This is a refinement that we have developed on the pm discussions and I would like to share it more widely.
First a bit of background. In a lecture on biodiesel processing at last years conference, the speaker mentioned that glycerol actually helps the transesterification reaction in the early stages, only becoming a hindrance in the later stages above 60% conversion. This got me interested and I began to look into it.
My theory of why this may be true is that when the catalyst (KOH) is added to the WVO the first and fastest reaction that occurs is neutralization of the FFAs. This produces soap and water.
The presence of this water ( along with any water in the WVO) creates conditions that are more favourable to saponification (oil + water + KOH = soap) than transesterification ( oil + methanol + KOH = biodiesel + glycerol). Only when the transesterification produces a significant amount of glycerol do the conditions change. The glycerol bonds to the available water suppressing further saponification.
The idea behind the Dry Start method is to make a small amount of glycerol available right from the start of processing so that saponification is reduced.
I prepared 2 litres of glycerol by first demething it and then drying it completely by heating it to 120°C until no bubbling was visible.
My normal supply of WVO is normally very high titration (25ml KOH +) but for these tests I blended some very good quality oil provided by Trigger to get the level of FFAs I wanted.
With my normal process I can successfully process 150litres of oil with a titration of up to 7mlKOH in 2 stages. total KOH 8 + titration grams per litre, total 16% methanol, first stage 80%1.5 hours, 2nd stage 20% 1hr.
If the oil is titration 7ml I get zero dropout on the 30/270 test.
If the oil is titration 8ml I will get 0.2 dropout
If the oil is titration 9ml I will get 1.2ml dropout and will require a 3rd stage.
I have now done 4 batches where just before I start processing I pour 2 litres of dried glycerol into the processor. Otherwise I process as normal.
batch no 1 Titration 7ml. 30/270 result 0 dropout.
Batch no 2 Titration 8ml. 30/270 result 0 dropout
Batch no 3 Titration 9ml. 30/270 result 0.1ml dropout
Batch no 4 titration 10ml. 30/270 result 0.5ml dropout.
All the batches were lower in soap than normal, passing the wash test in under 3 hours of my drywash rather than the normal 6.
I think these results are very encouraging. I chose to test using highish titration oil because the neutralization of the FFAs would produce more water and soap but I believe this technique can benefit any process. Its very easy, I dry 10 litres of glycerol in about half an hour and thats enough for 5 batches. I store them in 2 litre bottles until they are needed.
It would be interesting to hear of other brewers experience with this method.
I think if you just increase your methanol to the standard 20~22% you will find that all 3 of your samples from 7 to 9 KOH will pass in the first stage.
I have processed 11KOH oil in one stage on 2 different occasions with a 3/27 pass, just dont ask about my yield...
This is what prompted me to start with AE.
I wonder if the lower soap in your method is due to some of it getting bound with the glycerol, those of us that WBD also see way lower soap numbers than standard processing due to the soap getting bound up in the glycerin.
Simple schematic for a pump and heater control with a high limit
Sensor for the biodiesel/glycerin layer
This wasnt meant to be a discussion of my particular processing method. I have developed my process using extra processing time and minimising methanol because it costs less. Im well aware that using 20 or even 22% methanol will make it faster and easier.
The point of the experiment was to show that the addition of dry glycerol improved my process WITHOUT changing anything else. What Im hoping to find is that other homebrewers with different processes to mine will get a similar improvement using this technique.
The lower soap levels may well be connected to the extra 10% glycerol present. My theory about why this works is just that, a theory.
I do think the reason the 'dry start' method seems to work is because of the water absorbance properties of the added glycerol.
Possibly similar to the AAF method which has been proven to work well because (I think) glycerol is produced much earlier in the reaction, thus reducing saponification.
The big advantage of 'dry start' is that it can be used with ones processor system 'as is'. AAF requires a separate heating tank.
Simple schematic for a pump and heater control with a high limit
Sensor for the biodiesel/glycerin layer
Mr imakebiodiesel:I read almost all your notes, u are really good chemist even you are not.
My first question for you is; how are you demething raw glycerol? and the second question is;
how to perform the wash test?
Mr imakebiodiesel does not now post very often on here as he has other interests.
As per his 1st post he heats the glycerol to 120 degs C to evaporate the methanol and water.
is it gonna work with glycerol which have not been demethed and dried.I mean what happens if we just put fresh , even warm glycerol to another batch?Just trying to make a shortcut.Maybe is not necessary to demeth and dry glycerol for dry start. Any experiences?
Yes. This is called a "glycerine pre-wash" or "glycerine pretreatment"
You perform a first initial reaction using only glycerine that has not been de-methed.
People report this results in reduced titration of the WVO and a partial transesterification reaction occurring.
It also frequently releases biodiesel that is bound up in the glycerine layer and adds that to the WVO being processed.
Because you are performing a multi-stage reaction with some of the required chemicals already in the glycerine, this results in being able to reduce the amount of "new" chemicals you will need to use in the reaction compared to a single stage reaction.
Just a slight modification on what Mr Tilly has written.
The dry start method is different to a glycerol pre-treatment in that the 2 litres or so of glycerol that you add to the oil before the methoxide stays in the mix until removed with the 'new' glycerol formed from that reaction. The idea being that as the free fatty acids contained in the oil are neutralised the water formed in that reaction is absorbed in the added glycerol and as such less soaps are formed.
You can indeed use ex process glycerol without drying for this and it does work.
Hi Mr Dgs,
as I read bbarclay's post I did not think he was talking about the "dry start method"
EDIT After reading through the dry wash method again, I am not sure what it does?
I do not know what IMB's "normal" process is.This message has been edited. Last edited by: Tilly,
Tilly, You asked the question that I asked myself. How does the addition of "dry" glycerin byproduct improve processing? Obviously, it can't be due to the extra methanol because drying at 120 deg C until it stops bubbling effectively removes both the water and the methanol.
Second, under the anhydrous conditions any residual NaOH or Na-methoxide would like convert to Na-glycolate (Na-O-CH2-CHOH-CH2OH), which as indicated would rapidly react with any free fatty acids (HO-CO-R) to form soaps Na-O-CO-R and glycerin. This would increase the effective concentration of the base (Na+ or K+) but how much given the modest amount of "dry" glycerin added in comparison to the batch size.
Third, you might imagine that the "dry" glycerin might act as a processing aid early in the reaction cycle either by providing a second phase for the reaction products & soaps to move to or allowing the methoxide & methanol to mix with the unreacted oil but I've never noticed much problem mixing the oil and methanol-methoxide solution.
There is one possible mechanism that might explain the beneficial of "dry" glycerin. We know that glycerin, methanol & soap mixtures are not very soluble in either the unreacted oil or the later biodiesel-oil mixtures. In fact the separation glycerin soap byproduct from the bio-oil phase can be observed almost immediately. You might imagine that if the batch mixing isn't too efficient the glycerin/soap phase might sequester the residual water and newly created soaps separated by a interface. If the diffusion of water back through the interface was slow enough it might improve the reaction.
This raises a question. My reactor (and most others) drains from the bottom of the reaction tank into a pump and recirculates to the top. This has the effect of constantly mixing the settled out glycerin (& soaps & methanol) back through the batch.
What would happen if glycerin layer was allowed to simply settle and collect below the batch and any mixing or agitation be confined to the upper phase only?
You could imagine achieving a similar outcome by mixing in some of the methoxide-methanol mixture then allowing a certain percentage of glycerin layer to accumulate and then remove it (from the bottom) before adding the remaining catalyst & methanol.
I can easily try this with my current processor.
Hi Rick G,
My question was even more basic than that.
I did not see any meaningful comparison testing being performed.
I do not know what, if anything, was accomplished.
Did he perform 4 or 8 reactions all together?
If he was doing 150 litre batches of biodiesel the oil in each reaction would not have been identical so there is no meaningful comparison between each batch that I can see.
That has been done. Removing glycerine during the reaction does not work because the methoxide is also in the glycerine and you are removing chemicals from the reactor before the reaction is finished.
Unfortunatly the study that did this has been removed from the iowa website.
This is from the paper:
One of the objectives of the study was to
[i]”Set up a recirculating cyclonic separator to investigate potentially increased mixing during transesterification reactions and the effect on product quality. The addition of the cyclonic separator allowed a preliminary exploration into possibly enhanced mixing action and real time glycerin removal during the transesterification process.”
In the conclussion of part of the thesis they say
Separating a heavy phase stream from the transesterification process at steady state
causes decreased reaction rates. All tests took longer to reach acceptable total glycerol
levels of 0.15% as compared to reactions at the same temperature without separation.
Some of the reactions using the separator never reached completion and those that did
had higher total glycerol levels than reactions performed at the same temperature without
It is available on google
I'm curious so my last batch was a dry start. What I noticed is that the process seemed to take longer and my whole batch demeth stage took significantly longer. It seems that the dry glycerol absorbs a significant proportion of the methanol.
Glycerol and methanol are miscible in all proportions.
Thanks Tilly. My point really is that needed excess methanol may migrate to the added glycerol and pointing out that less methanol is recovered for a given distillation method (in my case not exceeding the highest batch distillation temperature that I cut off at)
I'll try to describe a chemical concept. Glycerine hydrogen bonds with methanol. There is a chemical force holding the two molecules together. Describing it in a vodoo chemistry manner, methanol doesn't want to separate from glycerine, they're holding to each other. There are types of pure molecules that don't stick together with some degree of force, like hydrocarbons (kerosene, naptha et cetera) the attraction between molecules is small, so their boiling point is much lower than if there was a stronger atraction between molecules. Since hydrogen bonding occurs between glycerine and methanol, the boiling point is probably elevated from what it would be if plain methyl biodiesel and methanol were boiled together. Anyway I wrote that to try to justify a suggestion for your whole batch demething. You might use a mechanical vacuum pump pulling some level of vacuum on the reaction product in your processor, while you distill out methanol in whole batch demething. The methanol would come out at a lower temperature. But an extreme vacuum might implode your processor, or some part of your processor system. It sure will collapse some types of plastic tubing. Thanks
|Powered by Social Strata|