Chapter Four · failure evidence
What Chemical Etching & Surface Micromachining got wrong, from 55 dissertations
Chemical etching and surface micromachining processes frequently suffer from physical and chemical defects that compromise device performance or cause structural destruction. Common failure mechanisms include capillary collapse during release, lateral undercutting, loss of chemical selectivity, surface roughness, and transport limitations in high-aspect-ratio features. These records come from PhD theses at 13 institutions, 2021 to 2026. Each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Surface roughness, sidewall defects, and polymer residues degrade device performance
Non-uniform etching and mask transfer errors cause striations, sloped profiles, and surface roughness across optical, electrical, and acoustic structures. Residues and etch-induced edge damage increase optical insertion losses and degrade electronic transconductance.
Tried and failed
hydrofluoric acid wet chemical etching applied to optical waveguide fabrication. Reason: isotropic etching degraded cross-sectional profile and poor selectivity reduced waveguide height
Integrated Visible-Light Platforms, Devices, and Systems for Augmented Reality and Beyond · MIT
Tried and failed
standard commercial foundry dry etching applied to sub-micron photonic crystal hole arrays. Reason: default etch parameters caused oblique sidewalls and incomplete etching at ~200 nm feature sizes
Resonant Spatial Light Modulation: Optical Programming and Sensing at the Fundamental Limit · MIT
Tried and failed
highly diluted selective chemical wet etching applied to semiconductor nanowire surface planarisation. Reason: diffusion-limited reactant transport caused severe surface roughness across macroscopic sample areas
Tried and failed
sidewall metallization liftoff after deep reactive ion etching applied to microfluidic channel electrodes. Reason: channel bottom and sidewall roughness and micromasking caused liftoff failure
CT On A Chip: Enabling High-Resolution Polar In-Situ Data Collection · Georgia Tech
Lost to a baseline
HEMT transconductance dropped sharply when finger width W was reduced below 2 µm due to sidewall edge damage from wet mesa etching, underperforming 10 µm width devices.
Cryogenic HEMT-based devices and circuits for quantum computing · EPFL
Lost to a baseline
Experimental waveguide insertion loss of ~7.2 dB was significantly worse than the simulated loss of ~2 dB due to etching roughness and residues
Flip-chip integrated terahertz quantum cascade lasers · DSpace-CRIS at TU Wien
Tried and failed
fluorine-based dry etching of microresonator coupling gaps applied to semiconductor optical microresonators. Outcome: worse than baseline. Reason: residual polymer deposition in submicron gaps could not be polished away, degrading the quality factor
Hybrid germanium-on-zinc selenide platform for photonic integrated circuits in the longwave infrared · UT Austin
Tried and failed
grayscale lithography combined with deep reactive-ion etching applied to silicon microchannel 3D profile patterning. Reason: sub-100nm resist thickness non-uniformities caused striations and major etch depth variations
Solid-State Flow Control for Ion Electrospray Propulsion · MIT
Tried and failed
semi-additive copper plating for transmission lines applied to high-frequency embedded die glass packaging. Outcome: worse than baseline. Reason: over-exposure and over-etching caused metal strip edge roughness, leading to excess transmission line loss
Glass Based Packaging for Beyond 5G Communication · Georgia Tech
Tried and failed
grayscale lithography mask transfer via deep reactive-ion etching applied to 3D microstructuring of silicon. Outcome: unstable. Reason: surface roughness and lack of process control during high-selectivity etching
Reactive ion etching techniques for uncooled pyroelectric detectors · Cranfield
Lost to a baseline
Scattering Q of rectangular OMCs from quasi-isotropic etching reached 4.71 x 10^4, an order of magnitude lower than RIBAE with HSQ-alumina (2.50 x 10^5).
Quantum Optomechanics with Color Centers in Diamond · Harvard
Lost to a baseline
Suspended AlN-on-diamond acoustic waveguides exhibited higher loss (-50 dB transmission) compared to ridge acoustic waveguides (-22.4 dB transmission) due to modal mismatch and bottom surface roughness from quasi-isotropic etching.
High-Performance Photonic and Acoustic Interfaces for Qubits · Harvard
Capillary forces, internal stress, and stiction cause structural collapse during release
Capillary surface tension during drying and unmitigated residual thin-film stresses lead to stiction, warping, and cracking in suspended membranes and cantilevers. Agitation during wet etching and liquid catalyst removal can also shatter or collapse delicate free-standing microstructures.
Tried and failed
stress-engineered suspension supports applied to suspended microcavity release etching. Reason: structural collapse caused by stiction failure during release etch
Resonant Spatial Light Modulation: Optical Programming and Sensing at the Fundamental Limit · MIT
Tried and failed
full-wafer wet etching without pre-dicing applied to silicon micro-oscillator fabrication. Reason: released structures stuck together and shattered during solution agitation
Considered and rejected
Considered and rejected: Rejected surface micromachining in favor of SOI bulk micromachining due to high residual stresses causing buckling/cracking and limited displacement.
Development of MEMS-Based Devices for Characterizing 2D Nanomaterials at Low Temperatures · IRIS - UNITN - prod
Tried and failed
deep anisotropic wet etching of thin membranes applied to silicon dioxide membrane fabrication. Outcome: unstable. Reason: long wet etching enlarged pinholes, causing thin membranes to break during post-etch drying
A flexible interconnection for integrated circuits · Iowa State
Tried and failed
wet chemical under-etching for suspended membrane release applied to thin-film diamond photonic crystal membranes. Outcome: unstable. Reason: Capillary forces during wet etching caused thin membranes to collapse and stiction to the underlying substrate
High-Performance Photonic and Acoustic Interfaces for Qubits · Harvard
Tried and failed
frontside isotropic chemical etching for membrane release applied to suspended piezoelectric thin-film resonators. Outcome: unstable. Reason: unrelieved internal stress caused severe film cracking and substantial structural warpage after release
AlScN-Based Acoustic Resonators Utilizing Piezoelectric and Ferroelectric Effects · Georgia Tech
Tried and failed
anisotropic chemical etching for free-standing microstructures applied to flexible thin-film cantilever fabrication. Reason: free-standing cantilever structures suffered mechanical failure during anisotropic wet etching inside square mask windows
A flexible interconnection for integrated circuits · Iowa State
Tried and failed
evaporated metal oxide sacrificial layer wet release applied to metallic nanostructure transfer to flexible substrates. Reason: thin-film stress caused partial or entire structure dislocation during wet chemical etching
Tried and failed
thin film membrane release with wet etching applied to suspended silicon dioxide membranes. Outcome: unstable. Reason: capillary surface tension during drying and internal stress cracked fragile thin membranes
A flexible interconnection for integrated circuits · Iowa State
Considered and rejected
Considered and rejected: Rejected liquid/wet etching of RuO2 (e.g., ceric ammonium nitrate) for HAR catalyst removal due to capillary-force-induced pattern collapse during liquid drying
Reducing manufacturing constraints of metal-assisted chemical etching · UT Austin
Excessive lateral undercutting and isotropic etching destroy structural anchors and adjacent features
Uncontrolled lateral etch propagation strips protective hard masks and destroys nearby suspension springs, waveguides, or structural anchors. Isotropic etching profiles also cause internal fluid leakage in channels and electrical shorting between closely spaced electrodes.
Tried and failed
wet chemical etching of chromium hard masks applied to oxide thin film ion milling. Reason: isotropic wet etching caused severe dimensional bias and feature degradation
Infrared electro-optical modulators leveraging plasmonic and nonlinear materials · UT Austin
Tried and failed
isotropic wet etching with hard mask applied to submicron semiconductor multilayer patterning. Reason: excessive lateral undercut stripped the protective hard mask entirely at small feature sizes
Mid-infrared ballistic metamaterials and subdiffraction limited photonic funnels · UT Austin
Tried and failed
fine nanoscale finger projections on patterned electrodes applied to single-electron transistor gate electrodes. Reason: narrow feature geometry caused lift-off defects and sustained damage from reactive-ion etching undercutting
Nanofabrication, measurement, and microscopy of ~10 nm scale single electron transistors · Imperial
Tried and failed
deep reactive ion etching applied to bulk refractory metal microchannels. Reason: severe sidewall undercutting caused internal leakage between adjacent fluid channels
Design, fabrication and test of the iridium catalysed electrolysis (ICE) thrusters · Imperial
Tried and failed
isotropic wet chemical etching of thin dielectric applied to metal-insulator-metal multilayer stack. Reason: severe undercutting of the dielectric spacer layer caused electrical shorting between adjacent metal electrodes
A Microwell-Based Impedance Sensor In Microneedle Shape For Cytokine Detection · Penn
Considered and rejected
Considered and rejected: Rejected isotropic wet etching for etching sub-micron pattern features in SiO2 in favor of dry reactive ion etching (RIE with CHF3/Ar) to maintain feature profile and avoid electrical shorting.
Building and Characterizing Graphene Nanomechanical Resonator Networks · Scholars' Bank
Tried and failed
isotropic wet etching for structural suspension applied to integrated photonic waveguides and resonators. Reason: the unmasked wet etchant destroys adjacent narrow waveguide structures during sacrificial layer release
A Heterogeneous Integrated Photonic Platform for High Sensitivity Sensors · Georgia Tech
Tried and failed
continuous isotropic gas etching for cavity release applied to suspended membrane over micropillar arrays. Reason: etched structural pillars laterally causing premature membrane release instead of clearing cavity
Capillary-Driven Condensation for Heat Transfer Enhancement in Steam Power Plants · MIT
Tried and failed
vapor hydrofluoric acid cavity release etching applied to silicon-on-insulator buried oxide sacrificial layer. Reason: defective or low-quality oxide layer caused uncontrolled lateral etch propagation destroying structural anchors
Tried and failed
segmented deep silicon dry etching applied to thick silicon wafer handle layer. Reason: cumulative lateral under-etching destroyed delicate adjacent suspension springs
Inadequate chemical selectivity causes substrate attack, mask stripping, and electrode corrosion
Insufficient etch selectivity attacks underlying bonded substrates, titanium adhesion layers, and metallic electrodes during patterned etching. Missing dopant etch stops and porous mask structures also allow aggressive etchants to dissolve target materials and strip protective masks.
Tried and failed
selective wet chemical etching of semiconductor heterostructures applied to low-arsenic quaternary III-V microcavities. Reason: insufficient etch selectivity due to chemical compositional similarity causes surface damage and degrades optical performance
Scalable development of multiplexed microparticle technologies for optical single-cell barcoding · MIT
Tried and failed
selective chemical wet etching for hard mask removal applied to heterogeneous bonded semiconductor substrate. Outcome: unstable. Reason: the acid etchant attacked the underlying bonded interface and substrate material causing severe undercutting
Hybrid germanium-on-zinc selenide platform for photonic integrated circuits in the longwave infrared · UT Austin
Tried and failed
dopant-selective wet etching using polymer masks applied to degenerately phosphorus-doped silicon. Reason: phosphorus-doped silicon lacked etch-stop behavior, leading to non-selective polymer mask removal
Bottom-up Process Development for High Throughput Electronic Device Fabrication · Georgia Tech
Tried and failed
dopant-selective wet chemical etching applied to n-type patterned silicon substrates. Outcome: did not generalise. Reason: phosphorus doping lacked the etch selectivity contrast observed with boron doping
Engineering a self-aligned metal-oxide-semiconductor gate stack for nano-modular device fabrication · Georgia Tech
Tried and failed
selective chemical etching of mold templates applied to intermetallic and semimetal nanowires. Reason: acid selectively etched constituent elements from target compounds causing structural degradation and dissolution
Thermomechanical Nanomolding of Electronic Materials · Cornell
Tried and failed
two-step chemical wet etching applied to piezoelectric thin films on titanium electrodes. Outcome: unstable. Reason: acid etchants corroded the underlying titanium adhesion layer, causing electrode delamination
Additive Fabrication of Functional Microfluidic Devices · EPFL
Tried and failed
PECVD silicon nitride etch mask applied to long anisotropic wet etching of silicon. Reason: Porous microstructure and poor step coverage caused mask stripping in concentrated KOH.
Modeling, simulation, and fabrication of a PZT valveless micropump · Iowa State
Considered and rejected
Considered and rejected: Rejected SiO2 passivation layers patterned by wet etching because the wet etchant attacked and dissolved the metallic Al electrodes.
Considered and rejected
Considered and rejected: Rejected pure wet chemical etching of PZT with hot HCl due to Ti adhesion layer corrosion
Additive Fabrication of Functional Microfluidic Devices · EPFL
Considered and rejected
Considered and rejected: Cl2 ICP-RIE dry etching for Al0.6Sc0.4N patterning was rejected in favor of pure physical Ar Ion Beam Etching (IBE) due to very poor selectivity to photoresist (1:6) and significant etch rate degradation with high Sc doping.
High Scandium concentration AlScN resonators for 5G applications · EPFL
Transport limits and aspect-ratio-dependent etching impede material clearing in narrow gaps
Diffusion-limited reactant penetration and excessive fluid backpressure prevent the complete removal of sacrificial materials in narrow microchannels and trenches. Aspect-ratio-dependent etching rates cause uneven clearing that leads to insoluble complex accumulation or over-etching and breakage of fine features.
Tried and failed
selective wet chemical etching of oxide cladding applied to sub-200 nm slot waveguide gaps. Reason: insufficient etchant penetration or excessive difficulty clearing narrow sub-200 nm high-aspect-ratio trenches
Scalable Packaging and Integration Solutions for Next-Generation Photonic Systems · MIT
Tried and failed
single-run deep reactive-ion etching applied to high-aspect-ratio silicon microstructures with varied gap sizes. Reason: aspect-ratio-dependent etching caused narrow gaps to clear too slowly, leading to over-etching and breakage of fine features
Tried and failed
dry etching of atomic layer deposited alumina applied to bottom contacts in high-aspect-ratio microchannels. Outcome: no signal. Reason: dry etching failed to clear alumina from bottom pads in high aspect ratio channels, blocking electrical contact
Amorphous Silicon Based Microchannel Plates for Time-of-Flight Positron Emission Tomography · EPFL
Tried and failed
injection-based solvent etching of sacrificial material applied to high-resistance heterogeneous microchannel networks. Reason: fluid shunting and excessive backpressure caused structural cracking in high-resistance and mismatched-diameter channels
Negative Enrichment of Circulating Tumor Cells Via 3D Printed Microfluidic Device · Georgia Tech
Tried and failed
continuous single-solvent alkaline wet etching applied to polymer-masked silicon substrates. Reason: poor polymer solvation and accumulation of insoluble potassium-silica complexes at the interface
Bottom-up Process Development for High Throughput Electronic Device Fabrication · Georgia Tech
Tried and failed
sacrificial isotropic gas etching for suspended membranes applied to integrated microelectronic device electrodes. Reason: incomplete sacrificial layer removal left residues and caused stress-induced cracking in the suspended metal film
Vacuum Transistors Based on III-Nitrides Field Emitter Arrays with Self-Aligned Gate · MIT
Catalyst detour and non-uniform density cause morphological failures in metal-assisted chemical etching
Catalysts in metal-assisted chemical etching can wander during vertical descent, causing adjacent high-aspect-ratio holes to merge and destroy nanoscale bridges. Suboptimal etching temperatures and non-uniform catalyst distributions also produce unwanted porous layers and distorted etch depths.
Tried and failed
gas-phase metal-assisted chemical etching applied to silicon micro- and nanostructuring. Reason: sub-optimal temperatures caused porous layer formation and minimal vertical etching rather than directional removal
Tried and failed
metal-assisted chemical etching with thin metal catalyst applied to dense high-aspect-ratio hole arrays. Reason: catalyst detoured during vertical descent causing adjacent nanoscale holes and bridges to merge
Integrated photonic devices for trace gas sensing and spectroscopy · UT Austin
Tried and failed
metal-assisted chemical etching applied to integrated optical waveguides and resonators. Reason: variable catalyst density caused non-uniform etch depths and distorted microstructures
Hybrid germanium-on-zinc selenide platform for photonic integrated circuits in the longwave infrared · UT Austin
Prolonged through-wafer etching causes mask erosion and wafer brittleness
Deep through-wafer etching often erodes masks before the target depth is reached, leading to process failure. Extended wet etching also induces severe wafer brittleness that dramatically reduces manufacturing yield.
Tried and failed
single-sided wet etching through wafer applied to silicon microfluidic device fabrication. Reason: prolonged etching caused severe wafer brittleness and very low manufacturing yield
Tried and failed
single-sided through-wafer deep reactive-ion etching applied to silicon microfluidic channel fabrication. Reason: mask erosion degraded the etch mask before full wafer penetration
Paving the Way for Reliable Electrospray Propulsion through Materials Science and Design · MIT
Considered and rejected
Considered and rejected: Rejected thick photoresist masks (~2 µm) for submicron etching because high-aspect-ratio nanoscale resist features could not achieve sufficient etch resistance.
Scaling In Lithium Niobite: Synaptic Devices for Neuromorphic Computing · Georgia Tech
Left open by the authors
Problems the authors named and did not get to.
Left open
Develop a drift compensation method and refine the soft sensor skin fabrication process to reduce inter-unit repeatability errors. Blocker: Requires multi-material 3D printing hardware, wet etching/lamination lab equipment, and physical sensor fabrication
A data-driven method to reduce excessive contact pressure of hand orthosis using a soft sensor skin · Imperial
Left open
Identify physical failure mechanisms underlying resistance slope changes in Kirigami interconnects during cyclic stretching up to one million cycles. Blocker: Requires fabricated micro-kirigami interconnect samples and physical cyclic stretch-testing apparatus with in situ microscopy or electromechanical characterization.
Left open
Replace toxic HF in MACE etching using low-concentration solutions, KOH, or mechanical roughening to fabricate porous silicon contacts. Blocker: Requires a wet chemistry and semiconductor fabrication lab with chemical etching apparatus.
Left open
Develop a low-electrolyte-concentration cleaning etch to eliminate non-uniform underside surface profiles and residual substrate material on SiC membranes. Blocker: Requires a wet lab, chemical etching equipment, and physical SiC membrane samples.
Integrating Spin Defects with Thin-Film Silicon Carbide Devices · Harvard
Left open
Investigate substrate undercut etching techniques on commercial CMOS nodes to improve thermal dissipation and tuning efficiency of Bragg grating filters. Blocker: Requires access to advanced semiconductor fabrication facilities and cleanroom etching tools.
CMOS Fabrication-Compatible Bragg Grating Devices and Applications · Georgia Tech
Left open
Modify the microfabrication etching process to prevent etching away metal pads and connectors on the photonic chip. Blocker: Requires access to a semiconductor cleanroom and microfabrication wet/dry etching equipment
MEMS Simulations and Waveguide Loss Analysis forPhotonic Integrated Circuits · MIT
Left open
Calibrate silicon dioxide etching to reduce sidewall transfer lithography nanoscale feature widths below 100 nm. Blocker: Requires a semiconductor cleanroom and physical nanofabrication/etching equipment (RIE/DRIE).
STL NEMS fabrication, design, and inspection · Imperial
Left open
Optimize the sacrificial etching process to undercut the InP substrate without damaging QCL and QCD structures. Blocker: Requires a physical semiconductor fabrication cleanroom and wet/dry etching lab equipment
Near- and mid-infrared integrated photonic bio-chemical sensing devices with on-chip spectrometers · UT Austin
Left open
Resolve fabrication obstacles including lagging, notching, knifing, and passivation during high-aspect-ratio DRIE of 4H-SiCOI substrates. Blocker: Requires a semiconductor cleanroom and DRIE etching equipment to conduct physical microfabrication experiments
Applications of New Materials Systems and Machine Learning Techniques for compensation and Calibration of MEMS · Georgia Tech
Left open
Perform selective etching of binary nanosphere monolayers using HF or toluene to tune fill-factor and pitch without reactive ion etching. Blocker: Requires a chemistry wet lab and hazardous chemical handling (HF, toluene) for physical nanosphere etching.
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.