Reference library

How to Test for Purity: HPLC, MS, Karl Fischer

Purity testing takes three measurements: HPLC area percent, mass spec identity, and Karl Fischer water content, plus why an HPLC-only CoA is not enough.

Testing a reference material for purity means running three independent measurements and refusing to treat any one of them as the answer. Reversed-phase HPLC with UV detection at 214 nm gives you chromatographic purity as an area percentage — typically reported as ≥95%, ≥98%, or ≥99%. Mass spectrometry (ESI or MALDI-TOF) confirms the molecule is the one named on the label, matching the theoretical monoisotopic mass within roughly ±0.5 Da. Karl Fischer titration and counterion analysis tell you what fraction of the vial's mass is actually the compound rather than water and trifluoroacetate salt. A certificate of analysis showing only the first of those three has told you about one-third of the story, and it is the third that is easiest to inflate.

What "99% pure" actually measures

Chromatographic purity is an area-normalized number. The instrument integrates every peak the detector sees, divides the main peak area by the total, and reports a percentage. That definition carries two limits worth stating plainly.

The first is the detector. At 214 nm you are watching the peptide bond itself, so nearly every peptide-related impurity — deletion sequences, truncations, oxidized methionine — shows up. At 280 nm you are watching tryptophan, tyrosine, and phenylalanine. A COA reporting 99.2% at 280 nm on a sequence with a single tyrosine has effectively hidden every impurity lacking an aromatic residue. Check the wavelength before you check the number.

The second is the gradient. A 5–95% acetonitrile ramp run over 8 minutes will co-elute a des-Gly deletion product with the parent peak and report the sum as one 99% peak. A 20–50% ramp over 40 minutes on a 4.6 × 250 mm, 5 µm C18 column at 1.0 mL/min separates them, and the same lot suddenly reports 96.4%. Neither chromatogram is fraudulent. One is just uninformative. If a supplier will not send the raw chromatogram — the actual trace with the time axis, gradient table, and integration marks, not a cropped screenshot — assume the method was chosen to flatter the result.

A purity percentage without the wavelength, the gradient, and the column dimensions is a marketing claim wearing a lab coat.

Mass spectrometry answers identity, not purity

MS is the cheapest insurance against the worst failure mode: a correctly-purified, well-characterized batch of the wrong compound. ESI-MS on a 30-residue peptide should land within 0.5 Da of theoretical; MALDI-TOF in linear mode runs looser, ±1–2 Da, which is fine for a 3,500 Da target and marginal above 10,000 Da.

What MS does not do is distinguish isomers. A D-amino acid substituted for its L counterpart at position 4 has identical mass. Aspartate that has rearranged to isoaspartate during synthesis or storage has identical mass. Both are common, both change the molecule, and both pass an MS identity check without comment. Detecting them requires chiral amino acid analysis after hydrolysis or a specific isoAsp assay, which almost no supplier runs and almost no buyer requests. If isomeric identity is material to the work, that test is on you to commission.

The number almost no COA shows you

Here is the gap that separates buyers who have run this analysis from buyers who have not. Chromatographic purity and net peptide content are different measurements, and only one of them tells you how much compound is in the vial.

A lyophilized peptide purified by RP-HPLC comes out as a TFA salt. Trifluoroacetate binds to basic residues and typically accounts for 5–20% of the dried mass, scaling with arginine and lysine count. Lyophilized powder is hygroscopic and commonly holds 3–10% water by Karl Fischer, sometimes more if the vial has been opened in humid air. Add residual salts from the mobile phase and it is entirely ordinary for a material with 99% chromatographic purity to be 72–85% peptide by mass.

That is not a defect. It is the normal physical state of the material, and every honest supplier knows it. The problem is that a vial labeled "5 mg, 99% pure" is understood by most buyers to contain 4.95 mg of compound when it may contain 3.8 mg. Across a 12-month program consuming 200 vials, that discrepancy is the difference between the mass balance closing and a year of quietly irreproducible numbers.

Ask for net peptide content by quantitative amino acid analysis or nitrogen determination. If the supplier does not have it, at minimum ask for the Karl Fischer water figure and the counterion form. Our own approach to what appears on a COA is documented on the quality page, and the comparison in our high purity peptides review walks through what different vendors do and do not disclose.

What independent verification actually costs

Sending a sample to a contract lab is the only way to convert a supplier's claim into your own data. Approximate US commercial pricing, per sample:

  • RP-HPLC purity with a documented method: $75–$250, turnaround 3–10 business days
  • ESI-MS or MALDI-TOF identity confirmation: $100–$200, often bundled with HPLC for $150–$300 total
  • Karl Fischer water determination: $60–$120, needs roughly 10–20 mg of material
  • Quantitative amino acid analysis for net peptide content: $150–$300, 5–10 business days
  • ICP-MS heavy metals panel: $150–$400
  • GC headspace residual solvents against USP <467> limits (acetonitrile 410 ppm, methanol 3,000 ppm, dichloromethane 600 ppm): $150–$250
  • LAL endotoxin testing: $75–$150

A full independent workup therefore runs $600–$1,400 per lot and consumes 30–50 mg of material. On a $180 vial, that arithmetic is obviously upside down for a single purchase — which is the honest lead-in to the next section.

When independent testing is not worth the money

Plenty of pages on this topic imply every buyer should verify every lot. That advice is wrong often enough to be worth contradicting directly.

  • Single-vial exploratory purchases. If you are spending $150–$400 to characterize a $95 vial you will consume once in a pilot experiment, the testing budget has exceeded the scientific value of the answer. Buy from a supplier with a lot-specific COA, keep the retain sample, and spend the money on replication instead.
  • When you cannot act on a bad result. Testing is only worth it if a failing number changes something — a return, a supplier switch, a re-run. If the material is already consumed and the study is already written, you have bought an expensive regret. Our returns terms exist precisely so a failed verification has somewhere to go.
  • When you lack a reference standard. Comparing your lot against nothing tells you it is 97.6% pure by one lab's method. It does not tell you whether that is better or worse than what you had six months ago. Verification becomes useful at the third or fourth lot, when you have a baseline.
  • Low-value, well-characterized, high-volume commodity materials where the supplier already runs the assays under a quality system and the per-lot cost of testing exceeds 30–40% of the material cost.
  • When the assay does not address your actual risk. Endotoxin testing on a material used in a cell-free binding assay answers a question nobody asked.

The defensible middle path for most programs: verify the first lot from any new supplier in full ($600–$1,400, once), then run HPLC-only spot checks on roughly 1 in 5 subsequent lots at $75–$250 each. That holds annual QC cost near $1,200–$2,500 on a program spending $20,000–$40,000 on materials — 4–8% overhead, which survives a budget review.

Failure modes that a clean COA does not catch

  • The COA is for a different lot. Generic, undated COAs with no lot number are the single most common problem. A COA that does not name the specific lot on your vial is documentation of a stranger's material.
  • Time-of-manufacture data on a two-year-old lot. A certificate dated 26 months ago describes the powder as it was, not as it is. Deamidation, oxidation, and aggregation are real over that interval, particularly for sequences containing Met, Cys, or Asn-Gly.
  • Cold-chain gaps in transit. Lyophilized material stored at −20°C is generally stable for 24–36 months, but a package that sat 4 days at 90°F on a loading dock has an undocumented thermal history. Ask what the shipping method actually was rather than assuming.
  • Integration games. Excluding a shoulder from the integration window, or setting a threshold at 0.5% so that four 0.4% impurities vanish, moves a 97.9% lot to 99.1% without touching the sample.
  • Retention time reported without a system suitability injection. Without a known standard run on the same column the same day, the retention time is not comparable to anything.

Building a supplier qualification file that scales

Treat purity testing as a records problem rather than a chemistry problem, because after the first two lots that is what it becomes. For each lot, keep the lot number, the raw chromatogram file, the MS spectrum, the Karl Fischer figure, the receipt date, and the storage location. Six lots in, you can plot purity against date and see drift that no single COA would ever reveal — a supplier sliding from 98.7% to 96.2% over 14 months is visible in the trend and invisible in any individual document.

Set a re-test interval rather than testing on suspicion: HPLC re-check at 12 months for material held at −20°C, or at 6 months for anything held at 4°C. Budget one 15–25 mg retain per lot for that purpose at purchase time, not later, when the vial is nearly empty.

All materials discussed here are supplied under research use only terms, and volume buyers coordinating multi-lot qualification can arrange consistent lot sourcing through wholesale rather than assembling a program from scattered single-vial orders. If you need a specific assay run before committing to a lot, ask before you order — that request is far cheaper answered in advance than litigated afterward through contact.


Supplied strictly for in-vitro laboratory research by qualified researchers and institutions. Not a drug, food, cosmetic or dietary supplement. Not for human or veterinary consumption. Not evaluated by the FDA.

Every batch ships with its own certificate.

Purity, identity and lot number, documented for the exact vial you receive.

How we test