Sputtering

The sputtering (from english to sputter = atomize), also sputtering called, is a physical process whereby atoms in a solid state ( target ) by bombardment with energetic ions (mainly noble gas ions) are released and pass into the gas phase.

This effect is used, for example, in surface physics for the preparation of extremely pure surfaces, for the analysis of the chemical composition of surfaces, e.g. B. Secondary ion mass spectrometry (SIMS), for secondary neutral particle mass spectrometry (SNMS) or as a sputter ion source for particle accelerators . The process is of great technical importance as a partial step in sputter deposition, a fine vacuum-based coating technology belonging to the group of PVD processes . Here it is used to atomize a material, which is then deposited on a substrate and forms a solid layer. In the field of coating technology, sputter deposition is often referred to as “sputtering”.

In electron tubes and gas discharge tubes such as glow lamps and Nixie tubes , sputtering is one of the undesirable effects that limit the service life. The impact of ions on the electrodes affects them. In addition, the removed electrode material is deposited on the inside of the glass bulb and reduces its transparency.

Gottfried K. Wehner in the USA and Vera Yevgenevna Jurassowa in the USSR were pioneers in the study of the phenomenon .

Basics of the sputtering process

When bombarding a surface with ions , different effects can occur , depending on the ions used and their kinetic energy :

• Material from the bombed target ( cathode ) is removed. This is the sputtering described here .
• The ions are built into the target material and may form a chemical bond there. This effect is called (reactive) ion implantation .
• The ions condense on the bombarded substrate and form a layer there: ion beam deposition .

If material removal is intended, the ions must have a certain minimum energy. The impacting ion transfers its impulse to atoms of the bombarded material, which then - similar to billiards  - trigger further collisions in a collision cascade . After several collisions, some of the target atoms have an impulse that points away from the interior of the target. If such an atom is close enough to the surface and has a sufficiently high energy, it leaves the target.

The sputtering yield essentially depends on the kinetic energy and mass of the ions as well as on the binding energy of the surface atoms and their mass. In order to knock an atom out of the target, the ions have to generate a material-dependent minimum energy of typically 30–50  eV . The yield increases above this threshold. The initially strong rise flattens out quickly, however, since with high ion energies this energy is deposited ever deeper in the target and thus hardly reaches the surface. The ratio of the masses of ion and target atom determines the possible momentum transfer. For light target atoms, the maximum yield is achieved when the mass of target and ion approximately match; however, as the mass of the target atoms increases, the maximum yield shifts to ever higher mass ratios between ion and target atom.

Ion bombardment not only generates neutral atoms, but also secondary electrons and, to a lesser extent, secondary ions and clusters of different masses ( secondary ion mass spectrometry ). The energy distribution of the dissolved atoms has a maximum at half the surface binding energy at a few eV, but drops only slowly at high energies, so that the average energy is often an order of magnitude higher. This effect is used in analysis methods in surface physics and thin film technology as well as for the production of thin films (sputter deposition ).

If an atomically thin layer is removed by sputtering, the number of particles sputtered over time can be estimated using the following equation: ${\displaystyle t}$

${\displaystyle N(t)=N_{\text{max}}\left(1-\exp \left(-{\frac {YI_{\text{P}}}{eN_{\text{max}}}}\,t\right)\right)}$

with e: elementary charge , : number of particles on the surface (about 10 ¹⁵  cm ^-2 ) : sputtering yield (number of sputtered particles per incident particle) : primary current. ${\displaystyle N_{\text{max}}}$ ${\displaystyle Y}$ ${\displaystyle I_{\text{P}}}$

Sputter deposition

Scheme of a sputter deposition system (reactive HF sputtering)

General

For sputter deposition, also known as sputter coating, a substrate is brought close to the target so that the atoms knocked out can condense on it and form a layer. The gas pressure in the process chamber must be so low ( vacuum ) that the target atoms reach the substrate without colliding with gas particles. This means that the mean free path of the gas particles must at least correspond to the distance between the target and the substrate. At a distance of approx. 30 cm, the gas pressure should therefore not be greater than approx. 2 × 10 ⁻⁴  mbar (high vacuum), otherwise the layer properties will be poorer.

In most applications, a direct current gas discharge (DC sputtering) is used as the ion source . If a magnet is also attached under the target, it is called magnetron sputtering . All conductive materials can be deposited in this configuration. In contrast to thermal evaporation, for example, there is no segregation of alloys . The adhesion of the layers is usually better than with vapor-deposited layers, and large areas, e.g. B. architectural glass, be coated homogeneously. For this application magnetron - cathode used m with a length of 3.5. Integrated circuits are also metallized on wafers in this way . In these applications, metal layers that are as pure as possible are normally desired. Therefore, in these cases, high-purity noble gases, usually argon , are used to avoid oxidation of the layers.

The coating of architectural glazing or absorbers of thermal solar collectors consists of layer systems in which transparent and partially absorbent materials, which are often not or not sufficiently electrically conductive, are also used. Here, a reactive gas, usually nitrogen or oxygen, can be added to the inert gas such as argon in order to deposit their compounds. In this case one speaks of reactive sputtering.

Other non-conductors, whose reactive sputter deposition is not possible or practicable, can usually be deposited with high-frequency or ion beam sputtering; however, the advantage of the large-area homogeneity is largely lost.

variants

Sputter deposition has been one of the most important coating processes for decades. During this time, different variants were developed with which the layer properties can be specifically influenced or the range of materials that can be deposited can be increased. The most important basic forms are briefly described below. It should be noted that there are mixed forms of all variants, e.g. B. RF magnetron sputtering as a mixture of high frequency and magnetron sputtering.

The most important basic forms are:

1. DC sputtering
2. RF sputtering
3. Ion beam sputtering
4. Magnetron sputtering
5. Reactive sputtering

DC sputtering

In DC sputtering, a direct voltage of several hundred volts is applied between the target and the substrate to be coated, which is why it is also called direct voltage sputtering. The target forms the negative and the substrate forms the positively charged electrode. By impact ionization of the atoms of the inert gas used (e.g. argon ), a plasma (an argon low-pressure plasma ) forms in the gas space , the components of which are negatively charged electrons and positively charged gas ions, such as Ar ⁺ , due to the applied direct voltage in the direction of the substrate or . of the target are accelerated. A permanent stream of positive ions now hits the target; Hence the English name of the process direct current sputtering , dt. direct current or DC sputtering. When impacting the target, particles are knocked out of the target by impulse transfer, which move away from the target in the direction of the substrate and are deposited there as a thin layer (deposition).

The main disadvantage of this method is that it can only be used for electrically conductive target materials. In the case of electrically insulating materials, the constant supply of new charge carriers would cause both the target and the substrate to become electrically charged, so that the DC voltage field would be compensated and the sputtering process would be impeded, since subsequent ions would be electrically repelled. In addition, the achievable sputter rates and thus also the coating rates are relatively low, since only a few sputter gas ions arise in the low-pressure plasma used.

The variant described above with two electrodes is also known as DC diode sputtering. In addition, there are other forms such as DC triode sputtering, in which the target is arranged as a third electrode outside the plasma space. In this way, the plasma generation and the sputtering process can be decoupled.

RF sputtering

In high- frequency sputtering ( RF sputtering for short, radio frequency sputtering , RF sputtering ), instead of the direct electrical field, a high-frequency alternating field is applied (usually with the free radio frequency of 13.56 MHz or a multiple thereof). The high-frequency voltage source required for this is connected in series with a capacitor and the plasma. The capacitor is used to separate the direct voltage component and to keep the plasma electrically neutral.

The alternating field accelerates the ions (mostly argon ions) and the electrons alternately in both directions. From a frequency of around 50 kHz, the ions can no longer follow the alternating field due to their significantly lower charge-to-mass ratio. The electrons oscillate in the area of ​​the plasma and there are more collisions with argon atoms. This causes a high plasma rate, one consequence of which is the possible pressure reduction to 1–20 m Torr (approximately 10 ⁻¹ −10 ⁻²  Pa) with the same sputtering rate. This enables the production of thin layers with a different microstructure than would be possible at higher pressures. The positive ions move through a superimposed negative offset voltage on the target in the direction of the target and, as in DC sputtering, detach atoms or molecules from the target material through collisions. The subsequent sputter deposition corresponds to that of other sputtering processes (see above).

Advantages:

• This also enables insulators (e.g. aluminum oxide or boron nitride ) and semiconductors to be sputtered
• the substrate heats up less
• Due to the oscillating electrons, the sputtering rate is around 10 times higher than with DC sputtering at the same chamber pressure.

Disadvantage:

• Relatively low coating rates
• HF generation is more complex than a DC voltage source
• With large rectangular cathodes (larger than 1 m), irregularities in the plasma density (layer thickness distribution) can occur

A variant of HF sputtering is what is known as bias sputtering. The substrate holder is not kept at ground potential, but rather a mostly negative electrical potential (−50 to −500 V). This results in an increased bombardment of the substrate with argon ions. This bombardment can, on the one hand, detach loosely bound impurities from the surface, and on the other hand, it brings additional energy into the deposited layer. This back-sputtering effect makes it possible to positively influence the layer properties or to improve the deposition in trenches.

Magnetron sputtering

Schematic structure of magnetron sputtering

Whereas with simple cathode sputtering, only an electric field is applied, with magnetron sputtering an additional magnetic field is arranged behind the cathode plate. Due to the superimposition of the electric field and magnetic field, the charge carriers no longer move parallel to the electric field lines, but are deflected onto a spiral path (more precisely: helical line ) (see Lorentz force ) - they now circle over the target surface. This extends their path and increases the number of collisions per electron. The electron density is highest at the point where the magnetic field is parallel to the target surface. This causes a higher ionization in this area. Since the ions are hardly deflected by the magnetic field due to their mass, most of the sputtering on the target takes place in the area below. The erosion trenches typical of magnetron sputtering are formed there on the target.

The effectively higher ionization capacity of the electrons leads to an increase in the number of noble gas ions and thus also in the sputtering rate. Since more target material is atomized, this leads to significantly higher coating rates with the same process pressure. Since the layer growth and thus the layer properties depend primarily on the process pressure in addition to the temperature, the process pressure can be set up to one hundred times lower than with conventional cathode sputtering with the same growth rates. This leads to less scattering of the material on the way to the substrate and to a denser (less porous) layer.

Magnetron sputtering is the most commonly used method in microelectronics for producing metal layers.

The Hochenergieimpulsmagnetronsputtern ( high-power impulse magnetron sputtering , HiPIMS) is a more advanced procedure, the effect of pulse-like discharge ( to reach services used) is greater than 1 MW to a markedly increased degree of ionization. The high degree of ionization can significantly change the properties of the growing layer via a changed growth mechanism and leads, for example, to higher adhesive strength. ${\displaystyle t\ll 1\ {\text{µs}}}$

Reactive sputtering

In reactive sputtering, one or more reactive gases (e.g. oxygen or nitrogen ) are added to the inert working gas ( Ar ) . The gases react on the target, in the vacuum chamber or on the substrate with the atomized layer atoms and form new materials. The resulting reaction products are then deposited on the substrate surface. For example:

${\displaystyle \mathrm {4\,Al(Target)+3\,O_{2}\longrightarrow 2\,Al_{2}O_{3}(Schicht)} }$

Reactive sputtering can be combined with other variants such as DC, HF or magnetron sputtering. It is mainly used to deposit oxides (SiO ₂ , Al ₂ O ₃ , ZnO), nitrides (Si ₃ N ₄ , TiN, AlN) and oxynitrides (e.g. SiO _x N _y ). In reactive sputtering, the layer properties can be well influenced by the gas mass flow, among other things. In addition to oxygen or nitrogen, simple molecules such as water vapor , ammonia , hydrogen sulfide , methane or tetrafluoromethane are also used as reaction gas, for example for cadmium sulfide , polytetrafluoroethylene or carbides such as tantalum carbide .

Since the added gases are ionized like the working gas in the plasma space, on the one hand the reactivity and thus the reaction rate with the atomized target material increases, on the other hand the ions of the reaction gas are also accelerated in the direction of the target. When they hit the target, like argon ions, they not only knock particles out of the target, but are also built into the target to a certain extent. This leads to contamination of the target. A problem that arises when the target is to be used for the deposition of other material compositions.

Ion beam sputtering

In ion beam sputter deposition (IBSD), a beam of noble gas sputtering ions (argon, krypton , xenon ) is directed onto the target from an ion source - the incident ion beam causes atomization. Ion beam sputtering offers the possibility of setting the particle energies in a targeted and energetically narrow band - the kinetic energies of the layer-forming particles are higher than with alternative vacuum coating techniques such as evaporation or magnetron sputtering. This makes it possible to achieve a more uniform condensation of the material vapor and so u. a. to produce dense, smooth and defect-free layers.

With a so-called assist ion beam it is still possible to influence the growing layer or to initiate an additional reactive process (see ion beam-assisted deposition ).

Atomic beam sputtering

In order to avoid electrostatic charges, insulating materials can also be sputtered with the aid of atomic beams , which can be generated with a capillary nitrone, for example .

application

The sputtering effect is used in material processing and analysis for cleaning materials and samples. Sputter cleaning can remove small particles and organic contamination from the surface, so that a subsequent coating process can be reproducible. The same applies to the cleaning of samples using surface-sensitive measurement techniques, for example photoelectron spectroscopy (UPS, XPS, etc.) by briefly bombarding them with argon ions. However, the sputtering effect can also be used to obtain information from deeper areas. In photoelectron spectroscopy and secondary ion mass spectrometry (SIMS), for example, a depth profile of layer systems can be determined by alternating or simultaneous sputtering and measurement - but effects such as preferential sputtering (different sputtering rates for atoms of different weight) must be taken into account.

Sputter deposition is one of the standard coating techniques and has many uses in industry. The materials or material systems that can be used for the coating differ greatly in the special process used. In general, however, the range of possible materials is very large. With “classic” (passive) sputter deposition processes, mainly metals are deposited, for example titanium , tungsten and nickel , but alloys with nickel-aluminum (NiAl) and non-metals such as silicon or silicon dioxide (SiO ₂ ) are also possible. Reactive processes, on the other hand, enable the deposition of metal compounds, for example metal oxides such as aluminum oxide (Al ₂ O ₃ ), with high precision in the stoichiometry of the layer by installing additional components from the gas space . In semiconductor and microsystem technology, the process is mainly used for the production of thin layers, for example aluminum or titanium nitride. However, sputtered thin films are also used in other industrial areas, for example in material or surface finishing (e.g. mirrors, car headlights, car rims ) or in optics as a functional layer (e.g. thin-film polarizer , heat protection glass ).

disadvantage

In general, when cleaning the surface of sensitive materials (e.g. graphite single crystals) by sputtering, it should be remembered that the sputtering will (partially) destroy the crystal structure. If the sputtering is too strong, i.e. if the energy of the ions is too high, there is also the risk that the dirt atoms that have not yet been removed are introduced into the surface ( knock-on effect ).

literature

• PJ Martin: Ion-based methods for optical thin film deposition . In: Journal of Materials Science . tape 21 , no. 1 , 1986, pp. 1-25 , doi : 10.1007 / BF01144693 .
• Markus Bautsch: Scanning tunnel microscopic examinations of metals atomized with argon , Chapter 2.4: Atomization of surfaces by particle bombardment , sub-chapter: Atomization rate - Ablation rate - Microstructure formation - Secondary effects . Verlag Köster, Berlin 1993, ISBN 3-929937-42-5 , pp. 18-27.

Web links

• Lecture notes on thin film technology ( Memento from February 1, 2012 in the Internet Archive ) - Institute for Microsystems Technology, Freiburg (PDF file; 2.9 MB)

Individual evidence

1. ↑ Hwaiyu Geng: Semiconductor Manufacturing Handbook . Surendra Kumar, 2005, ISBN 978-0-07-144559-7 , pp. 13.14 ff .
2. ↑ Dietrich Widmann, Hermann Mader, Hans Friedrich: Technology of highly integrated circuits . Gabler Wissenschaftsverlage, 1996, ISBN 978-3-540-59357-7 , p. 32-33 .
3. ↑ Wolfgang Bergmann: Material technology 2: Material production - material processing - material application . Hanser Verlag, 2009, ISBN 978-3-446-41711-3 , pp. 215 .
4. ^ Hari Singh Nalwa: Handbook of thin film materials: Deposition and processing of thin films . Academic Press, 2002, ISBN 978-0-12-512908-4 , pp. 416-419 .
5. ↑ cf. Diederik Depla, Stijn Mahieu: Reactive Sputter Deposition . Springer, 2008, ISBN 978-3-540-76662-9 .
6. ↑ Peter Gawlitza, Stefan Braun, Andreas Leson, Sebastian Lipfert, Matthias Nestler: Production of precision layers using ion beam sputtering . In: Vacuum in research and practice . tape 19 , no. 2 , 2007, p. 37-43 , doi : 10.1002 / vipr.200700310 .
7. ↑ Dietrich Widmann, Hermann Mader, Hans Friedrich: Technology of highly integrated circuits . Gabler Wissenschaftsverlage, 1996, ISBN 978-3-540-59357-7 .
8. ^ Diederik Depla, Stijn Mahieu: Reactive Sputter Deposition . Springer, 2008, ISBN 978-3-540-76662-9 , pp. 172 .