What “traceable” has to mean for a septum, in numbers

A customer called me about a septa lot that had passed incoming inspection at 40 punctures and then started coring at injection 22 on a 6890. The certificate was clean. The bag label matched the carton. What did not match was the resin lot: the molder had switched from a 55 Shore A bromobutyl compound to a 52 Shore A compound from a second supplier three weeks earlier, and nothing on the septa bag said so. That is the failure mode this article is about. Raw material traceability for hplc septa is not a certificate problem. It is a handoff problem, and the handoff that breaks most often is the one between the resin silo and the molding log.

I use a simple test. Pick a septum out of a vial on your autosampler. Write down the lot code printed on the bag. Now ask your supplier to walk that code back to four things: the resin batch, the glass tubing lot (for the vial, not the septum, but the two travel together), the molding run with its cure and post-cure parameters, and the carton packing record. If any of those four is missing, the chain is broken and the certificate is decoration.

Here is what a complete chain looks like at each link, with the data I expect to see:

• Resin batch: manufacturer name, batch or lot code, date of manufacture, cure system, hardness (Shore A), specific gravity, and the supplier’s own CoA. For bromobutyl that is typically a 50 to 58 Shore A range; for silicone, 40 to 60 Shore A. The batch code must be a printable string, not a “see invoice.”
• Compounding record: if the molder compounds in-house, the mix sheet with curatives, filler load, and mixing temperature. If they buy pre-compounded stock, the stock lot code and the date it entered the freezer or the controlled store.
• Molding run: press ID, cavity count, cure temperature (typically 160 to 175 °C for bromobutyl), cure time, post-cure where applicable, and the operator or shift code.
• Finishing and packing: wash or extraction step, drying, visual AQL result, bag seal date, carton date, and the quantity per bag.

Every one of those has to be keyed to the same lot code the customer sees. If the molder assigns the customer lot at packing and the molding log uses a different internal code, you have two chains, not one. That is the most common break I find, and it is invisible until something goes wrong.

The four handoffs where the chain actually breaks

In fifteen years of auditing this, the breaks cluster in the same four places. They are not exotic.

Handoff 1: resin receipt to molding

The receiving clerk logs the resin batch on the goods-inward sheet. The molder pulls from the same drum three weeks later and writes the internal mix code on the press card. Nobody joins the two. When I ask for the resin batch behind a finished lot, the answer is “we used what was in the drum.” That is not traceability, it is a guess with a date stamp.

The fix is a one-page resin issue log: date, drum or box ID, resin batch, quantity issued, press ID it went to. It takes the material handler ninety seconds per issue. I have never seen a molder refuse once they understand it is the difference between a recall and a quarantine.

Handoff 2: the glass tubing lot behind the vial

Septa travel with vials, and a vial problem looks like a septum problem more often than people admit. If a batch of vials has a 0.1 mm bore variance in the neck ID, the septum sits at a different compression and the coring count moves. You cannot chase that without the glass tubing lot. Tubing lots carry a draw date and a dimensional certificate; the vial maker should be able to name the tubing lot for any vial lot. Ask for it once and see what happens. If the answer takes more than a day, the vial supplier is buying on price and not tracking draw lots.

Handoff 3: molding lot to finishing

Septa get washed, dried, and inspected after molding. If the wash batch mixes two molding lots, the customer lot is now a blend and the molding parameters for half the bag are unknown. I have seen this on silicone septa where the molder ran two presses into one wash basket to keep the line moving. The fix is a rule: one wash batch, one molding lot. No exceptions, even if it means a half-empty basket.

Handoff 4: packing to carton

This is where the customer-facing lot code is born, and it is where the chain most often gets severed from everything above. The packer writes the carton label from a work order, not from the molding log. If the work order was raised against a forecast rather than a specific molding run, the lot code on your bag traces to a plan, not a product.

A comparison: what a closed chain versus an open chain looks like on paper

The table below is the checklist I use in a supplier audit. It is not a certificate list; it is a data-join list. The right column is what I expect to see when I ask for the record set behind a single bag.

If a supplier can close all six rows for a single bag in under an hour, the chain is real. If it takes a week and three emails, it is reconstructed, and a reconstructed chain is a story, not a record.

Field case 1: the coring count that moved 18 punctures

Symptom. A pharma QC lab running a 9 mm bromobutyl septa on a 7693 autosampler saw coring start at injection 22 instead of the usual 40. The septa were from a lot that had passed the incoming puncture test at 40. The customer was ready to blame the needle.

What the measurement showed. We pulled the molding log and found the lot had been molded at 168 °C for 6 minutes instead of the specified 172 °C for 7 minutes. The lower cure gave a slightly softer part; hardness measured 51 Shore A against a 55 Shore A spec. The softer compound deformed more under the needle and cored earlier. The incoming test had passed because it was run on a hand-tightened vial, not on the autosampler tray where the gripper compresses the cap to a different torque.

The fix. The molder re-qualified the cure window and tightened the press card to a 171 to 174 °C band with a 7-minute minimum. The customer added a hardness check to incoming inspection at 55 ± 3 Shore A. Coring moved back to 38 to 42 punctures. The lesson is not about cure temperature; it is that the customer lot could not be traced to the cure parameters until we forced the join.

Field case 2: the silicone lot that bled siloxane at 260 °C

Symptom. A contract lab running GC-MS on EPA method 8270 saw siloxane peaks at m/z 73, 147, and 207 in blanks. The septa were silicone, rated to 250 °C, and the inlet was at 280 °C. The lab assumed the septa were out of spec.

What the measurement showed. The septa were within their temperature rating. The problem was the post-cure. The molding log showed a 4-hour post-cure at 200 °C; the resin supplier’s data sheet called for a 6-hour post-cure at 200 °C for this compound to drive off low-molecular-weight siloxanes. The molder had shortened it to hit a delivery date. The lot code on the bag traced to the molding run, but the post-cure record was not part of the traceability set, so nobody saw it until the blanks failed.

The fix. The molder added post-cure time and temperature to the molding log and to the lot record. The lab moved to a lower-bleed septa for the 280 °C inlet, which is the right call anyway at that temperature. The traceability fix was adding one field to a form. The chromatography fix was a different part. Both were needed.

This is why I keep saying the traceability chain for a consumable has more links than the certificate. The certificate covers the resin. It does not cover what the molder did to it.

How to audit a supplier’s chain in one afternoon

You do not need a three-day audit to find out whether a supplier can trace a septum. You need one bag and four questions, asked in this order:

1. Hand them a bag and ask for the resin batch code behind it. Time the response.
2. Ask for the molding run record for that lot, including cure temperature, cure time, and post-cure.
3. Ask which wash batch the lot went through, and whether that wash batch contained any other molding lot.
4. Ask for the packing record and confirm the carton label was written from the molding log, not a work order.

If all four come back inside an hour with matching codes, the chain is closed. If any one comes back as “we’ll get back to you,” you have found the break. I have run this on maybe forty molders over the years. Roughly a third close all four. Another third close three and fail on the wash batch. The rest fail on the first question.

For the vial side of the same audit, the questions are the same shape: tubing lot, draw date, dimensional certificate, and the neck ID range. The HPLC vial and its vial closures have to travel with the septa lot if you want to chase a seal problem to its root. And if you are sending samples out for confirmation, the receiving lab’s own chain matters too; analytical testing labs that run EPA methods will ask you for the septa lot when a blank fails, and you want to answer in minutes, not days.

The resin batch is the link most suppliers cannot name

I want to dwell on this because it is the one that surprises procurement people. Most septa molders buy pre-compounded rubber, not raw resin. The compounder’s lot code is what the molder should record, and the compounder should be able to name the underlying resin batch. That is a two-step chain, and it is fine as long as both steps are recorded.

What I find instead is a molder who records the compounder’s lot on the receiving sheet and then loses it at the press. The press card says “bromobutyl” and the date. That is the break. When a customer asks for the resin batch, the molder calls the compounder, who may or may not still have the batch on file after six months. If the answer comes back “we can’t confirm,” you have a lot of septa with no upstream. Quarantine it or use it on non-critical methods and write the decision down.

For silicone, the same logic applies but the compounder is often the resin maker, so the chain is shorter. For PTFE-faced septa, you now have two materials and two chains; the PTFE film has its own lot and its own thickness tolerance, typically 0.05 to 0.13 mm, and a film lot change can move the puncture count even when the rubber is identical. Ask for both.

What to put in your own receiving record

You cannot fix a supplier’s chain, but you can make your side of it airtight. Here is the receiving record I recommend for any lab that runs critical methods:

• Septa lot code as printed on the bag, transcribed exactly, not paraphrased.
• Carton lot code, if different. If they differ, flag it and ask why.
• Date received and date the bag was opened.
• Resin or compound batch code, if the supplier provides it. If not, write “not provided” and note the supplier.
• Hardness check result at 55 ± 3 Shore A (bromobutyl) or the spec for your material.
• Puncture count on the first vial from the bag, run on the actual autosampler, not a hand-tightened vial.
• Blank run result at your inlet temperature, at the start of the lot and again at the end.

That record takes ten minutes per lot and it is the difference between a two-hour investigation and a two-week one. It also gives you something to hand an inspector that is not a certificate. On the certificate question, the sibling article on this site already covers the record set an inspector asks for, so I will not repeat it here.

The decision rule for a supplier who cannot close the chain

Not every lab needs a closed chain. A method development lab running a few hundred injections a week on a 250 °C inlet can live with an open chain and a good incoming test. A GMP release lab running USP methods on a 280 °C inlet cannot. The decision rule I use is this: if a failed blank or a coring event would cost you more than a day of investigation, demand the closed chain. If it would cost you an hour, an open chain with a hardness check and a puncture count is enough.

For the labs that need the closed chain, the ask is specific and it is not a certificate. It is the resin batch code, the molding run record with cure and post-cure, the wash batch identity, and the packing record, all keyed to the same lot code on the bag. Ask for it before you place the first order, not after the first failure. A supplier who can produce it in an hour has a system. A supplier who cannot has a story, and you will hear it at the worst possible moment.

If you are sourcing septa and want a supplier who can walk a lot code back to the press card, start with the HPLC septa range and ask for the four records above on a sample lot. For the chromatography methods that sit behind the septa decision, the vendor technical libraries at Agilent, Thermo Fisher, and Restek are worth a read, and the method source for waste and materials work is EPA SW-846. None of them will tell you the resin batch behind your septa. That is your supplier’s job, and it is the one question worth asking before you sign the PO.

References & further reading

Related resources from our lab consumables network:

• Hplc Septa at hplcsepta.com
• Vial Closures at vialsclosures.com
• Analytical Testing at testinganalytical.com
• Hplc Vial at vialforhplc.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