Nuclear Science and Techniques

LOW ENERGY ACCELERATOR, RAY AND APPLICATIONS

Cooling enhancement of planar balanced magnetron cathode

M'hamed Salhi，

Seddik El Hak Abaidia，

Brahim Mohamedi，

Sofiane Laouar

Nuclear Science and Techniques

Vol.28, No.8

Article number 111

Published in print 01 Aug 2017

Available online 29 Jun 2017

DOI：10.1007/s41365-017-0271-2

7300

In physical vapor deposition (PVD) on a magnetron cathode, temperature of sensitive components must be kept under threshold limit, so as to ensure the cathode reliability, the process reproducibility, and the best quality of thin films. This can be achieved by an adequate design to enhance the dissipation of heat generated at the cathode. In this paper, temperature distribution and streamlines velocity of the cathode coolant inside a cathode magnetron are analyzed by using CFD solver ANSYS FLUENT in the single phase method in combination with k-ε standards turbulent model. The results show that the design is appropriate under the calculation parameters, and for high heat densities some improvements are necessary to enhance heat dissipation and keep temperature under the threshold limit.

Magnetron cathodePVDTemperature thresholdCFDANSYS FLUENT

1 Introduction

Magnetron sputtering configuration (MSC) is a common method in physical vapor deposition (PVD) processes for a wide range of industrial coatings. It allows obtaining high quality functional thin films with appropriate reproducible properties [1,2]. Under a magnetic field and an oscillating electric field, electrons generated in glow discharge plasma per ions bombardment of target surface (cathode) follow a spiral closed path and are maintained close to the cathode surface. This high flux of electrons creates high density plasma and high temperature [3] by means of electron-atom collisions and ionization. This, in turn, leads to higher sputtering rates and, therefore, higher deposition rates at the substrate (anode). The main factors limiting the deposition rate are thermal conductance of the target, efficiency of the cathode cooling, melting point of the target, and the sputtering yield [4].

Among MSCs of different types, the planar balanced or unbalanced magnetron source is the simplest and commonly used [5-9]. An excessive local heating of the principal components (namely the target and magnets), under effect of the delivered power, can take place if the heat dissipation and thermal contact between the components are not suitable. Without an adequate cooling of the cathode, structures of the magnets can be affected by overheating (over the Curie temperature), resulting in a disorder of the uniform distribution of the field lines and consequently the deviation of the trajectories of plasma electrons occurs. On top of that, the erosion of the target may occur in a non-uniform manner [10].

Kelly and Arnell [2] reported that the target could reach a local temperature of 400°C which could damage the target and magnets. Lake and Harding [11] designed an adequate cooling system to ensure a good heat dissipation and a better thermal compatibility between the mechanical components of the magnetron source. Takatsuji et al. [12] found that the flow rate of target cooling in a DC magnetron sputtering affected greatly the crystalline structure of deposited aluminum film, with a low texture in (111) orientation at low flow rate, and at low flow rate of cooling the high temperature cathode caused a transformation of certain components and degassing of the sputtering chamber walls. Baek and Kim [13] used two shapes of cooling water path in rectangular magnetron source for enhancing heat dissipation of the target. Doerner et al. [14] found that whatever the magnetic properties of magnets, an over-threshold temperature increase of the cathode often led to an increase in deposition rate, consequently the sputtering of cathode atoms becomes uncontrolled and the deposited films is not in desired quality. Recently Caillard et al. [15] used the calorimetric method to study the energy transfer to the substrate during a cathode magnetron sputtering of a pure nickel target, and showed that the rate of pulverization and the coefficient of secondary ion emission increased at the Curie temperature (T_C=358°C). From this temperature threshold the discharge illuminates in visible and infrared ranges.

The present paper is concerned with computational fluid dynamic (CFD) to study magnetron cathode (MC). The focus is placed on the numerical simulation of the pre-conditioning of single phase flows in an annular path target cooling, titanium or nickel as appropriate, under heated wall conditions, with the target being a heat source of the magnetron cathode. CFD technique shows increasing promise for simulating vapor deposition processes. The commercial CFD solver ANSYS FLUENT is employed as the computational platform and the single phase method in combination with standard k-ε turbulent model is used. This model is the most general among all the models of flow.

2 Overview of the magnetron cathode design

A planar balanced magnetron cathode was designed at UR-MPE of Boumerdes University with the contribution of Nuclear Research Centre of Birine, in order to build a PVD bi-cathode magnetron system [16]. The cathode has a plane circular geometry with outer diameter of 90 mm (Fig. 1). NdFeB magnets with magnetic field of 400 mT intensity are arranged in such a way that one pole is positioned at the target central axis and the second pole is formed by a ring of magnets around the target outer edge. Targets of Φ50 mm are fixed tightly at the upper side of magnets and a 3-mm soft iron, as magnetic yoke, is placed at the lower side of the magnets in order to confine and convoy the magnetic field. The three components (target, magnets, and magnetic yoke) are logged in a stainless steel container. Thus, a tight annular cavity is formed between the interior surface of the outer magnets, the external surface of the central magnet, the lower surface of targets and the upper surface of the magnetic yoke. This cavity has two holes ofΦ6 mm for two flexible drains to feed and evacuate cooling water, which enters in the cathode through the inlet pipe and circulates the cavity to extract the heat generated in the target and magnets, and leaves through the outlet pipe.

Fig. 1

(Color online) Schematic cross section of the planar balanced magnetron cathode

3 Mathematical modelling

3.1. Governing equations

To study characteristics of the cooling effect on the target at different inlet velocities and heat fluxes, the governing equations of continuity, momentum and energy are expressed as follows [17,18]:

\nabla (ρ \vec{ν}) = S_{m},

(1)

where $S_{m}$ is the mass added to the continuous phase from the dispersed second phase and any user-defined sources.

\nabla (ρ \vec{ν} \vec{ν}) = - \nabla p + \nabla (\bar{\bar{τ}}) + ρ \vec{g} + \vec{F},

(2)

where, $p$ is the static pressure, $\bar{\bar{τ}}$ is the stress tensor, $ρ \overset{⇀}{g}$ is the gravitational body force, and $\overset{⇀}{F}$ is the external body forces.

\nabla (\vec{ν} (ρ E + p)) = \nabla (k_{eff} \nabla T + ({\bar{\bar{τ}}}_{eff} \vec{v})) + S_{h},

(3)

where, $k_{eff}$ is the effective conductivity; the first two terms on the right-hand side areenergy transfer due to conduction and viscous dissipation, respectively; and $S_{h}$ is the volumetric heat sources.

3.2. Turbulence model

Two-equation turbulence models allow the determination of both turbulent length and time scale by solving two separate transport equations. The standard k-ε model in ANSYS FLUENT falls within this class of models and has become the workhorse of practical engineering flow calculations since it was proposed by Launder and Spalding.

The turbulence kinetic energy, k, and its rate of dissipation,ε, are obtained from the following transport equations [17,19]:

\frac{\partial}{\partial t} (ρ k) + \frac{\partial}{\partial x_{i}} (ρ k u_{i}) = \frac{\partial}{\partial x_{j}} [(μ + \frac{μ_{t}}{σ_{k}}) \frac{\partial k}{\partial x_{j}}] + G_{k} + G_{b} - ρ ε - Y_{M} + S_{k},

(4)

and

\frac{\partial}{\partial t} (ρ ε) + \frac{\partial}{\partial x_{i}} (ρ ε u_{i}) = \frac{\partial}{\partial x_{j}} [(μ + \frac{μ_{t}}{σ_{ε}}) \frac{\partial ε}{\partial x_{j}}] + C_{1 ε} \frac{ε}{k} (G_{k} + C_{3 ε} G_{b}) - C_{2 ε} ρ \frac{ε^{2}}{k} + S_{ε},

(5)

where, G_k is the generation of turbulence kinetic energy due to the mean velocity gradients, calculated as described in modeling turbulent production in the k-ε models; Y_M is the generation of turbulence kinetic energy due to buoyancy, calculated as described in Effects of Buoyancy on Turbulence in thek-ε Models; $Y_{M}$ is the contribution of the fluctuating dilatation in compressible turbulence to the overall dissipation rate, calculated as described in effects of compressibility on turbulence in the k-ε models; $C_{1 ε}$ , $C_{2 ε}, C_{3 ε}$ and Cμ are constants;σ_k $σ_{k}$ and $σ_{ε}$ are the turbulent Prandtl numbers for k and ε, respectively; and $S_{k}$ and $S_{ε}$ are user-defined source terms.

The turbulent (or eddy) viscosity is computed as follows:

μ_{t} = ρ C_{μ} \frac{k^{2}}{ε} .

(6)

4 Numerical solution

4.1. Model and grid systems

The MC cooling water passes through a soft iron plate inΦ6 mmpipes and flows into the space interposed between the two magnets which lead to the titanium or nickel target.For the numerical simulation, after designing and drawing the MC using Solidworks 3D CAD software, we adopted the commercial CFD simulation code ANSYS FLUENT (version 14.5) as CFD approach. For perfect reliability of the results, an extensive test for the confirmation of grid independence of the model were carried out by increasing the mesh density and adopting various mesh grading until further refinement showed a difference of less than 1% in two consecutive sets of results [19] (Fig. 2). Given the geometry simplicity, and to enable saving mesh elements and reduce the error, the choice has been made on the structured grid contained 2405001 quad nodes and 2301728 elements.A 3D steady state numerical model of heat transfer with four resolutions to ensure mesh independenceis adopted. The problem is processed as a single-phase turbulent flow.

Fig. 2

(Color online) Meshing of the coolant inside magnetron cathode

4.2. Numerical method and boundaries conditions

In this paper, the turbulence phenomena is described by a classical standardk-εmodel in combination with "standard wall treatment" for the near-wall treatment, the value of y⁺ between 1.5 and 5.17, as appropriate, is considered reasonable for the selected near-wall treatment approach. The choice of these models was based on their wide availability in broader literature devoted to this field, and after several convergence tests by combination of various models of turbulence. Table1 lists correlated information for the numerical study.

Magnetron cathode numerical model input

Models		Status
Solver	Time	Steady
	Type	Pressure based
	Velocity formulation	Absolute
	Gravity(Y-direction)	−9.81 m/s²
Models	Energy	Active
	Multiphase	Off
	Viscous	Standardk-ε
	Near wall treatment	Standard
Solution Methods	Scheme	SIMPLEC
	Gradient	Least square cell based
	Pressure	Standard
	Momentum	Second order upwind
Turbulent kinetic energy	First order upwind
	Turbulent dissipation rate	First order upwind
	Energy	Second order upwind

The Ti and Nitargets were numerically analyzed at cooling velocities of 0.01, 0.05 and 0.1 m/s, each at heat fluxes of 12738 and 25476 W/m²(totaling 12 simulations). We used the CFD code ANSYS Fluent 14.5 to simulate the flow around the target. The equations governing the flows were discretized with the finite volume method (FVM) which is well suited for simulating various types of conservation laws. The SIMPLEC algorithm (semi-implicit method for pressure-linked equations-consistent) wasused to solve the Navier-Stokes equations. The standard k-ε model was used to analyze the turbulent flow. For a better accuracy, no-slip boundary conditions for velocity at the walls were assumed.The boundary conditions for the simulation are given inTable 2, and material properties, in Table 3.

Initial conditions

Parameters	Values
Inlet velocity(m·s⁻¹)	0.01, 0.05, 0.1
Inlet water temperature(K)	303.15
Wall heat flux(W·m⁻²)	12738, 25476
Pressure(Pa)	1001325

Material properties

Properties	Water	Ti	Ni
Density(kg·m⁻³)	998.2	4500	8900
Specific heat(J·kg⁻¹·K⁻¹)	4182	523	460.6
Thermal conductivity(W·m⁻¹·K⁻¹)	0.6	21.9	91.74
Viscosity (10⁻⁶kg·m⁻¹·s⁻¹)	1003	—	—

5 Results and Discussion

The temperature contours and axial temperature profiles (Fig. 3) show that the Ni and Ti target temperatures at local positions can exceed the Curie temperatures of the magnets used.

Fig. 3

(Color online) 3D temperature contours (a) and axial profile (b) of Ti and Ni targets at different inlet coolant velocities and heat flux densities.

At the flow velocity of 0.01 m/s and heat density of 25476 W/m², the Ti and Ni targets reached maximum temperatures of 509K and 482K, respectively, in the surrounding area of the outlet water orifice. At 0.05 m/s and 25476 W/m², the target temperatures were favorable to have an appropriate sputtering yield but still above the optimal functional temperature of the magnets. Then, 0.1 m/s is an improvement flow velocity for cooling of the sensitive components of the magnetron cathode, with maximum temperatures of the Ti and Ni targets being 347K and 338K, respectively. The temperatures can be decreased more by heat dissipation between the cathode and sputtering chamber.

Figure 4 shows the velocity streamlines of the cooling water path in the embedded blank between the two magnets and the target at flow rate of 0.1 m/s. The cooling fluid was roughly slow-moving close to the contact surfaces of the peripheral magnet, and the outside surfaces of central magnets were under supplied. Clearly, boundary layers were growing in near surface regions. This is a result of the momentum exchange occurred between the mainstream fluid and coolant in the shear layer.

Fig. 4

(Color online) Velocity streamlines of the coolant water path in magnetron cathode.

The results show that for increasing power supply of the cathode, the inlet coolant temperature should be decreased, otherwise the sensitive components would attain the limit functional temperature. An improvement of the cathode design is necessary for depositing thin films with high powers. It is suggested that this improvement may be done by increasing the volume of the embedded blank, by changing the flow path inside the cathode and by improving the heat dissipation.

These can be done by:

(1) Enlarging the annular cavity of circulation of cooling water in contact with permanent magnets by reducing thickness of the stainless steel container and/or diameter of the central permanent magnet, and thickness of the annular permanent magnet, while preserving the magnetic field intensity and the configuration of magnetic field lines;

(2) Modifying the inlet and outlet sections of cooling water into a divergent or convergent section, instead of the uniform circular section;

(3) Using compatible materials with good thermal conductivity in particular the material of the container of the cathode;

(4) Reducing the target area and optimizing the cathode-anode distance; and

(5) Connecting only the target to the power supply, not the container, magnetic yoke and target cover.

6 Conclusion

The cooling effect on nickel and titanium targets and permanents magnets in dual cathode magnetron sputtering system was simulated at different water coolant inlet velocities and power densities. The results show that the magnets reached Curie temperature, and the temperature of both targets rises abruptly at low inlet velocity and high power density. At flow velocity of 0.1 m/s, an improvement of cooling of the sensitive components was observed. Flow path of cooling water allowed to evidence that there was roughly slow moving and undersupplied area. Parametric study will be done to further enhance heat removal from magnetron cathode and overcome cooling deficiencies at high power density, considering the cathode design, thermal hydraulic parameters of the coolant water, thermal gradient and mechanical deflections, sputtering conditions and quality of the obtained thin films.

References

[1]

K. Wasa, S. Hayakawa, Handbook of Sputter Deposition Technology: Principles, Technology and Application. Park Ridge (New Jersey, USA), Noyes Publications, 1992.

[2]

P.J. Kelly, R.D. Arnell,

Magnetron Sputtering: A review of recent developments and applications

. Vacuum, 2000, 56(3): 159-172. Doi: 10.1016/S0042-207X(99)00189-X