Published on

Focused Ion Beam (FIB) Technology:Principles and Development History

Authors

Since the 20th century, nanotechnology has rapidly developed as an emerging field of science and technology. Currently, nanotechnology stands as one of the representative frontiers of science and technology in the 21st century. The structural unit size of nanomaterials approaches the coherence length of electrons and the wavelength of light, exhibiting unique properties such as surface and interface effects, size effects, and quantum size effects. These materials possess novel characteristics in electronics, magnetism, optics, mechanics, and hold immense potential for high-performance device applications.

Developing novel nanostructures and devices with specific properties necessitates the advancement of high-precision, multidimensional, and stable micro-nanofabrication technologies. The range of micro-nanofabrication processes is extensive, including commonly used techniques such as ion implantation, photolithography, etching, and thin-film deposition. In recent years, due to the trend towards miniaturization in modern processing technologies, focused ion beam (FIB) technology has increasingly been applied in micro-nanostructure manufacturing across various fields, becoming an indispensable technology in micro-nanofabrication.

FIB technology, developed on the basis of conventional ion beam and focused electron beam systems, fundamentally operates similarly to the latter. In comparison to electron beams, FIB scans the sample surface using ion beams generated from an ion source that are accelerated and focused. Due to the significantly larger mass of ions compared to electrons—even the lightest ions such as H+ are over 1800 times heavier than electrons—ion beams not only achieve imaging and exposure akin to electron beams but also enable sputtering of atoms from solid surfaces, serving as a direct writing tool. Moreover, FIB can be combined with chemical gases to induce atomic deposition onto sample surfaces, making it a versatile tool in micro-nanofabrication.

This paper primarily introduces the basic principles and developmental history of FIB technology.

Ion Source

Unlike the electron gun in electron beam systems that generates accelerated electrons through electron optical systems, FIB employs an ion source. The ion source is the core of the FIB system. Early ion sources were initially developed through advancements in mass spectrometry and nuclear physics. Following the 1960s, further development of ion sources was driven by ion implantation processes in the semiconductor industry. These ion sources can be broadly categorized into three types based on their operating principles:

(1) Electron bombardment ion sources utilize electrons emitted from a hot cathode. These electrons are accelerated to bombard gas molecules inside the ion source chamber, causing them to ionize. Such ion sources are predominantly used in mass spectrometers, characterized by low beam currents and minimal energy dispersion.

(2) Gas discharge ion sources generate ions from plasma discharge, such as glow discharge, arc discharge, and spark discharge ion sources. These ion sources produce high beam currents and are commonly used in nuclear physics research.

(3) Field ionization ion sources utilize a strong electric field near a needle-like electrode to ionize gas atoms adsorbed on the surface of the needle. These ion sources are mainly employed in field ion microscopy.

With the exception of field ionization ion sources, all these ion sources ionize gases in a large spatial area (ionization chamber) and extract ions through small apertures. Consequently, these ion sources have low ion current density, large ion source areas, are unsuitable for focusing into narrow beams, and therefore are not suitable as ion sources for FIB. In the 1970s, Clampitt and others developed the liquid metal ion source (LMIS) during research on cesium ion sources used in satellite thrusters.

fig1
Figure 1: LMIS Basic Structure

LMIS Basic Structure

A tungsten wire with an initial diameter of approximately 0.5 mm is electrochemically etched to form a needle with a tip diameter as small as 5-10 μm. Subsequently, molten liquid metal adheres to the tip of the needle. When a strong electric field is applied, the liquid metal at the tip forms an extremely small Taylor cone (approximately 5 nm), with an electric field intensity reaching up to 10^10 V/m. Under such high electric fields, metal ions on the surface of the liquid tip evaporate in a phenomenon known as field evaporation, generating an ion beam. Despite the ion current being only a few microamperes due to the LMIS's small emission area, the current density can reach approximately 10^6 A/cm², and its brightness is around 20 μA/Sr, which is 20 times that of field ion sources.

The development and widespread application of FIB systems owe much to the research on LMIS. The ion emission from LMIS involves a complex dynamic process where the shape of the liquid metal emission tip changes with variations in the electric field and emission current. Continuous replenishment of the liquid metal is crucial to maintain stable emission, making the process a synergistic interplay between electrohydrodynamics and field ion emission. Analytical studies have identified three key conditions for stable LMIS emission:

(1) The emission surface must have a specific shape to form a defined surface electric field.

(2) The surface electric field must be sufficient to sustain a specific emission current and liquid metal flow rate.

(3) The surface flow rate must be adequate to sustain the material flux loss corresponding to the emission current, thereby maintaining the emission surface's specific shape.

From a practical standpoint, ensuring good wetting between the liquid metal and the tungsten needle tip during LMIS fabrication is critical. This complete and continuous attachment ensures optimal liquid metal flow, necessary for both forming the emission liquid tip and supplying the liquid metal continuously.

Experimental findings reveal additional characteristics of LMIS:

(1) There exists a critical emission threshold voltage, generally above 2 kV; beyond this threshold, the emission current increases rapidly.

(2) LMIS exhibits a relatively large emission angle. The natural emission angle of the ion beam is typically around 30º, increasing with the ion beam's intensity, which reduces beam utilization efficiency.

(3) The angular current density distribution is relatively uniform.

(4) LMIS exhibits significant ion energy dispersion (chromatic aberration), typically around 4.5 eV, which increases with the ion beam's intensity due to severe space charge effects at the emission tip. This high-density ion repulsion at the emission tip results in substantial energy dispersion. The most effective method to reduce chromatic aberration is by reducing the emission current; however, once below 2 μA, further reductions in chromatic aberration become challenging, maintaining it around 4.5 eV. Continued reduction leads to unstable LMIS operation, resulting in pulsed emission. Large energy dispersion increases chromatic aberration in ion optical systems, exacerbating beam spot dispersion.

(5) Mass spectrometry analysis of LMIS indicates that at low beam currents (≤ 10 μA), nearly 100% of the ions are single-charged. With increasing beam current, the proportion of multi-charged ions, molecular ions, ion clusters, and charged metal droplets increases, posing challenges for focused ion beam applications.

These characteristics indicate that, for practical applications, LMIS should operate under conditions of low beam current, ideally below 10 μA. Under these conditions, ion energy dispersion and divergence angles are minimized, maximizing beam utilization efficiency.

Initially developed using liquid gallium as the emission material due to its low melting point of 29.8°C, LMIS has several advantageous properties: minimal volatility, heavy atomic nucleus, good adhesion to tungsten needles, and excellent oxidation resistance. Over the years, LMIS has expanded to include various other materials besides gallium, such as Al, As, Au, B, Be, Bi, Cs, Cu, Ge, Fe, In, Li, Pb, P, Pd, Si, Sn, U, and Zn. Some of these can be directly used as single-source materials, while others require fabrication into eutectic alloys to reduce the melting points of refractory metals. Different elements' ions can be separated using EXB separators from alloy ion sources. Elements such as As, B, Be, and Si in alloy ion sources can directly dope semiconductor materials. Despite the increasing variety of ion sources available today, gallium's exceptional properties continue to make it the most widely used ion source, with some advanced models even employing isotopically pure gallium.

Focused Ion Beam (FIB) systems

Focused Ion Beam (FIB) systems, fundamentally similar to electron beam exposure systems, consist of components such as ion emission sources, ion columns, workstations, vacuum systems, and control systems. Just as the core of an electron beam system is its electron optical system, the core component for focusing ion beams into fine beams is the ion optical system. The most basic difference between ion optics and electron optics lies in the fact that ions have a charge-to-mass ratio much smaller than electrons. Therefore, magnetic fields cannot effectively control the movement of ion beams. Currently, focused ion beam systems only employ electrostatic lenses and electrostatic deflectors. Electrostatic lenses have a simple structure, produce no heat, but have significant optical aberrations.

fig2
Figure 2: Diagram of Focused Ion Beam (FIB) System

Typical Focused Ion Beam (FIB) System Structure

A typical Focused Ion Beam (FIB) system consists of a two-stage lens system. Ion beams generated from a liquid metal ion source are shaped into a tiny apex under the influence of an external electric field (suppressor) and a negative electric field (extractor) that guides the metal apex to produce the ion beam. Initially, after passing through the first aperture, the ion beam is focused by the first-stage electrostatic lens, and a primary octupole deflector is used to adjust the ion beam to minimize image dispersion. Following a series of variable apertures, the ion beam spot size can be flexibly adjusted. Subsequently, a secondary octupole deflector scans the ion beam according to the defined processing patterns. By using blanking deflectors and blanking aperture masks, ion beam blanking can be achieved. Finally, through the second electrostatic lens, the ion beam is focused to a very fine spot size, achieving a resolution of approximately 5 nm. The focused ion beam bombards the sample surface, generating secondary electrons and ions that are collected and imaged by corresponding detectors.

In solid materials, ions, like electrons, undergo a series of scattering events where they progressively lose energy before eventually coming to rest within the material. This scattering process includes both elastic and inelastic interactions: elastic scattering changes the ion's flight path without energy loss, whereas inelastic scattering involves energy loss due to the comparable mass between ions and atoms in the solid material.

There are two primary mechanisms through which ions lose energy in the material:

  1. Nuclear energy loss occurs when ions collide with atomic nuclei in the solid material, transferring some energy to the nuclei. This can cause atoms to displace or even be ejected from the material surface entirely, a phenomenon known as sputtering, which is exploited in FIB processes for etching purposes.

  2. Electronic energy loss involves transferring energy to electrons surrounding the atomic nuclei, leading these electrons to either emit secondary electrons or to be stripped from the atoms, ionizing them and resulting in the emission of secondary ions.

Ion-Atom Collisions in Solid Materials Analysis

As charged particles, ions undergo a series of scattering events within solid materials, similar to electrons, where they continuously lose energy and eventually come to rest within the material. This process involves both elastic and inelastic scattering: elastic scattering does not lose energy but alters the ion's flight direction within the solid. Inelastic scattering results in energy loss because ions, with masses comparable to those of atoms in the solid material, transfer energy during collisions with atomic nuclei or electrons.

There are two main reasons for ion energy loss in materials. Firstly, nuclear loss occurs when ions collide with atomic nuclei in the solid material, transferring energy that displaces atoms or ejects them completely from the surface—a phenomenon known as sputtering, utilized in FIB processing for etching. The second type of loss is electronic loss: ions transfer energy to electrons surrounding atomic nuclei, causing these electrons to either be excited and emit secondary electrons or be stripped from the vicinity of the solid atom, ionizing it and generating secondary ion emission.

Ion scattering processes can be simulated using Monte Carlo methods, a process similar to simulating electron scattering:

(1). Calculate the total scattering cross-section using atomic nucleus differential scattering cross-sections to determine the probability of ion collisions with a specific solid material atom.

(2). Randomly select scattering angles and mean free paths, calculating nuclear and electronic energy losses from scattering.

(3). Track ion scattering trajectories until ions lose all their carried energy and come to rest at a specific location within the solid material, becoming implanted ions. This process assumes the substrate material is an amorphous material with randomly arranged atoms. In practical applications, however, substrates often consist of crystalline materials such as silicon single crystals. Compared to amorphous materials, crystals have crystallographic directions with low-index planes where atomic arrangement is sparse, allowing ions to penetrate deeper by "channeling effect," increasing penetration depth by several orders of magnitude.

History and Current Status of FIB

Ion beam applications have a century-long history since Thomson invented the gas discharge ion source in 1910. However, the true application of FIB began with the invention of LMIS, as briefly discussed earlier.

In 1975, Levi-Setti, Orloff, and Swanson developed the first FIB system based on field emission technology, utilizing a gas field ionization source (GFIS).

Also in 1975, Krohn and Ringo produced the first high-brightness ion source: the liquid metal ion source (LMIS), marking the beginning of a new era for FIB technology—the LMIS era.

In 1978, researchers at Hughes Research Labs in California built the first LMIS-based FIB system.

In 1982, FEI manufactured the first focused ion beam column.

In 1983, FEI produced the first electrostatically focused electron beam column. The same year, Micrion was founded, focusing on developing focused ion beam systems for mask repair applications. In 1984, Micrion collaborated with FEI, with FEI supplying components to Micrion.

In 1985, Micrion delivered its first focused ion beam system.

In 1988, the first dual-beam system combining focused ion beam with scanning electron microscopy (FIB-SEM) was successfully developed. In this configuration, ion and electron beams are installed at a certain angle, enabling simultaneous imaging with electrons and processing with ions. By tilting the sample stage, the sample surface can be oriented perpendicular to either the electron or ion beam. Since then, almost all FIB systems developed are combined with SEM, forming dual-beam systems known as FIB-SEM.

In the 1990s, FIB dual-beam systems transitioned from laboratories to commercial applications.

fig3
Figure 3: Typical FIB-SEM Dual-Beam Equipment Schematic

In 1999, FEI acquired Micrion and integrated its product line and operations.

In 2005, ALIS (Advanced Lithography and Imaging Systems) was established, and the following year, ZEISS acquired ALIS.

In 2007, Carl Zeiss introduced the first commercial He+ microscope. Helium ion microscopy uses helium ions as the ion source, which, despite causing minimal sputtering at high magnification and long scan times, inherently damages samples much less than gallium ions. Helium ion microscopy can achieve higher resolution images compared to SEM (Scanning Electron Microscopy) and offers excellent material contrast due to the ability to focus into smaller probe sizes.

In 2011, Orsay Physics released a Xe plasma source suitable for FIB-SEM. The Xe plasma source operates by ionizing inert gas through high-frequency vibration and then extracting and focusing the ion beam using an extraction electrode. Unlike liquid gallium ion sources, the Xe plasma source can achieve a maximum beam current of up to 2 μA after passing through apertures, significantly enhancing the micro-machining capabilities of FIB. Processing speeds with Xe plasma FIB can be up to 50 times faster than liquid gallium FIB, making it highly practical for microscale processing, often reaching several hundred micrometers in size. In recent years, both Thermo Fisher and TESCAN have introduced Xe-FIB products.

Today, FIB technology has made tremendous advancements, integrating seamlessly with various detectors, nanomanipulators, and testing devices. It has evolved into a versatile processing and characterization tool that combines micro-area imaging, processing, analysis, and manipulation capabilities. FIB systems are extensively used in industries such as semiconductor manufacturing, nanoscale research, life sciences, and earth sciences. Leading manufacturers of FIB instruments include Thermo Fisher Scientific (represented by the Helios series), Carl Zeiss (represented by the Crossbeam series), TESCAN from Czech Republic, Hitachi from Japan, and Raith from Germany.