Hydrogen Machines at Home: How They Work, What Studies Used, and Safety

By Hydrogen Studies Editorial Team · Updated

Hydrogen Studies is funded by Echo Technologies LLC, which sells hydrogen water products and hydrogen machines, including inhalation machines. Echo doesn't decide which studies we cover or how we describe them — see our editorial policy.

Short answer: A hydrogen machine is a device that makes hydrogen gas, usually by splitting water with electricity. A hydrogen inhalation machine sends the gas to a nasal cannula or mask; a hydrogen water machine dissolves it into drinking water. Machines in human studies differed enormously: by our rough estimate, the hydrogen people actually breathed ranged from about 0.1% to more than 10% of the air they inhaled. Hydrogen is flammable in air at concentrations from about 4% to 75%. Premixed gas below 4% can't burn, but the pure hydrogen or hydrogen–oxygen gas that many home machines make can ignite, or explode, near a flame or spark. No hydrogen machine has been shown to treat or prevent any disease.

What is a hydrogen machine?

"Hydrogen machine" is a loose label for several kinds of devices. Most make hydrogen gas (H₂) on the spot by electrolysis, and they differ mainly in what comes out and how you take it in.

Type What comes out How studies used it Main safety issue
Hydrogen inhalation machine (pure hydrogen generator) Nearly pure hydrogen; one home machine in a trial was rated at 99.99% [1] Nasal cannula at home, 300 mL/min, 1 hour twice a day [1] Pure hydrogen mixes with room air as it leaves the tube and passes through the flammable range near the nose [2]
Hydrogen–oxygen ("oxyhydrogen") machine About 66% hydrogen and 33% oxygen, straight from splitting water [3, 4] Nasal cannula, nasal mask or face mask, from about 30 mL/min to 6 L/min of total gas [3, 4, 5, 6, 7] The 2-to-1 mix is flammable and can detonate at any flow rate if ignited [2, 8]
Premixed low-concentration gas Hydrogen premixed at 2% to 3% in air or oxygen [9, 10, 11] Hospital studies, including through a ventilator [10] Below 4%, the mixture itself can't burn; one trial stored its hydrogen as 4% in nitrogen so the cylinders held no flammable gas [10]
Hydrogen water machine or water ionizer Water with dissolved hydrogen; ionizers also make alkaline water [12] Drinking Low or falling hydrogen output, very high pH and electrode wear [12]

Some home inhalers dilute hydrogen with room air inside the machine. Engineers at a company that makes this kind of inhaler argue that where the dilution happens matters: a machine that dilutes hydrogen right where it's made doesn't hold a flammable mix inside, while one that carries concentrated hydrogen through tubing before diluting it can leak it inside the case [8].

A hydrogen machine is not an oxygen concentrator. A concentrator filters oxygen out of room air (the control device in one hydrogen trial used a molecular sieve) [3]. Hydrogen–oxygen machines add only a little oxygen: for an adult breathing quietly, 1 L/min raises inhaled oxygen from about 21% to 21.4%, and 6 L/min to about 23% [2].

Portable hydrogen water bottles and magnesium tablets are covered separately in what is hydrogen water? and hydrogen tablets: side effects and evidence.

How does a hydrogen machine make hydrogen?

Most hydrogen machines use electrolysis: an electric current runs through water between two electrodes. Hydrogen forms at the negative electrode (the cathode) and oxygen forms at the positive electrode (the anode). If the water contains chloride, chlorine compounds also form at the anode [12].

What happens next depends on the design:

  • Membrane machines (PEM or SPE). A proton exchange membrane, also called a solid polymer electrolyte, separates the two electrodes so hydrogen and oxygen leave through different outlets. The home inhalation machine in one post-COVID trial used this design [1]. Newer neutral-pH hydrogen water devices also use proton exchange membranes [12].
  • Hydrogen–oxygen machines collect both gases together. That gives two parts hydrogen to one part oxygen, about 66.7% hydrogen [2, 3].
  • Water ionizers use a membrane to keep the acidic, chlorine-containing water from the anode apart from the alkaline, hydrogen-rich water you drink from the cathode [12].

Inhalation machines are usually rated by gas flow in milliliters per minute (mL/min) and by the percentage of hydrogen. For hydrogen–oxygen machines, only two-thirds of the rated flow is hydrogen: a 3 L/min machine makes about 2 L/min of hydrogen and 1 L/min of oxygen [3].

What matters in your body is the fraction of hydrogen in the air you actually breathe in, which researchers call FiH₂ [2]. Flow rate alone can't tell you that. It depends on how fast you breathe, whether you breathe through your mouth, how open your nose is and how well the cannula fits. In a 2026 modeling paper, the same 400 mL/min flow could give less than 1% or more than 10% hydrogen depending on the situation [2].

The same paper offers a rule of thumb for an average adult sitting quietly with a nasal cannula: divide the hydrogen flow by 30 L/min. By that rule, 300 mL/min of hydrogen gives roughly 1% and 600 mL/min roughly 2% [2]. It's a rough estimate from a computer model, not a measurement.

How much hydrogen did human studies use?

The amounts varied far more than the label "hydrogen machine" suggests. Here's what ten human studies delivered. Where a study gave only a machine flow rate, we estimated the inhaled hydrogen with the rule of thumb above.

Study Who took part How hydrogen was given Hydrogen in inhaled air
Gao 2025 [6] 66 adults with sleep problems (41 on hydrogen–oxygen) Hydrogen–oxygen machine, 60 mL/min by nasal tube, 1 hour twice a day for 7 days About 0.1% (our estimate)
Liu 2022 [5] 60 adults aged 50–70 with high blood pressure (30 on hydrogen–oxygen) Hydrogen–oxygen machine, about 30–60 mL/min by nasal cannula, 4 hours a day for 2 weeks About 0.2–0.4% (authors' estimate)
Botek 2022 [1] 50 adults recovering from COVID-19 Pure hydrogen machine, 300 mL/min by nasal cannula at home, 1 hour twice a day for 14 days About 1% (our estimate)
HYBRID II, Tamura 2023 [10] 73 adults in a coma after cardiac arrest (39 on hydrogen) Premixed hydrogen with oxygen through a ventilator for 18 hours 2% (set)
Cole 2021 [9] 8 healthy adults 2.4% hydrogen in medical air at 15 L/min through a high-flow cannula, 24–72 hours 2.4% (set)
Ono 2017 [11] 50 adults with a recent stroke (25 on hydrogen) 3% hydrogen, 1 hour twice a day for 7 days 3% (set)
Yoritaka 2021 [13] 20 adults with Parkinson's disease 6.5% hydrogen in air at 2 L/min, 1 hour twice a day for 16 weeks 6.5% in the gas supplied
Zheng 2021 [3] 108 adults with a COPD flare-up (54 on hydrogen–oxygen) Hydrogen–oxygen machine, 3 L/min (2 L/min hydrogen) by nasal mask, 6–8 hours a day for 7 days About 7% (our estimate)
Guan 2020 [4] 90 adults in hospital with COVID-19 (44 on hydrogen–oxygen) Hydrogen–oxygen machine, 6 L/min (about 4 L/min hydrogen) by nasal cannula, about 7.7 hours a day About 13% (our estimate)
Chen 2026 [7] 53 infants with epileptic spasms (27 on hydrogen–oxygen) Hydrogen–oxygen machine, 3 L/min by face mask, 1 hour four times a day for 14 days Not estimated (the adult rule of thumb doesn't apply to infants)

"Set" means the gas was premixed to that concentration. Our estimates assume an adult breathing quietly and could be off by a factor of two or more in either direction [2].

Three things stand out:

  • Doses spanned roughly 100-fold, from about 0.1% to about 13% inhaled hydrogen. Two studies that both used "hydrogen–oxygen inhalation" differed by about 100 times in hydrogen flow (40 mL/min vs 4 L/min) [4, 6].
  • The hospital studies with the tightest control used premixed gas at 2% to 3% [9, 10, 11]. The 2026 modeling paper concluded that about 1% to 4% inhaled hydrogen is the practical range suggested by current evidence, and that going above 4% adds fire risk without proven extra benefit [2].
  • Some studies went above 4%. By our estimates, two machine studies exceeded 4% inhaled hydrogen, and one premixed study supplied 6.5% [3, 4, 13]. Above 4%, the gas around the nose is in the flammable range [2].

Is hydrogen flammable?

Yes. Hydrogen gas burns in air at concentrations from about 4% to 75% by volume [14, 15]. Below about 4% there isn't enough hydrogen to keep a flame going; above 75% there isn't enough oxygen. In the middle of that range, mixtures of roughly 15% to 59% hydrogen can detonate, exploding with a shock wave instead of just burning [14].

Other properties make hydrogen easy to underestimate:

  • It ignites very easily. Igniting hydrogen in air takes about 0.02 millijoules of energy, roughly ten times less than methane, propane or gasoline. Even a static spark from your body in dry air can be enough [14].
  • You can barely see it burn. Hydrogen burns with a pale blue flame that's nearly invisible [15, 16].
  • It rises and spreads fast, but it can build up indoors. Hydrogen is about 7% as dense as air, so it disperses quickly outdoors, but leaked hydrogen can collect in an enclosed space [14, 16].
  • It isn't toxic, but it can push out oxygen. In an enclosed space, a large leak can displace the oxygen you need to breathe [14, 15].

Why 1–4% hydrogen mixtures can't burn

A mixture below the roughly 4% lower flammability limit can't carry a flame at room temperature; the flammable range widens as temperature rises [14]. That's why tightly controlled hospital studies used hydrogen at 1.3% to 3% [9, 10, 11, 17], and why one trial stored its hydrogen as 4% in nitrogen, so the cylinders never held a flammable mix [10]. The 2026 modeling paper recommends keeping both the gas that leaves the device and the air you breathe below 4%, which in practice means premixed gas delivered through a face mask [2].

Why hydrogen–oxygen machines are an explosion risk

Splitting water gives exactly two parts hydrogen to one part oxygen, sometimes called oxyhydrogen [2, 8]. Nothing dilutes it, so the gas is flammable at any flow rate and can detonate if a flame or spark reaches it [2].

  • In ignition tests, a 66% hydrogen–oxygen inhaler detonated with a loud blast that threw off its lid and wiring [8].
  • Japan's Consumer Affairs Agency has recorded accidents, including a hydrogen–oxygen inhaler whose lid flew off with a bang that left the user's ears ringing, and a leaking home hydrogen generator that exploded, damaging a bathroom door and injuring the person [8].
  • The 2026 modeling paper notes documented explosions of commercial inhalers that caused facial fractures, a torn airway and internal bleeding [2]. A 2026 review describes a published case of an explosion inside a patient's airway [18].

Makers of hydrogen–oxygen machines describe safeguards such as flame arresters, liquid-blocking parts, filters, a mist to reduce static, and hydrogen leak sensors [19]. In the COPD trial, a test of the machine running for 2 hours in a closed room found the room's hydrogen level never rose above 0.8% [3]. That's reassuring about the room, but the gas coming out of the machine is still a flammable mix [2].

What about pure hydrogen machines?

Pure hydrogen is above the 75% upper limit, so the gas inside the tube can't burn on its own. But once it leaves the cannula it mixes with room air, passes through the flammable range and can form small flammable pockets around the nose [2]. In ignition tests, pure-hydrogen inhalers exploded when lit after running for several minutes; the testers traced this to hydrogen leaking and mixing with air inside the machine [8]. For pure hydrogen through a nasal cannula, the 2026 modeling paper calls 300 to 600 mL/min, and no more than 1,200 mL/min, a "defensible, but not a risk-free" range [2].

A note on who wrote these papers: the ignition tests and accident reviews come from employees of a Japanese company that sells low-concentration inhalers [8, 18], and the safeguard review comes from a company that makes hydrogen–oxygen machines [19]. Each has a commercial stake. Their descriptions of hydrogen's basic behavior match the government data above [14, 15].

Safety steps if you use a hydrogen inhalation machine

  • Keep it away from flames and sparks: cigarettes, candles, gas stoves, heaters and fireplaces. Don't smoke while using it [14, 15].
  • Use it in a well-ventilated room, not a closet, small bathroom or car, where leaked hydrogen can build up [14, 16].
  • Check the tubing. Make sure it's firmly attached and not cracked. Hydrogen leaking inside machines was the main cause of the explosions described above, and a loose tube is one way leaks start [8].
  • Follow the maker's instructions for water, cleaning and replacement parts, and switch the machine off when you're done.
  • Stop if you feel dizzy or short of breath. Gas that displaces oxygen can lower blood oxygen [2]. In one trial, an hour of hydrogen inhalation slightly lowered healthy young women's oxygen saturation, from 96.7% to 95.9% [20].
  • Ask a clinician first if you have lung or heart disease, are pregnant, or are thinking of using a machine with a child. See hydrogen inhalation side effects for what trials reported.

What should you look for in a hydrogen inhalation machine?

We don't test or rank hydrogen machines, and this page doesn't recommend brands. Hydrogen Studies has no gas-testing lab, and it's funded by a company that sells hydrogen machines. Here is what you can check yourself, from the label, the manual and the maker's documents:

  • Hydrogen output, verified independently. Ask for the hydrogen flow in mL/min (for hydrogen–oxygen machines, the hydrogen part only) and the hydrogen percentage, measured by an independent lab that names its method. Device output varies, and the 2026 modeling paper advises verifying gas composition rather than trusting company claims [2].
  • How much hydrogen you'd actually breathe. Divide the hydrogen flow by 30 L/min for a rough adult estimate, and compare it with the 1% to 4% that most clinical studies delivered [2].
  • The type of gas. From a fire-safety standpoint, premixed gas below 4% through a face mask is the safest setup; pure hydrogen creates flammable pockets at the nose; hydrogen–oxygen gas is flammable at any flow [2].
  • How it keeps hydrogen contained. Ask whether hydrogen is diluted at the point where it's made or carried through tubing first, whether there's a hydrogen leak sensor with automatic shut-off, and what happens if a tube comes loose [8, 19]. A flame arrester can stop a flame from traveling back through the tubing, but it doesn't prevent an explosion from hydrogen that leaks inside the machine [8].
  • Electrolysis design. A membrane (PEM or SPE) machine keeps hydrogen and oxygen apart [1]. If a machine mixes them, ask what explosion safeguards it has [19].
  • What else comes out. Ask for independent test results on anything besides hydrogen in the gas or water, such as chlorine compounds (which form from chloride at the anode) and metals from electrode wear [12].
  • Water and upkeep. Follow the maker's water requirements. For water ionizers, output depends on the minerals in your water and on clean electrodes (see the next section) [12]. Beyond the price, ask about replacement parts and cleaning; price doesn't tell you the hydrogen output.
  • Registration isn't approval. "FDA registered" doesn't mean FDA approved: a registration listing doesn't denote approval, clearance or authorization, and FDA doesn't issue registration certificates [21]. According to the 2026 modeling paper, hydrogen inhalation therapy is still investigational in the United States; Japan's health ministry authorized it as an "Advanced Medical Care" treatment specifically for post-cardiac arrest syndrome; and one hydrogen inhalation device has Class III medical-device approval in China [2]. Be wary of any machine sold with disease-treatment claims.

What about hydrogen water machines and ionizers?

Water machines don't carry the same fire risk as inhalation machines, but they have their own limits. A 2022 review found:

  • Output is often low and can fall fast. New water ionizers typically make 0.2 to 1.2 mg/L of dissolved hydrogen. Scale on the electrodes can cut that to almost nothing within two weeks until the machine is cleaned with citric acid [12].
  • Ionizers need minerals. Source water needs roughly 50 mg/L of minerals for an ionizer to make meaningful hydrogen [12].
  • Very alkaline water is a concern. Japanese and Korean rules cap ionized drinking water at pH 9.8. Above pH 10, dangerously high blood potassium has been reported in some people with poor kidney function [12].
  • Electrodes can wear. Platinum-coated electrodes can shed tiny metal particles, more so at high settings, slow flow and mineral-rich water [12].
  • Some sales demos mislead. Holding a flame under running ionized water shows undissolved hydrogen escaping, not dissolved hydrogen, and ORP meters can't measure hydrogen concentration [12].

Neutral-pH membrane devices avoid most of the high-pH problems, though metal contamination can't be completely ruled out [12]. For concentration targets, see hydrogen water ppm levels; for trial doses, how much hydrogen water per day; for drinking side effects, hydrogen water side effects.

What the evidence doesn't show

  • The right dose. Few studies measured the hydrogen people actually breathed, doses varied about 100-fold, and the 1% to 4% range rests mostly on animal studies and a few hospital trials [2].
  • Long-term home safety. Most safety data come from supervised studies lasting hours to weeks [9, 10]. We found no long-term studies of daily home use.
  • Which machines are better. We found no independent human studies comparing brands or designs. Some trials were run with, or co-authored by, the companies that made the machines [6].
  • How often home machines catch fire. Accident reports come mainly from Japan's consumer agency, summarized by a company with a competing product [8, 18]. There's no systematic count.
  • Disease treatment. No hydrogen machine has been shown to treat or prevent a disease. Some trials found no benefit, including a Parkinson's disease pilot and the infant epilepsy trial [7, 13], and the largest cardiac arrest trial missed its main goal [10]. For the broader evidence, see our evidence-graded guide to hydrogen benefits; our methodology explains how we weigh studies.

Frequently asked questions

What is a hydrogen inhalation machine?

It's a device that makes hydrogen gas, usually by electrolysis of water, and delivers it through a nasal cannula or mask. Some make nearly pure hydrogen, some make a hydrogen–oxygen mix of about 66% hydrogen, and hospital studies used premixed gas at 1.3% to 3%. Output and safety differ a lot between these types.

Is hydrogen gas flammable or explosive?

Both. Hydrogen burns in air at concentrations from about 4% to 75%, and mixtures of roughly 15% to 59% can explode. It ignites far more easily than gasoline vapor, and its flame is nearly invisible. Mixtures below 4% can't burn, but pure hydrogen and hydrogen–oxygen gas from electrolysis machines can.

Is a hydrogen machine safe to use at home?

In short, supervised studies, inhaled hydrogen was generally well tolerated, but home use hasn't been studied long term, and the gas many machines make is flammable. Keep machines away from flames and sparks, use them in a ventilated room, check the tubing, and talk to a clinician first if you have lung or heart disease, are pregnant, or are considering it for a child.

What flow rate of hydrogen did studies use?

It varied widely: from about 40 mL/min of hydrogen from a small hydrogen–oxygen machine to about 4 L/min from a large one, plus premixed gas at 1.3% to 3% in hospital studies. By a common rule of thumb, 300 mL/min of pure hydrogen gives an adult roughly 1% inhaled hydrogen.

Do hydrogen therapy machines work?

They do make hydrogen, but whether breathing it improves health is unproven. Trials were small and short and results were mixed: some reported benefits on symptoms or lab markers, while others, including trials in Parkinson's disease and infant epilepsy, found none. No machine is proven to treat or prevent a disease.

Sources

  1. Botek M, et al. Molecular hydrogen positively affects physical and respiratory function in acute post-COVID-19 patients: a new perspective in rehabilitation. International Journal of Environmental Research and Public Health. 2022. PubMed · Our summary
  2. LeBaron TW, et al. Respiratory-physiology modeling of therapeutic hydrogen inhalation: defining the fraction of inspired hydrogen (FiH₂) and flow-rate requirements. Respiratory Research. 2026. PubMed · Our summary
  3. Zheng ZG, et al. Hydrogen/oxygen therapy for the treatment of an acute exacerbation of chronic obstructive pulmonary disease: results of a multicenter, randomized, double-blind, parallel-group controlled trial. Respiratory Research. 2021. PubMed · Our summary
  4. Guan WJ, et al. Hydrogen/oxygen mixed gas inhalation improves disease severity and dyspnea in patients with coronavirus disease 2019 in a recent multicenter, open-label clinical trial. Journal of Thoracic Disease. 2020. PubMed · Our summary
  5. Liu B, et al. The effect of a low dose hydrogen-oxygen mixture inhalation in midlife/older adults with hypertension: a randomized, placebo-controlled trial. Frontiers in Pharmacology. 2022. PubMed · Our summary
  6. Gao YH, et al. Effect of hydrogen-oxygen inhalation on sleep disorders and abnormal mood: a single-blind, randomized controlled trial. Medical Gas Research. 2025. PubMed · Our summary
  7. Chen C, et al. A randomized controlled double-blind study on the brain protection of infantile patients with epileptic spasm syndrome through atomized inhalation of hydrogen-oxygen gas. Frontiers in Neurology. 2026. PubMed · Our summary
  8. Ichikawa Y, et al. Guidelines for the selection of hydrogen gas inhalers based on hydrogen explosion accidents. Medical Gas Research. 2023. PubMed · Our summary
  9. Cole AR, et al. Safety of prolonged inhalation of hydrogen gas in air in healthy adults. Critical Care Explorations. 2021. PubMed · Our summary
  10. Tamura T, et al. Efficacy of inhaled hydrogen on neurological outcome following brain ischaemia during post-cardiac arrest care (HYBRID II): a multi-centre, randomised, double-blind, placebo-controlled trial. eClinicalMedicine. 2023. PubMed · Our summary
  11. Ono H, et al. Hydrogen gas inhalation treatment in acute cerebral infarction: a randomized controlled clinical study on safety and neuroprotection. Journal of Stroke and Cerebrovascular Diseases. 2017. PubMed · Our summary
  12. LeBaron TW, et al. Electrolyzed-reduced water: Review II: safety concerns and effectiveness as a source of hydrogen water. International Journal of Molecular Sciences. 2022. PubMed · Our summary
  13. Yoritaka A, et al. Randomized double-blind placebo-controlled trial of hydrogen inhalation for Parkinson's disease: a pilot study. Neurological Sciences. 2021. PubMed · Our summary
  14. US Department of Energy (College of the Desert). Hydrogen Fuel Cell Engines and Related Technologies, Module 1: Hydrogen Properties (Rev 0, December 2001). energy.gov
  15. NOAA Office of Response and Restoration. Hydrogen — chemical datasheet. CAMEO Chemicals. cameochemicals.noaa.gov
  16. US Department of Energy, Hydrogen and Fuel Cell Technologies Office. Safe use of hydrogen. energy.gov
  17. Katsumata Y, et al. The effects of hydrogen gas inhalation on adverse left ventricular remodeling after percutaneous coronary intervention for ST-elevated myocardial infarction — first pilot study in humans. Circulation Journal. 2017. PubMed · Our summary
  18. Ichikawa Y, et al. Preventable in-body hydrogen explosions from high-concentration H₂ inhalers in Japan — switch to safe, low-concentration hydrogen therapy. International Journal of Risk & Safety in Medicine. 2026. PubMed · Our summary
  19. Lin HY, et al. A narrative review of hydrogen-oxygen mixture for medical purpose and the inhaler thereof. Medical Gas Research. 2020. PubMed · Our summary
  20. Grepl P, et al. Sixty-minute inhalation of molecular hydrogen decreases blood oxygen saturation but does not alter autonomic cardiac regulation at rest in healthy females: a randomized, double-blind, placebo-controlled crossover study. Canadian Journal of Physiology and Pharmacology. 2026. PubMed · Our summary
  21. US Food and Drug Administration. Are there "FDA registered" or "FDA certified" medical devices? How do I know what is FDA approved? fda.gov

How we wrote this: drafted with AI assistance from the sources above and checked against each source by the Hydrogen Studies editorial team. Accident and fire-safety figures come from US government sources and the papers cited. This is general information, not medical or safety-engineering advice — follow your device maker's instructions and talk to your clinician before using hydrogen for a health condition.