Andrew,

I must admit the single slide is out of context by being presented singlely. the full presentation

He has just finished talking about the quantitave possibilties for measuring turbididty in the pHLip test, and he has moved to how sensitive the human eye is at detecting the changes in the results caused by slight changes in glycerides near ASTM.

Rick,

For reference, the free glycerin in both of our recent tests was 0.001%.

For all, my screen name login is herman (an old nickname from high school - I'm tall), but my real name is Graham.

Graham

Rick,

Also, our use of the pHLip test has been problematic for us. I like the test for its simplicity, but there's something that we don't yet understand about how to use it correctly. I'm trying to follow the instructions correctly, but what tends to happen is that we have had samples that look fantastic (as good or better than the reference) with the pHLip test after 10 minutes and still look fine after 30 minutes and an hour, but then you'd fail them if you look at the mirror interface the next day. The ASTM testing in each case has given us a "fail" on Free and Total Glycerin, when we think we've passed according to the pHLip Test I've sent comments directly to Randall about this for him to investigate this.

Graham

Graham,

I've talked a lot about this issue at this forum, and at this point, I use the 27/3 test during the process, the pHlip test at the end (but just for water, soap, and acidity), and then Gas Chromatography.

All 3 of these tests are equally important to us. The 27/3 tells us early on whether reprocessing may be necessary. The pHlip test tells us (on fuel that has already passed 27/3) whether the fuel is essentially "finished". And the G.C. (and other ASTM tests, some of which we can do, some of which we can't, yet) makes it official.

Graham,

Can give an example of the GC data from one that passed phlip, then failed ASTM for total glycerol on the GC? What is the biggest margin that one of these failed by?

Andrew

Andrew,

The following batch passed pHLip after 10 minutes and 30 minutes but failed after 24 hours.

Total Glyc 0.642%, Free Glyc 0.028%, Mono 1.25%, Di 0.989%, Tri 1.373%.

This is the widest margin that a test failed GC but passed pHLip.

A second sample also showed similar results:
Total Glyc 0.613%, Free Glyc 0.001%, Mono 1.16%, Di 1.292%, Tri 1.138%.

In none of the pHLip tests have we seen any turbidity or change in color of the cherry red phase, so we know we have done OK on separating out catalyst and soaps.

I am wondering if the pHLip test has some time sensitive aspect to its performance.

In response to Kumar, I am now adopting the same process for the use of the 27/3 test for in process analysis. We will be investing in a Near Infrared device soon that will hopefully allow us to correlate all three (NIR, pHLip, 27/3) with ASTM methods.

Herman,

Thanks for posting your GC results. It has to make one wonder about the pHlip test. I already had my doubts...

I am sorry to keep bugging you, but I am pretty sure I don't understand your results.

Part of my recent GC data looks like:

quote:

Batch 3
17% Methanol 9g/L base KOH
Total Mass Percent
Monoglycerides 0.1046
Diglycerides 0.0274
Triglycerides 0.1216
Total Bound Glycerin 0.2536
Free Glycerin 0.0012
Total Glycerin 0.2548% ASTM is .24%

In this case, the total of the MG, DG, TG, plus the Free G add up to the total glycerin.

Yours, from what I see, don't add up.

quote:

Total Glyc 0.642%, Free Glyc 0.028%, Mono 1.25%, Di 0.989%, Tri 1.373%.

Is it that your MG, DG, and TG #'s include the fatty acid portion?

Is there some other explanation?

Thanks,

Andrew

I think you're right, Andrew.

The ASTM method factors the GC results for the mass percent of the glycerin component only in the mono's, di's, and tri's only, not the mass including the FFA as the spec limits are the amounts of free and total glycerin - as you correctly pointed out to me in person last July in Golden.

BTW, I intend to give you a call in the next week or so when I get back home.

Best Regards,

Bob in Berthoud (but not at the moment)

[QUOTE]Originally posted by RickDaTech:
Thank you for posting this Herman. Your data has confirmed some of my suspisions. I don't think this test can be used to determine if a sample will meet ASTM spec. I do think it will indicate "near ASTM" which your sample was. You didn't mention the free glycerin results. So far we have not had any indication of how sensitive this test is to free glycerin.

Rick,

I've found that the methanol test does not show glycerine. I did a 25/225 test it was cloudy with no dropout after 1 hour. Then I added 25ml acid seperated glycerine. I shook it very hard and waited another hour. there was absolutely no dropout. It was done with Biodiesel and methanol made from koh and new oil 500ml/125ml/3.5g. It just becomes a little darker color with the same amount of haze in the methanol.

Logan Vilas

Graham,

Sorry, but I've been busy this week and missed your post.

I have a number of issues with the instructions that are put with the kit. They are not really complete. For instance they really don't tell you how hard to shake things up. They don't tell you to perform the test at room temperature with room temperature biodiesel. They dont tell you the test is extremely time sensitive. They don't tell you that the test is only valid 10 minutes after shaking. They dont tell you that you can't shake the test vial a second time. They don't tell you that your sample biodiesel will degrade over time in the vial and give false negatives. They don't tell you the test will work on B99 but not on B20.

From my experience with the pHLip test, the one with .642% total glyc should have easily failed the pHLip test right away. The biodiesel should have been so turbid that you could not see the mirror finish or tell if anything was on it. Do you have a sample of that biodiesel still available? If you have any left that failed GC but passed pHLip, I would love to get some of it to test further.

Logan Vilas,

Thank you for that information. I suspected as much, but never tested for it. If free glycerine dropped out on the 27/3, then it would not work on unwashed biodiesel.

Herman,

Can you confirm that explanation for why the #'s don't add up?

Did the data you got include the #'s for the glycerol portion only on mono, di & tri?

Andrew

Logan,

Free glycerin won't show up in the 27/3 test because, like biodiesel, it dissolves in methanol. But if washing is done remotely well (a lot easier than achieving full conversion), then free glycerin is not what we are worried about, right?

Kumar,

You are completely right. I started useing this test from JTF over 1 year ago. I just read this thread last night and saw no one ever posted a test on glycerine. On page 15 rick asked if someone knew wether it would show glycerine or not. So I gave it a quick test just to be positive for myself.

Logan

Andrew,

Sorry for the radio silence, I have been at the NBB Conference with no time for web access.

As per the post by Bob in Berthoud, I believe that the results that I receive for MonoG, DiG and TriG are expressed as weight % glycerides (including the mass of fatty acids), whereas your results for MonoG, DiG and TriG are converted to the equivalent amount of bound glycerin by subtracting the fatty acids. I am convinced that my results do not add up to total glycerin as it still has the weight % included of the fatty acids which are still attached to the glycerin backbone. This is confusing I'm sure, as I would have expected the results for ASTM testing to be expressed on the same basis regardless of which lab is used.

I have developed a calculation for determining the average molecular weight of the fatty acids in my vegetable oil, based on the fatty acid profile provided by our supplier. Soon we will get smart and analyze this on each batch of oil so that we have correct calculations of fatty acid average molecular weight. I use the average molecular weight of the fatty acid to then calculate the average molecular weight of MonoG, DiG and TriG. Using this data allows me to convert the weight % of MonoG, DiG and TriG to an equivalent amount of bound glycerin which I then compare to the bound glycerin I get from the lab (which is in turn calculated by subtracting Free Glycerin from Total Glycerin). While my calculation has several inaccuracies (fatty acid profile, errors in labe results, averaging of molecular weights) it gets me within 2-3% of the bound glycerin calculated based on lab results.

I also use the same information to calculate % conversion of bound glycerides to free glycerin as another measure of reaction conversion. I have calculated that the bound glycerin in an unreacted sample is 10.72 wt% (based on our fatty acid profile), so all I do is determine conversion based on how much bound glycerin remains in our reacted product.

Our latest results came back as follows:

0.001 mass% Free Glycerin
0.903 mass% Monoglyceride
0.246 mass% Diglyceride
0.026 mass% Triglyceride
0.274 mass % Tot. Glycerine

My bound glycerin is 0.273% which is the weight % of glycerin still attached to fatty acids. Using the above data and my fatty acid average molecular weight, my calculated amounts of bound glycerin attached to MonoG, DiG and TriG are 0.239, 0.038, and 0.003 respecively for a total of 0.279% (an error of 2% compared with the lab results). If you want more details of how I perform the entire calculation, I'll share it with you.

Interestingly, I have just looked at our retained sample from the 3/27 test from the above batch and I now have a few hundred miniscule (pinhead) drops of precipitate in the bottom of the jar. Enough to be visible, but not enough to coalesce together in a large mass. My conclusion is that the 3/27 test IS sensitive to monoglycerides, but only if you leave the sample to sit for a long time (up to a week). For people like me, this is too long for the purposes of daily quality control.

Rick/Kumar/Logan Vilas,
I also agree with the comments about free glycerin being soluble in methanol. I have demonstrated this on numerous occasions in our lab. The 3/27 test will not pick up free glycerin. All of the last three samples sent to the lab have been 0.001% FG indicative that our settling, centrifuge and dry wash resin are doing their job well. Glycerin is very soluble in methanol, which is the reason we are all learning to flash off our methanol in order to improve glycerin separation. Once the methanol is gone, the glycerin has nowhere to go but down.

Graham

Graham,

Do you have a sample of the biodiesel from that left? If so Why not try a 25/225 test or 100/900 test? I think with the larger volume of insoluable materials they will be more visable and possiably combine to dropout quicker then the 3/27 test. It is not a waste because you can put the entire test into your next batch of methoxide. It woun't affect the batch unless you are doing batches under 10 liters.

I have only been here a short time, but have read a large number of postings. I have not noticed anyone directly say why they are useing the smaller batch except to save a little money. But if you reuse the material from the test then there is no loss. Even if you don't reuse the money a 25/225 test would cost about 50 cents each at a cost of 4.00 a gallon for methanol. I currently pay 2.87.

Logan Vilas

quote:

Interestingly, I have just looked at our retained sample from the 3/27 test from the above batch and I now have a few hundred miniscule (pinhead) drops of precipitate in the bottom of the jar. Enough to be visible, but not enough to coalesce together in a large mass. My conclusion is that the 3/27 test IS sensitive to monoglycerides, but only if you leave the sample to sit for a long time (up to a week). For people like me, this is too long for the purposes of daily quality control.

I've seen the same thing, but I've been reluctant to conclude the late fallout to be MG. There is the possibility that it could be something else.

Results from two of my recent batches show that this test is very dependant on tempurature and that washed fuel tests better unwashed. By better I mean less fallout.

Both of these batches were would fail ASTM by a small margin. Batch 2 is slightly less fully converted.

Batch 1:
Tested .2548% glycerol by GC and is now my refference sample. Unwashed, this fuel fails with a tiny drop at 68f, but washed it passes at 68f, although somwhat cloudy. At 55f, the washed sample fails with a small drop.

This is the batch that I reported on recently that was made with 17% methanol.

Batch 2:
Unwashed, this fuel fails with a small drop at 68f, but washed it passes at 68f, although quite cloudy. At 55f, the washed sample fails with a small-medium drop.

This fuel was made with 16% methanol.

In both cases, the unwashed fuel was allowed to settle around 1 week. I waited 30 min. to read the tests.

I was surprised by this outcome, since the trace of methanol in the unwashed fuel should make it pass easier. The best explanation I have is that the washing process does remove some of the underreacted components, along with the soap. Other ideas?

Bob has reported that his fuel tests better on GC washed compared to unwashed as well.

Andrew

Thank You Andrew,

This is very interesting. The fact that solubility limits change with temperature are common knowledge. It was one of my early objections to the test back in the first couple of pages. I also noted some considerable differences between performing the test with hot (130f) biodiesel and room temp (70f) biodiesel. In that case the difference was dramatic. A difference in results between washed and unwashed is something that makes sense in theory, it's just no one has shown it yet. you wash out lots of stuff out of biodiesel and all that stuff is bound to have an effect on the test.

This test does set a standard. It's a higher standard than "did I get separation?" It is not ASTM and passing or failing is only an indication of the possiblity of passing or failing ASTM D6584. It's close enough for homebrew most of the time, but commercial guys relying on it will end up getting bitten in the backside.

Rick,

Yes, I agree. However, since I have a stable feedstock and refference samples of several batches which were GC tested, I can peg where I am pretty well.

Now I like doing the test cold, so the samples fail. That way, I can compare by how much they fail.

Andrew

Hello Rick
I do not think people really expect GC results out of a test that costs about 5 Cents.
Of course the instructions do say to perform the procedure at Room Temperature

But you are correct, as with the Phlip test which costs $5 and does not guarantee ASTM Quality biodiesel, passing the Warnqvest test does not guarantee ASTM conversion.

I feel that if a commercial manufacturer is relying on either the Warnqvest Test or the Phlip test as his sole means of determining whether he has ASTM standard fuel then his operation is not nearly up to scratch and deserves to get bitten on the backside.

As far as I know, there is only one correct testing regime for checking whether your fuel is of ASTM standard or not, and I do not believe either the Phlip test or the Warnqvest test is mentioned in any of these tests.