Journal of ASTM International, Vol. 5, No. 3 Paper ID JAI100464 Available online at www.astm.org

Mokhfi Takarli1 and William Prince-Agbodjan1

Temperature Effects on Physical Properties and Mechanical Behavior of Granite: Experimental Investigation of Material Damage ABSTRACT: In this paper, the effect of thermal treatment on physical and mechanical properties of a granitic rock is experimentally investigated. The open porosity, gas permeability, P-wave velocity 共and their attenuation兲, ultimate strength, and Young’s modulus are measured on samples heated at temperatures ranging from 105°C to 600°C. First, results show good correlations between the evolution of physical properties and the amount of the damage induced by the thermal treatment. Second, the mechanical parameters are shown to be dependent on the microcracks’ density in the samples. The effect of temperature on the failure process in granite is also investigated using strain gauge measurements and permeability evolution in a uniaxial compressive test. The results show that the extent of the crack closure stage depends on the initial crack density and that the crack thresholds, which characterize the failure process of the rock under compressive loading, decrease with the thermal treatment. KEYWORDS: granite, thermal damage, physical properties, mechanical properties, failure process

Nomenclature

Latin Symbols

ka P Q VP E

⫽ ⫽ ⫽ ⫽ ⫽

Apparent permeability coefficient Pressure Volumetric gas flow rate P-wave velocity Young’s modulus

Greek Symbols


␾ ␳ ␮ ␴ ␧

⫽ ⫽ ⫽ ⫽ ⫽ ⫽

Diameter Porosity Density Viscosity Stress Deformation

Introduction It is well known that physical and mechanical properties of rocks depend strongly on their void network. In the natural state, crystalline rocks commonly display complex composite microcrack systems, which are formed during different geological processes and under varying conditions 关1兴. In granitic rocks, microcracking occurs with thermal contraction during cooling, tectonic stresses during stages of deformation, and stress relaxation during uplift and unroofing 关2兴. Schild et al. 关3兴 show that most micro-cracks in granite occur along grain boundaries in quartz polycrystals and along cleavage planes in feldspars and micas. Manuscript received February 16, 2006; accepted for publication February 15, 2008; published online March 2008. 1 Civil Engineering Department, INSA de Rennes, 20, Avenue des Buttes de Coësmes, Rennes, Bretagne 35043 France. Copyright © 2008 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.

2 JOURNAL OF ASTM INTERNATIONAL

FIG. 1—Photomicrograph of fresh granite. In many situations, such as extraction of geothermal energy, nuclear waste disposal, and fire in tunnels and in monuments, grantic rocks can be subjected to temperature elevation. Increasing temperature of this type of rock causes thermal cracking even if only a small temperature gradient is applied. This is due to the fact that thermal stresses arising from differential thermal expansion between mineral grains produce new micro-cracks or contribute to the opening of pre-existing ones 关4,5兴. In geothermal and nuclear applications, thermal induced cracks can influence the local stress state at different depths, the transport phenomena in hydrothermal systems, and the formation of porous networks in crustal rocks. In the case of granitic elements subjected to fire conditions, thermal cracking can induce weakening of rock; therefore, a fundamental understanding of thermal cracking mechanisms becomes crucial. In all these situations, the problem is basically the same; i.e., how do changes in temperature influence the physical and mechanical properties of rock? The aim of this study is to characterize the thermal damage in granite by laboratory measurements of key physical and mechanical properties of the rock. We are also interested in the evolution of the failure process induced by the heating treatment. In the first section of the paper we present the grantic rock studied and the heating procedure. In the second section the measurement of open porosity, gas permeability, and P-wave velocity 共and its attenuation兲 on fresh and thermally damaged granite are presented. The P-wave velocity test provides an accurate estimation of the total damage in the material while the permeability measurement gives indications of a specific material damage, which corresponds to porous network evolution. This network constitutes a possible pathway for aggressive solutions penetration. It is the combination of the P-wave velocity and permeability measurements that allow us to make a distinction between overall damage 共micro-crack density兲 and interconnected microcracks, which influence the fluid flow in the material. In general, measurements show good correlations between the evolution of the physical properties and the amount of damage induced by the thermal treatment. In Sec. 3, we investigate the effect of the thermal damage on the mechanical properties by measuring the ultimate strength and Young’s modulus in uniaxial compressive test. These mechanical parameters are shown to be dependent upon the density of micro-cracks induced by the thermal treatment. Finally, the evolution of the failure process in granite with temperature elevation is studied by analyzing strain-stress and permeability-stress curves. The paper ends with some conclusions. Sample Preparation and Heat Treatment Rock Description and Sample Preparation The specimens of this study are granite, from the Pyrénées 共France兲, with average grain size ranging from about 0.5 mm to 3 mm 共Fig. 1兲. The mineralogical composition of these specimens is determined using the feldspar coloration method proposed by Bailey and Stevens in 1960 and then modified by Laniz in 1964 关6兴. This method consists in a double coloration of the feldspar surface with a specific chemical treatment. The plagioclase and K-feldspar are colored in red and yellow, respectively. The determination of the mineral proportion in the rock is carried out using image treatment of the colored section of the samples. The specimens used show little variation in mineralogical composition. They are composed of 42 % quartz, 46 % plagioclase, 8 % K-feldspar, and 4 % mica 共Fig. 2兲. Samples subjected to thermal treatment were cored in the same orientation from a granitic rock. This procedure helped us to obtain cored samples with almost the same physical properties and provided a high level of replication under controlled test conditions. The sample shape was a cylinder of 40.0 mm in

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 3

FIG. 2—Mineralogical composition determination of studied granite, from feldspar coloration method and image analysis. The section diameter of the samples is
40 mm. diameter and 60.0 mm in length. The sample end surfaces were made parallel to within 1 / 100 mm. Fifteen samples with equal density were finally selected to perform our study. Three of the selected samples were naturally micro-fractured 共see Fig. 3兲. Heat Treatment Granite samples are heated in a furnace at ambient pressure with a rate of 1 ° C / min until a nominal temperature was reached. The low rate of heating is used to ensure that cracking events result only from the temperature effect and not due to thermal gradients across the sample. Samples are heated to 200, 300, 400, 500, and 600° C, held for 2 h and then cooled down at 1 ° C / min to room temperature. The adopted heating procedure does not generate a temperature gradient within the sample exceeding 0.9° C / cm 关7兴. All the treated samples were put into a dessicator during the period preceding the test. Physical Properties of Fresh and Heated Granite Laboratory measurements conducted on our set of samples are open porosity, permeability, velocity, and attenuation of ultrasonic waves. The initial state of tested specimens is characterized by measurements on samples only heated to 105° C 共see Table 1兲. Preheating to 105° C is necessary to eliminate humidity in the samples. Porosity Porosity is one of the basic physical properties of rocks. It can be classified according to different types such as absolute or total porosity, open porosity, and effective or connected porosity. The total porosity ␾ is the fractional volume of all void space inside a porous material. The open porosity ␾o is defined as the proportion of voids that are accessible to fluids. The effective or connected porosity ␾c is the volume fraction of pore spaces that are fully interconnected and contribute to fluid flow through the material, excluding dead-end or isolated pores that are not part of flow path. This last porosity is commonly quantified by permeability measurement 关8兴. Here, we investigated the open porosity, which is generally

FIG. 3—Picture of naturally micro-fractured sample.

4 JOURNAL OF ASTM INTERNATIONAL TABLE 1—Measured physical properties for fresh samples at 105° C (1 to 12: intact samples; 13 to 15: naturally micro-fractured samples). Specimen 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Mean value Standard deviation

␳, kg/ m3 2704 2692 2692 2694 2689 2689 2689 2698 2687 2695 2692 2689

␾ o, % 0.65 0.65 0.65 0.66 0.64 0.66 0.67 0.68 0.68 0.70 0.68 0.69

Q, ml/min 0.50 0.61 0.68 0.79 0.81 0.82 0.90 0.99 1.01 1.02 1.05 1.24

k a, m 2 4.64⫻ 10−18 5.66⫻ 10−18 6.42⫻ 10−18 7.42⫻ 10−18 7.56⫻ 10−18 7.71⫻ 10−18 8.43⫻ 10−18 9.31⫻ 10−18 9.47⫻ 10−18 9.57⫻ 10−18 9.86⫻ 10−18 1.16⫻ 10−17

VP, m/s 5185 5089 4869 4937 5013 5023 4930 4930 4861 5008 5103 4928

2692 2695 2687 2692 0.12 %

0.74 0.72 0.77 0.68 3.88

1.39 1.62 1.83 1.02 26.83 %

1.29⫻ 10−17 1.52⫻ 10−17 1.72⫻ 10−17 9.53⫻ 10−18 26.78 %

5159 5010 5116 5011 1.63 %

inferred from successive weight measurements of dry, water saturated, and immersed samples. To saturate the specimens, they were placed under vacuum for 24 h. Next, water was slowly introduced into the vacuum vessel until the specimens were covered. The vacuum was maintained for another 24 h, and after releasing the vacuum the specimens remained submerged in water at atmospheric pressure for 24 h 关9兴. The porosity of the fresh granite ranges between 0.64 % and 0.77 %. The porosity data are reported in Table 1. For the heated samples, Table 2 shows only a slight increase in porosity between 105° C and 500° C. This is explained by the fact that granite exhibits only minor structural modifications in this range of temperature, primarily due to the opening of pre-existing micro-cracks and the nucleation of new cracks as well as their extension at increasing temperature 关10兴. The most important change appears between 500° C and 600° C, where a significant increase of the open porosity is induced. This is explained by the connection of the discontinuities created during the previous phase, and by an increase in cracks number and an extension of crack size. These last phenomena are caused by the anisotropic expansion linked to the quartz transition ␣ / ␤ that occurs at 573° C and room pressure 关11兴. Permeability Permeability is one of the most important parameters used to characterize porous material. For rock materials, it depends on the rock porosity, pore throat geometry, tortuosity, and pore connectivity. If the percolating fluid is reactive to the rock, other factors 共wettability of the fluid, swelling of clays and other minerals, chemical interaction between the rock and the fluid兲 affect the permeability. For these reasons we use an inert gas 共helium兲 as the flowing fluid for the permeability measurement. The test was performed using a simple gas-flow method under a constant pressure difference across the specimen 关12兴. In order to eliminate any pore water, specimens were dried at a temperature of 105° C for two days. This period is sufficient to obtain a constant weight of the specimens. The specimens were then set in the TABLE 2—Porosity and density change in initially intact granite under temperature elevation. Temperature, °C 105 200 300 400 500 600

Porosity, % 0.68 0.71 0.81 0.91 1.10 2.85

Density, kg/ m3 2687 2678 2671 2662 2650 2613

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 5

FIG. 4—Gas flow rate evolution toward (establishing of the steady state flow). desiccator for 24 h prior to permeability measurement. The permeability measurements were carried out in an air-conditioned room 共20⫾ 1 ° C兲. The relative pressure 共Pi − Patm兲 and the confining pressure applied to the sample were 0.5 MPa and 0.9 MPa, respectively. The measurement of the volumetric gas flow rate was done when the steady state flow was reached. This condition was checked by measuring continuously the gas flow 共see Fig. 4兲. The steady state flow condition is considered to be reached if during a 10 min period, the variation of the flow rate does not exceed 2 %. In fresh granite, it can be noted that the time needed to attain the steady state flow varied from 30 min to 1 h depending on the sample permeability. The apparent permeability coefficient ka 共m2兲 is calculated from the Darcy relationship 共Eq 1兲 for laminar flow of compressible viscous fluid through a porous material 关12兴: ka =

Q 2 ␮LPatm A 共P2i − P2atm兲

共1兲

where Q is the volumetric gas flow rate 共m3 / s兲 experimentally measured at Patm 共atmospheric pressure equal to 105 Pa兲, Pi the applied gas pressure 共6 ⫻ 105 Pa兲, ␮ the viscosity of the helium gas 共2 ⫻ 10−5 Pa· s at 20° C兲, L the length of the sample 共m兲, and A the cross-sectional area 共m2兲. As shown in Table 1, the density, the porosity, and the P-wave velocity variations in fresh samples are less than 4 %, while the permeability variations reach 26 %. This result demonstrates the complexity of the porous network of the rock. Indeed, the permeability is not only affected by the total porosity of the rock but also by the number, the geometry, the size, and especially the interconnectivity of the pores 关13兴. Some studies 关13,14兴 have been carried out with simultaneous measurement of porosity and permeability. They show how difficult it is to relate these two macro-structural parameters using a simple relation. The flow phenomena in rocks are complex and all the pores do not contribute in the same way to the fluid circulation. In fact, only the pores that are fully interconnected contribute to the fluid flow through the material. Generally, the permeability decreases with the decrease of porosity. This is in agreement with our results. The variation of the gas permeability with the open porosity of the 15 specimens tested in this study is shown in Fig. 5. In order to compare how permeability changes with temperature for each sample, the permeability temperature has been normalized with respect to the initial value measured at 105° C 共Fig. 6兲. Intact samples, with initial porosity ranging between 0.64 % and 0.70 %, exhibit a small increase in permeability from 105° C to 300° C. This variation becomes more significant between 300° C and 500° C. Above 500° C, a rapid increase of permeability probably induced by the quartz ␣ / ␤ transition can be seen. For naturally micro-fractured samples with larger initial porosity 共0.72– 0.77 % 兲, the gas permeability evolution is similar except at 200° C, where the measured permeability exhibits a certain decrease 共⬎20 % 兲. Each stage of the permeability evolution results from a specific change of the porous network 共crack closure, crack nucleation, and crack extension兲 during the heating treatment 关10兴. Above 200° C, Skinner 关15兴 showed that the increase of porosity, throat diameter, and permeability in granitic rocks are mainly controlled by the thermal expansion of quartz. Below 200° C, our study shows that the variation of permeability is controlled by the initial micro-crack network of the samples. Thus, we observe an increase in permeability for the intact samples and a decrease for the micro-fractured samples. Géraud 关10兴 showed

6 JOURNAL OF ASTM INTERNATIONAL

FIG. 5—Relationship between open porosity and gas permeability on fresh granite. that granite with a larger crack density exhibits a significant decrease of both connected porosity volume and throat diameter when temperature is less than 200° C. These phenomena induce a decrease of the permeability in heated granite. Géraud 关10兴 concluded that, within this temperature range, the evolution of porosity volume and throat diameter are mainly controlled by the initial network. P-wave Velocity and Attenuation The P-wave velocity VP is calculated using the transmission travel time of an acoustic pulse along the axial direction of the samples. The experimental setup includes a waveform generator, two piezoelectric transducers 共with a resonant frequency of 500 kHz兲 mounted on sample holder, and a numerical oscilloscope board connected to a computer. A constant pressure is systematically applied 共with a constant mass兲 to ensure a tight contact between the rock specimen and the transducers. Preliminary tests showed that gel spreading onto the surfaces of the sample affect the measurement of gas permeability. For this reason, we use water as coupling agent and checked that this procedure did not reduce the sensitivity of P-wave velocity measurement. The travel time was measured using a program developed under LABVIEW system with a resolution of 50 ns. For unheated granite, the travel time ranges between 11.60 ␮s and 12.40 ␮s, and the relative error is about 0.1 ␮s, which corresponds to a VP resolution less than ⫾53 m / s. There is not a significant difference in P-wave velocity between intact samples and initially micro-fractured samples. This result can be explained by the fact that measurements were performed in parallel with the micro-fracture orientation. P-wave velocity is also measured for every maximum temperature treatment and the results are plotted in Fig. 7. As expected, there is a negative correlation between velocity and the damage induced by the heat treatment. All the samples tested show a weak decrease 共about 12 %兲 in velocity between 105° C and

FIG. 6—Gas permeability evolution according to the thermal treatment.

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 7

FIG. 7—P-wave velocity of granite according to the heat treatment.

300° C. The variation becomes greater at 500° C 共about 30 %兲. Between 500° C and 600° C, the velocity magnitude decreases dramatically 共more than 64 %兲. These results are in agreement with the porosity and permeability measurements presented in the previous section. In the case of initially micro-fractured samples, it has been shown that there was no increase in the permeability when the samples are heated at 200° C. The simultaneous decrease of P-wave velocity and permeability in theses samples can be explained by the fact that micro-cracks created at this temperature level are located mainly at the grain boundary and are not necessarily connected to the already existing pore network. The P-wave velocity measurement is affected by the total material damage, while the permeability measurement is affected by a specific material damage that corresponds to porous network evolution. The permeability value is only controlled by the amount of open and connected porosity. Géraud 关11兴, who studied the structural evolution of four granite samples heated from 20° C to 700° C, mentioned that micro-fractured samples exhibited a decrease in open porosity at low temperatures 共⬍200° C兲. Thus, permeability decrease observed in initially micro-fractured samples is probably induced by a reduction of open porosity in the material. The amplitude spectrum of granite is derived from the transmitted waveform by application of the fast Fourier transformation. Figure 8 shows the evolution of the amplitude spectra for each temperature treatment. Firstly, we can observe a reduction in the central amplitude. Contrary to P-wave velocity, there is a positive correlation between attenuation and induced thermal damage. Secondly, we observe a high attenuation filtering phenomenon that affects the high frequencies and preserves the low ones. This phenomenon characterizes the interaction between high frequency waves and heterogeneities. In our case, micro-cracks have a size d, which is larger than the wavelength ␭ 共2␲d  ␭兲 关15兴.

FIG. 8—Evolution of the amplitude spectrum according to the thermal treatment.

8 JOURNAL OF ASTM INTERNATIONAL TABLE 3—Mechanical properties of fresh and heated samples. Temperature, °C 105 300 500 600

Stress at failure, MPa 244共⫾2.40兲 224共⫾2.30兲 194共⫾9.75兲 128共⫾6.40兲

Young’s modulus, GPa 75共⫾2.25兲 62共⫾1.86兲 54共⫾2.70兲 28共⫾1.68兲

Mechanical Properties and Failure Process of Fresh and Heated Granite Mechanical Properties All specimens were tested in uniaxial compression at a constant displacement rate of 0.002 mm/ s. Axial and circumferential strains were measured with two strain gauges bonded directly to the specimen. The first one is parallel to the specimen axis and the second one is placed in the circumferential direction. In order to calculate the volumetric strain 共␧v兲, the following equation is used: ␧v = ␧a + 2␧r

共2兲

where ␧a and ␧r are, respectively, the axial and the radial strain. Two samples were tested for each temperature peak 共105, 300, 500, and 600° C兲 to determine the critical stress level and the elastic modulus 共E兲 “the tested samples were initially intact.” The results show a small decrease of the uniaxial compressive strength from 105° C to 300° C 共8 % 兲. The strength decrease is greater at 500° C 共20 % 兲 and becomes very significant at 600° C 共47 % 兲. The elastic modulus evolution presents the same trend, but the decrease is more pronounced 共17, 28, and 62 % respectively for 300, 500, and 600° C兲. The values of the uniaxial compressive strength and the elastic modulus are reported in Table 3. Failure Process The deformation and fracture characteristics of brittle rocks have been studied by a number of researchers 关16,17兴. These works lead to the same conclusion: the failure process can be divided into a number of stages based largely upon the stress-strain characteristics displayed during uniaxial or triaxial laboratory tests. Based on the stress-strain behavior of a loaded material, Brace 关16兴 and Bieniawski 关18兴 defined these stages as follows: crack closure; linear elastic deformation; crack initiation, and stable crack growth; crack damage and unstable crack growth; failure and post-peak behavior 共see Fig. 9兲. In a first approach,

FIG. 9—Stress-strain diagram showing the elements of crack development (after Martin [23]). Note that only the axial 共␧axial兲 and lateral 共␧lateral兲 strains are measured; the volumetric strain and crack volume are calculated.

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 9

FIG. 10—Axial and radial strain of heated granite in uniaxial compression test. we used this definition to analyze the effect of temperature elevation on the failure process taking place in our granite samples. Crack closure occurs during the initial stage of loading 共␴ ⬍ ␴cc, see Fig. 9兲 where pre-existing cracks, oriented at a certain angle with respect to the applied load direction, close. ␴cc corresponds to the stress at cracks closure and ␧cc 共see Fig. 10兲 is the corresponding inelastic deformation. During crack closure, the stress-strain response is nonlinear, exhibiting an increase in axial stiffness. The extent of this nonlinear region is dependent on the initial crack density and geometric characteristics of the crack population 关19兴. Figure 10, which gives the evolution of the axial and lateral stress-strain curves with the temperature treatment, shows that the axial deformation corresponding to the crack closure increases progressively between 105° C and 500° C and more significantly at 600° C. This result is in agreement with the gas permeability and P-wave velocity measurements, which show that the thermal treatment induces new micro-cracks in the granitic specimens. Once the majority of existing cracks have closed, linear elastic deformation takes place. Crack initiation 共␴ci兲 represents the stress level when microfracturing begins and is marked as the point where the lateral strain-stress curve departs from linearity. Crack propagation can be considered as being either stable or unstable. Unstable crack growth occurs at the point of reversal in the volumetric strain-stress curve and is also known as the point of critical energy release or crack damage stress threshold ␴cd 关20兴. Unstable cracking continues to the point where numerous micro-cracks have coalesced and the rock can no longer sustain an increase in load. In Fig. 10, we show that the lateral strains of treated granite are, generally, less affected compared to the axial strains. In general, the overall shape of the stress-deformation curves remains mostly the same. By contrast, at 600° C, and at the beginning of loading, negative lateral strains are observed. Homand and Houpert 关21兴, have observed the same phenomenon at temperatures above 400° C, and explained it by the high sensitivity of the lateral stress-strain curves to the cracking evolution. Figure 11 represents the calculated volumetric stress-strain curves of the heated samples.

FIG. 11—Volumetric strain of heated granite in uniaxial compression test.

10 JOURNAL OF ASTM INTERNATIONAL

FIG. 12—Stress-volumetric gas flow rate curves of the heated samples. In the previous paragraphs, it has been shown how the deformation and the failure process in granite occur in a certain number of stages. Electric resistance strain gauges were used to measure slight change in sample deformations, which are related to closing and opening cracks. This method is completed by a measurement of gas permeability evolution under a uniaxial compressive test. To measure the volumetric gas flow rate in the stressed samples, a specific permeability cell was used for this test. Permeability-stress curves showed four stages of evolution of the connected cracks 共see Fig. 12兲 defined as follows: • Crack closure 共␴ ⬍ ␴k-s兲 occurs at the initial stage of loading. During this stage, the permeability decreases progressively until a maximum number of pre-existing cracks are closed at the ␴k-s stress level. This stress level is generally larger than the one obtained by the measurement of the axial deformation 共␴cc兲. In thermally damaged samples, the value of the permeability at the end of this stage 共␴k-s兲 is higher than the initial value 共no heated sample; see Fig. 12兲. This phenomenon can be explained by the fact that micro-cracks induced by the temperature elevation do not have a preferential orientation and the mechanical loading closes only those oriented at a certain angle with respect to the applied load direction. • Permeability stabilization 共␴k-s ⬍ ␴ ⬍ ␴k-coal兲: the volumetric gas flow rate remains constant during mechanical loading. In this phase, we consider that there is combination of crack initiation, crack growth, and crack closure. • Crack interconnection and crack growth 共␴k-coal兲 represent the stress level where the permeability increases with the compressive loading. • Unstable crack growth 共␴k-pi兲 and failure. In this last stage, permeability increases rapidly. In Fig. 12 it can be shown that the overall shape of the stress-permeability curves remains similar. In Fig. 13, we represent the evolution of the different micro-crack thresholds, obtained by the analysis of stress-strain curves and stress-permeability curves, with the temperature elevation between 105° C and 600° C. We can observe that the majority of the crack thresholds 共␴k-s, ␴cd, ␴k-coal, and ␴k-pi兲 are affected by the thermal damage of the material and especially above 500° C with the quartz transformation. Indeed, the thermal micro-cracks induced in the material enhance the interaction probability between micro-cracks. This interaction increases with the initial crack density. However, the ␴ci stress level that characterizes the crack initiation remains approximately constant between 105° C and 500° C. Within this range of temperature, there is no mineralogical transformation and the matrix resistance of the material is not really affected. Discussion In the rock deterioration process, voids, especially cracks in crystalline rocks, are the main factors controlling the intensity of the physical and chemical damage suffered by rocks when subjected to new environmental conditions. Thus, it is important to know how the void proportion in the rock can change according to the service stresses, and the effect of new micro-discontinuities on the rock properties. In the case of thermal stresses, our study shows that temperature increase in granite enhances open porosity, connected porosity, and overall porosity inside the material. These micro-structural modifications were

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 11

FIG. 13—Micro-crack thresholds change according to the thermal treatment. detected and quantified, respectively, by the following measurements: adsorption capacity by water saturation technique, gas permeability, P-wave velocity, and attenuation. In Table 4, it can be seen that permeability and open porosity increase with temperature elevation. This result highlights an extension of the porous network induced by the thermal micro-cracking. The new micro-cracks can accelerate the degradation process of granite when it is subjected to water penetration. Indeed, it is known that water is the main weathering agent and that open micro-cracks are the way for it to penetrate inside the rock. The fluid reacts with the rock at the rock-water interface, producing mineral dissolution and/or precipitation reactions 关22兴. In silicate rocks, water reacts with silicate to break the bond, inducing profound chemical effects on the physical properties, and on the micro-structure of the rock 关23兴. If the thermally damaged rock is also subjected to low temperature conditions, the effect of the freezing/thawing of water can be significant. Water stored in the porous network expands 9 % in volume and induces locally large tensile stresses which are responsible for material damage. When the material is thawed, water flows through the porous network and accelerates the damage 关24兴. All thermally induced micro-cracks are not necessarily interconnected and cannot be detected by porosity and permeability measurements. The unconnected discontinuities do not take part in material degradation during fluid-rock interaction. However, they can affect the mechanical behavior of the rock. The ultrasonic wave propagation in the rock interacts with theses discontinuities but also with the connected ones. Therefore, P-wave velocity decrease and attenuation increase observed in heated samples 共see Table 4兲 can be related to the overall damage of the material. In addition, as the ultrasonic wave propagation is related to some mechanical properties of materials, in particular to the elastic modulus, then the drop in longitudinal wave velocity involves a drop in the elastic modulus of the rock. In Table 4, it can be seen that the drop in VP and E during the thermal treatment are approximately the same. In geothermal systems, when the heated water is returned to the surface, the rock of the upper layers can be subjected to temperature elevation. Our result shows that this temperature elevation causes development of cracks and induces permeability augmentation in the material. Crack and permeability increase can enhance the efficiency of the geothermal system but can also affect the mechanical behavior at TABLE 4—Physical and mechanical properties change in initially intact granite under temperature elevation. The measurements are normalized with respect to the initial state.

Temperature, °C 105 200 300 400 500 600

Porosity 0.68 % 1.04␾105°C 1.19␾105°C 1.34␾105°C 1.62␾105°C 4.19␾105°C

Permeability 5.99⫻ 10−18 m2 1.15ka-105°C 1.72ka-105°C 2.86ka-105°C 6.26ka-105°C 224ka-105°C

P-wave velocity 5013 m / s 0.95VP-105°C 0.86VP-105°C 0.78VP-105°C 0.70VP-105°C 0.33VP-105°C

Stress at failure 244 MPa 0.92␴105°C 0.80␴105°C 0.52␴105°C

Young’s modulus 75 GPa 0.83E105°C 0.72E105°C 0.37E105°C

12 JOURNAL OF ASTM INTERNATIONAL

different depths. Indeed, it has been shown that the mechanical parameters 共strength and Young’s modulus兲 and the crack thresholds, which characterize the failure process, are affected by temperature elevation. For nuclear waste disposal application, temperature elevation can lead to material degradation and thus to leakage of nuclear radiation into the environment. The permeability measurement in thermally damaged rock allows the evaluation of this type of degradation; in other words, the evaluation of the transport properties of the rock. However, we must keep in mind that our permeability data are not an intrinsic property of the rock. Permeability was measured using a compressible fluid and it depends on the pressure level at which the gas is percolated. Conclusions In this paper, we have experimentally investigated the effects of temperature elevation on key physical properties and mechanical behavior of natural granite. Regarding the physical properties, open porosity, permeability, and P-wave velocity 共and attenuation兲 were measured in fresh conditions and after temperature treatments at 200, 300, 400, 500, and 600° C. The evolution of these properties under the thermal treatments was associated with the micro-crack network development and the micro-structural transformations in the rock specimens. Results showed that granite is progressively damaged until 500° C; the damage becomes more significant with the quartz ␣ / ␤ transformation at 573° C. Indeed, the following observations have been made: • The porosity increases slightly between 105° C and 500° C. This is due to the opening of preexisting micro-cracks and probably nucleation of new cracks and their extension at increasing temperature. The greatest change appears between 500° C and 600° C. • The permeability changes in three stages: The volumetric gas flow rate increases slightly from 105° C to 300° C, significantly between 300° C and 500° C, and dramatically above 500° C. It has been also shown that initially micro-fractured samples exhibit a permeability decrease at 200° C, which is probably induced by a partial closure of the pre-existing cracks due to the mineral dilatation. • The P-wave measurements are in good agreement with the porosity and permeability evolution. Thus, the heating treatment causes the decrease of the velocity and the increase of the attenuation. Regarding the mechanical behavior, we have investigated the effect of the thermally induced damage on the macroscopic properties of the tested granite and its effect on the failure process. The results showed that the ultimate strength and Young’s modulus in uniaxial compressive test decrease with temperature elevation. The failure process was characterized by analyzing the strain-stress and permeability-stress curves, which showed that: • The extent of the crack closure stage, which depends principally on the initial crack density in the rock, increases with temperature elevation. This is in compliance with the physical measurements, which show that the thermal treatment induces new micro-cracks in the granitic specimens. • The crack thresholds, which characterize the failure process of the rock under compressive loading, are also affected. For example, the crack damage stress threshold 共␴cd兲, which is also considered as the creep strength of the rock, decreases with temperature elevation. References 关1兴 关2兴 关3兴 关4兴 关5兴

Vollbrecht, A., Rust, S. K., and Weber, K., “Development of Microcracks in Granites During Cooling and Uplift: Examples from the Variscan Basement in NE Bavaria, Germany,” J. Struct. Geol., Vol. 13, No. 7, 1991, pp. 787–799. Kowallis, B. J. and Wang, H. F., “Microcrack Study of Granite Cores from Illinois Deep Borehole UPH-3,” J. Geophys. Res., Vol. 88, 1983, pp. 7373–7380. Schild, M., Siegesmund, S., Vollbrecht, A., and Mazurek, M., “Characterization of Granite Matrix Porosity and Pore-Space Geometry by in situ and Laboratory Methods,” Geophys. J. Int., Vol. 146, 2001, pp. 111–125. Heard, H. C., “Thermal Expansion and Inferred Permeability of Climax Quartz Monzonite to 300°C and 27.6 MPa,” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., Vol. 17, 1980, pp. 289–296. Wang, H. W. and Heard, H. C., “Prediction of Elastic Moduli via Crack Density in Pressurized and

TAKARLI AND PRINCE-AGBODJAN ON THERMAL DAMAGE IN GRANITE 13

关6兴 关7兴

关8兴

关9兴

关10兴 关11兴 关12兴 关13兴 关14兴 关15兴 关16兴

关17兴 关18兴 关19兴 关20兴 关21兴

关22兴

关23兴 关24兴

Thermally Stressed Rock,” J. Geophys. Res., Vol. 90, No. B12, 1985, pp. 342–350. Laniz, R. E., Stevens, R. V., and Norman, M. B., “Staining of Plagioclase Feldspar and Other Minerals with F.D. and C. red no. 2,” U.S. Geol. Survey Prof. Paper 501-B, 1964, pp. 152–153. Glover, P. W. J., Baud, P., Darot, M., Meredith, P. G., Boon, S. A., LeRavalec, M., Zoussi, S., and Reuschle, T., “␣/␤ Phase Transition in Quartz Monitored Using Acoustic Emissions,” Geophys. J. Int., Vol. 120, 1995, pp. 775–782. Wang, R. L., Mair, R. W., Rosen, M. S., Cory, D. G., and Walsworth, R. L., “Simultaneous Measurement of Rock Permeability and Effective Porosity Using Laser-Polarized Noble Gas NMR,” Phys. Rev. E, Vol. 70, 2004, 026312. Sousa, L., Suarez, L., Calleja, L., Ruiz, V., and Rey, A., “Influence of Microfractures and Porosity on the Physico-Mechanical Properties and Weathering of Ornamental Granites,” Eng. Geol. (Amsterdam), Vol. 77, 2005, pp. 153–168. Géraud, Y., “Variations of Connected Porosity and Inferred Permeability in a Thermally Cracked Granite,” Geophys. Res. Lett., Vol. 21, No. 11, 1994, pp. 979–982. Géraud, Y., Mazerolle, F., and Raynaud, S., “Comparison Between Connected and Overall Porosity of Thermally Stressed Granites,” J. Struct. Geol., Vol. 14, No. 8/9, 1992, pp. 981–990. ASTM Standard D4525, “Standard Test Method for Permeability of Rocks by Flowing Air,” Annual Book of ASTM Standards, ASTM International, West Conshohocken, PA, 1990. Brown, K., Fluids in Deforming Sediments, in the Geological Deformation of Sediments, A. Maltman Ed., Chapman & Hall, London, 1994, pp. 239–244. De Marsily, G., Hydrogéologie Quantitative, Masson, Paris, 1994, 219p. Skinner, B. J., “Thermal Expansion,” in Handbook of Physical Constants, S. P. Clark, Ed., Geol. Soc. Am., Boulder, CO, 1966, pp. 75–96. Brace, W. F., “Brittle Fracture of Rocks,” in State of Stress in the Earth’s Crust, Proceeding of the International Conference, Santa Monica, American Elsevier Publishing Co., New York, 1964, pp. 110–178. Eberhardt, E., Stead, D., and Stimpson, B., “Quantifying Progressive Pre-peak Brittle Fracture in Rock During Uniaxial Compression,” Int. J. Rock Mech. Min. Sci., Vol. 36, 1999, pp. 361–380. Bieniawski, Z. T., “Mechanism of Brittle Rock Fracture. Part I: Theory of the Fracture Process,” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., Vol. 4, No. 4, 1976, pp. 395–406. Eberhardt, E., Stead, D., Stimpson, B., and Read, R. S., “Identifying Crack Initiation and Propagation Thresholds in Brittle Rock,” Can. Geotech. J., Vol. 35, No. 2, 1998, pp. 222–233. Martin, C. D., “The Strength of Massive Lac du Bonnet Granite Around Underground Openings,” Ph.D. thesis, Department of Civil Engineering, University of Manitoba, Winnipeg, MB., 1993. Homand-Etienne, A. and Houpert, R., “Thermally Induced Microcracking in Granites: Characterization and Analysis,” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., Vol. 26, No. 2, 1989, pp. 125–134. Sausse, J., Jacquot, E., Fritz, B., Leroy, J., and Lespinasse, M., “Evolution of Crack Permeability During Fluidrock Interaction: Example of the Brézouard Granite 共Vosges, France兲,” Tectonophysics, Vol. 336, 2001, pp. 199–214. Lockner, D. A., “Rock Failure,” in: Rock Physics & Phase Relations, A Handbook of Physical Constants, T. J. Ahrens, Ed., 1995, American Geophysical Union, Washington, DC, pp. 127–147. Muneo, H., “Micromechanical Analysis on Deterioration Due to Freezing and Thawing in Porous Brittle Materials,” Int. J. Eng. Sci., Vol. 36, No. 4, 1998, pp. 511–522.