Article pubs.acs.org/IECR
Characterization of Modified Phenol Formaldehyde Resole Resins Synthesized in Situ with Various Boron Compounds Yubo Chai,†,‡ Junliang Liu,† Yong Zhao,‡ and Ning Yan*,‡ †
Research Institute of Wood Industry, Chinese Academy of Forestry, Beijing, China 100091 Faculty of Forestry, University of Toronto, 33 Willcocks Street, Toronto, Ontario, Canada M5S 3B3
‡
S Supporting Information *
ABSTRACT: In this study, three different boron compounds were used together with alkaline catalyst to synthesize phenol formaldehyde (PF) resole resins in situ. The resin curing behavior, molecular structure, bonding performance, and properties of resin-impregnated wood were investigated. Results showed that boron compound-modified PF resins had a lower degree of polymerization than the control PF resin made in the laboratory. The curing kinetics, molecular structure, and functional groups of the modified resins varied depending on the type of boron compounds used. The thermal stability of cured modified PF resins was slightly lower than that of laboratory-made control PF resin. Boron compound-modified PF resins exhibited dry and wet bonding strengths comparable to the those of the laboratory-made control PF resin. Wood impregnated using modified PF resins had comparable dimensional stability, mechanical properties and improved fire resistance than the wood impregnated using lab made control PF resin regardless the type of boron compounds used.
1. INTRODUCTION Phenol formaldehyde (PF) resole resins are widely used as wood adhesives for producing various wood-based products for exterior applications, such as oriented strand board (OSB), plywood, and laminated veneer lumber (LVL). PF resole resins are obtained via phenol and formaldehyde undergoing addition and condensation polymerization reactions under alkaline conditions. PF resole resins demonstrated excellent bonding performance, water resistance, and durability.1,2 Compared to amino plastic resins such as urea−formaldehyde (UF) resin, PF resin had a higher bonding efficiency but a slower cure speed. The slow resin cure rate would lead to lower production efficiency and higher production costs. This could be a major factor limiting the application of PF resins as impregnation resins and in the production of particleboard and medium-density board.3 Many studies have reported on the attempts to accelerate the PF resin’s curing rate for wood-adhesive applications. Bivalent metal oxides,4,5 carbonates,6−9 and esters3,10−14 were found to be effective in the acceleration of PF resin curing. The acceleration mechanism was associated with the bivalent metallic ions forming complexes with phenols and formaldehyde. The acceleration effect was determined by the metal exchange rate in solution and the stability of the complexes.4,5 The cure acceleration mechanism of carbonates and esters was found to be complex. So far, there was no conclusive mechanism presented in the literature. Organic or inorganic carbonates and esters accelerated the PF resin curing process differently. Organic carbonates both increased the curing speed © 2016 American Chemical Society
and improved the strength of the cured resin, whereas inorganic carbonates accelerated the curing but contributed little to the hardened network strength. The reactivity of phenolic rings was reported to increase by adding carbonates or esters. Meanwhile, carbonates and esters also affected the resin structures, such as the distribution of ortho−para to para−para linkages and resin curing kinetics. In addition, the mechanism involving the formation of an ortho carbonyl or carboxyl group by the phenate ion on the phenolic ring was reported. These groups rearranged and shifted to sites other than the ortho position. Anhydride-like bridges formed temporarily and contributed to oligomeric structure with or without methylene bridges. As a result, the PF resin cure process was accelerated.3,6−14 With the increasingly stringent demand for fire safety, it is required that wood and wood composites be treated with fire retardants such as boron compounds15 to achieve the necessary flame-resistant properties. To make wood composite panels with desirable termite- and decay-resistant properties, boronbased chemicals, such as boric acid, borax, zinc borate, ammonium pentaborate, and mixtures of boron compounds, are routinely added.15−18 The boron-based chemicals are either directly applied to wood composites or blended with wood strands or adhesives prior to hot pressing. Boron-based chemicals affected the resins curing kinetics, and the adverse Received: Revised: Accepted: Published: 9840
June 4, 2016 August 18, 2016 September 1, 2016 September 1, 2016 DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research
heated to 85 °C and stayed at this temperature for an hour after the remaining amount of sodium hydroxide was added. The resulting resin mixture was cooled to room temperature and used for subsequent analysis as the laboratory-made control PF resin. PF resins with different boron compounds were synthesized following this procedure. Boric acid, borax, and zinc borate were added to the 40% sodium hydroxide solution during resin synthesis. The total amount of boron compounds added was 2 wt % of the final resins. The resulting resins were denoted as PF-BA, PF-BX, and PF-ZB for boric acid-modified PF, boraxmodified PF, and zinc borate-modified PF resins, respectively. All chemicals were supplied by Caledon Laboratory Chemicals, Canada, and were used as received. 2.2. Resin Characterization. 2.2.1. pH, Solids Content, Gel Time, and Molecular Weight. Both the pH and viscosity of the resins were tested at 25 °C. A Brookfield rotary viscometer was used for the viscosity measurements. ASTM D 3529 was followed to determine the solid content of the resin. The gel time was measured similarly to what was reported in our previous studies.35,36 Resin (5 g) in a test tube (16 mm diameter) was heated in an oil bath at 100 ± 1 °C. The gel time was recorded as the time duration from the test tube entering the oil bath to the resin starting to gel by forming a string when lifting a glass rod from the resin. The gel time is reported as the average value from three replicate measurements. Matrix-assisted laser desorption/ionization-time-of-flight/ time-of-flight spectrometer (MALDI-TOF/TOF, Applied Biosystems, Framingham, MA, USA) was used to evaluate the weight-average molecular weight (Mw), number-average molecular weight (Mn), and polydispersity index (Mw/Mn) of the PF resins. Low linear mode with a mass range from 60 to 2000 Da and a laser intensity of 6000 arbitrary units was used to obtain resin mass spectra. Trifluoroacetic acid (0.1%) and DHB (2,5-dihydroxybenzoic acid) in 50% aqueous ethanol were mixed to form a 10 mg/mL concentration of matrix solution. One microliter of the premixed resin and the matrix mixture in a ratio of 1:4 (v/v) was air-dried on the MALDI target, and each spectrum was based on the total number of ions collected from 500 laser pulses. Mané et al.’s methods were used to calculate the average molecular weight and polydispersity index.29 2.2.2. Resin Curing Behavior and Curing Kinetics. The resin curing behavior was investigated using high-pressure pans (DSC-Q1000, TA Instruments, USA). Dynamic scans were conducted with multiple heating rates of 5, 10, 15, and 20 °C/ min, starting from room temperature to 250 °C. The curing activation energy was calculated using the Kissinger method31
effect of borates on the mechanical and physical properties of the wood composite panels bonded using PF resins was frequently reported, including a lower internal bond (IB), increased thickness swelling (TS), and a decreased wood failure rate.19−24 It was due to the resin started to gel caused by the boron ion reacting with the methylol groups before an effective bond can be developed. Meanwhile, the low solubility and poor dispersion of boron compounds in PF resol resin typically associated with post-blending methods tend to downgrade the mechanical properties of the resulting wood composites. Adding hydroxyl-rich organic flow agents such as PEG to the PF resins could potentially minimize the adverse effect caused by borates;23,24 however, it adds extra cost to the manufacturing of the PF resin and wood composite panels. Instead of using boron compounds as additives or physically post-blending with PF adhesives, boron compounds could be chemically incorporated into phenolic resin structures and improve their physical, thermal, and mechanical properties.25−28 The resulting arofene phenolic resins with superior heat, chemical, and water resistance could be found in many applications, including brake pads, metal plates, grinding wheels, lamination, and so on. The reaction mechanism to produce PF resin is step-growth polymerization, therefore, subtle variations in reaction conditions could have profound effect on the resulting resins’ structure and performance.2 The recent structural investigation of boron-modified phenol formaldehyde resins showed that the PF resin was cross-linked via tetragonally coordinated boron species (BIV). The boronbased cross-links included monoester and diester complexes represented by six-membered spirocyclic borate anions. During the thermal curing process, the monoester and diester borate complexes underwent additional transformation in which the spirocyclic borate anions were more tightly incorporated into the polymer matrix via additional N-type cross-links. A 11B−11B double-quantum correlation MAS NMR experiment showed that the majority of the monoester and diester borate complexes (ca. 80%) were uniformly distributed within and effectively isolated by the polymer matrix, with an average 11 B···11B interatomic distance greater than 6 Å. The final properties of the boron-modified PF resins were enhanced with increased cross-link density, which will be used as novel binders for soft magnetic composites.28 However, there are few literature-reported PF resol resin in situ syntheses with boron compounds for wood-adhesive applications. It is also unknown how the boron compounds affect the resin’s molecular structure and curing behavior. There are no previous studies investigating the properties of wood impregnated using boron-compound-modified PF resins. Therefore, in this study, three different boron compounds boric acid, borax, and zinc boratewere added to an alkaline catalyst to synthesize in situ PF resole resin for wood-adhesive applications. The resulting modified PF resins were used to impregnate low-density and low-quality plantation poplar wood. The resin molecular structure, curing behavior, resin bonding performance, and properties of the impregnated wood were investigated.
⎛ ϕ ⎞ ⎛ RA ⎞ E 1 ⎟ + ln⎜ ln⎜⎜ 2 ⎟⎟ = − ⎝ E ⎠ R T p ⎝ Tp ⎠
(1)
where ϕ is the heating rate (K/s), Tp is the peak temperature (in Kelvin) at the given heating rate, A is the pre-exponential factor, R is the ideal gas constant, and E is the activation energy. A randomly selected sample with three replicates was used for the testing. There was less than 1 °C variation among the replicates for the onset and peak temperature measurements. 2.2.3. Thermal Gravimetric Analysis (TGA) of the Cured Resins. All of the resins were first oven-cured at 80 °C for 48 h prior to being ground into fine powders smaller than a 100mesh screen and were tested in a thermogravimetric analyzer
2. EXPERIMENTAL SECTION 2.1. Resin Synthesis. The solution mixture of phenol, 37% formaldehyde, and 40% sodium hydroxide was stirred at 40 °C for 15 min in a three-necked flask with a reflux condenser first. The mixture was then heated to 65 °C in 30 min and was kept at this temperature for another 30 min. The mixture was then 9841
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research
not be required to acquire meaningful ratios. However, differences in the nuclear Overhauser effect (NOE) and relaxation time should be considered. For both polymer and small molecules, the signal-to-noise ratio might affect the accuracy if a species or end group has an exceptionally low concentration. 2.2.7. PF-Resin-Impregnated Wood and Its Properties. 2.2.7.1. Wood Impregnation. Samples with dimensions of 400 mm × 40 mm × 40 mm (longitudinal direction × radial direction × tangential direction) were cut from low-density plantation poplar wood (Populus ussuriensis). The laboratorymade control PF resin, boric acid-modified PF (PF-BA), boraxmodified PF (PF-BX), and zinc borate-modified PF (PF-ZB) resins were diluted to 30% solids content for wood impregnation. Wood samples were treated using the vacuum-pressure method. Each impregnation condition contained 15 wood samples as replicates. The treatment process was as follows: −0.095 MPa vacuum was first applied to the wood samples for 40 min, and then the wood samples were impregnated with PF resin solutions under the vacuum condition. After the vacuum was released, 0.8 MPa of pressure was applied and held for 2 h. Once the treatment was completed, the impregnated wood samples were oven-dried at 60 °C for 72 h and then at 80 °C for 72 h and at 100 °C for 48 h. The weight percent gain (WPG) was calculated using eq 2
(TGA, Q500, TA, USA). A platinum pan holding 10 mg of each cured sample was heated from room temperature to 800 °C at 10 °C/min under an N2 atmosphere. 2.2.4. Evaluation of Bonding Strength. Wood strips with dimensions of 3 mm thickness, 25.4 mm width, and 108 mm length were cut from birch veneer. The wood grain was parallel to the wood strips’ length direction. The two-layer birch wood veneer samples were glued together using four different types of resins, i.e., laboratory-made control PF resin, boric acidmodified PF (PF-BA), borax-modified PF (PF-BX), and zinc borate-modified PF (PF-ZB) resins. The resins were applied to one side of the birch strip and covered an area of 25.4 mm × 25.4 mm (Figure 1) with a controlled spread rate of 0.025−
Figure 1. Lap shear specimen dimensions.
0.035 g/cm2 on the basis of the solids. Another birch strip without resins was overlapped with the area of the birch strip coated with resins to form the two-layer lap shear sample. The lap shear sample was then subjected to hot pressing at 160 °C with thickness control of 4.5 mm for 300 s. After hot pressing, the samples were cooled and conditioned and tested for shear strength on a Zwick universal test machine (Zwick/Z100, Zwick Roell Group, Germany). The standard lap shear test methods in ASTM D5868 were followed. The crosshead speed was set at 1.3 mm/min. The average value of the shear strength is reported on the basis of a minimum of 10 replicates. 2.2.5. Water Resistance Tests. To evaluate the durability of adhesives, a water-soaking-and-drying (WSAD) test and a boiling-water test (BWT) were conducted. The test procedures were based on voluntary standard PSl-95 published by the U.S. Department of Commerce through the Engineered Wood Association (Tacoma, WA). For the WSAD test, the lap shear samples were first soaked in water at room temperature for 24 h and then air-dried with good circulation (fume hood) at room temperature for another 24 h, after which the samples were tested for shear strength. For the BWT test, the samples were first boiled in water for 4 h and then oven-dried at 63 ± 2 °C for 20 h. After that, the samples were boiled again in water for another 4 h, cooled using tap water, and then immediately tested for shear strength when they were still wet. The resulting shear strength was defined as the BWT/W strength. 2.2.6. Resin Molecular Structure. Sample preparation was conducted according to previous studies.31 The PF resins were first dissolved in D2O. The liquid-state 13C NMR spectra of these samples were acquired using a Unity 500 spectrometer. The experimental conditions are the following: a pulse angle of 60° (8.3 μs), a relaxation delay of 10 s, and gated Waltz-16 1H decoupling during the acquisition period. Each spectrum was accumulated from about 40 000 scans. Sodium 2,2-dimethylsilapentane-5-sulfonate (DSS) was used as the internal standard for the measurement of the 13C chemical shifts. Usually quantitative 13C NMR requires long relaxation delays to achieve sufficient signal-to-noise ratios and results in extremely long acquisition times. Gated decoupling 13C NMR using only a short delay between pulses could be used for the ratio determination of small molecules and for end-group analysis of polymer. If certain precautions are taken, then long delays will
WPG =
G2 − G1 × 100% G1
(2)
where G1 is the weight of the oven-dried sample before treatment and G2 is the weight of the oven-dried sample after treatment. 2.2.7.2. Properties of Impregnated Wood. On the basis of the requirements of different national test standards, the treated and untreated wood samples were cut into different dimensions. According to GB/T1934.2-2009, the dimensional stability of the samples was evaluated by measuring the moisture absorption (MA) and antiswelling efficiency (ASE), which were calculated using eqs 3 and 4, respectively, m − m0 MA = 1 × 100% m0 (3) ASE =
SC − ST × 100% SC
(4)
where MA is the moisture absorption rate, m1 is the weight of the sample after moisture conditioning, and m0 is the weight of the oven-dried sample. ASE is the antiswelling efficiency of samples, SC is the volume-swelling rate of untreated samples, and ST is the volume-swelling rate of treated samples. In addition, the air density, modulus of elasticity (MOE), modulus of rupture (MOR), and oxygen index of untreated and treated wood samples were tested according to GB/T 19332009, GB/T 1936.1-2009, GB/T 1936.2-2009, and GB/T 2406-2009, respectively.
3. RESULTS AND DISCUSSION 3.1. Resin Properties. The basic properties of different resins are shown in Table 1. The resins’ solid contents ranged from 52.26 to 52.82%. The pH value of resins ranged from 10.04 to 10.42. The viscosity of resins ranged from 401.2 to 586 cps. The laboratory-made control PF resin had a slightly higher 9842
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research Table 1. Properties of the Resins solids content (%)
PF control PF-BA
52.75
pH
viscosity (cps)
Mw (Da)
Mn (Da)
PI
10.14
468.5
1347
727
1.85
52.64
10.04
401.2
1048
654
1.60
PF-BX
52.26
10.42
586.2
1178
668
1.76
PF-ZB
52.82
10.11
535.6
1222
708
1.73
resin type
gel time at 100 °C 15 min 16 s 20 min 53 s 19 min 28 s 16 min 25 s
average molecular weight and polydispersity index compared to those of the boron compound-modified PF resins. Although all of the resins were cooked following the same procedure, the average molecular weight of the boron compound-modified PF resins varied depending on the type of boron compounds. The gel time of PF resins with boron compounds was generally longer than that of the laboratory-made control PF resin. The boric acid-modified PF (PF-BA) had the longest gel time, followed by the borax-modified PF (PF-BX) and then the zinc borate-modified PF (PF-ZB). 3.2. Resin Curing Behavior and Curing Kinetics. All of the resins demonstrated a single exothermic peak at different heating rates in their DSC curing curves, indicating that condensation reactions dominated the resin curing process (Figure 2). The addition reactions associated with the formation of methylol phenol and oligomers may have been completed in large proportion during the resin synthesis stage.32−36
Figure 3. DSC curing curve of the laboratory-made control PF resin at different heating rates.
suggesting the boron compound-modified PF resins will start to cure at a lower temperature than the laboratory-made control PF resin without boron compounds. There was no significant difference in the onset temperature of PF-BA, PFBX, and PF-ZB. The peak temperature of boric acid-modified PF, borax-modified PF, and zinc borate-modified PF resin was close to that of the laboratory-made control PF resin. The curing activation energy, pre-exponential factor, and enthalpy of different resins are listed in Table 2. Boric acidmodified PF resin and borax-modified PF resin had a lower curing activation energy than did the laboratory-made control PF resin. It suggested that boric acid-modified PF and boraxmodified PF resin were more thermodynamically favorable for curing than was the laboratory-made control PF resin. However, the pre-exponential factor of boric acid-modified PF resin and borax-modified PF resin was lower than that of the laboratory-made control PF resin, indicating that boric acid and borax retarded the resin curing reaction kinetic rate. Similar results were observed for PF resins postblended with boron compounds.15−19 On the opposite side, zinc borate-modified PF resin demonstrated a slightly higher curing activation energy than did the laboratory-made control PF resin. However, the higher pre-exponential factor of zinc borate-modified PF resin indicated that zinc borate slightly accelerated the resin curing reaction kinetic rate when there was sufficient thermal energy available. This is consistent with previous studies in which the PF resin cure speed was accelerated by bivalent metallic ions.4,5 In addition, boric acid-modified PF, borax-modified PF resin, and zinc borate-modified PF resin demonstrated a lower curing reaction enthalpy than did the laboratory-made control PF resin. A similar trend was observed for PF resins postblended with ammonium pentaborate and wood powder as additives.16,17 Because the enthalpy of the resin curing reaction could be used as an indicator of the completion of the reaction, our results suggested that the curing degree of the boron compound-modified PF resins was lower than that of the laboratory-made control PF resin, and the cross-linked structure of boron compound-modified PF resins would differ from that of the laboratory-made control resin. Meanwhile, the endothermic effect from the boron compounds could inhibit the curing process of the resulting PF resins and lead to an unchanged or slightly higher curing peak temperature and lower curing reaction enthalpy compared to those of the laboratory-made control PF resin. The detailed cross-linked
Figure 2. DSC curing curve of different resins (10 °C/min).
It can be found that the curing curves of the resin shifted to higher temperature when the heating rate increased (Figure 3). The onset and peak temperatures of all of the resins increased with the increasing heating rate, which was consistent with previous studies on PF resins.32−36 Because the heating rate did not affect the actual resin curing temperatures, for further comparison the onset and peak temperatures obtained from dynamic DSC runs at different heating rates were extrapolated to the heating rate of zero.32−36 The curing temperatures of different resins with variable heating rates are listed in Table 2. Boron compound-modified PF resins exhibited different cure temperatures from the laboratory-made control PF resin. Boric acid-modified PF (PF-BA), borax-modified PF (PF-BX), and zinc borate-modified PF (PF-ZB) resin had a lower onset temperature than the laboratory-made control PF resin, 9843
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research Table 2. Curing Characteristics of Different Resins
heating rate (°C/min) 0 5 10 15 20 curing activation energy (kJ/mol) pre-exponential factor (/s) enthalpy ΔH (J/g)
laboratory-made control PF resin
boric acid-modified PF resin
onset temp (°C)
peak temp (°C)
onset temp (°C)
peak temp (°C)
onset temp (°C)
peak temp (°C)
onset temp (°C)
peak temp (°C)
110.4 115.7 122.5 129.3 133.1
129.7 136.7 145.7 151.2 159.7
98.1 105.5 116.7 124.1 131.5
131.3 138.0 149.4 156.3 162.7
101.7 108.4 116.4 124.5 129.8
127.6 137.7 147.7 158.4 168.0
97.0 103.3 116.0 118.2 128.6
129.2 133.8 145.2 151.7 154.5
borax-modified PF resin
zinc borate-modified PF resin
83.39
77.22
60.83
86.98
1.39 × 107 300.8
1.84 × 106 260.3
1.31 × 104 249.9
4.61 × 107 289.2
Figure 4. Liquid-state 13C NMR spectrum of the laboratory-made control PF resin.
Figure 5. Liquid-state 13C NMR spectrum of the boric acid-modified PF resin.
structure of boron compound-modified PF resins is currently under investigation. 3.3. Liquid-State 13C NMR Spectra of Different PF Resins. The liquid-state 13C NMR spectra of the laboratorymade control PF resins and boron-modified PF resins are shown in Figures 4−7. The chemical shifts for the resins are assigned according to the literature.31,37−40 The chemical shifts at 166−168 ppm representing carbonyl groups demonstrated similar intensities in the laboratory-made control PF resin and boron compound-modified PF resins. The
formation of carbonyl groups was attributed to the oxidation of the phenolic rings to quinone structures during and after resin synthesis. Phenoxy carbons had chemical shifts at 153.5−163.9 ppm. The changes in polarity of the phenoxy groups would cause changes in the resin’s pH. As a result, the chemical shifts of phenoxy carbons could vary. The chemical shifts between 158.0 and 159.5 ppm were from the para-alkylated phenolic groups; their intensities were higher than those of chemical shifts of ortho-alkylated phenolic groups between 153.5 and 157.2 ppm. Because of the changes in polarity of the phenoxy 9844
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research
Figure 6. Liquid-state 13C NMR spectrum of the borax-modified PF resin.
Figure 7. Liquid-state 13C NMR spectrum of the zinc borate-modified PF resin.
a chemical shift at 81.7 was absent in all of the resins. The peak around 83−86 ppm was from formaldehyde oligomers. The chemical shifts of 92.1−93.7 ppm due to the ortho- and paraPh−CH2OCH2OCH2OH were present in all of the resins. There were no peaks found between 69 and 73 ppm, suggesting that no methylene ether bridges are formed between the phenolic rings in the PF resins synthesized under the current experimental conditions with and without boron compounds. It was consistent with the results in a previously published article.31 The chemical shifts at 29.2−29.6 ppm for the ortho−ortho linkage of methylene groups (Ph−CH2−Ph) between phenols were not observed for all of the resins, whereas the ortho−para and para−para linkages of methylene groups with chemical shifts at 34−38 ppm and 39−43 ppm were observed. The presence of sodium hydroxide in PF resin synthesis did not favor the formation of the ortho−ortho methylene linkages. The peaks at 60.0−67.0 ppm were associated with the methylol groups (Ph−CH2OH) in the PF resins, with 60.0−63.0 ppm for ortho linkage and 64.0−67.0 ppm for the para linkage. Methylol groups affected the resin stability and kept the polymeric molecules soluble in the aqueous medium. Although there were no significant differences in the chemical shifts between the laboratory-made control PF resin
groups, the chemical shifts of phenoxy carbons including both para and ortho carbons varied with the changes in the resin’s pH. The peaks at 114−119 ppm related to unsubstituted ortho phenolic carbon and the peaks at 120−124 ppm associated with unsubstituted para phenolic carbon were observed in the laboratory-made control PF resin and boron compoundmodified PF resins. The peaks of unsubstituted para phenolic carbon of boric acid-modified PF resin shifted to downfield. The possible reason could be attributed to the lower pH value from the neutralization between boric acid and sodium hydroxide. The boron compound-modified PF resins had a different peak intensity of unsubstituted para phenolic carbon compared to that of the laboratory-made control PF resin, indicating that boron compounds affected the reaction between formaldehyde and the para aromatic carbon of phenol. The chemical shift at 48.7 ppm was associated with methanol (CH3OH). Methanol is known as a stabilizer in formaldehyde solution. Meanwhile, the Cannizzaro reaction of formaldehyde under strong alkaline conditions could be the reason for the formation of methanol in PF resin synthesis. The methyl carbon in hemiformal (CH3OCH2OH) has a chemical shift at 55.2 ppm.. The chemical shift of the methylene carbon was at 89.3 ppm (CH3OCH2OH). The unreacted formaldehyde with 9845
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research Table 3. Ratios of the Relevant Functional Groups Relative to Phenolic Rings sample laboratory-made control PF boric acid-modified PF borax-modified PF zinc borate-modified PF
para/ortho (−CH2OH)
p−p/o−p link (−CH2−)
0.38
1.03
0.13
3.30
0.26
1.67
0.21
3.99
0.15 0.18
2.38 1.69
0.19 0.15
4.70 3.44
unsubstituted/substituted hydrogen (−H/−CH2OH) methylol/methylene-(−CH2OH/−CH2−)
Figure 8. Possible molecular structures existing in boron compound-modified PF resins.
Figure 9. Thermal stability of different postcured resins.
and boron compound-modified PF resins, quantitative 13C NMR analysis indicated that the resin structure was indeed changed by the addition of different boron compounds in the resin synthesis. Table 3 shows the variation of the functional groups on the phenolic rings for the PF and boron compound-modified PF resins studied. Generally, the para position of the phenolic ring has a slightly higher relative reactivity toward formaldehyde than the ortho position. Compared with the laboratory-made control PF resin, boron compound-modified PF resins had a lower ratio of para/ortho methylol groups, suggesting that the addition of boron compounds in PF resin synthesis promoted the methylolation between formaldehyde and the ortho position of the phenolic ring. The decrease in the para/ortho ratio of methylol groups was caused by the decrease in the free ortho positions and increased reaction with para positions. Meanwhile, a higher ratio of the para−para/ortho−para
methylene linkage was observed for boron compound-modified PF resins than for the laboratory-made control PF resin. It indicated that boron compounds promote condensation between para methylol groups. The unsubstituted/substituted hydrogen (−H/−CH2OH) ratio of the laboratory-made control PF resin was lower than that of boron compound-modified PF resins. The possible reason could be that the addition of boron compounds to the PF resin synthesis reduced the substituted hydrogen and retarded the reaction between phenol and formaldehyde. Meanwhile, the laboratory-made control PF resin demonstrated a lower methylol/methylene (−CH2OH/−CH2−) ratio, which was used as an indicator of the degree of polymerization. The boron compound-modified PF resins had a lower degree of polymerization than the laboratory-made control PF resin. These differences possibly explained the results of DSC in which although boric acid and borax reduced the curing 9846
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research Table 4. Weight Losses of Different Resins in the Thermal Degradation Process weight loss in a different temperature range (%)
total weight loss (%)
resin type
RT−200 °C
200−400 °C
400−600 °C
600−800 °C
RT−800 °C
curing reaction enthalpy (J/g)
control PF boric acid-modified PF borax-modified PF zinc borate-modified PF
11.2 9.8 11.2 11.9
9.8 10.8 10.4 10.2
14.3 18.1 18.2 13.1
12.9 13.5 16.4 16.5
48.2 52.2 56.2 51.7
300.8 260.3 249.9 289.2
Figure 10. Shear strength of lap shear samples bonded with different resins. (Columns sharing the same letter are statistically insignificantly different on the basis of ANOVA and Tukey’s test (α = 0.05.)
than for the laboratory-made control PF resin and zinc boratemodified PF resin. The weight loss of borax-modified PF resin and zinc borate-modified PF resin from 600 to 800 °C was higher than that of boric acid-modified PF resin and laboratorymade control PF resins. The addition of boron compounds to the PF resin in situ synthesis did not significantly change the thermal degradation of cured PF resins at a temperature lower than 400 °C. However, different types of boron compounds had different effects on the resulting cured PF resins’ thermal stability when the degradation temperature was higher than 400 °C. From the previous DSC analysis, the curing reaction enthalpy of the boron compound-modified PF resins was lower than that of the laboratory-made control PF resin, suggesting a lower curing degree for the boron compound-modified PF resins. Meanwhile, the cross-linked structure of cured boron compound-modified PF resins differed from that of the laboratory-made control resin. A higher curing reaction enthalpy indicated a more stable cured resin structure, which explained the lower total weight loss for the laboratory-made control PF resin from thermal degradation compared to that of boron compound-modified PF resins. Although boron compounds were widely used as fire retardants to postblend with wood adhesives, the final cross-linked structure of cured resins will dominate their thermal degradation behavior and degradation kinetics. Meanwhile, the fire-retardant effects of boron compounds under a nitrogen atmosphere could be different from those in an air environment. The detailed crosslinked structures, thermal degradation kinetics of the cured PF resins, the resin’s structural changes during thermal degradation, and the effects of boron compounds on the cured PF resins’ thermal stability under both nitrogen and air atmospheres are currently under investigation. 3.5. Bonding Strength. The bonding strength of different resins is shown in Figure 10. The boron-modified PF resins had a dry bonding strength comparable to that of the laboratorymade control PF resin. There was no significant difference in
activation energy of the resulting PF resins, they still retarded the resin curing process according to a lower pre-exponential factor. The possible molecular structure existing in boron compound-modified PF resins is shown in Figure 8. However, given the complexity of boron-containing PF, more detailed structural analysis in the future is needed to be certain of the structural diagram of the resins. In future studies, it would be interesting to conduct 11B NMR analysis to better understand the different boron compound-modified PF resins for woodadhesive applications. 3.4. Thermal Degradation of Cured PF Resins. The structure and cross-linking density of cured PF resins determined their thermal stability.41−43 Thermal stability is an important property closely related to a resin’s durability. The thermal degradation curves of different PF resins are shown in Figure 9. In general, the weight loss of the cured PF resins from room temperature to 200 °C was mainly due to moisture evaporation, dehydration, and the evaporation of smallmolecular-weight compounds. The cured resin underwent further degradation from 200 to 400 °C. The major polymer decomposition took place when the temperature increased from 400 to 600 °C. When the temperature further increases from 600 to 800 °C, collapse and carbonization will occur in the cured resin network along with the graphitization process.41−43 In general, boron compound-modified PF resins had a slightly higher total weight loss than the laboratory-made control PF resin up to 800 °C. Borax-modified PF resin had the highest total weight loss, followed by boric acid-modified PF resin and zinc borate-modified PF resin (Table 4). Boric acidmodified PF resin had a slightly lower weight loss than the borax-modified PF, zinc borate-modified PF, and laboratorymade control PF resins from room temperature to 200 °C. All of the resins had similar weight losses from 200 to 400 °C. Boric acid-modified PF resin and borax-modified PF resin had similar weight losses from 400 to 600 °C, which were higher 9847
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research Table 5. Physical and Mechanical Properties of Untreated and Treated Wooda WPG (%) untreated wood treated wood by laboratory-made control PF treated wood by PF-BA treated wood by PF-BX treated wood by PF-ZB a
0 47.6 49.3 44.2 45.7
(3.1) (2.5) (3.7) (2.9)
density (g/cm3) 0.38 0.54 0.57 0.51 0.52
(0.02) (0.03) (0.03) (0.04) (0.03)
MA (%) 9.32 5.83 5.62 6.47 6.15
ASE (%)
(0.55) (0.48) (0.52) (0.61) (0.50)
NA 50.72 53.51 47.64 48.65
(4.11) (3.79) (3.95) (4.01)
MOE (GPa)
MOR (MPa)
8.57 (0.60) 12.92 (0.81) 13.64 (0.78) 11.75 (0.85) 12.26 (0.69)
56.84 92.36 96.58 89.45 91.22
(5.35) (6.77) (6.96) (5.88) (6.01)
oxygen index (%) 24.5 37.8 47.9 48.6 50.2
(4.7) (3.9) (3.2) (3.5) (4.0)
Values in parentheses are standard deviations.
modifiers for low-density and low-quality wood impregnation are systematically investigated. Results showed that the modified PF resins had a slightly lower average molecular weight and polydispersity index and a longer gel time than the laboratory-made control PF resin without boron compounds. The addition of boron compounds affected the curing kinetics of the resulting PF resins. Boron compound-modified PF resins started to cure at a lower temperature than the laboratory-made control PF resin. Different boron compounds had different effects on the resin curing kinetics. Boric acid-modified PF resin and boraxmodified PF resin retarded the resin curing process, and zinc borated modified PF resin slightly accelerated the resin curing process compared to the laboratory-made control PF resin. The liquid-state 13C NMR showed that the addition of boron compounds to PF resin synthesis promoted the methylolation between formaldehyde and the ortho position of the phenolic ring and condensation between para methylol groups. Boron compound-modified PF resins demonstrated a lower degree of polymerization than the laboratory-made control PF resin. The resins’ molecular structure, functional groups, and cross-linked structure after curing varied depending on the type of boron compounds. The post-curing thermal stability of boronmodified PF resins was slightly lower than that of the laboratory-made control PF resin. Lap shear tests showed that boron compound-modified PF resins had dry and wet bonding strengths comparable to those of the laboratory-made control PF resin. Wood impregnated using boron compounds modified PF resins (PF-BA, PF-BX and PF-ZB) had comparable dimensional stability, mechanical properties and improved fire resistance than the wood impregnated using the lab made control PF resin regardless the type of the boron compounds used. The boron compound-modified PF resins could be used as an effective and multifunctional modifier for making value-added wood products with higher performance and sustainability. In future studies, it would be interesting to conduct 11B NMR analysis to better understand the different boron compound-modified PF resins for wood-adhesive applications. Meanwhile, the detailed cross-linked structures, thermal degradation kinetics of the cured boron compound-modified PF resins, resin’s structure changes during thermal degradation, and effects of boron compounds on the cured resins’ thermal stability under both nitrogen and air atmospheres are also worthy of further investigations.
the dry bonding strength among the boric acid-modified PF, borax-modified PF, and zinc borate-modified PF resins. After the water-soaking-and-drying (WSAD) treatment and the boiling water treatment (BWT), no sample delamination was found. The wet bonding strength of all of the resins was lower than the drying bonding strength. The laboratory-made control PF resin had a wet bonding strength similar to that of the zinc borate-modified PF resin and a slightly higher wet bonding strength than the boric acid-modified PF resin and boraxmodified PF resin. However, the ANOVA analysis showed that there was no statistically significant difference in the wet bonding strength of all of the resins. 3.6. Impregnated Wood and Its Properties. The properties of the treated and untreated wood are shown in Table 5. After PF resin impregnation, the treated wood gained about 45−50 percent of its weight on average (WPG%). The density of the treated wood also increased. The PF resin-treated wood had a lower moisture absorption (MA) compared to that of the untreated wood. The antiswelling efficiency (ASE) of the treated wood ranged from 47 to 53%. It indicated that the PF resin-impregnated wood had a higher dimensional stability than the untreated wood. Meanwhile, the MOE and MOR of PF resin-impregnated wood increased significantly compared to those of the untreated wood. The wood impregnated using boron compound-modified PF resins (PF-BA, PF-BX, and PFZB) had comparable dimensional stability (MA and ASE) and mechanical properties (MOE and MOR) to the woodimpregnated using the laboratory-made control PF resin without boron compounds. In addition, laboratory-made control PF resin-impregnated wood had a higher oxygen index than did the untreated wood, suggesting the improvement in fire resistance after impregnation. Wood samples treated with boron compound-modified PF resins had a higher oxygen index than those treated with the laboratory-made control PF resin, indicating that the boron compound-modified PF resins had a higher fire resistance than the laboratory-made control PF resin without boron compounds. There were no significant differences in the dimensional stability, mechanical properties, and fire resistance of wood impregnated using different boron compound-modified PF resins.
4. CONCLUSIONS In this study, different boron compounds were used together with an alkaline catalyst to synthesize phenol formaldehyde (PF) resole resins in situ for wood-adhesive and woodimpregnation applications. Compared to conventional postblending PF resole resins with boron compounds, the in situ synthesis of PF resin with boron compounds overcame the issues related to low solubility and poorer mechanical properties. The effects of boron compounds on the novel PF resins’ structure and performance as well as their application as
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.iecr.6b02156. 9848
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research
■
Assignments of chemical shifts for the liquid-state 13C NMR spectra of different boron compound-modified PF resins and laboratory-made control PF resins (PDF)
(16) Gao, W.; Cao, J.; Li, J. Effect of ammonium pentaborate on the cure kinetics of aqueous phenol formaldehyde resin in the presence of wood. Iran. Polym. J. 2010, 19, 959. (17) Gao, W.; Cao, J.; Li, J. Effect of ammonium pentaborate on curing of aqueous phenol formaldehyde resin. Iran. Polym. J. 2010, 19, 255. (18) Gao, W. 13C CP/MAS NMR analysis of cure characteristics of phenol formaldehyde resin in the presence of wood composite preservatives and wood: effect of ammonium pentaborate and copper compounds. Iran. Polym. J. 2012, 21, 283. (19) Gao, W.; Cao, J.; Li, J. Some physical, mechanical properties and termite resistance of ammonium pentaborate-treated strand board. Wood Res. 2010, 55, 61. (20) Ayrilmis, N. Effect of fire retardants on internal bond strength and bond durability of structural fiberboard. Build Sci. 2007, 42, 1200. (21) Winandy, J. E.; Wang, Q. W.; White, R. H. Fire-retardanttreated strandbord: properties and fire performance. Wood Fiber. Sci. 2008, 40, 66. (22) Ayrilmis, N.; Kartal, S. N.; Lanfenberg, T. Physical and mechanical properties and fire, decay, and termite resistance of treated oriented strandboard. For. Prod. J. 2005, 55, 74. (23) Lee, S.; Wu, Q.; Smith, W. R. Formosan subterranean termite resistance of borate-modified strandboard manufactured from southern wood species: a laboratory trial. Wood Fiber. Sci. 2004, 36, 107. (24) Lee, S.; Wu, Q.; Strickland, B. Influence of flake chemical properties and zinc borate on gel time of phenolic resin for OSB manufacturing. Wood Fiber. Sci. 2001, 33, 425. (25) Abdalla, M. O.; Ludwick, A.; Mitchell, T. Boron modified phenolic resins for high performance applications. Polymer 2003, 44, 7353. (26) Gao, J.; Liu, Y.; Yang, L. Thermal stability of boron-containing phenol formaldehyde resin. Polym. Degrad. Stab. 1999, 63, 19. (27) Liu, Y.; Gao, J.; Zhang, R. Thermal properties and stability of boron containing phenol-formaldehyde resin formed from paraformaldehyde. Polym. Degrad. Stab. 2002, 77, 495. (28) Kobera, L.; Czernek, J.; Strečková, M.; Urbanova, M.; Abbrent, S.; Brus, J. Structure and distribution of cross-links in boron modified phenol formaldehyde resins designed for soft magnetic composites: A multiple-quantum 11B-11B MAS NMR correlation spectroscopy study. Macromolecules 2015, 48, 4874. (29) Mané, C.; Sommerer, N.; Yalcin, T.; Cheynier, V.; Cole, R. B.; Fulcrand, H. Assessment of the molecular weight distribution of tannin fractions through MALDI-TOF MS analysis of protein-tannin complexes. Anal. Chem. 2007, 79, 2239. (30) Kissenger, H. E. Reaction kinetics in differential thermal analysis. Anal. Chem. 1957, 29, 1702. (31) He, G. B.; Yan, N. 13C NMR study on structure, composition and curing behavior of phenol−urea−formaldehyde resole resins. Polymer 2004, 45, 6813. (32) He, G. B.; Riedl, B.; Aït-Kadi, A. Model-free kinetics: Curing behavior of phenol formaldehyde resins by differential scanning calorimetry. J. Appl. Polym. Sci. 2003, 87, 433. (33) Lei, Y.; W, Q.; Lian, K. Cure kinetics of aqueous phenolformaldehyde resins used for oriented strandboard manufacturing: Analytical technique. J. Appl. Polym. Sci. 2006, 100, 1642. (34) Lei, Y.; Wu, Q. Cure kinetics of aqueous phenol-formaldehyde resins used for oriented strandboard manufacturing: Effects of wood flour. J. Appl. Polym. Sci. 2006, 102, 3774. (35) Zhao, Y.; Yan, N.; Feng, M. Characterization of phenolformaldehyde resins derived from liquefied lodgepole pine barks. Int. J. Adhes. Adhes. 2010, 30, 689. (36) Zhao, Y.; Yan, N.; Feng, M. Bark extractives-based phenolformaldehyde resins from beetle-infested lodgepole pine. J. Adhes. Sci. Technol. 2013, 27, 2112. (37) Kim, M. G.; Amos, L. W.; Barnes, E. Study of the reaction rates and structures of a phenol-formaldehyde resol resin by carbon-13 NMR and gel permeation chromatography. Ind. Eng. Chem. Res. 1990, 29, 2032.
AUTHOR INFORMATION
Corresponding Author
*Tel:+1-416-946-8070, +1-416-978-3834. E-mail: ning.yan@ utoronto.ca. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge financial support from the Special Fund for Forest Scientific Research in Public Welfare (201404501) provided by the State Forestry Administration of the People’s Republic of China and NSERC-discovery, the Natural Sciences and Engineering Research Council of Canada.
■
REFERENCES
(1) Pizzi, A. Wood Adhesives Chemistry and Technology;. Marcel Dekker: New York, 1993; Vol.1. (2) Pilato, L. Phenolic Resins: A Century of Progress; Springer: New York, 2010. (3) Feng, M.; He, G.; Anderson, A. W. Effects of esters and resorcinol on the phenolic resins as adhesives in medium density fiberboard manufacturing. Wood. Fiber. Sci. 2010, 42, 192. (4) Fan, D. B.; Chang, J. M.; Li, J. Z. On the structure and cure acceleration of phenol-urea-formaldehyde resins with different catalysts. Eur. Polym. J. 2009, 45, 2849. (5) Zhang, Y.; Zhang, Y.; He, L.; Zhou, Z. Cure rate of phenol formaldehyde resol resins catalyzed with MgO. J. Adhes. Sci. Technol. 2007, 21, 833. (6) Kamo, N.; Okamura, H.; Higuchi, M.; Morita, M. Condensation reactions of phenolic resins V: Cure-acceleration effects of propylene carbonate. J. Wood. Sci. 2004, 50, 236. (7) Park, B. D.; Riedl, B.; Hsu, E.; Shields, J. Differential scanning calorimetry of phenol−formaldehyde resins cure-accelerated by carbonates. Polymer 1999, 40, 1689. (8) Park, B. D.; Riedl, B. 13C NMR study on cure-accelerated phenol−formaldehyde resins with carbonates. J. Appl. Polym. Sci. 2000, 77, 841. (9) Tohmura, S.; Higuchi, M. Acceleration of the cure of phenolic resin adhesives. VI. Cure acceleration action of propylene carbonate. Mokuzai Gakkaishi. 1995, 41, 1109. (10) Zhao, C.; Pizzi, A.; Garnier, S. Fast advancement and hardening acceleration of low-condensation alkaline PF resins by ester and copolymerized urea. J. Appl. Polym. Sci. 1999, 74, 359. (11) Zhao, C.; Pizzi, A.; Kuhn, A.; Garnier, S. Fast advancement and hardening acceleration of low condensation alkaline phenol−formaldehyde resins by esters and copolymerized urea. II. Esters during resin reaction and effect of guanidine salts. J. Appl. Polym. Sci. 2000, 77, 249. (12) Lei, H.; Pizzi, A.; Despres, A.; Pasch, H.; Du, G. Ester acceleration mechanism in phenol formaldehyde resin adhesives. J. Appl. Polym. Sci. 2006, 100, 3075. (13) Pizzi, A.; Stephanou, A. On the chemistry, behavior and cure acceleration of phenol formaldehyde resins under very alkaline conditions. J. Appl. Polym. Sci. 1993, 49, 2157. (14) Mirski, R.; Dziurka, D.; Łecka, J. Properties of phenol formaldehyde resin modified with organic acid esters. J. Appl. Polym. Sci. 2008, 107, 3358. (15) Lei, Y.; Wu, Q. Cure kinetics of aqueous phenol formaldehyde resins used for oriented strandboard manufacturing effect of zinc borate. J. Appl. Polym. Sci. 2006, 101, 3886. 9849
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850
Article
Industrial & Engineering Chemistry Research (38) Vázquez, G.; López-Suevos, F.; Villar-Garea, A.; GonzálezAlvarez, J.; Antorrena, G. 13C NMR analysis of phenol−urea− formaldehyde prepolymers and phenol−urea−formaldehyde−tannin adhesives. J. Adhes. Sci. Technol. 2004, 18, 1529. (39) Holopainen, T.; Alvila, L.; Rainio, J.; Pakkanen, T. T. Phenol− formaldehyde resol resins studied by 13C NMR spectroscopy, gel permeation chromatography, and differential scanning calorimetry. J. Appl. Polym. Sci. 1997, 66, 1183. (40) Luukko, P.; Alvila, L.; Holopainen, T.; Rainio, J.; Pakkanen, T. T. Effect of alkalinity on the structure of phenol-formaldehyde resol resins. J. Appl. Polym. Sci. 2001, 82, 258. (41) Chen, Y. F.; Chen, Z. B.; Xiao, S. Y.; Liu, H. B. A novel thermal degradation mechanism of phenol-formaldehyde type resins. Thermochim. Acta 2008, 476, 39. (42) Chetan, M. S.; Ghadage, R. S.; Rajan, C. R.; Gunjikar, V. G.; Ponrathnam, S. Thermolysis of orthonovolacs. Part 1. Phenolformaldehyde and m-cresol formaldehyde resins. Thermochim. Acta 1993, 228, 261. (43) Chetan, M. S.; Ghadage, R. S.; Rajan, C. R.; Gunjikar, V. G.; Ponrathnam, S. Thermolysis of orthonovolacs. Part 2. Phenolformaldehyde and α-naphthol-formaldehyde resins. J. Appl. Polym. Sci. 1993, 50, 685.
9850
DOI: 10.1021/acs.iecr.6b02156 Ind. Eng. Chem. Res. 2016, 55, 9840−9850