Splitless Injection Issues - Larger injection, less signal

[mbanks]

I'm developing a GC splitless method for trace impurities in BPA samples that is giving me a problem I can't quite figure out.

Some basic info:
Agilient 7890 with MSD (to be transferred to FID later)
Column: HP-1 ms, 60m x 0.25um x 250um
Oven: 40 C (held during splitless time), slow ramp (5C to 100C, then 10C/min to 300)

Injector: 0.3uL into .9mL single goose-neck with wool into splitless injector. 2min splitless time, 350C temp. Flows are set so 0.9mL/s for the sake of the MSD vacuum, comes out to about 15psi in injector.

Sample info: ~4wt% BPA in acetonitrile w/ BSTFA:TMCS for silylation

This issue that I am having is that the more sample I inject, the worse signal I am getting from my lighter eluting peaks. My lighter peaks are getting less signal the more sample I inject, while the heavier peaks grow larger (as I would expect). Right now, even going from 0.3uL to 0.5uL results in less signal for the lighter peaks, though using the vapor calculator, both should have no issue as far as overflowing the injector.

What can cause this? Where are these compounds eluting? Is this some problem with my solvent trap? I'd like to inject as much sample as possible given practical limitations, but am not sure why I am losing my lighter compounds. Any advice much appreciated.

[Peter Apps]

Welcome to the forum.

Solvent effects are a possibility - expand the vertical scale on the chromatogram from 0.5 ul and see if there is a shallow step a few tens of seconds wide in front of each lighter peak - that is where your lost peak area is sitting.

What size syringe are you using, and are you doing fast, cold needle injections or letting the needle warm up with a pre-injection dwell ?

What is a "solvent trap" in this context ?

Peter

[mbanks]

Welcome to the forum.

Solvent effects are a possibility - expand the vertical scale on the chromatogram from 0.5 ul and see if there is a shallow step a few tens of seconds wide in front of each lighter peak - that is where your lost peak area is sitting.

What size syringe are you using, and are you doing fast, cold needle injections or letting the needle warm up with a pre-injection dwell ?

What is a "solvent trap" in this context ?

Peter

Looking at the chromatogram, I'm not seeing anything in front of these peaks. Though the S/N for these peaks isn't large, maybe this effect is masked in the baseline?

I'm using a 10uL syringe with fast injection and no dwell. Would using something different help?

Solvent trap is not a very good phrase I used. More like solvent focusing. Since i'm starting well below the boiling point of Acetonitrile (80C), I should get some solvent focusing for the lighter compounds. I'm thinking solvent may be affecting these peaks. Since the heavier peaks have much higher boiling points than the start temp (>150C), these may be fully cold trapped and not affected by the solvent.

Assuming it is like you suggest and the additional solvent is causing the peaks to front, what should I do to minimize this? Longer hold at lower temp to allow more solvent to evaporate off?

[Rndirk]

I'm also thinking in the direction of your light compounds eluting together with a big solvent tail. It would actually be more helpful to see a chromatogram with FID, because your MSD masks the acetonitrile. What's the boiling point of your first compound?

One way of having an indication when the acetonitrile reaches the detector is looking at the vacuum; you could see a dip there.

I suggest to increase the flow and lower the splitless time.

Flows are set so 0.9mL/s for the sake of the MSD vacuum, comes out to about 15psi in injector. .

I guess this is 0.9 mL/min. I don't understand your reasoning behind this rather low flow. If you put the right dimensions of your column in the software, and specify that the outlet is MSD, it takes the vacuum into account while calculating the pressure.

I would also increase the starting oven temperature. You're right about solvent focusing, but for acetonitrile 60-65°C should be OK. You might be condensing too much.

You can also check a recent topicof mine, i was having a similar problem with early compound eluting with solvent.

[mbanks]

I would also increase the starting oven temperature. You're right about solvent focusing, but for acetonitrile 60-65°C should be OK. You might be condensing too much.

Would you explain this a bit more? I'm not too experienced with splitless or solvent trapping. Is condensing too much solvent a bad thing? Will this cause broader peaks? What should I do to optimize using a solvent trapping method?

I could also stand to flow a bit faster, i think I can go up to 4mL/min with our pump (turbo), so I'll probably shoot for a more efficient flow. Thanks for pointing this out.

How do you mean the MSD masks the solvent tail? Wouldn't I be able to see that? First boiling compound is fairly high, it's TMS derivitized phenol, boiling point around 160C.

How would lowering the splitless time help me, other than transferring less sample onto the column? Shouldn't the refocusing largely mitigate any peak broadening due to long splitless time? I chose a long splitless time to reduce weight discrimination as much as possible, as I have a wide rage of MW at low levels.

[Peter Apps]

Let's go with the simple things first. A 10 ul syringe is very inaccurate at volumes below 1 ul, so a 5 ul would be better. The smaller early peaks with 0.5 ul cannot be due only to inaccurate injection volume though, because the late peaks were bigger.

Unless you definitely need solvent effects to focus early eluters it is best to increase the column temperature to deliberately avoid them and test whether thermal and stationary phase focussing is adequate.

I would certainly run the gas flow at the column's optimum flow - you are running slower than the 35 cm/s optimum linear velocity, and the only effect is to waste time.

If your only analyte is BPA (BisPhenol A ?) then why are you seeing early and late peaks ? What else is in there ?

Peter

[mbanks]

If your only analyte is BPA (BisPhenol A ?) then why are you seeing early and late peaks ? What else is in there ?

Peter

I'm looking for trace impurities in the BPA (correct, bisphenol A) sample. There are some lighter than BPA impurities, such as phenol (a reactant in the process to make BPA) and others, an isomer of BPA (p,p-BPA is the desired product, some o,p-BPA is also formed), and heavier than BPA contaminants, such as Trisphenol (BPX) and other heavier molecules in small quantities (including some color bodies that are of special interest).

So what you end up seeing is a chromotogram with a very large, broad BPA peak (being high purity BPA sample) and small peaks for contaminants. Separation has not actually been too much of an issue, even for the isomers of BPA. I can get a good chromatogram at 0.3uL injection, but would like to inject more (say 1.0uL) to get better S/N for the smaller peaks.

I will certainly increase the flow, no point is making the runs longer than needed, but am still vexed by my inability to shoot anything but the smallest injections.

I haven't tried shooting higher than the bp of acetonitrile. I'm worried the lighter peaks will be very broad without solvent refocusing. TMS-Phenol is around 160C, so starting at say 90-100C will result in broad peaks I fear.

[Peter Apps]

Don't worry about wider early peaks until you see wider early peaks - thermal and statioanry phase focussing works wonders, and by trying to add solvent focussing before you even have a chromatogram that shows you need it is making things unneccesarily complicated.

Try a default splitless injection; 1 ul, 1 min splitless, column start temp 80C. Shorten the splitless time to 30s, to see if peaks get smaller. If not use 30s, if they do go back to 1 min. If the early peaks are wide drop colum temperature to 60C to see if they improve.

I just love simple methods - have you tried without derivatization ? Columns these days are so well deactivated that you can get phenols through them as sharp peaks with no problems.

Peter

[mbanks]

Don't worry about wider early peaks until you see wider early peaks - thermal and statioanry phase focussing works wonders, and by trying to add solvent focussing before you even have a chromatogram that shows you need it is making things unneccesarily complicated.

Try a default splitless injection; 1 ul, 1 min splitless, column start temp 80C. Shorten the splitless time to 30s, to see if peaks get smaller. If not use 30s, if they do go back to 1 min. If the early peaks are wide drop colum temperature to 60C to see if they improve.

I just love simple methods - have you tried without derivatization ? Columns these days are so well deactivated that you can get phenols through them as sharp peaks with no problems.

Peter

Thanks for the advice. Seems like everything I was seeing about splitless made it sound like some refocusing was necessary, but I'll certainly give it a shot like you describe. Simple is always preferable : )

Unfortunately, derivatization is necessary. Some of the higher boilers will not fly without it. The high temperature necessary in the inlet causes decomposition of some of the analytes and heavy fouling of the liner, so quantification is not reliable.

Thanks for all the help, this thread has been very useful.

[Rndirk]

I would also increase the starting oven temperature. You're right about solvent focusing, but for acetonitrile 60-65°C should be OK. You might be condensing too much.

Would you explain this a bit more? I'm not too experienced with splitless or solvent trapping. Is condensing too much solvent a bad thing? Will this cause broader peaks? What should I do to optimize using a solvent trapping method?
.

Having solvent eluting at the same time as your analytes affects the ionization, lowering the signal. With MS detection, you typically have a "solvent delay" setting in your method during which the filament (i'm assuming you run EI-MS) is turned off. Measuring during the solvent peak lowers the lifetime of the filament.

I could also stand to flow a bit faster, i think I can go up to 4mL/min with our pump (turbo), so I'll probably shoot for a more efficient flow. Thanks for pointing this out.

A simple rule called 'speed optimized flow', if you use helium gas, is flow (ml/min) = 8 x internal diameter.

So for your column 2 ml/min

How do you mean the MSD masks the solvent tail? Wouldn't I be able to see that? First boiling compound is fairly high, it's TMS derivitized phenol, boiling point around 160C.

How do you use the MSD? SIM? Or scan mode between which range? If you don't measure for acetonitrile, you don't see it (="masked"). But it can still impact the detection of your compounds (see above).

How would lowering the splitless time help me, other than transferring less sample onto the column? Shouldn't the refocusing largely mitigate any peak broadening due to long splitless time? I chose a long splitless time to reduce weight discrimination as much as possible, as I have a wide rage of MW at low levels.

Your inlet temperature is more than high enough too avoid discrimination and you also use glass wool. After a certain time it's better to start purging since stuff (most likely solvent) that hangs around in the liner only gives rise to tailing and broadening.

[James_Ball]

If you want to optimize the solvent effect(trapping focusing or whatever term we want to call it) try using a starting temp about 25C below the boiling point of the solvent, hold the oven at that temp until you go split on the inlet, then do a fast 40C/minute or faster ramp up to about 5C above the boiling point of the solvent, then go to the slow ramp for analyte elution. This will allow the solvent to elute in a more narrow band and cause less problems at the MS. You will also need to experiment with the splitless(purge) times to optimize the recovery versus tailing.

https://cn.agilent.com/cs/library/userm ... 041007.pdf

This is a good guide to optimizing inlets and other information can be found if you google splitless optimization.