In this video, Sunny discusses water testing in Bangkok, using brick clay and alternate burnout material and flow rate testing. Enjoy!

Please click on ‘Ceramic Water Filters Video Update’ to view video.

Figure 1 shows a range of different colours associated with a wide range of temperatures to which the surface has been subject, with that range being a function of packing, fuel type, the direction and strength of any prevailing wind, as well as the duration of the firing. This temperature variation indicates a potential problem for the production of a strong, permeable, sintered ceramic filter, as it indicates lack of uniformity in density and permeability, with probable variation to post-sintered strength.

Those areas that are black and grey, [2]. [3], [4], [5], indicate that the surface temperature at those points can be regarded as not having exceeded 700°C, with carbon deposited on the surface by combustion remaining on those surfaces as a consequence. It can reasonably be assumed that beneath those regions combustion of organic materials contained within the pot’s cross section will also be incomplete, with corresponding consequences for both permeability of the structure as well as development of an approximately uniform density of the sintered structure. Permeability will be variable across the walls of the pot because of the probable non-removal of organic material, while density will also be lower there than in surrounding areas where higher temperatures will have produced a more open void fraction because of more extensive, hotter sintering and contact-point fusion of constituent particles.

The rim of the pot [1] shows an approximately even colour that is associated with both an oxidising flame and surface temperatures sufficiently higher than 700°C to ensure the combustion of carbon on at least the surface of the vessel. It is probable that the core of the vessel’s wall will still contain residual carbon with reduced permeability as a consequence. See Fig 3, [5].

Macro voids are seen on the surface of the rim. These will be a consequence of the making process and the combustion of organic material during the firing.

It may be assumed that the voids that can be seen on the rim’s surface, [4], [8], will exist with variation to their distribution though out the pot’s cross section and it was to reopen that void fraction that I suggested scraping the surfaces of the pot when dry. But regardless of the efficacy of that, if the carbon is not removed in the firing, [5], the pot will not be able to achieve uniform maximum permeability.

Macro voids are seen on the surface of the rim. These will be a consequence of the making process and the combustion of organic material during the firing.

It may be assumed that the voids that can be seen on the rim’s surface, [4], [8], will exist with variation to their distribution though out the pot’s cross section and it was to reopen that void fraction that I suggested scraping the surfaces of the pot when dry. But regardless of the efficacy of that, if the carbon is not removed in the firing, [5], the pot will not be able to achieve uniform maximum permeability.

Figure 3, [4], [6] shows the broken cross section of a pot broken in transit, with two layers of oxidised sintered clay in which carbon from organic material additions has been successfully removed. The differences in the thicknesses of the external and internal oxidised layers can be attributed to differing time periods over which the respective layers were exposed to flame >700°C, and in the case of the internal layer, because of its relative inaccessibility to flame path and direct heat.

When fired in a conventional kiln a ceramic object is exposed to increasing heat that is delivered via flame pathways, where combustion gases and flame will move though a pack, flowing around and over surfaces in much the same way that water moves in its passage over and around rocks. Water is said to follow the path of least resistance and this is similarly true with flame, so the packing of objects then becomes a determining factor in flame-path development and direction. A second factor to be considered is that the temperatures achieved in these low-temperature Lao firings are significantly lower than those temperatures, ≈1150-1200°C in a conventional kiln, at which heat ceases to be delivered via flame pathways and where the whole ceramic-pack mass assumes a much narrower temperature-range variation and energy is distributed through out the pack by radiation. This means that the following factors then achieve importance greater than they would in a conventional chamber: the placement of objects relative to each other and the consequent development of flame pathways, the volume of fuel available and its water content, prevailing wind speed and constancy, the total elapsed time of the firing and the relative humidity of the air available for combustion.

So, in summary:

• The pots need to be fired hotter and longer
• Their method of packing needs to be such that more of the internal surfaces of the forms are available to flame pathway development, this can be facilitated by stacking the forms on their rims, so that the interior of the forms are open and not sitting like upside-down cups.
• The firing will need more fuel and that might have to be added during the course of the firing, with it not being assumed that fuel heaped over the objects being fired at the onset will be adequate and not need replenishment (this might already be being done, I have no way of knowing.)
• The wall thickness of the filter structure might have to be reduced and made more uniform, so that uniform combustion of carbon is achieved

By Dr. Tony Flynn.

So when we say ‘we have tested another batch of filters’ what do we actually mean? What do we do? Please allow me to elaborate on how we create a rudimentary testing lab in very rustic conditions. What follows below is an ordered list of the steps that we follow.

1. Collect some dirty contaminated water. It must have faecal coliform bacteria in it. We have experimented with: cow poo (fresh), water form a dirty puddle by the side of the road, green swamp water, and water with a few drops of juice from raw pork meat (if smell was any indicator, this one is by far the most contaminated). The aim is to have a sample of water highly concentrated with bacteria but that is still relatively clear. This is because our testing laboratory has informed us that clear water is easier for them to test. Yes facilities are limited here!
2. Buy five of the highest quality bottles of drinking water we can find. These are used for transporting test samples, providing a check that the laboratory is applying adequate hygiene practices, and that the filters are fully sterilised.
3. Sterilise the filters, and the jar that will be used to collect the water from the filter. Filter and collection jars are also boiled in between collecting samples.
4. Move the filters hygienically into the filtering harness you have seen in the pictures.
5. The filter is now ready for use. The first sample passed through is straight from the bottle of clean drinking water. Once analysed, this will tell us if we are properly sterilising the filters.
6. Once the required amount has been collected in the cup at the base of the filter this is poured into one of the, empty but still uncontaminated, drinking water bottles.
7. The next sample is the dirtiest contaminated water we can find.
8. The last sample tested is dirty water double filtered by two filters stacked one atop the other.
9. The five samples are labeled and marked as below:

• A straight clean sample that has not been through the filter. This allows us to confirm that the testing laboratory is reliable. We know that if a clean sample comes back as having bacteria in it there is a fault with the laboratory’s testing process.

• A straight dirty sample that has not been through the filter. This allows us to register a baseline for the number of bacteria in the water.

• A clean sample that has been through the filter. For reasons stated above.

• A dirty sample that has been through the filter. Percentage filtration rates are then obtained by comparing the bacteria number in this sample with the bacteria number in the dirty sample baseline above.

• A dirty sample that has been double filtered.

The five samples are then placed in a refrigerated bag and transported to the laboratory for immediate testing.

When each filter is tested we also time how long a measured amount of water takes to pass through so we can calculate flow rates and record them for future reference.

So that is how we are testing the filters currently. Needless to say, we will send the filters off to an International quality testing laboratory to confirm that the filters are adequate before they are ever used in villages.

Sabaidi!

There are a number of factors to consider when producing this variety of clay pot water filter. Please allow me to talk a little about these factors and illustrate some of the aspects we must consider when designing a working filter.

Thickness of filter: This is a crucial factor as there is a trade-off between a thick filter, which will remove more bacteria, but at a very slow flow rate; and a thinner filter that traps fewer bacteria, but allows water to flow faster.

Firing the filter: The thicker a filter is the more difficult it is to fire. Thicker filters must be sun-dried for at least a week before they can be fired. This is to remove all the moisture from the wet clay. Water causes the clay to crack, or even more spectacularly – explode, during firing

Clay-to-coffee ratios: Tony Flynn has found that an approximate 50/50 ratio of coffee-to-clay is good, but that 40/60 and 60/40 combinations have also proven effective. We need to experiment to find the optimal ratio.

Coffee grounds: The coffee grounds in Laos are coarser than those found in Australia. We need to experiment with locally available methods to grind them finer. Ultimately, we are seeking to replace coffee with something that is even more widely available. One step at a time though, we’ll use coffee as that has already been demonstrated to work.

Type of clay: The Lao potting techniques have a few unique characteristics, namely that pottery is fired using rice straw and without a kiln. Pieces are exposed directly to the naked flame, thus experiencing a rapid rise in temperature. The temperature climbs much faster than during a kiln firing. To alleviate the effects of thermal shock, a special mix of pre-fired clay is mixed with the wet clay used to make pieces of pottery. When done by a skilled potter, adding this pre-fired clay, allows the pieces to undergo rapid temperature changes without cracking or exploding.

Adding pre-fired clay may alter the filtration properties of the filters. We will investigate the possibility of producing a filter without pre fired clay. Also, we will determine the impact of pre fired clay on filtration rate, effectiveness, and taste of the filtered water. Steps will be taken to explore methods to reduce any negative impacts the pre fired clay has on the filter characteristics.

Density of the filter: Our discussions with Tony Flynn indicated that the density of the filter is very important in ensuring it functions correctly, as a filter that is too dense will not allow water to pass through. In Laos, pottery pieces are shaped by placing the piece on a wooden anvil and beating it with a paddle. This compacts the pots and may require adaptation to produce a low density filter.

Our approach in effectively implementing the use of these filters in rural Laos and ensuring potable water has a few stages: Firstly, we will experiment with the pottery process and materials to create a filter that reliably removes bacteria. Once we have a working filter we will adjust various aspects to meet the needs of users, such as size, shape, durability and flow rates. Finally, we will simplify and standardise the production process, while seeking to make it as universally applicable as possible. We will also investigate potential alternative materials.