Simply put there are two competing reactions. Neutralisation and saponification. I would guess neutralisation is faster, but that's a guess. I don't know how to improve your caustic stripping procedure. Someone might determine how to improve it experimentally.
I thought about your caustic stripping some. With high ffa vegetable oil you use sodium or potassium hydroxide in water to make the ffa into soap. Then the soap is removed from the remaining vegetable oil, leaving low ffa oil to make the methyl biodiesel from. A problem is the potassium hydroxide with water also reacts with triglycerides to make soap (saponification). The potassium hydroxide react with ffa to make soap and water (neutralisation). In neutralization an acid reacts with a base to form a salt (soap) plus water. Neutralisation doesn't require water initially to work, generally speaking. Saponification does require some water, if I recall correctly. So, on a very small scale, neutralise the ffa in a minimum of water, dry the vegetable oil, remove the soap, and you've got low titration vegetable oil to make methyl biodiesel from. And isopropyl alcohol doesn't transesterify with vegetable oil very much. So, it might be more trouble and expense than its worth. With vegetable oil that titrates at 19 (KOH) dissolve 19 grams of potassium hydroxide in a minimum of dry isopropyl alcohol, add the isopropyl alcohol with the caustic dissolved in it to the high ffa vegetable oil at say 40-50 degrees centigrade, mix strongly long enough allowing the neutralisation reaction to take place. Remove the isopropyl alcohol and water from the treated vegetable oil, leaving potassium soap in vegetable oil. At that point I don't know if the soap settles out, but all the ffa should be made into soap in the neutralisation step. Now the titration should be low. I don't know if it would be worth it. I don't know if you've got the equipment to do a test batch of small size. I think this would work, but it cost extra time and isopropyl alcohol.
Thats a fascinating insight into the reaction. If I read you correctly the neutralization is likely to happen first and then as it produces more water the saponification gets going. So the aim is to minimize the water in the reaction and stop as soon as the neutralization is complete.
That makes perfect sense and fits in with the use of glycerol. In the method proposed so far the High FFA oil is treated with NaOH and just enough water to dissolve it. The mixture is circulated thoroughly for a short time and then the glycerol is added which both attracts the water and creates a bottom layer that the soaps can settle into.
Your idea of using Iso propyl may be uneconomic but trying it on a small scale would provide us with valuable information about the kinetics of the reactions. When you say 19gms of KOH do you mean 19gms per litre of oil?
I mean 19 grams per litre KOH for high ffa vegetable oil that titrates at 19 (KOH). I wrote KOH partly because NaOH doesn't dissolve as much in isopropyl alcohol. Maybe I can find some high ffa oil, I'm not sure. I usually use new vegetable oil for my test batches.
I tried an experiment with isopropyl alcohol. I processed a litre of oil that titrated at 19ml KOH by the same method as my first attempt on pg 1 but I replaced the 40ml of water with 40ml of isopropyl. First thing I noticed is that NaOH hardly dissolves in isopropyl at all. I pressed ahead anyway but next day the titration was still 19ml KOH.
NaOH and isopropyl dont mix.
I used only 3.5gms per litre of NaOH which worked the first time. However your figure of 19gms makes sense, we add an extra gram on top of the minimum amount for every 1 ml KOH titration when processing ordinarily.
yeah i definitely wouldn't add water in this process. methanol is probably the logical choice here. i'd dissolve KOH (maybe NaOH) in a minimum of methanol then mix in with caustic glycerin from a previous batch. with the right amount of KOH, you would have a final titration of 0, which should be the goal. that way, most of the water produced (neutralisation = water) would get separated out in the glycerin drainage instead of being produced during transesterification. maybe keep the mixing to a minimum to avoid saponification. temperature probably 50-60*C.
feasibility would depend on downstream process logistics as well(methanol recovery), which would vary greatly among producers. i like it because you're not adding any cost to the process, unless you don't have methanol recovery.
In a previous experiment I got 3.2 grams of potassium hydroxide to dissolve in 148 millilitres of isopropyl alcohol (anhydrous). It did dissolve. I don't know what the minimum amount of isopropyl alcohol it would take to dissolve 19 grams of potassium hydroxide.This message has been edited. Last edited by: WesleyB,
It was NaOH I tried to dissolve in isopropyl and as you predicted, it didnt dissolve very well.
Rosedales method involves only 3.5gms per litre and 40ml of water, he has used it many times and gets good conversion. I can confirm that it works as my one litre batch worked beautifully. He has tried using methanol and found no advantage.
Im going to do a number of 1 litre batches with oil that titrates at 19ml KOH.
1. Repeat the original batch , 3.5 gms NaOH, 40ml water, 20% glycerol.
2. Substitute methanol , 3.5gms NaOH, 40ml methanol, 20% glycerol.
3. Substitute KOH, 5.5gms KOH, 40ml water, 20% glycerol.
4. substitute KOH and methanol, 5.5gms KOH, 40ml methanol, 20% glycerol.
5. substitute KOH and isopropyl, 5.5gms KOH, 40ml isopropyl, 20% glycerol.
It will be interesting to see which will work best.
Today; I dissolved 4 grams 90% KOH in 100 ml anhydrous isopropyl alcohol. Melted some stearic acid, poured about 14 grams of liquid stearic acid into the KOH/ isopropyl alcohol solution. The stearic acid turned solid within less than 10 seconds. I stirred the mixture with a glass rod, stoppered the flask and left it to sit. I felt the bottom of the flask about 20-30 minutes later and it was warm. An exothermic reaction was occurring. The last time I stirred it it made a kind of slush consistency. I suppose that making soap from all the free fatty acids in titration 19 vegetable oil might make a slushy consistency vegetable oil that the soap wouldn't precipitate from. The potassium soap I made is a solid at room temperature of about 25 degrees centigrade. I have read that some potassium soft soap is liquid. Stearic acid is also called octadecanoic acid, it is one of the fatty acids in vegetable oil.
Good to see some new experimentation taking place on the forum. I think you guys are right on the mark.
Some background on what has gone on before. Caustic stripping has been limited in the commercial biodiesel world to low titration oils. That's because feedstock is so expensive and even ffa is feedstock. Anyway, these references might give you some ideas to work with.
From my websites are Caustic stripping with Soda Ash and The Glycerin Treatment, and Water Washing with baking soda.
There are also a couple of old threads on this forum with more classical approaches to caustic stripping like this one.
The original purpose of caustic stripping was to bleach oil and remove water and particulates of relatively low titration oils. They used no more than a 6% w/w NaOH / water solution. More than that would start breaking down the triglycerides.
Keep up the experiments
thanks Rickdatech, these are really useful links, there is no point in reinventing the wheel. The essence of Rosedales idea is to combine caustic and glycerol stripping. The articles you provided pointed out the fact that the water and soap produced by caustic stripping was difficult to remove, the glycerol solves that problem.
I begin to see why so little caustic soda is used . You want enough NaOH/water to get neutralization going but not so much that saponification and transesterification can happen. this may also be the reason that Rosedales processing times are so short.
Im running out of the oil that titrates at 19ml KOH so I had to decide to do only 2 of the modified experiments.
One litre of oil heated to 50C, added 5.5gms KOH dissolved in 40ml of water, mixed for 10 mins, added 20% glycerol, mixed for 20 mins, settled for 2 hours. Result titration 9ml KOH.
One litre of oil heated to 50C, added 5.5gms KOH dissolved in 40ml methanol. mixed for 10mins, added 20% glycerol, mixed for 20 mins, settled for 2 hours. Result titration 6.5ml KOH
KOH works but not as well as NaOH
Methanol is better than water.( at least with KOH)
Using methanol does not lighten the colour of the oil.( hardly matters)
The obvious next step is to repeat the original experiment with NaOH but use methanol instead of water. I may just be able to scrape up enough to do that tomorrow.
If using methanol with NaOH was even as good as with water it might have the advantage of not needing to be dried before processing. Fingers crossed.This message has been edited. Last edited by: imakebiodiesel,
I managed to do 2 more tests today.
Just for clarity I will present yesterdays results here as well. All tests were 1 litre of t19 oil heated to 50C. the hydroxide was added and mixed for 10minutes. 200ml of glycerol was added and mixed for 20 minutes. After settling for 2 hours the oil was titrated.
5.5g KOH + 40ml water = t9
5.5g KOH + 40ml methanol = t6.5
3.5g NaOH + 40ml water = t7
3.5g NaOH + 40ml methanol = t9
Note that with NaOH the water mixture performed a little better, this may be because NaOH does not dissolve very well in methanol.
Note also that the results were not as dramatic as my first batch. this is because I only settled these samples for 2 hours. The titration drops further in 24 hours but there is no improvement after that.
Conclusion. The KOH/methanol performed best. Oil treated in this way may not need to be dried before processing.
Based on this result my latest recommended method for caustic stripping t19 oil is as follows...
1. fill your processor with 150 litres of high titration oil.
2. Heat and circulate the oil to 50C.
3. Add 825 gms of KOH and 6 litres of methanol and mix vigorously for 10 minutes.
4. Add 30 litres of glycerol and mix vigorously for 20 minutes.
5. Settle for 24 hours and drain off glycerol.
6. Titrate and process using 6 litre less methanol than usual.
Note, high titration oil can be very dark and difficult to distinguish from glycerol. Mark the level inside your processor with a pencil before adding the glycerol and then drop the level to the mark when draining the glycerol.
I have no more high titration oil left so if any one wants to donate 150 litres of bad oil to research Ill be happy to process it. Likewise I would like to hear from anyone who tries this method.
Hi John you started out with 150lt of oil, what volume of oil were you left with after the caustic stripping?
Its hard to tell Dean, as I havnt been able to do a full sized batch yet. There is certainly going to be a loss as a good proportion of the FFAs will convert to soap and drop into the glycerol along with some of the water produced. On the other hand the 6 litres of methanol and any biodiesel in the glycerol should float up into the oil layer.
Sorry John I had it in my head you'd done a full batch. It will be interesting to see just how much is lost to soap. Obviously it makes sense to have the oil as dry as possible too, more water means more soap. As you say most if the methanol should stay with the oil so no loss there.
That is the downside of this method, the FFAs are lost as soap. In the AE method the FFAs are converted directly to biodiesel. This is the reason that the commercial producers have no interest in caustic stripping. In the commercial world feedstock is expensive and even FFAs are feedstock.
However the convenience of caustic stripping makes it attractive to homebrewers.
my plan is to recover the stripped ffa's and feed them through AE. AE takes a couple hours (hoping stripping also can be done in 2 hours), so i think i can alternate stripping and AE to produce a batch per hour. i'm a little worried the extra ffa could overwhelm the AE stage, but perhaps with extra methanol it will work.
yeah i often struggle with drainage points. my strategy recently is to draw off a 500 mL sample right after mixing is shut off. whatever % (v/v) settles in my sample beaker is the % that i drain from the reactor. i don't have to be too accurate though because i end up processing the glycerin anyway to recover ffa, methanol, and product.
most of the methanol will end up in the glycerin. methanol and oil do not like each other at all, which is why the 3/27 works so well. good news is that most of the bio should migrate to the oil layer, so that could make up for the lost ffa.
thanks very much for posting your results. looks promising.
Im not sure that alternating caustic stripping and AE is going to work. The two common reasons for failure of AE is water and residual hydroxide both of which will be produced by caustic stripping. I could be wrong but I think you will have to choose one method or the other.
yeah i probably wasn't very clear. the AE and stripping would be done in separate reactors both feeding a first stage trans reactor. this is a 200 L/ hr commercial design, albeit unproven as yet. advantage of this (instead of more AE capacity) would be that i wouldn't have to upgrade my preheat/drying tank (currently 100 L/hr) since i'm hoping the glycerin will strip some/most of the water.
might well be simpler to just add more AE + preheat capacity tho. anyway, we are kinda off topic.
|Powered by Social Strata||Page 1 2 3 4 5 ... 11|