How to test for microplastics: detection methods, size limits and what a lab report shows

How microplastics are detected, step by step

A lab test for microplastics has five parts. The lab collects the sample without adding plastic, isolates the particles, identifies the polymer, reports a count or a mass, and checks the result against blanks. The method sets two things: the smallest particle the lab can see, and whether you get a count or a mass.

Step	What the lab does	Why it matters
Collect	Glass or stainless steel containers, and a field blank where the method needs one	Plastic containers and open air add particles
Isolate	Filter or sieve water; dissolve or digest the rest of a food or powder	Particles must lie on a filter, clear of material that looks like plastic
Identify	Match each particle’s infrared or Raman spectrum to a polymer library, or heat the residue and read the breakdown products	Only a spectrum or a pyrolysis fingerprint shows that a particle is plastic
Report	Particles per litre by polymer and size class, or micrograms per litre by polymer	Counts and masses answer different questions and are not directly comparable
Check	Procedural blanks, spiked recovery samples, a detection limit	Lab air, water and clothing carry microplastics too

The steps are the same for drinking water, food and supplements. Isolation is the step that changes most, and for most products it is the hardest.

Microplastic detection methods compared

Method	What it measures	Smallest particle	Result	Main limit
µ-FTIR (infrared microscope)	Infrared spectrum of each particle	About 10 µm	Count, size, shape, polymer	Slow per particle; misses the smallest particles
Raman microspectroscopy	Raman spectrum of each particle	About 1 µm, lower with care	Count, size, shape, polymer	Fluorescence can hide the spectrum; long run times
LDIR (laser direct infrared imaging)	Infrared spectra from a quantum cascade laser	About 20 µm in published work	Count, size, polymer	Misses the smallest particles, like µ-FTIR
Pyrolysis GC/MS	Breakdown products of each polymer, heated to 600 °C	No size floor; needs microgram amounts of polymer	Mass per polymer	No count, size or shape
Microscope or Nile red stain	What a particle looks like, or whether it takes up a dye	Set by the lens	Suspected particles	Cannot confirm a polymer

µ-FTIR

An infrared microscope records a spectrum from each particle on a filter, and software matches it with a library of polymer spectra. The diffraction limit holds its spatial resolution to about 10 µm. Focal plane array detectors take thousands of spectra at once, so a lab can image a whole filter. In California’s 22-lab method study, FTIR named the polymer correctly for 95% of the particles analysed. It took 12 minutes per particle on average (standard deviation 9).

Raman microspectroscopy

Raman uses a visible laser, so it resolves particles down to 1 µm and, with care, to about 300 nm. Its weak point is fluorescence from minerals, organic residues and microbes, which can swamp the spectrum. In the same study, Raman named the polymer for 91% of particles. A 2016 comparison of one marine sample found about 35% fewer microplastics by FTIR imaging than by Raman imaging, especially below 20 µm.

LDIR

Laser direct infrared imaging uses a quantum cascade laser. It scans the whole filter at one wavenumber to find particles, then collects a wider spectrum from each one. A 2024 study of soft drinks used it for particles from 20 to 500 µm.

Pyrolysis GC/MS

ASTM D8401 heats microgram amounts of the isolated material to 600 °C, separates the fragments by gas chromatography and identifies each polymer by its marker compounds in the mass spectrometer. The result is a mass per polymer. Particle size does not limit it; mass does. The method works on microgram amounts, so small particles register only when enough of them are present. It gives no count, size or shape.

Microscopes and stains

Under a microscope, an analyst counts suspected particles by colour and shape. Nile red, a fat-soluble dye that biologists use to stain lipids in cells, makes many plastics glow under blue light and speeds up screening. Both find candidates, and neither shows what a particle is made of. Spectroscopy has to follow.

How accurate microplastic testing is: a 22-lab study

To build its drinking water methods, California ran a method study: 22 labs in six countries received blind samples of clean water and a lab blank. Each sample held known microplastics (PE, PS, PVC and PET, in four size classes from 1–20 µm to over 500 µm), plus natural hair, fibres and shell fragments. The results:

• Mean recovery by stereomicroscope was 76%, and 92% for the three larger size classes. Recovery of the 1–20 µm class was poor.
• FTIR and Raman named the polymer for 95% and 91% of particles.
• Blank samples held 91 particles of lab contamination on average (standard deviation 141).
• By eye, labs reported between 1 and 51 of the 80 natural particles per sample as microplastics (mean 16). Less experienced labs reported higher concentrations.

The study found good recovery and identification above 50 µm. California’s monitoring plan follows that: its first phase measures particles larger than 20 or 50 µm, depending on the method. When the plan was written in 2022, no commercial supplier offered proficiency testing samples for microplastics in water. Results also differ between labs for other analyses, as our guide on why labs disagree shows for peptides.

Blanks and contamination control

Microplastics can reach a sample from lab air, water, reagents and clothing, so a result is only as good as its blanks. A consensus paper from 12 European labs sets minimum rules for clean-water analysis:

• At least one lab blank per series or per day, for no more than 10 samples. The series fails if a blank exceeds the detection limit.
• The detection limit is the mean of 10 blanks plus three standard deviations.
• Glass and stainless steel tools in place of plastic ones, and cotton or antistatic lab coats.
• Spectral libraries that include proteins and cellulose, so they are not mistaken for polyamide (nylon).
• Three replicate samples as good practice.

California’s monitoring rules require raw data to be uploaded without blank correction, and all blank contamination to be reported.

Blanks decide how far a result can be read. A 2024 PNAS study used stimulated Raman scattering microscopy, which can detect particles as small as 100 nm. It estimated about 240,000 plastic particles per litre of bottled water, about 90% of them nanoplastics. A published letter argued that the study’s purified-water blank, taken through every step, showed about the same plastic level as the bottled water and should have set the detection limit. The authors replied that the purified water was itself contaminated, so they used filters exposed to the same lab air as their blank. The exchange shows why a report must state which blank it used and what that blank contained.

How to test for microplastics at home, and what a kit cannot show

A home test can filter water and show the filter under a magnifier, a phone microscope or a dye under blue light. That shows particles. It cannot show which particles are plastic. In the 22-lab study, trained analysts with stereomicroscopes still reported cotton, cellulose, fur and shell as suspected microplastics. Fibres from clothes and from the air also land on any open filter. Home kits for peptides have the same problem: see peptide test kits.

What you can do at home is collect a sample a lab can use:

• Use the lab’s own clean containers, glass or stainless steel, and its instructions on volume.
• Keep plastic funnels, tubing and lids away from the sample where you can.
• Ask whether to take a field blank: a container of clean water opened at the sampling point for the same time.
• Record the date, the volume and the source.

Microplastic testing in water: standards, California and the EPA

Document	What it covers	Size range	Result
ISO 16094-2:2025	Drinking water and other low-solids water, by microscopy with vibrational spectroscopy	1–5,000 µm	Count by size class; 10 named polymers
ASTM D8332-20 and D8333-20	Collecting and preparing water samples for Raman, infrared or pyrolysis GC/MS	Set by the filter or sieve	Sample ready for analysis
ASTM D8401-24	Pyrolysis GC/MS of water samples	No lower particle size	Mass per polymer
California methods (2022 Policy Handbook)	Drinking water, by infrared or Raman spectroscopy	Phase one: above 20 µm or 50 µm, depending on the method	Count by size class and polymer

The ten ISO polymers are PE, PP, PET, PC, PS, PTFE, PVC, PA, PMMA and PU.

California was the first government to require microplastic monitoring in drinking water. Its plan set phase one, on source waters, for autumn 2023 to autumn 2025. Phase two, on treated drinking water and particles both smaller and larger than 20 µm, is set for autumn 2026 to autumn 2028. The state’s accreditation programme offers accreditation for four size classes: above 500 µm, 500 to 212 µm, 212 to 20 µm and 212 to 50 µm.

The US EPA put microplastics on its draft sixth Contaminant Candidate List on 2 April 2026, and says that listing is not regulation. Its proposed UCMR 6 rule of 1 July 2026 left microplastics out. The reason it gave: there is no validated EPA or consensus drinking water method with the quality control, accuracy and precision the monitoring needs. The EPA said it will keep evaluating ASTM D8332-20 and D8333-20.

What a microplastic lab report should show

Item	What to look for
Method	µ-FTIR, Raman, LDIR or pyrolysis GC/MS, and the standard or procedure followed
Size range	The smallest size counted, and the filter pore size
Polymer library	Which polymers can be named, and the match threshold (a hit quality index of 70% has been suggested as a cut-off)
Results	Particles per litre or per gram by polymer and size class, or mass per polymer
Blanks	Each blank’s count by polymer, beside the sample result
Detection limit	How it was set; results below it are not reported as numbers
Recovery	Spiked samples in the same kind of material
Subsampling	The share of the filter measured, if not all of it
Raw data	Spectra or pyrograms for the identified particles

“Not detected” means nothing was found above the method’s size limit and detection limit. It says nothing about smaller particles.

What drives the cost of microplastic testing

California’s method study found that microplastic analysis costs more, and takes longer, than tests for many regulated and unregulated contaminants. Without prices, these are the factors that change the work:

• Lower size limit. In clean drinking water, a 5 µm limit can mean up to 10,000 particles of all kinds per litre. At 1 µm, more than a million can be found.
• Share of the filter measured. Measuring every particle costs more than measuring a representative part.
• Quality control. Blanks, replicates and spiked recovery samples are extra analyses.
• Sample clean-up. Food, powders and biological material must be digested or dissolved first.
• Count and mass. Spectroscopy and pyrolysis GC/MS together cost more than either alone.

Supplements, peptides, vials and packaging

The standards above are written for water. A capsule, powder or peptide vial must first be dissolved or digested, so that any plastic is left on a filter. The lab must then show, with spiked samples, that the clean-up neither destroys plastic nor leaves residue that looks like it. Peptides need one more check. Amino acids in a peptide are joined by amide bonds, the bond found in proteins and in polyamide (nylon), and the consensus paper asks for libraries with proteins so they are not mistaken for polyamide. Our guide to microplastics in supplements covers products in more detail.

Injectable medicines already have a particle test, and it is a different test. USP <788>, harmonised with the European and Japanese pharmacopoeias, counts subvisible particles by light obscuration, or by microscope where that method does not suit. For containers under 100 mL, the average must not exceed 6,000 particles of 10 µm or larger and 600 of 25 µm or larger per container. It counts particles of any material and does not say which are plastic.

Chemicals from vials, stoppers, pens and packaging are a separate question with separate tests: see extractables and leachables, BPA testing and PFAS testing.

What HK Analytical does today

HK Analytical is adding microplastic, PFAS and BPA testing. Write to contact@hkanalytical.com to hear when it starts. You can also reach us through the contact page.

Today we test peptides: identity (LC-MS), purity (HPLC), content per vial, blind GLP-1 identification, endotoxin (LAL), Sterility (USP <71>, 14 days), residual solvents, counter-ion (TFA, acetate), water content (Karl Fischer) and vial-to-vial consistency, in any mix per sample. The test pages explain each one. Every number comes with its raw data, in a signed record you can check on /verify, and the example report shows the format. We do not hold ISO/IEC 17025 accreditation. Quotes for current tests are on request at /quote.