methanol/isopropyl separation on db5-ht?

[AZBiodiesel]

Hey guys,

I was wondering, is it possible to separate methanol and isopropanol in the gas phase on an agilent db5-ht?
I don't think it is, since it is not a polar column? I see a good methanol peak which flies through after about a minute, but with the isopropyl it just doubles the peak pretty much, i don't get separation.

I am doing headspace analysis with gas tight syringes. I have made standards at 0.01%, 0.2%, 0.5% and 1% methanol and get a nice linear curve and r squared of 0.994, but my response factors seem out of whack, so i am trying to use an internal standard. However I am limited to only using a db5-ht for the time being. Should I try for another compound to use as the internal standard?

I am taking 2mL sample, heating in a 12 mL vial to 80 deg C for 5 minutes, and injecting 200uL of headspace directly on the column. Helium at 3mL/min, injector port at 70degC , oven isothermal 70 deg C for 6 minutes. Should I try a cooler oven temp and then ramp up temp slower? I am also trying to do these tests in the shortest time as possible.

Thanks!

[chromatographer1]

Film thickness is everything and of course, the length and ID of your column.

If you are using a 0.53mm ID 15 meter column with 0.16µm film, BAD CHOICE.

You may wish to increase the temperature of your injector to 120°C - 140°C

Lower your oven to 50°C isothermal.

Use a low volume injection liner and inject slowly.

I think if you check the Supelco or Restek or Alltech web sites you will find chromatograms of methanol and ipa on a methyl silicone phase column for comparison.

best wishes,

ROd

[Peter Apps]

Presuming that you made you standards in a matrix that matches you samples, and since you get a good (enough ?) linear fit, you might not need an internal standard at all. I'm not sure what you mean by "response factor" - methanol gives a lower FID signal than the same mass of hydrocarbon, and lower than the same mass of isopropanol, because it has proportionally less C-C and C-H in it.

Given that you cannot change the column, low oven temperatures are your friend.

Peter

[AZBiodiesel]

thanks for the help guys!

Ok on the math side I was under the assumption that to determine the quantity of an unknown analyte, you have to first run a standard and get the response factor= (standard peak area)/(concentration of analyte in standard). This is a single point standard. I am using a concentration near to what I think it should be of the unknown analayte.
Once i get the response factor, I take the area of the unknown sample's peak and divide it by the response factor... But when I look at the response factors from 0.01% to 0.5% they all give different percentages for the unknown analyte amount. Shouldn't they all be roughly the same? I am seeing differences 0.2%, 0.4% and 4% all for the same sample.

So I thought I would need an internal standard to correct for the variation in headspace sampling as well as the GC. I want to test how reproducible my headspace sampling methods are as well as get an accurate number of %methanol.

On the sampling side I am using our biodiesel as my matrix its water washed multiple times and heated to 100 deg C for 30 min. I then spike this with methanol. I do all injections with a heated needle and prime it 3 times before piercing the headspace vial, and then prime it 3 times in the vial itself. I then inject.

I have tried your method chromatographer but I am getting random peaks and noisy baseline when i heat my inlet to 120 degC, could be contamination/septa bleed?

[Peter Apps]

Just so that I am absolutley sure - your are trying to measure how much methanol there is in you biodiesel ?

Given that you have a linear fit of peak area vs amount of methanol the response factor must be the same at the different levels of your calibration. If the response factor (signal/analyte mass) changed with analyte mass the fit could not be linear.

Since you already have a 4-point calibration, calculate the equation for the line of best fit (peak area = x times analyte mass + y, where x is the slope of the line and y the intercept), then run a sample and calculate analyte mass from that equation.

Peter

[AICMM]

AZbiodiesel,

The HT series have almost no film thickness to them at all. This is so they can go to very high temperatures to measure high MW hydrocarbons. Plus, they're non-polar phases. I would say that the only way you are going to separate meoh and IPA is with cryo at something like -20. Thus, your other idea of picking some other constituent other than IPA might work but, to get your separation, you are going to end up with a component that is not anywhere near the volatility of meoh. Is that a good choice for IS, I don't know.

For what it is worth, you could consider an external iso-thermal oven (a valve oven works great for this) for meoh analysis on the same GC.

Best regards,

AICMM