Why the break-even point is a number, not an opinion

Twelve years ago I ran a cost review on our consumables spend and found we were buying 40 mL amber vials by the pallet while a 500 mL borosilicate bottle sat on the same bench getting washed three times a week. The vials cost us $1.14 each landed. The bottle cost $9.80 and had been in service for six years. Neither number told the truth about which was cheaper. What decided it was a column I had never bothered to build: the break-even injection count. Once I built it, half of our “obviously disposable” items turned out to be reusable, and three of our “obviously reusable” bottles turned out to be losing us money every time they went through the wash.

This is the guide I wish I had then. It is not about which bottle is better in the abstract. It is about the specific point, in volume and in cycles, where the cost model flips from single-use to reusable, and the operational constraints that move that point by a factor of three in either direction.

Every procurement team I have worked with has a default position on single-use versus reusable glassware. It is usually inherited from a previous lab manager and it is usually wrong for at least a third of the items on the shelf. The reason is simple: the break-even calculation has six inputs, and most people only look at two.

The six inputs are:

• Unit cost of the single-use bottle (landed, including freight and disposal)
• Unit cost of the reusable bottle (purchase price amortized over service life)
• Cost per wash cycle (labor, water, detergent, energy, validation)
• Failure rate per cycle (breakage, contamination, out-of-spec volume)
• Cost of a failure (lost sample, rerun, investigation)
• Cycle count achievable before retirement (not the same as cycle count before it breaks)

Get all six right and the answer is usually within 10%. Get three of them right, which is typical, and you can be off by 300%. I have seen labs argue for two years about whether to switch from reusable to single-use when the answer was a one-page spreadsheet.

The baseline numbers I use for a first pass

Here is the working model I start with before I have site-specific data. It is deliberately conservative.

Run the arithmetic at 40 cycles and the reusable bottle costs about $6.80/40 = $0.17 per use in amortized purchase, plus $0.95 per wash, plus failure costs of around $0.10 per cycle. That is $1.22 per use. The single-use bottle is $0.48. At 40 cycles, the disposable wins by a factor of 2.5.

Push to 60 cycles and the reusable drops to $1.13. Still losing. So why does anyone reuse glass at all?

Because the $0.95 wash cost is not fixed. It is the number most sensitive to scale, and it is the number most labs get wrong.

Where the wash cost actually lives, and why it dominates

The $0.95 per-cycle figure I used above assumes a lab washing 40 to 60 bottles per week in a shared glasswasher. Push that to 400 bottles per week and the per-bottle cost falls to roughly $0.28. Push it to 1,200 per week and you are at $0.14. The labor component amortizes; the water, detergent, and energy components scale linearly but are small.

I have watched a 4-bottle-a-week lab try to justify reusable glassware on environmental grounds and fail the cost model by 400%. I have watched a 300-bottle-a-week lab switch to single-use and lose $18,000 a year. The difference is throughput, not philosophy.

The three throughput thresholds

From my own records across four facilities, the flip point sits at these approximate volumes for a 125 mL bottle handled at ambient temperature with aqueous or low-hazard chemistry:

• Under 20 uses per week: single-use wins on cost by a wide margin. The wash cycle never amortizes.
• 20 to 80 uses per week: it depends on your wash setup. A validated in-house glasswasher that runs at least three loads a week can make reusable competitive. A hand-wash protocol almost never does.
• Over 80 uses per week: reusable wins unless the chemistry is incompatible with cleaning (see below). The per-cycle cost drops below $0.40 and stays there.

These are not universal. Change the chemistry, the temperature, or the required cleanliness, and the thresholds move. But they are the right starting point, and they are the numbers I use to open the conversation.

What actually disqualifies a reusable bottle, and why it is not breakage

Breakage is the failure mode everyone worries about. It is almost never the reason a reusable bottle gets retired. In my records, borosilicate bottles break at 0.5% to 1.2% per wash cycle for a lab with a decent rack system and trained staff. That is a rounding error in the cost model.

The real disqualifiers are three:

1. Cleaning validation for trace analytes

If your bottle holds a standard at 1 ng/mL and your wash protocol leaves 0.05 ng/mL of the same compound on the glass, you cannot reuse it for that standard. Period. The wash is not validated, and validating it costs more than a decade of single-use bottles.

This is the single biggest reason reusable glassware looks cheaper on paper and fails in practice. The cost of a validated wash protocol for trace organic analytes at sub-ppb levels runs into five figures and requires re-validation any time the detergent, water source, or drying step changes. For a lab running EPA SW-846 methods at low detection limits, single-use is often the only defensible choice for anything that touches a calibration standard.

2. Chemical compatibility with the cleaning agent

Caustic wash at pH 13 will etch borosilicate over time. The etch is slow (roughly 0.02 mm per 100 cycles on the interior surface) but it changes the surface chemistry. Etched glass adsorbs more, and adsorption is the enemy of low-concentration work. I have seen a bottle that passed visual inspection fail a recovery check at 85% because the interior had been etched by three years of hot caustic wash.

If your chemistry requires caustic cleaning, budget for 25 to 30 cycles, not 60. That shifts the break-even point up by a factor of two.

3. Closure and thread wear

The GL45 screw thread is the part that fails first. A borosilicate bottle with a GL45 neck and a PP cap will survive 200 to 400 cap-on/cap-off cycles before the thread engagement degrades enough to fail a leak test. If you are autoclaving, the cap liner (usually PTFE-faced silicone) compresses and loses seal integrity after 15 to 25 autoclave cycles at 121 °C. After that, the cap leaks under vacuum and the bottle is no longer fit for sterile or anaerobic work.

This is why I always price the cap separately. A bottle is not one consumable; it is a bottle plus a closure plus a liner, and the closure has a shorter service life than the bottle by a factor of five to ten.

The single-use side has hidden costs too

Single-use bottles are not free of operational cost. They are free of wash cost, which is different.

Inventory and shelf life

A 125 mL HDPE bottle has a shelf life of roughly 3 to 5 years from manufacture for general use, and 2 years for sterile or certified-clean grades. That is a real constraint if you buy by the pallet. I have seen labs write off $4,000 of expired single-use bottles because they over-ordered to hit a freight break. The 12% freight savings cost them 30% in spoilage.

Leachables and extractables

HDPE and PET both leach at the low-ppb level. For most reagent storage this is irrelevant. For trace organic analysis it is not. If you are storing a standard in a plastic bottle for more than a few days, you need to know what the bottle is contributing. The HPLC autosampler vial world has spent 20 years characterizing this; the reagent bottle world has not, and most suppliers will not give you an extractables profile. Ask anyway. The ones who can answer are the ones worth buying from.

Regulated waste stream

Single-use bottles that have held hazardous reagents become regulated waste. The disposal cost per bottle is small ($0.05 to $0.10 typically) but it is not zero, and it scales linearly with volume. For a lab using 50,000 bottles a year, that is $2,500 to $5,000 in disposal alone. Add it to the unit cost before you compare.

A decision checklist you can run in an afternoon

Here is the checklist I actually use. It takes about two hours to fill in for a single bottle type and it will tell you which way the model flips.

Step 1: Characterize the use case

• What is the highest-concentration analyte the bottle will hold? (If it is a standard above 1 µg/mL, reusable is usually fine. Below 100 ng/mL, single-use is usually required.)
• What is the cleaning compatibility? (Aqueous, solvent, caustic, acid, or none.)
• What is the temperature range? (Ambient, refrigerated at 4 °C, frozen at -20 °C, autoclaved at 121 °C, or dry heat at 180 °C.)
• What is the closure requirement? (GL45, GL32, GPI 24-400, crimp, or screw-thread with a specific liner.)

Step 2: Count the real cycle volume

• How many bottles do you fill per week? (Actual count, not budgeted count.)
• How many of those are the same chemistry and could share a wash protocol?
• How many wash loads per week can you realistically run? (Most labs overestimate by 50%.)

Step 3: Price the wash honestly

• Labor: how many minutes per bottle to load, unload, inspect, and log? (Typical: 1.5 to 3 minutes per bottle fully loaded.)
• Utilities and detergent: $0.10 to $0.25 per bottle for a well-run glasswasher.
• Validation: amortized over the number of bottles washed per validation cycle. (This is the number that kills the case for low-volume labs.)

Step 4: Price the failure

• What does a contaminated batch cost? (Rerun + investigation + potential OOS report.)
• What is the historical contamination rate for reusable bottles in your lab? (If you do not know, use 0.15% as a starting assumption.)

Step 5: Run the arithmetic at 30, 45, and 60 cycles

If the reusable option is not cheaper at 60 cycles, it will not be cheaper. If it is cheaper at 30, it is cheaper by a wide margin. The middle band is where you need site-specific data.

Troubleshooting: when the model says one thing and the lab does another

The most common failure I see is a lab that has run the numbers, decided reusable, and then watched costs go up anyway. Here is the checklist I run when that happens.

• Are the bottles actually being washed at the rate you assumed? A bottle that sits in a tote for three weeks waiting for a full load is not being amortized. Count the actual wash cycles per bottle per year. I have seen labs assume 40 and achieve 11.
• Are you counting the right bottles? The 500 mL bottle used once a month is not the same as the 125 mL bottle used daily. Segment by size and frequency before you compare.
• Is the wash validated? If not, you are carrying a contamination risk you have not priced. Add 0.5% to the failure rate and rerun the model.
• Are you pricing the cap? A bottle with a worn cap is a bottle that fails leak testing. Price the cap separately at $0.80 to $1.40 each and assume 15 to 25 autoclave cycles per cap.
• Are you counting the storage cost? Reusable bottles need rack space, drying space, and a wash area. That is floor space, and floor space has a cost. At $40 per square foot per year in a typical lab, a 20-square-foot wash area costs $800 a year before you wash anything.

Where the model flips in practice: three real cases

Case 1: A contract analytical lab running 400 samples a week, all aqueous, all at ppm concentrations, all using 125 mL amber bottles. They were buying 600 single-use bottles a week at $0.48 landed. That is $14,976 a year. They switched to 200 reusable borosilicate bottles at $8.20 each, ran a validated wash at $0.31 per bottle, and replaced 15% of the fleet per year. Annual cost: $2,460 in amortized purchase plus $6,448 in wash plus $1,640 in replacement. Total $10,548. They saved $4,428 a year and the model flips at week 12.

Case 2: A university research lab running 15 samples a week, one standard at 500 ng/mL, one solvent at 4 °C. They were buying 60 single-use bottles a week at $0.42. That is $1,310 a year. They tried reusable and found the wash was not validated for their analyte, so they had to run a solvent rinse by hand at 4 minutes per bottle. That is 4 hours a week of technician time at $38 an hour fully loaded, or $7,904 a year. The model never flips. Single-use is correct here, and the environmental argument does not survive contact with the labor cost.

Case 3: A QC lab running 1,200 bottles a week, all the same chemistry, all aqueous, all at ambient temperature. They built a dedicated wash line with a validated protocol and a per-load cost of $0.14. Reusable wins by a factor of three and the model flipped at week 6. This is the case where reusable glass is genuinely the right answer, and it is the case most people cite when they argue for reuse. It is also the case that applies to maybe 15% of labs.

What I actually recommend, by category

Here is my working recommendation, and I will defend it. It is based on the six-input model above and on 15 years of watching labs get this wrong in both directions.

Use single-use when any of these are true

• You use fewer than 20 bottles per week of that size and chemistry.
• You store a standard below 100 ng/mL and you do not have a validated wash for that analyte.
• Your chemistry requires caustic cleaning (which etches borosilicate and shortens service life to 25 to 30 cycles).
• You need a certified-clean or sterile bottle and you cannot validate the wash to that standard.
• You need traceability to a lot number for a regulated submission and your reusable bottles are not lot-tracked.

Use reusable when all of these are true

• You use more than 80 bottles per week of that size and chemistry.
• You have a validated wash protocol for the analytes involved, or the analytes are at ppm concentrations where carryover is not a concern.
• You have rack and drying space, and a technician whose time is genuinely available (not borrowed from another task).
• You track bottle service life and retire at the validated cycle count, not at the point of breakage.
• You price the cap and liner separately and replace them on the autoclave cycle count.

Use a hybrid when the volume is in the middle

For 20 to 80 uses per week, the right answer is usually a hybrid: single-use for standards and trace work, reusable for buffers, diluents, and wash solutions at ppm concentrations. This is the case that most labs land in and most procurement reviews miss because they treat “reagent bottle” as one line item.

If you are running HPLC or LC-MS work, the same logic applies at the vial level, and the numbers are different because the vial volume is smaller and the per-use cost is much lower. The sample vial cost model flips at a much higher injection count than the reagent bottle model, which is why most labs correctly use single-use vials and incorrectly use single-use bottles. The two are not the same decision.

Specifications that matter when you do go reusable

If you decide reusable is right for a given use case, here are the specifications I would not compromise on. These are the ones that determine whether the bottle actually achieves the 40 to 60 cycles the cost model depends on.

• Material: Borosilicate 3.3 (ISO 3585). Not soda-lime. Soda-lime leaches sodium and silicate into aqueous contents at the ppm level and etches faster under caustic wash.
• Neck finish: GL45 for 100 mL and above, GL32 for 50 to 100 mL, GL25 for smaller. Confirm the thread is molded, not cut. Molded threads hold torque better and leak less.
• Cap: PP with a PTFE-faced silicone liner for general use; PTFE-lined for solvent; butyl for some aqueous applications. Replace the liner every 15 to 25 autoclave cycles.
• Torque: 1.2 to 1.5 N·m for a GL45 cap with a fresh liner. Over-torque compresses the liner and shortens its life; under-torque fails the leak test. A torque wrench is $60 and it will save you more than that in failed leak tests.
• Graduation accuracy: Class A if you are using the bottle to measure; Class B if you are only storing. The cost difference is real and the accuracy difference is 2x.
• Autoclave rating: Confirm the bottle is rated for 121 °C at 15 psi for 20 minutes, and confirm the cap is rated for the same. Many bottles are rated and their caps are not.

For the analytical work that sits downstream of the bottle, the reference material from Agilent, Thermo Fisher, and Restek is worth bookmarking. None of them will tell you how to run the cost model, but they will tell you what the bottle has to survive on the analytical end, and that is half the input.

The one number I would put on a sticky note

If you take one number from this article, take 80 uses per week. Below that, single-use is almost always the right answer for a 125 mL bottle. Above it, reusable almost always wins, provided the wash is validated and the cap is priced separately. The 20 to 80 band is where you need to do the work, and it is where 60% of labs actually sit.

The second number is 25 autoclave cycles. That is the realistic service life of a PTFE-faced silicone liner at 121 °C, and it is the number that turns a “reusable” bottle into a single-use bottle with extra steps if you are not tracking it. A bottle with a worn liner is a bottle that will fail a leak test, and a failed leak test in a regulated lab is a $40 rerun plus an investigation.

The reagent bottle decision is not a philosophy question. It is an arithmetic question with six inputs and a break-even point that moves by a factor of three depending on throughput. Run the arithmetic. Segment by chemistry and size. Price the cap. Then decide.

And if you are buying chromatography consumables alongside the bottles, the same segmentation logic applies to the whole order. The chromatography products catalog has the same split between single-use and reusable, and the break-even points are different for every line item. Treat them separately. The lab that treats “consumables” as one budget line is the lab that gets the cost model wrong.

Takeaway: the single-use versus reusable reagent bottle decision flips on throughput, and the number is roughly 80 uses per week for a 125 mL bottle at ambient temperature with aqueous chemistry. Below 20 uses per week, single-use wins by a factor of two or more. Between 20 and 80, you need a validated wash and a real cycle count before you can decide. Above 80, reusable wins unless the chemistry requires caustic cleaning or the analyte concentration is below 100 ng/mL. Price the cap separately, retire at the validated cycle count rather than at breakage, and run the six-input model before you commit. The bottle itself is the smallest number in the equation.

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:

• 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