OAK Liquid Tree — Algae + HEPA Air Purifier for Offices & Schools
Clean air you can measure, in rooms up to 1,000 sq ft
OAK puts two engines in one 1.6-metre column. An H13 HEPA filter takes particles out of the air. Sixty litres of living spirulina take up carbon dioxide, which no filter can do. Sensors on the unit log CO₂, PM2.5, PM10, temperature and humidity to a cloud dashboard, so you can see the effect and export it for ESG or facility reporting.
Buying for an office, school or clinic? Book a 15-minute demo or ask for a quote for several units. We are happy to introduce you to our reference sites, including Gammon Construction, HKUST, Hong Kong Metropolitan University and ISF Academy.
What OAK does
- Removes 99%+ of PM2.5 and PM10 through its H13 HEPA stage, at a 380 m³/h clean air delivery rate: about 1.9 air changes an hour in an 800 sq ft office.
- Takes up CO₂ biologically: about 20.6 kg a year, roughly what one mature tree absorbs (the USDA Forest Service benchmark is 22 kg a year).
- Reduces odours and VOCs through an activated-carbon layer and the algae culture.
- Shows its work: an on-unit screen plus a cloud dashboard with full history and CSV export.
- Needs no installation: it stands on castors and plugs into a wall socket. No ducting, no building works, below 35 dB.
Where it fits
Offices and meeting rooms, classrooms and libraries, clinic waiting areas, consulting rooms and pharmacies, lobbies and receptions, and large living rooms.
OAK is a building air-quality appliance, not a medical device. It is not intended for operating theatres, isolation rooms or other spaces where air handling is part of a validated clinical process.
Specifications
| Dimensions | 400 × 400 × 1,600 mm, freestanding on castors |
| Culture | 60 L live spirulina (Arthrospira platensis) in a sealed bioreactor |
| Filtration | H13 HEPA with activated-carbon layer, 380 m³/h CADR (True HEPA option) |
| Coverage | Up to 1,000 sq ft (93 m²). Busy spaces of that size may need two units for particle clearance |
| Sensors | CO₂, PM2.5, PM10, temperature, humidity |
| Data | On-unit screen; cloud dashboard with history and CSV export |
| Noise | Below 35 dB |
| Power | 90–120 W on a standard 220–240 V, 50 Hz socket. Ask us about 110 V markets |
| Care | A 5-minute refill each month; HEPA filter replaced once a year |
| Warranty | 12 months from commissioning |
Smoke-chamber test results
We sealed a room of 20–30 m³, taped every gap and vent, then filled it with smoke and CO₂ until every sensor read its maximum. Six runs plus controls, with dual sensors logging PM2.5, PM10, CO₂, formaldehyde, benzene, TVOC, AQI, temperature and humidity roughly every five seconds. Tested 3 July 2026. This is our own data, not a third-party certification; raw logs are available on request.
| Pollutant | Peak (smoke) | After OAK | Reduction | Guideline |
|---|---|---|---|---|
| PM2.5 | 999 µg/m³ | 13 µg/m³ | −98.7% | 15 µg/m³, 24-hour (WHO 2021) |
| PM10 | 999 µg/m³ | 18 µg/m³ | −98.2% | 45 µg/m³, 24-hour (WHO 2021) |
| Formaldehyde | 0.374 mg/m³ | 0.008 mg/m³ | −97.9% | 0.1 mg/m³, 30-minute (WHO) |
| TVOC | 0.94 mg/m³ | 0.03 mg/m³ | −96.8% | 0.5 mg/m³ (common IAQ target) |
| CO₂ | 2,016 ppm | 441 ppm | −78.1% | ~430 ppm outdoor air |
Roughly eleven minutes from ignition to clean air, and about three minutes once the smoke source stopped. In the control run with OAK switched off, the same smoke load stayed pinned at the sensor ceiling for the whole test.
Why the algae matters
A HEPA filter traps particles. It cannot remove CO₂. OAK's algae bioreactor can, and the test runs separate the two jobs cleanly:
- Algae only (bubbler on, fan off, no smoke): CO₂ drawn down to 463 ppm while particle levels stayed flat.
- No device, elevated CO₂: CO₂ held at around 2,100 ppm with no drawdown at all.
- Full unit under smoke load: particles fell almost vertically as the HEPA stage caught up, while CO₂ and combustion gases declined steadily: two engines working in parallel, each on the pollutants it is built for.
Chamber results vs real rooms
The figures above come from a sealed chamber, deliberately loaded far beyond anything a real room would see. That is what makes the mechanism visible, but a real room is a different problem. In an open space of roughly 800–1,000 sq ft, with normal air exchange and occupancy, expect up to 30% reduction in indoor CO₂. Both numbers are real; they describe different conditions, and we would rather show you both than quote only the flattering one.
About these figures: several sensors saturate at their ceilings (PM2.5 at about 999 µg/m³), so the real smoke load was higher than the peaks shown and the particle reductions above are conservative.
