GC Troubleshooting

• Evaluating The Problem
• Checking The Obvious
• Ghost Peaks or Carryover
• Excessive Baseline Noise
• Baseline Instability or Disturbances
• Tailing Peaks
• Split Peaks
• Retention Time
• Change in Peak Size
• Loss of Resolution
• Condensation Test
• "Low-Bleed" Columns - Fact or Fiction?

Evaluating The Problem
The first step in any troubleshooting effort is to step back and evaluate the situation. Rushing to solve the problem often results in a critical piece of important information being overlooked or neglected. In addition to the problem, look for any other changes or differences in the chromatogram. Many problems are accompanied by other symptoms. Retention time shifts, altered baseline noise or drift, or peak shape changes are only a few of the other clues that often point to or narrow the list of possible causes. Finally, make note of any changes or differences involving the sample. Solvents, vials, pipettes, storage conditions, sample age, extraction or preparation techniques, or any other factor influencing the sample environment can be responsible.

Back to Top

Checking The Obvious
A surprising number of problems involve fairly simple and often overlooked components of the GC system or analysis. Many of these items are transparent in the daily operation of the GC and are often taken for granted (set it and forget it). The areas and items to check include:

• Gases - pressures, carrier gas average linear velocity, and flow rates (detector, split vent, septum purge).
• Temperatures - column, injector, detector and transfer lines.
• System parameters - purge activation times, detector attenuation and range, mass ranges, etc.
• Gas lines and traps - cleanliness, leaks, expiration.
• Injector consumables - septa, liners, O-rings and ferrules.
• Sample integrity - concentration, degradation, solvent, storage.
• Syringes - handling technique, leaks, needle sharpness, cleanliness.
• Data system - settings and connections.

Back to Top

Ghost Peaks or Carryover
System contamination is responsible for most ghost peaks or carryover problems. If the extra ghost peaks are similar in width to the sample peaks (with similar retention times), the contaminants were most likely introduced into the column at the same time as the sample. The extra compounds may be present in the injector (i.e., contamination) or in the sample itself. Impurities in solvents, vials, caps and syringes are only some of the possible sources. Injecting sample and solvent blanks may help to find possible sources of the contaminants. If the ghost peaks are much broader than the sample peaks, the contaminants were most likely already in the column when the injection was made. These compounds were still in the column when a previous GC run was terminated. They elute during a later run and are often very broad. Sometimes numerous ghost peaks from multiple injections overlap and elute as a hump or blob. This often takes on the appearance of baseline drift or wander.

Increasing the final temperature or time in the temperature program is one method to minimize or eliminate a ghost peak problem. Alternatively, a short bake-out after each run or series of runs may remove the highly retained compounds from the column before they cause a problem. Performing a condensation test is a good method to determine whether a contaminated injector is the source of the carryover or ghost peaks.

Back to Top

Excessive Baseline Noise

Back to Top

Baseline Instability or Disturbances

Back to Top

Tailing Peaks

Back to Top

Split Peaks

Back to Top

Retention Time

Back to Top

Change in Peak Size

Back to Top

Loss of Resolution

Back to Top

Condensation Test
Use this test whenever injector or carrier gas contamination problems are suspected (e.g., ghost peaks or erratic baselines).

• Leave the GC at 40-50°C for 8 or more hours.
• Run a blank analysis (i.e., start the GC, but with no injection) using the normal temperature conditions and instrument settings.
• Collect the chromatogram for this blank run.
• Immediately repeat the blank run as soon as the first one is completed. Do not allow more than 5 minutes to elapse before starting the second blank run.
• Collect the chromatogram for the second blank run and compare it to the first chromatogram.
• If the FIRST chromatogram contains a substantially larger amount of peaks and baseline instability, then that is an indication that there is contamination upstream of the capillary column (ie. contaminated inlet, dirty carrier gas, etc.).
• If BOTH chromatograms contain few peaks or very little baseline drift, it can be assumed that the carrier gas and/or inlet are relatively clean.
• If BOTH chromatograms contain a significant amount of noise and/or baseline drift, then that usually is an indication that the detector or detector gases are contaminated.
  

.

Back to Top

"Low-Bleed" Columns - Fact or Fiction?
Prof. Walt Jennings
Cofounder, J&W Scientific Incorporated

Several manufacturers offer "low bleed" columns. In some cases, these are merely selected from the standard production process, but in other cases the columns are actually "synthesized" for low bleed. In recent years, it has been established that where functional groups (i.e. phenyl) are inserted into the polysiloxane chain as aryl inclusions, as opposed to being attached to the chain as pendant groups, the resultant phase possesses increased thermal and oxidative resistance. Columns coated with such phases emit lower levels of bleed signal and are capable of going to higher temperatures. The increased thermal resistance is apparent only at temperatures above ca. 300 degrees. While some users can reap the benefits of these developments, others find little or no improvement.. their bleed signals are still too high.

True column bleed, of course, comes only from the column. What the user perceives as bleed is usually the total signal reaching the detector, which is the summation of signal from the septum (this gives a typical silicone mass spectrum), the injector, and the detector, all of which is usually blamed on the column.

It is good procedure to first check the detector. Disconnect and remove the column, and place an undrilled cap on the column attachment fitting. Activate the detector, and note the signal at 50 degrees. Increase the oven temperature to 320 degrees, and again note the signal. On a pristine detector, the FID signal will increase by one to two picoamps. If the increase exceeds this level, attention should be directed to cleaning the detector, make-up gas and hydrogen lines. Once the detector signal falls to an acceptable level at 320 degrees, attention should be directed to the injector. If the injector liner is visibly soiled, the injector should be cooled, dissembled and interior cavities scrubbed with solvent and natural bristle brushes or cotton swabs. After assembling the injector, a "jumper tube" (one to three meters of uncoated fused silica or steel tubing) is then used to connect the injector directly to the detector. The injector heater should be energized, and the oven set at 320 degrees. Any increase in "bleed" signal over that observed with the detector alone must come from the front end of the instrument, and may originate with the septum, the carrier gas line, in-line regulators, valves, or flow controllers.

Wrap a new septum in aluminum foil, ensuring that one face is smooth, and install this, smooth side down. If the signal emanating from the jumper tube is decreased, it indicates a need for better quality septa. If the signal is still high, materials entrained in the carrier gas may have deposited in lines, valves, or regulators, which should be dissembled and cleaned or replaced.

When the combined signal from the injector and detector falls to an acceptable level (one to two picoamps @ 320 degrees on an FID), the user is ready to install and reap the benefits of a true low-bleed column. The bleed rate of conventional columns is normally high enough to mask signal from the injector and detector unless these latter are heavily contaminated. With low bleed columns, the signal from the injector and detector assumes increased importance. This spurious signal is not infrequently limiting, and is usually (and incorrectly) perceived as "column bleed".

Back to Top