CO2 enrichment: feeding the plant carbon, safely
HomeEnvironment & climateCO2 enrichment: feeding the plant carbon, safely
Environment & climate · CO2

CO2 enrichment: feeding the plant carbon, safely

Carbon dioxide is the raw material a plant turns into sugar. Under strong light an indoor crop can use far more CO2 than the air provides, so adding it can lift yield by around a third. But CO2 only helps in the light, it is easy to waste, and at the wrong concentration it will erase the plants' returns or kill the people in the room. This is how it works, how much to add and when, and how to do it without hurting anyone.

Core concept to safe setup7 diagramsPeer-reviewed + safety standards · 38 sources~30 min read
01 · Read this first

CO2 is food, but only in the light

A plant builds itself out of air. The carbon in every leaf, stem and flower comes from carbon dioxide (CO2) pulled out of the room and welded to water using the energy in light. Give a bright canopy more CO2 than the ~420 ppm in normal air and it can build faster, which is why growers add it. But CO2 enrichment is one of the easiest things in the room to get dangerously wrong.

The five things this paper comes down to
  • CO2 is only used in the light. In the dark the plant does the opposite, it breathes CO2 out. Injecting at night is pure waste and a safety risk.
  • It only pays under strong light. CO2 and light limit each other. Under weak light, adding CO2 does almost nothing[1].
  • There is a target band, roughly 1,000–1,500 ppm. Below it you leave yield on the table, above it you waste gas for no gain[6].
  • You have to seal the room. In a vented room the CO2 you pay for is blown straight outside[12].
  • It can kill people. The gas that grows the plant is odourless and, at high concentration, lethal, and a sealed enrichment room or a drying room can reach that concentration[16].

Everything below builds on those five. If you only take one thing away: enrich in the light, seal the room, and put a CO2 alarm on the wall.

02 · The vocabulary

The words you need

CO2 (carbon dioxide)The gas plants turn into sugar using light. Normal outdoor air is about 420 ppm today[3].
ppmParts per million. 1,000 ppm means one part CO2 in a thousand of air (0.1%). All the numbers in this paper are in ppm, and 10,000 ppm = 1%.
PhotosynthesisThe light-powered reaction that combines CO2 and water into sugar. This is the only process that pulls CO2 into the plant, and it needs light to run.
RespirationThe plant burning that sugar to live and grow, which releases CO2 back out. It runs day and night, in every living cell[8].
C3 plantCannabis is a C3 plant, the plant type whose photosynthesis is held back by today's CO2 level, which is exactly why enrichment helps it.
PhotorespirationA wasteful side-reaction where the plant's carbon-fixing enzyme grabs oxygen instead of CO2. More CO2 (and cooler leaves) starves this waste out[2].
Compensation pointThe CO2 level (~50 ppm for a C3 plant) at which photosynthesis exactly cancels respiration and the plant makes no net gain[2].
Saturation pointThe CO2 level above which extra CO2 stops helping. For enrichment this sits near 1,000–1,500 ppm.
NDIR sensorNon-dispersive infrared, the only cheap sensor type that actually measures CO2 correctly. Cheap ‘MQ’ gas sensors do not[22].
TWA / STEL / IDLHWorker-safety limits: the 8-hour average you may breathe (TWA), the 15-minute short spike (STEL), and the level that is Immediately Dangerous to Life or Health (IDLH)[18].
03 · The mechanism

How CO2 becomes flower

Inside a lit leaf, an enzyme called Rubisco grabs CO2 out of the air and fixes it into sugar. The trouble is Rubisco is sloppy: it also grabs oxygen by mistake, and when it does it kicks off photorespiration, a reaction that burns energy and throws carbon away. The more CO2 there is relative to oxygen, the more often Rubisco does its real job instead of the wasteful one[2].

Today's air, at ~420 ppm CO2[3], is not enough to keep Rubisco busy. A C3 plant like cannabis is CO2-limited: raise the CO2 and photosynthesis climbs. In the one classic cannabis gas-exchange study, lifting CO2 from 350 to 750 ppm raised leaf net photosynthesis by about 50%[1], and a later study across four high-THC cultivars found gains of 38–48% going from 390 to 700 ppm[4].

More CO2, more photosynthesis (cannabis leaf)Cannabis leaf net photosynthesis, relative to the 350 ppm baseline = 100. Starve it of CO2 and it halves; enrich it and it climbs by half.0428512817050250 ppm100350 ppm150750 ppm
Figure 1. Cannabis leaf net photosynthesis against CO2. Dropping to 250 ppm cuts it ~50%; raising to 750 ppm lifts it ~50%[1]. The response is steep at first, then flattens as other limits take over.

The climb does not go on forever. As CO2 rises, photosynthesis stops being limited by CO2 and starts being limited by how fast the leaf can use the light, so the curve flattens into a saturation point. For most greenhouse C3 crops that plateau is around 1,000–1,300 ppm[7].

The rule that decides whether CO2 is worth anything

CO2 and light limit each other, so the weaker one caps the plant (Liebig's law of the minimum). Adding CO2 under weak light is like flooring the accelerator with the handbrake on, nothing happens. Cannabis flower yield keeps climbing with light all the way to very high intensity[5], so CO2 only earns its keep once the light is already strong.

04 · Day and night

The plant breathes both ways

Here is the fact that trips up beginners. Photosynthesis, the CO2-consuming reaction, only runs in the light. Respiration, the CO2-releasing one, never stops, day and night, in every living cell[8]. During the light period the big inward pull of photosynthesis swamps the small outward push of respiration, so the canopy is a net CO2 sink. When the lights go off, photosynthesis stops dead and only respiration is left, so the same canopy flips to a net CO2 source.

Day and night: the plant breathes both waysSchematic canopy net CO2 exchange, lights on 06:00-18:00. Below zero = net uptake (light); above zero = net release (dark).-24-16-8080h06121824hnet CO2 flux
Figure 2. The canopy's net CO2 flux over a day. In the light it pulls CO2 in hard; in the dark it gives a little back. The deep daytime trough dwarfs the shallow night-time bump, and that gap is the whole point of the next section.

Respiration also speeds up when it is warm, roughly doubling for every 10 °C rise (a Q10 of about 2)[10], and in soil or coco there is extra CO2 from microbes and roots in the root zone. But none of that changes the headline: the plant only takes CO2 in while the lights are on. That single fact is why you inject in the light and never in the dark.

05 · The mass balance

Day use vs night release: they don't balance

Growers often ask: if the plant uses CO2 all day and gives it back all night, doesn't it even out? No, and the reason it doesn't is the reason enrichment works. The carbon a plant fixes in the light does not all get burnt again. Across a season only about half of the carbon captured by photosynthesis is respired back out; the rest, roughly 46%, is locked up as sugar, stem, leaf and flower[9] (the respired fraction is put at 30–60% by crop-physiology reviews[8]). That retained carbon is the crop. So the CO2 pulled in during the day always exceeds the CO2 breathed out at night, by the amount that became plant.

You can see the asymmetry in a sealed room. With the lights on and no CO2 added, a hungry canopy strips the air below ambient fast: a tightly sealed glasshouse can fall to ~200 ppm within hours, which itself cuts canopy photosynthesis by about 26%[11]. Overnight the same room drifts back up as the plants respire, but the rise is gentle and leakage caps it.

A sealed room with no added CO2, over 24 hoursSealed room, no CO2 added. Lights on 06:00. Photosynthesis crashes CO2 toward ~200 ppm; overnight respiration lifts it back over ambient.01803605407200h06121824hroom CO2 (ppm)
Figure 3. Room CO2 in a sealed, unenriched flower room. The lights-on crash is deep and fast; the overnight recovery is slow and modest. Enrichment exists to fill that daytime hole[11].
The numbers, for a mid-size room
  • Lights-on drawdown. A ~45 m² canopy pulling 15–25 µmol CO2 m²/s strips a sealed ~160 m³ room at roughly 350–600 ppm per hour at first, taking 420 down to ~200 ppm in 20–40 minutes, then slowing as CO2 runs low.
  • Lights-off rise. The same canopy respiring in the dark adds only about 25–70 ppm per hour, a few hundred ppm over the whole night before leakage stops it.
  • The gap is the crop. The daytime pull is many times the night-time push, because about half the fixed carbon stays in the plant instead of being breathed back out[9].

This is the deeper answer to ‘how much do I need?’. Enrichment is not topping up a tank that empties overnight, it is holding CO2 up during the light hours against a canopy that would otherwise strip the room bare.

06 · The dose

How much, and what it buys you

The best cannabis-specific yield data come from controlled work at Utah State University: raising a sealed room from ambient (~420 ppm) to about 1,200 ppm lifted dry flower yield by roughly 40%, and 1,200 ppm captured about 95% of the achievable gain, so there is little point going higher[6]. General greenhouse crops tell the same story: tomato yield rises up to ~80% at 1,000 ppm[7], and most C3 crops saturate around 1,000–1,300 ppm[7].

Where to hold CO2 while the lights are onHold roughly 1,000-1,500 ppm during the light period. Below ambient the plant starves; above ~1,500 ppm you pay for gas the plant can't use.sub-ambient: starvedambient: gains beginenrichment targetdiminishing / wasteful200 ppm1100 ppm2000 ppm
Figure 4. The working band. Aim for roughly 1,000–1,500 ppm in the light. The plateau above ~1,500 ppm is real, so pushing to 2,000 ppm mostly buys wasted gas and a bigger safety and cooling burden[6].
What CO2 does and doesn't change

CO2 is a yield lever, not a potency lever. It grows more flower mass, so grams of cannabinoid per plant go up, but it does not meaningfully raise %THC or %CBD, and its effect on terpenes is minor. Expect a bigger harvest, not stronger flower.

About the ‘39% biomass, 43% flower’ figures you'll see quoted

Those exact numbers, and most ‘20–40% uplift’ claims, circulate on vendor blogs without a controlled trial behind them. The trustworthy primary source is the Utah State work above, which lands around a ~40% flower-yield gain at 1,200–1,400 ppm[6]. Treat anything more precise from a seller as marketing until you see the study.

07 · Getting it in

Delivery and dosing

There are four common ways to put CO2 in a room, and for a sealed indoor cannabis room one of them clearly wins:

For a sealed licensed room, compressed CO2 on a sensor-driven controller is the standard choice.
MethodHow it worksWatch out for
Compressed / bottled CO2Cylinder → regulator → solenoid valve, gated by an NDIR controllerCleanest option, no heat or byproducts, precise. Preferred for sealed rooms. Cylinders run out, so plan swaps
Propane / gas burnerBurns fuel to make CO2 (~3 lb CO2 per lb fuel)Cheap gas, but adds heat and water vapour, and a poor flame makes carbon monoxide, ethylene and NOx that harm plants and people. Needs a CO alarm
Fermentation / yeastSugar + yeast → CO2 + alcoholTiny, uncontrolled output. Hobby scale only
Dry iceSolid CO2 sublimes into gasUncontrolled, short-lived, and a cold-burn / confined-space hazard
The dosing formula

To raise a sealed room, work in milligrams: mass of CO2 (mg) = room volume (m³) × rise wanted (ppm) × 1.8. The 1.8 is the mass of CO2 in a cubic metre per ppm at room temperature (CO2 is 1.799 g/L at 25 °C, so 1 ppm = 1.8 mg/m³)[18]. Worked example: a 30 m³ room from 420 to 1,200 ppm (a 780 ppm rise) needs 30 × 780 × 1.8 = 42,120 mg ≈ 42 g of CO2, about 23 litres of gas. That is only the one-time charge, a room that leaks needs continuous top-up on top of it.

Timing and distribution. Inject only during the light period, typically starting about an hour after lights-on and stopping before lights-off[12]. Drive it from an NDIR CO2 sensor at canopy height holding a setpoint, not a blind timer. Because CO2 is denser than air it sinks, so run the supply line above the canopy and use horizontal airflow (HAF) fans to mix it down and thin the still, CO2-starved layer that forms right at the leaf[14].

Closed-loop CO2 control, lights-on only1NDIR sensorreads CO2 atcanopy2Controllercompares tosetpoint3Solenoidopens if belowtarget4InjectionCO2 from thecylinder5HAF fansmix down to theleafThe loop runs only during the photoperiod. In the dark it stays shut, the plant isn't using CO2.
Figure 5. A sensor-driven enrichment loop. The controller injects to hold the setpoint during the light period and shuts off in the dark.
08 · Holding it in

Sealed vs vented: why the room matters more than the gas

Here is where most CO2 money is wasted. An exhaust fan replaces room air with ~420 ppm outdoor air, so any CO2 you have added decays straight back toward ambient. The loss follows a simple curve: the richer the room and the faster the air changes, the faster the CO2 leaves. Run an exhaust fan and the injector at the same time and you are heating money and blowing it out the wall.

Even a room with the fans off leaks. In measured commercial greenhouses, more than half of the injected CO2 was lost to structural leakage, with CO2 use efficiency below 50–60% at a 1,000 ppm setpoint[12]. That is why enrichment demands a sealed room: instead of exhausting for heat and humidity, you handle them inside the room, with a recirculating mini-split air conditioner for heat and a standalone dehumidifier for moisture, so the door stays shut and the CO2 stays in.

Enrichment and exhaust ventilation are opposites. Pick a sealed design before you buy a gram of CO2.
Vented roomSealed room
Heat / humidityRemoved by exhausting air outsideRemoved in-room by AC + dehumidifier
CO2 enrichmentPointless, blown out with the exhaustHolds, so enrichment works
Air exchangeHigh, many changes per hourLow, only leakage
When to use itNo CO2, or CO2 only to prevent daytime depletionAny serious enrichment programme
Purge in the dark

You still need to flush the room sometimes, to reset humidity or clear stale air. Do it at lights-off, when the plants aren't using CO2 anyway, so you never vent gas you just paid to inject. CO2 pools low, so a floor-level exhaust clears it fastest.

09 · Fresh air

Stale air: why a sealed room still needs a purge

You seal the room to hold CO2 in, and then you're told to let fresh air in. That sounds like a contradiction, and it is where the filter question comes up: if carbon and HEPA are scrubbing the air non-stop, why exchange it at all? Because filters clean the air, they do not refresh it. What makes air ‘stale’ is mostly the plants' own invisible waste gases, plus their heat and moisture, and a carbon-and-HEPA stack removes almost none of that.

Start with what those filters actually catch, and what they miss entirely:

Neither filter refreshes the air. Between them they miss every gas that builds up in a sealed room, plus the heat and the moisture.
FilterRemovesDoes NOT touch
HEPAParticulates ≥0.3 µm: mould spores, dust, pollen (99.97%)[25]Any gas at all, water vapour, heat, CO2, ethylene, oxygen
Activated carbonOdour and heavier VOCs / terpenes, by adsorptionCO2, humidity, heat, oxygen; and it grips ethylene only weakly and saturates[26]

The gas that forces the issue is ethylene (C2H4). It is a plant hormone the plants make and breathe out themselves, a little all the time and a lot more when they are stressed, wounded, defoliated or flowering[30]. And plants react to it at almost unbelievably low levels: sensitive crops show damage at 10 ppb (0.01 ppm), and greenhouse practice is to hold it under about 20 ppb[28][29].

In a sealed space it builds up from the plants alone. NASA sealed a chamber of wheat, soybean, lettuce and potato and watched ethylene climb to 40–120 ppb purely from the plants' own breath, enough to visibly deform the wheat[30]. In a controlled tomato study a steady 20 ppb cut fruit yield to about half, and 40 ppb cut it by roughly 90%, while the foliage still looked nearly normal[31]. Cannabis is not exempt: it has a fully working ethylene-signalling system, so responsive that ethylene is what flips flower sex[32].

Ethylene: damage starts at parts per billionPlants respond at 10 ppb and a sealed room self-generates 40-120 ppb. Carbon and HEPA don't reliably remove it - air exchange or a KMnO4 scrubber does.clean airkeep below ~20 ppbchronic: stunting, bud abortionacute: epinasty, abscission0 ppb200 ppb400 ppb
Figure 7. The ethylene damage scale, in parts per billion. A sealed room of plants climbs into the chronic-damage band on its own[30][31].
Why the filters don't save you here

Ethylene is a tiny, light molecule, about 0.4 nanometres across. HEPA cannot catch a gas at all, and activated carbon holds ethylene only weakly and fills up fast[26]. A carbon-and-HEPA stack can run all day and ethylene will still creep up from the plants faster than the carbon retains it. The only reliable fixes are to exchange the air or fit a dedicated ethylene scrubber, potassium-permanganate media or a UV / TiO2 photocatalytic unit, which chemically destroy ethylene instead of briefly holding it[27].

The oxygen myth

Growers often think they purge ‘so the plants get oxygen.’ They do not need it from you. Room air is ~21% oxygen (about 209,000 ppm); the level even flagged for human workers is 19.5%, which means losing ~14,500 ppm of oxygen, and a plant canopy removes nowhere near that[33]. Room-air oxygen is a non-problem. The oxygen that does matter is in the root zone, dissolved in the water and held in the substrate's air-filled porosity, and that is an irrigation and substrate question, not an air-exchange one. The one thing that genuinely lowers a room's oxygen and raises its CO2 is people working in it.

Ethylene is the sharpest reason to exchange air, but stale air is also carrying:

  • Humidity. Plants return most of their irrigation water to the air as vapour[34]. That latent load is removed by a dehumidifier, not a filter, and it keeps coming after lights-off when transpiration outruns the dehumidifier.
  • Heat. Removed by the air conditioner, not the carbon can.
  • The plants' own terpenes and other VOCs, which carbon catches for a while, then passes straight through once it saturates.
  • Mould spores. HEPA on the recirculation loop genuinely lowers the airborne spore count[38], but it only cleans the air that reaches it. Deep in a fat canopy the humidity runs 15–25% higher than your room sensor reads[35], and that stagnant, humid pocket is where bud rot starts. Exchanging air, and moving it through the canopy rather than just filtering the room average, is what lowers disease[36] (detail in the mould and airflow papers).
So how do you square it with holding CO2?
  • Purge at lights-off. The plants aren't using CO2 in the dark, so a timed fresh-air exchange during the dark period dumps the accumulated ethylene, VOCs and humidity without venting the CO2 you paid for.
  • Or scrub instead of vent. A potassium-permanganate or photocatalytic ethylene scrubber lets a genuinely sealed room hold its CO2 and still kill ethylene[27].
  • Don't confuse circulation with exchange. HAF fans mix the room and thin the boundary layer at the leaf, which helps CO2 reach the stomata[14], but they only stir the same air. Only fresh air (or a scrubber) changes what is in it.

How much exchange? Enough to hold ethylene and humidity down, not so much that you dump CO2 and dry the media out. Enclosed plant environments run anywhere from well under 1 to ~15 air changes an hour; a cannabis micropropagation study landed on about 4.4 changes an hour as optimal, with more than that drying the substrate and stressing the plants[37] (that was at plantlet scale, so treat it as a direction, not a flower-room setpoint). ‘Sealed’ never means ‘never exchanged.’

10 · The knock-on effects

CO2 changes everything else in the room

Raising CO2 does not act alone, it pulls on the rest of the room. This is the same ‘inputs travel in convoys’ idea from the grow-room systems paper. Change CO2 and three other things want to move with it:

Raise CO2, then ask what must move with it. The temperature and water changes are the ones growers most often miss.
LeverWhat elevated CO2 doesWhat to do about it
TemperatureShifts the plant's ideal temperature up a few degrees[13]Run warmer, ~28–30 °C instead of ~25 °C, if light supports it. Cannabis photosynthesis peaks near 30 °C[1]
Stomata & waterPartly closes the leaf pores: ~42% less stomatal conductance, ~29% less transpiration per leaf[1]Per leaf it drinks less, but a bigger, hotter canopy usually drinks more overall. Re-check your VPD target, irrigation volume and feed EC
LightNothing, unless light is already high[5]Only enrich rooms running strong light. Otherwise CO2 is wasted
FeedLess transpiration can let salts concentrate at the rootWatch runoff EC and adjust feed strength
Water use: efficiency goes up, volume usually goes up too

Elevated CO2 partly closes the stomata, so a leaf loses less water per unit of leaf for the same photosynthesis, and its water-use efficiency roughly doubles, about twice the growth per litre transpired[1]. But that is efficiency, not less water, and in a real enriched room the total usually moves the other way. CO2 grows a bigger, faster canopy with far more leaf area, and the room is run hotter and brighter to exploit it[13], which more than cancels the per-leaf stomatal closure. So in practice total water use and humidity load commonly go up, not down, most growers running CO2 end up feeding and dehumidifying more. Don't assume a wetter substrate, and re-tune irrigation by watching VWC, dryback and runoff on the actual crop rather than copying an old schedule.

11 · The hidden hazard

The drying room: CO2 with no light to burn it

A drying room is the dark side of everything above, literally. You fill a sealed, dark, low-airflow room with tens or hundreds of kilos of freshly cut biomass, and that biomass keeps respiring for days after harvest, releasing CO2 the whole time, with no photosynthesis to soak any of it back up[15].

Cut leaf and flower is a fast-respiring tissue, in the same postharvest class as leafy greens and cut herbs, and warmth speeds it up (that Q10-of-2 again)[15]. A drying room held at 15–18 °C with the air barely moving is a near-perfect CO2 trap. Sealed, planted greenhouses already climb to 600–1,000+ ppm overnight from living plants[11]; a room packed with cut biomass and almost no ventilation can go substantially higher, into the thousands of ppm and, in the worst unventilated cases, potentially toward percent-level concentrations.

Treat an unventilated drying room as a confined space

CO2 is heavier than air, so it pools low, exactly where a person stands, and you cannot smell it. A sealed drying room that has been shut overnight can hold a dangerous atmosphere at floor level. Ventilate it before anyone enters, fit a CO2 alarm, and never let someone walk into a closed, full drying room to ‘just check on it’ without air exchange running.

An honest limit on the numbers

No peer-reviewed study has directly measured CO2 inside a cannabis drying room. The mechanism (respiring biomass in the dark) and the human-safety limits below are rock solid, but the exact ceiling a given room reaches is an engineering inference, not a measured cannabis value. The right response is not to guess, it is to put a meter in your own drying room and find out. The full drying environment is covered in the harvest, dry, trim & cure paper.

12 · The part that can kill you

CO2 and people: the numbers that matter most

Read this section even if you skip the rest. Your enrichment setpoint of 1,000–1,500 ppm is well below any worker limit and safe for brief occupancy. The danger is not the setpoint, it is a leaking regulator, an overfilled room, a burner fault, or a sealed drying room, any of which can reach lethal CO2 while you notice nothing.

Enrichment sits far below the worker limitYour grow-room setpoint (green) is 3-5x below the 5,000 ppm worker exposure limit. The problem is never the setpoint, it's a leak that runs away past it.outdoor airenrichment targettoward the 5,000 ppm 8-hour worker limit0 ppm2500 ppm5000 ppm
Figure 6. The normal enrichment band against the occupational limit. Correctly run, enrichment is nowhere near the danger zone, which is exactly why a runaway leak is so easy to ignore until it isn't[17].

Every major safety authority agrees on the headline number: the 8-hour average a worker may breathe is 5,000 ppm (0.5%), from OSHA in the US[17], NIOSH and ACGIH[18], and WorkSafe New Zealand[19]. Above that, effects climb steeply:

The CO2 effect ladder. Note the jump: the enrichment band is safe, but the levels a leak can reach in a sealed room are not.
CO2 levelWhat happensReference point
~420 ppmNormal outdoor airBaseline[3]
1,000–1,500 ppmTypical enrichment target, safe for short work3–5× below the worker limit
1,000–2,500 ppmMeasurable dip in concentration and decision-makingCognition, not danger[20]
5,000 ppm (0.5%)8-hour worker exposure limit (TWA)OSHA / NIOSH / ACGIH / WorkSafe NZ[19]
30,000 ppm (3%)15-minute short-term limit (STEL); headache, fast breathingNIOSH / ACGIH / WorkSafe NZ[18]
40,000 ppm (4%)IDLH, immediately dangerous to life or healthEscape-impairing[18]
50,000 ppm (5%)Hypercapnia and respiratory acidosis within ~30 minDirect poisoning, not just low oxygen[16]
70,000–100,000 ppm (7–10%)Unconsciousness within minutesCannot self-rescue[16]
>300,000 ppm (>30%)Loss of consciousness in secondsFatal-scene cases 14–26% CO2[16]
CO2 is a poison, not just a smothering gas

It is tempting to think CO2 just crowds out oxygen. It does that, but it is also directly toxic: above ~5% it acidifies the blood, and forensic reviews conclude the cause of death in CO2 incidents is the poisoning itself, not the lack of oxygen[16]. That is why it drops people so fast they can't open a door, and why rescuers who rush in without protection are so often the second victims. A sealed, high-CO2 room is a confined space. Ventilate and test the air before entry, every time.

13 · Measuring and alarming

Sensors, alarms and the MQ trap

You cannot manage or survive CO2 you cannot measure, and the sensor market is full of parts that don't actually measure it. The correct technology is NDIR (non-dispersive infrared), which reads CO2 by its specific infrared fingerprint[22]. The cheap ‘MQ’ metal-oxide sensors sold as CO2 sensors are a trap:

If a ‘CO2 node’ is built on an MQ-135 or MQ-5, it is not measuring CO2. Only trust NDIR for control and for life safety.
SensorDoes it measure CO2?Verdict
NDIR (SCD30, SCD41, MH-Z19…)Yes, infrared, CO2-specificCorrect for both control and safety[22]
MQ-135No, it infers ‘equivalent CO2’ from a mix of VOCsNot a real CO2 reading, unsuitable[23]
MQ-5No, it's built for LPG / combustible gasWrong gas entirely[23]
The sealed-room calibration trap

NDIR sensors drift, so they self-calibrate by assuming the lowest CO2 they ever see is fresh 400 ppm air (Automatic Baseline Calibration). In a continuously enriched sealed room the sensor never sees 400 ppm, so ABC slowly mis-calibrates it downward[22]. Disable ABC and calibrate manually against fresh air or a reference gas on a schedule.

Two sensors, two jobs. Put a control sensor at canopy height to run the setpoint, and a separate life-safety sensor low on the wall (~30 cm off the floor) because CO2 settles. For larger systems the International Fire Code (Section 5307) makes this mandatory: any installation over 100 lb of CO2 needs gas detection that alarms at 5,000 ppm, alarms hard at 30,000 ppm, and automatically shuts off the CO2 and starts ventilation[21]. Even below that threshold, an alarm plus an occupancy interlock is cheap insurance.

14 · Straight talk

Is it worth it, and when it isn't

CO2 is one of the highest-return upgrades in a sealed, high-light room, and a waste of money in any other[24]. The gas itself is cheap; the room around it is what decides the payback.

worth it
Enrich when…
  • The room is sealed (AC + dehumidifier, not exhaust)
  • Light is already strong (high PPFD / DLI)[5]
  • Temperature, water and feed are dialled in
  • You have a CO2 alarm and interlock fitted
skip it
Don't bother when…
  • The room is vented, you'll blow it outside[12]
  • Light is low, CO2 can't be used[1]
  • You'd push past ~1,500 ppm chasing more[6]
  • You can't monitor it safely yet
The usual CO2 failures. Almost all of them are the room, the light or the sensor, not the CO2.
ProblemLikely causeFix
Added CO2, no extra yieldLight too low, or the room isn't sealedRaise PPFD and seal the room before blaming the gas
CO2 won't hold at setpointRoom leaks, or exhaust is running during injectionSeal leaks; only exhaust at lights-off
Plants wilting / substrate staying wetEnriched plants transpire less, old irrigation is now too muchRe-tune VPD and cut irrigation volume[1]
Leaf damage with a burnerIncomplete combustion making CO / ethylene / NOxService the burner, fit a CO alarm, or switch to bottled CO2
Sensor reads lower over weeksABC mis-calibrating in a sealed roomDisable ABC, calibrate manually[22]
The mindset
  • Light first, CO2 second. CO2 amplifies strong light; it can't replace it.
  • Seal before you enrich. An unsealed enrichment room is a money leak.
  • Measure with NDIR, alarm for people. Control and safety are two sensors, not one.
  • Enrich in the light, purge in the dark. The plant only breathes CO2 in while the lights are on.

Read the grow-room systems, airflow and harvest and dry papers alongside this one, CO2 only makes sense as part of that whole machine.

Related papers

References

  1. Chandra S, Lata H, Khan IA, ElSohly MA (2008). Photosynthetic response of Cannabis sativa L. to variations in photosynthetic photon flux densities, temperature and CO2 conditions. Physiol. Mol. Biol. Plants 14(4):299-306. https://pmc.ncbi.nlm.nih.gov/articles/PMC3550641/
  2. Tolbert NE, Benker C, Beck E (1995). The oxygen and carbon dioxide compensation points of C3 plants: possible role in regulating atmospheric oxygen and carbon dioxide concentrations. Proc. Natl. Acad. Sci. USA 92(24):11230-11233. https://doi.org/10.1073/pnas.92.24.11230
  3. NOAA Global Monitoring Laboratory (2024). Trends in atmospheric carbon dioxide, global monthly/annual means (2024 global mean 422.8 ppm). (non-peer-reviewed source) https://gml.noaa.gov/ccgg/trends/global.html
  4. Chandra S, Lata H, Khan IA, ElSohly MA (2011). Photosynthetic response of Cannabis sativa L., an important medicinal plant, to elevated levels of CO2. Physiol. Mol. Biol. Plants 17(3):291-295. https://doi.org/10.1007/s12298-011-0066-6
  5. Rodriguez-Morrison V, Llewellyn D, Zheng Y (2021). Cannabis yield, potency, and leaf photosynthesis respond differently to increasing light levels in an indoor environment. Front. Plant Sci. 12:646020. https://pmc.ncbi.nlm.nih.gov/articles/PMC8144505/
  6. Westmoreland FM (2023). Environmental Physiology of Medical Cannabis. PhD dissertation, Utah State University (Bugbee Crop Physiology Lab): CO2 enrichment to 1,200-1,400 ppm raised dry flower yield ~40%. (non-peer-reviewed source) https://digitalcommons.usu.edu/etd2023/243/
  7. Doddrell NH, Lawson T, Raines CA, Wagstaff C, Simkin AJ (2023). Feeding the world: impacts of elevated [CO2] on nutrient content of greenhouse-grown fruit crops and options for future yield gains. Horticulture Research 10(4):uhad026. https://doi.org/10.1093/hr/uhad026
  8. Amthor JS (2024). After photosynthesis, what then? Importance of respiration to crop growth and yield. Field Crops Research 321:109638. https://www.sciencedirect.com/science/article/pii/S0378429024003915
  9. Collalti A, Prentice IC (2019). Is NPP proportional to GPP? Waring's hypothesis 20 years on. Tree Physiology 39(8):1473-1483 (mean NPP:GPP ~0.46). https://doi.org/10.1093/treephys/tpz034
  10. Atkin OK, Tjoelker MG (2003). Thermal acclimation and the dynamic response of plant respiration to temperature. Trends in Plant Science 8(7):343-351 (respiration Q10 ~2). https://doi.org/10.1016/S1360-1385(03)00136-5
  11. Agriculture & Horticulture Development Board (AHDB). CO2 enrichment: best-practice guide (a sealed glasshouse can fall to ~200 ppm, cutting canopy photosynthesis ~26%). (non-peer-reviewed source) https://horticulture.ahdb.org.uk/knowledge-library/co2-best-practice-guide-background
  12. Wang X, Lv J, Shi X, et al. (2022). CO2 enrichment in greenhouse production: towards a sustainable approach. Frontiers in Plant Science 13:1029901 (CO2 use efficiency below ~50-60% from leakage). https://doi.org/10.3389/fpls.2022.1029901
  13. Lv Z, Zhu Y, Liu X, et al. (2022). Elevated [CO2] raises the temperature optimum of photosynthesis and thus promotes net photosynthesis of winter wheat and rice. Physiologia Plantarum 174(6):e13757. https://doi.org/10.1111/ppl.13757
  14. Kitaya Y, Shibuya T, Yoshida M, Kiyota M (2004). Effects of air velocity on photosynthesis of plant canopies under elevated CO2 levels. Adv. Space Res. 34(7):1466-1469. https://doi.org/10.1016/j.asr.2003.08.031
  15. Kader AA, Saltveit ME (2002). Respiration and gas exchange. In Postharvest Technology of Horticultural Crops, 3rd ed. Univ. of California ANR Publication 3311. (non-peer-reviewed source) https://escholarship.org/uc/item/8q37j80t
  16. Permentier K, Vercammen S, Soetaert S, Schellemans C (2017). Carbon dioxide poisoning: a literature review of an often forgotten cause of intoxication in the emergency department. Int. J. Emerg. Med. 10:14. https://doi.org/10.1186/s12245-017-0142-y
  17. US OSHA. Annotated Permissible Exposure Limits, Table Z-1, carbon dioxide (CAS 124-38-9): PEL 5,000 ppm 8-hour TWA. (non-peer-reviewed source) https://www.osha.gov/annotated-pels/table-z-1
  18. US NIOSH / CDC. Pocket Guide to Chemical Hazards and IDLH documentation, carbon dioxide: REL 5,000 ppm TWA, STEL 30,000 ppm, IDLH 40,000 ppm; 1 ppm = 1.80 mg/m3. (non-peer-reviewed source) https://www.cdc.gov/niosh/idlh/124389.html
  19. WorkSafe New Zealand (2025). Workplace Exposure Standards and Biological Exposure Indices, 15th ed., carbon dioxide: WES-TWA 5,000 ppm, WES-STEL 30,000 ppm. (non-peer-reviewed source) https://www.worksafe.govt.nz/topic-and-industry/monitoring/workplace-exposure-standards-and-biological-exposure-indices/all-substances/view/carbon-dioxide/
  20. Azuma K, Kagi N, Yanagi U, Osawa H (2018). Effects of low-level inhalation exposure to carbon dioxide in indoor environments: a short review on human health and psychomotor performance. Environment International 121:51-56. https://doi.org/10.1016/j.envint.2018.08.059
  21. International Code Council (2021). International Fire Code, Section 5307, carbon dioxide systems: gas detection alarming at 5,000 ppm (low) and 30,000 ppm (high) for installations over 100 lb CO2, with automatic shutoff and ventilation. (non-peer-reviewed source) https://codes.iccsafe.org/s/IFC2021P2/part-v-hazardous-materials/IFC2021P2-Pt05-Ch53-Sec5307.3.2
  22. Sensirion. SCD30 / SCD4x NDIR CO2 sensor datasheets and SCD30 field-calibration application note (ASC/ABC assumes periodic exposure to ~400 ppm fresh air). (non-peer-reviewed source) https://sensirion.com/products/catalog/SCD30
  23. Zhengzhou Winsen Electronics. MQ-135 and MQ-5 metal-oxide gas sensor datasheets (MQ-135 infers equivalent-CO2 from mixed VOCs; MQ-5 targets LPG / combustible gas). (non-peer-reviewed source) https://www.winsen-sensor.com/d/files/MQ-5.pdf
  24. Oklahoma State University Extension. Greenhouse carbon dioxide supplementation (HLA-6723). (non-peer-reviewed source) https://extension.okstate.edu/fact-sheets/greenhouse-carbon-dioxide-supplementation.html
  25. US EPA. What is a HEPA filter? A HEPA filter removes at least 99.97% of airborne particles at 0.3 um (particulates only, not gases). (non-peer-reviewed source) https://www.epa.gov/indoor-air-quality-iaq/what-hepa-filter
  26. Souza et al. (2024). Ethylene elimination using activated carbons impregnated with copper oxide. Molecules 29(12):2717 (plain activated carbon held ~1,111 ug ethylene/g; ethylene kinetic diameter ~3.9 A). https://doi.org/10.3390/molecules29122717
  27. Alvarez-Hernandez MH, et al. (2024). Postharvest handling of ethylene with oxidative and absorptive means (review). PMC10933227 (KMnO4 scrubbers ~40-80 mg ethylene/g, irreversible; activated carbon 11-78 mmol/kg; photocatalytic TiO2/UV). https://pmc.ncbi.nlm.nih.gov/articles/PMC10933227/
  28. Abeles FB, Morgan PW, Saltveit ME (1992). Ethylene in Plant Biology, 2nd ed. Academic Press (response threshold ~10 ppb; half-maximal response 0.1-1 ppm).
  29. Cornell University Greenhouse Horticulture. Ethylene in the greenhouse: symptoms, detection and prevention (chronic damage from ~10 ppb; acute epinasty/abscission/chlorosis above 0.1 ppm). (non-peer-reviewed source) https://greenhouse.cornell.edu/crops-culture/ethylene-in-the-greenhouse-symptoms-detection-prevention/
  30. Wheeler RM, Peterson BV, Sager JC, Knott WM (1996). Ethylene production by plants in a closed environment. Advances in Space Research 18(4-5):193-196 (sealed NASA CELSS chamber reached 40-120 ppb ethylene from plant emission alone; visible wheat epinasty). https://doi.org/10.1016/0273-1177(95)00877-h
  31. Hudelson TJ, et al. (2023). Elevated atmospheric ethylene and high temperature independently inhibit fruit set but not vegetative growth in tomato. HortScience 58 (continuous 20 ppb cut fruit yield to 37-51% of control; 40 ppb to 4-11%; vegetative growth reduced <10%). https://journals.ashs.org/hortsci/view/journals/hortsci/58/3/article-p247.xml
  32. Monthony AS, et al. (2026). Sex-specific ethylene responses drive floral sexual plasticity in Cannabis sativa. The Plant Journal (cannabis has an intact, sensitive ethylene-signalling pathway).
  33. US OSHA. 29 CFR 1910.146, permit-required confined spaces: an oxygen-deficient atmosphere is below 19.5% O2 (normal air is ~20.9%). (non-peer-reviewed source) https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.146
  34. HPAC Engineering. Latent loads matter: HVAC for cannabis grow facilities (transpiration returns most irrigation water to room air as vapour, the dominant dehumidification load; filters do not remove it). (non-peer-reviewed source) https://www.hpac.com/industrial/article/21270796/latent-loads-matter-hvac-for-cannabis-grow-facilities
  35. Zhang D, et al. (2020). Substantial differences occur between canopy and ambient climate: quantification of interactions in a greenhouse-canopy system. PLoS ONE 15(5):e0233210 (in-canopy RH ~15-25% higher than surrounding air). https://doi.org/10.1371/journal.pone.0233210
  36. Baptista FJ, et al. (2012). Effect of nocturnal ventilation on the occurrence of Botrytis cinerea in Mediterranean unheated tomato greenhouses. Crop Protection (ventilation that lowers humidity reduced grey mould).
  37. Liang J, et al. (2026). CO2 and air-change-rate optimisation in photoautotrophic micropropagation of medicinal Cannabis sativa. Industrial Crops and Products (~4.4 air changes/hour optimal at 800 ppm CO2; higher exchange dried the substrate). Plantlet-scale.
  38. Punja ZK, Collyer D, Scott C, Lung S, Holmes J, Sutton D (2019). Pathogens and molds affecting production and quality of Cannabis sativa L. Front. Plant Sci. 10:1120. https://pmc.ncbi.nlm.nih.gov/articles/PMC6811654/

Citations marked in-text as [n] map to this list. Peer-reviewed sources except where noted. Cannabis tissue culture is strongly genotype-dependent, verify dilutions, hormone doses and local regulations against the primary sources before relying on them.