Large RT shift first injection & gradual shift to normal RT

[jjp295]

We utilize Shimadzu GC-2010 GC-FIDs and using a supelco SP-2560 column. We notice that we experience a significant retention time shift with injection 1 and gradually the retention time will shift back to normal in about 10-20 injections. We mainly see this after the GC sits idle for extended period of time (>1 day). Basically every Monday at start of new work week we experience this on all of our GCs.

We utilize Gas Generators for all our gases but also have gas cylinders as back up and we experience this using both type of gas supplies.

What might be causing this to occur? How could we correct this issue?

Example of the shift is RT of 12.525min (injection 1) and then at injection 15 the retention time is at 12.575min

[antonk]

Slightly polar phase captures water while instrument not in use.

Try to insert moisture filter on carrier line. Or make 1 blank run after long idle time.

[jjp295]

We have a moisture and oxygen trap coming off our hydrogen generator going into the gas line. Those do not indicate any water or oxygen getting into the lines.

Even with 1 blank injection still see this shift in retention time and then it gradually shifts back to the normal retention time in about 10-20 injections.

[DR]

Leave the carrier running over the weekend if you are not doing so. Turn it down some if you want, but leave some flow... Might help.

[Jake]

As others have recommended, leave the instruments in a 'sleep' mode, where there is carrier always flowing. This shouldn't be an issue if you're using generators. We also leave the make up flow on to at a few mL/min to constantly sweep the FID to help keep dust from settling in it.

[Peter Apps]

What kind of analysis are you running that a 3s retention shift is "significant"?

[Consumer Products Guy]

What kind of analysis are you running that a 3s retention shift is "significant"?

Great question !

I did have a supervisor for a while who would've been upset with a 3 second retention time difference....



I always did at least one "start-up" injection and program through the oven/run cycle. I also changed the septum before each day's use.

[bcd_GCLCMSMS]

I also recommend leaving carrier (and other gases) running. If the carrier is expensive, such as He, consider switchover device UHP nitrogen when not in use.

[jjp295]

Over the weekend we usually run a cooldown method after all the samples have ran. This cooldown method reduces the flow of carrier gas through column to 1mL/min, turns off the FID gases and then cools FID/Injector/Column.

We integrate batches using Methods so this will pick the 27 peaks for us. However, with the shift the methods will not work as the parameters we set for the integration do not line up. Also, within our methods we set the retention times for the 27 peaks of interests and this shift creates issues as it will identify the wrong peaks or no peaks at all. This will create issues as we use that information to report results to clients.

Over this weekend I left a GC on with the correct carrier gas flow as well as heated up. I'll be running samples to confirm if this helps or if we still see this shift.

[jjp295]

Over the weekend I left one of our instrument on with normal carrier gas flow and temperatures. The FID flame was on as well. This morning I injected samples on this instrument and saw this initial shift left with a gradual shift back right to the normal peak retention time. Below is the list of injections with the retention time of our internal standard peak. This is a normal occurrence anytime our GC sit idle for >1day.


Ret. Time (min.)
Injection 1 9.951
Injection 2 9.990
Injection 3 10.003
Injection 4 10.010
Injection 5 10.016
Injection 6 10.020
Injection 7 10.022
Injection 8 10.024
Injection 9 10.026
Injection 10 10.026
Injection 11 10.028
Injection 12 10.028
Injection 13 10.029
Injection 14 10.029
Injection 15 10.029

[jjp295]

For the retention time shift here is the reason why it is significant. 1 picture is a sample at the start of the batch and 1 picture is the same sample at the end of a batch. The same integration method is ran through each sample. As you can see the sample at the start of the batch has no peaks integrated correctly and the identification of the peaks would be incorrect if the sample was integrated.

https://ibb.co/yfmvmMQ

https://ibb.co/TDXNWnH

Sorry for the url links to the picture instead of putting the pictures here but I could not get the image upload to work correctly.

[Peter Apps]

Have you tried increasing the width of the retention window in the integration method? - you have plenty of resolution to play with.

[jjp295]

Yes and is good suggestion we have tried. It will work for some peaks but not the peaks that elute close to each other. By setting a larger window it will also intergrate unwanted small peaks. Wo do not do this due to we have identify trans fatty acids which we group all together as 1 peak and our integration method does this. However, we cannot tell the inergration method to either identify the 1 large peak at each retention time or add all peaks at each retention time together as 1 peak area. We need the method to group all peaks around the retention time together due to the trans.

Also, since we set a retention time to identify the peaks we want the shift causes mis-identification of peaks.

[Jake]

Can you please share a bit more about the method? We routinely use a 2560 column without issues (other than the busted column that I posted about).

- What is the column length?

- Start temp., ramp rate, final temp. and post run.

- What is the matrix?

[jjp295]

Our column dimensions are 100m x 0.25mm x 0.2um. The injector is at 230C, FID is at 290C and the oven starts at 180C. We utilize constant linear velocity of 55cm/s for carrier gas flow and the carrier gas is Hydrogen. The following is the oven temperature program we utilize:

Rate (°C/min) Temperature (°C) Hold Time (min)
-_______________180_____________1.75
5_______________200_____________1.75
10______________240_____________10.5


The main 3 matrices we analyze are dried blood spots, red blood cells and plasma. The fatty acids are converted to FAMEs which then go into the hexane layer. We take an aliquot from this hexane layer to GC vial and this gets injected into our GC.