Effect of physical adhesion on mechanical behaviour of bamboo fibre reinforced thermoplastic composites

Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

Effect of physical adhesion on mechanical behaviour of bamboo fibre reinforced thermoplastic composites C.A. Fuentes a,∗ , L.Q.N. Tran a , M. Van Hellemont b , V. Janssens a , C. Dupont-Gillain c , A.W. Van Vuure a , I. Verpoest a a

Department of Metallurgy and Materials Engineering (MTM), Katholieke Universiteit Leuven; Leuven, Belgium Unit Matter, GROUP T - Leuven Engineering College; Leuven, Belgium c Institute of Condensed Matter and Nanosciences, Université Catholique de Louvain; Louvain-la-Neuve, Belgium b

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

 The wetting of bamboo fibres conforms well to the molecular-kinetic theory (MKT).  Meaningful information on the interfacial interactions was obtained.  X-ray photoelectron spectroscopy analysis was consistent with wetting measurements.  Surface energy components of bamboo fibres and matrices were matched.  Physical adhesion was improved as revealed by pull-out and three point bending tests.

a r t i c l e

i n f o

Article history: Received 24 September 2012 Received in revised form 7 November 2012 Accepted 7 November 2012 Available online 16 November 2012 Keywords: Interface Wetting Fibre-matrix bond Photoelectron spectroscopy (XPS) Bamboo Natural fibre composites Molecular kinetic theory Composites

a b s t r a c t Systematic experimental results describing the dynamic wetting properties of bamboo fibres were analysed by applying the molecular-kinetic theory of wetting. Results suggest that the bamboo fibre surface represents a well-defined system for wetting analysis. The surface free energy components were calculated according to the acid–base theory. These values were then used to calculate the theoretical work of adhesion, spreading coefficient, wetting tension, and interfacial energy. The wetting behaviour of various thermoplastic matrices (polypropylene, maleic anhydride-grafted polypropylene, polyvinylidene-fluoride, and polyethylene-terephthalate) was characterized. Surface chemical components were identified using XPS. Additionally, transverse 3-point bending tests and single fibre pull-out tests were performed. This integrated physical–chemical–mechanical approach was used to study the effect of adhesion on the mechanical strength of thermoplastic composites reinforced with bamboo, showing that increase in physical adhesion can explain the improved interfacial and longitudinal strength in bamboo polyvinylidene-fluoride (PVDF) composites compared to the other thermoplastic matrices used in this study. Surface energy components of bamboo fibres and PVDF were matched, resulting in an improvement of the physical adhesion. © 2012 Elsevier B.V. All rights reserved.

1. Introduction

∗ Corresponding author. Tel. +32 16 321448; fax +32 16 321990 E-mail address: [email protected] (C.A. Fuentes). 0927-7757/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfa.2012.11.018

The interaction between the reinforcing fibre and the matrix has a significant effect on the properties of the composite since the stress transfer and the load distribution efficiency at the interface is determined by the degree of adhesion between the components.

8

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

Natural fibre composites with a thermoset matrix have already shown notable mechanical properties due to chemical bonding at the interface and low viscosity of the resins, allowing good impregnation of the fibres by the resin [1]. However, problems arise over their end-of-life environmental impact since they are neither biodegradable nor recyclable. The use of thermoplastics as matrices for natural fibre composites represents an approach with lower environmental impacts. However, the hydrophilic nature of most natural fibres reduces their potential as reinforcing agents due to low interfacial interactions with typically hydrophobic thermoplastic matrices. Other issues with thermoplastic matrices are the lack of occurrence of covalent bonding and the high viscosity of molten thermoplastics (lowering the speed of wetting), making maximization of the physical interactions indispensable to obtain better composites. When physical interactions are predominant over chemical interactions, then the interfacial strength of a composite can be related to the surface energies of a fibre and a matrix in a given system. On top of this, an effect of mechanical interlocking may be expected. The contact angle is a quantitative measure of solid–liquid molecular interactions and thus provides information on the surface energy of solids. Well described wetting parameters, and in particular advancing and receding contact angles, must be obtained in order to correctly evaluate surface energy components by means of surface energy component theories, which are based on the theoretical Young contact angle assuming that an equilibrium state can be reached [2,3]. Theoretical models describing the wetting kinetics of synthetic materials show a dependency of the contact angle on the wetting velocity. Therefore, a similar logical dependency of dynamic contact angles must be found as a necessary condition for the characterization of the wetting properties of natural fibres [4]. In a previous study [5] we showed that the high concentration of lignin on the surface of technical bamboo fibres is responsible for their wetting properties, and that the large fluctuations during wetting between various bamboo fibres are related to the surface topography irregularities of the fibres. If surface waviness, roughness and liquid penetration are minimized, then the wetting behaviour of bamboo fibres can be studied. In order to reduce these surface irregularities, a procedure based on an autoclave treatment was used, smoothening the lignin surface layer. In this study, the results of systematic experiments, describing the dynamic wetting properties of bamboo fibres with different liquids, were analysed by applying the molecular-kinetic theory of wetting. Compliance with this theory may ensure that the bamboo fibre surface represents a well defined system for wetting analysis. In this way, meaningful information on the interfacial interactions was obtained in order to conduct a study of physical adhesion, in terms of surface free energies, of bamboo fibres with different thermoplastics.

constant speed to drive the contact line at a given velocity (either in an advancing or in a receding direction), forming a stable dynamic contact angle  (v). If the driving force for wetting is taken to be the out-of-balance surface tension force (cos 0 −cos), then the relationship between v and q is given by:



v = 2K0 

sinh



2 (cos 0 − cos ) 2kT

(1)

where  is the average length of each molecular displacement, k is the Boltzmann constant, T the absolute temperature,  the surface tension of the liquid, and  0 the static contact angle at v = 0. The complete derivation of this model is presented by Blake [3]. 2.2. Surface energy components Van Oss, Chaudhury and Good [8] developed a model that considers the acid-base interaction between molecules. In this model, the total surface energy is divided into a Lifshitz–van der Waals ( LW ), an acid ( + ) and a base ( − ) component. The relation between the surface energy components of different liquids and a solid with respect to the contact angle  is given in Eq. (2) in its matrix notation [9]. Ax = b,

(2)

where

⎡



LW l,1



+ l,1

− l,1

⎢  ⎢ + ⎢  LW l,2 l,2 A=⎢ ⎢ ⎢ ... ... ⎣  sLW

− l,2

...



+ l,m

LW l,m

⎡



− l,m

⎤

⎡ l,1 (1 + cos 1 )/2 ⎥ ⎥ ⎢ ⎥ l,1 (1 + cos 2 )/2 ⎥,b = ⎢ ⎢ ⎥ ⎣ ⎥ ... ⎦

⎤ ⎥ ⎥ ⎥, ⎦

l,1 (1 + cos m )/2

⎤

⎢ ⎥ + ⎥  s ⎣ ⎦ 

x=⎢

s−

The results depend on the number and the choice of liquids since small variations in experimental measurements can induce very large differences in the numerical results. According to Della Volpe and Siboni [9], it is possible to obtain reliable results by selecting a group of liquids with the lowest possible value of the so-called “condition number”. For a determined case (when m = 3 in Eq. (2)), the condition number is defined as cond(A) = ||A||||A−1 ||

(3)

where A denotes the matrix associated to Eq. (2) and   designates the matrix norm ||A|| =

n

i,j=1

|Ai,j | . A low condition number value

2. Theoretical background

typically corresponds to a combination of a dispersive, an acidic and a basic liquid. The complete mathematical approach proposed by Della Volpe and Siboni is shown in [10,11].

2.1. Molecular-kinetic theory

2.3. Wetting parameters

According to the molecular-kinetic theory of wetting (MKT) [3,6], the displacement of the contact line at a given velocity v depends on the frequency of forward and backward molecular displacements within this three phase zone, K+ and K– , respectively. These displacements occur randomly but progressively in the direction of the moving contact line [7]. At equilibrium, v = 0, the net rate of displacement is zero, so that K+ = K– = K0 , where K0 is the equilibrium displacement frequency. In forced wetting, the substrate is put in contact with the liquid until a meniscus is formed and then the substrate is moved at

The thermodynamic work of adhesion, Wa , refers to the work required to disjoint a unit area of the solid-liquid interface [12]. The latter implicitly relates Wa with adhesive strength, thus several studies are reported in the literature showing the correlation of Wa with interfacial strength [13–15]. Wa is defined in terms of surface energies by the Dupré equation (Eq. (4)), which can also be expressed in terms of the contact angle by substituting Young’s equation in Dupré’s definition (second equality of Eq.(4)). Wa = s + l − sl = l (1 + cos )

(4)

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

The spreading coefficient, S, represents the ability of a liquid to spread over a surface. During the wetting process of a smooth surface, a droplet with a finite contact angle minimizes the free energy of the system (partial wetting) if s < sl + 1 . On the other hand, if s > sl + 1 , the droplet spreads out on a smooth surface to form a liquid layer (total wetting) [16]. The relation between S and the interfacial strength was proposed by Wu [17], who suggests that interfacial strength is maximum at maximum S. S = s − (l + sl )

(5)

The wetting tension, F, can be defined as the work done in eliminating a unit area of the solid-liquid interface while exposing a unit area of the solid–vacuum interface [12]. Mittal [18] and Gray [19] proposed that maximum adhesive strength would be obtained at maximum F. Eq. (6) defines F as the difference of the surface free energies at the solid–vapor and solid–liquid interfaces. When the Young’s equation is used, the wetting tension can be obtained in terms of the contact angle and the surface energy of the liquid (second equality of Eq. (6)). F = s − sl = l cos 

(6)

The interfacial energy,  sl , is defined as the reversible work necessary to increase the common boundary area of two contacting surfaces by unit area. Mittal [18] studied the relation of  sl with the strength of adhesive bonds and suggested that the lower  sl , the higher the joint strength. The interfacial energy can be expressed in terms of Lifshitz–van der Waals ( LW ) and acid–base ( AB ) components as sl = sLW + lLW + 2[(s+ s− )1/2 + (l+ l− )1/2 − (sLW lLW )1/2 − (s+ l− )1/2 − (s− l+ )1/2 ]

(7)

The four wetting parameters discussed above have been considered as criteria for optimum adhesion by several authors. Connor et al. [20] performed an algebraic analysis of the conditions to maximise Wa , S, and F, while minimizing  sl , in order to show which criterion is more representative of adhesion. By differentiating the various parameters, it was shown that F is maximum when Wa is highest within the region where spontaneous wetting occurs, i.e. s ≥ 0 .Then it was concluded that the latter situation corresponds to the optimum condition for an adhesive to adhere optimally to a given substrate. For the case of composites, the term “adhesive” corresponds to the polymer matrix and “substrate” to the reinforcing fibre.

Bamboo fibres were further put in an autoclave to smoothen the lignin at the fibre surface under 3 bars of pressure at 150 ◦ C for one hour (see [5] for more details on this procedure). The thermoplastic films were washed with a detergent (RBS-35 from Chemical Products) at a concentration of 4% v/v in water during one hour under magnetic stirring, and next rinsed in ultrapure water at 90 ◦ C for one hour. The cleaned films were then dried under vacuum at 90 ◦ C for two hours and then conserved under silicagel. 3.3. Contact angle measurements Advancing and receding contact angles were measured under controlled conditions, temperature of 20 ◦ C and humidity of 60%, with a Krüss K100 tensiometer using the Wilhelmy technique. This method is shown to be reliable for the study of wetting of natural fibres [5]. Dynamic contact angle measurements at a given speed were performed with a 0.05 mm data sampling step. Accordingly, the data for a single 10 mm length sample represent 200 values. The average and standard deviation values reported in this study were calculated from the data of all the samples measured at a given speed. 3.4. Determination of the fibre perimeter The method applied to determine the fibre perimeter is based on the same principle as the Wilhelmy technique. In this case, however, the perimeter is sought. When a liquid with 0◦ contact angle is used, the Wilhelmy equation, F = p l cos, becomes: F = pl

3.1. Materials Technical bamboo fibres of the species Guadua angustifolia were mechanically extracted from bamboo culms in the Department of Metallurgy and Materials Engineering at KULeuven. Polypropylene (PP), maleic anhydride grafted polypropylene (MAPP, Bynel 50E725), polyethylene terephthalate (PET), and polyvinylidene fluoride (PVDF, Solef 1008) polymer films were obtained from Propex, Dupont, Goodfellow, and Solvay respectively.

(8)

where F is the measured force, p is the fibre perimeter, and l the liquid surface tension. At relatively low speeds, very low surface tension hexane is assumed to have a contact angle of 0◦ with virtually all substrates [5]. 3.5. Fitting procedure MKT The parameters of the wetting kinetics are obtained by curve fitting from the correlation plot of experimental dynamic contact angle values and wetting velocity. The procedure followed by Vega et al. [2] was adapted in fitting the data using the molecular-kinetic theory. Accordingly, Eq. (1) can be simplified to

v=A 3. Materials and methods

9

sinh[B(C − cos )]

(9)

where A = 2K0  , B = 2 /2kT , and C = cos 0 are the independent parameters. Eq. (9) was used as the regression model to fit each set of experimental data. More information regarding this procedure can be found in our previous publication [5]. 3.6. Surface free energy calculation The calculation of the acid–base surface free energy components was performed by using SurfTen 4.3 software developed by Claudio Della Volpe and Stefano Siboni, at the University of Trento, Italy, to solve Eq. (2) and (3).

3.2. Materials preparation For the MKT fitting process, the technical bamboo fibres that were examined underwent the following preparation procedure: After being selected (by means of an optical microscope), the fibres were cleaned, first with warm water for one hour (90 ◦ C), then with ethanol before being dried in a vacuum oven at 80 ◦ C for one hour.

3.7. Surface characterization: X-ray photoelectron spectroscopy (XPS) XPS analyses were performed on a Kratos Axis Ultra spectrometer (Kratos Analytical – Manchester – UK) equipped with a monochromatized aluminium X-ray source (powered at 10 mA and

10

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

15 kV). More information regarding the XPS analyses procedure can be found in Ref. [5]. 3.8. Composite interfacial bond tests Unidirectional composites were prepared by compression moulding of stacks of prepregs consisting of technical bamboo fibres, sandwiched between thermoplastic films. The applied pressure was 20 bar and temperatures of 175 ◦ C and 200 ◦ C were used. Flexural three point bending tests in transverse direction were performed on a universal testing machine (Instron 4426) based on ASTM D790, to obtain an estimation of the bond strength in tensile mode. Single fibre pull-out specimens were prepared by using the compression moulding technique with a specially designed mould under 10 bar and 200 ◦ C for all samples. The embedded length of the fibres into the obtained block of polymer is controlled by drilling a hole through the fibre and matrix. Pull-out test were performed on a low-force universal testing machine (Instron 5943). 4. Results and discussion 4.1. Dynamic wetting The contact angle behaviour (advancing, receding and hysteresis) is determined by molecular interactions between the liquid and the solid at the three-phase contact line. The relation between contact angle and surface energies is determined by the Young’s equation and it only holds for smooth, homogeneous, flat, and undeformable surfaces. However, natural fibres’ surfaces are rough, heterogeneous, and susceptible to liquid sorption and diffusion; hence, contact angle measurements on natural fibres usually result in a range of angles, different from the advancing, receding or equilibrium [4,5]. Our purpose was to investigate whether it is possible to obtain reliable parameters of the wetting dynamics and quasi-equilibrium (advancing and receding static contact angle) in order to ensure a meaningful interpretation of the wetting data. Therefore, a correlation of experimental dynamic contact angle values and wetting velocity of autoclave-treated bamboo fibres in different test liquids was performed, using the Wilhelmy method and the MKT analysis. The set of liquids that were used corresponds to a broad range of surface tensions (polar and non-polar), with the intention of performing a proper surface energy components analysis and MKT

Table 1 Physical characteristics of test liquids: density , dynamic viscosity , surface tension , and molecular mass of the bulk liquid mL . Test liquid

 (g/cm3 )

 (mPa/s)

(mN/m)

mL (u)

Water Diiodomethane Ethylene glycol Benzyl alcohol

1.00 3.40 1.11 1.04

1.0 2.8 16.1 8.0

72.8 50.8 47.7 39.0

18.0153 267.8356 62.0681 108.1385

Table 2 MKT wetting parameters for water, diiodomethane, ethylene glycol and benzyl alcohol on bamboo technical fibre.

Water Diiodomethane Ethylene glycol Benzyl alcohol

K0 (×106 )

 (nm)

 0 (◦ )

R2

0.059 ± 0.006 0.069 ± 0.003 1.618 ± 0.205 0.118 ± 0.080

0.824 ± 0.012 2.422 ± 0.021 0.603 ± 0.004 2.158 ± 0.018

60.2 ± 2.2 47.9 ± 1.1 42.8 ± 0.8 28.0 ± 1.5

0.90 0.96 0.99 0.98

wetting parameters characterization (see Table 1). All fluids had relatively low viscosity to exclude further viscosity related kinetic effects. It was also necessary to minimize surface waviness and roughness of bamboo fibres in order to maintain the stability of the contact line. These irregularities are largely reduced by autoclave treatment and subsequent smoothening of the lignin surface layer, as confirmed by our previous study [5]. The results indicate that the advancing angles are speeddependent in all the test liquids conforming well to the prediction of the MKT over the entire experimental speed range (see Fig. 1). The quality of the fit to Eq. (1) reveals good agreement between the experimental and the calculated values of the dynamic contact angle , with coefficient of determination (R2 ) values above 0.90 (see Table 2). The MKT parameters are presented in Table 2. The advancing static contact angle with water is a bit high suggesting a largely non-polar surface. However, the obtained jump frequency K0 is low, suggesting a more polar surface, if compared with published values for water on Nylon [2] and PET [3,6] (three and one order of magnitude larger respectively). That is to say, the equilibrium frequency K0 of the random molecular displacements is lower possibly due to strong interactions if compared with Nylon and PET. Since the measured contact angles represent just the advancing front, which is possibly underestimating the polar content, the small K0 value would also suggest a small receding contact angle

Fig. 1. The dynamic contact angle as a function of wetting velocity for water (open squares), ethylene glycol (filled triangles), diiodomethane (open tilted squares), and benzyl alcohol (filled squares). The theoretical curves were obtained by nonlinear regression of experimental data using Eq. (1). Each point represents the average contact angle of 10 fibres. The angles were measured at different speeds ranging from 0.15 to 200 mm/min. The inset shows the low velocity data only.

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

for a bamboo–water system. This could not be measured due to the effect of sorption and diffusion during the receding process, resulting in a situation of irregular or zero receding contact angles [5]. As can be seen in Table 2, the value of the average length of molecular displacement, , is in the same order of magnitude when comparing between the four different liquids. Concerning the variation of K0 , which is more pronounced than for , Vega et al. [2] studied the wetting behaviour of Nylon with five different liquids and saw that K0 increased by a factor of 70 with no obvious trend. The origin of these variations remains to be investigated. There are no references of MKT parameters for liquids on bamboo fibres, and thus the obtained values cannot be compared. In order to discuss our obtained values, the link between the work of adhesion (Eq. (4)) and the so-called contact-line friction coefficient ( 0 ) [2,7] was evaluated when the system was near to equilibrium. That is to say, Eq. (1) reduces to a linear expression for small values of the function sinh when the contact line velocity is close to zero:

v=

l (cos 0 − cos ) 0

K0 =

T 2 Wa exp l

T



(11)

When 0 = T/K0 3 is substituted in Eq. (11), it can be seen that 0 increases with both liquid viscosity and Wa :



 2 Wa = s l exp

T 

Fig. 2. Plot of logarithm of the contact-line friction coefficient corrected for the liquid viscosity versus the work of adhesion. Straight line: y = 0.169x − 6.861, R2 = 0.97. The work of adhesion Wa was calculated by using Eq. (4) and the contact angles of Table 2.

(10)

According to Eq. (10) [7], 0 can be calculated by fitting dynamic data or by using the values of K0 and  which were obtained previously by fitting the experimental dynamic contact angle values and wetting velocity of autoclave-treated bamboo fibres (see Table 2). Blake and De Coninck [3,7] proposed a relation of the equilibrium frequency K0 with the thermodynamic work of adhesion, Wa (see Eq. (11)). Later studies found in the literature have shown experimental evidence of the proposed link between K0 and Wa [2,21].



11



0

(12)

where 1 is the volume of the “unit of flow”, which is usually equal to the molecular volume of the liquid (1 = mL /). According to Eq. (12), a plot of ln( 0 /) against Wa = l (1 + cos ) should give a straight line of positive slope, if  and the ratio l /3 stay relatively constant from one liquid to another (see Fig. 2) [7]. Although the rough assumption was made of a relatively constant  and ratio l /3 for all liquids used (particularly for benzyl alcohol), and the irregularities of bamboo fibre’s surface, the agreement appears to be acceptable. The linear regression corresponds well to the relation established in Eq. (11) with a coefficient of determination (R2 ) value of 0.967 and a positive slope of 0.169. The results support the correctness of MKT wetting parameters obtained for bamboo fibre, and may ensure a meaningful interpretation of wetting data. A proper characterization of the wetting behaviour of bamboo fibres, as discussed before, was possible due the reduction of the

effects of liquid absorption and perimeter variation, which are normally inherent to natural fibres. It was shown in our previous study [5] that liquid absorption and perimeter variation play a minimum role during the measurement of advancing contact angles by using the Wilhelmy method on autoclave treated bamboo fibres. 4.2. Surface energy components The main goal behind the wetting kinetic modelling was not to obtain the static advancing contact angle, but to show that the bamboo fibre’s surface represents a well defined system. Since the latter has been established, the static advancing contact angles that were obtained in the previous section (see Table 2), can now be used to determine the surface energy components of the bamboo fibres (although the lowest speed contact angle data are as reliable as the static angle obtained by using the complex MKT fitting). The selected liquid triplet, which contains a dispersive as well as an acidic and a basic liquid, presents a condition number of 5.4 according to Eq. (3). This value is low if compared with other triplets found in the literature review [9–11], corresponding to a well-balanced set of liquids. The higher the condition number, the less acceptable the triplet selected for the calculation of surface energy components [10]. Table 3 shows the calculated surface energy components values according to the acid–base model. Given the scale used (see Table 3), the Lewis basicity of the material is approximately as high as the one of water while its acidity is about 170 times smaller than that of water. Moreover, the Lifshitz-van der Waals component is higher than the one of water. Accordingly, the fibre’s surface is relatively hydrophobic and a Lewis base; this means that a strong acidic Lewis matrix should give the best adhesion performance. Wetting measurements relate the acidic or basic nature of the investigated surface to the magnitude of the work of adhesion, Wa . Accordingly, the calculated Wa (see Table 3) is in agreement with the obtained surface energy components for bamboo,

Table 3 Surface energy components of used liquids according to the scale proposed by Della Volpe and Siboni [11] and the obtained surface energy components of bamboo. The work of adhesion Wa (absolute value) is calculated by using Eq. (4) and bamboo as substrate.

Water Diiodomethane Ethylene glycol Benzyl alcohol Bamboo

 tot (mJ/m2 )

 LW (mJ/m2 )

 + (mJ/m2 )

 − (mJ/m2 )

Wa (mJ/m2 )

72.8 50.8 48.0 39.0 38.8

26.25 50.80 33.90 – 35.4

48.50 0.00 0.97 – 0.28

11.16 0.00 51.60 – 10.13

108.97 84.86 83.22 73.43 –

12

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

Table 4 Relative atomic percentages (excluding hydrogen), O/C ratio, and decomposition of C 1s peaks obtained by XPS on bamboo technical fibres (Each value represents the average of 8 samples). C (%)

N (%)

O (%)

74.29 ± 1.54

22.86 ± 0.25

Si (%)

1.75 ± 0.71

Binding energy (eV)

O/C

0.63 ± 0.43

0.31 ± 0.02

showing a large Wa with a relatively acidic liquid (water), and a smaller Wa with a relatively basic liquid (ethylene glycol). The wettability of bamboo by a nonpolar liquid (diiodomethane) arises solely from Lifshitz–van der Waals interactions. As shown in Table 3, the Wa of the diiodomethane–bamboo system is almost the same as in case of ethylene glycol, even though the latter possesses polar groups. This observation might confirm the smaller presence of acidic groups on the surface of bamboo if compared with the Wa obtained with water, since the basic surface of bamboo will interact more strongly with liquids containing acidic groups. There is no information regarding the LW and acid–base components of benzyl alcohol, but it is known that alcohols are weak acids as they lose H+ from the hydroxyl group. Accordingly, the wettability of bamboo by benzyl alcohol might be noticeable. As can be seen in Table 2, the contact angle of the system is the smallest among the rest of probe liquids, showing a considerable interaction. However, the total surface energy of benzyl alcohol is also small (see Table 3) leading to a relatively small work of adhesion (see Table 3). Relative atomic percentages, O/C ratio, and decomposition of C 1s peaks obtained by XPS are shown in Table 4. The organic functional groups of lignocellulosic materials identified by XPS [22,23] are consistent with our wetting results, showing both acidic and basic groups (see Table 4). Surface of bamboo technical fibres is composed of lignin mainly instead of cellulose or hemi-cellulose like in case of most other natural fibres (3). Accordingly, the oxygento-carbon atomic ratio (O/C) value of 0.31 (see Table 4) obtained for the surface of bamboo fibres is within the range of that reported for lignin (0.24–0.36) [5,24,25] and far different from that of cellulose with an O/C ratio of 0.83 [24]. The different contributions of functional groups to the shape of the C1s peak have been described in the literature for lignocellulosic materials [22–25]: C–(C,H) linkages of lignin, hemicelluloses and extractives (C1); C OH groups of cellulose, hemicelluloses, lignin and extractives, as well as C O C linkages of (hemi)cellulose and

284.8

286.3

287.6

289.0

C1 (%)

C2 (%)

C3 (%)

C4 (%)

(C (C,H))

(C O)

(C O)

(O C O)

58.02 ± 3.07

28.80 ± 2.34

7.63 ± 1.31

5.59 ± 0.41

extractives (C2); C O groups in lignin and extractives, as well as O C O linkages in cellulose and hemicelluloses (C3); COOH groups of hemicelluloses, as well as COOC and COOH groups of extractives (C4). According to the literature [23,26,27], these four different components of the carbon peak are related to the acid–base properties of materials within a measured depth of 5–10 nm: C1 may be recognized as basic due to its low binding energy, C2 can be defined as acidic owing to the C OH and C O C groups, C3 is considered basic since carbons doubly bound to oxygen are electron donors, and C4 can be judged acidic owing to the presence of carboxyl groups. The high level of C1 component compared to the others is in agreement with the rather basic character of bamboo fibres deduced from the surface energy components (see Table 3). 4.3. Adhesion optimization For the thermoplastic matrix films, advancing and receding contact angles were measured at speeds of 1,0 mm/min with the same set of liquids used to characterize the bamboo surface. The average of the cosines of the advancing ( adv ) and receding ( rec ) angles to estimate the cosine of the equilibrium angle ( eqw ), was used in order to characterise the polar and non-polar surface energy components of the analysed surfaces (see Table 5) [28]. cos equ = 0.5 cos adv + 0.5 cos rec

(13)

In Table 6, the surface energy components of the thermoplastic surfaces, calculated from the equilibrium contact angles (using Eq. (2)), are presented. As expected, PVDF possesses higher acidity due to the different electro-negativities of carbon, fluorine, and hydrogen. In particular, the strong inductive effect of the fluorine atoms polarizes the electronic distribution of partially fluorinated polymers [29]. The results for PET are in accordance with published literature [9,30] wherein it is classified as a basic polymer due to its ester functional groups.

Table 5 Advancing, receding and equilibrium contact angles of probe liquids on thermoplastic surfaces: water (WT), ethylene glycol (EG), diodomethane (DIO), polypropylene (PP), maleic anhydride grafted polypropylene (MAPP), Polyvinylidene fluoride (PVDF), Polyethylene terephthalate (PET). Velocity: 1.0 mm/min. Each value represents the average contact angle of 8 samples. PP

MAPP

PVDF

PET

liquid

adv

rec

equ

adv

rec

equ

adv

rec

equ

adv

rec

equ

WT EG DIO

97.8 ± 1.6 71.1 ± 0.6 67.9 ± 1.2

74.1 ± 1.5 49.4 ± 1.7 46.1 ± 1.7

86.0 ± 1.1 60.8 ± 0.8 57.7 ± 1.0

99.7 ± 1.3 84.8 ± 1.3 77.9 ± 0.9

64.2 ± 0.9 41.7 ± 0.8 38.1 ± 0.5

82.3 ± 0.8 65.3 ± 0.8 60.1 ± 0.5

85.5 ± 0.9 54.1 ± 1.2 63.6 ± 0.4

68.9 ± 1.2 32.1 ± 1.6 46.5 ± 1.8

77.3 ± 0.7 44.2 ± 0.9 55.5 ± 0.8

78.4 ± 0.2 49.5 ± 0.3 38.1 ± 0.1

53.8 ± 0.1 16.8 ± 0.8 25.7 ± 0.2

66.6 ± 0.1 36.5 ± 0.3 32.4 ± 0.1

Table 6 Surface energy components of bamboo fibre and thermoplastic films. Material

 tot (mJ/m2 )

Bamboo PP MAPP PVDF PET

38.82 30.87 29.07 34.64 44.96

± ± ± ± ±

0.76 0.52 0.37 0.53 0.20

 LW (mJ/m2 ) 35.44 29.90 28.52 31.16 43.20

± ± ± ± ±

0.06 0.47 0.31 0.47 0.18

 ab (mJ/m2 ) 3.37 0.97 0.56 3.48 1.76

± ± ± ± ±

0.57 0.20 0.20 0.23 0.11

 + (mJ/m2 ) 0.28 0.12 0.02 0.93 0.15

± ± ± ± ±

0.06 0.05 0.01 0.10 0.02

 − (mJ/m2 ) 10.13 1.97 3.72 3.26 5.21

± ± ± ± ±

1.25 0.31 0.32 0.23 0.08

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

13

Table 7 Wetting parameters with bamboo fibre as the substrate. Material

Wa (mJ/m2 )

PP MAPP PVDF PET

68.81 66.55 74.52 83.14

± ± ± ±

0.97 0.99 0.99 0.86

S (mJ/m2 ) 7.06 8.41 5.23 -6.78

± ± ± ±

For the case of PP, we find a deviation in the magnitude of the polar surface energy component, which was expected to be zero since pure PP is a nonpolar polymer. This effect could be related to aging processes or surface contamination [31]. For the case of MAPP, the determined surface energy shows an expected increment of polar components due to the presence of maleic anhydride [31]. The work of adhesion (Wa ), spreading coefficient (S), wetting tension (F), and interfacial energy ( sl ) for bamboo fibre as substrate were calculated by using Eq. (4), (5), (6), and (7) respectively (see Table 7). Fig. 3 shows a plot of the wetting parameters as a function of the matrix surface energy, keeping the bamboo surface energy constant and using the definition of Good–Girifalco to evaluate the interfacial energy. This theory for the estimation of interfacial energies has been replaced by more modern approaches as the van Oss, Chaudhury and Good theory which was used in this study; however, it is still acceptable for the purpose of showing the variation of the wetting parameters solely as a function of the surface energy of the matrix ( l ). As can be seen in Fig. 3, Wa and F are (largely) increasing while increasing the surface

Fig. 3. Wetting parameters, Wa , F, S and  sl as a function of surface energy of the matrix and using bamboo fibre surface as substrate, and the position of matrices’ surface energies used in this study. The interfacial energy is evaluated by using the 0.5 definition of Good and Girifalco [20,32]: sl = s + l − 2 (s l ) and an interaction parameter ˚ = 1.

0.97 0.99 0.99 0.86

 sl (mJ/m2 ) 0.89 1.34 -1.06 0.64

± ± ± ±

0.32 0.51 0.35 0.36

F (mJ/m2 ) 37.93 37.48 39.88 38.18

± ± ± ±

0.33 0.52 1.30 0.36

energy of the matrix. However, S is going down in case the surface energy of the matrix is higher than 10 mJ/m2 approximately and becomes negative at around 40 mJ/m2 . The latter would mean that the molten matrix would have difficulties to spread on the fibre’s surface during the impregnation process. Looking at the results in Table 7, PET presents the highest Wa due to its high total surface energy and significant polar component (see Table 6), but S is negative which hinders the wetting of bamboo fibres in the molten polymer. Since PET possesses a higher total surface energy than bamboo fibre, a positive S value is not possible (see Fig. 3). The latter represents a serious problem for making fibre reinforced composites considering that a negative S corresponds to partial spreading state on a flat surface. This situation is even worse on a thin fibre since the curvature of the fibre inhibits the wetting power of the molten polymer; in this case the threshold between wetting and non-wetting cannot be at S = 0, but at a positive value [33]. As expected the results for PP and MAPP present a relatively low Wa and F due to their almost null polar component. It has to be noted that these values are only related to intramolecular physical forces and do not relate to covalent bonding (which is important in the case of MAPP). On the other hand, PVDF shows a higher Wa than PP and MAPP, as well as a positive S value and a low  sl , helping to achieve a better wetting of the molten polymer on the bamboo fibre. Furthermore, the F value is the highest corresponding to the situation where Wa is maximum within the region where spontaneous wetting occurs [20]. This strong interfacial interaction is a consequence of electron donor–acceptor interactions due to the presence of a high acid component in PVDF and a high basic component on the bamboo fibre’s surface; as well as a relatively high total surface energy of both the matrix and the substrate. Since polar interactions are electron donor-acceptor interactions, strong interfacial interactions occur only when one phase has basic and the other has acidic sites. Fig. 4 shows the transversal (left) and longitudinal (right) properties of unidirectional MAPP, PP and PVDF bamboo fibre composites made with the same procedure either at 175 ◦ C or 200 ◦ C. PET–bamboo composites could not be prepared due to the high melting point of the polymer (∼260 ◦ C). Depending on the time of

Fig. 4. Left – Transversal flexural properties of MAPP, PP and PVDF bamboo fibre uni-directional composites at 175 ◦ C and 200 ◦ C. Right – Longitudinal properties of MAPP, PP and PVDF bamboo fibre UD composites at 175 ◦ C and 200 ◦ C.

14

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

Fig. 5. Apparent shear strength versus embedded length for PVDF–bamboo fibre (left) and PP–bamboo fibre (right) composites. The curves were obtained by fitting the experimental data with the Greszczuk model [36].

exposure, bamboo fibres start to degrade at temperatures above 200 ◦ C [34]. The transverse 3-point bending test provides a direct estimation of interfacial tensile strength and practical adhesion if the fibre volume fraction is high enough (otherwise the matrix has a great influence). For this study the fibre volume fraction was set at 40%. The values in Table 7 correlate well with the obtained results for the transversal properties, shown in Fig. 4-left. Theoretically, the work of adhesion is slightly higher for PP than MAPP. The small difference in interface tensile strength for MAPP at 200 ◦ C could be explained by the fact that MAPP is capable of forming covalent bonds with the fibre (when activated at temperatures above 170 ◦ C), while PP can only establish physical adhesion. In this study, maleic anhydride grafted to PP is less effective than what other authors reported for different natural fibres. The latter could be related to the fact that the surface of bamboo technical fibres is covered with lignin instead of cellulose or hemi-cellulose like most other natural fibres, presenting a lower concentration of hydroxyl groups [5]. As predicted by Table 7, transversal bending results for PVDF are the best as can be seen in Fig. 5-left, proving that the adhesion has really been improved. There is a difference in interface tensile strength of PVDF–bamboo composites made at different temperatures; this is related to processing conditions since the melting point of PVDF is around 170 ◦ C. At 175 ◦ C, the viscosity of the molten polymer is still relatively high, producing a poorer impregnation. As can be seen in Fig. 4-right, the longitudinal flexural strength of PVDF-bamboo composites is the highest, reaching a value of 220 MPa (∼170 MPa for MAPP and PP). If the latter is compared with the value calculated theoretically (314 MPa) by using the “rule of mixtures” and assuming a tensile strength of 700 MPa for bamboo fibre and 57 MPa for PVDF, the composite reaches 70% of the theoretical strength. In the same fashion, the measured stiffness for the composite (15 GPa) reaches 86% of the theoretical value (17 GPa). The experimental values of flexural strength and Young’s modulus obtained for the PVDF–bamboo system are comparable to the values reported in the literature for epoxy–bamboo composite (78% for flexural strength and 95% for Young’s modulus [1]), indicating good adhesion and fibre alignment. For epoxy–bamboo UD composites typically a transverse bending strength somewhat above 30 MPa is found, compared to nearly 25 MPa for PVDF–bamboo; the transverse bending strength is a good indicator of mode I interface strength (tension loading) and shows that the adhesion in the case of PVDF is close to well-bonding epoxy. While 3-point bending tests evaluate the interfacial properties of composites which are often affected by processing conditions (fibre volume fraction, voids, fibre alignment, and inhomogeneties), micromechanical tests could let us evince the importance of physical interactions at the interface in a more clear way. Accordingly, single fibre pull out tests were performed on PP–bamboo and

Table 8 Results for interfacial shear strength from single fibre pull-out tests. Samples

Ultimate interfacial shear strength (MPa)

Apparent shear strength at 0.6 mm (MPa)

PVDF–Bamboo PP–Bamboo

7.5 3.4

8.4 3.2

PVDF–bamboo composites with the aim to compare both nonpolar and polar systems. Fig. 5-left shows the dependence of the apparent shear strength on embedded fibre length for the PVDF–bamboo fibre system. If compared with PP (Fig. 5-right), the observed pattern for PVDF shows a more brittle interface fracture behaviour since there is no constant function of apparent shear strength versus embedded length [35]. On the contrary, Fig. 5-right reveals a more constant shear strength indicating a less brittle interface for PP and bamboo fibre. The apparent shear strength was calculated according to Eq. (14), assuming that the interfacial loading is a shear mode, there is no friction, the shear stress along the fibre is uniform, and the fibre cross section is circular.

app =

F DL

(14)

where F is the maximum force recorded, D is the diameter of the fibre, and L the embedded length. Two different interfacial shear strength values for PVDF and PP are shown in Table 8. The second column corresponds to the value of apparent shear strength at an embedded length of 0.6 mm, while the first column corresponds to the maximum interfacial shear strength. The latter value is obtained by fitting the experimental data with the model proposed by Greszczuk [36], which provides an extrapolation to zero embedded length, and assumes that the shear stress is small compared to the matrix and that the tensile stress in the matrix is insignificant compared to the fibre. It is evident that higher values are obtained when PVDF is used as a polymer matrix, as predicted by surface energy calculations and confirmed by transverse 3-point bending tests. 5. Conclusions Surface energy components and wetting parameters obtained from contact angle measurements are useful and valuable tools for evaluating the compatibility of natural fibres and various matrices for making composites. In this way, surface energy components of bamboo fibres and thermoplastic matrices were matched, resulting in the improvement of the physical adhesion of bamboo fibre composites, confirmed by 3-point bending and pull-out test results.

C.A. Fuentes et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 418 (2013) 7–15

As predicted, PVDF–bamboo fibre composites present the best combination of wetting parameters, showing a high work of adhesion, as well as a positive spreading coefficient helping to achieve a better wetting of the molten polymer on the bamboo fibre. Accordingly, the wetting tension value is the highest of the four systems presented in this study. This is a consequence of strong physical interactions at the interface due to the presence of a high acid component in PVDF and a high basic component on the bamboo fibre’s surface; as well as a relatively high total surface energy of both the matrix and the substrate. In the case of natural fibres, measuring the contact angle is problematic, and so their wetting behaviour is difficult to study. If bamboo fibres are correctly characterized (in this study, by means of the molecular-kinetic theory of wetting), then its surface represents a well defined system and so its wetting behaviour can be studied and a meaningful interpretation of wetting data is ensured. Acknowledgments Courtesy of Professor Della Volpe we were able to discuss and better understand the surface energy component results. Our thanks to Yasmine Adriaensen and Michel Genet for help with XPS measurements, and Elisa Melcón for help with pull-out test. We would also like to acknowledge the financial support of K.U. Leuven (SBA Scholarship). References [1] L. Osorio, E. Trujillo, A. Van Vuure, I. Verpoest, Morphological aspects and mechanical properties of single bamboo fibers and flexural characterization of bamboo/epoxy composites, J. Reinf. Plast. Compos. 30 (2011) 396–408. [2] M. Vega, C. Gouttiere, D. Seveno, T. Blake, M. Voue, J. De Coninck, A. Clarke, Experimental investigation of the link between static and dynamic wetting by forced wetting of nylon filament, Langmuir 23 (2007) 10628–10634. [3] T. Blake, Dynamic contact angles and wetting kinetics, Wettability 49 (1993) 251. [4] S. Barsberg, L.G. Thygesen, Nonequilibrium phenomena influencing the wetting behavior of plant fibers, J. Colloid Interface Sci. 234 (2001) 59–67. [5] C.A. Fuentes, L.Q.N. Tran, C. Dupont-Gillain, W. Vanderlinden, S. De Feyter, A.W. Van Vuure, I. Verpoest, Wetting behaviour and surface properties of technical bamboo fibres, Colloids Surf. Physicochem. Eng. Aspects 380 (2011) 89–99. [6] T.D. Blake, The physics of moving wetting lines, J. Colloid Interface Sci. 299 (2006) 1–13. [7] T. Blake, J. De Coninck, The influence of solid-liquid interactions on dynamic wetting, Adv. Colloid Interface Sci. 96 (2002) 21–36. [8] C. Van Oss, R. Good, M. Chaudhury, The role of van der Waals forces and hydrogen bonds in "hydrophobic interactions" between biopolymers and low energy surfaces, J. Colloid Interface Sci. 111 (1986) 378–390. [9] C.D. Volpe, S. Siboni, Some reflections on acid-base solid surface free energy theories, J. Colloid Interface Sci. 195 (1997) 121–136. [10] C. Della Volpe, D. Maniglio, S. Siboni, M. Morra, Recent theoretical and experimental advancements in the application of van Oss-Chaudury-Good acid-base theory to the analysis of polymer surfaces I. General aspects, J. Adhes. Sci. Technol. 17 (2003) 1477–1505.

15

[11] C. Della Volpe, S. Siboni, Acid–base surface free energies of solids and the definition of scales in the Good-van Oss-Chaudhury theory, J. Adhes. Sci. Technol. 14 (2000) 235–272. [12] J.C. Berg, Role of acid-base interactions in wetting and related phenomena, Wettability (1993) 76–148. [13] D. Packham, Work of adhesion: contact angles and contact mechanics, Int. J. Adhes. Adhes. 16 (1996) 121–128. [14] M. Nardin, J. Schultz, Relationship between work of adhesion and equilibrium interatomic distance at the interface, Langmuir 12 (1996) 4238–4242. [15] E. Pisanova, S. Zhandarov, E. Mäder, How can adhesion be determined from micromechanical tests? Composites Part A: Appl. Sci. Manuf. 32 (2001) 425–434. [16] D. Bonn, J. Eggers, J. Indekeu, J. Meunier, E. Rolley, Wetting and spreading, Rev. Mod. Phys. 81 (2009) 739. [17] S. Wu, in: Polymer interface and adhesion, CRC, New York, 1982. [18] K. Mittal, The role of the interface in adhesion phenomena, Polym. Eng. Sci. 17 (1977) 467–473. [19] V. Gray, Surface aspects of wetting and adhesion, Chem. Ind. 23 (1965) 969–978. [20] M. Connor, J.E. Bidaux, J.A.E. Manson, A criterion for optimum adhesion applied to fibre reinforced composites, J. Mater. Sci. 32 (1997) 5059–5067. [21] D. Duvivier, D. Seveno, R. Rioboo, T.D. Blake, J. De Coninck, Experimental evidence of the role of viscosity in the molecular-kinetic theory of dynamic wetting, Langmuir 27 (2011) 13015–13021. [22] M. Kazayawoko, J. Balatinecz, R. Sodhi, X-ray photoelectron spectroscopy of maleated polypropylene treated wood fibers in a high-intensity thermokinetic mixer, Wood Sci. Technol. 33 (1999) 359–372. [23] Q. Shen, P. Mikkola, J.B. Rosenholm, Quantitative characterization of the subsurface acid-base properties of wood by XPS and Fowkes theory, Physicochem. Eng. Aspects Colloids Surf. 145 (1998) 235–241. [24] L.S. Johansson, J. Campbell, K. Koljonen, P. Stenius, Evaluation of surface lignin on cellulose fibers with XPS, Appl. Surf. Sci. 144 (1999) 92–95. [25] L.Q.N. Tran, C. Fuentes, C. Dupont-Gillain, A. Van Vuure, I. Verpoest, Wetting analysis and surface characterisation of coir fibres used as reinforcement for composites, Colloids Surf. Physicochem. Eng. Aspects (2011). [26] F.M. Fowkes, Quantitative characterization of the acid-base properties of solvents, polymers, and inorganic surfaces, J. Adhes. Sci. Technol. 4 (1990) 669–691. [27] L. Lavielle, J. Schultz, K. Nakajima, Acid-base surface properties of modified poly (ethylene terephthalate) films and gelatin: Relationship to adhesion, J. Appl. Polym. Sci. 42 (1991) 2825–2831. [28] C. Andrieu, C. Sykes, F. Brochard, Average spreading parameter on heterogeneous surfaces, Langmuir 10 (1994) 2077–2080. [29] S. Lee, J.S. Park, T.R. Lee, The Wettability of Fluoropolymer Surfaces: Influence of Surface Dipoles, Langmuir 24 (2008) 4817–4826. [30] M. Gheorghiu, M. Pascu, G. Popa, C. Vasile, V. Mazur, Poly, (Ethylene Terephthalate) Films with Different Content of Acid–Base Functionalities. I. Surface Modifications, Int. J. Polym. Mater. 40 (1998) 229–256. [31] I. Aranberri-Askargorta, T. Lampke, A. Bismarck, Wetting behavior of flax fibers as reinforcement for polypropylene, J. Colloid Interface Sci. 263 (2003) 580–589. [32] L. Girifalco, R. Good, A theory for the estimation of surface and interfacial energies. I. Derivation and application to interfacial tension, J. Phys. Chem. 61 (1957) 904–909. [33] F. Brochard, Spreading of liquid drops on thin cylinders: The “manchon/droplet" transition, J. Chem. Phys. 84 (1986) 4664. [34] J. Vertommen, Flexural properties of thermoplastic bamboo fibre composites consolidated at different (high) temperatures under inert atmosphere Master Thesis, Lessius Mechelen University College, Mechelen, Belgium, (2011). [35] A. Stamboulis, C. Baillie, E. Schulz, Interfacial characterisation of flax fibrethermoplastic polymer composites by the pull-out test, Die Angew. Makromol. Chem. 272 (1999) 117–120. [36] L. Greszczuk, Theoretical studies of the mechanics of the fiber-matrix interface in composites, Interfaces in Composites 452 (1969) 42.