I've been working on ethyl esters the last few days for a company in Fiji, which makes it's own ethanol using solar heat and other great things (like solar-dehydrated zeolite). I've been pretty successful with it using glycerine from methyl batches to precipitate ethyl ester reaction glycerine, and sometimes using water in combination with glycerol (I even made two-stage waste oil ethyl esters, which I'd assumed is a hard thing to do. Glycerine is amazing stuff).
Anyway, I realized I don't actually know what the 30% ethanol recommendation is based on. I know it's all about molar ratios and excess and so forth, but I don't know what it "should" be. I have limited internet time to research this for the rest of this week.
What does 'the literature' say about ideal ethanol ratios?
The factors at play are:
-enough excess alcohol to push the reaction to completion
-not too much alcohol so as to allow the glycerine to precipitate.
Since I seem to have taken care of problem #2 by using glycerine from prior batches, what's the theoretical correct amount of excess needed for issue #1?
MarkThis message has been edited. Last edited by: Shaun,
1.44 times the Methanol amount (I searched under "neutral", as author, and "ethanol").
Sorry, that's the theoretical quantity. I'm looking for actual experiences, as with ethanol there is more than one factor going on (ie solubility issues, and the water problem) that affects the result.
Anyone know what the published scientific literature says about this? Do they just go with the same conversion as Neutral's using here, or has research found something different is needed?
I think I didn't make myself clear- I know that 30% is about 1.4 times more than 22%. However, we arrived at 22% kinda empirically- in terms of molar ratios and excess, is that still the correct one for ethanol? Ethanol introduces some other problems into the picture, so I feel that empirically we're going to find out something different is ideal. Not sure if that's true, just a gut feeling based on the additional variables.
(See edits, I screwed up on the molecular weight of a triglyceride the first time - thanks for the reality check, dropout)
This may already be in everyone's head, but just to be clear:
22% by volume as expressed in molar ratios is:
100ml oil = .1 moles (molecular weight of a triglyceride = molecular weight of an average free fatty acid times 3, so 280 x 3, plus molecular weight of glycerin which is 92. So, 280 x 3 + 92 = 932 grams/mole for triglyceride. Also, average triglyceride density is about .92 g/cm3)
22ml methanol - .54 moles (given 32.04 as molecular weight, density = .7918 g/cm3)
.54 moles meth. / .1 moles oil = 5.51 molar ratio
Using that same molar ratio of 5.51, you end up using 31.74% ethanol (see attached for specifics, ick, I can't attach spreadsheets???? ).
The real useful question is, is that ratio the same in order to make good bio using ethanol?... I have done a number of minibatch tests with 95% ethanol 5% methanol and found that they all separated relatively quickly and easily - however, they were pretty underconverted (the first lab wash made 90% mush, 10% bio). My tests were done with chicken fat of low ffa (homebrew titration of 3). It would be interesting to get a graph of reaction rates and settle times using various concentrations of ethanol/methanol (95/5, 90/10, 80/20, etc). We may find that adding even a small amount of methanol might improve reaction rates and settle times, and reduce the likelihood of emulsions. Any word on that?
It seems like SOP is: 1) get good conversion on your system using ethanol, 2) if you have emulsion problems, try water or water/glycerin solution, 3) if emulsion problems are too much to handle, add methanol with your ethanol to the point of reducing those problems.
ethanol_calc.txt (311 Bytes, 51 downloads)
In a conversation with a prof from Brazil who does a lot of biodiesel research I was given to understand that in Brazil they use 9:1 to 12:1 molar ratios for ethanol. I will email him and report back.
friend, sweetie, honey, dearest (I really do call him those things in real life, by the way, I"m not 'just' begging here)-
any chance I could convince you to run a couple of test batches with varying molar ratios and then do some SafTests on whatever you can lab-wash of the result?
mmm... I'm going to call it begging, but no need to split hairs -
We'll see what I can do this weekend, I'm thinking 32% ethanol with a base of 9, 11, and 13 g/L WVO titration 3.5 or less, 130F water bath, 20 min intermittent shaking - no guarantees but I'll try?
Rachel gets to decide how you pay back the lab for the saftests (ouch).
OK, here's something from another thread:
TurbinePowered, who uses E-85 to process his ethyl esters, said that he can use 22% total alcohol (ie more E-85, amount adjusted so that ethanol content is 22% just like with methanol).
His batches wash well, which is the only way (with ethyl esters)that we have to view conversion and not a very good way since poorly converted oil doesn't always emulsify in the wash and well converted fuel can be high-soap, high-emulsions.
Here's something that occurred to me:
I'm becoming obsessed with the idea that co-solvency is an important factor in two-stage acid-base process (ie some folks, like the BioPro people, as well as Appleseed two-stage acid-base users, seem to see VERY high conversion biodiesel when doing the acid esterification process first, and I am starting to think it's because of the cosolvent effect of the esters formed in the first stage that then affect the mixibility of methanol/oil in the second stage). If I'm right, then the cosolvent effect of ethyl esters should be greater. So maybe TurbinePowered can get away with what is a theoretically lower excess of alcohol because the cosolvent effect of the ethanol on the glycerine/ethanol/unreacted oil mixture is so huge that it overcomes the need for so much excess. It's like having better mixing in a reactor.
In Dr Pepper batches of ethyl esters, you don't really see anything separating and can get away with shaking the bottle fewer times over the course of an hour, and that's got to be good for the ability of the alcohol to contact the oil.
The Saftests are cheap, the GC tests I can't afford yet, I"ll wait till the college has the chemicals to do that (ie months)
thank you darling.
But... can you vary the ethanol as well/instead? It's easier for us to test the effects of various catalyst quantities (ie soap production).
Maybe go on the safe side and seriously overdose the catalyst, yield be damned. To get the stuff to separate, dump some glycerine on top. It might contaminate your minibatch with methyl esters, but that shouldn't matter.
Try 250 ml glycerine into 1.3 (or wahtever) liter of unseparated 'biodiesel'
Ok, I'll do a constant base of 10, and a molar ratio of 1.73:1 (equivalent to 22% meth, or 31.74% eth), 1.89:1 (equivalent to 24% meth or 34.63% eth), and 2.2:1 (equivalent to 28% meth or 40.4% eth).
I'll get it to settle, no worries -
I wish ethanol testing for US production was more useful - for all my searching I haven't found anything that we can buy which isn't a) highly regulated and difficult to buy and b) way overpriced. Fuel grade ethanol is a buck something a gallon, but you can't use the stuff for biodiesel production, you can only use it for direct, on road use. And blends of methanol/ethanol are denatured (therefore not considered drinkable and taxed differently) but use industrial grade ethanol which is much more expensive. When will someone provide biodiesel producers with low cost ethanol/methanol blends???? ARRGH!
Done ranting, sorry.
So Caustic, if I read your previous post correctly, you were showing an ethanol percentage of 34.63% (or 35% rounded up) as the amount to have a complete transesterfication?
Illegitimi Non Carborundum
Well, I guess I'm just saying that if you want to use the same molar ratio that we typically use for methanol (which, at 22% by volume is 1.73:1 alcohol to oil as I calculated it), then you should be using 31.74% by volume ethanol. If there is a way to post a scalc spreadsheet I'd put it up so you can see the calculations, but I think know if I can post that type of file.
Obviously, questions still remain if that will be good enough, or if we need more catalyst, more mixing, forced settling, etc. in order to hit ASTM spec.
Well, I did 3 minibatches with 31%, 34%, and 40% ethanol with a base catalyst of 10 using waste oil that titrated at 7ish (FFA% of 3.85). The ethanol moisture tested at 1700ppm. After making the minibatches, none of them separated, so I added some water to force separation, which worked fine. However, upon lab washing in preparation for the Saftest I (actually, Link) found that it emulsified so quickly and easily that most of the sample was lost to emulsion on the first and second washes. Only the one with the 40% ethanol had enough non-emulsified sample after lab washing to test. Because they were obvious failures (if you lab wash and get huge emulsions, that's usually because of monos and dis, not as much because of soap), I only actually Saftested the 40% ethanol sample, which failed miserably at over 1 (spec is .24).
On the question of whether Saftest works for ethyl esters, my guess is yes, but I don't know for sure. Bob has Saftested ethyl esters and gotten what he considered reasonable results, but did not do third party testing to check.
My guess is that I did not dissolve the catalyst sufficiently in the ethanol. I shook it for quite a long time, but it's possible that it was not long enough... I can't imagine what else I could have done wrong - besides varying ethanol amounts, I used 250ml oil, 4.28g catalyst, 20 minutes of shaking at 130 degrees (in water bath). A good test would be to use methylate, and see if I'd get a better results.
I have done similar tests in the past with 95% ethanol 5% methanol and I feel like I've gotten significantly better results. Any suggestions as to why I got such a bad failure after overkilling both the alcohol % and catalyst would be appreciated.
When ever I tried the molar correct amount of ethanol, the batches turned out to be total failures. I know that ethanol is a very strong solvent and I am sure that has something to do with it. I know it defies logic, buy you might want to try 22%, maybe vary the base amounts of caustic. Whenever I try to go by the book with ethanol, even with clean oil, I fail miserably. Right now, I'm doing 22% with 70% E85 and 30% methanol. I'm using a base of 9-10g/L and making great fuel.
Try using no methanol and see what happens - my guess is your low amount is based more on the fact that you're using a percentage of the alcohol as methanol than on the fact that less ethanol is better.
I have used E85 only in jar tests and it does work. I use E85 because I have no access to 200 proof ethanol. I mix in 30% methanol in 35 gallon batches to provide some security against unforeseen variances in the E85, which is pretty common when dealing with ethanol. When I prefect the caustic amounts using strictly E85, then I will make the plunge on the full batches.
Most people swear that you must use the correct molar amount, but factoring in the solvency of ethanol plus the fact that our use of 22% methanol is excessive in order to swing the reaction to one side will most likely result in an excess of ethanol even if we use only 22%...again the solvency probably factors in. If you use 30+ percent of ethanol, the excess will probably inhibit separation. The following are unproven theories of mine, but I believe that the excessive ethanol will impede a biodiesel reaction due to it's solvency. I suspect that the extra ethanol will either quickly emulsify with the unreacted oil, preventing it from joining up with the ethoxides or that ethanol might impede the reaction by emulsifying with the broken off FFA before it has a chance to chemically join up with the ethoxide. I'm open minded enough to be corrected, but 22% ethanol works for me in the jar tests.
I had a major "Eureka!" moment tonight thinking about this thread, and while I might be offbase, here's my thinking about what happened to Caustic's test batches:
He's using SafTest to test conversion.
SafTest is an enzymatic test which gives results in terms of total glycerol: this means the number it spits out includes the glycerol portion of tri, di, mono, AND free glycerol (I think).
A gas chromatograph running the ASTM 6584 test gives several separate numbers:
'free glycerol' and a list of the individual percentages of tris, dis, and monos, and then, confusingly, a calculated count of the glycerol portion of all tris, dis, and monos, and free glycerine all combined. The number that the SafTest provides should be the same as the GC free and total glycerine number (or it should be close, depending on the interpreting of the GC chromatogram).
If I understand right, the SafTest works by converting di,tri, etc to a free glycerol, then reacting the mixture with something that gives a purple color whose intensity varies based on the amount of free glycerine that's THEN present, then using a machine to read the exact color of purple and to calibrate that to a percentage glycerol. It does not differentiate between free and total glycerol. So I think if you start with high free glycerine biodiesel you'll end up with a higher number out of the SafTest.
I just did soap tests in Fiji on my two-stage ethyl esters and found FIVE TIMES the soap content in my batches that had settled for a few days versus the same exact oil when reacted with methanol .
The reason for this is of course that ethanol is a great cosolvent that keeps glycerine droplets from settling, and soap along with them. I had no way to test the glycerine content but I'm pretty sure it'll be abnormally high.
I have noticed that when I have had emulsions with biodiesel, it takes a larger-than-usual amount of water to wash out the soap in the biodiesel.
Ordinarily Piedmont Biofuels, for whom Caustic works, are really diligent about soap-testing their production biodiesel. However, I wonder if Caustic did soap tests while he washed for his lab batch, as there is no reason for him to do so- there's no reason to think that meeting a specific soap standard should matter in the SafTest or a GC test.
What I think could have happened to Greg's batches is:
-that he perhaps could have had under-washed fuel at the time he tested the stuff.
-I also am absolutely certain (because of the color of some of my ethyl esters wash water) that along with high soap levels, ethyl esters has a higher amount of glycerine suspended in the ester layer than methyl esters at the same stage of settling.
-The SafTest doesn't make distinctions between free glycerine and bound glycerine.
-perhaps the SafTest machine was turning out unusually high total glycerine counts because it was also reading a lot of unwashed free glycerine that had come from the sample.
-normally, free glycerine is fairly easy to wash out, in fact, homebrew fuel is usually extremely low in free glycerine compared to commercial fuel
-commercial producers sometimes have higher free glycerine counts to worry about when trying to pass ASTM. I have some suspicions that this is because:
-they sometimes wash after a shorter settling time than homebrewers
-then it apparently becomes harder to get all of the free glycerine out of the biodiesel, compared to homebrew fuel which practically NEVER has a free glycerine problem
-perhaps that situation is analagous to what Greg's been dealing with when trying to lab-wash a batch of ethyl esters which started with a higher-than-normal free glycerine count
-I already know that Dr-Pepper-style 'lab washes' aren't as effective as a good mistwash or bubblewash or circulated wash that involves breaking up the water into smaller droplets to increase surface area
-this also argues for my idea that he may have tested underwashed fuel
I'm sure the SafTest inventors considered this scenario for methyl esters, but I wonder if our ethanol-based fuel just falls outside the range of the 'normal' that they were designing for (my 5X soap counts certainly did!) and therefore requires much more washing before using the SafTest.
hmmm... so -
- ethanol bio has more free glycerin in it because it's a good cosolvent.
- ethanol bio also has more soap in it for the same reason
- therefore, you're more likely to make an emulsion (because of the high soaps), and more likely to underwash (because of the emulsion), and since you're starting with high glycerin anyway it might be easy to not get all the free glycerin out.
Well, I've done a crapload of labwashes on a variety of different fuels. I know how methanol bio lab washes, and I know that emulsions almost never happen with ASTM spec fuel, and they almost certainly do happen with off spec fuel... in fact, I can pretty closely predict the Saftest result now just based on how things lab wash. Even very soapy (high titration oil) bio will lab wash normally (no big emulsions) if it's well converted.
So in all my other experience, large emulsions means failure i.e. off spec fuel - and this ethanol batch was a large (almost complete) emulsion on every lab wash. We only had a tiny amount of sample left after 5 or 6 washes to be able to test.
Also, glycerin very much prefers to dissolve in water than bio. That we know because most of the time on ethanol minibatches, we have to add water just to get the stuff to separate. Even soap (which is probably more likely to stick in the bio layer) comes down significantly from just the first wash. Using chicken fat, from wash 1 - 4, soaps will go from 12,000ppm - 800ppm - 100ppm - 50ppm. Glycerin is probably an even more pronounced drop over the first few washes, so the reality is the vast majority of the glycerin is leaving with the first couple washes - and even if there's a lot of it, within 3 or 4 washes it's probably mostly gone.
Honestly though, I've never made an ethanol batch that didn't lab wash with large emulsions (and I have made them prior to this), so you're statement could still hold true.
The easiest test for this is to GC the sample.
Another good, though not perfect, test for what I think I did wrong is to try using potassium methylate as the catalyst to see if I really just didn't adequately dissolve my KOH in my ethanol. It adds an extra factor (methanol), but it assures that the catalyst is fully dissolved.
|Powered by Social Strata||Page 1 2 3|