We Tested 7 Bottled Water Brands for Microplastics. Here's How Much Plastic You're Drinking.

Liquid Death spends millions telling you their aluminum cans are better than plastic bottles.

We tested it. They're not.

We bought 7 of the most popular bottled water brands at a grocery store in Los Angeles — Dasani, Liquid Death, Fiji, Evian, Crystal Geyser, San Pellegrino, and Path — and ran every one through our fluorescence microscopy lab.

We weren't just counting particles this time. For the first time ever, we identified what type of plastic was in each one.

The result: all 7 brands tested positive for microplastics. Every single one. And the dominant plastic in all of them was PET — the exact same plastic the bottles are made from.

Your water bottle is dissolving into your water. And so is your aluminum can.

The full ranking

Here's how all 7 brands stacked up, ranked by fluorescence signal strength. Higher means more plastic detected:

Dasani came in worst at 143. Path came in lowest at 106. But even Path — the "cleanest" brand we tested — still showed a clear PET signal.

There's no winner here. Just degrees of contamination.

And look at this: Liquid Death (aluminum can) scored 141. Dasani (plastic bottle) scored 143. Virtually identical. The can didn't help.

Why aluminum cans are a lie

This is the finding that surprised us most.

Liquid Death and Path both come in aluminum cans. Their whole brand is built on "not plastic." Premium. Clean. Metal.

But when we ran them through our process, they matched the exact same PET signature as every plastic bottle brand we tested.

Why? Because every aluminum beverage can has a thin plastic lining on the inside. It's there to prevent the metal from reacting with the liquid. And that lining is made from PET-based epoxy.

The can looks like metal on the outside. But your water is still sitting against plastic on the inside. And that plastic is shedding particles into your water — just like a bottle does.

Here's the part that gets us: our classifier identified Liquid Death and Path as PET without being told what container they came from. It just looked at the fluorescence data and said "water bottle plastic." It didn't know about the lining. It just saw the signal.

All 7 brands. Same plastic type. Whether it came from a $4 aluminum can or a $1 plastic bottle.

How we identified the type of plastic

Until now, identifying what type of plastic was in a water sample required a machine called an FTIR spectrometer. Those cost $50,000 and up. The labs that have them charge $598 to $800 per sample.

We figured out how to do it at a fraction of that cost using a fluorescent dye called Nile Red. When you stain plastic particles with Nile Red and shine different wavelengths of light on them, different types of plastic glow in different ways.

Here's the same water sample photographed under three different lights:

Each wavelength reveals different information about the plastic particles on the filter. By measuring the color intensity across all three channels, we build a unique fingerprint for each plastic type.

Here's what those fingerprints look like:

Water bottle plastic (PET) has the strongest UV response. Shopping bag plastic (PE) barely glows under blue light. Nylon fibers have the weakest UV signal. Styrofoam sits in the middle. Each one has a pattern you can't confuse with the others.

Right now we can identify 4 types of plastic in your water:

When we tested this classification system on known plastic samples, every single type separated cleanly from every other type. Six out of six possible pairings. Zero overlap.

See those orange triangles on the right? Those are the 7 bottled water brands. Every one of them clusters with the PET training samples. Not near polystyrene. Not near nylon. Right on top of water bottle plastic.

This approach isn't new science. Researchers like Meyers et al. (2022) demonstrated polymer identification through fluorescence microscopy with 88% accuracy. The NSW EPA in Australia validated Nile Red methods against traditional FTIR testing. What's new is that we're doing it at a price regular people can actually afford.

We had to build a new process to do this

Our standard process — diluted Nile Red on a nylon filter — is what we use every day to count particles in customer samples. It works. We still use it in production.

But counting particles and identifying what type of plastic they are require different levels of sensitivity. When we tried to run polymer identification with the standard process, the signal was too faint to distinguish between plastic types. Everything looked the same.

So we developed a second process: undiluted Nile Red on a PCTE filter. It's not a replacement — it's a different tool for a different job. The counting process tells you how many particles are in your water. The polymer ID process tells you what kind of plastic they are.

Here's the same Crystal Geyser sample run through both:

On the left: the standard counting process on a nylon filter. Clean, precise, optimized for particle detection. On the right: the polymer ID process on a PCTE filter with undiluted Nile Red. The stronger staining produces a 1,000x signal boost — enough to measure the fluorescence fingerprint of the plastic and classify it by type.

We also had to add a second light source. Under blue light alone, PET and polystyrene look too similar. Adding UV light revealed a 46-point gap between them that was completely invisible before. Different wavelengths reveal different things.

What we can say — and what we can't

What we can say: All 7 brands tested positive for microplastic particles. The dominant plastic type in every sample matched PET (polyethylene terephthalate) — the same plastic used in water bottles and aluminum can linings. The container is the most likely source of contamination.

What we can't say yet: We can identify the dominant plastic type in a sample, but we can't break it down particle by particle yet. We can say "your water is mostly PET." We can't say "you have 8 PET particles and 3 nylon fibers." That's coming.

These are results from an ongoing study, not a peer-reviewed paper. We're publishing them because transparency matters more than perfection. Our process is based on peer-reviewed methods (Meyers et al. 2022, Leonard et al. 2022 from the UCLA Ozcan Lab) and validated against established approaches, but our classifier is still being trained with more samples every week.

This is Round 1 of 3

We're running the same 7 brands through 3 rounds of testing over the next few months:

• Round 1 (March 2026): Fresh off the shelf — you're reading it.
• Round 2 (April 2026): Same batch, stored at room temperature for 30 days.
• Round 3 (May 2026): Same batch, stored in a hot car for 30 days.

The question we're answering: does water get worse the longer it sits in plastic? Does heat accelerate it?

If you've ever left a water bottle in your car on a hot day and drank it later — Round 3 is the one you'll want to see.

We'll publish each round's results at thewatermap.com/bottled-water.

What you can do

We're not going to tell you to stop drinking water. And we're not a filter company — we don't sell solutions.

What we do is give you the data so you can make your own decisions.

Some people see their results and realize their tap water with a basic filter is cleaner than the bottled water they've been buying. Some people discover their tap is worse and invest in better filtration. Some people just want to know.

If you want to know what's in your water — tap, filtered, or bottled — you can order a test kit. It's $99 with free shipping, and you get results in 48 hours.

All of our data is public. Every test we run goes on the map. You can see results from homes and brands across LA at thewatermap.com.