- Email: [email protected]
Modification of fast-growing Chinese Fir wood with unsaturated polyester resin: Impregnation technology and efficiency
Results in Physics 6 (2016) 543–548
Contents lists available at ScienceDirect
Results in Physics journal homepage: www.journals.elsevier.com/results-in-physics
Modification of fast-growing Chinese Fir wood with unsaturated polyester resin: Impregnation technology and efficiency Qing Ma, Zijian Zhao, Songlin Yi ⇑, Tianlong Wang Beijing Key Laboratory of Wood Science and Engineering, MOE Key Laboratory of Wooden Material Science and Application, Beijing Forestry University, 35# Qing Hua East Road, Beijing 100083, China
a r t i c l e
i n f o
Article history: Received 24 June 2016 Accepted 19 August 2016 Available online 26 August 2016 Keywords: Unsaturated polyester resin (UPR) Impregnation Vacuum Curing weight percent gain distribution
a b s t r a c t In this study, Chinese Fir was impregnated with unsaturated polyester resin to enhance its properties. Samples 20 mm 20 mm 20 mm in size were split into different sections with epoxy resin and tinfoil and subjected to an impregnation experiment under various parameters. Vacuum degree was 0.04 MPa, 0.06 MPa or 0.08 MPa and vacuum duration was 15 min, 30 min, or 45 min. The results indicated that impregnation weight percent gain is linearly dependent on curing weight percent gain. Vacuum duration appears to have less influence on the curing weight percent gain than vacuum degree, and impregnation was most successful at the transverse section compared to other sections. The optimal impregnation parameters were 30 min modification under 0.08 MPa vacuum followed by 120 min at atmospheric pressure for samples 200 mm 100 mm 20 mm in size. Uneven distribution of weight percent gain and cracking during the curing process suggested that 30 min post-processing at 0.09 MPa vacuum was the most effective way to complete the impregnation process. The sample’s bending strength and modulus of elasticity increased after impregnation treatment. Bending strength after impregnation without post-processing reached 112.85%, but reached 71.65% with vacuum-processing; modulus of elasticity improved 67.13% and 58.28% without and with post-processing, respectively. Ó 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Introduction As one of the most common fast-growing tree species in China, Chinese Fir is an important potential solution to the problematic shortage of wood resources. A kind of softwood, however, it exhibits several drawbacks which limit its practical application [1]. A number of modification technologies have been proposed to enhance the physical, chemical, and biological properties of the wood, such as heat or pressure treatment, surface coating, or impregnation with reactive materials [2,3]. Among the available modification methods, chemical impregnation under vacuum or pressure may be the best way to improve the specific properties of the wood cell wall material [2]. Curing the wood with resins appears to be especially beneficial. Unsaturated polyester resin (UPR) is a popular thermosetting material [4] generally prepared from a saturated acid and an unsaturated acid condensed with dihydric alcohols, which forms a three dimensional structure via cross linking with a vinyl monomer [5,6]. It can be used either as pure resin or with fillers, or reinforced, respectively [6]. As a reactive material, unsaturated polyester resin ⇑ Corresponding author. E-mail address: [email protected] (S. Yi).
has been shown to effectively enhance wood performance [3] and prolong the service life of wood for several years [7]. Mechanical loss is a drawback to impregnation modification in general [8], but less so when utilizing UPR. In this study, a general-grade UPR was prepared in our laboratory, which is prepared from 1,2-propylene glycol (PG), maleic anhydride (MA), and phthalic anhydride (PA) [4]. The primary goal was to establish an impregnation technique well-suited to the modification of Chinese Fir. Fir samples 200 mm 100 mm 20 mm in size was used, as per the typical dimensions of impregnation specimens, and partitioned them in order to observe the relationship among elements distributed within the different partitions. Finite element analysis [9,10] and mathematical analysis were utilized to split the samples into 50 impregnation elements (each 20 mm 20 mm 20 mm in size) which ultimately fell into four distinct categories, as discussed below. Materials and methods Materials The specimens for this study were prepared from three-yearcut, fast-growing Chinese Fir (Cunninghamia lanceolata (Lamb.)
http://dx.doi.org/10.1016/j.rinp.2016.08.017 2211-3797/Ó 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
544
Q. Ma et al. / Results in Physics 6 (2016) 543–548
Hook). The logs were cut from a forest in Guangxi Province, China, with diameter at breast height of 30–40 cm and age of 20–25 years. The logs were sawn into planks, and air-dried to 12–13% moisture content, then the sapwood near the butt portion was processed into samples. The test wood was of an average absolute dry density of 0.33 g/cm3 and porosity of approximately 77.62%. (For details of the Computational Method, refer to Siau’s 1984 study [11]). As mentioned above, the modifier used in this study was a homemade, general-grade UPR, comprised of 1,2-propylene glycol (PG), maleic anhydride (MA) and phthalic anhydride (PA) synthetic materials [4] as well as styrene as a final cross-linking agent. The average molecular weight of the homemade UPR is shown in Table 1. Methods Dimensions of the impregnation specimens The dimensions of the impregnation specimen for performance measurement was 200 mm long by 100 mm wide by 20 mm thick, as shown in Fig. 1(I). To determine the optimal impregnation process parameters for the UPR into Chinese Fir blocks of this size, impregnation was processed with this size in general [12–16] and various parameters were set and tested the consumption of the resin and the blocks accordingly through repeated screening experiments. The results indicated that the impregnation process needed to be simplified, which is the method by building an
analogy with mesh via finite element analysis [9,10] followed by mathematical analysis. When the impregnation process was conducted on the ‘‘big” (200 mm 100 mm 20 mm) block, there were six impregnation directions for UPR as shown in Fig. 1(I). Due to the regular shape of the impregnation specimen, this created fifty smaller samples (or ‘‘elements”) 20 mm 20 mm 20 mm in size; the corresponding mesh is shown in Fig. 1(II). Wood is a highly anisotropic material [17], so the elements themselves featured notable differences in structure [18], as well as mechanical properties. The mechanical property particularly affected in this case, permeability, is mainly influenced by fiber orientation [17,19]. Combining the location of the elements on the specimens, the corresponding impregnation section, and the fiber orientation of the impregnation section, it can be divided into four types of elements which are depicted in Fig. 1(III); the distinct impregnation section and fiber orientation are depicted in Fig. 1(IV). Element (a) was comprised of four impregnation sections: one transverse section, two tangential sections, and one radial section. Elements (b) and (c) were comprised of only three impregnation sections: two tangential sections and a radial section (b) or transverse section (c). Element (d) was comprised of only two tangential sections. Experimental methods The UPR thermosetting resin was first utilized to treat the element blocks via vacuum impregnation with various parameters
Table 1 Average molecular weight of the experimental UPR.
UPR
Mn
Mw
Mz
Mz + 1
Polydispersity
Mz/Mw
Mz + 1/Mw
1189
1377
1663
2056
1.157472
1.208365
1.493613
Mn: number-average molecular weight; Mw: weight-average molecular weight; Mz: Z-average molecular weight.
Fig. 1. Element partition and classification of the specimen.
545
Q. Ma et al. / Results in Physics 6 (2016) 543–548
elasticity Ew was calculated with the Eq. (4). Again, three replicate specimens of each group were tested.
Table 2 Impregnation experiment parameters. Iteration
Vacuum degree Vacuum duration
1
2
3
rbw ¼ ð3Pmax lÞ=ð2bh2 Þ
ð3Þ
0.04 MPa 15 min
0.06 MPa 30 min
0.08 MPa 45 min
Ew ¼ ð23Pl Þ=ð108bh f Þ
3
ð4Þ
(i.e., different vacuum degrees and vacuum durations) (Table 2). There were nine replications run in total to complete the fullfactor experiment. The vacuum duration began from the vacuum degree and was run until reaching the target value. Before the impregnation process, the elements were subjected to surface treatment. Different impregnation sections, in regard to both amount and direction, were treated corresponding to the different locations of elements. Closed sections of each of the four elements were blocked with epoxy resin and tinfoil, oven-dried at 103 °C for 8 h, then oven-dried once more at 80 °C for 6 h after the epoxy resin cured. The weight (W0) was then measured just prior to vacuum impregnation. The samples were then submerged in the modifier and subjected to vacuum treatment. The samples were impregnated in UPR for two more hours at atmospheric pressure after a 30 min vacuum duration, then the surface residues were removed and impregnation weight (W1) was measured. Samples impregnated with UPR were not cured straight through the previous step but instead during a 48-h repose. The impregnated samples were cured at 80 °C for 6 h and 103 °C for 8 h, then their curing weight (W2) was measured; impregnation weight percent gain (WPG1) and curing weight percent gain (WPG2) were determined via Eqs. (1) and (2), respectively.
WPG1 ¼ ðW 1 W 0 Þ=W 0 100%
ð1Þ
WPG2 ¼ ðW 2 W 0 Þ=W 0 100%
ð2Þ
The impregnation method for the subsequent experiments was determined according to these results through variance analysis. After selecting the appropriate method, 200 mm 100 mm 20 mm samples were impregnated as follows. The samples were cured and then sawn into smaller blocks according to the distribution of the elements shown in Fig. 1(II) to obtain the distribution of the curing weight percent gain. Density was measured first according to the weighed mass and the volume determined via the drainage method. Although the 200 mm 100 mm 20 mm samples exhibited favorable impregnation weight percent gain when impregnated via the selected process, the curing process remained uncertain; it was found that the impregnated samples cracked easily during the curing process. To homogenize the weight percent gain of the ‘‘big” samples appropriately, a post-processing technique under a vacuum 0.09 MPa for 30 min was added during which the samples were not submerged in the modifier. The 300 mm 20 mm 20 mm samples was impregnated both with and without post-processing, then measured their respective bending strength and modulus of elasticity according to Chinese standard ISO [20,21]. All the samples were measured under the conditions of equilibrium moisture content of 12%, 20 °C temperature and 65% relative humidity to constant weight. The bending strength of the samples was tested with a mechanical testing machine. The testing speed of the indenter was set to 5 mm/min. The strength rbw was calculated with Eq. (3). Three replicate specimens of each group were tested. The modulus of elasticity of the samples was also tested with a mechanical testing machine at indenter speed of 3 mm/min. After determining the difference of the load bound based on the linear relationship of the load-deformation diagram, the modulus of
3
where Pmax is the fracture load (N), l is the span between the two supporting seats (mm), b is the width of specimen (mm), h is the height of the specimen (mm), P is the difference in the load bound based on the linear relationship, and f is the deformation of the load bound based on the linear relationship. Results and discussion Effects of various vacuum degrees and durations to various elements The impregnation weight percent gains (WPG1) and curing weight percent gains (WPG2) were calculated for all four elements across nine 9 experiments. Because these properties are shortterm, the impregnation weight percent gain was of less importance than the curing weight percent gain. The two weight percent gain values are linearly dependent – nearly proportional, in fact – as is shown in Fig. 2. Accordingly, the subsequent analysis of the various vacuum degrees and durations in terms of the weight percent gain was mainly focused on the curing weight percent gains (WPG2). As mentioned above, nine experiments were run with various vacuum degrees and durations on the four kinds of fast-growing Chinese Fir elements. The resulting variance analysis for the curing weight percent gains (WPG2) is provided in Table 3. As shown in Table 3, compared to duration, vacuum degree exerted a far greater impact on the curing weight percent gain especially to Element (a) and Element (c). The shared characteristic of these two elements was the transverse impregnation section, so it is reasonable to infer that vacuum degree had more significant influence on the transverse section than the tangential or radial sections. This observation can also be attributed to the fact that the longitudinal permeability of softwoods is much greater than the transverse permeability [11]. For samples with certain section orientation, vacuum degree can be controlled to result in excellent permeability. The external vacuum contributes to pressure gradients in the wood [22], and samples with good permeability better
Fig. 2. Correlation between impregnation weight percent gain and curing weight percent gain.
546
Q. Ma et al. / Results in Physics 6 (2016) 543–548 Table 3 Variance analysis of curing weight percent gains (four types of elements).
Element (a) Element (b) Element (c) Element (d)
Vacuum degree Duration Vacuum degree Duration Vacuum degree Duration Vacuum degree Duration
Sum of squares
df
Mean square
F
Sig.
2.321 0.110 0.093 0.003 2.506 0.043 0.022 0.032
2 2 2 2 2 2 2 2
1.161 0.050 0.046 0.001 1.251 0.022 0.011 0.016
39.586 0.185 2.719 0.076 125.485 2.159 1.292 1.944
0.000 0.832 0.080 0.927 0.000 0.130 0.287 0.158
generate a gradient receptive to impregnation; it was found that an increase in vacuum degree was more beneficial to the pressure gradient than an increase in vacuum duration. The function of vacuum duration was fairly negligible, especially for samples with good permeability. This can be attributed to the favorable impregnation efficiency facilitated by appropriate vacuum degree having ensured favorable permeability.
weight percent gain, but the respective influence of these three methods on the four elements was inconsistent. Lengthier duration did not, per the expectations, result in high curing weight percent gain, so Method 9 was excluded from further analysis. Method 8 resulted in better impregnation effect than Method 7, so Method 8 (impregnation under 0.09 MPa for 30 min followed by 120 min at atmospheric pressure) was employed for subsequent experiments.
Impregnation method selection Curing weight percent gain distribution of the samples
Again, vacuum degree exerted greater influence on the curing weight percent gain after impregnation than vacuum duration, so the next step to attempt was to ensure similar curing weight percent gain with the briefest vacuum duration possible. An appropriate weight percent gain is a guarantee of favorable properties in the modified material, thus, the experiments were aimed to sink an impregnation method that resulted in high curing weight percent gain. Among the four types of elements, Element (a) exhibited the maximum weight percent gain followed by Element (c). The UPR distribution in these two elements was similar, as well; the modifier nearly penetrated both. When the two elements were returned to the big sample, nearby Elements (b) and (d) naturally transformed into Elements (a) and (c); this made the UPR distribution in the big sample relatively more homogeneous than the UPR distribution in the individual elements. As shown in Fig. 3, a stepwise improvement was observed in average curing weight percent gain which further implied that vacuum duration had little influence on the resulting properties of the sample. The final three tested impregnation methods (vacuum degree of -0.08 MPa) were preferable in regard to the curing
Samples 200 mm 100 mm 20 mm in size were impregnated with a pre-determined initial weight to measure only the curing weight percent gains of the whole blocks. According to the respective WPG2 values of the different elements, the distribution of the curing weight percent gain on the samples was expectedly inhomogeneous. To determine the distribution on the whole breadth of the sample, the sample was split into smaller blocks approximately 20 mm 20 mm 20 mm in size again. Because the initial weights of these blocks were unknown, the relationship between their density and curing weight percent gain was ascertained based on the smaller blocks observed in previous experiments. The epoxy resin and tinfoil that were used to block the impregnation process were removed prior to measuring the small samples’ weights and volumes via water displacement method. The relationship between the curing weight percent gain and the density (D) of the samples is shown in Fig. 4. It was found that density was linearly associated with curing weight percent gain
Fig. 3. Average curing weight percent gain of the four elements among nine impregnation methods.
Fig. 4. Relationship between curing weight percent gain and density of the samples.
Q. Ma et al. / Results in Physics 6 (2016) 543–548
(R2 = 0.94507). The approximate curing weight percent gain of the sawn blocks was calculated with the Eq. (5).
WPG2 ¼ ðD 0:32814Þ=0:00416 100%
ð5Þ
The distribution of the curing weight percent gain of the samples was uneven, as shown in Fig. 5; curing weight percent gain decreased gradually as the distance from the transverse section increased. The lower end in Fig. 5(II) can be attributed to outflow of the resin from the sample during the curing process. This created non-uniform distribution of the properties of the modified wood as well as severe end cracking during the curing process, which necessitated post-processing to reduce and homogenize the weight percent gain. Vacuum processing can be utilized to extract a portion of the modifier in the cell lumen, so a post vacuum processing of 0.09 MPa for 30 min was added to reduce the risk of cracking in the proposed impregnation technique. It should be noted that because of the differences of the samples themselves, the weight percent gains of the two test groups were not identical but did have the same magnitude (i.e., comparison of the two groups was still reasonable). The distribution of the curing weight percent gain of the samples with post-processing was shown in Fig. 6. The curing weight percent gain of the sample ends was still higher than the inner parts after post-processing, but the homogeneity of the samples was improved significantly and there was no cracking in the sample. By comparison among the distributions shown in Figs. 5 and 6, plus the curing results, it was determined that the optimum impregnation process included a vacuum degree of 0.08 MPa for 30 min followed by 120 min at atmospheric pressure under the modifier, then post-processing with a vacuum degree of 0.09 MPa for 30 min without the modifier.
547
Modification effect on bending strength and modulus of elasticity Bending strength and modulus of elasticity are two important properties of wood materials that are likewise affected by different modification methods [23–25]. Fig. 7 shows the average bending strength and modulus of elasticity of the three impregnation states tested in histogram form; the error bars are also marked. It was found that both the bending strength and modulus of elasticity of the samples increased significantly with the two modification methods it was used – especially for samples impregnated
Fig. 7. Bending strength and modulus of elasticity of samples with different impregnation processes.
Fig. 5. Distribution of curing weight percent gain of the samples without post-processing.
Fig. 6. Distribution of curing weight percent gain of post-processed samples.
548
Q. Ma et al. / Results in Physics 6 (2016) 543–548
without post-processing – but that greater increase in properties came with greater volatility. Compared to the non-impregnated samples, the bending strength of the samples impregnated without post-processing increased by 112.85%; the improvement after impregnation with post-processing was 71.65%. There was less difference in terms of modulus of elasticity, with improvements of 67.13% and 58.28%, respectively. Although the effects of the impregnation without post-processing were better, the deviations were larger. The curing weight percent gain of Sample (c) was greater than that of Sample (b), but the homogeneity of the distribution of Sample (b) was better. The homogeneity of UPR distribution, which is also directly related to the stability (i.e., performance) of the impregnated wood, was better in the post-processed sample as well. Conclusion The most notable conclusions of this study can be summarized as follows. 1 Impregnation weight percent gain is linearly dependent upon curing weight percent gain. Vacuum degree exerted greater effect on the curing weight percent gain (and thus the impregnated weight percent gain) than vacuum duration. Elements with transverse sections manifested greater curing weight percent gain and were affected more by vacuum degree than elements with other sections. 2 The optimal modification process included impregnation at 0.08 MPa vacuum for 30 min, 120 min at atmospheric pressure with submersion in the modifier, and post-processing under a 0.09 MPa vacuum for 30 min without the modifier. The distribution of the curing weight percent gain in postprocessed samples was optimal compared to samples that were not post-processed. 3 The impregnation modification facilitated substantial increase in bending strength and modulus of elasticity; the improvement was greater without vacuum processing than that with postprocessing (112.85% increase in bending strength without post-processing; 71.65% increase with post-processing). Modulus of elasticity improved to approximately equal extent without or with vacuum treatment (67.13% and 58.28%, respectively). Acknowledgements The paper was supported by the Special Fund for Forest Scientific Research in the Public Welfare (201404502) and the Key Discipline Construction Projects of Beijing—‘‘Efficient utilization of fast growing wood”. References [1] Cao Y, Lu J, Huang R, Jiang J. Increased dimensional stability of Chinese fir through steam-heat treatment. Eur J Wood Wood Prod 2012;70(4):441–4. http://dx.doi.org/10.1007/s00107-011-0570-y.
[2] Mathias LJ, Lee S, Wright JR, Warren SC. Improvement of wood properties by impregnation with multifunctional monomers. J Appl Polym Sci 1991;42 (1):55–67. http://dx.doi.org/10.1002/app.1991.070420107. [3] Mahmoud A, Eissa A, Omar M, El-Sawy A, Shaaban A. Improvements of white pine wood properties by impregnation with unsaturated polyesters in admixture with styrene. J Appl Polym Sci 2001;82(6):1410–6. http://dx.doi. org/10.1002/app.1978. [4] Zhang C, Huang JY, Liu SM, Zhao JQ. The synthesis and properties of a reactive flame-retardant unsaturated polyester resin from a phosphorus-containing diacid. Polym Adv Technol 2011;22(12):1768–77. http://dx.doi.org/ 10.1002/pat.1670. [5] Haq MIU. Applications of unsaturated polyester resins. Russ J Appl Chem 2007;80(7):1256–69. http://dx.doi.org/10.1134/S1070427207070464. [6] Fink JK. Reactive polymers fundamentals and applications: a concise guide to industrial polymers. William Andrew 2013;1–6:31–3. [7] Islam MN, Ando K, Yamauchi H, Kobayashi Y, Hattori N. Comparative study between full cell and passive impregnation method of wood preservation for laser incised Douglas fir lumber. Wood Sci Technol 2008;42(4):343–50. http:// dx.doi.org/10.1007/s00226-007-0168-z. [8] Sun W, Shen H, Cao J. Modification of wood by glutaraldehyde and poly (vinyl alcohol). Mater Des 2016;96:392–400. http://dx.doi.org/10.1016/ j.matdes.2016.02.044. [9] Nairn JA. A numerical study of the transverse modulus of wood as a function of grain orientation and properties. Holzforschung 2007;61(4):406–13. http://dx. doi.org/10.1515/HF.2007-079. [10] Tankut N, Tankut AN, Zor M. Finite element analysis of wood materials. Drvna Industrija 2014;65(2):159–71. [11] Siau JF, Transport processes in wood, Springer Science & Business Media 2012; 25–30, 44–46. [12] Shi Z, Fu F, Wang S, He S, Yang R. Modification of Chinese fir with alkyl ketene dimer (AKD): processing and characterization. BioResources 2012;8 (1):581–91. [13] Awoyemi L, Westermark U. Effects of borate impregnation on the response of wood strength to heat treatment. Wood Sci Technol 2005;39(6):484–91. http://dx.doi.org/10.1007/s00226-005-0001-5. [14] Kajita H, Imamura Y. Improvement of physical and biological properties of particleboards by impregnation with phenolic resin. Wood Sci Technol 1991;26(1):63–70. [15] Gabrielli CP, Kamke FA. Phenol–formaldehyde impregnation of densified wood for improved dimensional stability. Wood Sci Technol 2010;44(1):95–104. http://dx.doi.org/10.1007/s00226-009-0253-6. [16] Lande S, Westin M, Schneider M. Properties of furfurylated wood. Scand J For Res 2004;19(sup5):22–30. [17] Müller U, Ringhofer A, Brandner R, Schickhofer G. Homogeneous shear stress field of wood in an Arcan shear test configuration measured by means of electronic speckle pattern interferometry: description of the test setup. Wood Sci Technol 2015;49(6):1123–36. http://dx.doi.org/10.1007/s00226-0150755-3. [18] Lakes R. Materials with structural hierarchy. Nature 1993;361(6412):511–5. [19] Dinwoodie J. Timber—a review of the structure-mechanical property relationship. J Microsc 1975;104(1):3–32. http://dx.doi.org/10.1111/j.13652818.1975.tb04002.x. [20] Chinese standard. GB/T 1936.1-2009. Method of testing in bending strength of wood. [21] Chinese standard. GB/T 1936.2-2009. Method for determination of the modulus of elasticity in static bending of wood. [22] Koumoutsakos A, Avramidis S, Hatzikiriakos SG. Radio frequency vacuum drying of wood. I. Mathematical model. Drying Technol 2001;19(1):65–84. http://dx.doi.org/10.1081/DRT-100001352. [23] Gaff M, Gašparík M, Boru˚vka V, Haviarová E. Stress simulation in layered wood-based materials under mechanical loading. Mater Des 2015;87:1065–71. http://dx.doi.org/10.1016/j.matdes.2015.08.128. [24] Srivaro S, Matan N, Lam F. Stiffness and strength of oil palm wood core sandwich panel under center point bending. Mater Des 2015;84:154–62. http://dx.doi.org/10.1016/j.matdes.2015.06.097. [25] Guo X, Cao J, Peng Y, Liu R. Incorporation of microencapsulated dodecanol into wood flour/high-density polyethylene composite as a phase change material for thermal energy storage. Mater Des 2016;89:1325–34. http://dx.doi.org/ 10.1016/j.matdes.2015.10.068.
Contents lists available at ScienceDirect
Results in Physics journal homepage: www.journals.elsevier.com/results-in-physics
Modification of fast-growing Chinese Fir wood with unsaturated polyester resin: Impregnation technology and efficiency Qing Ma, Zijian Zhao, Songlin Yi ⇑, Tianlong Wang Beijing Key Laboratory of Wood Science and Engineering, MOE Key Laboratory of Wooden Material Science and Application, Beijing Forestry University, 35# Qing Hua East Road, Beijing 100083, China
a r t i c l e
i n f o
Article history: Received 24 June 2016 Accepted 19 August 2016 Available online 26 August 2016 Keywords: Unsaturated polyester resin (UPR) Impregnation Vacuum Curing weight percent gain distribution
a b s t r a c t In this study, Chinese Fir was impregnated with unsaturated polyester resin to enhance its properties. Samples 20 mm 20 mm 20 mm in size were split into different sections with epoxy resin and tinfoil and subjected to an impregnation experiment under various parameters. Vacuum degree was 0.04 MPa, 0.06 MPa or 0.08 MPa and vacuum duration was 15 min, 30 min, or 45 min. The results indicated that impregnation weight percent gain is linearly dependent on curing weight percent gain. Vacuum duration appears to have less influence on the curing weight percent gain than vacuum degree, and impregnation was most successful at the transverse section compared to other sections. The optimal impregnation parameters were 30 min modification under 0.08 MPa vacuum followed by 120 min at atmospheric pressure for samples 200 mm 100 mm 20 mm in size. Uneven distribution of weight percent gain and cracking during the curing process suggested that 30 min post-processing at 0.09 MPa vacuum was the most effective way to complete the impregnation process. The sample’s bending strength and modulus of elasticity increased after impregnation treatment. Bending strength after impregnation without post-processing reached 112.85%, but reached 71.65% with vacuum-processing; modulus of elasticity improved 67.13% and 58.28% without and with post-processing, respectively. Ó 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Introduction As one of the most common fast-growing tree species in China, Chinese Fir is an important potential solution to the problematic shortage of wood resources. A kind of softwood, however, it exhibits several drawbacks which limit its practical application [1]. A number of modification technologies have been proposed to enhance the physical, chemical, and biological properties of the wood, such as heat or pressure treatment, surface coating, or impregnation with reactive materials [2,3]. Among the available modification methods, chemical impregnation under vacuum or pressure may be the best way to improve the specific properties of the wood cell wall material [2]. Curing the wood with resins appears to be especially beneficial. Unsaturated polyester resin (UPR) is a popular thermosetting material [4] generally prepared from a saturated acid and an unsaturated acid condensed with dihydric alcohols, which forms a three dimensional structure via cross linking with a vinyl monomer [5,6]. It can be used either as pure resin or with fillers, or reinforced, respectively [6]. As a reactive material, unsaturated polyester resin ⇑ Corresponding author. E-mail address: [email protected] (S. Yi).
has been shown to effectively enhance wood performance [3] and prolong the service life of wood for several years [7]. Mechanical loss is a drawback to impregnation modification in general [8], but less so when utilizing UPR. In this study, a general-grade UPR was prepared in our laboratory, which is prepared from 1,2-propylene glycol (PG), maleic anhydride (MA), and phthalic anhydride (PA) [4]. The primary goal was to establish an impregnation technique well-suited to the modification of Chinese Fir. Fir samples 200 mm 100 mm 20 mm in size was used, as per the typical dimensions of impregnation specimens, and partitioned them in order to observe the relationship among elements distributed within the different partitions. Finite element analysis [9,10] and mathematical analysis were utilized to split the samples into 50 impregnation elements (each 20 mm 20 mm 20 mm in size) which ultimately fell into four distinct categories, as discussed below. Materials and methods Materials The specimens for this study were prepared from three-yearcut, fast-growing Chinese Fir (Cunninghamia lanceolata (Lamb.)
http://dx.doi.org/10.1016/j.rinp.2016.08.017 2211-3797/Ó 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
544
Q. Ma et al. / Results in Physics 6 (2016) 543–548
Hook). The logs were cut from a forest in Guangxi Province, China, with diameter at breast height of 30–40 cm and age of 20–25 years. The logs were sawn into planks, and air-dried to 12–13% moisture content, then the sapwood near the butt portion was processed into samples. The test wood was of an average absolute dry density of 0.33 g/cm3 and porosity of approximately 77.62%. (For details of the Computational Method, refer to Siau’s 1984 study [11]). As mentioned above, the modifier used in this study was a homemade, general-grade UPR, comprised of 1,2-propylene glycol (PG), maleic anhydride (MA) and phthalic anhydride (PA) synthetic materials [4] as well as styrene as a final cross-linking agent. The average molecular weight of the homemade UPR is shown in Table 1. Methods Dimensions of the impregnation specimens The dimensions of the impregnation specimen for performance measurement was 200 mm long by 100 mm wide by 20 mm thick, as shown in Fig. 1(I). To determine the optimal impregnation process parameters for the UPR into Chinese Fir blocks of this size, impregnation was processed with this size in general [12–16] and various parameters were set and tested the consumption of the resin and the blocks accordingly through repeated screening experiments. The results indicated that the impregnation process needed to be simplified, which is the method by building an
analogy with mesh via finite element analysis [9,10] followed by mathematical analysis. When the impregnation process was conducted on the ‘‘big” (200 mm 100 mm 20 mm) block, there were six impregnation directions for UPR as shown in Fig. 1(I). Due to the regular shape of the impregnation specimen, this created fifty smaller samples (or ‘‘elements”) 20 mm 20 mm 20 mm in size; the corresponding mesh is shown in Fig. 1(II). Wood is a highly anisotropic material [17], so the elements themselves featured notable differences in structure [18], as well as mechanical properties. The mechanical property particularly affected in this case, permeability, is mainly influenced by fiber orientation [17,19]. Combining the location of the elements on the specimens, the corresponding impregnation section, and the fiber orientation of the impregnation section, it can be divided into four types of elements which are depicted in Fig. 1(III); the distinct impregnation section and fiber orientation are depicted in Fig. 1(IV). Element (a) was comprised of four impregnation sections: one transverse section, two tangential sections, and one radial section. Elements (b) and (c) were comprised of only three impregnation sections: two tangential sections and a radial section (b) or transverse section (c). Element (d) was comprised of only two tangential sections. Experimental methods The UPR thermosetting resin was first utilized to treat the element blocks via vacuum impregnation with various parameters
Table 1 Average molecular weight of the experimental UPR.
UPR
Mn
Mw
Mz
Mz + 1
Polydispersity
Mz/Mw
Mz + 1/Mw
1189
1377
1663
2056
1.157472
1.208365
1.493613
Mn: number-average molecular weight; Mw: weight-average molecular weight; Mz: Z-average molecular weight.
Fig. 1. Element partition and classification of the specimen.
545
Q. Ma et al. / Results in Physics 6 (2016) 543–548
elasticity Ew was calculated with the Eq. (4). Again, three replicate specimens of each group were tested.
Table 2 Impregnation experiment parameters. Iteration
Vacuum degree Vacuum duration
1
2
3
rbw ¼ ð3Pmax lÞ=ð2bh2 Þ
ð3Þ
0.04 MPa 15 min
0.06 MPa 30 min
0.08 MPa 45 min
Ew ¼ ð23Pl Þ=ð108bh f Þ
3
ð4Þ
(i.e., different vacuum degrees and vacuum durations) (Table 2). There were nine replications run in total to complete the fullfactor experiment. The vacuum duration began from the vacuum degree and was run until reaching the target value. Before the impregnation process, the elements were subjected to surface treatment. Different impregnation sections, in regard to both amount and direction, were treated corresponding to the different locations of elements. Closed sections of each of the four elements were blocked with epoxy resin and tinfoil, oven-dried at 103 °C for 8 h, then oven-dried once more at 80 °C for 6 h after the epoxy resin cured. The weight (W0) was then measured just prior to vacuum impregnation. The samples were then submerged in the modifier and subjected to vacuum treatment. The samples were impregnated in UPR for two more hours at atmospheric pressure after a 30 min vacuum duration, then the surface residues were removed and impregnation weight (W1) was measured. Samples impregnated with UPR were not cured straight through the previous step but instead during a 48-h repose. The impregnated samples were cured at 80 °C for 6 h and 103 °C for 8 h, then their curing weight (W2) was measured; impregnation weight percent gain (WPG1) and curing weight percent gain (WPG2) were determined via Eqs. (1) and (2), respectively.
WPG1 ¼ ðW 1 W 0 Þ=W 0 100%
ð1Þ
WPG2 ¼ ðW 2 W 0 Þ=W 0 100%
ð2Þ
The impregnation method for the subsequent experiments was determined according to these results through variance analysis. After selecting the appropriate method, 200 mm 100 mm 20 mm samples were impregnated as follows. The samples were cured and then sawn into smaller blocks according to the distribution of the elements shown in Fig. 1(II) to obtain the distribution of the curing weight percent gain. Density was measured first according to the weighed mass and the volume determined via the drainage method. Although the 200 mm 100 mm 20 mm samples exhibited favorable impregnation weight percent gain when impregnated via the selected process, the curing process remained uncertain; it was found that the impregnated samples cracked easily during the curing process. To homogenize the weight percent gain of the ‘‘big” samples appropriately, a post-processing technique under a vacuum 0.09 MPa for 30 min was added during which the samples were not submerged in the modifier. The 300 mm 20 mm 20 mm samples was impregnated both with and without post-processing, then measured their respective bending strength and modulus of elasticity according to Chinese standard ISO [20,21]. All the samples were measured under the conditions of equilibrium moisture content of 12%, 20 °C temperature and 65% relative humidity to constant weight. The bending strength of the samples was tested with a mechanical testing machine. The testing speed of the indenter was set to 5 mm/min. The strength rbw was calculated with Eq. (3). Three replicate specimens of each group were tested. The modulus of elasticity of the samples was also tested with a mechanical testing machine at indenter speed of 3 mm/min. After determining the difference of the load bound based on the linear relationship of the load-deformation diagram, the modulus of
3
where Pmax is the fracture load (N), l is the span between the two supporting seats (mm), b is the width of specimen (mm), h is the height of the specimen (mm), P is the difference in the load bound based on the linear relationship, and f is the deformation of the load bound based on the linear relationship. Results and discussion Effects of various vacuum degrees and durations to various elements The impregnation weight percent gains (WPG1) and curing weight percent gains (WPG2) were calculated for all four elements across nine 9 experiments. Because these properties are shortterm, the impregnation weight percent gain was of less importance than the curing weight percent gain. The two weight percent gain values are linearly dependent – nearly proportional, in fact – as is shown in Fig. 2. Accordingly, the subsequent analysis of the various vacuum degrees and durations in terms of the weight percent gain was mainly focused on the curing weight percent gains (WPG2). As mentioned above, nine experiments were run with various vacuum degrees and durations on the four kinds of fast-growing Chinese Fir elements. The resulting variance analysis for the curing weight percent gains (WPG2) is provided in Table 3. As shown in Table 3, compared to duration, vacuum degree exerted a far greater impact on the curing weight percent gain especially to Element (a) and Element (c). The shared characteristic of these two elements was the transverse impregnation section, so it is reasonable to infer that vacuum degree had more significant influence on the transverse section than the tangential or radial sections. This observation can also be attributed to the fact that the longitudinal permeability of softwoods is much greater than the transverse permeability [11]. For samples with certain section orientation, vacuum degree can be controlled to result in excellent permeability. The external vacuum contributes to pressure gradients in the wood [22], and samples with good permeability better
Fig. 2. Correlation between impregnation weight percent gain and curing weight percent gain.
546
Q. Ma et al. / Results in Physics 6 (2016) 543–548 Table 3 Variance analysis of curing weight percent gains (four types of elements).
Element (a) Element (b) Element (c) Element (d)
Vacuum degree Duration Vacuum degree Duration Vacuum degree Duration Vacuum degree Duration
Sum of squares
df
Mean square
F
Sig.
2.321 0.110 0.093 0.003 2.506 0.043 0.022 0.032
2 2 2 2 2 2 2 2
1.161 0.050 0.046 0.001 1.251 0.022 0.011 0.016
39.586 0.185 2.719 0.076 125.485 2.159 1.292 1.944
0.000 0.832 0.080 0.927 0.000 0.130 0.287 0.158
generate a gradient receptive to impregnation; it was found that an increase in vacuum degree was more beneficial to the pressure gradient than an increase in vacuum duration. The function of vacuum duration was fairly negligible, especially for samples with good permeability. This can be attributed to the favorable impregnation efficiency facilitated by appropriate vacuum degree having ensured favorable permeability.
weight percent gain, but the respective influence of these three methods on the four elements was inconsistent. Lengthier duration did not, per the expectations, result in high curing weight percent gain, so Method 9 was excluded from further analysis. Method 8 resulted in better impregnation effect than Method 7, so Method 8 (impregnation under 0.09 MPa for 30 min followed by 120 min at atmospheric pressure) was employed for subsequent experiments.
Impregnation method selection Curing weight percent gain distribution of the samples
Again, vacuum degree exerted greater influence on the curing weight percent gain after impregnation than vacuum duration, so the next step to attempt was to ensure similar curing weight percent gain with the briefest vacuum duration possible. An appropriate weight percent gain is a guarantee of favorable properties in the modified material, thus, the experiments were aimed to sink an impregnation method that resulted in high curing weight percent gain. Among the four types of elements, Element (a) exhibited the maximum weight percent gain followed by Element (c). The UPR distribution in these two elements was similar, as well; the modifier nearly penetrated both. When the two elements were returned to the big sample, nearby Elements (b) and (d) naturally transformed into Elements (a) and (c); this made the UPR distribution in the big sample relatively more homogeneous than the UPR distribution in the individual elements. As shown in Fig. 3, a stepwise improvement was observed in average curing weight percent gain which further implied that vacuum duration had little influence on the resulting properties of the sample. The final three tested impregnation methods (vacuum degree of -0.08 MPa) were preferable in regard to the curing
Samples 200 mm 100 mm 20 mm in size were impregnated with a pre-determined initial weight to measure only the curing weight percent gains of the whole blocks. According to the respective WPG2 values of the different elements, the distribution of the curing weight percent gain on the samples was expectedly inhomogeneous. To determine the distribution on the whole breadth of the sample, the sample was split into smaller blocks approximately 20 mm 20 mm 20 mm in size again. Because the initial weights of these blocks were unknown, the relationship between their density and curing weight percent gain was ascertained based on the smaller blocks observed in previous experiments. The epoxy resin and tinfoil that were used to block the impregnation process were removed prior to measuring the small samples’ weights and volumes via water displacement method. The relationship between the curing weight percent gain and the density (D) of the samples is shown in Fig. 4. It was found that density was linearly associated with curing weight percent gain
Fig. 3. Average curing weight percent gain of the four elements among nine impregnation methods.
Fig. 4. Relationship between curing weight percent gain and density of the samples.
Q. Ma et al. / Results in Physics 6 (2016) 543–548
(R2 = 0.94507). The approximate curing weight percent gain of the sawn blocks was calculated with the Eq. (5).
WPG2 ¼ ðD 0:32814Þ=0:00416 100%
ð5Þ
The distribution of the curing weight percent gain of the samples was uneven, as shown in Fig. 5; curing weight percent gain decreased gradually as the distance from the transverse section increased. The lower end in Fig. 5(II) can be attributed to outflow of the resin from the sample during the curing process. This created non-uniform distribution of the properties of the modified wood as well as severe end cracking during the curing process, which necessitated post-processing to reduce and homogenize the weight percent gain. Vacuum processing can be utilized to extract a portion of the modifier in the cell lumen, so a post vacuum processing of 0.09 MPa for 30 min was added to reduce the risk of cracking in the proposed impregnation technique. It should be noted that because of the differences of the samples themselves, the weight percent gains of the two test groups were not identical but did have the same magnitude (i.e., comparison of the two groups was still reasonable). The distribution of the curing weight percent gain of the samples with post-processing was shown in Fig. 6. The curing weight percent gain of the sample ends was still higher than the inner parts after post-processing, but the homogeneity of the samples was improved significantly and there was no cracking in the sample. By comparison among the distributions shown in Figs. 5 and 6, plus the curing results, it was determined that the optimum impregnation process included a vacuum degree of 0.08 MPa for 30 min followed by 120 min at atmospheric pressure under the modifier, then post-processing with a vacuum degree of 0.09 MPa for 30 min without the modifier.
547
Modification effect on bending strength and modulus of elasticity Bending strength and modulus of elasticity are two important properties of wood materials that are likewise affected by different modification methods [23–25]. Fig. 7 shows the average bending strength and modulus of elasticity of the three impregnation states tested in histogram form; the error bars are also marked. It was found that both the bending strength and modulus of elasticity of the samples increased significantly with the two modification methods it was used – especially for samples impregnated
Fig. 7. Bending strength and modulus of elasticity of samples with different impregnation processes.
Fig. 5. Distribution of curing weight percent gain of the samples without post-processing.
Fig. 6. Distribution of curing weight percent gain of post-processed samples.
548
Q. Ma et al. / Results in Physics 6 (2016) 543–548
without post-processing – but that greater increase in properties came with greater volatility. Compared to the non-impregnated samples, the bending strength of the samples impregnated without post-processing increased by 112.85%; the improvement after impregnation with post-processing was 71.65%. There was less difference in terms of modulus of elasticity, with improvements of 67.13% and 58.28%, respectively. Although the effects of the impregnation without post-processing were better, the deviations were larger. The curing weight percent gain of Sample (c) was greater than that of Sample (b), but the homogeneity of the distribution of Sample (b) was better. The homogeneity of UPR distribution, which is also directly related to the stability (i.e., performance) of the impregnated wood, was better in the post-processed sample as well. Conclusion The most notable conclusions of this study can be summarized as follows. 1 Impregnation weight percent gain is linearly dependent upon curing weight percent gain. Vacuum degree exerted greater effect on the curing weight percent gain (and thus the impregnated weight percent gain) than vacuum duration. Elements with transverse sections manifested greater curing weight percent gain and were affected more by vacuum degree than elements with other sections. 2 The optimal modification process included impregnation at 0.08 MPa vacuum for 30 min, 120 min at atmospheric pressure with submersion in the modifier, and post-processing under a 0.09 MPa vacuum for 30 min without the modifier. The distribution of the curing weight percent gain in postprocessed samples was optimal compared to samples that were not post-processed. 3 The impregnation modification facilitated substantial increase in bending strength and modulus of elasticity; the improvement was greater without vacuum processing than that with postprocessing (112.85% increase in bending strength without post-processing; 71.65% increase with post-processing). Modulus of elasticity improved to approximately equal extent without or with vacuum treatment (67.13% and 58.28%, respectively). Acknowledgements The paper was supported by the Special Fund for Forest Scientific Research in the Public Welfare (201404502) and the Key Discipline Construction Projects of Beijing—‘‘Efficient utilization of fast growing wood”. References [1] Cao Y, Lu J, Huang R, Jiang J. Increased dimensional stability of Chinese fir through steam-heat treatment. Eur J Wood Wood Prod 2012;70(4):441–4. http://dx.doi.org/10.1007/s00107-011-0570-y.
[2] Mathias LJ, Lee S, Wright JR, Warren SC. Improvement of wood properties by impregnation with multifunctional monomers. J Appl Polym Sci 1991;42 (1):55–67. http://dx.doi.org/10.1002/app.1991.070420107. [3] Mahmoud A, Eissa A, Omar M, El-Sawy A, Shaaban A. Improvements of white pine wood properties by impregnation with unsaturated polyesters in admixture with styrene. J Appl Polym Sci 2001;82(6):1410–6. http://dx.doi. org/10.1002/app.1978. [4] Zhang C, Huang JY, Liu SM, Zhao JQ. The synthesis and properties of a reactive flame-retardant unsaturated polyester resin from a phosphorus-containing diacid. Polym Adv Technol 2011;22(12):1768–77. http://dx.doi.org/ 10.1002/pat.1670. [5] Haq MIU. Applications of unsaturated polyester resins. Russ J Appl Chem 2007;80(7):1256–69. http://dx.doi.org/10.1134/S1070427207070464. [6] Fink JK. Reactive polymers fundamentals and applications: a concise guide to industrial polymers. William Andrew 2013;1–6:31–3. [7] Islam MN, Ando K, Yamauchi H, Kobayashi Y, Hattori N. Comparative study between full cell and passive impregnation method of wood preservation for laser incised Douglas fir lumber. Wood Sci Technol 2008;42(4):343–50. http:// dx.doi.org/10.1007/s00226-007-0168-z. [8] Sun W, Shen H, Cao J. Modification of wood by glutaraldehyde and poly (vinyl alcohol). Mater Des 2016;96:392–400. http://dx.doi.org/10.1016/ j.matdes.2016.02.044. [9] Nairn JA. A numerical study of the transverse modulus of wood as a function of grain orientation and properties. Holzforschung 2007;61(4):406–13. http://dx. doi.org/10.1515/HF.2007-079. [10] Tankut N, Tankut AN, Zor M. Finite element analysis of wood materials. Drvna Industrija 2014;65(2):159–71. [11] Siau JF, Transport processes in wood, Springer Science & Business Media 2012; 25–30, 44–46. [12] Shi Z, Fu F, Wang S, He S, Yang R. Modification of Chinese fir with alkyl ketene dimer (AKD): processing and characterization. BioResources 2012;8 (1):581–91. [13] Awoyemi L, Westermark U. Effects of borate impregnation on the response of wood strength to heat treatment. Wood Sci Technol 2005;39(6):484–91. http://dx.doi.org/10.1007/s00226-005-0001-5. [14] Kajita H, Imamura Y. Improvement of physical and biological properties of particleboards by impregnation with phenolic resin. Wood Sci Technol 1991;26(1):63–70. [15] Gabrielli CP, Kamke FA. Phenol–formaldehyde impregnation of densified wood for improved dimensional stability. Wood Sci Technol 2010;44(1):95–104. http://dx.doi.org/10.1007/s00226-009-0253-6. [16] Lande S, Westin M, Schneider M. Properties of furfurylated wood. Scand J For Res 2004;19(sup5):22–30. [17] Müller U, Ringhofer A, Brandner R, Schickhofer G. Homogeneous shear stress field of wood in an Arcan shear test configuration measured by means of electronic speckle pattern interferometry: description of the test setup. Wood Sci Technol 2015;49(6):1123–36. http://dx.doi.org/10.1007/s00226-0150755-3. [18] Lakes R. Materials with structural hierarchy. Nature 1993;361(6412):511–5. [19] Dinwoodie J. Timber—a review of the structure-mechanical property relationship. J Microsc 1975;104(1):3–32. http://dx.doi.org/10.1111/j.13652818.1975.tb04002.x. [20] Chinese standard. GB/T 1936.1-2009. Method of testing in bending strength of wood. [21] Chinese standard. GB/T 1936.2-2009. Method for determination of the modulus of elasticity in static bending of wood. [22] Koumoutsakos A, Avramidis S, Hatzikiriakos SG. Radio frequency vacuum drying of wood. I. Mathematical model. Drying Technol 2001;19(1):65–84. http://dx.doi.org/10.1081/DRT-100001352. [23] Gaff M, Gašparík M, Boru˚vka V, Haviarová E. Stress simulation in layered wood-based materials under mechanical loading. Mater Des 2015;87:1065–71. http://dx.doi.org/10.1016/j.matdes.2015.08.128. [24] Srivaro S, Matan N, Lam F. Stiffness and strength of oil palm wood core sandwich panel under center point bending. Mater Des 2015;84:154–62. http://dx.doi.org/10.1016/j.matdes.2015.06.097. [25] Guo X, Cao J, Peng Y, Liu R. Incorporation of microencapsulated dodecanol into wood flour/high-density polyethylene composite as a phase change material for thermal energy storage. Mater Des 2016;89:1325–34. http://dx.doi.org/ 10.1016/j.matdes.2015.10.068.