Journal of Non-Crystalline Solids 353 (2007) 308–312
www.elsevier.com/locate/jnoncrysol
Characterizations of a-Se based photodetectors using
X-ray photoelectron spectroscopy and Raman spectroscopy
K. Okano
a,b
, I. Saito a,*, T. Mine b, Y. Suzuki c, T. Yamada d, N. Rupesinghe a,
G.A.J. Amaratunga a, W.I. Milne a, D.R.T. Zahn c
a
d
Department of Engineering, University of Cambridge, Electrical Engineering Division, 9, JJ Thompson Avenue, Cambridge CB3 0FA, UK
b
Department of Physics, International Christian University (ICU) 3-10-2 Osawa, Mitaka, Tokyo 181-8585, Japan
c
Halbleiterphysik, Institut fuer Physik, Technische Universität Chemnitz Reichenhainer Str.70/72, D-09126 Chemnitz, Germany
Diamond Research Centre, National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Umezono, Tsukuba, Ibaraki, 305-8568, Japan
Received 13 October 2004; received in revised form 14 June 2006
Available online 17 January 2007
Abstract
The ‘PLUMBICON’ was one of the first successful imaging tubes using amorphous selenium (a-Se) and many followed. Significant
properties of a-Se based imaging tubes have been rediscovered through the invention of the ‘HARP (high-gain avalanche rushing amorphous photoconductor)’, but its operational mechanism and the physics, however are yet poorly described. Previously, we have fabricated photodetectors using nitrogen (N)-doped diamond as a cold cathode and a-Se as a photoconducting target, which successfully
responded to light illumination, The device performance,in this case, deteriorates after continuous use largely due to the degradation
of a-Se. In this paper, a-Se and amorphous arsenic selenide (a-As2Se3) films have been deposited Stoichiometry has been determined
by XPS (X-ray photoelectron spectroscopy) followed by Raman spectroscopy characterization. We have found that even an extremely
weak incident laser power causes sample degradation during signal accumulation. We speculate that either the incident laser itself and/or
the temperature rise due to illumination causes the phase transition in a-Se films. In addition, when As is added into the film, the phase
transition leading the degradation is hardly observed, implying that As affects the formation of crystalline Se making chemical bonds in
the crystallographic network stronger.
2006 Elsevier B.V. All rights reserved.
PACS: 78.30.L; 78.30; 79.60.H; 79.60
Keywords: Amorphous semiconductors; Crystallization; Devices
1. Introduction
There have been many interesting attempts to use amorphous selenium (a-Se) based targets for photodetectors,
camera tubes and other photoelectric conversion devices.
Among them, ‘PLUMBICON’, which was presented in
1963, was the first successful photoconductor-type camera
tube having blocking layers [1]. In this device, a-Se was
deposited on top of n-type semiconductors such as SnO2
or CdSe in order to form the blocking layer. Although
*
Corresponding author.
E-mail address: [email protected] (I. Saito).
0022-3093/$ - see front matter 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jnoncrysol.2006.11.007
the sensitivity of the device was fairly high, the band gap
of a-Se is slightly wider for detecting red light and therefore
it did not exhibit sufficient sensitivity. In addition, crystallization occurs under room temperature, which lowers the
resistivity of a-Se resulting in noise on the monitor screen.
Small amounts of tellurium (Te) and arsenic (As) were
introduced in the target (named ‘SATICON’) in order to
overcome these disadvantages [2]. However, since all of
these devices consists of thermal cathodes as the electron
source, it was impossible to miniaturize and thus, most of
the consumer applications have been taken over by small
and compact CCDs. Nevertheless, one of the largest
advantages in using an a-Se based photoconductor, is that
K. Okano et al. / Journal of Non-Crystalline Solids 353 (2007) 308–312
it exhibits extremely high quantum efficiency (approx. 10).
In this case, the device is known as the HARP (high-gain
avalanche rushing photoconductor), in which the target is
operated in the avalanche breakdown mode [3,4]. This
extremely high quantum efficiency, which has been never
achieved using other image (detectors including CCDs)
provides extremely high sensitivity that enables image collection even under the darkness of starlight without the
need for external light.
In our previous studies, we have fabricated photodetectors using nitrogen (N)-doped diamond as a cold cathode
and amorphous selenium (a-Se) as a photoconducting target [5–7]. When the a-Se target is illuminated with light,
electron–hole pairs (EHP) are generated within the film,
which changes the surface potential of the cathode side of
the target. As this surface potential increases the extraction
voltage also increases and builds up electron emission from
the diamond surface and as a whole, enhances the electric
current within the circuit. The result indeed demonstrated
a photo-response according to the light illumination for a
certain period. However, when the device is used for hours
continuously, the performance was found to deteriorate
largely due to the degradation of the a-Se target as mentioned in the literature. There are many optical studies of
chalcogenide glasses in order to understand crystallographic characteristics, including those associated with
phase transition [8–12], but as far as the authors know,
there are hardly any reports focusing on utilizing a-Se films
for photoelectric conversion devices.
In this report, a-Se and a-As2Se3 films have been deposited and characterized by Raman spectroscopy followed by
the determination of their stoichiometry using XPS (X-ray
photoelectron spectroscopy). We have found that even
under the illumination of a laser at an extremely weak
incident laser power, the film may be degraded after long
exposure time, i.e. numerous accumulations. In addition,
when a certain amount of As is added during the film
deposition using an As2Se3 source, the phase transition
leading the degradation is hardly observed implying that
adding As in the a-Se structure plays an important role
to strengthen the chemical bonds in the crystallographic
network.
2. Experimental
The amorphous selenium (a-Se) and the amorphous
arsenic selenide (a-As2Se3) films are prepared under vacuum evaporation. According to the literature, it seems to
the authors that amorphous selenium films must be carefully prepared under accurate control in order to exhibit
the same material parameters. It is, thus vital to provide
the preparation details as concise as possible in order to
maintain the reproducibility especially for the fabrication
of devices in the future. Micro Cover Glass (Mastunami
Ind. Ltd. No. 2) with 17.5 mm in diameter and 0.2 (plus/
minus 0.05) mm in thickness was used as a substrate. The
glass plate is cleaned first by acetone, then methanol, and
309
finally by diluted water, each in an ultrasonic bath for
5 min. The glass plates are then placed onto a sample
holder as shown in Fig. 1. A Mo boat is charged with
grains of either selenium or arsenic selenide, and is set in
between two electrodes located inside the cylinder. A crystal oscillator is set aside the sample to measure the deposition rate and the film thickness. The system is evacuated
under the order of 10 6 Torr. The evaporating source is
preheated according to the following procedure. The current is gradually increased to 50 A within approximately
4 min, and kept on 50 A for another 4 min. The current
is then dropped to 30 A for 1.5 min. After this preheating
treatment, the shutter that covered the source during preheat is now opened for evaporating onto both the glass
plate and the crystal oscillator. The evaporation rate is kept
at 3.3 nm/s and the shutter is closed when the film thickness
reaches 2 lm. The system is kept under vacuum for another
1 h for cooling down of the boat and the evaporation
source. Finally, Nitrogen gas is introduced into the system
for re-pressurising.
X-ray photoelectron spectroscopy (XPS) determined the
stoichiometries of the samples. XPS measurements were
carried out on a-Se and a-As2Se3 films deposited on glass
substrates using JEOL JPS-9010MX with a non-monochromated AlKa X-ray source. The excitation power was
500 mW, and no slit for microanalysis was used so that
the detection area is as wide as 10 mm in diameter, which
almost covers the whole area of the sample. XPS spectra
were recorded in the binding energy range from 0 to
400 eV at the base pressure of 5 · 10 10 Torr.
The a-Se and a-As2Se3 films deposited on glass substrates were, then, characterized using Dilor XY800, setup
to micro-Raman spectroscopy. It is well known that one of
the advantages of using Raman spectroscopy is that this
technique is non-destructive when the incident laser power
is carefully chosen. The Kr+ laser with a wavelength of
647.1 nm was used for the excitation light and the laser
was focused onto the sample to a spot of approximately
100 lm diameter. The laser power on the sample surface
was carefully calibrated to be 0.7 mW, which was confirmed to be almost the minimum power for detecting aSe. Raman scattering was measured in a backscattering
geometry and a parallel polarization configuration with a
spectral resolution of 2.2 cm 1.
3. Results and discussion
3.1. XPS
The XPS for the a-Se film contains peaks around
285 eV, 230 eV, 165 eV and 55 eV which can be assigned
as C1s, Se3s, Se3p1/2 and Se3d3/2, respectively as shown
in Fig. 2. The spectrum for the a-As2Se3 film shown in
Fig. 2 exhibits 210 eV, 145 eV and 48 eV peaks, which
can be assigned as As3s, As3p1/2 and As3d3/2, respectively
in addition to the peaks observed for the a-Se films. The
results clearly indicates that using As2Se3 as the doping
310
K. Okano et al. / Journal of Non-Crystalline Solids 353 (2007) 308–312
Crystal Oscilator
Glass Plate
Sample
Mounter
Shutter
Cylinder
Boat
Diffusion
Pump
Monitor
Power
Source
Rotary Pump
Shutter Axis
Fig. 1. A schematic diagram of the a-Se evaporating system.
obtained ratio (0.42) as the stoichiometry of 2:3. The peak
intensity ratio was obtained as 0.45 for the deposited film
and the stoichiometry was confirmed to be almost equivalent to the source material (2:3). The weak peak of As3s
around 210 eV for the a-Se film should be the result of contamination since all the films are deposited in the same vacuum chamber. In addition, the stoichiometry of the aAs2Se3 film was also confirmed to be 2:3 by calculating
the value using photoemission cross-sections reported in
the literature [13]. We are, thus, confident to say that films
deposited using As2Se3 as a starting material, contain
almost the same amount of As compared to the source
material while those deposited using only Se contain only
negligible amount of As.
3.2. Raman
Fig. 2. XPS spectra of a-Se (blue: a-Se, red: a-As2Se3). (For interpretation
of the references in color in this figure legend, the reader is referred to the
web version of this article.)
source gives rise to the addition of As into the deposited
films.
Furthermore, the stoichiometry was evaluated by comparing the peak intensity ratio of the As3s to the Se3s
peaks. First, we have measured the XPS spectrum of
As2Se3 that was used as source material and defined the
The Raman spectra of a-Se films are shown in Fig. 3(a).
In this figure, the lowest spectrum is a sum of the first three
accumulations. The following spectra represent the situations after adding three accumulations to the previous
ones. The highest spectrum is the result of after more than
50 accumulations, and no more significant change occurs
due to the accumulations. The bottom spectrum of
Fig. 3(a) indicates a broad peak around 250 cm 1 and a
peak relatively sharp peak around 235 cm 1. It shows that
the intensity and the shape of the peak changes as the accumulation proceeds, and the broad peak around 250 cm 1
becomes weaker while the peak around 235 cm 1 splits into
sharp peak at 233 cm 1 and a shoulder at 237 cm 1.
According to the literature, the peak observed around
K. Okano et al. / Journal of Non-Crystalline Solids 353 (2007) 308–312
311
235 cm 1 on the a-Se spectra, superimposed to the t-Se
peak, it is technically impossible to separate and therefore
is not subtracted. However, we have confirmed by using a
different system on the same spot that this is indeed a
235 cm 1 t-Se peak and therefore the plasma line has little
effect on our obtained results.
The Raman spectra observed for As2Se3 film are shown
in Fig. 3(b). This figure also shows the effect of accumulation adding three accumulations for obtaining each spectrum starting from the lowest and the highest spectrum
indicates the one after more than 50 accumulations. The
plasma line is subtracted and smoothed from the obtained
data. It clearly indicates that all the obtained spectra consist of a broad peak around 235 cm 1, which can be
assigned to a-As phase [15]. In contrast with the spectra
shown in Fig. 3(a), the results shows no significant change
as accumulation progresses, which implies that no phase
transition occurs largely because of the presence of As
atoms added during the deposition. The noise level of the
spectra, however, is much higher for the set of spectra
obtained for As2Se3 films. This is likely to be caused by a
reduced cross-section of the Raman scattering.
From these results, it can be said that photon–atomic
interaction between the incident Kr+ laser and a-Se films
turns Se/As atoms to excited states and/or produces joule
heat. On the other hand, from our previous results, we have
confirmed that heat induces degradation of a-Se whereas
a-As2Se3 films do not [16]. From the Raman spectroscopy
results, it is clear that a-Se is vulnerable to a phase transition caused by either ‘optical excitation’ or joule heating
while As atoms within the crystallographic network of
the As2Se3 films tend to strengthens the bonding and prevents such degradation.
4. Conclusion
Fig. 3. Raman spectra of (a) a-Se and (b) a-As2Se3.
250 cm 1 can be assigned to the broad amorphous phase of
selenium, while that at 235 cm 1 corresponds to the polycrystalline trigonal Se (t-Se) phase as a result of photo-crystallization. The t-Se feature becomes more pronounced and
two resolved bands appear at 237 cm 1 and 233 cm 1
[12,14]. Although a plasma line could be observed around
From these results, we have confirmed that adding of As
atoms in a-Se films plays an important role preventing the
photo-induced phase transition from amorphous to polycrystalline phase. However, the film deposited using As2Se3
as the source material, only exhibits a broad peak of a-As
and no peak corresponding to a-Se in Raman spectra. For
further development as a photodetector, the relationship
between crystallization must be clarified. From the results
of the Raman spectroscopy, incident Kr+ laser induced
phase transition of a-Se but not a-As2Se3. This fact indeed
indicates that the crystallographic network of a-As2Se3 is
stronger than that of a-Se. However, the photoconductivity, which we believe to be the most important parameter
for photoconductor application, may become much smaller
compared that of to a-Se films [16]. In order to develop
long lifetime photoconductors with ultra high sensitivity
in the future, the result might imply that the amount of
additional As should be optimized so that it prevents
the phase transition without drastically decreasing the
photoconductivity.
312
K. Okano et al. / Journal of Non-Crystalline Solids 353 (2007) 308–312
Acknowledgments
The present work was financially supported in part by
the Grant-in-Aid for Scientific Research (A) (# 13305006)
and Academic Frontier Project both from the Ministry of
Education, Science, Sports and Culture, Japan.
References
[1] J.A. Hall, Imaging tubes, in: W.L. Wolfe, G.J. Zissis (Eds.), The
Infrared Handbook, Environmental Research Institute of Michigan,
1978 (Chapter 13).
[2] T. Fujita, N. Goto, Television 28 (1974) 879 (in Japanese).
[3] N. Egami, K. Goto, M. Nanba, S. Okazaki, A camera tube with a
filled emitter arrays, Tech. Report IEICE ED94-103, 1994, p. 1 (in
Japanese).
[4] M. Nanba, T. Yamagishi, S. Okazaki, K. Tanioka, N. Egami, F.
Sato, The first study of a field emitter array’s flat image sensor with
HARP target, Tech Report IEICE ED96-43, 1996, p. 13 (in
Japanese).
[5] Y. Suzuki, H. Yamaguchi, Y. Okamura, K. Okano, IEEE Electron
Dev. Lett. 24 (2002) 16.
[6] Y. Suzuki, K. Oonuki, H. Yamaguchi, Y. Okamura, K. Okano,
Electron. Lett. 38 (2002) 1711.
[7] K. Oonuki, Y. Suzuki, H. Yamaguchi, Y. Okamura, K. Okano, J.
Vac. Sci. Tech. B 21 (2003) 1586.
[8] M. Gorman, S.A. Solin, Solid State Commun. 18 (1976) 1401.
[9] P.J. Carroll, J.S. Lannin, Solid State Commun. 40 (1981) 81.
[10] A.A. Baganich, V.I. Mikla, D.G. Semak, A.P. Sokolov, A.P.
Shebanin, Phys. Stat. Sol. 166 (1991) 297.
[11] Z.S. El Mandouh, Fizika A 2 (1993) 35.
[12] V.V. Poborchii, A.V. Kolobov, K. Tanaka, Appl. Phys. Lett. 72
(1998) 1167.
[13] J.H. Scofield, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129.
[14] H. Wilhelm, W. Richter, U. Rossow, D. Zahn, D.A. Woolf, D.I.
Westwood, R.H. Williams, Surf. Sci. 251&252 (1991) 556.
[15] D. Drews, A. Schneider, D.R.T. Zahn, D. Wolffram, D.A. Evans,
Appl. Surf. Sci. 104&105 (1996) 485.
[16] I. Saito, K. Oonuki, T. Yamada, M. Aono, T. Butler, N. Rupsinghe,
G.A.J. Amaratunga, W.I. Milne, K. Okano, Jpn. J. Appl. Phys. 44
(2005) L334.