As expected, initial treatments provided highly variable retentions depending on the product type (Douglas-fir, spruce or radiata pine; Table 2). The preservative solution appeared to preferentially penetrate along pre-existing cracks in the wood and along the bond-lines, particularly in the panels made with refractory species (Douglas-fir and spruce; Fig. 4). The distributions of borate shown in Fig. 4 are following treatment and without any diffusion storage or subsequent wetting. If these treated samples were exposed to increased moisture, the yellow (unprotected) areas could be expected to turn red (protected) with time (Cabrera & Morrell, 2009), as long as adequate overall retention was achieved.
Table 2
Summary of preservative retention, swelling and mechanical properties
Species
|
Treatment
|
N
|
SBX Retention (kg/m3)i,ii
|
Thickness Swelling (%)i,ii
|
Rolling Shear Strength (MPa)i,ii
|
Rolling Shear Stiffness (MPa)i,ii
|
Douglas-fir
|
None
|
10
|
--
|
--
|
1.70 ± 0.11a
|
195 ± 32ab
|
Initial
|
10
|
0.25 ± 0.04a
|
0.40 ± 0.49a
|
1.59 ± 0.13a
|
154 ± 20b
|
Optimized
|
10
|
1.04 ± 0.19b
|
0.80 ± 0.61b
|
1.68 ± 0.13a
|
246 ± 64a
|
Spruce
|
None
|
10
|
--
|
--
|
1.28 ± 0.11ab
|
172 ± 23a
|
Initial
|
10
|
2.48 ± 1.04a
|
0.93 ± 0.59a
|
1.10 ± 0.07b
|
131 ± 33ab
|
Optimized
|
18iii
|
1.12 ± 0.19b
|
1.63 ± 0.65b
|
1.29 ± 0.08a
|
124 ± 18b
|
Radiata Pine
|
None
|
10
|
--
|
--
|
2.06 ± 0.23a
|
294 ± 59a
|
Initial
|
10
|
9.11 ± 2.09a
|
2.77 ± 1.24a
|
1.18 ± 0.38b
|
149 ± 51b
|
Optimized
|
6iiii
|
1.80 ± 1.03b
|
0.51 ± 0.95b
|
2.68 ± 0.26a
|
188 ± 53ab
|
iMean values shown with 90% confidence intervals
iiValues within a species and measurement group with a different letter are significantly different (p < 0.05, ANOVA with Tukey HSD test for multiple comparisons)
iiiExtra samples of this treatment were tested accidentally
iiiiSamples of this treatment were limited due to limited supply of CLT panels
|
There is currently no standard for preservative treatment of CLT; however, the New Zealand (where subterranean termites do not occur) Standard NZS3640 (New Zealand Standard 3602, 2003) for lumber is approximately 1.0 kg/m3 SBX, as is the ‘biological reference value’ (similar to a ‘toxic threshold’ for various fungi) (Lloyd, 1997) - suggesting that it might be suitable for CLT to prevent beetles, drywood termites and fungi. Interestingly this retention may also mitigate subterranean termite activity over time (Jones, 1991), although it does not prevent damage in small test specimens according to EN 599 (DIN, 2014). The American Wood Protection Association (AWPA) standard for lumber for interior applications (UC2; (AWPA, 2022)) requires much higher retentions (2.7 or 4.5 kg/m3 SBX), but this is intended to protect against subterranean termites (non-Formosan and Formosan, respectively), in addition to decay and other insects. For this study, we targeted the lower biological reference value and the New Zealand Standard, based on the assumptions that CLT buildings are constructed above-grade and that separate subterranean termite control measures will be implemented according to building codes (e.g., soil termiticides). By this reasoning, the main threat to CLT will be non-subterranean insects and fungi, for which 1.0 kg/m3 SBX will provide effective control.
With reference to the 1.0 kg/m3 target, the initial treatments undertreated the Douglas-fir while over-treating the spruce and radiata pine samples. The ‘optimized’ treatments came close to achieving the target, across all the species. This potential for adjustment is noteworthy, given that Douglas-fir and spruce are refractory species, while radiata pine in this test was mostly highly permeable sapwood.
Dimensional changes in the length and width were negligible, as expected for a cross-laminated product in which the longitudinally oriented lamellae restrain swelling in that direction. Thickness swelling was measurable and mostly positively correlated to the amount of preservative retained (Table 2). The apparent increase in spruce swelling in the optimized samples (with lower retention) is counterintuitive and we believe may be an artifact of a different person making the measurements of those samples. Standard tolerances for thickness variation of CLT panels are 2% (FPInnovations, 2019); the average thickness swell of the samples receiving optimized treatments was below that level, suggesting that the thickness swelling observed here would be acceptable in practice. Delamination was not measured but appeared to be minimal across all samples.
The rolling shear test typically resulted in failure near one or both bond lines, usually crossing over the middle lamella (Fig. 5). Initial treatments that resulted in overtreatment led in some cases to dramatic reductions in mechanical properties (e.g., radiata pine rolling shear strength; Table 2). In contrast, the optimized treatments did not lower the strength values of any species tested. Reductions in rolling shear strength, associated with very high treatment (water-based treatment) retentions, may result from the wetted wood swelling, and the differential swelling across- and along-the grain causing stresses on the bond lines. Stiffness values were less consistent in response to treatment retention levels and poorly correlated to strength values (R2 = 0.31, simple linear regression). Guidance in EN 789 (European Committee for Standardization (CEN), 1996) notes that the variability of rolling shear stiffness obtained from planar shear testing is high.
Overall, these results suggest that meaningful retention levels of borate treatment can be achieved using vacuum, that the process can be adjusted to the species being treated, and that the impacts on the mechanical properties can be minimized.
Vacuum bag treatment demonstration
The vacuum bag trial was successful. It provided high retention (1.52 kg/m3 SBX) and good penetration (as indicated by red color between interior lamellae and in internal checks; Fig. 6) in a panel made with a refractory species (Douglas-fir) in which the lamellae were edge-glued in addition to being face-glued. This result suggests that vacuum treatment of full-size panels in plastic bags is practically possible, even in less permeable panels. In our trial, some of the treatment liquid entered the vacuum port which was close to the surface of the panel. Thus, relocating the vacuum port and/or installing a vacuum trap may be necessary to prevent problems with treating solution entering the vacuum pump.
Purpose-built, rigid vacuum treatment tanks may be more practical in commercial settings, but flexible bags may provide useful test chambers for the development of treating processes with reasonably large samples.