171501-2
Y. Ohtsu and H. Fujita
Appl. Phys. Lett. 92, 171501 ͑2008͒
FIG. 2. ͑Color online͒ Plasma density and electron-neutral mean free path as
a function of Ar gas pressure.
was also calculated. Especially, an absolute value of the
plasma density for Ar gas pressure of 5 Pa, as a reference
value, was estimated from an electron saturation current den-
sity of characteristics of current voltage and is about 8
ϫ1010 cm−3. It is seen that the plasma density remarkably
increases with the increasing gas pressure from 1 to 3 Pa and
then changes its order, while the electron-neutral mean free
path is inversely proportional to the gas pressure and ap-
proaches the hole size at the gas pressure of 15 Pa. This
result indicates that the hollow cathode effect is achieved at
the pressure range of 3–15 Pa where the mean free path is
comparable to the hole size.
FIG. 4. ͑Color online͒ ͑a͒ Characteristics of probe current density and volt-
age and ͑b͒ semilog plot of electron current density vs probe voltage at z
=5 and 15 mm for r=0.
celeration due to oscillation of electrons in the sheath field of
hole.
The plasma parameters such as density and temperature
of electrons near the rf electrode were measured in front of
hole ͑r=0 mm͒ and flat ͑r=5 mm͒ in the rf electrode. Figure
3 shows characteristics of probe current density and voltage
at r=0 and 5 mm for z=5 mm. Here, Ar gas pressure and rf
power are 5 Pa and 200 W, respectively. It was obviously
found that probe current density jp at z=0 mm is higher than
that at z=5 mm. That is, it indicates that the hollow cathode
discharge is performed in a hole. After an analysis of the
probe characteristics, it was revealed that for z=5 mm high-
energy electrons with more than 30 eV existed, whereas for
z=0 the electron energy distribution had an almost Maxwell-
ian profile. The high-energy electrons are generated by ac-
Figure 4 shows probe characteristics at z=5 and 15 mm
for r=0. It is seen that the plasma density at z=15 mm is
higher than that at z=5 mm. In fact, the plasma density in-
creased from 1.2ϫ1010 cm−3 to 8ϫ1010 cm−3 with the in-
creasing distance ͑z=5→15 mm͒ from the rf multihollow
electrode. One of the reasons for the increase in the plasma
density was the ionization between the high-energy electrons
and neutral atoms because the energy of high-energy elec-
trons detected at z=5 mm decreased from 30 to 10 eV. The
other reason is the diffusion of electrons from the next holes.
It is found that the multihollow electrode contributes to the
increase in the plasma density.
The effect of secondary-electron emission to achieve
high-density plasma production was examined by biasing the
substrate. Figure 5 shows plasma density as a function of
substrate biasing voltage Vb for z=5 mm and r=0. Here,
plasma density was also estimated from the ion saturation
current of the negatively biased probe, as shown in Fig. 2.
The dashed line corresponds to the absolute value of plasma
density for Vb=0, which was estimated to be about
1010 cm−3 from the characteristics of current voltage. The
difference of scale in Figs. 2 and 5 for the relation between
the arbitrary units and the reference line to the quantified
value is due to the difference in electron temperature at z
=5 and 15 mm. Plasma density increases with the substrate
biased voltage Vb and reaches about 1011 cm−3 at Vb=
−800 V. This indicates that increasing Vb contributes to ion-
ize neutral atoms by inelastic collisions of secondary-high-
energy electrons.
As mentioned in the above section, high-density plasma
production is performed with the effect of hollow cathode
FIG. 3. ͑Color online͒ Characteristics of probe current density and voltage
at r=0 and 5 mm for z=5 mm.
discharge and secondary-electron emission. The film prepa-
132.206.7.165 On: Mon, 24 Nov 2014 15:08:00
Ohtsu
