what shall do if I have components retained?

[moonchips]

Please give me some suggestions about the following issue.

About my retained components, I have 3 options:
1) flush it out for each single injection (I wasted time to flush out retained components, and I do not know if 5min is enough to equilibrate between mobile phase change)
2) flush it out for every N injections (I can save some time, and re-equilibration time can be set longer than 5min. But is this practical? I might have ghost peaks)
3) use a guard column (question: what about those components I am looking at? Retention times of them will change, and maybe I cannot save much time. Do I get ghost peaks?)

[karan755]

First two options are OK, If you are not interested in retained components.

Guard columns don't change the retention time due to very short length. Components retained in analytical column will not be retained in guard column. However you can check wether these comonents retained by injecting the sample using guard column only.

[moonchips]

I saw this from Agilent. It seems a guard can help to retain material that is permenantly retained in analytical column.

Question: What does the filter do that the guard-column does not?

Answer: Nothing really - both a filter and a guard can capture small particulates that are either coming from your sample, mobile phase or are from wear-and-tear of instrument seals and gaskets. A guard, on the other hand, traps sample material that is strongly retained on your column, material that would not elute from the column under the experimental conditions defined by the method. These strongly retained materials can build over time at the head of the column and generally lead to high pressure and/or poorly shaped peaks. In some cases, these retained materials cannot be removed even after extensive washings with strong solvent. By using a guard, your column is protected from material that is permanently retained.

[Uwe Neue]

It says: "permanently retained", and not "late eluting peaks".

If you are running an isocratic method, and it runs to k = 10, i would not be surprised if a "late eluting compound" shows up at k = 15 or 20. So it will show up in the middle of your next run.

If you are talking about a gradient method, then you can get away with only an occasional flushing of the column.