Breathalyzer: How Does It Work? Sensors, the 2100:1 Ratio and What It Measures

Updated 26 Oct 2026 · Written by Daniel Brooks · Medically reviewed by Dr. Maya Chen, PharmD

A breathalyzer doesn't look at your blood at all. It measures the alcohol in a sample of air from deep in your lungs, then uses a fixed conversion factor to estimate what your blood alcohol concentration (BAC) probably is. How well it does that depends on the sensor inside, how you blow, how long ago you drank and how well the device has been looked after. This guide walks through how a breathalyzer works from the moment alcohol hits your stomach to the number on the screen, compares the three sensor types you'll run into, and explains what can push a reading up or down.

The short answer

When you drink, alcohol (ethanol) is absorbed into your bloodstream, mostly through the small intestine. Blood carries it to your lungs, where a small share of it crosses from the blood into the air in the tiny air sacs (alveoli). When you breathe out, that alcohol leaves with your breath. A breathalyzer captures a sample of that air, measures how much alcohol is in it, and multiplies by a standard ratio to report an estimated BAC.

So a breathalyzer works in three steps:

1. Collect a sample of deep-lung (end-expiratory) air.
2. Measure the alcohol in that air with a sensor: a fuel cell, a semiconductor or an infrared detector.
3. Convert the breath alcohol figure into an estimated blood alcohol number using a partition ratio, usually 2100:1 in the United States.

Each step has its own sources of error, which is why two devices, or two blows a few minutes apart, can give different numbers.

What a breathalyzer actually measures

People often say a breathalyzer "measures your BAC". Strictly, it measures breath alcohol concentration (BrAC), the amount of ethanol in a set volume of exhaled air. The BAC on the screen is a calculated estimate built on top of that measurement.

This is more than a technicality. Breathalyzers don't directly measure blood alcohol. They rely on a physical rule (Henry's law) that says when a liquid and a gas sit together in a closed space, a volatile substance like alcohol spreads between them in a predictable proportion at a given temperature. Your lungs are a close real-world version of that setup: blood on one side of a thin membrane, air on the other.

The 2100:1 partition ratio

In the US, breath testing assumes that 2,100 milliliters of deep-lung air hold about the same amount of alcohol as 1 milliliter of blood. That's the 2100:1 ratio. A device reads the alcohol in your breath, applies that ratio and shows a figure in grams per deciliter (the familiar 0.02, 0.05, 0.08 style numbers).

The catch is that 2100:1 is an average. Forensic research on breath testing has found that the real ratio varies from person to person, and even within one person over a single evening. Someone whose true ratio is lower than 2100:1 will tend to read high on a breath test compared with a blood draw, and someone with a higher ratio will tend to read low. Body temperature, breathing pattern and where you are in the drinking curve also move the ratio around.

Because of this, many US states write their impaired-driving laws in breath terms directly (grams of alcohol per 210 liters of breath) as well as blood terms. That way an evidential breath result stands on its own rather than as a guess at blood. This is also why we treat every home reading as an estimate, never a promise of what a police or workplace device would show.

Step 1: getting a deep-lung sample

The first air you breathe out comes from your mouth, throat and windpipe, where little gas exchange happens. The air at the end of a long, steady breath comes from the alveoli and best reflects the alcohol in your blood. Good breath testers are designed to capture that end-of-breath air:

• Minimum blow time or volume. Many devices beep or count down and won't test until you've blown long enough. If you stop early, you'll often get an error or a low reading.
• Pressure or flow sensors. Better units check that you're blowing hard enough and steadily enough.
• Mouthpieces. A mouthpiece keeps the sample path consistent and keeps saliva out of the sensor.

Why mouth alcohol throws everything off

If there's alcohol sitting in your mouth, from a sip you just took, a burp, or some mouthwashes and breath sprays that contain alcohol, the device can't tell it apart from lung alcohol. That residual "mouth alcohol" can make a reading jump well above your true level for a short while. It usually clears within about 15 to 20 minutes.

That's why police and workplace breath-testing protocols build in a waiting or observation period before an evidential test, and why every serious home device tells you to wait at least 15 to 20 minutes after your last drink, food, smoke or mouthwash before testing. If you skip the wait, you're mostly measuring your mouth.

Step 2: the sensor - three ways to measure breath alcohol

Here's where devices differ the most. There are three main sensor technologies in breath testers today, plus the original chemical method that started it all.

The original chemical breathalyzer

The first practical roadside device, the Drunkometer, was developed by Rolla Harger in the early 1930s. The device trademarked as the "Breathalyzer" came from Robert Borkenstein in 1954. It bubbled breath through a solution of potassium dichromate and sulfuric acid. Alcohol reacted with the orange dichromate and turned it green, and the change in color, read by a photocell, showed how much alcohol was present. These chemical units are history now, but the name stuck, and people still call any breath tester a "breathalyzer".

Fuel-cell sensors

A fuel-cell sensor is a small electrochemical cell with two platinum-coated electrodes separated by an electrolyte. When breath passes over the active electrode, any ethanol in it is oxidized (broken down, ending up mostly as acetic acid). That reaction frees electrons, which flow through a wire to the other electrode as a tiny electric current. The more alcohol in the sample, the larger the current. The device measures the current and converts it into a BrAC, then a BAC estimate.

Why people like fuel cells:

• Alcohol-specific. They respond strongly to ethanol and much less to other compounds that can show up in breath, such as acetone.
• Stable. Readings stay consistent over many tests between calibrations.
• Proven. Most police handheld preliminary breath testers and many workplace screening devices use fuel cells.

Trade-offs: fuel-cell units cost more, and the sensor slowly loses sensitivity with use and age, so it needs periodic recalibration or replacement. Many makers suggest recalibration about every 6 to 12 months or after a set number of tests. Check the manual for your specific model.

Semiconductor (metal-oxide) sensors

Budget keychain testers and many cheap personal breathalyzers use a semiconductor sensor, usually a heated bead of tin dioxide. In clean air the bead has a certain electrical resistance. When alcohol molecules land on the hot surface, they react with oxygen there and the resistance drops. The device reads that change and converts it into a BAC estimate.

Semiconductor sensors are cheap and small, but they have real weaknesses:

• Less specific. They can react to things that aren't ethanol: acetone (higher in people on very low-carb diets or with uncontrolled diabetes), cigarette smoke, some aerosols and solvents.
• Drift. Readings can wander with temperature, humidity and sensor age.
• Warm-up time. The bead has to heat up first, and low batteries can affect the reading.

A semiconductor unit is fine for a rough "am I clearly drinking or not" check. It's not the tool you want if you need tight numbers near a limit.

Infrared spectroscopy

The big desktop machines used for evidential testing at police stations mostly work by infrared (IR) absorption. Alcohol molecules absorb infrared light at particular wavelengths, linked to the way their carbon-hydrogen and carbon-oxygen bonds vibrate. The instrument shines IR light through a chamber filled with your breath and measures how much light is absorbed. More alcohol means more absorption.

Evidential IR instruments often check more than one wavelength so they can spot substances like acetone that absorb in some of the same regions. Some combine an IR detector with a fuel cell for a second, independent reading. These machines are large, expensive and regularly checked against known alcohol standards. You won't find one in a consumer product.

Sensor types in one minute

If you only remember one thing: fuel cells oxidize alcohol and measure the current, and they're the solid choice for handhelds. Semiconductor sensors read a resistance change on a heated bead, which is cheap but less specific. Infrared machines measure how much light alcohol absorbs, and they're the precise, expensive desktop instruments used for evidence.

Step 3: from breath reading to BAC number

Once the sensor has a signal, the device's software takes over. It compares the signal to its calibration curve, a stored relationship between sensor output and known alcohol concentrations set during factory or service calibration. That gives a breath alcohol value. The device then applies the partition ratio and shows the result as an estimated BAC.

A few details matter here:

• Calibration ages. If the calibration is old, the curve no longer matches the sensor, and readings drift. Fuel-cell sensors usually read low as they age, which is the more dangerous direction.
• Display rounding. Some devices show two decimals, others three. Some round, others truncate.
• Accuracy specs. A spec like "±0.005 at 0.050" tells you the sensor's performance at one test point in the lab. Real use adds technique and timing error on top. Treat any single reading as a number with a band of uncertainty around it.

What can make a breathalyzer read wrong

Most bad readings come from the person using the device, not the hardware. The usual causes:

Testing too soon after drinking

Mouth alcohol is the big one, covered above. There's also the absorption phase. BAC usually keeps rising for a while after your last drink, often peaking somewhere around 30 to 90 minutes later depending on food, how fast you drank and your body. A low reading ten minutes after your last drink tells you very little about where you'll be in an hour.

Breathing technique

• Holding your breath before blowing tends to raise the reading, because the air warms up and picks up more alcohol from the lungs.
• Hyperventilating (several fast, deep breaths) before blowing tends to lower it.
• Short or weak blows may not reach deep-lung air.

For an honest number, take a normal breath and give a long, steady blow until the device signals it's done.

Burping, reflux and vomiting

These can bring alcohol from the stomach back into the mouth and throat. Evidential protocols restart the waiting period if any of these happen, and you should do the same with a personal tester.

Other compounds

Fuel cells and IR machines handle these better than semiconductor units, but cheap sensors can respond to acetone, smoke, paint or solvent fumes, and some sprays. Test in clean air, away from cigarette smoke and cleaning products.

Temperature, altitude and the device itself

Very cold or hot devices, low batteries, a dirty or wet mouthpiece, and an overdue calibration can all push numbers off. Store the device as the manual says, let it reach room temperature, replace mouthpieces and keep up with calibration.

Preliminary vs evidential breath tests

You may hear about two kinds of police breath tests:

• Preliminary breath tests (PBTs) are the handheld devices used at the roadside, usually with fuel-cell sensors. They help an officer decide whether there's probable cause for an arrest.
• Evidential breath tests are done on larger, regularly certified instruments (often IR-based) under a set procedure, with an observation period and duplicate samples. These are the results used in court.

In the US, the National Highway Traffic Safety Administration (NHTSA) publishes conforming products lists for evidential breath measurement devices and for alcohol screening devices that meet its model specifications. Workplace testing programs regulated by the Department of Transportation use devices from those lists. We don't give advice on beating any police, DOT or court-ordered alcohol test, and nothing here should be used that way.

How to use a personal breathalyzer properly

If you own a breathalyzer to keep tabs on your own drinking, a careful routine gives you the most useful number:

1. Wait 15 to 20 minutes after your last drink, food, cigarette, vape or mouthwash.
2. Let the device warm up and reach room temperature. Check the battery.
3. Use a clean mouthpiece if your model has one.
4. Breathe normally, then give one long, steady blow until the device signals it has a sample. Don't hold your breath first.
5. Repeat after a few minutes. If two readings disagree by a lot, wait and test again.
6. Write it down with the time. Readings over time tell you more than any single blow.
7. Keep up with calibration as the maker recommends.

And the most important rule: a personal breathalyzer reading is never a green light to drive. Impairment starts below the legal limit, readings carry error, and BAC can still be rising. If you've been drinking, use a ride service, a designated driver or wait it out. Our BAC calculator can give a rough sense of how long alcohol may take to clear, but it's an estimate too.

Choosing a breathalyzer: what the sensor tells you

If you're shopping for a breath tester, the sensor type is the single best clue to how much you can trust it:

• For dependable personal readings, a fuel-cell unit is the better pick. TestClear lists the AlcoMate AL9000 as a fuel-cell breathalyzer; check the product page for its accuracy spec and calibration options.
• For a cheap rough check, a budget unit like the Alcoscan AL2500 costs less. Check its listing for the sensor type and accuracy claims before you rely on it near any limit.
• For single-use checks, disposable alcohol test kits give a simple color-change answer without the upkeep.

Whatever you choose, look for: the sensor type stated plainly, a stated accuracy figure with the test point, a recalibration or sensor-replacement program, mouthpieces, and clear directions on waiting times.

How the body clears alcohol (and why readings fall slowly)

Breath readings drop as your liver breaks down alcohol, and nothing you can do speeds that up in a meaningful way. Coffee, cold showers, food after the fact and exercise don't change how fast alcohol leaves your blood. On average, the body clears alcohol at roughly 0.015 g/dL per hour, though people vary. So if you reach 0.08, it can take something like five hours or more to get back to zero, and longer for some people.

That's also why a breathalyzer is a useful learning tool: watching your own readings fall slowly over a few hours can make the "wait it out" message real. For how long alcohol can show on different tests, see how long alcohol stays in your system for a breathalyzer and alcohol in urine vs blood.

Common myths about how breathalyzers work

• "Mints, gum or pennies will beat it." No. Strong mints or gum don't remove alcohol from deep-lung air, and some breath products contain alcohol and can raise a reading. The copper-penny idea is an old myth with no basis.
• "Breathalyzers measure how drunk you feel." They measure alcohol in breath, not impairment. Tolerance changes how drunk you feel, not your BAC.
• "A low reading right after drinking means I'm fine." BAC often keeps rising after the last drink. Test again later.
• "All breathalyzers are the same." Sensor type, calibration and design make a big difference in accuracy.

Keep reading

For more on alcohol testing, see our guides to how long alcohol stays in your system after your last drink and our how drug tests work hub.

Information only, not medical or legal advice. Breathalyzer readings are estimates and never mean it's safe to drive. We do not help with DOT, federal, safety-sensitive or court-ordered tests. Tampering with a test can be illegal and usually counts as a refusal. If drinking is becoming hard to control, the free, confidential SAMHSA National Helpline is available 24/7 at 1-800-662-4357. In a crisis, call or text 988.