A Comparative Study of Proton Radiation Damage in p- and n-Channel CCDs

TL;DR: In this paper, a comparative study of n-channel and p-channel charge coupled devices (CCDs) was conducted under the same conditions, and it was shown that p-CCDs are more radiation hard than conventional nchannel devices as they are not affected by the dominant electron trapping caused by the displacement damage defect the phosphorus-vacancy.

Abstract: It has been demonstrated that p-channel charge coupled devices (CCDs) are more radiation hard than conventional nchannel devices as they are not affected by the dominant electron trapping caused by the displacement damage defect the E-centre (phosphorus-vacancy). This paper presents a summary of the results from a comparative study of n-channel and p-channel CCDs each type operated under the same conditions. The CCD tested is the e2v technologies plc CCD47-20, a 1024 × 1024 frame transfer device with a split output register, fabricated using the same mask to form n-channel and p-channel devices. The p-channel devices were irradiated to a 10 MeV equivalent proton fluence of 4.07×10 10 protons.cm -2 and 1.35×10 11 protons.cm -2 , an n-channel CCD was irradiated to a 10 MeV equivalent proton fluence of 1.68×109 protons.cm-2, however due to time constraints the n-channel device was not characterised, n-channel comparisons are instead made using a CCD02. As expected the p-channel CCD demonstrated improved radiation tolerance when compared to the n-channel CCD, at -90 °C there is an approximate ×7 and ×15 improvement in tolerance to radiation induced parallel and serial CTI respectively for equivalent pixel geometries.

Summary

1. INTRODUCTION

• The concentration of phosphorous is two orders of magnitude lower than that in n-channel devices; therefore the formation of E-centre defects will be negligible [5].
• The parallel and serial buried channel widths are 17.5 µm and 36.0 µm respectively.
• The image area was irradiated leaving the store region un-irradiated to act as a control, serial CTI would not be measured as the serial register was not irradiated.
• The entire active area of six p-channel devices and one n-channel device were irradiated using 63 MeV protons at the Paul Scherrer Institut (PSI) in Switzerland, with one n-channel and p-channel device held as a control.

2. EXPERIMENTAL ARRANGEMENT

• The CCD47 was housed inside the vacuum test facility shown in Figure 2.
• Initially a potential mirror was created to convert the potentials provided by the XCAM Ltd. USB2REM1 camera drive box to those required to operate the p-channel CCD, given in Table 2.
• Mirroring the clock and reset potentials introduced a large amount of additional system noise, therefore, the mirror on the clock and reset potentials was removed and the ground referenced to 12.0 V to provide the required potentials.

3.1 Dark Current

• The mean dark current was measured across the surface of the p-channel devices using six sets of images, each taken using a 30 s integration period, the results as a function of temperature for three of the devices are shown in Figure 3.
• Srour demonstrated that a trap with an activation energy of 0.63 eV, attributed to the divacancy, is responsible for the temperature dependence of dark current associated with thermally generated charge [6].
• The data were plotted in an Arrhenius plot, shown in Figure 4, with a line of best fit drawn through data points above -55 °C.
• The feature formed below -90 °C is believed to show the limit of the measurement technique, where insufficient dark current is generated to be measured, requiring a longer integration period.

3.2 Charge Transfer Inefficiency

• The initial testing was performed using the p-channel control CCD held at -110.0 °C using an X-ray density of one event per 700 pixels, conducting fifteen readouts in 10 °C intervals, the results for parallel CTI are displayed in Figure 4.
• To increase the rate of data collection and the number of data points the X-ray density was increased to one event per 250 pixels, requiring only six readouts taken at 5 °C intervals, the results for parallel CTI are displayed in Figure 5.
• The temperature dependence of parallel CTI pre and post irradiation is shown in Figure 6, the trend for the irradiated and un-irradiated devices both demonstrate an increase in CTI as the device is cooled, possibly due to removal of thermally generated charge keeping the traps filled.
• The slow parallel line transfer, at 50 kHz, does not benefit from the increased emission time.
• The p-channel CCDs tested exhibited comparable parallel CTI, however, the serial CTI varied between the un-irradiated device and devices irradiated with the same proton fluence, as illustrated in Figure 7 where the CTI of a CCD irradiated with 4.07×1010 protons.cm-2 was measured to be lower than the un-irradiated control.

4. CONCLUSIONS

• The p-channel CCD47s tested demonstrate a clear improvement in tolerance to radiation induced CTI, showing an improvement of ×7 and ×15 for parallel and serial CTI respectively for equivalent pixel geometries when compared to an n-channel CCD.
• A program is required to improve the base parallel and serial CTI, possibly through the use of a bulk (float zone) devices which have demonstrated base CTI equivalent to n-channel devices, demonstrated by Bebek et. al. [7] and Dawson et. al. [11].
• Unfortunately the devices were not available for testing prior to the irradiation so no comparison could be made on the rate of hot pixel generation.
• The potential mirror will also be developed to improve its noise performance.

Figures

Fig. 5. Parallel CTI measured using the p-channel control as a function of temperature with different Mn-Kα event densities

Fig. 6. Parallel CTI of the un-irradiated and irradiated p-channel devices as a function of temperature

Fig. 7. Serial CTI of the un-irradiated and irradiated p-channel devices as a function of temperature

Table 1. Irradiation Characteristics for p-channel and n-channel CCD47-20s

Fig. 1. A photograph of the p-channel CCD47-20 [a] and the n-channel CCD02 [b]

Fig. 8. Trend in parallel CTI at -90 °C based on the experimental data showing the base CTI and the effect of increased proton fluence

Fig. 9. Comparison of the p-channel CCD47 with an n-channel CCD02 at -90 °C showing the ×7 improvement in radiation tolerance and the operational suggestions for a p-channel CCD47 at its current level of performance

Table 1. Irradiation Characteristics for p-channel and n-channel CCD47-20s

Fig. 2. A photograph of the vacuum test facility

Fig. 4. Dark current as a function of 1000/T for three p-channel devices pre and post irradiation

Fig. 3. Dark current as a function of temperature for three p-channel devices pre and post irradiation

Full text

Open Research Online
The Open University’s repository of research publications
and other research outputs
A comparative study of proton radiation damage in p-
and n-channel
Conference or Workshop Item
How to cite:
Gow, J.; Murray, N. J.; Holland, A. D.; Burt, D. and Pool, P. (2009). A comparative study of proton radiation damage
in p- and n-channel. In: Proceedings of SPIE: UV, X-Ray, and Gamma-Ray Space Instrumentation for Astronomy
XVI, 3 Aug 2009, San Diego, USA.
For guidance on citations see FAQs.
c
 SPIE - The International Society for Optical Engineering
Version: Accepted Manuscript
Link(s) to article on publisher’s website:
http://dx.doi.org/doi:10.1117/12.826866
Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyright
owners. For more information on Open Research Online’s data policy on reuse of materials please consult the policies
page.
oro.open.ac.uk

A Comparative Study of Proton Radiation Damage in p- and n-
Channel CCDs
J. Gow*
a
, N. J. Murray
a
, A. D. Holland
a
, D. Burt
b
, P. Pool
b
a
e2v centre for electronic imaging, The Open University, PSSRI, Milton Keynes, MK7 6AA, UK
b
e2v technologies plc, 106 Waterhouse Lance, Chelmsford, Essex, CM1 2QU, UK
ABSTRACT
It has been demonstrated that p-channel charge coupled devices (CCDs) are more radiation hard than conventional n-
channel devices as they are not affected by the dominant electron trapping caused by the displacement damage defect the
E-centre (phosphorus-vacancy). This paper presents a summary of the results from a comparative study of n-channel and
p-channel CCDs each type operated under the same conditions. The CCD tested is the e2v technologies plc CCD47-20, a
1024
×
1024 frame transfer device with a split output register, fabricated using the same mask to form n-channel and
p-channel devices. The p-channel devices were irradiated to a 10 MeV equivalent proton fluence of 4.07
×
10
10
protons.cm
-2
and 1.35
×
10
11
protons.cm
-2
, an n-channel CCD was irradiated to a 10 MeV equivalent proton fluence of
1.68
×
10
9
protons.cm
-2
, however due to time constraints the n-channel device was not characterised, n-channel
comparisons are instead made using a CCD02. As expected the p-channel CCD demonstrated improved radiation
tolerance when compared to the n-channel CCD, at -90 °C there is an approximate ×7 and ×15 improvement in tolerance
to radiation induced parallel and serial CTI respectively for equivalent pixel geometries.
Keywords: CCD, Proton radiation damage, p-channel, n-channel, charge transfer efficiency, displacement damage
hardened
1. INTRODUCTION
The work presented in this paper has been carried by the e2v centre for electronic imaging at the Open University to
investigate the post proton irradiation performance of n-channel and p-channel CCDs fabricated using the same mask set.
It is essential for a space mission that the selected detector meets the performance requirements over the mission
duration, accounting for performance loss due to the space radiation environment [1]. Displacement damage caused by
protons within the Earths radiation belts and from cosmic rays (solar and galactic) generate stable lattice defects within
the silicon, creating energy levels within the silicon band-gap. The three main stable defects, for n-channel CCDs, are the
phosphorous vacancy (E-centre) at 0.44 eV, the oxygen-vacancy (A-centre) at 0.18 eV (both as a result of impurity
atoms within the lattice), and the divacancy (J-centre) at 0.40 eV and 0.25 eV consisting of two adjacent vacancies. The
E-centre is located at near mid-gap, the ideal location for electron trapping. An n-channel device uses phosphorus as the
dopant atom for the buried channel; it was suggested and demonstrated that using boron to create a p-channel device
would reduce the post irradiation increase in charge transfer inefficiency (CTI) [2].
It is difficult to compare CTI results taken using different operating conditions and equipment setups, as described
elsewhere in this volume [3]. The aim of this work was to provide a definitive comparison of n-channel and p-channel
devices. The devices were to be operated using the same drive voltages and timings, held in the same position within the
test facility and exposed to the same incident X-ray flux, however due to time constraints the n-channel CCD47 was not
tested. Results found using a CCD02 characterised using comparable voltages, timings, and X-ray flux [4] were used to
provide the n-channel comparison. The X-ray method was selected as the measurement technique for CTI, using Mn-K
α
X-rays at 5,898 eV (
~
1600 e
-
), measured between -40.0 °C and -110.0 °C, with the dark current measured between
0.0 °C and -110.0 °C.
*jpdg3@open.ac.uk; phone +44 (0) 1908 852 769; fax +44 (0) 1908 858 052; www.open.ac.uk/cei

The CCD tested was the e2v technologies plc front illuminated CCD47-20, a frame transfer device with an image and
store format of 1024 by 1024 with 13 µm square pixels, illustrated in Figure 1a. The parallel and serial buried channel
widths are 8.5 µm and 24.0 µm respectively. The p-type buried channel was doped with boron, the epitaxial layer with
phosphorous (20 to 100 Ω.cm), and the substrate with antimony (< 20 mΩ.cm). The dopants were selected to provide
comparable resistivity and properties to those found in n-channel devices, for ease of manufacture [5]. The concentration
of phosphorous is two orders of magnitude lower than that in n-channel devices; therefore the formation of E-centre
defects will be negligible [5]. The CCD02 is a front illuminated device, with an image format of 385 by 578 with 22 µm
square pixels, illustrated in Figure 1b. The parallel and serial buried channel widths are 17.5 µm and 36.0 µm
respectively.
Fig. 1. A photograph of the p-channel CCD47-20 [a] and the n-channel CCD02 [b]
A proton irradiation was originally conducted, using two p-channel devices and one n-channel device, at the Kernfysisch
Versneller Instituut (KVI) in the Netherlands. The image area was irradiated leaving the store region un-irradiated to act
as a control, serial CTI would not be measured as the serial register was not irradiated. A pre-irradiation characterisation
was not performed and post-irradiation it became apparent that the irradiated devices, and un-irradiated devices from the
same wafer, exhibited defects which flooded the active area with excess charge. As a result testing moved to six
p-channel devices which had been irradiated as part of an e2v study into p-channel and n-channel devices, funded by
ESA. The entire active area of six p-channel devices and one n-channel device were irradiated using 63 MeV protons at
the Paul Scherrer Institut (PSI) in Switzerland, with one n-channel and p-channel device held as a control. The irradiation
details, including the 10 MeV proton fluence, is summarised in Table 1. Post irradiation an anneal stage at 100 °C, with
the irradiated devices unbiased, was performed for a duration of 168 hours the p-channel devices exhibited negligible
change in CTI post anneal [5].
Table 1. Irradiation Characteristics for p-channel and n-channel CCD47-20s
Device Type
Beam Energy
(MeV)
Flux
(p.cm
-2
.s
-1
)
Fluence
(p.cm
-2
)
10 MeV equivalent
fluence (p.cm
-2
)
×3 p-channel 63 1.70×10
8
2.70×10
11
1.35×10
11
×3 p-channel 63 1.40×10
8
8.12×10
10
4.07×10
10
×1 p-channel Not irradiated
×1 n-channel 44 1.95×10
7
3.03×10
9
1.68×10
9
×1 n-channel Not irradiated
2. EXPERIMENTAL ARRANGEMENT
The CCD47 was housed inside the vacuum test facility shown in Figure 2. A vacuum pump was used to evacuate the air
in the chamber with testing occurring at a pressure of < 10
-5
mbar. The CCD being tested was clamped onto a copper
cold bench connected to a CryoTiger® refrigeration system (PT-30) capable of cooling the detector to around -120 °C or
153 K. A resistive heater was employed in thermal contact between the copper cold bench, allowing a maximum
operating temperature of -40 °C before overloading the cooling capacity of the CryoTiger®. To allow for warmer
operating temperatures the data were recorded as the CCD cooled, with the heater at maximum to reduce the cool down
rate. The temperature can be controlled to within ± 0.1 °C using a feedback system, comprising a Lakeshore 325
temperature controller, platinum resistance thermometer (PRT), and the heater. The temperature of the CCD ceramic was
[a] [b]

measured using a 1,000 Ω PRT (it is assumed the device silicon is in good thermal contact with the ceramic). An Oxford
Instruments XTF5011/75-TH X-ray tube (tungsten anode) was used to fluoresce a manganese target held at 45° to the
incident X-ray beam to provide a known energy (5,898 eV) for calibration and CTI measurements, with an X-ray density
of approximately one event every two hundred and fifty pixels.
Fig. 2. A photograph of the vacuum test facility
Initially a potential mirror was created to convert the potentials provided by the XCAM Ltd. USB2REM1 camera drive
box to those required to operate the p-channel CCD, given in Table 2. Mirroring the clock and reset potentials introduced
a large amount of additional system noise, therefore, the mirror on the clock and reset potentials was removed and the
ground referenced to 12.0 V to provide the required potentials. The total noise measured at -110 °C reduced from 200 e
-
r.m.s. to 20 e
-
r.m.s after the removal of the potential mirror. A headboard provided local low-pass filtering for the D.C.
bias connections and pre-amplification of the output, with the clocking and biasing provided by an XCAM Ltd.
USB2REM1 camera drive box in conjunction with USB2 v1.15 drive software. The devices were operated using a line
transfer rate of 50 kHz, and a readout rate of 500 kHz.
Table 1. CCD47-20 operational voltages
Clock/Bias Description Bias Used (V)
Vss Substrate 0.0
Vog Ouput gate 3.0
Vrd Reset drain 17.0
Vod Output drain 32.0
IØ Image/Store clock 12.0
RØ Register clock 12.0
ØR Reset gate 12.0
V
ABD
Anti blooming drain 28.0
3. RESULTS AND DISCUSSIONS
3.1 Dark Current
The mean dark current was measured across the surface of the p-channel devices using six sets of images, each taken
using a 30 s integration period, the results as a function of temperature for three of the devices are shown in Figure 3.
Srour demonstrated that a trap with an activation energy of 0.63 eV, attributed to the divacancy, is responsible for the
temperature dependence of dark current associated with thermally generated charge [6]. The data should be proportional
to exp(-0.63/kT), where k is Boltzmann’s constant and T is the temperature. The data were plotted in an Arrhenius plot,
Rotary target wheel for future
CTI measurement at multiple
X-ray energies
X-ray tube, allowing the
incident X-ray flux to
be controlled
Cold end to provide
cooling to -120 °C

shown in Figure 4, with a line of best fit drawn through data points above -55 °C. The feature formed below -90 °C is
believed to show the limit of the measurement technique, where insufficient dark current is generated to be measured,
requiring a longer integration period. The activation energy was calculated to be 0.61 eV un-irradiated, 0.62 eV
irradiated with 4.07×10
10
protons.cm
-2
, and 0.63 eV irradiated with 1.35×10
11
protons.cm
-2
, the results are comparable to
those found by Srour [6], Bebek et. al. [7], and Spratt [8]. This suggests that for the p-channel devices tested the
divacancy is the dominant source of thermally generated dark current pre and post irradiation. The increase in dark
current as a function of proton fluence is approximately linear. Based on the fits to the data the increase after irradiation
at 21 °C (at this temperature the device was saturated i.e. ADC max out) was calculated to be
~
1.4 nA.cm
-2
.krad
-1
,
similar to that found in the p-channel CCD02 [9] and other e2v n-channel CCDs [10].
Fig. 3. Dark current as a function of temperature for three p-channel devices pre and post irradiation
Fig. 4. Dark current as a function of 1000/T for three p-channel devices pre and post irradiation
0.01
0.1
1
10
100
1000
3.7 3.9 4.1 4.3 4.5 4.7 4.9 5.1 5.3 5.5 5.7
1000/T (1000.K
-1
)
Dark Current (e
-
.pixel
-1
.s
-1
)
Series4
2.70E+11
Series5
Un-Irrad
Series6
8.12E+10
Expon. (2.70E+11)
Un-irradiated control
4.07×10
8
p.cm
-2
1.35×10
8
p.cm
-2
Best fit to un-irradiated data
Best fit to 4.07×10
10
p.cm
-2
data
Best fit to 1.35×10
11
p.cm
-2
data
Limit of measurement
technique
1
10
100
1000
-75.0 -70.0 -65.0 -60.0 -55.0 -50.0 -45.0 -40.0 -35.0 -30.0
Temperature (
o
C)
Dark Current (e
-
.pixel
-1
.s
-1
)
Series4 8.12E+10
Series5 2.70E+11
Series6 Un-Irrad
Series7 Expon. (Un-Irrad)
Expon. (8.12E+10) Expon. (2.70E+11)
Un-irradiated control
4.07×10
8
p.cm
-2
1.35×10
8
p.cm
-2
Best fit to un-irradiated data
Best fit to 4.07×10
10
p.cm
-2
data
Best fit to 1.35×10
11
p.cm
-2
data

Frequently asked questions

Q1. What are the contributions mentioned in the paper "A comparative study of proton radiation damage in p- and n-channel" ?

This paper presents a summary of the results from a comparative study of n-channel and p-channel CCDs each type operated under the same conditions. 

Q2. What are the future works mentioned in the paper "A comparative study of proton radiation damage in p- and n-channel" ?

Future work will include completing the testing with the n-channel CCD47, measuring the parallel CTI as a function of temperature as only the n-channel device irradiated at KVI is available for testing. Despite the initial poor CTI the large improvement in tolerance to radiation induced CTI still makes these devices, at their current level of performance, suitable for use in hostile radiation environments indicating that p-channel devices will have a large part to play in the future of CCDs in space. The potential mirror will also be developed to improve its noise performance.