Is FIREClean really just Crisco? I headed to the lab to find out. The results will be out soon, but for now here’s what I did in the lab and the background science behind the tests.
Rocking my ESS Crossbows and Noveske Rifleworks hat in the lab
Before I begin, I need to thank some people for making this project possible. First and foremost is my academic advisor at Worcester Polytechnic Institute, Professor Drew Brodeur. He not only got me access to the labs, but gave me some extremely helpful advice on my methods for testing. If I didn’t have him helping, this project would never have happened.
Next is Professor John MacDonald of WPI, who was extremely helpful in reviewing my planned methods and giving suggestions. I had sent a short email explaining the project and expected something along the lines of “Sure, that would work fine.” What I got was five paragraphs detailing what would work and what I should focus on, as well as other professors that could be helpful to talk with.
Some other key individuals are Andrew Butler and Daryl Johnson, also both of WPI, who allowed me to use the research NMR labs and helped me along the way. It’s not every day you get access to $3 million worth of fancy equipment, and they were extremely kind for letting me run NMR samples, as well as helping get the data in a useable form.
Outside of my college, I also need to thank Curtis from the VSO Gun Channel. He is an experienced chemist and took a lot of time to discuss my methods, suggest other tests to do, and be another opinion in evaluating the data.
Finally, thank you to Andrew Tuohy for inspiring this project in the first place, getting me the samples, and doing most of the work outside of the lab. I’m assuming most of my readers came here from his blog, but if you haven’t seen it go check out vuurwapenblog.com
Now, on to the reason you’re here…
Several weeks back Vuurwapen Blog posted about the Fireclean controversy. For those that missed it, several people claimed that the popular firearm lubricant and cleaner is actually just repackaged Crisco brand canola oil. Several Youtube videos show Fireclean and various oil samples placed on a stove smoking at the same point, leading many to cite these videos as “proof” that Fireclean is a common cooking oil.
The problem is, any decent scientist knows that smoke point of an organic compound is not nearly enough proof to claim that two samples are the same. So Andrew Tuohy of Vuurwapen Blog talked with an organic chemist at the University of Arizona to conduct Infrared Spectroscopy of Fireclean and two types of Crisco oil. As I am currently an undergraduate studying chemistry, this testing greatly interested me. His final conclusion was that Fireclean is “probably a modern unsaturated vegetable oil,” but not Crisco. Without doing more testing, I couldn’t say for sure. Along with several comments asking to test X brand of firearms lubricant next time, I got an idea.
As far as I know Mr. Tuohy is not a chemist (if he is, I need to figure out how to use my degree to make it so I can shoot guns all day), so I assume that he doesn’t have access to the equipment or skills to do this testing every day. But I’m a chemistry student. I pay absurd amounts of money in tuition to a small technical school, so I can easily get full access to some of the most complicated and stupidly expensive chemistry equipment you can imagine.
And I did…
After I read the original post of the IR spectra, I sent an email to my academic advisor asking if I could get access to one of the instrument rooms in our labs for a personal project. Once I confirmed that I could do IR testing, I emailed Mr. Tuohy to see if he was interested in the project. A few weeks later, I got a box with 18 small vials, each labeled with only a number. During the weeks I was waiting for the samples I sent several other emails to professors and others seeking help with the project. I had asked if I could use the NMR lab that is used for teaching the experimental chemistry labs, but was told it was for class use only. Instead, I was informed I would have to use the much nicer NMR lab in our research facility… It’s not often that you ask to use a $400,000 piece of equipment, only to get told you have to use the $3,000,000 thing that does the same task even better. Oh well, if I have to…
I should note that neither myself nor Mr. Tuohy knew what sample numbers are any given oil. I was told that the pair of #6 and #8 are the Fireclean/Criso samples, but as of writing this I don’t know which one is which. I was also asked to compare a few other pairs or groups of samples. From what I understand, Andrew gave 18 labeled vials to a friend to randomly fill with 18 different oils. That friend then told him the pairs to be compared. This double blind methodology ensured that there was no bias by anyone involved in the testing.
The samples as I got them from Andrew Tuohy
When the samples arrived the only change I made before the trip to the lab was to add labels to the caps in addition to the labels Andrew put on the side of the vials. This made it easier to see the sample numbers while they were in the carton that Mr. Tuohy shipped them to me in. Then I grabbed my safety glasses and it was off to the lab.
You know you’re a chemistry nerd when the highlight of your fall term of junior year is spending about five hours in the chemistry labs during finals week…
The first thing I did was start running IR spectroscopy on the samples. Unfortunately the first day worth of testing (samples 1-9 of the 18 total) didn’t save, so I had to redo them the next day. That’s why the data Mr. Tuohy will post (or maybe already has) shows the sample names 1 through 9 as Sample#X_001 instead of Sample#X like the following ones do. I don’t want anyone claiming I faked the data, so I’m just putting that out there now. All of the spectra sent to Mr. Tuohy were taken on the same day in the lab.
The Perkin Elmer IR Spectrometer. The underclassmen never clean up after themselves...
The IR spectra were taken on a Perkin Elmer Infrared Spectrometer with an Attenuated Total Reflectance attachment. The ATR attachment means that instead of spreading the sample between a pair of sodium chloride plates and securing it in a holding device (a time consuming process for a lot of samples) I could simply deposit a drop or two of the oil (or spread a small blob for the grease samples) on a small metal plate with a lens in the middle, swing a small probe attachment over it, and screw the probe down to secure it. Using the ATR I was able to run a sample in about two minutes, as opposed to 10-15 minutes for the NaCl plate method we used in my general chemistry classes.
As you can see, it doesn't take much for an IR scan with the ATR. The amount used here is several times the amount needed for a proper scan.
One thing Curtis from the VSO Gun Channel was sure to remind me of was the possibility of cross contamination between each sample. For this reason I used acetone and kimwipes (fancy science tissues that don’t leave any residue. We use them in lab a lot for cleaning sensitive equipment like lenses) to triple clean both the plate and probe before each sample was run. I also changed out the nitrile gloves I was wearing frequently when I got any oil on them, ensuring that the spectra seen are only of that sample.
Cleaning the IR Spec. I went through lots of Kimwipes and gloves...
How it works:
IR spec. uses infrared light to bend and stretch the bonds in organic molecules. Different bonds have different strengths, so it will take different amounts of energy to cause a change. Infrared light is in the proper energy range that it adds the right amount of energy without breaking the bonds. Depending on the two atoms bonded, a given wavelength of IR light will cause bonds to stretch, bend, rock, or twist. By looking at the wavelength of light that the sample absorbs, we can know the energy level that is going into changing the bonds. This is measured as the percentage of light transmitted through the sample versus the wavenumber (1/cm, which can be used to calculate the wavelength and frequency of the light). A “peak,” or low spike in the IR spectrum indicates that the sample absorbs at that wavelength, and therefore a bond is doing something with that energy. The peaks of common bonds are well studied, so by looking at the wavenumber of the peak we can accurately predict what bonds are in the sample being studied.
Note that all of the spectra have a lighter line and a darker line. The lighter one is the original spectrum of that sample and the darker one is after adjusting to the baseline. Before any testing was done I did a “background scan,” essentially scanning without a sample in place so that I could “blank out” any peaks caused by outside sources. The important things to compare are the darker lines, as they have been corrected by the computer program for the background and other possible errors.
Scan in progress. From here the spectrum was adjusted to a baseline and saved. Some spectra were then overlaid to compare sets of lubricants.
What this means for us:
By looking at the peaks in different spectra, we can see what bonds the samples have in common. I’ll be doing a more in depth post after my results are posted on Vuurwapen Blog, so for now I’ll keep it general. I don’t want to ruin the surprise before Andrew can post the results. We already know all of the samples are hydrocarbons, so we can easily expect a major peak around the 3000-2900 cm-1 range from the C-H bond stretch. If we see another peak in the spectrum of a sample, we can check the literature values for that wavenumber and determine what chemical bonds are there.
Nuclear Magnetic Resonance Spectroscopy:
NMR is a complicated but extremely useful method of analysis. Several of my professors have described it as “the gold standard in analytical chemistry,” and for good reason. When combined with other methods of analysis it allows a chemist to determine the molecular structure of organic molecules.
How it works: Nuclear Magnetic Resonance is one of the more complex methods to explain. I spent most of a general chemistry class learning how it worked and how to evaluate the results. I’ll do my best to explain it here without making it too complicated. NMR was actually the basis from which we get MRI imaging. Magnetic Resonance Imaging in hospitals uses the basis of NMR scanning to build a computer image of certain parts of the human body. So the next time you see a movie or TV show in a hospital where something metal goes flying across the room and sticks to the MRI machine, you can sort of understand how that machine functions. Yes, the magnets are that strong, but the nicer equipment usually has shielding to prevent flying metal parts like you see in the movies.
Nuclear Magnetic Resonance, like the name suggests, uses magnets to look at the nucleus, or core, of an atom. Certain isotopes (forms of the same element that have different masses) of varying elements can be used for NMR because they have an odd number of protons and neutrons. Neutrons and protons have “spin” values of plus or minus ½, and will pair up when possible, a +½ “up spin” pairing with a -½ “down spin” and resulting in no net spin. The reason that NMR requires an odd number of protons and neutrons is that the imaging requires a net spin that isn’t zero to properly image the sample.
The NMR is essentially a giant electromagnet. It uses liquid helium to cool large coils of wire down to the point that they become superconductors. Liquid nitrogen is then used to help insulate this setup, because liquid helium is extremely expensive compared to liquid nitrogen. Depending on the exact setup the nitrogen levels must be topped off every week or so, with the liquid helium levels being maintained every few years. This giant magnet is powerful enough to align the spin of the nucleus in the sample to be either spin up or spin down relative to the magnetic field.
The down spin state is slightly lower energy, and therefore more stable. But there’s not much difference in energy level, so even the relatively low power of radio waves can cause the nucleus to flip from spin down to spin up. Much like a ball balanced at the top of a hill, the up spin state is unstable and would easily “roll down the hill” to the down spin state. Much like IR spectroscopy used IR waves and looked at what wavelengths of light were absorbed, NMR transmits a range of wavelengths of radio waves through the sample and looks at what wavelengths are absorbed and used to flip the spin states.
The NMR spectrum that results is shown as a series of peaks. The X axis (horizontal axis for those of you that don’t remember high school math classes) is in units of parts per million, or ppm, and calculated based on the strength of the particular NMR used. So the raw data from a 300 MHz NMR is going to look different than the raw data from a 700 MHz NMR.
The Y axis (vertical axis) of an NMR spectrum is not measured in any standard measurement, instead it is relative to the specific spectrum. Each peak on the graph is shifted to the left based on the atoms it’s bonded to, and the “shift” as measured in ppm will be consistent across all NMR spectra for that sample, but the height of the peaks will vary from NMR to NMR. The relative peak height will be consistent though. Given a single NMR spectrum with one peak being twice the height of another peak, the taller peak indicates that “type” of atom occurs twice as often in the molecule analized. When looking at the NMR spectra Andrew Tuohy posts please keep in mind that you need to compare the relative peak height between the two spectra, instead of directly comparing the peak height between any two peaks across the different spectra.
A note on NMR solvents:
As stated, NMR detects atoms with odd numbers of protons and neutrons. Hydrogen-1 (known as proton NMR) and carbon-13 (or C13 NMR) are the most common, but other possibilities include nitrogen-17, oxygen-19, and many others. Unfortunately, hydrogen-1, also known as protium, composes approximately 99.9885% of all naturally occurring hydrogen. This means that using standard solvents will mask any chemical shifts from the actual sample. Therefore we must use special solvents that don’t contain protium. To get around this, scientists have made solvents using hydrogen-2, known as deuterium (represented by a D). Many common lab solvents have been made using deuterium instead of protium. Water for NMR analysis is D2O or “heavy water.” Other options for solvent include deuterated acetone, deuterated THF, and the solvent I used for this analysis, deuterated chloroform or CDCl3. No, chloroform is nothing like in the movies. You won’t suddenly pass out from a tiny bit of it, although it doesn’t exactly smell good…
I love the smell of Chloroform in the morning... Note that deuterated solvents are absurdly expensive. This bottle was 100mL and cost my school $60. Good thing I had free use of the labs
To prepare the samples I zeroed out a small beaker and a new NMR tube on an analytical balance. I then added as close as I could get to 0.25 g of sample #6 to the NMR tube. Approximately 0.6 mL of deuterated chloroform was measured out in a 5 mL graduated cylinder and added to the NMR tube via a pasteur pipet. The tube was then capped, labeled, and inverted several times to ensure the oil sample was fully in solution. This process was then repeated with sample #8 using new pipets. In order to get a “look” at the carbon structure of the oil samples, I decided to use C13 NMR. This type of NMR uses naturally occurring carbon-13 as the detected atom in the structure. For this analysis I used the 500 MHz NMR that my school has in their research labs, and the computer program converted the units to ppm. The images I sent to Andrew Tuohy of the data were the exact PDFs I was sent by Daryl Johnson, one of the people in charge of the NMR lab at my school.
I don’t want to beat Andrew to the punch, so I’m going to wait for my analysis of the data until he has posted the results. I sent him the IR spectra, several IR overlays to compare sets of lubricants that he had requested I look at, and the two NMR spectra. Keep an eye on his blog for the full data set. Shortly after he publishes the results I’ll post my analysis as a chemist. Please note that I have no opinion on this issue. I have used FIREClean in my rifles in the past, and really don’t care if it is or isn’t canola oil. What matters to me are facts, which is why I used science. Regardless of the outcome, some people will disagree and get pissed off. To those people, I suggest you start studying for your PhD, then go buy your own NMR and run your own tests. The collective group of experts consulted for this projects totals over a hundred years of experience in chemistry, so I have full confidence in my results. Until someone shows otherwise with hard scientific facts, I will take my results as fact.