I think it will be very interesting to see what happens with WVO.
Just to be clear, the oil you will be using has not had any type of pretreatment, neutralization, enhancement, etc performed to it?
I find the water and soap numbers very enlightening
To be honest I expected the soap levels to be much higher in all the samples. But the results are what they are, and I am confident that I did the tests carefully.
What do you make of the soap being so low? Has it all dropped into one of the two byproduct bands we discussed?
As for the oil in my processor, it is completely un-neutralised and has not been pretreated in any way. I have tipped the contents of the wvo cubies straight in, settled a few days, drained off 10 litres of watery gunk and then have kept warming and bubbling the mix using an aquarium bubbler in my processor. This is now well mixed together and dry, so I will take 3 litres and use it for re-running the test.
PS. I am in the process of redoing the 10/90 tests on the bio I made from new oil. I will photograph the dropout in the tube and post the result when done.
There was no water present and probably very little FFA so there was probably very little soap produced.
The few times I titrated new oil the colour change came after the first or second drop of titrating fluid
Of course one of the claims for this procedure is that it is foolproof and no titration required.
Just that there is no claim that it will actually produce very high conversion biodiesel
I titrated new corn oil to get a 0.6 Sodium Hydroxide titration number. It was not zero.
Here are the second 10/90 dropout test results for the new oil that was reacted with the control, 'quicklime-dried' and 'cement-dried' methoxides.
Thank you for your earlier comments Wesley. I have tried to carry out these retests under as standardised conditions as possible:
1. Same amount of agitation after adding the sample to the methanol (10 hard shakes)
2. Same amount of time for the settlement to occur (6 hours)
3. As constant a temperature as possible (17.5C in my cellar)
Methoxide that has been 'dried' with quicklime in the test produced 0.2ml less dropout than the dropout in the control and 'cement dried' samples.
I have started the second experiment running to test the Imisides technique, this time using WVO instead of new oil.
When I removed the oil for the samples from my processor, I assumed it was dry, however it certainly wasn't clear - Imisides specifies that the oil used must be clear (his indicator for "dry enough").
I heated the oil up to 180C in a pan for 40 minutes to make sure it was as dry as possible. I have just checked it this morning, and it is now very clear, even at 17C.
I will process at the temperature in my cellar and report back when I have some results.
image1-3.JPG (45 Kb, 12 downloads) Dried and clear WVO
Best of luck.
I suppose it is a lot to expect it to work exactly as M I says when even the new oil didn't fully convert.
It wasn't a total failure but using @ 14.3 gms KOH in a litre of new oil without achieving full conversion was a bit disappointing.
The most suprising thing for me was the lack of the insoluble calcium soaps. With our 1st full scale tests, I'm wondering if the residual soap after the glycerol pre-wash was the cause, converting some/all of these potassium soaps to calcium soaps virtually before the transesterification process had started.
Yes, that was a surprise. No suspended calcium soap. After all the horrible stuff we created in the full size batches at Christmas. Hopefully the wvo test will be a better test. Will post as soon as the data is ready.
Thank you for continuing to report your findings.
Do you think this Tiny difference of 0.2g fallout is significant?
If you do, what do you think it means?
I am looking forward to your continued experimentation with this procedure, keep up the good work.
You seem to have never been able to understand that there is not a Stoicheometric amount of KOH required in the reaction. All that is required is sufficent KOH to keep the reaction going. It is the methanol that is required in excess, except of course you claim to achieve a full conversion using 12% methanol.
There is your problem.
You were combining two different kinds of magical procedures.
Only one magical procedure per reaction please
Re the by-product in the pic, the two separate layers, can you suck some out of each layer and see how it behaves in water? Does either act like soap?
You do things your way and I will do things my way.
Remember what Bob told us,
We are here to discuss the ins and outs of making biodiesel, We don't have to agree.
3/27; 90/10 or however you want to portion the test as I read it is an almost immediate indication of how well the transesterfication process has proceeded. 3 mls of biodiesel will dissolve into 27 mls of alcohol. The test result should be read at room temp within 15 minutes? Where does it say that 3 mls of biodiesel will dissolve into 27 mls of alcohol permanently? Is it possible that too much time has lapsed here and the results are skewed? Another part of the test was that your sample have a very clear phase, you certainly achieved that!
" I don't know what I don't know until I know"
1994 GMC 6.5 Tubo 2005 Dodge ram 3500, 3 VW's 2000, 2002, 2005.
So quicklime had same soap as control but less fallout
Cement had same fallout as control but double the soap
was the soap layer visibly bigger in the cement than the control?
Interesting stuff David, I'm surprised 15g KOH didn't give full conversion on new oil, I'm surprised you got clear bio, Interested to see how the wvo tests compare.
It is the use of 143ml methanol in a single stage reaction that would be more likely to be the problem.
As well,the time restriction for the temperature the reaction was run at would be a problem.
I know some people pretend to perform the reaction at room temperature using 120ml methanol in a single stage in 45 minutes, however, once we actually get into the real world, 14.3% methanol in a single stage reaction run for an hour starting at 30deg C is very unlikely to result in a pass on the 3/27 test.
Here are the results of the reactions I have carried out on wvo that titrated at 3.5koh using the Imisides method
3 x 1 litre samples of wvo were reacted using methoxides "dried" using quicklime and cement as well as a control sample.
The bio phase and the byproduct phases were homogenous, by which I mean, no banding of the byproduct phase was visible. It was all one colour, and the bio was all one colour.
The bio produced after this 1st stage reaction was clear in the case of the control and quicklime samples and crystal clear in the case of the cement sample.
The control and cement samples produced 173ml of byproduct
The quicklime sample produced 161ml of byproduct
10/90 Dropout Test
Soap Test (titration amount of HCl/ppm)
I am a bit perplexed by these results, especially comparing them to the earlier results obtained using new oil. The 10/90 test has resulted in less dropout in each case for wvo as opposed to new veg oil. The wvo I used was pretty rum stuff, and contained a proportion of palm and animal fat. I would have expected that the new oil would be much easier to convert. Having said that, the new oil did produce less byproduct (120 to 125ml) as opposed to wvo (161 to 173ml).
As for the validity of the Imisides method, well, there is clearly a difference between the results, so adding quicklime to methoxides does have some beneficial effects upon the rection:
1. In the case of wvo, over a single stage, it produces 12ml less byproduct per litre of oil reacted as compared to a control sample.
2. In the case of wvo, whilst the 10/90 test is not a proportional test of conversion, when used in this context to compare samples within one experiment, I think we can say that quicklime treated methoxide does produce slightly greater conversion of feedstock to bio.
3. In the case of wvo, there is a marked decrease in the amount of soap produced when compared to the control sample
The Imisides method does, therefore, seem to have some merit. But for me, as a small scale home producer, it does not bring sufficient benefits to make me wish to use it for my own process.
I think it is worth us considering exactly what is really different about the quicklime treated methoxide that brings about the measurable changes I have observed in the course of these experiments. Is it reducing the water in the methoxide, or is the quicklime reacting with the methanol to produce calcium methoxide?
If the quicklime is only a drying agent and nothing else is happening chemically, it could be that the Imisides method would be of most use when using methanol that contains a high starting proportion of water - for example recovered methanol. I'm lucky that I have a good supply of nice quality virgin methanol, but not everyone can obtain this in their part of the world.
I'm attaching a picture of the sample bottles that I took this morning. If there is any further testing that might be useful, please let me know. I will keep them for a few weeks then add them to a batch in the reactor.
Thank you for your interest
image.jpg (44 Kb, 11 downloads) Sample bottles
Well done with all the testing you have done David.
There is certainly a lot to take in here. I found a reduction in the water present in the potassium methoxide dried with CaO of 85% using my s/b.
I also found a greater reduction in the glycerol layer. As said the main drawback (as we both found when we did our full scale batches) was the cloudy bio which we have concluded were calcium soaps.
However your tests have generally produced clear biodiesel. I am assuming when you did your full size batch your oil (like mine) was glycerol pre-treated.
Could this be the reason we were left with cloudy biodiesel.
Thank you for continuing to post your test results.
Exactly my thoughts, it all gets back to both time and money.
I can achieve the same results in a morning without buying CaO or Cement so I am not sure why a person would want to go to all this trouble.
And, once again, claims to the contrary, the normal reaction without drying does progress quite nicely at room temperature.
I found the main drawback was the huge amount of extra time spent for no benefit.
What testing did you do to conclude it was calcium soap?
|Powered by Social Strata||Page 1 2 3 4 5 6 7|