Jump to content

User:Dshashkov/sandbox

From Wikipedia, the free encyclopedia

Accelerated Neutral Atom Beam (ANAB)

[edit]
A commercial nAccel100 ANAB tool

Accelerated Neutral Atom Beam (ANAB) is a particle beam technology used for nano-scale surface modification of various materials. It is capable of angstrom-level smoothing of surfaces, creating a nano-scale surface texture, layer-by-layer material removal (sputtering or etching) with minimal subsurface damage, and bio-activation of surfaces, making underlying material more bio-compatible. It is also used to chemically alter the surfaces through monolayer or sub-monolayer coating or ultra-shallow doping. ANAB is a derivative, but distinct, technology from its predecessor technology, GCIB [Gas Cluster Ion Beam].

Technology

[edit]

ANAB Overview

[edit]

In ANAB, the beam consists of electrically neutral gas molecules or atoms that have been accelerated to energies of 10-100 eV/atom, corresponding to velocities of 5-15 km/sec. Energy of individual atoms in ANAB exceeds sputtering threshold of most materials. As a result, collision of individual atoms with a solid surface results in controlled removal (sputtering and/or etching) of the surface. Due to relatively low energy per ANAB atom, compared to other particle beam and plasma based surface modification methods, such as reactive ion etching (RIE), focused ion beam (FIB) etc., ANAB displaces minimal amount of material on the surface, usually limiting the sub-surface damage to no more than 2-3 nm.

Principle of ANAB operation

Cluster Formation

[edit]

In ANAB, initial stages of forming the beam are identical to Gas Cluster Ion Beam (GCIB) technique. Pressurized gas is expanded through a small nozzle into vacuum to form gas clusters that are typically comprised of a few hundred to several thousand atoms or molecules. The nozzle is specially shaped (typically conical convergent-divergent[1] or de Laval shape) to achieve supersonic gas velocity and facilitate adiabatic cooling. Such conditions result in localized temperature near the nozzle exit of just a few Kelvin, which in turn facilitates cluster formation via gas supersaturation and condensation mechanism. Clusters in upper part of the nozzle form by monomer addition mechanism, which in lower parts of the nozzle is complemented by cluster-cluster aggregation. This results in relatively broad cluster size distribution at larger sizes. Majority of clusters, however, are in the range of 500-2,000 atoms per cluster.

Clusters exiting the nozzle tend to concentrate near the central axis of the nozzle and typically account for only 1-2% of the total gas volume exiting the nozzle. To minimize contamination of the cluster beam with the residual non-clustered gas, clusters are directed through a small conical-shaped skimmer into a second vacuum chamber, while most of the non-clustered gas emerging from the nozzle is removed by a nozzle chamber pump.

Cluster Ionization

[edit]

After exiting the nozzle chamber, cluster beam enters the ionizer chamber, where electrically neutral clusters become positively charged by electron impact ionization. An ionizer typically uses tungsten filament as a thermionic source to generate thermal electrons at the cathode. Electrons emitted from the filament are accelerated by a voltage applied between the filament and a cylindrical anode, causing them to collide with gas clusters passing through the center of the anode. These collisions cause ejection of a single or multiple electrons from some clusters, making them positively charged. If sufficient electron flux is created in the ionizer, multiple cluster-electron collisions become prevalent. As a result, the cluster beam exiting the ionizer contains a mix of neutral, single-charged (+1), and double-charged (+2) clusters, with charges above (+2) becoming more rare.

Cluster Acceleration, Dissociation and Deflection

[edit]

Upon ionization, charged clusters are accelerated through a high electrical potential (20-50 kV). The resulting energy of the clusters corresponds to velocities of 5-20 km/sec and to energy per atom of 10 eV - 100 eV.

As accelerated gas cluster ion travels further through vacuum, it will experience collisions with residual gas species in its path, leading to gradual cluster dissociation, sometimes described as cluster evaporation. There are two main sources sources of residual gas, namely the background gas present in any vacuum system, and so-called "dead gas". Dead gas represents atoms that emanated from the nozzle but did not cluster, or clusters that successfully formed but did not ionize. Residual gas species carry thermal or near-thermal energy, and thus are nearly stationary in comparison with rapidly moving charged clusters. Upon collision, significant amount of kinetic energy is absorbed by the cluster, making it thermodynamically unstable. As a result, multiple gas atoms are released from the cluster with each collision, making residual cluster smaller and in some cases, leading to its complete disintegration. Neutral gas atoms released from the cluster possess minimal kinetic energy and continue traveling in the same direction and with the same velocity as the cluster from which they emerged. It is believed that in each dissociation event, electric charge stays with the original diminished cluster, and in case of complete cluster disintegration, is left on individual dissociated atoms.

In case of GCIB, such cluster dissociation is undesirable and is minimized by lowering the residual gas pressure as much as possible through additional vacuum pumping below the acceleration electrode. Conversely, cluster dissociation as an essential step in forming ANAB. Cluster dissociation is maximized by optimizing the residual gas pressure between the acceleration electrode and the specimen. If this residual pressure is too low, cluster dissociation is too low and ANAB intensity is minimized. If it is too high, secondary collisions become prevalent, both reducing the beam energy increasing its divergence. Upon cluster dissociation, all charged species in the ANAB are electrostatically removed from the beam by applying horizontal electric field using deflector electrode.

Neutral Nature of ANAB

[edit]

Electrically neutral nature of ANAB results in several beneficial properties vs. other surface treatment techniques that utilize charged particles.

  1. As charged species (ions, electrons, clusters etc.) repel each other, all other techniques suffer from widening known as "beam blowup", especially at lower energies. This makes anisotropic processing problematic, especially when high-aspect ratio structures are involved. ANAB, on the other hand, maintains its low beam divergence over extended distance, as neutral atoms to not repel each other.
  2. In other ion-based techniques, beam blowup can be partially mitigated by flux neutralization, but not completely eliminated. Flux neutralization is usually achieved using a secondary source of the opposite polarity (e.g. compensating electron flux used in conjunction with positively charged ion beam). ANAB does not require flux neutralization.
  3. When processing dielectric materials, charged beams require complete neutralization of the specimen surface, to avoid charge accumulation that can lead to capacitive discharge and damage to the specimen.

Summary of ANAB Beam Parameters

[edit]
Parameter Value Comment
Gas species Ar, O2, N2, CH4, B2H6, PH3, AsH3, SiH4, GeH4, NF3, CF4, SF6, etc. Inert gas (Ar, Ne, Kr, etc.) can be mixed with other gases to create reactive ANAB mixtures
Base pressure <10-6 Torr
Beamline pressure 5*10-5 Torr
Accelerating voltage 5 kV - 50 kV
Cluster temperature 35 K At nozzle exit
Beam diameter 1 cm
Beam divergence 0.15 deg
Beam flux 1015 - 1016 atoms/cm2/sec Equivalent of 0.1 - 1.0 mA current
Energy/atom 10 eV - 100 eV
Atom velocity 5 - 15 km/sec

ANAB-Material Interaction

[edit]

ANAB impacts solid material surfaces in multiple ways described below.

Sputtering

[edit]

ANAB energy per atom (10 - 100 eV) exceeds sputtering threshold of most materials. As a result, sputtering rates between 0.1 nm/sec and 10 nm/sec have been demonstrated. In case of chemical compounds or heterogeneous materials exposed to ANAB, differential sputtering rates between different chemical species result in gradual depletion of the top material layer with regards to the species with lower sputtering threshold. In case of organic molecules, carbon-rich layer is often formed on the surface due to lower sputtering rate of carbon and higher energy of carbon-carbon bonds. Table below illustrates the range of sputtering rates as a function of material and ANAB energy.

ANAB Sputtering Rates, Argon Beam
Beam energy Material Sputter rate
30 kV Amorphous C 0.025 nm/sec
30 kV Si 0.4 nm/sec
30 kV SiO2 0.5 nm/sec
50 kV Amorphous C 0.09 nm/sec
50 kV Si 7.5 nm/sec
50 kV SiO2 9.6 nm/sec

Similar to Gas Cluster Ion Beam, ANAB sputtering is anisotropic. Relatively little sputtering occurs in the direction of the incident neutral beam, while maximum sputtering intensity is observed at approximately 60 degrees to the incident beam direction. Preferential ejection of sputtered material at large polar angles, so-called lateral sputtering, enables smoothing nature of ANAB sputtering, as illustrated below.

Angular distribution of sputtered species, Argon ANAB (30 kV beam) on Cu substrate[2]

Etching

[edit]

When reactive gases are used to create ANAB, higher material removal rates have been demonstrated (up to 30-50 nm/sec), due to combined kinetic and chemical impact of ANAB species. By tuning the chemical composition of the beam, differential etching rates can be achieved even for materials with similar sputtering thresholds, resulting in selective etching. Combined with highly parallel nature of ANAB beam, this property makes ANAB particularly useful for etching semiconductor heterostructures. Cross-sectional TEM (XTEM) image shown below shows removal of 20 nm thick silicon nitride (SiN) layer deposited by atomic layer deposition (ALD) from a patterned Si/SiO2 surface using reactive-beam ANAB.

Sequence of ANAB etching of 20 nm thick silicon nitride (SiN) cap layer deposited by ALD on a patterned Si/SiO2 surface. From left: initial structure; structure after 5, 7 and 9 ANAB beam passes. Complete removal of the top SiN layer is achieved after 9 ANAB passes, while preserving sidewall SiN layer and with minimal rounding of corners. Courtesy of Neutral Physics Corporation.

Doping

[edit]

When reactive gas species are used in ANAB, their interaction with the surface may cause them to penetrate just below the surface, resulting in ultra-shallow doping. The depth profile in this case is limited to amorphization depth of ANAB, typically 2-3 nm. This phenomenon is demonstrated below, showing Secondary Ion Mass Spectrometry (SIMS) profile of silicon (Si) substrate treated with GCIB and ANAB utilizing 30 kV beams and Argon-1% B2H6 mixture as a source gas. While the profile shows boron penetration up to 10 nm below the surface, the actual concentration profile is likely even more shallow, since depth profiling in SIMS at shallow depths is prone to atomic mixing of the top 10 nm below the surface known as "knock-on" effect[3].

Boron concentration profile obtained by Secondary Ion Mass Spectroscopy (SIMS) from the shallow-doped silicon wafer created by GCIB (red curve) and ANAB (blue curve). In both cases, 30 kV beam energy and argon - 1% B2H6 source gas was used.

Smoothing

[edit]

Due to lateral sputtering nature of ANAB, sub-micron features on the surface of the specimen can be removed, resulting in low levels of surface roughness not routinely achievable with any other surface polishing techniques. In an example below, a glass substrate for Extreme Ultraviolet Lithography (EUV) photomask has been treated with ANAB. As a result, peak-to-valley roughness Rz has been reduced from 11.2 nm to 1.2 nm, while the average roughness Ra was reduced from 1.4 nm to 0.13 nm.

Atomic Force Microscopy (AFM) image of the mask blank for Extreme Ultraviolet (EUV) lithography in its initial condition (after mechanical polishing). Field of view: 1 μm x 1 μm. Peak-to-valley roughness Rz = 11.2 nm, average roughness Ra = 1.2 nm.
Atomic Force Microscopy (AFM) image of the mask blank for Extreme Ultraviolet (EUV) lithography after ANAB treatment with an Ar beam. Field of view: 1 μm x 1 μm. Peak-to-valley roughness Rz = 1.4 nm, average roughness Ra = 0.13 nm.

Nano-texturing

[edit]

In addition to smoothing effect of ANAB, the beam interaction with the substrate can result in an opposite effect. Under certain exposure conditions, nanotexture is created on the surface, with random peaks and valleys spaced 20-60 nm apart. It is assumed that this texture results from a similar mechanism as in GCIB, where individual clusters create nano-craters on the surface similar to meteorite craters on the moon. In ANAB, the clusters are dissociated into individual neutral atoms. However, due to very low ejection energy, upon cluster dissociation the individual atoms in ANAB continue to travel in close proximity to each other as an unbound "cloud" that is not significantly larger than the original cluster from which these atoms originated. As a result, their collective impact on the surface produces similar nano-sized craters, however less pronounced in comparison with GCIB.

Atomic Force Microscopy (AFM) image of the surface or polyether ether ketone (PEEK). Field of view: 1 μm x 1 μm. Left: original (as-machined) surface. Right: ANAB-treated surface exhibiting a 20 nm - 60 nm nano-texture (Ar source gas, 30 kV beam energy, 2.5*1017 atoms/cm2 dose)

Amorphization

[edit]

When a crystalline specimen surface is impacted by ANAB, top 2-3 nm will be amorphized by energetic ANAB species causing physical displacement of material on the surface and breakage of the regular crystalline order. Amorphization in ANAB produces thinner and more uniform amorphous layers in comparison with Gas Cluster Ion Beam (GCIB). In a cross-sectional TEM (XTEM) images below, amorphous layers created by GCIB and ANAB on mono crystalline silicon reveal that in GCIB, the depth of the amorphous layer varies from 6.3 nm to 8.1 nm, while in case of ANAB a uniform 2.3 nm thick amorphous layer is created.

Amorphous layer on Si created by Gas Cluster Ion Beam (GCIB). Thickness 6.3 nm - 8.1 nm.
Amorphous layer on silicon created by ANAB. Thickness 2.3 nm. Interface between amorphous and crystalline regions in Si is atomically flat.

Amorphization depth in ANAB depends on the nature of the specimen and ANAB beam characteristics. XTEM image below shows comparison of amorphized layers on silicon created by ANAB at different beam energies. As beam energy increases, the thickness of amorphous region increases from 2.0 nm to 3.2 nm.

High resolution TEM images of amorphous layer on the surface of silicon created by ANAB treatment at various beam energies using Ar source gas. Process 1: 10 kV beam, 250 sccm gas flow; Process 2: 10 kV beam, 200 sccm gas flow; Process 3: 20 kV beam, 200 sccm gas flow; Process 4: 30 kV beam, 200 sccm gas flow.

Surface energy and chemistry modification

[edit]

ANAB interaction with a specimen produces rearrangement of the chemical bonds on the surface as well as the breakdown of the specimen crystallinity (amorphization). Both processes result in increase of the surface energy of the specimen or "surface activation" similar to the effect of plasma.

Improved bio-compatibility

[edit]

XYZ

Applications

[edit]

Blah Blah

See Also

[edit]

Gas Cluster Ion Beam (GCIB)

Ion Implantation

Reactive Ion Etching (RIE)

Focused Ion Beam (FIB)

Atomic Layer Etching (ALE)

Supermirror

Biomaterial Surface Modifications

References

[edit]

Blah Blah

[edit]
  1. ^ "Nozzle Design". spaceflightsystems.grc.nasa.gov. Retrieved 2018-10-11.
  2. ^ Kirkpatrick, A.; Kirkpatrick, S.; Walsh, M.; Chau, S.; Mack, M.; Harrison, S.; Svrluga, R.; Khoury, J. (2013). "Investigation of accelerated neutral atom beams created from gas cluster ion beams". Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms. 307: 281–289. doi:10.1016/j.nimb.2012.11.084. ISSN 0168-583X.
  3. ^ "Secondary Ion Mass Spectrometry Theory Tutorial". pprco.tripod.com. Retrieved 2018-10-23.