Wood

Jul 30, 2013 - This project was supported by the National Research Initiative of the USDA ... Matuana , L. M.; Park , C. B.; Balatinecz , J. J. Proces...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/IECR

Strategy To Produce Microcellular Foamed Poly(lactic acid)/WoodFlour Composites in a Continuous Extrusion Process Laurent M. Matuana* and Carlos A. Diaz School of Packaging, Michigan State University, East Lansing, Michigan 48824, United States ABSTRACT: In microcellular foaming process of wood/plastic composites, the alterations in melt viscosity induced by the addition of wood-flour into the plastic matrix play an important role during the growth and stabilization of nucleated cells. This study examined the effects of wood-flour and low molecular weight rheology modifier addition levels on the melt flow index, shear viscosity, and extensional viscosity of PLA/wood-flour composites. From the results, an effective strategy is proposed to successfully and continuously produce microcellular foamed PLA/wood-flour composites with cellular structures similar to those achieved in neat PLA foams. It consists of matching the composites’ melt index value (or melt viscosity) to that of neat PLA by using a rheology modifier. Lowering the melt index of the composites from the matching value resulted in unprocessable foaming conditions owing to the increased melt viscosity, whereas cell structures slightly deteriorated by increasing the melt index of the composite above its optimum value.



wood-flour led to finer cell morphologies in foamed composites.8 On the basis of the literature review, lower processing temperatures during foaming seem to be the appropriate conditions for producing microcellular foamed composites with homogeneous structures because these conditions prevent the devolatilization of wood compounds that could adversely affect the uniformity of the cellular structure. However, the extrusion of PLA/wood-flour composites at lower processing temperature may be difficult given the increased viscosity of the melt when wood particles are present in the formulation.3,11 Our previous study showed that microcellular foaming of neat PLA in an extrusion process requires high pressure in the system5 and the addition of wood-flour into the formulation further increases this pressure. In some cases, this overpressure may lead to unstable processing conditions or exceed operational limits. For this reason, when wood-flour is incorporated into a matrix, a change in the processing conditions (i.e., processing temperature increase) and/or modification in the formulation is necessary to provide the melt with the desired viscosity for processing. Increasing the melt temperature is known to reduce the melt viscosity, as described by an Arrhenius relationship.5,16 Nevertheless, higher processing temperatures could accelerate the degradation of wood-flour and PLA matrix as well as promote the release of volatiles from wood fflour, which are detrimental to the foaming process. Therefore, an alternative approach should be considered to successfully manufacture microcellular foamed PLA/wood-flour composites in a continuous extrusion process. Selection in this study designated controlling the melt viscosity of the composites through the use of a low molecular

INTRODUCTION Melt viscosity is a vitally important parameter in plastic foaming, particularly during cell growth and stabilization.1−7 The presence of wood-flour particles in the formulation affects the microcellular extrusion process of polymers because of their effect on the melt rheology,8 cell nucleation type (heterogeneous nucleation),9 and foamability.7 In the molten state, wood-flour increases the viscosity of the resin,8,10,11 affecting the operating conditions in a foam extrusion system as well as bubble formation, because high melt viscosity induces high pressures in the extrusion system5 and also offers higher resistance to cell growth.1−5 It has been also shown that woodflour particles can act as heterogeneous nucleation sites,9 which can affect the overall nucleation rate. Additionally, when exposed to heat, wood-flour releases water vapor and volatile components that can negatively affect the development of the cellular structure in foamed composites.12−14 The motivation for the inclusion of wood-flour into plastic matrices has been described by several investigators.1,3,7,8,11−14 Foaming of plastic/wood-flour composites at processing temperature above 170 °C has been shown to adversely affect the development of the cell morphology. Guo and co-workers foamed HDPE with 50 wt % wood-flour using the volatiles generated by the wood-flour as foaming agent.12 They pointed out that when wood-flour is part of the formulation, higher processing temperatures in the extrusion barrel contribute to its degradation by releasing water vapor and volatiles in the system. Other investigators reported similar results for foamed wood/plastic composites based on polyvinyl chloride13 and polystyrene14 matrices. A processing temperature of 170 °C or lower was suggested to suppress adverse effects from the volatiles generated.12,14 Above this processing temperature, these residues are detrimental to the foam morphology and should be kept to a minimum. Additionally, higher processing temperatures could initiate thermal degradation of PLA, causing chain scission.15 In a subsequent study, Guo and coworkers evaluated the influence of fiber size on the foamability of HDPE/wood-fiber composites and showed that finer sizes of © 2013 American Chemical Society

Received: Revised: Accepted: Published: 12032

June 20, 2013 July 15, 2013 July 30, 2013 July 30, 2013 dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

die was kept constant to maintain a constant melt temperature for accurate viscosity measurements. All of the heating zones were set to 170 °C. The rotational screw speeds varied from 10 to 80 rpm to generate different shear rates during experiments.5,11 Experiments followed ASTM standard D5422 without injecting CO2 in the system. The melt density used in the viscosity calculations was measured as described in a previous work.4 Apparent shear stress (τa) and apparent shear rate (γ̇a) were calculated on the basis of the following equations

weight rheology modifier in the formulation as a strategy to reduce the processing temperatures during foaming. Low molecular weight additives are known for their low viscosity and plasticizing effects, a desirable characteristic to control the viscosity of a blend. It was hypothesized that by controlling the viscosity of the composite melt through changes in the formulation, the foaming process could be carried out under processing conditions similar to those used in the manufacture of microcellular PLA without raising the processing temperature. Processing at lower temperatures would favor the foamability of the composites. Therefore, the aim of this study was to produce microcellular structures in PLA/wood-flour composites using processing conditions not detrimental to the thermal stability of both wood-flour and PLA matrix, while preventing the release of volatiles from wood-flour. Particular emphasis was placed on examining the effects of wood-flour content and the concentration of the low molecular weight rheology modifier on the melt flow index, shear viscosity, and extensional viscosity of the composites in order to identify their relative amounts suitable for microcellular extrusion foaming of PLA/wood-flour composites.

τa =

ΔP 4(L /D)

(1)

γȧ =

32Q πD 3

(2)

with ΔP as the pressure drop across the capillary, L/D the length-to-diameter ratio of the capillary insert, Q the volumetric flow rate of the polymer melt, and D the diameter of the capillary. Bagley and Rabinowitsch corrections account for the excess of pressure drop due to entrance flow and the shear thinning effect of pseudoplastic fluids, respectively. By using capillaries with different L/D ratios, the pressure drop at the entrance can be isolated and included in the calculation of shear stress (τw) as an “end effect” (e) as follows:21



EXPERIMENTAL SECTION Materials. A semicrystalline grade of polylactic acid (PLA 2002D) from NatureWorks was used as matrix with an approximate D-lactic content of 4% and a melt flow index between 5 and 7 g/10 min (210 °C/2.16 kg).4 The wood-flour used was pine flour (120 mesh size, grade 12020), supplied by American Wood Fibers (Schofield, WI). To promote heterogeneous nucleation, talc (Mistron Vapor-R) from Luzenac Corp., with a median particle size of 2 μm and a specific surface area of 13.4 m2/g, was used. The rheology modifier utilized in the study was Epolene E-43, a low molecular weight maleic anhydride-modified polypropylene, supplied by Eastman Chemical Co. (Kingsport, TN). Its characteristics were as follows: 8% maleic anhydride, melting temperature of 155 °C, approximate weight average molecular weight of 11 200 g/mol, and a viscosity of 0.3 Pa·s (at 190 °C). This low molecular weight maleated polypropylene was used because of its low viscosity and affinity for the cellulosic substrate and the polymer.17−20 Carbon dioxide (CO2) with a purity of 99.5%, supplied by Airgas, was used as physical foaming agent. Rheological Measurements. Melt flow index (MFI) and online capillary rheometry were the complementary techniques used to assess the effects of wood-flour and rheology modifier (E-43) contents on the rheological behaviors of PLA/woodflour composites. Melt Flow Index (MFI). MFI was measured according to the procedure outlined in ASTM D1238 (Procedure A) using a melt indexer (model LMI 4000) by Dynisco Polymer Testing Inc. The melt temperature was set at 210 °C, and an applied dead load of 2.16 kg (including the piston) was employed. A melt time of 2 min was set in order to avoid PLA thermal degradation.4 Shear Viscosity. Online capillary rheometry measurements were performed on an Intelli-Torque Plasticorder torque rheometer (C.W. Brabender Instruments Inc.) equipped with a 19.1 mm single screw extruder (L/D ratio 30:1) to determine the shear viscosity of the melts. A capillary die with three different inserts (L/D ratios of 10:1, 15:1, 20:1) were use for the experiments. The temperature profile in the extruder and

τw =

ΔP 4(L /D + e)

(3)

The end correction (e) was calculated from the plot of pressure (ΔP) versus L/D for each processing screw speed using a linear extrapolation. The absolute value of L/D at the point of interception (ΔP = 0) is the end correction e. The apparent shear rate, which gives only Newtonian behavior (constant viscosity), required a correction for shearthinning fluids (pseudoplastic fluids). The Rabinowitsch correction was done to calculate the true shear rate at the wall (γ̇w) of the capillary by using the following equation γẇ =

3n + 1 γ̇ 4n a

(4)

with n as the flow behavior index obtained as the slope of the linear plot of log τw vs log γ̇a. The true melt viscosity (η) was then calculated, from the corrected shear stress and shear rate, as follows: η=

τw γẇ

(5)

Extensional Viscosity. Two methods were employed to characterize the effects of wood-flour and rheology modifier addition levels on the extensional viscosity of PLA: (1) online capillary rheometry and (2) dynamic rotational rheometer. From the entrance pressure drop in the online capillary measurements, apparent extensional viscosity was calculated using Cogswell’s equations.21−23 The extensional strain rate (ε̇A) and the extensional shear stress (σE) are given by εȦ = 12033

4γȧ 2ηa 3(n + 1)ΔPe

(6)

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

3 (n + 1)ΔPe (7) 8 where γ̇a is the apparent shear rate, ηa the apparent shear viscosity, and ΔPe the pressure drop at the capillary entrance, calculated from the end effect as

Foam Characterization. Densities measurements were carried out according to ASTM standard D792 (buoyancy method), for unfoamed (ρu) and foamed (ρf) samples. The results were then used to calculate the void fraction (Vf) using the following equation:1−5

σE =

ΔPe = 4eσw

(8)

Vf =

Finally, the apparent extensional viscosity, which is the ratio of the extensional shear stress to the extensional strain rate, can be calculated as ηEC =

9(n + 1)2 (ΔPe)2 σE = εȦ 32ηa γȧ 2

(ρu − ρf ) ρu

× 100 (10)

Morphology of foamed samples was analyzed via scanning electron microscope (SEM JEOL JSM-6400, 10 kV). Samples were fractured in liquid nitrogen and coated with gold prior to the test. The images obtained aided the characterization of cellpopulation density (No)1,5

(9)

Direct measurement of the extensional viscosity was done using the extensional viscosity fixture (EVF) on the ARES extensional rheometer from TA Instruments (New Castle, DE). Rectangular samples with 18 × 10 × 0.78 mm dimensions were molded at 180 °C. Extensional viscosity was measured at a fixed Henky strain rate (0.1 s−1) as a function of time at 160 °C. Further discussion of the method can be found elsewhere.24,25 Compounding of Materials. PLA in powder form (ground from pellets) and wood-flour used in the formulations were oven-dried at 55 °C for 8 h and 105 °C for 48 h prior to use, respectively. PLA, wood-flour (0−30 wt %), rheology modifier (1.3−6 wt %), and 0.5 wt % of talc were dry-mixed in a kitchen mixer (Blender MX1050XTS from Warning Commercial Xtreme) at 22 000 rpm for approximately 45 s. Extrusion Foaming. Figure 1 shows a schematic of the foaming system used in this study. The process used a 19.1 mm

⎛ nM2 ⎞3/2 ⎡ 1 ⎤ No = ⎜ ⎥ ⎟ ⎢ ⎝ A ⎠ ⎣ 1 − Vf ⎦

(11)

with n as the number of cells in the micrographs and A and M as the area and the magnification factor of the micrograph, respectively. Vf is the void fraction described in eq 10. The average cell sizes were measured using image analysis software (UTHSCSA Image Tool). After a proper spatial calibration of the micrograph, the diameters of at least 100 randomly selected cells were manually measured, and the average cell size was calculated. Breaking the cell size values into classes allowed the making of histograms and cumulative percentage plots. Statistical Analysis. Design-Expert v.7 software from the Stat-Ease Corp. (Minneapolis, MN) was used to perform statistical analyses. A two-sample t test and Duncan’s multiple range tests were employed to determine the statistical differences among the variables investigated at a 95% significance level.



RESULTS AND DISCUSSION Rheological Characterization of Unfilled PLA and PLA/Wood-Flour Composites. Plots of MFI as a function of rheology modifier (E-43) contents for composites with different amounts of wood-flour are shown in Figure 2. Two distinct trends are observed in this figure. First, increasing wood-flour content reduced the MFI of the composites, regardless of rheology modifier content. The results suggest that increasing wood-flour content makes the composite melts Figure 1. Schematic of the single-screw extrusion foaming system.

single screw extruder (L/D ratio 30:1, C.W. Brabender Instruments Inc.) with a mixing screw powered by a 3.73 kW (5 hp) driver with speed control gearbox (Brabender Prep Center). A positive displacement syringe pump (Teledyne Isco., model 260D) delivered the gas into the extrusion barrel. A diffusion-enhancing device (static mixer Omega FMX8441S) attached to the extruder ensured the creation of a polymer/gas solution with a rapid pressure drop induced by a nozzle die with a 0.5 mm diameter. All foaming experiments were performed with gas injected at a constant volume of 5 mL/min (approximately 5 wt %) and screw rotational speed of 80 rpm. The processing temperatures was set in the extruder barrel 170−170−165 °C from hopper to extruder’s exit, 165 °C in the static mixer, and 150 °C in the nozzle die. Samples were collected after the process reached steady-state conditions and were cooled by air.

Figure 2. Effects of wood-flour and rheology modifier contents on the MFI of PLA 2002D at 210 °C. The dashed line represents the MFI of neat PLA 2002D. The 10, 20, and 30% in the legend refer to woodflour contents. 12034

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

valuable insights on the importance of melt rheology on the foamability of PLA/wood-flour composites. Let us consider first the effects of wood-flour content and E43 concentration on the shear viscosity of the composites. The true viscosities of selected formulations were measured as a function of shear rate. Figure 3 shows the double-logarithmic

more viscous. Interactions between wood-flour particles and the matrix may account for this decreased trend,8 since studies have shown that solid wood-flour particles embedded in the polymer matrix increase the friction toward the flow of the melt, resulting in higher viscosity.26 Second, the MFI of the composites increased significantly with the rise in rheology modifier content, irrespective of wood-flour addition levels. Low molecular weight additives such as E-43 used in this study can produce a plasticizing effect by separating the polymer chains and increasing their flexibility27 in addition to inducing a lubricating effect, which reduces the friction between the polymer and the wood-flour.28 This plasticizing/lubricating effect enhances the flowability of the composite melts. In the range studied, the effect of rheology modifier content on MFI of the melts was linear for all three wood-flour contents (Figure 2). Noticeably, the slope of the linear relation was steeper as the wood-flour content decreased to zero, meaning that the efficiency of the additive was higher in the melts with lower wood-flour contents. In other words, an equal increment in rheology modifier content produced a higher increase in MFI for melts with lower wood-flour contents than for the melt with higher wood-flour contents. These results indicate that there may be a significant interaction between the wood-flour particles and the rheology modifier (E-43). It is also noted that by drawing a horizontal line parallel to the x-axis at MFI of ∼6.9 g/10 min, which corresponds to the MFI of unfilled PLA (dashed line in Figure 2), three intersection points with the data can be obtained depending on wood-flour contents. The first intersection point is detected at the MFI curve of the composites with lower wood-flour content (10%), and the next two are seen at MFI curves of the composites with higher wood-flour contents (20% and 30%). These intersection points, which can also be predicted by the regression equations derived from the MFI−rheology modifier content curves (Figure 2), correspond to the optimum amount of rheology modifier needed in the composites to alter the value of their melt indexes (or equivalently the melt viscosity) to equal (or be closer to) that of neat PLA (Table 1). Results

Figure 3. True viscosity vs true shear rate of PLA, PLA with 20 wt % wood-flour, and PLA with 20 wt % wood-flour and 2 wt % rheology modifier.

plots of the true viscosity versus the true shear rate of the three formulations. All melts showed shear thinning behavior and obeyed the power law as described by the following equation5 η = Kγẇ n − 1

where K and n are constants representing the melt viscosity coefficient and power law flow index, respectively.5 Since all melts in Figure 3 showed a shear thinning behavior and obeyed the power law equation, the curve fitting of this equation was performed to derive the constants representing the melt viscosity coefficient (K) and power law flow index (n) following the previously described approach.5 Compared to unfilled PLA, a shift in the viscosity curve upward was observed for the composite without rheology modifier (Figure 3). Additionally, an increase in the melt viscosity coefficient (K) and a decrease in flow index (n) when the wood-flour was present in the melt (Table 2) indicated a

Table 1. Linear Relationships of the Effect of Rheology Modifier Content on MFI for Each Wood-Flour Content

a

wood-flour content (wt %)

linear eqa

R2

predicted E-43 concn (wt %)

10 20 30

2.495x + 6.256 2.006x + 3.504 1.279x + 0.959

0.966 0.982 0.996

0.253 1.688 4.637

(12)

Table 2. Melt Viscosity Coefficient (K) and Power Law Flow Index (n) for Neat PLA, PLA with 20 wt % Wood-Flour, and PLA with 20 wt % Wood-Flour and 2 wt % Rheology Modifier

x = E-43 content.

regression parameters: η = Kγ̇n−1

listed in Table 1 clearly indicate that the amount of rheology modifier (E-43) needed in the composites to match the MFI of unfilled PLA was a strong function of wood-flour content. This amount increased as the wood-flour content increased in the composites. Although melt flow index is related to the melt viscosity,4 it does not provide the actual flow properties encountered during processing operations. Consequently, viscosity characterization at typical extrusion shear rates (100−1000 s−1)5,21,29 is important to understand the flow of the material during foaming. Additionally, since both shear and extensional viscosities play a key role in the dynamics of bubble growth during foaming process, their assessment could provide

material composition

K (kPa·sn)

n

R2

neat PLA PLA + 20% WF PLA + 20% WF + 2% E-43

30.6 51.7 11.4

0.318 0.257 0.456

0.9951 0.9999 0.9999

stronger shear thinning effect,5,30 implying that the composite melt was more viscous than the unfilled matrix. Shear thinning responds to the alignment and disentanglement of the polymeric chains when subjected to shear; adding wood-flour into a polymer matrix is known to accentuate this behavior, because the fibers also align and induce higher local shear rates in the thin polymeric layer between wood-flour particles, thus the observed reduction in the flow index (n).30 Conversely, the 12035

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

(eqs 6−9). Irrespective of the formulation, as the extensional strain rate increased, the extensional viscosity decreased in a linear manner; thus, a power law relation persists, similar to that of the shear viscosity (Figure 3). Neat PLA showed the lowest extensional viscosity. When wood-flour was added in the formulation, the extensional viscosity increased, confirming the resistance to flow effect of wood-flour shown in Figures 2 and 3. Similar results were reported by Li and Wolcott in their study on the rheology of HDPE/wood-flour composite melts that also showed a strong dependence of extensional viscosity on woodflour content, with extensional viscosity of melts increasing with increasing wood-flour loading in the composites.10 While the shear viscosity results (Figure 3) supported the MFI data (Figure 2), suggesting that the inclusion of rheology modifier into the composite reduces its viscosity, a different trend was obtained from the extensional viscosity measured from online capillary rheometry. The addition of rheology modifier appeared to have no significant effect on the extensional viscosity of the composite melts, as seen by the overlapped viscosity curves of the composites with and without rheology modifier (Figure 4). In contrast, Li and Wolcott reported a slight decrease in the extensional viscosity when a rheology modifier (maleated polyethylene) was added into HDPE/wood-flour flour composite.10 Since the flow that takes place in MFI measurements involves contracting flow through a die, MFI values are affected by both shear and extensional viscosities. It is believed that adding a rheology modifier to the composite lowered the shear viscosity, but the extensional viscosity remained at the same level as for the composite without modifier. This could explain why, although the extensional viscosity of neat PLA and the composite containing rheology modifier did not overlap (Figure 4a), similar values of MFI were obtained for these two formulations. Additionally, some of the small differences may be due to the fact that a concentration of 2 wt % E-43 was used instead of 1.7 wt % E43, as suggested by the linear regression (Table 1). It should be pointed out that the values for extensional viscosity were 1 order of magnitude higher than those of shear viscosity. This difference has been observed for other plastics in the power law region.22 Extensional viscosity measurements in a rotational rheometer were performed to corroborate the results from capillary rheometry (Figure 4b). Without rheology modifier, the extensional viscosity increased by adding wood-flour into the matrix (Figure 4), confirming the resistance to flow effect of wood-flour shown in Figures 2 and 3. The addition of rheology modifier (E-43) slightly reduced the extensional viscosity of the composite. As expected from the chain architecture of PLA, none of the formulation showed strain hardening. Similar results were reported by other investigators.25 Influence of Melt Rheology on the Cell Morphology of Foamed Composites. It was hypothesized that composites with MFI values similar to that of unfilled PLA will yield the best cellular morphologies in foamed samples. Thus, the values of predicted rheology modifier contents listed in Table 1 were used as a reference point for foam processing of composites. Composites containing 20% wt wood-flour and different amounts of rheology modifier (1.3−6 wt %) were foamed to understand the importance of the melt viscosity on the foamability of the composites. The range of addition levels of rheology modifier (E-43) was chosen to alter the MFI of the composites to the values below, equal to (or closer to), and

addition of rheology modifier into the composites lowered its melt viscosity curve downward (Figure 3) and made the composite less pseudoplastic (decreased K and increased n) (Table 2). This behavior was attributed to the plasticizing effect of the additive, which increased the mobility of the polymer chains and the flow rate of the melt.28 These results corroborate the MFI data shown in Figure 2. Nevertheless, it is worth mentioning that the melt viscosity curve of the composite with rheology modifier did not overlap (slightly lower) with that of unfilled PLA (Figure 3) because of the higher amount of rheology modifier (2 wt %) used in the composites instead of the predicted optimum amount needed to match the melt index of PLA, i.e., ∼1.7% (Table 1). Beside shear viscosity, elastic properties of the melt such as extensional viscosity also play an important role in foam processing. Extensional viscosity is responsible for entrance pressures when the polymer flows through contracting channels or tubes,31 affecting the pressures recorded during online capillary rheometry measurements and the pressures built in the extrusion system when foaming.5 Furthermore, as bubbles grow during foam processing, the polymer is stretched, making the elastic properties of the melt a decisive factor in controlling the cell morphology.6 The polymer resistance to stretching is proportional to the stretching rate, and this proportionality is referred as the extensional viscosity.32 Second, let us consider the effects of wood-flour and rheology modifier contents on the extensional viscosity of the composites (Figure 4). Figure 4a shows the double logarithmic plot of apparent extensional viscosity versus apparent strain rate calculated from the end effects using the Cogswell’s equations

Figure 4. Extensional viscosity measured by (a) online capillary rheometry and (b) ARES-EVF rheometer for PLA, PLA with 20 wt % wood-flour, and PLA with 20 wt % wood-flour and 2 wt % rheology modifier. 12036

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

Figure 5. Effect of rheology modifier content on cell morphology of foamed PLA samples containing 20 wt % wood-flour: (a) 2 wt %, (b) 4 wt %, and (c) 6 wt % E-43. Pictures originally taken at 350×.

above that of unfilled PLA (∼6.9 g/10 min). Therefore, adding 1.3, 2, 4, and 6 wt % of E-43 into the composites led to the melts with MFI values of 6.1, 7.5, 11.5, and 15.5 g/10 min, respectively. The foaming process was strongly affected by the melt flow indexes of the composites. The extrusion foaming process was unstable below the matching point (1.3 wt % E-43) due to the increased melt viscosity (or lower MFI value of 6.1 g/10 min). The composite melt experienced contraction flow through the static mixer and the nozzle (Figure 1), generating entrance pressures that led to a rise in pressure above the operational limit to a point that the extrusion system was shut down. In addition, the nozzle used for these experiments was particularly small (0.5 mm in diameter), and without the proper lubrication provided by the rheology modifier, the wood-flour particles may have agglomerated, leading to the nozzle being blocked. By contrast, the foaming process was successful in composites with MFI values closer to and above that of unfilled PLA (i.e., 2, 4, and 6 wt % E-43). Foamed composites with uniform and homogeneous cell structures having approximately 10 μm in average size were obtained when the MFI of the composites was closer to the matching point (Figure 5a). It is believed that the melt viscosity of the composites containing 2 wt % E-43 was low enough and suitable to allow cell formation and growth but high enough to prevent cell coalescence. Cell structures of foamed composites slightly deteriorated above the matching points (4 and 6 wt % E-43), showing large cracks and cell coalescence (Figures 5b,c). When the melt indexes of composite increased above the matching point, the melt viscosity was lowered too much (higher MFI) and provided insufficient melt strength to trap the growing bubbles; thus, cells coalesced.6 The void fraction of foamed samples decreased significantly as the amount of rheology modifier (E-43) increased in the composites (Figure 6), which softened the resin, thus favoring cell coalescence.1 As anticipated from cell coalescence, the cellpopulation density in foamed composites decreased and the average cell size increased with the modifier content (Table 3). Our experimental results not only indicate that the melt rheology of the composites is a critical parameter for their foamability in a continuous extrusion process but also validate the adequacy of our proposed strategy of tailoring the melt index (or melt viscosity) of the composites to equal that of unfilled PLA. Unprocessable foaming conditions resulted and the composites could not be foamed when their melt indexes were below that of unfilled PLA, whereas cellular structures meeting the characteristics of microcellular plastics were not achieved in the composites with melt indexes above that of

Figure 6. Effect of rheology modifier content on void fraction of foamed PLA samples containing 20 wt % wood-flour. The means with different letters denote that the difference between two treatments is statistically significant (p < 0.05).

Table 3. Effect of Rheology Modifier Content on CellPopulation Density and Cell Size of Foamed Composites Containing 20 wt % Wood-Flour E-43 content (wt %)

cell-population density (109 cells/cm3)a

av cell size (μm)

2 4 6

0.67 ± 0.106A 0.28 ± 0.094B 0.08 ± 0.043C

9.2 15.5 18.3

a The different superscript letters denote that the difference between two treatments is statistically significant (p < 0.05).

unfilled PLA. Microcellular structures were successfully produced in the composites only when its melt index matched (or was closer to) that of unfilled PLA. Effect of Wood-Flour Content on the Cell Morphology of Foamed Composites. Microcellular extrusion foaming of PLA/wood-flour composites with various wood-flour contents was performed. These composites contained optimum amounts of rheology modifier specific to each wood-flour content in order to alter their melt indexes to the value closer or equal to that of unfilled PLA. Consequently, 0.3, 2, and 5 wt % E-43 were added to the formulations containing 10, 20, and 30 wt % wood-flour, respectively. It should be mentioned that the amounts of the modifier used in the composites were slightly higher than the optimum values predicted in Table 1 to ensure processability of the composites. The cellular morphology of foamed composites is illustrated in Figure 7. More uniform microcellular structures were 12037

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

Figure 8. Effect of wood-flour content on void fraction of foamed PLA/wood-flour composites samples. The differences in means were not statistically significant.

Figure 7. Effect of wood-flour content on cell morphology of foamed PLA samples: (a) 0 wt %, (b) 10 wt %, (c) 20 wt %, and (d) 30 wt % wood-flour. Pictures originally taken at 350×

increasing the wood-flour content in the PLA matrix tends to decrease noticeably the void fraction of PLA in foamed samples.3 This tendency was attributed to the expected effect of adding wood-flour in the matrix, as it not only reduces the cellpopulation density, as mentioned above, but it also increases the melt viscosity of the matrix, providing resistance to cell growth in the polymer matrix.1,3,33 Unlike the previous study, the results of this investigation clearly suggest that the void fraction of the composites is insensitive to the amount of woodflour, as long as the melt viscosity (or melt index) of the composites is modified to equal or be closer to that of unfilled matrix. Matching the melt index of the matrix provides the composites with melt viscosity suitable to allow bubble nucleation and growth without severe coalescence. Figure 9 illustrates the cell size distribution of the microcellular foamed PLA and PLA/wood-flour composites

achieved in the composites with up to 20 wt % wood-flour (Figure 7b,c), as in the case of foamed neat PLA (Figure 7a). Above this concentration, foam structures begun to slightly deteriorate (Figure 7d), showing large cells with irregular shape, probably attributable to cell coalescence. The cell-population density decreased gradually with increasing amounts of wood-flour (Table 4). This decreased Table 4. Effect of Wood-Flour Content on Cell Population Density and Cell Size of Foamed Samples wood-flour content (wt %)

rheology modifier (E-43) content (wt %)

0 10 20 30

0 0.3 2 5

cell population density (109 cells/cm3)a 1.46 1.02 0.67 0.26

± ± ± ±

0.54A 0.30AB 0.11B 0.08C

av cell size (μm) 7.4 8.9 9.3 11.3

a The different superscript letters denote that the difference between two treatments is statistically significant (p < 0.05).

trend in the number of cells nucleated per unit volume was expected, because the volume fraction of the matrix in the composites, i.e., the plastic part that is available to dissolve the gas and form bubbles, decreased by increasing wood-flour content in the composites.3,33 In other terms, since the plastic component available to dissolve the gas and form bubbles is reduced as the amount of wood-flour increased, the cellpopulation density was reduced at higher wood-flour contents. Nucleation mechanisms during foaming may also account for this reduced cell-density population. Without wood-flour (Figure 7a) the bubble nucleation is mainly homogeneous.5 For small amounts of wood-flour, i.e., 10 wt % (Figure 7b), it is believed that this nucleation mechanism remained the dominant one. However, at higher wood-flour concentrations, i.e., 30 wt %, bubbles may have nucleated at the interface of the wood-flour particles and the plastic, reducing the overall nucleation rate and thus lowering cell-population density. Figure 8 shows the void fraction of the foamed composites, which remained unaffected by the addition of up to 20 wt % wood-flour to the matrix. Although the void fraction appeared to slightly decreases above 20 wt %, the difference was not statistically significant. We reported in a previous study that

Figure 9. Effect of wood-flour content on cell size and cell size distribution of foamed samples. The 0, 10, 20, and 30% in the legend refer to wood-flour contents.

as a function of wood-flour content. Notice that more than 90% of the cells in unfilled PLA had cell sizes below 10 μm, with a high degree of cell uniformity given by the steepness of the curve. Foamed composites containing 10 and 20 wt % showed similar cell size distributions, with about 60% of the cells being smaller than 10 μm and 90% smaller than 15 μm. Above 20 wt % wood-flour, the cell size distribution in the composites 12038

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

Notes

became broader and less uniform (less steep curve), indicating larger cells in the samples. Higher wood-flour concentration may have provided sites for undissolved gas pockets during processing, leading to uncontrolled cell growth and overall deterioration of the morphology.34 Additionally, wood-flour could release volatiles, which may have played a deleterious effect on the cell structure, especially at the highest wood-flour content.12 Overall, the average cell sizes of foamed samples decreased by increasing the wood-flour content in the matrix (Table 4), in agreement with the cell-population density results (Table 4).

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This project was supported by the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service (grant number 2006-35504-17414).





(1) Matuana, L. M.; Park, C. B.; Balatinecz, J. J. Processing and Cell Morphology Relationships for Microcellular Foamed PVC/WoodFiber Composites. Polym. Eng. Sci. 1997, 37, 1137. (2) Matuana, L. M. Solid State Microcellular Foamed Poly(lactic acid): Morphology and Property Characterization. Bioresour. Technol. 2008, 99, 3643. (3) Matuana, L. M.; Faruk, O. Effect of Gas Saturation Conditions on the Expansion Ratio of Microcellular Poly(lactic acid)/Wood-Flour Composites. eXPRESS Polym. Lett. 2010, 10, 621. (4) Matuana, L. M.; Faruk, O.; Diaz, C. A. Cell Morphology of Extrusion Foamed Poly(lactic acid) using Endothermic Chemical Foaming Agent. Bioresour. Technol. 2009, 100, 5947. (5) Matuana, L. M.; Diaz, C. A. Study of Cell Nucleation in Microcellular Poly(lactic acid) Foamed with Supercritical CO2 through a Continuous Extrusion Process. Ind. Eng. Chem. Res. 2010, 49, 2186. (6) Diaz, C. A; Matuana, L. M. Continuous Extrusion Production of Microcellular Rigid PVC. J. Vinyl Addit. Technol. 2009, 15, 211. (7) Mengeloglu, F.; Matuana, L. M. Foaming of Rigid PVC/WoodFlour Composites through a Continuous Extrusion Process. J. Vinyl Addit. Technol. 2001, 7, 142. (8) Guo, G.; Lee, Y. H.; Rizvi, G. M.; Park, C. B. Influence of Wood Fiber Size on Extrusion Foaming of Wood Fiber/HDPE Composites. J. Appl. Polym. Sci. 2008, 107, 3505. (9) Rodrigue, D.; Souici, S.; Twite-Kabamba, E. Effect of Wood Powder on Polymer Foam Nucleation. J. Vinyl Addit. Technol. 2006, 12, 19. (10) Li, T. Q.; Wolcott, M. P. Rheology of Wood Plastics Melt. Part 1. Capillary Rheometry of HDPE filled with Maple. Polym. Eng. Sci. 2005, 45, 549. (11) Shah, B. L.; Matuana, L. M. Online Measurement of Rheological Properties of PVC/Wood-Flour Composites. J. Vinyl Addit. Technol. 2004, 10, 121. (12) Guo, G.; Rizvi, G. M.; Park, C. B.; Lin, W. S. Critical Processing Temperature in the Manufacture of Fine-Celled Plastic/Wood-Fiber Composite Foams. J. Appl. Polym. Sci. 2004, 91, 621. (13) Matuana, L. M.; Mengeloglu, F. Manufacture of Rigid PVC/ Wood-Flour Composite Foams Using Moisture Contained in Wood as Foaming Agent. J. Vinyl Addit. Technol. 2002, 8, 264. (14) Rizvi, G.; Matuana, L. M.; Park, C. B. Foaming of PS/Wood Fiber Composites using Moisture as a Blowing Agent. Polym. Eng. Sci. 2000, 40, 2124. (15) Garlotta, D. A Literature Review of Poly(lactic acid). J. Polym. Environ. 2001, 9, 63. (16) Muksing, N.; Nithitanakul, M.; Grady, B. P.; Magaraphan, R. Melt Rheology and Extrudate Swell of Organobentonite-Filled Polypropylene Nanocomposites. Polym. Test. 2008, 27, 470. (17) Kazayawoko, M.; Balatinecz, J. J.; Matuana, L. M. Surface Modification and Adhesion Mechanisms in Woodfiber−Polypropylene Composites. J. Mater. Sci. 1999, 34, 6189. (18) Li, Q.; Matuana, L. M. Surface of Cellulosic Materials Modified with Functionalized Polyethylene Coupling Agents. J. Appl. Polym. Sci. 2003, 88, 278. (19) Li, Q.; Matuana, L. M . Effectiveness of Maleated and Acrylic Acid-Functionalized Polyolefin Coupling Agents for HDPE−WoodFlour Composites. J. Thermoplast. Compos. Mater. 2003, 16, 551. (20) Carlborn, K.; Matuana, L. M. Composite Materials Manufactured from Wood Particles Modified through a Reactive Extrusion Process. Polym. Compos. 2005, 26, 534.

CONCLUSIONS The presence of wood-flour particles in the formulation affects the microcellular foaming of plastics because of their effect on the melt rheology, which plays an important role during cell growth and stabilization. This study was aimed to produce microcellular structures in PLA/wood-flour composites using processing conditions not detrimental to the thermal stability of both wood-flour and PLA matrix, while release of volatiles from wood-flour was prevented. Particular emphasis was placed on examining the effects of wood-flour content and low molecular weight rheology modifier addition levels on the melt flow index, shear viscosity, and extensional viscosity of the composites in order to identify their relative amounts suitable for producing microcellular PLA/wood-flour composites in a continuous extrusion process with supercritical carbon dioxide as foaming agent. The following conclusions were drawn from the experimental results: Melt index of PLA decreased as the wood-flour content increased in the matrix. However, this increased melt viscosity induced by the resistance to flow effect of wood was offset by incorporating various concentrations of the rheology modifier into the composites. Regression equations derived from the melt index versus rheology modifier content curves of composites with different wood-flour contents assisted in predicting the optimum amount of rheology modifier needed to alter the value of the melt index (or equivalently the melt viscosity) of the composites to equal (or be closer to) that of neat PLA. An effective strategy is proposed from the experimental results to successfully produce microcellular foamed PLA/ wood-flour composites with uniform and homogeneous cellular structures similar to those achieved in neat PLA foams. Matching the MFI of the composites to that of neat PLA matrix resulted in the best cell morphologies. Lowering the MFI of the composites from the matching point values resulted in unprocessable foaming conditions, whereas increasing it resulted in foamed composites with poor cell morphologies. Matching the MFI (or viscosity) of the composite melt to that of the neat PLA by means of low molecular weight rheology modifier proved to be an effective way and a suitable approach to facilitate the continuous extrusion microcellular foaming process of the composites. Using this approach, microcellular foams with uniform morphologies were successfully produced in PLA/wood-flour composites in concentrations up to 20 wt % wood-flour.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*Tel: (517)-353-4616. Fax: (517)-353-8999. E-mail: [email protected]. 12039

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040

Industrial & Engineering Chemistry Research

Article

(21) Dealy, J. M.; Wissbrun, K. F. Melt Rheology and Its Role in Plastics Processing; Van Nostrand Reinhold: New York, 1990. (22) Huang, J. C.; Leong, K. S. Shear viscosity, Extensional Viscosity, and Die Swell of Polypropylene in Capillary Flow with Pressure Dependency. J. Appl. Polym. Sci. 2002, 84, 1269. (23) Tzoganakis, C.; Vlachopoulos, J.; Hamielec, A. E.; Shinozaki, D. M. Effect of Molecular-Weight Distribution on the Rheological and Mechanical Properties of Polypropylene. Polym. Eng. Sci. 1989, 29, 390. (24) Hadinata, C.; Boos, D.; Gabriel, C.; Wassner, E.; Rullmann, M.; Kao, N.; Laun, M. Elongation-Induced Crystallization of a High Molecular Weight Isotactic Polybutene-1 Melt Compared to ShearInduced Crystallization. J. Rheol. 2007, 51, 195. (25) Zhu, W. L.; Wang, J.; Park, C. B.; Pop-Iliev, R.; Randall, J. Effects of Chain Branching on the Foamability of Polylactide. ANTEC Technical Papers. Soc. Plast. Eng. 2009, 41−45. (26) Azizi, H.; Ghasemi, I. Investigation on the Dynamic Melt Rheological Properties of Polypropylene/Wood Flour Composites. Polym. Compos. 2009, 30, 429. (27) Harper, C. A. Handbook of Plastics, Elastomers and Composites, 4th ed.; McGraw-Hill: New York, 2002. (28) Li, H. J.; Law, S.; Sain, M. Process Rheology and Mechanical Property Correlationship of Wood Flour−Polypropylene Composites. J. Reinf. Plast. Compos. 2004, 23, 1153. (29) Giles, H. F.; Wagner, J. R.; Mount, E. M., Extrusion: The Definitive Processing Guide and Handbook; William Andrew Pub.: Norwich, NY, 2005. (30) Klyosov, A. A. Rheology and a Selection of Incoming Plastics for Composite Materials. In Wood−Plastic Composites; John Wiley & Sons, Inc.: Hoboken, NJ, 2007. (31) Pandey, A.; Lele, A. Exploring the Utility of an Axisymmetric Semi-Hyperbolic Die for Determining the Transient Uniaxial Elongation Viscosity of Polymer Melts. J. Non-Newton. Fluid Mech. 2007, 144, 170. (32) Throne, J. L. Thermoplastic Foam Extrusion; Carl Hanser Verlag: Munich, 2004. (33) Matuana, L. M.; Park, C. B.; Balatinecz, J. J. Characterization of Microcellular Foamed PVC/Cellulosic-Fibre Composites. J. Cell. Plast. 1996, 32, 449. (34) Lee, J. W . S.; Wang, K. Y.; Park, C. B. Challenge to Extrusion of Low-Density Microcellular Polycarbonate Foams Using Supercritical Carbon Dioxide. Ind. Eng. Chem. Res. 2005, 44, 92.

12040

dx.doi.org/10.1021/ie4019462 | Ind. Eng. Chem. Res. 2013, 52, 12032−12040