A research team has demonstrated an integrated manufacturing technique capable of producing large-scale solid-state nanopore arrays on silicon wafers while tuning channel geometry through thermally driven atomic migration. Termed the FAST method, the four-stage approach combines nanotemplate creation, thin-film deposition, array pattern formation, and three-dimensional channel shaping to generate thousands of uniform nanopores simultaneously across broad surfaces. Experimental measurements confirmed that the process achieves channel lengths down to four nanometers, while establishing compact distributions packing over twenty-one hundred nanopores within a scanning electron microscope field of view measuring 21.4 square micrometers. Furthermore, the overall patterning area can extend beyond one hundred square millimeters by adjusting particle size and distribution parameters. During subsequent thermal processing, controlled atomic movements induce an asymmetric morphological change where pore tops widen while bottoms narrow into conical geometries, establishing an efficient manufacturing pathway for scalable solid-state nanopore platforms.
Nanopores are nanoscale apertures ranging between one and one hundred nanometers in diameter, serving as vital tools for sensing biomolecules and investigating nanofluidic transport. Commercial sensing instruments have historically relied on biological nanopores consisting of naturally occurring protein channels embedded within lipid membranes. Although commercial instruments deploy assemblies incorporating over two thousand biological channels to enhance throughput, these biological molecules remain fragile, sensitive to environmental conditions, non-reusable, and costly to configure into scalable arrays. Solid-state alternatives formed in engineered materials offer enhanced physical robustness, superior chemical stability, adaptable operational conditions, and direct compatibility with microelectronic manufacturing workflows. Nevertheless, earlier solid-state fabrication has been predominantly confined to single-nanopore protocols executed on serial laboratory instruments. Conventional serial techniques such as focused ion beam milling, electron beam lithography, laser drilling, and controlled dielectric breakdown entail steep instrumentation costs, slow throughput, and tight material restrictions, severely impeding broader commercial adoption of solid-state nanopore technology.
Efforts to fabricate solid-state nanopore arrays have previously achieved limited progress, frequently restricted to sequential demonstrations creating four, nine, or thirty-six pores across localized areas. Existing collective techniques such as electron beam deposition and chemical wet etching routinely struggle to preserve uniform dimensions, leaving broad diameter dispersions that undermine operational consistency. To overcome these serial bottlenecks, the researchers designed the FAST protocol to execute high-throughput collective processing across four systematic stages. The integrated workflow coordinates polystyrene nanosphere template self-assembly, physical vapor deposition of matrix materials, selective template dissolution, and controlled post-deposition thermal annealing. In contrast to conventional serial drilling systems that process single apertures sequentially, this hybrid platform enables parallel generation of ordered pore patterns over large wafer areas. Economic modeling indicates that the collective method could reduce fabrication expenses to roughly ten euros for a five-millimeter square chip, allowing the parallel fabrication of up to two thousand chips in a three-hour run.
The initial phase of the FAST sequence requires constructing an ordered monolayer template using polymeric nanospheres on commercial silicon wafers. Silicon substrates measuring ten millimeters square were first cleaned using ultrasonic agitation in acetone, ethanol, and distilled water for thirty minutes each, followed by a thirty-minute oxygen plasma cleaning step and nitrogen drying. To resolve aggregation difficulties associated with nanospheres under one hundred nanometers in diameter, the team investigated spin coating dynamics across various rotational velocities, acceleration profiles, and dwell intervals. Experimental trials demonstrated that rotational acceleration constitutes the decisive parameter governing structural uniformity: acceleration rates below five hundred revolutions per minute per second produced disordered or multilayer deposits, whereas higher accelerations produced consistent monolayers. Using polystyrene nanospheres with nominal diameters of one hundred nanometers, the team established stable monolayer packing across areas exceeding one hundred square millimeters, yielding a baseline sphere diameter of 98.8 nanometers with a 1.3-nanometer standard deviation.
For dense arrays where nanospheres touch directly, creating open spacing is essential to permit physical vapor deposition around polymer templates without encapsulation. The authors utilized oxygen plasma etching operated at fifty watts to systematically trim the polystyrene spheres prior to thin-film coating. Chemical cleaving driven by reactive oxygen radicals severed polymer chains, gradually shrinking particle diameters while expanding inter-particle gaps. Exposure for sixty seconds reduced mean particle diameters to 81.0 nanometers, ninety seconds reduced them to 74.6 nanometers, and one hundred and twenty seconds achieved 68.9 nanometers, while maintaining tight dimensional tolerances within plus or minus two percent. The relative shrinkage rate slowed progressively over extended etching periods, pointing toward a process threshold beyond which over-etching causes particle deformation and detachment. Following controlled shrinkage, physical vapor deposition deposited silicon dioxide thin films across the prepared templates. To ensure that polymer spheres remained exposed for chemical lift-off, deposition thickness was restricted below half the nanosphere diameter.
Array formation was completed by stripping the embedded polystyrene spheres from the deposited silicon dioxide thin films. Immersing the coated wafers in toluene solvent with ultrasonic assistance for thirty to sixty minutes dissolved the polymer cores cleanly, leaving ordered through-membrane nanopores across the silicon support. Oxygen plasma etching also proved viable for sphere removal. Scanning electron microscopy confirmed that over ninety percent of the treated wafer areas formed highly uniform nanopore distributions characterized by sharp pore boundaries and consistent spatial intervals. High-density arrays fabricated with twenty-nanometer silicon dioxide films on separate deposition tools exhibited mean diameters of 66.6 nanometers and 69.4 nanometers, while a four-nanometer thick membrane yielded average pore diameters of 64.9 nanometers. By diluting the initial nanosphere solutions up to ten or twenty-five times with deionized water, the researchers also produced loose arrays averaging forty-seven nanometers in diameter and ultra-sparse configurations featuring single isolated pores.
Following pattern release, the researchers examined thermal annealing to adjust the three-dimensional geometry of the open silicon dioxide channels. Nanopatterned films were heated in a furnace from room temperature to twelve hundred degrees Celsius for durations of ten, twenty, and thirty minutes, alongside an extended two-and-a-half-hour run, followed by natural cooling. Classical capillary models describe thermal pore evolution through surface free energy minimization driven by viscous flow, which predicts that apertures expand if their radius exceeds half the membrane thickness. Because the experimental pore diameters exceeded twice the film thickness, classical theory dictated uniform lateral enlargement. Instead, scanning electron micrographs revealed an inhomogeneous cross-sectional response along the pore axis: the upper pore mouth expanded outwards while the lower channel constricted, transforming cylindrical cavities into conical geometries. As thermal processing advanced, the bottom aperture closed completely while the top channel widened into a shallow surface crater, leaving only a subtle concave depression after two and a half hours.
To identify the driving forces behind this asymmetric geometric modification, the authors first verified the chemical stability of the annealed thin films. X-ray photoelectron spectroscopy evaluated elemental silicon-to-oxygen ratios before and after high-temperature annealing, confirming that the atomic composition remained constant near a one-to-two stoichiometry and eliminating oxidation-induced volume expansion as a potential cause. The authors then constructed dynamic molecular models within Materials Studio, employing the COMPASS force field and an NVE thermodynamic ensemble across thermal fields reaching thirteen hundred degrees Celsius. The simulations replicated dynamic pore contraction from 10.450 angstroms down to 3.721 angstroms across a one-thousand-picosecond window. Concurrently, surface atoms adjacent to the pore rim collapsed inward to produce a distinct circular depression surrounding the central aperture, closely mirroring experimental observations. The molecular simulations demonstrated that complete pore closure was prohibited in computational space because interatomic repulsion prevented spontaneous covalent bond formation without additional external driving energy.
The investigation attributed the asymmetric channel evolution to directed atomic migration propelled by an energy potential gradient across the pore depth. By tracking probe silicon dioxide molecules interacting with specific atomic sites along the channel wall, the researchers calculated the variation in interaction energies from the top entrance to the lower interface. Computed interaction energies declined monotonically from 5799 kilocalories per mole at the topmost atomic position to 3983, 3623, and 1422 kilocalories per mole at sequential depths toward the bottom boundary. This energy gradient demonstrated that atoms near the upper surface inhabit higher energy states with lower interaction barriers, whereas atoms situated near the underlying substrate remain structurally constrained in lower, more stable energy configurations. Driven by this potential gradient and facilitated by physical substrate restrictions, mobile surface atoms migrate downward along the pore walls toward lower-energy sites, accumulating at the base to close the lower channel while simultaneously excavating the upper entrance into an expanded conical mouth.
Despite demonstrating high repeatability and robust geometric control, the researchers noted key technical boundaries governing the current implementation of the FAST platform. While the approach reliably regulates three-dimensional channel profiles down to complete bottom closure, processing beyond established thermal or etching thresholds increases dimensional variations, with pore diameter inconsistencies reaching twenty-five percent in unoptimized regimes, though optimal conditions maintained dispersion below ten percent. Minor physical defects also concentrated near wafer perimeters due to mechanical stress during wafer cutting, which the team suggests addressing through refined dicing and customized deposition tooling. Downstream applications requiring open nanofluidic access through the support wafer can incorporate selective backside chemical etching to reveal freestanding nanopore channels with customizable lengths. By replacing serial beam drilling with scalable colloidal lithography, thin-film deposition, and directed atomic migration, the FAST methodology establishes a viable pathway toward the cost-effective mass production of structured solid-state membranes for chemical catalysis and nanofluidic filtration devices.
Subject of Research: Scalable fabrication and three-dimensional geometric regulation of solid-state nanopore arrays
Article Title: Fabrication of Large-Scale Nanopore-Array Patterns with Controllable 3D Geometry via Atomic Migration
Article References: Liu, H., Cui, F., Fu, R., & Zhang, J. (2025). Fabrication of Large-Scale Nanopore-Array Patterns with Controllable 3D Geometry via Atomic Migration. Nanomanufacturing and Metrology, 8(1), Article 30. https://doi.org/10.1007/s41871-025-00279-4
Image Credits: AI Generated
DOI: 10.1007/s41871-025-00279-4
Keywords: Nanomanufacturing, Solid-State Nanopores, Nanopore Arrays, FAST Process, Atomic Migration, Spin Coating, Polystyrene Nanospheres, Oxygen Plasma Etching, Physical Vapor Deposition, Silicon Dioxide Thin Films, Thermal Annealing, Conical Nanopores, Molecular Dynamics Simulation, Materials Studio, Surface Free Energy, Substrate Restriction, Silicon Wafers, Thin Film Metrology
Cite Scienmag News
Denise Maddox. (November 21, 2026). Hybrid Manufacturing Method Forms Scalable Nanopore Arrays Using Directed Atomic Migration. Scienmag. https://scienmag.com/hybrid-manufacturing-method-forms-scalable-nanopore-arrays-using-directed-atomic-migration/
Denise Maddox. "Hybrid Manufacturing Method Forms Scalable Nanopore Arrays Using Directed Atomic Migration." Scienmag, 21 November 2026, https://scienmag.com/hybrid-manufacturing-method-forms-scalable-nanopore-arrays-using-directed-atomic-migration/. Accessed 21 November 2026.
Denise Maddox. "Hybrid Manufacturing Method Forms Scalable Nanopore Arrays Using Directed Atomic Migration." Scienmag. November 21, 2026. https://scienmag.com/hybrid-manufacturing-method-forms-scalable-nanopore-arrays-using-directed-atomic-migration/








