A batch of 2 mL screw-top vials came off the line at 4,100 units per hour, and every tenth vial measured 12.25 mm OD instead of 12.00 mm. The shift supervisor caught it at station four, not at final inspection. That is the difference between a factory that controls the line and one that just boxes product. When you ask how a chromatography supplier is manufactured and what the factory controls, the answer is not a brochure. It is a station map with specific instruments, setpoints, and rejection rules. Here is what actually happens on the floor.

Most people assume the cap or the septum is the hard part. It is not. The glass or polymer body is where 80% of dimensional drift originates, and it is the first station where closed-loop control matters. A chromatography supplier that runs a modern line starts with the body former, whether that is a glass tube converter for 9 mm screw vials or an injection molder for 12 mm crimp vials. The control points are not optional.

Glass conversion controls the ID first, OD second

For Type I borosilicate glass converted from tubing, the inner diameter at the neck is the first controlled dimension. We hold 9 mm vial neck IDs to 7.60 mm ±0.05 mm. If the ID drifts high, the seal compresses unevenly against the septum. If it drifts low, autosampler needles bind. The OD follows from the tubing wall thickness, which is checked at the incoming glass stage with a laser micrometer. We reject tubing lots above 1.05 mm wall thickness for 9 mm vials because the OD tolerance stack becomes unrecoverable downstream.

Injection molding uses cavity pressure, not shot weight

For polymer vials and caps, the factory controls cavity pressure transducers inside each mold cavity. Shot weight is a lagging indicator. Cavity pressure is real-time. A polypropylene 12 mm crimp cap mold runs at 350 to 420 bar peak cavity pressure. If the pressure drops 15 bar across a cavity, that cavity is flagged before the part ejects. We have seen molds where one cavity ran 18 bar low and produced caps that looked fine but cracked under a 45 N crimp force. Closed-loop cavity pressure would have caught it at cycle 200, not at the customer’s autosampler.

Septum punching is where puncture count is decided, not tested

The septum station is the least understood part of the line. A OEM vial buyer will ask about PTFE/silicone thickness and shore hardness. A factory engineer asks about punch tool clearance and die wear. The puncture count you see on a datasheet is a downstream consequence of what happens at the punch press.

Die clearance sets the edge quality

For a 9 mm PTFE/silicone septum, the punch die clearance is held at 4% to 6% of material thickness. For a 1.3 mm thick septum, that is 0.052 mm to 0.078 mm. Tighter than 4% and the die galls the PTFE layer. Looser than 6% and you get a ragged edge that sheds particles into the sample. We measure die clearance every 5,000 punches with a feeler gauge, not every shift. That is the cadence that keeps puncture counts above 50 for a 55 Shore A silicone core without overbuilding the septum.

Puncture count is a function of silicone crosslink density

The silicone layer is compression molded and post-cured. Post-cure temperature and time are the two parameters that move puncture resistance. We run 200°C for 4 hours for standard septa. If the oven drifts to 190°C, the crosslink density drops and puncture count falls from 50 to 35. If it runs 210°C, the silicone hardens and the needle force goes up 20%. This is checked at the oven with a thermocouple data logger, not by testing finished septa. That is the difference between controlling the line and inspecting the result.

Cap torquing is a station, not a shipping instruction

Every cap gets torqued on the line. The torque value is not a suggestion printed on the box. It is a setpoint on a servo-driven capping head with a torque transducer. For 9 mm screw caps, we run 0.11 N·m to 0.14 N·m. For 12 mm crimp caps, the crimp force is 45 N ±5 N. The transducer logs every cap. If the torque drifts outside the band for three consecutive caps, the line stops.

Why three caps, not one

One cap outside the band is usually a bad cap or a piece of debris. Three in a row means the capping head has drifted or a jaw is worn. We replace capping jaws every 200,000 cycles. That is a number a vial closures buyer should ask about. If a supplier cannot tell you their jaw replacement interval, they are not controlling torque. They are hoping the caps happen to be tight enough.

Torque verification is destructive, so it is sampled

You cannot torque-test a cap and then ship it. The removal torque test destroys the seal. So the factory pulls one capped vial every 500 units and measures removal torque on a calibrated torque tester. The removal torque for a 9 mm screw cap should be 0.08 N·m to 0.12 N·m. If the removal torque drops below 0.06 N·m, the seal is not reliable for a 60°C headspace run. If it exceeds 0.16 N·m, the analyst will need a tool to open it, and that is a complaint waiting to happen.

Washing and silanization are chemical controls, not rinses

The wash line is where many suppliers cut corners because the vials look clean. Clean is not the same as low-activity. For trace analysis, the factory runs a three-stage wash: alkaline detergent at 60°C, deionized water rinse, and a final rinse with 18.2 MΩ·cm water. The conductivity of the final rinse is monitored continuously. If it rises above 1.0 µS/cm, the line stops and the DI system is checked.

Silanization is a timed reaction, not a dip

For deactivated vials, the silanization step is a vapor-phase reaction with dimethyldichlorosilane at 70°C for 90 minutes. The time and temperature are the control points. A dip in a silane solution at room temperature gives uneven coverage. We have measured RSD values for basic compounds at 8% to 12% on dipped vials versus 3% to 5% on vapor-phase treated vials. The difference is not the chemistry. It is the control of the reaction conditions.

Dimensional verification happens at three gates, not one

The factory does not wait until the end of the line to measure vials. There are three dimensional gates. Gate one is after body forming. Gate two is after capping. Gate three is a final audit before packing. Each gate has a different measurement scope and a different rejection rule.

Gate one: body dimensions

At gate one, a vision system measures OD, ID, height, and neck finish on every vial. The system runs at 200 parts per minute and rejects any vial outside the drawing tolerance. For a 2 mL 9 mm screw vial, the tolerance band is 12.00 mm ±0.10 mm OD and 32.00 mm ±0.25 mm height. The vision system also checks for cracks, chips, and bubbles larger than 0.5 mm. This is not a sampling plan. It is 100% inspection at line speed.

Gate two: cap and septum seating

Gate two checks the capped assembly. The vision system measures cap height above the vial rim and septum compression. For a 9 mm screw cap, the cap should sit 0.3 mm to 0.5 mm above the vial rim when properly torqued. If the cap sits higher, the septum is not compressed enough. If it sits flush, the cap is over-torqued and the septum is crushed. The system also checks for cross-threading by measuring the cap angle relative to the vial axis. Anything over 2 degrees is rejected.

Gate three: final audit

Gate three is a manual audit using calibrated gauges. An operator pulls 20 vials per lot and measures OD, ID, height, and cap torque with hand instruments. The results are recorded on a control chart. If any dimension shows a trend, the line is stopped and the upstream station is adjusted. This is not the primary control. It is a check that the automated gates are still calibrated. The automated gates catch drift in real time. The audit catches calibration drift in the gates themselves.

Cleanroom class and particle counts are line controls, not marketing

The cleanroom classification is a physical property of the line, not a sticker on the box. A analytical testing lab that buys vials for LC-MS at 1 ppb detection limits needs ISO Class 7 or better. The factory controls this with particle counters mounted at the filling and capping stations. The counters sample continuously and alarm if the 0.5 µm particle count exceeds 352,000 per cubic meter for ISO Class 7.

Particle shedding is a material property, not a wash issue

Some vials shed particles because the glass was not properly annealed or the polymer was not fully cured. The factory checks this with a particle shed test on finished vials. A vial is filled with filtered water, capped, and agitated for 60 seconds. The water is then passed through a particle counter. A good 2 mL vial sheds fewer than 10 particles larger than 10 µm. A bad lot sheds 50 to 100. The fix is upstream: slower annealing ramp or longer polymer cure. You cannot wash away a material defect.

Closed-loop versus end-of-line checks: the actual split

Here is the honest breakdown of what is controlled in real time versus what is only checked at the end. A chromatography supplier that tells you everything is 100% inspected is either lying or running a very slow line. The reality is a mix.

Parameter Control type Setpoint or tolerance Check frequency
Vial OD (9 mm) Closed-loop vision 12.00 mm ±0.10 mm 100% at gate one
Vial ID (neck) Closed-loop vision 7.60 mm ±0.05 mm 100% at gate one
Cap torque (9 mm) Closed-loop servo 0.11 to 0.14 N·m 100% at capping station
Septum puncture count End-of-line sample ≥50 punctures 5 septa per lot
Septum shore hardness End-of-line sample 55 ±5 Shore A 5 septa per lot
Final rinse conductivity Closed-loop sensor ≤1.0 µS/cm Continuous
Particle shed (10 µm) End-of-line sample ≤10 particles per vial 10 vials per lot
Silanization coverage End-of-line sample RSD ≤5% for basic compounds 3 vials per lot
Cleanroom particle count Closed-loop counter ISO Class 7 limit Continuous at fill/cap

The pattern is clear. Geometric dimensions and mechanical forces are closed-loop. Chemical properties and material performance are sampled at the end because they are destructive or slow to measure. A buyer who understands this split asks the right questions. A buyer who does not gets a datasheet and a promise.

Field case one: the 60°C headspace leak that was not a cap problem

A contract lab running U.S. EPA SW-846 method 8260 for volatiles started seeing low internal standard response on 20% of their samples. They blamed the caps. We pulled the retained samples and measured cap removal torque. All were within 0.08 to 0.12 N·m. The caps were fine. The problem was the vial neck ID. A batch of vials had drifted to 7.72 mm ID, which is outside the 7.60 mm ±0.05 mm spec. At 60°C, the glass expanded and the septum lost compression. The fix was not a tighter cap. It was a supplier that controlled neck ID at gate one. The lab switched to a line with closed-loop vision on the neck, and the problem disappeared.

Field case two: the autosampler needle that bent at 4,000 injections

Another lab reported bent autosampler needles on a 100-position tray. The needles were bending when they hit the bottom of the vial. We measured the vial height and found a batch at 32.4 mm instead of 32.0 mm. That 0.4 mm was enough to change the needle penetration depth. The lab had been using the same instrument method for years. The vial height drift was new. The supplier had changed glass tubing sources and the new tubing had a slightly different cut tolerance. The fix was a height check at gate one with a 0.25 mm tolerance and a supplier audit of the tubing source. The needles stopped bending.

What the factory cannot control, and what that means for you

There are limits. The factory cannot control how you store vials after delivery. It cannot control the solvent you use or the autosampler needle gauge. It cannot control the analyst who overtightens caps by hand. What it can control is the dimensional and mechanical consistency of every unit that leaves the line. That is the product. The rest is your process.

When you evaluate a chromatography and sample handling supplier, ask for the station map. Ask which parameters are closed-loop and which are end-of-line sampled. Ask for the setpoints and the rejection rules. A supplier that can answer those questions has a controlled line. A supplier that cannot is selling boxes.

Decision rule for procurement

Buy from a supplier that controls vial OD, neck ID, and cap torque in real time, and that samples septum puncture count and silanization coverage at the end of the line. Reject any supplier that claims 100% inspection of chemical properties or that cannot state their capping jaw replacement interval. The line controls the product. The datasheet just describes it. For a deeper look at what the instrument side expects from consumables, check the column and accessory specifications from the major instrument manufacturers. For method-specific vial requirements, the technical resources at Restek are a useful cross-reference. But the factory floor is where the decision gets made. That is where the 12.25 mm vial gets caught at station four, or it does not.

References & further reading

Related resources from our lab consumables network:

External references and standards: