Massive solvent peak tailing

[Mendicant]

Good day, separation scientists.

I have worked for a small laboratory almost fresh out of college and have spent the past couple years blindly chromatographing.

Recently I attended Agilent's Practical Gas Chromatography, Course Number R1915A taught by Dr. Polite in Chicago, and that guy really knows his stuff. Much was revealed.

On to my first quandary.

These methods were run on a Shimadzu GC 2010 Plus (our only Shimadzu). Inside are two almost identical columns. The front inlet is your standard split/splitless, and the back inlet is a PTV. On the front, here is an example of an isothermal method using a 25:1 split ratio.



On the back, here is an example of what is essentially an on column method using the PTV.



The issue is that I need to be able to quantify that standard that is coming out at seven and a half minutes, and I still have some significant peak tail from the solvent at that time.

Let me know if you have any suggestions on improvements.

Thank you kindly, for any input.

[Blazer]

I took Dr. Polite's class a few years ago myself. Good instructor, fun class. The hands on section was great since I was a novice GC user but had been using LC for years. Learned a lot. Even got to watch a Blackhawks game and pick up a jersey for my fiancee. Patrick Sharp is her favorite player. But I digress. Let's see if we can get you to look through your notes and solve this one a bit on your own (with some guidance).

My first question for you: are you troubleshooting a method that used to work just fine or are you trying to develop a method?

If it's the former, has anything changed about your system or method since it was run last?
If it's the latter, what do your notes from Dr. Polite's class say about peaks that elute early? What does that mean with respect to your temperature gradient?

[Mendicant]

My first question for you: are you troubleshooting a method that used to work just fine or are you trying to develop a method?

If it's the former, has anything changed about your system or method since it was run last?
If it's the latter, what do your notes from Dr. Polite's class say about peaks that elute early? What does that mean with respect to your temperature gradient?

I am troubleshooting a method that, as far as I know, has always done this for years. I just want to make the data more legitimate.

As we learned in class, in order to increase the retention time of peaks whereby the retention factor (k') is less than 1, the initial temperature of the program shall be decreased in order for the analyte to increase the time spent interacting with the stationary phase. However, I do not believe this is the case in this example. My T0 is 20 seconds while my first peak of interest is 7.7 minutes. The real problem seems to me is that my solvent peak spends 15 minutes measurably tailing. What might I be able to do to reduce this phenomenon?

[Peter Apps]

One of the things that they apparently did not teach on the course was that to troubleshoot a method we need all the details of the operating conditions; flows, pressures, temperatures, times. Otherwise we are just guessing.

Peter

[dblux_]

...
What might I be able to do to reduce this phenomenon?

make use of PTV mode - solvent vent

[Mendicant]

One of the things that they apparently did not teach on the course was that to troubleshoot a method we need all the details of the operating conditions; flows, pressures, temperatures, times. Otherwise we are just guessing.

Peter

PTV2
PTV2: 350C
Injection Mode: Direct
Sampling Time: 1.00 min
Carrier Gas: He
Flow Control Mode: Flow
Total Flow: 4.4 mL/min
Purge Flow: 0.0 mL/min
Split Ratio: 4.0


Column
Column Length: 30.0m
Inner Diameter: 0.32 mm ID
Flim Thickness: 0.25u
Max temp: 350C
Temperature: 80 °C, 20C/min, 140 °C, 5C/min, 335 °C, hold 20 min

FID2
Temperature: 375C
Sampling Rate: 40 msec
Stop time: 63 min
Delay Time: 0.00 min
Makeup Gas: He
Makeup flow: 30.0 mL/min
H2 Flow: 50.0 mL/min
Air Flow: 400.0 mL/min

[Peter Apps]

As far as I can tell you are using the PTV to do a hot splitless injection - there is nothing in the method about PTV start temperature or heating rate. The main problem is that you do not open the split at any stage, which causes the long tail on the solvent peak. You need to open the split / purge / vent (depending on what it is called on your particular system) after 1 min with a flow of 30 ml/min. That should get rind of the solvent tail. The adjust the time at which the split opens to give you the required peak size and repeatability for the standard.

Peter

[dblux_]

In fact this is split injection.
FID parameters are irrelevant for troubleshooting this case.
It's a problem with sample introduction. Thus information about your solvent and analyte would be more appropriate to diagnose solvent tailing.
Make use of solvent vent mode - that's what PTV is for.

[Peter Apps]

In fact this is split injection.
FID parameters are irrelevant for troubleshooting this case.
It's a problem with sample introduction. Thus information about your solvent and analyte would be more appropriate to diagnose solvent tailing.
Make use of solvent vent mode - that's what PTV is for.

I agree that solvent venting might be useful (though not much more useful than a simple split/splitless injection since the injection volume is only 1 ul ), but solvent vent is only possible if the inlet temperature is programmed - which it appears not to be.

If this is a split injection, then where is the solvent tail coming from ? If the Shimadzu PTV has a septum purge than that accounts for the 4 ml/min extra flow. Although it says a split ratio of 4:1, the vent flow is zero - is this vent flow only for solvent venting ? If it is split injection, then increasing the split ratio is the first step.

Crud in the inlet might cause this - check that the liner is clean and that there are no fragments of septum or ferrule lying in the bottom of the inlet body.

What does "direct injection" mean - are you using an inlet liner that seals to the column (a Uniliner) ? If so is it drilled top or bottom ?

Peter

[KM-USA]

If this has worked OK before, if me I would remove the inlet end of the column and reinstall. Agree, extreme high probability the issue is at the inlet.

[umrebel1303]

I would echo what some of the other folks have said. Most of the time, issues with large tailing solvent peaks is focused around the inlet. This could be in the form of leaks, dead space due to improper column placement, poor column cuts, dirty liners, degraded/dirty front portion of the column, etc. So, my first advice would be to look at historical data and make sure that this issue has been present since the start. If it hasn't, go back and figure out what changed before the issue.

I'm not sure what solvent you are using, but another cause of large solvent fronts can be tied to backflash. If you aren't familiar with the term, this is when the solvent expansion surpasses the capacity of the liner that you are using in your system. If the solvent is expanding beyond your liner, this can lead to solvent ending up in your gas lines, purge vents, etc which will effectively contaminate your runs. Here is a link to a calculator:

http://www.restek.com/images/calcs/calc_backflash.htm

I noticed in your method that you don't seem to have any purge flow, which if I remember correctly with Shimadzu systems is their designation for septum purge flow. If you have checked the other parameters, this may be something you can try increasing/adding. If you are having any solvent backflash this can help sweep it away from the head of the liner/injection port.

[Mendicant]

In fact this is split injection.
FID parameters are irrelevant for troubleshooting this case.
It's a problem with sample introduction. Thus information about your solvent and analyte would be more appropriate to diagnose solvent tailing.
Make use of solvent vent mode - that's what PTV is for.

I agree that solvent venting might be useful (though not much more useful than a simple split/splitless injection since the injection volume is only 1 ul ), but solvent vent is only possible if the inlet temperature is programmed - which it appears not to be.

If this is a split injection, then where is the solvent tail coming from ? If the Shimadzu PTV has a septum purge than that accounts for the 4 ml/min extra flow. Although it says a split ratio of 4:1, the vent flow is zero - is this vent flow only for solvent venting ? If it is split injection, then increasing the split ratio is the first step.

Crud in the inlet might cause this - check that the liner is clean and that there are no fragments of septum or ferrule lying in the bottom of the inlet body.

What does "direct injection" mean - are you using an inlet liner that seals to the column (a Uniliner) ? If so is it drilled top or bottom ?

Peter

Interestingly enough, I did try to change the liner out yesterday partly to see what it looks like. I already had the AOC off because I was trying to solve an issue with it bending all the needles we put in, and I figured what the heck. Unfortunately, unlike the front inlet, I was unable to yank the liner out. Tech support at shimadzu indicated that it may have been a case where someone overtightened the nut, causing some of the o-ring to melt down around the liner and sealing it into place. I was worried the amount of force I was applying to the forceps might shatter the glass, so I had to back off a bit. I'm still contemplating how I'll get it out safely.

But that is a separate issue. I'll work on screwing around with this "vent flow" thingie in the meantime.

[Peter Apps]

I wouldn't bother doing anyting until you get the liner out and replace it with a clean one - obviously the inlet has not been properly maintained, and this is by far the most likely (but possibly not the only) cause of the solvent tail.

Peter

[AICMM]

The question I have is, is the Shimadzu PTV plumbed like a 5890 or more like a 6890? Huh you say? On a 5890, the way they were plumbed you could measure split flow with the valve on or off
and they would be the same (made for challenging troubleshooting.) But on a 6890, you only measure split flow when the splitter is on. So.... try measuring your split flow on your PTV while making a run. If it 6890 like, you will see a big jump in flow when you start the vent/split flow.

Best regards,

AICMM

[Mendicant]

A big thank you to everyone who had contributed. The previous method is in magenta and the current method is overlaid in black. (Note that they are two slightly different samples.)

It seems that the main problem was that the purge flow was set to zero. Because this is an on column injection, I programmed the method such that it would wait to start the purge flow until after the PTV had finished volatilizing some of the heavier organics. While I was messing with it, I also figured out how to ramp the PTV temperature.

Between these two changes, it really improved the integration. I assume the tailing was occurring on other peaks in addition to the solvent peak. And I'm pretty sure the PTV also helped get a little more separation on some of the doublets in there that I was looking for.

Furthermore, a little more digging revealed that the purge was completely off for *every* method that was on that particular GC. It was less noticeable because all the other methods have a pretty heavy split ratio, but adding that purge (in addition to just being sound practice) may have a very slight improvement on the rest of the methods too.

The powers-that-be were happy with the progress, and I had a lot of fun.

Good work people.

Still haven't gotten that inlet out, though. Oh well. Onto the next machine!