I am trying to develop a GC-FID method for analyzing glycol that comes from glycol dehydration plant (mostly ethylene glycol and triethylene glycol). I am currently using an Agilent DB-Wax-UI column and using 2-ethyl-hexanol as the internal standard. I am getting +/-10% error, which is huge. Can anyone provide any idea regarding how to decrease the error? Thank you very much!
General way to do this is divide and conquer. What is the error on replicate injections? Move back one step in your protocol. WHat is the error there? Find the places where your error increases the most and focus on them.
Thank you Scott, I have 2-5% error with injection from an autosampler. But I have internal standard, should that not correct for the injection error?
Thank you again!
“Error” and large variation in results have many causes. Please explain how you measured this error? How many replicates did you use? Investigation of any problem requires that we first understand what method you are using. Please provide the detailed GC method used. Please include any sample prep procedures too, as these are just as important as the analysis. Can you tell us if this same method has ever worked for you in the past, and provided much better results (~ 1-2% RSD)? This woudl help put your question into context and allow us to have a discussion about which areas may need to be addressed.
Hi Bill, thank you very much for your reply. My error is relative to my theoretical sample amount, over 2 or 3 repeats. Please see my GC condition in my other comment. I was having this problem for a while and I’m also getting “front inlet pressure shut down” error, could that be a faulty septum?
It sounds like you may need to address the instrument issues first. Has maintenance been done?
Thank you very much Bill, I will check the maintenance status of the instrument.
Thank you for your reply. My GC conditions are 0.1 uL injection, 250C inlet temperature, 2:1 split ratio, 5 mL/s for flow rate, oven temperature: 50C for 2 min then to 230C at 10C/min. FID detector operates at 400C.
Tao – in addition to my previous comments.
Most split injections work better at >20mls/min in addition 0.1ul injection would cause larger errors. Suggest 0.5ul injection at 10:1 split ratio. This should be equivalent to your current column loading. This should improve your repeatability.
Not sure why you needs FID temperature of 400C?
Hope you do not have a regular vessel graphite ferrule for the jet and it is graphite. 300C should be more than sufficient.
Thank you very much, Gary. I will definitely try that.
Excellent question. Usually, internal standard will correct for injection error. The exception to this rule is when you are on the edges of your linearity. If bias creeps in from high or low signal of IS or analyte, it may not be corrected for.