PFA vs Glass Vials: Which for Trace Metal Analysis?

# PFA vs Glass Vials: Which for Trace Metal Analysis?

I’ve run enough ICP-MS sequences to know that the vial is not always the quiet bystander it looks like. If your lab reports metals down to ppb or ppt levels, the container can add or subtract more signal than your instrument does. And if you run ppm-level work, you can probably ignore all of this and save your budget.

For sub-ppb trace metal work (ICP-MS, GFAAS, low-level ICP-OES), PFA vials win: borosilicate glass leaches boron, sodium, silicon, and potassium straight into your blanks. For routine ppm-level or non-metal work, standard glass is fine and far cheaper. That’s the whole decision in two sentences.

Here’s how I actually make the call for each method, plus the cost math nobody shows you until the quote arrives. The details below include one experiment that cost my old lab a week of re-runs.

Why borosilicate glass works against you at trace levels

Borosilicate glass is roughly 80% silicon dioxide, with boric oxide, sodium oxide, and potassium oxide mixed in. It’s excellent glass for chromatography. It’s a terrible container when the elements it’s made of are your analytes.

In dilute acid, which is exactly how metals samples and standards are stored, the glass surface slowly releases boron, sodium, silicon, and potassium into solution. A Chinese research group measured this directly: boric acid solutions left in borosilicate volumetric flasks for one hour already showed boron contamination in the low µg/L range, and the contamination climbed with acid strength and holding time. Pharma packaging studies tell the same story, with boron, sodium, and silicon detectable by ICP-MS after weeks of contact.

For context, modern ICP-MS (inductively coupled plasma mass spectrometry, which measures elements by ionizing them and sorting the ions by mass) routinely quantifies single-digit ng/L. Your glass vial quietly contributes a competing background right in that range.

I learned this the expensive way. Years ago I inherited a boron method where the calibration blank drifted upward all afternoon and the QC recovery climbed past 110% by run fifteen. The cause was a borosilicate volumetric flask used to hold the intermediate standard. Every hour it sat there, the standard got “richer.” We re-ran the whole week’s samples after switching to PFA for standards.

What makes PFA different

PFA (perfluoroalkoxy) is a fluoropolymer, the same chemical family as Teflon. Its carbon backbone is fully wrapped in fluorine atoms, and fluorine doesn’t want to exchange with anything. There is no boron, sodium, or potassium in the material to leach, and manufacturers mold it from high-purity resin with near-zero metal content.

Three practical consequences:

1. Blank levels. PFA labware gives you blanks near the instrument’s true background instead of a glass-derived offset. This is the difference between a 0.01 µg/L boron blank and a “where did that come from” boron blank. It’s also why certified materials like NIST SRM 1643f ship with strict container guidance; at ng/L levels the container is part of the answer.

2. HF and strong acid resistance. PFA handles hydrofluoric acid, aqua regia, and concentrated nitric acid, which matters if you digest samples in the same vessels you measure from. Glass etches in HF on contact. Geochemistry clean labs have built their whole workflow around Savillex PFA labware for decades, and their catalog shows why: vials rated for service from -200°C to 260°C with almost no extractables.

3. Lower adsorption losses. The walls can grab analytes too, not just donate them. Mercury is the famous case. One review tabulated decades of studies where mercury standards at µg/L levels lost 40 to 100% of their concentration to container walls within days, in glass and plastic alike, depending on preservation. Fluoropolymer surfaces are consistently among the least adsorptive options, which is why EPA mercury methods call for glass or fluoropolymer bottles.

Quick comparison:

The cost question, honestly

A 15 mL PFA vial runs about $46 a pop from Savillex, and even their 1.5 mL autosampler vial is listed at $13.57. The equivalent borosilicate autosampler vial costs cents. That’s the 10-50x premium people quote, and it’s real.

So don’t pay it everywhere. I’d spend it in three places only: calibration standards, digests you can’t repeat, and any method with a reporting limit below roughly 1 µg/L. That covers environmental work under EPA Method 6020B, which targets sub-µg/L elements in waters and digests, and pharma labs chasing the low limits in USP elemental impurities testing. One ruined sequence costs more in instrument time than a rack of PFA vials. For everything else, the money belongs elsewhere in the lab.

When glass is perfectly fine

If your samples sit in the tens of µg/L to mg/L range, glass leaching is a rounding error. A µg/L of boron means nothing when your sample contains 500 µg/L.

Same goes for analytes that aren’t elements. Running HPLC, GC, or dissolved organics? The vial material conversation changes completely, and glass is usually the answer. I keep borosilicate in our HPLC room and nobody has lost sleep over it. If you want to see how the materials compare for LC work, I covered it in our guide to HPLC vial materials, glass vs plastic.

One more glass-friendly case: non-metal elements measured where the container doesn’t interfere. Carbon, chlorine, nitrogen work in glass all day.

Cleaning and acid-washing protocols

PFA survives cleaning cycles that would slowly strip glass. A common protocol: soak new PFA ware in 10-20% high-purity nitric acid at 50-60°C for a day or two, rinse with ultrapure water, then repeat before first use. For Hg work, labs often add a dilute HCl or bromine step because mercury clings to previously wetted surfaces.

Glass can be acid-washed too, and for ppm work it usually is. The catch is that no cleaning removes the composition of the glass itself. You can scrub borosilicate until the acid is fresh, and it will still donate boron into your next blank. Cleaning removes what sits on the surface, not what the surface is made of.

Autosampler compatibility

This is where good intentions die. Your ICP-MS or GFAAS autosampler tray was designed around specific vial geometry, and PFA vials don’t clone standard 12 x 32 mm dimensions. Check the tray type and any required adapters before ordering. Savillex publishes which vials fit Agilent and CETAC trays right in the product listing.

Also think about the whole wetted path. A PFA vial feeding a glass transfer tube or a glass spray chamber reintroduces the surfaces you paid to avoid. For the lowest blanks, I’d match the entire sample contact path to the analyte list, which is the same logic behind the consumables we laid out in ICP-MS sample preparation.

One small warning: PFA is translucent, not clear. You won’t see the meniscus the way you do in glass, so fill by weight or by marked graduation rather than by eye.

Conclusion

The honest default: if your reporting limits are below about 1 µg/L, buy PFA for standards, digests, and final samples, and treat the cost as insurance on instrument time. If you run ppm-level metals, organics, or routine QC, borosilicate glass remains the right tool and the cheaper one. The mistake I see most often is not choosing the wrong material, it’s applying one material to every job out of habit. Sit down with your method’s reporting limits and your element list, and let those two facts decide. For HF digests, mercury work, or any boron-adjacent method, glass shouldn’t even be in the conversation. And if you’re weighing polypropylene as a middle ground, it works for many general dilutes, though it has its own leaching profile, which we examined in our polypropylene vial leaching data and in what MS certified vials actually guarantee.

Frequently Asked Questions

Does borosilicate glass really leach boron into samples?

Yes, and it’s measurable at ng/L concentrations in dilute acid within an hour. The effect scales with acid strength, contact time, and temperature, which is exactly the condition profile of metals samples waiting for analysis.

Can I use PFA vials with hydrofluoric acid?

You can, and that’s one of PFA’s strongest use cases since HF digests are common in geochemistry and materials work. Glass etches on contact with HF, so any HF-containing solution needs fluoropolymer containment end to end.

How many times can I reuse a PFA vial?

Labs commonly reuse PFA vials for years with a proper nitric acid cleaning cycle between uses. Inspect for crazing or discoloration, and dedicate cleaned vials to either high-level or low-level work so you don’t carry contamination backwards.

Are PFA vials compatible with my ICP-MS autosampler?

Some are, but the geometry is not standardized like borosilicate autosampler vials, so check the tray model and adapter requirements before you buy. Savillex lists Agilent and CETAC compatibility in their product descriptions.

Is polypropylene a good cheaper alternative to PFA?

For general dilute-acid work at moderate trace levels, high-purity polypropylene is a reasonable middle ground and much cheaper. It falls short for HF, for mercury, and for the lowest blanks, where PFA is still the right call.