cold reactions or Drying Potassium methoxide.

[dgs]

Happy New Year to you all.

OK, I've now completed the first process of road fuel with the lime dried methoxide.

I used 200 litres of my best oil, it had been glyc washed (no enhancement used for this) and a 10/90 test showed 9.5mls dropout. I'm not sure what this actually means as I have done many 10/90's on untreated oil and many times had a result of less than 10mls.

I ended up doing 3 reactions to get a clear 10/90 pass. Starting temperature was about 32degs and end temperature just under 40degs (the pump heats up the mix without the immersion)

Total methanol used 24litres. Total methoxide volume nearer to 25 litres. Total KOH used 1540gms (7.7gms/litre)

Glycerol extracted 16 litres. I don't think there is any doubt the reduction in glycerol volume is because of the vast reduction in soap level.

I have now finished de-mething the batch and I am left with 204 litres before water washing. There is still some lime in suspension, so I will have to see how the washing goes.
Although the KOH used isn't excessive it isn't particularly low, it is certainly higher than I thought it would be, however the process has used a minimum of methanol, 12% (the expensive part)

After looking at the cubies of methoxide I mixed for the last batch, about 80% of the bottom settled layer of lime has gone as solid as a brick, presumably the part that has reacted with the water, now Ca(OH)2
After seeing this I decided not to use the small 60 litre tank to mix the methoxide, I had visions of trying to remove the solid lime out of the cone. So I will continue to mix in cubies and discard when the solid layer gets too deep.

The quick lime I bought was £10 from a local company and this will dry the methoxide for at least 5 processes, so it adds 1p/litre to the bio cost. Having said that, as there seems to be an increase in yield there is probably a net gain.

I am going to set up my smaller 120 litre processor for the next 2 batches and use the same oil from the storage tank. I will do 2x identical processes, one with the dried methoxide and one with normal. This should give a better comparison of exactly how much better the process works.

[Glycer-rides]

Excellent and interesting work, Dave.

[dgs]

Thanks, Glycer-rides,
David Shinn is also trying this method and he will make a post on here in due course.

[Chug]

was the soap ppm level similar to 1700 before?

and 21 ltrs glyc last time with 2 stage process, 16 ltrs this time with 3 stage process, I don't think a third stage would have produced 3 litres glyc on the previous batch do you, so showing the difference between good oil against poor oil

[dgs]

I think you are correct Chug.

Regarding the soap levels, I have done literally hundreds of soap tests over the years and this latest batch was very high at >5000ppm after the last reaction and today 3200ppm. But I don't think it's soap that I am testing. I think the alkalinity I am titrating is due to the suspended lime in the bio.

The reason I say that is when I de-methed I usually get about 6 to 7 litres of glycerol/soap falling after a few hours. With this batch I got 1 ish litres.

Also David Shinn (we PM each other all the time re this process) has reported that he has had the same minimum amount of glycerol fall after de-mething and is in the process of pump washing with no emulsions.

I think David was probably a little more carefull to reduce the amount of lime added with the methoxide than I was. It can be avoided by having enough methoxide mixed so that the bottom layer isn't poured in. I assume this dried methoxide will 'keep' for a few weeks.

Julian asked could the lime be used for drying the finished bio. Whilst there is little doubt it does dry it very effectively if the lime was left to settle it would take a long time. Or a centrifuge maybe.

I am just hoping that the water washing removes all traces of the lime.

[DavidShinn]

Thanks, Glycer-rides,
David Shinn is also trying this method and he will make a post on here in due course.

Hello!

I am a friend of dgs from Leeds, and as he says, I will post my own results when I have finished pump washing and drying my batch. it will be two or three days, as I am now back at work.

Best wishes,

David

[Chug]

Welcome David, are you doing similar sized barches to Dave? 

I tried some experiments with quicklime a few years back but only 25 L batch sizes, hats off to Dave for going in on full size batches.

[countrypaul]

Another possibility if there is lime left in the mix, is that the soap being formed is calcium soap which tends to be very insoluble and would precipitate out much more easily than potassium or even sodium soap. If this happened, then you might be left with much less soap/glycerol present than normally expected - which could be a partial explanation for why you seem to get so little glycerol. Also calcium soap due to its insolubility is much less likely to act as an emulsifier.

Another factor to bear in mind is that calcium is bivalent, so whilst one mole of quicklime weighs about 74g it will require 2 moles of acid to neutralise it - I don't know how you are testing and calculating the soap content, but this could cause significant discrepancy. For example, if you are relying on a mw of 39 for potassium but actually have calcium (mw 40) then you will be nearly a factor of 2 out (due to calcium being bivalent).

Paul

[dgs]

Thanks for that Paul.

I calculate by multiplying the titration by 320 so does this mean the soap (if it is calcium soap) is about half of the amount calculated.

I will put up a full post when I have totally finished the batch, but after drying it is still cloudy at the moment. For the last 2 hours of the drying cycle I added some magnesol to hopefully remove anything in suspension. I know magnesol can be a difficult thing to remove but when it is all settled (i will leave it 2 days) Then I will centrifuge it and hopefully end up with a clear product.

The water washing went really well, hardly any soapy water and no mono's at all.

[dgs]

I have now 'sort of' completed the batch. I'm afraid to say that things haven't gone well. I will explain.

The water wash went "seemingly" very well. Even after the 1st wash there was hardly any soapy water. I washed 4 times and on the 2nd wash the water was clear. I dried the batch and noticed before heating that there was something suspended in the bio. After drying the bio looked the same ( like bio with lots of suspended water) I centrifuged 70 litres of it and although the centrifuge caught what looked like caramel coloured soap/glycerol it did not separate it hardly at all. As the density of this contaminant seems to be the same as bio the centrifuge won't work.

It seems that whatever it is, it is not water soluble and has virtually the same density as bio.
I did a small test filtering it through woodchips and although not clear the biodiesel clarity has considerably improved.

At the moment it is all (70lts) filtering through some fresh oak chips and it seems as though it is now clear.
I intend to keep filtering it until the soap test is in spec. I've not checked it yet.

So, some questions. What is this contaminant, is it as Paul suggests and is calcium derived soap/glycerol which seems difficult to remove. Why did it form in the 1st place, was it me not being careful enough so pouring some lime in with the methoxide, or just adding too much lime in the 1st place. Any thoughts would be appreciated.

I think David, although having problems is having more success than myself.

[dgs]

After filtering it twice through oak chips (once and then once through another tub with new) it has come out OK. In fact it is beautiful and clear. Soap test is zero, so I'm sure it has also removed any glycerol present.

The chips seem to be the only thing to get rid of the suspended calcium soaps as I have found water washing, magnesol, centrifuging doesn't work.

To try and get to the bottom of this I have today finished processing a new batch. I used (roughly) 25% of the lime dried potassium methoxide and 75% ASM.  I was careful not to tip any lime out of the dried methoxide cubie. The last 10/90 showed a clear pass, but at the bottom of the tube, is sure enough a residue that I would suspect is calcium soap. I am not worried now as I know what to do to remove it.

The batch is de-mething and my intention is to then let it settle for several days, running off the glycerol in ever decreasing amounts and then without water washing or drying just run it through the oakchips. I will then test for soap and water.

I'm just wondering about the calcium soaps. I wonder if the massive overdose of CaO I gave the methoxide
has produced a lot of calcium methoxide. If this is the case if the minimum of CaO was used to dry the methoxide, without it reacting with the methanol it would result in dry methoxide without it being able to produce the calcium soaps. Maybe Countrypaul would know.

Chug, when you tried this did you get the cloudy bio due to the calcium soap.

As usual using the dried methoxide + ASM has resulted in a vast decrease in glycerol volume.

[countrypaul]

Thanks for that Paul.

I calculate by multiplying the titration by 320 so does this mean the soap (if it is calcium soap) is about half of the amount calculated.

I will put up a full post when I have totally finished the batch, but after drying it is still cloudy at the moment. For the last 2 hours of the drying cycle I added some magnesol to hopefully remove anything in suspension. I know magnesol can be a difficult thing to remove but when it is all settled (i will leave it 2 days) Then I will centrifuge it and hopefully end up with a clear product.

The water washing went really well, hardly any soapy water and no mono's at all.

Dave,

OK, your calculations are exactly as per the wiki (forgot to check that earlier) however this may be a little more complicated than anticipated. Rather than using 320 you should be using 602 for calcium soap. This will give you a techinically correct value for calcium soap, however it will be only half the value it would be for sodium or potassium soap due to calcium being bivalent. Therefore to compare the soap level in terms of the amount of fatty acid present in the soap you should double the figure you then get for calcium. It appears the original ASTM definition is for the amount of sodium  or potassium in terms of ppm and this has been extrapolated to ppm of soap. Using the value I gave would give you the correct value for calcium (not sure if that is in the ASTM standard).

Looks like I may have to trawl through infopop to find the original maths and assumptions behind them to fully understand why the figures supplied are used. After correcting the figure above (602 for calcium) it became obvious that the oleate salt is used for the soap calculation. Should not matter too much since everyone uses the same figures, but could possibly be out a significant amount if someone is using say beef fat. However the test is actually for sodium/potassium rather than soap per se so referencing a standard soap makes a lot of sense even if it is a little misleading in terms of what it is now used for.

Sorry for the earlier mistake (I subtracted KOH MW rather than just K).

Paul

[countrypaul]

A simple test to see whether there is calcium present would be a flame test on some of the bio (or better still the contaminants). If you get a piece of nichrome wire and heat it in a hot flame such as a blowlamp it should glow red and the flame may go from blue to orange. If you dip the wire into the bio/contaminant then put it in the blowamp flame with luck it will show up differnt colour. Green would indicate copper present, violet/purple would indicate potassium and brick red would indicate calcium present (for reference sodium would be orange and therefore almost impossible to distanguish from background). Not sure if esting the wood chips would be valid but try some unsed ones and if they just come out orange then try some used ones and look for the distinct colours.

Another test you might try with the contaminated bio (you did keep a sample didn't you  ) is to wash it with some dilute hydrochloric acid (brick acid), if it is calcium soap, I would expect most of it to get removed as the soap goes back to ffa and the calium gets washed out as calcium chloride. If you use sulphuric acid it will produce calcium sulphate which is not water soluble (and I doubt bio soluble) but might be so fine that it makes the bio appear cloudy.

It is possible that in adding the quicklime to the KOMe to take out any residual water that some of the quicklime is converted to CaOMe that may be methanol soluble. Again you could try a flame test and the solution, but be especially careful if you do as methanol fires are extreme. So a small sample bottle near the flame so that you can dip the cold wire into it at most. the colour might also be difficult to be sure of as there will be significant colouring from the potassium.

Obviously the flame test is not quantitative, but a negative for calcium might help rule that out.

Another obvious feature is that because the oil has been glyc washed and hopefully has a titration value very close to zero using almost anhydrous methaoxide for the reaction will result in virtually zero soap production. Any soap present in the oil from the gly washing process could still be converted from potassium to calcium soap however.

Paul

[dgs]

Thanks for that Paul, I will do a flame test tomorrow and report.

I still have 2x cubies of the original 110 litres that I set aside for the heating. I noticed today that one has now clear-ish bio and a small layer of off white deposit in the bottom.

I have ended up woodchip filtering the 90 or so litres from this batch. It is now filtering for the 3rd time and the bio is lovely and clear with zero soap.

[countrypaul]

If you can test some of that white deposit in the bottom I would be very intrested in whether that is calcium soap.