Professional Documents
Culture Documents
Kinetic modeling of Langmuir probe characteristics in a laboratory plasma near a conducting body
Phys. Plasmas 24, 012901012901 (2017); 10.1063/1.4972879
Strongly emissive plasma-facing material under space-charge limited regime: Application to emissive probes
Phys. Plasmas 24, 013506013506 (2017); 10.1063/1.4973557
An analytical model of multi-component radio frequency capacitively coupled plasma and experimental
validation
Phys. Plasmas 24, 013503013503 (2017); 10.1063/1.4973233
PHYSICS OF PLASMAS 24, 013507 (2017)
The behavior of floating potential in plasma can also 1 eAw ð xÞ
Ue ð xÞ ¼ ene ðxÞveth e
¼ enes vth exp ; (1)
provide valuable information about the negative ion parame- 4 KB Te
ters. Fernandez Palop et al.28 studied the variation in floating
potential in iodine plasma by varying the concentration of 1 eAw ð xÞ
U ðxÞ ¼ en ðxÞvth ¼ ens vth exp ; (2)
iodine I2 in an Ar/I2 discharge. Sherifian and Shokri29 have 4 KB T
reported the variation in floating potential for a large range Uþs ¼ nþs uBm : (3)
of electronegativity and an analogous work has also been
reported on the estimation for floating potential.30 From the In the above equations, the subscript, s, denotes the density
diagnostic perspective to measure electric field/potential of the individual species at the sheath edge, where the super-
inside plasma, the floating potential of the cold probe is rela- scripts have their usual meanings. Here, veth & v th are the
tively robust than that of the hot emissive probes since most mean velocities of electrons and negative ions
electronegative gases are highly corrosive. However, this 1 1
comes with a limitation that in most laboratory/industrial 8KB Te 2 8KB T 2
veth ¼ & vth ¼ : (4)
plasmas, foreseeable presence of non-thermal electrons can pme pm
drastically shift the floating potential to greater negative val-
ues, than that predicted based on single Maxwellian On the other hand, the modified Bohm speed uBm for the pos-
electrons. itive ions1 is expressed as
In the light of the above facts, the floating potential " #12
when considered carefully can be potentially useful for KB Te ð1 þ aÞ
validating analytical models concerning sheaths in low pres- uBm ¼ : (5)
mþ ð1 þ caÞ
sure electronegative discharges. Besides, it is also the aim
to establish the effectiveness of floating potential measure- For calculating floating potential, we apply the quasi-
ments to infer parameters such as a and c in electronegative neutrality condition at the sheath edge and the flux conserva-
plasmas. The basic motivation is therefore focused to under- tion can be expressed as
stand the behavior of floating potential in electronegative
discharges. In this paper, the floating potential of a cylindri- Uþs ¼ Ues þ Us ; (6)
cal probe immersed in electronegative oxygen discharge has
nþs ¼ nes ð1 þ aÞ: (7)
been investigated. In conjunction, parametric studies have
been performed to study the influence of electronegative The above equation can be solved to obtain an expression of
parameters a and c on the floating potential. The analytical the form
curves are compared with the experimental results to esti-
mate the range of negative ion temperatures in the discharge. P ¼ evw þ Qecvw : (8)
In Section II, analytical behavior of floating potential
in the presence of negative ions has been presented. The In the above expression, P, Q, & vw are given by P ¼
description about the experimental set-up is provided in 3 1 1
ðl2pðð1þa Þ 2 2
a 2 eAw
Section III. The experimental results including the phenome- þ 1þacÞÞ ; Q ¼ ðcl Þ ; vw ¼ KB Te and the constants have
nological model to determine a using the Langmuir probe been defined as follows:
have been presented in Section IV, followed by comprehen-
ns Te mþ m
sive discussion in Section V. Finally, Section VI ends with a a¼ ;c¼ ; lþ ¼ & l ¼ : (9)
nes T me me
brief summary and conclusion.
For oxygen plasma, Oþ
2 and O have been considered to be
II. FLOATING POTENTIAL IN THE PRESENCE the most dominant ionic species.33 Therefore, lþ and l can
OF NEGATIVE IONS
be calculated by considering mþ ¼ 32 amu and m ¼ 16 amu,
Consider an electrode immersed in plasma consisting of respectively. Equation (8) can be solved to obtain the absolute
electrons, positive ions, and negative ions. The electrode is value of floating potential, vw normalized to Te. The plot of vw
spontaneously charged to a negative potential by influx of as a function of a, for c ¼ 1, 10, and 100 has been shown in
thermal electrons. In the immediate vicinity of the electrode, Fig. 1. The figure shows that above a threshold value ac , the
the negative potential is shielded by a sheath of positive absolute wall potential vw tends to fall monotonically on
ions. The region outside the sheath is quasi-neutral with a increasing a. Below this range a < ac , the vw is observed to
weak electric field in the pre-sheath pointing from the bulk increase in the case of c ¼ 10 and 100.
towards the sheath edge. These positive ions prior to entering In Fig. 2, the maxima value of vw corresponding to ac
the sheath are accelerated in the electric field to reach the has been plotted for a range of c. It is found that vwc
characteristic Bohm speed uBm . In quasi-neutral equilibrium, increases monotonically as c values are increased from 1–70,
both electrons and negative ions are assumed Boltzmann whereas, ac tends to saturate at larger values of c. The overall
throughout the pre-sheath as well as inside the sheath.31,32 trend in vc and ac can be explained as follows:
Following Liebermann and Lichtenberg,1 the individual When a < ac (i.e., a is below 50%), the electron flux to
fluxes of electrons, negative ions, and positive ions at the the probe dominates over the negative ion flux by a factor
pffiffiffi
wall/probe surface can be written as follows: 170
a c. Therefore, the negative ion flux plays a weaker role
013507-3 A. K. Pandey and S. K. Karkari Phys. Plasmas 24, 013507 (2017)
FIG. 6. (Plot of bi-Maxwellian electron temperature (in eV) with power for
FIG. 7. Plot of electron to ion saturation current ratios in (a) oxygen and
4 Pa pressure using Oxygen gas).
(b) argon; Pressure ¼ 4.0 Pa.
voltages. The dc discharges are primarily sustained by ion C. Expression of floating potential
impact secondary electrons, which emerge with energies
corresponding to cathode fall potential. These electrons are The floating potential is straightforward to obtain in
eventually thermalized by making ionizing collisions with electropositive plasma by equating the electron and positive
the background neutrals and constitute the bulk electron pop- ion flux, which reduce to a simple expression as given in
ulation in the discharge. A small fraction of energetic elec- Ref. 36
trons are randomized by suffering multiple reflections from 0 sffiffiffiffiffiffiffiffiffiffiffi1
the walls and remain at higher temperatures than the back- Ies 2me pA
ground bulk population. Finally, the electrons which did not Vf ¼ Vp Te ln ¼ Vp Te ln@0:6 : (10)
Iþs mþ
suffer collisions with the target atoms are likely to contribute
as energetic beams impinging the probe/substrate. In the above equation, Ies and Iþs are the saturation currents
In the present set-up, the characteristic geometry of the corresponding to electron and positive ions. The above equa-
source is such that the secondary electrons emitted from the tion holds for moderately electronegative plasmas as well,
cathode plates are mainly constrained within the annular since the negative ion contribution can be largely ignored as
region (Fig. 4). However, there exists a finite probability that compared to the electrons at the floating potential. However,
a fraction of these electrons can reach the central region after contribution of non-thermal electrons (secondary electrons,
suffering collisions at the edge. Since the majority popula- beams) as well as hot electrons can modify the floating
tion of electrons in the bulk are at a low temperature, there- potential depending on their respective magnitudes.
fore the electron impact detachment, which is considered The probe characteristic in Fig. 5 shows a bi-Maxwellian
to be dominant loss mechanism than ion-ion recombination nature. From the analysis of I(U), the hot component of
losses, can be considered to be small/or negligible [see electrons (Teh 5 eVÞ is found to be almost one order in
Section V]. Therefore, the probability of negative ion forma- magnitude higher than the bulk electrons (Teb 0:6 eVÞ [c.f.
tion in the central region is relatively higher than the active Fig. 5]. The hot electrons in the discharge mainly contribute
region between the cathode plates. to vibrational excitation of oxygen molecules, whereas low
temperature electrons constitute the major bulk population of
B. Saturation current ratios electrons [Ref. 1-Chap. 10].
Consider the case of electronegative plasma comprising
In the I (U) characteristics in Fig. 5, the saturation T
of different negative species. The net saturation current Ies
region defines the maximum positive/negative fluxes T b h h
at Vp can be expressed as, Ies ¼ Ies þ Ies þ Is , where, Ies ,
absorbed by the probe surface. In electropositive plasma b
Ies , and Is are the saturation currents for hot electrons, bulk
like argon, the negative flux to the probe is entirely consti-
electrons, and the negative ions, respectively. To find
tuted by electrons. The ratio between the electrons to posi- T
whether the contribution of hot electrons in Ies is significant,
tive ion flux is therefore reduced to a constant factor R,
we can express
which is solely determined by their respective mass ratio
hence independent of external discharge parameters. h Vp Vf
However in the case of electro-negative plasma, this rela- Ies ¼ Iþs exp : (11)
Teh
tive ratio will change due to contribution of negative ions.
This fact has been illustrated in the plot of R versus dis- In the above equation, the contributions due to bulk electrons
charge power in Fig. 7. The value of R remains almost con- and negative ions have been neglected at the floating potential.
stant in the case of argon, whereas it tends to increase in the Therefore, (11) gives an upper estimate of hot electron contri-
case for oxygen. bution to the probe. With this approximation, the resultant
013507-6 A. K. Pandey and S. K. Karkari Phys. Plasmas 24, 013507 (2017)
h
value of Ies has been estimated in the range of 103 mA. This Total h h
32
Ies Ies Ies Teb
is approximately three orders in magnitude smaller than the R p ; A b
; B b
;
T
Iþs Ies Ies Teh
total saturation current Ies 1 mA. Therefore, we have approx- qffiffiffiffiffiffi
imated the floating potential Vf Vfb to be largely determined h
12 2lþ
Ies Teb p
by the bulk electrons C b
; a : (15)
Ies Teh 1:2
!
b
b Ies In the above equation, all the terms have their usual mean-
Vf ¼ Vp Teb ln : (12)
Iþs ings. The factor p has been introduced as a correction term in
the electron to ion saturation current ratio as discussed in
In the case of moderately electronegative plasma, it is rea- Section IV C. The cubic Equation (14) gives three roots of a,
sonable to use the electron saturation flux in (12), to obtain in which only the real and positive value has been consid-
Vfb . However, as reported by various authors, the electron ered. Since the contribution of the hot electrons is small,
h
saturation current is generally underestimated due to differ- Ies 0, therefore the above equation simplifies to
ent external factors,41 such as the return current area of the " # " pffiffiffiffiffiffiffi#
2 2 2
probe, presence of energetic electrons in the discharge, and a a T 2a Teb
R2 a3 þ 3R2 pffiffiffiffiffiffiffiffiffiffiffi a2
the external magnetic field. These factors can reduce the net l Teb l T l
electron saturation flux to the probe surface; thus it will " pffiffiffiffiffiffi#
impact the electron to positive ion saturation ratios. In order a2 Teb 2a2 T
þ 3R 2
pffiffiffiffiffiffiffiffiffiffiffiffi a þ ½R2 a2 ¼ 0: (16)
to avoid relying on the exact saturation current ratio to deter- T Teb l
mine the negative ion density, it is recommended to compare
the electron to ion saturation ratio with a known electroposi- In obtaining (14), we considered the modified Bohm speed,
tive gas, i.e., argon. This procedure has been adopted by var- which comprises bulk electrons, hot electrons, and the nega-
ious authors.41,42 In the experiment, the ratio R for argon, tive ions based on Boyd and Thompson18
which is expected to be 180 has been found to be underesti- ð
eTe 1 1 d
mated by a factor 4.7 [c.f. Fig. 7]. Hence, this factor can be nþs f ðvÞdv Te nb ðxÞ þ nhe ðxÞ þ n ð xÞ At A¼0 :
m þ 0 v2 dA e
used for rectifying the underestimated value of electron satu-
ration current in oxygen. (17)
V. DISCUSSION
The experimental results/analysis presented in Section IV
highlights the significance of negative ions on the floating
FIG. 8. plot of a and normalized value of floating potential vw ¼ Vf =Teb as potential of a probe in electronegative plasma. The important
a function of discharge power for 4, 5, 6, & 7 Pa. aspect has been centered on the effectiveness of using this
simplistic method to determine the basic parameters viz, a
discharge power. This suggests that the experimental range and c with the help of the Langmuir probe. The estimation of
lies above the critical value ac (c.f. Fig. 1). a can be influenced by several unknown factors such as effec-
tiveness of the electron to ion saturation ratio and the uncer-
tainty in determining negative ion temperature. It is also well
E. Comparison of the experiments with the theoretical
known that the floating potential is susceptible to the presence
model
of energetic electrons in the discharge.
The normalized value of the floating potential vw as a In the present experiment, the hot electrons have been
function of a has been shown in Fig. 9. The floating potential observed in the probe characteristics; however their popula-
vw obtained from the theoretical model is dependent on T tion is extremely small to cause a significant variation in the
(or c). On the other hand, the experimental value of vw has floating potential. The negative ion parameter a in the exper-
been measured directly. However to derive a, the informa- iment has been determined from the I(U) characteristics
tion about T or c is required [c.f. (14)]. Therefore in Fig. 9, using a phenomenological model. In this model, the electron
a series of theoretical curves corresponding to different val- to positive ion saturation ratio is provided as an input param-
ues of c are plotted, whereas the experimental data are shown eter to calculate the cubic equation in a [c.f. (14)]. To solve
this equation, additional information about negative ion tem- recombination. This reaction mechanism is higher than the
perature T is required. Here, T is an unknown parameter, detachment loss observed at low pressures.
which is related to the electronegative parameter c ¼ Te =T . In our experimental system, the plasma volume V ¼ 1:60
However, an optimal value of T was obtained in Fig. 9 by 104 cm3 and the loss area A ¼ 5:80 102 cm2 . The range
comparing the vw versus a plot with the theoretical curve. of a obtained in our case is found to be consistent with
The obtained negative ion temperatures and the range of a those reported by previous authors.37–40 In Reference 1 viz.,
are found to be in reasonable agreement with those reported a > K ng =K6 where K ¼ 4K5 K7 V=c v A; where ng is neu-
by other authors.40,41 Therefore, the procedure not only trals gas density, V is discharge volume, c is loss probability
allows one to estimate the negative ion temperature range, for excited species on the wall, v is average velocity of
but it is also useful as a tool to verify the analytical models, excited neutrals, and A is the loss surface. Therefore, the RHS
provided the external plasma parameters are known. of inequality in our case is approximately 0:07. This indi-
Using the above method, the variation in a as a func- cates that our operating range lies in the low pressure regime
tion of discharge power has been presented in Fig. 8. The (or high a), where the recombination loss dominates over the
observed fall in a on increasing the discharge power is detachment loss of negative ions.
consistent with the earlier results.33,37,38 The negative ions An important feature observed in Fig. 1 is the increasing
are created /lost by a series of reactions. Table I based on magnitude of vw versus a in the weakly electronegative
Ref. 40, provides the dominant reaction mechanism, regime (a < ac ). This increase has been attributed to reduc-
responsible for the production / loss of negative ions. The tion in the Bohm flux of positive ions entering the sheath
process also includes the production of excited neutrals, edge (c.f. Fig. 3). In order to balance the net electron flux
which participate in the generation and loss of negative with the positive ions, a higher repelling/negative potential is
ions. The production of oxygen negative ions is mainly necessary at the probe surface. Hence, the magnitude of
attributed to dissociative attachment of the ground state floating potential will increase on introducing a small con-
and excited neutrals in the discharge (rate constants K1 to centration of negative ions. In contrast to this, at higher elec-
K4 ), whereas these ions are lost via: (1) electron impact tronegativity (a > ac ), the floating potential vw is found to
detachment having rate constant K5 , (2) recombination of reduce monotonically as the negative flux to the probe sur-
negative with positive ions having rate constants K6 and face is mainly contributed by negative ions than electrons.
K8 , and (3) excited neutral impact detachment having rate Therefore, ion flux/ion bombarding energy at the substrate
constant K7 . can be controlled by introduction of negative ions.
The observed decrease in negative ion density in Figure
VI. SUMMARY AND CONCLUSION
8 on increasing the discharge power can be attributed to elec-
tron impact detachment as the electron density increases In this paper, a parametric study of floating potential has
with the discharge power. Another important factor responsi- been performed as a function of electro-negativity parameter
ble for the loss of negative ions is caused by ion-ion a, while the results have been compared with the experiment
TABLE I. Dominant reaction mechanism for production and loss of negative ions as well as the production of excited neutrals, which participate in the genera-
tion and loss of negative ions.
n production mechanism
1. e þ O2 ! Oð3 PÞ þ O K1 ¼ 8:8 1017 expð4:4=Te Þ 1:32 1020
2. O þ O2 ! O2 þ O K2 ¼ 2:0 1016 ð300=Tg Þ0:5 2:2 1016
3. e þ O2 ða1 Dg Þ ! Oð3 PÞ þ O K3 ¼ 4:19 1015 Te 1:376 exp ð5:19=Te Þ0:5 6:05 1017
4. e þ O2 ! Oþ þ O þ e K4 ¼ 7:10 1017 Te 0:5 exp ð1:7=Te Þ0:5 9:17 1018
n loss mechanism
5. e þ O ! e þ Oð3 PÞ þ e K5 ¼ 5:47 1014 Te 0:324 exp ð2:98=Te Þ 1:12 1016
6. þ
O2 þ O ! e þ Oð3 PÞ þ O2 K6 ¼ 2:6 1014 ð300 K=Tg Þ0:44 2:07 1014
7. O þ Oð3 PÞ ! e þ O2 K7 ¼ 1:6 1016 ð300 K=Tg Þ0:5 1:27 1016
8. Oþ þ O ! 2Oð3 PÞ K8 ¼ 4:0 1014 ð300 K=Tg Þ0:43 4:0 1014
n* production mechanism
9. e þ O2 ! e þ O2 ðv ¼ 1Þ K9 ¼ 2:80 1015 expð3:72=Te Þ 1:64 1018
10. e þ O2 ! e þ O2 ðv ¼ 2Þ K10 ¼ 1:28 1015 expð3:67=Te Þ 8:30 1019
11. e þ O2 ! e þ O2 ðv ¼ 3Þ K11 ¼ 5:00 1016 expð3:6=Te Þ 3:73 1019
12. e þ O2 ! e þ O2 ðv ¼ 4Þ K12 ¼ 2:00 1016 expð3:5=Te Þ 1:82 1019
13. e þ O2 ! e þ O2 ða1 Dg Þ K13 ¼ 1:37 1015 expð2:14=Te Þ 1:89 1017
14. e þ O2 ! Oð3 PÞ þ O K14 ¼ 8:80 1017 expð4:4=Te Þ 1:32 1020
15. e þ O2 ! Oð3 PÞ þ Oð3 PÞ K15 ¼ 6:86 1015 expð6:29=Te Þ 2:35 1020
16. e þ O2 ! Oð3 PÞ þ Oð1 DÞ K16 ¼ 1:80 1013 expð18:33=Te Þ 2:15 1029
17. e þ O2 ! Oð1 DÞ þ Oð1 DÞ K17 ¼ 1:44 1016 expð17:25=Te Þ 1:49 1031
18. e þ O2 ! 2e þ Oþ þ Oð3 PÞ K18 ¼ 1:88 1016 Te1:699 expð16:81=Te Þ 1:45 1031
013507-9 A. K. Pandey and S. K. Karkari Phys. Plasmas 24, 013507 (2017)
9
conducted in weakly magnetized oxygen plasma. The overall P. Chabert, N. Plihon, C. S. Corr, and J. L. Raimbault, Phys. Plasmas 13,
behavior of the floating potential has been characterized into 093504 (2006).
10
W. Oohara, D. Date, and R. Hatakeyama, PRL 95, 175003 (2005).
two regimes with regard to a critical value ac . In the lower 11
C. G. Pantano and T. E. Madey, Appl. Surf. Sci. 7, 115 (1981).
range when a < ac , the theoretical values show an increasing 12
T. Yakabe, D. Fujita, and K. Yoshihara, Appl. Surf. Sci. 241, 127
trend with a. This observation has been attributed to reduc- 13
(2005).
M. Matsuoka, Y. Hoshi, and M. Naoe, J. Appl. Phys. 63, 2098 (1988).
tion in the positive ion flux to the probe surface owing to a 14
J. I. Fernandez Palop, J. Ballesteros, V. Colomer, M. A. Hernandez, and
fall in the Bohm speed entering the sheath edge. When the A. Dengra, J. Appl. Phys. 77(7), 2937 (1995).
value of a is sufficiently large, the flux balance between the 15
J. I. Fernandez Palop, J. Ballesteros, M. A. Hernandez, and R. Morales
positive and negative charge species arriving at the wall sur- Crespo, J. Appl. Phys. 91(5), 2587 (2002).
16
J. I. Fernandez Palop, J. Ballesteros, M. A. Hernandez, and R. Morales
face is predominantly determined by negative ions than elec-
Crespo, Plasma Sources Sci. Technol. 16, S76–S86 (2007).
trons. In this case, the magnitude of potential tends to fall 17
K.-U. Riemann, J. Phys. D: Appl. Phys. 24, 493 (1991).
monotonically as a was increased. 18
R. L. F. Boyd and J. B. Thompson, Proc. R. Soc. London, Ser. A 252, 102
The theoretical predictions have been experimentally (1959).
19
T. E. Sheridan, P. Chabert, and R. W. Boswell, Plasma Sources Sci.
validated in the regime, a > ac over a limited range of a. A
Technol. 8, 457–462 (1999).
phenomenological model has been provided to interpret the 20
N. St. J. Braithwaite and J. E. Allen, J. Phys. D: Appl. Phys. 21, 1733
value of a and the corresponding values of floating potential. (1988).
21
The result shows excellent agreement with the theoretical R. N. Franklin and J. Snell, J. Phys. D: Appl. Phys. 31, 2532 (1998).
22
N. St. J. Braithwaite, T. E. Sheridan, and R. W. Boswell, J. Phys. D 36,
curves. Based on this model, the value of T has been found
2837 (2003).
to be in range of 0.05 eV to 0.07 eV. In principle, if the 23
M. Bacal, Rev. Sci. Instrum. 71(11), 3981 (2000).
24
correction factor is benchmarked against another diagnostics, N. Sirse, N. Oudini, A. Bendib, and A. R. Ellingboe, Plasma Sources Sci.
for example, pulse laser photo-detachment,40 then the above 25
Technol. 25, 04LT01 (2016).
N. Oudini, N. Sirse, R. Benallal, F. Taccogna, A. Aanesland, A. Bendib,
method can be highly robust for determining the negative and A. R. Ellingboe, Phys. Plasmas 22, 073509 (2015).
ion temperature accurately. This shall be considered in our 26
H. Amemiya, J. Phys. Soc. Jpn. 57(3), 887–902 (1988).
27
future study. N. Sirse, S. K. Karkari, and M. M. Turner, Plasma Sources Sci. Technol.
24, 022001 (2015).
28
ACKNOWLEDGMENTS J. I. Fernandez Palop, J. Ballesteros, V. Colomer, and M. A. Hernandez,
J. Appl. Phys. 80(8), 4282 (1996).
29
The author thanks Dr. M. Bandyopadhyay and Dr. S K M. Sharifian and B. Shokri, J. Plasma Phys. 77(part 3), 307–314 (2011).
30
H. Shindo and Y. Hosiike, Jpn. J. Appl. Phys. 30(1), 161–165 (1991).
Mishra for their valuable suggestions regarding the 31
R. N. Franklin and J. Snell, J. Plasma Phys. 64(part 2), 131–153 (2000).
manuscript and Mr. Dhrumil Patel for his technical support. 32
Y.-C. Ghim and N. Hershkowitz, Appl. Phys. Lett. 94, 151503 (2009).
33
J. T. Gudmundsson, I. G. Kouznetsov, K. K. Patel, and M. A. Lieberman,
1 J. Phys. D: Phys. 34, 1100–1109 (2001).
M. Lieberman and A. Lichtenberg, Principles of Plasma Discharges and
34
Materials Processing (Wiley, New York, 1994). W. Oohara and R. Hatakeyama, Phys. Rev. Lett. 91(20), 205005 (2003).
2 35
G. Carter and J. S. Colligan, Ion Bombardment of Solids (American M. A. Mujawar, S. K. Karkari, and M. M. Turner, Plasma Sources Sci.
Elsevier, New York, 1968). Technol. 20, 015024 (2011).
3 36
L. I. Maissel and R. Glang, "Applications of sputtering to the deposition of R. L. Merlino, Am. J. Phys. 75(12), 1078 (2007).
37
films," in Handbook of Thin Film Technology (McGraw-Hill, New York, H. Amemiya, Vacuum 58, 100–116 (2000); in Proceedings of the 11th
1970), Chap. 4. International School on Vacuum, Electron and Ion Technologies, 20–25
4
V. S. Smentkowski, Prog. Surf. Sci. 64, 1–58 (2000). September 1999, Varna, Bulgaria.
5 38
M. Bacal, “Volume production of hydrogen negative ions,” Nucl. Instrum. M. Roberto, M. J. Verbonceour, P. Verdonck, and E. Cizzoto, Proceedings
Methods Phys. Res. B, Beam Interact. Mater. At. 37/38, 28–32 (1989). of the SBMICRO 2006. ECS Trans. 4(1), 563 (2006).
6 39
L. G. Christophorou and R. J. Van Brunt, IEEE Trans. Dielectr. Electr. E. Stoffels, W. W. Stoffels, D. Vender, M. Kando, G. M. W. Kroesen, and
Insul. 2, 952 (1995). F. J. Hoog, Phys. Rev. E 51(3), 2425 (1995).
7 40
J. Mankowski, J. Dickens, IEEE Member, M. Kristiansen, and Life G. Roberson, M. Roberto, J. Verboncoeur, and P. Verdonck, Braz. J. Phys.
Fellow, IEEE Trans. Plasma Sci. 26(3), 874 (1998). 37(2A), 457 (2007).
8 41
A. Aanesland, A. Meige, and P. Chabert, “Electric propulsion using ion- H. Amemiya, Jpn. J. Appl. Phys. 30(10), 2601–2605 (1991).
42
ion plasmas,” J. Phys: Conf. Ser. 162, 012009 (2009). M. Bowes and J. W. Bradley, J. Phys. D: Appl. Phys. 47, 265202 (2014).