If you have just installed the column, there should be instructions with the column for conditioning it (and some of us don't bother with that these days, as columns come pretty clean.)
If you are trying to bake off residue from an analysis - it depends on what it takes to get rid of the residue. And, if you are trying to get rid of column bleed, baking may just make things worse.
Inlet temperature: There are two schools on operation of the GC. Some of us run the inlet a bit cooler than the maximum temperature used in the analysis - assuming that it is easier to replace an inlet liner than a column and it takes more temperature to move junk across a stationary phase than across a liner (even with glass wool). Others run the inlet up to be sure everything gets onthe column. - So the first question is how have you been operating the instrument? If you have beek keeping the inlet cool, change the inlet liner and keep the inlet cool. If you have been running the inlet hot, change the liner and keep the inlet hot. (Clean the inlet by changing the liner - not by pushing garbage through the column)
The column - If you have been running an analyisis with an upper temperature on the column that is below the temperature required to elute some copunds, run the temperature high enough to elute the compunds - and do it every run rather than collecting stuff on the column. That stuff becomes part of the stationary phase for the next injectoon(s) And, as the stationary phase changes, so does your separation.
If you have materials that will not elute in reasonable time at the upper limit of your column. You may need to remove a portion from the front end of the column. For some, this works. For the work I've done, I've discovered that it is more effective to simply replace the entire column. (When I kill a column, I kill it dead!)
Detector temperature - this scares me a bit. If you will be pushign through heavy stuff that leaves you needing to raise the temperature of a mass spec to avoid fowling -- expect ot have to clean the ion source anyhow. And this gets into the question of how long to bake: Not long. And, you will need to moitor what is coming off to know when to stop (thus detector temperature close to normal operating temperature - or just a bit warmer).
Run the column up in temperature to close to the max. If you think there is a lot of stuff to bake off, hold it a while ( 10 min to 1 hour - you learn this from experience with your analysis) and then make a run on your instrument - like an analytical run, but no injection. Watch the background. If it low enough, you baked the column long enough. If the background seems to be going up or will not fall, you may have a column that has reached the end of its life. If running the column up to close to the maxiumum and holding for 5 to 10 minutes at the end of a run is not enough to keep your column in shape, look at your sample prep -- you are putting things on your column that will kill the column.
