Because of discussions about real-world quality control issues which came up in the discussion-based Advanced Topics biodiesel classes I've done recently, I'm spending some extra time this fall in doing some testing of the finer points of the 3/27 test to try to learn more about it's parameters.
At the moment I'm trying to figure out what exactly happens across various fatty acid profiles and with various cold temperatures at which the test can be done. I have access to some very limited laboratory use periodically, and hope to correlate some of the cold-temperature physical properties of the 3/27 to GC or other conversion testing.
However, a funny thing happened on the way to the forum:
It just so happens that the day I went to the lab to make up my Big Testing Series samples, I had just filled up with what turned out to be bad fuel from a Co-op That Shall Not Be Named (CTSNBN). This group remains Not Be Named because I don't want to embarass these people and they actually do try really hard to make quality fuel and use 3/27, pHlip test, and sometimes even do GC testing using an outside lab, something that few homebrewers have done.
They work with fairly difficult feedstock Their feedstock varies wildly, depending on what members (or the cat) drag in, and is sometimes pretty high in FFA. They do two-stage base/base 80/20 processing and sometimes make really well-testing fuel as some of the GC samples reveal- they've been well within ASTM before when using better quality feedstock.
However, their oil is sometimes high in FFA, they do not do acidbase processing to deal with it, they dont' have really good dewatering for the bad oil, and they react at lower temperatures than desirable. I now suspect that we all should use more KOH than just 7 g+ purity compensation, too, and they aren't doing this yet.
At one of my recent classes I made a full-size batch and counted off 6 major quality control things (temperature, water, bad oil, measurements, etc) that I did wrong in front of the class, yet my big batch still made good biodiesel that passed 3/27 at 64F, and I used this to illustrate the point that 2-stage base-base biodiesel process can be used to make up for a huge array of quality control 'sins'. It looks like the Co-op That Shall Not Be Named must have done at least 7 things wrong because their fuel turned out to be just bad enough to cause problems, and unfortunately 2-stage base process did not save them there.
The first sign of trouble was that the fuel was refusing to clear fully after days of active drying (ie circulating through a sprayer). I'd never seen that before and tried some 3/27 tests.
It appeared to pass at about 72F and failed at about 66F. Bad news- this
is the exact situation we're afraid of, people not checking their 3/27
test at cold enough temperatures, and getting a false pass.
I actually spent many hours testing various other home brew samples and most of the other WVO-based samples didn't have a pass/fail point at quite this high of a temp, which gives me more confidence in the test-ie most commonly it seemed that the fuel could pass at 60F and fail at 40F (I tested most of the samples at 65F or lower, picked that temp just because that's what my ice bath maintained well, then challenged the same sample again at much lower temps to see what could cause it to fail).
Most of the other samples, if they clearly passed at 60F or so, also passed at lower temperatures , at least till I got down to the 40F's.
I tested about 10 samples so it was good to see that this wasn't a common problem. These were mostly samples made in production runs by various homebrewers I'd recently met.
I'm currently doing a project to characterize cold-temperature anomalies with 3/27 and to try to send samples to a lab to get some more correlation between it and GC or other conversion tests. I've been talking up labs as I often do, trying to find a way to do this testing without paying Magellan $89 a sample for GC testing. I've found a few educational institutions where I may be able to do this.
The obvious problem with the Fuel That Shall Not Be Named made by the Co-op That Shall Not Be Named is that the fuel was borderline hazy- like it's almost dry but not quite. Definitely not what I"d call 'bright and clear and dry' but what some people would call 'clear enough' if they didn't know better.
offtopic diversion: Earlier that day I was at a
laboratory chatting with an ASTM petroleum diesel committee guy, and they certainly had problems with ASTM settling on a real definition and real test for 'bright and clear' (which I think is in the gasoline spec or something)- clarity is a bit subjective, even to chemists on the ASTM committees. Homebrewers with any level of experience do know what bright and clear and dry means, but newbies can be confused by something that's 'almost dry' but not quite sparkling clear. I got to argue briefly with a couple of fuel-hungry co-op members who put this slightly macho peer pressure on me ('that looks good enough to ME, I'd use that in MY Truck' said the men, the implication being that I was being a wimp. I gave in to the peer pressure and used it in my truck. Bad idea.)
Anyway, the lack of 'clear drying' appearance of the Fuel That Shall Not Be Named caused me to take some stuff to the lab and try to test it.
It failed 3/27 if it was chilled slightly below room temp. I was really careful in how I tested the temperature- I had a laboratory chilling device that was an ice bath (looked like a big insulated crockpot so it'd stay at temp for a while) in which the ice (or ice and salt) chilled four little wells and you'd stick the sample tubes into the wells to get air-cooled (ie the samples weren't in the ice itself, just in cold copper wells).
My technique was to leave the samples (and sealed prepared methanol vials) in the chiller wells for 10 minutes till the temp
stabilized, and to use a thermometer stuck in the biodiesel sample only
(the methanol was sealed) to tell me when that temp was stable. Then I"d
do the test. I also had corks with thermometers stuck in the middle,
which I could use to actually take the temperature of the 3/27 test
itself, without evaporating significant methanol. (evaporating the methanol supposedly makes the test fail prematurely)
I was using a pipette
pump to transfer my 3 ml of biodiesel, and a large syringe to transfer
27 ml of methanol. My methanol was reagent grade, from a bladder
(meaning it couldn't evaporate easily or get contaminated with water). I
was transferring it using the syringe really quickly. I also measured
the weight of the methanol and fuel samples using a very accurate
balance accurate to .001 gram (not quite analytical balance but not the
average homebrew equipment).
While doing this I found that a pure sample of "Co-op That Shall Not Be
Named" fuel had gelled completely in the chiller. This is after I'd pumped it into my truck, and just as cold weather was coming. Uh oh. I looked at the temp
and it was turning to slush at something like 64F, which is exceedingly
high for biodiesel. Palm oil or animal fat biodiesel clouds in the
45F-52F range from what I've seen- this was much, much worse.
Incidentally it's probably really well-washed (soap tests to follow this
weekend to make sure)
I'll post my results on the other samples at my blog (girlmark.com/blog) another time- I was graphing
them all on paper and havent typed up the results yet.
I picked the 3 most interesting samples and got them tested with other
tests. We used the SAFTest, an enzymatic test that gives total glycerol
results similar to the total glycerol GC results, but does not separate
out individual MG, DG, and TG values. We also got to run one on a GC.
One Infopop homebrewer's acid-base fuel, which had started life as 27 titration(homebrew titration) slop, got an incredibly good .10 total glycerol on the SafTest (max allowable by ASTM D6751 GC method is a .24, and there's a huge difference between .10 and .24). This stuff was the fuel that did the best on 3/27 tests, by the way. The 3/27 test on it was a very clear, colorless, with no dropout- looked just like methanol with no biodiesel added to it. It also stayed clear down to 39F which is as cold as my ice-water challenge bath could go.
Another sample that failed 3/27 'barely' (ie it dropped material at a
light temp challenge of 60F, passed, with a very cloudy appearance at
our room temperature of 72F, but could have been oxidized/polymerized,
so the failing could have reflected polymerization rather than
conversion. The sample was quite old and therefore we suspected
polymerization). It tested out at .25 on the SafTest. It sorta
correlates to the 3/27 behavior, but it's not possible to tell what we
were really seeing on 3/27, because of the age (we assume that 3/27 also
tests for polymerization). I also didn't do a good job on this
particular sample when running my 3/27 tests- I had taken the 72F test
and just chilled it to
60 and stopped testing that particular vial when I got it to fail, so I
didnt' really give it long enough at 72F to see if it would have
eventually dropped out material at 72F also. I'll do so this weekend
with a fresh sample and a fresh sample of methanol. However it was
definitely not a clear pass. There is nothing wrong with the way we did
the SafTest on this, by the way, just that my 3/27 testing wasn't
comprehensive enough.
A third sample was the Goo That Shall Not Be Named. It came up as a .30
on the SafTest. It was run on a GC today but I don't have the results
yet- the tech said that from the chromatogram it looked like it was
really high in di and triglycerides with some monoglycerides, but I
won't know till next week when the tech comes back to work and
integrates the chromatogram more carefully than ChemStation does.
Interesting thing to gather from this:
Assumign soap wasnt the problem (it shake-tests really clean, nothing in
the water):
-it doesn't take much di/monoglycerides to cause a serious problem with
cold-flow properties at times (see postscript below!). At .30 total
glycerol, this particular goo turns very opaque and cloudy at about 65 F
and becomes a complete gel at 45. Incidentally,
I've seen chromatography results (from Magellan) for other 'goo-causing'
bad commercial fuel, and one of the ones I recall was a .34 total
glycerol, and clogged fuel filters all over the bay area.
I had run 15 gallons of this week's goo into my van, with nothing else
in the tank, and it caused the engine to lope. I happened to discover
the gelling problem that night in the lab, and immediately got some
diesel, so I'm now driving on Goo-50 instead of Goo-100. This morning I
had a huge cloud of black smoke on startup, and the engine is running
with some surging (filter clogging?) and a couple of times today, when
starting from a cold engine, I could smell my exhaust strongly for the
first time in months- my GMC 6.5 is recently rebuilt and has barely any
odor even on petrodiesel.
-it was really obvious from a 60F 3/27 test that the stuff had something
very wrong with it. It caused an overall clouding- very, very white,
with oily granules suspended in the methanol. I'd never seen that
before, it's not like the 'snow' you get with very soapy samples.
-However, this fuel looked cloudy-but-no-dropout (after 30 minutes) at
72F. The folks who made that fuel had done 3/27 tests and they had
called it a 'pass'- obviously, they hadn't chilled the sample.
I tend to drop my 3/27 samples (methanol vial and biodiesel sample
before mixing) into the fridge for a few minutes before testing, just to
minimize this possibility. In the absence of a fridge get a cup of ice.
To be really exact use a thermometer... we don't know yet if there's a
magic science about 68F, Andrew's recommended temperature for the test,
but my experiments at lower temps seem to back up that there's a
correlation between better conversion and ability of the test to avoid
dropout at lower temps.
More to come soon...
postscript:
Just wanted to clarify something important:
when I said this:
> >
> >
> > -it doesn't take much di/monoglycerides to cause a serious problem with
> > cold-flow properties. At .30 total glycerol, this goo turns very opaque
> > and cloudy at about 65 F and becomes a complete gel at 45. Incidentally,
> > I've seen chromatography results (from Magellan) for other 'goo-causing'
> > bad commercial fuel, and one of the ones I recall was a .34 total
> > glycerol, and clogged fuel filters all over the bay area.
> >
> > .
I do not mean that .30 total glycerol will always cause filter clogging
and gelling at brisk autumn temperatures, or that .30 total glycerol can
be always detected by the kind of extremely odd 3/27 behavior I
described (the white color of the whole vial and the granules of fat
that didn't seem to settle easily are very unusual). Mostof the time
really bad fuel just drops out a big blob of oil, with the methanol
eventually clearing.
I actually think that what the Co-op That Shall Not Be Named is
experiencing is the unusual, once-in-a-blue-moon phenomenon described
variously as crystalization of monoglycerides and other similar stuff. I
heard there was a recent AOCS paper on this phenomenon that was a
response to Randall Von Weidel and the Austrian researchers' insistance
that sterols primarily cause this filter clogging, and that new study
found that it was in fact monoglycerides, not sterols, primarily, that
causes this filter-clogging thickening of part of the biodiesel (I feel
vindicated as I've been railing about this phenomenon for years, and
have primarily found it in commercial fuel where someone cut corners...
somewhere). I believe I heard, if I understood right, that they also
found that what randall calls "near-spec" is a serious problem
(Randall says that you don't have to be far off-spec to see filter
clogging/material dropout/goo problems, which is also what I've found
upon testing various bad fuel made by commercial producers and
bootleggers in California that all caused similar problems no matter the
season).
I also suspect that the goo phenomenon can be caused by more than one
factor and that what we see is more than one kind of phenomena.
I also hope my truck continues to start in the morning, as I'd hate to
pump out 30 gallons of half-petrodiesel, half-goo.
Mark
http://www.girlmark.com/blog/index.php?p=190
Linear Mode
