«CURE KINETICS OF WOOD PHENOL-FORMALDEHYDE SYSTEMS By JINWU WANG A dissertation submitted in partial fulfillment of the requirements for the degree of ...»
ABSTRACT The cure kinetics of pure phenol formaldehyde (PF) resins have been investigated extensively. However, it is unclear whether the obtained kinetic parameters can be used to describe and predict cure development in the presence of wood, since wood is known to affect the cure kinetics of PF resins. In this research, differential scanning calorimetry (DSC) was used to investigate the influence of wood and wood constituents on the cure development and kinetics of PF resins. Mixtures of PF resin with southern yellow pine, extracted southern yellow pine, southern yellow pine extractives, cellulose, lignin, and hemicelluloses were evaluated in terms of heat of reaction, onset and end cure temperatures and activation energies using the Borchardt-Daniels nth order (nth-BD) model, Kissinger equation, and Vayazovkin model-free kinetics.
DSC analysis showed that the curing behavior of the PF resin did not change significantly when the wood content was below 20%. When the wood content was above 35%, the thermogram shapes and derived kinetic parameters changed significantly. One new exotherm appeared at low temperatures (ca. 50°), while the main reaction exotherms were not affected. The time-to-completion for wood/PF mixture was also not affected by the presence of wood. When conversion above 70% is of interest in practical application, the models working for pure resin can also be appropriate as a predictor for PF/wood mixture. The mixture of southern yellow pine extractives at 35 % content level released similar heat in reactions with pure PF resin, while all other substrates reduced the reaction heat significantly, suggesting that the resin did not reach the same cure extent as PF alone. Cellulose and xylan did not change the cure kinetics for PF curing. The catalytic effect of the presence of wood may come from interactions between PF and lignin.
Key words: Differential scanning calorimetry (DSC); kinetic models; southern yellow pine; wood constituents; phenol formaldehyde resin.
INTRODUCTIONThe cure behavior and kinetics of neat phenol-formaldehyde (PF) resins have been investigated for decades (Prime 1998). Wang et al. (2005, 2006) demonstrated that both model-fitting kinetics and model-free kinetics (MFK) are practical models for describing and predicting the cure of neat PF resins. At the same time, wood has long been known to influence the cure mechanisms and kinetics of PF resins (Chow 1969; Pizzi et al. 1994; He and Riedl 2004; He and Yan 2005).
In the presence of wood, the activation energy of the curing process can be considerably altered (Chow 1969; Pizzi et al. 1994; He and Yan 2005). Some wood species have no impact on the PF cure, while others increase or decrease the activation energy, accelerating or retarding the cure (Mizumachi and Morita 1975). He and Riedl (2004) reported that wood accelerated the substitution reactions (hydroxymethylation) and retarded the condensation reactions in the curing processes, further separating the two reactions into two exotherms on a DSC thermogram. Most often, a decrease in activation energy is observed in the presence of wood (Mizumachi and Morita 1975; Pizzi et al. 1994). The decrease of activation energy has been ascribed to catalytic activation induced by secondary interactions, such as dipolar forces or hydrogen bonds, between the lignocellulosic substrate and the PF resin (Pizzi et al.1994; He and Riedl 2004). Others have postulated that covalent bonds between wood carbohydrates and the resin could form, thus depressing the cure activation energy (Chow 1969; Kottes-Andrews et al. 1986). Furthermore, He and Riedl (2004) have shown that wood reduces the total heat of reaction, resulting in lower degrees of cure in PF resins cured in the presence of wood. The wood particles were proposed to physically separate the resin species, hence decreasing their mobility and probability to connect to each other. In line with these results, Laborie and Frazier (2006) noted that PF carbons exhibited lower cross polarization rates in the presence of wood than in the neat state, supporting the thesis of a lower crosslink density and degree of cure for PF resins cured in the presence of wood.
As a result, researchers have evaluated the impact of specific wood constituents on PF cure. Barry et al. (1993) noted a correlation between the heat of reaction of PF cure and the lignin content, with higher lignin content corresponding to lower heat of reaction. Chow (1969) reported that the PF chemical conversion decreased with increasing hydroxyl content in the mixture, either due to higher cellulosic content or higher hydroxyl groups per pyranose unit in the cellulosic compounds. Chow (1969) postulated that the activation energy for forming a resin-carbohydrate bond was less than that for a resin-resin bond, suggesting preferential reaction of phenolics with carbohydrates over self-condensation. Recent research also suggests that the low pH value of wood and extractives is the main cause for changes in activation energy (He and Riedl 2004, He and Yan 2005).
In most of these studies, activation energies were simply computed from the Kissinger equation, which is generally valid for simple reactions where iso-conversional cure peaks are obtained (Kissinger 1957). For complex, multiple reactions as in the case of PF resins (Detlefsen 2002), the Kissinger equation and the nth-BD method are inadequate to disclose mechanisms and may be inaccurate (Wang et al. 2005, Vyazovkin & Sbirrazzuoli 2006).
On the other hand, MFK is well-suited to portray the kinetics of complex reactions such as the cure of PF resins (Wang et al. 2005). MFK does not assume any definite form of the reaction and allows for variations in activation energy throughout the reaction process. As a result, the cure of PF resins in the presence of wood has been recently characterized with MFK using the Kissinger-Akahira-Sunose (KAS) algorithm (He & Riedl 2004). The study demonstrated that the activation energyconversion curve for 50% PF/spruce mixtures had a similar pattern to that of pure PF resin. In the early stage of curing, an increase in Eα with conversion was ascribed to consecutive and competitive reactions, after which a decrease in Eα was ascribed to a diffusion-controlled regime. In contrast to the cure thermogram of the neat PF resin, He and Riedl (2004) and He and Yan (2005) reported that a small peak appeared in the lower temperature range between 60 and 100°C in the DSC curves of PF resin/wood mixtures. This small exotherm was ascribed to hydroxymethylation reactions accelerated by the wood, while the main condensation reactions were retarded. The activation energy-changing pattern obtained by KAS was sensitive to this small exotherm, and a slightly slower increase rate in activation energy below 10% degree of cure was noted in comparison to neat PF (He & Riedl 2004).
The objective of this research was twofold: 1) to evaluate and compare the effectiveness of the Kissinger equation, nth BD, and model-free Vyazovkin method for characterizing the cure kinetics of PF resins in the presence of wood and wood constituents, and 2) to utilize the best-suited kinetic model to determine the influence of wood and wood constituents on the cure of PF resins.
EXPERIMENTALMaterials A commercial PF resole was obtained from Georgia-Pacific Company and stored at -20oC until use. The resin had a weight-average molecular weight (Mw) of 6576 g/mol, polydispersity (Mw/Mn) of 1.72. Gel permeation chromatography analysis showed that the resin has three fractions: a main fraction of high molecular weight species and two small fractions of low molecular weight material, monomers and dimers. The resin solid content is 45.0% and pH in the 11.0-11.5 range. In addition, elemental analysis showed the presence of 3.7 wt % nitrogen for the resin, indicating that urea was present (Wang et al. 2005).
Southern yellow pine (SYP) strands with 5% moisture content were obtained from the Huber Wood Engineering. The wood strands were ground into wood flour, passing through 60 mesh screen which corresponded to particles up to 250 µms. In order to evaluate the effect of extracted wood and extractives on the resin cure, part of the SYP was soxhlet extracted with acetone and then water for 72 hours each.
Measured extractive content was 3.9% for SYP. The flours of SYP and extracted SYP as well as SYP extractives were then dried under a vacuum at 60 ˚C until constant weight and stored in a desiccator with drierite until use. Cellulose powder was made by grinding filter paper (Whatman) into flour passing through 60 mesh with a Wiley mill machine. Unsulfonated kraft lignin (Indulin® AT) and birchwood xylan powder were obtained from Westvaco and Sigma Aldrich respectively, and used as received.
Mixtures of wood/ PF and wood constituents/ PF were prepared as follows:
2.0 g of liquid resin was manually mixed in a vial with 0.50, 1.1, 2.0, 4.7 g wood flour to yield mixtures having PF: wood weight % ratios of 80/20, 65/35, 50/50 and 30/70% wood content based on total weights (i.e. liquid resin +wood flour). Similarly, mixtures were prepared with cellulose, lignin, xylan, SYP extractives, and extracted SYP in a 65/35 PF: substrate % weight ratio. The time between the preparation of the mixtures and DSC analysis of all replicates did not exceed 12 hours and did not affect the DSC thermograms.
In order to investigate the effects of wood particle size on the effects of curing, PF resin was sprayed on both sides of a 1mm-thick oven-dried wood strip with an air brush, resulting in a 70% wood content level. A small disk was trimmed from the resinated wood strip for DSC scanning.
Differential scanning calorimetry A Mettler-Toledo DSC 822e was used to perform dynamic and isothermal cure experiments. For dynamic tests, a baseline for moist wood was first obtained by placing 10 mg SYP flour with a moisture content of 25% and 125% into high pressure crucibles and performing a heating scan at 10 °C/min in the DSC. Then wood/PF mixtures were evaluated. Approximately 13.5mg of PF/wood mixtures or neat PF resin were placed in a 30µl high pressure gold-plated crucible. Dynamic temperature scans were conducted at 4 heating rates 2, 5, 10, and 20 °C/min from 25°C to 250°C.
For each heating rate, three replicate measurements were performed. The resinated disk was scanned at one heating rate of 5 °C/min from 25°C to 250°C. DSC thermograms were then processed with the Mettler-Toledo STARe V7.2 software to extract the degree of cure, α, reaction rate, dα/dt, and corresponding temperature, Tα, in the 0≤ α ≤0.99 range. Both α and dα/dt were determined at a specific cure time, t, by normalizing the partial heat of reaction, ∆H(t), and heat flow, dH/dt, respectively by the total heat of reaction ∆H.
The neat resin was also characterized under isothermal cure conditions. An isothermal cure was performed for different time periods in the DSC cell that had been preheated at 120°C. Following the precure of the neat resin sample in the DSC cell at 120°C, the sample crucible was quickly removed from the DSC, quenched in liquid nitrogen, and rescanned in the DSC at 10 °C/min to determine residual cure.
Kinetic models An Arrhenius equation can be used to describe the cure reaction rate of
In Eq. (36), α is the conversion and a function of time (t), f(α) the reaction model, T the temperature, A the pre-exponential factor, E the activation energy and R the gas
constant. Using the nth order Eq. (36) can be rearranged into Eq. (37):
Kinetic parameters, A, E and n can be extracted by the Borchardt-Daniels procedure with the values of α and dα/dt and corresponding temperature from one DSC dynamic scan (ASTM E 2041, nth-BD).
The kinetic parameters can also be determined from multiple heating rate scans. The peak temperature (Tipeak) dependency on heating rate (βi) can thus be used to calculate the activation energy. Assuming an iso-fractional peak temperature, a linear regression of ln (βi/T2ipeak) against 1/Tipeak across several heating rates yields the activation energy with Kissinger Eq. (38) (Kissinger 1957).
With this method, when the peak temperatures are closer to each other across a series of heating rates, the calculated activation energies will increase in value. The nth-BD method gives cure kinetic parameters for the overall cure process at each heating rate, while the Kissinger equation estimates activation energy for specific peaks (Wang et al. 2006).
The patterns of activation energy against the degree of cure of the curing process can be evaluated by model-free kinetics of Vyazovkin method. Wang et al.
(2005) reported this method as the most appropriate for gaining insight on the cure mechanisms of commercial PF resoles. The method has been described previously (Vyazovkin 2001; Wang et al. 2005) and summarized as follows. In the Vyazovkin method, n scans are performed at different heating programs, Ti(t). The activation
energy at a specific degree of cure is obtained by minimizing the function ϕ(Eα):
Eq. (40) can be solved numerically by integrating the experimental data within small time intervals ∆α. The ‘I’ values are then substituted into ϕ(Eα), and this function is minimized by Brent’s method (Brent 1973) leading to Eα. The procedure is repeated for distinct values of α. A new parameter Cv(α) can be created that complements Eα in
fully describing the cure kinetics:
Model-free kinetics does not assume any definite form of the reaction model and allows for variations in activation energy as the reaction progresses.