Loss of Column performance

[Gardner]

I am analyzing a Glycol Phosphite Mixture in Corn Oil using a Phenomenex ZB-1701 column with methylene chloride as the diluent. The method requires the column to exceed the maximum temperature (300 C) by going to 320 C and holding for 5 minutes. This occurs each run (the runs are in duplicate). After analyzing my standard curve and system suitability (approximately 19 hours of run time or 27 run cycles), my sample recoveries are all approximately 18% too low. My within run standards are experiencing the same problem. I am assuming a loss of performance due to column bleed due to thermal breakdown. Am I on the right track,or does someone else have another idea?

[neval]

i have some problems about column effıcıency.i working with oils...and in the last times in the stearic acid value is low.but ı raise the tempr.for the stearic acid and i saw that the value is approx.the real value...but ı think it is not enough.because when i injected the sample 3-4 times i saw that all fatty acid values(oleic,linoleik) were lowered.i think also,this is because the column efficiency..

[chromatographer1]

You are on the right track.

A limit is meant to be a limit, not a suggestion. A programmed temperature limit is meant to be passed through briefly at best, not held isothermally.

I would suggest that if you need a moderate polar column that exceeds 300°C you try a 20 or 35% phenyl column, or extend your analysis time at a lower temperature.

Certainly 300°C is stretching the cyanopropyl thermal limit. Chemistry is chemistry and thermal decomposition happens according to the chemistry involved.

You most certainly lost a large portion of your liquid phase and or partially stripped the column exposing bare silica. You could try removing 20% of the front end of your column to regain some possible use of the column.

But please keep the temperature at 280° for isothermal work unless you want to buy new columns often.

[Gardner]

I must have been channeling your thoughts. I am thinking of trying a DB-35ms or DB-17 column. Either of these will maintain the polarity, and give me the thermal limit that I need. I appreciate the response. I'm tired of looking at the data and seeing the same peak area regardless of the input. It's been virtually the same for at least ten injections.