If we were to react the oil twice then centrifuge it (with luck), the byproduct should be fully removed. However, would the separated ester be alkaline?

I mean would the spin separated ester be neutral enough to need little or no washing?

I've been offered a small centifuge designed for dewatering machine tool cutting oils. I don't really want to waste time and cash if it won't allow me to avoid washing the ester.

Dave UK

A centrifuge can only accelerate the settling of materials of different densities. If two or more substances are in solution the mixture will be homogeneous and, from the point of view of a centrifuge, one material and hence inseparable. Thus methanol and free fatty acids can never be separated this way.

In the case of soap and water the situation is different. Both dissolve in the ester layer to some extent. But both can be present at a level in excess of that which will dissolve, giving a cloudy or hazy appearance. A centrifuge will do an excellent job of removing the excess but not the dissolved portion. The amount left in solution will depend on temperature: the lower the temp the less left in. The amount of soap left in will depend also on the amount of methanol in solution: more methanol will hold more soap.

So the short answer is no.

The best use of a centrifuge is to use it after washing out the soap. The BD will now be free of methanol and soap but will contain water in a fine haze. The centrifuge will remove this haze immediately but note that some dissolved water will remain, which may reappear as haze if cooled.

If you let a sample of BD stand for a week or two it will be perfectly clear and just the same as the centrifuged sample. The main advantage of a centrifuge therefore is to avoid the necessity for having large settling vessels for the BD but in a small scale system the settling vessel may be more economic than a centrifuge. A settling vessel takes no energy.

[This message was edited by neutral on 08 April 2002 at 08:46 AM.]

Thanks Neutral,

So I could use the centrifuge to speed up removal of (most but not all of) the second stage byproduct. After that I would still have to wash the ester.

I had a feeling this would be the case, but I was not considering dissolved soaps, more unused catalyst from the second stage reaction. I was also thinking of distillation recovering methanol before spinning the mixture - thickens the byproduct.

The second stage feedstock could be 10% unreacted oil, but most people over-egg the catalyst - presumably to get around the clutter effect of the ester from the stage 1 reaction.

This excess has to go somewhere and presumably can't be spun out. Acid washing after spinning out the soap being the only option.

At least if it's spun before washing, the FFA soaps should have gone so little risk of sending a fractional percentage of FFA back into the ester.

I agree totally about conditioning the washed ester. I found that after a week my ester was amazingly clear - even if it was still over the final wash water.

I put test samples in the fridge with heat dried control samples. When chilled, the non heated samples only became slightly cloudy so there was very little water in there.

Anyone thought about, or tried, water misting whilst centrifuging? I haven't thought of a good reason it wouldn't work, and would speed up the process considerably. For that matter, anyone tried a "flow-through" centrifuge - like a cream separator, instead of batch spinning?
JohnO
(the difficult part might be fitting the spray nozzle inside a Dr Pepper bottle

Dave UK

I'm not sure I understand all of your post, but here are some comments:

>"So I could use the centrifuge to speed up removal of (most but not all of) the second stage byproduct. After that I would still have to wash the ester."

If by byproduct you mean the glycerine layer there is little to be gained by using a centrifuge as it settles well anyway, being fully settled in the time it would take for you to run a batch through the centrifuge.

Neither settling nor centrifuging will remove dissolved soap. To do that you need a water wash. By the time the soap is reduced to low levels the glycerol will be totally removed as it has a high affinity for water. It is more appropriate to use water to remove the glycerine/soap and the centrifuge to then remove the water.

>"The second stage feedstock could be 10% unreacted oil, but most people over-egg the catalyst - presumably to get around the clutter effect of the ester from the stage 1 reaction."

If you do a single stage reaction and don't skimp the methanol the conversion will be quite adequate, certainly not 10% unreacted, perhaps 2%. Why bother with a second stage? If you are going to recover the methanol why skimp it?

>"This excess [catalyst] has to go somewhere and presumably can't be spun out. Acid washing after spinning out the soap being the only option."

Most of the catalyst you put in turns into soap which is water soluble and comes out in the wash. It has some affinity for the ester layer however, so takes a bit of effort: three washes usually being needed. In contrast the excess lye, if present, has a huge affinity for water and will be all gone after the first wash. Why anybody worries about lye being in their BD after washing is a mystery. There is certainly no need to add acid.

>"At least if it's spun before washing, the FFA soaps should have gone so little risk of sending a fractional percentage of FFA back into the ester."

If you spin it you will only remove the excess soap that was not in solution. If you add acid all the dissolved soap will convert into FFA and will finish up in the BD.

>"...after a week my ester was amazingly clear - even if it was still over the final wash water. I put test samples in the fridge with heat dried control samples. When chilled, the non heated samples only became slightly cloudy so there was very little water in there."

Glad to hear you did that experiment. It confirms the theory. Allowing dry air to waft over the surface of the BD helps. I think it gets the water content lower than settling alone.

[This message was edited by neutral on 08 April 2002 at 10:15 PM.]

My last post clouded the issues

Neutral thanks for your comments. Most of your points are well taken and I certainly agree with you. I was trying to be concise but accidentally raised more questions. Thankyou indeed for your point about dissolved soaps - I was not aware that soaps could dissolve in methyl ester.

I have made a number of 40 gallon batches using a mechanical mixer in an oil barrel. With my feedstock, I've found that a first reaction allowed to settle and separated followed by a second reaction seems to make the best ester. Reaction 1 settled very fast but reaction 2 needed 24 hours greater excess methanol diluted the byproduct.

I have always water washed using the air bubble method with 3 water changes. I used acid in the first wash to reduce the emulsion created. Some say this can release FFA from soaps suspended in the unwashed ester. I reasoned that as there is very little soap present and that most of the FFA came out at stage 1 there should not be much to release.

My question was sent because I am trying to think up a cheaper way to make a commercial sized plant. I'm miles off going into business, but keep pegging away.

In UK there is not much point in doing much less than 20 tonnes per day. Premises, labour costs and overheads generally mean anything smaller won't make enough profit to expand. Whatever I do I want the plant to be inexpensive and small so it can more easily expand.

I could simply scale up the barrel method with proper commercial reactors, but if I can come up with a semi continuous process then I'd gain from lower operating costs. With variable quality feedstock like waste oil, a fully continuous process is not really possible.

But, design benefits in one area cause problems in another. I'm working out the likely most effective ways to get things done. The centrifuge question came about from one line of thought.

If you'd like to discuss offline Email me at ~iodiesel.tiscali.co.uk (to avoid spam I've swapped the ~ for an obvious letter)

Dave UK: I'd be happy to help, in any way I can, design a continuous BD processor (I'm a Manufacturing Engineer, and process improvement is my daily job). My company isn't interested in pursuing BD, so it's simply a hobby/passion for my free time. Contact me off-line at jouster@NO_SPAMred.gd-ots.com

cheers,
JohnO

Dave UK

As I mentioned on another thread a good place for a centrifuge is for treating the starting material if you are thinking of going commercial. On a larger scale you cannot be very selective and will have to deal with dirty and wet oil.

Re your comments about the two stage reaction >"Reaction 1 settled very fast but reaction 2 needed 24 hours greater excess methanol diluted the byproduct."

This time penalty would turn me away from a two stage process, seeing that one stage works so well by gc analysis, provided the NaOH is correct and enough methanol (25% by volume) is used. The high methanol use is not a penalty if recovered, and it would be in a commercial situation.

Re acid:

>"I used acid in the first wash to reduce the emulsion created. Some say this can release FFA from soaps suspended in the unwashed ester. I reasoned that as there is very little soap present and that most of the FFA came out at stage 1 there should not be much to release."

Clearly you put the acid in because it reduced emulsion. There is only one way that acid can reduce emulsion: it is by turning soap into FFA. Significant soap must be there or you wouldn't be wanting to add acid. It all depends on the FFA you are prepared to accept in the product. The ASTM standard does not seem to be very strict on this point. But why put acid in when it makes so little difference to the washing when washing is done by a good method. Many people report problems with bubble washing. Mist washing is the way to go. See Biodiesel's posts.