Accepted Manuscript

Mechanical and physical stability of polyhydroxyalkanoate (PHA)-based wood plasticcomposites (WPCs) under natural weathering

Clement Matthew Chan, Steven Pratt, Peter Halley, Desmond Richardson, AlanWerker, Bronwyn Laycock, Luigi-Jules Vandi

PII: S0142-9418(18)31597-6

DOI: https://doi.org/10.1016/j.polymertesting.2018.11.028

Reference: POTE 5704

To appear in: Polymer Testing

Received Date: 28 September 2018

Accepted Date: 19 November 2018

Please cite this article as: C.M. Chan, S. Pratt, P. Halley, D. Richardson, A. Werker, B. Laycock,L.-J. Vandi, Mechanical and physical stability of polyhydroxyalkanoate (PHA)-based wood plasticcomposites (WPCs) under natural weathering, Polymer Testing (2018), doi: https://doi.org/10.1016/j.polymertesting.2018.11.028.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Material Properties

Mechanical and physical stability of polyhydroxyalkanoate (PHA)-

based wood plastic composites (WPCs) under natural weathering

Clement Matthew Chana, Steven Pratta, Peter Halleya, Desmond Richardsonb, Alan Werkera,c,

Bronwyn Laycocka, Luigi-Jules Vandia

a School of Chemical Engineering, The University of Queensland, St Lucia, QLD, Australia.

b Norske Skog Paper Mills (Aust) Ltd, Boyer, TAS, Australia

c Promiko AB, Sweden

Keywords:

Wood flour; Polyhydroxyalkanoate (PHA); Biocomposites; Mechanical properties; UV stability;

Weatherability

ABSTRACT

This study investigated the effects of natural weathering on the physical and mechanical

properties of biodegradable composites based on poly(3-hydroxybutyrate-co-3-hydroxyvalerate)

(PHBV) and wood flour (WF). Composite samples with three different wood contents (0, 20 and

50 wt%), as well as polylactic acid (PLA) and polyethylene (PE) composites containing 50 wt%

of WF, were placed on an outdoor inclined rack for 12 months. Neat PHBV and the 20% WF

samples showed minimal mould growth and little loss in mechanical properties over the 12-

month period. By contrast, the tensile strength of the composites (of all polymer types) made

with 50% WF dropped gradually. This decrease was interpreted to be due to the evolution of

surface/bulk defects brought about by the degradation of wood particles from moisture-induced

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

fungal (mould) attack. Overall, the wood content and the accessibility of wood particles inside

the matrix control the stability of the composites under natural weathering.

1. Introduction

In recent years, the applications of bio-derived and biodegradable polymeric materials have

expanded to address concerns over the accumulation of plastic waste in landfill and in the

environment, particularly the ocean. The use of such polymers as the matrix in wood plastic

composites (WPCs) is one of the growing areas of interest because of the expanding WPC

market [1] and the cost competitiveness achieved by taking advantage of cheap, abundant and

biodegradable wood fibres as structural fillers [2].

Among all biodegradable polymers, polyhydroxyalkanoates (PHAs) have been of particular

interest, as these polymers can be synthesised intracellularly by a range of bacteria and archaea

from renewable/waste resources, and are biodegradable under ambient conditions [3]. PHAs are

a family of polyesters which have been shown to have similar physical and mechanical

properties to polypropylene [4]. PHAs also have a lower melt viscosity than traditional

polyolefins, which can be an advantage in the processing of composites [5].

Extensive research is being conducted to understand the physical and mechanical behaviour of

PHA/wood composites. Among all the PHAs, poly(3-hydroxybutyrate) (PHB) and its copolymer

poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) are the most common polymer matrices

used in WPCs. Several successful attempts have also been made to improve the composite

mechanical properties through compatibilisation techniques including maleic anhydride grafting

[6, 7] and wood pretreatment [6-8]. Although such studies have shown the great potential of

PHA-based WPCs, a current lack of understanding of the material stability while in service

remains an obstacle to suitable applications and commercial developments.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

Given the market growth in exterior applications, the durability of WPCs such as moisture

resistance, weatherability and resistance to fungal attack need to be explicitly assessed. Therefore,

numerous studies have been targeted at understanding the weatherability of WPCs through

natural and accelerated weathering [9-13].

It is well known that the hygroscopic nature of wood causes deterioration of mechanical

properties of WPCs through moisture absorption. For example, studies on PP-based WPCs [14,

15] have shown that a higher wood content leads to higher total water absorption and faster

mechanical deterioration under water immersion. This deterioration is due to the incomplete

encapsulation of wood filler by the polymer matrix at such higher wood contents [14]. The

decolouration of PP/wood flour composites (from brown to chalky white) was also observed

during accelerated weathering [9, 12] and attributed to photo bleaching of lignin components in

these WPCs [12]. Loss in mechanical properties [11], a higher degree of surface roughness [9]

and surface cracking [10] on UV exposure were also reported. The type of polymer matrix [10]

and processing methods [13] also play a role in the weatherability of the resulting WPCs.

To date, however, very little research has been conducted on the durability of prototype

PHA/wood composites. Christian et al. [16] and Srubar et al. [17] investigated the moisture-

induced mechanical deterioration of PHBV/oak wood flour composites and found that the

addition of wood flour accelerated the deterioration of mechanical properties due to a higher

degree of wood swelling, which induced macroscopic surface cracking. However, there have

been no studies on the effect of outdoor (non-soil) weathering on the durability of PHA-based

WPCs [18].

Given the growing interest in fully biodegradable PHA-based WPCs, a more thorough

understanding of the physical and mechanical stability of PHBV/wood composites under outdoor

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

weathering is, therefore, required. In this work, composite samples based on PHBV and Pinus

radiata wood flour with three different wood content levels (0, 20 and 50 wt%) were placed on

an outdoor inclined rack at a fully monitored field site in sub-tropical Queensland (27.5o S,

152.9o E). Polylactic acid (PLA) and polyethylene (PE) composites containing 50 wt% WF were

also included as reference materials. The objectives for the material environmental testing was to

evaluate the effect of outdoor (natural, non-soil) exposure on the material properties of the

composites over an extend period of one year, and to understand the factors that control the

stability of the composites.

2. Experimental

2.1. Materials

PHBV with 1 mol% HV content in powder form was supplied by TianAn Biopolymer, China

under the trade name of ENMAT Y1000. PLA (IngeoTM 2003D) and HDPE (EL-Lene H5818J)

pellets were obtained from NatureWorks and SCG Chemicals respectively. Pinus radiata wood

flour (WF) sieved to below 300 µm in particle size was obtained from Micromilling, Australia.

2.2. Composite extrusion

The composite materials were prepared by a two-step process, consisting of dry mixing and melt

compounding. The formulations with associated sample identification are summarised in Table

1. PHBV and WF were first dried separately at 105oC in a vacuum oven at a gauge pressure

of -80 kPa for 24 h. The dried PHBV and WF were then premixed according to the formulations

shown in Table 1 by manual stirring using a Homemaker double-blade kitchen stick blender at

200 rpm for 2 mins (150 g per batch). The pre-mixed formulations were then fed to a co-rotating

twin screw extruder (Eurolab 16 (ThermoScientific), diameter: 16 mm, L/D ratio: 40:1). The

optimised extrusion parameters were determined based on a preliminary study [19]. Flood

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

feeding and a decreasing temperature profile with a maximum barrel temperature of 180oC and a

die temperature of 160oC were implemented. The screw profile consisted of forward conveying

elements only, with no mixing zones. The screw speed was maintained at 100 rpm, which

resulted in a residence time of around 1 min. A slit die with cross sectional dimension of 25 by 2

mm was placed at the die to yield rectangular strips. The rectangular strips were then melt-

pressed at 180oC with 5 tonnes of force for 2 mins to produce consistent flat sheets with a

thickness of 1.5 mm.

The extruder set-up was modified for extrusion of the dried PLA/dried wood flour and PE/dried

wood flour mixes, due to their different melt temperatures and viscosities. Two gravimetric

spring feeders were calibrated to deliver a 50/50 polymer to wood ratio by weight into the screw.

A screw speed of 100 rpm, a maximum barrel temperature of 210oC and a die temperature of

160oC were employed for both PLA and PE composites. A kneading section that consists of five

30o forward, three 60o forward and four 90o forward elements was included to facilitate mixing

between the polymer pellets and wood flour. The rectangular strips were then melt-pressed at

200oC with 5 tonnes of force for 2 mins to produce consistent flat sheets with a thickness of 1.5

mm.

Table 1 Composite formulations with associated samples identifications

Samples PHBV (wt%) PLA (wt%) HDPE (wt%) Wood flour (wt%)

PHA0W 100 -- -- --

PHA20W 80 -- -- 20

PHA50W 50 -- -- 50

PLA50W -- 50 -- 50

PE50W -- -- 50 50

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

2.3. Experimental set-up

The melt-pressed sheets were laser cut to produce 70 x 15 mm rectangular specimens for outdoor

exposure. The specimens were clipped, as shown in Figure 1, onto an unbacked 45o inclined

aluminium rack according to ASTM G7, with one surface facing the sun and the other in the

shade. The rack was placed facing the equator at a fully monitored field trial site in sub-tropical

Queensland (27.5o S, 152.9o E). On-site weather data including temperature range, total rainfall

and solar radiation were recorded. Specimens were collected after 1, 2, 3.5, 6 and 12 months for

physical and mechanical characterisations.

Figure 1 Unbacked 45o inclined aluminium exposure rack used in this study (with samples

attached)

2.4. Colour measurement

The colour analysis of all samples was performed using Adobe Photoshop software. The samples

were scanned using a document scanner with a white background. The colour of the scanned

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

images was determined by using the eyedropper tool with a 31 by 31 average across the sample

according to the CIE 1976 L*, a*, b* colour space. L* (from 0 to +100) represents lightness: an

increase in L* indicates a lighter colour. a* and b* (from -300 to +300) represent colour

components from green to red and from blue to yellow respectively. The mean values were

obtained from the analysis of 5 areas on each of 3 replicate samples.

2.5. Scanning electron microscopy (SEM)

Samples were imaged by JOEL 6460 and JOEL 6610 scanning electron microscopes (the 6460

broke during the study) under secondary electron mode. The top and bottom surfaces were

exposed for imaging. They were coated with ~30 nm of Iridium and vacuum dried at room

temperature for at least 24 hours before imaging. Image acquisition was done on the Iridium-

coated samples at 5 kV accelerating voltage and approximately 10 mm working distance.

2.6. Bio-sequencing

The samples after 6 months outdoor exposure were subjected to bio-sequencing 16S rRNA gene

sequencing. Samples were cut into small pieces of less than 0.5 mm in length. They were then

placed into a PowerBead tube and secured horizontally using a PowerLyzer® 24 Bench Top

Bead-Based Homogenizer and vortexed at maximum speed for 45 seconds. DNA extraction was

performed using the Fast DNA®SPIN Kit for Soil (MO BIO Laboratories, Inc.). Extracted DNA

was sent to the Australian Centre for Ecogenomics (ACE) sequencing facility for Illumina iTag

sequencing and data processing. Returned 16S rRNA gene sequences that met the quality

threshold were clustered using a 95% cut-off and used for community composition analyses.

2.7. Differential scanning calorimetry (DSC)

A differential scanning calorimeter Q2000 (TA Instruments) under a constant nitrogen flow of

50 mL/min was used to determine the thermal properties of the polymer composites. Samples of

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

2 to 3 mg were placed in a sealed aluminium pan and analysed using standard DSC heating and

cooling scans. Each sample was heated from 25oC to 185oC at 10oC/min and kept isothermal for

0.1 min, and then cooled to -70oC at 10oC/min. The melting temperature, Tm, and enthalpy of

fusion, ∆Hm, were determined from the first heating cycle. The crystallisation temperature, Tc,

was determined from the first cooling cycle. Two replicates were assessed to confirm the

reproducibility of the thermal properties.

2.8. Mechanical testing

Prior to any characterisation, the samples were dried in a vacuum oven at 60oC and -80 kPa

gauge for 24 hours to remove any absorbed moisture. The final drying before mechanical testing

provided a common basis and conditions from which to compare the influence of environmental

exposure on material properties. Tensile mechanical tests were performed according to ASTM

D638 on an Intron 5584 with a 1-kN electronic load cell. Sections of extruded rectangular strips

were laser cut into Type V dumbbell-shaped specimens. Tests were performed at a rate of 2%

strain/min until fracture. An extensometer was used to obtain an accurate strain value across the

narrow region. Five replicas were performed for each sample.

3. Results and Discussion

3.1. Physical appearance

Figure 2 shows the visual appearances of both surfaces (FRONT: facing the sun and BACK: in

the shade) of the unaged and weathered composite samples at five different exposure times (1, 2,

3.5, 6 and 12 months). Whitening was clearly observed on the surface exposed to UV sunlight

for all samples including neat PHBV (Figure 2 top row). Such a bleaching effect on all samples

suggests that both PHBV and wood flour contain chromophores that can be modified through

UV radiation.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

The colour changes through UV sunlight exposure were quantified using the CIE Lab colour

scale, with Figure 3 showing the changes in colour properties (∆L*, ∆a*and ∆b*) of the surface

facing the sun for all samples. The changes were calculated by subtracting the colour values of

the weathered samples from those of the unaged samples.

As can be seen in Figure 3a, the ∆L* values increased for all composites after 1 month but then

decreased afterwards, except for the sample containing 20 wt% wood flour (PHA20W), which

showed a gradual increase in lightness throughout the 12-month period. Unlike PP, which

showed no changes in colour under UV exposure [9, 12], neat PHBV showed an increase in the

∆L* value (lighter in colour) after 1 month and then stabilised. The ∆a* values (Figure 3b)

decreased (i.e. the redness was reduced) gradually throughout the first 2 months for all samples

and then stabilised. There was less of a change for neat PHBV than for the composites.

In terms of yellowness, all samples including neat PHBV showed steep decreases in yellowness

over the first 3.5 months of UV exposure, with limited differences between samples (Figure 3c).

The UV-activated reduction reaction of the quinone structures in lignin has been proposed to

result in a decrease of yellowness, and thus a decrease in b* values [20], which could explain the

observed decrease in yellowness in the wood composite samples. A similar trend of decrease in

b* values for neat PHBV suggests that similar types of chromophores could be present in both

the PHBV and wood flour. The photo-bleaching reaction could also be similar; however, a

deeper study has to be conducted to fully explain the yellowish colour in processed PHBV,

which may be related to residual biomass components or to by-products formed during extrusion

[9, 12].

The colour values for all samples remained steady after 2 months of UV sunlight exposure i.e.

the photo-bleaching reactions were complete. The weather data showed a total accumulated solar

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

radiation of 1040 MJ/m2 was applied to all samples over this time. It could be inferred that a

post-manufacture UV-treatment of samples by exposure to 1040 MJ/m2 irradiation could be used

for achieving a whiter colour of PHBV and its composites with wood. However, this does not

take into account any damage caused using such a process.

The physical appearances of the composites were affected by aging (Figure 2 bottom row), with

black regions associated with mould growth being found on the shaded surface of all of the

composites containing 50 wt% of WF within 2 months. All samples had experienced natural wet-

dry cycles through rainfall (accumulated rainfall of 890 mm throughout the period) and sunlight

(day/night) cycles. More moisture was naturally retained on the shaded surface leading to

increased susceptibility to mould growth on this side. On the other hand, very few signs of mould

growth were observed on the PHA0W and PHA20W samples after 12 months of weathering.

Since PHBV is a hydrophobic material, the limited mould growth on PHBV supports a proposal

that the growth of mould requires a moist environment and the supply of substrate and/or niche

microenvironments such as found in the unprotected wood phase [21]. The absence of evident

mould on the PHA20W composites surface implies that, at the lower wood flour content, the

PHBV matrix sufficiently encapsulates the wood particles acting as something of a barrier to the

fungi [14].

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

Figure 2 Visual appearances of all unaged and weathered samples at five different exposure times (1, 2, 3.5 and 6 months) FRONT

(top row): the surface facing the sun; BACK (bottom row): the surface in the shade

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

Figure 3 Changes in CIE Lab colour parameters of the surface facing the sun of all composites

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

3.2. Scanning electron microscopy (SEM)

Figure 4 shows the SEM micrographs of the shaded (mouldy) surfaces of all samples before and

after 12 months of weathering. No observable micro-organisms were evident on the surface of

neat PHBV (PHA0W) after 12 months of natural weathering. When comparing the lower

magnification images between PHA20W and PHA50W, a denser network of fungal hyphae cells

was observed for the PHA composites with higher wood content. A network of long fungal

hyphae cells with an average diameter of around 5 µm, including a colony of ~3 µm-diameter

spores, was observed on the surfaces of all composites containing 50 wt% WF. Such a network

of long, filamentous fungal hyphae and spherical spores is commonly seen in decayed woody

materials [21, 22].

Surface roughness with visible divots was observed before weathering for the composites

containing 50 wt% WF due to incomplete mixing or imperfect wetting of the WF during

processing. It can be seen from the SEM images that the fungal cells attacked the accessible

wood particles created by processing and created deeper divots on the surfaces. Thus, the fungal

cells could access deeper wood flour particles inside the composite matrix and create voids.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

Figure 4 SEM micrographs of the surface in the shade of all samples (in rows) before weathering

(left column) and after 12 months at 100x magnification (middle column) and after 12 months at

higher 1000x magnification (right column), showing the cell structure of the as-colonised fungal

hyphae.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

3.3. Bio-sequencing

The organisms present on the samples after 6 months of natural weathering were identified using

16S rRNA gene sequencing, and the relative abundance profiles (Order level) of all the samples

are presented in Figure 5. As expected, fungi had the highest relative abundance in all

composites. PHA composites, both 20W and 50W, showed a lower relative abundance of fungi

(70%) when compared to PLA50W and PE50W, which had 87% and 81% abundance,

respectively. Instead, PHA composites had a higher relative abundance of Rhizobiales, which are

nitrogen-fixing bacteria that are symbiotic on plant roots. Although the abundance profiles of

PHA20W and PHA50W were similar, it was observed that the absolute amount of rRNA

obtained from PHA20W is lower than for PHA50W.

By contrast, neat PHBV had a much lower relative abundance of fungi (21%) alongside a mix of

different bacteria (including Bacillales and Rhizobiales), where the total bacteria shared a higher

proportion of the RNA than fungi. The lower relative abundance of fungi for all PHBV

containing samples suggests that the PHBV matrix may show potential for fungal resistance and,

in the present work, the neat PHBV was relatively more susceptible to bacterial enzymatic

attack.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

Figure 5 Abundance profiles of all samples after 6 months of natural weathering (order level).

Orders with relative abundance less than 1% were combined into “others”.

3.4. Weight loss and Molecular weight

Based on the visual evidence of fungal network on the surface of the samples, it is hypothesised

that the fungal cells could attack the wood flour and promote wood decay. To test this

hypothesis, the dry weight loss values of the samples after 12 months of weathering were

determined using the following equation:

%������ℎ�� �� =�����������ℎ������12� ��ℎ� − �������������ℎ�

�������������ℎ�× 100%

The % dry weight loss values are given in Table 2. Minimal weight loss was observed from neat

PHBV. The weight loss increased with increasing wood content, which correlates with the

observed denser fungal cell network in samples with higher wood content. When comparing the

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

composites with 50 wt% WF, higher weight loss values were found in PLA50W and PE50W

composites when compared to PHA50W. This indicates that more wood particles were decayed

by fungal attack for the PLA50W and PE50W samples, which is consistent with the observed

higher relative abundance of fungi in the bio-sequencing results.

Table 2 Dry weight loss of all samples after 12 months of natural weathering (data presented as

mean value ± 95% confidence interval)

Sample Dry weight loss after 12 months (%)

PHA0W 0.22 ± 0.05

PHA20W 3.3 ± 0.3

PHA50W 4.9 ± 0.2

PLA50W 6.5 ± 0.2

PE50W 6.6 ± 0.1

To assess whether the natural weathering caused degradation of remaining PHBV chains, the

molecular weight values were characterised. The molecular weight values of PLA50W were also

determined as reference. The number-average molecular weight (������), weight-average molecular

weight (�������) and the polydispersity (PDI) are shown in Figure 6(a), 6(b) and 6(c), respectively.

Both the PHA0W and PHA20W showed slight decreases in both ������ and ��

����� over time. These

results indicate that the combination of UV exposure, rainfall cycle and fungal growth led to a

higher degree of polymer chain scissions (either at the surface or throughout). On the other hand,

while the ������ values for the PHA50W samples also decreased slightly throughout the period, the

������� values remained roughly the same, with ��

���� being more sensitive to changes to the

proportion of low molecular weight chains in the distribution, and hence indicating the possible

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

loss of some of this fraction. PHA50W had a lower ������� initially when compared to PHA0W and

PHA20W, possibly due to the the presence of moisture in the wood, which could facilitate

hydrolytic chain scissioning of the PHA during processing [23]. The ������ and ��

����� values of PLA

in PLA50W were lower than for the PHA50W initially but also showed slight decreases in ������

and ������� over time, meaning that chain scissions also occurred in PLA composites. No obvious

trends were observed from the PDI values due to the complexity and randomness of the chain

scission.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19

Figure 6 The (a) number-average molecular weight (������), (b) weight-average molecular weight (��

�����) and (c) polydispersity (PDI) of

all unaged and weathered samples against exposure times

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

20

3.5. Thermal properties

The melting and crystallisation behaviour of PHBV and its composites were analysed in order to

fully understand the effect of weathering/UV exposure on the polymer matrix. The peak melting

temperature (Tm) values of all samples are presented in Figure 7(a). The peak melting

temperatures of PHBV and its composites remained relatively constant throughout the 12-month

exposure. These trends suggest that the bulk crystal structures of the PHBV matrix were not

significantly influenced by the exposure to natural UV during weathering. It is proposed that any

surface changes induced by cross-linking may affect the mechanical properties but the bulk

properties will remain unchanged. It has already been shown that the bulk properties of PHBV

such as melting point are stable under UV exposure, even with an observed decrease in

molecular weight from 350 to 130 kDa [24], unless the polymer was exposed to high energy

sources such as gamma [25] or electron irradiation [26]. The UV exposure to sunlight may not be

sufficient to influence the lamellar structure of PHBV, and thus the peak melting temperature

[24]. Under stronger energy sources, PHBV could undergo chain scission, accompanied by

cross-linking mainly in the amorphous region, which leads to change in the bulk crystal structure

and lowers the peak melting temperature [26].

The enthalpy of melting (∆Hm) values from the 1st heating scan of all samples are shown in

Figure 7(b). No clear trends were observed over the exposure time. ∆Hm has been shown to be

directly proportional to the ‘equivalent weight’ crystallinity of PHBV, which can be calculated

by dividing the experimentally measured ∆Hm by the ∆Hm of the 100% crystalline crystal [27].

Therefore, the result indicates that the bulk crystallinity of the PHBV was not affected during the

period of natural weathering.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

21

The melt crystallisation temperatures (Figure 7(c)) of all PHBV samples, both neat and

composite, increased after 1 month of weathering and then stabilised. In other words, a

nucleating effect was introduced to the polymer in the melt after 1 month of natural weathering.

Exposure to UV could have a nucleating effect on PHBV, with trace cross-linked particles acting

as seeds to promote polymer crystallisation. Such a nucleation effect of in-situ cross-linked

particles has been reported in a study with isotactic PP [28]. Higher chain mobility through

random chain-scission reactions could also explain the phenomenon.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

22

Figure 7 The (a) peak melting temperature, (b) enthalpy change during melting and (c) melt crystallisation temperature values of all

unaged and weathered PHBV and PHBV/WF composite samples at different exposure times.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

23

3.6. Tensile properties

The tensile strength and modulus, and strain at break values for the composites over time are

presented in Figure 8(a), 8(b) and 8(c), respectively. Although photo-bleaching of the PHBV was

observed, the mechanical properties were maintained over the course of natural environmental

exposure (non-soil sunlight and rainfall) for a 12-month period. This agrees with the literature,

which suggested that the biodegradation, and thus mechanical deterioration, of PHBV requires

the presence of an active microbial community such as found in soil, landfills, and marine

environments [29]. Interestingly, the tensile modulus of neat PHBV increased and its strain at

break decreased during the first 2 months of weathering. Similar mechanical changes were also

reported in an accelerated weathering study on PHBV under long-wave irradiation (340 nm) [24].

The UV-induced cross-linking was usually associated with the above mechanical changes for PE

[30, 31]. However, aliphatic polyester such as PHA and PLA have been shown to undergo both

chain scission and chain recombination rather than cross-linking [32, 33]. The constant tensile

strength suggests that the chain recombination mechanism is more dominant than chain scission

throughout the exposure time period.

On the other hand, the mechanical strength (Figure 8a) and stiffness (Figure 8b) of the

composites containing 50 wt% wood decreased throughout the 12-month weathering period. It

was observed that PLA50W suffered a more substantial drop in the tensile strength throughout

the exposure time when compared with PHA50W and PE50W.

Figure 9 shows a schematic diagram illustrating the underlying factors that contribute to the

worsened mechanical properties. The mechanical deteriorations observed from the composites

containing 50 wt% wood are considered to be due to the introduction of surface/bulk defects

brought about by both the swelling and shrinkage of wood during the wet-and-dry cycles and the

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

24

observed mould growth. Such defects could act as initiation points for crack propagation through

the bulk, leading to earlier onset of mechanical failure. In addition, the mechanical properties of

WPCs are well-known to be governed by the strength of the polymer-wood interface [2]. On

weathering, the as-colonised fungal hyphae cells, which were presumably associated with the

wood particles, could loosen the wood-polymer interfaces through enzymatic erosion or

mechanical action. Thus, the load in this case would not be effectively transferred within the

matrix and this would explain the observed deterioration of mechanical properties with time.

In contrast, the mechanical strength and stiffness of PHA20W were maintained for the first 3.5

months but then decreased slightly, and less substantially than for PHA50W, over the period

from 3.5 to 12 months. These observations support the hypothesis that, at low wood content, the

wood particles are sufficiently sealed by the PHBV matrix and, in this way, shielded from fungal

colonisation [14].

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

25

Figure 8 The (a) tensile strength, (b) tensile modulus and (c) tensile strain at break of all unaged and weathered samples at different

exposure times showing the mean values with error bars representing 95% confidence interval (PE50W samples at the 3.5 and 6

month time point were lost, therefore not shown in the above graphs).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

26

Figure 9 Schematic diagram showing the natural weathering of the composites with 50 wt% wood flour

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

27

4. Conclusions

The effect of natural weathering on neat PHBV and PHBV/WF composites was studied, using

PLA/WF and PE/WF composites as reference materials. Real-time sunlight UV exposure

introduced a bleaching effect to all samples, including neat PHBV, but had no significant effect

on mechanical properties. On the other hand, mould grew slowly on the shaded surface of all of

the composites containing 50 wt% WF, independent of the polymer type. The trapped moisture

and the bulk supply of substrate (the unprotected wood phase) contributed to the mould growth,

which led to weight loss from fungal attack. Neat PHBV showed no sign of mould growth and

negligible dry weight loss. A less dense fungal cell network and lower dry weight loss values

were observed on the PHA20W composite when compared to PHA50W. At lower wood content,

the PHBV matrix acted as a barrier, partially encapsulating the wood particles. Interestingly,

lower relative abundance of fungi was shown in all PHBV containing samples, suggesting that

PHBV does not promote fungal colonization relative to woody materials. PHA depolymerase is

expressed by bacteria, and so the polymer is more susceptible to bacterial enzymatic attack. Neat

PHBV and PHBV/20 wt% WF plaques showed stable mechanical properties over the 12-month

period. However, the tensile strength and modulus of the composites (all polymer types) made

with 50 wt% dropped after 12 months of weathering. Such decreases were considered to be due

to the introduction of surface/bulk defects and an anticipated worsened wood-polymer interface

brought about by fungal attack of wood. Overall, it is concluded that wood content and the

accessibility of wood particles inside the matrix control the stability of the composites under

natural weathering. Future work will be directed to investigate the presumed anti-fungal

properties of PHA and the effect of PHA coating on the stability of the mechanical properties of

PHA WPCs following outdoor aging.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

28

5. Acknowledgement

The authors thank the Australian Research Council for funding through Linkage Grant

LP140100596. CC thanks the University of Queensland for the tuition waiver and the Research

Training Program from the Australian Government for scholarship during this study. The authors

also thank: the Centre of Translational Polymer Research group for lab support; Dr. Paul Evans

and Ms. Pawarisa Luangthongkam for aid with bio-sequencing work; the facilities, and the

scientific and technical assistance of the Australian Microscopy & Microanalysis Research

Facility at the Centre for Microscopy and Microanalysis, the University of Queensland for their

expertise in SEM imaging.

6. Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the

data also forms part of an ongoing study.

7. Reference

[1] Marketsandmarkets, Wood Plastic Composite Market by Type (Polyethylene,

Polyvinylchloride, Polypropylene, Others), Application (Building & Construction Products,

Automotive Components, Industrial & Consumer Goods, Others), Region (North America,

Europe, Asia-Pacific, RoW) - Global Trends & Forecast to 2021, 2016.

[2] K.G. Satyanarayana, G.G.C. Arizaga, F. Wypych, Biodegradable composites based on

lignocellulosic fibers—An overview, Prog. Polym. Sci., 34 (2009) 982-1021.

[3] K. Sudesh, G. Abe, Y. Doi, Synthesis, structure and properties of polyhydroxyalkanoates:

biological polyesters, Prog. Polym. Sci., 25 (2000) 1503-1555.

[4] G. Chen, Polyhydroxyalkanoates, in: R. Smith (Ed.) Biodegradable polymers for industrial

applications, CRC Press, Cambridge, 2005, pp. 32-56.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

29

[5] P. Gatenholm, J. Kubát, A. Mathiasson, Biodegradable natural composites I. Processing and

properties, J. Appl. Polym. Sci., 45 (1992) 1667-1677.

[6] W.V. Srubar, S. Pilla, Z.C. Wright, C.A. Ryan, J.P. Greene, C.W. Frank, S.L. Billington,

Mechanisms and impact of fiber–matrix compatibilization techniques on the material

characterization of PHBV/oak wood flour engineered biobased composites, Composites Sci.

Technol., 72 (2012) 708-715.

[7] D.N. Thompson, R.W. Emerick, A.B. England, J.P. Flanders, F.J. Loge, K.A. Wiedeman,

M.P. Wolcott, Final report: Development of renewable microbial polyesters for cost effective

and energy-efficient wood-plastic composites, Idaho National Laboratory, Idaho Falls, Idaho,

2010.

[8] A. Gregorova, R. Wimmer, M. Hrabalova, M. Koller, T. Ters, N. Mundigler, Effect of

surface modification of beech wood flour on mechanical and thermal properties of poly (3-

hydroxybutyrate)/wood flour composites, Holzforschung, 63 (2009) 565-570.

[9] C. Badji, L. Soccalingame, H. Garay, A. Bergeret, J.-C. Bénézet, Influence of weathering on

visual and surface aspect of wood plastic composites: Correlation approach with mechanical

properties and microstructure, Polym. Degradation Stab., 137 (2017) 162-172.

[10] J.S. Fabiyi, A.G. McDonald, M.P. Wolcott, P.R. Griffiths, Wood plastic composites

weathering: Visual appearance and chemical changes, Polym. Degradation Stab., 93 (2008)

1405-1414.

[11] F.P. La Mantia, M. Morreale, Accelerated weathering of polypropylene/wood flour

composites, Polym. Degradation Stab., 93 (2008) 1252-1258.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

30

[12] Y. Peng, R. Liu, J. Cao, Y. Chen, Effects of UV weathering on surface properties of

polypropylene composites reinforced with wood flour, lignin, and cellulose, Appl. Surf. Sci., 317

(2014) 385-392.

[13] N.M. Stark, L.M. Matuana, Characterization of weathered wood–plastic composite surfaces

using FTIR spectroscopy, contact angle, and XPS, Polym. Degradation Stab., 92 (2007) 1883-

1890.

[14] N. Stark, Influence of Moisture Absorption on Mechanical Properties of Wood Flour-

Polypropylene Composites, J. Thermoplast. Compos. Mater., 14 (2001) 421-432.

[15] M. Tajvidi, S.K. Najafi, N. Moteei, Long-term water uptake behavior of natural

fiber/polypropylene composites, J. Appl. Polym. Sci., 99 (2006) 2199-2203.

[16] S.J. Christian, S.L. Billington, Moisture diffusion and its impact on uniaxial tensile response

of biobased composites, Composites Part B: Engineering, 43 (2012) 2303-2312.

[17] W.V. Srubar, S.L. Billington, A micromechanical model for moisture-induced deterioration

in fully biorenewable wood–plastic composites, Composites Part A, 50 (2013) 81-92.

[18] O. Faruk, A.K. Bledzki, H.-P. Fink, M. Sain, Biocomposites reinforced with natural fibers:

2000–2010, Prog. Polym. Sci., 37 (2012) 1552-1596.

[19] C. Chan, L.-J. Vandi, S. Pratt, P. Halley, D. Richardson, A. Werker, B. Laycock, Processing

and characterisation of polyhydroxyalkanoate (PHA)-based wood plastic composites: effect of

non-reactive additives, Appita, 69 (2016) 352-360.

[20] M. Muasher, M. Sain, The efficacy of photostabilizers on the color change of wood filled

plastic composites, Polym. Degradation Stab., 91 (2006) 1156-1165.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

31

[21] K. Xu, J. Feng, T. Zhong, Z. Zheng, T. Chen, Effects of volatile chemical components of

wood species on mould growth susceptibility and termite attack resistance of wood plastic

composites, Int. Biodeterior. Biodegrad., 100 (2015) 106-115.

[22] E.C. Hammer, Z. Balogh-Brunstad, I. Jakobsen, P.A. Olsson, S.L.S. Stipp, M.C. Rillig, A

mycorrhizal fungus grows on biochar and captures phosphorus from its surfaces, Soil Biol.

Biochem., 77 (2014) 252-260.

[23] K.J. Kim, Y. Doi, H. Abe, Effect of metal compounds on thermal degradation behavior of

aliphatic poly(hydroxyalkanoic acid)s, Polym. Degradation Stab., 93 (2008) 776-785.

[24] L. Wei, A.G. McDonald, Accelerated weathering studies on the bioplastic, poly(3-

hydroxybutyrate- co -3-hydroxyvalerate), Polym. Degradation Stab., 126 (2016) 93-100.

[25] H. Mitomo, Y. Watanabe, I. Ishigaki, T. Saito, Radiation-induced degradation of poly(3-

hydroxybutyrate) and the copolymer poly(3-hydroxybutyrate-co-3-hydroxyvalerate), Polym.

Degradation Stab., 45 (1994) 11-17.

[26] A. Bergmann, J. Teßmar, A. Owen, Influence of electron irradiation on the crystallisation,

molecular weight and mechanical properties of poly-(R)-3-hydroxybutyrate, J. Mater. Sci., 42

(2007) 3732-3738.

[27] H.-L. Chen, J.C. Hwang, Some comments on the degree of crystallinity defined by the

enthalpy of melting, Polymer, 36 (1995) 4355-4357.

[28] Y. Zhang, P. Tiwary, J.S. Parent, M. Kontopoulou, C.B. Park, Crystallization and foaming

of coagent-modified polypropylene: Nucleation effects of cross-linked nanoparticles, Polymer,

54 (2013) 4814-4819.

[29] K.C. Batista, D.A.K. Silva, L.A.F. Coelho, S.H. Pezzin, A.P.T. Pezzin, Soil biodegradation

of PHBV/peach palm particles biocomposites, J. Polym. Environ., 18 (2010) 346-354.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

32

[30] I.A. Hussein, Rheological investigation of the influence of molecular structure on natural

and accelerated UV degradation of linear low density polyethylene, Polym. Degradation Stab.,

92 (2007) 2026-2032.

[31] M. Raab, L. Kotulák, J. Kolařík, J. Pospíšil, The effect of ultraviolet light on the mechanical

properties of polyethylene and polypropylene films, J. Appl. Polym. Sci., 27 (1982) 2457-2466.

[32] R.K. Sadi, G.J.M. Fechine, N.R. Demarquette, Photodegradation of poly(3-

hydroxybutyrate), Polym. Degradation Stab., 95 (2010) 2318-2327.

[33] P. Stloukal, V. Verney, S. Commereuc, J. Rychly, L. Matisova-Rychlá, V. Pis, M. Koutny,

Assessment of the interrelation between photooxidation and biodegradation of selected

polyesters after artificial weathering, Chemosphere, 88 (2012) 1214-1219.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Highlights

• UV exposure resulted in surface whitening of PHA and their composites.

• Neat PHA was physically and mechanically stable upon natural weathering.

• PHA composites behaved similarly to PLA and PE composites upon natural

weathering.

• Wood is the species responsible for mechanical deterioration due to mould growth.

• Partial sealing of wood flour by hydrophobic PHA provided protection from mould.