Influence of mineralogy on swelling and consolidation of ...€¦ · tial hazards with indigenous soils containing expansive clay minerals. Heave and settlement are the most important

Influence of mineralogy on swelling andconsolidation of soils in eastern Saudi Arabia

Shahid Azam

Abstract: The mineral calcium sulphate transforms from anhydrite through bassanite to gypsum and then can convertback to bassanite and anhydrite. This transformation of calcium sulphate causes volume change that adds to the poten-tial hazards with indigenous soils containing expansive clay minerals. Heave and settlement are the most importantgeotechnical problems associated with many argillaceous soils in eastern Saudi Arabia. Mineralogical evolution andengineering behaviour of such soils are governed by local geology and severe climatic and environmental conditionsprevalent in the region. Based on laboratory investigations, this paper discusses the influence of mineralogy on swellingand consolidation of expansive clay, calcium sulphate forms, and their mixtures. Mineralogy is investigated usinggeotechnical index properties and thermal and X-ray diffraction analyses. Swelling and consolidation characteristics aredirectly determined and are studied in conjunction with microstructural assessment. Results indicate that under the hotand humid climate prevalent in eastern Saudi Arabia, form changes in calcium sulphate can take place in the timeframe of engineering importance. The volume-change behaviour of local expansive clay depends on the form andamount of calcium sulphate present in the soil.

Key words: calcium sulphate, expansive clay, swelling, consolidation.

Résumé : Le minéral de sulfate de calcium se transforme d’un anhydride en passant par un basanite et en gypse etpeut alors se convertir à rebours vers un basanite et un anhydride. Cette transformation du sulfate de calcium produitun changement de volume qui augmente le potentiel de risque rencontré avec les sols indigènes contenant des miné-raux argileux gonflants. Le soulèvement et le tassement sont les deux plus importants problèmes géotechniques reliés àplusieurs sols argileux en Arabie Saoudite de l’est. L’évolution minéralogique et le comportement mécanique de telssols est régi par la géologie locale et par des conditions climatiques et environnementales sévères dominantes dans larégion. En partant d’études en laboratoire, cet article discute de l’influence de la minéralogie sur le gonflement et laconsolidation de l’argile gonflante, des formations de sulfate de calcium et de leurs mélanges. La minéralogie estétudiée au moyen des propriétés d’indice, et des analyses de diffraction thermique et aux rayons-X. Les caractéristiquesde gonflement et de consolidation sont déterminées directement et sont étudiées par rapport à l’évaluation microstructu-rale. Les résultats indiquent que dans les conditions humides et chaudes qui prévalent dans l’Arabie Saoudite de l’est,les changements de forme du sulfate de calcium peut se produire dans une période de temps significative pour lestravaux d’ingénieurs. Le comportement de changement de volume de l’argile gonflante locale dépend de la forme etde la quantité de sulfate de calcium présentes dans le sol.

Mots clés : sulfate de calcium, argile gonflante, gonflement, consolidation.

[Traduit par la Rédaction] Azam 975

Introduction

Swelling and consolidation of soils are generally associ-ated with expansive clay minerals such as smectite and illite(Slater 1983), whereas non-clay minerals are considered in-ert (Azam and Al-Shayea 1999). Calcium sulphate is anexception to this generalization because it undergoeshydration–dehydration that results in transforming anhydrite(CaSO4) through bassanite (CaSO4·0.5H2O) to gypsum(CaSO4·2H2O), and vice versa (Abduljauwad et al. 1998).According to Blatt et al. (1980), gypsification of anhydrite

results in swelling that can be as high as 62%, when basedon molar volumes. Likewise, complete dehydration of gyp-sum leads to a volume reduction of 38% for the same theo-retical volumes (Zanbak and Arthur 1986). Geology,climate, and environment of eastern Saudi Arabia provide anatural setting for the occurrence of such phenomena.

Structural damage caused by expansive clays is well docu-mented in the literature (Chen 1988). Gypsification of anhy-drite is known to create floor heaving in tunnels and uplift indams (Deer et al. 1972), and dehydration of gypsum leads tofracturing due to settlement (Ko et al. 1995). In easternSaudi Arabia, the hazard caused by both these soil types iscombined in many argillaceous soils possessing mixed min-eralogy (Abduljauwad et al. 1998). Such damage includescracks in masonry fences, grade beams, and members ofreinforced concrete, uplift of floating slabs on grade, andheave of pavements and walkways. Distress to both struc-tures and pavements is multiplied when the underlying soils

Can. Geotech. J. 40: 964–975 (2003) doi: 10.1139/T03-047 © 2003 NRC Canada

964

Received 22 May 2002. Accepted 15 July 2003. Published onthe NRC Research Press Web site at http://cgj.nrc.ca on24 September 2003.

S. Azam. Department of Civil and EnvironmentalEngineering, University of Alberta, Edmonton, AB T6G 2G7,Canada (e-mail: [email protected]).

I:\cgj\CGJ40-05\T03-047.vpSeptember 19, 2003 8:34:03 AM

Color profile: DisabledComposite Default screen

are periodically and (or) differentially wetted (Azam andAbduljauwad 2000).

Soil behaviour depends on the type and amount of miner-als present therein. Evaluation of the form changes in cal-cium sulphate and the interaction of these forms with clayminerals is key to the fundamental understanding of thevolume-change behaviour of local soils. Based on laboratoryinvestigations, this paper discusses the influence of mineral-ogy on swelling and consolidation of expansive soils fromeastern Saudi Arabia and highlights the relative significanceof the form and amount of calcium sulphate on potential vol-ume change. Results of geotechnical index properties anddifferential thermal (DT) and X-ray diffraction (XRD) anal-yses are used to assess the mineral composition of clay andcalcium sulphate forms. Swelling–consolidation test resultsare studied in conjunction with soil microstructure obtainedfrom scanning electron microscopy (SEM).

Mineralogical evolution

The evolution of various minerals in the argillaceous soilsof eastern Saudi Arabia was governed by several interactinggeologic, climatic, and environmental factors. Local expan-sive soils are derived from calcareous limestone, marl, shale,and chert of the Tertiary and the Quaternary. Proximity to al-kaline waters of the Persian Gulf, extreme degree of aridity,and restrained leaching resulted in the formation of expan-sive clay minerals such as smectite and illite (Abduljauwad1994). The development of diagenetic minerals like calciumsulphate within the host clay was based on dolomitisation ofcalcareous mudstone (Akili and Ahmed 1983). In the mud-stone containing aragonite and high-Mg calcite, Mg2+ re-placed Ca2+ in the crystal lattice of aragonite (Cook et al.1985). The displaced Ca2+ combined with SO4

2– from theseawater and precipitated as gypsum (Blyth and deFreitas1984). Subsequent dehydration of the gypsum led to the de-velopment of bassanite and anhydrite.

Dehydration of gypsum took place as a result of increas-ing overburden pressure due to burial (Blatt et al. 1980).Likewise, rehydration of bassanite and anhydrite was not anuncommon occurrence because of uplift and erosion coupledby daily and seasonal variations in temperature and relativehumidity. This rendered the distribution of various calciumsulphate forms in the clay matrix extremely variable inboth the vertical and horizontal directions (Azam andAbduljauwad 2000). The present-day clay sediments of Al-Qatif, Al-Hassa, and Ummus Sahik contain all forms ofcalcium sulphate in quantities ranging from 5 to 85%(Abduljauwad et al. 1998). Although dehydration of gypsumis a geologic process, hydration of bassanite and anhydritecan occur in these argillaceous soils in the time frame ofengineering importance under the hot and humid climateprevalent in eastern Saudi Arabia (Azam and Abduljauwad2000).

Experimental program

The aim of the experimental program was to understandthe swelling and consolidation of local expansive clays con-taining variable amounts of calcium sulphate forms. For thispurpose, expansive clay was obtained from Al-Qatif, and

gypsum, bassanite, and anhydrite were obtained from acalcareous deposit in Dhahran; both towns are located ineastern Saudi Arabia between latitudes 26°35′N and 26°40′Nand longitudes 49°34′E to 49°37′E. To maintain their initialwater contents, all samples were wrapped with cheeseclothand painted with molten wax. Thereafter, these sampleswere stored at a temperature of 25 °C and relative humidity(RH) of 50 ± 10%.

Laboratory investigations were conducted in stages. Pre-liminary evaluation of the mineralogy and the behaviour ofclay and of calcium sulphate forms was made using geo-technical index properties. Detailed mineral assessment wasbased on DT and XRD analyses. Results of these two analy-ses were also used to choose critical temperature and relativehumidity values representing various stages of form changesin calcium sulphate.

Swelling and consolidation behaviour of field samples ofclay and calcium sulphate forms was directly determined us-ing a conventional oedometer test. The same test was alsoconducted for various synthetic clay – calcium sulphate mix-tures. Prepared by blending pulverized materials on a dry-weight basis, these samples were statically compressed inthe oedometer. To evaluate the influence of fabric on thevolume-change behaviour, SEM was conducted on fieldsamples and on selected synthetic mixtures both before andafter water addition in the swelling–consolidation tests.

Geotechnical index properties

Geotechnical characteristics of different materials wereassessed using index properties. Preliminary evaluation ofthe mineralogy and behaviour of clay and calcium sulphateforms was based on water content (w), specific gravity (Gs),dry unit weight (γd), and consistency limits (liquid limit wland plastic limit wp). For consistency limits, all materialswere pulverized to obtain the fraction passing 425 µm(American Society for Testing and Materials (ASTM) sieveNo. 40).

Table 1 summarizes the geotechnical index properties ofthe clay and calcium sulphate and indicates that the field wa-ter content of clay is 41.8%, which corresponds closely to itsoptimum water content as determined by ASTM test methodD698-00a (Abduljauwad 1994). The w of various calciumsulphate forms is actually the water of crystallization. Thisparameter was determined at different temperatures, as dis-cussed later in the DT and XRD analyses.

The low field γd of clay reported in Table 1 should be at-tributed to its high void ratio of 1.56. Such a high void ratiofor the clay is derived from the high amount of small clayparticle sizes associated with various expansive clay miner-als. On the contrary, the difference in the field γd for the cal-cium sulphate forms is mainly due to the variation in Gs.Knowing the percentage of calcium sulphate (C), mixturesof clay and calcium sulphate were compacted to theweighted average of field γd of the constituents according tothe following equation:

[1] γd mixture = (1 – C)γd clay + Cγd calcium sulphate

This linear relationship provides a common basis for acomparison among different compacted clay – calcium sul-phate mixtures and among the field and compacted speci-

© 2003 NRC Canada

Azam 965



mens. The significance of dry unit weight on the swellingand consolidation behaviour of the investigated soils is de-scribed later in the paper.

The high wl and wp values of the clay are attributed to thepresence of expansive clay minerals. Facilitated by ex-changeable cations and large specific surface areas, theseminerals adsorb high amounts of water (Mitchell 1993). Onthe same note, the consistency limits of the calcium sulphateforms increase with dehydration of the mineral. Table 1 fur-ther indicates that wl of anhydrite equates to that of eithergypsum or bassanite when their respective waters of crystal-lization are added to their observed wl. This indicates that,given full access to water, both anhydrite and bassanite hy-drate to gypsum.

Differential thermal (DT) analysis

DT analysis was aimed at investigating water adsorptionby clay and hydration–dehydration in calcium sulphate. Inaddition, the analysis was also used to identify critical tem-peratures pertaining to mineral changes in the latter. For thispurpose, a Simultaneous Thermal Analyser (Netzcsh ModelSTA 429) was used. The analyses were conducted on 100 mgpowdered samples, each placed in an alumina (Al2O3) cruci-ble; the temperature was raised from 20 to 1000 °C at a uni-form rate of 10 °C/min.

Figure 1 depicts the results of the thermal analysis of theclay in the form of DT and thermogravimetric (TG) curvesand indicates that the dehydration of clay occurs in morethan one stage. The first and the largest low-temperature en-dothermic peak between 20 °C and 300 °C (DT curve) cor-responds to the removal of adsorbed or interlayer water. Thisis accompanied by an 11% mass reduction (TG curve),which indicates the presence of clay in the sample (Macken-zie 1992). The flat exothermic peak between 300 °C and500 °C in the DT curve is produced by palygorskite claymineral. The second endothermic peak and the correspond-ing reduction in mass (3%) between 500 °C and 600 °Cpertain to OH-ion expulsion from the mineral lattice (dehy-droxylation). The third endothermic peak and the reductionin mass (1%) between 800 °C and 900 °C are due to thedecomposition of carbonate ions associated with dolomitepresent in the calcareous expansive clay (Abduljauwad1994).

Figure 2, which gives the results of the DT analysis ofcalcium sulphate, indicates a single dehydration reactiongiven by the double endothermic peak between 130 °C and220 °C (DT curve). Such dehydration is accompanied by a

21% reduction in mass as shown by the TG curve that isparallel to the abscissa up to a temperature of 1000 °C. Itfollows that gypsum remains unaltered up to a temperatureof 130 °C, whereas anhydrite is the stable form of calciumsulphate beyond 220 °C. This is separately confirmed by ad-ditional water content data (not provided in this paper) forcalcium sulphate determined at temperatures of 25, 50, 100,

© 2003 NRC Canada

966 Can. Geotech. J. Vol. 40, 2003

Material

PropertyASTM testmethod Clay Gypsum Bassanite Anhydrite

w (%) D2216-98 41.8 20.9* 6.2* 0.0*Gs D854-00 2.82 2.31 2.77 2.92γd (g/cm3) D2937-00 1.1 1.5 1.7 1.8wl (%) D4318-00 174.6 42.8 54.5 61.3wp (%) D4318-00 63.7 36.2 40.1 43.9

*Water of crystallization.

Table 1. Geotechnical index properties of expansive clay and calcium sulphate forms.

Fig. 1. Thermal analysis of expansive clay.

Fig. 2. Thermal analysis of calcium sulphate.



125, 150, 200, and 300 °C, with a heating time ranging from24 to 96 h. These observations are consistent with other pub-lished data such as those of Kosztolanyi et al. (1987), whoreported that the temperature range of gypsum to anhydriteform transformation is 130–180 °C and that the transforma-tion is complete by 200 °C.

The DT curve of calcium sulphate also illustrates a smallendothermic peak at 200 °C, which pertains to the conver-sion of gypsum to bassanite. Further, the tiny hump in theDT curve at a temperature of 350 °C indicates the conver-sion of soluble anhydrite to insoluble anhydrite (Mackenzie1992).

Heating of the investigated clay results in releasing anamount of water equal to 11% of the total mass of the clay.This is the adsorbed or interlayer water, imbibition and liber-ation of which depend only on water availability. On thecontrary, transformation of gypsum to anhydrite is associ-ated with liberating water of crystallization that amountsto 21% of the total mass of calcium sulphate. The reverseprocess termed gypsification of anhydrite is governed bytemperature and relative humidity. Therefore, the interactionbetween clay adsorption and hydration–dehydration of cal-cium sulphate is key to the understanding of swelling andconsolidation in natural argillaceous soils present in easternSaudi Arabia. Study of this interaction is also vital becauseof the nonuniform distribution of calcium sulphate in thehost clay matrix and form changes of the mineral in bothvertical and horizontal directions (Azam and Abduljauwad2000).

X-ray diffraction (XRD) analysis

XRD analysis was conducted to determine mineral com-position of the clay and calcium sulphate under different en-vironmental conditions. The analysis was carried out using aPhilips diffractometer (model Pw 173/10), which generatedradiations at 40 kV and 40 mA. The scanning speed was0.01° 2θ/s and the angle scanned was 4°–80°. Randomly ori-ented samples were prepared by manual grinding air-driedmaterial in a porcelain mortar and pestle to powder form andsubsequently pressing the material lightly into rectangularmetal holders, 20 mm × 10 mm in size. To identify mineraltypes, qualitative assessment was made by comparing thediffraction pattern of each sample with the standard patternsprepared by the Joint Committee of Powder Diffraction DataService (JCPDS). The amount of minerals present in thesamples was quantitatively estimated using areas under thepeaks.

Figure 3a gives XRD results for the investigated clay andindicates the presence of various clay mineral species at 2θ <10°. Quantitative XRD analysis revealed that major constitu-ents of the clay are smectite (50 ± 5%), illite (20 ± 5%), anddolomite (10 ± 5%), and palygorskite, quartz, and gypsumcollectively measured 15 ± 5%.

Three sets of gypsum samples were heated for 48 h totemperatures of 130, 200, and 300 °C, respectively. Each setwas then cooled at room temperature (25 °C) under relativehumidities of zero (desiccator), 50 ± 10% (open air), and100% (water tub). Figures 3b, 3c, and 3d show XRD pat-terns of various calcium sulphate forms and indicate typicalsharp peaks for gypsum, bassanite, and anhydrite at 2θ =

12°, 30°, and 26°, respectively. Similar diffraction patternswere obtained for all the other samples, each treated under adifferent temperature and relative humidity.

Table 2 provides a summary of the quantitative estimatesof the XRD results for calcium sulphate. These results indi-cate that samples cured in water were completely convertedback to gypsum, irrespective of the heating and cooling tem-peratures. This is because of the high water affinity of bassa-nite and anhydrite, both of which readily take up anyavailable water. Conversely, no rehydration was observed forsamples heated to 300 °C, provided that the relative humid-ity did not exceed 50 ± 10%. This permanent change is dueto changes in the crystal structure of calcium sulphate at ele-vated temperatures.

Table 2 indicates that samples heated to 130 and 200 °Cand subsequently cooled in a desiccator displayed moreanhydrite (less bassanite) than those in the open air. This isbecause of the higher moisture availability in the latter envi-ronment. For all these samples, the marginal variation in theamounts of anhydrite and bassanite suggests experimentalerrors during sample transfer from the oven to the desicca-tor. Despite such errors, the amount of bassanite was alwayshigher than that of anhydrite for all the samples, and noneshowed any trace of gypsum. This means that moisture firstleads to partial rehydration prior to gypsification.

Removal of water of crystallization from gypsum dependson the origin, previous history, grain size, texture, lattice im-perfections, and presence of impurities in the mineral (Deeret al. 1972). This study suggests that in local soils, water ofcrystallization of gypsum cannot be removed at the oventemperature of 105 ± 5 °C, and therefore the water contentshould be determined at this temperature (Azam andAbduljauwad 2000). Results of this study further indicatethat hydration of bassanite and anhydrite can take place inthe time frame of engineering importance under the tempera-ture and relative humidity conditions of eastern Saudi Ara-bia.

Swelling–consolidation and soilmicrostructure

Knowledge of the potential volume-change characteristicsin soils at the outset of any engineering construction projectis obligatory because of their relationships to heave dueto expansion and settlement due to compression. Data onvolume-change characteristics for this study were collectedby direct measurement of soil deformation in the laboratoryoedometer test. Fundamental understanding of soil behav-iour during the oedometer test was provided by visually ob-serving the soil microstructure.

Swelling–consolidation tests were performed in fixed-ringoedometers on field samples of clay and calcium sulphateforms and their synthetic mixtures. The specimens were pre-pared according to the ASTM standard test method D4546-96 for one-dimensional swell or settlement potential ofcohesive soils and standard test method D2435-96 for one-dimensional consolidation properties of soils. Field sampleswere tested under in situ conditions as given in Table 1, andmixtures were compacted to the weighted average of field γdof the constituents according to eq. [1].

© 2003 NRC Canada

Azam 967



A morphological study was conducted on thin sections cutfrom the oedometer samples before and after water floodingin the swelling–consolidation tests; SEM specimens were

oven-dried at 105 ± 5 °C. Despite some shrinkage in thehigh clay content materials, oven-drying minimizes micro-structural changes, as the shorter time requirement of this

© 2003 NRC Canada


Fig. 3. Mineralogy of expansive clay and calcium sulphate forms. A, anhydrite; B, bassanite; G, gypsum; I, illite; P, palygorskite; Q,quartz; S, smectite.



method ensures a reduced particle rearrangement when com-pared with that for air-drying (Mitchell 1993; Tovey andWong 1973). The analyses were carried out on the JEOLscanning electron microscope (model JSM-840), which pro-vided fabric appraisal and gave an elemental descriptionusing an energy-dispersive X-ray analyser (EDXA). Eachsample was held in an aluminium sample holder andsputter-coated with a fine gold film. The micrographs weretaken in split-screen mode with enlargements of 1000 (gen-eral) and 6000 (enlarged inset) times.

Figure 4 presents swelling–consolidation test results forclay and calcium sulphate in the form of void ratio versus ef-fective stress and shows that the specimens are initially sub-jected to a seating stress of 7 kPa, which closely correspondsto the in situ effective stress acting on these field samples.All the samples exhibit an increase in void ratio whenflooded with water and a subsequent decrease in void ratiowith the application of load.

Table 3 gives volume-change characteristics of clay andcalcium sulphate forms and indicates that both the amountof expansion and the amount of compression depend on thetype of material. Among the calcium sulphate forms,anhydrite has the highest swelling potential of 9.32%, fol-lowed by bassanite and then gypsum. Because of the com-plete confinement and rigidity of the ring, this swellingpotential measured in the conventional oedometer is only15% of the maximum theoretical value of 62% estimated byBlatt et al. (1980). The cited theoretical estimate is based onmolar volumes and an open system that allows freeexit–reentry of water during hydration. Similar estimatedvalues of volume change in systems with partial and no wa-ter access fall below this boundary value (Zanbak and Arthur1986). Despite the effect of apparatus confinement and rigid-ity, the use of identical oedometer rings provided a bettercomparison of different materials. Table 3 indicates that theswelling potential of anhydrite is one-fourth that of the ex-pansive clay, which measured 34.6%.

© 2003 NRC Canada

Azam 969

Temperature (°C): 130 200 300

Relative humidity (%): 0 50 100 0 50 100 0 50 100

Gypsum — — 100 — — 100 — — 100Bassanite 56 59 55 59 — — — —Anhydrite 41 38 — 45 41 — 100 100 —

Table 2. Mineralogical composition of calcium sulphate forms.

Fig. 4. Void ratio – pressure (effective stress) relationship of ex-pansive clay and calcium sulphate forms.

Fig. 5. Swelling potential of expansive clay and calcium sulphateforms.

Material

Characteristic Clay Gypsum Bassanite Anhydrite

Initial void ratio (ei) 1.70 0.28 0.50 0.61Void ratio after swelling (es) 2.60 0.31 0.62 0.76Swelling potential (∆e/(1 + ei)100) 34.30 2.34 8.00 9.32Compression index (Cc) 1.166 0.005 0.217 0.267

Table 3. Volume-change characteristics of expansive clay and calcium sulphate forms.



© 2003 NRC Canada


Fig. 6. Microstructure of expansive clay and gypsum. The SEM micrographs to the right in (a) and (c) are enlargements of the rectan-gles shown in the SEM micrographs to the left. EDX, energy dispersive X-ray.



© 2003 NRC Canada

Azam 971

Fig. 7. Microstructure of bassanite and anhydrite. The SEM micrographs to the right are enlargements of the rectangles shown in theSEM micrographs to the left.



Figure 5 gives the rate of swelling for clay and calciumsulphate in the form of swelling potential versus time andshows that, with the exception of gypsum, the swelling-potential curves comprise three distinct zones. For clay,bassanite, and anhydrite, the initial swelling is generatedquickly due to sample hydration. Next, a high rate of in-crease in swelling potential is observed in the primary stagefollowed by a low rate in the secondary stage.

The high swelling potential of the clay is due to thepresence of smectite as shown in Fig. 6a, which gives theSEM micrograph after water addition in the swelling–consolidation test. The clay structure emerges as a crustyweb due to shrinkage of the specimen during sample prepa-ration. The enlarged inset shows honeycombed smectitemorphology formed by vaguely contoured foliated sheetswith involute edges. These sheets interact face-to-face andface-to-edge, forming enclosed rounded cells up to 10 µm insize. This flocculated structure corresponds to a void ratioincrease upon water imbibition and explains the high com-pression index (Cc) of the clay depicted in Table 3. Theenergy-dispersive X-ray (EDX) spectrum shown in Fig. 6bportrays typical peaks for Si, Mg, Al, and Ca, which are theelements constituting calcium-rich smectite. Similarly, the Kpeak is attributed to the presence of illite in the sample.

The SEM micrograph of gypsum (Fig. 6c), taken after wa-ter addition in the swelling–consolidation test, shows poorlydeveloped rosettes of lath-like gypsum crystals. These iso-metric euhedral crystals with well-defined boundaries areaggregated in separate but mutually conjoined colonies. TheSEM micrograph shows no sign of gypsum dehydration dueto load application. Therefore, the compressibility of thissample can be envisaged to be much less than the volume re-duction of 38% due to complete dehydration of gypsumquoted by Zanbak and Arthur (1986). The EDX spectrum(Fig. 6d) reveals the nearly equal amounts of Ca and S.

Figure 7 provides evidence of form changes due to wateraddition which explains the volume-change behaviour ofbassanite and anhydrite. Before water addition in the swelling–consolidation test, the SEM micrograph of bassanite (Fig. 7a)shows a complex configuration without any definite structural

motif. Under similar conditions, the SEM micrograph ofanhydrite (Fig. 7c) illustrates a compact soil fabric compris-ing platy anhydrite crystals. The corresponding SEM micro-graphs of the same samples of bassanite and anhydrite afterwater addition in the swelling–consolidation tests are givenin Figs. 7b and 7d, respectively. Both of these micrographsshow lath-like euhedral gypsum crystals formed by hydrationof the samples. The latter micrograph shows that gypsumcrystals are oriented in all directions, thereby creating largevoids. The high void ratio of this specimen after water addi-tion is responsible for its relatively high compressibility, asindicated in Table 3.

Figure 8 compares the swelling potential of various fieldsamples of clay and calcium sulphate and their syntheticmixtures. Figure 8 depicts a decrease in the swelling poten-tial of the clay with an increase in the amount of calciumsulphate because of the lower expansion capabilities of cal-cium sulphate. Moreover, the clay–gypsum curve plots be-low the clay–anhydrite curve, as gypsification of anhydrite isaccompanied by swelling.

Figure 8 also depicts the effect of sample preparation onvolume-change behaviour by comparing field samples ofclay and calcium sulphate with their recompacted mixtures.The use of in situ γd for field samples and the recompactionof synthetic mixtures to the weighted average of field γd ofthe constituents were based on previous experience with lo-cal soils. Azam and Abduljauwad (2000) noted that geology,climate, and the environment prevalent in eastern Saudi Ara-bia govern calcium sulphate deposition in expansive clays.This deposition is mostly in granular form, which necessi-tates that, for the time frame of engineering importance, theγd of the clay remain constant. Conversely, due to the highsolubility of calcium sulphate, the clay matrix may undergocompressibility. The associated increase in γd, however,would generally take place in geologic time because of theextreme degree of aridity of the area. The field γd of a soilcontaining both clay and calcium sulphate falls betweenthese boundary conditions and is best described by eq. [1].

Using different values of γd for each sample, Fig. 8 showsan initial sharp decrease in swelling potential of the claywith an addition of up to 20% calcium sulphate. Such a de-crease amounts to 38% when gypsum replaces the clay and56% when anhydrite replaces the clay. This reduction ispartly attributed to the loose soil fabric in the recompactedmixtures compared to the undisturbed clay microstructurethat is developed over a long time under natural conditions.Attom et al. (2001) measured a 30% decrease in the swellingpotential of various clay soils due to static compaction. Sim-ilar reasons elucidate the decrease in the swelling potentialbeyond 80% calcium sulphate content. A lower decrease of43% for the clay–anhydrite curve compared with 88% forthe clay–gypsum curve is attributed to the high measuredswelling potential of the field anhydrite sample. This ispartly due to a dense anhydrite fabric (Fig. 7c), which re-sulted from the dehydration of gypsum due to burial (Blatt etal. 1980).

The swelling-potential curves of both clay–gypsum andclay–anhydrite mixtures are primarily explained by micro-structural assessment. Due to gypsification of anhydrite bywater imbibition and the relatively well defined zones inthe swelling-potential curve of the clay–gypsum mixtures

© 2003 NRC Canada


Fig. 8. Swelling potential of clay, calcium sulphate forms, andtheir mixtures.



© 2003 NRC Canada

Azam 973

Fig. 9. Microstructure of clay–gypsum mixtures. The SEM micrographs to the right are enlargements of the rectangles shown in theSEM micrographs to the left. A, clay–gypsum crumbs; B, pore spaces; C, embedded gypsum; D, dolomite; E, partially parallel cleav-age laths or striated–furrowed gypsum crystals; F, bridging between the clay; G, cross-linked gypsum crystals; H, filamentouspalygorskite; I, crystalline gypsum; J, unexpanded illite crystals.



(Fig. 8), morphological examinations were carried out onlyfor clay–gypsum mixtures.

The initial sharp decrease in swelling potential of the claywith an increase in the amount of calcium sulphate is mainlyattributed to the low volume-change capability and cementa-tion of calcium sulphate. Figure 9a, which shows the SEMmicrograph of a clay–gypsum mixture containing 10%gypsum, indicates the formation of clay–gypsum crumbs(A) and interconnected pore spaces (B). Crumb formation isattributed to the cementatious properties of embedded gyp-sum (C) randomly scattered throughout the clay matrix. Themicrograph also shows the occasional presence of dolomite(D), thereby confirming the XRD results of the clay, whichshowed 9% dolomite. Dolomite is distinguished by its anhe-dral shape, which is characterized by relatively rounded andirregular crystal boundaries.

High amounts of calcium sulphate and the associated highcementation result in an increase in crumb size. Figure 9b,which shows the SEM micrograph of a clay–gypsum mix-ture containing 20% gypsum, captures part of a large-sizecrumb. This micrograph shows partially parallel cleavagelaths or striated–furrowed gypsum crystals (E), leading tosignificant bridging between the clay (F). The crumb sizeachieved at 20% calcium sulphate is the threshold size, be-yond which further reduction in swelling potential does notoccur.

Crumbing is not favoured at higher calcium sulphate con-tents owing to depleted cementation arising from the re-duced water content of the mixture. Figure 9c gives the SEMmicrograph of a clay–gypsum mixture containing 80% gyp-sum. This micrograph shows cross-linked gypsum crystals(G) and filamentous palygorskite (H) but negligible crumb-ing.

The drop in swelling potential at calcium sulphate con-tents higher than 80% is primarily attributed to its low voidratio. Under such conditions, the calcium sulphate matrix in-hibits expansion of the clay particles. Here calcium sulphateparticles with cemented end junctions squeeze the soft clayparticles. Figure 9d gives the SEM micrograph of a clay–gypsum mixture containing 90% gypsum. Alongside crystal-line gypsum (I), this micrograph also shows unexpandedillite crystals (J) arranged face to face in a highly accordantfabric.

Summary and conclusions

Knowledge of the potential volume-change characteristicsin soils at the outset of any engineering construction projectis obligatory because of their relationships to heave due toexpansion and settlement due to compression. Fundamentalunderstanding of the swelling and consolidation behaviourof soils is provided by studying the associated mineralogy.This is especially true for many argillaceous soils in easternSaudi Arabia, where geology, climate, and the environmentgovern their mineralogical evolution and engineering behav-iour. To correlate soil mineralogy with volume-change be-haviour of local soils, detailed laboratory investigations werecarried out. Geotechnical index properties and differentialthermal (DT) and X-ray diffraction (XRD) analyses wereused to assess soil mineralogy, whereas swelling and consol-idation characteristics were directly determined and studied

in conjunction with microstructural assessment. The findingsof this study are as follows:(1) Swelling and consolidation of indigenous soils are influ-

enced by form changes in calcium sulphate that can takeplace in the time frame of engineering importance underthe hot and humid environment prevalent in easternSaudi Arabia.

(2) Gypsum is stable up to 130 °C, bassanites are stablefrom 130 to 220 °C, and anhydrite is stable beyond220 °C at RH ≤ 50%. Gypsification is independent ofthe heating temperature and takes place at RH = 100%.

(3) Among calcium sulphate forms, anhydrite has the high-est volume-change capability, followed by bassanite andthen gypsum. The swelling potential and compressionindex of anhydrite are about one-fourth that of the in-vestigated clay.

(4) Addition of up to 20% calcium sulphate inhibits theswelling potential of the investigated clay by crumb for-mation owing to cementation and bridging of the for-mer. Swelling potential of soils with more than 80%calcium sulphate is reduced due to squeezing of the clayparticles.

Acknowledgments

The author is grateful to King Fahd University of Petro-leum and Minerals for providing laboratory and computerfacilities. Thanks to Dr. J.D. Scott, Professor Emeritus ofCivil Engineering at the University of Alberta, for reviewingthe manuscript.

References

Abduljauwad, S.N. 1994. Swelling behaviour of calcareous claysfrom the Eastern Province of Saudi Arabia. Quarterly Journal ofEngineering Geology, 27: 333–351.

Abduljauwad, S.N., Azam, S., Al-Shayea, N.A., and Al-Amoudi,O.S.B. 1998. The behaviour of geological formations containingexpansive clay and calcium sulphate phases. In Proceedings ofthe 8th International Congress of the International Associationof Engineering Geology, Vancouver, B.C., 21–25 Sept. 1998.Edited by D. Moore and O. Hungr. A.A. Balkema, Rotterdam,The Netherlands. Vol. 1. pp. 169–175.

Akili, W., and Ahmed, N. 1983. The sabkhas of eastern Saudi Ara-bia: geotechnical considerations. In Proceeding of the 1st SaudiEngineering Conference, 22–24 May 1983, Jeddah, Saudi Ara-bia. Vol. 1. King Abdul Aziz University Press, Jeddah, SaudiArabia. pp. 300–322.

Attom, M.F., Abu-Zreig, M.M., and Obaidat, M.T. 2001. Changesin clay swelling and shear strength properties with differentsample preparation techniques. Geotechnical Testing Journal,24: 157–163.

Azam, S., and Abduljauwad, S.N. 2000. Influence of gypsificationon engineering behavior of expansive clay. Journal of Geo-technical and Geoenvironmental Engineering, ASCE, 126:538–542.

Azam, S., and Al-Shayea, N.A. 1999. Clayey behaviour of calcare-ous soil. In Engineering for Calcareous Sediments: Proceedingsof the 2nd International Conference, Manama, Bahrain, 21–24 Feb. 1999. Vol. 1. Edited by K. Al-Shafei. A.A. Balkema, Rot-terdam, The Netherlands. pp. 199–208.

Blatt, H., Middleton, G., and Murray, R. 1980. Origin of sedimen-tary rocks. 2nd ed. Prentice-Hall Inc., New York.

© 2003 NRC Canada




Blyth, F.G.H., and deFreitas, M.H. 1984. A geology for engineers.7th ed. Edward Arnold, London, U.K.

Chen, F.H. 1988. Foundation on expansive soils. Elsevier ScientificPublishing Company, Amsterdam, The Netherlands.

Cook, D.J., Randazzo, A.F., and Sprinkle, C.L. 1985. Authigenicfluorite in dolomitic rocks of the Floridan aquifer. Geology, 13:390–391.

Deer, W.A., Howie, R.A., and Zussman, J. 1972. Rock formingminerals. Lowe and Brydon Printers Limited, London, U.K.

Ko, S., Olgaard, D.L., and Briegel, U. 1995. Transition from weak-ening to strengthening in dehydrating gypsum: evolution ofexcess pore pressure. Geophysical Research Letters, 22:1009–1012.

Kosztolanyi, C., Mullis, J., and Weidmann, M. 1987. Measure-ments of the phase transformation temperature of gypsum–anhydrite included in quartz by microthermometry and Ramanmicrobe techniques. Chemical Geology, 61: 19–28.

Mackenzie, R.C. 1992. The differential thermal investigation ofclays. Mineralogical Society, Clay Minerals Group, London,U.K.

Mitchell, J.K. 1993. Fundamentals of soil behaviour. 2nd ed. JohnWiley & Sons, Inc., New York.

Slater, D.E. 1983. Potential expansive soils in Arabian Peninsula.Journal of Geotechnical Engineering, ASCE, 109: 744–746.

Tovey, N.K., and Wong, K.Y. 1973. The preparation of soils andother geological materials for the SEM. In Proceedings of theInternational Symposium on Soil Structure, Gothenburg, Swe-den. pp. 59–67.

Zanbak, C., and Arthur, R.C. 1986. Geochemical and engineeringaspects of anhydrite/gypsum phase transitions. Bulletin of theAssociation of Engineering Geologists, 23: 419–433.

© 2003 NRC Canada

Azam 975



Documents

Influence of mineralogy on swelling and consolidation of ...€¦ · tial hazards with indigenous soils containing expansive clay minerals. Heave and settlement are the most important