A stability run on a 0.1 N sodium hydroxide mobile phase failed its system suitability after 11 injections. The RSD on peak area hit 4.8%, the limit was 2.0%. The column was fine, the pump was fine. The culprit was a 1 L borosilicate reagent bottle that had been leaching sodium and silicate into the mobile phase for six weeks. That bottle was never part of the method validation, but it should have been. When I talk about USP requirements for reagent bottle in a validated method, I am talking about the chapters that reach the consumable, the tests that actually apply, and the line where USP stops and your own specification begins.
Most people assume USP has a dedicated chapter for reagent bottles. It does not. What it has is a set of chapters that apply to plastic packaging systems, glass containers, and the materials that contact pharmaceutical products. The one that matters for a reagent bottle is USP <661.1> Plastic Materials of Construction. If your bottle is glass, the relevant chapter is USP <660> Containers—Glass. These chapters exist to control what the container contributes to the contents, and that contribution is exactly what a validated method does not need.
USP <661.1> sets limits on extractables for plastic materials. The test uses model solvents under defined conditions: purified water at 70°C for 24 hours, 0.9% sodium chloride at 70°C for 24 hours, and isopropanol at 25°C for 24 hours. The resulting extract is analyzed for heavy metals, non-volatile residue, and UV-absorbing substances. The limits are specific. Non-volatile residue must not exceed 12 mg/L for the water extract. Heavy metals must be below 1 ppm. UV absorbance at 220 to 350 nm must not exceed 0.2 AU for the isopropanol extract. If a bottle material passes these tests, it is considered suitable for pharmaceutical contact.
But here is the part most lab managers miss. Passing USP <661.1> does not mean the bottle is inert in your specific mobile phase. The chapter uses model solvents, not your actual eluent. A bottle that passes water and isopropanol extraction can still leach something into a 0.1% trifluoroacetic acid solution or a 50 mM phosphate buffer at pH 3.0. The chapter gives you a floor, not a ceiling. Your validated method has to set the ceiling.
USP <660> applies when your reagent bottle is glass, and the type matters more than you think
For glass reagent bottles, USP <660> classifies containers into Type I, Type II, Type III, and NP (non-parenteral). The classification is based on the hydrolytic resistance of the glass surface. Type I is borosilicate, Type II is soda-lime glass with a dealkalized inner surface, Type III is soda-lime glass, and NP is soda-lime glass that only passes the powdered glass test. Each type has a specific test method and limit.
The Type I test is the surface glass test. A container is filled with purified water and autoclaved at 121°C for 60 minutes. The water is then titrated with 0.02 N sulfuric acid. Type I glass must consume no more than 1.0 mL of acid per 100 mL of extract. Type II and Type III use the same surface test but allow higher acid consumption, up to 8.5 mL for Type III. The difference is not cosmetic. A Type III soda-lime reagent bottle will leach sodium, calcium, and silicate into an aqueous mobile phase over time. If your method uses an aqueous buffer at pH above 7, that leaching changes the pH of the mobile phase and can cause baseline drift or ghost peaks.
For any validated method that runs aqueous mobile phases, I specify Type I borosilicate glass for all reagent bottles. The cost difference is real, about 2 to 3 times the price of soda-lime, but the failure cost is a rejected run and a revalidation. That math is not close.
The chapters that reach the bottle indirectly: <621>, <1058>, and <1225>
USP <621> Chromatography is the chapter that governs your HPLC and GC methods. It does not mention reagent bottles by name, but it sets the system suitability requirements that your bottle can destroy. Peak area RSD, retention time RSD, tailing factor, and resolution are all calculated from injections. If the mobile phase composition changes between injections because the bottle is interacting with the eluent, those numbers drift. The chapter does not tell you how to fix a bad RSD. It tells you the limit, and the limit is 2.0% for peak area RSD across five replicate injections in most methods. A leaching bottle can push that number past the limit without any other component failing.
USP <1058> Analytical Instrument Qualification covers the instrument, not the consumable, but it has a phrase that applies: the analytical system includes everything that touches the sample or the mobile phase. The reagent bottle is part of the system. If you change the bottle supplier or material without qualifying the change, you have changed the system. That is a finding in an audit, and it is a legitimate one.
USP <1225> Validation of Compendial Procedures is the chapter that defines what a validated method requires. It lists the parameters: accuracy, precision, specificity, detection limit, quantitation limit, linearity, range, and robustness. The reagent bottle affects precision and robustness. Precision is measured as RSD across replicates. Robustness is measured by deliberately varying method parameters and observing the effect. If you validate a method using one bottle material and then run it with a different material, you have changed a robustness parameter without testing it. That is a validation gap.
What USP leaves to your own specification: the three things you must define
USP does not specify the cap liner material for a reagent bottle. It does not specify the thread finish. It does not specify the cleaning procedure. It does not specify the maximum storage time for a prepared mobile phase. Those are all your responsibility. A validated method that references USP but does not define these three things has holes in it.
First, the cap liner. A polypropylene cap with a PTFE-faced silicone liner is standard for chromatographic reagent bottles. But PTFE is not universally inert. It resists most aqueous and organic solvents, but it can adsorb nonpolar analytes from a headspace. For trace analysis of volatile organics, a PTFE liner with a thickness of at least 0.010 inches is the minimum I specify. Thinner liners can develop pinholes after repeated tightening, and a pinhole exposes the silicone layer, which leaches siloxanes into the headspace. Those siloxanes show up as m/z 73, 147, and 207 peaks in GC-MS. If you have ever seen a series of peaks at those masses that you could not identify, check your cap liners.
Second, the bottle cleaning procedure. USP does not tell you how to clean a reagent bottle before use. A new borosilicate bottle has residual manufacturing lubricants, dust, and possibly mold release agents on the surface. My standard procedure is a 2% Alconox solution at 50°C for 30 minutes, followed by three rinses with deionized water, one rinse with methanol, and air drying. For trace metal analysis, the final rinse is 5% nitric acid, then three deionized water rinses. The cleaning procedure has to be written into the method or the method is not reproducible.
Third, the mobile phase storage time. A prepared buffer in a reagent bottle has a finite lifetime. Phosphate buffers grow mold after 3 to 5 days at room temperature. Acetate buffers degrade faster. Methanol-water mixtures can absorb carbon dioxide from the air, which shifts the pH. The method must state the maximum storage time and the storage conditions. I use 48 hours at 4°C for most aqueous buffers, and 7 days at room temperature for pure organic solvents in sealed bottles. Those numbers are not from USP. They are from running the method and measuring the drift.
How the bottle material changes extraction profiles: a comparison that matters
The choice between glass and plastic for a reagent bottle is not a preference. It is a chemical decision. The table below is the decision checklist I use when writing a method. The numbers come from extraction tests I have run and from published extractable studies on common bottle materials.
| Parameter | Type I borosilicate glass | Soda-lime glass (Type III) | HDPE | PTFE |
|---|---|---|---|---|
| Extractable metals (µg/L in water, 70°C/24h) | <10 | 50–500 | <5 | <1 |
| UV-absorbing extractables at 220 nm (AU) | <0.05 | <0.05 | 0.10–0.30 | <0.02 |
| pH shift in unbuffered water after 7 days | <0.2 | 0.5–1.5 | <0.1 | <0.1 |
| Solvent compatibility | All common HPLC solvents | All common HPLC solvents | Poor with chlorinated solvents | Excellent, but expensive |
| Reusable after cleaning | Yes, indefinitely | Yes, but surface degrades | Limited, 5–10 cycles | Yes, but surface scratches |
| Cost per 1 L bottle (USD, approximate) | 8–15 | 3–6 | 2–4 | 40–80 |
The decision rule is simple. If the method uses aqueous mobile phases, use Type I borosilicate. If it uses aggressive organic solvents like methylene chloride or tetrahydrofuran, use glass, not HDPE. If it uses hydrofluoric acid or strong alkali, use PTFE or polypropylene. The extraction profile of the bottle has to match the chemistry of the method. The reagent bottle you choose is part of the method, not a peripheral supply item.
Case 1: the phosphate buffer that grew a chromatographic shoulder
A QC lab running a USP monograph method for a generic drug noticed a shoulder appearing on the main peak after the mobile phase had been in use for three days. The shoulder was not present on day one. The retention time of the main peak shifted by 0.18 minutes, and the tailing factor went from 1.1 to 1.6. The analyst changed the column, replaced the guard column, and re-made the mobile phase. The shoulder persisted.
The measurement that identified the problem was a pH check on the mobile phase in the reagent bottle. The pH had drifted from 3.0 to 3.4 over three days. The bottle was a soda-lime glass bottle that had been in use for over a year. The inner surface had been etched by repeated exposure to the phosphate buffer, and the glass was slowly releasing sodium ions into the solution. The sodium raised the pH, which changed the ionization of the analyte, which created the shoulder. The fix was replacing all mobile phase bottles with Type I borosilicate glass and writing the bottle material into the method. The shoulder disappeared and did not return.
Case 2: the GC-MS siloxane peaks that came from the cap liner
A contract lab running a residual solvent method by GC-MS kept seeing a cluster of peaks at m/z 73, 147, and 207 in the blank injection. The peaks were small, about 0.5% of the main peak area, but they interfered with the quantitation of a low-level impurity. The analyst baked the column, changed the septum, cleaned the inlet, and replaced the liner. The peaks stayed.
The measurement that identified the problem was a headspace extraction of the reagent bottle cap. The bottle was a 100 mL borosilicate bottle with a polypropylene cap and a PTFE-faced silicone liner. The liner had been tightened and loosened over 30 times, and the PTFE face had developed a pinhole. The silicone underneath was exposed to the solvent vapor in the headspace, and it was releasing cyclic siloxanes. Those siloxanes showed up as the m/z 73, 147, and 207 peaks. The fix was replacing the cap with a new one that had a 0.020-inch thick PTFE liner and writing a replacement interval of 20 tightening cycles into the method. The peaks disappeared.
How to write the bottle specification into a validated method
The method document has to state the bottle parameters that affect the result. I use a specification block that includes the material, the size, the thread finish, the cap liner, the cleaning procedure, and the storage conditions. A typical block looks like this:
- Bottle material: Type I borosilicate glass, USP <660> compliant
- Bottle size: 1 L, with a GL45 thread (ISO 4796-2)
- Cap: Polypropylene, with a PTFE-faced silicone liner, 0.020-inch PTFE thickness
- Cleaning: 2% Alconox at 50°C for 30 minutes, three deionized water rinses, one methanol rinse, air dry
- Mobile phase storage: 48 hours at 4°C for aqueous buffers, 7 days at room temperature for organic solvents
- Cap replacement: Every 20 tightening cycles or when the liner shows visible wear
This block is not optional. It is the part of the method that connects the USP chapter requirements to the actual consumable in the lab. Without it, the method is not reproducible, and reproducibility is what validation is for.
Where USP stops and your QMS begins: the audit-facing answer
When an auditor asks how your reagent bottles comply with USP, the answer is not that they are USP-compliant. There is no such thing as a USP-compliant reagent bottle. The answer is that the method specifies the bottle material, the cleaning procedure, and the storage conditions, and that the bottle material meets the extractable limits in USP <661.1> for plastics or the hydrolytic resistance limits in USP <660> for glass. The specification is in the method, the method is validated, and the validation data shows that the bottle does not contribute interfering substances at the method’s detection limit.
That answer works because it is true and because it shows you understand the boundary. USP tells you what the bottle material must not contribute. Your method tells you what it must contribute, which is nothing detectable. The gap between those two statements is where most problems live.
For sourcing, a HPLC autosampler vial has the same material requirements as a reagent bottle, but the volume is smaller and the surface-to-volume ratio is higher, which makes extractables more concentrated. The same logic applies to a sample vial used for storage before injection. For the full range of chromatography products, the material specification has to be consistent across every component that touches the sample or mobile phase. Agilent publishes extractable data for their GC columns and accessories, and Thermo Fisher has similar data for their chromatography and sample handling products. Use those data sets when you need to compare materials. For methods that reference EPA test procedures, the SW-846 methods have their own container requirements, and they do not always match USP. Restek has published technical notes on chromatography consumables that cover liner selection and extractable profiles in more detail than most USP discussions.
Troubleshooting checklist: when the bottle is the suspect, not the column
Before you replace a column or revalidate a method, run this checklist. It is ordered by probability, and it takes about 30 minutes.
- pH drift check: Measure the pH of the mobile phase in the bottle at the start and end of a run day. A drift of more than 0.1 pH units in an aqueous buffer points to the bottle or the cap liner.
- Blank injection scan: Run a blank injection and look for peaks that were not there when the method was validated. If the peaks appear in the blank, the bottle or cap is contributing extractables.
- Liner inspection: Remove the cap and inspect the PTFE liner under a light. Look for pinholes, cracks, or discoloration. Replace the cap if the liner shows any visible defect.
- Bottle surface check: Hold the empty bottle up to a light. Look for haze, etching, or residue on the inner surface. Haze in a glass bottle means the surface has been chemically attacked.
- Extraction test: Fill the bottle with the actual mobile phase, cap it, and store it at the method temperature for 24 hours. Inject the stored mobile phase and compare it to a freshly prepared batch. Any new peak is a bottle contribution.
- Retention time check: Compare the retention time of a standard injection made from the stored mobile phase to one made from fresh mobile phase. A shift of more than 0.05 minutes suggests a pH or composition change.
If the checklist points to the bottle, replace the bottle with a Type I borosilicate bottle, replace the cap, and re-run the check. If the problem persists, the bottle is not the cause. But in my experience, the bottle is the cause more often than most analysts want to admit.
The decision rule
Specify Type I borosilicate glass with a GL45 thread and a 0.020-inch PTFE-faced liner for every reagent bottle in a validated method that uses aqueous mobile phases. Write the cleaning procedure, the storage time, and the cap replacement interval into the method document. Verify the bottle material against USP <660> for glass or USP <661.1> for plastic, and run a 24-hour extraction test with the actual mobile phase before the method goes into routine use. If the extraction test shows any peak above the method’s reporting threshold, change the bottle material. That is the whole decision rule, and it is the only one that keeps a reagent bottle from becoming the reason a validated method fails.
References & further reading
Related resources from our lab consumables network:
- Reagent Bottle at reagentbottleforsale.com
- Hplc Autosampler Vial at hplcautosamplervials.com
- Sample Vial at samplevialprice.com
- Chromatography Products at chromatographyselling.com
External references and standards: