terms in both of them have to be equal. Consequently, the quantity of free pore water We/|c„=i =a , the average thickness of the adsorbed water around the clay grains ta\cu=\kPa = $, and the adsorbed water quantity around the clay grains Wea\cu=\kPa = ta\cu=lkPa ' ^SC = P ' 4sC • In determining the influence of mineralogical properties of soils on the intergrain water content w at the effective stresses a' in the e soils, Dolinar et al. (2004) proceeded from the dependencies: we=iea^ (10) where ie andje are soil dependent parameters. In accordance with previous findings they can be expressed in dependence on the size and quantity of clay minerals in soil by Eqs. (11) and (12), where ro = 33.46, 0 = 1.39, Y = 0.05, X = 0.27 andp is the portion of the clay minerals in the sample. ie=P(o+QASe (11) Je=be=yAscx (12) It should be noted, however, that in practical applications it is impossible to determine the structural formula of an individual clay mineral, which is needed for an accurate calculation of interlayer water. It is also difficult to determine an accurate quantitative mineral composition. The quantity of water in soils can also be influenced by organic substances, which are often found in soils, but their influence has not yet been investigated. So, the aim of the investigations described in this article, was to verify practical applications for the above-mentioned findings. Five samples of heterogeneous soils were tested for this purpose, in which the geomechanical properties were determined both experimentally and on the basis of mineralogical properties. This allowed a comparison between the measured and calculated values. Investigation of natural soils Experimental determination of parameters ae, be, ie, and je We have determined the parameters ae and be experimentally with five measurements of the undrained shear strength cu (kPa) at different intergrain water content we (%). The relationship between c and w is linear in a r u e double logarithmic scale, therefore ae (%) is the water content at the undrained shear strength cu = 1 kPa and be is the slope of the linear function, which enables easy determination of both parameters. The liquid limit LL and the plastic limit PL, which correspond to the undrained shear strength c = 2.66 kPa and c = 266 kPa, were also uu determined in the samples. The quantity of intergrain water we in the saturated clays was measured under different effective stresses using an oedometer consolidation test. The initial moisture content in the samples was near the liquid limit. The corresponding water content after consolidation of the clays was determined at axial stresses a' = 50 kPa, 100 kPa and 200 kPa. The relationship between the intergrain water quantity and the effective stress is completely linear when both variables are shown in a double logarithmic scale. In this case the parameter ie represents the water content in the soil at a' = 1 kPa and the parameter je the slope of the linear function. In determining parameters ae, be, ie and je it was necessary to consider only the intergrain water content in the soils. The interlayer water in the tested samples, however, resulted from the presence of Ca- montmorillonite. The procedure of calculation is shown below. The calculation of interlayer water in an expanding mineral An example of the interlayer water content calculation is given for the sample of pure montmorillonite with the calcium exchangeable cation. The montmorillonite structure consists of an octahedral sheet sandwiched between two silica sheets. The layers formed in this way are continuous in the a and b directions and stacked one above the other in the c direction. Bonds between layers are weak, thus water or other polar molecules can enter between the unit layers causing the lattice to expand in the c direction. The basal spacing (spacing between the centre of two neighbouring layers) in the c direction which is d = 0.96 nm for dry calcium montmorillonite (dried at 105 °C) rises to d. = 1.54 nm after the adsorbtion of water (Grim, 1962). In case of calcium exchangeable cation in montmorillonite the adsorption of water between layers is completed already at very low water content and the basal spacing is then practically constant. The volume of interlayer water V J Wl can thus be calculated using the equation (Fink and Nakayama, 1972): V . = Asr(d2-dJ (13) where As. (m2/g) is the internal specific surface of montmorillonite. With our samples we considered the internal specific surface of montmorillonite grains As. = 626.80 m2/g and adequate mass portions of this mineral in the individual soils. Determination of mineralogical properties The bulk mineral composition and the clay fraction of the samples were scanned by Table I. Mineral composition of the powder bulk samples. Tabela I. Mineralna sestava uprašenih vzorcev. Sample Mineral composition (%) and specific surface (m /g) 1 + Rxl Ch + Rx2 Ka CaM Q PI Mic Cal G/H As 1 25 8 5 14 34 9 5 0 0 30.1± 0.4 2 35 14 0 0 25 3 0 22 1 28.5± 0.4 3 28 16 0 0 42 9 4 0 0 16.7 ±0.1 4 35 0 12 34 18 0 0 0 1 54.1 ±0.3 5 34 18 0 0 43 3 3 0 0 32.6 ± 0.2 LEGEND: I illite Ch chlorite Rxl mixed layer clays illite/montmorillonite Rx2 mixed layer clays chlorite/montmorillonite Ka kaolinite CaM Ca - montmorillonite Q PI Mic Cal G/H quartz plagioclase microcline calcite goethite/hematite Table 2. Mineral composition of the fraction < 0.002 mm. Tabela 2. Mineralna sestava frakcije < 0.002 mm. Sample Illite, MLC illite / Ca-montmorillonite, Ca-montmorillonite (%) I I1/EI1 I1/EI2 I1/EI3 I1/EI4 I1/EI5 I1/ZI6 11/217 I1/ZI8 CaM 1 4 0 0 0 60/2 0 0 23/2 0 6 2 7 0 82/6 0 0 55/5 0 0 0 2 3 2 0 0 78/1 0 0 42/1 0 0 1 4 6 0 0 78/4 0 0 42/3 0 0 16 5 7 84/5 0 78/4 0 0 42/3 0 15/2 0 LEGEND: I illite CaM Ca-montmorillonite II illite in mixed layer clays illite/montmorillonite 211 -18 percentage of mixed layer clays illite/montmorillonite in soil Table 2 (continued). Mineral composition of the fraction < 0.002 mm. Tabela 2 (nadaljevanje). Mineralna sestava frakcije < 0.002 mm. Sample Kaolinite, Chlorite, MLC kaolinite/Ca-montmorillonite, MLC chlorite/Ca-montmorillonie [%] Ka Ch K/SKao C/EChl C/£Ch2 C/£Ch3 < 0.002 mm 1 1 0 0 0 0 0 14.54 2 3 2 0 0 68/1 0 26.21 3 0 3 0 76/2 0 63/2 12.84 4 2 0 90/2 0 0 0 33.02 5 4 0 0 0 0 0 24.96 LEGEND: Ka kaolinite Ch chlorite K kaolinite in mixed layer clays kaolinite/montmorillonite C chlorite in mixed layer clays chlorite/montmorillonite EKao percentage of kaolinite in mixed layer clays kaolinite/montmorillonite in fraction < 0.002 mm EChl -Ch3 percentage of mixed layer clays chlorite/montmorillonite in soil Table 3. Chemical composition. Tabela 3. Kemična sestava. Chemical composition (%) 1 2 Sample 3 4 5 Si02 64.15 46.83 72.10 51.71 67.48 ai2o3 15.83 13.45 12.09 20.38 12.77 Ti02 0.89 0.61 0.78 0.80 0.88 Fe203 5.09 5.02 4.92 9.06 6.70 FeO 1.1 1.0 1.0 0.3 0.5 MnO 0.03 0.07 0.05 0.18 0.18 MgO 1.97 2.27 1.07 1.66 1.62 CaO 0.87 12.42 0.41 1.21 0.40 Na20 0.97 0.15 1.19 0.16 0.28 k2o 2.20 2.75 1.95 2.52 2.64 P2O5 0.13 0.12 0.15 0.13 0.07 Cr203 0.017 0.015 0.012 0.024 0.027 TOT/C 0.27 3.03 0.40 0.41 0.30 TOT/S 0.01 0.06 0.03 0.06 <0.01 Organic matter 2.37 2.02 1.26 2.70 1.82 Table 4. Portion p of the clay minerals in the soil, parameters a and b , interlayer water w, intergrain water at the liquid limit w , liquid limit LL, intergrain water at the plastic limit w^pL, plastic limit PL, placticity index 1'. Tabela 4. Delež glinenih mineralov v Zemljini p, parametra a in b , medslojna voda w, medzrnska voda na meji židkosti w , meja židkosti LL, medzrnska voda na meji plastičnosti w , meja plastičnosti PL, indeks plastičnosti 1. Sample 1 2 3 4 5 C E C E C E C E C E P 0.52 0.49 0.44 0.89 0.52 ae (%) 47.32 50.67 44.73 49.67 31.36 34.22 83.55 85.82 49.79 48.94 be 0.149 0.164 0.149 0.148 0.133 0.113 0.152 0.152 0.152 0.151 w,. (%) 4.05 1.02 0.55 8.14 0.93 We\LL (%) 40.90 43.15 38.66 42.98 27.61 30.64 72.00 73.96 42.91 42.22 LL (%) 44.95 47.20 39.68 44.00 28.16 31.19 80.14 82.10 43.84 43.15 (%) 20.59 20.25 19.47 21.71 14.92 18.22 35.75 36.78 21.31 21.07 PL (%) 24.64 24.30 20.49 22.71 15.47 18.77 43.89 44.92 22.24 22.00 iP (%) 20.31 22.90 19.19 21.29 12.69 12.42 36.25 37.18 21.60 21.15 Note: ae = 33.70 + 0.99-As; be= 0.05• Asc027 ;Asc=p-As; = ae ■ 2.66~b<; LL = W(|£l + w,; weipl = ae' 266 ; PL = we+ w(.; Ip = we LJ — we\PL = LL — PL because the interlayer water quantity is equal at the liquid and plastic limits. Table S. The specific surface As and the portion p of the clay minerals in the soil, the water content w before and after the test, the calculated interlayer water quantity w, the measured and calculated intergrain water quantity w after consolidation under different axial stresses s', the calculated and experimentally determined soil dependent parameters i and j . Tabela S. Specifična površina As in delež glinenih mineralov p v Zemljini, vsebnost vode w pred in po preiskavi, izračunana količina medslojne vode w., merjena in izračunana količina medzrnske vode w po konsolidaciji pri različnih efektivnih napetostih s', izračunana in eksperimentalno določena parametra i in j . Sample 1 2 3 4 5 As (m2/g) 30.1 28.5 16.7 54.1 32.6 P 0.52 0.49 0.44 0.89 0.52 w (%) before the test o'= 50 kPa 49.36 45.92 31.20 70.13 37.77 a'=100 kPa 49.36 44.43 30.38 69.35 37.70 a'=200 kPa 44.99 44.05 27.95 66.22 35.25 w (%) after the test a'=50 kPa 37.47 36.76 25.82 66.04 34.72 a'=100 kPa 33.67 34.23 22.56 60.12 31.51 a'=200 kPa 30.85 32.68 23.42 55.16 27.87 wt (%) 4.05 1.02 0.55 8.14 0.93 E - we = w - wi (%) a'=50kPa 33.42 35.74 25.27 57.90 33.79