The 10/90 test -Dropout versus Oil

[dgs]

Hi Everyone,
Just thought I would share some 10/90 tests with you, I posted these on VOD a while ago.




I have done a few dozen dropout tests over the last 2 to 3 weeks and unfortunatley cannot get any agreement between dropout and triglycerides.
I did manage to get the test to work by leaving it until the methanol phase was clear, then removing the methanol with the dissolved bio and re-adding methanol. I had to do this twice on some tests before I got a reading that was reasonably close to the amount of actual oil, so it was totally impractical.

I have therefore produced the following graph, so by looking along the vertical axis at a certain amount of dropout , then the actual amount of oil can be found.

There are so many variables with this test that I don't propose the graph to be deadly accurate.
As you can see the line will not intersect all the points. Some of the tests were done several times with only slightly different results.

The graph however will be far more accurate than taking the dropout 'as is'.

Click on graph to enlarge.

[dgs]

Just a few words of explaination re the above post. Each test was done at least twice.

A very rough estimate is that the dropout amount from the 10/90 test over estimates the amount of oil by @ 100% which leaves us with more questions. If we add further chemicals to fully convert a batch using the graph, we will only add @ 50% of the reqd amount. It therefore seems that although the total chemical amount won't change, the intermediate amounts added are too high for the amount of oil present. Does this mean the base amount changes as the reaction proceeds?

[Tony]

Very interesting that it is considerably less than we thought.  I would love to find a equation to fit the curve in order to produce a javascript tool for the "Tools" menu that calculates a good measure of actual conversion.  Please can I have the source data?

For a couple of years now I've been adding excess catalyst to no ill effect, despite the "you'll make soap or jelly" scaremongering - with dry oil I've not seen any ill effect, either with KOH or NaOH.  From what you say it seems so has everyone else, unwittingly, for second stage, which should always be nice and dry thanks to stage one glycerol.

This tallies with what I've observed in other ways too, I've seen a partial conversion pushed through to clear pass with less methanol in the second stage than I thought would be required to do so.

Excellent work dgs!

[dgs]

Thankyou Tony,
The test data was as follows;
I mixed up samples by pipette in a 200ml volumetric flask, they consisted of different %'s of new soy oil which were 2% 4% 7% 10% 15% 20% These were mixed with fully converted biodiesel (when I say fully converted I mean clear 10/90) Which maybe explains the zero dropout at @ 1.5% oil)

I then simply did a dropout test in duplicate on each sample. The 7% sample was tested several times as it doesn't exactly fit the curve.

Although interesting it throws up the anomoly re chemicals to add at any given dropout.
For instance on a 100 litre batch if we had a 2.0ml dropout we would normally add base amount for 20 litres of oil. However, at 2.0mls dropout we only have 9% unreacted oil which would not convert on @ 50% of the chemicals required for a 2.0 ml dropout.

It is almost as if we are working with some sort of double error which compensates for itself.

[Tony]

Thanks - did you see Drum's thread on the same subject?

http://www.biopowered.co.uk/forum/index.php/topic,2480.0.html

[dgs]

Yes I did see that. Although it is a mathmatical model some of the results are quite close to the graph.

[Drum]

Although interesting it throws up the anomoly re chemicals to add at any given dropout.
For instance on a 100 litre batch if we had a 2.0ml dropout we would normally add base amount for 20 litres of oil. However, at 2.0mls dropout we only have 9% unreacted oil which would not convert on @ 50% of the chemicals required for a 2.0 ml dropout.

It is almost as if we are working with some sort of double error which compensates for itself.

Hi guys been away for a while. Just seeing this now. Dave you have kinda nailed it with your last sentence. The reason why we cannot convert on 'half' the chemicals for a given dropout is that we are not properly accounting for the di's and mono's present in the mix. I did a quick calculation which assuming I am not missing something very obvious, indicates that we are underestimating the chemical required to finish off the di's and mono's and much of this is unseen because the dropout contains very little of the monos which actually constitute the largest mass fraction of glycerides in the bio in the latter stages of the reaction. Combine this with the fact, as I see it, that equal mass fractions of triglycerides, diglycerides and monoglycerides take approximately the same amount of methanol to convert. I mean if your bio mix contained 1kg of tri's, 1kg of di's and 1kg of mono's then each of these components would use up equal amounts of methanol to convert to biodiesel. Well not exactly the same: I calculated that for rapeseed/canola that the ratio of methanol required to convert equal masses of tri:di:mono would be in the parts ratio 100:98:91. Combine this with the result in my model indicating that if you have 1ml dropout in the 10/90 then your mix contains, by mass, approx 3% tri, 7% di and 7% mono. If this is accurate then in terms of converting every -glyceride molecule to FAME, this has a triglyceride equivalence of about 16% mass fraction. If this is the case then the 1:1 dropout rule we have been using for stage 2 completion has not been overdosing at all!! Or if it is overdosing it is down to our built in base amount actually being higher than necessary. Comments and criticism welcome.

[dgs]

Hi Drum,
No criticism at all, what you are saying is totally logical. You could well have found the answer here. The reason why we are not overdosing with chemicals is that a soluable proportion of the di's and nearly all the mono's still have to be converted even though they arn't showing up in a 10/90.

As regards the base amount, I'm sure you are correct in that the amount isn't exactly as stated.

Didn't the chemist 'Neutral' come up with the base amount many years ago? not that I would criticise his work. Over the years he has done a lot for the advancement of biodiesel.

Well done Drum, that is such a logical answer, staring me in the face!

[Chug]

Neutral determined with gc testing that 3.5g base for NaOh that we all used, although it was making biodiesel(ie able to seperate glyc) was not enough to reach ASTM spec and the amount needed to reach astm pass was 5g. Along with 20% meth.

I still have the graphs for these that Sam Ley drew up from Neutrals results somewhere.

[Chug]

here ya go, Sam Ley's graphs of Frank's GC results.

[dgs]

Thanks for that Chug, most interesting.

[Drum]

Very interesting Chug. Do you know what, presumeably fixed, KOH level was used in the second graph or indeed the, presumeably fixed, methanol level used in the first graph? Also were the tests carried out with fresh oil? What about temperature equivalence, processing time and mixing capacity variables? Have you any data on those?

I have been playing with trying to reduce chemical usage recently and have been working with constant 17% methanol (in my setup just sufficient for a zero dropout 10/90 pass). When I reduced the KOH to 6+t g/litre of 90% KOH I was only getting 98% conversion which encouragingly for me agrees pretty well with the graph! I have been able to get a zero dropout pass with 17% CH3OH asnd 6.5+t g/l. This was on very good oil to begin with; quite dry (800-1000 ppm) and titrating at around 0.4-0.8.