Low VOC drying of lumber and wood panel products. Progress report Number 9 [January 1999]

Abstract

SummaryResults &om a multi-year study show that a significant part of the extensive variability observed in OSB flake dryer emissions can be traced to physiological effects, and the rest can be attributed to handling and other factors. Low-headspace treatment of lumber was scaled up to the 50 kg level. The amount of turpentine collected was of the same magnitude as that released upon drying lumber. For the process to be economical, the wood must first be brought to about 95°C with steam, and then processed with RF. Attempts to remove VOCs fkom OSB through low-headspace by placing a curtain over the furnish failed because of leaks. A more rigid container will be required. 0RF-treatment does not alter the gas permeability of lumber. 0 Effect of seasonal variation in southern pine terpenes on dryer emissionsThis study was begun in 1996 because a southern pine oriented strand board (OSB) mill reported wide variability in Method 25A VOC emissions measured in the field. Two measurements could not be reproduced even when taken on the same day; it was unknown whether this result was due to random variability in the raw material itse& systematic seasonal variations, or to handling. An additional issue was that the sporadic measurements typically made in the field may include unrecognized seasonal effects, atypical of average annual emissions. The mill was interested in alternative means of making measurements and sent samples of fiesh flakes to the Institute of Paper Science and Technology for analysis. These samples were wrapped, stored cold, and dried in a 130°C tube furnace equipped to monitor total hydrocarbon emissions (Banerjee et al., 1998). The results of laboratory analyses fiomthe first two year's samples demonstrated that the VOC concentrations varied by as much as 500% @om -300 pg/g to 1500 g/g on a wet basis), as shown in Figure 1.Consistent with the mill's previous Method 25A measurements, significant variation was apparent even within single sampling dates. There appeared to be some degree of interactive relation with the ambient temperature and rainfall data (Figure 2), but (at least in part because of the data variability) it was not clear that there was a s i m c a n t causeand-effect relationship. It