ANALYSIS OF WOOD PARTICLE DRYING FOR ROTOMOLDING APPLICATION

Nicolas Ward-Perron, Denis Rodrigue

Université Laval, Department of Chemical Engineering, Quebec City, Quebec, Canada, G1V 0A6.

Abstract compared to other processes, we wanted to investigate the

possibility to introduce the wood particles directly in the
This study focused on estimating the moisture content mold via simple dry-blending with the polymer powder
inside wood particles when dry-blended with polyethylene without the need to dry them first. The idea is to determine
powder to produce wood/polymer composites (WPC) for if the initial period in the molding cycle before the material
application in a rotomolding process. The effects of starts melting and sticking to the mold walls in a rotating
different parameters on the dynamics of drying were mold is long enough to remove the humidity inside the
evaluated: initial moisture content of the wood particles, natural fibres, or at least decrease it to a reasonable level.
total thickness of the powder bed, temperature ramp and Since a humidity balance is used in this study (static mode,
wood concentration in the mixture. As expected, higher no particle motion), it is believed that the results obtained
initial humidity took longer to obtain equilibrium moisture will give the maximum humidity inside the particles since
and higher material thickness led to slower drying the rotating motion of the mold should improve
dynamics (time to equilibrium). On the other hand, substantially the heat and mass transfer coefficients
increasing the temperature ramp decreased drying time. through convection.
From the results obtained, it is clear that drying the wood Based on the hypothesis described above, the
particles before blending with the polymer is not necessary objective of this work is to study the drying dynamics of
for rotomolding applications as most of the humidity will wood powder via different temperature ramp, powder bed
have left before the polymer starts melting and sticks to the thickness, initial wood moisture content and wood
mold walls. concentration in the wood/polymer mixture produced via
simple dry-blending.
Introduction
Materials
Wood is a material that can absorb a substantial
amount of moisture from air humidity. Furthermore, the The wood particles used were sawdust provided by
average saturation point can be as high as 30%wt. This residues of the cutting and transformation equipments of
high water content can be a problem when molding the Department of Forestry and Wood Sciences at
polymer composite parts since this humidity will be Université Laval. Upon receipt, the material was sieved to
released and produce gas at the processing temperatures keep only particles between 150 and 500 microns. The
used for most thermoplastic resins (T > 100oC). As a result, polymer powder used was a grade of linear medium
substantial amount of uncontrolled porosity can be density polyethylene (LMDPE): HIVAL 103538 supplied
produced leading to lower mechanical properties, poor by Ashland Inc. (Ashland, Canada). This polymer has a
surfaces quality of the molded parts and other processing density of 938 kg/m3 (ASTM D792A) and a melt flow
problems (degassing, degradation, discoloration, etc.). It is index of 3.5 g/10 min (ASTM D1238). By using standard
therefore important to minimize the amount of moisture differential scanning calorimetry (DSC), the melting point
making its way inside a molded polymer part. For this was found to be 128oC. The particle size distribution was
reason, most of the processes using wood or any other measured by sieving with different openings. The average
material of natural origin need to dry the particles before particle size was found to be 300 microns with 90% of the
introducing them in the molding process through melt particles between 150 and 500 microns.
compounding (twin-screw extrusion, thermo-kinetic mixer,
etc.). Nevertheless, drying equipments can be relatively Experimental
expensive in terms of capital and operation costs.
Recently, natural fibres were used to produce Sample Production
composites parts using rotational molding. Although some Two types of wood samples were tested. The “dry”
studies used a melt pre-compounding step to blend the condition refers to wood particles that were left at room
fibres and the polymer resin [1-2], others used the simple temperature (23oC) and humidity (55% RH) for two weeks
dry-blending of the polymer powder and natural fibres to to attain complete equilibrium. For the “wet” condition, the
produce their composites and eliminated the costly and wood particles were brought to a high level of moisture by
time-consuming step of melt compounding [3-4]. placing them in a closed vessel with a humidity saturated
Nevertheless, since the material can experience relatively atmosphere (23oC, 100% RH) using the setup shown in
high temperatures for long times in rotomolding when Figure 1. A minimum of 15 hours (overnight) was used to
get the equilibrium. Using both conditions, the “dry” case (wt. %)
would represent standard industrial operating conditions, Dry wood fibers 5.5
while the “wet” one would be close to maximum humidity Wet wood fibers 16.9
in the wood as an extreme case (worst case scenario). LMDPE 0.2

Drying
Once the initial moisture content was known, the
evolution of humidity with time was determined. First,
Table 2 shows typical results for the time required for
some samples to reach 1%wt. moisture, which is
considered as a reasonable value to process wood/polymer
composites without having moisture problems. In Table 2,
red cells indicate that the sample never reached 1%
moisture, while yellow cells indicate that the sample
Figure 1: Experimental setup used to produce the “wet” reached 1% moisture only during the cooling period.
wood conditions.
Table 2: Time required to reach 1%wt. moisture in the
For each test, only the required amount was removed samples. The mass, material thickness in the pan, initial
at a time and stored in a sealed bag to blend with the moisture, and wood concentration are also given.
polymer powder (dry-blending of the wood and
12.5°C/min 10°C/min 8.33°C/min
polyethylene by hand). To test the effect of material Sample (8 min) (10 min) (12 min)
thickness, different amounts (weights) were placed in small
aluminum pans to get different powder bed heights (to 3g dry (~4 mm) 00:04:20 00:04:45 00:05:30
simulate different wall thicknesses). Also, different 6g dry (~8 mm) 00:06:40 00:07:25 00:08:25
concentrations of wood in the polymer matrix were tested: 9g dry (~12 mm) - 00:09:59 00:10:55
5, 10, 15, 20, 30, 40, and 50%wt. The polyethylene and 3g wet (~5 mm) 00:07:25 00:08:30 00:08:00
wood powders were also analyzed alone. 6g wet (~9 mm) - - -
9g wet (~13 mm) - - -
Moisture Analysis 6g dry 5% (~5 mm) 00:03:40 00:02:55 00:03:15
The tests were conducted using a MX-50 (A&D 6g dry 10% (~5 mm) 00:03:40 00:03:20 00:03:20
Company, Tokyo, Japan) moisture analyzer. First, the 6g dry 15% (~5 mm) 00:03:55 00:03:55 00:04:05
initial moisture content of the samples was determined 6g 5% wet (~5 mm) 00:03:40 00:04:15 00:06:20
after heating at 100°C until no significant changes in mass 6g 10% wet (~5 mm) 00:03:40 00:05:35 00:04:30
were observed. Thereafter, the drying dynamic curves were 6g 15% wet (~5 mm) 00:06:50 00:05:25 00:05:25
conducted by heating from 50°C to 150°C under different 20g 5% dry (~20 mm) 00:14:35 00:13:05 00:14:35
times to produce linear temperature ramps of 50, 25, 12.5, 20g 10% dry (~20 mm) - - 00:19:50
10, 8.33, 6.67 and 3.33°C/min to be representative of 20g 15% dry (~20 mm) - 00:15:25 00:19:05
typical rotomolding conditions. Finally, after each 20g 5% wet (~20 mm) - - 00:21:05
experiment, heating was stopped and it was possible to 20g 10% wet (~20 mm) - - -
follow the mass as a function of time for the cooling cycle 20g 15% wet (~20 mm) - - -
up to room temperature.
From the results of Table 2, several conclusions can be
Results and Discussion obtained. As expected, material thickness was the main
factor affecting the drying of the samples. For example,
Initial Moisture Content doubling the thickness of wood from 4 mm to 8 mm
The average moisture content of each material taken increased drying time by approximately 60%.
alone is shown in Table 1. As expected, almost no Nevertheless, for almost all the “dry” conditions tested up
humidity is coming from the polymer itself (less than to 5 mm of material thickness, the time to reach 1%
0.2%), while the wet wood (16.9%) had three times the humidity was less than 7.5 min, which is a typical time
humidity of the dry one (5.5%). From these results, it will before the material should start sticking to the mold surface
be assumed in the analysis of the drying kinetics of the in a rotational process. Since rotating the mold will
wood/polyethylene blends that moisture release will be improve heat and mass transfer, it is believed that the time
totally associated to the wood content only. to achieve 1% humidity should be much less than the times
reported in Table 2 because they were obtained under static
Table 1: Average initial moisture content for the different conditions (no particle motion). Figure 2 presents some
materials tested. typical results of the effect of powder bed thickness with
Material Moisture respect to temperature ramp. It seems that the curves are
parallel to each other and temperature ramp higher than heat and mass transfer (porous bed of solid particles). In
10oC/min will always give less than 1% humidity in less general, the drying time for the polyethylene/wood
than 10 min when the bed thickness is less than 12 mm. mixtures were roughly increased by 50% due to transfer
limitation by the polymer particles.
9g (~12 mm)
6g (~8 mm)
15g 50% (~20 mm)
3g (~4 mm) 6
20:00 16g 40% (~20 mm)
Time to reach 1% moisture (mm:ss)

18g 30% (~20 mm)
5
20g 20% (~20 mm)
15:00
20g 15% (~20mm)
4
20g 10% (~20 mm)

Moisture (%)
10:00 20g 5% (~20 mm)
3

05:00
2

00:00
1
0 2 4 6 8 10 12
heating cooling
Temperature ramp (°C/min) 0
04:00 08:00 12:00 16:00 20:00 24:00
Figure 2: Time required to reach 1%wt. moisture as a Time (mm:ss)
function of temperature ramp for 3, 6 and 9 g dry wood.
Figure 4: Evolution of moisture content for polyethylene-
wood mixtures (5 to 50% wood) for a heating ramp of
The effect of the temperature ramp can also be seen in
8.33°C/min (heating and cooling time: 12 min).
Figure 3. As expected, fixing the composition and amount
of material, increasing the temperature ramp increases the
18 9g wet (~13 mm)
drying dynamics. In this particular case, moisture content
6g wet (~9 mm)
below 1% can be achieved in less than 10 min for all the 16
3g wet (~5 mm)
conditions tested. For typical rotomolding processes, the 14
temperature ramp is around 10oC/min which gives less 12
9 dry (~12 mm)

than 1% humidity in less than 5 min. 6 dry (~8 mm)

Moisture (%)

10
3 dry (~4 mm)
8
6 3 g (~4 mm) 50°C/min (2 min)
6
3 g (~4 mm) 25°C/min (4 min)
4
5 3 g (~4 mm) 12.5°C/min (8 min)
2
3 g (~4 mm) 10°C/min  (10 min)
Moisture  (%)

4 0
3 g (~4 mm) 8.33°C/min (12 min)
‐2
heating cooling
3 g (~4 mm) 6.67°C/min (15 min)
3 00:00 04:00 08:00 12:00 16:00 20:00
3 g (~4 mm) 3.33°C/min (30 min)
Time (mm:ss)
2 Figure 5: Evolution of moisture content for dry and wet
wood with varying material thicknesses. Temperature ramp
1 of 8.33°C/min (heating and cooling time: 12 min each).
0
Finally, the effect of initial moisture content can be
00:00:00 00:10:00 00:20:00 00:30:00
Time (mm:ss) seen in Figure 5. As expected, higher initial humidity will
increase substantially the drying time of the samples. Wet
Figure 3: Evolution of moisture for 3 g of wood (about 4 wood, having 3 times the initial moisture of dry samples,
mm) at different temperature ramps. took approximately twice as long to dry when used alone.
Figure 5 also suggests that a combination of material
Another information from Table 2 is the fact that thickness and initial moisture can modify the drying time
increasing the wood content in a polyethylene-wood since the 5 mm wet sample dried as fast as the 8 mm dry
mixture led to increase drying time, but the effect is to a sample and faster than the 12 mm dry sample, even though
lesser degree than for bed thickness as seen in Figure 4. As its initial moisture content was three times greater.
expected, higher wood concentration imposes a higher
amount of humidity in the mixture, which needs more time Moisture absorption
to be removed. In this particular case, some instability can To complete our analysis, moisture re-absorption tests
be seen in the results which can be associated with the were performed from a cooling period at the end of the
homogeneity of the powder mixture which will influence heating cycle. During the cooling period, if the sample
dried completely in the first step during heating, significant  A higher wood content in the polyethylene-
moisture was re-absorbed by the wood particles due to a wood mixture increases the drying time, but
larger concentration gradient at the particle/air interface. the effect is less important than for material
As it was the case for the drying period, moisture thickness and heating speed.
absorption increased with smaller bed thickness (Figure 6).  Wood particles reabsorb moisture quickly, up
For example, the 4 mm dry sample re-absorbed 50% of its to 50% of their equilibrium value in about 30
initial moisture in 30 minutes, with almost 30% the first 10 minutes.
minutes of cooling. This implies that moisture absorption
is a fast process indicating that wood drying is not effective An average heating period for rotomolding is about 20
at long time and the dried material must be used minutes. The results of this study indicate that the majority
immediately and cannot be kept without careful handling. of the samples reached 1% moisture or less in about 10
minutes. We can therefore conclude that during the actual
6 rotomolding process, moisture has been completely
9 g (~12 mm)
6 g (~8 mm)
removed from the composite powder blend before the end
5
3 g (~4 mm) of the heating cycle.
4
Acknowledgements
Moisture (%)

Financial support from the National Science and

2
Engineering Research Council of Canada was appreciated.
1 Technical support from the Research Centre on Advanced
Materials (CERMA) was also highly useful.
0

‐1 heating cooling References

00:00 10:00 20:00 30:00 40:00 50:00 00:00

Time (mm:ss) 1. B. Wang, L. Tabil, and S. Panigrahi, Sci. Eng. Comp.

Mater., 15, 43 (2008).
Figure 6: Evolution of moisture content for dry wood with 2. B.A.C. Siaotong, L.G. Tabil, S.A. Panigrahi, and W.J.,
varying material thicknesses. Initial heating ramp was Crerar, J. Nat. Fibers, 7, 289 (2010).
3.33°C/min (heating and cooling time: 30 min). 3. G.W.G. McDowell, J.F. Orr, J. Kissick, and R.J.
Crawford, SPE-ANTEC Tech. Papers, 47, 1281 (2001).
Conclusions 4. B. Wang, S. Panigrahi, L. Tabil, and W. Crerar, J.
From the results obtained, it is clear that under most of Reinf. Plast. Comp., 26, 447 (2007).
the conditions tested, it is believed that negligible moisture
is present inside wood particles when simply dry-blended
with the polyethylene powder and introduced inside the Key Words: Wood, polyethylene, composite, moisture,
mold of a rotational molding machine. As discussed rotomolding.
previously, the results obtained can be considered as the
maximum moisture content since all the experiments were
performed under static conditions; i.e. no particle motion
inside the humidity balance. As the particles will tumble
inside the bi-axially rotating mold, it is believed that
improved heat and mass transfer will occur to remove even
faster the moisture in order to produce good quality parts
without voids, defects and inhomogeneities.
Based on the results obtained and for the range of
conditions tested (wood contents between 5 and 50%,
temperature ramps between 3 and 50oC/min, powder bed
thicknesses between 4 and 20 mm, and initial wood
moistures between 5 and 17%), the following conclusions
can be made:

 Drying time (time to reach 1% humidity)

increases with sample thickness.
 Increasing the heating speed (temperature
ramp) decreases the drying time.