IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS

Computer simulation of metals ionization process during electron-beam

welding
To cite this article: R V Rodyakina et al 2020 IOP Conf. Ser.: Mater. Sci. Eng. 759 012022

View the article online for updates and enhancements.

This content was downloaded from IP address 181.214.179.221 on 08/02/2020 at 12:20

EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

Computer simulation of metals ionization process during

electron-beam welding

R V Rodyakina, A V Shcherbakov and D A Gaponova

National Research University “Moscow Power Engineering Institute”, 14,

Krasnokazarmennaya str., Moscow, 111250, Russia

E-mail: [email protected]

Abstract. One of the most actual problems during electron beam welding of metal
materials is the determination of the laws of primary electrons interaction with a gas
(vapor) flow, taking into account existing and generated (by charged particles) electric
fields. The paper is devoted to the development of computer simulation method for the
process of electron beam passage through a layer of evaporated metal. The analysis of
simulation results showed a significant effect of ionization process on electric field in
a gas-vapor channel. It was established that at the initial stage of the process, a region
with a positive potential is formed in the lower part of gas-vapor channel, and in its
upper part, on the contrary, potential decreases.

1. Introduction
Currently, many areas of machine-building production are intimately associated with making such
products that are subject to increased requirements for strength, ductility, corrosion resistance, etc.
One of the promising areas that can successfully overcome different technological difficulties arising
during the process of such products making is electron beam processing. Electron beam heats metal
(during electron deceleration), causes its melting and evaporation.
One of the most actual tasks is to determine the laws of interaction of primary electrons with a gas
(vapor) flow, taking into account existing electric fields and generated by charged particles. Therefore,
this paper is devoted to the development of computer simulation method for the process of electron
beam passage through a layer of evaporated metal.

2. Method of computer simulation

As the initial data for modeling, it is necessary to set medium density or concentration of particles . It
is known [1] that density of particles above the surface of penetration channel is about 10 17 per 1 cm3,
and above the surface of molten metal (at the bottom of the channel), it approaches to the density of
particles in the solid state, i.e. it is about 1024 in 1 cm3. The computer simulation process was
performed for aluminum.
In addition to ionization, the process of electron passage through a vapour layer will be
accompanied by energy losses for other types of interactions, for example, excitation of phonons,
plasmons, deceleration radiation generation and recombination processes. In addition, it is possible to
change the direction of primary electrons motion due to elastic interactions [2]. The purpose of this

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

study is to analyze ionization processes, potential distribution and ionization losses.

In the simulation process, we shall consider inelastic interaction cross-section — ionization cross-
section of gas (vapour) molecules by electron impact. As is known, ionization cross-section of an atom
or molecule by electron impact is a function of electron energy [3]. Typically, ionization cross-section
is about 10–16–10–17 cm2. The dependence of ionization cross-section of aluminum atom (σAl) on
electron energy used in the proposed model is shown in Figure 1. The model diagram is shown in
Figure 2. Target 1 (product) of aluminum has a cylindrical shape, its lateral and lower faces are
grounded (their potential is taken equal to zero). The gas-vapor channel in the model is approximated
by normal distribution function (item 2 in Figure 2). The distribution of vapors density spreading in
vacuum 3 in the channel and beyond it is specified. Potential on the channel wall and on the upper
surface of cylinder is determined by field potential created by electron beam 4. Electron flux density
obeys the normal distribution function. Ionization processes occur in the vapor-gas channel, and
charged particles move under the action of Lorentz forces in the field, created by the particles
themselves. This field does the work and changes trajectories of the particles.
10

6
i

1 10 100 1000 10000

Ti

Figure 1. Dependence of ionization cross-section for aluminum atom on electron energy:

Тi – electron energy values, eV; σi – values of ionization cross-section, × 1016, cm2 [7]

The calculation of electric field is carried out in an axially symmetric formulation. For this case, the
equation can be used
 2U 1 U  2U  (r , z )
+ + 2 =− , (1)
r 2
r r z 0
where U is electric potential (V); r and z are radial and axial coordinates, respectively (m), 0 is
permittivity of vacuum (F∙m–1); (r,z) is volume density of spatial charge created by charged particles
beam in the zone of field action (C∙m–3). Its equation can be written as
dQ
= , (2)
dV
where dQ is charge part, distributed over the elementary volume dV. A cylindrical grid is
superimposed on two-dimensional computational region, cells of which are numbered along r-axis as
i = 0,1,2 ... imax, and along z-axis as j = 0,1,2 ... jmax (Figure 2). The sizes of cylindrical grid cells are r
and z, respectively. In accordance with selected grid scheme, calculation values Ui,j are performed in
consecutive order for all nodes of the grid and are stored in the data array U(i,j). A finite-difference
approximation of computational region was used, for which equation (1) can be written as
U i +1, j − 2U i , j + U i −1, j 1 U i +1, j − U i , j U i , j +1 − 2U i , j + U i , j −1 i , j
+ + =− . (3)
r 2
i r r z 2
0

2
EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

Figure 2. Diagram of two-dimensional model of the process of metal vapor ionization in a gas-vapor
channel: 1 - grounded product; 2 - gas-vapor channel; 3 - vacuum, metal vapor; 4 - electron beam;
5 - equipotentials; elements of the computational grid and diagram of electron impact ionization
process are shown on the right

This equation was solved by the method of Euler. Then, using grid function U(i,j) obtained, vector
field of electric field strength E (V/m) was calculated by numerical differentiation
E = −gradU , (4)
as well as projections of this vector on the coordinate axes Er and Ez for each node of the
computational grid. The acceleration components, acquired by a particle when moving in an
electrostatic axially symmetric field due to the Coulomb force, are determined from the equatuion
 qEr
 a r =
m ,
 qE z
(5)
a z =
 m
where ar and az are the radial and axial components of acceleration (m∙s–2); q is particle charge (C);
m is particle mass (kg). The task of particle trajectory calculating is solved with a fixed time
integration step (which was taken equal to 10–12 s). Equations for the projections of speeds increments
vr, vz and coordinates increments Sr, Sz of the particle in the time interval (1, 2) are written as
follows:
2
q 2

vr =  Er dt , Sr =  vr dt , (6)

m 1 1
2 2
q
v z =  E z dt , S z =  vz dt . (7)
m 1 1
The distribution of current density in the product was calculated on the basis of Ohm's law in
differential form
j =E, (8)

3
EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

where j is current density vector (A∙m–2),  is specific conductivity of product’s material (aluminum),
Ohm–1∙m–1.
The most difficult stage of the problem to solve is simulation of ionization process, which is
accompanied by energy exchange between particles and appearance of a free electron. In accordance
with definition of the term “interaction cross-section” [4], mean free path of an electron λi between
ionization acts is defined as
1
i = , (9)
 i
where  is concentration of aluminum atoms (m–3). Concentration of atoms is related to vapor
pressure by ideal gas law
P
= , (10)
kT
where T is vapor temperature, k is Boltzmann constant. At the current stage of research, a linear
approximation of pressure distribution function over the channel depth P(z) was used, which is agree
with published works [5].
To take into account random nature of vapour particles distribution, the Monte Carlo method [6]
was used. In this case, path si, which electron passes between ionization acts, will be determined by
the equation
si = −i ln(1 − R1 ) , (11)
where R1 is a dimensionless random value, uniformly distributed in the range from 0 to 1. As a
result of ionization, a pair of such particles, as positive ion and slow electron, appears; their
trajectories will also need to be calculated in each time interval.
The initial energy of secondary electron Ws (eV), which appeared as a result of ionization, is
proposed to determine by the equation
  W p − J  
Ws =   tg  R2 arctg   , (12)
  2  
where Wp is energy of bombarding (primary) electron (eV); J is ionization potential;  is tabular
value, obtained by comparing theoretical and experimental data (for aluminum this value is equal
2.7 eV); R2 is a dimensionless random value, uniformly distributed in the range from 0 to 1.
In the case considered above, primary electron energy loss due to ionization Wpi (eV), was
calculated by the equation
W pi = J + Ws . (13)
Path that an electron travels between elastic interaction acts se is determined by the equation
se = −e ln(1 − R3 ) , (14)
where R3 is a dimensionless random value, uniformly distributed in the range from 0 to 1.
In order to study ionization processes, calculations in accordance with expressions (1)–(14) were
implemented in the Elion program module in C++.

3. Results of simulation and their discussion

At the current stage of research, the most interesting was information on the distribution of electric
field potential, created by particles in the penetration channel and near it, as well as spatial distribution
of particle ionization density acts. Figure 3 shows the results of computational experiments, realized
on a personal computer with an IntelCorei7-6700 HQ processor.
Figure 3a shows the case corresponding to the absence of aluminum vapor concentration (it is
equal to zero) in the region above penetration channel. This situation corresponds to the initial stage of
electron beam impact on the target (in the real case, penetration channel is not formed yet, and sample

4
EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

surface is flat). The beam current is 18 mA. It is seen that a region with the smallest electric potential
is in the beam center. In the approach to the bottom of the channel, potential increases, approaching
almost to zero, since potential difference between the channel bottom surface and the lowest
(grounded) plane of the product is negligible for this case. The minimum potential value under
electron beam steady action on the product without taking into account beam interaction with metal
vapors is equal to –1.8 V. In the case when vapor pressure distribution of aluminum along Z-axis is
specified, the picture changes radically: potential distribution will be completely different after a time
on the order of 0.05 ns. Figure 3b shows such a case. Vapor pressure of aluminum in the lowest part of
the channel is 30 Pa, and at the point z = 0 (at the distance of 2 mm above the product surface) it is
equal to 0.1 Pa. The dependence P(z) at this stage of the research was assumed to be linear.

a)

b)
Figure 3. Electric field distribution: а) in the channel in the absence of metal vapor,
b) taking into account linear distribution of vapor pressure along coordinate z.

5
EBW2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 759 (2020) 012022 doi:10.1088/1757-899X/759/1/012022

Ionization of neutral aluminum atoms leads to the appearance of positively charged ions and slow
electrons. Due to the fact that electron beam has a current density distribution over its radius, most of
ions appear precisely in the center of the beam, and at the initial stage of ionization process, shown in
Figure 3,b, slow electrons will also appear near the center of the beam. Under the action of positive
ions field, localized mainly in the center of the beam, peripheral slow electrons will drift in their
direction, creating a sharp minimum of potential in the center.
In the lower part of the channel, on the contrary, a region with a positive space charge will be
formed, as evidenced by the fact that potential here will assume a positive value. It can be predicted
that, under the action of primary and slow secondary electrons, a region with a potential gradient
sufficient to accelerate ions in the direction of electron gun will be formed, and forming plasma will
correspond in its characteristics (density and temperature) to a glow-discharge plasma. However,
processes of ion acceleration simulation using the developed program, run on a computer with the
above-cited characteristics, may take considerable time (up to several days). This is due to a
significant difference in the masses of ions and electrons. The most obvious solution of this problem is
transition to parallel computing technology.

4. Conclusion
1. It is shown that, using proposed complex physical and mathematical model, it is possible to study
plasma formation process during the influence of electron beam on metals.
2. Analysis of the first simulation results has been carried out. There was shown a significant effect of
ionization process on electric field distribution in gas-vapor channel. It was established that at the
initial stage of the process, a region with a positive potential is formed in the lower part of gas-vapor
channel, and in its upper part, on the contrary, potential decreases.
3. It is shown that for effective application of the developed model and studying of the drift process of
metal ions, having a significant mass in comparison with electrons, using of parallel computing
technology is necessary.

5. Acknowledgments
The research was carried out in National Research University “Moscow Power Engineering Institute”
with the financial support of the Russian Science Foundation (project No. 18-19-00652).

References
[1] Novokreshchenov V V, Rodyakina R V and Karimbekov M A 2017 Solid State Phenomena 265
237–45
[2] Nazarenko O K, Kaydalov A A and Akopyants K S 1973 Automatic welding 8 75—6.
[3] Bondarenko G G, Kristia V I and Fisher M R 2005 Interaction of Ions with the Surface: The
XVII Int Conf (Moscow) vol 2 331–4.
[4] Nazarenko O K, Kaydalov A A and Kovbasenko S N 1987 Electron-beam Welding (Kiev:
Naukova Dumka) p 256
[5] Rai R. 2008 Modeling of Heat Transfer and Fluid Flow in Keyhole Mode Welding (A Diss. in
Materials Science and Engineering) (USA: The Pennsylvania State University) p 206
[6] Kvlividze V A and Krasilnikov S S 1985 Introduction to the physics of atomic collisions
(Moscow: Publishing House of Moscow State University) p 224
[7] Kim Y K and Stone P M 2001 Phys. Rev. A 64 052707