I tried an experiment today, which I would ask others to try, to verify it is repeatable.

I made a 1 litre test batch of biodiesel, using 22% methanol and 6 grams NaOH to oil which titrated 1.

I reacted for 1 hour at 60 C

I added 5% prewash, mixed and then IMMEDIATELY did methanol removal on a 300gram sample, without first separating the glycerol. I added the prewash in the hope that the water would inhibit reverse reaction from methyl esters back to glycerol esters.

I ran to 95 degrees C with constant stirring.

When no more methanol would boil off, I allowed the mixture to cool whilst still stirring and aerating.

My 300gram sample now weighed 252 grams.

It separated into biodiesel and glycerol layers very quickly, within a minute.

I did a 27/3 test on the biodiesel and it passed perfectly - no dropout noticed at all, crystal clear.

This is exciting - it suggests we can perform methanol distillation of the whole batch before glycerol separation, which would mean a great improvement in recovery efficiency.

Please try, please report.

If someone has access to a GC, it would be interesting to know what the total glycerol content is, to dermine how much back-reaction, if any, has occurred.

Thanks to juliena and locknut for giving me the idea to try this.

http://biodiesel.infopop.cc/eve/forums/a/tpc/f/71960555...261011391#4261011391
http://biodiesel.infopop.cc/eve/forums/a/tpc/f/44410893...751078291#1751078291

Graham,

Another break through??

How long did it take to drive off the methanol? About how long did you allow it to cool? And, approximate temperature of the ingredients used in the 2/27 test?

Thanks.

Wow - this does look very promising.... In one fell swoop, you reduce the energy required to purge the biodiesel and glycerine of methanol, plus you don't need thousands of Jolts to drop out the glycerine! Dang it Graham, if you make this process any more efficient, there'll be nothing for the operator to do!

If you post me a sample (say 10ml) of the product, I'll run it through a GC for you.

Cheers

Nick

Hi Producer

I've added some weight info to my post - started with 300grams sample, finished with 252 grams.

I used a small 500 watt immersion heater to drive off the methanol over a 25 minute period (Same kind you take in the car to boil a cup of water)

I have just run another 300ml up to 160 degrees C, had 249 grams remaining.

Both were cooled over a 1/2 hour to 45 minute period, I guess.

The 27/3 was done at around 18 C.

I added another dose of biodiesel and still get no dropout - that's 27/6. Crystal clear.

The glycerol is pretty solid.

Hi Nick - Wow! thanks! I'll get it packed and posted tomorrow.

Graham,
Good morning.
I tried this same idea several weeks back on a full batch but did not have positive results. However, based on your right up I see that I did not get the batch up to a high enough temp. I have a batch (135l) that is ready to test. It has reacted and had the 5% pre wash and has now been sitting for about 2 days at that stage because I have not been able to get back to it. Will the fact that it has been sitting so long have any effect on the outcome?

quote:

Originally posted by GrahamLaming:
When no more methanol would boil off, I allowed the mixture to cool whilst still stirring and aerating.

Is it important to continue to stire and aerate while cooling or can the mixture be pumped into a settling tank after meth recovery? Do you think the soap falls out with the gly?

Hi Bill

I'd think your 2 day old batch should be fine.

Why not re-stir it with the pump, and take off a sample and heat that to try it?

I aerated simply to get all the methanol out I could.

The 160C test was to be as brutal as I could, to see if heat would cause the reaction to back-track enough for me to detec with the 27/3 test. - I'd never use that temperature in my processor, it wouldn't handle it. 95C I could just about achieve.

I have noticed a couple of really, really small droplets in the 27/3 test of the 160C batch - but I mean really small, less than a pinhead.

I haven't done any soap tests on the biodiesel yet. I'll do that shortly.

The biodiesel is full of soap.

filtering or settling should take care of that though right?

Hi Shawn - I believe so.

But, I think the soap in my sample affected the 27/3 test.

I removed all the soap from a sample and redid the 27/3 test, which failed.

I'm going to redo the experiment and compare the 27/3 results for the batch, taking one half and processing traditionally, another half and doing meth separation all together.

The comparison will be more meaningful, I hope.

If they both pass, you can, you probably already know, re-run the 27/3 test colder to force them to fail. This will allow you to see if they are identical or not.

Andrew

OK, I re-ran the tests twice tonight on 2 separate batches and didn't get the same effect as yesterday.

I think I must have done something strange to the batch yesterday.

Today both batches behaved similarly and both 27/3 tests, before and after washing were the same for both batches, whether having methanol distilled before or after separation.

500ml starting liquid (biodiesel and glycerine combined) reduced to 440ml after distillation, of which 50ml was glycerol, 390ml biodiesel.

Interestingly, I see no discernible difference in dropout between the biodiesel distilled with glycerol present or that distilled after separation.

Nick, I'll send you unwashed samples of biodiesel from the combined distillation and from the biodiesel only distillation, for GC comparison. Seems a little more soap in the BD only distillation sample.

Neither fully passed the 27/3 test, but not far off.

quote:

Originally posted by GrahamLaming:
. . . started with 300grams sample, finished with 252 grams.
. . .I have just run another 300ml up to 160 degrees C, had 249 grams remaining.

Graham, I think your results show that distilling the methanol from the combined ME + Gly mixture causes the reaction to reverse.

The data contained in your first post was used as the basis to calculate the theoretical volatiles that would be present at the end of a complete reaction.

The calcuation was based on: 1 liter of uvo at a SG of 0.92, 220 ml of MeOH at a SG of 0.794, titration value of 1, and 6 gm/l of NaOH.

Theoretically a complete reaction would yield 923 gm of ME, 95.7 gm of Glyc, and 74 gm of XS MeOH. You added 5%, or 50 gm of water to the mixture as a prewash.

You then withdrew a 300 gm sample and evaporated the volatiles. The final weights of your two samples were 252 gm and 249 gm, showing an average loss of about 50 gm/sample.

The 300 gm sample, theoretically, would have contained about 242 gm of ME, 25 gm of Gly, 19.4 gm of XS MeOH and 11 gm of H2O. The ME and Gly are considered non-volatile and equal 267 gm. The XS and H2O are volatile and total 31 gm.

Therefore, you recovered about 19 gr more volatiles than one would expect to be recoverable.

So, assuming your numbers are accurate and your experimental procedure is good, and I have never seen less in your work, then what can explain the extra volatiles? (Of course the first thing to verify is my math.)

Assume all the NaOH was consumed by making soap. You did say you had a lot of soap in your sample. 6 gm (6/40 moles) of NaOH would convert FFA to soap and produce a by product of 6/40 moles of H2O or 2.7 gms of H2O from 1 liter of wvo. This would only amount to about 1 gm (2.7gm x 0.3) more water in your 300 gm sample. Another 18 gm of volatiles must still be "found".

The reverse reaction to form MG is:

(1)Gly + (1)ME <- catalyst-> (1)MG + (1) MeOH.

The conversion of 0.45 moles (18 gm/ 40 gm/mole for MeOH) of ME would release 18 gms of MeOH and could account for the extra volatiles you observed. (In actuality some extra DG and maybe some TG would be formed from the ME as all three reverse reactions are possible.)

These calculations are only approximate, but they do describe one mechanism to explain your results.

The GC results will help to provide more information. If this analysis is correct, then your GC results will probably show a much higher percentage of MG and to a lesser extent DG, than one would observe in a sample that failed because of underconversion instead of failure due to a reverse reaction.

Hi Producer

What you say sounds exactly right.

I'm doing a new batch right now with virgin canola oil, because my WVO has been variable in water content and I seemed to make more soap than I expected in those test batches.

I want a good pass on the traditional biodiesel, to compare with the combined distilled process result.

I'll post the results of the virgin oil later.

(PS - this is a strange experience using 'bought' oil. I feel I have sinned )

I have just distilled methanol from 2 samples of combined biodiesel and glycerol.

I took methanol from one batch without 5% water added, and from another with 5% prewash.

The difference is remarkable.

With no water, no glycerol drops out after distillation.

With water, there is a rapid glyceriol dropout.

Graham
I was only able to get the temp up to 170F which is 20F short of the goal of 190F. However, it did seem to be a bit clearer than normal. I drained off the gly and moved the remaining 130L of unwashed BD to a wash/dry tank. It was still "steaming" a little and had a strong acrid smell.
ME recovery did not do well at all. Got only an ounce or 2 of brown liquid. Got a lot of white vapor coming out of the condensor.

Have to take the boys to soccer, more later.

Bill

quote:

Originally posted by GrahamLaming:
The difference is remarkable.

With no water, no glycerol drops out after distillation.

With water, there is a rapid glyceriol dropout.

Something is causing the difference.

Even though the results are not as we would have wanted, let's assume the presence of the 5% water prewash did limit the effects of the reverse reaction.

Now, let's assume with no water to limit the reverse reaction the "non-prewashed" batch produced much, much more MG and DG than the prewash batch.

Glycerol is insoluble in Methyl Ester. DGs and MGs having limited solubility in both ME and Glc may act as a cosolvent.

If this is the case, then the glycerol, ME, DG, and MG would all stay in solution with each other and nothing would settle out.

Just a theory.

quote:

Originally posted by GrahamLaming:
I have just distilled methanol from 2 samples of combined biodiesel and glycerol.

I took methanol from one batch without 5% water added, and from another with 5% prewash.

The difference is remarkable.

With no water, no glycerol drops out after distillation.

With water, there is a rapid glyceriol dropout.

...which suggests that your theory that the 5% water prewash inhibits the reverse reaction is correct. It would be interesting to run two more batches without distillation, one with 5% water prewash and the other without. Then carefully water wash and dry all four samples. The yeild should give you and indication of reaction completion since you will lose some to emulsion if mono and diglycerides are present. Be interested to see the 3/27 results on the samples too. The GC will confirm but the above should give us a guide in the meantime.

Cheers

Nick

There has been some droput, 10 hours later, but the biodiesel is much darker than the combined water batch, suggesting suspended glycerol is still present.

The soap has dropped from the watered batch - it is now perfectly clear and first measurements with bromophenol blue show less than 100ppm soap.

This is much faster than I see with biodiesel made from tallows and partially hydrogenated oils. I suspect the lower density of canola-based biodiesel allows the soaps to settle out faster.

I therefore won't water-wash the sample for the GC - I'd rather leave all MG DG and TG in the sample.

Hi Graham,

Nice to see your excellent post mate.

I'm up north in Bingley and am just starting off biodiesel processing. I wanted to know whether you have a reasonably priced methanol recovery unit available for sale as it seems a bit complicated for me to make! If so please contact me via is_auii@yahoo.co.uk. I look forward to your reply.