Summary of the results of the diffusion experiment

The following graphs show cleaned up traces for the various test specimens, together with the trace for the empty chamber, whose water absorption/desorption in a cyclic RH is negligible. The values are hourly averages.

The best way to understand these strange graphs is to equate the amount of water exchanged with the specimen during a RH cycle with the amount of water exchanged with a column of air of the same cross section as the specimens. For example, for the bricks masked on one side the water exchanged during the daily cycle is about 8 g per 10% RH change. The change from 50% to 60% RH requires adding 1.5g/m³ into space. So eight grams of water is equivalent to 5.3 m³ of space rising 10% in RH. The area of specimen is 0.2 m² so the column of the same volume at this cross section will be 27 m high. Given a 96 hour cycle the column will be almost twice as high, so the equivalent air column (strictly a space column because air does not influence RH) will be about fifty metres. Another way of looking at it is that if this buffer material is the wall of a space one metre deep, it will take fifty air changes instead of one to get the RH in that space a good way towards the RH of the incoming air. This humidity buffer figure of merit is dependent on the temperature and also on the rate of change of RH and on the geometry of the specimen. It is not a material constant.

Perforated unfired brick with free air flow through the holes

Eight perforated unfired bricks, open both sides. Exposed area about 0.4 sq. m. The right side of the graph is the weight change of the water tank in the empty chamber.

Wood

End grain wood tiles, 5 cm thick, 0.2 sq.m exposed one side. Note that the 24 hour cycle does not use the full moisture capacity of the wood, even though it is cut to give the highest diffusion rate into the cells.

Perforated unfired brick with sides and back sealed

Eight perforated bricks sealed on all but one side. 0.2 sq m exposed. Notice that the 24 hour cycle transfers about half as much water as the 96 hour cycle. The geometry of this specimen is exactly comparable with the end grain wood tiles. The brick performs better as a humidity buffer.

Finally, in this set, the same bricks were exposed to a sudden RH drop after ten days at a constant high RH. This gives close to the maximum available water and allows calculation of the effective diffusion coefficient. The irregularity in the water content trace is almost certainly a real effect, given the precision of the weighing device as shown in the empty chamber.

Discussion of the diffusion rate experiments

The brick performs well, compared with the wood, which represents the best performance to be expected of a cellulosic material. A more diffusion open cellulosic, such as paper insulation, would respond quicker but the lower density will give a poor value of exchangeable water per unit surface area of wall.

The geometry of the brick is probably responsible for the peculiar response to a sudden change of RH. The thin lamellae between perforations give a rapid response compared to the massive ends and sides of the brick, whose relatively slow response give the gently sloping curves away from the moment of RH change. A clever pattern of perforations could give a combination of rapid buffering of the daily vapour flux combined with a long term buffering effect against change of air mass from continental dry to Atlantic humid. The perforations could also be optimised for compressive strength, allowing greater unsupported height to a buffer wall.

I suggest that we concentrate now on optimising the brick performance and test various smooth finishes which still allow rapid diffusion through to the perforations. We could also investigate the possibility of forced or convective circulation through a wall with perforations aligned vertically.

On the instrumental side, the weighing precision needs to be improved. We are making some mechanical changes before the next run. The RH precision is as good as the dew point sensor allows. The temperature control is perfect.