Consolidation of Timber Degraded by Fungi in Buildings: an Experimental Approach

Transcription

1 Consolidation of Timber Degraded by Fungi in Buildings: an Experimental Approach Henriques, D.F. Department of Civil Engineering, Instituto Superior de Engenharia de Lisboa, R. Conselheiro Emídio Navarro, 1, Lisbon, Portugal ( Nunes, L. Structures Department, Laboratório Nacional de Engenharia Civil, Av. do Brasil, 101, Lisbon, Portugal ( de Brito, J. Department of Civil Engineering and Architecture, Instituto Superior Técnico, Universidade Técnica de Lisboa, Av. Rovisco Pais, , Lisbon, Portugal ( Abstract The present paper concerns the theme of maintenance/rehabilitation of timber structural elements in heritage buildings that have been degraded by rot fungi. This type of degradation usually results from the unwanted but continuous presence of water in timber elements, very frequently permeating through damaged roofs and facades. In the past, it was common practice to replace degraded timber with new elements of the same or similar material. However, the option of maintaining in the building the original timber (even though deteriorated) has gained considerable importance in recent years. Therefore, a research programme on the consolidation of timber through impregnation of polymeric products is under development, looking for ways to regain the former original physical and mechanical conditions. This technique has been largely used in restoration of works of art but very little in civil engineering. The laboratorial work, developed using degraded maritime pine specimens, is presented. The test-specimens were either prepared in the laboratory, at different decay levels by varying the exposure time to brown-rot fungi. Six commercial low-viscosity products, specific for the impregnation of timber deteriorated by rot fungi, were tested: four epoxy-based (R, E, Lw, Li) and two acrylics (PB72, B98). The efficiency of these products was assessed through compression tests parallel to the grain and hardness tests, involving both impregnated and non impregnated wood. The relationships between the level of degradation of the timber (mass loss), the penetration capacity of the polymer and the consolidation efficiency on test-specimens are presented and discussed. Test results indicate a reasonably good strengthening effect of some of the products tested, the best value being a 39% increase in compressive strength and 91% increase in surface hardness by comparison with timber under the same circumstances but not consolidated. Keywords: timber, degraded wood, heritage buildings, consolidation, rehabilitation 424

2 1. Introduction The process of consolidating degraded timber by impregnation consists in forcing a specific fluid material to penetrate it, which when hardened will give back its integrity and promote an improvement of the physical and mechanical characteristics (Wang & Schniewind, 1985; Schaffer, 1984; Sakuno & Schniewind, 1990 and Loferski, 2001). The efficiency of the impregnation is affected from the depth achieved, the nature of the product employed and its permanence on the cells and cellular walls of the material (Nero et al, 2004). It has been common practice to replace degraded timber in buildings with elements of the same or other material, but in the future the option of maintaining the original timber, even if deteriorated, will gain importance. As a matter of fact, extensive replacement of elements not only is expensive but it is often also unnecessary, it changes the aspect of the structure, it disrupts the normal use of the building and it results in a loss of historical identity (Nunes et al, 2005; Lavisci et al, 2001). Consolidation of timber through impregnation of polymeric products is a technique already widely applied in the conservation and restoration of works of art but not so in civil engineering. Despite the large number of application in practical cases, quantitative data on the effect of strengthening by consolidants are still very much unavailable (Nunes et al, 2005). Concerning the conservation and restoration of works of art areas, before the development of synthetic materials a great number of other materials, such as animal glues and natural resins, were already used as consolidants, each of them with its own drawbacks (Wang & Schniewind, 1985; Lehmann et al, 2005). Thermo-hardening synthetic resins, such as epoxy and methyl methacrylate, were first used as consolidants for deteriorated timber in the middle of the XX century (Schaffer, 1974), with the limitation of its irreversibility, among others (Wang & Schniewind, 1985; Schaffer, 1984). The researchers interest then turned to thermoplastic polymers whose application was reversible, easily penetrated the micro-porosity of the timber and enhanced the strength the degraded timber (Wang & Schniewind, 1985). Among the thermoplastic products polyvinyl butyral (Butvar ) and acrylic resin (Acryloid or Paraloid ) were the ones that presented the better results in terms of enhancement of mechanical strength (Wang & Schniewind, 1985; Nakhla, 1986). As regards the buildings rehabilitation area epoxy-based products have been largely used in the last decades as glues for structural strengthening of degraded timber (Lavisci et al, 2001; Pizzo et al, 2002). Advances in epoxy resin formulations coupled with adhesive repair techniques developed for concrete and other materials, have fostered the application of wood structures (Loferski, 1999). Among them the low viscosity impregnants stand out essentially used for non-structural repair, to consolidate deteriorated wood members such as window and door frames, cornices, capitals and other decorative architectural components (Loferski, 1999). This paper presents the first part of a study whose aim is the application in the rehabilitation of structural elements in buildings of acrylic products used in conservation and restoration and of epoxybased products used essentially in decorative elements. 425

3 2. Materials 2.1 Wood samples 15 x 25 x 50 mm specimens cut from dry maritime pine (Pinus pinaster Ait.) were grouped in series of 10 similar specimens, each series representative of the sample as a whole. The selection of specimens for testing complied with the following physical demands: exclusively sapwood, free from defects, maximum dimensional deviation of 0.5 mm in any of the faces, annual growth rings between 2.5 and 8 per 10 mm, proportion of late wood in the annual rings not exceeding 30% of the whole as described in EN113, and density of the specimens 605 ± 20 kg/m 3. The specimens were then exposed to the cubic rot fungus Coniophora puteana (Schumach.) P. Karst, for periods of 4, 8 and 12 weeks. Various levels of degradation of the specimens were thus obtained, translated into mass loss determined as a percentage of the ratio between the lost dry mass and the initial dry mass of the non-degraded specimen. Before and after the degradation process, the specimens were conditioned in an environment with 20 ± 2 ºC and 65 ± 5% RH, and weighing was performed only after mass stabilization. Mass loss levels between 3% and 25% were obtained for the whole sample, considered representative of the degradation level of timber in buildings in a situation where it can be recuperated (Figure 1). Figure 1: Non-degraded mini-specimen (centre) and with various levels of mass loss: 5%, 10%, 15%, 20% and 25% 2.2 Products of consolidation To select the consolidants the following characteristics were considered: penetration capacity of the timber, mechanical strength, durability, short-term reversibility, ease of application, low toxicity, good aesthetical aspect, low price, and facility of acquisition (Nakhla, 1986). Four consolidants are thermo-hardening epoxy-based products (R, E, Lw and Li), made of a resin and a hardener. The polymerization of the product is initiated immediately after mixing and homogenizing 426

4 of the two components, its pot life defined as the period in which each of these products remains workable. Product E has a resin from diglycidyl ether of bisphenol A (DGEBA) and a hardener made of aliphatic and cycloaliphatic amines. Product R has a resin from bisphenol A (DGEBA) and from bisfenol F (DGEBF) and the hardener is based on aliphatic amines. Products Lw and Li s composition is not referred available. The pot life, is relatively low: at 22 ºC, 50 min for R, 40 min for E, 2 h for Lw, and 4h for Li. Two commercial thermoplastic acrylic-related products were also included, Paraloid B72 and Butvar 98, referred here as PB72 and B98 currently used in conservation and restoration of works of art. They are made available as solid granulate and need to be dissolved in order to obtain a liquid solution. The penetration capacity of the resin strongly depends on the solvent as well as on its proportion in the composition (Lehmann et al, 2005; Wang & Schniewind, 1985). Several preliminary selection tests were performed, from which acetone was chosen as solvent to the product PB72 at a ratio in mass of 80:20 (acetone:pb72) (Henriques et al, 2009). A pot life of 50 min at 22 ºC was obtained. To dilute B98 resin results from Sakuno & Schiewind (1990) tests were used, and two solvents were used: ethanol and toluene at a 40:60 proportion and 15% concentration in mass. A pot life of around 1.5 h at 22 ºC was obtained. 3. Methods 3.1 Consolidation test The specimens were stabilized in a conditioned room and they had a moisture of 12.9%, n=6, at the treatment date. Consolidation was performed by immersion of the blocks in the different consolidants for 15 min, considered long enough after preliminary tests made with various consolidants. The amount of polymer absorbed was determined by weight using the difference in mass that each specimen stabilized in a conditioning room with 20 ± 2 ºC and 65 ± 5% RH presents before and after consolidation 3.2 Mechanical tests Consolidated specimens were tested after stabilization of the mass in a conditioned room. The axial compression test (parallel to fibres) was performed in accordance with NP618:1973 until rupture of each specimen, and the corresponding value was presented in N/mm 2. The static hardness test (perpendicular to fibres) followed ISO 3350:1975, consisting on driving in a semi-sphere with a radius of 5.64 mm at constant speed and registering the strength in N at which rupture occurs or at the one needed to totally insert the semi-sphere. For each type of mechanical test and consolidant, a series of 10 degraded specimens and 3 non-degraded specimens were tested. Furthermore, a similar set of non consolidated control specimens was submitted to each of the mechanical tests (Figures 3 and 4). 427

5 4. Results and discussion 4.1 Quantity of polymer absorbed Figure 2 presents the mass values of polymer absorbed versus the loss of mass of each specimen. As predicted, the mass of product absorbed by the non-degraded specimens is lower than that of the degraded specimens. Nunes et al (2005) refer that this can be explained by the fact that fungal attack decreases wood cells strength and cohesion, thus leading to a number of internal cracks in the wood as it dries and a larger internal surface. However, there is no clear correlation between the level of decay and product absorption. There is however a distinct difference in the quantities absorbed by the timber after stabilization of each consolidant used, as seen in Figure 2. Both for the two acrylic products (PB72 and B98) and the epoxy-based one (Li) the quantities absorbed are very small (in all cases lower than 2 g per specimen). These three products contain a very high proportion of volatile components (around 80% for PB72, around 85% for B98, and around 80% for Li, according to the results obtained). The quantities absorbed by the timber for both epoxy-based products R and E (between 3 g and 8 g per degraded specimen) are much higher than the cases mentioned before and moderately higher for epoxy-based product Lw (between 2 g and 4 g per degraded specimen). The volatile components contents of these last three products are insignificant. 4.2 Consolidation effect Each specimen provided a pair of values for each type of mechanical test with which graphs were prepared for each consolidant and a regression line was obtained (Figures 3 and 4). The low dispersion of the mechanical results, as translated by R 2 values very close to one, provides a high degree of confidence in the results obtained. The control line obtained by correlation between compressive strength in the non-consolidated degraded specimens and corresponding mass loss highlights the great loss of strength due to the increase in the timber s degradation, evident in both test types. Nunes et al (2005) in their experimental work also observed that even a small mass loss (e.g. 10%) can result in a severe reduction (of almost 50%) of both compressive strength parallel to fibres and surface hardness (perpendicular to fibres), and that a similar trend is found for the influence of mass loss in both types of test. The numerical translation of the diagrams in Figures 3 and 4 is presented in Table 1 as the percentiles of strength increment promoted by each consolidant in relation to the control specimen, determined according to equation (1), whose ordinates y is obtained from the respective exponential regression lines presented in Table

6 8 Asorption (g) % 5% 10% 15% 20% 25% Mass loss(%) R E Lw Li PB72 B 98 Figure 2: Mass of consolidant absorbed by each specimen versus its mass loss. R 2 R = 0,86; R 2 E = 0,95; R 2 Lw = 0,81; R 2 Li = 0,83; R 2 PB72 = 0,90; R 2 B98 = 0,69; R 2 Control = 0,93 55 Compressive strength (N/mm 2 ) % 5% 10% 15% 20% 25% mass loss (%) R E Lw Li PB72 control B98 Expon. (Lw) Expon. (control) Expon. (E) Expon. (R) Expon. (Li) Expon. (PB72) Expon. (B98) Figure 3: Compressive strength versus mass loss. R 2 R = 0,87; R 2 E = 0,87; R 2 Lw = 0,92; R 2 Li = 0,83; R 2 PB72 = 0,90; R 2 B98 = 0,77; R 2 Control = 0, Resistance to indentation (N) % 5% 10% 15% 20% 25% mass loss (%) R E Lw Li PB72 control B98 Expon. (Lw) Expon. (control) Expon. (E) Expon. (R) Expon. (Li) Expon. (PB72) Expon. (B98) Figure 4: Resistance to indentation versus mass loss. 429

7 Where: yi, x yc, x I i, x = 100 (1) y c, x I i,x - increment x promoted by consolidant I in relation to the control; y i,x - ordinate x of the regression line of consolidant i; y c,x - ordinate x of the control regression line. Table 1: Mechanical strength increments of the consolidated sets in relation to the control set, for various levels of mass loss, according to equation (1) Mass loss R E Lw Li PB 72 B 98 (%) Compressive strength increment in relation to the control specimen (%) Surface hardness increment in relation to the control specimen (%) , , , , , The analysis of Figures 3 and 4 and Table 1 indicate that almost all consolidation products tested have strengthening effects, increasing those increments with the level of decay. In terms of compressive strength both the epoxy-based (E and Lw) products and two acrylic (PB72 and B98) ones show reasonable consolidation effects (between 27% and 39% in relation with the control specimens), for a timber mass loss of 20%. In terms of surface hardness consolidant E clearly stood out with an increment of 20%, while the others showed lower increments: R, Lw and PB72 between 20% and 27% and Li and B98 between 3% and 13%. 4.3 Absorption versus consolidation effect Each polymer presents a particular performance in what concerns the relationship between the quantity of resin retained by the timber and the mechanical performance as consolidant. 430

8 Acrylic consolidant PB72, even though presenting a small quantity absorbed by the timber, provides good consolidation increments translated in increments of 27% for compressive strength and surface hardness, for a mass loss of 20%. Acrylic consolidant B98 shows an even lesser quantity absorbed by the timber, promotes decreasing and small surface hardness increments as degradation levels increases, but has a good performance in terms of compressive strength: increments of 37% for a mass loss of 20%. In what concerns epoxy-based consolidants, Li shows a small quantity absorbed by the timber and provides low mechanical performances: increments of 15% and 13% in compressive strength and surface hardness respectively, for a mass loss of 20%. Consolidant R shows great quantities absorbed but relatively low mechanical performances: increments of -3.8% and 20% in compressive strength and surface hardness respectively, for a mass loss of 20%. Consolidant Lw, with a reasonable capacity of retention within the timber, also presents reasonable increments of compressive strength and surface hardness (30% and 22% respectively) for the maximum value of mass loss considered. Finally, consolidant E shows great quantities absorbed by the timber and very good mechanical performances: increments of 39% and 91% in compressive strength and surface hardness respectively, for a mass loss of 20%. However it is concluded that there is no direct relationship between the quantity of polymer absorbed by the degraded timber and its efficiency as consolidant, for the present set of products tested. 4.4 Consolidation effects from this and the previous set of tests These results were compared with those of another set of tests developed by the same authors (Henriques et al, 2009) using the same methods, the same materials and the same degradation process of the specimens, but with a larger range of timber densities before degradation (650 ± 65 kg/m 3 ). The results dispersion in the mechanical tests performed (compressive strength and surface hardness), translated in not very high R 2 values, led the authors to repeat the tests with a narrower range of densities (605 ± 20 kg/m 3 ). The differences obtained are highlighted in Table 2, which presents (for the control specimens and those using epoxy-based consolidant E and acrylic consolidant PB72) the values obtained in the tests, the percentile increments in relation to the control specimens and the values of R 2 for each set of results obtained. The lower values of mechanical strength obtained in the present study as compared with those of Henriques et al (2009) were expected as they are a direct consequence of the lower density of the timber used (650 ± 65 kg/m 3 versus 605 ± 20 kg/m 3 ). Mateus (1961) obtained an average of compressive strength for Pinus pinaster sound wood of 43.5 N/mm 2 for the density of 605 kg/m 3 and 48.1 N/mm 2 for the density of 650 kg/m 3. The analysis is centred on the increments of the consolidated timber s mechanical strength versus the equivalent values of the non-consolidated timber and on the measurement of the results scatter, R 2. For the compressive strength test the lower increment values are largely compensated by the dispersion decrease (increase of R 2 ), providing greater reliability in the analysis of the results. For the surface hardness test the increments were lower and the increase of R 2 is very small. 431

9 This apparent discrepancy is due to the fact that the surface hardness increases more with the increase in density (due to the boundary conditions) of the consolidated specimens than the compressive strength parallel to the fibres (Madsen et al, 1982). Table 2: Comparison between the results obtained in the previous study and the present one, in terms of compressive strength and surface hardness. Compressive strength Control (non consolidated) Epoxy E Acrylic PB72 Mass loss (%) Henriques et al (2009) Present study Strength(N/mm 2 ) Increment (%) R Strength.(N/mm 2 ) Increment (%) R Surface hardness Control (non consolidated) Epoxy E Acrylic PB72 Mass loss (%) Henriques et al Hardness (N) (2009) Increment (%) R Present study Hardness(N) Increment (%) R Conclusions From this study the performance of products E, Lw and PB72 stand out. In particular, product E, made of a resin from diglycidyl ether of bisphenol A (DGEBA) and a hardener made of aliphatic and cycloaliphatic amines, even though not recuperating the initial timber characteristics, provides it with a good consolidation capacity, shown by the significant increase of the mechanical characteristics (compressive strength and surface hardness). Of these three consolidants E is the one that is absorbed in greater quantity by the timber. The results in terms of mechanical strength shown in Table 2 confirm that one should work with a narrow range of initial density values. 432

10 Acknowledgements To Fundação para a Ciência e Tecnologia (FCT) for the PhD scholarship awarded to M. Dulce Franco Henriques (ref. SFRH/BD/44216/2008). References Henriques D, Nunes L, Brito J de, (2009), Test of consolidation products for wood degraded by fungi (in Portuguese). 3 rd Meeting on pathology and rehabilitation of buildings - PATORREB 2009, Porto University, Portugal, pp ISO (1975): Bois - Détermination de la dureté statique. O. I. de Normalisation, Suisse. Lavisci, P; Beri, S; Pizzo, B; Triboulot, P; Zanuttini, R. (2001): A shear test for structural adhesives used in the consolidation of old timber. Holz als Roh - und Werkstoff, 59, pp Lehmann, E; Hartmann, S; Wyer, P. (2005): Neutron radiography as visualization and quantification method for conservation measures of wood firmness enhancement. Nuclear Instruments and Methods in Physics Research A, 542, pp Loferski, JR (1999): Technologies for wood preservation in historic preservation. Archives and Museum Informatics 13, pp Madsen, B; Hooley, RF; Hall, CP (1982): A design method for bearing stresses in wood. Canadian Journal of Civil Engineering, 9, pp Mateus, THE (1961): Basis for the design of timber structures (in potuguese). National Laboratory for Civil Engineering, Lisbon, Portugal. 306 pp. Nakhla, S. M. (1986): A comparative study of resins for the consolidation of wooden objects. Studies in Conservation, 31, pp Nero, GJM; Duarte, AP; Bordado,JC. (2004): Principles for understanding timber impregnation. Evaluation of the efficiency of impregnation (in Portuguese). 1 st Iberian Congress Timber in Construction, Minho University, Guimarães, Portugal, pp NP618. (1973): Axial compression test (in Portuguese). IGPAI, Lisbon, Portugal. Nunes, L; Cruz, H; Duarte, AP; Bordado, JC; Nero, GJM. (2005) Wood consolidation by impregnation with reactive polymers. International Conference on Conservation of Historic Wooden Structures, Firenze, Italy, pp Pizzo, B; Rizzo, G; Lavisci, P; Megna, B; Berti, S. (2002): Comparison of thermal expansion of wood and epoxy adhesives. Holz als Roh - und Werkstoff, 60, pp Sakuno, T. & Schniewind, A. P. (1990): Adhesive qualities of consolidants for deteriorated wood. JAIC, 29(1), art. 3: EUA, pp Schaffer, E. (1974): Consolidation of painted wooden artifacts. Stud. Conserv. 19, pp Wang, Y. & Schniewind, A. P. (1985): Consolidation of deteriorated wood with soluble resins. JAIC, 24(2), art. 3: EUA, pp