Heavy Metals and Elemental Impurities: How ICP-MS Testing Works Under ICH Q3D

A certificate of analysis answers one question well: how much of what's in the vial is the compound you ordered? Research-grade peptides are sold for research use only, and the purity figure on that certificate is measured against the peptide itself. Metals are a separate question. Trace elements arrive with the raw materials, the reagents, the water, the glassware, and — in some synthetic routes — from a catalyst that was put there on purpose. None of them show up in an HPLC purity number, because HPLC was never looking for them.

That second question has its own regulatory framework: ICH Q3D, Elemental Impurities. It redefined what a "heavy metals" line on a certificate is supposed to mean — and what to ask when one shows up.

What ICH Q3D Actually Changed

In one sentence: the rules stopped asking are there metals in here? and started asking which metals, how much, and where did they come from?

What Q3D displaced was a colour comparison. Treat the sample with a sulfide reagent, then hold the resulting precipitate against a lead standard and judge which is darker. One pass-or-fail verdict, for a loosely defined group of elements, operator-dependent enough that the answer could shift with the analyst. The ICH Q3D guidance published by the FDA frames the modern approach as a deliberate replacement for it.

Q3D does three things at once. It evaluates the toxicity data for each element of concern, converts that into a permitted daily exposure — a numeric ceiling, per element — and wraps both in a risk-based control strategy, so whether an element needs routine measurement depends on whether the process could plausibly have introduced it. The guideline also names the tools it expects: validated instrumental procedures, specifically ICP-MS and ICP-OES.

Multi-element instruments won that role for a practical reason. Older techniques read one element at a time — a survey of cadmium and lead in herbal tinctures used graphite-furnace atomic absorption, one element per run. Fine for a two-element check. Impractical for the seven-to-sixty-element panels a modern risk assessment calls for. The architecture mirrors the ICH Q3C framework for residual solvents, applied to a different impurity class.

The 24 Elements, Sorted Into Four Classes

Not every element gets equal attention. Q3D ranks 24 of them on two criteria at once — how toxic they are, and how likely they are to turn up in the first place.

• Class 1 — avoid. Arsenic, cadmium, mercury and lead: significant human toxicants with little or no legitimate role in pharmaceutical manufacturing, so there's no good reason for them to be present at all.
• Class 2A — assess every time. Cobalt, nickel and vanadium, common enough in natural materials and process equipment that they're evaluated in every risk assessment whatever the synthetic route.
• Class 2B — only if deliberately introduced. Low natural abundance and low probability of travelling along with other materials. This is where several catalyst metals live.
• Class 3 — route-dependent. Relatively low toxicity by the oral route, with correspondingly high ceilings, but still worth considering for parenteral and inhalation products.

The adopted European text of ICH Q3D (R2) carries the same classification — the guideline runs across ICH member regions, not one jurisdiction. The consequence shows up in the literature: published ICP-MS methods converge on the same seven-element core, because Class 1 plus Class 2A is mandatory regardless of route. A validated method for a finished tablet product targeted exactly cadmium, lead, arsenic, mercury, cobalt, vanadium and nickel.

Permitted Daily Exposure, and Why the Route Changes the Number

A permitted daily exposure, or PDE, is a ceiling on total daily intake of one element. First-time readers of these tables are usually caught out by the same thing: the identical element gets a different ceiling depending on how it enters the body.

Here is the contrast for the seven mandatory elements, in micrograms per day. Oral, from that tablet study: cadmium 5, lead 5, arsenic 15, mercury 30, cobalt 50, vanadium 100, nickel 200. Parenteral, from a parenteral-product study: cadmium 2, lead 5, arsenic 15, mercury 3, cobalt 5, vanadium 10, nickel 20.

Mercury drops by a factor of ten. So do cobalt, vanadium and nickel. The reason is absorption: the oral route doesn't deliver everything that's swallowed, because gastrointestinal uptake of most of these elements is incomplete, and the parenteral route bypasses that filter. Where route-specific toxicity data was thin, Q3D derived the parenteral and inhalation ceilings from the oral ones using modifying factors based on oral bioavailability — which is why the ratios look so systematic.

A ceiling on intake still has to become a limit on concentration in the material. Q3D offers four conversions: Option 1 assumes a 10-gram daily intake and derives one permitted concentration for every component; Option 2a uses the product's actual maximum daily intake; Option 2b apportions the allowance across components; Option 3 tests the finished product directly against a PDE-derived acceptance criterion.

How ICP-MS Measures Parts Per Billion

The instrument does three things in sequence. The prepared solution is nebulised into an argon plasma hot enough to strip electrons from the atoms in it; the resulting ions are pulled into a mass analyser, which separates them by mass-to-charge ratio and counts them. Because each element is counted at its own mass, one run can cover a very wide panel — one published survey quantified 62 elemental impurities in a single workflow.

The catch is that ICP-MS can't look at a powder. The sample has to be a solution first, which means acid digestion — and digestion is the slow step, not the instrument time. Closed-vessel microwave digestion runs at up to 100–150 bar and 180–240 degrees Celsius and takes at least two hours; the parenteral-product team built a graphite digestion specifically to avoid that, finishing in about 30 minutes at far lower pressure.

Interference is the harder problem. The plasma doesn't only ionise the sample — it also forms polyatomic species from the acid matrix and the argon itself, some landing on exactly the same masses as the target elements. Validated methods handle this physically rather than in software. Both studies above ran a collision cell flooded with helium: larger polyatomic ions collide more often and lose more energy than smaller analyte ions of the same nominal mass, so an energy barrier downstream rejects them.

What that buys, for the mandatory seven, is quantitation in the sub-microgram-per-litre range: 0.25 micrograms per litre for cadmium and lead, 0.75 for arsenic, rising to 8 for nickel. Sensitivity is matrix-dependent, though — a caveat that matters more than it sounds.

What Makes an ICP-MS Method Valid Under USP <233>

An instrument reading is only evidence if the method behind it has been shown to recover a known quantity of the element from this particular matrix. USP <233> defines how that gets demonstrated.

It starts by converting a PDE into a working concentration limit — the J-value — by dividing the permitted daily exposure by the product's maximum daily intake and applying the procedure's dilution factor. The method is then challenged by spiking samples with known amounts of each element at 0.5 and 1.5 times that J-value. Acceptable recovery is 70–150%.

That band is wide by chromatography standards, and the width is informative: pulling a trace element quantitatively out of a digested solid matrix is genuinely harder than assaying a compound already in solution. Real methods land well inside it — the tablet method reported correlation coefficients of 0.9996 to 1.0000, accuracy from 91.8% to 103.6%, and relative standard deviation no greater than 1.8%.

Then there's the rule that actually decides how much testing happens: the 30% control threshold. Where measured levels stay consistently below 30% of the applicable PDE across batches, process control can substitute for routine release testing. In the parenteral study, every elemental impurity in all six production batches came in under that bar, and the authors concluded no additional specification was required — risk-based logic putting the testing burden where the risk is.

Catalyst Residues Are a Different Story

Lead and cadmium arrive by accident. Palladium and platinum were invited.

That distinction drives how the two groups get handled. Environmental contaminants enter with raw materials, process water and equipment wear — diffuse sources, hard to predict. Catalyst metals such as palladium, platinum, rhodium, ruthenium, iridium and osmium are introduced deliberately during synthesis, so the route chemistry tells you in advance which ones could be there. You screen for what the process could plausibly have left behind, not blindly.

Palladium is the hard case on both counts. Work on quantifying trace palladium in active pharmaceutical ingredients notes that no universally effective removal protocol exists, because individual molecules bind palladium differently — one industry case study cited there needed 8–10 weeks to assess and clear palladium after a palladium-catalysed step. It's awkward to measure, too: routine ICP-MS quantitation of palladium in an organic matrix sits near 5 ppm, reaching 0.1–1 ppm only with extensive sample preparation. Orders of magnitude coarser than the cadmium figures, in the same technique on the same instrument class.

Solid-phase peptide synthesis isn't usually a palladium-heavy route, so that element is often a non-issue for peptides. The reasoning transfers regardless: ask what the process uses, then ask what it leaves. The organic-impurity parallel is residual TFA from the synthesis and purification route.

Reading Elemental Data on a Certificate of Analysis

Four questions make an elemental section useful rather than decorative.

• Which elements were measured? A named seven-element Class 1 and 2A panel is meaningful. "Heavy metals: pass" is the old visual test in spirit, whatever instrument produced it.
• What was the quantitation limit for each? A "not detected" is only as strong as the limit behind it, and those limits vary by more than an order of magnitude.
• Which exposure route were the limits drawn against? Parenteral ceilings are tighter, sometimes tenfold.
• Was the method validated in this matrix? Recovery figures from a spiked sample of the same material type are what make the numbers transferable.

Where these impurities originate is less often the process than people assume. The 62-element survey found raw material grade dominated: one grade of a common excipient carried cadmium at 0.6 ppm and iron up to 424 ppm, while another grade of the same material was markedly more consistent batch to batch.

A clean elemental panel is necessary but not sufficient, and one study makes that vivid. Accredited independent testing of three grey-market research peptide products ran identity, content and ICP-MS elemental screening together. The elemental findings were reassuring — lead at a trace 0.7 milligrams per kilogram, well below thresholds of toxicological concern. The content findings were not: the three vials, all labelled 10 mg, held 51.3%, 165.0% and 190.0% of their stated peptide content. Elemental screening alone would have passed all three. Pair it with the rest of the panel — net peptide content and HPLC purity measure two different things, amino acid analysis as an independent content check, and Karl Fischer water content for what the rest of the powder is.

Frequently Asked Questions

What is the difference between a heavy metals test and elemental impurities testing?

The older heavy metals test was a colorimetric check: treat the sample with a sulfide reagent, then compare the darkness of the precipitate against a lead standard by eye. One pass-or-fail answer for an undefined group of metals, and notoriously operator-dependent. Elemental impurities testing under ICH Q3D replaces it with element-specific instrumental measurement — usually ICP-MS — against a numeric limit per element.

How sensitive is ICP-MS for heavy metals in a peptide?

For the Class 1 and Class 2A elements, validated methods routinely reach quantitation limits in the sub-microgram-per-litre range — one published method reports 0.25 micrograms per litre for cadmium and lead. Sensitivity is element- and matrix-dependent, though. Palladium in an organic matrix is far harder: routine quantitation sits near 5 ppm, reaching 0.1 to 1 ppm only with extensive sample preparation.

Why are the limits lower for parenteral-grade material than for oral products?

Because the oral route doesn't deliver everything that's swallowed. Gastrointestinal absorption is incomplete for most of these elements, and the parenteral route bypasses that filter entirely. ICH Q3D accounts for this with route-specific limits — cadmium drops from 5 micrograms daily to 2, mercury from 30 to 3, nickel from 200 to 20.

Does a certificate of analysis always include elemental impurities data?

No, and that's worth checking rather than assuming. ICP-MS elemental screening is a separate test with its own sample preparation and cost, so it isn't automatically bundled with identity and purity work. When it is present, look for which elements were measured, the quantitation limit for each, and whether a result reads as a real number or as "not detected".

What to Take Away

Elemental impurity control moved from a visual pass-or-fail impression to a quantified, element-by-element exposure argument, and ICP-MS made the shift possible. But the instrument is the easy part — the real work sits in the risk assessment that decides which elements are worth looking for, and in the acid digestion that gets a solid into a form the plasma can read at all.

So the useful habit when reading a certificate of analysis is a small one: stop accepting "heavy metals: pass" and start asking which elements, measured to what limit, against which route's ceiling. That question has a real answer now, and a lab that ran the test properly can give it to you.

For research use only. Not for human or animal consumption of any kind. The information in this article is for educational purposes only and is not intended to diagnose, treat, cure, or prevent any disease. The statements made have not been evaluated by the U.S. Food and Drug Administration. These products are NOT FDA APPROVED. Please consult with a licensed healthcare professional before making any decisions regarding your health or research.

Optides LLC is a chemical supplier. Optides LLC is not a compounding pharmacy or chemical compounding facility as defined under 503A of the Federal Food, Drug, and Cosmetic Act. Optides LLC is not an outsourcing facility as defined under 503B of the Federal Food, Drug, and Cosmetic Act.