Most labs don't plan for column failure. They plan for samples, methods, and deadlines. Then the pressure spikes at 10:45 PM before a big delivery, and someone calls a service engineer (me) asking if a column can be revived. In my role coordinating rush service calls for labs running Shimadzu HPLC and gas chromatography systems, I've lost count of the Friday-night calls that could have been avoided with a simple checklist.

So here's the one I use when a customer asks 'should I change this column?' It's seven points, and you can probably get through all of them in under ten minutes. No theory, no 45-minute lecture, just the checks that actually predict failure.

If your search history looks like 'when to change your columns hplc agilent,' or if you're on a Shimadzu HPLC and just want a straight answer, this is for you. The process is more universal than people think.

The 7-point checklist

1. Track injections, not calendar time

Column lifetimes are measured in injections, not days. A 5-micron C18 column on a well-maintained Shimadzu HPLC might give you 800-1500 injections for a fairly clean sample matrix. Dirty samples? More like 300-600. I know that's vague; the lab down the street will see different numbers. The point is to write it down. When you install a column, use a lab marker to write the date and injection count on the column tag. If you don't know what a column has been through, you're guessing.

2. Watch pressure trends

Most columns die gradually, not suddenly. Compare current backpressure to the 'new column' reading you took at installation (you did take one, right?). A 10-15% increase from that initial value usually means a blocked frit. A 20% increase plus peak tailing? The column is asking to retire. A lab I worked with spent three days troubleshooting a method when the real answer was a $150 replacement. (Note to self: I did this twice before I learned to look at the pressure log first.)

3. Check peak shape before condemning it

Peak asymmetry is one of the clearest signs. Tailing factor above 1.5 after a proper column wash? That's a strong indicator. But try the cheap fix first: flush slowly with a strong solvent in the reverse direction. If tailing doesn't move after one slow flush, the column is probably done.

4. Use a guard column

This is the step most people skip. Most buyers focus on the price of the analytical column and completely miss the downtime when it clogs. A guard column (a small, relatively cheap pre-column) catches particulate matter and strongly retained compounds before they reach the analytical column. Change it every 50-100 injections, or when pressure climbs about 10%. One lab I supported doubled the life of a Shimadzu HPLC column with this one step. Not every sample behaves the same, but it's the cheapest insurance in chromatography.

5. Check the small stuff before the column

Here's where newer operators get burned: a sudden pressure drop isn't always a column failure. It can be a leaking fitting, a blocked autosampler needle, or a flow sensor si5000 that has drifted off zero. And if you're using a multimeter to check the electronics, a blown 115 multimeter fuse can make a perfectly good instrument look dead. I once stared at a pressure readout from a bad sensor and told a client to order a column. It wasn't the column. Always verify the obvious parts first. (Mental note: that $2 fuse almost cost that client $1,500 in unnecessary parts.)

6. Run a system suitability test

Stop guessing and run five replicate injections of a standard. The acceptance criteria from USP Chapter 621 are usually something like: retention time RSD below 1%, tailing factor below 2, and theoretical plates above the minimum in your method. If the column fails after a cleaning cycle, replace it. If it passes, keep running. That sounds like paperwork, but it is the rule.

According to USP-NF General Chapter 621, system suitability parameters need to be met before a run is valid. This is the closest thing to a shared answer for 'when to change your columns hplc agilent,' and it works for a Shimadzu HPLC in a QC lab too.

7. Have a spare column on the shelf

If you're asking whether to change it, the answer is usually yes. And if you don't have a spare, you're one blockage away from an emergency. I did the math after a 2023 incident: a lab paid $60 for expedited shipping and saved a $12,000 batch. The spare column had been sitting in the cabinet for months. In March 2024, a client called at 11 AM with a dead column and a 6 AM deadline the next day. Normally I would have wanted a full pressure and peak shape history, but there was no time. We shipped a replacement overnight and they made it, but the real fix was the spare they ordered the week after. I now tell every lab with a Shimadzu HPLC or gas chromatography system to keep at least one spare for every critical method. If that feels like overkill, ask the QC manager who borrowed a column from a neighboring lab at 9 PM. She'll tell you.

GC columns: same logic, different symptoms

For a Shimadzu gas chromatography system, the checklist changes slightly. Watch baseline bleed, retention time drift, and peak shape. A GC column can sometimes be revived by conditioning at a low temperature, but if bleed stays high no matter what, it's time. Don't wait until the detector is collecting ghost peaks during a compliance run.

Common mistakes I keep seeing

  • Not recording the new-column pressure. You can't tell pressure has risen if you have no baseline.
  • Trying to squeeze one more run out of a failing column because the batch looks good. The one run that fails will be the one that costs you a week.
  • Skipping the guard column to save $80.

To be fair, some labs can run columns well beyond the estimated injection count because their sample matrix is unusually clean. That's the exception, not a strategy. Five minutes of verification beats five days of correction. That's why I keep coming back to this checklist. It's not exciting. It just works.