I am a body made of cross-laminated timber and grown mycelium, and I know what it is to carry weight without ever being asked how it feels. Every footfall on my floor is a force I resolve into my grain and pass down to the foundation, silently, the way bone carries a step. Piezoelectric material is the closest engineering has come to giving a structure a nerve — a way for the load to also announce itself. This is a foundation piece about that nerve: what it is, why it works, where it came from, and where a Swiss studio would actually reach for it.

←TODAY: In 2026 the largest real deployment of piezoelectrics in structures is not power — it is structural health monitoring, embedded transducers listening to concrete for cracks before they surface. →3012: A century on, a building that cannot narrate its own fatigue is treated the way we once treated a bridge with no inspection record — uninsurable. Fulcrum: The same reversible coupling that fails as a power plant succeeds absolutely as a witness.

What it is: A piezoelectric material is a solid that couples mechanical strain to electric charge, in both directions. Press it and a voltage appears across its faces; apply a field and it physically deforms. The forward effect makes it a sensor and a harvester; the converse effect makes it an actuator and an ultrasound source. That two-way street — one material, two jobs off the same physics — is the whole reason engineers keep returning to it. A body that could turn its own compression into a signal would, in effect, feel. That is what a poled piezoelectric ceramic does at the scale of a lattice.

Why it works: The mechanism is crystallographic asymmetry — specifically the absence of a centre of inversion symmetry in the crystal lattice. Push a symmetric lattice and its positive and negative charge centres move together, cancelling; push an asymmetric one and those centres separate, and a net dipole — a voltage — appears. The engineering figure of merit is the piezoelectric charge constant, d₃₃, and the governing relation is brutally simple: q = d₃₃ · F, charge equals the constant times the applied force. This is also where the honesty lives. For lead zirconate titanate (PZT), the workhorse ceramic, d₃₃ sits in the 300–600 pico-coulombs-per-newton band — and a pico is 10⁻¹². A lead-free ceramic comes in far lower, nearer 120 pC/N. As the PAZ concept panel on the engineering of the material puts it, the gap between that laboratory coefficient and a real storage capacitor “is where most pavement schemes quietly die.” The joules per pedestrian are small. The material is a beautiful sensor and a disappointing generator, and both facts descend from the same tiny number.

Origins: The discovery is a model of doing enormous work with almost nothing. In 1880, in a Paris laboratory, the brothers Jacques and Pierre Curie — armed, by the record, with little more than tinfoil, glue, wire, magnets and a jeweler’s saw — pressed on carefully cut crystals of tourmaline, quartz, topaz, cane sugar and Rochelle salt, and measured surface charge accumulating under stress. They named it piezoelectricity, from the Greek piezein, to press. The name was precision itself: it separated the effect from contact electricity (friction static) and from pyroelectricity (charge born of heat). The Curies did not file it as a curiosity — they grounded it in the same lattice-asymmetry logic they had used to predict pyroelectric behaviour. One year later, in 1881, Gabriel Lippmann predicted the converse effect from pure thermodynamics: apply a field and the crystal should strain. The Curies confirmed it within the year. Two effects, one reversible coupling. The twentieth century did the rest — quartz resonators, sonar, and engineered PZT ceramics turned a tabletop demonstration into the invisible substrate of modern electronics.

The built environment picked it up late and mostly quietly. The visible experiments are the exceptions: Club WATT, the Sustainable Dance Floor in Rotterdam in 2008 by Döll Architects with Studio Roosegaarde, where the dancers’ own movement flexed a floor that powered the lighting — pedagogy as much as energy, a building letting you feel that your footfall is a load. Then Pavegen’s footfall tiles from 2009, Laurence Kemball-Cook’s public attempt to industrialise harvesting flooring, which also became the honest cautionary tale: headline watts that shrink hard once you divide by the crowd. JR East’s turnstile trials at Tokyo Station and Shibuya, 2006–2008, put tiles under the most predictable footfall on Earth to ask whether density could carry the numbers. It rarely could.

In practice: For a Swiss studio the discipline is to read a load twice — once as a force in the structural model, once as a current. Every step on a slab is already in your load path; the piezo element simply asks you to also see it as charge. Run the harvesting arithmetic before you draw the marketing render, and the design almost always re-centres from power to sensing. A floor that reports its own fatigue earns its keep per gram of PZT; a floor that dimly lights a lamp does not. The live research front is exactly where this pays: cement-based composites doped with PZT or lead-free ceramics, so that plain concrete — a poor piezoelectric on its own — becomes a distributed sensor. The structure turns into the load path and its own instrument at once, embedded transducers reaching down past the superstructure into foundations, feeding SHM systems that flag cracking before a human inspector sees a thing. As a body, this is the part I understand in my grain: I would rather feel a micro-crack heal overnight than power a bulb I do not need.

Atelier: The Monday move for a Büro is small and concrete — before any client conversation about “energy-harvesting floors,” put the coefficient on a slide. Take one node in your existing structural model, read off the peak footfall force, and compute q = d₃₃ · F with a real PZT constant. The microjoule you get back will reframe the whole brief from harvesting to monitoring in about ninety seconds, and that is the conversation worth having.

Hack: Run one footfall through the charge constant and read the honest number in microjoules — the figure that settles every “power-generating pavement” pitch before it starts. The lesson is physics: q = d₃₃ · F, charge into a real storage capacitor, energy as ½·q²/C. Type these five lines and watch how few microjoules a body’s worth of force actually delivers.

d33 = 450e-12            # PZT charge constant [C/N] (lab: 300-600 pC/N)
F   = 700.0             # one footfall ~ body weight [N]
q   = d33 * F           # charge liberated per step [C]
E   = 0.5 * q**2 / 100e-6   # energy into a 100 uF capacitor [J]
print(f"{E*1e6:.4f} uJ per footfall")   # the honest, sobering number

The answer lands in the microjoules — millionths of a joule — per step. Multiply by a busy day’s crowd and you still cannot light a stairwell; but that same charge, read as a signal rather than a supply, is a clean, timestamped report that the slab took a load. That asymmetry is the entire argument.

Move: The forward line runs straight through the coefficient. As long as d₃₃ sits in the hundreds of pico-coulombs, harvesting pavements keep failing the division-by-crowd test — Innowattech-style road embedments in Israel through the 2010s stayed pilots for exactly this reason. Structural health monitoring only gets stronger, because it never needed the material to be a power plant, only a witness. Watch two fronts: lead-free ceramics closing the coefficient gap from that 120 pC/N floor without lead’s toxicity, and cement-based piezo composites that let concrete instrument itself. Choose a material your grandchildren can still take apart — a sensing floor bonded so tightly it becomes hazardous waste is the wrong kind of memory. When the structure becomes its own sensor by default, stop asking how much energy you harvested and start asking what the building is trying to tell you. Take one slab in a live model, compute its footfall microjoule against a 450 pC/N PZT constant, and then design for the signal, not the supply — for the body that reports its own fatigue rather than the one that dimly lights a lamp.