Biodiesel stability test results surprise researchers

NREL Chemist Earl Christensen provided preliminary data from an ongoing stability study of B100 and biodiesel blends
By Ron Kotrba | February 13, 2013

National Renewable Energy Laboratory Chemist Earl Christensen presented preliminary results from an ongoing B100 and biodiesel blends stability study at the National Biodiesel Conference & Expo in Las Vegas last week. The purpose of the study, Christensen said, is to determine what it will take to store neat biodiesel and various blends for up to one to three years. I attended the session last week, and this morning I followed up with Christensen on the telephone.

For B100, only two samples were tested, and both samples were dosed with antioxidant at the production site, a typical procedure. The feedstocks used to produce the biodiesel samples were not known. The aging tests were done in accordance with ASTM D4625 and held at 43 degrees Celsius (110 F) for weeks at a time. In this study, 13 weeks simulates one year.

“One of the samples only took five weeks, and once it dipped below 2.5 hours on the Rancimat test, we dosed it back with antioxidant to reach the original Rancimat time,” Christensen said. Once the first sample failed the Rancimat after five weeks, the researchers added 200 ppm of TBHQ and 500 ppm of BHT to two different aliquots to “reset” them, and they each lasted another five weeks before they dipped down to 2.5 hours. The other sample, however, lasted 26 weeks in an underground storage simulation (two years) with no adulteration save the antioxidant it was treated with at the biodiesel production site. After lasting 26 weeks, when it fell to 2.5 hours in Rancimat, that same sample was retreated and stable for another 13 weeks, “which was really surprising,” Christensen said, adding that it took so long to go out of spec again that it delayed some of their other projects.

New B100 samples were drawn for the blend tests. Two biodiesel samples were taken prior to additization at the production site. Once the unadditized samples were obtained by NREL, the researchers dosed each with enough antioxidant to pass the three-hour and six-hour Rancimat tests. Then each of those was blended at 20 percent with two different types of diesel fuels, one hydrotreated and the other hydrocracked, to make up a total of eight B20 samples.

“Where we’re at right now, going to a three-year simulation (39 weeks), we’re only two-thirds into the test, so the data is preliminary,” Christensen said. “What it implies though is at two years of simulated underground storage, almost every [B20] sample remained on spec for Rancimat.” Two of the eight tested B20 samples were off spec for the Rancimat test but Christensen said both of those were blended with biodiesel samples dosed to pass the three-hour test and blended in two different diesel fuels. “Essentially after two years, almost every B20 sample is still an on-spec fuel,” he said. The ones that dipped below the Rancimat threshold could be reset with additional antioxidant dosing.

Eight B5 samples under the same conditions were tested, and every one of those has remained on spec for the two simulated years.

One of the things Christensen noted in his presentation is that the same bio with the same treat rate can perform very differently depending on which diesel fuel—hydrotreated or hydrocracked—it is blended with. Christensen said he can’t really comment on that right now because he doesn’t really have an explanation for it yet, and NREL is doing more detailed characterization of the diesel fuels to try and understand the differences between them and what sort of impact their properties may have on the stability of final biodiesel blends.

One of the comments Christensen made during his presentation is that, for blends, Rancimat remains the more conservative value. This morning I followed up with him on exactly what he meant by this.

“In a B100, if you measure peroxide value, it is [typically a] very good measurement to tell you when the onset of spoilage, or rancidity, is occurring because when you begin to see peroxides form, you know oxidation is occurring,” he said. “In a blend, however, we weren’t able to detect any change in peroxide value despite the one biodiesel sample going off spec for Rancimat, which was kind of surprising. We expected to be able to detect an increase in peroxide value based on oxidation occurring, but for the blends, the Rancimat remains the more sensitive, or conservative, analysis.”  

Christensen concluded our call this morning by saying, “We still have a lot of work to do in analyzing the data, this is all preliminary, but we’re learning longer-term storage is most likely capable, and two years of storage for biodiesel blends is very achievable, it’s just a matter of understanding the initial stability that you have and how stable a biodiesel product you have begun with. At this point we can’t say with certainty what the specifics are going to be, we have to wait for all the data. And it’s a limited study, so we can’t draw too wide of conclusions from it.”