IEEE Transactions on Dielectrics and Electrical Insulation Vol. 24, No.

6; December 2017 3425

Development of Recycling Method for CD-ROM

Using Pulsed Power
Tomohiko Yamashita, Naofumi Yoshihara
Graduate School of Science and Technology
Kumamoto University
Kurokami 2-39-1, Chuo-ku, Kumamoto, Japan

Hamid Hosseini, Takashi Sakugawa and Hidenori Akiyama

Institute of Pulsed Power Science
Kumamoto University
Kurokami 2-39-1, Chuo-ku, Kumamoto, Japan

ABSTRACT
This paper describes a new recycling method for metal-coated plastics. Pulsed power
technology was used to remove a metal layer (aluminum) from CD-ROM. The progress
of removal and the state of discharge were investigated in detail. In addition, the
shadow picture method including a digital image processing was used to estimate
removal amount. A pulsed power generator using magnetic pulse compression (MPC-
PPG) with maximum storage energy of 40 J/pulse was used. Pulsed power was applied
to concentric ring electrodes placed on the CD-ROM. Separation between electrodes
was about 35 mm. Results show that almost all of the metal layer was removed under
application of about 30 shots of pulse power. A strong correlation between the removed
area measured by the shadow picture method and the removed mass was obtained,
allowing the removed area and mass to be measured instantly by the shadow picture
method. Furthermore, it was revealed that the discharge onset voltage increases with
gap length between the electrodes and the aluminum layer left on the surface. Due to
the surface flashover voltage between ring electrodes of about 80 kV, a pulsed power
source output voltage larger than 80 kV is required.
Index Terms: Recycling, CD-ROM, pulse power, magnetic pulsed compression
circuit.

1 INTRODUCTION processing of their wide surfaces for collection of valuable

metals is difficult. Therefore, development of efficient
ELECTRIC and electronics industries have caused an recycling technology that enables the processing of composite
increase in e-waste (electronic waste), whose disposal has materials containing small amounts of valuable contents
become a serious global problem. Improper disposal of waste, remains undeveloped.
including hazardous substances, has become a major
At present, the vast majority of e-waste including metal-
environmental issue which affects human health. Also, as coated plastics is recycled by crushing and dissolving.
electronic components include valuable materials such as iron, However, this method requires a large-scale facility and high
aluminum, and rare metals, recycling of e-waste is of great cost. Here, we focus on metal and plastic separation using a
importance. relatively simple and cheap method, pulsed power discharge.
Recycling technology development is driven by the value of Pulsed power, generated by temporal compression of a
materials contained in electronic components if they are large amount of power, has been applied to various fields such
effectively recyclable. While literature reports various as environmental technology, biotechnology, recycling and
technologies related to physical separation [1, 2], the medical technology [3 - 5]. Energy-efficient pulsed power
development of recycling technology for composite materials generators are viable for small-scale plants. In particularly,
such as metal-coated plastics has lagged behind for two Magnetic Pulse Compression-Pulsed Power Generator (MPC-
reasons: these contain few valuable materials, and the PPG) has advantages of compactness and easily adjustable
repetition frequency.
Manuscript received on 18 May 2017, in final form 31 August 2017,
accepted 1 October 2017. Corresponding author: T. Yamashita. We have previously introduced pulsed power technology to

DOI: 10.1109/TDEI.2017.006820
3426 T. Yamashita et al.: Development of Recycling Method for CD-ROM Using Pulsed Power

the recycling of metal-coated plastics and have reported on

separation processes of plastic and metal from CD-R [6]. Our
present study utilizes a CD-ROM, which has a slightly
different structure from the CD-R used before. Progress of
metal removal and discharge states were observed. Especially,
the discharge onset characteristics were investigated in detail
using a model electrode. In addition, a method to measure
separation rates instantaneously was investigated. The major
results using the CD-R in our previous work are referenced
for comparison with the results using the CD-ROM.

2 EXPERIMENTAL SETUP AND

PROCEDURE

2.1 PULSED POWER GENERATOR AND

EXPERIMENTAL SETUP
Figure 2. Schematic diagram of experimental setup for removal of
metal layer from CD-ROM.
Figure 1 shows the main circuit of our MPC-PPG pulsed
power generator, whose maximum storage energy is 40
J/pulse. The MPC-PPG is basically a combination of the C-L- 2.2 STRUCTURE OF COMPACT DISC AND
C parallel circuit and boosting transformer. After the primary ELECTRODE ARRANGEMENT
capacitor C0 (8.9 μF) is charged, the Gate Turn-Off (GTO)
Figure 3 shows the structures of both a CD-ROM and CD-
thyristor turns on. The capacitor C1 (200 nF) is then charged
R. The thickness and outer diameter of compact discs are
through the Pulse Transformer PT1. The Saturable Reactors
commonly 1.2 mm and 120 mm, respectively. CD-ROMs are
SR1 and SR2 are used as magnetic switches, and the SR1 is
composed of plastic substrate, metal layer, and a protective
also used to protect the GTO thyristor. The energy stored in
layer including the printed layer (lacquer layer and label
the capacitor C1 is transferred to the load through the
layer). Data are recorded by making pits in the metal layer
transformer PT2. This allows an output voltage of about 90-
(aluminum) on the plastic substrate [8]. CD-Rs are composed
120 kV to be applied to the load.
of plastic substrate, organic dye layer, reflective layer (silver),
and protective layer that includes the printed layer (lacquer
layer and label layer) [8]. In the case of CD-Rs, data are
recorded by burning organic dye (phthalocyanine) using a
laser, and the data are replayed by reading reflectance changes
[9]. As the material of the metal layer, gold, silver, or an alloy
is used for CD-Rs, and aluminum is used for CD-ROMs. The
plastic substrate is commonly made of polycarbonate, and the
protective layer including the printed layer is made of UV
curing resin, etc.
Figure 1. Main circuit of Magnetic Pulse Compression – Pulsed
Power Generator [7].

Figure 2 shows a schematic diagram of the experimental

setup for the removal of the metal layer from a CD-ROM. The
CD-ROM was set in an acrylic box in order to collect
fragments separated by pulsed power application. Pulsed
(a) CD-ROM
power was applied to concentric ring electrodes set to the CD-
ROM. The applied voltage between the electrodes was
measured with a high voltage probe (EP-150K, Nissin Pulse
Electronics Co., Ltd.), and the current through the electrodes
was measured with a current monitor (Model-110A, Pearson
Electronics) using an oscilloscope (DPO 5104B, Tektronix
Inc.). Progress of metal layer removal and discharge state
were observed with a camera (Model D40, Nikon, F-number:
(b) CD-R
f/36, exposure time: 1 s). In addition, emission spectrum from
the discharge was observed by a spectroscope (PMA-12 Figure 3. Structures of CD-ROM and CD-R [9].
C10029-01, Hamamatsu Photonics K.K.).
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 24, No. 6; December 2017 3427

Figure 4 shows a schematic configuration of the discharge

electrodes. Concentric ring electrodes were attached to the
label and lacquer side of a CD-ROM and set along the inner
and outer edges. The inner and outer electrodes were high
voltage and ground electrodes, respectively. These electrodes
were made of copper wire with 1.6 mm in diameter. The gap
between the two electrodes was about 35 mm.

(a) Before. (b) After first shot.

(c) After second shot. (d) After third shot.

Figure 4. Schematic configuration of concentric ring electrodes.

3 EXPERIMENTAL RESULTS AND

DISCUSSION
(e) After 4th shot. (f) After 5th shot.

3.1 PROGRESS OF THE REMOVAL OF METAL

Figure 5 shows the progress of the removal of the metal

layer with increased shot number under application of
pulsed power at 35.3 J/pulse. The white broken line shows
the ring electrodes. Figures 5a and 5b show the state of the
CD-ROM before treatment and after the first shot,
respectively. The metal layer near the electrodes was
(g) After 6th shot. (h) After 7th shot.
removed from the substrate after the first shot.
Figure 5c shows the state of the CD-ROM after the
second shot. The metal layer was removed around the parts
where the metal layer was removed by the first shot.
Further shots gradually removed the metal layer around
both electrodes. The removal of the metal layer progressed
with increased shot number along the electrodes, as shown
in Figures 5d-5f. By 5 shots, the aluminum layer around the
high voltage electrode was completely removed from the
polycarbonate substrate, as shown in Figure 5f. (i) After 17th shot. (j) After 30th shot.
After 5 shots, the separation extended toward the ground
Figure 5. Progress of the removal of metal layer.
electrode, as shown in Figures 5g-5j. Finally, the aluminum
layer was almost completely removed from the CD-ROM
by 30 shots. after treatment [6]. This did not affect recyclable amounts
Cases of both the CD-ROM and CD-R showed but made visualization more difficult.
similarities in removal progress. However, metal removed
areas of the CD-ROM resulted in clearer images compared 3.2 SHADOW PICTURE METHOD TO ESTIMATE
with the CD-R because the former lacks an organic dye AMOUNT OF REMOVAL
layer. In the case of the CD-R, the organic dye was burnt In order to evaluate the progress of the removal of the metal
and left on the substrate, and a brown surface was observed layer, two quantitative measurement methods were used: a
3428 T. Yamashita et al.: Development of Recycling Method for CD-ROM Using Pulsed Power

mass measuring method and an area measuring method. In the 3.3 STATE OF DISCHARGE AND WAVEFORMS OF
mass measuring method, removal amount was calculated by VOLTAGE AND CURRENT
weight change of the CD-ROM or by the removed mass. In
Figure 8 shows variations in discharge state and voltage and
the shadow picture method, the removed area was measured
optically using the software "Image J". The threshold of

(a) First shot.

Figure 6. Progress of the separation measured by a subtraction method

and a shadow picture method.

brightness was set carefully. Both methods were applied to (b) Second shot.
evaluate the removal of the metal layer from a CD-ROM, and
the correlation between the two methods was examined.

Figure 6 shows the progress of metal removal measured

by each method. No significant difference was found
between the two methods. The discontinuity of removed
mass is caused by measuring error because the resolution of
the weight measurement is 0.01 g. The removal of the metal
layer was in proportion to the number of shots until about (c) Third shot.
the 12th shot, after which removal tended to saturate. A
removal rate of almost 90% was achieved at about the 17th
shot.

(d) 7th shot.

Figure 7. Relation between removed mass and removed area. (e) 17th shot.

Figure 7 shows the relation between removed area and

removed mass as measured by the two methods mentioned
above. A strong correlation was obtained in the range up to
90% of removal. This shows that removal progress can be
estimated by the shadow picture method developed in this
study, a far easier method than the removed mass method. As a
computer program can instantaneously measure removal using (f) 30th shot.
the shadow picture method, it is practical for removal
evaluation use in industries. Figure 8. State of discharges and voltage and current waveforms.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 24, No. 6; December 2017 3429

current waveforms. Figure 8a shows those at the first shot. the voltage peak causing surface discharges on the
White light emission was observed near the high voltage and polycarbonate increased with the number of shots, as shown in
ground electrodes. The spectrum of the light emission at the Figures 8d-8f.
first shot was observed with a spectroscope, as shown in As shown in Figure 8f, surface flashover took place on the
Figure 9. Light emission related to the metal layer (spectrum polycarbonate substrate between the electrodes. At this point,
of aluminum) was observed. It indicates that a discharge took metal layer had been almost completely removed from CD-
place between the electrodes and the aluminum layer, and that ROM. The value of the voltage peak reached about 80 kV.
the protective layer and the printed layer (lacquer layer and
label layer) were broken down. In addition, strong spectrum 3.4 DISCHARGE ONSET VOLTAGE AND
emission was also observed around the 650 nm. It is identified MECHANISM
as Hα (656.3 nm), which comes from the insulation layer by Figure 10 shows the relation between the value of the first
vaporization and/or burning. voltage peak and the shot number. The values were measured
three times under identical experimental conditions. The value
at the first shot was about 10 kV. In the region from the 2nd to
the 6th shots, the value was almost the same. After the 7th
shot, the value began to increase with the shot number.
Finally, the value reached about 80 kV at the 30th shot. Value
variations will be discussed in three stages below, from the 1st
shot; the 2nd to 6th; and from the 7th to 30th.

Figure 9. Emission spectrum at the first shot.

In addition, another discharge bridging between electrodes

was also observed, as shown in Figure 8a. Two types of
surface discharges with red-purple light emission were
observed. One was a surface discharge with relatively strong
light emission, which was a surface flashover between the
electrodes. The other one was a streamer-like surface
discharge with a weak light emission. Voltage and current at Figure 10. Relation between discharge onset voltage and the number of
the first shot arose simultaneously, as shown in Figure 8a. shots.
Voltage peak, indicated by an arrow, is about 10 kV, which
A. First Shot
corresponds to the breakdown voltage of the insulation layer.
Detailed discussion is presented in Section 3.4. When a voltage was applied between the electrodes on the
untreated CD-ROM, the first voltage peak was about 10 kV.
Figure 8b shows the state of the discharge and the
Since the insulation layer at both sides of the electrodes
waveforms at the second shot. White light emission was
should be taken into consideration, 5 kV was applied to the
observed near the electrodes close to places of light emission at
the first shot. Expansion of area where metal was removed was
observed, as shown in Figure 8c. The aluminum layer around
the high voltage electrode was completely removed within 7
shots.
As shown in Figure 8d, the aluminum layer around the high
voltage electrode was completely removed, at which point the
residual metal layer became isolated from the high voltage
electrode. Surface discharges on the polycarbonate between
electrodes and the tip of metal layer were caused by the peak
voltage of about 20 kV.
After the 7th shot, the removal of the metal layer gradually
developed with the repeated application of pulse power.
Consequently, the gap length between the high voltage
electrode and the tip of metal layer also gradually increased
with the progress of the removal of the metal layer. As a result, Figure 11. Dependences of voltage and resistance on time at the first shot.
3430 T. Yamashita et al.: Development of Recycling Method for CD-ROM Using Pulsed Power

insulation layer (protective layer and printed layer) between

the electrode and the metal layer. The withstand field strength H.V.
of insulation layer (mainly UV curing resin) is in the range of electrode
12-15 kV/mm [10]. The thickness of the insulation layer CD-ROM
including the metal layer was about 30 μm from the
measurement. Since the thickness of the metal layer is several Gap GND
tens nm, the thickness of the insulation layer is estimated to be electrode
about 30 μm. Then, the static breakdown voltage is estimated
to be about 0.36-0.45 kV. Since the insulation layer exists at Copper
both electrodes, the static breakdown voltage of the insulation tape
layer is estimated to be about 0.72-0.90 kV. The value of the
first voltage peak (5 kV) was higher than the static breakdown
voltage of the insulation layer. The value is reasonable if
taking spark time lag into consideration. Also, light emission Figure 12. Model electrode system to measure discharge onset voltage
related to the metal (aluminum) was observed, as stated above. characteristics.
Therefore, it is considered that the value of the first voltage
corresponds to the breakdown voltage of the insulation layer. separation is caused by the vaporization of the metal layer
Another cause for the first voltage valuation can be derived and/or by the shock wave. They are accompanied with the
from the voltage and current waveforms. Figure 11 shows the discharge. The surface discharge is important to bridge the
temporal variations of the voltage and the resistance between gap between the edge of the exposed metal layer and the
the electrodes. Resistance was calculated by dividing voltage electrode. The gap length between the high voltage electrode
by current. The rapid decrease in resistance corresponds to the and the exposed metal layer increased with the progress of the
first voltage peak. Therefore, the value of the first voltage removal of the metal layer. Since the discharge initiates at the
peak corresponds to the breakdown voltage of the insulation gap, the discharge onset voltage increases with gap length.
layer. The variation in resistance before the first voltage peak Therefore, the first peak voltage increased with the increase of
is caused by a leakage current passing through the saturable shot number. In order to confirm this hypothesis, an additional
reactor SR2. Although the inductance of SR2 is quite large experiment was conducted.
before switching, a small current flows. In addition, since the Figure 12 shows a schematic diagram of a model electrode
electrodes form a capacitor, the capacitor is charged by the system. In Figure 13, circles indicate dependence of discharge
leakage current, and the voltage between the electrodes onset voltage on the gap length. The gap length between the
increases. After the rapid decrease, resistance variation grows tip of aluminum and the electrodes on the CD-ROM is from
moderate. Residual resistance is supposed to be the resistance Figure 5, and the first voltage peak is from Figure 10.
of the metal layer through which the discharge current flows. Dependence of the first peak voltage on gap length for the
As mentioned in Section 3.3, two types of surface treated CD-ROM is shown by squares. Figure 13 shows that
discharges were observed at the first shot. In this case, the initial peak voltage corresponds to discharge onset voltage at
metal layer of the CD-ROM can be regarded as the back side the shortest gap between electrodes and the tip of aluminum of
electrode. However, the potential of the back side electrode is the treating on CD-ROM.
floating. Generally, the surface flashover voltage with back
side electrode is considerably low compared with the case
without back side electrode.

B.2nd to 6th shot

The value of the first voltage peak in the region from the
2nd to the 6th shot was stable. The value likely corresponds to
the surface discharges on the polycarbonate between the
electrodes and the exposed metal layer. The gap length
between the high voltage electrode and the exposed metal
layer did not vary, resulting in a stable value.

C. 7th to 30th shot

Figure 13. Relation between discharge onset voltage and gap length.
The value of the first voltage peak increased with the shot
number after the 7th shot. The metal layer was gradually
removed by the repeated application of pulse power. Surface
4 CONCLUSIONS
discharges on the polycarbonate between electrode and the tip
of metal layer were caused after the metal layer around the This study authors investigated separation of metal from a
high voltage electrode was completely removed. The metal CD-ROM using pulsed power in order to develop a new
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 24, No. 6; December 2017 3431

recycling technology for metal-coated plastics. Obtained [9] D. Wohrle, G. Schnurpfeil, S. G. Makarov, A. Kazarin, and O. N.
Suvorova, “Practical applications of phtalocyanines - from dyes and
results are as follows.
pigments to materials for optical, electronic and photo- electronic
devices”, Macroheterocycles, Vol. 5, No. 3, pp. 191-202, 2012.
1) The metal layer (aluminum) and insulation layers [10] H. Nishiyama, T. Fujisawa, K. Kishi, T. Hibino, Y. Arai, and K. Kojima,
(protective and printed layers) were progressively separated “Three Bond Technical News - Light-Curing Resins”, Three Bond CO.,
LTD., Vol. 44, pp. 1-10, 1995.
from the CD-ROM polycarbonate substrate with increased
pulsed power shot number. Almost all of the aluminum layer
was separated from the polycarbonate substrate by 30 shots
of pulsed power with a storage energy of 35.3 J/pulse. Tomohiko Yamashita (S’16) was born in Nagasaki,
Japan, in 1990. He received the B.SC. degree from
Kyushu Sangyo University, Fukuoka, Japan, in 2015,
2) The progress of the separation was as follows. At the first and the M.E. degree from Kumamoto University,
shot, breakdown of the insulation layer took place near the Kumamoto, Japan, in 2016, where he is currently
two electrodes. In the region from the 2nd to the 6th shots, working toward the Ph.D. degree in the Graduate
School of Science and Technology. His current
separation progressed around the electrodes. After that, research interests are application of pulsed power.
discharge took place via isolated metal layer, and the
separation extended toward ground electrode. Finally,
almost all of the aluminum layer was separated from the
Naofumi Yoshihara was born in Saga, Japan, on
CD-ROM at the 30th shot. May 1993. He received the B.E. degree from
Kumamoto University, Kumamoto, Japan, in 2016,
3) The breakdown voltage of the insulation layer was about where he is currently working toward the M.E. degree
in the Graduate School of Science and Technology.
10 kV. Puncture of the insulation layer was confirmed by
the measurement of the emission spectrum of aluminum.
The discharge onset voltage increased with the increase of
the gap length between the electrodes and the aluminum
layer. The surface flashover voltage between the two ring
electrodes was about 80 kV. The output voltage of the Hamid Hosseini (M’09) received the M.Sc. (Eng.)
pulsed power generator larger than 80 kV is required to degree from Sharif University of Technology, in
remove the entire aluminum layer. 1990 and the Ph.D. degree from Tohoku University,
Sendai, Japan, in 1999. He was a Lecturer at Sharif
University of Technology, from 1990 to 1996. He
4) Separation progress was measured quantitatively by the served as Research Associate and Associate
developed shadow picture method. This method allows Professor at Tohoku University, from 1999 to 2006.
From 2006 to 2007, he was a Visiting Scientist at the
instantaneous measurement and shows that 90% separation Department of Bioengineering, University of
was achieved by pulsed power of about 17 shots. Washington, Seattle, USA. In 2008, he joined
Kumamoto University, Japan, where he is currently a Professor of the
Bioelectrics Department, Institute of Pulsed Power Science. His research
REFERENCES interests are biomedical application of pulsed power, shock waves,
[1] T. Oki, “Physical separation technology to support the strategic therapeutic ultrasound, and bioelectrics.
development of urban mining”, Synthesiology, Vol. 6, No. 4, pp. 238-
245, 2013.
[2] J. Oki, K, Torihara, Y. Kadowaki, and N. Mishima, “A study on Takashi Sakugawa (M’04) received the M.E.
separating characteristics of metals towards remote recycling”, CIRP, degree from Kyushu University, Fukuoka, Japan, in
13th Global Conf. Sustainable Manufacturing - Decoupling Growth from 1989, and the Ph.D. degree from Kumamoto
Resource Use, Ho Chi Minh, Vietnam, Vol. 40, pp. 274-279, 2015. University, Kumamoto, Japan, in 2004. He was with
[3] S. Gnapowski, H. Akiyama, T. Sakugawa, and M. Akiyama, “Effects on the Central Research Laboratory, Meidensha
pulsed power discharge in water on algae treatment”, IEEJ Trans. Corporation, Tokyo, Japan, from 1989 to 2004. From
Fundam. Mater., Vol. 133, No. 4, pp. 198-204, 2013. 2004 to 2008, he was an Associate Professor with
[4] S. Katsuki, N. Nomura, H. Koga, H. Akiyama, I. Uchida, and S. Abe, the Corporative Research Center, Kumamoto
“Biological effects on narrow band pulsed electric Fields”, IEEE Trans. University, where he has been a Professor with the
Dielectr. Electr. Insul., Vol. 14, No. 3, pp. 663-668, 2007. Institute of Pulsed Power Science, Kumamoto
[5] T. Nagashima, H. Akiyama, T. Namihira, and S. Katsuki, “Recycle of University
metal-plating plastics by pulse arc discharge”, J. Plasma Fusion Res.,
Vol. 8, pp. 1471-1476, 2009.
[6] T. Yamashita, H. Akiyama, S. H. R. Hosseini, and T. Sakugawa, Hidenori Akiyama (M’87-SM’99-F’00) received
“Separation process plastic and metal from CD-R using pulsed power”, the Ph.D. degree from Nagoya University, Nagoya,
IEEE Conf. Int’l. Power Modulator High Voltage Conf., San Francisco, Japan, in 1979. He was a Research Associate with
CA, USA, 2016. Nagoya University from 1979 to 1985. In 1985, he
[7] M. Akiyama, Z. Yang, S Gnapowski, S. H. R. Hosseini, and H. Akiyama, joined as a Faculty Member with Kumamoto
“Observation of underwater streamer discharges produced by pulsed University, Kumamoto, Japan, where he is currently
power using high-speed camera”, IEEE Trans. Plasma Sci., Vol. 42, No. an Emeritus Professor. Also, he is a President of
10, pp. 3215 – 3220, 2014. Fusion Techs Company. Dr. Akiyama received the
[8] Byers. Fred R, “Care and handling of CDs and DVDs – A guide for IEEE Major Education Innovation Award in 2000,
librarians and archivists”, NIST Special Publication, Vol. 500, No. 252, the IEEE Peter Haas Award in 2003, the Germeshausen Award in 2008, and
pp. 1- 40, 2003. the IEEE Leon Kirchmayara Graduate Teaching Award in 2013.