CO2 Sensors in Open-Plan Offices: Ventilation Assumptions vs What Monitors Actually See

Walk into a modern office with exposed ductwork, living walls, and a little e-ink tile flashing “CO₂ 720 ppm,” and you might assume someone finally wired cognition to the HVAC brain. Often the truth is messier: a consumer-grade NDIR sensor bolted where the facilities manager could hide cable, feeding a dashboard nobody trained on, while the air handler still chases a schedule written when half the desks were storage.

Carbon dioxide is a useful proxy for human occupancy and dilution ventilation effectiveness, but it is not a synonym for “healthy air.” This article unpacks what open-plan CO₂ monitors actually measure, where building standards assume averages that reality violates, and how to read numbers without turning ppm into astrology.

Facilities teams already juggle comfort calls, sustainability targets, and capital renewals; this is not a pile-on. It is a translation layer between glossy sensor marketing and the ductwork you cannot see.

Why CO₂ became the office dashboard darling

Outdoor background CO₂ sits near 420 ppm and climbs slowly with global trend; indoors, humans exhale roughly forty grams per person-hour at rest—more when you are stressed in a budget meeting. That makes indoor concentration a rough count of exhaled air accumulated against whatever fresh air the system delivers. Compared with expensive per-person airflow meters, a $200 sensor network feels like democratized building science.

ASHRAE and cognate standards historically sized ventilation using occupancy tables and minimum outdoor air rates. Post-pandemic interest in “more fresh air” collided with energy codes and heat waves, so operators hunt levers that look responsive without ripping chillers out. CO₂ feedback loops promise dynamic control—more people, more air—except open plans scramble the mapping between sensor location and what each breathing human receives.

What the sensor sees: NDIR limits and calibration drift

Most affordable units use non-dispersive infrared absorption tuned to the 4.26 µm CO₂ band. They are solid when calibrated, but they drift with temperature swings, barometric pressure changes, and age. Auto-calibration routines that assume periodic exposure to “outdoor-like” 400 ppm can hallucinate if the device never sees fresh air—think a sensor trapped in a storeroom that thinks midnight is outdoors. Cross-sensitivity to high humidity events can spike readings briefly; alcohol-based cleaners can do worse.

Placement amplifies error. Ceiling returns pull air that mixed across zones; a desk sensor reads local exhalation plumes. A monitor behind a monitor might read 950 ppm while three meters away feels fine. Open plans are gradient fields, not bathtubs.

Ventilation assumptions that monitors expose but do not fix

Variable-air-volume systems rebalance dampers constantly; a CO₂ tile cannot tell you whether outdoor air fraction actually rose or the box reheated return air to mask poor delivery. Economizers stuck shut, dirty filters, and mis-linked BACnet schedules all produce “mysterious” plateaus where ppm climbs despite “max ventilation” on the slide deck. The sensor is a symptom flashlight, not a wrench.

CO₂ versus risk from respiratory pathogens

During COVID-era discourse, ppm became a moral speedometer. Lower is better for diluting airborne contaminants in general, but CO₂ is not a direct measure of viral load or filtration efficiency on its own. HEPA in-room cleaners can slash particle counts without moving CO₂ much if they recirculate. Conversely, aggressive outdoor air that drops ppm can spike energy and humidity problems that indirectly harm comfort. Policy needs stacks of controls, not one KPI.

What building operators should log alongside ppm

• Outdoor air dampers and fan speeds correlated with timestamps.
• Occupancy estimates from badging or Wi-Fi density—not perfect, but explanatory.
• Filter pressure drop and last change date.
• Indoor temperature and RH because thermal comfort changes how people tolerate stuffiness.

What desk workers can reasonably infer

If your zone reads north of 1,000 ppm sustained and you feel groggy, opening a conference door or taking a walk is rational even before you diagnose the AHU. If numbers bounce wildly every ten minutes, suspect firmware auto-calibration or a device too close to the coffee grinder. Trend lines beat snapshots.

Green building certifications and the paperwork gap

LEED and cousins love measurable IAQ indicators, but audits are periodic while leases are daily. A spike during an all-hands Friday may never hit the commissioning report. Persistent employee logging—ethical, anonymized—sometimes tells truer stories than BAS trend exports nobody downloads.

Future sensors: multisensor boxes and the VOC confusion

Vendors bundle total VOC proxies that react to perfume and printer toner. Useful for odor events, dangerous if blended into a single “air quality score” that hides CO₂. Demand raw channels and documentation of fusion algorithms; black-box indices are marketing, not maintenance.

Setpoints: where “green” meets cognitive performance folklore

You will see blog posts claim sharp thinking dies above 1,000 ppm. Laboratory studies vary wildly with task, baseline ventilation, and whether subjects were blinded. The pragmatic office takeaway is trend and slope: a slow climb from 650 to 1,100 ppm across a morning suggests under-dilution for headcount; a spike after thirty people enter a glass box in five minutes might be physics, not negligence. Avoid turning ppm into a shame sticker on Facilities’ door.

Commissioning versus operations handoff

New buildings undergo TAB—testing, adjusting, balancing—then reality arrives: walls move, sensors get bumped, setpoints “optimized” for complaints about cold ankles. CO₂ logging during the first year of occupancy catches drift that commissioning missed because the tenant mix was hypothetical. Treat year-one IAQ data as part of lease acceptance, not decoration.

Remote work hybrids and ghost schedules

Hybrid attendance makes occupancy-driven ventilation smarter on paper and harder in practice. If Monday is full and Friday is empty but the AHU still pre-cools like 2019, you waste energy without helping Tuesday’s air. Integrate calendar density or desk booking APIs where privacy rules allow; otherwise CO₂ feedback is the coarse correction loop you actually have.

Legal and HR edges nobody wants in email

Workers citing ppm in accommodation requests can force conversations employers are unprepared for. Document what the sensor is not: not medical diagnosis, not proof of mold, not substitute for industrial hygiene when someone smells chemical off-gassing. Point people to qualified IH pros when symptoms persist regardless of ppm.

Energy codes tightening the vice

More outdoor air fights heat domes and utility bills. CO₂-based demand control ventilation (DCV) can save energy when occupancy swings, but sensors must be maintained or the control loop optimizes fiction. Budget calibration labor the way you budget filter changes.

DIY literacy for curious employees

Portable monitors let you walk gradients and learn your floor’s personality. Share findings with facilities as data, not accusations. Photograph sensor model numbers; firmware updates sometimes fix wandering baselines. If leadership buys wall tiles for vibes alone, polite questions about maintenance contracts extend their useful life.

Outdoor air fraction: the number hides behind the ppm

Indoor CO₂ drops when outdoor air dilutes exhaled breath. But two rooms at the same ppm can differ wildly in particle counts, ozone from copiers, or formaldehyde off-gassing from new furniture. PM2.5 sensors add another channel; so do relative humidity probes that explain why “fresh” air feels swampy in August. Multivariable thinking is harder than glancing at a traffic-light widget—do it anyway before redesigning work from home policy around a single tile.

Stack effect and tall buildings

In towers, elevator shafts and stairwells can pump air vertically, stratifying CO₂ by floor even when central plants behave. A monitor on level 42 may disagree with level 12 not because anyone failed, but because infiltration paths differ. Troubleshooting sometimes means walking the mechanical floors, not refreshing a browser tab.

Meeting rooms: the turbocharger nobody models

Glass boxes with twelve people and one undersized supply diffuser spike faster than open areas where air mixes laterally. If your app only shows floor averages, meeting rooms stay invisible until someone passes out—not literally, hopefully, but focus loss is measurable. Per-room sensors are cheaper than lost sales calls; penny-wise building standards sometimes skip them anyway.

Night flush and unoccupied setbacks

Many systems drop outdoor air overnight to save energy, then purge before occupancy. CO₂ should fall during those purges; if it does not, dampers merit inspection. Morning peaks that lag calendar by an hour often trace to purge schedules misaligned with early birds opening the office at six.

Psychology: numbers change behavior even when wrong

Visible ppm can nudge people to crack windows or skip the stuffiest huddle room—even when the sensor is miscalibrated. That placebo has limits: chronic false highs train cynicism; false lows breed complacency. Accuracy matters because trust is part of the safety system.

Closing frame

CO₂ monitors in open offices are worth installing—if you treat them as one instrument in an orchestra, not a soloist. Ventilation assumptions baked into codes assumed tidier rooms than glass barns full of hot desks. Sensors reveal those mismatches; people still have to act on the plot.

When the next heat dome hits, the honest conversation will still be about trade-offs—not about whether the little screen was pretty.