ICP-MS vs AAS for Supplement Heavy Metal Testing — Which Method?

When you submit a supplement for heavy metal testing, the lab will typically run either ICP-MS or AAS. These two methods dominate the industry, but they are not interchangeable. Your choice affects detection limits, the number of metals screened, cost per sample, and whether your COA holds up under Amazon compliance review or FDA scrutiny.

Quick answer

ICP-MS runs all four heavy metals (arsenic, cadmium, lead, mercury) simultaneously at parts-per-billion (ppb) detection. AAS tests one metal at a time at parts-per-million (ppm) or ppb levels depending on the atomization technique. ICP-MS costs $175-350 per sample. Flame AAS costs $75-150 per sample. Graphite furnace AAS falls in between. For Amazon compliance, California Prop 65, and USP <2232>, ICP-MS is the safer choice. For routine in-house QC on a limited budget, AAS can work if you know which metals matter most for your raw material.

What each method actually measures

Both ICP-MS and AAS measure elemental metal concentrations in a digested sample. The sample is first acid-digested (typically nitric acid and hydrogen peroxide in a microwave digestion system), then the liquid is introduced into the instrument. The difference is how each instrument detects the metals.

ICP-MS uses an argon plasma torch at roughly 6,000-10,000 K to ionize the sample. The ions are then separated by mass-to-charge ratio in a quadrupole mass spectrometer and counted by an electron multiplier detector. It measures all metals in a single run. Detection limits typically reach 0.01-0.1 ppb for most elements.

AAS measures light absorption by ground-state atoms. A hollow cathode lamp specific to each element emits light at the element's characteristic wavelength. The sample is atomized either in a flame (flame AAS, roughly 2,300-2,800 degrees C) or in a graphite tube furnace (GF-AAS, up to 3,000 degrees C). Each metal requires a separate lamp and a separate run. Flame AAS detection limits are typically 0.01-1 ppm. GF-AAS reaches 0.01-1 ppb for some elements, but only one element at a time.

Detection limits compared

The table below shows typical method detection limits (MDL) for the four heavy metals most relevant to dietary supplements:

For context, USP <2232> sets limits of 10 mcg/day for lead and 15 mcg/day for arsenic for most supplements. California Prop 65 requires warning labels above 0.5 mcg/day of lead. These levels are low enough that AAS, particularly flame AAS, may not reliably quantify them.

⚠️ Note

If you sell in California or on Amazon, use ICP-MS. Flame AAS detection limits may fall above Prop 65 thresholds, meaning you cannot demonstrate compliance below actionable exposure levels.

Real cost comparison

Here is what you can expect to pay at a typical ISO 17025-accredited lab in the US:

The per-metal cost is always lower with ICP-MS when screening four or more elements. If you only need one metal — say, lead on a single-ingredient calcium powder — flame AAS may be the cheapest option.

When AAS is good enough

AAS makes sense in these scenarios:

1. Single-metal screening. If your raw material risk assessment flags only one metal of concern (e.g., lead in a calcium carbonate mined from a known deposit), AAS is cost-effective.

2. High-concentration products. Mineral supplements where the target metal is present at percent levels (e.g., zinc at 15 mg per tablet). Flame AAS handles high concentrations well without the dilution and matrix effects that can complicate ICP-MS.

3. In-house QC with limited budget. If you test every incoming lot and your lab already owns an AAS instrument, the per-sample consumable cost is lower than ICP-MS. Some contract manufacturers run AAS for routine screening and send a subset of lots to an external lab for ICP-MS confirmation.

4. Mercury by cold vapor AAS. Cold vapor AAS is a specialized technique that achieves excellent mercury detection (0.01-0.1 ppb) and is often cheaper than ICP-MS for mercury-only testing. It is recognized by USP <2232> for mercury analysis.

When you need ICP-MS

ICP-MS is the right choice when:

1. You sell on Amazon. Amazon's Supplement Compliance Requirements reference USP <2232> and ISO 17025 testing. While Amazon does not mandate a specific method, ICP-MS data will pass documentation review more smoothly than AAS, especially for multi-metal panels. See our guide on Amazon approved labs.

2. You sell in California. Prop 65 thresholds for lead (0.5 mcg/day MADL) and cadmium (4.1 mcg/day MADL) are low enough that flame AAS may not provide defensible data. ICP-MS detection limits are 10-100x lower.

3. You export to multiple markets. The EU, Canada, and Australia all maintain heavy metal limits. ICP-MS provides a single data set that satisfies most regulators.

4. Your raw material has multi-metal risk. Botanical ingredients grown in polluted soil, marine-sourced ingredients (seaweed, kelp), and mineral clays often carry arsenic, cadmium, lead, and mercury simultaneously. ICP-MS screens all of them in one run.

5. You want speciation data. Certain metals require species identification. Arsenic in rice protein or seaweed may be predominantly organic arsenic (arsenobetaine), which is less toxic. ICP-MS coupled with HPLC can separate and quantify arsenic species. AAS cannot do speciation.

How labs report results

Regardless of method, your COA should show:

• The method used (e.g., EPA 6020B for ICP-MS, EPA 7061A for GF-AAS arsenic)
• The detection limit and quantitation limit for each element
• Results in appropriate units (mcg/g, ppm, or mcg/serving)
• Whether results are reported on a per-sample or per-serving basis
• The lab's ISO 17025 scope of accreditation for the method and matrix

A common mistake is receiving a COA that reports "ND" (not detected) with a detection limit higher than your regulatory threshold. For example, if your Prop 65 lead limit is 0.5 mcg/day, a detection limit of 1 ppm (roughly 1 mcg/g) on a 1 g serving means you cannot demonstrate compliance. Always check the detection limit, not just the result.

💡 Note

When requesting quotes, specify "USP <2232> by ICP-MS with reporting limits below applicable Prop 65 thresholds." This tells labs you need pharma-grade detection, not just a pass/fail screen.

Which labs use which method

Most commercial supplement testing labs offer both methods. Eurofins, SGS, Intertek, Alkemist Labs, and smaller ISO 17025 labs all run ICP-MS routinely. Flame AAS and GF-AAS are more common in older labs, in-house manufacturer labs, and university-affiliated labs.

Ask the lab directly: "Do you run USP <2232> heavy metals by ICP-MS, and is this method on your ISO 17025 scope for dietary supplement matrices?" If the answer is no to either question, consider another lab. See our lab comparison guide.

Method selection by product type

Related guides

• Heavy metal testing guide — USP <2232> explained
• Heavy metal testing for supplements — what to test and when
• Supplement testing cost guide — 2026 pricing
• How to find a supplement testing lab
• Finished product testing requirements

FAQ

Q: Will Amazon accept an AAS heavy metal COA?

A: Amazon does not specify an approved method list. An AAS COA from an ISO 17025-accredited lab may pass review. However, if the detection limits are higher than your product's Prop 65 or USP thresholds, the COA may be rejected or flagged for additional testing. ICP-MS is safer for Amazon documentation.

Q: How much does ICP-MS cost vs AAS?

A: ICP-MS for a standard USP <2232> panel (As, Cd, Pb, Hg) costs $175-350 per sample. Flame AAS for the same four metals costs $75-150 total. GF-AAS costs $150-250. The per-metal cost is lower with ICP-MS when testing four or more elements.

Q: Can I use AAS for Prop 65 compliance testing?

A: Possibly, but risky. Prop 65 lead MADL is 0.5 mcg/day. A 1 g serving with 0.5 ppm lead hits that threshold exactly. Flame AAS detection limits (0.05-0.2 ppm) may not distinguish between 0.3 ppm (compliant) and 0.5 ppm (warning required). GF-AAS or ICP-MS is strongly recommended.

Q: What is speciation and when do I need it?

A: Speciation identifies the chemical form of a metal. For example, inorganic arsenic (toxic) vs. arsenobetaine (relatively non-toxic, common in fish). Seaweed, rice-based ingredients, and fish oils may need arsenic speciation. AAS cannot provide speciation. ICP-MS coupled with HPLC can.

Q: How long does heavy metal testing take?

A: Standard ICP-MS heavy metal panels take 3-5 business days at most commercial labs. AAS takes 5-7 business days because each metal requires a separate run. Speciation analysis adds 2-5 business days. Rush service (1-2 business days) is available at most labs for a surcharge of 50-100%.

Quick Reference

Lab Category: Elemental Analysis / Heavy Metals

Methods covered:

Sample requirements: 5-50 g of finished product or raw material per sample.

Turnaround: 3-7 business days standard; 1-2 business days rush available.

Accreditation: ISO 17025 required for regulatory and Amazon use. Verify the method is on the lab's scope for your matrix type.

Pricing:

Country/Region: United States (FDA 21 CFR 111, USP <2232>, Prop 65), European Union (EC 1881/2006), Canada (NHPD QARR), Australia (TGA requirements).

Get lab testing quotes →