CRUDE OIL 101
THE PROBLEM WITH FLUORESCENCE
A common method of sensing oil in water is using the principle of fluorescence. Fluorescence sensors shine light at certain wavelengths and measure the wavelengths that are emitted back by objects in the light path. Different objects have different fluorescence characteristics, which is why it's important to check the wavelengths that a fluorescence sensor produces ("excitation" wavelengths) and measures ("emission" wavelengths). Fluorescence sensors for oil usually produce light at wavelengths around 300-400 nm, and identify the presence of oil when they receive wavelengths between 400-600 nm.
Unfortunately, this isn't a perfect method. Other things (or more scientifically, "fluorescent Dissolved Organic Matter", or "fDOM") fluoresce at the same wavelengths as crude oil. This means that there is a risk of overestimating the amount of oil in the water - maybe some of it is actually bacteria, or algae.
In controlled experiments, fluorescence measurements are taken before the oil is added, and this baseline is subtracted from the fluorescence measurements of oil. Aside from the fact that we often don't have the opportunity to do this during a real-life oil spill, this method assumes that the baseline is true everywhere. This is not a great assumption since the ocean is a dynamic environment and is therefore constantly moving the very things that are the source of our baseline fluorescence measurement. Without an additional means of verification, such as simultaneously using other measurement methods (which introduce their own uncertainties) or taking water samples and measuring the actual oil content in the lab later (which kind of defeats the purpose of continuous, real-time mapping), you can't be 100% sure that all of the fluorescence is from oil.
So, for now, we'll just be looking at relative oil concentrations - something along the (admittedly very unscientific) lines of "no oil, a little oil, lots of oil". Fortunately, in the real world of oil spill response, this type of 'broad' characterization is enough: we can still differentiate on the order of two micrograms of oil per litre of seawater, or two parts of oil per one billion parts of seawater.
'DO-IT-YOURSELF' CRUDE OIL
Importing samples of crude oil - even for academic research - isn't a very quick or easy process. Fortunately, there's a relatively quick (and dirty) way to approximate crude oil ("sludge") with just two ingredients that are easily sourced: diesel (available at a gas station) and road asphalt (there's usually plenty of loose pieces in ditches). Here is how we made a batch of sludge that we used to get an idea of what our sensors would show during the actual experiment.
IT IS NOT RECOMMENDED TO TRY THIS PROCESS, WHICH IS EXTREMELY MESSY.
ALWAYS TAKE APPROPRIATE PRECAUTIONS WHEN WORKING WITH VOLATILE/HAZARDOUS MATERIALS.
Repeat process, filtering out solids, until desired density is achieved
Dissolve small pieces of asphalt in diesel
Obtain diesel and asphalt
(fresher = better)
... and there you have it: "crude oil".
mass: 707 g
volume: 0.820 L