Characterization and acceptance testing of fully depleted

thick CCDs for the Large Synoptic Survey Telescope

Ivan V. Kotov

, Justine Haupt

, Paul OConnor

, Thomas Smith

, Peter Takacs

, Homer

Neal

, and Jim Chiang

Brookhaven National Laboratory, Upton, NY 11777, USA

SLAC, Menlo Park, CA 94025, USA

ABSTRACT

The Large Synoptic Survey Telescope (LSST) camera will be made as a mosaic assembled of 189 large format

Charge Coupled Devices (CCD). They are n-channel, 100 micron thick devices operated in the over depleted

regime. There are 16 segments, 1 million pixels each, that are read out through separate ampli?ers. The

image quality and readout speed expected from LSST camera translates into strict acceptance requirements for

individual sensors.

Prototype sensors and preproduction CCDs were delivered by vendors and they have been used for developing

test procedures and protocols. Building upon this experience, two test stands were designed and commissioned

at Brookhaven National Laboratory for production electro-optical testing.

In this article, the sensor acceptance criteria are outlined and discussed, the test stand design and used

equipment are presented and the results from commissioning sensor runs are shown.

Keywords: CCD, Characterization, calibration

1. INTRODUCTION. LSST SENSOR REQUIREMENTS

The quality of a sensor can be characterized by a set of parameters describing sensor performance in speci?c areas.

The output ampli?er and signal routing are characterized by: gain, noise, linearity and cross talk. The front

surface processing and design are responsible for full well capacity and pixel response non-uniformity (PRNU).

The silicon bulk and back surface quality translates into quantum e?ciency (QE), number of bright and dark

defects, dark current, charge transfer e?ciency (CTE), and number of traps. The sensor thickness and applied

electrical ?eld determine the point spread function (PSF). Requirements on sensors follow from survey goals

and design. In three words, the survey can be described as wide, deep and fast. Other leading considerations

are the image quality and survey e?cacy. Simply put, images should not be degraded by the camera and

sensor performance. For example, on sky time should be spent on making exposures so sensor readout time

should be minimized. These survey level parameters can be satis?ed by applying speci?c requirements on sensor

parameters. The LSST sensor speci?cations are summarized in table 1 together with results obtained from the

?rst article sensor. More discussions on the electro-optical testing can be found in Ref.

{

and the camera design

overview in Ref.

In the table

, speci?cation de?nitions are:

? bright pixel - pixel with dark current exceeding 5e

/s;

? dark pixel - pixel with sensitivity to 500 nm light less than 80% of the segment mean;

? bright column - column with more than 20 contiguous pixel with dark current exceeding 5e

/s;

? dark column - column with more than 100 contiguous pixels with sensitivity to 500 nm light less than 80%

of the segment mean.

Both serial and parallel CTE are measured using the Extended Pixel Edge Response, EPER, method. All

measurements are conducted at a sensor temperature of -95 C.

Further author information: (Send correspondence to Ivan Kotov)

E-mail: kotov@bnl.gov, Telephone: 1 631 344 2615

Table 1. LSST sensor speci?cations and electro-optical test results for an E2V-CCD250 sensor

Description

Speci?cation

Measurement

Read noise

<8e-

4.72 - 5.04e-

Blooming full well

<175ke-

115-120 ke-

non-linearity

<2%

max deviation 1.7%

cross talk

< 1:9 ? 10

? 3

not measured

PRNU

< 5%

max variation 4.2% at 350 nm

u band QE

> 41%

55.4%

g band QE

> 78%

87.4%

r band QE

> 83%

92.1%

i band QE

> 82%

95.5%

z band QE

> 75%

91.9%

y band QE

> 21%

35.4%

Bright pixels

Dark pixels

486

Bright columns

Dark columns

Traps > 200 e-

Cosmetic, sum of above

< 0:5% defective pixels 0.08%, 12504 defective pixels

Dark current 95-th percentile < 0:2 e-/s

0.02 e-/s

Serial CTE

>0:999995

0:999947+= ? 9 ? 10

? 6

Parallel CTE

> 0:999997

0:9999995+= ? 6 ? 10

? 6

Point spread function, sigma

< 5 micron

3.64 micron

2. TEST STAND HARDWARE AND CONTROL SOFTWARE

To perform the acceptance test, the following equipment is needed: the Dewar housing CCD under test; CCD

controller; a monochromatic light source with wavelength range 300-1100 nm; photo diode (PD) to monitor the

light ?ux; retractable

F e source with motion controller; picoammeters; voltage source; cryogenic controller;

pressure and temperature sensors. The test stand layout is shown in Fig.

. The Dewar with a 210 mm inner

diameter and a 180 mm window aperture is used in the horizontal orientation. Vacuum is obtained using a

dry diaphragm pump and a turbo pump attached directly to the chamber plumbing. The vacuum pressure

is monitored by a Granville-Phillips Micro-Ion gauge. The residual gas is monitored using an MKS VQM835

ion trap mass spectrometer. The CCD is cooled using a Brooks Polycold PCC Compact Cooler system. The

temperature is measured using pt100 RTDs. A CryoCon Model 24C temperature controller monitors the RTDs

and maintains the temperature set point. Inside the Dewar, ?ex cables bring signals from the CCD to hermetic

DB50-style connectors on two opposite ?anges. Video signals (OD/OS pairs) are routed to four feedthroughs

(4 channels per feedthrough) and the clocks are brought to a separate feedthrough. Clock signals are passed

between the CCD two sides on a separate internal cable. As dictated by CCD type, rigid-?ex cables with JFET

bu?ers placed on the CCD end are installed.

The CCD controller, Archon, is supplied by Semiconductor Technology Associate, Inc. The controller is

located atop the Dewar, allowing a short, 3' cable run. A custom transition board routes all signals to ?ve

DVI-I cables connected to preampli?ers and a clock board. Boards are attached to the feedthrough connector on

the air sides. Preampli?ers are custom 4-channel circuits on 100 ? 100 mm printed circuit cards. The on-CCD

Figure 1. Test stand layout photo.

source follower load is a current limiting diode providing 3.6 mA. Video signals are decoupled by a 2200 pF NPO

capacitor and clamped to ground by an analog switch. The ?rst stage of preampli?cation uses FET, LSK389

inputs to a low-noise op-amp, OPA 1612 which provides a gain of 2 - 4 with appropriate gain resistor choice.

This is followed by a quasi-di?erential driver. Filters are also included for reset, guard, and substrate biases

and the output gate bias is bu?ered. The boards have a ground plane to which the local substrate point on the

CCD is tied. Each preampli?er section has its own linear voltage regulators and power planes. The preampli?er

power (+= ? 7.5V @ 200 mA) is supplied from Archon. Clock signals are brought through a separate PCB which

provides TVS diode clamps to protect the CCD.

F e X-ray exposure device is installed inside the Dewar to provide an in-situ X-ray source for calibration

purposes. The 5?Ci

Fe source is mounted directly on a motor armature that moves the source over the sensor

during exposure and returns it to an edge location behind the shield. On the other end, the armature has 2

permanent magnets moving over a stator printed circuit board. The current applied to the printed "coils" drives

the magnets. The elimination of conventional windings in the stator makes this arrangement intrinsically vacuum

compatible. The armature pivots on ball bearings coated with Dicronite, a Tungsten Disul?de based ultra-high

vacuum compatible dry lubricant. The actuation of the XED is controlled via a programmable digital signal

from the Archon Controller.

The sensor specs call for good sensor illumination uniformity or an accurately measured illumination pro?le.

We have chosen to do both: limit illumination non-uniformity by design means to less than 1% variation across

the sensor area and measure the illumination pro?le. The light source is a 300W Xe arc lamp Newport 6258 in

a Newport 66485-300XF-R1 housing with an aspheric F/1 lens. For the wavelength selection, a Cornerstone 260

1/4m monochromator with motorized slits, CS260-RG-1-MT-D, is used. The gratings are: 1200 l/mm, 350 nm

blaze, range 200-1400 nm, Newport 74063; 600 l/mm, 400 nm blaze, range 250-1300, Newport 74066. They are

installed on a triple grating assembly together with a 50 ? 50 mm mirror, Newport SP45701-4682. The light from

the lamp is focused onto the monochromator input slit using an o?-axis parabolic mirror, Newport 50331AL.

Both input and output slit widths are set to 240 ?m. Mounted in front of the monochromator are a 25 mm fast

shutter Newport 76994, with exposure frequencies up to 40 Hz, and a motorized 6 position ?lter wheel, Newport

74041. The ?lter wheel houses 2 order-blocking long-pass ?lters with 275 nm and 550 nm cut-on wavelength

and neutral density ?lters. The output of the monochromator is directly coupled to a 6" integrating sphere,

Labsphere 4P-GPS-060-SF. A silicon photodiode Hamamatsu S2281-04 is mounted in an auxiliary port of the

integrating sphere and is read out by a Keithley 6487 picoammeter.

The illumination uniformity was achieved by placing the sensor at a distance of ˘ 550 mm from the output

port of the integration sphere. The dark tunnel with single ba?e is used to block the room light from entering

the sensor and suppress scattered light. The measured light level variation is 0.3% over 35 mm of travel in

good agreement with design parameters (for the uniformity calculation see equation 2-37 in reference

). This is

adequate for our testing purposes.

All instruments are connected to a Linux PC and controlled by the Java based LSST Camera Control

System, CCS. CCS provides the device control and monitoring layer. An application for communications with

the Archon Controller and an application for controlling the auxilliary devices (monochromator, lamp, back bias

voltage source, photodiode picoammeter, environment monitor and power distribution unit) receive scripting

commands over a UDP messaging bus, JGroups.

The acquisition scripts are written in a JAVA implementation

of Python, Jython.

The scripts send commands and receive responses using messages over the JGoups bus

with the CCS applications. Many of the CCS commands are higher level commands involving coordination of

multiple devices. Images are written using the FITS ?le format which also includes metadata on the setting,

operating and environmental conditions. A trending database application is used to keep time histories of all

essential parameters. On the instrument side, communication goes through RS232, RS485 or Ethernet ports.

On the PC side, USB and Ethernet ports are used. The USB-RS232 and USB-RS485 adapters eliminate the

need for additional interfaces.

Each sensor is registered in the LSST electronic traveler system and a traveler type for each type of operation

is used to guide and track all activities on the sensor. For electro-optical testing, the traveler follows all steps from

the preparation before accessing the sensor to its storage after completion of testing including: the equipment

checks before sensor installation, the sensor insertion into the Dewar, pumping and cooling, validation of the

readiness of the sensor for testing, an automated 12 hour long series of acquisitions and analyses of the data

terminating in a complete test report, the warm-up of the Dewar and ?nally the sensor extraction, storage

and declaration of suitability of the sensor for use in an raft tower module. During the acquisition sequence,

which is typically run unaided overnight, the quantities of exposures and corresponding FITS ?les produced are

shown in Table

and Fig

. The FITS ?les have multiple-extensions starting with a primary header containing

essential meta-data including sensor identi?ers, environment conditions and conditions of test stand devices.

This is followed by 16 image extensions (one per segment), an extension for test conditions, an extension for

CCD operating conditions, and an extension for monitoring photo diode light curve data. For photo diode

calibrations, an extra similar extension is added to register the calibrated photo diode data (see calibration

procedure description in section

2.2

Figure 2. Electro-optical test output ?le count

Table 2. Electro-optical test output ?le count

TEST

EXPOSURE TYPE COUNT

Dark

bias

Dark

dark

bias

Flat

bias

Flat

?at

142

bias

155

?at

persistence bias

persistence dark

persistence ?at

pre?ight

?at

SFlat

?at

Trap

bias

Trap

pocket pump

2.1 Hardware calibration

QE measurements are most demanding in terms of equipment calibration. All instruments used for these mea-

surements have to be calibrated periodically. These instruments are: picoammeters (in our case Keithley model

6487) to measure the monitoring PD current and calibrated PD current; monitoring PD (Hamamatsu S2281-04)

mounted on the integrating sphere; NIST calibrated PDs (Hamamatsu S2281-01) and monochromator (Corner-

stone 260). What quantities are involved in QE calculations and how their accuracy a?ects the total QE accuracy

can be seen from the following equations (also discussed in

{

). The CCD QE is the ratio of the number of

electrons, N

detected by CCD to the number on photons, N

fallen onto the CCD

QE=

(1)

The number of electrons is calculated as

=G ? A;

(2)

where G is conversion gain in e ? =adu and A is measured signal in adu. The number of photons N

is calculated

R(?)

(?)

;

(3)

where Q

is the charge measured by monitoring PD; q

is the electron charge; R(?) is the ratio of monitoring

PD to calibration standard PD currents, and quantum e?ciency QE

(?) is related to spectral responsivity as

(?) =

S(?)

;

(4)

where S(?) is the calibration standard PD sensitivity provided by NIST in our case; ? is the wavelength and hc

is Plank's constant multiplied by speed of light. The total error is a sum of contributing factor errors

?QE

;

(5)

These equations shows that in terms of equipment calibration we need to measure the ratio R(?), have the

monochromator output wavelength well de?ned and measure monitoring PD charge accurately.

The Keithley 6487 picoammeter is factory calibrated and comes with calibration certi?cate. Provisions are

made for periodic factory recalibration. The accuracy in 2 nA current range is 0.3% and 0.1% above 20?A.

Monochromator calibrations are done using Xe spectral lines in two wavelength ranges: 800-900 nm and 400-

500 nm. First a scan through the 800 to 900 nm range was performed with 0.3 nm increments and narrow 48?m

slit setting. At each step, 3000 monitoring PD readings are taken during 50 sec using the Keithley picoammeter.

For the ?rst 10 seconds the shutter was closed, the next 25 seconds the shutter was open, and for the remaining

time (about 15 seconds) the shutter was closed again. An example of one such sequence at 823 nm is shown in

Fig.

Figure 3. PD current, X and Y axis represent time in sec and current in pA, respectively.

The signal area minus the baseline was determined by integration with respect to time using a rectangle

approximation method. This gives us the total accumulated charge over the entire 25 seconds of exposure.

The baseline is measured using data points when the shutter was closed. Accumulated charges plotted against

respective wavelengths are shown in Fig.

This graph gives us the relative intensity of the light coming from the Xenon lamp at various wavelengths

which can be compared to NIST spectral line data. The relatively isolated peak at approximately 840 nm was

?tted with a Gaussian (in red). The Gaussians center was found to be at 839.977 nm. According to the NIST

database entry for Xenon I ,

this spectral line actually is at 840.919, which means that the monochromator is

o? by 0.942 nm. This was checked for several other peaks at other wavelength and they all show about the same

shift value. Two monochromator parameters, o?set and factor, needed to be adjusted. The necessary wavelength

calibration was performed following the procedure from Cornerstone 260 manual.

To perform our calibration, we used the o?set command to change the o?set by the amount determined by the

spectral line at 840nm. The 800-900 nm range scan performed after the o?set calibration showed the agreement

with NIST database within 0.02 nm. Then the factor calibration accuracy was assessed. The wavelength range

400-500 nm was scanned to check how peaks match the NIST data after o?set correction. The lamp spectra in

this range and the 473 nm line ?t are shown if Fig.

Figure 4. The lamp spectra in 800-900 nm range.

We determined that the match is good enough to justify not modifying the factor parameter.

2.2 Monitoring photo diode calibration

The monitoring PD is calibrated in situ using NIST calibrated photo diodes placed at the end of the dark tunnel.

The circular precision aperture is mounted in front of the calibrated PD and it de?nes the active area. The

aperture diameter was measured on OGP coordinate measuring machine with positional accuracy of ˘ 2?m

and found to be 8:975mm (correspondingly the area is 63:259 mm

). The calibrated PD is positioned behind

the Dewar window at the location of the CCD. The calibration PD substitutes the CCD. The ratio R(?) of

monitoring PD current to calibrated PD current is measured as a function of the wavelength. The ?nal scan is

shown in Fig.

. The ?nal ratio measurements are done in 10 nm steps with slits in nominal position. We used

3 NIST calibrated PD for these measurements. This gives us a sense of these measurements accuracy. We found

that repeat ability is better than 0.2% over 350 950 nm range. This is in agreement with accuracy of calibrated

PDs. Their relative expanded uncertainty is 0:2% in the visible wavelength range 450 - 950 nm and goes up to

3% at 1100 nm and to 0.7% at 350 nm.

3. X-RAY DATA SET AND MEASUREMENTS

The

Fe Mn K?;? lines provide the absolute system gain calibration. X-rays absorbed in silicon generate

electrons in amounts proportional to the energy. For example, in Si at -100C K? X-ray produces 1602 electrons

(assuming pair creation energy is w=3.68 eV per e/h pair

{

). The system gain knowledge is needed for most

of the measurement tasks data analyses. Therefore, an electro-optical test, EOT, normally starts with the

F e

data acquisition and analysis. In the acquisition part, 25 bias frames and 25 exposures are taken. The exposure

time is 3.5 sec with 0.5 sec allocated for the arm to detract before readout starts. The total exposure time plus

readout time is under 5 minutes.

An example of a

Fe X-ray spectra is shown in Fig.

for a high statistics run, ten times more exposures

than usual. This plot includes X-rays hits from all segments and was produced with the gain calibration taken

Figure 5. The lamp spectra in 400-500 nm range. Observed peak: 473.139nm, NIST line: 473.415nm, di?erence: 0.276nm.

into account. The Mn K?; ? lines are dominant with negligible background. The K? peak to background ratio

is 10

. In this situation the gain estimate is rather insensitive to the background shape and is robust. The

statistical accuracy in the K? peak position ?t can be roughly estimated using the peak width and the square

root of the number of hits. The number of K? hits in a nominal EOT run varies from segment to segment in

the range 2 ? 10

? 9 ? 10

. The Mn K?;? line width is what is expected for 5e

readout noise. It should

be noted that X-ray hits typically occupy 3 ? 3 pixels and consequently the noise in the total signal is 15e

The total line width is the sum of natural line width and signal noise in quadrature and is about 20e

. Thus,

the peak position accuracy is around 10

? 4

. It means that the segments relative calibration is at the 10

? 4

level.

Systematic errors dominate the absolute gain measurement accuracy. Most of the systematic uncertainty comes

from the uncertainty in the pair creation energy and is ˘ 1% percent.

There are peaks in the low energy part of

Fe X-ray spectra. The ?t of these peaks and their identi?cation

are shown in Fig.

. The Si K? X-ray has a 30 micron mean free path in silicon and as a result it can escape

from Mn K?;? hits. This process leads to three peaks: Si K?, Mn K? escape and Mn K? escape. Other

peaks are ?uorescence of materials activated by Mn K?;? X-rays. These lines can be used to check system

linearity at low signal level, see .

The M n K? events are also used for point spread function measurements by analyzing the charge distribution

in X-ray clusters (for details see ,

2324

When the gain has been determined, the system noise is measured using bias frames or overscan regions (see

discussions in

). The gain values are essential input for QE and blooming full well measurements.

3.1 CTE measurements with X-rays

The EPER method used for CTE measurements is very strait forward to apply but it also limited to interrogating

just a few rows or edge pixels. The X-ray method presented in

reveals CTE problems at any location in the

active area. The E2V-CCD250 sensor featured in table

fails our serial CTE requirements. Further studies

Figure 6. The wavelength dependence for the ratio of monitoring PD current to calibrated PD current, ?nal measure-

ments.

are needed to understand CTE problems. In addition to the standard EOT analysis package, we have developed

specialized analyses for detailed studies ,

An X-ray method

was applied to E2V-CCD250-088

data. The poor CTE manifests itself as a signal dependency on the number of transfers. In our case, the signal

is the Mn K? line. In the serial direction, no dependency was found. In the parallel direction, some segments

show a signal change with the number of transfers as can be seen in Fig.

The detailed analysis of the Mn K? position in di?erent areas is shown in Fig.

for segments 0 and 12, as

an example.

4. CONCLUSIONS

Two test stands were commissioned at BNL for LSST electro-optical sensor testing. Nineteen sensors from 2

vendors were tested as part of commissioning and initial operations. The system noise is 2e

. The system ability

to measure sensor noise at 4-5e

level is demonstrated. The periodic equipment calibration and precision gain

measurements using

Fe allows us to perform absolute QE measurements with a precision of a few percent. In

addition to the standard analysis package, the analysis tools for detailed sensor studies were developed.

ACKNOWLEDGMENTS

The authors acknowledge the help of e2v technologies, Silicon Technology Associates and Imaging Technol-

ogy Laboratory in sensor prototype development. We acknowledge the help of P.Kuczewski, J.Kuczewski and

G.Fraizer with test stand assembly. We acknowledge the help of Eric Aubourg, Anthony Johnson, Heather Kelly,

Etienne Marin-Matholaz, Dmitry Onoprienko, Owen Saxton, Massimiliano Turri and Francoise Virieux with pro-

duction acquisition software development. We acknowledge the help of Rebecca Coles with the pretesting needed

to develop the methodology for the QE measurements. We thank Peter Doherty for his important contributions

to the testing procedure and access to his test setup for early testing of the software. We acknowledge the

Figure 7.

55Fe

spectra, Mn K?;? lines ?t is shown by red curve.

help of Richard Dubois, Joanne Bogart, Warren Focke and Anders Borgland with the eTraveler development.

Many thanks to Jon Thaler, Stuart Marshall and Seth Digel for their input to the CCS framework design and

documentation. We also would like to recognize the input of Bill Wahl, Chris Stubbs, Michelle McQueen and

Matthew Rumore. This manuscript has been co-authored by employees of Brookhaven Science Associates, LLC.

LSST project activities are supported in part by the National Science Foundation through Governing Coopera-

tive Agreement 0809409 managed by the Association of Universities for Research in Astronomy (AURA), and the

Department of Energy under contract DE-AC02-76-SFO0515 with the SLAC National Accelerator Laboratory.

Additional LSST funding comes from private donations, grants to universities, and in kind support from LSSTC

Institutional Members.

REFERENCES

[1] P.O'Connor, V.Radeka, J.Frank, J.C.Geary, D.K.Gilmore, I.Kotov, P.Takacs, and J.A.Tyson, \Characteri-

zation of prototype LSST CCDs," Proc. SPIE 7021 , 702106 (2008).

[2] V.Radeka, J.Frank, J.C.Geary, D.K.Gilmore, I.Kotov, P.O'Connor, P.Takacs, and J.A.Tyson, \LSST sensor

requirements and characterization of the prototype LSST CCDs," Jinst 4 , P03002 (2009).

[3] I.V.Kotov, A.I.Kotov, J.Frank, P.Kubanek, M.Prouza, P.O'Connor, V.Radeka, and P.Takacs, \CCD charac-

terization and measurements automation," Nuclear Inst. and Methods in Physics Research A695 , 188{192

(2012).

[4] Doherty, P. E., Antilogus, P., Astier, P., Chiang, J., Gilmore, D. K., Guyonnet, A., Huang, D., Kelly, H.,

Kotov, I., Kubanek, P., Nomerotski, A., OConnor, P., Rasmussen, A., Riot, V. J., Stubbs, C. W., Takacs,

P., Tyson, J. A., and Vetter, K., \Electro-optical testing of fully depleted CCD image sensors for the large

synoptic survey telescope camera," Proc. SPIE 9154 , 915418 (2014). doi:10.1117/12.2056733.

[5] Kahn, S. and et.al, \Design and development of the 3.2 gigapixel camera for the Large Synoptic Survey

Telescope," Proc. SPIE 7735 , 77350J (2010).

Mn K

esc

Mn K

Ti K

esc

Mn K

Ca K

Cl K

Si K

Cluster Total Amplitude, e-

500

1000

1500

2000

Number of clusters

2000

4000

6000

Figure 8. Low energy part of

55Fe

spectra. Present lines are (left to right): Si K?; Cl K?; Ca K?; Mn K? escape; Ti

K?; Mn K? escape.

[6] Mahajan, V., [Optical Imaging and Aberrations: Ray geometrical optics, Parts 1-2], SPIE Press. (1998).

[7] \Jgroups - the jgroups project."

http://www.jgroups.org

[8] \The jython project."

http://www.jython.org

[9] Larason, T. C. and Houston, J. M., \Spectroradiometric detector measurements: Ultraviolet, visible, and

near-infrared detectors for spectral power." NIST Special Publication 250-41 (2008)

http://www.nist.gov/

calibrations/upload/sp250-41a.pdf

. (Accessed: 29 April 2016).

[10] Jacquot, B. C., Monacos, S. P., Hoenk, M. E., Greer, F., Jones, T. J., and Nikzad, S., \A system and

methodologies for absolute quantum e?ciency measurements from the vacuum ultraviolet through the near

infrared," Review of Scienti?c Instruments 82 , 043102 (2011).

[11] Lesser, M., \Silicon sensor quantum e?ciency, re?ectance, and calibration," Proc. SPIE 9154 , 915411

(2014).

[12] NIST, \Atomic Spectra Database Lines Form." NIST Database

http://physics.nist.gov/PhysRefData/

ASD/lines_form.html

. (Accessed: 26 April 2016).

[13] J.R.Janesick, [Scienti?c charge-coupled devices], vol. xvi of SPIE Press., Bellingham, Wash. (2001).

[14] B.G.Lowe and R.A.Sareen, \A measurement of the electronhole pair creation energy and the Fano factor

in silicon for 5.9 kev X-rays and their temperature dependence in the range 80 270 K," Nuclear Inst. and

Methods in Physics Research A576 , 367{370 (2007).

Figure 9. Mn K?; ? amplitude dependence on the number of parallel transfers (horizontal axis). The number of electrons

measured for X-ray cluster in a.d.u. (1 a.d.u is ˘ 1 e-) is along vertical axis. Parallel CTE has a problem in segments

0,1,2,3,4,8,9,10,11,12.

[15] F.Scholze, H.Henneken, P.Kuschnerus, H.Rabus, M.Richter, and G.Ulm, \Determination of the electronhole

pair creation energy for semiconductors from the spectral responsivity of photodiodes," Nuclear Inst. and

Methods in Physics Research A439 , 208{215 (2000).

[16] R.C.Alig, S.Bloom, and C.W.Struck, \Scattering by ionization and phonon emission in semiconductors,"

Phys. Rev. B 22 , 5565 (1980).

[17] P.Lechner, R.Hartmann, H.Soltau, and L.Strder, \Pair creation energy and Fano factor of silicon in the

energy range of soft X-rays," Nuclear Inst. and Methods in Physics Research A377 , 206{208 (1996).

[18] Lechner, P. and Strder, L., \Ionization statistics in silicon x-ray detectors new experimental results,"

Nuclear Inst. and Methods in Physics Research A354 , 464{474 (1995).

[19] G.W.Fraser, A.F.Abbey, A.Holland, K.McCarthy, A.Owens, and A.Wells, \The X-ray energy response of

silicon. part A. theory," Nuclear Instruments and Methods in Physics Research A350 , 368{378 (1994).

[20] A.Owens, G.W.Fraser, A.F.Abbey, A.Holland, K.McCarthy, A.Keay, and A.Wells, \The X-ray energy re-

sponse of silicon B. measurements," Nuclear Instruments and Methods in Physics Research A382 , 503{510

(1996).

[21] Mazziotta, M., \Electronhole pair creation energy and Fano factor temperature dependence in silicon,"

Nuclear Instruments and Methods in Physics Research A584 , 436{439 (2008).

Figure 10. Parallel CTE plot for segment 0 and 12.

[22] I.V.Kotov, J.Haupt, P.Kubanek, P.O'Connor, and P.Takacs, \X-ray analysis of fully depleted CCDs with

small pixel size," Proc. SPIE 9154 , 9154H{6 (2014).

[23] I.V.Kotov, J.Frank, A.I.Kotov, P.Kubanek, P.O'Connor, V.Radeka, and P.Takacs, \Charge Di?usion Mea-

surement in Fully Depleted CCD using Fe-55 X-rays," Nuclear Inst. and Methods in Physics Research A695 ,

296{297 (2012).

[24] I.V.Kotov, A.I.Kotov, J.Frank, P.Kubanek, P.O'Connor, V.Radeka, and P.Takacs, \X-ray analysis of fully

depleted CCDs with small pixel size," Proc. SPIE 8453 , 84531B (2012).

[25] I.V.Kotov, A.I.Kotov, J.Frank, P.O'Connor, V.Perevoztchikov, and P.Takacs, \CCD Base Line Subtraction

Algorithms," IEEE Trans. Nucl.Science 57, No.4 , 2200 { 2204 (2010).

[26] I.V.Kotov, A.I.Kotov, J.Frank, P.Kubanek, M.Prouza, P.O'Connor, V.Radeka, and P.Takacs, \Study of

pixel area variations in fully depleted thick CCD," Proc. SPIE 7742 , 774206 (2010).

[27] I.V.Kotov, P.O'Connor, and Murray, N., \Pocket pumped image analysis," Nuclear Inst. and Methods in

Physics Research A787 , 26{33 (2015).