Recent changes in Western Australian Government policy and bans on
harvesting old-growth native forests have increased the need for
plantation-based timber resources. A significant number of paulownia
(Paulownia fortunei) trees have been planted on farms in several States
for sawlog production, and these logs can supplement other hardwood
timber resources for non-structural products. FPC Timber Technology was
commissioned by Environmental Forest Farms to assess the processing
characteristics and physical properties of paulownia logs from a managed
plantation in Queensland, and was asked to suggest what products could
be produced from the species.
Materials and methods
Log storage and handling
Environmental Forest Farms supplied three logs 3.6 m long from
Queensland for the assessment; the small-end diameters under bark were
290-390 mm. Growth rings indicated that the logs were cut from trees 10
y old. On arrival at FPC Timber Technology, the logs were measured and
stored in a 'wet' shed in an environment maintained at or
above 85% relative humidity. Prior to sawing the logs were docked to a
length of 2.5 m so that the sawn timber would fit into FPC Timber
Technology's heat-and-vent research kiln. The ends of the logs and
their off-cuts were sealed using coloured paints to enable boards cut
from them to be identified, and to prevent end splitting. Log volume was
calculated using the Smalian method, i.e. length multiplied by the mean
cross-sectional area of the large and small ends.
Conversion of logs to sawn timber
The logs and their off-cuts were sawn using a
'Wood-Mizer' hydraulic LT40HD portable horizontal bandsaw mill
fitted with a 0.88 x 31.75 mm blade, sharpened with a 9[degrees] hook
and a 0.6 mm set. A backsawing pattern was initially used to cope with
any stress-related distortion that might occur, but because movement of
the log during sawing was only slight, a mixed pattern was also used to
improve recovery. Random-width pieces, 25 mm, 40 mm and 50 mm thick were
sawn from the 2.5-m logs and sections of 100 mm x 100 mm were sawn from
the log off-cuts to provide a range of section sizes for assessment and
utilisation. Recovered timber was tallied and the fraction of log volume
converted to green sawn timber was calculated. This recovery figure
includes boxed heart and wane.
Dipping for protection against lyctid attack
The green sawn timber was block-stacked and dipped in an 0.2%
solution of boric acid for protection against lyctid attack, and then
stored under cover for three weeks to allow diffusion and penetration of
the chemical prior to drying.
After three weeks the green sawn timber was strip-stacked and a 600
kg concrete weight was placed on top of the stack to minimise any
distortion of the timber during drying. The moisture content (mc) of two
sample boards was measured using the oven-dry method and the timber was
dried in an electric heat-and-vent kiln.
Mass loss and kiln conditions were periodically measured and
recorded during drying. Drying continued for 45 d until the sample
boards reached 12% mc. The moisture content of all boards was then
checked using a Wagner '1601-3' hand-held electromagnetic
moisture meter. The dry sawn timber was tallied, because the client did
not require it to be dressed. The percentage of log volume converted to
dry sawn timber was estimated, and dry dressed recovery was estimated as
though the boards had been dressed to standard industry sizes.
Drying quality class
The 100 mm x 100 mm pieces from the log off-cuts were re-sawn into
backsawn and quartersawn 100 mm x 25 mm sample boards using a Wadkin
'Power Band' re-saw and planed to 87.5 mm x 22 mm (for later
use in test panels) using an SCM 'Superset 23' foursider
planer. The drying quality of these boards was assessed using Drying
Quality Class 'E' as defined in AS/NZS 4787:2001 (Standards
Australia/Standards New Zealand 2001). The Standard requires the
individual assessment of moisture content variation, residual drying
stresses, checking and collapse, and the assignment of a Drying Quality
Class to each assessment. Distortion and checking could also have been
assigned a Drying Quality Class using the standard, but were assessed
later using the FIFWA Industry Standard (FIFWA 1992).
Density and hardness
The air-dry density of a test piece from each sample board was
measured using the method specified in AS 1080.3-1981 (Standards
Association of Australia 1981) for prismatic test pieces. The method
involved weighing and accurately measuring the dimensions of the air-dry
test pieces and calculating density. The hardness of the samples was
then assessed using a modified Janka method.
The Janka method given in ASTM D 143 (American Society for Testing
and Materials 1994) requires incremental measurement of the force
required to embed to half its diameter a steel hemisphere of diameter
0.444 in. into the face, edge and end grain of a 2 in. x 2 in. x 6 in.
test block at a strain rate of 0.25 in. [min.sup.-1]. In this trial, a
vice with a modified screw was used to embed a 0.444 in. steel
hemisphere into samples 22 mm thick. The maximum load required to embed
the hemisphere was determined from the maximum torque applied at the
screw with a torque wrench. The rate of indentation was controlled
manually by the speed of rotation of the wrench handle to achieve
movement at the jaws of about 0.25 in. [min.sup.-1]. The depth of
indentation was measured at the jaws using a Vernier caliper. The test
was repeated five times on each sample and the results were averaged.
The results for air-dry density and air-dry hardness are reported in
comparison to published data for some WA commercial species.
Grade and graded recovery
The sample boards were visually graded to Industry Standard for
Seasoned Sawn and Skip-Dressed WA Hardwoods (FIFWA 1992), which
describes Prime, Standard and Reject grades.
Description of appearance
The colour of the sample boards was described using CSIRO's
colour features (Ilic 1990). The method involved matching the base
colour, and other colour features such as flecks, with standard colours
and descriptions given in the Macro Key. Texture, grain and figure were
then described using CSIRO Trade Circular No. 43 (CSIRO Division of
Forest Products 1960). The circular contains standard descriptions for
grain (referred to as the prevalent direction of the wood fibres along
the length of each board), texture (referred to as the size and position
of the wood elements), and figure (described as the patterns in each
board due to structural features, including knots and growth rings).
Fabrication of test panels
Two replicate panels were constructed from the dressed sample
boards from each log, thus producing six test panels. Each panel
comprised quartersawn and backsawn pieces. The machining processes that
were later carried out were therefore replicated twice on both backsawn
and quartersawn end, face or side matched timber sections from each log.
The sections in each panel were edge-glued using Selleys
'Aquadhere' (a proprietary polyvinyl-acetate (PVA) adhesive)
with an alternating backsawn/quartersawn assembly pattern. The gluing
process involved applying a film of Aquadhere to the edges of the timber
within 24 h of planing, followed by clamping for 48 h. Timber shims were
used to prevent the sash clamps from indenting the edge pieces of
After 48 h the panels were trimmed to 610 mm in length using a
Wadkin 'CP380' panel saw. The 610 mm x 610 mm dimensions of
the panels enabled the use of standard machining templates. The off-cuts
from the panels were used for assessing glue strength in accordance with
AS/NZS 1328.1:1998 (Standards Australia/ Standards New Zealand 1998).
The glue strength test involved preparing notched samples that were
split along the glue line using a bolster and masonry hammer. The
separated surfaces were visually assessed and the extent of wood failure
was determined in 5% increments on a scale ranging from 0% to 100%.
Machining of test panels
The panels were then planed with and against the grain using a
Robinson 36 in. x 7 in. thicknesser adjusted to remove 1 mm of thickness
with each pass. The quality of the planed surfaces was graded using the
FPC Timber Technology visual grading rules for assessing machining
quality (Box 1). The grading rules rank the quality of the planed
surface on a scale ranging from 1 (a high quality surface that is
virtually free from any defect associated with the particular machining
or finishing process) to 5 (a surface that would be rejected as unfit
The panels were then sanded using a Worldmax 'Sheng Feng A
376' belt sander fitted with a 120 grit belt, using the method and
grading system that had been used for planing. The panel arrises were
pencil rounded using a Makita '361 2BR' 6 mm router fitted
with a quarter rounding bit of radius 5 mm, and the quality of the
machined surfaces was assessed using the FPC grade rules. The
double-grid pattern of trenches, 6 mm wide by 5 mm deep, shown in Figure
1 was then cross-cut and ripped into the panels using an Omga 'RN
600' radial arm saw. The quality of the cuts was graded using the
FPC grade rules given in Box 1. Using standard guides and the Makita
router, the circle and cross patterns, 12 mm wide by 5 mm deep shown in
Figure 1 were routed into the panels; results were also graded using the
FPC grade rules.
Holes were bored to 5 mm depth in the panels in the pattern shown
in Figure 1, using an 'F&R Tough' pedestal drill-press
fitted with a 6 mm brad-point bit. This procedure was repeated using a
12 mm brad-point bit and 16 mm, 19 mm and 25 mm saw-tooth bits. The
bored panels were then graded using the FPC grade rules.
The results of the machining trials were compared with those from
assessment of local commercial species.
Finishing of test panels
At the completion of the machining trials two panels were sanded by
hand with 240 grit paper to produce clean, smooth surfaces suitable for
finishing. One panel was stained using Wattyl 'Craftsman fade
resisting interior woodstain--Dark Mahogany' and finished with
Wattyl 'Isoguard--Clear' in accordance with the
manufacturer's instructions. The other panel was finished using
only the Wattyl 'Isoguard--Clear'. The quality of the finished
surfaces was assessed using the FPC grade rules (Box 1). A painting
treatment was not included because this would have concealed the grain
and figure in the timber.
[FIGURE 1 OMITTED]
Results and discussion
The dimensions of the logs are presented in Table 1, and the
appearance of the logs can be seen in Figure 2.
As shown by Table 1 and Figure 2, the logs were regular in form,
though tapered, and comparatively large considering the tree age. The
green logs were fairly heavy owing to the very high moisture content of
the wood, and needed careful handling to avoid mechanical damage because
of the thin bark. Overall green sawn recovery (including heart-in
timber) from the logs as a fraction of log volume is given in Table 2.
Although the recovery figures were high, log A recorded the lowest
recovery because a conservative sawing pattern was initially adopted to
accommodate any stress-related distortion. Growth stresses were minimal,
however, and logs B and C were both back- and quartersawn, and higher
recoveries were achieved.
A smooth sawn finish was obtained using the Wood-Mizer portable
bandsaw mill. The logs were easy to saw when the feed speed was fairly
high, but the band tended to wander when the feed speed was slow. A band
with an increased hook angle and set should handle a faster feed rate
than that used in our study.
The drying schedule used for drying the timber to 12% mc is given
in Table 3. A comparatively conservative schedule was chosen for safety
because FPC Timber Technology did not have experience with the species.
The timber was air dried for one week prior to kiln drying, during which
time its moisture content decreased by about 19%. It was evident that
the initial moisture content of the timber was very high and that
moisture loss occurred very quickly, particularly in the first week of
drying. The entire drying cycle took 45 d, but this could be reduced by
using a less conservative schedule. The volumetric loss in shrinkage was
estimated as 6.0%, based on the dry and green sawn recoveries in Table
Sample board moisture content and the in-kiln conditions maintained
during drying of the timber are given in Table 4. Some minor problems
arose during drying. The jarrah strip sticks used to separate the timber
caused slight staining, but dressing the boards removed the marks; the
use of light-coloured strip sticks is recommended. No difficulties were
experienced with the timber that was dried after sawing, but off-cuts
that were left to stand for eight weeks prior to drying were affected by
sap-staining fungi. Haslett et al. (1992) recommended using an
anti-sapstain dip before air drying.
[FIGURE 2 OMITTED]
The drying qualities in respect of moisture content, residual
drying stress, checking and collapse were all assessed as Class
'E' of AS/NZS 4787:2001 (Standards Australia/Standards New
Zealand 2001), i.e. the timber would be suitable for uses where movement
in service must be minimised.
In Table 5 average values from this assessment of air-dry density
and air-dry hardness are compared with data for some Western Australian
(WA) commercial species. The average air-dry density of the paulownia
timber was 260 kg [m.sup.-3], compared to 260-330 kg [m.sup.-3] quoted
by Rao et al. (1986) for a range of Chinese species, and compared with
820 kg [m.sup.-3] for jarrah and 480 kg [m.sup.-3] for radiata pine. As
density is a good indicator of strength, it can be assumed that the
strength of the paulownia would be significantly lower than that of the
commercial species. In view of this, paulownia timber would generally be
unsuitable for use in structural applications.
Using the vice test, the average Janka hardness of paulownia was
0.5 kN, compared with hardness of 1.3 kN for New Zealand-grown P.
tomentosa (Haslett et al. 1992). Hardnesses of three other WA-grown
species also assessed using the vice are compared with data from Bootle
(1983) in Table 5. These data confirm that paulownia timber has low
resistance to indentation.
All the timber was Prime Grade, with none in Standard or Reject
Grades, according to the FIFWA Industry Standard (FIFWA 1992),
indicating that timber of particularly high quality was produced.
Similar studies at FPC Timber Technology using eucalypt species have
produced significantly less Prime Grade timber.
The grain, texture, colour and figure of the timber (Table 6) were
described using CSIRO's Macro Key for Hardwood Identification (Ilic
1990) and the CSIRO Trade Circular No. 43 (CSIRO Division of Forest
Products 1960). Although the timber might be considered bland in
comparison to that of eucalypts, it is particularly well suited for use
in picture frames or other products where the inclusion of knots and
other natural characteristics might detract from the performance of the
In the glue test using AS/NZS 1328.1:1998 all replicates exhibited
100% wood failure, which means that both the adhesion between the wood
and the glue, and cohesion within the glue, exceeded the strength of the
wood perpendicular to the grain. These data indicate that the timber can
be glued successfully, as noted by Rao et al. (1986).
All replicates of all treatments in the machining study rated well
when assessed using the FPC Timber Technology visual grading rules (Box
1), although the routing was slightly lower in quality than the other
operations. When the test pieces were routed against the grain, there
was some tearing of fibres and minor chipping. This poor finish was
removed by planing and sanding, with a concomitant loss of thickness.
Methods for improving routing performance therefore require
investigation. The use of sharp high-speed tools is crucial when
machining paulownia, as discovered during the drilling tests: poor
results were obtained with the drill set on a slow speed (300 rpm), but
results on the same samples were good at a higher speed (6000 rpm).
The average machining quality of the six test panels for each
operation is compared in Table 7 with the results from previous trials
at FPC Timber Technology on commercial species using the same methods,
machinery and grade rules. The results show that the timber machined
well notwithstanding a systematic problem with routing, and produced
better results than the major commercial species previously assessed.
Haslett et al. (1992) rated machining properties of paulownia as similar
to those of radiata pine.
The assessment of finishing quality indicated that the timber
finished very well with the clear finish (Finish Grade 1), but did not
stain very well, exhibiting unstained residual patches (Finish Grade 3).
Although a poor rating for staining was given by FPC Timber Technology,
good staining properties have been reported in the production of timber
blinds (Gosatti (1), pers. comm.). The quality of the finish was not
compared to that of commercial species tested by FPC Timber Technology
because the other species were not assessed using the same grade rules.
Although painting properties were not assessed because viewing the grain
and figure was considered important, the timber should paint well with
opaque coatings, after filling, because it has a coarse texture and is
largely defect free. Trials would be required to confirm this
Although logs from only three Queensland-grown Paulownia fortunei
trees were assessed, the trial gave a good indication of the potential
of paulownia timber. The logs had definite taper but were otherwise of
good form, and based on the 10-y age estimate the trees had grown
rapidly when compared with logs of other plantation species. The logs
were unexpectedly heavy due to high water content, and care was needed
to avoid damage because the thin bark did not offer much protection. The
logs were not prone to end splitting.
The timber dried quickly, though it was susceptible to sap stain.
Sawn sections were largely free of characteristics such as knots, and
the wood had a moderately coarse uneven texture. High wood quality
combined with low growth stresses resulted in high recovery. The timber
machined well in most treatments, notwithstanding a need for sharp tools
and suitable cutter and feed speeds. More work is required to improve
The timber finished well with a clear coating, but staining was
less successful. The light colour of the timber means that it has
potential for staining, but further assessment of this property is
required. The timber should paint well after its coarse texture is
filled, but this expectation requires confirmation. The timber glued
well with water-based PVA adhesive, but suitability of other glue
formulations should be assessed. Natural durability was not assessed.
Due to its low density the timber was deemed unsuitable for use in
furniture where moderate to high strength and hardness are required.
Inappropriate uses would include chair legs and tabletops. The timber is
also unsuitable for use in building structure and decoration where
strength and hardness are required. Inappropriate building uses include
flooring, skirting, stair treads and bench tops.
There is, however, a wide range of applications to which the
physical and aesthetic qualities the timber are well suited. Rao et al.
(1986), Haslett et al. (1992) and Rural Industry Business Services
(1998) also discussed a range of possible products from paulownia, and
there is general agreement about possible uses. Outdoor use is an option
if effective preservative treatment is possible. Potential uses include,
but are not limited to, wall and ceiling linings, wall claddings and
trim (if treated), eaves linings, decorative interior and exterior
mouldings, picture frames, timber blinds, doors and window joinery. The
timber also seems well suited for use as a substrate for veneered
products including doors and tabletops and as a general carcass material
in situations where it would not be over stressed and weight savings
would result from its use.
Revised manuscript received 11 August 2004
American Society for Testing and Materials (1994) Standard Test
Methods for Small Clear Specimens of Timber. ASTM Standard D 143.
American Society for Testing and Materials, West Conshohocken.
Bootle, K.R. (1983) Wood in Australia--Types, Properties and Uses.
McGraw-Hill Book Company, Sydney.
CSIRO Division of Forest Products (1960) Figure in Timber. Trade
Circular No. 43 (Revised). CSIRO, Melbourne.
FIFWA (1992) Industry Standard for Seasoned, Sawn and Skip-Dressed
WA Hardwoods. Forest Industries Federation WA Inc., Western Australia.
Haslett, A.N., Young, G.D., Simpson, I.D. and Britton, R.A.J.
(1992) Paulownia: the timber of the future? What's New in Forest
Research No. 225. New Zealand Forest Research Institute.
Ilic, J. (1990) The CSIRO Macro Key for Hardwood Identification.
Division of Forestry and Forest Products, CSIRO, Melbourne, pp. 16-27.
Rao, A.N., Hsuan, K., Corlett, R., Dkanarajan, G. and Sastry, C.B.
(eds) (1986) Paulownia in China: Cultivation and Utilisation. Asian
Network for Biological Sciences and International Development Research
Rural Industry Business Services (O'Sullivan, M. and Switala,
J.) (1998) Paulownia--A Commercial Overview. Department of Primary
Industries, Queensland, Q198043. 30 pp.
Standards Association of Australia (1981) Methods of Testing Timber
--Determination of Density. AS 1080.3-1981, 4 pp.
Standards Australia/Standards New Zealand (1998) Glued-Laminated
Structural Timber. AS/NZS 1328.1:1998.
Standards Australia/Standards New Zealand (2001) Timber--Assessment
of Drying Quality. AS/NZS 4787:2001.
(1) Mr Ardino Gosatti, Inglewood Products Group, Malaga, WA 6090
M. Beel [1,2,3], S. Davis , J. Murphy [1,4] and P. Piper [1,4]
 Forest Products Commission, 64 Weir Road, Harvey, WA 6220,
 Email: firstname.lastname@example.org
 Present address: Housing Industry Association, Osborne Park, WA
 Present address: Forest Products Commission, Bunbury, WA 6230,
Box 1. FPC Timber Technology Visual Grade Rules for Assessing
Machined surfaces are visually assessed on a scale from 1 to 5 using
the following criteria:
1 Very good describes a high-quality surface that is virtually free
from any defect associated with the particular machining or
2 Good describes a good-quality surface that is free from most
defects associated with the machining or finishing process.
3 Fair describes a surface that is acceptable, but not good.
4 Poor describes an unacceptable surface that would have to be
reworked or a reduced value negotiated.
5 Very poor describes a surface that would be rejected as unfit
for use. If reworking is possible, it probably would not be
warranted due to the extent of the work required.
Table 1. Log length, small- and large-end diameter under
bark and log volume
Length Sedub Ledub (mm Log volume
Log (mm) (mm) (mm) [m.sup.-1]) ([m.sup.3])
A 2400 390 525 54 0.403
B 2400 320 475 62 0.309
C 2400 290 410 48 0.238
sedub = small-end diameter under bark; ledub = large-end
diameter under bark
Table 2. Green sawn, dry and estimated dry dressed volume
Green sawn Dry
Volume Recovery Volume Recovery
([m.sup.3]) (%) ([m.sup.3]) (%)
0.524 55.0 0.493 51.8
Table 3. Schedule used for drying paulownia boards
Technology at FPC Timber
Moisture temperature Relative Air speed
content (%) ([degrees]C) humidity (%) (m [s.sup.-1])
Green 45 88 1
50 45 84 1
40 50 76 1
30 55 67 1
20 60 50 1
Table 4. Sample board moisture content and in-kiln conditions during
drying of paulownia at FPC Timber Technology
Date Moisture Temp. Rel.
(d/mo/y) Sample content (%) ([degrees]) hum. (%)
19/04/01 S1 155.4 45 88
S2 144.9 45 88
27/04/01 S1 94.1 45 88
S2 82.9 45 88
30/04/01 S1 85.6 45 88
S2 64.0 45 88
02/05/01 S1 58.1 45 84
S2 51.4 45 84
07/05/01 S1 44.4 50 76
S2 45.1 50 76
14/05/01 S1 30.6 50 76
S2 32.5 50 76
23/05/01 S1 23.8 50 76
S2 26.2 50 76
27/05/01 S1 12.8 50 50
S2 12.3 50 50
Table 5. Air-dry density and air-dry hardness of paulownia
and selected commercial species
density (kg Hardness (kN) hardness (kN)
Species [m.sup.-3]) (vice method) (Bootle 1983)
Paulownia 260 0.5 Not available
Jarrah 820 8.4 8.5
Karri 900 9.2 9.0
Radiata pine 480 3.1 3.3
Table 6. Aesthetic properties of paulownia assessed using CSIRO's
Macro Key for Hardwood Identification and Trade Circular No. 43
Texture Moderately coarse
Colour Heartwood straw to yellow-brown
Figure Backsawn boards were lightly figured;
quartersawn boards had no figure
Table 7. Machining properties of paulownia and other species assessed
using the same equipment, procedures and grading methods. The values
are scores on an empiric scale ranging from 1 (very good) to 5 (very
poor) (see Box 1).
Species Ripping Cross-cutting Planing
Paulownia 1 1 1
Jarrah 1 2 2
Marri 1 2 3
Karri 1 3 2
Species Sanding Routing Boring
Paulownia 1 2 1
Jarrah 1 2 1
Marri 1 2 1
Karri 1 2 1