Oil Testing -- grading the molecules

A batch of oil product can be tested at many points along its way from refinery to end-user, each time it changes hands or is blended into another slightly different permutation.

Inspectors go out into the field and take a sample from an onshore tank, or barge or ship, in what looks almost like an old jam jar, and deliver it to the lab. Once inside, the sample is tagged before being subjected to a vast array of tests that range from low to high tech, simple to vastly complex.

For light products, specifications are set by the government, and are quite black-and-white: meet the specs, and you can sell the oil. Don't meet the specs, you need to reblend. For heavy products, which are often changing hands between sets of blenders, a buyer will usually accept a slightly off-spec delivery, for a price discount. Also, what may not meet the specs for one end-user, may be just fine for another one.

Among the interesting lessons we learned:

--There are multiple ways to test for the same specification. For instance, aluminum and silica had 3 different methods. The different methods come back with slightly different results, and take different amounts of time. To our non-surprise, sellers typically want the quickest test done.

--Tests are conducted multiple times, to measure repeatability (whether the same person at the same machine can get the same result from the same sample) and reproducibility (whether a different person at a different machine can get the same result from the same sample). There is a margin of error on every test.

--Different samples from the same tank can come back with vastly different results. Why? Oil, especially No. 6 oil, is not totally homogeneous when different components are blended together. There can be striations -- that is, layers of different quality. Samples are generally taken from the bottom third, middle and upper third of a tank, in order to take a representative summary. But as the gentlemen from Intertek pointed out, the striations are not always perfectly horizontal, so quality can differ across the same level of the tank.

--Some of the testing methods used are upward of 100 years old. This shocked us at first, but makes sense now that we have seen the lab area itself. Since testing is mostly a chemistry experiment, it is no surprise that some of the most accurate, widely accepted testing methods involve a lab technician eyeballing a hydrometer, or timing the pace of residual fuel ooze with a stopwatch. The lab has computerized X-ray scanners sitting next to Bunsen burners and test tubes that would be home in any 10th grade chemistry classroom.

--There is an actual replicated car engine in the back of the lab, in which they test octane by simply running a sample of gasoline to check on engine knock.

There is no doubt as to the rigor of the testing regimen; some of the tests take up to 12 hours to complete. And the role of the tester will continue to be absolutely vital for continued smooth(ish) running of the physical market, as an arbiter of quality. As specifications continue to get tighter on both clean and dirty products, the testing companies will likely play an even larger role, with even more pressure for fast, accurate results.

And just remember, it ain't 87 octane, 13.5 RVP conventional gasoline until these guys say it is.