Ghost peaks that scale with tray occupancy point at the consumable, not the

A 30 m x 0.25 mm x 0.25 µm DB-5 column that had run clean blanks for two years suddenly showed a repeating train of peaks at retention times of 14.2, 15.8, 17.1, and 19.6 minutes, each at roughly 0.3 to 0.8 pA full scale, appearing only when the autosampler tray held a full batch of 2 mL vials with pre-slit silicone/PTFE caps. Empty needle positions were clean. That single comparison, tray loaded versus tray empty, is the fastest way to separate chromatography supplier ghost peaks and extractables from contamination that came in with the sample. What follows is how I trace that class of artifact back to the consumable lot, the material, and the temperature that released it.

Ghost peaks that scale with tray occupancy point at the consumable, not the column

The diagnostic logic is simple and I have used it for fifteen years. If the artifact disappears when the tray is empty but returns when it is loaded, the source is something in the vial, the cap, the septum, or the septa of the inlet. If it persists with an empty tray and a blank injection, the source is upstream of the sample: carrier gas, inlet liner, septum, or column. If it only appears in sample injections and never in solvent blanks, the sample matrix is releasing it. Three tests, three different answers.

Extractables from consumables have a signature. They appear at consistent retention times across every vial in a lot, they scale with oven temperature (higher final hold, larger peak), and they often come in homologous series because what is bleeding is a polymer additive or a plasticizer with a distribution of oligomers. Siloxane trains at m/z 73, 147, 207, 281, and 341 in GC-MS are the classic fingerprint of a silicone septum. Phthalate ions at m/z 149 are the fingerprint of a plasticizer from a cap liner or a bag. If you see either pattern, stop looking at the column and start looking at what touched the sample.

The primary query here, chromatography supplier ghost peaks, extractables and contamination, is really a question about which material releases what, at what temperature, and how you prove it with a blank and a placebo. That is the whole article. The chromatography supplier you buy from either controls this or does not, and the difference shows up in a chromatogram, not a certificate.

Which polymer bleeds which compound at which temperature

Materials do not fail generically. Each polymer family has a characteristic extractable profile and a temperature above which it starts to give it up. I keep this table taped inside a lab notebook because it answers 80% of the calls I get.

Two numbers matter most. Silicone starts to surrender siloxanes in a meaningful way around 200 °C, and a general-purpose septum in a hot inlet at 280 °C is a siloxane generator no matter who made it. If your method runs the inlet at 280 °C, you need a platinum-cured low-bleed septum or a PTFE-faced one, and the price difference per hundred is trivial against the cost of a re-run. The second number is the plasticizer onset temperature: PVC and adhesive layers start releasing phthalates and acrylates below 100 °C, which means the vial sitting on an autosampler tray at ambient for six hours is already a source if the cap has a glued liner.

Why the temperature that matters is often the tray, not the oven

People forget that the autosampler tray, the sample compartment, and the transfer line are all part of the thermal history. A vial capped at 09:00 and injected at 15:00 has spent six hours at ambient with the septum in contact with headspace vapor. If the headspace contains solvent, the septum is being extracted in situ, and the extractables drop back into the vial. I have seen a 0.5 pA ghost peak at 11.3 minutes grow from run 1 to run 60 in a sequence because each vial sat longer than the last. The fix was not the column. The fix was a run time under four hours and a septum with a PTFE face that does not swell in the solvent.

Blank and placebo injections, run in the right order, isolate the source in under an hour

Most labs run a solvent blank and call it contamination control. A solvent blank tells you almost nothing about consumables because the consumable has to be in the path. The sequence below takes about 45 minutes and it will tell you which component is guilty.

1. Instrument blank, no vial, no injection. Empty tray. This captures column bleed, carrier gas, and inlet residue. If peaks appear here, the consumable is upstream of the sample and you are looking at the inlet septum or liner, not the vial.
2. Solvent blank through a new vial from the suspect lot. Same solvent, same volume, capped with the suspect cap. If the ghost peaks appear, the vial or cap is releasing them. If they do not, the vial is clean under this method.
3. Solvent blank through a vial from a known-good lot. If the suspect lot shows peaks and the known-good lot does not, you have a lot-level problem and you can call the supplier with data.
4. Placebo injection. Matrix without analyte, prepared in the suspect container and processed through the full sample prep. This is the only test that catches extractables generated during prep: a filter, a pipette tip, a transfer bag.
5. Repeat the placebo with a delay. Hold the capped vial on the tray for the full sequence duration, then inject. This catches the time-dependent extraction that a fresh vial hides.

Run these in that order because each one narrows the search. If step 1 is clean and step 2 shows peaks, you have skipped the entire inlet and column question. If step 1 shows peaks, you can ignore every vial in the batch. I have watched analysts spend a week on a column that was fine because they never ran step 1.

Reading the blank the way a mass spec reads it

If you have MS, extract the suspect peak and look at the ion pattern before you touch anything. Siloxane at 73/147/207 means silicone, full stop. Phthalate at 149 means plasticizer, which means a flexible polymer somewhere in the path. A homologous series spaced by 14 Da is an oligomer, which means a polyolefin. A single sharp peak with a strong molecular ion and no series is more likely a solvent impurity or a carryover from the previous injection. That ion pattern decides whether you replace a septum, a cap lot, or a filter, and it takes ten minutes.

If you only have FID, use retention index matching. Run a standard alkane mix and compute Kovats indices for the ghost peaks. Siloxane D5 sits near 990 to 1000, D6 near 1180, and the phthalate series is unmistakable above 1500. A retention index is enough to classify the compound class even without a mass spec, and it is defensible in a deviation report.

Field case: the 0.9 pA peak that only appeared after lunch

A contract lab running EPA 8270 semivolatiles on a 30 m x 0.25 mm x 0.25 µm column reported a ghost peak at 16.4 minutes, 0.9 pA, appearing in every sample injected after 13:00 but never in the morning. The morning blanks were clean. The afternoon blanks were not. The column had been in service three weeks.

The measurement that identified it: they ran the delayed-placebo test above and reproduced the peak with a solvent blank that had sat capped on the tray for five hours. The peak was absent at one hour. They then swapped caps between two lots and the peak followed the cap lot, not the vial body. The cap was a laminated silicone/PTFE closure with an acrylic adhesive layer, and the solvent was dichloromethane, which swelled the adhesive and extracted acrylate monomers over time. The ion at m/z 55 and 69 confirmed it. The fix was a vial closure with a molded PTFE-faced liner and no adhesive layer, and a sequence cap of four hours. The peak never returned. The lesson: the closure is a chemical component, not a lid, and the adhesive is the part nobody specs.

Field case: the siloxane train that came from a bag, not a septum

A environmental lab running volatiles by purge and trap saw a repeating siloxane train at 73/147/207 in every sample, at roughly 0.2 to 0.4 pA, with the largest peak at 21 minutes. They replaced the inlet septum, the liner, and finally the column. No change. The instrument blank was clean, which ruled out the inlet. The solvent blank through a fresh vial was clean, which ruled out the vial and cap.

The measurement that found it: they ran the placebo through the full prep, including the transfer step, and the train appeared. The samples were being transferred from a 40 mL VOC vial into a 5 mL autosampler vial through a length of silicone pump tubing that had been in service for months. Silicone tubing is a siloxane reservoir, and every transfer pulled a fresh slug of it into the vial. The fix was PTFE-lined transfer tubing and a replacement schedule of 30 days. The cost of the tubing change was under $200. The cost of the investigation was eleven working days.

This is the pattern I see most often: the consumable that is not on the sample path in the analyst’s mental model is the one on the sample path in reality. Transfer lines, filter housings, pipette tips, and septa all touch the sample. The analytical testing lab that maps every contact point before it starts swapping parts finds the source faster every time.

What to demand from a supplier before the lot reaches your bench

You cannot test your way out of a bad lot one vial at a time. The supplier has to control extractables at the material level, and you should ask for the specific evidence. Here is the decision checklist I use when qualifying a new source.

The extractable profile line is the one that matters and the one suppliers resist. A supplier who cannot produce a blank chromatogram of their own closure at your method temperature is asking you to be the QA lab. A supplier who can is telling you they have already run the test. That is the difference between a catalog and a qualified source, and it is why the OEM vial route often beats a relabeled import: the molding source is known and the resin lot is traceable.

Temperature ratings on a datasheet are not the same as extractable ratings

A cap rated to 300 °C means the polymer does not melt or deform at 300 °C. It does not mean the polymer is clean at 300 °C. Those are two different specifications and suppliers conflate them constantly. A silicone septum rated to 300 °C will still bleed siloxanes at 250 °C. Ask for the extractable rating, which is a chromatographic number, not a thermal one. If the datasheet only gives a maximum use temperature, treat it as a mechanical limit and assume nothing about cleanliness.

A short isolation procedure you can run today

When a ghost peak appears and you suspect consumables, do this in order and stop as soon as the source is isolated.

• Step 1: Instrument blank with an empty tray. Clean means the inlet and column are not the source.
• Step 2: Solvent blank through a suspect-lot vial and cap. Peaks here mean the vial or cap is the source.
• Step 3: Swap only the cap to a known-good lot. If the peak disappears, the cap lot is guilty. If it stays, the vial body is.
• Step 4: If steps 2 and 3 are clean, run the placebo through the full prep. Peaks here mean a filter, tip, or transfer line.
• Step 5: Confirm with MS ion pattern or retention index before you write the deviation.
• Step 6: Quarantine the lot, document the ion pattern and the lot number, and send both to the supplier.

Two operational precautions cut the recurrence rate. First, cap the sequence at four hours for solvent-heavy methods and re-inject rather than let vials sit. Second, keep a known-good reference lot of vials and caps in a sealed bag and use it as your control every time you qualify a new lot. The reference lot is your baseline, and without it you are comparing a suspect lot to a memory.

For method-level guidance on container and handling, the source documents are worth reading directly: U.S. EPA SW-846 treats the container as a method parameter, and the major instrument vendors publish consumable compatibility tables that list temperature and solvent limits per material. Agilent’s GC consumables documentation, Thermo Fisher’s sample handling guidance, and the Restek technical library all publish blank chromatograms and material compatibility data. Use them as a cross-check against what your supplier claims, because the vendor who makes the instrument has no incentive to hide a bleed problem in a septum.

The decision rule is this: a ghost peak that scales with tray occupancy or sequence time is a consumable extractable until proven otherwise, and you prove it with an instrument blank, a suspect-lot blank, a cap swap, and a placebo, in that order, before you replace a single piece of hardware. The ion pattern tells you the material class, the lot number tells you who to call, and the blank chromatogram on the supplier’s datasheet tells you whether you should have called them before you bought the lot.

References & further reading

Related resources from our lab consumables network:

• Chromatography Supplier at chromatographysupplier.com
• Oem Vial at oemsamplevials.com
• Vial Closures at vialclosures.com
• Analytical Testing at labanalyticaltesting.com

External references and standards:

• Agilent chromatography consumables and columns
• Thermo Fisher chromatography and sample handling
• U.S. EPA SW-846 test methods for waste and materials
• Restek technical resources for chromatography