Do essential oil diffusers pollute your indoor air?

An ultrasonic diffuser shakes oil and water into a fine mist; heat diffusers warm the oil so it evaporates. Either way, the oil's volatile terpenes (limonene, pinene, linalool, eucalyptol) enter room air.

What comes out. In a chamber test, 15 minutes of ultrasonic diffusing released milligrams of the oil's main terpene, such as 2.6 mg of d-limonene from lemon oil, plus particles 10 to 100 nanometres wide. Particle output ranged from about 0.1 to 3 mg an hour by oil, and tap water instead of deionized water gave five times more¹. In an office, limonene rose about 3-fold and the pinenes 13- to 14-fold within two hours of switching on a diffuser, then declined². A home simulation with tea tree oil found terpene levels from a heat diffuser above long-term exposure limits, while an ultrasonic one stayed below them⁹. Another lab measured only 0.6 ppb of oil vapour from an ultrasonic unit, soon falling below 0.1 ppb¹⁰. Output varies with the device, oil, water and room.

What happens next. Terpenes react with ozone to form formaldehyde and strong airway irritants⁶.

Who is affected. In 29 adults with asthma, inhaling perfume reduced lung function compared with saline³. That was perfume, not a diffuser, but it shows scent can trigger airways. In 59 students, diffuser emissions made responses faster but more impulsive, with worse memory sensitivity. Lemon oil and unscented grapeseed oil had similar effects, which points to particles rather than scent¹¹.

Do they help? An overview of systematic reviews found most were poor quality and the evidence not convincing for any condition⁴. Newer meta-analyses are more positive: 20 trials (1,717 people) found aromatherapy cut preoperative anxiety⁵, and 34 studies found it improved sleep, most for lavender¹². See lavender for that oil.

Ranked ways to cut the risk

1. Skip it, or keep sessions short; the anxiety trials found 20 minutes or less worked better⁵.
2. Ventilate. Emissions peak when a diffuser starts, so open a window².
3. Use distilled or deionized water in an ultrasonic unit¹.
4. Do not pair it with an ozone-generating air cleaner⁶.
5. Keep it out of infants' rooms and away from pets, and keep bottles out of reach^7,8.

More on oils themselves: essential oil safety and candles and air fresheners.

Emissions. Schwartz-Narbonne and colleagues ran an ultrasonic diffuser with lemon, lavender, eucalyptus and grapeseed oils in a chamber. Fine-particle (PM1) emission rates were about 2 mg per hour for lemon oil, 0.1 for lavender and eucalyptus, and 3 for grapeseed, with particles mainly 10 to 100 nm across for the essential oils. Tap water raised PM1 fivefold for eucalyptus oil, and modelling suggested ordinary use can add substantially to indoor primary and secondary particles¹. Kim and Kwon ran one commercial diffuser in an office for two weeks. Emissions were highest at the start and fell over time. Within 2 hours limonene rose from 1.8 to 5.8 µg/m³, α-pinene from 0.17 to 2.2 and β-pinene from 0.12 to 1.7. An unidentified compound spiked up to 698-fold, though the authors could not exclude sampling artifacts². Zisook and colleagues simulated a household and compared ultrasonic and heat diffusers with tea tree oil. Total VOCs ran about one order of magnitude above a 200 to 300 ppb background with the ultrasonic unit and two with the heat diffuser. With the heat diffuser, α-pinene and several other terpenes exceeded long-term derived no-effect levels, and α-pinene topped the European INDEX project limit. Terpene levels with the ultrasonic unit were about ten times lower and under the limits⁹. By contrast, Itoh and colleagues measured total VOCs from four oils in room air diffused by an ultrasonic unit at only 0.6 ppb, falling below 0.1 ppb¹⁰. All of these are single-device, single-room experiments.

Chemistry. Wolkoff and colleagues used a mouse airway-irritation test. Mixtures of ozone with α-pinene or limonene, each well below its own no-effect level, cut the respiratory rate by 30 to 50%, and the identified irritants could not explain it, so strong unidentified irritants probably formed. The authors said the effects of ozone-emitting equipment should be re-evaluated⁶.

Human responses. Kumar and colleagues challenged 29 adults with asthma and 13 without using perfume strips, perfume-impregnated paper, isopropyl alcohol and saline. Perfume caused significant FEV1 falls in the asthma group, chest tightness and wheeze in 20.7%, and exacerbations in 36%, 17% and 8% of patients with severe, moderate and mild asthma³. Du and colleagues randomized 59 students, double-blind, to four scenarios in a simulated office: lemon oil versus deionized water, and two particle levels from unscented grapeseed oil. Diffuser emissions shortened reaction time but worsened response inhibition and memory sensitivity, and lemon and unscented oil acted alike¹¹. Neither study tested people with asthma using a diffuser at home.

Benefits. Lee and colleagues found ten systematic reviews of aromatherapy, mostly of poor quality, and judged the evidence unconvincing for any condition⁴. Guo and colleagues pooled 20 randomized trials of 1,717 adults: aromatherapy lowered preoperative anxiety (standardized mean difference −0.57, 95% CI −0.75 to −0.39), and sessions of 20 minutes or less worked better than longer ones, which argues against leaving a diffuser running⁵. Cheong and colleagues pooled 34 studies of aroma inhalation for sleep problems and reported a benefit, largest for single lavender oil¹².

Harms. Posadzki and colleagues reviewed 42 case reports and series covering 71 patients. Effects ranged from mild to severe and included one death; dermatitis was the most common, and lavender, peppermint, tea tree and ylang-ylang were most often responsible. How often harm happens is unknown⁸. For pets, Khan and colleagues reviewed 443 cases of exposure to 100% tea tree oil (337 dogs, 106 cats), mostly on the skin or by mouth. Drooling, lethargy, weakness, poor coordination and tremors began within 2 to 12 hours and lasted up to 72, and younger, lighter cats were at greater risk⁷. Those doses were far larger than diffused vapour, and no study has measured diffusing near pets.

What does not line up. Gas-phase levels ranged from under 1 ppb¹⁰ to above exposure limits⁹, so device type matters. The asthma data come from perfume, not diffusers³. The cognitive effect was the same with and without scent, so it may owe more to particles than to the oil¹¹. And the benefit meta-analyses^5,12 sit beside a low-quality overview⁴.

Bottom line (emerging evidence). Diffusers measurably add terpenes and ultrafine particles, mostly modestly, and the gaps in health evidence are wide. Any calming benefit is small and uncertain, so use one briefly and ventilate, or skip it if you have asthma, a baby, or cats.