Tested values carry a method, a tolerance, and a batch record

Last fall I watched a QC lab reject 4,000 autosampler vials because the lot drifted outside a 1.2% RSD window on a residual solvent method. The purchasing manager pulled up the supplier spec sheet and pointed at the line that said “silanized.” That word was doing a lot of work. The sheet did not say how the silanization was verified, what surface energy threshold was used, or whether that batch had been tested at all. It was a nominal claim dressed up like a tested specification. If you read chromatography products datasheet specifications the way you read a certificate of analysis, you will keep making that same mistake. Most spec sheets are marketing documents with a few real numbers buried in them. The skill is knowing which numbers were earned in a QC lab and which ones were typed by a product manager.

A tested specification has three things attached to it. There is a test method, usually a reference to an ASTM, ISO, USP, or internal SOP number. There is a tolerance band, like ±0.05 mm or ≤1.5% RSD. And there is a batch or lot record that ties the number on the sheet to the physical goods in your hand. If any of those three is missing, you are looking at a nominal value.

Take a 9 mm screw-top vial. The sheet says 12 × 32 mm. That sounds tested. But the outer diameter and height are almost always nominal because they are controlled by the glass tubing drawing process, not by post-forming inspection. A reputable manufacturer will still measure them, but the tolerance is often ±0.25 mm on diameter and ±0.5 mm on height. The dimension that matters for sealing is the neck finish, specifically the GPI 9-425 thread profile. That one is tested because a 0.1 mm deviation in thread pitch will cause the cap to cross-thread or fail to seal at the rated torque. When I evaluate a vial datasheet, I look for the GPI reference first. If it is absent, I assume the neck was not verified against the standard and I move on.

For septa, the tested line is puncture count. A sheet that says “resilient” or “multi-injection” without a number is nominal. A sheet that says 50 punctures at 23 gauge with a 22° bevel needle is tested, because it specifies the needle geometry, the insertion angle, and the count. The needle matters enormously. A 23 gauge needle with a sharp point will make a clean cut. A blunt or slightly hooked needle will core the septum. I have seen a “50 puncture” septum fail at 12 injections because the lab switched from a standard 23 gauge cone-tip to a cheaper 23 gauge bevel that was not specified on the sheet. The septum was fine. The test conditions did not match the use conditions.

Nominal values are design targets, not batch guarantees

A nominal value is the number the product was designed around. It is the center of a distribution that may or may not be controlled. The classic example is pore size on a filter membrane. A sheet that says “0.45 µm” is giving you the design target. The actual pore size distribution is a curve. A good datasheet will tell you the bubble point in psi or bar, because that is the tested value that correlates with the largest pore in the membrane. For a 0.45 µm PTFE membrane, the bubble point should be around 13 to 15 psi when wetted with isopropanol. If the sheet lists only the pore size and no bubble point, you cannot tell whether the lot has a tail of oversized pores that will let particulate through.

The same logic applies to surface area on silica columns. A sheet that says “200 m²/g” is nominal unless it also gives the BET method reference and a tolerance, usually ±10%. Silica surface area drifts between production lots because it depends on the hydrolysis and condensation conditions. A column packed with 180 m²/g silica will not behave the same as one packed with 220 m²/g silica, even if both sheets say “200.” The carbon load, usually expressed as a percentage, is another one. A C18 column with 12% carbon load is tested, but the test is often done on the bulk silica before bonding, not on the packed column. The packed column performance is verified by a chromatographic test, which is a different line on the sheet.

The chromatographic test is the only line that predicts your separation

Every column datasheet has a section that lists retention time, plate count, tailing factor, and resolution for a specific test mix. This is the most valuable part of the sheet and the most frequently ignored. The numbers are tested, but only for the conditions printed next to them. A sheet that says N ≥ 80,000 plates/meter for naphthalene on a 5 µm C18 column at 1.0 mL/min with 70:30 acetonitrile:water is a real measurement. But it tells you nothing about how the column will perform on your method with your mobile phase and your analyte.

What you are looking for is the tailing factor, usually specified as Tf ≤ 2.0 or As ≤ 1.5 for a basic compound. Silica with residual silanols will tail basic analytes. A manufacturer that tests with amitriptyline or pyridine and reports the tailing factor is showing you that their endcapping works. A manufacturer that tests only with neutral compounds like toluene or naphthalene is hiding something. I have rejected columns from three different suppliers because the sheet showed a beautiful plate count on naphthalene but said nothing about basic compound tailing. When I ran propranolol on those columns, the tailing factor was 2.8 to 3.2. The sheet was honest. It just was not useful.

For GC columns, the equivalent line is the Grob test or a simplified version of it. A tested sheet will list the retention indices and peak shapes for a mix that includes a base, an acid, and a polar compound. The acid and base peaks tell you about active sites in the tubing. If the sheet only lists a hydrocarbon mix, the column may have active sites that will destroy your polar analytes. The Restek technical resources have good examples of what a full Grob test looks like and which peaks indicate specific surface problems.

Temperature and pressure ratings: tested to failure or copied from a competitor

Vial and cap temperature ratings are a mess. A sheet that says −60°C to 200°C for a PTFE/silicone septum is giving you the material limits, not the assembly limits. The PTFE face can handle 200°C. The silicone core can handle 200°C. But the cap, the vial neck, and the way they interact at temperature are a different story. A polypropylene cap will soften above 135°C. If you run a headspace method at 150°C with a polypropylene cap, the cap will deform and the seal will fail, regardless of what the septum material is rated for.

The tested value you want is the seal integrity at temperature. Some manufacturers run a helium leak test on capped vials at 150°C for 30 minutes and report the leak rate in cc/sec. That is a real test. A sheet that just lists the material temperature range is nominal. The same applies to pressure ratings on HPLC fittings. A fitting rated to 1,200 bar was tested to that pressure with a specific ferrule material and tubing diameter. Change the ferrule from stainless steel to PEEK and the rating drops. Change the tubing from 1/16 inch to 1/32 inch and the rating changes again. The sheet should specify the test conditions. If it does not, the number is a design estimate.

How to spot a sheet that hides its test method

There are four patterns I look for when a datasheet feels evasive. First, the sheet uses adjectives where numbers should be. “High purity,” “low bleed,” “excellent inertness,” and “superior reproducibility” are nominal claims. They mean nothing without a measurement. Second, the sheet lists a test but no tolerance. “Purity 99.9%” without a ± or a ≥ is not a specification. It is a marketing average. Third, the sheet references a standard but not a method. “Meets USP requirements” is common on vial sheets. The USP chapter for containers is a system test, not a parts list. It does not tell you which specific test was run or what the pass criteria were. Fourth, the sheet gives a number that is physically impossible to verify on the finished product. “Residual silanol 1.5 µmol/m²” is a nice number, but it is measured on the bulk silica, not on the packed column. The packed column may have a completely different surface chemistry after the bonding and endcapping steps.

When I see any of these patterns, I ask the supplier for the certificate of analysis for the specific lot I am buying. The COA is where the tested values live. The datasheet is where the nominal values live. If the supplier will not provide a COA before purchase, I assume the tested values do not exist. That is not always fair, but it is a useful filter when you are comparing three suppliers and one of them cannot produce a lot-specific document.

A decision checklist for reading any chromatography products datasheet

Here is the checklist I use when I sit down with a new spec sheet. It is not exhaustive, but it catches the most common deceptions.

Field case one: the silanized vial that was not silanized

A contract lab running a pesticide residue method by GC-MS started seeing carryover peaks for chlorpyrifos at 2 to 5% of the previous injection. The method used 2 mL amber vials with a silanized surface. The supplier sheet said “silanized for low adsorption.” The lab had been buying that part number for two years without issue. The new lot was different.

The measurement that identified the problem was a surface energy test using a dyne pen set. The good vials from the previous lot gave a surface energy below 30 dynes/cm on the inner wall. The new lot gave 42 to 45 dynes/cm, which is close to untreated borosilicate glass. The silanization had either been skipped or was incomplete. The fix was to return the lot and switch to a supplier that provides a lot-specific surface energy value on the COA. The lesson: a nominal claim on a datasheet will not catch a batch that missed a processing step. Only a tested value with a batch record will.

When I evaluate silanized vials now, I ask for the dyne level and the test method. A sheet that says “silanized” without a dyne number is a nominal claim. A sheet that says ≤32 dynes/cm by ASTM D2578 is tested. The difference is the difference between a product that works and a product that costs you a re-run.

Field case two: the 0.45 µm filter that was not 0.45 µm

A pharmaceutical QC lab was filtering a viscous oral suspension before HPLC analysis. The method called for a 0.45 µm PTFE syringe filter. The lab switched to a lower-cost supplier and immediately saw column backpressure spikes and a shortened column lifetime. The filter sheet said “0.45 µm” and nothing else.

The measurement that identified the problem was a bubble point test on the new filters. The old supplier’s filters had a bubble point of 14.2 psi in isopropanol, consistent with a true 0.45 µm membrane. The new supplier’s filters had a bubble point of 8.5 psi, which corresponds to a membrane with a maximum pore size closer to 0.8 µm. The filter was passing particulate that the old filter was catching. The fix was to switch back to the original supplier and add a bubble point specification to the purchase order.

This is a common pattern with vials with caps and filters sourced from distributors who do not manufacture the product. The distributor’s datasheet is often a copy of the manufacturer’s nominal values with the tested values stripped out because the distributor does not have access to the batch records. When you buy through a distributor, ask for the manufacturer’s COA, not the distributor’s spec sheet.

What the test method reference tells you about the manufacturer’s QC program

The presence or absence of a test method reference on a datasheet is a signal about the manufacturer’s quality program. A company that tests its products will tell you how it tests them. A company that does not test will hide behind nominal values and vague standards references.

Look for these specific references. For vials and closures, ISO 8362 covers injection containers for injectables and specifies dimensions, materials, and performance requirements. For glass vials used in chromatography, the USP 660 chapter covers glass containers for pharmaceutical use and includes tests for surface hydrolytic resistance and arsenic release. For septa, ISO 8871 covers elastomeric closures for injectables and includes tests for fragmentation, self-sealing, and penetrability. A datasheet that cites these standards is telling you that the manufacturer has a QC lab that runs these tests. A datasheet that does not cite any standard is telling you that the manufacturer may not have a QC lab at all.

The same logic applies to column manufacturers. A C18 column datasheet should reference the USP L1 classification and give a chromatographic test with specific conditions. If the sheet says “C18” without a USP classification or a test chromatogram, the column may be packed with a silica that has not been fully characterized. The Agilent chromatography consumables pages are a good reference for what a complete column datasheet looks like. Compare any sheet you are evaluating against those.

How to compare two datasheets when the numbers look the same

Two suppliers can list identical specifications and produce completely different products. The difference is in the test conditions and the tolerance bands. Here is how I compare them.

First, look at the tolerance width. A sheet that says 12.00 ±0.05 mm for a vial diameter is a much tighter spec than one that says 12.0 ±0.3 mm. The tighter tolerance costs more to produce, and it tells you the manufacturer is controlling the drawing process closely. Second, look at the test conditions. Two sheets may both say “puncture count 50.” One specifies a 23 gauge needle at a 90° insertion angle. The other does not specify the needle at all. The first is a tested value. The second is a guess. Third, look at the batch size. A manufacturer that tests every batch and provides a COA with each shipment is doing more QC than one that tests once a year and files the results. The batch size is not always on the datasheet, but the COA will tell you.

When I compare chromatography products from different suppliers, I create a side-by-side table with the tested values only. I ignore the nominal values because they are not differentiating. Two suppliers can both say “0.45 µm” and “silanized” and “low bleed.” The tested values are where the real differences show up. A supplier with a bubble point of 14.2 psi is making a better filter than one with a bubble point of 8.5 psi, even if both sheets say 0.45 µm.

The purchase order is where tested values become enforceable

The final step in reading a chromatography products datasheet is translating what you read into a purchase order. If the sheet says puncture count ≥50 with a 23 gauge needle, put that on the PO. If the sheet says bubble point ≥13 psi in isopropanol, put that on the PO. If the sheet says tailing factor ≤1.5 for amitriptyline, put that on the PO. The PO is the only document that creates a legal obligation for the supplier to deliver what the sheet promised.

Most labs skip this step. They buy from a catalog, the catalog references a datasheet, and the datasheet has a disclaimer at the bottom that says “specifications subject to change without notice.” That disclaimer means the supplier can change the product and you have no recourse. When you put the tested values on the PO, you are creating a contract. If the lot fails, you can reject it. If the supplier will not accept a PO with tested values, that tells you something about their confidence in their own product.

For sourcing from Aijiren or any other manufacturer, the same rule applies. Ask for the COA before you commit to a volume purchase. A manufacturer that tests its products will send the COA without hesitation. A manufacturer that hesitates is telling you that the tested values on the sheet may not match the lot you are about to receive.

The decision rule

When you read a chromatography products datasheet, separate every line into one of two buckets. Tested values have a method, a tolerance, and a batch record. Nominal values have none of those. Buy based on the tested values. Put the tested values on the purchase order. Reject lots that fail the tested values. If a sheet has no tested values for the parameter you care about, ask for the COA. If the COA does not exist, find another supplier. The Thermo Fisher chromatography catalog and the EPA SW-846 methods both assume that the consumables you use meet their stated specifications. Your method validation depends on that assumption. If the sheet is lying by omission, your method is not validated. The fix is not to trust the sheet more. The fix is to demand the test data behind it.

References & further reading

Related resources from our lab consumables network:

• Chromatography Products at chromatographyselling.com
• Aijiren at aijirentech.com
• Vials With Caps at vialswithcaps.com
• Vial Supplier at vialsupplier.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