Focused Ion Beam Induced Deposition

• FIBID is a direct-write, maskless nanofabrication process where a focused ion beam induces precursor decomposition to deposit functional materials locally.
• The technique employs primary ions, secondary electrons, and excited surface atoms to control deposition versus sputtering and determine composition and morphology.
• FIBID enables precise nanoscale patterning for applications in nanoelectronics, superconducting devices, catalyst placement, and advanced AFM probe fabrication.

Focused Ion Beam Induced Deposition (FIBID) is a direct-write, maskless nanofabrication technique in which a focused ion beam, typically in the keV energy range, is scanned over a surface supplied with a gaseous precursor through a gas injection system, so that localized ion-stimulated decomposition leaves non-volatile fragments as a deposit while volatile byproducts are pumped away. In contrast to the more familiar subtractive use of focused ion beams, FIBID is an additive process that enables site-selective deposition of metals, oxides, and related functional materials, and it now spans electrical contacting, superconducting nanostructures, catalyst placement, AFM-tip fabrication, and three-dimensional heterostructure nanoprinting (Höflich et al., 2023).

1. Fundamental process and growth physics

The elementary picture of FIBID is the interaction of adsorbed precursor molecules with the primary ion beam and the secondary species generated by the ion-solid collision cascade. A generic representation given for the process is

$$\text{adsorbate} + (\text{PI}, \text{SE}, \text{ESA}) \xrightarrow{\text{FIBID}} \text{deposit fragments} + \text{volatile fragments},$$

where PI denotes the primary ion, SE secondary electrons, and ESA excited surface atoms (Höflich et al., 2023). This framing is important because FIBID is not governed solely by direct ion impact; secondary electrons and excited surface atoms also contribute to precursor dissociation.

The process is intrinsically kinetic. The surface coverage $\Theta$ is described by the rate equation

$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$

with $\nu_{\rm gas}$ the impingement rate, $\nu_{\rm des}$ the desorption rate, $\nu_{\rm dis}$ the dissociation rate, and $D$ the surface diffusion coefficient. The deposition or etch flux is given by

$$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$

and the local dissociation rate by

$$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$

These expressions make explicit that net growth is controlled by precursor supply, surface transport, and beam-induced dissociation, while competing sputtering can reduce or even reverse net deposition (Höflich et al., 2023).

That competition between deposition and sputtering is central to FIBID. In practical terms, the beam not only decomposes precursor molecules but also removes deposited and substrate atoms. A recurring theme across the literature is therefore that morphology, composition, and functional performance emerge from a balance among dissociation efficiency, ion implantation, sputter yield, and the volatile or non-volatile character of the precursor fragments.

2. Ion species, sources, and precursor chemistry

Commercial and research FIB platforms use multiple ion species. The roadmap identifies Ga$^+$ as the most widely used ion because of mature instrument availability, while He$\Theta$0 and Ne$\Theta$1 from Gas Field Ion Sources, as well as Xe$\Theta$2 and more specialized Si$\Theta$3 and Co$\Theta$4 sources, extend the accessible resolution and interaction regimes (Höflich et al., 2023). Ion choice is not a secondary detail: it directly alters interaction volume, sputter yield, implantation, backscatter, and contamination.

The precursor space is similarly diverse. Common examples summarized for FIBID include MeCpPtMe$\Theta$5 for Pt, Au(hfac)Me$\Theta$6 for Au, W(CO)$\Theta$7 for W, Co$\Theta$8(CO)$\Theta$9 for Co, Nb(NMe$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$0)$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$1(N-t-Bu) for Nb, TMOS for SiO$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$2, and phenanthrene and related compounds for carbonaceous or polymeric structures (Höflich et al., 2023). Several case studies illustrate how this chemistry maps onto device classes. Trimethylplatinum, $$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$3, was used for Pt deposition with both Rb$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$4 and Ga$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$5 beams (Li et al., 2022). W(CO)$$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$6 was used for amorphous superconductors, W-based inducers and leads, W-C nanoSQUIDs, and organometallic AFM tips (Weirich et al., 2014). TEOS was used to deposit silicon oxide catalyst pads for single-walled carbon nanotube growth (Sadki et al., 2018).

Helium and neon FIBID have introduced a qualitatively different materials regime. He-FIB offers sub-nanometer probe sizes of approximately $$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$7 nm, enabling nanowires down to $$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$8 nm diameter and pillars as slender as $$\frac{\partial\Theta}{\partial t} = \nu_{\rm gas}(1-\Theta) - \nu_{\rm des}\Theta - \nu_{\rm dis}\Theta + D\nabla^2\Theta,$$9 nm in width but microns in height; Ne-FIB has a slightly larger probe size of approximately $\nu_{\rm gas}$0 nm but higher sputter yield, which is advantageous for post-growth trimming and milling (Allen, 4 Oct 2025). Because He and Ne are gases, their use avoids metallic contamination of insulating structures, and the lower backscatter of He ions reduces halo effects. At the same time, Ga-FIBID remains highly effective for metallic and superconducting nanodevices, and in some cases can yield higher metal content than lighter ions for the same precursor family (Allen, 4 Oct 2025).

3. Process metrology, compositional analysis, and precursor screening

A major development in FIBID methodology is the move from post hoc characterization toward integrated process analytics. A 2024 pathway for precursor evaluation uses supported thick precursor layers on Si(111), real-time SEM backscattered-electron monitoring during Ga-FIB irradiation, SEM-EDX hyperspectral mapping, and machine-learning decomposition of the hyperspectral data by Non-Negative Matrix Factorization (NMF) (Jany et al., 2024). In this workflow, the BSE intensity tracks metal enrichment because image contrast is proportional to the mean atomic number. The “sputtering point” is identified as the point of maximal metal enrichment before sputtering dominates.

At that optimal point, EDX hyperspectral data are acquired and decomposed by NMF into chemically distinct components corresponding to substrate and deposit. The NMF-separated EDX signal is then quantified by the EDX ZAF standardless protocol to obtain atomic percentages, while performance metrics include metal content, Ga content, carbon and oxygen fractions, deposited volume per dose, and the normalized fluence per height,

$\nu_{\rm gas}$1

together with the Precursor Score,

$\nu_{\rm gas}$2

This framework was proposed specifically to pre-screen and rank candidate precursors for FIBID and FEBID without immediately resorting to full gas-injection-system development or laborious TEM-based analysis (Jany et al., 2024).

The importance of ion-species-dependent process analytics is illustrated by a direct comparison of Pt FIBID with Rb$\nu_{\rm gas}$3 and Ga$\nu_{\rm gas}$4 ions under similar beam energies and currents (Li et al., 2022).

Property	Rb$\nu_{\rm gas}$5 FIBID-Pt	Ga$\nu_{\rm gas}$6 FIBID-Pt
Beam conditions	8.5 keV, 7.0 pA	8.0 keV, 8.5 pA
Deposition rate $\nu_{\rm gas}$7	$\nu_{\rm gas}$8 $\nu_{\rm gas}$9m$\nu_{\rm des}$0/nC	$\nu_{\rm des}$1 $\nu_{\rm des}$2m$\nu_{\rm des}$3/nC
Composition	C:O:Pt:Rb = 25:20:49:5	C:O:Pt:Ga = 22:14:37:27
Resistivity $\nu_{\rm des}$4	$\nu_{\rm des}$5 $\nu_{\rm des}$6cm	$\nu_{\rm des}$7 $\nu_{\rm des}$8cm
Grain size	$\nu_{\rm des}$9 nm	$\nu_{\rm dis}$0 nm

For these Pt deposits, the deposition rate was defined as

$\nu_{\rm dis}$1

and the resistivity as

$\nu_{\rm dis}$2

The central result is that although the Rb$\nu_{\rm dis}$3 deposit had higher Pt content and much lower primary-ion content than the Ga$\nu_{\rm dis}$4 deposit, its resistivity was more than an order of magnitude higher. The reported explanation is that the Rb$\nu_{\rm dis}$5-deposited films contained somewhat higher C and O impurities, and the higher C and O content caused the increased resistivity (Li et al., 2022). This directly corrects a common simplification: lower implanted-ion contamination does not, by itself, guarantee superior electrical transport.

4. Electronic transport, disorder, and superconductivity

FIBID has become an established route to superconducting and metallic nanostructures, but the transport properties are strongly conditioned by disorder and by the exact non-metal residue incorporated during growth. In the amorphous superconductor $\nu_{\rm dis}$6 grown by Ga-FIBID from $\nu_{\rm dis}$7, the maximum critical temperature was $\nu_{\rm dis}$8 K, resistivities were in the range $\nu_{\rm dis}$9–$D$0 $D$1cm, and all samples exhibited a negative $D$2 in the normal state between $D$3 and $D$4 K, indicating localization and proximity to a disorder-induced metal-insulator transition (Weirich et al., 2014). The upper critical field was analyzed with

$D$5

A feedback-controlled optimization based on in-situ conductance monitoring and a genetic algorithm identified beam parameters that maximized conductivity during growth. The optimized sample showed the highest $D$6 of $D$7 K and the sharpest superconducting transition (Weirich et al., 2014).

A related but distinct use of FIBID is the direct writing of superconducting electrodes and voltage probes for hybrid transport experiments. In proximity-effect measurements on Cu and Co nanowires, amorphous W “inducer” electrodes fabricated by Ga-FIBID from $D$8 had composition W: 47 at.%, O: 8 at.%, C: 30 at.%, Ga: 16 at.%, with superconducting transition temperatures from $D$9 to $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$0 K and upper critical fields up to $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$1 T. Pt voltage leads deposited by FIBID had Pt: 32 at.%, O: 5 at.%, C: 53 at.%, Ga: 10 at.% and remained metallic rather than superconducting (Kompaniiets et al., 2013). This geometry enabled spatially resolved measurements over $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$2–$$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$3 $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$4m, from which a superconducting proximity length of about $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$5 $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$6m at low temperatures was inferred for crystalline Cu and Co. In polycrystalline Co, the long-range proximity effect remained insensitive to magnetic fields up to $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$7 T, which the authors interpreted as indicative of spin-triplet pairing (Kompaniiets et al., 2013).

FIBID has also been used to fabricate complete superconducting quantum devices. Ga$$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$8 FIBID of W-C nanoSQUIDs from $$F_{\rm dep,etch} = m \cdot \nu_{\rm dis} \cdot \Theta \cdot n_s,$$9 produced resist-free direct-write Dayem-bridge devices with a loop area of $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$0 and $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$1 nm-wide nanowires serving as Josephson junctions, written in less than $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$2 minutes (Sigloch et al., 2022). The devices showed $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$3 K, $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$4A at $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$5 K, $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$6, and flux-to-voltage transfer coefficients up to $$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$7. The current-flux relation was written as

$$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$8

with screening parameter

$$\nu_{\rm dis}({\bf x}) = \int_0^\infty \sigma(E) f(E,{\bf x}) \, dE.$$9

and the transduction figure of merit as

$^+$0

The reported correlation was that the very high transfer coefficient tracked the high normal-state resistivity of W-C (Sigloch et al., 2022).

5. Three-dimensional nanoprinting and specialized architectures

Recent work has expanded FIBID from planar direct writing to three-dimensional nanoprinting of heterostructures and architected internal morphologies. Using He- and Ne-FIBs, sequential switching between W(CO)$^+$1 and PMCPS enabled in-situ fabrication of metal-dielectric heterostructures, including metallic-insulator-metallic trilayers, metallic nanopillars inside dielectric nanosleeves, dielectric sleeves around metallic pillars, and capped hybrid nanopillars (Allen, 4 Oct 2025). Patterning modes such as spot, annulus, and filled circle were used to create rings, wells, sleeves, and pillars.

A particularly distinctive capability of gaseous-ion FIBID is the formation of engineered internal voids. He-FIBID produced hollow-core nanopillars through concurrent focused milling during deposition, while increased ion dose generated internal nanovoids whose depth depended on ion species: larger singular voids at greater depths for He, and distributed porosity closer to the surface for Ne, consistent with the different implantation depths and lateral scattering of the ions (Allen, 4 Oct 2025). The same study reported isotope-resolved satellite deposition in Ne-FIBID, where a main nanopillar from $^+$2Ne and a smaller adjacent “satellite” from $^+$3Ne reflected the natural abundance ratio of approximately $^+$4.

Advanced microscopy was used to resolve the resulting microstructures. STEM-XEDS comparisons of W(CO)$^+$5 deposits showed that W content increased with ion mass in the order Ga $^+$6 Ne $^+$7 He, while 4D-STEM combined with automated Bragg-disk detection and NMF showed that He-FIBID metallic pillars were polycrystalline with an average grain size of approximately $^+$8 nm and axial symmetry in the grain map. Dielectric PMCPS deposits were described as amorphous with typical compositions around $^+$930 at.% Si, 25 at.% C, and 40 at.% O (Allen, 4 Oct 2025).

Another specialized architecture is the direct growth of AFM probes on delicate microcantilevers. Focused helium ion beam induced deposition and FEBID from W(CO)$\Theta$00 produced high-aspect-ratio nanopillars with tip radii below $\Theta$01 nm, lengths of $\Theta$02–$\Theta$03 nm for He-FIBID and $\Theta$04–$\Theta$05 nm for FEBID, and no damage to the underlying cantilever (Allen et al., 2023). TEM revealed a tungsten-rich core with a thin amorphous carbonaceous surface layer of approximately $\Theta$06–$\Theta$07 nm. The reported compositions were W $\Theta$08 at.%, C $\Theta$09 at.%, O $\Theta$10 at.% for He-FIBID and W $\Theta$11 at.%, C $\Theta$12 at.%, O $\Theta$13 at.% for FEBID. Post-growth helium-ion milling reduced the tip radius from approximately $\Theta$14 nm to approximately $\Theta$15 nm. These organometallic tips were used for high-speed AFM of DNA and nucleoproteins in liquid and withstood UV-ozone cleaning better than conventional carbonaceous tips (Allen et al., 2023).

6. Application domains, trade-offs, and outlook

FIBID is used wherever nanometer-scale site specificity outweighs the limitations of serial writing. In nanotube synthesis, Ga-FIBID of silicon oxide catalyst pads from TEOS enabled lithography-free positioning of single-walled carbon nanotubes at arbitrary substrate locations, with the positional precision limited only by the focused ion beam diameter, typically “on the order of ten nanometers” (Sadki et al., 2018). Pad geometry controlled nanotube yield: $\Theta$16m pads of $\Theta$17 nm thickness typically yielded one SWCNT per pad, while $\Theta$18m pads of the same thickness yielded “tens” of SWCNTs per pad, and pads thinner than $\Theta$19 nm generally yielded no nanotubes (Sadki et al., 2018). This is a clear example of FIBID functioning as a deterministic catalyst-placement technique rather than as a final structural material.

Across applications, the same trade-offs recur. The roadmap emphasizes that FIBID deposits often consist of metallic nanoscale crystallites in a carbonaceous or amorphous matrix, with typical metal contents for room-temperature organometallic growth often below $\Theta$20 at.% for metals (Höflich et al., 2023). Electrical, magnetic, and mechanical properties are therefore coupled to precursor chemistry, ion species, beam parameters, and sputtering. Resolution also trades against rate: Ga$\Theta$21-FIBID typically yields lateral features of approximately $\Theta$22–$\Theta$23 nm, He$\Theta$24-FIBID can reach approximately $\Theta$25 nm, but lighter ions also alter deposition yield and growth mode (Höflich et al., 2023). The serial nature of FIBID further means that for large-area or high-density patterning, conventional lithography remains faster and more cost-effective (Sadki et al., 2018).

Several misconceptions can be addressed directly from the literature. First, FIBID is not restricted to protective capping or contact repair; it is now used for superconducting devices, catalyst definition, AFM-tip renewal, and multimaterial 3D nanoprinting (Sigloch et al., 2022). Second, lower implanted-ion content is not synonymous with better functional performance, as shown by the Rb$\Theta$26 versus Ga$\Theta$27 Pt comparison where lower primary-ion contamination coincided with substantially higher resistivity because of higher C and O content (Li et al., 2022). Third, He- and Ne-based FIBID do not simply replace Ga-FIBID; they reduce metallic contamination in insulators and enable high-fidelity 3D architectures, but Ga-FIBID remains central for many metallic and superconducting applications (Allen, 4 Oct 2025).

The current outlook is defined by precursor design, predictive modeling, cryogenic growth, and higher-throughput screening. The roadmap identifies cryo-FIBID as a new development for higher purity and lower dose, while the precursor-screening pathway based on BSE, EDX hyperspectral data, NMF, and the precursor score $\Theta$28 provides a route to faster materials selection (Höflich et al., 2023). This suggests that the future of FIBID lies less in a single optimal beam or precursor than in a tightly coupled ecosystem of source engineering, surface-chemistry control, in-situ analytics, and application-specific optimization.