Blog

Chromatographic Symptoms

These are symptoms you may observe in the chromatograms generated by data
acquisition. In general, these symptoms do not prevent you from operating
your GC/MSD system. They indicate, however, that the data you are acquiring
may not be the best data obtainable. These symptoms can be caused by
instrument malfunctions but are more likely caused by incorrect
chromatographic technique.

Two of the symptoms: If sensitivity is low and If repeatability is poor, also
apply to mass spectral data.

No peaks

If an analysis shows no chromatographic peaks, only a flat baseline or minor
noise, run one of the automated tune programs. If the MSD passes tune, the
problem is most likely related to the GC. If the MSD does not pass tune, the
problem is most likely in the MSD.

Passes tune

• Incorrect sample concentration
• No analytes present
• Syringe missing from the ALS or not installed correctly
• Injection accidentally made in split mode instead of splitless mode
• Empty or almost empty sample vial
• Dirty GC inlet
• Leaking GC inlet*
• Loose column nut at the GC inlet*
* These could cause a fault condition in the GC that would prevent
the GC from operating.

Does not pass tune

• Calibration vial is empty
• Excessive foreline or analyzer chamber pressure

Very dirty ion source
• Calibration valve is not working correctly
• Bad signal cable connection
• Filament has failed or is not connected correctly
• Bad ion source wiring connection
• Bad detector wiring connection
• Failed electron multiplier horn

Peaks are tailing

• Active sites in the sample path
• Injection is too large
• Incorrect GC inlet temperature
• Insufficient column flow
• GC/MSD interface temperature is too low
• Ion source temperature is too low

Peaks are fronting

• Column film thickness mismatched with analyte concentration (column
overload)
• Initial oven temperature is too low
• Active sites in the sample path
• Injection is too large
• GC inlet pressure too high
• Insufficient column flow

Peaks have flat tops

• Insufficient solvent delay
• Incorrect scale on the display
• Injection is too large
• Electron multiplier voltage is too high

Baseline wanders

• Insufficient carrier gas supply pressure*
• Malfunctioning flow or pressure regulator*
• Intermittent leak in the GC inlet*
* These could cause a fault condition in the GC that would prevent
the GC from operating.

Retention times for all peaks drift – shorter

• Column has been shortened
• Initial oven temperature was increased
• Column is getting old

Retention times for all peaks drift – longer

• Column flow has been reduced
• Initial oven temperature was decreased
• Active sites in the sample path
• Leaks in the GC inlet*
* This could cause a fault condition in the GC that would prevent the
GC from operating.

How do I know if I have an air leak?

Large air leaks can be detected by vacuum symptoms: loud gurgling noise
from the foreline pump, inability of the turbo pump to reach 95% speed or, in
the case of smaller leaks, high pressure readings on the high vacuum gauge
controller.

The mass flow controller is calibrated for methane and the high vacuum gauge
controller is calibrated for nitrogen, so measurements are not accurate in
absolute terms:

Familiarize yourself with the measurements on your system under operating
conditions. Watch for changes that may indicate a vacuum or gas flow
problem.

Always look for small air leaks when setting up methane flow. Run the
methane pretune, starting with a good PCI tune file (Figure 2). The abundance
of m/z 19 (protonated water) should be less than 50% of m/z 17 for acceptable
PCI performance. For NCI, the abundance of m/z 19 (protonated water)
should be less than 25% that of m/z 17. If the MSD was just pumped down, look
for the abundance of m/z 19 to be decreasing

There should not be any peak visible at m/z 32 (O2). This almost always
indicates an air leak.

Special NCI notes

Since NCI is so extremely sensitive, air leaks that are not detectable in EI or
PCI can cause sensitivity problems in NCI. To check for this kind of air leak in
NCI, inject OFN. The base peak should be at m/z 272. If the abundance of
m/z 238 is much greater than that of m/z 272, you have an air leak.
How do I find the air leak?
1 See Figure 3 and Table 6.
2 Look for the last seal that was disturbed.
• If you just pumped down the MSD, press on the sideplate to check for
proper seal. Poor alignment between the analyzer and the GC/MSD
interface seal can prevent the sideplate from sealing

• If you just replaced the reagent gas bottle or gas purifier, check the
fittings you just opened and refastened.
3 Check for tightness of seals at GC inlet and interface column nuts. Ferrules
for capillary columns often loosen after several heat cycles. Do not
overtighten the interface nut.
4 If any of the fittings inside the flow module (VCR fittings) were loosened
and then retightened, the gasket must be replaced. These gaskets are good
for one use only.
5 Remember that most small air leaks visible in CI mode are located in either
the carrier gas or reagent gas flow paths. Leaks into the analyzer chamber
are not likely to be seen in CI because of the higher pressure inside the
ionization chamber.
6 Half-split the system.
• Close valves starting at the gas select valves (Gas A , then Gas B), then
close the shutoff valve. See Figure 3 and Table 6.
• Cool and vent the MSD, remove the GC column, and cap off the interface.
If you use argon or other introduced gas to find air leaks, this does not work
well for the reagent gas flow system. It takes as long as 15 minutes for the peak
to reach the ion source if the leak is at the inlet to the flow module.

I've spent a lot of time under the hood as they say, tearing what little hair I have left out trying to troubleshoot weird, bizarre and downright impossible problems on GCMS equipment.  GCMS equiment is some of most complicated commercially sold devices a person can deal with.  Going down the wrong path can not only be costly but effect you lab productivity.

So, what have I learned over the years?  Here's a general list of troubleshooting tips.

Rule 1: “Look for what has been changed.”

Many problems are introduced accidentally by human actions. Every time any
system is disturbed, there is a chance of introducing a new problem.
• If the MSD was just pumped down after maintenance, suspect air leaks or
incorrect assembly.
• If the reagent gas bottle or gas purifier were just changed, suspect leaks or
contaminated or incorrect gas.
• If the GC column was just replaced, suspect air leaks or contaminated or
bleeding column.
• If you have just switched ion polarity or reagent gas, suspect the tune file
you have loaded in memory. Is it the appropriate file for your mode of
operation?

Rule 2: “If complex isn’t working, go back to simple.”

A complex task is not only more difficult to perform, but also more difficult to
troubleshoot as well. For example, CI requires more parts to work correctly
than EI does.
• If you’re having trouble with NCI, verify that PCI still works.
• If you’re having trouble with other reagent gases, verify that methane still
works.
• If you’re having trouble with CI, verify that EI still works.

Rule 3: “Divide and conquer.”

This technique is known as “half-split” troubleshooting. If you can isolate the
problem to only part of the system, it is much easier to locate.
• To isolate an air leak, select Shutoff valve. If abundance of m/z 32 decreases,
the problem is not in the flow module.

Running analytical instruments and keeping them running can cost up to 10-20% of the initial purchase price of an instrument every year.  There are several ways to reduce these costs while not impacting the productivity of the instrument.

Training

Operators can do many of the tasks that the manufacturer's or third party service charge up and beyond $300 / hour.  These tasks include maintence, cleaning source or trap components, changing consumable

Method Optimization

Many customers setup an application, certify it according to their QC requirements and then move onto other projects.  In real world samples, matrix, especially biological samples, can be very dirty and eventually build up residue in the inlet and other areas of the GC/MS. Careful attention should be paid the application in the initial year it is running.  At some point the contamination becomes so bad that either the calibration or tune fails the operator can't fix, and a service call is placed. 

Preventing or minimizing this situation can be minimized by the chemist who initially sets it up. 

- Clean up of the sample before it gets into the instrument.  For many types of analysis (purge and trap, air monitoring) this is not possible but for others it can make a big difference.

- Implementing back flushing on the MS.  Not all applications work well with back flush, it adds more hardware and connections to system.  However, when properly implemented, it can keep sample matrix from ending up in your expensive MS.

Arclabs will post blogs