EEE4463: MEMS Devices and Applications

Course Description

EEE 4463 MEMS Devices and Applications covers micro-electro-mechanical systems — devices in which mechanical structures a few micrometres across are fabricated by semiconductor processes and integrated with electronics. Accelerometers, gyroscopes, pressure sensors, microphones and digital micromirrors are all MEMS devices, and they are among the highest-volume products in the semiconductor industry.

The Statewide Course Numbering System titles it MEMS Devices and Applications and describes it as covering “micro-electro mechanical systems devices, microfabrication, sensors and actuators, bulk and surface micromachining, optical MEMS, instrumentation and applications.” The statewide prerequisite is EEL 3306 or consent of instructor.

Two Florida institutions carry it, both at 3 credits:

• University of Central Florida — MEMS Devices and Applications, matching the statewide title, described as covering “Micro-Electro Mechanical Systems devices, microfabrication, sensors and actuators, bulk and surface micromachining, optical MEMS, instrumentation and applications,” with a prerequisite of EEE 3350 Semiconductor Devices. UCF lists it as offered occasionally rather than every year.
• Florida International University — MEMS I.

⚠ FIU’s title implies a sequence the other institution does not run. “MEMS I” suggests a two-course treatment, and a student transferring from FIU may have covered only the first half of what a single-course version compresses into one term — while a student transferring to FIU may find the follow-on course assumes material their single course did not reach. This is an emphasis and pacing difference rather than a subject divergence, but it is worth checking a topic list rather than assuming coverage.

⚠ A note on availability. UCF’s “occasional” offering is a real planning constraint: a student who needs this course for a concentration should confirm the schedule well in advance rather than assuming an annual offering.

Learning Outcomes

Required Outcomes

• Explain how physical behaviour scales with dimension, and identify which forces dominate at the micrometre scale and which become negligible.
• Analyse mechanical microstructures — cantilevers, fixed–fixed beams, membranes — for stiffness, deflection under load and resonant frequency.
• Analyse damping in microstructures, including squeeze-film and slide-film damping, and determine quality factor.
• Describe bulk micromachining, including anisotropic wet etching of silicon and the crystallographic planes that determine the resulting shapes.
• Describe surface micromachining with sacrificial layers, the release step, and the stiction problem and its mitigation.
• Analyse capacitive transduction for both sensing and actuation, including parallel-plate and comb-drive geometries, and derive the pull-in instability.
• Analyse piezoresistive transduction and design a piezoresistive sensor with appropriate bridge configuration.
• Analyse piezoelectric, thermal and electromagnetic transduction and select an appropriate mechanism for a stated application.
• Analyse the operating principle and performance limits of MEMS accelerometers, gyroscopes, pressure sensors and microphones.
• Describe optical MEMS devices — micromirrors, scanners, optical switches, tunable filters — and their applications.
• Design interface and instrumentation electronics for a MEMS sensor, addressing the very small capacitance changes involved and the parasitic capacitance that competes with them.
• Identify noise sources in a MEMS sensor system, including Brownian mechanical noise, and estimate the resolution limit they impose.
• Describe packaging requirements for MEMS, including vacuum encapsulation and the protection of moving structures.

Optional Outcomes

• Perform finite-element simulation of a microstructure for modal and static analysis.
• Build lumped-element and equivalent-circuit models of electromechanical systems.
• Analyse RF MEMS: switches, resonators and tunable capacitors.
• Analyse microfluidic devices and lab-on-a-chip systems.
• Analyse energy harvesting and MEMS power generation.
• Analyse reliability and failure mechanisms including fatigue, wear, creep and charging.
• Complete a design project taking a MEMS device from specification to simulated verification.

Major Topics

Required Topics

• Scaling laws — how force, mass, stiffness, resonant frequency, thermal time constant and surface-to-volume ratio vary with linear dimension; why surface forces dominate body forces at small scale, and why this makes gravity nearly irrelevant and stiction a central problem.
• Mechanics of microstructures — stress and strain, Young’s modulus and Poisson’s ratio for silicon and thin films, beam bending and boundary conditions, plate and membrane deflection, residual stress and stress gradients in deposited films.
• Dynamics — the mass–spring–damper model, resonant frequency, quality factor, squeeze-film and slide-film damping, and the frequency response of a micromechanical resonator.
• Bulk micromachining — anisotropic wet etching in KOH and TMAH, crystallographic etch rates and the resulting {111}-bounded shapes, etch stops (boron and electrochemical), and deep reactive ion etching with the Bosch process.
• Surface micromachining — structural and sacrificial layers, polysilicon processes, the release etch, stiction and its mitigation by critical-point drying, self-assembled monolayers and dimples; wafer bonding.
• Capacitive transduction — the parallel-plate capacitor as sensor and actuator, the comb drive and its linear force, electrostatic force derivation, the pull-in instability and the one-third gap result, and differential sensing.
• Piezoresistive transduction — the piezoresistive effect in silicon, gauge factor, crystallographic orientation dependence, Wheatstone bridge configurations, and temperature compensation.
• Other transduction mechanisms — piezoelectric (PZT, AlN, ZnO), thermal (bimorph and thermal expansion actuators), electromagnetic and magnetostrictive, and shape-memory actuation; comparison on force, stroke, speed, linearity and power.
• Inertial sensors — the MEMS accelerometer (proof mass, suspension, capacitive readout, open- and closed-loop operation); the vibratory gyroscope and the Coriolis principle, drive and sense modes, and quadrature error.
• Pressure sensors and microphones — membrane mechanics, piezoresistive and capacitive readout, absolute and differential configurations, and MEMS microphone architecture.
• Optical MEMS — digital micromirror devices, scanning micromirrors, optical switches and crossconnects, tunable Fabry–Pérot filters, and grating light valves.
• Instrumentation and interface electronics — capacitance-to-voltage conversion, charge amplifiers, switched-capacitor readout, synchronous detection, parasitic capacitance and shielding, and sigma-delta closed-loop sensing.
• Noise and resolution — Brownian mechanical noise and its relation to damping, electronic noise in the readout, and the resolution floor of the combined system.
• Packaging — the special requirements of devices with moving parts: cavity packages, vacuum sealing and getters, stress isolation, and media protection.

Optional Topics

• Finite-element simulation: modal, static and coupled electromechanical analysis.
• Lumped-element and equivalent-circuit modelling of electromechanical systems.
• RF MEMS: switches, resonators, filters and tunable capacitors.
• Microfluidics, micropumps, microvalves and lab-on-a-chip.
• Energy harvesting: piezoelectric, electrostatic and thermoelectric microgenerators.
• Reliability: fatigue in polysilicon, wear, dielectric charging, and long-term drift.
• Bio-MEMS and implantable devices.
• A design project on a MEMS sensor or actuator.

Resources & Tools

• Microsystem Design (Stephen Senturia) is the standard text for this course and is the most rigorous treatment of the electromechanical modelling. Foundations of MEMS (Chang Liu) is the most widely adopted undergraduate text and is more accessible; Fundamentals of Microfabrication and Nanotechnology (Madou) is the encyclopaedic process reference; An Introduction to Microelectromechanical Systems Engineering (Maluf and Williams) is a compact and readable overview.
• Finite-element simulation — COMSOL Multiphysics is the dominant tool for MEMS because coupled electromechanical, thermal and fluidic physics is exactly what it is built for. CoventorWare and IntelliSuite are MEMS-specific alternatives; ANSYS also has MEMS capability.
• MATLAB for lumped-element modelling, frequency response and noise calculations — and it is worth building a mass–spring–damper model by hand before opening a finite-element package, because the lumped model is what gives you intuition.
• Layout — KLayout, L-Edit or AutoCAD for mask design.
• Foundry process design kits — the MEMSCAP PolyMUMPs and SOIMUMPs multi-user processes are the standard route for student designs to be actually fabricated, and their design handbooks are free and instructive even if you never submit a design. The NSF National Nanotechnology Coordinated Infrastructure network provides cleanroom access.
• Reference material — the Journal of Microelectromechanical Systems (the field’s principal journal), Sensors and Actuators A/B, the IEEE MEMS and Transducers conferences, and the IEEE Electron Devices and Sensors Councils.

Career Pathways

• MEMS design engineer and MEMS process engineer (SOC 17-2071, Electrical Engineers; SOC 17-2141, Mechanical Engineers — MEMS roles are filled from both disciplines).
• Sensor design engineer — a large employment category, since MEMS sensors are in every smartphone, vehicle and wearable device.
• Test and characterisation engineer for MEMS devices; reliability engineer.
• Systems engineer for inertial navigation — particularly relevant to Florida’s aerospace and defence sector, where MEMS inertial measurement units are ubiquitous in guidance systems and unmanned vehicles.
• Packaging engineer — MEMS packaging is a specialised discipline in its own right, because the package must protect a moving structure while exposing it to the quantity being measured.
• Biomedical device engineer — microfluidics, implantable sensors and lab-on-a-chip.
• Major MEMS employers include Bosch (the largest MEMS manufacturer globally), STMicroelectronics, TDK InvenSense, Analog Devices, Knowles, Texas Instruments (digital micromirror devices) and Honeywell. Florida settings include L3Harris (Melbourne and Palm Bay — inertial and RF MEMS for defence systems), Lockheed Martin (Orlando), Northrop Grumman (Melbourne — navigation systems), Honeywell (Clearwater — a significant inertial navigation site), SkyWater Technology (Kissimmee), and the Space Coast launch and satellite sector, where MEMS inertial sensors are standard flight hardware.
• ⚠ MEMS is a specialised field and dedicated design roles are concentrated at a modest number of companies, most outside Florida. The sensor-systems and instrumentation skills from this course transfer much more broadly, however, and are what most graduates actually use.

Special Information

⚠⚠ MEMS is mechanical engineering, and the prerequisite does not say so

This is the most important thing to know before registering. UCF requires EEE 3350 Semiconductor Devices; the statewide record lists EEL 3306 or consent of instructor. Both are electrical prerequisites, and they establish the fabrication and electronics background.

⚠ Neither names the mechanics that the course actually uses from week two. Beam bending, stress and strain, Young’s modulus, resonance, damping and quality factor are the working vocabulary of MEMS design, and an electrical engineering student may never have met any of it. The description’s phrase “micro-electro mechanical systems” is not decorative.

What is actually needed — and what is not. You do not need a full mechanics of materials course. You do need: stress and strain and their relationship through Young’s modulus; the deflection of a cantilever and a fixed–fixed beam under load; the mass–spring–damper second-order system and its resonant frequency and quality factor; and the idea of a mode shape. An electrical engineering student who has taken a signals and systems course already understands the second-order system — it is the same mathematics with mechanical labels, and recognising that early removes most of the apparent difficulty.

Preparation that pays: a few hours on beam bending formulae and on the physical meaning of stiffness before the term begins. Students who arrive expecting a devices course and meet a mechanics problem in week two are the ones who fall behind.

The consent-of-instructor route is genuinely used here and works in both directions: mechanical engineering students are frequently admitted and bring the mechanics but need the electronics. If you are outside electrical engineering and want this course, ask.

⚠ Course-code variation across Florida

MEMS is taught narrowly in Florida and under several numbers:

• EEE 4463 — UCF (MEMS Devices and Applications) and FIU (MEMS I).
• EEE 4463L — Florida Polytechnic University carries a course under this laboratory-suffixed number at 3 credits, titled simply MEMS — an unusual pairing of an L suffix with a 3-credit value, which suggests a laboratory-heavy delivery rather than a 1-credit adjunct. Confirm the format with the department if this is your institution.
• EEE 4222 Resonant MEMS — University of Florida, which requires EEE 3396C or instructor permission and focuses specifically on resonant devices.
• EEE 4421C Introduction to Nanofabrication — FIU and Florida Poly, covering the fabrication side.
• Comparable material appears under EML (mechanical engineering) prefixes at several institutions.

⚠ This is a prefix-divergence case as well as a numbering one. A student who took MEMS under an EML number has covered the subject, but a receiving programme naming EEE 4463 will not match it automatically. Read target-programme prerequisites literally and carry the syllabus.

Position in the curriculum

EEE 4463 is a senior-level elective following a devices course. At UCF it sits downstream of EEE 3350 alongside the semiconductor electives; at FIU it follows the devices and nanofabrication sequence. It pairs naturally with EEE 4421C (nanofabrication), which supplies the process detail, and with instrumentation or analogue design coursework, which supplies the readout electronics.

Difficulty and time commitment

The characteristic difficulty is breadth across disciplines: a single MEMS design problem requires mechanics (will the beam deflect enough?), electrostatics (what actuation voltage?), fabrication (can this be built?), electronics (can the capacitance change be measured?) and noise analysis (is it above the Brownian floor?) simultaneously. No single part is hard; holding all of them at once is.

Plan on eight to ten hours a week, more where a simulation-based design project is set — coupled electromechanical finite-element analysis has a substantial learning curve and the models take time to converge.

Articulation and transfer

SCNS records EEE 4463 as guaranteed to transfer to an institution offering the same course. Two Florida institutions carry the exact number, both at 3 credits. The course is upper-division and carries no general-education or Gordon Rule designation.

FE exam relevance

The NCEES Fundamentals of Engineering (Electrical and Computer) exam does not cover MEMS. The Mechanical FE covers the underlying mechanics, which is a reminder of the cross-disciplinary nature of the field but not a reason to take this course for FE purposes. This is an elective taken for its own value.

AI Integration

MEMS design is a field where simulation is central and where machine learning has begun to accelerate a genuinely slow design loop.

Where AI is used in the discipline. Surrogate modelling is the significant one: coupled electromechanical finite-element simulations are slow, so training a model on a set of them and using it for design-space exploration turns an overnight sweep into a real-time one. Machine learning is also applied to MEMS sensor signal processing — activity recognition from accelerometer data, sensor fusion for inertial navigation, and drift compensation in gyroscopes are all now routinely learned rather than hand-designed. That second category is worth noticing, because it is where most graduates will actually encounter it: the MEMS device produces the data that the learned model consumes.

⚠ The standard caution applies to surrogates: they are valid only within the parameter range they were trained on and extrapolate confidently and wrongly outside it. The physics-based simulation remains the ground truth.

Where a general-purpose assistant helps in coursework. Explaining the pull-in instability or why squeeze-film damping depends so strongly on gap; deriving or checking a scaling relationship; generating MATLAB code for lumped-element models and frequency response; setting up a COMSOL study; and explaining an unfamiliar process step or device acronym.

⚠ Where it fails, and why the failure coincides with this course’s subject. The characteristic error of an AI tool asked a MEMS question is to apply macroscale intuition to a microscale problem. That is precisely the error the first week of this course exists to eliminate. At the micrometre scale, surface forces dominate body forces: gravity is essentially irrelevant to a polysilicon cantilever, while van der Waals and capillary forces are strong enough to permanently stick a released structure to the substrate. A model reasoning from everyday mechanical experience will weigh these exactly backwards — worrying about the weight of a proof mass and omitting stiction, when stiction is the failure mode that actually destroys devices.

The scaling laws are the whole point of the opening material, and they are counter-intuitive by construction. A generated answer that sounds mechanically sensible is, at this scale, a warning sign rather than a reassurance.

Two further failures. Models routinely quote material properties for bulk silicon when thin-film values differ substantially — deposited polysilicon has residual stress, stress gradients and a fracture strength quite unlike a bulk wafer, and using bulk values gives a beam the wrong stiffness and the wrong resonant frequency. And they propose device geometries that cannot be fabricated, ignoring the constraint that every MEMS structure must be producible by a real process sequence with real etch selectivities and real aspect-ratio limits. Manufacturability is not a detail in MEMS; it is the primary design constraint, and it is invisible to a purely geometric answer.

The engineer’s responsibility. A MEMS design is a claim that a structure can be built by a specific process and will behave as predicted at a scale where intuition is unreliable. The habit worth forming is to ask of every design: what process makes this, what are the film properties in that process, and what dominates at this dimension? The scaling analysis is cheap and it is the check that catches the most expensive errors.

Academic integrity. UCF’s Rules of Conduct and FIU’s academic integrity policy both cover AI-generated work. Design analysis and simulation interpretation are normally expected to be your own even where coding assistance is permitted, and generated simulation results are data fabrication — treated more seriously than plagiarism. Ask before you rely on a tool, and disclose its use where the syllabus requires it.