Elemental Impurities and ICP-MS: The Contaminant a Purity Percentage Cannot See
A certificate of analysis reporting 99.56% purity by HPLC-UV has said something precise and narrow: that when the material was pushed through a reversed-phase column and the eluate watched at around 214 nanometers, the main peak accounted for 99.56% of the absorbance recorded. That wavelength is chosen because the peptide bond absorbs there. It is a peptide detector.
A metal atom is not a peptide. Palladium, ruthenium, copper and nickel carry no amide chromophore, do not elute as a resolved peak in a reversed-phase gradient, and do not contribute meaningfully to absorbance at 214 nm. A vial could carry a substantial metal burden and the purity figure would not move. The percentage is not reassuring about metals — it is silent about them, which is a different thing and a more important one.
What follows is what fills that silence: where metals enter peptide manufacture, what framework exists for judging them, what instrument measures them, why a low-micromolar metal load is enough to move an in-vitro endpoint, and why our own GHK-Cu turns the question inside out.
Where the metal comes from
Three routes, and they are not equally likely.
Deliberate catalysis. Modern peptide chemistry uses transition metals on purpose. Palladium(0) complexes remove allyl and Alloc protecting groups, which is how orthogonal side-chain deprotection is done on resin when a selective cyclization or a branch point is needed. Ruthenium carbene catalysts drive the ring-closing metathesis that forms the hydrocarbon staple in stapled-peptide work. Copper(I) catalyzes the azide-alkyne cycloaddition used to build triazole linkages. In each case the metal is introduced at high concentration, does its job, and must then be taken out again by scavenging, chromatography, or both. Removal is never perfectly complete; it is a matter of orders of magnitude.
Process contact. Reactors, transfer lines, filter housings and lyophilizer shelves are alloys, and acidic cleavage cocktails and low-pH mobile phases are exactly the conditions under which alloys shed chromium, nickel and iron.
Reagents and water. Resins, coupling reagents, salts and process water each carry their own trace element profile, and a synthesis consuming liters of solvent per gram of product concentrates whatever those inputs contained.
The first route matters most for peptides, because it is the one where the metal arrived at gram scale rather than at parts-per-billion.
The framework: ICH Q3D
The governing document is the ICH guideline Q3D(R2) on elemental impurities, Step 4, March 2022. It sorts elements into four classes and sets a permitted daily exposure — a PDE, in micrograms per day — for each element on each route of administration.
- Class 1 is arsenic, cadmium, mercury and lead. The guideline states these “require evaluation during the risk assessment, across all potential sources of elemental impurities and routes of administration.” No exemption.
- Class 2A is cobalt, nickel and vanadium, which have a “relatively high probability of occurrence” and are likewise always assessed.
- Class 2B is silver, gold, iridium, osmium, palladium, platinum, rhodium, ruthenium, selenium and thallium. These “may be excluded from the risk assessment unless they are intentionally added during the manufacture of drug substances, excipients or other components of the drug product.”
- Class 3 is barium, chromium, copper, lithium, molybdenum, antimony and tin. For oral routes they are ignored unless intentionally added; for parenteral and inhalation routes “the potential for inclusion of these elemental impurities should be evaluated during the risk assessment, unless the route specific PDE is above 500 µg/day.”
Read those exemption clauses against peptide chemistry and the picture sharpens. Palladium and ruthenium sit in Class 2B — normally ignorable. But a synthesis using a palladium deprotection or a ruthenium metathesis has intentionally added them, and the exemption evaporates. The parenteral PDE for each of palladium, platinum, ruthenium, rhodium, iridium and osmium is 10 µg/day. Copper sits in Class 3 at a parenteral PDE of 300 µg/day — below the 500 µg/day cut-off, so for a parenteral product copper must be evaluated whether or not anyone added it.
The concentration arithmetic follows from the PDE. Under the guideline’s Option 1, which assumes a daily intake of the product of 10 grams or less, the permitted concentration is simply the PDE in micrograms per day divided by the daily amount in grams. Ten micrograms per day of palladium against a one-gram daily amount is 10 ppm; against a ten-gram daily amount it is 1 ppm. The ppm number is not a property of the element. It is a property of the element and the dose together, which is why quoting a bare “1 ppm palladium limit” without stating the assumed intake is meaningless.
What Q3D does not cover
Q3D applies to “new finished drug products,” and its scope explicitly reaches products containing purified proteins, polypeptides and synthetic polypeptides. It excludes herbal products, radiopharmaceuticals, vaccines, cell metabolites, DNA products, allergenic extracts, whole blood and cellular blood components, and gene, cell and tissue therapies.
Research-grade material sold for in vitro and animal-model work is not a finished drug product, so Q3D imposes no obligation on it. That is a jurisdictional fact and should be stated as one rather than dressed up as a quality argument in either direction. The useful question for someone running an experiment is not whether a limit binds, but whether the number is large enough to change what the assay reads.
The instrument
Inductively coupled plasma mass spectrometry is the method of record. A liquid sample is nebulized into an argon plasma running at several thousand kelvin, which atomizes and ionizes essentially everything in it. The resulting ions are extracted into a mass analyser and counted by mass-to-charge ratio. Because the plasma destroys all molecular structure, ICP-MS reports total elemental content and nothing about chemical form — it counts copper atoms without caring whether they arrived as free cupric ion or as part of a complex. Sensitivity runs to nanograms per liter for most elements, several orders below the concentrations that matter here.
Its older sibling, ICP optical emission spectrometry, reads light emitted by excited atoms rather than counting ions — more tolerant of dissolved solids, cheaper to run, and two to three orders of magnitude less sensitive. The compendial pair describing both is USP General Chapters <232> on limits and <233> on procedures.
The practical constraint for peptides is sample preparation: a lyophilized peptide must be digested before it can be nebulized, and the digestion has to break metal-ligand bonds without introducing metal from the acids used to do it.
Why a low metal load is not a rounding error
The reason to care is not regulatory. It is that trace metals are biologically active at concentrations well below anything that looks like poisoning.
Tsai and colleagues, reporting in Biological Trace Element Research (2026; 204:5236-5247, DOI 10.1007/s12011-026-05027-3), exposed RAW264.7 macrophages to soluble copper(II) chloride and measured a battery of endpoints at once. The 24-hour IC50 for viability was 13.24 µM. Below that threshold, at 5 to 20 µM, they recorded increased reactive oxygen species, mitochondrial depolarization, cytochrome c release and caspase activation, with genotoxic markers — micronucleus formation and comet tail moment — rising in parallel from about 10 µM. The authors were explicit that their highest concentration, 40 µM, sat above the IC50 and represented non-specific terminal cytotoxicity rather than mechanistic injury.
That is the whole argument in one experiment. The concentrations that moved ROS, mitochondrial potential and DNA damage sat below the concentration that killed the cells. A metal carried in with a peptide can leave a viability assay looking normal while shifting exactly the redox and apoptotic readouts that much peptide research uses as its primary endpoint. Attributing a mitochondrial or oxidative effect to a peptide without knowing the material’s elemental content leaves an uncontrolled variable in the experiment.
The case that inverts the question: GHK-Cu
Everything above treats metal as contamination. Our GHK-Cu 50MG breaks that frame, and it is worth working through because the failure mode is a category error rather than an arithmetic one.
GHK-Cu is the glycyl-histidyl-lysine tripeptide in complex with copper(II). The complex has a molecular weight of roughly 402 g/mol against 340.38 g/mol for the free tripeptide, so copper accounts for about 15.8% of the mass by design. In a 50 mg vial that is approximately 7.9 mg of copper — around 26 times the 300 µg/day parenteral PDE that Q3D assigns to copper considered as an impurity.
Applied naively, an elemental impurities specification would fail this product on its active constituent — not a defect in the product or the guideline, but what happens when a specification written for adventitious metal is pointed at a coordination complex. For GHK-Cu, elemental analysis is not an impurity assay at all. It is a stoichiometry and identity assay, and it should answer different questions: is the copper-to-peptide ratio the expected 1:1, and how much of the copper is bound rather than free in solution.
ICP-MS alone cannot answer the second question, because the plasma destroys the very speciation being asked about. Separating bound from free copper needs a chromatographic front-end — LC-ICP-MS — or a method sensitive to coordination, such as the shift in the complex’s visible absorbance band. GHK-Cu’s stability constant is high, log10 around 16.4, which makes substantial free copper unlikely in a well-made batch; but “unlikely” is an inference from chemistry, not a measurement on a lot.
What our certificates carry, and what they do not
The certificate behind GHK-Cu 50MG is report DBAV-GHK-Cu-50-062226 from Testides Analytical, an independent third-party laboratory, stating 99.56% purity by HPLC-UV. Our certificates follow the same pattern across the catalog: chromatographic purity, content mass and fill accuracy, all by HPLC-UV.
They do not carry an ICP-MS section and they do not carry an elemental row. Nothing on them speaks to palladium, ruthenium, nickel or chromium, and in the GHK-Cu case nothing on them confirms the copper stoichiometry either — the 99.56% is a chromatographic figure about the peptide, not a statement about the metal. If the elemental content of a batch is load-bearing for an experiment, the honest position is that our certificate does not release the material against it, and the number would have to be generated by the laboratory using it or commissioned separately.
Saying so is more useful than implying coverage that does not exist. A certificate is a record of the tests that were run, and reading it well means reading the absent rows as carefully as the present ones.
All materials described here are supplied for research use only. They are not for human use, not for veterinary use, and not for diagnostic use. All studies referenced in this article were conducted in cultured cells or animal models; nothing here describes or implies use in people.
References
- ICH. Guideline for elemental impurities, Q3D(R2), Step 4, 8 March 2022. International Council for Harmonisation. Element classification, parenteral permitted daily exposures, Option 1 concentration calculation, and scope.
- Tsai PK, Wu SW, Lee YJ, and colleagues. Multi-endpoint characterization of soluble copper(II) chloride effects on oxidative stress, mitochondrial function and genotoxic markers in RAW264.7 macrophages. Biological Trace Element Research, 2026; 204:5236-5247. DOI 10.1007/s12011-026-05027-3.
- United States Pharmacopeia. General Chapter <232>, elemental impurities — limits, and General Chapter <233>, elemental impurities — procedures. Compendial framework for ICP-MS and ICP-OES determination.
- Coordination chemistry and stability constant of the glycyl-histidyl-lysine copper(II) complex, as summarized in the published literature on the copper tripeptide.