help with 5890 GC (on column, FID, MSTFA, ASTM D6584, etc)

[jf]

hi everybody, first post to this forum. I have a bunch of questions related to our 5890gc. we are chemical engineers, not analytical chemists, so we have a good understanding of the gc, but little practical experience. Were trying to get this up and running and need some pointers on what is 'normal'. We had an incident which might have damaged our column, need some help getting going.

a little background:
astm d6584 uses mstfa to derivatize fatty-acid methyl esters, glycerol, and mono,di, and triglycerides. The goal is determination of free and total glycerin in biodiesel. Here is some more info:

http://www.sigmaaldrich.com/supelco/bul ... 107943.pdf


We are using manual injection with cool-on-column injector, and a 5m guard column

the method recommends injector temp of 50C with presumably temperature tracking oven temp

the person who sold us the gc was highly critical of the method and recommended injection at 100C with a 40C/min injector zone temp ramp rate to 380C. I assume this was to assure complete volatilization of the solvent to prevent tailing and protect the column

the oven temperature ramps at the following rates
hold at 50C for 1 min
15C/min to 180
7C/min to 230
30C/min to 380

FID detector is operated at 380C

we are using helium for the carrier gas

we created our standard and stock solutions last week, and shot a few to attempt to establish our calibration. Of our 4 standards (glycerol, mono-, di-, and tri-olein), only monoolein came out with a decent correlation coefficient. glycerol was particularly hard to detect as it was obscured by the tail of the heptane peak.

We decided to re-run the injections the next day, but we ran into an 'EPPB system shutdown'. It took us a few days to figure out that this was likely due to the injector septa, which we replaced. in the meantime, there was a sample that got left in there with no He purge (EPC shut itself down), and i'm afraid it damaged our column.

I made some new standard solutions this week, but everytime i inject a sample, we only get a few short broad peaks and the tailing of the heptane peak has gotten worse. I'm not seeing any of the internal standards or correlation standards. To try to figure out whats going on, I did a blank run:



here is a run with internal standards:



note that the attenuation was set to zero on this, i put a marker in to indicate the time and peak height at that point.

I'm not sure if this is due to poor column resolution due to contamination or maybe our mstfa got deactivated. (I purged the mstfa bottle with helium the best i could and kept it in the refrigerator all week. While not a perfect purge, i would expect it to keep longer than a week under those conditions, and at least show some activity.)

so in summary i have 4 possible thoughts here:

1. column/injector is badly contaminated, thus little to no resolution

2. the stock solutions i made last week (glycerol/monoolein/diolein/triolein/butanetriol/tricaprin, all in pyridine) have broken down and are unusable (these were stored in septa vials in the freezer. while not perfectly purged, i expect some resolution, so i'm gonna go out on a limb and say that #2 is less likely

3. MSTFA got de-activated, despite my best efforts

4. some combination of #1 and #3

I have a bunch more questions, but for now, if i could get some assistance in cleaning injector, and/or baking/trimming/saving my column and somehow verifying i have useful standards, that would be a great starting point.

thanks in advance!

[jf]

also, the column we are usin gis a j&w db-5, part number 123-5711

j&w's catalog lists this as having a maximum temp rating of 350C. I remember when the unit was sold to us, i was told "there really are no columns rated higher than that".

if i do bake this column, should i bake it at 350 or 380? and for how long?

[Bruce Hamilton]

Final temperature for column - you have the high temperature version, 123-5711, which is rated to 400C. Conditioning the column at High temerature should only happen when the baseline is too high at elevated temperatures.

You need to perform some on-column injections of purchased standards or Internal standards, or some other simple molecules to confirm that you have all of the GC parameters set correctly. Once you know that, you can worry about the derivatisation.

The column should not die unless you let it see air whilst at high temperature. If you think it's dead, prepare and analyse the test mix using the same condistions as the CoA. If the peaks tail badly, or don't resolve, your column may be dead.

Use the MSTFA to derivatise a standard compound, such as glycerol, monostearin, or the specified Internal Standards. If the monostearin peak doesn't appear, or is badly tailing, work on reducing the active sites of your GC system by repeated injections of the sample - a few injections should result in inceased peak size and improved peak shape.

Only condition the column at high temperature if you suspect rubbish has accumulated on the column and needs to be removed. If you are only injecting a few standards, that should not be a problesm.

Bruce Hamilton

[jf]

You need to perform some on-column injections of purchased standards or Internal standards, or some other simple molecules to confirm that you have all of the GC parameters set correctly. Once you know that, you can worry about the derivatisation.

Use the MSTFA to derivatise a standard compound, such as glycerol, monostearin, or the specified Internal Standards. If the monostearin peak doesn't appear, or is badly tailing, work on reducing the active sites of your GC system by repeated injections of the sample - a few injections should result in inceased peak size and improved peak shape.

Bruce Hamilton

ok, i will try to derivatize some new standards in the morning and re-inject, perhaps i didnt shake them long enough. I have a ton of glycerol so i can remake the glycerol/pyridine stock solution from scratch if needed.

which 'GC parameters' are you referring to? the flows are within a defined range, im not sure what i would change to sharpen up the peaks

Any recommendation on how to reduce the solvent tailing?

[jf]

You need to perform some on-column injections of purchased standards or Internal standards, or some other simple molecules to confirm that you have all of the GC parameters set correctly. Once you know that, you can worry about the derivatisation.

Bruce Hamilton

unfortunately, even the internal standards require derivitization.

something has definitely changed. our peaks were fairly sharp on the first day

[Consumer Products Guy]

We use trimethylsilylation (but use BSTFA) to assay for stuff like glycerin, monoglycerides, diglycerides, and triglycerides all the time here. Hell, we've even published a handful of them. When we look for glycols, propylene glycol, glycerin, sorbitol, etc., we generally use a DB-5 or DB-1 type column, works fine, peaks are nice and sharp. When we want to look at the mono, di-, and triglycerides we generally use a column designed for high temperature use, like metal capillary from Quadrex, 10 m with thin film, which can go 400 C easy. Of course, the triglycerides are not derivatized here, just the others. It seems like the inlet temperature needs to be higher (do you really have a temperature-programmable inlet?). I'd probably use like a 325 C inlet temperature and keep it here.

[jf]

(do you really have a temperature-programmable inlet?). I'd probably use like a 325 C inlet temperature and keep it here.

it starts at 100C and then ramps to 380 at 40C/min

the following is from Gerstel's 'Sample introduction techniques for capillary gas chromatography':

"There are many opinions on the selection of optimum injection temperature for on-column injection. Good results are generally obtained wth an inlet temperature between 20C below and 10C above the boiling point of the solvent. At higher temperatures, selective vaporization from the needle can take place"

the astm method says 50C, our gc salesman says 100, you say 325. what are the advantages/disadvantages of these extremes? i would think selective vaporization would cause residue buildup on the syringe over time.

[Bruce Hamilton]

If you want to report results to the ASTM method, you must follow the method, including using the correct columns and conditions. I believe it was revised this year ( just added precision statement I think ), so you should be work from -07a version if you plan to report results to the method.

They specify cool on column injection because that works with most of the oils used to prepare the FAMES. Highly-unsaturated derived components may polymerise and/or degrade at higher temperatures.

My suggestion about the parameters was to ensure that the peaks were eluting at close to the expected retention window and are of suitable shape etc. Parameters covers everything from column perforamnce to GC conditions to integration settings.

If you don't have any known good derivative standards, you can make some. It doesn't really matter which TMS reagent you use for that experiment, MSTFA is used because it, and it's byproduct, ( N-methyltrifluoroacetamide ) are volatile and don't elute in the regions of interest.

You can also often add 1% of TMCS to the MSTFA to help many reactions. Whichever derivatisation method method you use, ensure that you keep moisture out.

If you started with sharp peaks at the expected retention, and the shape has degraded quickly, I'd guess that you have injected rubbish onto the column - conditioning may or may not work, but I would check the column against the COA performance before doing anything dramatic.

The problem appears to be that aren't able to demonstrate that the derivatisation has worked, or that you system is correctly set up. You need to fix the latter to work on the former.

Bruce Hamilton

[Peter Apps]

What a pleasure to have a question with all the technical detail, and some chromatograms included .

A quick and straightforward fix might be to change the guard column: with luck any crud should be sitting there rather than on the analytical column - that's why they are called guard columns.

The idea behind cool on column injection is to avoid the poor repeatability and molecular weight discrimination that you get with vapourizing injections. The "cool" is the important part.

Peter

[chromatographer1]

There are a couple of options for you to evaluate:

Two metal columns and one FSOT column have proven themselves for this analysis in my experience.

One is from Varian, the second is from Restek: Biodiesel Select and MTX-500, the first is a 5% phenyl methyl silicone, the second is a Carborane methyl silicone. Both will work well. I have also tested a SGE FSOT high temperature 5% Phenyl which gave excellent results.

The triglyceride analyte will tend to degrade if a high injection temperature is used. That is why the test uses a cool on-column technique.

It sounds like you are trying a heated injector splitless injection technique.

Do not do this. Follow the method. If you do not have a cool on-column injector get one and the special needle it requires.

best wishes,

Rod

[Consumer Products Guy]

Make some changes, one at a time, see what (1) helps, (2) hurts, or (3) has no effect. Just because a procedure is "ASTM" or "USP" doesn't mean that it's the best there is, could be outdated, obsolete, or just something that's "OK" but there are no resources to do a round-robin to update it, or they're fearful that people would need to upgrade their equipment. Look at USP web site, they list that they have a gap because their test procedure for sodium chloride is "not stability-indicating".

[jf]

nope, its cool-on-column with the direct manual injection

It sounds like you are trying a heated injector splitless injection technique.

Do not do this. Follow the method. If you do not have a cool on-column injector get one and the special needle it requires.

best wishes,

Rod

[chromatographer1]

Some comments:

If your vials were capped with a septum that had been pierced then enough water can get into the vial to consume the silylating reagent.

Then your glycerol peak can look like a blob or disappear completely.

Having the pyridine sitting in your column with no gas flow did not help that methyl silicone column's health.

Perhaps you got a piece of septum into your column. This shut down your EPC as it could not meet the flow rate if the column was blocked.

Do you have the proper flow rate established? You can shoot a syringe of natural gas to determine if the flow rate is acceptable.

The best thing to do is to get a new column and start over. Be sure you make a good connection for your guard column and analytical column connector and NO GRAPHITE particles into your column.

Good luck. Biodiesel analysis is not a trivial task, but I am sure you will achieve success.

best wishes,

Rod

[jf]

heres what i did:

created the stock solutions (glycerol, mono-, di-, and tri-olein, butanetriol, and tricaprin in pyridine). these are stored in septa vials in the freezer. they were blanketed with helium, but stored with pierced septa for a few days before i wised up

i am using these to create the standards (uL quantities of each of the above stocks+MSTFA>silate>add heptane>inject

the original MSTFA vial is brought out of the fridge, brought to room temperature before opening to prevent condensation, then blanketed with helium and re-fridged

do you think the stock solutions could have gone bad? we are in a very dry climate, i wouldnt expect the pyridine to absorb much water.

i'm leaning toward column contamination. I'm trying another injection shortly and will report back

Some comments:

If your vials were capped with a septum that had been pierced then enough water can get into the vial to consume the silylating reagent.

Then your glycerol peak can look like a blob or disappear completely.

[jf]

its very odd, i know my column has the correctly set flows and the FID is working, but every standard solution looks just like a blank run. I am getting no peaks at all other than the initial solvent peak. Note that the scale on the following chromatograph is zoomed, the top of the y-axis corresponds to a signal of 11500




I suppose i should order new standards and a new column? does this definitvely indicate bad column or bad standards?