Problem reproducibility

[Damien]

Hi

We have a new GC in my lab with an autosampler and I made several runs to check the reproducibility of measurements. Unfortunately, the area ratio for 5 consecutive runs is: 1.65; 1.65; 1.52; 1.53 and 1.67. I am not able to understand what I am doing wrong. Do you have any idea from where could my error come? Please help me

Here is some features of my method:
Varian 3900 GC
Autosampler varian CP-8400

Product: 4-methoxybiphenyl boiling point 157°C
Standard: diethyleneglycol di n-butyl ether boiling point 256°C
Solvent: dichloromethane

Syringe size: 10µL
Injection mode: Std split/splitless
Injection volume 0.1 or 1µL

Injector type 1177
Temperature 300°C
Split ratio 20 or 50

Injector EFC type 1
Constant column flow 1.0 mL/min

FID detector
Temperature 300°C

Type 11 detector EFC
Make up flow 29 mL/min
H2 flow 30mL/min
Air flow 400mL/min

Cheers

[Bernd]

Hello Damien,

first a lot of questions: is the 2nd peak the problem or all (by area or peak shape)? The column is correctly installed? Have You some deactived glass whool inside the liner? And excuse the next question is the injector tight (septum, column, nut with vespel ferrule)?
What are really doing, split or splitless?, Your concentrations are in ppm or higher? Did You have Galaxy or Star?

Regards
Bernd

[Peter Apps]

And yet more questions; what column are you using, and what is the temperature programme ?, what are your carrier gas flow and pressure settings ? What autoinjector settings are you using - plunger speeds, pre and post injection dwell etc ?. What is the temperature of the laboratory ?
How repeatable are the peak areas ? please tell us the actual peak areas for the ratios that you have already posted.

Forget about doing 0.1 ul injections with a 10 ul syringe. It would be better if the internal standard was more simialr to the analyte in terms of boiling point.

Peter

[Damien]

Hello

Thanks for your help. Here is the answer for Bernd

It seems to me there is no problem with any peaks. The shape is right (no shoulder) and the width ½ is about 1.5-2.1 seconds.
As the GC is new, the column has been installed by the varian technician.
This morning I changed the septum, cleaned the glass insert which was without glass wool and reinstalled the glass insert with deactivated glass wool. The septum purge is set to 2mL/min (checked this morning after the cleaning). I followed the instruction from the manual. so I think everything is OK. The septum, column, nut with vespel ferrule are tight (neither too much, nor too less).
The software is Varian Star version 6.41

I am using split injection with a ratio of 50 (I tested it with a ration of 20 without improvement)

Injector Type 1177
------------------
Oven Power: On
Temperature: 300 C
Time Split Split
(min) State Ratio
----------------------
Initial On 50

Here is for Peter

The column is a Varian factorfour VF-1ms.

My temperature program is

Column Oven
-----------
Coolant: Off
Enable Coolant at: 50 C
Coolant Timeout: 0.20 min
Stabilization Time: 2.00 min
Temp Rate Hold Total
(C) (C/min) (min) (min)
--------------------------------
70 0.0 2.00 2.00
180 10.0 2.00 15.00

My washing program is

Column Oven
-----------
Coolant: Off
Enable Coolant at: 50 C
Coolant Timeout: 0.20 min
Stabilization Time: 2.00 min
Temp Rate Hold Total
(C) (C/min) (min) (min)
--------------------------------
250 0.0 10.00 10.00

My carrier gas is helium and the pressure settings are
Air: 3,4 bar
Helium : 5 bar
Hydrogen : 3,5 bar

The autoinjector setting are:
Fill speed of the syringe: 2µL/sec
inject speed of the syringe: 50µL/sec
pre inj delay 0 sec
post inj delay 0 sec

sample air gap: no
air plug after sample 1µL

solvent penetration depth :90%
sample penetration depth :90%

clean volume 5µL
clean strokes 1
clean drawup speed 5µL/sec

clean mode pre inj solvent flushes 3
clean mode post inj solvent flushes 3
clean mode pre inj sample flushes 3

the temperature of the laboratory is about 20°C

Here are the area of the two compounds for each run with the ratio 4-mbp/DGBE
1 DGBE 69497
2 4-mbp 114770
ratio 1.65

1 DGBE 73403
2 4-mbp 121111
ratio 1.65

1 DGBE 61721
2 4-mbp 93880
ratio 1.52

1 DGBE 65234
2 4-mbp 99683
ratio 1.53

1 DGBE 70544
2 4-mbp 118045
ratio 1.67

I only look at the ratio in order to avoid any problem with the volume injected. I think even if the injection is about 0.1µL, the ratio should be the same.

When I started this study, I prepared 6 solutions to draw a calibration curve. I stopped to analyse all of them when I saw that ratio were not reproducible in order to focus only on one solution.

Cheers

[Peter Apps]

Hi Damien

Thanks for the details. The peak areas are more variable than the ratios, which points either to an inconsistent injection volume, or an inconsistent split ratio. Since you only tell me the type of column, not its length, diameter and phase thickness I have to guess from its being installed with a new GC that it is the default 30 m, 0.25 mm, 0.5 um. I cannot believe that the inlet pressure at the head of the column is 5 bar, surely this is the pressure in the supply line to the GC ?! Assuming, again, that the inlet pressure is around 7psi, or that you have set a flow rate of 1 ml/min, a 50:1 split ratio should work OK, and the settings for the autosampler also look OK. With a lab temperature of 20C I doubt that ypou are getting evaporation of the solvent, or bubbles in the syringe, but you should check that repeated injections from a single vial are not causing a vacuum inthe sample vial. The temperature programme is odd - why hold at 180 for 15 min when you could continue the ramp ? - and looks rather short and cool for the high boiling internal standard, which might suggest that there is in fact something odd about the inlet pressure and gas flows. If you inject some lighter gas or methane how long does the peak take to appear ?

I suspect that you are getting inconsistent split ratios due to a pressure surge in the inlet as the very volatile solvent is injected. Flash vaporization is, at best, eratic. Do you see any improvement with glass wool in the inlet ?

Since this is a new GC why not call the service rep back to check it out ?

Peter

[Damien]

Hi peter

The column is a 15m lenght x 0.25 mm ID x 0.25µm thickness. I set a constant column flow of 1mL/min.
I will check tomorrow for bubbles during the sequence of injection.
I hold for 2 min at 180°C and the total time of the run is 15 min (2 min at 70°C + 11 min of temperature ramp + 2 min at 180°C)
At the beginning, I made several runs up to 300°C with nothing else coming out from the column after 150°C so I changed the final temperature for 180°C. I thought that, once vaporized, the compound of a boiling point of 256°C is so dilute that is in the gas phase even at low temperature.

This afternoon, the results were with an injection of 0.1µL of another solution with a concentration of about 2 µg/µL of DGBE and 4µg/µL of 4-mbp
1st area 2nd area ratio
83158 190940 2.29
96119 223666 2.33
89650 209146 2.33
105407 249985 2.37
116671 251085 2.15

the retention time of DGBE (diethyleneglycol di n-butyl ether boiling point 256°C) is 9.81 min
the retention time of 4-mbp (4-methoxybiphenyl boiling point 157°C) is 12.26 min

I am waiting for an answer from varian...

cheers

[sassman]

As already mentioned, do not try to inject 0.1uL with a 10 uL syringe. It is too small to be reproducible. You should inject 1uL instead. Also (as already mentioned) your internal standard and analyte are very different. Why not use something more similar?

[Peter Apps]

Hi Damien

The results now look better, was this with glass wool in the inlet ? Do not inject 0.1 ul, 10ul syringes cannot deliver this volume repeatedly no matter what the manufactuers might say, and differences in volume of solvent and solvent vapour will give differences in split ratio and discrimination.

Is there a reason why you use an internal standard that is so different from the analyte ?

Peter

[oncoming storm]

Damien,

I have reproduceability issues with my Varian 3800/ CP-8400 as well. My column dimensions are the same only it's an EC-5 column and I am using methanol as my solvent. Part of the problem is a variability in lab temperature. We have a very bad heating/ac system. I hope that is not an issue with your lab.
What I have found helps is to use the volatile injection mode or a user defined mode. You may have to play with this to get the best parameters for your analytes and solvent. I have also found that with smaller volumes, I usually use 0.5 or 1 uL injections, a 5 uL syringe works better. We get our syringes from SGE. They are cheaper than most and seem to work very well.
After talking with the Varian service rep, we learned that the 3800 can't handle a split ratio of more than about 1:50 with a megabore column. The instrument can't hold the higher pressure required to keep the ratio consistent. Using a 0.35 or smaller capillary column it can hold the pressure. This may be different with a 3900 but it's something to ask about. Your split ratio is 20-50 so it should be ok but I would use 20. You may also want to increase your column flow or change the makeup flows. A higher linear velocity (70 cm/sec or higher) seems to help.

Good luck