By Brett Miller & Dan Cassens
This small experiment included two separate tests. Both tests included 4-inch solid rift cut white oak flooring and 4-inch rift cut engineered wood flooring with a 4-mil sawn wear layer and an 11-ply platform. The material was part of both flooring experiments being conducted at Purdue University.
The first test consisted of one piece of each type of flooring. Both test pieces measured 4 inches wide on the face and about 33.5 inches along the length. The material was conditioned at 75 degrees Fahrenheit and 35 percent relative humidity. These conditions should result in an equilibrium moisture content of about 6.8 percent for solid wood. An inch-long piece from each sample was weighed and set aside to determine moisture content on an oven dry basis. The two samples were weighed as each set of dimensional change measurements were taken and then oven-dried and weighed at the end of the experiment. The length of each piece was measured to 1/1000 of an inch in one marked location with a digital caliper. Likewise, the width was measured in three marked locations and the thickness in three marked locations. Marking the spot for measurement insured that the same spot was measured each time.
The moisture content samples were only an inch along the grain whereas the samples used for width, thickness, and length measurements were 33.5 inches long. Because of their short length, it is possible that the moisture content samples responded more quickly to environmental changes than the larger samples. If this did happen, it would make the dimensional changes given in Table 1 somewhat conservative. That is if the large samples had been given more time, the dimensional change would have been somewhat larger.
Table 1 presents the results for both the solid and engineered wood flooring specimens. The specimens were measured May 1, 2016, while they were in the conditioning room. The specimens were then moved to outside, but covered, storage. Four additional sets of measurements were taken through Aug. 25. Note that the initial moisture content of the solid flooring was 7.4 percent and the engineered flooring was 7.0 percent. In just eight days of outside, but covered exposure, the moisture content of both samples jumped to more than 12 percent. From May 9 through June 10, the moisture content actually decreased. Little rain was received during this period, so in spite of it being spring and early summer in Indiana, we had low relative humidity resulting in a declining moisture content for the samples. The last set of measurements were taken Aug. 25. Beginning in July, rainfall and relative humidity increased, and the moisture content of the samples went to more than 14 percent.
The manufacturing histories for the engineered and solid flooring are unknown, so it is difficult to make definitive statements. The somewhat consistently lower moisture content for the engineered flooring as compared to the solid flooring is probably the result of a reduction in hygroscopicity. Shmulsky and Jones (2016) indicate that “wood that has been subjected to temperatures in excess of 100 degrees centigrade for long periods becomes less hygroscopic; i.e., it equalizes at a lower moisture content than normal wood.” These authors indicate that this phenomenon is more important in fiberboard and particleboard products because they are typically subjected to higher temperatures during manufacturing than plywood.
The 4-inch-wide solid oak flooring expanded in width 0.064 inches or about 1/16 of an inch in going from 7.4 percent to 14.8 percent moisture content. This is 1.6 percent and very close to what would be predicted using Wood Handbook (2010) numbers. On the other hand, the engineered flooring expanded 0.02 inches or just 0.5 percent. This low expansion is due to the cross laminates in the base material restraining each other.
Thickness swell is another important property. The solid flooring increased 0.013 inches or 1.7 percent as the moisture content increased from 7.4 percent to 14.8 percent, while the engineered flooring increased 0.019 inches or 2.8 percent. Again, Shmulsky and Jones (2016) indicate that if high pressures are used in the manufacturing process, thickness swelling can be more for plywood as compared to solid wood. Thickness swell is likely to be even more in fiberboard and particleboard products.
Longitudinal shrinkage is the last property to consider. Normal wood moves very little along the grain, and any movement in that direction is generally ignored. However, it has been shown that juvenile wood in the softwood species can move enough along the grain to be troublesome in the wood truss industry. For both the solid flooring and the engineered flooring, expansion was just a little more than 3/100 of an inch or 0.1 percent.
These measurements are for small samples of flooring. Flooring systems often cover larger areas and these small changes can become significant and should be accommodated by design features.
The larger flooring experiment at Purdue University included installing the 4-inch solid white oak and 4-inch engineered white oak over an OSB subfloor (provided by Huber Engineered Woods). Additional materials were donated by Fortifiber, MAPEI, Allegheny Mountain Wood Flooring, and WD Flooring. These panels were installed per NWFA Guidelines by Dean Hultman, Owner, Hultman Flooring (who is an NWFA Certified Installer). This installed panel then became the door to an environmental chamber. This allowed the testing to include different controlled conditions on both sides of the installed floor (as the door). The intention was to recreate the moisture gradient we often find within an installed flooring system over a crawlspace or unconditioned basement. The wood flooring was laid perpendicular to the length of the OSB panel.
The RH in the environmental chamber ranged from 60-85 percent during the course of two months. The RH conditions outside of the environmental chamber (in the lab) ranged from 35-55 percent throughout the same period. Board width measurements were taken weekly and recorded.
Measurements showed the wood flooring swelled as predicted based on the gain in moisture. What we did not expect was that the ends of the solid planks also opened up, resulting in end-joint gaps. After a thorough analysis, and taking into account the orientation of the wood floor in relation to the OSB, we determined that these gaps opened up in response to the subfloor swelling from the gain in moisture from underneath. This was backed by the fact that both OSB and plywood shrink/swell approximately twice as much in their width as they do in length.
These two samples of solid and engineered wood flooring responded in a very predictable way. Moisture content change due to exposure to non-controlled environmental conditions resulted in significant dimensional change. The moisture content and amount of dimensional change depends on whether the sample was solid wood or engineered flooring. It is this moisture content and dimensional change of the flooring that is often focused on when problems develop. However, the visible flooring is only part of an entire system composed of joists or sleepers, subfloor, vapor retarders, and fastening systems (adhesive and/or mechanical). Once assembled, the entire system begins to function as a unit. The subfloor is likely to be plywood or a reconstituted product, namely an oriented strand board (OSB). The properties of the reconstituted materials and impacts on the entire system vary depending on the manufacturer.
Furthermore, all of these basic building materials are often subjected to non-controlled environmental conditions during storage and construction. Dimensional change (shrinkage) can then occur as the members adjust to climate-controlled conditions typical of an occupied house or commercial building. This change can affect the flooring itself.
Daniel Cassens is Professor Emeritus at Purdue University. Brett Miller is Vice President, Education and Certification, National Wood Flooring Association. Neil Osborn and Dan Bollock also assisted with the study. Neil Osborn was an undergraduate student in the Department of Forestry and Natural Resources, Purdue University at the time of the study. Dan Bollock is a Wood Laboratory Technician in the Department of Forestry and Natural Resources, Purdue University.
Shmulsky, Rubin, and P. David Jones. 2016. Forest Products and Wood Science – An Introduction, Sixth Ed. Wiley-Blackwell. West Sussex, UK. [477 p.]
U.S. Department of Agriculture, Forest Service. 2010. Wood Handbook: Wood as an Engineering Material. Gen. Tech. Rep. FPL-GTR-190. Madison, Wisconsin: Forest Products Laboratory. [508 p.]
APA Technical Topics. 2016. Moisture-Related Dimensional Stability, TT-028C. 5pp.