Cheap, Green Solutions to Pollutant Degradation
Advisor: Moshofsky
Internship Location: Cheap, Green Solutions to Pollutant Degradation
Onsite Mentor: Dr. Kevan Wang, Product Development Manager at NVIGEN
Hello, my name is Christopher Xu, and my project regards the degradation (essentially removal) of a simulated pollutant (Methylene Blue dye) with a photocatalyst (something that speeds up a reaction under light) made from copper and chitosan, where chitosan is a derivative of chitin, a polysaccharide found in crab and shrimp shells. Studies have been conducted with this catalyst with the same pollutant but only under lab conditions, meaning neutral DI water. These studies miss a crucial point: while this catalyst may work under lab conditions, there is no telling how they might work under harsher conditions such as those in the ocean: specifically, pH, salinity and hardness. I will investigate how the catalyst performs under these conditions in hopes to gain further insight into when it can be used for pollutant removal most efficiently. Furthermore, I intend to gain insight on the main drivers of reaction. To do this, I will run radical scavenger tests or introduce certain chemicals into the reaction to take away one of the reactants, then investigate if there was a change in reaction rate. For the experimental design, I will have 2 controls, and 3 experimental groups. The two controls will be DI water (to compare against current literature to ensure I synthesized the catalyst correctly) and a solution of copper ions (to make sure that it is the complex, not just the copper that is degrading the pollutants). The three experimental groups would be pH, salinity and hardness, each with a low, medium and high value. I would expect the pH group to show the catalyst being most efficient in the medium value, and the salinity and hardness to be most efficient in the low group. This project will allow us to determine the best time to use a copper-chitosan complex to remove pollutants, knowing what salinity, hardness and pH are best, and what groups are essential for driving the reaction.
Project Blog
Week 1: - Context, Literature Review and Initial Tests
In this senior project, I will be investigating the effect of testing a photocatalyst made of copper, an abundant metal in the Earth's crust, and chitosan, a deacylated (acetate groups removed) derivative of chitin, which is found easily in crustacean shells in different ocean conditions, namely, pH, hardness, and alkalinity. As the ocean's dynamics continue to change (I have done data collection on these variables in tide pools two summers ago), the efficiency of such catalysts decrease. If students have taken Honors or AP Chemistry, they'd know that catalysts often work in an ideal pH or temperature range. If the environment is too extreme, the catalyst's efficiency drastically decreases. Therefore, I am doing this experiment to see if this cheap and abundant photocatalyst (Cu-CS complex) can be used across all these environments to successfully degrade a simulated pollutant (Methylene Blue).
As a student who has taken Honors, AP and Inorganic Chemistry, I am fluent in all lab techniques and procedures, such as vacuum filtration and titrations, all of which are required in a lab. Furthermore, this project serves as an extension of my research with UCSD two summers ago, that dealt with heavy metal effect in bioluminescence. I would expect to see very high amounts of degradation in the control (DI water), as most literature reports this as about 90% degradation within 30 minutes. Across the other three, alkalinity, acidity and salinity, I'd expect to see the most neutral values to exhibit the highest amount of degradation (i.e. low alkalinity, low salinity, neutral pH), as this introduces the least number of external variables (i.e. Hydrogen ions and alkaline metal ions for acidity and alkalinity respectively), allowing for no inhibition of this process.
Week 2: - Preliminary Tests
Last week, I did literature review and formulated a rough procedure. This week, I set to testing this procedure. The first barrier I had to overcome was dissolving Cupric Oxide (CuO) to isolate copper ions. The papers I had read up to that point mentioned a concentration of 1% volume to volume ratio of acetic acid (found in vinegar) to water, but I found that the concentration was too weak. To solve this, I initially tried a "stress test", using the most concentrated acid I could find to attempt and dissolve the acid. The result I observed, after letting the mixture settle, was a clear and dark blue solution -- perfect. Going off of another source, I tried this again with 0.3 Molar (diluted) acetic acid, which worked as well, but only after I boiled it on the stir plate's max setting (to increase reaction rate). After cooling to room temperature, I dissolved 1 gram of chitosan into the solution. This is where trial one went wrong: after stirring for two hours, I observed two chunks of slimy substance, which upon neutralization via 1M NaOH and drying in an oven at 80 degrees Celsius overnight, resulted in a slimy substance the next day -- not the result I wanted. On the second attempt, I adapted my approach, slowly adding chitosan and gently heating to allow small amounts of chitosan to slowly dissolve before adding more, resulting in a more continuous, gel-like solution. After adding NaOH and neutralizing and drying, I found a better result, with a rubbery film over a near transparent liquid (residue of copper that isn't needed). The rubbery solid was put in a desiccator over the weekend and then ground with a mortar and pestle to get a flaky, dark teal substance -- the catalyst.
This week was mostly for troubleshooting and making adjustments to the literature procedure to see if it'd work. The literature used sonication equipment, muffle ovens, and vacuum ovens, none of which I had. I had to test if each of my substitutes would work, and if so, perfect them. And with the week over, I have my conclusion: replace muffle ovens for the thermally decomposed product -- it will reduce to reactivity but will suffice for me. Sonication -- instead of adding chitosan in one go and sonicating it, I will slowly add it, while gently heating it to prevent burning the chitosan while allowing a greater reaction rate while using a magnetic stirrer on the highest setting. Vacuum oven -- instead of 1 night, I will use the desiccator for a whole weekend to remove moisture. Next week, I hope to start preliminary testing on Methylene blue, as there is a lot of ambiguity in that procedure too, namely, the concertation of Hydrogen peroxide used to generate the ROS (reactive oxygen species) to start the degradation.
Week 3: - First trial, optimization
Although this week was mostly at Disneyland, I did manage to get some work done. To recap, on Monday last week (which is when I wrote the blog) I obtained a flaky and dark teal result, which was the catalyst. I decided to crush it instead of break it into chunks for two main reasons: one was for easy measurement: literature mentioned masses of around 0.2 to 0.4 grams, and it would be easier to measure out small flakes than large chunks. Second, smaller flakes would allow for a greater surface area of the catalyst to be exposed to the solution, allowing for more interaction with the dye and the catalyst, potentially allowing more degradation.
The methylene blue solution in stock was liquid, while the literature used a solid that was dissolved. I found that a vast majority of the solution's constituents wouldn't inhibit reaction rate, though the existence of some alcohols in the solution would compete for the OH radical groups. However, the concentration of the alcohol I deemed to be sufficiently low enough for there to be little effect on the rate. As for H2O2 concentration, the paper didn't say the exact concentration, probably since they referenced another paper (locked behind an expensive paywall) but it would present its own issue. Normally, H2O2 is good for the system: it generates ROS allowing for the removal of dye. However, it forms an "inverse parabola" relationship with quantity: too little, and not enough ROS are generated and therefore not much degradation happens, while too many leads to competition in the active sites. Therefore, I chose three values, all from other literature that don't necessarily deal with Cu-CS but with similar photocatalysts: 5mM, 10mM and 20mM, to see which one works the best. Next week, I hope to finalize my procedure and within a week or two, meet with Mr. Wang.
Week 4: - A New Take
Last time, it took me one week to get catalyst, drying, dehydrating and all. The result? 0.8 grams. A pretty solid percent yield, all things considered (80%), as I had to replace so many steps with my own due to not having access to advanced equipment. However, this is far too slow, even if each trial only uses 0.2g (half the paper's) of catalyst. Therefore, this week, I decided to try to double everything: double the volume of acetic acid, double the initial copper and chitosan. I mixed in the copper separately, so it didn't take forever to mix. Then, I slowly added chitosan until I ended up at a smooth, thick solution I got in the first trial. Upon adding the NaOH, though, I saw some different results, as it seemed like the moment a drop of NaOH would hit the solution, it would precipitate some copper (indicated by the solid blue on top of the solution). After drying (for much longer this time, due to much more water being there TO dry) I was left with the same result: a slimy film that would need to be dehydrated in the desiccator before crushing.
Next week, I hope to start running trials with degradation and figure out the optimal concentration of Hydrogen Peroxide (if you don't remember, in a previous blog, I mentioned that Hydrogen Peroxide was required to generate the ROS to start the reaction, but too much would lead to it inhibiting the reaction instead). I also need to learn how to use the UV light.
Week 5: - Formulation of a Procedure
So, this week, it was mock AP week, meaning I couldn't go into the lab and get much work done. I mostly stayed home and read papers on the photocatalytic degradation step, as the paper I originally took the procedure from was extremely vague in the sense that they didn't provide me with any concentrations or numbers. For instance, the paper mentioned the "addition of 1mL H2O2", without telling me concentration. As I said in a previous post, this was a problem, due to there being a peak at which H2O2 drives the reaction forward without inhibiting it. Upon reading a few sources, I found that most value lie between 1-20mM. I settled on 10mM, and how I'll reach such a small concentration is by serial dilution.
If any readers have taken Honors Chemistry, surely you remember the lab where you diluted Kool Aid down by orders of magnitude of 10 each time. I intend to do the same thing with Hydrogen Peroxide. I will first get to 0.1M, by using 1mL of my stock (30%, which is around 9.8M, which I rounded up to 10M for simplicity) and 99mL of water. This gets me to 100mM. Then, I dilute this tenfold with 10mL of this new solution with 90mL water, getting to my desired volume of 10mM. As for the other concentrations, I also calculated it. In a previous post, I also mentioned that the Methylene Blue I'm using is different from the papers, as they used a powder, but I'm using a solution. My solution is approximately 0.23% dye, meaning about 2.3 grams per mL. If I want the 2mg they mentioned in the paper, I will need approximately 0.87mL of solution, or, rounded for simplicity, 0.85mL. Adding this to about 9.15mL of water will get me to what the paper used: 2mg in 10mL.
The UV light also produced another issue. While the manual I found online says it supports bulbs at 200 ish nm and 300 ish nm (I need 365nm). However, as the lamp I have uses a bulb, I am unsure if it has one of the two specifications, as whoever used it prior to me totally could have bought a new bulb and swapped it in. Therefore, I am essentially running of an educated guess: I can see the light (I didn't stare at the light though don't worry), so it can't be too low. The fluorescence emitted by paper matches the images I found online of people using lights of similar wavelength.
Week 6: Tests WITHOUT a centrifuge
Over the past weeks, I have done everything to build up to this week. I have obtained a UV light, my catalyst, and all the numbers required to make my solutions. Now, I have initial results from basic equipment. I learned a few things: first of all, I should wait a bit to see if my reading goes up at all. After inserting a cuvette into the UV-Vis, I notice that it takes a while for the absorbance reading to settle at a slightly consistent value. Of course, my setup is much more sensitive to light, as the spectrophotometer I'm using is open to light, making it possible that the lights of the room is introducing noise into my data. However, I did see a clear (but small) decrease in concentration, indicating that I synthesized my catalyst right. Of course, the values I'm getting now are quite close to each other, likely due to not centrifuging and a tiny flake or two in my sample scattering the light, leading to incorrect readings.
Next week, the biggest goal is to secure a centrifuge. I need a way to get rid of the flakes/powder inside my solution and a centrifuge is the best way to separate the actual solution from the solid. This obviously entails me learning how to use a centrifuge. I have used one two summers ago, but I wasn't the one operating it. I do know that I need to balance it, with "dummy" tubes with water, as I will only be centrifuging one at a time. As for where I will be obtaining this centrifuge, I have contacted the biology department, and Ms. Pochiraju has agreed to lend me their centrifuge. With this, I can run an official DI water trial, and record data and start graphing. I also should start finding out the values I'll use for hardness, pH and salinity, based on the values of these in different types of water (ponds, oceans, tidepools, etc.)
Week 7: - First "Successful" Trial
All right, so last week I mentioned I ran some tests without a centrifuge. And so, the results were quite rough. The samples I was drawing had "impurities" in the sense that they would lead to incorrect readings from the spectrophotometer. The worst case of this was one test, where the absorbance actually went up quite a bit after allowing "degradation" to happen. Logically, this made no sense, as some dye should have gone. Also, there physically should have been no way for absorbance to go up since that would indicate a higher concentration afterward, even though I added nothing. So, with this in mind, as well as repositioning the UV light, I was ready for my first "real" trial.
About the UV light: in the past, I would shine it down onto the solution. However, this presented a problem: despite stirring, only the top layer would receive the light. Think about it: when stirring, a vortex is created, but the solution doesn't necessarily "flip" upside down. This is a very watered-down explanation of why this is bad, but ultimately, it just means that the lower layers of the solution will receive less UV than the upper layers, leading to less efficient degradation. So, by moving the light to shine on the side, not only do I bring the light closer to the solution, but since the solution is stirred circularly, more of the solution receives the light, leading to more "uniform" generation of ROS.
The last modification I made was about the sample. In UV-Vis, the spectrophotometer outputs a graph, with absorbances across the visible light spectrum. Usually, you'd want your peak absorbance to be at around 1. 1 is optimal, since too low would mean most light passes through, leading to unreliable data, and too high would mean nearly no light reached the detector, allowing for noise to dominate your data. Therefore, I reduced my sample from 1 mL of solution and 2mL of DI water to 0.5 mL of solution and 3 mL of DI water, allowing me to reach a peak absorbance of 1, leading to more accurate data.
With all these adjustments, I ran the test, and I found a visible drop (in the data, not the naked eye) in absorbance, allowing me to finally confirm my catalyst is usable. Next week, I will be meeting with Kevan, and running my control group, while also researching further on the ocean's range of pH, salinity and hardness. My current research is pretty good, but it only is in the Bay of Bengal. I would like to find more sources after this one to find tests on a wider scale (all oceans).
Week 8: - Testing onsite
So, after last week's "successful" trial, I have started testing onsite. I met with Kevan initially on Wednesday, and I got experience in a much more advanced lab. For instance, I didn't have to manually shake my aliquots, but instead I could use a vortexer to quickly mix up my sample. The UV-Vis was also a lot more sensitive than what we had at school: I had to consider the tiny amounts of leaching from my catalyst, and the 1-to-31 diluted 10mM introduction of H2O2 into my solution, all of which I didn't ever have to consider with the schools very much insensitive spectrophotometers. On a side note, though, the computer that runs the program is a Windows XP. Anyways, onto my trials.
On Wednesday, I was figuring some things out, such as how to use the spectrophotometer, and what volumes to use, since the pipettes in the lab went up to 1mL. I essentially halved all the volumes, so while I might have been using a 0.5mL of my solution and 3mL of water in the cuvette before, I started using 0.25mL and 1.5mL instead, to fit in the 2mL centrifuge tubes. However, the results weren't the greatest. I saw null results everywhere, which leads me to believe that there might be some noise coming in due to the dye staining the cuvettes. Next time, I intend on running a control, in the sense that I don't even start the reaction, I just put the catalyst and dye together, but with no peroxide to generate ROS. That way, if the dye is truly staining, I will be able to see and compare the data sets to see change. Next week, I will also start making solutions to simulate ocean conditions. I found another good source and have some numbers to aim for now.
Week 9: - Simulating Ocean Conditions
Okay, so after testing last week, and although not getting the greatest results, at the end of the day, the purpose is to see if there is a statistically significant difference between the behavior in ocean conditions and behavior in "lab" conditions. I found a few more sources. Unfortunately, due to time concerns, I have made the decision to reduce the number of groups for salinity and hardness from low, high and neutral to normal and neutral, neutral being the control of DI water. This was made mostly due to me trying to get my catalyst perfectly aligned with the papers, which, likely due to my limited resources, I was unable to do. So, I need to get on with the next step of my experiment with testing ocean conditions. Tomorrow, I intend on meeting Kevan again to start running some tests.
A bit of a look into how my next week will look: I intend on going to Kevan twice, once on Tuesday and once on Thursday, where Tuesday will be the reduced salinity and hardness tests, while Thursday will be the pH tests. The pH tests will still be the same, as I will be running a low, neutral and high pH. On Wednesday, I will go to school and make my pH solutions with numbers I will calculate later on. And then, on Friday, I will do numbers and finish up my presentation. On the topic of the presentation, I have already began making it. Most of the slides are pretty simple: background, methods, and my motivations to do this. The only thing I have after Friday is the data, analysis, and room for improvement section and I'll be done.
Week 10: - Finalizing, Paper, and Presentation
Well, now I've got all my data. Working with Kevan was pretty interesting but quite monotonous, all I did was mix my catalyst into a solution of water and premade stock for each ocean condition and let it sit for an hour, while every ten minutes I'd get up and sample it. I decided to not run a UV-Vis test between each sampling time because 10 minutes was just too short. When I initially tried doing so, centrifuging, diluting, measuring and cleaning just took too long, especially cleaning, since I didn't have a sink I could rinse the cuvette out with. Normally, I'd quickly stick the cuvette in the sink and use the tap to quickly clean it, but I didn't have the ability to do so at NVIGEN, so I ended up waiting until I had all my samples before measuring each one. While this may have caused some additional degradation while it sat, I did cover them in aluminum foil to reduce light exposure, which, in theory, would stop the reaction.
Anyways, with my data, I ended up plotting the difference between my experimental group and my control group, because it showed the trend more cleanly than separately plotting the points. That said, though, my statistical test was still with the two samples. I ended up using the Mann-Whitney U test. Those students who have taken AP Statistics might know about such tests, and while a 2 sample t test would be the most obvious course of action, my sample size, 6, was just too small for the t test to reliably give me a result. Moreover, my data was anything but normal. Therefore, this was perfect for the Mann-Whitney U, or the Wilcoxon ranked sum test. This test is great for small data sets, and is intentionally made for non normal data. Even better, it only works on continuous data, and with my absorbance measurements being of that type, it was perfect. My stats ended up showing a statistically significant difference in the saline and hardness solution groups, but no statistically significant difference in the pH groups.
As I have finished my paper, now all I need to do is polish my presentation and I will be good to go.
Week 11: - Presentation, Final Thoughts
Well it's all done now. The presentation was pretty easy to make, it was mostly just rehearsing and practicing, including some mock presentations with Dr. Mo. For my graphs, I decided to go with differences in absorbance. This was so I could show a better downward trend, showing a difference. This was then confirmed with a Mann-Whitney U test so I could find if a statistically significant difference existed between the groups. These are all the big things that happened to my project in the week. Now, onto reflection.
To be honest, this project taught me a lot about lab work. In the past, I have always followed a pre-tested lab procedure with appropriate equipment, and got a more or less expected result. This project made me do everything but that. I was using a more or less pretested procedure, but without equipment and without an expected result. I had to get used to seeing really bad results, since it is merely a part of research. Furthermore, chemistry work turned out to be a lot more draining and boring than I was used to. A lot of the labs I did in chemistry courses could be completed in a period. Only when I got to senior year was when I started doing multi-day labs. However, this project forced me to wait an hour before I could run stats, and these stats had to be collected across five different groups which was a total of five hours for just one trial. This is what I meant by not having enough time to run a second trial for any group during my presentation. The time it took to schedule with Kevan, run the tests and then do stats was simply too long to fit more than one trial in. All in all though, this was a pretty fulfilling experience and I got to learn a lot, especially how to cope with failure which I'm sure will help as I go to college.