Applied Radiation and Isotopes 68 (2010) 2214–2217

Contents lists available at ScienceDirect

Applied Radiation and Isotopes

journal homepage: [Link]/locate/apradiso

The effect of temperature changes on electrical performance

of the betavoltaic cell
Guanquan Wang n, Rui Hu, Hongyuan Wei, Huaming Zhang, Yuqing Yang,
Xiaoling Xiong, Guoping Liu, Shunzhong Luo
Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics, Mianyang, Sichuan 621900, China

a r t i c l e in f o a b s t r a c t

Article history: There is a significant relationship between temperature and electrical performance of a betavoltaic cell.
Received 5 November 2009 Two silicon diodes used as energy conversion devices of betavoltaic cells were irradiated by Ni-63, and
Received in revised form the relationships between the temperature and the electrical performance such as Voc, Isc, and Pmax were
23 May 2010
examined. Isc increased very little as temperature increased but Voc decreased considerably. The changing
Accepted 7 June 2010
values of Voc were  3.1 and  3.0 mV/K, respectively, in the temperature range 233.15–333.15 K. As a
result of this, Pmax and Z also decreased markedly.
Keywords: & 2010 Elsevier Ltd. All rights reserved.
Temperature effect
Electrical performance
Betavoltaic cell

1. Introduction affects the characteristic parameters of the semiconductor; for

instance diffusion length (L), intrinsic carrier concentration (ni)
Isotope batteries could be a reliable power supply for military and mobility (m) are all functions of T (Chen and Wang, 1999). L
equipment and scientific research of space, the deep ocean and mostly affects the short circuit current (Isc) of the betavoltaic cell
other special environment, due to their excellent characteristics and ni mostly affects the leakage current (I0) of the semiconductor
such as high-energy density, long life, management-free operation diode, further affecting the open circuit voltage (Voc) of the
and anti-jamming properties. There are more than 10 kinds of cell. Based on these issues, we are interested in examining
energy conversion techniques used in isotope batteries, including the effect of temperature changes on electrical performance of
direct charge collection cells (Ronen et al., 2004), betavoltaic cells the betavoltaic cell.
(Guo and Lal, 2003; Kosteski et al., 1998; Lee et al., 2009; Clarkson
et al., 2007), radioisotope thermoelectric generators (Guenther
et al., 2001), reciprocating electromechanical cells (Duggirala et al., 2. Experiment
2008) and radiophotovoltaic cells (Sychov et al., 2008). The study
of on-board micro-electromechanical system (MEMS) power The material of the diode PN junction is silicon. The main
sources has aroused long-term and worldwide interest since the structure parameters of the two kinds of PN junctions are shown
1980s, due to trends in micromation and machine-electron in Table 1. The nickel-63 source of 2.96  108 Bq is a round, thin,
incorporation in the MEMS fields. The betavoltaic cell, mentioned foil of steel plated with Ni-63.
above, transmits electricity irradiated by the radioisotope because The Ni-63 source was placed coaxially on the device, and all of
of the particularity of the semiconductor diode. The betavoltaic these were placed in temperature controlled surroundings. The Isc,
cell can be very tiny and light, due to its high-energy density, and Voc and illuminated I–V characteristics of the devices were
it is easy to incorporate within MEMS because its main component monitored as a function of temperature. All measurements were
is a semiconductor. made in air.
The semiconductor diode as the energy conversion device is the
key part to the electrical performance of the betavoltaic cell, so
3. Results and discussion
electrical performance changes with quality of the semiconductor
material of the diode. Temperature (T) is an important factor that
The two devices were irradiated by the Ni-63 source and the
electrical performance was measured in the temperature range
n
Corresponding author. Tel.: +86 8162496852. 233.15–333.15 K. The I–V characteristics of the two devices are
E-mail address: [Link]@[Link] (G. Wang). shown in Figs. 1 and 2. The two I–V characteristic curves have

0969-8043/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2010.06.011
 G. Wang et al. / Applied Radiation and Isotopes 68 (2010) 2214–2217 2215

Table 1
Structure parameters of the PN junctions

Top (P type) Basic (N type)

Na (cm  3) Thickness (mm) Nd (cm  3) Thickness (mm)

Device 0 1.0  1019 1.20 1.0  1012 180

Device 1 1.0  1019 0.50 1.0  1017 150

Fig. 3. Experimental relationships between Isc and T.

Fig. 1. I–V characteristics of device 0.

Fig. 4. Experimental and predicted relationships between Voc and T.

Fig. 2. I–V characteristics of device 1.

similar shapes and gradual changing trends. With increaseing

temperature, the slopes of the curves increase and the stable
current range is diminished gradually. The decrease of fill factor
(FF) is the embodiment of this effect. The experimental and
predicted relationships between T and Isc, Voc, FF, the maximum
output power (Pmax) and energy conversion efficiency (Z) are
shown in Figs. 3–7. In general the trends in these relationships are
comparable for the two devices. The electrical performance of the
two devices at 233.15 and 333.15 K are given in Table 2.

3.1. Isc Fig. 5. Experimental relationships between FF and T.

The temperature dependence of Isc arises primarily with L,

which can be expressed as (Chen and Wang, 1999)
D kT
L ¼ ðDtÞ1=2 ð1Þ ¼ ð3Þ
m q
where D is the diffusivity and t the carrier lifetime. D can be
where subscripts n and p are the signs of the N and P type regions,
expressed by m and the Einstein relation
respectively. Since m varies little at low temperatures, the net
m1n,p effect of D on L is small. It is assumed that the temperature
mn,p ¼ m0n,p þ ðT=300Þan,p ð2Þ
1 þðNa,d =Nn,p Þgn,p dependence of t is determined by single-level recombination
2216 G. Wang et al. / Applied Radiation and Isotopes 68 (2010) 2214–2217

Subsequently it is possible to predict that Isc increases very

little with temperature at low temperature and the experimental
results shown in Fig. 3 support this. Therefore the temperature
dependence of Isc will be neglected in the discussion.

3.2. Voc

The relationship between Voc and T is noticeable. Because Isc

varies very little, the temperature effect of electrical performance
of the betavoltaic cell comes just from the change of Voc due to the
changes of the characteristic parameters of semiconductor diode
such as ni, mn, p and I0 with T. Other electrical output results (for
example Pmax, Z) of the cell change accordingly.
When To700 K, the characteristic parameters of semiconductor
diode such as ni and I0 can be given by the relations with T and Na,d:
ni ¼ 3:87  1016 T 1:5 expð7:02  103 =TÞ ð6Þ
Fig. 6. Experimental relationships between Pmax and T.
 
Dn Dp
I0 ¼ sqni 2 þ ð7Þ
Ln Na Lp Nd
Voc can be given by the relation with Isc, I0 and T:
 
AkT Isc
V oc ¼ ln þ1 ð8Þ
q I0
The temperature dependence of Voc arises with I0 primarily. The
magnitude of I0 is determined primarily by ni with certain values of Na
and Nd. From Eqs. (6) and (7), the exponential increase in ni with
temperature results in the exponential increase in I0, so Voc must
decrease with increase in temperature. Based on Eqs. (1)–(3) and (6)–
(8), we can predict the relationship between Voc and T. The predicted
relationships are shown in Fig. 4. The Voc of device 1 is higher than
that of device 0 due to differences in the structure. The Voc of the two
devices linearly decreases with increase of T in the temperature range
we examined. The changing values are  3.7 mV/K (device 0) and
 3.3 mV/K (device 1), calculated by curve fitting of the predicted
relationships. Fig. 4 also shows the experimental results of the
Fig. 7. Experimental relationships between Z and T. relationships between Voc and T. The experimental results and
predicted relationships have similar trends. The changing values
Table 2 are  3.1 mV/K (device 0) and 3.0 mV/K (device 1) calculated by
Electrical performance of the two devices at 233.15 and 333.15 K curve fitting of the experimental results, which are both smaller than
the predicted result. In addition, the experimental values of Voc are
T (K) Isc (nA) Voc (V) FF Pmax (nw) Z (%)
smaller than the predicted ones. All of these differences may be
Device 0 233.15 24.0 0.313 0.735 5.52 0.670 attributed to the inevitable shunt resistance leaking out some current.
333.15 24.5 0.0123 0.266 0.0803 0.00975
3.3. Pmax and Z
Device 1 233.15 23.3 0.463 0.829 8.94 1.09
333.15 24.1 0.156 0.580 2.18 0.265
Isc and Voc parameters describe the basic electrical perfor-
mance. Isc increases very little with T as has been discussed in
statistics, so that t in the N and P type regions can be given as Section 3.1. Voc clearly decreases linearly as T increases, as has
   been analyzed in Section 3.2. Considering that FF of the I–V
ET EF characteristic curves decreases with increase of T, the relations of
tn ¼ tp0 1 þ exp ð4Þ
kT Eqs. (9) and (10) (Guo and Lal, 2003) result in decrease in Pmax and
Z with increase of T:
 
ET þ EF 2Ei Pmax ¼ FF Isc Voc ð9Þ
tp ¼ tn0 þ tp0 exp ð5Þ
kT
Pmax
Z¼  100% ð10Þ
The temperature dependence of lifetime arises from the P0
exponential terms in Eqs. (4) and (5). Since the exponential terms
are almost invariable in the case of moderate temperature, t is
approximately constant. 4. Conclusion
Another factor related to Isc is the generation rate. The
generation rate increases slightly with temperature. This increase There is a significant relationship between electrical performance
is due to the decrease in band gap, and the consequent increase in of the betavoltaic cell and temperature. The electrical performance
the number of electron–hole pairs that are generated by the beta according to the parameters Voc, Pmax and Z clearly decreases with
particles. However the effect is small in the temperature range increase in temperature in the case of silicon diodes irradiated by
considered in this paper and can be neglected. Ni-63. In the temperature range 233.15–333.15 K, the changing
 G. Wang et al. / Applied Radiation and Isotopes 68 (2010) 2214–2217 2217

values of Voc of the two betavoltaic cells are  3.1 and  3.0 mV/K Clarkson, J.P., Sun, W., Hirschman, K.D., et al., 2007. Betavoltaic and photovoltaic
and these changes in Voc bring about the decreases of Pmax and Z energy conversion in three-dimensional macroporous silicon diodes. Phys.
Status Solidi A 204 (5), 1536–1540.
because Isc varies very little with temperature. Duggirala, R., Li, H., Lal, A., 2008. High efficiency radioisotope energy conversion
using reciprocating electromechanical converters with integrated betavoltaics.
Appl. Phys. Lett. 92, 154104.
Acknowledgements Guenther, B.D., Weller, H.R., Godwin, J.L., 2001. Search for nuclear isotopes for use
in a nuclear battery. J. Propul. Power 17 (3), 540–546.
Guo, H., Lal, A., 2003. Nanopower betavoltaic microbatteries. In: Proceedings of the
The authors gratefully acknowledge financial support for this 12th International Conference on Solid State Sensors, Actuators and Micro-
work from the Science and Technology Development Foundation systems. IEEE, Boston.
Kosteski, T., Kherani, N.P., Gaspari, F., et al., 1998. Tritiated amorphous silicon films
of the China Academy of Engineering Physics (CAEP, 2007A02001). and devices. J. Vac. Sci. Technol. A 16 (2), 893–896.
Lee, S.K., Son, S.H., Kim, K.S., et al., 2009. Development of nuclearmicro-battery
References with solid tritiumsource. Appl. Radiat. Isot. 67, 1234–1238.
Ronen, Y., Hatav, A., Hazenshprung, N., 2004. 242mAm fueled nuclear battery. Nucl.
Instrum. Methods A 531, 639–644.
Chen, Z.M., Wang, J.N., 1999. Basic Material Physics for Semiconductor Devices. Sychov, M., Kavetsky, A., Yakubova, G., et al., 2008. Alpha indirect conversion
Science Publishing Company, China. radioisotope power source. Appl. Radiat. Isot. 66, 173–177.