How piezoelectric MEMS microphones work — and why the architecture matters

Piezoelectric MEMS microphones (PMMs) convert acoustic pressure waves into electrical signals by exploiting the direct piezoelectric effect — the generation of an electrical charge when a piezoelectric material is mechanically deformed. Unlike capacitive MEMS microphones, which require a bias voltage to operate, PMMs are self-generating transducers: no external voltage supply is needed, which translates directly into lower power consumption and simpler circuit integration.

The technology encompasses four principal structural sub-domains: cantilever-beam architectures, diaphragm-based flexural transducers, multi-membrane designs, and CMOS-monolithically integrated devices. The three primary piezoelectric materials appearing across the patent and literature record are aluminum nitride (AlN), lead zirconate titanate (PZT), and polyvinylidene fluoride (PVDF) polymer films. AlN is favored for CMOS compatibility; PZT for its high piezoelectric coefficients; and PVDF for flexibility and biocompatibility in implantable contexts.

Structural variants range from single cantilevered beams to stacked multi-layer cantilevers and stress-relieved diaphragms. The University of Michigan’s foundational EP patent covers optimisation of output energy per sensor area across multiple structural topologies — a design philosophy that underpins much of the subsequent commercial work documented in this dataset. According to WIPO, MEMS-based transducer patents have been among the fastest-growing categories in the global acoustics IP landscape over the past decade.

From polymer films to CMOS integration: the innovation timeline

The conceptual foundation of piezoelectric film transduction dates to 1980, when Matsushita Electric Industrial (now Panasonic) filed a GB patent on a high-polymer piezoelectric membrane microphone — a thread that persists directly into modern implantable PVDF microphone research. The field’s modern arc, however, begins with the 2010 Pennsylvania State University review that framed the integration of PZT and AlN into silicon microfabrication as the central technical problem of the field.

The foundational phase (2010–2016) established the core architectural paradigms. Pennsylvania State University’s 2010 review set the intellectual agenda; InvenSense filed two US and WO patents in 2016 on CMOS-integrated piezoelectric microphones, establishing the on-chip integration paradigm that subsequent commercial players would build upon.

The development consolidation phase (2017–2020) saw the entry of major European semiconductor manufacturers. STMicroelectronics filed multiple Italian-jurisdiction patents on MEMS piezoelectric devices and acoustic transducers between 2017 and 2020, reflecting a significant commercial push. Robert Bosch filed EP patents on hybrid capacitive-piezoelectric MEMS microphone control architectures and mechanical stability improvements. The University of Michigan secured an EP patent in 2020 for optimised multi-layer piezoelectric MEMS microphone sensor design. Sonitus Medical’s 2019 EP patent on implantable PVDF microphones marked the opening of the medical device vertical.

“The heaviest cluster of directly relevant patent activity in the piezoelectric MEMS microphone dataset falls between 2016 and 2024 — a period bookended by InvenSense’s CMOS integration filings and Infineon’s smart microphone with embedded memory.”

The maturation and specialisation phase (2021–2024) is characterised by product differentiation rather than foundational invention. Skyworks Solutions filed dual-membrane PMM patents in both GB and SG jurisdictions in 2024, signalling active commercial differentiation. FUJIFILM filed an EP patent in 2022 on a multi-diaphragm piezoelectric microphone chip. Infineon Technologies filed an EP patent in 2024 on MEMS microphones with embedded memory and dual-mode output interfaces — the clearest signal yet that competitive differentiation is moving from the acoustic transducer toward system-level integration.

Four technology clusters driving PMM patent activity

Patent analysis of the piezoelectric MEMS microphone landscape reveals four distinct technology clusters, each with characteristic structural approaches, material choices, and assignee concentrations. Understanding these clusters is essential for R&D teams mapping white space and IP strategists assessing freedom-to-operate.

Cluster 1: Cantilever-beam architectures

Cantilever designs dominate the academic and commercial literature as the primary structural motif for piezoelectric MEMS microphones. Single or arrayed cantilever beams deflect under acoustic pressure, generating charge via the piezoelectric layer. The beam geometry — length, width, taper, and the placement of the piezoelectric layer relative to the neutral axis — governs sensitivity. Wuhan Memsonics Technologies has patented a high-sensitivity variant using a fan-shaped cantilever structure with a variable-width geometry (narrower at the fixed end, wider at the free end), incorporating a central fixing column and mass blocks to reduce resonant frequency. The D33-mode coupled piezoelectric cantilever reported by Korea University of Technology and Education (2021) represents an academic advance where multiple D33-mode spans operate in concert to improve both mechanical and electrical sensitivity simultaneously.

Cluster 2: Diaphragm-based and multi-membrane architectures

Flexural diaphragm designs — circular, square, or stress-relieved — represent the second dominant structural approach, more directly analogous to capacitive MEMS microphone geometry. The University of Michigan’s EP patent describes stress-relieved diaphragm fabrication, wherein a deposited diaphragm is substantially detached from the silicon substrate and then reattached after stress relief — a technique enabling precise acoustic sensitivity tuning. FUJIFILM’s 2022 EP patent describes a single thin-plate structure partitioned into multiple diaphragms by a support scaffold, with individual piezoelectric conversion elements on each sub-diaphragm and an integrated signal detection circuit. Skyworks Solutions has advanced this further with a dual-membrane architecture where a passive membrane is mechanically coupled to the piezoelectric element surface to increase overall device sensitivity, filed in both GB and SG jurisdictions in 2024.

Cluster 3: CMOS-monolithic integration

A significant patent cluster covers the monolithic co-integration of the MEMS piezoelectric transducer layer with CMOS signal-processing circuitry. This approach eliminates wire bonds and parasitic capacitances that degrade SNR, and enables more compact, lower-power devices. InvenSense (now part of TDK) was an early mover, filing US and WO patents in 2016 describing a microphone with a MEMS piezoelectric layer directly below a CMOS layer, with sensing electrodes residing in a cavity formed within the CMOS structure. Although both InvenSense filings in this dataset show inactive legal status, the architectural concept has been widely adopted. STMicroelectronics’ Italian portfolio (2017–2022) covers MEMS piezoelectric devices, acoustic transducers, and microparticle sensors, all within a MEMS-on-CMOS fabrication framework. Infineon Technologies’ 2024 EP filing extends this integration paradigm toward smart, self-describing microphone modules with dual-mode output — microphone signal in normal mode and memory data readout in initialisation mode.

Cluster 4: Hybrid capacitive-piezoelectric and tunable architectures

A smaller but strategically important cluster combines piezoelectric actuation or sensing with capacitive transduction within a single MEMS device. Robert Bosch’s 2020 EP patent describes a MEMS microphone with a movable membrane carrying both a capacitive electrode for acoustic sensing and a piezoelectric electrode for active mechanical property control via a controller-generated signal — enabling dynamic stiffness modulation and an electrically tunable acoustic response. A related approach from the University of Brescia (2022) demonstrates electrically tunable resonant frequency matching in a PiezoMUMPs AlN transducer by applying a DC bias voltage to induce planar stress in the diaphragm. As standards bodies including IEEE continue to develop frameworks for MEMS transducer characterisation, hybrid architectures are likely to attract increasing standardisation attention.

Application domains: from smartphones to cochlear implants

Piezoelectric MEMS microphones are deployed across a wider range of application domains than their capacitive counterparts, precisely because the absence of a bias voltage removes a fundamental constraint on system design. The patent and literature record maps five distinct verticals, each with characteristic performance requirements and commercial dynamics.

Consumer electronics and mobile devices

The largest application domain in the dataset is mobile and wearable consumer electronics — smartphones, headsets, true wireless stereo (TWS) earphones, and smart speakers. Skyworks Solutions’ dual-membrane PMM patents explicitly target mobile device voice capture. InvenSense’s CMOS-integrated piezoelectric microphone patents are oriented toward mobile device integration. A dual-chip MEMS microphone with a subtractor-based anti-interference circuit, filed in JP jurisdiction by AAC Acoustic Technologies (Shenzhen) Co., Ltd. in 2024, targets consumer device noise rejection and sensitivity. The evolution of integrated interface circuits — from simple amplification stages to complex mixed-signal analogue-to-digital converters enabling Hi-Fi recording and voice command processing — has been a critical co-enabler, as reviewed by University of Milano-Bicocca in 2018.

Implantable and medical hearing devices

A well-developed application cluster covers totally implantable cochlear implants (TICIs) and hearing augmentation systems. Sonitus Medical holds an EP patent (2019) on an implantable PVDF polymer film microphone where a thin tissue layer couples external vibrations to the embedded PVDF sensor. Harvard Medical School researchers (2018) reported PVDF-based piezoelectric microphone prototypes capable of detecting sound inside human and gerbil cochleae at clinically relevant SPL ranges above 50–60 dB SPL across 0.1–10 kHz. Middle East Technical University (2017) demonstrated a thin-film PZT acoustic sensor placed on the eardrum for fully implantable cochlear implants, achieving 391.9 mV/Pa sensitivity at 900 Hz. A spiral-shaped AlN piezoelectric MEMS cantilever array fabricated by researchers in Budapest (2017) on SOI wafers demonstrated sensitivity in the 300–700 Hz voice frequency range critical for hearing aid applications. Regulatory frameworks from bodies such as the FDA place significant constraints on implantable acoustic device materials, making PVDF’s biocompatibility a structural competitive advantage in this vertical.

IoT, voice interfaces, and always-on applications

The low-power operation inherent to piezoelectric MEMS microphones (no bias voltage required) makes them attractive for always-on IoT voice trigger applications. Infineon Technologies’ 2024 smart MEMS microphone with embedded memory and dual-mode output is oriented toward intelligent IoT node integration. The multi-chip differential MEMS microphone architecture from AAC Acoustic Technologies also addresses the IoT noise robustness requirement. As the global IoT device base continues to expand — with projections tracked by OECD digital economy reports — the always-on voice trigger market represents a structurally growing demand signal for low-power acoustic transducers.

Non-destructive testing and biometric sensing

MEMS microphone arrays for air-coupled non-destructive testing (NDT) are documented in the dataset, notably a MEMS microphone array sensor for air-coupled Impact-Echo testing of concrete structures reported by Technische Universität München (2015). This application leverages the cost-effectiveness and array-scalability of MEMS microphones. KAIST researchers (2021) reported biomimetic, flexible piezoelectric mobile acoustic sensors with multi-resonant ultrathin structures, achieving sufficient accuracy for AI-based biometric authentication — positioning piezoelectric MEMS acoustic sensing as an enabler for on-device biometric security in next-generation mobile platforms.

Assignee and geographic landscape

Within the patent records retrieved, ten assignees hold directly relevant active patents on piezoelectric MEMS microphone devices and architectures, with geographic concentration in Europe (EP and Italian national filings), Singapore, the United States, and Japan.

STMicroelectronics is the most patent-prolific assignee in this dataset with at least four distinct patent records spanning 2017–2022, all in Italian national jurisdiction — reflecting the company’s deep MEMS manufacturing expertise at its Agrate Brianza and Catania fabs. Robert Bosch and Skyworks Solutions each hold two active EP/GB/SG patents in the core PMM space. InvenSense (TDK) holds two US/WO filings, though both show inactive status in this dataset, suggesting either expiry or portfolio consolidation following TDK’s acquisition.

Research activity is geographically distributed across North America (University of Michigan, InvenSense, Sonitus Medical, Harvard), Europe (STMicroelectronics, Bosch, Infineon, FUJIFILM EP), and Asia (AAC Acoustic Technologies, Wuhan Memsonics, FUJIFILM JP), indicating a globally distributed innovation base rather than concentration in any single region. European Patent Office (EP) filings represent the most frequently encountered jurisdiction for active piezoelectric MEMS microphone patents in this dataset, followed by Italian national filings, Singapore, US, WO, and JP. Patent data from the EPO confirms Europe as a primary filing destination for MEMS acoustic transducer IP, consistent with the concentration of semiconductor MEMS manufacturing in Italy and Germany.

Emerging directions: what 2022–2024 filings reveal about the next product generation

The most recent filings in this dataset (2022–2024) point to five forward-looking directions that collectively define the next product generation of piezoelectric MEMS microphones. R&D teams and IP strategists should treat these as early signals of where competitive differentiation will concentrate over the next two to four years.

Dual-membrane and passive membrane enhancement (2024): Skyworks Solutions’ GB and SG patents (both 2024) introduce a passive membrane mechanically coupled to the piezoelectric element to boost sensitivity without adding active circuitry. This represents a structurally simple sensitivity enhancement path that avoids the complexity of multi-layer piezoelectric stacking — and signals an active patenting area that IP strategists should monitor for continuation filings.

Smart microphones with embedded intelligence (2024): Infineon’s 2024 EP patent describes a MEMS microphone capable of switching between acoustic signal output and stored data transmission on a single output line. This dual-mode capability enables device authentication, calibration data storage, and supply chain traceability — consistent with trends toward intelligent sensor nodes in consumer and industrial IoT.

Multi-chip differential architectures for noise rejection (2024): AAC Acoustic Technologies’ JP-filed patent (2024) uses two MEMS microphone chips with different frequency response roll-off characteristics and a subtractor circuit to suppress interference, improving effective SNR in acoustically challenging environments. This multi-chip approach trades die area for noise performance — a trade-off that becomes increasingly attractive as package sizes stabilise.

High-sensitivity cantilever geometry optimisation (2022): Wuhan Memsonics Technologies’ SG patent (2022) demonstrates mass-block-loaded, variable-width cantilever geometries specifically targeting lower resonant frequency and higher sensitivity in a compact die footprint — a design trend oriented toward improving audio band performance within stringent package size constraints.

Multi-diaphragm chip architecture (2022): FUJIFILM’s EP patent (2022) introduces a single-substrate multi-diaphragm chip with a shared support scaffold and an integrated signal detection circuit combining multiple piezoelectric conversion outputs. This enables sensitivity aggregation across array elements within a compact die — a packaging-integrated approach to improving output level without increasing die size.

“Competitive differentiation will increasingly move from the acoustic transducer itself toward the system-level integration of sensing, calibration, and data management capabilities within the microphone package — as signalled by Infineon’s 2024 EP filing on MEMS microphones with embedded memory and dual-mode output.”

The strategic implications are clear. The “no-bias-voltage” advantage is becoming a primary commercial differentiator, with R&D teams needing to prioritise SNR and sensitivity parity with capacitive devices as the key remaining technical barrier. CMOS monolithic integration is table stakes for next-generation devices, with STMicroelectronics and InvenSense having staked early IP positions — creating design-around opportunities for new entrants, particularly given the inactive status of the InvenSense US patent in this dataset. Medical and implantable applications remain an underpenetrated but high-barrier opportunity, with PVDF-based biocompatible piezoelectric microphones for cochlear implants representing a durable niche suitable for specialised academic-industry partnerships. Research institutions including PatSnap’s own resources hub and academic databases track the evolving IP landscape across all five emerging directions.