A unified molecular sensing platform — MOF chemistry, FBAR transduction, and on-board TinyML — wrapped around an active gated getter that doesn't just watch contamination, it removes it.
Most environmental sensors stop at detection. Moseley closes the loop — we read molecules in parts-per-billion, infer their meaning on-device, then actively pull the bad ones out of the air. It's the same platform whether it's wafers in a fab or mangoes in a reefer.
Metal-organic frameworks are crystalline cages — metal nodes joined by organic linkers — with pores measured in ångströms. Change the linker, change which molecules fit.
We coat each FBAR with a target-selective MOF and group them into four sensing channels, with four sensors per channel for redundancy and cross-correlation. The result is an array that doesn't just see "something's there" — it sees what is there, and rejects what isn't.
The same chemistry doubles as the active capture material in the Gated Getter. Detection and remediation share a chassis.
A film bulk acoustic resonator (FBAR) vibrates at a precise frequency, governed by the mass on its surface. When a molecule binds to the MOF coating, the resonator gets microscopically heavier — and the frequency drops by a measurable amount.
Read that shift continuously, across 16 sensors arranged as four sensors per channel across four channels, and you get a real-time chemical fingerprint of the air. The redundancy is deliberate: four sensors per channel let the model cross-correlate within a channel and across channels, killing false positives that any single sensor would happily report.
No reagents. No wet chemistry. No drift past calibration. The same physics behind every smartphone RF filter — repurposed to keep your wafers and your produce safe.
An FBAR array doesn't directly say "ethylene is at 12 ppb." It produces a 16-dimensional vector of frequency shifts, drifting with humidity, temperature and resonator age. Turning that into a meaningful number is a machine-learning problem.
We do it at the edge. The 16-sensor FBAR array sits on the same PCBA as a low-power MCU inside the E-Canister and E-Nose label. The sensors read the air; the MCU runs the model. A compact neural network fingerprints the full 16-channel response, compensates for drift, and emits a calibrated risk score every second — all on the board, before anything touches the radio.
That matters because reefers cross oceans, fabs run on second-by-second budgets, and remediation has to fire before the cloud round-trip would even complete.
Detection without action is just a louder alarm. The Moseley Gated Getter (MGG) is the bit competitors don't have — a thermally-stimulated MOF surface that holds contaminants captive then releases them on cue, all inside the same chassis as the sensor.
The detect-predict-remediate stack is the constant. The packaging changes to match where contamination lives.
E-Canister — palm-sized cylindrical sensor + getter for FOUPs, EFEMs, load ports, refrigerated containers, cool rooms.
E-Patch — 1.4 mm flexible printed sensor for cartons and pallets. BLE telemetry, single-use, compostable.
E-Shield — humidity-hardened receptor variant for high-moisture environments.
One firmware. One cloud. One model architecture. The chemistry and the form factor swap.
Patent applications filed under Ambient IoT Pty Ltd — patents pending — covering the canister architecture, gated remediation, humidity drift mitigation, and receptor-inspired sensing.
Technical briefs, MOF datasheets, and integration specs available under NDA. Get in touch and we'll set up a conversation with the engineering team.
→ Talk to engineering