Heat Sink Material Selection

The heat sink would require a material which can easily absorb and store heat, this is known as thermal conductivity. A material with high thermal conductivity will allow heat to transfer more easily than a material with low thermal conductivity; most metals have a good thermal conductivity and most plastics have low thermal conductivity. A rule of thumb is any material that feels cool to the touch as room temperature is generally a good conductor of heat.

The particular properties desired for a greenhouse heat sink fall somewhere in between, as aluminium for example is a very good conductor of heat, but it is also quick an expensive material to use for this kind of project. Metals in generally are costly or have a higher value for recycling so are not as easy to come by, additionally you ideally want a material which can absorb heat over a longer period of time which will also release this heat at a slower rate (the ambient temperature will play a large part here).

Rocks will fall into the category of a slower to absorb heat to the core and release that energy, glass and concrete also make good heat sink materials as they have a relatively good thermal conductivity properties.

Material Testing

The three material chosen for testing where: Granite chips, glass chips and concrete chunks. The glass and concrete was taken from a local recycling centre with their permission and the granite chips were sourced from Jen’s driveway.

The material testing was designed to provide a basic profile of each material so that some conclusions could be drawn on the best material for the prototype system.

The test used a small plastic box which had a pipe in the base, air would be forced into the pipe using a small fan and allowed to vent through small holes in pipe at the base of the box. The air would then move through the test material and exit through a small vent in the top of the box. The box would be first left at room temperature and then cooled in a freezer providing a short accelerated temperature profile of the material. A temperature sensor would be used to log the material temperature thought the test and was buried in the material but only placed 3 cm below the surface.

The following images show the box and some of the material being tested.

The material and temperature sensor were allowed to settle for 10 minutes then the fan was started and the material was warmed at room temperature for 2 hours, then the material box was placed into the freezer and cooled for 2 hours with the fan on, for a further 10 minutes the fan was turned off at the end.

The next few images show the graphed results for the granite chips, glass chips and concrete chunks:

The first thing to note looking at the graphed results is the granite and glass have very similar profiles, this was expected as both materials have a very similar thermal conductivity values, list of material types and thermal metrics here. The granite and glass were also of a similar size pieces ranging from 1-3cm .

The concrete depending on the type also has similar thermal properties to the glass and granite, but the major difference here is the individual size of the material pieces. As the the concrete pieces are larger there is less effective surface area in contact with the air flow, therefore less heat can be absorbed and the core temperature takes longer to reach the same temperature of the outside of the material. Although the core temperature is not measured it can be seen that the concrete warms at a faster rate and cools more rapidly as well, but retains a fairly steady temperature for a period before falling below 0°C. Some of this is due to increased airflow through the concrete due to the materials irregular shape as it is less dense than the smaller pieces of granite and glass, the tests setup and length of time the test is run for could also be improved.

When the test temperature falls below -10°C it can be observed that there appears to be some anomalous data, this is due to the temperature sensor circuit. The circuit was designed around a PT1000 temperature sensor with a small window from -10°C to +60°C, the temperature circuit profile has areas of non-linearity at the edges of this window which can account for in the recorded values.