Paper wallpaper turns indoor humidity into power for low-energy electronics

Researchers at Binghamton University, State University of New York, have built a wallpaper that generates electricity from the water vapor in indoor air. Their paper, published in Advanced Energy Materials on 16 September 2026 (paper), describes a single sheet of treated paper that pulls moisture in at its edges and lets it evaporate at its center, producing a small, steady voltage. Tiled into an array of 1,596 units, the wallpaper ran a wireless keyboard while lowering indoor humidity from 38% to 32%.

Each unit produces microwatts, so this is no drop-in battery replacement. It is a clear attempt to treat a building surface as a power source for the low-energy electronics that IoT designers mount on walls.

What problem does it address?

Most small IoT devices still run on a primary battery. A coin cell is cheap and predictable, but it caps the battery life of the device, and in a large smart building deployment, replacing thousands of cells becomes a recurring maintenance job. Energy harvesting tries to remove or postpone that job by collecting power from the surroundings: light, heat, vibration, radio waves or, in this case, moisture in the air.

Moist-electric generators, or MEGs, convert ambient humidity into electricity. According to the abstract, their power output is inherently low, which makes outdoor use as a general energy source impractical. The authors argue that indoor spaces are a better match, because humidity there is stable and low-power electronics need only microwatt-to-milliwatt supplies. In the university’s announcement of 29 September 2026, Choi describes indoor humidity as staying between 30 and 60 percent. The release also notes that earlier MEG designs, built in vertical or horizontal layouts, struggle to scale to wall-sized arrays. A wall offers a large surface that is always exposed to room air, which is why the team chose wallpaper as the form factor.

What did the researchers do?

The work comes from the Bioelectronics and Microsystems Laboratory in Binghamton’s Department of Electrical and Computer Engineering, part of the Thomas J. Watson College of Engineering and Applied Science. The authors are PhD student Guangya Yuan, Yang Gao (PhD 2026) and Professor Seokheun Choi, whose group has a track record in paper-based electronics. The publisher’s metadata lists the US National Science Foundation as funder.

Each generator is one sheet of paper treated in three zones, an arrangement the authors call peripheral absorption and center evaporation. Glycerol at the edges captures water vapor from the air. A middle region of polyvinylpyrrolidone (PVP), a polymer, binds water, and according to the abstract its lower porosity and lower water activity pull moisture inward. The center is treated with wax, whose water-repelling pores block liquid but let vapor pass, so evaporation is confined to the middle without a solid seal. Moisture thus moves in one direction across a single sheet, and the release explains that this sets up an ion-concentration gradient, which produces a voltage and a small current. To scale up, units are connected through concealed holes in the paper, so all wiring sits on the back.

What did they find?

The abstract reports that a single unit delivers about 0.34 V and about 2.2 µW/cm² at 80% relative humidity, with stable output over 90 minutes. An array of 1,596 units powered a wireless keyboard and at the same time reduced indoor humidity from 38% to 32%. The release adds that the team plans to raise the power yield and make the wallpaper printable so it can be manufactured at scale. It names environmental sensors and wireless keyboards as the kind of small devices the arrays could run, and suggests the moisture-absorbing effect could reduce the energy that HVAC systems spend on dehumidification.

Why it matters

For people who design IoT hardware, the useful way to read this result is power per area of wall, not power per device. At a few microwatts per square centimeter, a harvester the size of a sensor enclosure will not keep one running. The idea becomes interesting when the harvesting surface is far larger than the electronics, which is exactly what wallpaper offers: a temperature, humidity or occupancy node on a wall could in principle draw from the surface around it instead of from a coin cell.

The most relevant comparison is with indoor photovoltaics, the most mature indoor harvester. Our October research roundup covered a new consensus method for measuring indoor solar cells. Light, though, is intermittent: offices go dark at night and on weekends, and many sensors sit in corridors, cupboards or behind furniture. Humidity does not switch off with the lights. A moisture harvester could complement a small solar cell rather than compete with it, covering the dark hours that today push designers toward extra energy storage or a backup battery.

None of this changes the firmware discipline that makes harvesting viable. A battery-free node still needs long deep sleep periods, a low duty cycle, and a power-management stage that stores energy until there is enough for a measurement and a short radio burst. A microwatt source only changes where the energy comes from. Our guide to IoT device types and constraints explains why energy is usually the first constraint in a sensor design.

The humidity reduction is a separate benefit that would interest people working on building operations and environmental monitoring, not only device makers. The study does not quantify any HVAC savings, so for now it is a possibility rather than a measured gain.

What it does not show yet

The headline per-unit figures were measured at 80% relative humidity, well above the 30 to 60 percent range Choi describes as typical indoors. The abstract does not report output across that normal range, so how much power a real wall would supply in a dry winter office remains open.

Stable output is reported over 90 minutes. That says nothing yet about behavior over months or years, about dust, cleaning, paint or seasonal swings in humidity, or about how the glycerol and polymer layers age after long exposure to moisture. Neither the abstract nor the release reports durability tests of that length.

The keyboard demonstration needed 1,596 units, and the abstract does not give the total array output, how long the keyboard ran, or whether energy was buffered in a capacitor or battery. Without those details the demo shows feasibility, not a power budget a product team could design to.

The drop from 38% to 32% humidity comes from the authors’ own setup, and the abstract does not describe the room or the test duration, so it cannot be extrapolated to whole buildings. The release says the team hopes to raise power yields and wants a fully printable design for mass production; both are aims, not results. Cost and installation are not addressed. This is also a lab study from a single group; independent replication would normally come before anyone plans a product around a new material.

Where to go next

• IoT device types and constraints: how power budgets shape device design.
• Research roundup, October 2026: indoor solar measurement and self-powered sensing glasses.
• How IoT works: from sensor to cloud and back: the path of one sensor reading.
• The smart home, explained: the low-power radios home sensors rely on.

Topics: research, energy harvesting, battery-free, smart building