A team put sensors in 87 homes and measured a gas stove on high emitting benzene at up to 25 times the electric rate, and the bedrooms were still above health limits hours after the flame went out

The blue ring on the gas stove lights instantly and burns clean to look at.

It is the picture the industry spent decades putting in front of people, and roughly 47 million American homes still cook this way.

Down the hall, two rooms away, a child’s bedroom door stands half open.

It estimated 16 to 69 excess leukemia cases a year, with a central figure near 43, among the 6.3 million people exposed to the highest emitting stoves.

There is a small pump on the nightstand, drawing air through a glass tube.

The burner has been off for hours.

The tube is still collecting something.

Why a flame makes a carcinogen out of clean fuel

Natural gas is mostly methane, and burning it is not the tidy chemical reaction the brochure implies.

A burner never achieves complete combustion. Some fraction of the fuel is rearranged rather than fully oxidized, and the carbon atoms come back together into heavier molecules.

Benzene is one of them, and it is the same chemistry that produces benzene in a refinery flare or an oilfield burn.

The difference is the distance. A flare does that a mile away behind a fence, and a cooktop does it at chest height in a room people sit in.

Benzene is a known human carcinogen, and the health agencies that set benchmarks for it do not identify a level below which exposure is considered safe.

Which is why the interesting question was never whether a flame makes it. It was how much reaches you.

What the hood over the stove does not catch

Building codes require combustion appliances to be vented outdoors. Furnaces, boilers and water heaters all have a flue.

A gas cooktop has a hood, and a hood is a fundamentally weaker device.

It only captures what rises into its footprint. Anything that spreads sideways at counter height before it gets there is simply in the room.

Many residential hoods also recirculate rather than vent, which moves air without removing anything from the house.

So the plume behaves like any other indoor gas. It drifts along hallways, under doors and into rooms where nobody is cooking.

The measurement that made this concrete was not taken at the stove. It was taken in the bedroom.

What the sensors and the model each found

The measurement work covered 87 homes in California and Colorado, with burners set to high and ovens to 350 degrees Fahrenheit.

Mean benzene emissions ran 2.8 to 6.5 micrograms a minute, which is 10 to 25 times the rate from electric coil and radiant stoves.

Neither induction cooktops nor the food itself emitted detectable benzene. Frying salmon or bacon produced none.

Bedroom concentrations rose above chronic health benchmarks in some homes and stayed there for hours after the burner was off. That work ran in an environmental chemistry journal in June of 2023.

A separate paper in 2025 took those measurements into a population model across 24 floor plans from 690 to 2,840 square feet, at low, medium and high use, ventilated and not.

It estimated 16 to 69 excess leukemia cases a year, with a central figure near 43, among the 6.3 million people exposed to the highest emitting stoves.

Where the numbers get softer

Those are two different studies, and the leukemia figure is modeled rather than counted.

It also applies to the top 5 percent of emitters. The other 95 percent of stoves are lower, and the model does not say they are harmless, only that this estimate is not about them.

The gradient matters more than the headline. Medium use without ventilation modeled around 10 cases a year, and low use about 1.5.

Children came out at 1.85 times the adult risk, which is body weight arithmetic. They breathe faster and weigh less, so the same air delivers a larger dose per pound.

The one thing a household cannot do is identify its own stove. Asked how to tell whether yours is a high emitter, the lead researcher answered that you do not.

Two identical models can differ by burner age, gas composition and air to fuel ratio, and no consumer sensor or flame color reports it.

What actually reduces it, in order

The measured interventions are ranked, and the order is not intuitive.

Opening windows all day removed 95 to 99 percent of the benzene. Even a modest routine, an hour in the morning and afternoon and two at night, cut levels by up to 42 percent.

A high efficiency vented hood brought concentrations down to about a quarter of the unventilated case, which is real and is not elimination.

An induction burner emits none at all, and the researchers are direct that replacing the appliance is the only route that removes the exposure rather than diluting it.

All of which sits at the end of a long chain, from wells nobody logged to deepwater pressure to a ring on a countertop.

The measured half of this is solid, the modeled half is an estimate, and the window is free.