Rules for solvent selection in GC analysis?

[sonicboom12345]

Hello all,

What are the rules of thumb for selecting a solvent?

For example, when working with a DB-5MS or a ZB-5MS column, which solvents are preferred and which are to be avoided? What about for a column with a polar stationary phase, such as PEG?

Is it safe to use a solvent like acetone on a non-polar column? How is one to determine which solvents may harm the interior coating?

Thank you!

[Don_Hilton]

You want to avoid corrosive solvents - they will destroy the column.

You would like the solvent to have a lower boiling point than the analytes of interst. This will allow you to get the solvent off the column before your analyte elutes -- and with a bit of luck you can get some nice solvent focusing out of the deal.

Acetone is fine, as are pentane, DCM (or methylene chloride), ethanol, ethyl acetate, THF, furan, benzene, and....

Methanol can be a bit tricky on a DB-1 or DB-5, but if you have the GC oven above the boiling point of the solvent it works well also.

The lower the molecuar weight, the greater the expansion volume. So if you are going to use a PTV inlet or use water with a standard split splitless inlet, you have to be careful of injection volumes.

[sonicboom12345]

You want to avoid corrosive solvents - they will destroy the column.

You would like the solvent to have a lower boiling point than the analytes of interst. This will allow you to get the solvent off the column before your analyte elutes -- and with a bit of luck you can get some nice solvent focusing out of the deal.

Acetone is fine, as are pentane, DCM (or methylene chloride), ethanol, ethyl acetate, THF, furan, benzene, and....

Methanol can be a bit tricky on a DB-1 or DB-5, but if you have the GC oven above the boiling point of the solvent it works well also.

The lower the molecuar weight, the greater the expansion volume. So if you are going to use a PTV inlet or use water with a standard split splitless inlet, you have to be careful of injection volumes.

Please explain what you mean by corrosive solvents. What is this determined by? Acidity/Basicity? Or are there some other rules at play? Can you give an example of some such corrosive solvents as you are talking about?

Also, what is it that makes methanol particularly "tricky?"

[Don_Hilton]

I would avoid solvent mixtures that contain strongly acidic or strongly basic compunds - for example a solvent mixture containing a significant quantity of HCl. And, this is a judgement call. In some derifitizations (such as when trimethylchlorosilane is used) you generate HCl in the derivitization - and ignore it. On the other hand if you are analyzing a reaction mixture that has been run in methanolic HCL, it is probably worth the time to get rid of the HCl, if possible.

Such a solvent may be your only option - at which point the colulmn will degrade after a number of injections. You do the calculation. If a $400 column will last for only 100 samples, that is a cost of $4 per sample - which may be a reasonable cost. And added in with solvents, supplies, etc., may not be really significant in the total cost of an analysis.

And, by the way, avoid solvents with non-volitile components. Thus a solution which contains NaOH is bad - not just because it is corrosive, but because the NaOH residues will accumulate in the liner and may accumulate on the head of the column as well.

On methanol being tricky - I've already told you the solution to the trick -- keeping the oven tempeture high enough. Methanol does not wet a DB-1 or DB-5 type column. So if you inject into a cool GC oven, the solvent condenses, but does not form a nice band at the head of the column like other solvetns. Rather it beads up and spreads along the column. The resulting chromatogram will show splilt and otherwise ugly peaks.

Other solvents form a nice band and focus the analytes in this liquid. And this is how you get solvent focusing. The solvent evaporates from the back of the band, making the band narrower as the solvent evaporates. The analytes are trapped in the liquid solvent until the band becomes so narrow that it goes away. Then the analytes start their journey down the column, startign in a narrow band, which offers the potential of narrow peaks at the end of the trip. Analytes in condensed methanol on the DB-1 or DB-5 column start scattered along a portion of the column.

[BMU_VMW]

Dear Don_Hilton,

You list DM as a good GC solvent.
In your opinion: would DM break down and form HCl when a liner is heated up to aprox 300°C ?
Would the HCl formed here have a significant effect on liner and column inertion (50µl / injection) ?

kind regards.

[sonicboom12345]

Pardon my ignorance, Mr. Hilton. You seem to be a well-informed authority on the subject, whereas I am but a neophyte! They taught us the fundamentals of chromatography in undergrad, but they never taught us anything about solvent focusing or expansion volumes!

Maybe you can clear a couple more things up for me.

My understanding of solvent focusing is that you *want* the oven temperature to be a good 20-25C below the boiling point of the solvent, so that the solvent condenses at the head of the column and forms a good trap for the analyte. If I am understanding you correctly, what you are saying is that this does not work for methanol, because it doesn't "wet" a DB-5 type column (i.e. it does not condense, for some reason??) So what you are saying is that if you use methanol on a DB-5, keep the starting oven temperature at or above methanol's boiling point, which is 65C.

I am actually interested in using methanol for one of my analyses. I am using a 60m ZB-5MS column with H2 carrier gas at a flow rate of 2.0mL/m. In this circumstance, what would you advise to be a good starting temperature for methanol? Conversely, what would be a good starting temp for DCM?

[Don_Hilton]

to BMU_VMW: I've probably made microliter injections of DCM into an inlet that hot and never given it a second thought.

An inlet at 300 and 50 microliters - that is a lot of temperature and a lot of solvent. With the volume, I assume this is not an injection into the normal split/splitless inlet unless you have a precolumn in a GC oven cool enough to condense the DCM and wide enough to handle the rapid gas transfer from the inlet.

You do not indicate wether you are talking about injection for a packed column or a capillary column. So, I 'll keep rambling on thinking in terms of cappilary.

Dow, on their website notes that DCM decomposes above 120 deg C in the presence of metal surfaces. With a 1 microliter injection made with a fast injection (typical of many autosamplers) the presence of the needle would not be a problem - it is not going to heat up that fast. To inject 50 microliters, a fast injection into a 300 degree inlet is going to generat a lot of vapor - and as fast as it is injected, so to avoid streaming vapors into the inlet you would, most likely have to do a slow injection - which would allow the needle to heat. This may or may not be sufficient to be a bit of a problem.

However I'd like to step back and ask why 300 degrees? I do have a few types of samples that have requried an inlet this hot, but some types of molecules are destroyed at this temperature. I'd rather to a cool injection in a PTV inlet, venting the solvent, and then ramp the temperature up, transfering the analytes to the column. This is easier on some of the inlet components and on the more labile analytes in the mixture.

to sonicboom12345: Methanol does condense, but in scattered droplets down a methyl slicone column, not a nice band. So, an injection above the boiling point of methanol works much better. It's been a while since I've set up anything using methanol as a solvent into the GC. I think I used something like 85 degrees. It may have been a bit cooler. DCM, I have injected as low as about 40 degrees.

When you do try for solvent focusing, keep the temperature below the boiling point of the solvent for a minute or so, allowing the sovent to gently evaporate and focus the band. Then, ramp up to what you need to elute your analytes.

Having said that, the longer the analytes stay on the column after the solvent elutes, the less solvent effect you have. So in the analysis of high boiling compunds you spend a lot of time cooling, focusing, and ramping up for little or no effect. In that case, inject into a hot oven - get the job done, and move on.

[BMU_VMW]

@ Don Hilton,

I'm sorry for not being to clear in my question.

We use a ptv injector at 50°C (we can't cool any lower); inject 50µl DM at once in a glass bead sinterd liner and than heat up to 300°C. Off course most of the DM is splitt off. there is a 5m pre-column and a 30m Rxi5 sil MS integra-guard.

** We use the 5m integra-guard to place it into the transfer line to avoid breakdown of column-phase in the transferline resulting in active places in the column. **

Question is: is the DM decomposing and forming HCl in this proces of heating and evaporating? and if so: would the amount of HCl formed be enough to have a negative influence on the chromatography within a span of about 200-300 injections ?

@ sonicboom: in our DM methode oven temp is at 40°C when injecting.

[Don_Hilton]

I doubt that you have any decomposition of the DCM to worry about. I presume that you hold the inlet at 50 degrees long enough for the liquid to evaorate - and whatever DCM is left quickly passes onto the column. Even if some DCM is present when the inlet reaches 300 degrees, I don't expect a problem.

[sonicboom12345]

How much similarity is there between the DB5 and the ZB5 column? Can methanol be expected to behave the same way?

[chromatographer1]

The columns SHOULD be very similar and act accordingly.

best wishes,

Rod