Construction and Building

MATERIALS

Construction and Building Materials 22 (2008) 573–579

www.elsevier.com/locate/conbuildmat

Thermal conductivity of cement composites containing rubber waste particles: Experimental study and modelling A. Benazzouk *, O. Douzane, K. Mezreb, B. Laidoudi, M. Que´neudec Laboratoire des Technologies Innovantes (EA 3899), Universite´ de Picardie Jules Verne, IUT De´partement Ge´nie Civil, Avenue des Faculte´s, 80025 Amiens, France Received 7 April 2006; accepted 19 November 2006 Available online 16 January 2007

Abstract In this paper an investigation of the thermal conductivity of a lightweight construction material containing rubber waste particles, is presented. Measurements were carried out in a dry state using a transient plane source (TPS) technique. To determine the effect of the rubber particles ratio on the thermal conductivity of a cement composite, 10%, 20%, 30%, 40% and 50% rubber particle ratios by volume as replacement to cement, were used. The experimental investigation revealed that the addition of rubber particles reduces the material unit weight, furthermore, thermal conductivity of the composite has been improved. The thermal insulating effect of rubber particles is most attractive and indicates a high and promising potential for development. Based on the self consistent method and assuming that the tri-phase composite consists of air, rubber particles and cement paste, thermal conductivity of the composite has been predicted as a function of the dry unit weight and formulations, using auto-coherent homogenisation model. The model requires the knowledge of rubber particles conductivity, which was experimentally measured using Horai and Simmons technique. A modelling application reveals good correspondence with the experimental results.  2007 Elsevier Ltd. All rights reserved. Keywords: Rubber wastes; Cement composite; Rubber particles; Thermal conductivity; Modelling; Self consistent technique

1. Introduction The reduction of energy consumption in construction, production of thermal insulation materials and the solution of environment problem by recycling of industrial, agriculture waste, and domestic waste are becoming greater problem. They are many lightweight composites that contain recycled fillers, including waste glass [1], fly ash [2], kraft pulps from sisal and banana waste [3], steel slag [4], lightweight crushed bricks, lightweight expanded clay aggregates [5], foam polystyrene and its waste [6]. Therefore, the development of composite construction materials with low thermal conductivity using these wastes will be an interesting alternative that would solve simultaneously energy and environment concerns.

*

Corresponding author. Tel./fax: +33 3 22 53 40 16. E-mail address: [email protected] (A. Benazzouk).

0950-0618/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2006.11.011

The amount of used tires produced in France is roughly 30,000,000 per year [7]. Landfill disposal which is the most prevailing method, will be drastically reduced in the near future, due to the recent introduction of European Union directives that include significant restrictions on this practice in favour of alternatives oriented toward materials and energy recovery. Furthermore, the disposal of used tires in landfills, stockpiles, or illegal dumping grounds, increases the risk of accidental fires with uncontrolled emissions of potentially harmful compounds. Innovative solution has to be developed to meet the challenge of tire disposal problem. Highway construction provides a significant market potential for waste tires recycling. Extensive studies have been conducted on waste tire modified Portland cement concrete [8–10]. The literature about the use of tire rubber particles in cement-based materials focuses on the use of tire rubber as an aggregate in concrete and evaluates only the mechanical properties. Results have indicated that

574

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579

Nomenclature E R T t

potential (V) resistance (X) temperature (C) time (s)

Greek symbols a temperature coefficient of the TPS sensor

rubberized concrete mixtures posses lower density, increased toughness and ductility, higher impact resistance, lower compressive and splitting tensile strengths, and more efficient sound insulation. However, some authors suggested that the loss in strength might be minimized by prior surface treatment of the tire rubber particles [11]. Previous study focused on the use of two types of rubber aggregates, as addition to cement paste, in order to develop a highly deformable material [12]. The types of the rubber aggregates were: compact rubber aggregates (CRA) and expanded rubber aggregates (ERA). Results have revealed the influence of rubber aggregate type on the material mechanical properties. A study of this composite has also demonstrated the importance of rubber particle type with respect to the hydraulic transport properties of the composite when coming into contact with water [13]. A recent study investigated the effect of powdered tire rubber as addition to cement paste on both the physico-mechanical and water absorption properties of the cement composite [14]. Results have indicated that, although the strength was reduced, the composite satisfies the basic requirement of construction materials, and could be used for load-bearing wall. In addition, the incorporation of rubber particles in cementitious matrix tends to restrict water absorption of the composite; sorptivity values decreases with increasing rubber particles ratio. Although several work has been done on the mechanical properties of composite containing rubber aggregates, no work has been previously reported on the thermal conductivity. In this work, the idea is to use rubber waste particles, as a raw material, to develop a lightweight construction materials with lower thermal conductivity so as to reduce heat transfer into building in order to decrease the energy consumption. An experimental test program was conducted mainly to investigate the effect of rubber particles addition on the thermal conductivity of composite, in dry state, using transient plane source (TPS) technique. The composite material was manufactured by reinforcing varying volume fraction of rubber particles in cementitious matrix. Assuming that the composite consists of three phases (air, rubber particles, cement paste), thermal conductivity has been predicted using auto-coherent homogenisation model [15]. This model requires the knowledge of rubber particles conductivity, which was experimentally measured

c k q h d e

constant depending on the TPS resistance thermal conductivity (W/mK) dry unit weight (kg/m3) volume fraction of rubber particles (%) volume fraction of cement (%) volume fraction of liquid (%)

using TPS methos, according to the Horai and Simmons technique [16]. 2. Materials and experimental testing 2.1. Materials Rubber particles used in this study has been obtained from mechanical shredding of rubber automotive industry waste. This waste comprises rubber particles of less than 1 mm in size and contains approximately 20% synthetic fibers by volume as well. The absolute density of this rubber waste particles is 430 kg/m3. The cement used was CPJ CEM II 32.5 in accordance with Standard NF P 15-301 [17]. Both the rubber particles and cement were initially dry-mixed in a laboratory mixer. The volume ratio of rubber ranged from 0% to 50% by volume as replacement to cement in mixtures. Total mixing water had been adjusted so as to achieve constant workability for all composites (i.e. a slump on the order of 90–100 mm). For each mixture, two cubic samples of 100 · 100 · 100 mm were prepared and moist-cured for 28 days at 20 ± 2 C and 98% relative humidity both before and after demolding. For thermal conductivity measurement in dry state, all the specimens were dried in a drying oven at 50 ± 2 C and weighed at 24-h intervals until the loss in weight did not exceed 1% in a 24-h. Then the surfaces of all the samples were polished to achieve smooth surfaces in order to maintain the proper contact between the TPS sensor and the specimen. It should be noted that the size of the sample must satisfy the condition of an infinite medium, that is the ‘‘probing depth’’, which indicates how far the heating pulse has propagated into the sample during the transient time, is less than the distance from the heater to the nearest boundary of the sample [18]. The physico-mechanical properties of this composite for various rubber particle volume ratios are listed in Table 1 [14]. 2.2. Test procedure Thermal conductivity measurement of the samples, at dry state, have been made at room temperature and normal

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579

575

Table 1 Properties of cement composite containing different rubber volume ratios Volume ratio of rubber particles (%)

Air-content (%)

Dry unit weight (kg/m3)

Elasticity dynamic modulus (GPa)

Compressive strength (MPa)

Flexural strength (MPa)

0 10 20 30 40 50

2.0 5.0 8.7 11.8 14.0 17.0

1910 1740 1620 1473 1300 1150

25.0 20.0 14.0 13.0 11.5 9.5

82.0 49.5 40.0 23.3 16.0 10.5

3.4 3.8 4.2 4.0 3.8 3.2

pressure using transient plane source (TPS) method, which is the subject of considerable research [19,20]. The experimental set-up is shown in Fig. 1. The technique is based on three-dimensional heat flow inside the sample, which can be regarded as an infinite medium by limiting the total time of transient recording. A disk-shaped TPS-element was placed between two cubic samples. In order to ensure a good thermal contact between the TPS sensor and the sample material, a chucking device was used. For measurement, the sample pieces containing TPS sensor were then introduced into the drying oven in order to control the experimental temperature. In addition, a thermocouple was kept just above the sample pieces to monitor the temperature of the sample. After achieving the isothermal conditions in the sample, a constant current pulse is passed thought the heating element. The temperature of the element is recorded simultaneously by recording its voltage increase. The duration of the experiment is about 360 s. The thermal conductivity was measured from four faces of each composite and calculated as their average value. In the TPS technique, the source of heat is a hot disc made out of a bifilar spiral, which also serves as a sensor of the temperature increase in the samples. In comparison with stationary or steady state methods, the advantage of transient methods is that some of them give a full set of thermophysical parameters within a single rapidly measurement, namely thermal conductivity, specific heat or

thermal diffusivity. It can measure solids and liquids with thermal conductivities ranging from 0.02 to 200 W/mK. Assuming that the conducting pattern is in y–z plane of a coordinate system placed inside an infinite solid, the timedependent resistance of TPS-element during the transient recording can be written as RðtÞ ¼ R0 ð1 þ aDT ðsÞÞ

ð1Þ

where R(t) is the resistance of the TPS sensor at time t,R0 is the resistance of the TPS sensor at time zero, a is the temperature coefficient of resistivity (TCR), DT(s) is the mean value of temperature rise in the TPS-element. Recording the change in potential difference DE(t) across the TPS sensor, the thermal conductivity has been calculated with a developed Matlab program using the following relation: k ¼ cDðsi Þ=DEðtÞ

ð2Þ

where c is the constant depending of the different resistances in the Wheatstone bridge, D(si) is the theoretical expression of time-dependent increase. The thermal conductivity of rubber particles was measured according to the Horai and Simmons technique [16]. This technique has already been used to estimate the thermal conductivity of solid mix phases within several types of composites [21]. In this method, the solid particles are mixed with a liquid, and then the thermal conductivity

Chucking device

TPS sensor Sample Sample

Computer

Climatic chamber

Power supply

Resistances

Power data acquisition

Fig. 1. Experimental TPS set-up.

576

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579

of this saturated mixture is measured using TPS technique. The thermal conductivity of the solid particles is thus evaluated using the arithmetic mean: kU þ kL ð3Þ 2 In accordance with Hashin and Strickman bounds [16], kU and kL can be written for two phases as follows: km ¼

kU ¼ kP þ

1 kl kP

e þ 1e 3kP

and

kL ¼ kl þ

1e þ 3ke l

ð4Þ

1 kP kl

where kP, km and kl are respectively the thermal conductivity of solid particles, to be determined, saturated mixture and liquid. e is the volume fraction of liquid. For the experimental determination, a plastic cell with a volume of approximately 450 cm3 was used. The rubber particles was mixed with glycerol and the TPS sensor was positioned between two layers of saturated mixes. In this manner, the thermal contact between the mixture and TPS-element is good. However, great care had to be taken to ensure that the TPS sensor remained horizontal. In order to minimize measurement errors, two glycerol-based rubber particle mixes containing various glycerol volume contents were prepared. For calculations, experimental thermal conductivity of pure glycerol liquid was determined. The thermal conductivity of rubber particles has been evaluated by substituting the conductivities of the mixture and glycerol in expression (3). For each material, the thermal conductivity was measured three times and the mean values are reported. 3. Experimental results and analysis

1.6

1.6

Thermal conductivity (W/mK)

Thermal conductivity (W/mK)

The experimental thermal conductivity values of the composite measured in dry state are listed in Table 1. The variation with respect to rubber particle volume content, is displayed in Fig. 2. It has been observed that the addition of rubber particles into the cement matrix reduces the thermal conductivity of the composite. Values decrease from 1.16, for the cement paste, to 0.47 W/mK for a specimen containing 50% rubber particles. This cor-

1.2 0.8 0.4 0

responds to a decrease of about 60%. The reduction of thermal conductivity of composite is due to the insulating effect of rubber particle, which has a lower thermal conductivity compared to that of cement matrix. It is evident from these results that the aggregates with less thermal conductivity produced the less conductive composite. Earlier investigations reveal that the type of aggregate would greatly influence the thermal conductivity of materials [21]. The corresponding relationship between thermal conductivity and dry unit weight of the composite is shown in Fig. 3. The thermal conductivity k (W/mK) decreases with decreasing unit weight q (kg/m3). The derived correlation is of the type: k = 0.1236 exp(0.0011q) (which yields a correlation coefficient of R2 = 0.98). The variation obtained is similar to that reported in previous work conducted on lightweight concretes [22,23]. The decrease in thermal conductivity is also related to air content in the matrix that results in less unit weight. The more the air voids ratio, the lighter the specimen and the lower its thermal conductivity. Table 1, which provides a list of air content values, measured using the pressure method (ASTM C457), indicates that the increasing of rubber volume ratio results in higher air contents, thereby decreasing the unit weight of the composite. Similar observations were also made by several authors [9,10]. This may due to the non-polar nature of rubber particles and their tendency to entrap air in their rough surfaces. Also when rubber is added to a mixture, it may attract air as it has the tendency to repel water, and then air may adhere to the rubber particles. The increases in air content with increasing rubber particles reduce the thermal bridges in the matrix and contribute to improving composite insulation. Fig. 4 shows an optical microscopy image of the matrix containing 50% rubber particles. In terms of bonding, we can observe a good cohesion between rubber particles and cement matrix, included a spherical cavity of air voids. The thermal conductivity of materials depends upon many factors, including their structure, material mixture proportioning, type of aggregate inclusions, density, porosity, etc. The thermal insulating performance of the composite containing rubber particles is also related to the porosity

0

10 20 30 40 Volume ratio of rubber particles (%)

50

Fig. 2. Variation of thermal conductivity of the cement composite with volume ratio of rubber particles.

0.0011

λ = 0.1236e R2 = 0.98 1.2 0.8 0.4

0 1000

1200

1400

1600

1800

2000

3

Dry unit weight (kg/m ) Fig. 3. Relationship between thermal conductivity and dry unit wight of cement composite.

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579

577

λ eq , Req λ 2 , R2 λ 1 , R1

λ 3 , R3

Fig. 4. Optical microscopy photo of the cement matrix containing rubber particles (magnification: 50·).

which plays an important part in heat transfer. The thermal conductivity of the composite containing 50% rubber particles is lower than that of expanded shale-based load-bearing insulated concrete with a thermal conductivity of 0.70 W/mK and a unit weight ranging between 1100 and 1300 kg/m3 [23]. The thermal insulating effect of rubber particles is most attractive and indicates a high and promising potential for development. 4. Modelling by auto-coherent homogenisation It is based on the self consistent method that leads to the characterisation of heterogeneous material from the characteristics and the concentration of each constituents. Developed for the mechanical characterisation (elasticity and viscoplasticity) of heterogeneous materials, il has been used for the characterisation of thermal conductivity of cellular concrete (in the case of bi-composite) [24] and then extended to tri-composite mediums (hemp concretes) [15]. This method is based on the following energy condition: the energy in the heterogeneous medium is the same to that in the equivalent homogeneous medium submitted to the same boundary conditions. The homogeneous material is assimilated to an assembly of spherical composite inclusions of variable sizes. The heterogeneous material could be constituted of either two or three components. The tri-composite model is applied to the equivalent homogeneous material with properties of keq (thermal con-

Fig. 5. Principal of self consistent scheme of a tri-composite cell.

ductivity) and qeq (unit weight) by assuming an internal sphere of air (k1, q1), surrounded by a spherical cell of rubber particles (k2, q2), and moreover surrounded by a shell of cement matrix (k3, q3) (see Fig. 5). Equivalent thermal conductivity of the composite is given as keq ¼1þ k3

h
h

13 þ

1þd3 k1 1d3 k3







k1 1 k2



k1 1 k2

ð5Þ 

2k2 þ1 k3

h and d are the volume fractions of rubber particles and cement, respectively. This approach requires knowledge of volume fractions and thermal conductivities of rubber particles and cement paste matrix. The volumic fractions are directly calculated from the mass of each component (of known unit weight). The thermal conductivity of rubber particles has been experimentally measured according to the method previously described. The corresponding results of different mixtures and glycerol are shown in Table 2. It should be noted that the average value obtained for glycerol is in agreement with the earlier findings [25,26]. The average thermal conductivity value of rubber particles, as calculated from expression (3), is equal to 0.19 W/mK, within ±5%. Thermal conductivity of rubber particles is over six times lower than that of the cement paste matrix (1.16 W/mK). The calculated values of thermal conductivity of composite, as obtained using auto-coherent model, are plotted

Table 2 Experimental results of thermal conductivity of different mixtures and glycerol Materials

Glycerol Rubber particles–glycerol mixture (40%) Rubber particles–glycerol mixture (60%) a,b

Thermal conductivity (W/mK) Measured-value

Theoretical-value

0.27 0.231 0.252

0.28a–0.29b – –

Thermal conductivity of the glycerol at 20 C [25] and 25 C [26], respectively. Measured thermal conductivity of mixture at 40% and 60% volume ratio of glycerol, respectively.

1, 2

578

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579

approach can furnish additional information regarding the nature of the pores and its distribution. The study of the thermal properties of this composite at different moisture contents is in progress.

Calculated values (W/mK)

1.4 1.2 1 0.8

References

0.6 0.4 0.2 0.2

0.4

0.6

0.8

1

1.2

1.4

Experimental values (W/mK)

Fig. 6. Comparison of the measured and calculated values of thermal conductivity of cement composite containing different rubber volume ratios.

against their corresponding experimentally obtained values in Fig. 6. It can be observed that the model overpredicts thermal conductivity of composite by about 3–14%, depending on the rubber particles ratio. The model calculation considered the shape of the aggregates to be spherical, which however is not the case for our composites. Moreover, the rubber inclusions are not composed solely of particles; they also contain a proportion of rubber fibers resulting from shredding, as well as textile fibers with a random orientation in the matrix. These fibers, of various chemical compositions, may exhibit a different thermal conductivity from that of the rubber particles. Nevertheless, the model for tri-composite approach is sufficiently accurate for predicting the influence of rubber particles on the thermal conductivity of cement composite. 5. Conclusion The work presented herein has focused on the effect of including rubber particles on the thermal insulating performance of cementitious composite. Results indicate a clear reduction in thermal conductivity of the composite with the addition of rubber particles. The properties of the composite, with 50% rubber particles, are as follows: thermal conductivity of 0.47 W/mK, compressive strength of 10.50 MPa, flexural strength of 3.25 MPa and dry unit weight of 1150 kg/m3. Thus, the potential for development, therefore, seems to be very promising. The material inclusion could provide multiple applications in particular to prevent heat transfer and to save energy for use in structure with low load-bearing capacities. Assuming that the material consists of a tri-composite mediums, the thermal conductivity was estimated using the auto-coherent model. Based on thermal conductivity measurement of rubber particles, the results obtained correspond to the values derived through experiment. The model overpredicts thermal conductivity of composite by about 3–14%, depending on the rubber particles ratio. However, from fundamental principal of heat transfer in porous media such as cementitious materials, porous media

[1] Duman V, Mladenovic A, Suput JS. Lightweight aggregate based on waste glass and its alkali-silica reactivity. Cement Concrete Res 2002;32:223–6. [2] Kearsley EP, Wainwright PJ. The effect of high fly ash content on the compressive strength of foamed concrete. Cement Concrete Res 2001;31:105–12. [3] Savastano Jr H, Warden PG, Couts RSP. Brazilian waste fibres as reinforcement for cement-based composite. Cement Concrete Comp 2000;22(5):379–84. [4] Yue YL, Li GZ, Xu XS, Zhao ZJ. Properties and microstructures of plan-fiber-reinforced cement-based composites. Cement Concrete Res 2000;30(5):1983–6. [5] Alduaij J, Alshalch K, Haque MN, Ellaythy K. Lightweight concrete in hot coastal areas. Cement Concrete Comp 1999;21(5–6):453–8. [6] Laukaitis A, Zurauskas R, Keriene J. The effect of foam polystyrene granules on cement composite properties. Cement Concrete Comp 2005;27:41–7. [7] Info-De´chets. Quelle filie`re globale pour la valorisation de pneus. Environ. Tech. 1998;181(Novembre):32–5. [8] Eldin NN, Senouci AB. Rubber tires particles as concrete aggregate. J Mater Civi Eng 1993;5(4):478–96. [9] Topc¸u IB. The properties of rubberized concrete. Cement Concrete Res 1995;25(2):304–10. [10] Fedroff D, Ahmed S, Savas DZ. Mechanical properties of concrete with ground waste tire rubber. Transport Res Rec 1996;1532:66–72. [11] Serge N, Joekes I. Use of tire particles as addition to cement paste. Cement Concrete Res 2000;30:1421–5. [12] Benazzouk A, Mezreb K, Doyen G, Goullieux A, Que´neudec M. Effect of rubber aggregates on the physico-mechanical behaviour of cement–rubber composites-influence of the alveolar texture of rubber aggregates. Cement Concrete Comp 2003;25:711–20. [13] Benazzouk A, Douzane O, Que´neudec M. Transport of fluids in cement–rubber composites. Cement Concrete Comp 2004;26:21–9. [14] Benazzouk A, Mezreb K, Douzane O, Que´neudec M. Physicomechanical properties and water-absorption of cement–rubber composites containing shredding rubber waste. Cement Concrete Comp, submitted for publication. [15] Arnaud L, Monnet H, Cordier C, Sallet F. Mode´lisation par homoge´ne´isation autocohe´rente de la conductivite´ thermique de be´ton de laine et de chanvre. In: Lallemand A, Leone JF, editors. Proceeding of the congre`s franc¸ais de thermique. Elsevier; 2005. [16] Horai K. Thermal conductivity of rock-forming minerals. J Geophy Res 1971;76(5):1278–308. [17] Norme AFNOR. Liants hydrauliques, ciments courants-compositions, spe´cification et crite`res de conformite´, 1994. [18] Gustafsson SE. Transient plane source techniques for thermal conductivity and thermal diffusivity measurements of solid materials. Rev Sci Instrum 1991;62(3):777–804. [19] Saxena NS, Pradeep P, Mathew G, Thomas S, Gustafsson M, Gustafsson SE. Thermal conductivity of styrene butadiene rubber compounds with natural rubber prophylactics waste as filler. Eur Polym J 1999;35:1687–93. [20] Mangal R, Saxena NS, Sreekala MS, Thomas S, Singh K. Thermal properties of pineapple leaf fiber reinforced composite. Mat Sci Eng A 2003;339:281–5. [21] Bouguerra A, Laurent JP, Goual MS, Queneudec M. The measurement of the thermal conductivity of solid aggregates using the transient plane source technique. J Appl Phys 1997;30:2900–4. [22] Al-Rim K, Ledhem A, Douzane O, Dheilly RM, Que´neudec M. Influence of the proportion of wood on the thermal and mechanical

A. Benazzouk et al. / Construction and Building Materials 22 (2008) 573–579 performances of clay–cement–wood composites. Cement Concrete Comp 1999;21:269–76. [23] Demurboga R, Gu¨l R. The effects of expanded perlite aggregate, silica fume and fly ash on the thermal conductivity of lightweight concrete. Cement Concrete Res 2003;33:723–7. [24] Boutin C. Conductivite´ thermique du be´ton cellulaire: mode´lisation par me´thode auto-cohe´rente. Mate´r Constr 1996;29:609–15.

579

[25] Chen Q, Li Y, Longtin JP. Real-time laser-based measurement of interface temperature during droplet impingement on a cold surface. Int J Heat Mass Transf 2003;46:879–88. [26] Sun J, Longtin JP, Irvine Jr TF. Laser-based thermal pulse measurement of liquid thermophysical properties. Int J Heat Mass Transf 2001;44:645–57.