Mixing vs. processing

[Julian]

I tried a little experiment today.  Processing a batch as normal, I took a sample just after the introduction of the methoxide on the first stage … you know, that gloopy mustard coloured stage that you pray will go away, and left it sitting in a sealed jar whilst the rest of the batch processed.

I left the batch processing for longer than normal (in all honesty I forgot about it), but 4 hr later I came back and the sample sitting in the jar looked, for all intents and purposes, like the sample I took from the processor.  I did a 3/27 test on both with the following results ... "mixed" sample 0.8ml dropout and processed sample 0.52ml drop out.

Given that I'm processing with Frankinpump I'm thinking that the energy used to process for all that time (perhaps a little excessive in this case) wasn't worth the diferance in conversion.

Maybe less processing and three stages with a small additional amount of chemicals would be more economical.  Or perhaps periodic mixing ... use the pump to draw in the chemicals and then run for 2 or 3 mins every hour ... this could quite easily be acheived on a plug in timer.

The batch was made with good quality glycerin washed oil starting at ambient temperature and, although Frankingpump took the processer up to 57°C, the "mixed" sample remained at 22°C.

I'll try and do some more tests using this idea and post up, but it would be great if others could have a play and see if they concur with my initial findings.

[dgs]

The gloopy mustard stage you mention is when the mix is immicible due to oil and methanol being present (ie not enough biodiesel produced to act as a cosolvent) so it's a suprise that both results are similar.

However as you did a 3/27 both results show a large unconverted percentage. Mixed sample 26.7% unconverted and processed sample 17.3%

Did you use any heat at all Julian as the increase of 35degs from just the pump is high even for a 4 hour process.

Multistage processing certainly reduces chemical requirements, a few years ago I did a 7 stage on glyc washed oil and used only 3.5gms/litre KOH to get a clear 10/90

Interesting stuff, let us know how you go on.

[Julian]

Dave, the heating effect of the pump will be proportional to the work it has to do.  I have a small throat in my venturi and the pump runs around 2 bar on converted bio, so that temperature rise is not unusual for my system.

The second stage of the batch went fine resulting in a clear 3/27.

Had a look at a couple of plugin timers I've got and both are rated at 2 amps on an inductive load, so no good for Frankinpump, but I'll try manually running it for brief periods on the next batch.

[Head Womble]

Use the timer to switch a SSR, that'll get around the amp rating.

[Julian]

Yup, that or a relay.

I was just looking for a quick lash-up as an experiment initially.

On my super-doper £4.46 Ebay watch I have a timer, so to try out the idea I'll set that and switch the pump manually.

[dgs]

Just a few thoughts here Julian, so hang in with me.

I suppose what you are saying is that when you have had so much conversion by only mixing for a short amount of time then the energy input (from mixing and heating) is disproportionally high to achieve a higher conversion when processing for an hour or so.

This all stems from the reaction being equilibrium. The only way I can see a process reaching completion in (say) 5 to 10 mins or so is to increase the chemicals. Doing this would leave the process open to possibly producing vast amounts of soap.

I remember several years ago doing 10/90's every 2 mins or so after the methoxide was added and I was surprised at the amount of conversion after just a few mins.

[Julian]

Yes, that sums it up pretty well.

I've always suspected the bulk of the conversion happens right at the start of proceeding but have never investigated.   Thoughts were to process over several hours (could even be days) using my normal 2 stage process, but only run Frankinpum say 5 mins in every hour.

Are you saying the process has to have vigorous and prolonged agitation to take it to completion?   I'm in no shape or form a chemist, but I'd have thought the components only need to come into contact to react which periodic "mixing" would achieve.


Pub quiz tonight so seeing my mate with a PhD in chemistry.  I'll see if he has any views.

[neisel]

Anyone conducted any experiments on the effect of introducing methoxide into the dry oil at different rates?

I don't have any figures but I have a strong suspicion that dosing very slowly (15 litres over 30 mins) yields a significantly better reaction.

[dgs]

I don't think your suspicions are correct neisel. Just look at the principal of the AAF method where all the methoxide has the oil added to it. My methoxide is added in about 15 seconds, I used to add it slowly but found there was no point.

[Glycer-rides]

I don't think your suspicions are correct neisel. Just look at the principal of the AAF method where all the methoxide has the oil added to it. My methoxide is added in about 15 seconds, I used to add it slowly but found there was no point.

Yeah, I second that, Dave. JRL championed doing the reverse, hence the AAF method.

And whilst it may pain us to say .... "he ain't not normally not wrong".

[neisel]

As I understand it the AAF method involves introducing oil to methoxide, not MO to oil.

I tried dumping 19 litres of MO into 145 litres of dry, hot oil once & had a very poor reaction. My last batch converted 145 litres of oil using just 19L of methanol (15 + 4) with 750 grms of KOH on the first stage & 140 grms of NaOH on the second, all introduced very slowly.

[Julian]

Well, I asked the question and the answer was "That's chemical engineering not chemistry", so no definitive answer I'm afraid.   But after perusing him, slow reactions in the lab are always stirred, and vigorous mixing is not necessary.

Best way is probably to try it and see what the results are.

Re speed of introduction, when I built my plant the accepted wisdom was to introduce the metoxide slowly, I've never deviated from that method and have found it quite satisfactory.  It makes sense to me as the methoxide stream will be better mixed with the oil.  At the end of the day I recon there probably isn't much difference between the two methods.

[dgs]

One of my many processors that I tried over the years was a 400 litre plastic tank (the one with the dished bottom instead of a proper cone) it was fitted with one of those 110lpm yellow and black pumps (closed impeller)

I sometimes had to help the pump by running glycerol out of the bottom tap and pouring it back into the top. The methoxide was introduced over a 20 min period. I used to need 12 to 13 gms/litre KOH for full conversion with this set up. What a poor combination of tank and pump it was.

I then went small, 120 litre tank fitted with a puisi panther and 3/8 eductor; what a difference! Normal amounts of KOH needed, methoxide poured into the top (not to be recommended) in less than 10 secs. Those panthers are powerful little things, this is the set up I used when I did my 3.5gm experiment.

The trouble with adding methoxide slowly is that the last few % of methoxide is added when most of the reaction is complete and as such the glycerol in the mix will try to absorb the methoxide before it can do its job. This is why multistage processing is much more efficient as regards chemicals.

If you have ever done a S2 with a considerable amount of S1 glycerol in the mix you will know what I mean, a very poor reaction.

[dgs]

I needed to do a process today, so thought I would give Julians idea a go.

196 litres of glyc washed oil that showed a conversion of 15% ( 8.5mls dropout using 10/90 )

I added the methoxide ( 700gms K + 1350mls ASM in 22 litres of methanol) and used the pump to mix for 5 mins only. Temp was 25degs.

10/90 showed 1.5mls dropout.

1 hour later I started the pump for another 5 mins then did another 10/90 which showed 0.5mls dropout.

1 hour later i started the pump again, just for 5 mins as before, after this the dropout was 0.25mls.

1 hour later I repeated this procedure and the dropout was 0.05mls,

After another cycle I got a clear pass with the 10/90

No heat was added and after the 5 hours the temp had risen to 30degs simply due to the hot sun shining on the side of the tank.

Off topic I know but tomorrow I'll do a stage 2 on this batch using about 0.3gms K/litre simply to convert the residual mono's that I know will still be there. Then I know it will water wash well with no emulsified layer.

[Julian]

Beat me to it … I loaded the oil for the next batch yesterday and I've yet to do the glycerine wash!

Must have saved a lot of electrickery. So if I've got it right 25 mins total processing, single stage giving a clear 3/27.  I've never acheived that with continious processing over several hours!


An interesting comparison would be to process for 25mins and let it sit for 5 hrs.