nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
1 Introduction
Modern technologies and their components, such as cell phones, laptops, and smartwatches, rely on cutting-edge technology to be manufactured. Therefore, they typically require small parts obtained from miniaturized wafers using thin-film technology [1,2,3,4,5].
In 1912, Pohl and Pringsheim accomplished thin films. They published a famous work on the production of mirrors using a process of metal vaporization such as silver (Ag) and aluminum (Al) in a high vacuum [6,7,8]. Later, with research and development in this field, the application of this domain was seen in the semiconductor industry, which represents the primary manufacturing technology, including telecommunication apparatus, integrated circuits (IC), transistors, solar cells, LEDs, photoconductors, magneto-memory optics, compact discs, and so on [9,10,11,12].
Several methods can be used to perform these miniaturization techniques where the sputtering process under the physical vapor deposition (PVD) system presents powerful results. For example, the magnetron sputtering technique allows metallic or non-metallic layers to be deposited on the substrate. Using a magnetron associated with the cathode creates a magnetic field that imposes the trajectories of the ejected electrons from the target, expanding the bombardment density [13,14,15,16].
Various simulation programs have been used to simulate the sputtering deposition process. Ready to cite an effective program based on the Monte-Carlo simulations SRIM [17].
During the simulation, the particle moved in a sequence of discrete steps, each of which collided with an atom of the target in an arbitrary mode, reproducing the statistical properties of the shutdown processes [18,19,20]. The composition development of binary thin films takes place experimentally and with Monte Carlo codes based on TRIDYN (DYNamic TRAnsport of Ions in the matter) [21] and TRIM (TRansport Ions in Matter) [17]. Another code, SIMTRA [22], was developed by Van Aeken for the trajectory simulation of gas-phase particles in a definable 3D configuration. The background molecules’ interatomic collision modeling, potentials, and thermal movement are included in the code [23,24,25].
The morphology of the layers deposited via the sputtering method relies on the material, vacuum chamber gas, and numerous deposition parameters, such as pressure, target-substrate distance, temperature, substrate bias, and chemical composition [26,27].
Deposition using the sputtering approach is a complex task; it is suitable for creating a process model for acquiring the physical values of sputtering. We used a model to simulate the sputtering and transport processes, as in our previous work [1,2,3,4,5].
It is widely known that the properties of thin-film materials depend on their sputtering state. However, the relationship between the plasma parameters of sputtering and the properties of thin films is unclear. The optimal sputtering conditions must be determined by preparing many samples under varying sputtering parameters. The results obtained help understand the effect of process parameters on transport, the sputter deposition process, and the resultant composition of films.
Several experimental studies have been conducted [28,29,30,31] showing the relationship between sputter deposition parameters and the quality of thin films obtained, thus validating the simulation models.
Previous works by different groups [32,33,34] do not permit target properties when utilizing metals, semiconductors, and dielectrics. Their posted results, primarily based on Monte Carlo models, are typically speaking, although it is imperative to recognize the distinction in using each. Furthermore, other groups [35,36,37] have demonstrated a close relationship between the transport conditions of the particles from the target to the substrate and their arrival energies.
Before deposition, the sputtered atoms underwent various binary collisions with plasma particles. This amount is vital because each collision induces a trajectory and energy variation. Therefore, the collisions with the background gas atoms were disregarded relative to the sputtered particles’ critical number of kinetic energies.
The morphology and quality of thin-film materials depend highly on the sputtering conditions. However, our group intends to examine the link between thin-film properties and plasma parameters. Thus, in earlier research cited in [1-3], we have shown the following:
-
In Ref. [1]: The simulation results obtained from a Monte Carlo code based on experimental and algorithmic calculations have been compared with those of other authors like Mahieu et al. [38]. They applied a Monte Carlo code to simulate the transport of atoms in DC magnetron sputtering, studied the thickness profiles, and simulated the power and path of sputtered particles arriving at the substrate.
Results: Target-substrate distances of 8 — 20 cm and a 0.3-1 Pa pressure range correspond well with the results above.
-
In Ref. [2]: The results acquired with analytical formulation primarily based on the Monte Carlo approach proposed by researchers such as Sigmund and Yamamura were compared with SRIM code simulation to validate models.
Results: The Yamamura model and SRIM simulation approximated the sputtering yield rate, mainly when no experimental data were available. SRIM was validated for energies above 100 eV but is not appropriate for less than this value.
-
In Ref. [3], the simulation results using the Monte Carlo code of the sputtering yield for metals and semiconductors were obtained by varying the energy and angles of incidence.
Results: As many ejected particles ensure a very thin and uniform film deposition, which we are looking forward to obtaining, we deduced the optimum energy and angle that should be applied in this way.
The present work studied the influence of two other parameters, temperature, and high pressure, on the sputtering process for silicon, germanium semiconductors, and copper metal. Continuing this goal, we started in our previous work by providing the necessary information through simulation and analytical computation, including all parameters that affect the desired thin-film realization that can be considered in experimental research or industry.
2 The sputtering process requirements
It is vital to understand the steps of the reactor equipment process to achieve a better deposition morphology of thin films, considering what occurs inside the vacuum chamber during the disposition cycle. Therefore, a preparation process was required before the deposition. Before its placement, cleaning the substrate outside the vacuum chamber in an ultrasonic bath resulted in outstanding film adhesion. This technique transforms electrical power into mechanical vibrations within a cleaning liquid. It is quick and less expensive because it does not require too much solvent and is easier to handle than an ionic bath [39,40,41].
The cleaning procedure takes a long time, which is a disadvantage in terms of industry efficiency because it boosts the final product costs. Controlling expenses requires managing the breakdown times and setup of the machine. Because this is a disadvantage to the industry, process parameter optimization is essential for reducing production times. The deposition rate is an important parameter that must be tuned to enhance the plasma density and energy available in the process. Therefore, it is necessary to consider all steps and parameters being studied to comply with the industry demands [42]. Furthermore, to obtain significant sputtering, it is necessary to go through the steps depicted in the following:
— The first step (adjustment) consisted of preparing the vacuum chamber, which is composed of a gradual increase in temperature created by tubular heating and a modular control system. Simultaneously, vacuum pumps were activated to reduce the pressure inside the chamber. For this, two pumps are used; the first produces a pressure of up to 10-5 bar, and the second (high vacuum) creates a pressure of 10 — 7 bar.
— The second step (engraving) is characterized by cathode cleaning, where the substrate is bombarded with ions from plasma etching to remove localized contamination on the substrate surface. This step is crucial before starting sputtering because it helps to increase adherence.
— Third step (gas integration): Before starting, we must integrate a gas that aims to bombard our target. Several gas choices are available, such as argon and xenon; the choice of gas is significant because it contributes to the sputtering efficiency.
— In the fourth step (coating), the sputtering process is started, and the material to be deposited is projected onto the substrate surface.
— The final step (ramp down) corresponds to the cooling of the chamber; a specific system must be applied to return the vacuum chamber to room temperature and ambient pressure [43,44,45].
3 Simulation protocol
The development of simulation models for sputtering with appropriate physical parameters can reduce time-consuming and costly trials in the operation and control of deposition systems used to fabricate thin films.
The simulation of film growth on time scales of seconds or minutes is possible using kinetic Monte Carlo algorithms [46,47].
This approach can model different surface processes such as nucleation, growth, post-deposition, and structural modification of films [48,49].
3.1 Simulation codes
The Sigmund-Thompson distribution describes the initial energy distribution well;
the angular distribution of the sputtered particles is considered as
The kinetic energy and the number of atoms arriving at the substrate were calculated using SRIM and SIMTRA.
The simulation was divided into two parts:
-
— The main goal of the first part is to calculate the sputtering yield Y(E) using the Monte Carlo simulation program SRIM, which uses the binary collision approximation (BCA) applied to ion-solid interactions [17]. This is an open-source program that is convenient for practical use.
Different angles of incidence were used (θ = [0o, 15o, 30o, 45o, 60o, 75o, 85o, and 89o]), and the sputtering yield was varied according to the applied energy (E = [100 eV, 1 KeV, 10 KeV, 100 KeV, and 1000 KeV]), and the target was bombarded by the vacuum chamber gas ions (argon and xenon) using the materials (Si, Ge , and Cu).
Each material must consider the incidence angle and bombardment energy to obtain the optimum results. Therefore, we note the bombardment energy and the incidence angle, which give the highest sputtering yield, named
-
— The transport from the target to the substrate of the sputtered atoms (obtained from the sputtering yield calculation by applying
Diverse temperatures (T = [100, 300, and 600 K]) and pressures (P = [0.5, 2, and 5 Pa]) were applied. The first parameter represents the three temperature states of the vacuum chamber in Kelvin, where 100 K
The main objective was to determine the effect of these two parameters on the number and energy of atoms that have reached the substrate. Moreover, deduce the temperature and pressure that should be applied to acquire the maximum number of atoms and hence the desired thin films.
3.2 Simulation model
As is most commonly used in industrial applications, we created a vacuum chamber with dimensions of 30 × 30 × 50 cm, a distance of 8 cm between the target and substrate, a circular shape of 2 cm in the magnetron target’s radius, and 7 cm in the substrate radius (as shown in Fig. 1).
Using SRIM, we injected approximately 105 argon and xenon ions
intended to bombard with the Eopt of energy and
The gas temperature and pressure were varied to determine their influence on the number of atoms reaching the substrate and their energies.
After these configurations, we can start the simulation of this model to calculate the total number of particles arriving at the substrate, their positions, and their energies. The results were saved to data files and represented as curves.
4 Results and discussion
4.1 Sputtering yield calculation
The sputtering process involves momentum transport in a collision cascade initiated by the incident particles in the solid surface layer. A surface atom is ejected if its energy exceeds its surface binding energy. To evaluate the number of atoms ejected under ion bombardment, we calculated the sputtering yield Y(E), which quantifies physical sputtering and is defined as the mean number of atoms removed from the surface of a solid per incident ion [2,3,4,52].
4.1.1 Sputtering yield depends on the bombardment energy and incidence angles
Figures 2 and 3 below represent sputtering yield rates obtained with SRTM code simulation for Si, Ge, and Cu materials bombarded with Ar and Xe vacuum chamber plasma gas, applying different energy and incidence angle values.
Figure 2 Sputtering yield as a function of the energy of the bombardment ions (Ar) for various incidence angles calculated by the MC method for a) Si, b) Ge, and c) Cu materials.
Figure 3 Sputtering yield as a function of the energy of the bombardment ions (Xe) for various incidence angles calculated by the MC method for a) Si, b) Ge, and c) Cu materials.
By looking at these two Figs. 2 and 3), we can deduce those observations:
Bombarding with Xe ions yielded a better sputtering yield than bombarding with Ar ions. Xe offers approximately 4-fold sputtered atoms compared to Ar and those for metals and semiconductors.
At [100 KeV/ 85o], for both Ar and Xe gases, the sputtering yield reached the maximum value, which represents the supreme value of the ejected atoms obtained from the sputtering process for different materials (Si, Ge, and Cu).
4.1.2 Comparison between different materials
In Fig. 4, we present the curve results of the sputtering yield calculations obtained by SRIM simulations. At an incidence angle of 85o, the sputtering yield varied according to the energy, and the target was bombarded with argon and xenon ions using the materials (Si, Ge, and Cu).
Figure 4 Comparison of sputtering yield on 85o of incidence angle as a function of the energy of the argon and xenon bombardment ions.
As illustrated in Fig. 4, increasing the
bombardment energy created a peak called
Copper is an excellent conductor, whereas germanium and silicon are semiconductors. Thus, the resistance of copper decreases, but that of germanium and silicon increases when they are subjected to an increase in the energy of the gas bombardment ions, which influences the atomic bond and leads to a significant difference between the ejected particles of metal (Cu) and semiconductors (Ge and Si).
The following regions can be developed for these curves:
— In region 1, sputtering at low energy (subthreshold): in this region, which covers the energy range from 0 to 80 eV, the bombardment energy is insufficient for the sputtering process to begin because either the argon or xenon ions have insufficient energy to reach the target or their energies are below the surface binding energy of the cathode materials.
— In region 2, knock-on sputtering, commercial and industrial sputtering applications are most interested in ions with energies in this region (approximately between 100 eV and one keV). The sputtering process began when the bombardment energy reached a threshold. Furthermore, the surface and near-surface atoms can be energetically removed from their equilibrium locations once the energy of the sputtering ions exceeds the surface binding energy of the cathode material. Increasing the bombardment energy accelerates this process. This area is known as the liner cascade regime [53,54,55].
— In region 3, nonlinear cascade sputtering appeared for ions with threshold energies above one keV. The incident ions are sufficiently strong to remove several cathode atoms. At approximately 80 — 100 KeV, the sputtering process reaches its limit. In this situation, the ions have high energy, and the striking particles penetrate deeply into the target, preventing the recoil atoms from escaping. Therefore, the addition of more energy reduced the yield. This regime is often not of industrial interest because of the high energy required and the high ejection energies of sputtered atoms.
Typical operating voltages used in commercial and industrial sputtering
applications are between 100 eV and one keV [53-55]. Therefore,
Table I Sputtering yield for 1 KeV of the energy of the bombardment ions (Ar and Xe) as a function of various incidence angles calculated by the MC method for Si, Ge, and Cu materials.
| Sputtering Yield | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 0 Deg | 15 Deg | 30 Deg | 45 Deg | 60 Deg | 75 Deg | 85 Deg | 89 Deg | ||
| Si | 0.5822 | 0.582 | 0.914 | 1.9 | 3.15 | 4.17 | 3.18 | 1.93 | |
| Ar | Ge | 1.89 | 2.08 | 2.18 | 2.89 | 3.44 | 3.56 | 2.65 | 1.75 |
| Cu | 3.71 | 3.83 | 4.31 | 5.15 | 5.91 | 5.29 | 3.79 | 2.67 | |
| Si | 0.399 | 0.404 | 0.807 | 1.79 | 3.5 | 5.48 | 4.76 | 3.07 | |
| Xe | Ge | 0.834 | 0.84 | 1.38 | 2.45 | 3.96 | 4.78 | 3.7 | 2.66 |
| Cu | 2.04 | 2.07 | 3 | 4.51 | 6.25 | 6.46 | 4.86 | 3.61 | |
Thus, we take the optimal values to
4.2 Influence of temperature and high pressure on the sputtered atoms arriving on the substrate using argon and xenon gas
4.2.1 Variation of temperature with a fixed pressure for argon ions
In the vacuum chamber, the target, which contains materials (Si, Ge, and Cu), is bombarded with argon gas ions with the optimal values of incidence angle and energy obtained from the SRIM simulation (75o and 1 KeV).
When the chosen pressure is 0.5 Pa, and with the temperature variation (100, 300 , and 600 K), the following figures represent the number of atoms that arrive on the substrate for each considered temperature, obtained with SIMTRA code simulation.
As shown in Fig. 5a), b), and c), the number of atoms that reached the substrate was different at each adjusted temperature. For example, when 100 K was applied, the total number of Si, Ge, and Cu atoms reaching the substrate was approximately 38000, 34000, and 37000, respectively. Applying 300 and 600 K increased the mobility and energy of the argon gas particles, bombarding the target with great energy on one side and expanding the number of ejected atoms on the other side (Table II).
Figure 5 Number of arriving atoms as a function of the atom’s kinetic energy for three different temperatures T = [100, 300, and 600 K] with Argon gas bombardment for a) Cu, b) Ge, and c) Si.
4.2.2 Variation of pressure with a fixed temperature for Argon ions
In this section, we investigate the effect of high pressure by applying three different pressures (0.5, 2, and 5 Pa) while maintaining the temperature inside the vacuum chamber at 100 K. Using the same gas (argon) and materials (Cu, Ge, and Si) to bombard the target, and the obtained results are presented in the Fig. 6.
Figure 6 Number of arriving atoms as a function of the atom’s kinetic energy for three different pressures P = [0.5 Pa, 2 Pa, 5 Pa] with Argon gas bombardment for a) Cu, b) Ge, and c) Si.
According to the above graphs (Fig. 6a), b), and c)), the pressure significantly influences the number of atoms arriving at the substrate. When 0.5 Pa was applied, approximately 38000, 34000, and 37000 arriving atoms for Si, Ge, and Cu were obtained.
Enlarging the vacuum chamber pressure to higher values produces many collisions and a massive drop in particle mobility. The ejected atoms collide with the argon ions, which decreases their kinetic energy and makes it difficult for them to reach the substrate. After applying pressures of 2 and 5 Pa, as shown in Table III, there was a substantial decrease in the number of atoms reaching the substrate, with approximately 50 % and 75 % of the total number of ejected atoms, respectively, and a significant decrease in their energies.
Table III The total number of material atoms arrived in the substrate depending on the pressure with argon gas ions bombardment.
| Temperature | Materials | Total number of atoms |
|---|---|---|
| 0.5 Pa | Si | 38315 |
| Ge | 33914 | |
| Cu | 37520 | |
| 2 Pa | Si | 19786 |
| Ge | 14050 | |
| Cu | 14934 | |
| 5 Pa | Si | 8651 |
| Ge | 6114 | |
| Cu | 7581 |
Low pressure inside the vacuum chamber will cause minor collisions between the particles; therefore, the ejected atoms will have a free path to reach the substrate.
4.2.3 Variation of temperature with a fixed pressure for Xenon ions
The same procedure will be carried out; temperature range T = [100, 300, and 600 K] and 0.5 Pa as a static vacuum chamber pressure; the only difference is that we will now be using xenon instead of argon as the bombardment gas.
Replacing argon with xenon as the vacuum chamber gas and applying the same procedure described in Sec. 4.2.1, the results shown in Figs. 7a)-c) indicate a considerable decrease in the number of atoms arriving on the substrate. For example, an applied temperature of 100 K yields approximately 28000 atoms for Si compared to 38000 for argon. Almost the same phenomenon occurs in other materials (Ge and Cu). Xenon ions are known for their high bombardment energies, which provide a massive number of atoms ejected and a good sputtering yield. However, these can cause reflection or trajectory changes in the ejected particles when they collide with those of Xe, so they will not be able to travel to the substrate. Therefore, we must always choose a target compatible with the chosen gas.
Figure 7 Number of arriving atoms as a function of the atom’s kinetic energy using three different temperatures T = [100, 300, and 600 K] with Xenon gas bombardment for a) Cu, b) Ge, and c) Si.
As shown in Table IV, increasing the temperature increased the total number of particles arriving at the final destination (substrate).
4.2.4 Variation of pressure with a fixed temperature for Xenon ions
As in Sec. 4.1.2, three different pressures P = [0.5, 2,5 Pa] and 100 K, were used as a fixed vacuum chamber temperature. Xenon replaces the argon gas used in the vacuum.
In the above figures (Figs. 8a)-c)), we can see that the application of xenon gas yields the same result as argon, except that the number of atoms that arrive at the substrate is dramatically diminished. As shown in Table V, for an applied pressure of 0.5 Pa, 28000 compared to 38000 for Si and approximately 26000, and 27000 compared with 34000 and 37000 for Ge and Cu, respectively.
Figure 8 Number of arriving atoms as a function of the atom’s kinetic energy for three different pressures P = [0.5 Pa, 2 Pa, 5 Pa] with Xenon gas bombardment for a) Si, b) Ge, and c) Cu.
Table V The total number of materials atoms arrived in the substrate depending on pressures with Xenon gas ions bombardment.
| Temperature | Materials | Total number of atoms |
|---|---|---|
| 0.5 Pa | Si | 28141 |
| Ge | 26940 | |
| Cu | 27676 | |
| 2 Pa | Si | 12224 |
| Ge | 11571 | |
| Cu | 1198 | |
| 5 Pa | Si | 5103 |
| Ge | 4628 | |
| Cu | 4986 |
When the pressure of 0.5 Pa was applied, the results proved that a large number of atoms arrived at the substrate compared to 2 and 5 Pa. In addition, most of these atoms arrived with energies in the range of 0.1 — 1 eV, as seen in Fig. 8, where the higher energy tail extends up to 100 eV but contains relatively few atoms. The findings, as mentioned earlier, are crucial since it is seen that energetic bombardment influences the characteristics, stress, microstructure, and surface roughness of the deposited films [32,56]. Therefore, high pressure considerably affected the morphology of the formed thin films, with a notable reduction in thickness and quality.
The SIMTRA simulation results were consistent with those reported in [32,57,58,59,60] regarding the impact of various vacuum chamber gas parameters on the kinetic energy of ejected atoms reaching the substrate.
4.3 Comparison between materials
To define the best parameters for improving the growth and quality of thin films, we compared the studied materials (Si, Ge, and Cu) for each influencing parameter.
According to the figures above (Figs. 9-11), the obtained results are as follows:
Si material gives the best result for arriving at the substrate atom numbers, followed by Cu and Ge in the last position, although Cu has the best sputtering yield. This implies that the Si semiconductor has better deposition efficiency than the other materials.
The desired thin films could be obtained at a low vacuum chamber pressure (in our case, 0.5 Pa).
In contrast, higher temperatures achieve the most important atom numbers at the substrate level. However, a severe amount arrives with high energy, which risks damaging the substrate or influencing the morphology of the film obtained.
Figure 9 Comparison between arriving Si, Ge, and Cu atoms as a function of the atom’s kinetic energy for 0.5 Pa and 100 K for a) Argon and b) xenon.
Figure 10 Comparison between arriving Si, Ge, and Cu atoms as a function of the atom’s kinetic energy for 0.5 Pa and 600 K for a) Argon and b) xenon.
5 Conclusion
The techniques for the deposition and coating of thin films are still in progress. However, deposition sputtering is one of the best-known methods to achieve the sound and desired quality of thin layers, depending on several factors.
Using the Monte Carlo codes such as SRIM and SIMTRA software described in this work allowed us to investigate the influence of temperature and high pressure on the magnetron sputtering process. We simulated several cases considering three materials (Si, Ge, and Cu) and two different bombardment gases (Ar and Xe) under temperature and high-pressure variations.
These findings demonstrate that the film characteristics are significantly influenced by the transport of sputtered atoms during the sputtering process. The acquired results are proven to help comprehend the impact of process parameters on transport and, as a result, in comprehending the sputter deposition process and the resulting film composition. They provided the following important information regarding the deposition of thin films:
— The increase in gas energy influences the atomic bonds of the materials, which leads to a significant difference between the ejected particles of metal (Cu) and semiconductors (Ge and Si).
— The increasing temperature values offer more kinetic energy to the particles, increasing their mobility inside the vacuum chamber and helping a large number of ejected atoms reach the substrate.
— The most important ejected atom number was determined when higher temperatures were applied. Nevertheless, the large number of atoms reaching the substrate when a high temperature is applied does not imply the obtention of the best thin film quality, awing that an essential number of atoms reaching the substrate with the highest energies (more than 20 eV) can ultimately damage the substrate structure or reflect, collide again with the ejected atoms or with the molecules of the gas, and influence the progress of the sputtering process. Low energies (in the range of 0.1 to 1 eV) were appropriate for building the finest films.
— In contrast, increasing the pressure from 0.5 to 5 Pa diminishes the particle’s kinetic energy and hinders their movement by creating more collisions and a reduced mean-free path for a significant number of ejected particles to reach the substrate and directly influencing the aspect of the thin film.
— Eventually, the plasma gas injected into the vacuum chamber was considered harmful. The wrong choice of gas will decrease the number of atoms arriving at the substrate.
-
— Argon is the plasma gas used to deposit thin films by sputtering rather than xenon in experiments and industries for the following reasons:
a Argon (Ar) is a noble gas recovered from the air, but its concentration is much higher (0.93%) and is not considered rare. Xenon (Xe) which comes under “rare gas”, is also a noble gas that can be obtained from air. Still, the economics of isolating them in air separation units depends on demand and pricing; their concentration in air is less than 20 parts per million (ppm).
b. When using xenon for sputtering, the ejected atoms from the target are approximately 40% more than argon, but most of them leave it with higher energy. Even if they collide with plasma particles, they reach the substrate with significant energy. Some are returned to the plasma, where others penetrate, resulting in lower film quality owing to less particle deposition than argon.
Adding this work to our previous studies, we can contribute to the thin film deposition field by optimizing those fundamental parameters (vacuum chamber gas, bombardment energy, incidence angle, target-substrate distance, gas pressure, and temperature) in the disposition of research and researchers.
