Our Experience Testing So Far...

Now that we have our prototypes, we’ve spent the last few weeks testing them! We’ve experienced lots of challenges and setbacks, but they have helped us learn so much and improve our prototypes.

Kiln:
At this point, we’ve conducted four different kiln “burn trials”. We followed the protocol explained in the last post, if you are curious about the exact measurements we took and information we recorded. To give a quick overview, the way we test our kiln is to find bricks/rack to place the kiln on (to allow for airflow through the holes on the bottom), fill the barrel with feedstock (corn husks, rice husks, corn cobs) about halfway, use a lighter to light the feedstock, place the top/chimney on the barrel, and watch! When the burn looks like it’s going out (or when we want to stop the burn), we use a hose to dump water in the kiln until the fire is out. Throughout our entire testing process, safety is paramount, and we make sure to take safety precautions such as wearing eye protection and being cautious of hot temperatures.

​Trial 1: 
Our first trial was more of a proof-of-concept, and helped us get a handle on the process of lighting and putting out the kiln. We conducted the test at the Lake Lag fire pit, and used corn husks from the farm as our feedstock. The fire was really easy to light (which is something that last year’s team struggled with, however we quickly encountered a couple of major issues: first, the burn was extremely dirty and there was a ton of smoke, and second, the chimney fell off :(. We did get some char out of the first trial, however, and it was very fine.​

Char Product of First Trial

We thought the dirtiness of the burn might have been impacted by preparation of feedstock, since the corn husks are super light and pack a lot of air between them. This air could contribute to the dirtiness of the burn. This informed changes we made for our second trial.

​Trial 2:
Our second trial was pretty similar to our first—we conducted the test at the Lake Lag fire pit, and also used corn husks. However, this time, we made sure to rip the corn husks up a lot more and push them down in the barrel to minimize air. We also reattached the chimney, using the same adhesive but giving it a lot longer to cure this time (one weekend), instead of testing 24 hours after attaching the chimney.
We followed the same procedure as the first trial. Unfortunately, we also experienced the same issue of the burn being very dirty and producing lots of visible smoke. We decided to put out the kiln before it could finish pyrolysis, because it was producing so much smoke.
One of the recurring questions that we’ve been facing in our burn trials is if we are actually performing pyrolysis, and if we are actually creating biochar or just charcoal. At the end of the day, the line between these two is so fine, especially for an imperfect (non-science-lab) kiln made out of repurposed materials. We have been using temperature and dirtiness-of-burn as indicators of pyrolysis (not just regular combustion), but this is imperfect and ultimately only lab-testing would be able to confirm, and this is something we will consider down the line. Now, our major priority is testing how our char/biochar works in a filter. 

Ripped Corn Husks (Feedstock) from Second Trial

Trial 3:
Our third trial took place at the Stanford O’Donohue Farm. This was closer to where we were storing our project materials and obtaining feedstocks for our burns. Following proper safety procedures, we placed our kiln on four heavy concrete spacers atop a gravel patch, as well as wet the ground with a hose beforehand. By this time, we had drilled side holes into the top of our kiln. We wanted to see whether this would make a difference in the burn’s smokiness.
Our burn was nonetheless smoky. However, one interesting insight was that when we removed the lid and chimney, the smoke became significantly clearer compared to before. This was a sign that insufficient airflow was likely a driving factor behind our dirty burn, rather than just the choice feedstock. We also observed a cleaner burn when testing the previous team’s smaller kiln prototype. It didn’t have a tall chimney, which allowed more air to reach the flame and reduced smokiness. In response to this, we quadrupled the amount of airholes on the kiln’s bottom face and tripled the number of side holes at the top. We hoped this would reduce smokiness for our fourth burn test.

With Lid in Third Trial (Very Smoky)

Without Lid in Third Trial (Very Little Smoke)

Trial 4:
Our fourth trial saw another corn husk burn, this time with more air holes. By this point, we had become familiar with the experimental design of a burn test, which made the procedure go very smoothly. The test was brief. Unfortunately, the resulting smoke from this burn was still opaque and dirty, meaning we had not added enough airholes to make a significant difference. The chimney also fell off during this burn test (the third time it did this across our trials), implying we needed to find a more robust means of attaching it. Rather than add even more additional airholes the next time, we decided to test our next burn with a heftier feedstock to see if that would make an unforeseen difference.

​Trial 5: 
On our fifth and most recent trial, we switched the burn’s feedstock to corn cobs instead of corn husks. Cobs had exponentially more mass and rigidity in their shape than husks. We filled the kiln with about 20 corn cobs, then interspersed husks between the top layers to act as kindling. We then began the burn by lighting one husk on fire, and dropping it into the middle of the kiln. 
In general, this burn brought some better news. It lasted 10 minutes, since we allowed it to go on for longer than previous ones. During the burn, temperatures reached 500 degrees Fahrenheit across all parts of the kiln, even peaking at 600 towards the top of it. The burn became less smoky as it went on, which was desired. A reduction of smokiness over time was something we observed in videos of successful kilns online, so this could be seen as a sign we were moving in the right direction. This could have also been due to the fact that we conducted this test without a soup can chimney, since we did not have time to attach it beforehand. This likely increased airflow to the feedstock, cleaning up the burn.
It was unclear to tell when to stop the burn, which is a question we should research the answer to. We quenched it when solid particulates began to rise out of the chimney. Afterward, we observed that our corn cobs had blackened completely and shrunken significantly in size. They were also much easier to snap in half than they were at the beginning - a sign that we changed the material in a promising way. Quality biochar is porous, which translates to less structural integrity. If it’s porous, it means it may be effective at filtering water.

Corn Cobs before Burning in Fifth Trial

Corn Cobs after burning in Fifth Trial

What now?
As it stands now, our kiln continues to face the issue of smoky burns and insufficient temperatures for pyrolysis. Our immediate goal is to talk with experts on kiln thermodynamics and biochar production so we can better understand why our models are performing the ways they are. We also plan to test our current TLUD design once more with corn cobs using a soup can chimney, to see if it affects airflow and smoke clarity, as well as with rice husks, to see if a tightly compacted, airtight fuel makes a difference in amounts of smoke produced. We also have ideas in mind for modifying our prototype. These include inserting an enclosed metal container filled with feedstock into the center of our kiln, then burning additional feedstock around the enclosed container to bake its contents. This would effectively convert our kiln into a retort. Finally, we may modify a paint can to serve as a second barrel in our TLUD, where we’d fill it with feedstock, insert it into our current kiln, and light it. This may help insulate the feedstock and emulate other conventional two-chamber TLUD designs.

Filter:
​
Trial 1: 
The test result for the filter part varies greatly from our initial expectations. We conducted the first trial during week 7, using our prototype with the PVC pipe attached on the side, the corn husk biochar created from our kiln, fine sand, and small pebbles. After 24 hours of incubation, we got the following result:

Results under UV light (showing E. coli)

Results under Regular Light (showing coliforms)

In the pictures above, the tray on the right contains our pre-filtered stock water, while the tray on the left contains water after filtration, both diluted by 1:100. As mentioned in our previous blog post, the pre-filtered stock water is a mix of Lake Lagunita water, Codiga waste water, and tap water.  The luminescent boxes indicate presence of E. coli, while yellow boxes indicate the presence of total coliforms. As the picture illustrates, our filter didn’t properly carry out its function of removing harmful bacteria. The difference in total coliform and fecal coliform levels are not significantly altered. 
Upon reflection, we proposed that there could be three reasons behind our initial failure: 

1. The end product of our kiln did not meet biochar standards
2. Biochar alone doesn’t have the ability to effectively remove E. coli
3. There’s a problem in the design of our prototype
In order to deduce the reason behind this failure, we did our second test in parallel with a control group, using our stacked bucket prototype and commercial biochar kindly provided to us by Yiran Li, a PhD student at the Luthy Lab. The commercial biochar is from douglas fir and is industrial grade, so we would expect that it would remove E. coli from the water samples. 
Test result can be seen here:

Numerical Results of Testing (Larger numbers equals larger presence of contaminants)

This picture contains the numerical result of the color of indicators, same as the first test, with TC indicating total coliform, FC indicating fecal coliform, large and small indicating the number of large and small chambers that indicates presence. 
While comparing the numerical result with the IDEXX reference sheet, we learned that there are around 400 MPN/100mL E. coli in the stock solution prior to filtration, 0 in the water after filtration through the second prototype (stacked bucket model using commercial biochar), and more than 24000 MPN/100 mL E coli. in the water after the second trial of filtration through the first prototype (side-outlet using kiln-produced biochar). 
The conclusions we can draw from these results are that:
1. High quality commercial biochar does have the ability to remove E. coli, 
2. E. coli actually proliferate in our first prototype, contaminating the water instead of filtering it, suggesting that there is a flaw in our design
After consulting Yiran, we learned that the growth in E. coli is not uncommon, because the bacteria feeds on dissolved organic carbon, which is abundant in the biomass, such as corn husk, that we used to produce biochar. Yiran mentioned that is will be difficult to filter significant amounts of E. coli out of our water sample without using chlorination or other more advanced methods. In the next few weeks, we will reevaluate how plausible it is for us to continue focusing on the complete removal of E. coli. It may be more realistic to focus on the removal of heavy metals instead. 

Thank you so much for reading our post! In conclusion, although we've encountered many challenges, we've learned so much from them and are better designers and engineers because of it!