9.6 Nanostructure fabrication techniques

Fundamentals of nanostructure fabrication

Nanostructure fabrication is the set of techniques used to build and pattern materials at the nanometer scale. These methods give researchers direct control over the size, shape, and composition of structures where quantum effects dominate, making them essential for both fundamental condensed matter research and device development.

The core challenge is bridging theory and experiment: theoretical predictions about low-dimensional physics only become testable when you can actually build the structures with sufficient precision.

Scale and dimensionality

Nanostructures typically range from 1 to 100 nm, a regime where dimensions become comparable to the de Broglie wavelength of charge carriers. At this scale, quantum confinement reshapes electronic, optical, and magnetic properties depending on how many dimensions are restricted:

• 0D structures (quantum dots) confine carriers in all three dimensions, producing discrete, atom-like energy levels
• 1D structures (nanowires, nanotubes) confine in two dimensions, leading to quantized conductance
• 2D structures (graphene, thin quantum wells) confine in one dimension, giving rise to unique band structures and transport behavior

As any dimension shrinks toward the de Broglie wavelength, confinement effects grow stronger and bulk approximations break down.

Material selection criteria

Choosing the right material is not just about the physics you want to study. You also need to consider practical fabrication constraints:

• Chemical composition sets the electronic band structure, optical response, and surface reactivity
• Crystalline structure determines symmetry-related properties and how defects form
• Process compatibility matters: the material must withstand the etching, deposition, or thermal steps involved
• Stability under operating conditions (thermal, mechanical, chemical) is required for reliable devices
• Scalability and cost become important if you ever want to move beyond single lab samples

Top-down vs. bottom-up approaches

These are the two broad strategies for making nanostructures, and most real fabrication workflows combine elements of both.

Top-down methods start with bulk material and carve features out of it:

• Rely on lithography to define patterns and etching to remove material
• Offer precise control over where features go and how large they are
• Resolution is ultimately limited by the lithography tool and the physics of material removal

Bottom-up methods build structures from atomic or molecular precursors:

• Include self-assembly, chemical synthesis, and template-assisted growth
• Can produce atomically precise features and complex 3D architectures
• The main difficulty is achieving large-scale order and integrating structures into devices

Lithography techniques

Lithography is the primary way patterns are defined at the nanoscale. Nearly every fabricated nanodevice starts with some form of lithographic patterning, making these techniques foundational to the field.

Photolithography principles

Photolithography transfers a geometric pattern from a photomask onto a light-sensitive photoresist coating on a substrate. The basic process follows these steps:

1. Spin-coat a thin layer of photoresist onto the substrate
2. Expose the resist to UV light through a patterned photomask
3. Develop the resist, dissolving away either the exposed regions (positive resist) or the unexposed regions (negative resist)
4. Transfer the pattern into the underlying material via etching or deposition
5. Strip the remaining resist

Resolution is fundamentally limited by the wavelength of light used. Modern deep-UV systems operate at 193 nm, and extreme ultraviolet (EUV) lithography pushes to 13.5 nm. Resolution enhancement techniques like phase-shifting masks and optical proximity correction extend the practical limits of each wavelength generation.

Electron beam lithography

Electron beam (e-beam) lithography is a direct-write technique: a focused electron beam scans across an electron-sensitive resist to draw patterns without a mask.

• Achieves sub-10 nm resolution, well beyond the limits of optical lithography
• Maskless operation makes it ideal for rapid prototyping and one-off designs
• The serial writing process is inherently slow, so throughput is low compared to optical methods
• Proximity effects arise because electrons scatter in the resist and substrate, broadening the exposed region. Correction algorithms adjust the dose pattern to compensate

E-beam lithography is the go-to tool for research-scale nanostructure fabrication where resolution matters more than speed.

X-ray lithography

X-ray lithography uses short-wavelength X-rays (0.4 to 4 nm) to expose resist, offering several advantages over optical methods:

• Produces high aspect ratio structures with nearly vertical sidewalls
• Less susceptible to diffraction-limited blurring than UV lithography
• Requires a synchrotron or similar high-brightness X-ray source
• Masks are complex to fabricate: thin membranes patterned with heavy-metal absorbers

The need for specialized sources and masks has limited widespread adoption, but the technique remains valuable for specific high-resolution, high-aspect-ratio applications.

Soft lithography methods

Soft lithography replaces rigid optical components with elastomeric stamps and molds, most commonly made from PDMS (polydimethylsiloxane). Several variants exist:

• Microcontact printing uses a PDMS stamp to transfer self-assembled monolayers (SAMs) onto a substrate
• Replica molding cures a polymer inside a PDMS mold to create 3D structures
• Capillary force lithography exploits surface tension to draw material into nanoscale mold features

These methods work on non-planar surfaces and with biological materials, and they're cost-effective for large-area patterning. Resolution is typically coarser than e-beam lithography but sufficient for many applications.

Thin film deposition

Thin film deposition builds up material layers with controlled thickness and composition. These techniques are essential for creating quantum wells, heterostructures, tunnel barriers, and other layered systems central to condensed matter research.

Physical vapor deposition

Physical vapor deposition (PVD) transfers material from a source to a substrate through a vacuum. The main variants differ in how they generate the vapor:

• Thermal evaporation heats the source resistively or with an electron beam until it vaporizes
• Sputtering bombards a target with energetic ions (usually Ar$^+$), ejecting atoms that deposit on the substrate
• Pulsed laser deposition (PLD) uses high-power laser pulses to ablate material from a target

PVD works with metals, semiconductors, and insulators. Film properties depend on deposition rate, substrate temperature, and chamber pressure.

Chemical vapor deposition

Chemical vapor deposition (CVD) forms films through chemical reactions of gas-phase precursors at or near the substrate surface.

• Thermal CVD drives reactions with heat, typically at atmospheric or reduced pressure
• Plasma-enhanced CVD (PECVD) uses a plasma to activate precursors, allowing deposition at lower substrate temperatures
• Metalorganic CVD (MOCVD) uses organometallic precursors and is the standard technique for growing compound semiconductors (III-V materials like GaAs, InP)

A key advantage of CVD is conformal coverage: films coat complex 3D topographies uniformly, which is difficult to achieve with line-of-sight PVD methods.

Atomic layer deposition

Atomic layer deposition (ALD) achieves the most precise thickness control of any deposition technique. The process works in a cycle:

1. Expose the substrate to precursor A, which chemisorbs as a single monolayer (self-limiting)
2. Purge excess precursor A and byproducts
3. Expose to precursor B, which reacts with the adsorbed layer to form the desired film material
4. Purge again
5. Repeat the cycle; each cycle adds roughly one atomic layer of material

Because each step is self-limiting, ALD produces highly conformal, pinhole-free films even on structures with extreme aspect ratios. It's widely used for depositing high-$\kappa$ dielectrics (like $\text{HfO}_2$), diffusion barriers, and protective coatings.

Molecular beam epitaxy

Molecular beam epitaxy (MBE) grows epitaxial films with atomic-layer precision under ultra-high vacuum ($< 10^{-10}$ torr).

• Thermal beams of atoms or molecules are directed at a heated crystalline substrate
• RHEED (reflection high-energy electron diffraction) monitors the surface in real time, allowing you to track growth layer by layer
• Produces atomically abrupt interfaces, which is critical for quantum wells and superlattices
• Growth rates are slow (roughly 1 monolayer per second), but the quality is unmatched

MBE is the technique of choice for fabricating the heterostructures used to study 2D electron gases, fractional quantum Hall states, and other low-dimensional phenomena.

Etching processes

Etching selectively removes material to define nanostructure geometry. The three key parameters to control are etch rate, selectivity (how much faster the target material etches compared to the mask or underlying layers), and anisotropy (directional vs. uniform removal).

Wet etching techniques

Wet etching immerses the substrate in a liquid chemical etchant. It comes in two flavors:

• Isotropic wet etching removes material at equal rates in all directions, producing rounded, undercut profiles
• Anisotropic wet etching exploits crystal-plane-dependent etch rates. For example, KOH etches Si{100} planes roughly 100× faster than Si{111} planes, creating well-defined V-grooves and pyramidal pits

Wet etching offers high selectivity and low equipment cost, but the undercutting in isotropic etching makes it difficult to control features below a few hundred nanometers.

Dry etching methods

Dry etching uses gas-phase species or plasma instead of liquid chemicals. The mechanisms range from purely physical to purely chemical:

• Ion milling (physical): energetic noble gas ions sputter material away. Highly anisotropic but non-selective
• Plasma etching (chemical): reactive radicals from a plasma react with the surface. More selective but tends to be isotropic
• Reactive ion etching (combined): uses both mechanisms simultaneously for anisotropic, selective etching

Dry etching is essential for transferring lithographic patterns into substrates with high fidelity, especially for features requiring vertical sidewalls and high aspect ratios.

Reactive ion etching

Reactive ion etching (RIE) deserves its own discussion because it's the workhorse of nanofabrication pattern transfer.

In RIE, a plasma generates reactive chemical species while an electric field accelerates ions toward the substrate. The directional ion bombardment enhances chemical reactions on horizontal surfaces (facing the ion flux) while leaving vertical sidewalls relatively untouched. This produces anisotropic etch profiles.

Key process parameters you can tune:

• Gas composition determines which materials are etched and the chemistry of volatile byproducts
• Chamber pressure affects the mean free path of ions and thus their directionality
• RF power controls ion energy and plasma density

Deep reactive ion etching (DRIE), using alternating etch and passivation cycles (the Bosch process), achieves aspect ratios exceeding 20:1 for MEMS/NEMS fabrication.

Focused ion beam milling

Focused ion beam (FIB) milling is a direct-write subtractive technique. A finely focused beam of ions (typically $\text{Ga}^+$) sputters material from the surface with sub-10 nm precision.

• Enables site-specific modification: you can cut, thin, or reshape individual nanostructures
• Simultaneous imaging (using secondary electrons from ion impact) allows real-time monitoring
• Gas-assisted FIB introduces reactive precursors to enhance etch selectivity or enable local deposition
• Commonly used for preparing TEM cross-sections and editing prototype devices

The main drawback is $\text{Ga}^+$ implantation, which can damage or dope the sample in the milled region.

Self-assembly techniques

Self-assembly exploits the natural tendency of certain components to organize into ordered structures through intrinsic interactions. This bottom-up approach can produce features at length scales difficult to reach with top-down lithography, and it does so in parallel across large areas.

Block copolymer self-assembly

Block copolymers consist of two or more chemically distinct polymer chains bonded together. Because the blocks are immiscible, they undergo microphase separation into nanoscale domains.

• Depending on the volume fraction of each block, the morphology can be spheres, cylinders, lamellae, or bicontinuous networks
• The domain spacing (typically 5 to 50 nm) scales with molecular weight
• Directed self-assembly (DSA) uses lithographically defined chemical or topographic guides to control the orientation and registration of the domains
• Applications include nanolithography masks, nanoporous membranes, and photonic crystals

Colloidal self-assembly

Colloidal particles (typically polymer or silica spheres, 10 nm to 1 μm) can organize into ordered arrays driven by interparticle forces: van der Waals attraction, electrostatic repulsion, and capillary forces during drying.

• Assembly methods include convective assembly (controlled evaporation), spin-coating, and electrophoretic deposition
• Produces 2D and 3D colloidal crystals that function as photonic crystals, plasmonic arrays, or templates for further fabrication
• These "artificial solids" also serve as model systems for studying phase transitions and collective behavior

DNA-guided assembly

DNA's programmable base-pairing (A-T, G-C) makes it a powerful tool for directing nanoscale assembly with sequence-level specificity.

• DNA origami folds a long single-stranded scaffold (typically ~7,000 bases) into a designed 2D or 3D shape using hundreds of short "staple" strands
• DNA tiles and bricks are modular building blocks that assemble into larger periodic or aperiodic structures
• Nanoparticles functionalized with complementary DNA strands can be positioned with nanometer precision
• Applications span plasmonic device engineering, molecular computing, and biosensing

Langmuir-Blodgett films

Langmuir-Blodgett (LB) deposition creates ordered molecular monolayers with precise control over packing and thickness.

1. Spread amphiphilic molecules on a water surface (the Langmuir trough)
2. Compress the monolayer with movable barriers to control molecular packing density
3. Dip a substrate through the compressed monolayer to transfer it as a uniform film
4. Repeat dipping to build up multilayer structures, one monolayer at a time

LB films are used to fabricate organic electronic devices, chemical sensors, and biomimetic membranes, and they provide model systems for studying 2D phase transitions.

Nanoimprint lithography

Nanoimprint lithography (NIL) physically stamps a pattern into a resist, bypassing the diffraction limits that constrain optical lithography. It combines high resolution (sub-10 nm) with high throughput, making it attractive for scaling nanostructure fabrication beyond the lab.

Thermal nanoimprint

1. Press a rigid mold (typically Si or quartz with nanoscale features) into a thermoplastic polymer resist
2. Heat the resist above its glass transition temperature ($T_g$) so it flows and fills the mold cavities
3. Cool below $T_g$ to solidify the pattern
4. Separate the mold from the substrate (demolding)
5. Remove the thin residual layer at the bottom of the imprinted features (usually by a short etch step)

Thermal NIL achieves sub-10 nm resolution over large areas. Challenges include thermal expansion mismatch between mold and substrate, mold wear over repeated cycles, and ensuring complete cavity filling.

UV-assisted nanoimprint

UV nanoimprint uses a transparent mold (quartz) and a liquid, UV-curable resist:

1. Bring the mold into contact with the liquid resist, which fills cavities by capillary action
2. Expose through the mold with UV light to crosslink (cure) the resist
3. Separate the mold, leaving the solidified pattern

Because it operates at room temperature, UV-NIL avoids thermal expansion issues and requires lower imprint pressures. It's widely used for patterning optical gratings, photonic crystals, and waveguides.

Roll-to-roll nanoimprint

Roll-to-roll (R2R) NIL adapts the nanoimprint concept for continuous processing on flexible substrates:

• A cylindrical mold or patterned sleeve imprints onto a moving web of material
• Enables fabrication of large-area nanostructured films at high speed
• Applications include flexible electronics, anti-reflective coatings for solar cells, and optical security features
• Maintaining pattern fidelity over meters of substrate while controlling web tension remains an engineering challenge

Scanning probe lithography

Scanning probe lithography (SPL) uses the nanometer-scale tip of a scanning probe microscope to write patterns directly on a surface. These techniques bridge top-down and bottom-up approaches, offering the ability to manipulate individual atoms and molecules.

Dip-pen nanolithography

Dip-pen nanolithography (DPN) uses an AFM tip coated with molecular "ink" to write patterns on a surface:

• A water meniscus forms naturally between the tip and substrate in ambient conditions
• Molecules transport from the tip through the meniscus to the surface
• Feature size (down to ~50 nm) depends on tip-substrate contact time, humidity, and ink properties
• Can deposit a wide range of materials: small molecules, polymers, proteins, nanoparticle suspensions
• Parallel tip arrays increase throughput for larger-area patterning

Scanning tunneling microscopy lithography

STM lithography exploits the atomically sharp tip and precise positioning of a scanning tunneling microscope:

• Atomic manipulation moves individual adsorbed atoms by adjusting the tip-sample distance and bias, famously demonstrated by spelling "IBM" with xenon atoms on nickel
• Field-induced deposition uses the intense electric field at the tip apex to decompose precursor gas molecules, depositing material locally
• Local oxidation applies a voltage bias in humid conditions to grow nanoscale oxide lines on semiconductor surfaces

These capabilities make STM lithography uniquely suited for creating and studying single-atom devices and quantum corrals.

Atomic force microscopy lithography

AFM lithography encompasses several tip-based patterning modes:

• Mechanical scribing physically scratches or indents the surface to create grooves and pits
• Local anodic oxidation uses the water meniscus between tip and surface as a nanoscale electrolyte cell, growing oxide lines (e.g., on Si or Ti) with widths below 20 nm
• Thermomechanical writing heats the tip to locally melt or decompose a polymer film, as demonstrated in IBM's "Millipede" data storage concept
• Bias-induced phase changes can switch materials between different structural or electronic phases at the nanoscale

Template-assisted synthesis

Template-assisted methods use a pre-existing structure with nanoscale features to guide the formation of new nanostructures. The template defines the geometry; you fill it, coat it, or etch through it.

Anodic aluminum oxide templates

Anodic aluminum oxide (AAO) templates are made by electrochemically anodizing aluminum foil, which spontaneously forms a self-ordered array of nanopores.

• Pores arrange in a hexagonal lattice with diameters tunable from ~10 to 400 nm
• Pore depth can reach 100 μm or more, giving very high aspect ratios
• Pore diameter and spacing are controlled by anodization voltage, electrolyte type (sulfuric, oxalic, or phosphoric acid), and temperature
• Filling the pores by electrodeposition or CVD produces arrays of nanowires or nanotubes
• Applications include magnetic storage media, thermoelectric devices, and photonic crystals

Nanosphere lithography

Nanosphere lithography (NSL) uses a close-packed monolayer of colloidal spheres as a deposition or etch mask:

1. Self-assemble a monolayer of polystyrene or silica spheres on a substrate
2. Deposit material (e.g., by evaporation) through the gaps between spheres
3. Remove the spheres, leaving a periodic array of triangular nanoparticles at the former interstitial sites

Using the sphere layer as an etch mask instead produces arrays of nanoholes. Double-layer sphere masks create more complex geometries. NSL is a simple, low-cost route to large-area periodic nanostructure arrays for plasmonics and surface-enhanced Raman spectroscopy (SERS).

Porous silicon templates

Porous silicon is formed by electrochemical etching of crystalline silicon in HF-based solutions:

• Pore morphology ranges from microporous (<2 nm) to macroporous (>50 nm), controlled by HF concentration, current density, and silicon doping level
• The porous network serves as a template for depositing or growing other materials inside the pores
• Silicon's compatibility with standard microelectronics processing is a major advantage
• Applications include photonic crystals, thermoelectric materials, lithium-ion battery anodes (where the porous structure accommodates volume expansion during cycling), and drug delivery platforms

Characterization of nanostructures

Fabrication without characterization is flying blind. You need to verify that your structures have the intended dimensions, composition, crystal structure, and electronic properties. The techniques below provide that feedback.

Electron microscopy techniques

• SEM images surface topography by scanning a focused electron beam and collecting secondary or backscattered electrons. Resolution reaches ~1 nm in modern instruments
• TEM transmits electrons through a thin specimen (<100 nm thick), providing atomic-resolution images and diffraction patterns for crystal structure analysis
• STEM combines scanning with transmission, enabling high-resolution imaging alongside elemental mapping via energy-dispersive X-ray spectroscopy (EDX)
• EELS (electron energy loss spectroscopy) in a TEM/STEM probes electronic structure, chemical bonding, and dielectric properties with nanometer spatial resolution
• FIB-SEM dual-beam systems use a focused ion beam to cut cross-sections for SEM imaging or to prepare thin lamellae for TEM

Scanning probe microscopy

• AFM maps surface topography by measuring forces between a sharp tip and the sample, achieving sub-nanometer height resolution
• STM tunnels electrons between a conductive tip and sample, imaging the local density of states at atomic resolution (works only on conductive surfaces)
• KPFM (Kelvin probe force microscopy) measures local surface potential and work function variations
• MFM (magnetic force microscopy) detects magnetic stray fields to image domain structures in nanomagnets
• Conductive AFM and scanning capacitance microscopy probe local electrical transport and capacitance, useful for mapping doping profiles and leakage paths

X-ray diffraction methods

• XRD identifies crystal phases and measures lattice parameters from Bragg peak positions
• GIXRD (grazing incidence XRD) enhances surface sensitivity by using a shallow incident angle, making it well suited for thin films and nanoparticle layers
• SAXS (small-angle X-ray scattering) probes structures at the 1 to 100 nm scale, providing information on particle size, shape, and spatial correlations
• XRR (X-ray reflectivity) measures thin film thickness, density, and interface roughness from interference fringes in the reflected intensity
• Synchrotron sources enable time-resolved and in-situ studies of nanostructure growth

Optical spectroscopy

• Photoluminescence (PL) spectroscopy probes radiative recombination, revealing electronic energy levels and defect states
• Raman spectroscopy measures inelastic light scattering from vibrational modes, providing information on crystal structure, strain, and composition
• UV-visible absorption spectroscopy determines optical bandgaps and plasmonic resonance positions
• FTIR identifies chemical functional groups through infrared absorption
• Time-resolved spectroscopy (e.g., ultrafast pump-probe) tracks carrier dynamics and energy transfer on femtosecond to nanosecond timescales

Applications in condensed matter physics

The fabrication techniques covered above aren't ends in themselves. They're tools for building the systems that let you test condensed matter theory and develop new devices.

Quantum dots and wires

Quantum dots confine electrons in all three dimensions, producing discrete energy levels analogous to artificial atoms. Quantum wires confine in two dimensions, and their conductance is quantized in units of $\frac{2e^2}{h}$.

• These structures enable the study of single-electron charging (Coulomb blockade), quantum coherence, and entanglement
• Size-tunable photoluminescence in quantum dots (smaller dots emit bluer light due to stronger confinement) is used in displays, biological imaging, and single-photon sources
• Semiconductor quantum dots are leading candidates for solid-state qubits in quantum computing

Plasmonic nanostructures

When light hits a metallic nanostructure (typically Au or Ag), it can excite collective oscillations of conduction electrons called surface plasmon resonances. These resonances concentrate electromagnetic fields into volumes far smaller than the diffraction limit.

• The resonance wavelength depends on particle size, shape, and local dielectric environment
• Near-field enhancement factors of $10^2$ to $10^4$ boost signals in surface-enhanced Raman spectroscopy (SERS)
• Applications include label-free biosensing, photocatalysis, and nanoscale optical circuits
• Arranged into periodic arrays, plasmonic nanostructures form the building blocks of metamaterials

Metamaterials

Metamaterials are artificially structured composites whose properties arise from the geometry of their subwavelength building blocks (meta-atoms) rather than their chemical composition.

• By engineering the shape, size, and arrangement of meta-atoms, you can achieve effective electromagnetic parameters not found in natural materials (e.g., negative refractive index)
• Applications include superlenses that beat the diffraction limit, electromagnetic cloaking devices, and perfect absorbers
• The concept extends beyond electromagnetics to acoustic and mechanical metamaterials
• Fabrication requires the lithographic and self-assembly techniques discussed earlier, often pushing resolution limits

Nanoelectronic devices

• Single-electron transistors exploit Coulomb blockade to switch with the addition or removal of a single electron, enabling ultra-low-power logic
• Resonant tunneling diodes use quantum well structures where electrons tunnel through double barriers, producing negative differential resistance useful for high-frequency oscillators
• Carbon nanotube and graphene devices explore ballistic transport, Klein tunneling, and valley-dependent physics
• Spintronic devices use electron spin rather than charge for information processing, with applications in magnetic memory (MRAM) and spin-transfer torque oscillators
• Molecular electronics investigates charge transport through single molecules contacted by nanoscale electrodes