HAP and VOC emissions from white fir lumber dried at high and conventional temperatures.

Abstract

Volatile organic compounds and hazardous air pollutant emissions occur during wood processing and must be understood so manufacturers can comply with the Clean Air Act Amendments (CAAA). Both types of emissions occur during lumber drying. White fir (Abies spp.) lumber was dried

at 82.2[degrees]C and 11 5.6[degrees]C, and EPA Method 25A was used to measure total hydrocarbon emissions, while the NCASI chilled impinger method was used for methanol and formaldehyde. Total hydrocarbon emissions more than doubled, from 0.052 to 0.141 kg/[m.sup.3], for wood dried at the higher temperature. Methanol emissions increased 240 percent, to 0.097 kg/[m.sup.3], and formaldehyde increased 470 percent, to 0.003 7 kg/[m.sup.3], A large mill might be considered a major source of pollution under Title III of the CAAA if drying white fir at a high temperature.

**********

As the Clean Air Act Amendments (CAAA) of 1990 continue to be phased in, industry continually needs new information. This information is mostly used for permits and often to simply show that emissions are small or under certain limits. Volatile organic compound (VOC) emissions have already received considerable attention from researchers, and provisions from Title III (Hazardous Air Pollutants or HAPs) of the CAAA are currently drawing attention to HAP emissions. The EPA is required to define the "maximum achievable control technology" (MACT) for HAPs for each industry The final rule for wood products is expected in the spring of 2003, after which the major sources will have 3 years to achieve compliance. Lumber kilns are not included under the rule; however, when co-located with other equipment, the kilns contribute to site emissions. Compliance will mean that fugitive emissions will have to be mini- mized and RTO-equivalent controls (RTOs are devices that incinerate emissions) placed on softwood veneer dryers, tube dryers, rotary dryers for green wood, and most presses at facilities where site emissions exceed 9,100 kg/yr. (10 tons/yr.) of any one HAP or 22,700 kg/yr. (25 tons/yr.) of combined HAPs.

As wood is heated, volatile materials present in the wood are driven off with the water. There are dozens of compounds in the resulting emission, including terpenes and resin acids. Some of these break down to form lower molecular weight compounds, mostly a range of one-to-five-carbon aldehydes and acids. Some breakdown of the other wood material may also contribute to the total emissions, especially at higher temperatures. Essentially all of the organic compounds in the press and dryer exhausts fit the federal definition of VOCs and a few are considered HAPs. The main HAPs emitted from wood dryers and presses are methanol, formaldehyde, and acetaldehyde. A significant amount of work has been done to learn about emissions from drying. An overview of the regulatory and research status for air emissions from the drying of wood can be found in a previous publication (Milota 2000a).

Much of the work done to date has been on VOCs for industry compliance with Title I (Ambient Air Quality) and Title V (Permitting) of the CAAA. Ingram et al. (1995) measured VOC emissions of 5.84 and 4.66 pounds per thousand board feet (lb./MBF) for southern pine lumber dried at 116[degrees]C (240[degrees]F) and 82[degrees]C (1 180[degrees]F), respectively. Thompson (1996), however, found no difference due to drying temperature. She did find that knotty wood had almost 100 percent higher emissions than clear wood. Wu and Milota (1999) demonstrated that VOC emissions increased about 12 percent for Douglas-fir, from 1.30 to 1.56 lb./MBF, when maximum drying temperature was increased from 71[degrees]C (160[degrees]F) to 93[degrees]C (200[degrees]F). Ingram et al. (1995) and Lavery (1998) identified terpene compounds in the exhaust; however, these only accounted for a portion of the total hydrocarbon emitted.

VOC emissions have also been quantified by the National Council for Air and Stream Improvement (NCASI 1996), including approximately 0.55 lb./MBF for true firs dried at 88[degrees]C (190[degrees]F). Rice and Zibilske (1999) used extraction to estimate VOC emissions of several northeastern species at less than 1.4 lb./MBF. Lavery and Milota (2000, 2001) developed and verified a small-scale drying method to estimate emissions from commercial kilns and tested this method on Douglas-fir and ponderosa pine. NCASI (with Mississippi State and Oregon State) used a similar procedure in the Georgia Pacific SEP study (Word 2001) and further verified the method. Shmulsky (2000) determined that season of harvest affected VOC emissions, but did not find an effect of material size. Methanol emissions of approximately 0.15 lb./MBF were reported for southern pine dried in steam-heated kilns at 113[degrees]C (230[degrees]F) by Word (2001). Formaldehyde emissions were approximately 10 percent of this.

The VOC concentration in stack gas is typically measured with a total hydrocarbon analyzer. This device uses flame ionization to measure the concentration of carbon atoms. The mass emitted is often reported "as carbon," meaning that if there were no effect of functional groups (such as aldehydes and alcohols) on the analyzer response, it would be the mass of the carbon portion of the VOC emissions. The moisture level in the sample gas also affects the analyzer reading (Milota and Lavery 1999). In contrast, the actual emitted mass of HAPs is reported.

The objective of this work was to determine the extent to which drying temperature affects emissions. White fir (Abies spp.) is a common name for a mixture of several species of true fir grown in the West. It is dried commercially at both conventional and high temperatures. For this study, white fir was kiln-dried at conventional (82.2[degrees]C; 180.0[degrees]F) and high (115.6[degrees]C; 240.0[degrees]F) temperatures, and measurements of HAPs (methanol and formaldehyde) and VOC were made.

Procedure

The logs for this study were harvested on June 19, 2000, in northeastern Oregon. Considering the region of harvest, the sample may have included white fir (Abies concolor), grand fir (A. grandis), noble fir (A. procera), and subalpine fir (A. lasiocarpa). The true firs cannot be distinguished by the wood anatomy. On June 21, the logs were sawn and side flitches were selected at random; 0.6 m (2 ft.) was removed from each end of each fitch, and the remainder was crosscut into 1.2-in-long (4-ft.) pieces. These were then wrapped as a unit in plastic and immediately shipped to the lab. On June 22, the pieces were sorted into two piles, depending on the number of knots in the piece. Packages containing four pieces, two from each pile, were then wrapped in plastic and stored in the freezer. The lumber was in the freezer on the third day after the trees were harvested. Each package contained wood with a similar quantity of knots; five packages were randomly selected for each kiln charge.

Prior to drying, the wood was thawed, 5 cm (2 in.) was trimmed from each end of each piece, and the remaining piece was weighed. After kiln-drying, each piece was ovendried, and its dimensions were measured, as were ring count and the knot area on one face. Initial and final moisture contents (MCs) were determined by ASTM D 4442 Method B (ASTM 1999).

A laboratory kiln was used to dry the wood (Fig. 1). The kiln was a cube of approximately 1.2 in (4 ft.) and was indirectly heated by steam. There was no steam spray for humidification. A minimum of 10 L/min. of clean, dry air was metered into the kiln to control the wet-bulb temperature and maintain pressure in the kiln at slightly greater than ambient pressure. The vent was a small pipe on the low-pressure side of the kiln. The airflow was 5.1 m/sec. through the 1.95-cm (0.75-in.) sticker slots. Each charge contained 0.145 [m.sup.3] (73.3 nominal BF) of 2-by-6 lumber.

The dry-bulb temperature was linearly ramped to 82.2[degrees]C (180[degrees]F) over a 3-hour interval for the conventional-temperature drying schedule and to 115.6[degrees]C (240[degrees]F) over a 6-hour interval for the high-temperature schedule. The dry-bulb temperatures were held at these values for the remainder of the schedule. The wet-bulb temperatures were similarly ramped to 54.4[degrees]C (130[degrees]F) and 82.2[degrees]C (180[degrees]F), respectively, and held at these values for the remainder of the schedule. The actual temperatures for two charges are shown in Figure 2. The wood was dried to an average MC of 15 percent.

The total hydrocarbon measurements were made with a JUM 3-200 total hydrocarbon analyzer, which drew a 2.8 L/min. sample. The sample to the analyzer was diluted to 15 to 20 percent moisture using heated, filtered air. All components of the sampling train were heated to prevent condensation of water or other compounds. The system was leak checked prior to each kiln charge by drawing a vacuum. The analyzer was calibrated at 3-hour intervals by using EPA protocol 1 span gas (propane in air), a certified mid-gas, and Grade 5 air as a zero gas. At each interval, the flow to the analyzer and dilution flow were measured using an NIST-traceable flowmeter. Calibration drift was less than 5 percent during any 3-hour interval. The quantity of hydrocarbon was calculated by converting the VOC concentration to a dry-gas basis and integrating this multiplied by the dry-gas flow rate from initial green MC to 15 percent MC. The method used was essentially EPA Method 25A (GSA 1996); the calculation procedure is fully describe d in Lavery and Milota (2000).

Methanol and formaldehyde were sampled by using the NCASI chilled impinger method. A 400 mL/min. sample of kiln exhaust was drawn through two midget impingers in a glycol bath at -1[degrees]C (30[degrees]F). The flow was held constant by a critical orifice in the line before a vacuum pump. The collection interval was either 3.0 or 4.5 hours, depending on the kiln humidity. To calculate the total methanol or formaldehyde released, the psychrometric relations and the ideal gas law were used to obtain the dry gas flows through the dryer and HAPs sample train. The quantity of HAPs collected in the sample train was then scaled up to the entire dryer based on a ratio of the two gas flows. This was done for each collection interval. The last interval was proportioned to 15 percent MC. The methanol concentration in the impinger catch was determined by gas chromatography. The formaldehyde concentration was determined by spectrophotometry after reaction with acetylacetone reagent.

Results

The wood properties for each charge are shown in Table 1. There were no statistical differences in wood properties among the charges. Heartwood percentage was not measured, but because the boards were sampled as side flitches off of a quad-band, they were from the outer portion or the tree, suggesting mostly sapwood. The heartwood in white fir is not readily distinguished from the sapwood.

The VOC, methanol, and formaldehyde emissions from initial MC to 15 percent MC are shown in Table 2. Between the conventional- and high-temperature schedules, average VOC emissions more than doubled. These are the sum of all hydrocarbon emissions as measured by the hydrocarbon analyzer and reported as carbon. Methanol emissions increased by over 240 percent and formaldehyde emissions by over 470 percent. HAP emissions were reported as the actual weights emitted. The increase in VOC emissions was greater than would be expected based on past work on southern pine, in which VOC emissions were either 25 percent greater (Ingram et al. 1996) or the same (Thompson 1996) at high temperatures as at conventional temperatures. Also, based on the 12 percent increase in VOC emissions observed in Douglas-fir from 71[degrees]C (160[degrees]F) to 93[degrees]C (200[degrees]F) (Wu and Milota 1999), one would not expect VOC emissions to double from 82[degrees]C (180[degrees]F) to 116[degrees]C (240[degrees]F). The values for V OCs were similar to those reported by NCASI, 0.55 lb./MBF for drying at 88[degrees]C (190[degrees]F).

The pattern of VOC release with respect to time can be seen in Figure 3. At each temperature, the rate of VOC emissions (as indicated by slope of curve) increased while the kiln was healing, peaked, then decreased for the remainder of the schedule. If Figure 3 were plotted with respect to wood MC instead of time, each curve would be slightly convex upward, meaning that more VOCs were released during a given change in MC when the wood was at low MC than when it was at high MC. The concentration of VOCs in the exhaust gas (not shown) increased from 15 ppm early to 121 ppm late in the schedule for high-temperature drying. For drying at conventional temperatures, the concentration decreased to a minimum of 8 ppm 12 hours after a small peak at 3 hours, then increased to 18 ppm for the remainder of the schedule. The concentrations were lower in conventional drying because the volume of air vented was much greater and the drying time longer.

The HAP emissions were much greater at the higher temperature; however, there is no prior data on HAP emissions from lumber as a function of temperature with which this new data can be compared. The quantity of methanol emitted at 116[degrees]C (240[degrees]F) was considerably higher than the 0.15 lb./MBF reported for southern pine (Word 2001) and twice that obtained for lodgepole pine (Milota 2000b, 2001) at similar temperatures. The formaldehyde emissions were similar to those reported for southern pine (Word 2001) and 1.5 times greater than those reported for lodgepole pine (Milota 2000b, 2001). The quantities of methanol (Fig. 3) and formaldehyde (not shown) released were nearly linear with respect to time. Because the drying rate slowed toward the end of drying, however, this means that more HAPs were released for a given change in MC at low MC. For example, 10 and 13 percent of the methanol was released between 20 and 15 percent MC for the high-and conventional-temperature schedules. This moisture chan ge represents only 4 to 5 percent of the total MC change. Mills should be careful about overdrying if the quantity of HAPs released is a concern.

The methanol reported represented the actual weight of methanol that left the kiln per kilogram of lumber dried. From Figure 3b it would appear that methanol comprised most of the VOC emissions. This was not true, however, because of four additional factors in measuring and reporting VOCs. To provide a true estimate of the mass of VOC, all four would have to be addressed as follows:

1. Due to moisture in the gas, the analyzer reading would have to be increased by about 12 percent (Milota and Lavery 1999).

2. VOCs are reported as carbon. Assuming that most of the hydrocarbon was not oxidized and because there are about two hydrogen atoms per carbon atom, this would necessitate increasing the VOC value by another 16 percent to get the mass of the whole molecule. Combined, these two factors would increase the VOC value to 0.18 kg/[m.sup.3] for the wood dried at high temperature. Of this, the methanol would then be 53 percent of the total.

3. There would be a reduction of approximately 35 percent in the analyzer response to methanol because of oxidation of the carbon atom due to substitution of a hydrogen for a hydroxyl group. Thus, the 53 percent methanol portion of the VOCs would have to be increased by 54 percent [100 x 0.35/(1 -0.35)] to make the total VOCs 0.24 kg/[m.sup.3]. Of this, the methanol would then be 40 percent of the total.

4. The molecular weight of methanol is 32, yet after the adjustment in factor 2, it is only counted as 14. Thus the 42 percent methanol fraction of the VOCs would have to be increased by another 129 percent to account for this, making the total VOCs 0.37 kg/[m.sup.3]. Of this, the methanol (0.097 kg/[m.sup.3]) would then be 26 percent of the total. Other compounds with alcohol, aldehyde, or acid groups would further reduce this percentage, also by increasing the total VOCs because of factors 3 and 4.

This discussion illustrates one of the weaknesses of Method 25A (NCASI 1996), particularly as the methanol becomes a larger fraction of the total VOC emissions.

Mills that emit more than 91,000 kg/yr. (100 tons/yr.) of VOC from their contiguous properties are considered major sources and may be required to file a Title V operating permit. No white fir mills are large enough to attain this level just from dry kilns alone, operated at conventional temperatures; however, at the higher temperature, a mill producing 647,200 [m.sup.3] (328 million board feet [MMBF]) of lumber per year could reach that level of VOC emissions. The actual production level of wood necessary to make a mill become a major source would be lower because all site VOC emissions are additive when determining when the 91,000 kg/yr. limit has been reached.

Mills that emit more than 9,100 kg/yr. (10 tons/yr.) of any one HAP or 22,700 kg/yr. (25 tons/yr.) of combined HAPs are considered major sources under Title III, and could be required to control emissions. The combined HAPs rule will not apply because 9,100 kg/yr. of methanol would be emitted before the combined limit would be reached. An annual production rate of only 94,000 [m.sup.3] (48 MMBF) of lumber at the higher drying temperature would put a mill at that limit. Again, the actual rate would be lower because all other sources of methanol would have to be considered. This would be particularly problematic if there were veneer or rotary dryers on the contiguous company property. At the lower temperature, the mill could dry 324,000 [m.sup.3] (164 MMBF) of white fir before reaching the major source category for HAPs.

Conclusions

The temperature at which lumber is dried may have a large effect on the HAP emissions. Further research is needed to determine whether this is unique to white fir or whether it is generally true across species. If it is generally true, mills could greatly reduce VOC and especially HAP emissions by not drying lumber at high temperatures. The actual VOC emissions may differ significantly from those measured by Method 25A (GSA 1996) if large amounts of methanol are present.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

Table 1

Properties of the wood in each charge.

Charge                Knots         Ring count  Initial MC  Ovendry mass
                ([in..sup.2]/face)  (no./in.)      (%)          (kg)

Conventional 1         1.7              12        122.0        53.37
Conventional 2         1.3              13        133.2        52.53
High 1                 2.2              15        126.3        52.95
High 2                 1.2              12        119.0        53.40

Table 2

Emissions from each charge. (a)

Charge                 VOC as carbon                   Methanol
                (kg/[m.sup.3])   (lb./MBF)  (kg/[m.sup.3])  (lb./MBF)

Conventional 1      0.051          0.22         0.022         0.096
Conventional 2      0.053          0.25         0.034         0.148
High 1              0.143          0.62         0.097         0.420
High 2              0.139          0.60         0.097         0.419

Charge                 Formaldehyde
                (kg/[m.sup.3])  (lb./MBF)

Conventional 1      0.0005       0.0022
Conventional 2      0.0008       0.0034
High 1              0.0036       0.0156
High 2              0.0038       0.0163

(a) Metric value based on 1.97 [m.sup.3] per MBF.

Literature cited

American Society for Testing and Materials (ASTM). 1999. Wood. ASTM 04.10. Book of Standards. ASTM, West Conshohocken, Pa.

Ingram, L.L., F.W. Taylor, V. Punsavon, and M.C. Templeton. 1995. Identification of volatile organic compounds emitted during the drying of southern pine in pilot and laboratory experiments. In: Proc. No. 7301. Measuring and Controlling Volatile Organic Compounds and Particulate Emissions from Wood Processing Operations and Wood-Based Products. Forest Prod. Soc., Madison, WI. pp. 35-40.

_____. _____. and M.C. Templeton. 1996. Volatile organic compound emissions from southern pine kilns. In: Proc. No. 7292. Drying Pacific Northwest Species for Quality Markets. Forest Prod. Soc., Madison, WI. pp. 41-45.

Lavery, M.R. 1998. Total hydrocarbon emissions from lumber dry kilns. M.S. thesis. Oregon State Univ., Corvallis, OR. 133 pp.

_____ and M.R. Milota. 2000. VOC emissions from Douglas-fir: Comparing a commercial and a laboratory kiln. Forest Prod. J. 50(7/8):39-47.

_____ and _____. 2001. Measurement of VOC emissions from ponderosa pine lumber using commercial and laboratory kilns. Drying Technology 19(9):2151-2173.

Milota, M.R. 2000a. Emissions from wood drying: The science and the issues. Forest Prod. J. 50(6): 10-20.

_____. 2000b. Small-scale kiln study utilizing ponderosa pine, lodgepole, pine, white fir, and Douglas-fir. Report. Intermountain Forest Assoc., Missoula, MT. 355 pp.

_____. 2001. Small-scale kiln emissions: Test considerations, comparison to largescale kilns, and work to date. In: Proc. of the National Council for Air and Stream Improvement, Inc., West Coast Regional Meeting, Portland, OR. NCASI, Research Triangle Park, NC. pp. 241-248.

_____ and M.R. Lavery. 1999. Effect of moisture on the reading of a total hydrocarbon analyzer. Forest Prod. J. 49(5):47-48.

National Council for Air and Stream Improvement, Inc. (NCASI). 1996. A small-scale kiln study on Method 25A Measurements of Volatile Organic Compounds from Lumber Drying. Tech. Bull. No. 718. National Council for Air and Stream Improvement, Inc., Research Triangle Park, NC. 27 pp.

Rice R.W. and L. Zibilske. 1999. Estimated VOC losses during the drying of five northeastern species. Forest Prod. J. 49(11/12):67-70.

Shmulsky, R. 2000. Influence of lumber dimension on VOC emissions from kiln-drying loblolly pine lumber. Forest Prod. J. 50(3):63-66.

Thompson, A.L. 1996. Volatile organic compounds emitted during the drying of southern pine lumber. M.S. thesis. Mississippi State Univ., Mississippi State, MS. 62 pp.

U.S. General Services Administration (GSA). 1996. Method 25A - Determination of total gaseous organic concentration using a total hydrocarbon analyzer. In: Code of Federal Regulations, Ch. 40, Pt. 60. GSA, National Archives and Records Service, Office of the Federal Register, U.S. Government Printing Office, Washington, DC. pp. 1070-1073.

Word, D. 2001. Review of industry and EPA activity to develop MACT standards for the wood panel industry. In: Proc. of the National Council for Air and Stream Improvements, Inc., West Coast Regional Meeting, Portland, OR. NCASI, Research Triangle Park, NC. pp. 226-230.

Wu, J. and M.R. Milota. 1999. Effect of temperature and humidity on total hydrocarbon emissions from Douglas-fir lumber. Forest Prod. J. 49(6):52-60.

Michael R. MiIota *

* Forest Products Society Member.

The author is Associate Professor, Dept. of Forest Products, Oregon State Univ., Corvallis, OR 97331-5751. The author extends his appreciation to the Intermountain Forest Association for their assistance with this work. This is paper 3479, Forest Research Lab., OSU. This paper was received for publication in June 2001. Article No. 9324.

[C] Forest Products Society 2003.

Forest Prod. J. 53(3):60-64.