Available online at www.sciencedirect.com

Construction and Building

MATERIALS

Construction and Building Materials 22 (2008) 668–674

www.elsevier.com/locate/conbuildmat

Evaluation of mechanical, physical and thermal performance of cement-based tiles reinforced with vegetable fibers Luiz C. Roma Jr., Luciane S. Martello, Holmer Savastano Jr.

*

Faculdade de Zootecnia e Engenharia de Alimentos, Universidade de Sa˜o Paulo (USP), Avenida Duque de Caxias Norte, 225, 13635-900 Pirassununga, Sa˜o Paulo, Brazil Received 10 March 2006; received in revised form 13 September 2006; accepted 14 October 2006 Available online 28 November 2006

Abstract The objective of this work was to analyze mechanical, physical and thermal performance of roofing tiles produced with several formulations of cement-based matrices reinforced with sisal and eucalyptus fibers. The physical properties of the tiles were more influenced by the fiber content of the composite than by the type of reinforcement. The type of the fiber was the main variable for the achievement of the best results of mechanical properties. Exposure to tropical climate has caused a severe reduction in the mechanical properties of the composites. After approximately four months of age under external weathering the toughness of the vegetable fiber–cement fell to 53– 68% of the initial toughness at 28 days of age. The thermal performance showed that roofing tiles reinforced with vegetable fiber are acceptable as substitutes of asbestos–cement sheets.  2006 Elsevier Ltd. All rights reserved. Keywords: Fiber–cement; Roofing tiles; Sisal fiber; Eucalyptus fiber; Thermal performance; Physical properties; Mechanical behavior

1. Introduction The design of a durable and low-cost fiber–cement for roofing is a technological challenge in developing countries. As increasing concerns are being associated to chrysotile fibers, new researches are now expected for the adaptation of available raw-materials and production systems to fit the consumer requirements at each particular application area [1–3]. Lee [4] showed that the 4 mm thick asbestos–cement corrugated sheet is the cheapest alternative for roofing and that it seems to be the main reason for its utilization in low-cost housing in countries as Brazil. However, the thermal insulation of this material is considered worse than the ceramic tile ones. The need of a low-cost alternative could justify the interest on roofing tiles produced out of recycled raw materials which show a lifelong acceptable performance. *

Corresponding author. Tel.: +55 19 3565 4153; fax: +55 19 3565 4114. E-mail address: [email protected] (H. Savastano Jr.).

0950-0618/$ - see front matter  2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2006.10.001

In tropical countries the wastes generated during the production of vegetable fibers can represent an important source of raw materials for the production of building components. The success of this application depends on the available amount of residues and their geographical dispersion related to the costs with transportation and local processing [5]. Guimara˜es [6], Agopyan [7] and Savastano Jr. [8] have reported several experiences about the use of cement-based matrices reinforced with natural fibers for the production of building elements as roofing tiles, wall panels and water tanks. In several countries these non-conventional fiber– cements are already included in programs of technological transference especially for cost-effective roofing systems as indicated by Saxena et al. [9], Gram and Gut [10] and Delvasto et al. [11]. The thermal performance of the roofs has been studied to mitigate the negative effects of solar radiation inside the houses in tropical countries. The reflectance of the tiles is one of the ways for measuring their thermal performance. The reflectance is the probability of a photo to be

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674

reflected on a surface. The spectrum of the solar radiation is divided in three distinct regions. The ultraviolet region (UV) can cause physical–chemical effects on materials as well as on living beings such as damages to cell structure. The visible region (V) comprises approximately 46% of the entire solar spectrum and can excite the human eye. The last one, the infrared radiation (IR) corresponds approximately to 43% of the solar spectrum and the nature of these radiations is mainly thermal [12]. The objective of this work was to analyze mechanical, physical and thermal performance of roofing tiles based on cement matrix reinforced with sisal and/or eucalyptus fibers in different compositions. The aging of the composite was evaluated by the exposition of the produced tiles to tropical weather for approximately four months during summer time. 2. Experimental study The present study was conducted in experimental scale in a tropical area of Sao Paulo State, Brazil during a typical summer season characterized by hot and wet climate. 2.1. Roofing tiles production More than 1500 roofing tiles of fiber–cement were fabricated with a matrix composed by Portland cement (PC) with pozzolanic addition CPII 32-Z type, ground blast furnace slag (GBFS) and silica fume (SF) reinforced with cellulose fibers of eucalyptus (Eucalyptus grandis) and/or sisal (Agave sisalana). The CPII 32-Z composite cement follows the Brazilian Standard NBR 11578 [13]. That type of Portland cement contains from 6% to 14% of pozzolanic material and up to 10% of finely ground limestone. Both fibers were used in the form of cellulose kraft pulps. The eucalypTable 1 General properties of the silica fume and cellulose fiber Property

Silica fume

Average diameter (lm)a Specific surface (m2/g)a Pozzolanic activity (mg/g)a Real density (g/cm3)a Phasea Average length (mm)b,c Thickness (lm)b,c Aspect ratiob,d Coarseness (mg/m)b

0.5

a b c d

Eucalyptus (Eucalytpus grandis)

Sisal (Agave sisalana)

0.66

1.66

10.9 61 0.107

13.5 123 0.163

tus fiber was a residue obtained from a papermaking plant. The chemical and physical characteristics of SF and fibers are shown in Table 1 and the chemical composition of GBFS was pointed by Oliveira [14]. The formulation of the matrix was fixed for the production of all the tiles (by mass composition): 59% CPII 32Z, 36% GBFS, 5% SF. The proportion of the superplasticizer admixture was fixed as 3% of the total mass of the solid content in the matrix. The water/binder ratio varied from 0.55 to 0.65 for obtaining a consistency index [15] between 210 and 250 mm for the fresh composite that is an acceptable range to facilitate the molding procedure. The different fiber contents (% of fiber by mass of solid components) used for the composite preparation were: 5% of eucalyptus fiber for E treatment, 2% of eucalyptus fiber and 1% of sisal fiber for ES treatment, and 3% of sisal fiber for S treatment. The following order of mixture of the raw material was adopted to obtain the optimum homogeneity: (i) pre-dispersion of SF in the entire amount of water during 60 min in a heavy-duty dough mixer; (ii) mixture of Portland cement and GBFS in a concrete mixer of vertical axis during 5 min at low velocity; (iii) addition of the slurry prepared in first step and further mixture for 10 min at low velocity and (iv) introduction of wet fibers and mixing for 5 min at high velocity. The tile production used the Parry Associates (UK) equipment, for molding and compaction by vibration. The dimensions of the tiles are 487 · 263 · 6 mm (frame measures) and in a very similar shape to the ceramic Roman tiles [11,16]. After approximately 24 h the tiles were removed from the molds and transferred to undergo a saturation cure by immersion in water during seven days. For the series tested with 28 days of age, the tiles were then kept in laboratory environment prior to mechanical and physical characterization. In the case of series tested with 155 days of total age, the tiles were transferred to a roofing system immediately after the saturation cure period. This roof is illustrated in Fig. 1 and was employed in an animal housing during 148 days. The weathered tiles were then tested in the same conditions of the non-aged ones. The weathering period was between

22.5 813.83 2.65 Amorphous

Microstructure Laboratory, Polytechnic School, USP. Savastano Jr, H. [8]. Standard error equivalent to 50% of the mean values. Average length/thickness.

669

Fig. 1. Roofing system with vegetable fiber–cement tiles.

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674

December 2002 and March 2003 during summer time in a rural area located in Pirassununga, Brazil (21 59 0 S of latitude and 627 m of altitude). In this four-month period, the total rainfall was measured at 563 mm, the maximum and minimum temperatures were of 35.4 C and 17.0 C, respectively, and average relative humidity of 87%. All the tiles were saturated in water 24 h prior to mechanical tests.

1.60 1.40 Bulk density (g/cm³)

670

2.2. Roofing tiles characterization

3. Results and discussion Water absorption, bulk density and permeable void volume are related to each other as depicted by Figs. 2–4. The

1.00 28 days 155 days

0.80 0.60 0.40 0.20

Six specimens were used for the three treatments for the determination of physical and mechanical properties in both ages under consideration.

0.00

E

ES Treatments

S

Fig. 2. Bulk density of the tiles for different treatments at 28 and 155 days of age.

W ate r ab so rp tion (% b y mass )

45 40 35 30 25

28 days 155 days

20 15 10 5 0

E

ES Treatments

S

Fig. 3. Water absorption of the tiles for different treatments at 28 and 155 days of age.

52 P er mea b l e v oi d v o lu m e ( % by m as s)

 Physical properties: Water absorption, bulk density and permeable void volume values were obtained from tested flexural specimens following the procedures specified in ASTM C 948-81 [17].  Mechanical properties: A three-point bend configuration was employed in the determination of maximum load and fracture energy. A span of 350 mm, corresponding to a span to depth ratio of approximately 40, and a deflection rate of 0.5 mm/min were used for all tests on an Emic model DL30000 universal testing machine equipped with load cell of 5 kN. Fracture energy was obtained by integration of the load–deflection curve to the point corresponding to a reduction in load carrying capacity to 50% of the maximum observed. For the purpose of this paper, ‘fracture toughness’ was measured as the fracture energy divided by specimen width and depth at the failure location. The mechanical test procedures are described in greater detail elsewhere [8].  Thermal performance: The temperature values of the downside surface of the roofing tiles were collected each 15 min, 24 h a day during the 23 hottest days of the summer period. The collection of these data was possible using an Onset HOBO model TMC1 – HA sensor with temperature range between 40 C and +100 C. The experiment compared the roofing tiles produced with sisal (S formulation) with asbestos–fiber corrugated sheets commercially available in the Brazilian market. Both types of tiles were installed in the roofing system (Fig. 1) and the thermal evaluation was coincident with the period of natural weathering.  Reflectance: This measurement was made using a spectrophotometer, which allows the evaluation of the results by the range of wavelength. This evaluation considered tiles of different materials: cement matrix with sisal fiber, asbestos–cement corrugated sheets, red ceramic tiles and corrugated sheets of galvanized steel with zinc coating.

1.20

50 48 46

28 days 155 days

44 42 40 38 E

ES

S

Treatments

Fig. 4. Permeable void volume of different treatments (28 and 155 days of age).

best results of bulk density (Fig. 2) at 28 days of age were reached by the ES and S treatments, whose values were 1.35 and 1.36 g/cm3, respectively. Such values were significantly higher (p < 0.01) than the correspondent result related to the E treatment that was equal to 1.17 g/cm3.

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674

Table 2 Values of reflectance for the ranges of solar radiation Material

Asbestos–cement Red ceramic Sisal fiber-cement Steel with zinc coating a b

Ultraviolet region. Infrared radiation.

Range UVa (%)

Visible (%)

IRb (%)

Total (%)

49.2 13.1 39.1 94.1

43.4 30.2 31.4 49.4

45.9 80.4 30.8 70.2

44.1 67.9 30.1 64.6

1200 1000 M a x i m u m l oa d ( N )

For the series tested with 155 days of age, the results of bulk density showed a similar behavior. The bulk densities of ES and S treatments (1.43 and 1.39 g/cm3, respectively) did not differ from each other and were superior (p < 0.01) to the density of E treatment (1.33 g/cm3). There was a significant increase of the bulk density just for E treatment what can be explained by the continued hydration and carbonation of the cement matrix. The high porosity of this composite seemed to facilitate the fast carbonation of the free lime available as a hydration product [18]. The lower results of water absorption (WA) at 28 days were related to the ES and S treatments (32.7 and 31.0% by mass, respectively) (Fig. 3). For the E treatment the WA reached 42.7%, which was considerably higher (p < 0.01) than the other results and would be an undesirable aspect for roofing applications. This can be understood by the larger fiber content of the E treatment (5% by mass) in comparison to the others (3% by mass). Additionally the number of filaments of eucalyptus fiber is expected to be higher due to the short average length of this hydrophilic fiber (Table 1), what also contributes to the higher WA by capillarity of the E treatment. After 155 days of total age the E formulation presented a WA value of 37.1% what is slightly over the limit of 37% by mass as recommended by the Brazilian specification ABNT NBR 7581 [19] for fiber–cement corrugated sheets. This result was worse than those of the ES (33.4%) and S (32.1%) formulations. The aging effect did not influence the WA for the ES and S treatments which maintained the similar values in both ages (28 and 155 days). Fig. 4 shows, that the highest value of permeable void volume registered to the E treatment (50.3%) was significantly different (p < 0.01) from those corresponding to ES (44.0%) and S (42.3%) treatments at 28 days. For 155 days of age the E treatment also presented the worse results of permeable void volume (49.5%) but at this time it was not significantly different (p < 0.01) from the correspondent value of the ES treatment (47.7%). The permeable void volume for the S treatment was equal to 44.8% and it did not differ from the ES treatment. The aging effect was not significant (p < 0.01) for none of the formulations. In the period from 28 to 155 days, a tendency of increase in the permeable void volume was observed for the ES and S treatments. It could be a consequence of the fast degradation of sisal fibers in the alkaline environment of Portland cement [20] or even due to the detachment of the cellulose

671

800 28 days 155 days

600 400 200 0 E

ES Treatments

S

Fig. 5. Maximum load for different treatments and ages.

fiber during wet–dry cycles attributed to its shrinkage into the cement matrix [21]. At 28 days of age the results of maximum load reached in the bending test were 533 N for E treatment, 761 N for ES and 1111 N for S series, respectively (Fig. 5). The thickness of the tiles was collected immediately after the bending test from the rupture section by three determinations per tile using calipers. The average values for the treatments were: 8 mm thick tiles for E treatment, 9 mm for ES treatment and 10 mm for S treatment. The best result (p < 0.01) was obtained for the formulation with 3% by mass of sisal fiber (S). These results indicate better capacity of reinforcement of the sisal fiber especially due to its higher ratio aspect (Table 1) as previously stated by Coutts [22]. All the results are in the acceptable range of load in accordance with Gram and Gut [10]. This specified range of load capacity is related to the thickness of the saturated tiles as follows: 425 N (8 mm thick tile), 553 N (9 mm) and 680 N (10 mm). At 155 days the treatments showed the following results: 297 N for the E treatment, 348 N for ES and 640 N for S, all of them under the limit proposed by Gram and Gut [10]. The higher result (p < 0.01) was again performed by the formulation reinforced with sisal fiber (S). All the treatments showed a drastic reduction (p < 0.01) of up to 54% of the maximum load in a period of only 148 days under natural weathering. Even with the high level of substitution of Portland cement by slag and pozzolanic additions in the matrix formulation, the degradation of vegetable fibers was very fast due to the alkaline environment combined with the hot and wet weather [21,23]. Toleˆdo Filho et al. [24] used a similar process for the fiber–cement production and they encountered considerable degradation of sisal fiber as reinforcement of ordinary Portland cement matrix with 40% by mass of GBFS replacement. The results of toughness (J/m2) are depicted in Fig. 6 and followed the same behavior pattern of the maximum load. The treatment with sisal reinforcement (S) performed the best result (p < 0.01) in both ages of 28 and 155 days (1152 and 612 J/m2, respectively) compared with the ES treatment (464 and 316 J/m2 at 28 and 155 days) and with

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674

In the analysis of thermal behavior the results of temperature on the downside surface of the roofing were divided in two periods, the first period between 5 pm and 7 am and the second one from 8 am to 4 pm. In the first period, both types of tiles had a similar performance without any significant differences (p < 0.01) in the temperature of the downside surface. In the second period, the hottest of the day, the S treatment with tiles reinforced by sisal fibers presented lower (p < 0.01) values of surface temperature in comparison with the asbestos–cement corrugated sheets. The difference between the average temperatures of both treatments reached 11.5 C at 12 pm (Fig. 7). The peak of the maximum temperature presented a delay of approximately 1 h when comparing to the curves of the surface temperatures for both tiles. Additionally the tiles of sisal fiber–cement performed considerably lower amplitude of temperature during that critical day. The best thermal performance of the S treatment during the day must be related to porosity and the water absorption results, since the humidity changes in sisal fiber– cement tiles are higher when compared to the asbestos– cement sheets ones. The water absorption for the corrugated asbestos–cement sheet used in this work was 15.5% by mass. Part of solar radiation absorbed during the day can be used for evaporating the water inside the pores and this possibly contributed for lowing the temperatures during the day. Bueno et al. [27] observed that moisture absorbing tiles (associated to higher porosity) had a better thermal performance during the day than less permeable ones under summer conditions. The values of reflectance are shown in Table 2. In the ultraviolet range, red ceramic showed low reflectance with 13.1%, while corrugated sheets of galvanized steel with zinc coating showed the best performance of 94.1%. The low reflectance on the ultraviolet region can contribute to a higher degradation of the material, since the absorbance of UV can cause alterations in the atomic structure of some materials as is the case of polymers in general [28]. The results of reflectance in the UV region for sisal fiber– cement (39.1%) and asbestos–cement (49.2%) indicated the favorable performance of these materials, showing that

1400

1000 800

28 days 155 days

600 400 200 0 ES Treatments

S

Fig. 6. Toughness for different treatments and ages.

0 22 :0

20 :0 0

18 :0 0

0 12 :0

0

10 :0 0

8: 0

00 6:

0 4: 0

2:

0:

00

50.0 45.0 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 00

Surface temperature (˚ C)

the E treatment (453 and 243 J/m2 at 28 and 155 days). For this period of exposition, only the S treatment showed a statistically significant reduction of toughness (p < 0.01) because of the dispersion of individual results (coefficients of variation of at least 12% for the analyzed series). Toleˆdo Filho et al. [25] and Bentur and Akers [26] noted similar embrittlement in aged vegetable fiber–OPC composites and found that it could be directly attributed to the petrifaction of the reinforcement through the migration of hydration products to the fiber lumens and pores. Fiber decomposition by alkaline environment and matrix micro cracking are other mechanisms of degradation that could also be mentioned [20]. In the initial age of 28 days the reinforcement of sisal fiber seemed to be more effective than the eucalyptus one. For the same fiber content the S treatment (3% by mass of sisal fiber) reached values of maximum load and toughness that were, respectively, 46% and 148% higher than the correspondent values of the ES treatment (2% of eucalyptus and 1% of sisal fiber). On the other hand the embrittlement of composite with sisal as the only reinforcement after the weathering attack was greater than that of the composite with hybrid (sisal and eucalyptus) reinforcement what suggests the more intense degradation of sisal fiber as stated by Agopyan [7].

0

E

16 :0 0

Toughness (J/m²)

1200

14 :0

672

Time Asbestos-fiber corrugated sheet

Tiles reinforced with vegetable fibers

Fig. 7. Downside surface temperature of tiles reinforced with vegetable fibers and of asbestos–cement corrugated sheet during 24 h.

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674

they did not present a potential for some deleterious effects by the incidence of UV radiation such as discoloration or degradation. Prado and Ferreira [12] obtained a similar performance of reflectance in the UV radiation for fiber– cement and metallic materials, with values around 25% and 70%, respectively. The results in the visible range showed values of reflectance between 43.4% and 31.4% for the fiber–cements under evaluation, with the best performance for corrugated sheets of galvanized steel with zinc coating (49.4%) and the lowest result for red ceramic (30.2%). The values found by Prado and Ferreira [12] for fiber–cement, red ceramics and metallic materials were 36%, 33% and 70%, respectively. The metallic materials are well known by their high reflection especially before the oxidation of the surface that usually increases with aging processes. The red ceramic and steel with zinc coating showed high reflectance for the infrared (IR) range with values around 80 and 70%, respectively, while the asbestos–fiber presented 45% and sisal fiber showed 30%. The nature of the IR is mainly thermal, and these results showed that the red ceramic had best performance. High values of reflectance in the visible and IR regions can contribute to better thermal performance of roofs [12]. The analysis of thermal results for sisal and asbestos fiber–cement did not favor the connection of the performance of reflectance with the temperature on the downside surface of the tile. Sisal fiber–cement showed lower IR reflectance and better performance than asbestos fiber– cement for the downside surface temperature. Therefore, other factors must have contributed for the better thermal performance of sisal fiber–cement compared to the asbestos–cement. One of these factors could be the design of the sisal fiber–cement tiles. This is the case of small tiles with higher incidence of unsealed joints, which promotes a better rate of ventilation and the consequent cooling of the roofing. Another explanation can be related to the higher thickness and porosity of tiles with sisal fiber (10 mm and 44.8% after aging, respectively) compared to the asbestos–cement (5.0 mm of thickness and 28.8% of permeable porosity). The higher porosity of sisal fiber–cement contributes to the lower transfer of heat and to the greater cooling of the roofing system in comparison to asbestos–cement. 4. Conclusions The thermal performance showed that roofing tiles reinforced with vegetable fiber are acceptable as substitutes of asbestos–cement sheets that are still in use in several developing countries. In the case of physical properties the fiber content was more important for the composite behavior than the type of reinforcement. The S and ES treatments with 3% of fiber reinforcement showed higher values of bulk density combined with lower water absorption and permeable void volume. For mechanical properties, the type of fiber was the main variable as the best results were

673

connected to the S composite reinforced only by 3% by mass of sisal kraft pulp. The superiority of the sisal fiber– cement was more evident for the toughness of tiles under bending solicitation. Exposure to tropical climate caused a severe reduction in the mechanical properties of the composites. This behavior can be attributed to alkaline attack and petrifaction of the natural fiber and progressive micro cracking of the cement matrix. The toughness of the vegetable fiber– cement fell to between 53% and 68% of that of non-aged composites after approximately four months under external weathering. The high porosity associated with water absorption of at least 30% by mass is expected to play a significant role in this undesirable behavior. The refinement of pore structure or the combined use of vegetable and synthetic fibers for reinforcement may be some effective approaches to material optimization. Acknowledgements The authors acknowledge the assistance given by the Laboratory of Building Systems of the Escola Polite´cnica, Universidade de Sa˜o Paulo (USP), Brazil, in the use of the spectrophotometer. They are also in debt with Dr. Neide Matiko Nakata Sato, of the Department of Civil Construction of Escola Polite´cnica, USP, for her help with the interpretation of the reflectance results. This research was conducted with financial support and grants kindly provided by the Fundac¸a˜o de Apoio a` Pesquisa do Estado de Sa˜o Paulo (Fapesp, process No. 2002/07362-0), Coordenac¸a˜o de Aperfeic¸oamento de Pessoal de Nı´vel Superior (Capes, process No. 0125/01-6 and scholarship to the first author) and Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq, process No. 305999/2003-6). References [1] Lola CR. Fiber reinforced concrete roofing sheets technology appraisal report. In: Third international symposium on developments in fiber reinforced cement and concrete, vol. 1, 1986; 5 p. [2] Giannasi F, The´baud-Noy A. Occupation exposures to asbestos in Brazil. Int J Occ Environ Hea 1997;3(2):150–7. [3] Harrison PTC, Levy LS, Pratrick G, Pigott GH, Smith LL. Comparative hazards of chrysotile asbestos and its substitutes: a European perspective. Environ Hea Perspect 1999;107(8):607–11. [4] Lee A. O impacto da substituic¸a˜o do telhado de fibrocimento. Monograph. Sa˜o Paulo, Brazil: Departamento de Engenharia de Construc¸a˜o Civil da Escola Polite´cnica Universidade de Sa˜o Paulo; 2000 [in Portuguese]. [5] John VM. Pesquisa e Desenvolvimento de Mercado para Resı´duos. In: Workshop Reciclagem e Reutilizac¸a˜o de Resı´duos como Materiais de Construc¸a˜o Civil, Sa˜o Paulo, 1996. Anais. Sa˜o Paulo, EP-USP/ ANTAC; 1997. p. 21–30 [in Portuguese]. [6] Guimara˜es SS. Vegetable fiber–cement composites. In: International symposium on vegetable plants and their fibers as building materials, Salvador. Proceedings. London: Chapman and Hall; 1990. p. 98–107. [7] Agopyan V. Materiais reforc¸ados com fibras para a Construc¸a˜o Civil nos Paı´ses em Desenvolvimento: o uso de fibras vegetais. EP-USP, Sa˜o Paulo (Livre–doceˆncia); 1991 204p [in Portuguese]. [8] Savastano Jr H. Materiais a` base de Cimento Amianto Reforc¸ados com Fibra Vegetal: Reciclagem de Resı´duos para a Construc¸a˜o de

674

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

L.C. Roma Jr. et al. / Construction and Building Materials 22 (2008) 668–674 Baixo Custo. Tese (Livre-doceˆncia). Escola Polite´cnica da Universidade de Sa˜o Paulo. Sa˜o Paulo, Brazil; 2000. 144p [in Portuguese]. Saxena M, Morchhale RK, Meshram AN, Khazanchi AC. Development of sisal cement composites as substitute for asbestos–cement components for roofing. In: International symposium on fiber reinforced cement and concrete, 4, Sheffield. Proceedings. London, E&FN Spon; 1992. p. 1140–51 (Rilem Proceedings, 17). Gram H-E, Gut P. Directives pour le controle de qualite. St. Gallen, Skat/BIT. 1994. 69p [Serie Pedagogique TFM/TVM: Outil 23, in French]. Delvasto SA, Gutierrez R, Lopez MM. Como hacer tejas con fibra de fique. 2nd ed. Cali, Departamento de Materiales de Ingenierı´a – Universidad del Valle; 1995. 16p (Programa Aprovechamiento Integral del Fique, 4) [in Spanish]. Prado RTA, Ferreira FL. Measurement of albedo and analysis of its influence the surface temperature of building roof materials. Energy Buildings 2005;37:295–300. ABNT – Associac¸a˜o Brasileira de Normas Te´cnicas. NBR – 11578. Cimento Portland composto: especificac¸a˜o. Rio de Janeiro: ABNT. 1991 [in Portuguese]. ´ gua do poro de pastas cimento de esco´ria. Sa˜o Paulo. Oliveira CTA. A Tese (Doutorado) – Escola Polite´cnica, Universidade de Sa˜o Paulo, Sa˜o Paulo, Brazil; 2000. 162 p [in Portuguese]. ABNT – Associac¸a˜o Brasileira de Normas Te´cnicas. ABTN NBR7215. Cimento Portland – Determinac¸a˜o da resisteˆncia a` compressa˜o. Rio de Janeiro: ABNT; 1996 [in Portuguese]. Savastano Jr H, Pimentel L. Viabilidade do aproveitamento de resı´duos de fibras vegetais para fins de obtenc¸a˜o de material de construc¸a˜o. Revista Brasileira de Engenharia Agrı´cola e Ambiental, Campina Grande, PB. 2000;4(1):103–10 [in Portuguese]. ASTM C 948-81. Test method for dry and wet bulk density, water absorption, and apparent porosity of thin sections of glass–fiber reinforced concrete, American Society for Testing and Materials.

[18] Taylor HFW. Cement chemistry. 2nd ed. London: Thomas Telford; 1997. [19] ABNT – Associac¸a˜o Brasileira de Normas Te´cnicas. NBR – 7581. Telha Ondulada de Fibrocimento. Rio de Janeiro: ABNT; 1993 [in Portuguese]. [20] Gram H-E. Durability of natural fibers in concrete. In: Swamy RN, editor. Natural fiber reinforced cement and concrete. Glasgow: Blackie; 1988. p. 143–72. [21] Mohr BJ, Nanko H, Kurtis KE. Durability of kraft pulp fiber– cement composites to wet/dry cycling. Cem Concr Compos 2005;25: 435–48. [22] Coutts RSP. Wood fiber reinforced cement composites. In: Swamy RN, editor. Natural fiber reinforced cement and concrete. Concrete technology and design, vol. 5. Glasgow: Blackie; 1988. p. 1–62. [23] Agopyan V, Savastano Jr H, Cincotto MA. Developments on vegetable fibre–cement based materials in Sao Paulo, Brazil: an overview. Cem Concr Compos 2005;27:527–36. [24] Toleˆdo Filho RD, Ghavami K, England GL, Scrivener K. Development of vegetable fiber–mortar composite of improved durability. Cem Concr Compos 2003;25:185–96. [25] Toleˆdo Filho RD, Scrivener K, England GL, Ghavami K. Durability of alkali-sensitive sisal and coconut fibers in cement mortar composites. Cem Concr Compos 2000;22(2):127–43. [26] Bentur A, Akers SAS. The microstructure and ageing of cellulose fiber reinforced cement composites cured in a normal environment. Int J Cem Compos Lightweight Concr 1989;11(2):99–109. [27] Bueno AD, Lamberts R, Philippi PC. Avaliac¸a˜o do comportamento higro-te´rmico de telhas ceraˆmicas em ambiente real. In: Fifth international seminar on structure masonry for developing countries, Floriano´polis, Brasil; 1994. 13p [in Portuguese]. [28] Savastano Jr. H, Warden PG, Coutts RSP. Performance of low-cost vegetable fiber–cement composites under weathering, CIB World Building Congress, Wellington, New Zealand; 2001. 11 p.