cellulose nanocomposite

Thermochimica Acta 540 (2012) 107–115

Contents lists available at SciVerse ScienceDirect

Thermochimica Acta journal homepage: www.elsevier.com/locate/tca

Thermal degradation kinetics of resole phenol-formaldehyde resin/multi-walled carbon nanotube/cellulose nanocomposite Byung-Dae Park a,∗ , John F. Kadla b a b

Department of Wood Science and Technology, Kyungpook National University, Daegu, 702-201, Republic of Korea Department of Wood Science, University of British Columbia, Vancouver, B.C., V6T 1Z4, Canada

a r t i c l e

i n f o

Article history: Received 7 January 2012 Received in revised form 16 April 2012 Accepted 18 April 2012 Available online 25 April 2012 Keywords: Thermal degradation kinetics Thermogravimetry Multi-walled carbon nanotube Resole phenolic resin Cellulose Nanocomposite Silanization

a b s t r a c t This study investigated thermal degradation kinetics of multi-walled carbon nanotubes (MWCNTs) reinforced resole phenol-formaldehyde (PF) resin/cellulose nanocomposite, using thermogravimetry (TG) analysis as a function of the content or surface treatment of MWCNTs with or without a surfactant. FT-IR spectroscopy showed that the oxidation provided hydroxyl or carboxyl groups with MWCNT’s surface while the silanization resulted in the silane attachment to MWCNT’s surface. Conventional TG (CTG) thermograms demonstrated six thermal degradation steps, corresponding to various components of the nanocomposite, and also showed that the use of surfactant hastened thermal decomposition of the nanocomposite. The activation energy (E) obtained by the Kissinger method slightly increased as the MWCNT content increased while that of the cellulose degradation was independent on the MWCNT content. Both the oxidation and silanization treatments of MWCNTs’ surface resulted in an increase of the E values compared to that of the control sample. The activation energy (E˛ ) based on the isoconversional method increased up to ˛ = 0.5, and then was rapidly elevated to fluctuations. The activation energy (Em ) based on the temperature modulated TG (MTG) analysis was within the range of the E value calculated by the Kissinger method for the thermal degradation of cellulose, a main component of the nanocomposite. These results show that MTG method provides similar activation energy to that of CTG method for thermal degradation of the nanocomposite, and indicate that MTG method be efficiently used to obtain activation energy without many scans from multiple heating rates. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Two types of carbon nanotubes (CNTs), i.e., single-walled CNTs (SWCNTs) and multi-walled CNTs (MWCNTs) are available in these days since the discovery of CNTs in 1991 [1]. SWNTs consist of a single sheet of graphene rolled seamlessly to form a cylinder with diameter of order of 1 nm and length of up to centimeters [2]. By contrast, MWCNTs consist of an array of such cylinders formed concentrically and separated by 0.35 nm, similar to the basal plane separation in graphite [1]. MWCNTs can have diameters from 2 to 100 nm and lengths of tens of microns. Owing to unique properties such as excellent strength, modulus, electrical and thermal conductivities along with a low density, CNTs resulted in a tremendous attraction for research on polymer-based nanocomposite [3,4]. For example, tensile properties of SWCNTs were measured inside an electron microscopy, and reported tensile moduli between 0.32 and 1.47 TPa and strengths between 10 and 52 GPa [5]. Tensile modulus and strength using bundles of MWCNTs prepared by chemical vapor

∗ Corresponding author. Tel.: +82 53 950 5797; fax: +82 53 950 6751. E-mail address: [email protected] (B.-D. Park). 0040-6031/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tca.2012.04.021

deposition (CVD) was reported as 0.45 TPa and ∼4 GPa [6]. Yu et al. [5] also employed the same method, and obtained modulus values of 0.27–0.95 TPa and tensile strengths in the range 11–63 GPa. However, there is a variation in these properties of CNTs upon fabrication methods such as laser ablation [7] and CVD [8,9], or, decomposition of CO [10]. Polymer-based nanocomposite is commonly defined as the combination of a polymer matrix and additives that have at least one dimension in the nanometer range. In order to exploit excellent mechanical properties of CNTs, varieties of polymer matrices were used to provide with reinforcements as reported in a review [11]. A quite number of works have been done on CNTs reinforced nanocomposite based on thermoset polymers even though it was much less than those on thermoplastic polymers [11]. Since the first report [12], epoxy resins have been widely studied as a potential matrix for CNTs-based nanocomposite due to their wide range of industrial uses [13–16]. Among thermosetting polymers, phenol-formaldehyde (PF) resin is one of the most common formaldehyde-based resins which provide outstanding performance in adhesion, high temperature resistance, flame resistance and electric insulation. Thus, PF resins have been used to improve mechanical properties, moisture

108

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

resistance, or fire resistance in many applications. For example, PF resin is impregnated into cellulose paper to fabricate decorative or protective laminates on the surface of wood-based composite products. These laminate papers are hot-pressed onto the surface of composite wood products to provide a hard surface finish. Excellent mechanical properties of MWCNTs are expected to provide a very effective reinforcement with PF resin that is impregnated into cellulose paper. A very limited work has been done on PF resin as a matrix for CNTs reinforcement because CNTs are extremely difficult to disperse in a polymer matrix due to its agglomeration caused by Van der Waals force [17]. For example, Yin et al. [18] used CNTs to reinforce novolac PF resin in powder form for the preparation of PF resin/graphite composites, and found that the surface modification of CNTs by introducing hydroxyl groups improved bending strength and conductivity of the composites. Liu and Ye [19] also modified MWCNTs by introduction of carboxyl group, benzene ring, or boric acid, and found an enhanced thermal stability of MWCNTs/PF resin nanocomposite. The improved thermal stability was ascribed to a better interfacial interaction between MWCNTs and PF resin matrix. In this study, we attempted to fabricate MWCNTs reinforced resole PF resin/cellulose nanocomposite as a potential candidate for surface laminates on wood-based composite panels. Therefore, this study focused on thermal stability of the nanocomposite in order to assess their feasibility as surface laminates for wood-based composite products. A brief summary of relevant thermal decomposition kinetics follows.

Assuming ˇ = constant, and rearranging in terms of two variables gives: d˛ A = exp f (˛) ˇ

 −E  RT



˛

g(˛) = 0

d˛ A = f (˛) ˇ

d˛ = k(T )f (˛) dt

(1)

where ˛ represents the extent of conversion (˛ = 0–1), t is time, k(T) is the rate constant, and f(˛) is the reaction model, which describes the dependence of the reaction rate on the extent of reaction. The extent of conversion, ˛, is experimentally determined by the below equation using TGA data. m0 − m(T ) m0 − m∞

(2)

Where m0 is the initial mass, m∞ is final mass, and m(T) is a mass at a temperature. In most cases the temperature dependence of k(T) can be satisfactorily described by the Arrhenius equation, whose substitution into Eq. (3) yields:



 f (˛)

(3)

where E is the activation energy and A is the pre-exponential factor. Various methods can be applied to find the kinetic parameters, particularly the E when kinetics follows a single equation expressed as the Eq. (1). For non-isothermal conditions, the Eq. (3) is converted into the Eq. (4) when the temperature is raised at a constant heating rate, ˇ:



ˇ

T˛

exp

 −E 

T0

RT

dT =

AE p(x) ˇR

(6)

p(x) = 0.0048e−1.0516x

(7)

From Eqs. (6) and (5), it follows ln(ˇ) = −5.331 − 1.0516

E  RT

+ ln

 AE  Rg(˛)

(8)

The E can be calculated from a plot of log ˇ versus 1/T. Since the multi-heating rate method uses the relationship between the peak temperature (Tp ) of the DTG peak and its corresponding heating rate. Another method by Kissinger [23] is based on the fact that at the Tp of a DTG peak, the maximum conversion rate is equal to zero:

dt

T =Tp

=0

(9)

According to Kissinger, appropriate approximations lead to the simple Eq. (10) derived from Eq. (9):

The single-step kinetic equation for polymer degradation is usually expressed as [20,21]:

d˛ E = A exp − RT dt



where x = E/RT. Since the p(x) function, the temperature integral, has no exact analytical solution, it can be solved using either approximations or numerical integration [21]. One of the simplest approximations by Doyle [22] is given as Eq. (7):

dt

2.1. CTG data

˛=

(5)

Integration of Eq. (5) involves solving the temperature integral in the following Eq. (6):

 d  d˛  2. Thermal degradation kinetic analysis

dT

d˛ E = A exp − RT dT

 f (˛)

(4)



ln

ˇ Tp2



=−

E + ln RTp

 AR  E

(10)

which is the equation of a straight line between ln(ˇ/Tp2 ) and −1/Tp . The E can be calculated from the slope and the pre-exponential factor from the intercept. Thus, the Kissinger method provides a single value of activation energy for the whole reaction process. The isoconversional kinetic analysis or so-called “model-free kinetics” method requires performance of a series of experiments at different temperature programs and yield the values of effective activation energy (E˛ ) as a function of conversion (˛). Coat and Redfern [24] suggested the following approximations for the p(x) function:



p(x) = x−2 e−x 1 −

2 x

 (11)

In general, the part of (1 − 2/x) of the Eq. (7) is neglected and an oversimplified approximation leads to Eq. (12), which is used [25]. p(x) = x−2 e−x

(12)

Submitting Eq. (12) into Eq. (6) and taking logarithm results in an isoconversional equation:



ln

ˇi 2 T˛,i



= − ln

E˛ A˛ R − RT˛,i E˛ g(˛)

(13)

The subscript i denotes different heating rates. The Eq. (13) provides 2 ) and 1/T . The E can be calculated a straight line between ln(ˇ/T˛,i ˛ ˛,i from the slope and the pre-exponential factor from the intercept. The above isoconversional method was proposed by Vyazovkin and Sbirrazzuoli [26], which was used in this study.

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

109

2.2. MTG data

3.2. Methods

By contrast, in early seventies, kinetics of MTG analysis was proposed by Flynn and Wall [27,28]. In MTG, a periodic temperature modulation is overlaid on the conventional linear heating rate experiments. A general temperature program of MTG is characterized by a sinusoidal modulation with amplitude (AT ) and period (P):

3.2.1. Surface modification of MWCNTs and FT-IR spectroscopy Prior to oxidation, the MWCNTs were purified by ultrasonication in toluene for 1 h, washing with acetone and distilled water, and drying in vacuum dryer for 24 h at 100 ◦ C. The MWCNT oxidation was done by dispersing 3 g of MWCNTs in 300 mL of H2 SO4 /HNO3 (3:1, v/v) solution, and by stirring for 3 h at 60 ◦ C. The solution was filtered, and then washed with water and acetone. The resulting oxidized MWCNTs (OX-MWCNTs) were then dried under vacuum at 80 ◦ C for 24 h. The APTMS was used for the surface modification of MWCNTs by a similar procedure to the published method [30]. 25 mg of OXMWCNTs was dispersed in 250 mL of ethanol via ultrasonication for 30 min. 25 mg of APTMS was slowly added to the OX-MWCNTs in ethanol, and then the reaction was conducted while stirring at 70 ◦ C for 4 h. The reaction product was then filtered, washed with water followed by acetone. The resulting silanized carbon nanotubes (SiMWCNTs) were dried under vacuum at 80 ◦ C for 20 h, and then prepared in a powdered form. FT-IR spectra were recorded on an attenuated total reflectance Fourier transform infrared (ATR-FTIR) instrument (Alpha-p, Bruker, Germany) with DTGS detector in the range of 400–4000 cm−1 . Spectral outputs were recorded in the transmittance mode as a function of wave-number and 100 scans with resolution of 4 cm−1 at room temperature.

T = T0 + ˇt + AT sin(ωt)

(14)

where T0 is the initial temperature, ˇ the average or underlying heating rate and ω is the period divided by 2␲. For an experiment of MTG by a sine wave, the Eq. (3) may be evaluated as the ratio for adjacent peaks, and valleys, of the periodic rate of reaction [29]. Therefore, the ratio of reaction rates, between adjacent peaks (p) and valleys (v) can be written as follows: (d˛p /dT ) Af˛p exp(−Em /Tp ) = )f (˛) (d˛v /dT ) Af˛v exp(−Em /Tv )

(15)

If the reacted fraction does not change significantly between adjacent half cycles, f(ap) ≈ f(av), and their ratio approaches unity, Eq. (15) can be simplified and solved for Em to yield Eq. (16). Em =

RTp Tv ln(d˛p /d˛p ) Tp − Tv

(16)

In MTG, the oscillatory temperature forcing function is defined by an average temperature T, the temperature amplitude A and its period or frequency. Because Tp = T + A and Tv = T − A, Eq. (16) may be rewritten as follows [19]: Em =

RL(T 2 − A2 ) 2A

(17)

where L = ln(d˛p /d˛v ), the ratio of the maximum and the minimum conversion rate during a cycle. Thus, the Eq. (17) allows a continuous calculation of Em . 3. Experimental 3.1. Materials MWCNTs synthesized by catalytic chemical vapor deposition (CVD) process was purchased from Carbon Nano-materials Technology Co., Ltd., Pohang, South Korea, and had a diameter range of 5–20 nm, length of ∼10 ␮m, aspect ratio of 500, purity of above 90 wt%, specific surface area (BET) of 100–700 m2 /g, and bulk density of 0.08–0.12 g/cm3 . 3-Aminopropyltrimethoxysilane (APTMS, H2 N(CH2 )3 Si(OCH3 )3 ) was used for the surface modification of MWCNTs. The following reagents were used without further purification: sulfuric acid (95%, Duksan Reagents Co, Korea), nitric acid (70%, Duksan Reagents Co, Korea), acetone (99.5%, Duksan Reagents Co, Korea), and ethanol (99.5%, Sigma–Aldrich). Sodium dodecyl sulfate (SDS: CH3 (CH2 )11 OSO3 Na, ≥99.0%, Sigma–Aldrich, Korea) was used as a surfactant. A commercial resole PF resin tailored for the impregnation of paper laminates was donated by a local resin company (Tae Yang Chemical Ltd. Co., Incheon, Korea). The resole PF resin had a formaldehyde/phenol mole ratio of 2.2 and non-volatile solids content of 48.2%, and was mixed with the MWCNTs. The cellulose paper used for the impregnation of MWCNTs reinforced PF resins was donated by a local low pressure laminate (LPL) manufacturer (Samsun Gratech Inc., Incheon, Korea). It was a grade of LPL with a weight of 80 g/m2 .

3.2.2. Preparation of resole PF resin/MWCNT/cellulose nanocomposite Either purified or treated MWCNTs particles were dispersed in distilled water with and without SDS (0.5 wt% in water), sonicated for 3 h, and then added into aqueous PF resins to have different levels of MWCNTs such as 0, 0.1, 0.5 and 1 wt% based on the nonvolatile solids content of PF resins. Mixtures of MWCNTs and PF resins were sonicated for 2 h before their impregnation into the cellulose paper. These mixtures were impregnated into the cellulose paper to obtain a target resin loading of 60%, which gave a paper weight of about 200 g/m2 . Excess resin was removed by squeezing with a roller. The resin impregnated papers were dried at room temperature for 3 h, and then cured at 80 ◦ C in a drying oven overnight. These nanocomposite samples were stored in a dessicator for future use.

3.2.3. TG analysis of resole PF resin/MWCNT/cellulose nanocomposite For CTG measurements, four different sample weights such as 20, 15, 10, and 5 mg were used for four heating rates of 2.5, 5, 10, 15 and 20 ◦ C/min to obtain thermal degradation kinetics for the nanocomposite. The samples were run from 30 ◦ C to 600 ◦ C under a nitrogen flow rate of 60 mL/min using a TG (Q500, TA Instruments, New Castle, DE, USA). MTG measurements were made with the same CTG with modulation option. The samples were also run from 30 ◦ C to 600 ◦ C using at a heating rate of 2 ◦ C/min, amplitude of ±0.5 ◦ C, and a period of 200 s. Since the sample weight influences thermograms of MTGA [31], different sample weights such as 5, 10, 15, and 20 mg were evaluated for resole PF resin/cellulose composites without MWCNTs. A mass of approximately 20 mg in an aluminum pan was selected for all MTGA experiments based on the fluctuations of continuous Em values measured.

110

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

also react with hydrolyzed silane, leading to the attachment of silane to the MWCNT surface [35].

4.2. Thermal kinetic analysis of CTG data

Fig. 1. FT-IR spectra of pristine and surface modified MWCNTs.

4. Results and discussion 4.1. FT-IR analysis FT-IR spectra of pristine and surface-modified MWCNTs were obtained to understand chemical changes on the surface of MWCNTs after either oxidation or silanization after the oxidation treatments. As shown in Fig. 1, two bands appeared at 3740 and 1630 cm−1 corresponding to free OH group due to atmospheric moisture present on pristine MWCNTs [30,32]. The band appeared at 2000–1900 cm−1 is due to various cumulated C C bonds [33]. When the MWCNTs were oxidized, two new bands appeared at 1789 and 1325 cm−1 are corresponding to C O from COOH group, and C–O from COOH, respectively [30,34]. In addition to common bands of pristine and oxidized MWCNTs, a new band occurred at 986 cm−1 belongs to Si–O of Si–OH when MWCNTs were treated with APTMS after their oxidation [35]. Based on the FT-IR spectra, a schematic diagram of chemical reactions occurred for treated MWCNTs are presented in Fig. 2. When MWCNTs were oxidized, hydroxyl or carboxyl groups are formed on their surface. These functional groups are reacted with hydroxylmethyl group (CH2 OH) of resole PF resins, leading to the attachment of phenolic molecules to the MWCNT. When the oxidized MWCNTs were treated with APTMS, these functional groups

Fig. 3 shows a typical thermogram of cTG at 5 ◦ C/min for the nanocomposite with 0.5% MWCNTs. Five derivative TG (DTG) peaks corresponding to five mass loss steps were appeared during thermal degradation of the nanocomposite. The first DTG peak (Tp1 ) around 75.7 ◦ C could be due to either the evaporation of water in the sample or thermal softening, or gelation of PF resins. The second DTG peak (Tp2 ) around 145.4 ◦ C could be resulted from the evaporation of water produced by the condensation reaction, or curing of UF resins. And the Tp3 could be caused by either from the degradation of small terminal groups (i.e., hydroxymetyl groups) or melting of the surfactant in the sample. In fact, the melting temperature of the surfactant (i.e., SDS) is 204–207 ◦ C. The most dominant DTG peak, Tp4 , is attributed to the decomposition of cellulose that is a main component of the nanocomposite. After decomposition of cellulose, the Tp5 could be due to the degradation of methylene linkages in cured PF resins, while the Tp6 could be from the decomposition of aromatic phenol rings of PF resins. However, the presence of a DTG peak around 200 ◦ C due to the surfactant, SDS, was inconsistent which was possibly due to a relatively small amount of the SDS compared to PF resins and MWCNTs. CTG thermograms of the nanocomposite as a function of the MWCNT content are shown in Fig. 4. As expected, the major decomposition peak occurred by cellulose at around 342 ◦ C. When MWCNTs were added into the nanocomposite, the mass loss (%) around 200 ◦ C was greater than that of the control sample. This became greater after the decomposition of cellulose, indicating a lower thermal stability for the nanocomposite with 0.1% and 0.5% MWCNTs, which could be due to the decomposition of the surfactant. However, the addition of 1% MWCNTs much improved thermal stability of the nanocomposite. Fig. 5 shows CTG thermograms of the nanocomposite with different surface treatments of the MWCNT (0.1 wt%). Thermal stability of the nanocomposite without the surfactant was much improved above 328 ◦ C which was the main decomposition temperature of cellulose. When the surfactant was used to disperse MWCNTs, the nanocomposite’s thermal stability significantly decreased above 328 ◦ C. This could be ascribed to high alkalinity of the surfactant (pH 7.2) and PF resins (pH 10.1). This high alkalinity facilitates to decompose cellulose component in the nanocomposite at high temperature, which subsequently reduce resistance to

Fig. 2. Schematic diagram of chemical reactions for surface modified MWCNTs.

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

111

Fig. 3. A typical CTG thermogram of the nanocomposite with 0.5% MWCNTs from 5 ◦ C/min scan.

Fig. 4. CTG results of the nanocomposite as a function of MWCNTs level at a heating rate of 5 ◦ C/min.

thermal degradation. When the OX-MWCNTs were dispersed in resole PF resin, thermal resistance of the resultant nanocomposite was slightly weakened below 300 ◦ C and above 400 ◦ C. It is not clear whether an interaction between the oxidized MWCNTs and surfactant accelerates thermal degradation of the nanocomposite, which requires further research work. However, thermal resistance of the nanocomposite was improved when the MWCNTS were treated with APTMS. This could be due to a chemical reaction between MWCNTs and APTMS as presented in Fig. 2. In order to evaluate thermal degradation kinetics of each component of the nanocomposite, four heating rates were employed, and a typical effect of heating rate is shown in Fig. 6. As expected, the DTG peaks moved to higher temperature as the heating rate increased from 2.5 to 20 ◦ C/min. So, these DTG peak temperatures were used to obtain the activation energy (E) values for thermal degradation of each component using the Eq. (10) of the Kissinger method. The calculated E values for pure cellulose paper and three DTG peaks and using the Eq. (10) were shown in Fig. 7. The E value of the thermal decomposition of pure cellulose paper was calculated

Fig. 5. CTG results of the nanocomposite with 0.1% MWCNTs, depending on various surface treatments (heating rate: 5 ◦ C/min).

Fig. 6. Typical CTG thermograms of the nanocomposite with 0.1% MWCNTs as a function of heating rates.

112

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

Fig. 7. Change of the E values of three DTG peaks (Tp3 , Tp4 , and Tp5 ) as a function of MWCNT content as well as cellulose paper.

Fig. 8. Comparison of the E values of three DTG peaks (Tp3 , Tp4 , and Tp5 ), depending on the MWCNT’s surface treatments for the nanocomposite (0.1% MWCNTs).

4.3. Isoconversional analysis of CTG data as 207.9 kJ/mol in the present work, and close to the reported value (145–200 kJ/mol) by Capart et al. [36] although the E values of cellulose thermal decomposition are in the range of 200–270 kJ/mol [36–39]. However, the E values of the nanocomposite were lower than that of the pure cellulose paper, which could be due to the present of PF resins. And the addition of MWCNTs was not significantly affected to the E values of cellulose decomposition in the nanocomposite. And, the addition of SDS also did not much affect the E of cellulose decomposition of the nanocomposite with 0.1% MWCNTs. These results indicate that thermal decomposition of cellulose in the nanocomposite was not much affected by the presence of MWCNTs and SDS. However, these two components did affect to the E values of Tp3 and Tp5 . As shown in Fig. 3, overall E values of the three DTG peaks are in the increasing order of Tp3 , Tp4 , and Tp4 , which indicates different amounts of thermal energies for the decomposition of each event. Fig. 7 also shows a significant influence of the surfactant (i.e., SDS) to the thermal degradation of the nanocomposite. In fact, when no SDS was used, the E values of both Tp3 and Tp5 for the nanocomposite were much greater than those of the counterparts at 0.1% MWCNTs content. This result suggests that the dispersion (or agglomeration) of MWCNTs in PF resin change thermal degradation behavior of the nanocomposite; the addition of SDS made MWCNTs dispersed well in PF resin, which makes it easy thermal degradation of the composites, but the agglomeration of MWCNTs without SDS requires much more energy for thermal degradation. As the MWCNTs content increased, the E value of the Tp3 decreased up to 0.5% MWCNTs followed by a slight increase while that of the Tp4 increased above 0.5% MWCNTs. We believe that these results are also related to the state of MWCNT’s dispersion in PF resin. A greater amount of MWCNTs (i.e., 1%) causes their agglomerations, which requires more thermal energy to be decomposed. Fig. 8 also shows E values of the nanocomposite, depending on the surface treatments of MWCNTs. The E values of the nanocomposite increased as the degradation temperature rose, regardless of surface treatments. The E values also increased when MWCNTs’ surface was modified by either oxidation or silanization. This could be due to a chemical reaction between MWCNTs and PF resins, or APTMS, which improved interfacial adhesion. However, the E value of the nanocomposite prepared with silane-treated MWCNTs slightly decreased when compared to those of the oxidized MWCNTs. This could be ascribed thermal degradation of APTMS at lower temperature.

An isoconversional method provides the effective activation energy (E˛ ) as a function of conversion (˛), which could lead to understand thermal degradation kinetics of the nanocomposite. Fig. 9 shows changes of E˛ values of the nanocomposite at different levels of MWCNTs content. When 0.1% MWCNTs was added, the E˛ increased for all degrees of conversions from the one of the control sample (0% MWCNTs). However, further addition of MWCNTs diminished E˛ at the degree of conversion up to 0.65. For all MWCNTs levels, E˛ values rapidly increased above the ˛ = 0.65. A large deviation of E˛ value versus ˛ suggests that multiple thermal degradation steps is dominant above the ˛ = 0.65. This is quite consistent with multiple degradation steps as shown in Fig. 3. And this result could also be due to the fact that much more energy was required to thermally degrade methylene linkages and aromatic rings of PF resins, leading to greater E˛ values. The isonconversional method was also employed to understand thermal degradation kinetics of the nanocomposite, depending on type of the surface modification of MWCNTs, and the result was presented in Fig. 10. When the surfactant (SDS) was not added, the E˛ between 0.1 and 0.5 degree of conversion of the nanocomposite (MWCNTs + No SDS) was dropped from those of the control

Fig. 9. Changes of E˛ values of resole PF resin/MWCNT/cellulose nanocomposite as the degree of conversion.

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

Fig. 10. Changes of Ea values of resole PF resin/MWCNT/cellulose nanocomposite, depending on surface modifications at 0.1% MWCNT.

ones (MWCNTs + SDS). However, the nanocomposite prepared by the silane-treated MWCNTs showed moderate increase in the E˛ value between 0.1 and 0.5 degree of conversion. The E˛ values were not obtained above 0.5 degree of conversion for the nanocomposite prepared with the silane-treated MWCNTs. However, the E˛ value rapidly rises above the ˛ = 0.65 for all samples, which is compatible with those of the nanocomposites with different MWCNT contents. These results also indicate that multiple thermal degradation steps occur above the ˛ = 0.65. As observed for the addition levels of MWCNTs, this result also suggests that more energy is required to thermally degrade methylene linkages, or benzene rings of PF resins. 4.4. Thermal kinetic analysis of MTG data In order to compare thermal degradation kinetics of the nanocomposite using CTG technique, MTG was employed to obtain a continuous measurement of activation energy (Em ). Fig. 11 displays a typical MTG thermogram of the nanocomposite with 0.1% MWCNTs and SDS. As shown Fig. 3, the weight loss curve of the nanocomposite was quite similar to the counterpart of CTG. Fig. 11

113

Fig. 12. Changes of Em values of resole PF resin/MWCNT/cellulose nanocomposite, depending on the content of MWCNTs.

also shows a continuous change of the Em value as thermal degradation temperature rose. Thermal degradation of the nanocomposite includes multi-step reactions corresponding to DTG peaks such as Tp3 , Tp4 , or Tp5 . As expected, the Em values are approximately corresponding to each reaction steps even though there is a temperature difference between the DTG peaks and Em peaks. For example, the DTG peak occurs at 305.2 ◦ C while the Em peak corresponding to this temperature appears at 319.5 ◦ C. Fig. 12 illustrates changes of the Em value of the nanocomposite at different levels of the MWCNTs added. When 0.1% and 0.5% MWCNTs were added, the Em at low temperature (i.e., around 100 ◦ C) was much greater than those of 1% MWCNTs were added. However, the Em value of Tp4 (i.e., cellulose degradation step) slightly increased as the content of MWCNTs rose. This result is quite consistent with the E value obtained by Kissinger method as shown in Fig. 5. The Em value obtained by MTG is in the range of the E value calculated by Kissinger method for the thermal degradation of cellulose that is a main component of PF resin/MWCNT/cellulose nanocomposite. These results indicate that both methods deliver similar activation energy for cellulose thermal decomposition of the nanocomposite, and are compatible with the result of polytetrafluorethylene [40].

Fig. 11. A typical MTG thermogram of the nanocomposite at 0.1% MWCNT content.

114

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115

Acknowledgment This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government MEST, Basic Research Promotion Fund (NRF-2010-013-F00005).

References

Fig. 13. Changes of Em values of resole PF resin/MWCNT/cellulose nanocomposite, depending on the surface modifications of MWCNTs (0.1%).

Fig. 13 displays changes of the Em value of the nanocomposite at different types of the surface modifications for MWCNTs. In most of cases, the Em values of the nanocomposite prepared by OX-MWCNTs without the surfactant were lower than those of the nanocomposite prepared with OX-MWCNTs and surfactant as well as silane-treated MWCNTs and surfactant. However, the Em values of cellulose pyrolysis were greater than those of the nanocomposite prepared by OX-MWCNTs without the surfactant. These results could be ascribed to chemical reactions between MWCNTs and PF resins or APTMS in the nanocomposite as presented in Fig. 8.

5. Conclusions This study investigated thermal degradation kinetics of MWCNTs reinforced resole PF resin/cellulose nanocomposite, using TG analysis as a function of the content or surface treatment of MWCNTs with or without a surfactant. The following conclusions were obtained:

1. FT-IR spectroscopy showed that the oxidation provided hydroxyl or carboxyl groups with MWCNT’s surface while the silanization resulted in the silane attachment to MWCNT’s surface. 2. The CTG analysis demonstrated six thermal degradation steps, corresponding to various components of the nanocomposite, and also showed that the use of surfactant quickened the thermal decomposition of the nanocomposite. The E values of the nanocomposite slightly increased as the MWCNT content increased while the one of cellulose degradation was independent of the MWCNT content. Both the oxidation and silanization of MWCNTs’ surface resulted in an increase of the E values compared to that of the control sample. 3. The E˛ values gradually increased up to ˛ = 0.5, and then were rapidly elevated to fluctuations at ˛ > 0.5. The Em value obtained by the MTG was within the range of the E value calculated by the Kissinger method for cellulose thermal degradation that is a main component of the nanocomposite. 4. These results show that MTG method provides similar activation energy to that of CTG method for thermal degradation of the nanocomposite, and indicate that MTG method could be efficiently used to obtain activation energy without many scans from multiple heating rates.

[1] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58. [2] S. Iijima, T. Ichihashi, Single-shell carbon nanotubes of 1-nm diameter, Nature 363 (1993) 603–605. [3] J.N. Coleman, U. Khan, W.J. Blau, Y.K. Gun’ko, Small but strong: a review of the mechanical properties of carbon nanotube–polymer composites, Carbon 44 (2006) 1624–1652. [4] T.-W. Chou, L. Gao, E.T. Thostenson, Z. Zhang, J.-H. Byun, An assessment of the science and technology of carbon nanotube-based fibers and composites, Compos. Sci. Technol. 70 (2010) 1–19. [5] M. Yu, O. Lourie, M.J. Dyer, T.F. Kelly, R.S. Ruoff, Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load, Science 287 (2000) 637–640. [6] S. Xie, W. Li, Z. Pan, B. Chang, L. Sun, Mechanical and physical properties on carbon nanotube, J. Phys. Chem. Solids 61 (2000) 1153–1158. [7] A.A. Puretzky, D.B. Geohegan, X. Fan, S.J. Pennycook, In situ imaging and spectroscopy of single-wall carbon nanotube synthesis by laser vaporization, Appl. Phys. Lett. 76 (2000) 182–184. [8] K. Hernadi, A. Fonseca, J.B. Nagy, D. Bernaerts, A. Fudala, A.A. Lucas, Catalytic synthesis of carbon nanotubes using zeolite support, Zeolites 17 (1996) 416–423. [9] G. Che, B.B. Lakshmi, C.R. Martin, E.R. Fisher, R.S. Ruoff, Chemical vapor deposition based synthesis of carbon nanotubes and nanofibers using a template method, Chem. Mater. 10 (1998) 260–267. [10] W. Zhou, Y.H. Ooi, R. Russo, P. Papanek, D.E. Luzzi, J.E. Fischer, Structural characterization and diameter-dependent oxidative stability of single wall carbon nanotubes synthesized by the catalytic decomposition of CO, Chem. Phys. Lett. 350 (2001) 6–14. [11] Z. Spitalsky, D. Tasis, K. Papagelis, C. Galiotis, Carbon nanotube–polymer composites: chemistry, processing, mechanical and electrical properties, Prog. Polym. Sci. 35 (2010) 357–401. [12] P.M. Ajayan, O. Stephan, C. Colliex, D. Trauth, Aligned carbon nanotube arrays formed by cutting a polymer resin–nanotube composite, Science 265 (1994) 1212–1214. [13] A. Allaoui, S. Bai, H.M. Cheng, J.B. Bai, Mechanical and electrical properties of a MWNT/epoxy composite, Compos. Sci. Technol. 62 (2002) 1993–1998. [14] J. Bai, Evidence of the reinforcement role of chemical vapour deposition multiwalled carbon nanotubes in a polymer matrix, Carbon 41 (2003) 1325–1328. [15] Y. Breton, G. Desarmot, J.P. Salvetat, S. Delpeux, C. Sinturel, F. Beguin, Mechanical properties of multiwall carbon nanotubes/epoxy composites: influence of network morphology, Carbon 42 (2004) 1027–1030. [16] X.D. Li, H.S. Gao, W.A. Scrivens, D.L. Fei, X.Y. Xu, M.A. Sutton, Nanomechanical characterization of single-walled carbon nanotube reinforced epoxy composites, Nanotechnology 15 (2004) 1416–1423. [17] N.G. Sahoo, S. Rana, J.W. Cho, L. Li, S.H. Chan, Polymer nanocomposites based on functionalized carbon nanotubes, Prog. Polym. Sci. 35 (2010) 837–867. [18] Q.K. Yin, A. Sun, L. Li, S. Shao, S. Liu, C. Sun, Study on carbon nanotube reinforced phenol-formaldehyde resin/graphite composite for bipolar plate, Power Sources 175 (2008) 861–865. [19] L. Liu, Z. Ye, Effects of modified multi-walled carbon nanotubes on the curing behavior and thermal stability of boron phenolic resin, Polym. Degrad. Stab. 94 (2009) 1972–1978. [20] J.D. Peterson, S. Vyazovkin, C.A. Wight, Kinetics of the thermal and thermooxidative degradation of polystyrene, polyethylene and polypropylene, Macromol. Chem. Phys. 202 (2001) 775–784. [21] S. Vyazovkin, N. Sbirrazzuoli, Isoconversional kinetic analysis of thermally stimulated processes in polymers, Macromol. Rapid Commun. 27 (2006) 1515–1532. [22] C.D. Doyle, Estimating isothermal life from thermogravimetric, J. Appl. Polym. Sci. 6 (1962) 639–642. [23] H.E. Kissinger, Reaction kinetics in differential thermal analysis, Anal. Chem. 29 (1957) 1702–1706. [24] A.W. Coat, J.P. Redfern, Kinetic parameters from thermogravimetric data, Nature 201 (1964) 68–69. [25] S. Vyazovkin, D. Dollimore, Linear and nonlinear procedures in isoconversional computations of the activation energy of nonisothermal reactions in solids, J. Chem. Inf. Comput. Sci. 36 (1996) 42–45. [26] S. Vyazovkin, N. Sbirrazzuoli, Confidence intervals for the activation energy estimated by few experiments, Anal. Chim. Acta 355 (1997) 175–180. [27] J.H. Flynn, L.A. Wall, A general treatment of the thermogravimetry of polymers, J. Res. Nat. Bur. Stand. A 70 (1966) 487–523. [28] J.H. Flynn, L.A. Wall, A quick direct method for determination of activation energy from thermogravimetric data, J. Polym. Sci. Part C-Polym. Lett. 4 (1996) 323–328. [29] J.H. Flynn, B. Dickens, Steady-state parameter-jump methods and relaxation methods in thermogravimetry, Thermochim. Acta 15 (1976) 1–16.

B.-D. Park, J.F. Kadla / Thermochimica Acta 540 (2012) 107–115 [30] J. Kathi, K.-Y. Rhee, J.H. Lee, Effect of chemical functionalization of multi-walled carbon nanotubes with 3-aminopropyltriethoxysilane on mechanical and morphological properties of epoxy nanocomposites, Composites: Part A 40 (2009) 800–809. [31] S. Kubo, J.F. Kadla, Thermal decomposition study of isolated lignin using temperature modulated TGA, J. Wood Chem. Technol. 28 (2008) 106–121. [32] T. Ramanathan, F.T. Fisher, R.S. Ruoff, L.C. Brinson, Amino-functionalized carbon nanotubes for binding to polymers and biological systems, Chem. Mater. 17 (2005) 1290–1295. [33] R.M. Silverstein, F.X. Webster, Spectrometric Identification of Organic Compounds, 6th edition, John Wiley & Sons Inc, New York, 1998, pp. 71–143. [34] S.M. Yuen, C.C.M. Ma, C.L. Chiang, J.A. Chang, S.W. Huang, S.C. Chen, C.Y. Chuang, C.C. Yang, M.H. Wei, Silane-modified MWCNT/PMMA composites – preparation, electrical resistivity, thermal conductivity and thermal stability, Compos. Part A: Appl. Sci. Manuf. 38 (2007) 2527–2535.

115

[35] P.C. Ma, J.K. Kim, B.Z. Tang, Functionalization of carbon nanotubes using a silane coupling agent, Carbon 44 (2006) 3232–3238. [36] R. Capart, L. Khezami, A.K. Burnham, Assessment of various kinetic models for the pyrolysis of a microgranular cellulose, Thermochim. Acta 417 (2004) 79–89. [37] M.J. Antal, W.S.L. Mok, G. Varhegyi, T. Szekely, Review of methods for improving the yield of charcoal from biomass, Energy Fuels 4 (1990) 221–225. [38] J.A. Conesa, J.A. Caballero, A. Marcilla, R. Font, Analysis of different kinetic model, in the dynamic pyrolysis of cellulose, Thermochim. Acta 254 (1995) 175–192. [39] A.K. Burnham, R.L. Braun, T.T. Coburn, E.I. Sandvik, D.J. Curry, An appropriate kinetic model for well-preserved algal kerogens, Energy Fuels 10 (1996) 49–59. [40] M. Schubnell, Comparison of activation energies obtained from modulated and conventional non-modulated TG, J. Therm. Anal. Calorim. 61 (2000) 1005–1011.