In early 2003, there were no established methods for the measurement of wax distribution in OSB and only recently had mills expressed concerns about how wax is distributed and its effect on panel properties, especially thickness swell and product consistency. To address these questions, a two-year project “Measurement of Wax Distribution on OSB” was initiated in April, 2003, as part of Forintek’s National Research Program (NRP).
In the first year, two technologies were selected for testing wax measurement, optical image analysis and near infra-red (NIR) spectroscopy. An online NIR (near infrared) sensor commercially available from Moisture Systems Corporation (Spectra-Quad 3000) was tested at the Alberta Research Council (ARC) for wax measurement of furnish samples blended at levels of 0, 0.5, 1.0, 1.5 and 2%, with both emulsified (e-wax) and slack waxes. Results showed very good correlations (R² > 0.9) between sensor measurements and wax level for both slack and e-wax. However, this method is limited because it only measures the relative level of wax application and not the evenness of wax dispersion on strands.
An optical, image analysis method for measuring wax distribution was also developed and evaluated. This is an off-line, benchtop system using a digital camera and lens to capture and analyze magnified images of wax dispersion on strand surfaces. To enable detection and measurement, a ultra-violet (UV) tracer is added to the emulsified or slack wax before spray application to the furnish. Furnish samples are then scanned by the imaging system using a UV light to fluoresce the wax/tracer coverage. Unlike the NIR system, this method is capable of measuring the relative evenness of wax coverage and spot size distributions.
In the second year of the project, pilot plant tests were carried-out at ARC to determine how different wax distributions affected panel properties and whether changes in distribution could be measured using the newly developed image analysis method. Good and poor wax distributions were produced by changing blending parameters to simulate variable mill conditions. Results showed that the poor blends produced panels with significantly higher thickness swell (TS) and water absorption (WA) compared to the better wax blends. For slack wax, WA and TS were 40% and 26% higher respectively for the poor wax blends. For e-wax, WA and TS were 26% and 13% higher respectively for the poor blends. Meanwhile, image analysis measurements showed that the wax coverage distribution of the poor blends had much higher COV values (coefficient of variation; calculated as standard deviation/average) compared to the good blends, for both wax types.
Finally, a mill trial was carried-out using the image analysis system to assess the coverage distributions for both e-wax and slack waxes. Results showed that both wax types were relatively easy to visualize and measure. Measurements also showed that wax coverage COV values were relatively high and similar to the poor pilot plant blends. This trial showed that the image analysis system was useful in gauging the mill’s wax distribution performance, and it’s implications to panel properties, by referencing to earlier pilot plant data.
In summary, this project developed practical methods for measuring wax distribution that can be used by OSB mills, and also showed that wax distribution should be monitored because it significantly influences panel properties, especially thickness swell and water absorption.