Tang, and Paul K. Chu
b)
Department of Physics & Materials Science, City University of Hong Kong, Kowloon, Hong Kong
Ping K. Ko
Hong Kong University of Science & Technology, Clear Water Bay, Kowloon, Hong Kong
Yiu-Chung Cheng
Hong Kong University, Hong Kong
Received 30 September 1998; accepted for publication 25 November 1998
Plasma immersion ion implantation is a burgeoning surface modication technique and not limited
by the line-of-sight restriction plaguing conventional beam-line ion implantation. It is therefore an
excellent technique to treat interior surfaces as well as components of a complex shape. To enhance
the implant uniformity and increase the thickness of the modied layer, we are using a high
frequency, low-voltage process to achieve high temperature and dose rate to increase the thickness
of the modied layer. The low voltage conditions also lead to a thinner sheath more favorable to
conformal implantation. In this article, we will describe our special modulator consisting of a single
ended forward converter with a step-up transformer. The modulator is designed to operate from 5
to 35 kHz and the output voltage is adjustable to an upper ceiling of 5000 V that is deliberately
chosen to be our voltage limit for the present experiments. We will also present experimental data
on SS304 stainless steel materials elucidating the advantages of our modulator and high frequency,
low-voltage experimental protocols. © 1999 American Institute of Physics.
S0034-6748 99 00803-5
I. INTRODUCTION
Plasma immersion ion implantation PIII is a edgling
technique to modify the surface of materials and industrial
components.
1,2
It has a number of advantages over conven-
tional ion beam ion implantation IBII such as high through-
put and no line-of-sight restriction. The surface properties of
many low-alloyed and microalloyed steels, stainless steel,
tool steel, etc. have been successfully improved using PIII.
3
One of the biggest advantages of PIII over IBII is the
ability to implant objects possessing an irregular shape with-
out beam rastering or target manipulation. However, the dose
uniformity may be less than desirable especially when treat-
ing interior surfaces.
46
Theoretically, conformal implanta-
tion can be achieved when the ion sheath is completely con-
formal around the target, but in practice, it is seldom the case
due to the irregular shape of most real objects. Besides, the
corners and edges pose a special challenge since the local
sheath is especially nonuniform.
7
It has been shown that the
ion impact angle near a corner is oblique and the ion ux,
which is enhanced above its planar value, peaks close to but
not at the corner.
8
Consequently, dose nonuniformity and
sputtering that affects the retained dose are quite severe, and
samples with sharp edges will exhibit considerable lateral
dose variation. For samples with a wedge shape, the lateral
variation in the implanted argon concentration can be as high
as a factor of 6 depending on the process parameters.
9,10
Recent theoretical and experimental data
11
have demon-
strated that the thinner sheath achievable at a lower implan-
tation voltage improves the lateral uniformity, but the thick-
ness of the treated layer is compromised.
The implantation depth by PIII is usually smaller than
that by IBII using beam-line ion implantation. Hence, it is
relatively difcult to achieve a modied layer that is thick
enough for real engineering applications. The reason is that
the implantation voltage in PIII is impractical above 100 kV
due to expensive instrumentation and arcing.
12
As an alter-
native, PIII can be carried out at an elevated temperature to
increase the diffusion depth of the implanted species. Previ-
ous work has demonstrated that elevated temperature PIII
ET-PIII works well for ferrous materials. The enhanced
nitrogen diffusion increases the thickness of the modied
materials and results in surface hardness and wear resistance
improvement superior to those achieved by conventional
IBII.
1315
Interestingly, recent reports
16
have shown that the im-
plantation voltage in ET-PIII may not be very crucial be-
cause different implantation voltage does not lead to a big
difference in the surface hardness or wear resistance of the
treated materials. Therefore, there are technical reasons fa-
voring the use of a low implantation voltage, although high-
energy bombardment does possess certain advantages, for
instance, removing oxide layers which can prevent nitrogen
a
Present address: Advanced Welding Production & Technology National
Key Laboratory, Harbin Institute of Technology, Harbin 150001, China.
b
Author to whom correspondence should be addressed; electronic mail:
paul.chu@cityu.edu.hk
REVIEW OF SCIENTIFIC INSTRUMENTS
VOLUME 70, NUMBER 3
MARCH 1999
1824
0034-6748/99/70(3)/1824/5/$15.00
© 1999 American Institute of Physics uptake or implantation through a surface layer. In order for
low energy PIII to work, the implantation current must also
be high. The high ion ux elevates the sample temperature
and promotes fast diffusion of the implanted species to attain
a thicker modied layer.
17,18
Hence, it can be envisioned that
a high frequency, low-voltage plasma immersion ion implan-
tation process will be suitable for the implantation of interior
surfaces or samples of an irregular shape. The high fre-
quency raises the dose rate and consequently the sample tem-
perature. Coupled with a low implantation voltage that re-
duces the sheath thickness, the lateral implant uniformity in
samples of a complex shape can be improved. It is apparent
that high frequency and low-energy are interdependent, and
both parameters must be optimized to yield satisfactory re-
sults. In this work, we concentrate on high frequency pulsing
at an implantation voltage less than 5 kV. Our newly devel-
oped power supply and its characteristics are also presented
in this article.
II. HIGH VOLTAGE PULSING
There are two primary methods to generate high-voltage
pulses. The rst means is to employ an onoff switch to
control the voltage. The second one is to use a switch in
conjunction with a pulse-forming network PFN to generate
the pulses. Both methods can be combined with a step-up
transformer to attain the nal required voltage. Generally
speaking, switching pulse generators are more suitable and
common in PIII.
19,20
A survey on the high-voltage pulsing
technology, including different pulse generator principles
and requirements, can be found elsewhere.
21
For lower volt-
age and peak power applications, solid state modulators are
usually better. The switching device can be made of compo-
nents from any high power device family. A comparison of
their operation characteristics indicates that the IGBT is
more desirable for the low-voltage, high frequency PIII pro-
cess.
III. PULSE GENERATOR DESIGN
A. Main power circuit
To satisfy the requirements of high frequency, low-
energy PIII, we use a single-ended forward converter as our
modulator due to its lower power. A forward converter re-
quires the polarity of the primary coil of the transformer to
be identical to that of the secondary one, and so the energy of
the load is obtained during the on state of the switching
device. For the yback converter, on the contrary, the power
supply provides the energy with the load during the off state
owing to the difference in the polarity of these two coils. The
maximum duty cycle of 50% for a single-ended forward con-
verter is not a shortcoming, as high frequency implantation is
usually coupled with small pulse widths. It is more desirable
to use a short implantation pulse to allow ample time for
plasma recuperation in between pulses. However, an auxil-
iary transformer core reset supply is required to attain the
optimal pulse transformer performance because of demagne-
tizing during the off cycle of the switching device. Thus, we
utilize a variation of the single-ended forward converter by
using two switching devices and two diodes, as shown in
Fig. 1. The advantage of this arrangement stems from the
fact that the IGBTs are subject to only half the maximum
voltage in comparison with the circuit using only a switching
device, i.e., only V
0
instead of two V
0
s. Moreover, the
power transformer T including T1, T2, T3, T4 requires no
reset winding and is thus simpler. The circuit operates in the
following mode. When the drive signals are imposed on the
two IGBTs 1MBH60D-100
22
at the same time, they are
turned on simultaneously. The primary current ows from
V0 through IGBT1, the primary winding of T, and IGBT2.
In this period, a back voltage is induced in the primary wind-
ing, and the forward converter transfers the signal to the
secondary side. The diode group D3 conducts, and the cur-
rent is delivered to the sample to carry out ion implantation.
When the drive signal is negative (
5 V), both switching
IGBTs are turned off. Since an abrupt interruption of the
primary current of T produces a high induced voltage, a path
is provided for the decay of this current and demagnetization
of T through diodes D1 and D2. To demagnetize completely,
FIG. 1. Schematic of the high frequency, low-voltage modulator.
FIG. 2. Drive circuit of IGBTs with IC EXB840.
TABLE I. Properties of EXB840 drive device for IGBT.
Specication
Rated value
Power supply voltage V
25
Input current of the inner optocoupler mA
10
Separation voltage kV
2.5
Working surface temperature (°C)
10
85
Storage temperature (°C)
25
125
1825
Rev. Sci. Instrum., Vol. 70, No. 3, March 1999
Tian
et al. T
on
must be smaller than T
off
, otherwise the IGBTs w