not able to separate benzene, isopropanol & n-propanol o

[Amos]

Hi I'm new to the GC. I need a method for calibrating a GC (preparing the standards) for a ternary mixture of benzene, isopropanol & n-propanol. At what temperature should my oven be when I iject this mixture after distillation? Thank you so much for your assistance.

[Rick Jagielski]

n-propanol has a Bp of 97.2 C and IPA has a Bp of 82.5 C,

You will have to have a column that can separate these two, there will be some differences in polarity, so the easiest way to go would probably be to check application notes for some of the more common column manufacturers.

If you have a column of the appropriate polarity, it should be a fairly easy separation based on a slow (4 - 10 C) ramp/min starting a decent interval below IPA's Bp, say 75C.

Adjust pressure and flow settings so that you get a good focus at the head of the column, and optimize your settings so that the peaks have good shape, the run time is acceptable ect.

You mention that this is analysis of a distallate, which component is the primary solvent? and are you trying to quantify all three, or just two as minor components in the third? This will determine how you mix your standard.

Rick

[Consumer Products Guy]

One can readily separate IPA (or ethanol) from n-propanol on a PEG-type capillary at low temperatures like 40C.

[Amos]

Hi Rick

I need to quantify all three componets. First I need to do the distillation and then analyse both the residue and the distillate so as to find the azeotrope (point) temperature & composition i.e where the vapour and the residue are at equilibrium and I'm doing it manually not using the head space analysis.

[chromatographer1]

A simple porous polymer column like Porapak Q will separate IPA, nPrOH, and BZ without trouble or much effort.

You may use glass or FSOT coated SS packed columns,1/4 or 1/8", or a even a capillary plot column will work.

Use an oven temperature of 180 to 220°C (isothermally is fine) and use carrier gas flow somewhere between 20 and 80cc/min.

[Rick Jagielski]

I personally have little experience with packed columns, I have separated these alchols on a capillary column.

I believe that the bianary azeotropic mixtures will optimize at the following ratios. Not sure how this will interact with a trianary mixture, I guess that is what your experment is for, of course exact compositions can vary as the temperature changes.

17 % 1-Propanol (nPA) & Benzene with a Bp77.1
33 % 2-Propanol (IPA) & Benzene with a Bp71.5

However knowing these values gives us a ballpark to start with for a standard prep, first pick a solvent that will not co-elute with your analytes but will solubilize all three analytes. (maybe something like octanol or decanol)

You want to inject an amount that will not overload your column with the highest concentration analyte. In this case benzene. Start by making a solution of benzene in your chozen solvent. Probably around 0.5mg/ml of solvent (you can do this on a mg benzene/mg solvent basis and increase your accuracy by avoiding volumetric measurements or use a dilution scheme to get a decent amount of material when you weigh it out). Run a few injections of this mixture checking for column overloading by watching what happens when you inject larger amounts, if you are seeing the peaks broaden or if you start seeing bad tailing reduce the injected amount until you have a pretty peak. You can watch peak height vs peak area ratios to see if the peak has stopped changing in height and started to broaden. Either decrease the injected amount, change split ratio, or change your concentration until your response is about 2/3 to 1/2 of the value where you see overloading occuring, or stop when you have a good sized signal that you feel comfortable with.

You can then run linearity to determine what concentration range would be best for your benzene peak, and your alchol peaks.

When you have determined what concentration will work well, prepare a standard with the optimum amount of benzene, and relative amounts of your two alchols say nPA @ 20% of the amount of benzene added, and IPA at 35% of the benzene concentration. (if you are planning on dealing with samples that may be close to 100% nPA or IPA you could mix them at equal concentrations and run them to test for linearity.

Prepare solutions (equal to the number of points you want usually n>5) ranging from 50% to 150% of these values and try running them with triplicate injections to see if you can get a linear curve.