HAPs release from wood drying

Article abstract of DOI:10.1021/es991083q

Hazardous Air Pollutant (HAP) profiles from drying softwood and hardwood flakes (for the manufacture of oriented strand board) are very similar, indicating that they originate through a common mechanism, the breakdown of wood tissue. Hence, the strategies employed to reduce VOC (volatile organic compound) emissions from hardwood can also be extended to decreasing HAPs from softwood. Drying aspen flakes in the field and in the laboratory gives rise to different VOC species, and direct extension of laboratory data to the field may prove difficult. Formaldehyde emissions from drying fresh aspen flakes are lower than those from stored material; the opposite effect occurs for methanol and the other aldehydes. HAPs evolved from drying pine flakes surge sharply at 5-10% moisture content during drying at 130-160 °C. Emissions of methanol, formaldehyde, pentanal, and hexanal all begin simultaneously, with the release of methanol and formaldehyde being the most sensitive to dryer temperature. Hence, the nature of the VOC mix is partly governed by the dryer temperature. Pine and aspen give rise to similar HAPs profiles during either drying or pressing flakes that are already dried.

Full text of DOI:10.1021/es991083q

Environ. Sci. Technol. 2000, 34, 2280-2283
by over 60%. These findings have been independently
confirmed for formaldehyde, whose evolution increases
rapidly with drying temperature (4).
HAPs Release from Wood Drying
L A W R E N C E P . O T W E L L A N D
M I C H A E L E . H I T T M E I E R
The VOC-dryer temperature curve is more gradual for
softwood since R- and â-pinene, the principal VOC con-
stituents, are released through vapor pressure and Henry’s
Law considerations, and not from wood breakdown (2).
Hence, while the dependence of the total VOCs from softwood
on dryer temperature may be quite moderate, that of the
smaller HAP subgroup should be just as sensitive as that
from hardwood. Given that HAPs are under regulation (5),
it is important to understand the factors that govern the
release of these compounds. The objective of this study was
to determine whether the HAPs from drying or pressing
softwood would follow the same mechanism as those from
hardwood. If so, then the strategies employed for reducing
HAPs from hardwood (3) should also be effective for softwood.
Georgia-Pacific Corporation, P.O. Box 105605,
Atlanta, Georgia 30348
U S H A H O O D A , H U I YA N , W E I S U , A N D
S U J I T B A N E R J E E *
Institute of Paper Science and Technology, 500 Tenth Street
NW, Atlanta, Georgia 30318
Hazardous Air Pollutant (HAP) profiles from drying
softwood and hardwood flakes (for the manufacture of
oriented strand board) are very similar, indicating that they
originate through a common mechanism, the breakdown
of wood tissue. Hence, the strategies employed to reduce
VOC (volatile organic compound) emissions from hardwood
can also be extended to decreasing HAPs from softwood.
Drying aspen flakes in the field and in the laboratory
gives rise to different VOC species, and direct extension
of laboratory data to the field may prove difficult. Formaldehyde
emissions from drying fresh aspen flakes are lower than
those from stored material; the opposite effect occurs for
methanol and the other aldehydes. HAPs evolved from
drying pine flakes surge sharply at 5-10% moisture content
during drying at 130-160 °C. Emissions of methanol,
formaldehyde, pentanal, and hexanal all begin simultaneously,
with the release of methanol and formaldehyde being
the most sensitive to dryer temperature. Hence, the nature
of the VOC mix is partly governed by the dryer temperature.
Pine and aspen give rise to similar HAPs profiles during
either drying or pressing flakes that are already dried.
Experimental Section
Pine flakes for the production of oriented strand board were
obtained from the Georgia-Pacific Dudley, NC, mill; aspen
flakes were provided by the Potlatch Grand Rapids, MN,
facility and the Georgia-Pacific, Woodland, ME, mill. The
flakes were dried at 105 °C for work that required dry furnish.
The tube furnace used for drying flakes in the laboratory has
been described earlier (2, 3). Briefly, air (0.25 L/ min) is
directed over 4-7 g of furnish placed in an aluminum foil
boat in a heated ceramic tube, and the VOCs are stripped
from the air stream in either chilled water (for the deter-
mination of methanol and formaldehyde) or in a methanol
trap (for collecting the other organics).
Formaldehyde recoveries were first established by heating
a known amount of formaldehyde (12.5 µg/ mL) in the tube
furnace at 160 °C for 0.5 h, collecting the emissions in 20 mL
of either chilled water or 1% sodium bisulfite solution, and
analyzing the trapped material by the chromotropic acid
method (6). Water was used in situations where the trapped
material was also analyzed by GC for other components; an
interference was observed in the presence of bisulfite.
Formaldehyde recoveries were 95% in water and 103% in
bisulfite. Methanol was determined by GC, and its recovery
through the oven/ trap assembly was 95%. Since formalde-
hyde was inseparable from methanol under our GC condi-
tions, the two were quantified together, and the methanol
was determined by subtracting out the independently
measured formaldehyde, after normalizing for differences
in GC response factors. Pentanal and hexanal were collected
in chilled methanol; their recoveries were 90-94%. All
emissions data are reported on a dry basis.
Flake temperature was measured during drying in a 160
°C oven by attaching a thermocouple to the surface and
measuring the time-temperature curve. The section of the
thermocouple not in contact with the wood was insulated.
Since flake weight loss and temperature could not be
simultaneously measured, equivalent batches of flakes were
dried under similar conditions for varying periods and a time-
MC curve was determined. The temperature-MC curve was
then obtained by combining the two sets of data.
VOCs were collected in the field (Woodland, ME, during
August 26-28, 1998) by inserting a heated sample probe into
the center of the dryer stack. A heated sample pump and line
was used to convey the gases to chilled impingers, each
containing 20 mL of water. Two impingers connected in
parallel were used; one contained organics-free water for
methanol collection, while 1% sodium bisulfite was used in
the other for trapping formaldehyde. Each collection was
Introduction
VOCs released from hardwood and softwood differ both in
the nature and in the quantities of material evolved.
Hardwood VOCs are principally degradation products that
arise from the thermal breakdown of wood tissue, including
lignin, cellulose, hemicellulose, and extractives. They include
methanol, formaldehyde, and other compounds designated
by the U.S. EPA as Hazardous Air Pollutants or HAPs (1).
Softwood releases a much larger quantity of VOCs, most of
which are terpenes (2). However, since wood tissue may also
be subject to thermal breakdown during softwood drying,
the same suite of VOCs evolved from hardwood may also be
released from softwood. In other words, in addition to its
predominantly terpene load, softwood VOCs may also include
HAPs.
We have shown that drying hardwood to final moistures
of 5% or more (as opposed to the much lower values that is
common industry practice) can lead to a dramatic decrease
in VOCs (3). Wood is evaporatively cooled by the departing
moisture as it dries. The cooling effect is lost late in the process
when the water is mostly gone, the wood tissue temperature
rises, wood decomposition begins, and VOCs are evolved.
The VOCs rise very sharply indicating that the decomposition
is associated with a high activation energy (3); increasing the
final moisture content (MC) from 2% to 5% can reduce VOCs
9
2 2 8 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 11, 2000
10.1021/es991083q CCC: $19.00
2000 Am erican Chem ical Society
Published on Web 04/20/2000

TABLE 1: VOC Emissions from Aspen
TABLE 3: Emissions from Aspen (µg/g, Dry Basis) after
Storagea
av (µg/g)a
std dev (µg/g)
days of
m ethanol
pentanal
100
70
9
20
3
storage methanol acetone pentanal hexanal formaldehyde
n
hexanal
30
Dried at 160oC
form aldehyde
11.3
53
10
30
55
126 (7)
100 (10) 20 (3)
100 (20) 15 (2)
25 (5)
42 (9) 120 (10) 11 (2)
4
2
2
unknownsb
5
19 (2)
11 (2)
40 (10) 43 (2)
26 (7)
35 (1)
a
b
n ) 5. Determ ined assum ing a GC response factor of pentanal.
Dried at 130 °C
10
30
59 (3)
10 (3)
19 (3)
4 (0)
2 (0)
5 (1)
27 (4)
15 (2)
1.0 (0.3)
3.3 (0.2)
2
2
TABLE 2: VOC Concentration (ppm) in the Stack Gases of the
Core Dryer at Woodland
a
Collected in water; the values in brackets are average deviations.
total
trapped
method
25A
VOC
double the Method 25A values on account of the terpenes
present.
run
formaldehyde methanol ethanol VOCs
Pentanal and hexanal, which constitute half of the total
emissions in laboratory work (Table 1), were not found in
the field (mdl: 4 ppm). Aldehydes have frequently been
measured in previous laboratory drying or pressing work.
For example, Carlson et al. found hexanal to be the major
VOC constituent during hot-pressing of aspen strand (9),
and Wang and Gardner (10) obtained similar results from
pressing Southern pine. Conversely, ethanol, a major com-
ponent of the field VOCs, was absent in the laboratory sample
but has been often reported in unpublished fieldwork (11).
Despite the differences in temperature profiles used in the
laboratory and in the field, it is surprising that the nature of
the VOCs changes so completely. For terpenes, differences
in drying temperature affect the quantity rather than the
nature of the material released.
The laboratory results are consistent, at least to the extent
that similar compounds have consistently been found across
different furnishes and processes, and it is likely that the
differences between laboratory and field measurements
originate from differences in drying conditions, which tend
to me more aggressive in the field. Large temperature and
humidity gradients exist in the field, and it is very difficult
to reproduce these in the laboratory. For example, aldehydes
are relatively unstable, and it is possible that they could have
degraded in the air stream. Differences may also be antici-
pated across samples taken from various commercial facili-
ties, where dryer temperatures and profiles can vary appre-
ciably.
To determine the effect of furnish storage time on
emissions from aspen, flakes were stored cold, and VOCs
were periodically measured from drying over 1 h at either
130 °C or 160 °C. The results, listed in Table 3, show that
drying stored flakes leads to more formaldehyde, suggesting
that formaldehyde is biologically or chemically generated
during storage. The generation of formaldehyde during lignin
biodegradation is well-known (12). In contrast, the emissions
of most of the other VOCs decrease upon storage, with the
decrease being most pronounced for pentanal and hexanal.
Corresponding Method 25A emissions also show a decrease
in VOCs upon storage.
Em issions from Green Pine. The surface temperature of
pine flakes during 160 °C drying is provided as a function of
MC in Figure 1. As discussed above, the temperature levels
off at about 90 °C during the evaporative cooling period and
then rises during the later falling rate period. Although
terpenes are the principal constituents of emissions from
pine, the compounds released from aspen should also be
found in pine VOCs, if indeed they also derive from the
breakdown of wood tissue.
1
2
3
4
2.93
4.62
2.89
4.20
4.12
5.53
8.61
12.1
8 (3)
5.49
6.41
5.90
9.22
7 (2)
12.5
16.6
17.4
25.5
29.2
23.9
37.3
50.8
av (av dev)
3.7 (0.8)
18 (5) 35 (6)
run in duplicate for 30 min. A second sampling assembly
was used to simultaneously record total hydrocarbon emis-
sions through a Method 25A determination (7). The flow
through each impinger was controlled at 0.4 L/ min, with the
flow meters connected after the impingers to prevent
deposition in the flow meter.
Although this was principally a drying study, some
measurements were taken during hot-pressing to determine
whether the VOC distribution would be similar. The VOCs
released during commercial pressing originate from both
the wood and the applied resin. To determine volatiles
released from wood alone, resin was not added, and flakes
were placed in a steel foil bag (0.002 in. thick 309 high-
chromium stainless steel) with inlet and outlet tube con-
nections. The bag was then placed between the top (200 °C)
and lower platens (90 °C) of an electrohydraulic press (8)
and pressed at 255 psi for 3 min to a thickness of 1/ 8-in.,
which corresponds to approximately 48 lb/ ft3. Air passed
through the bag at 0.5 L/ min was bubbled into a trap
containing either chilled water or methanol. Airflow was
continued for an additional 5 min after pressing to purge
any residual VOC from the bag. The trapped samples were
analyzed as described above.
Results and Discussion
Laboratory and Field Em issions from Green Aspen. Some
of the compounds released from aspen flakes during drying
at 160 °C for 2 h are listed in Table 1. We were principally
interested in methanol and formaldehyde on account of their
environmental importance. We included some of the alde-
hydes because they have been reported earlier in laboratory
work. We did not attempt to determine all the compounds
released from wood since our principal objective was to
develop methods to reduce emissions of only those com-
ponents that fall under MACT regulation.
Because the wood dries in about 20 min at 160 °C, Table
1 values include emissions from both green and dry wood.
This was done to ensure capture of compounds released
from both early and late drying. Field VOC data collected at
four 30-min intervals from the core dryer at the Georgia-
Pacific mill in Woodland, ME, are reported in Table 2.
Comparison of the Method 25A and compound-specific
results is qualitative, at best, for two reasons. First, the Method
25A response for methanol and formaldehyde is quite low,
which leads to a downward bias. Second, the mill includes
about 15% pine in its furnish (3) which would approximately
While the rise in VOCs observed from drying aspen is
sharp, that from pine is much more gradual (2), since the
terpenes, the major component of pine VOCs, are not
degradation products. However, the suite of compounds that
9
VOL. 34, NO. 11, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 8 1

from the same MC value suggests a common source, almost
certainly the thermal breakdown of wood tissue. The finding
that the VOC mix changes with temperature indicates that
the smaller compounds on one hand, and the larger
aldehydes on the other, arise from different pathways along
the degradation of wood tissue. The emission levels are closer
at higher moisture in the evaporative cooling region. At an
MC of about 50%, methanol ranges from 8 to 30 µg/ g, and
formaldehyde ranges from 2 to 4 µg/ g; a similar spread occurs
for pentanal and hexanal.
The Figure 2 result that methanol and the aldehydes are
released late suggests that they could be drastically reduced
if the temperature was reduced, and if drying was terminated
just before the surge, i.e., a slightly wetter furnish was
accepted. This practice would require a resin capable of
handling the higher moisture. Since this is already being
done at the Georgia-Pacific Woodland facility (3) it should
be easily extended to softwood. The release of terpenes, the
major constituent of softwood, will be reduced minimally,
but large reductions in HAPs could be taken at relatively low
cost. However, HAPs reduction will be difficult if the
temperature exceeds 160 °C, since the wood would still be
appreciably wet when HAPs evolution begins, i.e., the surface
will dry too rapidly.
Figure 2 profiles differ from those of the terpenes, where
two signals are observed (13). The first is sharp, arises early,
and originates from material present on the surface of the
wood. The second is broader and starts late in the process
when evaporative cooling tails off and the wood temperature
rises. These signals arise from physical processes and are
governed by vapor pressure and Henry’s Law considerations.
In contrast, the bulk of the VOCs in Figure 2 emerge late and
arise from chemical degradation of wood tissue. Hence, the
terpenes, on one hand, and HAPs (and related water-soluble
materials), on the other, are released through very different
mechanisms.
FIGURE 1. Surface temperature profile of pine flakes during drying
at 160 °C.
arise from the breakdown of pine tissue should be strongly
temperature-dependent, just as in the case of aspen. If so,
then drying softwood to lower wood tissue temperatures
would substantially reduce the amount of methanol and
aldehydes released, although the decrease in total VOC
emissions would be much more modest.
To establish that methanol and aldehydes emerge prin-
cipally during late drying, pine flakes of varying final moisture
were dried at different temperatures. The results shown in
Figure 2 confirm that the bulk of the release occurs when the
wood is almost dry and that overdrying should substantially
increase these emissions. The clear link between formalde-
hyde and the other traces in Figure 2 suggests that they arise
from the same process.
The VOC surge occurs slightly earlier (10% MC) at 160 °C
as compared to about 5% at 130 °C, since the VOCs start
being released at about 130 °C as the flakes warm to the oven
set temperature of 160 °C. The effect is exaggerated at 200
°C, where the VOC rise starts even earlier, since the flakes
start decomposing at temperatures well below 200 °C; i.e.,
the higher dryer temperature puts the surface tissue into the
falling rate zone, although appreciable moisture still remains
in the flakes. The position of the surge for methanol and
formaldehyde tracks those for the higher aldehydes, but much
more methanol and formaldehyde are evolved at the higher
temperature. Clearly, the VOC species produced will depend
on the drying conditions, with higher temperatures favoring
methanol/ formaldehyde emissions. That all of the VOCs surge
Em issions from Dry Pine and Aspen. Emissions collected
from dry pine and aspen flakes heated at 130, 160, and 190
°C, respectively, for 30 min are illustrated in Figure 3 and
demonstrate that the total water-soluble VOCs increase with
increasing dryer temperature. The VOC profiles between the
two species are remarkably similar, confirming that they arise
from the thermal degradation of wood tissue. A similar
correspondence between softwood and hardwood has been
observed in the field in an extensive multifacility study (14).
Since the furnish dries in about 20 min at 160 °C, it is clear
that overdrying will enhance methanol release. Means to
minimize this, e.g., through green-screening (to remove the
faster-drying fines), or drying to slightly higher moisture,
should reduce methanol emissions.
FIGURE 2. Profiles of selected constituents released from drying pine flakes at 130, 160, and 200 °C.
9
2 2 8 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 11, 2000

FIGURE 3. Water-soluble VOCs from dry pine and aspen.
TABLE 4: VOCs (µg/g) Collected in Water during 3 Minutes of Pressinga
acetone
methanol
formaldehyde
pentanal
hexanal
pinene
total
Water Trapb
pine
aspen
3 ( 1 (5)
1.2 ( 0.7 (8)
7 ( 6 (4)
4 ( 2 (8)
1.1 ( 0.5 (4)
0.5 ( 0.4 (7)
1.0 ( 0.3 (3)
1.1 ( 0.8 (8)
4 ( 2 (3)
6 ( 5 (7)
13 ( 3 (5)
15 ( 8 (8)
Methanol Trapc
pine
aspen
3.0
15.1
3.0
4.3
14
24
35
55
34
a
b
c
Increasing press tim e to 5 m in gave sim ilar results. The num ber of determ inations are in brackets. Single determ inations.
Em issions from Pressing Pine and Aspen. With the
exception of pinene, the similarity in dryer emissions between
softwood and hardwood extends to releases from pressing
Acknowledgments
This study was funded by the Electric Power Research
Institute, Southern Company, and Georgia-Pacific Corpora-
tion.
as shown in Table 4, indicating that they too arise from the
thermal degradation of wood. Formaldehyde emissions from
pressing either hardwood (9) or softwood (10) increase
linearly with press time and exponentially with temperature,
indicating that they originate from thermal degradation.
Modifying pressing conditions to minimize the temperature
rise should lead to reduced emissions from the wood furnish.
A potential concern with higher-moisture drying is whether
the emissions “saved” in the dryer are moved to the press
section. This should not be the case since the HAPs are
thermal degradation products, and the wood experiences a
lower overall temperature under higher-moisture conditions.
To confirm this, flakes dried to 3.9 and 7% MC were pressed
in duplicate, and the water-soluble emissions (methanol,
acetone, formaldehyde, pentanal, hexanal) were collected;
they were statistically indistinguishable at 23 and 18 µg/ g,
respectively.
Control Im plications. Although the VOC composition
changes with both species and drying conditions, methanol
and formaldehyde, the two major HAPs emitted from both
hardwood and softwood, are always released in both labora-
tory and fieldwork under conditions where wood tissue
thermally degrades. Since these HAPs constitute much of
the total VOC load from hardwood, drying to higher MC and
pressing higher-moisture flakes should reduce the quantity
of HAPs released. The Georgia-Pacific Woodland, ME, mill
has won regulatory approval to adopt this practice in lieu of
end-of-pipe control devices. We have now shown that a
similar strategy can also reduce HAPs from softwood,
although the total VOC load (mainly terpenes) will remain
relatively unaffected. The economic, energy, and environ-
mental tradeoff is under study.
Literature Cited
(1) Cronn, D. R.; Truitt, S. G.; Campbell, M. J. Atmos. Environ. 1983,
17(2), 201.
(2) Banerjee, S.; Su, W.; Otwell, L. P.; Hittmeier, M. E.; Nichols, K.
M. Environ. Sci. Technol. 1998, 32, 1303.
(3) Su, W.; Yan, H.; Banerjee, S.; Otwell, L. P.; Hittmeier, M. E.
Environ. Sci. Technol. 1999, 33, 1056.
(4) Barry, A. O.; Corneau, D. Holzforschung 1999, 53, 441.
(5) Code of Federal Register, 40 CFR Part 1998, 63, 71376.
(6) NIOSH Manual of Analytical Methods, 4th ed.; 1994.
(7) Method 25A - Determination of Total Gaseous Organic Con-
centration Using a Flame Ionization Analyzer; 40CFR Part 60,
August 1, 1995; pp 870-872.
(8) Banerjee, S.; Mahmood, T.; Phelan, P. M.; Foulke, R. W. Water
Res. 1998, 32, 258.
(9) Carlson, F. E.; Philips, E. K.; Tenhaeff, S. C.; Detelefson, W. D.
Forest Products J. 1995, 45(3), 71.
(10) Wang, W.; Gardner, D. J. Forest Products J. 1999, 49(3), 65.
(11) Georgia-Pacific Corporation, unpublished results.
(12) Kuwahara, M. In Lignin Biodegradation: Microbiology, Chem-
istry, and Potential Applications; Kirk, T. K., Higuchi, T., Chang,
H., Eds.; CRC Press: Boca Raton, FL, 1980; Vol. II.
(13) Banerjee, S.; Hutten, M.; Su. W.; Otwell, L. P.; Newton, L. Environ.
Sci. Technol. 1995, 29, 1135.
(14) Word, D. Presented at the American Forest and Paper Association
Symposium on VOCs from Wood Drying; Raleigh, NC, February,
1999.
Received for review September 20, 1999. Revised manuscript
received January 10, 2000. Accepted March 6, 2000.
ES991083Q
9
VOL. 34, NO. 11, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 2 8 3