Ghost peaks in HPLC and GC trace back to four material families, not one

A 20 µL injection of mobile A into a 2 mL vial that had been sitting in the autosampler tray for 18 hours produced a peak at 6.42 minutes with an area of 412 mAU·s. The same blank injected from a freshly opened vial gave 3 mAU·s at that retention time. The difference was not drift, not a lamp, not the column. It was the septa, and the ghost peak was the plasticizer package bleeding out of a 10 mm silicone-faced disc at 22 °C. That is the shape of most chromatography vial ghost peaks and extractables contamination I have chased in 15 years: a small, repeatable peak that appears only when the sample sits in contact with the wrong material for long enough, and disappears the moment you swap the consumable.

This article is about where those peaks come from, which compound classes bleed from which materials at which temperatures, and how to prove the source with blank and placebo injections instead of guessing. It is not about seal leaks or mechanical fit, which have their own failure signatures. Ghost peaks from vials are a materials problem, and they are traceable with a handful of controlled runs.

When a customer sends me a chromatogram with "unknown peaks at 4, 6.4, and 11 minutes, only in samples that sat overnight," I can usually narrow it to one of four sources before I ask for anything else. Each has a different thermal behavior and a different mass spectrum, and each is fixed by a different change to the consumable.

Septum extractables: the plasticizer and antioxidant package

A standard silicone/PTFE septum is not pure silicone. It is a compounded elastomer with a curing system, an antioxidant, and often a plasticizer or processing aid. Under HPLC conditions at ambient temperature, the low-molecular-weight fraction migrates into the sample by direct contact through the PTFE film and around the puncture. The classic offenders are phthalate plasticizers (dibutyl phthalate at m/z 149, diethylhexyl phthalate at m/z 149 with a different retention), 2,6-di-tert-butyl-4-methylphenol (BHT) and its oxidation products, and silicone oligomers (cyclic siloxanes at m/z 73, 207, 281, 355). In GC, the same septa off-gas into the inlet and produce the siloxane series that shows up as a rising baseline of peaks between 200 and 400 °C.

The temperature matters. A silicone septum at 22 °C in an HPLC tray bleeds slowly, and you see it after 12 to 24 hours of contact. The same septum in a GC inlet at 250 °C bleeds in minutes, and the bleed rate roughly doubles for every 10 to 15 °C above 200 °C. If your GC inlet is at 280 °C and your septum is a general-purpose silicone, you will see siloxane peaks in every blank. That is not a column problem, and no amount of conditioning will remove it.

Glass: sodium and boron leaching, and the alkaline surface

Type I borosilicate glass is the default for a chromatography vial, and it is the right default. But "Type I" is a hydrolytic resistance class, not a guarantee of zero surface activity. Soda-lime glass, which still shows up in low-cost vials, leaches sodium and calcium into aqueous samples, and the leached alkali raises the pH at the glass surface. For a pH-sensitive analyte or a compound that hydrolyzes, that local pH shift is a ghost peak generator. I have seen a 0.1% area peak at the void volume in a phosphate-buffered method that vanished when the customer moved from soda-lime to Type I borosilicate.

Borosilicate also has a surface silanol population. For basic compounds, those silanols act as cation exchange sites, and you get tailing and, occasionally, a late-eluting peak from material that adsorbed and then released. Deactivation (silanization) helps, but it is a surface treatment with a shelf life. A vial that has been stored open in a humid lab for a year has a different surface than one from a fresh case.

Caps and liners: the adhesive and the coating

The cap itself is usually polypropylene, which is relatively clean. The problem is the liner bond and any coating. A foil-lined or foil/silicone septum bonded with a pressure-sensitive adhesive can transfer adhesive residue to the sample if the puncture drags material into the vial. I have seen a peak at 8.1 minutes on a C18 method that matched an acrylic adhesive marker, and it only appeared when the same vial was punctured more than three times. The fix was a septum with a different bond chemistry, not a different cap color.

Polypropylene and the vial body in GC

In GC, the vial body is rarely the source because the sample is headspace or liquid injected through the septum. But in headspace GC, the polypropylene vial and the cap are both heated, and polypropylene at 80 to 120 °C releases oligomers and antioxidants. If your headspace oven runs at 90 °C and your blanks show a peak at m/z 207 or a series of alkane/alkene fragments, look at the vial and cap before the transfer line.

Which compound bleeds at which temperature: a working table

The table below is what I use when a customer sends a peak list and a method. It is not exhaustive, but it covers the material/compound/temperature combinations I have actually confirmed with a mass spec or a known standard. The temperature column is the point at which the bleed becomes visible in a blank at a typical injection volume, not the point at which the material fails.

Two things to notice. First, the HPLC failure mode is time, not temperature. A plasticized septum at 22 °C will not ruin a 30-minute run. It will ruin an overnight queue. Second, the GC failure mode is temperature, and it is fast. If you run a 280 °C inlet with a general-purpose septum, you are running a bleed experiment, not a blank.

Isolate the source with a blank ladder, not a single blank

The mistake I see most often is a single injection of mobile phase from one vial, followed by a shrug. One blank tells you there is a peak. It does not tell you where it came from. You need a ladder of blanks that changes one variable at a time, and you need to run them in the same sequence the sample sees.

The four-rung blank ladder

1. System blank. No vial, no injection, or a direct injection from the pump. If the peak is here, stop. It is the mobile phase, the needle, the loop, or the column, and the vial is innocent.
2. Fresh-vial blank. Mobile phase drawn from a vial opened that day, injected within 5 minutes. This is your baseline for the consumable.
3. Aged-vial blank. The same lot, filled and left in the tray for the same residence time as your sample queue (typically 12 to 24 hours). The difference between rung 2 and rung 3 is the extractables load.
4. Puncture-series blank. One vial, injected three to five times, with a blank between each. If the peak grows with puncture count, the septum or the liner bond is the source.

Run the ladder with the same needle, the same loop, and the same column you use for the method. A ladder run on a different instrument proves nothing about the instrument that is failing.

Placebo injections: the control that separates matrix from material

A placebo injection is the sample matrix without the analyte. If your method is a tablet assay, the placebo is the excipient blend dissolved in the same diluent. If your method is an environmental extract, the placebo is the clean matrix carried through the same extraction. The placebo tells you whether the ghost peak is from the vial or from the sample preparation, and it is the single most useful control in a contamination investigation.

The logic is simple. If the placebo from a fresh vial is clean and the placebo from an aged vial shows the peak, the vial is the source. If both placebos show the peak, the preparation is the source and the vial is a passenger. I have watched a lab spend three weeks on septa when the actual source was a plastic pipette tip used in the dilution step, and a single placebo injection would have caught it in an afternoon.

Confirming with a retention-time and mass match

Once you have a suspect peak, confirm it. On an HPLC with a diode array, pull the spectrum and compare it to a standard of the suspected plasticizer. On a GC-MS, match the mass spectrum to the siloxane or phthalate library entry. If you do not have a standard, the marker ions in the table above are enough to name the class. A peak at m/z 149 with a retention time that shifts with the septum lot is a phthalate, and you do not need a certificate to act on that.

Two field cases where the measurement named the culprit

Case 1: the 6.4-minute peak that only appeared after 16 hours

A contract lab running a reversed-phase method on a small-molecule API kept getting a peak at 6.42 minutes with an area of 300 to 450 mAU·s in the last third of every overnight queue. Fresh injections were clean. The system blank was clean. The column was new. The measurement that broke it open was the aged-vial blank: mobile phase in the same lot of vials, left in the tray for 16 hours, reproduced the peak at 410 mAU·s. A puncture-series blank showed the peak growing from 40 to 380 mAU·s across five punctures of the same vial. The septum was a general-purpose silicone with a phthalate plasticizer, and the mass spectrum at 6.42 minutes matched dibutyl phthalate.

The fix was a change of septum chemistry, not a change of autosampler. They moved to a PTFE-faced silicone septum with a low-extractable silicone compound and a documented extractables profile, and the peak dropped below the 10 mAU·s reporting threshold. The lesson: if the peak scales with tray residence time and puncture count, the septum is the source. If it scales with neither, look elsewhere. The HPLC septa selection is the lever here, and it is worth asking for the extractables data before you buy a case.

Case 2: the siloxane staircase in a 280 °C GC inlet

A environmental lab running SW-846 method 8270 on a GC-MS saw a rising series of peaks at m/z 73, 207, 281, and 355 in every blank, with the baseline climbing through the run. They had conditioned the column twice. The measurement that identified it was a blank run with the inlet at 250 °C versus 280 °C: the siloxane peak areas dropped by roughly 60% at the lower temperature, which is the signature of septum bleed, not column bleed. Column bleed does not change that much with a 30 °C inlet change. They swapped to a lower-bleed septum and a different inlet liner, and the staircase flattened. The SW-846 method does not specify the septum, which is exactly why the lab had to find it themselves.

A decision checklist for the bench

When a ghost peak shows up and the vial is a suspect, work this list in order. It takes about two hours and it will save you a week.

• Run the system blank first. If the peak is there, the vial is not the source. Move to the mobile phase, needle, and loop.
• Compare fresh-vial and aged-vial blanks. A peak that appears only in the aged vial is an extractables problem, and the residence time tells you the severity.
• Run a puncture series. Growth with puncture count points to the septum or the liner bond.
• Run a placebo. A clean placebo from a fresh vial and a dirty placebo from an aged vial isolates the material. A dirty placebo from both isolates the preparation.
• Check the inlet temperature against the septum rating. If the inlet is above 250 °C and the septum is general-purpose silicone, replace it before you do anything else.
• Check the glass class. If the peak is at the void volume in an aqueous method, confirm Type I borosilicate and consider a deactivated surface.
• Check the liner bond. If the peak appears only after multiple punctures, the adhesive is the suspect.
• Confirm with a standard or a mass match. Name the compound class before you change the consumable, so you can verify the fix worked.

For the consumable side, the chromatography vial you choose sets the glass class and the cap geometry, and the HPLC vial supplier's extractables documentation is what you actually need to compare. If you are building a method from scratch, the chromatography products catalog is a reasonable place to compare septum chemistries side by side. For instrument-side guidance, the Aijiren GC columns and accessories documentation and the Thermo Fisher chromatography technical notes both publish septum bleed data that is worth reading before you pick a lot. Restek also publishes bleed profiles that are useful for matching a suspected peak to a material.

What to specify so the problem does not come back

The fix is not a single part number. It is a specification you put on the purchase order and a control you run in the method. Here is what I put on the spec sheet when a lab asks me to solve this permanently.

• Septum: PTFE-faced silicone with a documented extractables profile, low-bleed grade, rated for the inlet temperature with at least a 20 °C margin. For HPLC, specify a septum that has been tested for phthalate and BHT extractables at 24-hour contact.
• Glass: Type I borosilicate, USP <660> hydrolytic resistance, with a deactivated surface for basic analytes. For ion chromatography, specify the sodium and calcium leachables limit.
• Cap: polypropylene, with the liner bond chemistry stated. If you puncture more than twice per vial, ask for a liner that holds its seal and does not transfer adhesive.
• Residence time: cap the autosampler queue at the validated residence time, and put an aged-vial blank at the end of every queue. If the aged blank shows the peak, the queue is too long or the septum is wrong.
• Torque: for screw caps, hand-tight plus a quarter turn is the working rule, and a torque of about 0.8 to 1.2 N·m is the range I have seen hold a seal without deforming the liner. Over-torque crushes the liner and increases extractables transfer, so tighter is not better here.
• Puncture count: replace the septum after three to five punctures for HPLC, and after one to two for a high-temperature GC inlet. That is the practical limit before the bleed and the coring risk both climb.

One more control that costs nothing: run a blank from a vial that has been in the tray for the full queue time, every day. If that blank stays under your reporting threshold, your consumable is doing its job. If it climbs, you have a materials problem and you have caught it before it reached a sample.

The decision rule is this: if a ghost peak scales with tray residence time or puncture count, the vial materials are the source, and the fix is a septum and glass specification, not an instrument service call. If it does not scale with either, stop blaming the vial and run the placebo.

References & further reading

Related resources from our lab consumables network:

• Chromatography Vial at chromvials.com
• Hplc Vial at hplcvialfactory.com
• Chromatography Products at chromatographyforsale.com
• Hplc Septa at hplcsepta.com

External references and standards:

• Aijiren 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