Determination of Residual Hexane in cocoa Powder

[dcav]

Hello, I'm studing Chemistry at University in Genoa (Italy)
First of all I want to apologize with everyone for my English, which probably isn't very good, but I hope you will understand me.
In these months I'm working for a company which produce cocoa butter extracting the cocoa seed with technical hexane, then they dry the residual and grind it to obtain cocoa powder defatted.
I have to improve their method to analyze the residual hesxane in the powder with headspace gas chromatography.
Now they put 5 grams of powder in a vial with ten ml of water and kept it closed at 80°C for an hour.
Then they inject 500 microliters of the headspace in a capillary columns ( i don't know the type yet, but the detector is FID).
They are using external standard calibration but there are few problems.
1) the hexane used in the industry shows four-five peaks and they refers only to the n-hexane peak.
2) their standard sample didn't contain powder but only solution of technical hesane in methanol
I suggested to try a matrix-matching method, preparing the standard by addiction of a known quantity of hexane in a "virgin" powder.
Maybe it could be also possible the use of an internal standard.
The problem is that is very difficult to completely remove hexane from the powder.
Unfortunately also a powders made without using hexane presents a peak with the same retention time.
I'd like to have some suggestions.
Overmore do you think that the addiction of water to the sample is strictly necessary?
Thank a lot!

[chromatographer1]

DCAV

Cocoa powder: sounds like fun. Yes I can understand you clearly.

You should find a new solvent.

Try dissolving your powder in dimethyl acetamide [DMAc] or a ratio of water-DMAc to dissolve. Once you can dissolve the powder completely you will have no problems with an external std as long as you run your stds in the same dissolution solvent (mix).

Hexane will be linear from combinations of these two solvents. Prove that is true by running several stds of hexane from the solvent without any powder added.

Please, use less sample. You can easily see 1 ppm from a 100mg sample of powder. Why use 5g? This is a common mistake. It does not improve reproducibility or sensitivity, in fact, it makes it worse.

I would suggest that if you can use 500µL or less of water (or water/DMAc), 25 to 100 mg of powder, and heat for only 8 minutes at 80°C with stirring you will be quite successful. Just be certain the powder is completely dissolved.

The best validation would be to make a solution of powder, completely dissolved, and heat to remove all hexane. Then make standards from this solution with 5 or 6 different hexane concentrations as well as a blank to show linearity.

Really, you are just wasting time in such excessive heating.

You may have a coelution of another analyte matching the retention time of hexane. Use a different column (or a MS if you use 100% DMAc as a solvent) to verify the peak is hexane.

good luck and best wishes,

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

814-359-5737 voice
814-359-5459 fax
rgeorge@sial.com

[dcav]

Thank you for your precious advices!
I will try to use a smaller sample and a shorter heating time, these were the conditions that they have used till today and I was thinking of changing just one variable at time.
I didn't think that there would be any solvent which could dissolve cocoa powder, because it is insoluble in water also at high temperature.
It is even a problem to get a complete wetting of the powder in short times.
i will search an adeguate solvent and let you know. Thnk a lot another time!
PS
maybe it sounds fun working with cocoa but i'm really starting to hate chocolate
Dario Cavallo

[yuri]

Dario

I just want to tell you that, in my opinion, you must considerate all the four-five peaks shown in your chromatogram.
Probably they are methyl pentanes, and their toxicity may be different from n-hexane, but FID response factor should be the same and you could quantify them without specific standards.

Good luck

[dcav]

thank you yuri,
I think i will consider all the peaks, that of course are three different metyl penthanes, even because their proportion varies between the pure industrial solvent and the sample analysed ( maybe due to different vapour tension?).
So you said that i can just make the sum of the areas, right?
There is another fact that i should consider, the peak with the same retention time that could be present in every cocoa powder and so interfer with the accuracy of the determination.
Maybe i can exclude this peak from the integration to be sure.
i was very surprised in knowing that there isn't any limitation of solvent in the powder by any law! Is it possible?

Thanks a lot

[chromatographer1]

hexane is not very toxic compared to many solvents

hexane present when the chocolate is made most likely is completed evaporated.

I never eat cocoa powder plain and dry.

You may wish to use dilute acid, sulfate, chloride, or carbonate solutions to achieve dissolution. Try to keep is as dilute as possible.

Keep in touch and let me know how it turns out please.

good luck and best wishes,

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

[dcav]

Dear Mr. Rodney
your right about thetoxicity, but i know that there is a limit for hexane in cocoa butter (1ppm) and it sounded strange to me that there is not a similar one for powder, it was just a curiosity.
Anyway, i'll try to dissolve the powder matrix, do you think that in this way i will get much better results?
Now i'm trying to make a calibration with internal standard ciclohexane in a "clean" powder the problem is in finding an adeguate solvent for the standard mixture, because directly with micropipette i didn't get reproducible data.
i'm using a paraffin oil.
Maybe i won't work on it for some days because this company is going to stay close for two weeks.
Thank you again
I'll let you know soon

[dcav]

Updating

The cyclohexane seems to be good in using as internal standard.
The problem is the solvent, i tried hexanol and paraffin oil, but both of them gave peaks at the end of the run that didn't "exit" totally in the run.
I will try with a seed oil, or maybe I could program the temp adeguatelly, maybe with a ramp after the cyclohexane peak.
Now the max temp in the column is 120°C, i don't know what temp i can get without any problem with the column.I let you know the type of column.

On the other and i made few scerening attempt of dissolving cacao powder in EtOH, THF, ethyl acetate and solforic acid. There were always a partial solubilisation, but i have to investigate more, changing conditions.
I'm sure that would be easier to prepare standard in a disssolved matrix but do you think that the chromatogram would be simplier?
Wouldn't be any product of degradation of the matrix (is quite complex: butter,cellulose, protein) like volatile fat acids or similar that would interfere?
Thank you very much for your kindness

[chromatographer1]

I do know that if you don't dissolve the powder the results you report will not have a high level of confidence, especially to a governmental oversight authority.

Anyone out there on the forum have a solvent or solvent mix they know will dissolve cocoa powder?

Using dimethylacetamide or dimethylformamide as a solvent will work well with hydrocarbon standards.

Enjoy your vacation and don't eat too much chocolate on the beach !

Rod

[dcav]

Hi everybody.Unfortunately i'm still at work!!!
The method developed with internal Standard seems quite good but should be improved.
At the same time we are trying to do an external calibration with hexane in cocoa butter added in cocoa powder.
The think that this would be the STD preparation nearest to the conditions in the real sample.(we found in litterature that hexane is adsorbed and binded to lypophilic regions)
we reduced the sample to 1 g, add water, homogenize with an high frequency bath (ultra-sounds also in english?)
the put at 80°C.The problem is that there is an upper pressure in the vial, because when i extract the siringe from the septum o hear a sound like a
little whistle, so i fear that i surely lose some of the air sampled.
I think it is due to the vapor pressure of water.
What should i do?
Use no water? less water?
Bring the sample to room temperature before injecting?
Thanks for the help

[dcav]

Hi everybody.Unfortunately i'm still at work!!!
The method developed with internal Standard seems quite good but should be improved.
At the same time we are trying to do an external calibration with hexane in cocoa butter added in cocoa powder.
The think that this would be the STD preparation nearest to the conditions in the real sample.(we found in litterature that hexane is adsorbed and binded to lypophilic regions)
we reduced the sample to 1 g, add water, homogenize with an high frequency bath (ultra-sounds also in english?)
the put at 80°C.The problem is that there is an upper pressure in the vial, because when i extract the siringe from the septum o hear a sound like a
little whistle, so i fear that i surely lose some of the air sampled.
I think it is due to the vapor pressure of water.
What should i do?
Use no water? less water?
Bring the sample to room temperature before injecting?
Thanks for the help

[chromatographer1]

Cocoa powder contains cocoa butter, does it not? Or is the butter already removed in the powder you are testing?

Try using no water and heating the sample to 110°C or until the sample is melted. I would heat for a longer time (15 min) since you have no water.

If it does not melt, then I would add a minimum of solvent (find a new solvent!) How about propylene glycol (@ 80°C) ?

Manual injections with a syringe are quite difficult to do well.

Do you have a HS autosampler?

Are you keeping the syringe warm between injections?

Be sure to flush the syringe well.

I would avoid the untrasonic bath, using a good solvent will alleviate that requirement. You release trapped solvent by dissolving (or melting) the sample completely.

Cooling the sample must be done in a consistent manner but can be done.

Not easy and prone to error, however.

good luck.

[dcav]

The powder used to make stanndards does not contain cocoa butter, it is defatted with supercritical CO2 ( so it doesn't contain neither hexane)
The sample analyzed contain ten percent of butter, the same quantity we add to standards.

I'll try use no water and also higher temp, but with higher temp or melting i fear for some interference (moreover i think that cocoa powder eould melt at a much higher temp)

For the solvent propylene glycol could be a good idea since it has a lower vapor pressure.

I predicted some difficulties wiith manual injectons, in fact i believe we can't reach a precision lower than 5-10%
i think they don't have an autosampler, if they had they were quite sadic in making me do it manually.

The syringe is kept warm and stripped with a vacuum pump between the injections, but not for the whole run just for about six minutes (half run)

Many thanks for your suggestions, you are very kind

[oscarBAL]

Hello, I think that Dcav don't have to apologize for his inglish, the mine is worse.
I just want to comment something in this in this topic, is a Doubt that I ever have about this.

FID, is a mass sencitive detector, that means that all the hexane isomers, would have very symilar Rf, I analized hexanes before and other isomeric sample, or even gasoline that is somthing diferent, and if you are validating a method if you are using and standar n-hexane and think that your limit of detection could be 100ppb for example, but is that representative? because 100ppb correspond to one single peake, and 100ppb of hexanes in your real sample is the results of a distribution of areas among all the isomer, I don't know if I can explain my self? you are "disolving" or "dividing" your signal 5 times or n times, so near to the detection limits yuo are loosing accuracy and precision?

I know that n-C6H14 is very good to calculate the technical hexanes in the sample when you are working in the midle of your curve, but what hapens in the border, or near to the detection or cuantification limits?

Sorry dcav for use your topic to expound my doubt in your topic.

SPME works good for this?

Thanks.

[dcav]

Hello i'm here again with new problem!!

If i understood what oscarbal said, i should specify that we do standards with technical hexane because is what we are looking for, then each peak is separated and we can consider n-hexane peak only or the sum of all the peaks, thi will be a further step for us!
Moreover they're looking for bigger concentration (from one to several ppm) so sensitivity is not a problem even considering only a peak.

I couldn't been able of analysing sample without water yet, but i repeated the sample with water, powder and butter added with hexane.
The problem is that data collected seems completely casual!!
There is a few correlation, average signal seems to grow with concentration but the spot on the plot ( refering to the single conc) are very disperded, there isn't reproducibility!

We will try with no water and with lower temp, but i don't think it is a pressure prblem, beause i think it would affect accuracy, not precision, is it right?

Could be a problem of dis-homogeneus standard (hexane in butter)?
The butter had a big viscosity, and i obtained better results a week ago with first trial of external calibration or internal standard using other solvents for veiculate and diluite the hexane.

Any ideas?

Thanks a lot