By Anonymous on Monday, June 14, 2004 - 01:08 pm:

Dear All:

I am developing a method for ethanol concentrations in azeotropic range i.e 95% ethanol.
We are using HP 5890, HP-5 column, FID and n-propanol as the internal standard.
Detector temp-240C Injector temperature-150 C and Injection volume-1 ul (manually)

At these high concentrations of ethanol the linearity is around 0.9. Our regression value improves (0.99) if the range of ethanol concentrations are increased i.e beginning from as low as 20% to a high value of 100% ethanol.

If any of you have experience in analyzing ethanol at azeotropic conditions, please post your valuable suggestions.

Thanks

-------------------------------------------------------------------------------------------------------
By Anonymous on Friday, June 18, 2004 - 06:22 am:

Is it possible that the concs you are injecting are simply too high to get a linear response? Your sample prep isn't clear, but maybe you need to reduce the conc. of ethanol (dilute them) in your samples before you inject them.

-------------------------------------------------------------------------------------------------------
By Anonymous on Friday, June 18, 2004 - 07:04 am:

Dear Anonymous:

Thanks for the suggestion. This is how i prepare my sample prep:

For each standard, the weight of analyte (ethanol) added ranges from 1.0000g to 1.1000g, the weight of internal standard n-propanol is approximately 0.2500g and amount of water added ranges from 0.1000 to 0.1100g. I use an analytical balance and vials to make the sample prep.

I have also tried diluting in the ratio of 1:10, 1:5 (water:ethanol) it didnt help much. I do observe that my peaks are tailing and there is little bit of overloading the column. I am using splitless type.

Any advice in changing the method of preparing the sample prep is welcome.
Thanks

-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Friday, June 18, 2004 - 09:58 am:

I would use a polar column (e.g. PEG-type) at about 40C and split injection from autosampler (0.5ul). Hp-5 is not great for stuff this polar, and I have tried it for this. I've found it to be best and most economical in the long run to use the appropriate and best-suited columns/equipment to do the job right. Trying to make do and economize by restricting to "on-hand" columns is not good science. That's my opinion. I would prepare ethanol standard by pipetting 10.00 ml 99.7+% ethanol standard such as SD-40B 200 proof (most ethanol is reported %v/v but you can convert to %w/w later) into a 100 ml VF partially filled with water, then mixing and diluting to volume with water. Make up corresponding internal standard stock solution by diluting 20 ml n-propanol to 500 ml with water. Make Rf standard by mixing 10.00 of the prepared ethanol solution and 25 ml of the prepared n-propanol solution. For samples, quickly weigh 1.0 gram sample and mix with 25.00 ml of the prepared n-propanol internal standard solution. This adds the internal standard and dilutes the alcohols to a reasonable level. There is no need to use splitless injection, that's for trace analysis, you have plenty of analyte. Keep injection volume to 0.5ul, water expands a lot in the inlet. Do this, it works.

-------------------------------------------------------------------------------------------------------
By ab on Friday, June 18, 2004 - 10:24 am:

Consumer Products Guy,
do you agree with Thanks about possibility of getting good results making this analysis in splittless mode?
Thanks, can you show us chromatogramm you are getting?

-------------------------------------------------------------------------------------------------------
By Anonymous on Saturday, June 19, 2004 - 10:06 am:

Dear Consumer Products Guy:
Your suggestion has been extremely valuable. I have changed my column to HP-FFAP (crosslinked FFAP).
I have a question related to preparing of Rf standard. When you say "Make Rf standard by mixing 10.00 of the prepared ethanol solution" I didnt understand how the standards are going to vary. If I have to make 80, 90, and 95 % ethanol solutions how do I do it with your method? Please bear with me for not getting the concept right. If you can kindly elaborate on this, it would be very helpful.

Thanks a lot

-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Wednesday, June 23, 2004 - 08:21 am:

From the stock solution of 10.00 ml 99.7+% ethanol diluted to 100 ml with water, take (for example) 7.00 ml, 8.00 ml, 9.00 ml, and 10.00 ml aliquots into separate vials, each containing 25.00 ml of the prepared n-propyl alcohol internal standard solution in water. Chromatogram attached, if I am successful (hey, I'm a Consumer Products Guy, not a Computer Guy !).
image/pjpegethanol.cdf
ethanol.jpg (17 k)

-------------------------------------------------------------------------------------------------------
By Anonymous on Thursday, June 24, 2004 - 12:19 am:


-------------------------------------------------------------------------------------------------------
By ab on Thursday, June 24, 2004 - 12:47 am:

Consumer Product Guy!
Yor peaks are overloaded. This effect slightly softened by tailing. I think the concentration is too high. Peaks are too wide, I would say. A width 18-24 sec is too much for capillary GC.

-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Thursday, June 24, 2004 - 08:36 am:

Yes, for this assay concentrations are quite high, but we're confined to staying close to an AOAC test method for cGMP requirements (just changing to capillary GC from the packed column specified in the AOAC). This is a 0.5ul injection on a 0.53mm i.d. column. Results are extremely reproducible, and frankly, we aren't staffed sufficiently to try to make everything perfect, we have to move on to other pressing assignments. Obviously, the integration parameters need to be tweaked for the above chromatogram as well, I just pasted this in from an old document. If we tried for perfection at all times, we'd never get anything done.

-------------------------------------------------------------------------------------------------------
By Anonymous on Thursday, July 1, 2004 - 10:13 pm:

Consumer Products Guy,

Again its a fundamental question (rather a trivial doubt)
I did exactly the same for stock and internal standard solution (like how you suggested). i.e.,
Stock solution = 10ml ethanol/100 ml of water (volumetric flask)
Concentration of stock solution is = 10% v/v * 0.789 (density of 99.9% v/v - ethanol)= 7.89 g/ml

Internal standard solution = 20ml n-propanol / 500 ml water (volumetric flask)

Standards: 8ml of ethanol stock solution + 25 ml of internal standard solution.

If I were to plot on my x-axis the concentration of ethanol, then its range differs than what i want.
for eg: (8ml ethanol stock solution * Concentration of stock solution)/ Total Mass of solution.
Total mass of solution = (mass of n-propanol solution + mass of ethanol solution taken)
.....gives a value of 1.95 w/w of ethanol ! ..

I would like to have on my x-axis the concentration of ethanol from 80 to 100....
Also,how can such a dilute stock solution (10%v/v) contribute to making high ethanol standards??

If i have understood ur method right, i must be having 80% ethanol concentration but i get a totally differnt value......I am sure I am missing something totally basic here...Please let me know how you would interpret this...

-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Friday, July 2, 2004 - 07:50 am:

For internal standard, all that is really relevant is the ratio of analyte to internal standard. For high ethanol levels, make up samples essentially the same as you do the standard. See AOAC Official Method 973.23, Official Methods of Analysis of AOAC International, 16th edition, 36.1.01.

-------------------------------------------------------------------------------------------------------
By ChemE on Wednesday, July 7, 2004 - 04:15 pm:

Dear All:
Thanks for all the advice, I have been able to develop a method that is fairly linear (r2 = 0.97). In particular, Thank you Consumer Products Guy...

-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Thursday, July 8, 2004 - 08:19 am:

You're welcome, ChemE. There's much to be said towards leaning to the practical, one can't be "perfect" in every assay, there's not enough time, business need, or manpower. Suitability is making sure the assay is suitable for the job at hand, and its requirements (who cares if there's no accuracy at ppm levels if your analytes are in the percent-plus range?). I've made a good living in this business with that philosophy.