Radiation Damage Studies of Vertex Detector CCDs

Jim Brau, Amsterdam, April 2, 2003

1

Nikolai Sinev and Jim Brau University of Oregon

April 2, 2003

Radiation Damage Studies of Vertex Detector CCDs

• First studies of radiation damage to spare CCD ladders of the SLD VXD3 were reported in– IEEE Trans. Nucl. Sci. 47, 1898 (2000)

• Now we are renewing these studies with– passive annealing investigation (of old damage)– post-mortem measurements of damage in VXD3 itself– new exposures with electrons


2

Radiation Hardness

Surface Damage from ionizing radiationhard to > 1 Mrad (acceptable for LC)(however, SLD VDX3 damage!)

Bulk Damageresults in loss of charge-transfer efficiency (CTE)

ionizing radiationdamage suppressed by reducing the operating temperature

hadronic radiation (neutrons)damage clusters complexes

Temperature (K)100 140 180 220 260 300C

harg

e T

rans f

er

Ineff

.

MS Robbins


3

SLD Experience during VXD3 commissioning,

•An undamped beam was run through the detector, causing radiation damage in the innermost barrel. •The damage was observed as the detector was operating at an elevated temperature (220 K). •Reducing to 190 K ameliorated the damage

VXD3 Experience on Radiation Damage

Temperature (K)100 140 180 220 260 300C

harg

e T

rans f

er

Ineff

.


4

Background estimates for the next Linear Collider have varied from 107 n/cm2/year to 1011 n/cm2/year

- 2.3 x 109 n/cm2/year (Maruyama)

Expected tolerance for CCDs in the range of 109-10

Increase tolerance to neutrons can be achieved throughimprove understanding of issues and sensitivityengineering advances

flushing techniquessupplementary channelsbunch compression & clock signal optimizationother ideas

Neutron Damage


5

Theory of Bulk DamageThe most important radiation damage in CCDs caused by heavy particles is displacement in the bulk silicon. 1 MeV neutrons can transfer up to 130 keV to PKA. Only 15 eV is needed to displace an atom from the lattice.

Example of simulated tracks of knock-out silicon atoms from a primary knock-out energy of 40 keV. (V.A.J.Van Lint, NIM A253, 453 (1987).)

Vacancy (V) and interstitial silicon (I) pairs are created as a result of atom displacement. More than 90% of such pairs recombine immediately. Those which are not recombined diffuse until they form complexes of two or more vacancies (V2 or V3) or vacancy-impurity (VP, V2O and so on). Such complexes are usually not mobile. Some of them are able to bind electrons, and the bound energy for some of these is about 0.35 - 0.5 eV below the conduction band. These may act as electron traps when empty. If the bound energy is close to the conduction band, (shallow traps) the lifetime of the bound state is so short, that the trapped electron will be released quickly and re-join the charge packet before the packet passes through the trap region. In this case no charge transfer inefficiency will be introduced by the defect.


6

Theory of Bulk Damage (cont.)

However, for the deeper levels (close to 0.5 eV below the conduction band) the lifetime of the bound state, which is:

is larger than the inter-pixel transfer time , so trapped electrons are removed from the charge packet and released after the packet passes through the trap region. This leads to charge transfer inefficiency. Such inefficiency may be cured, however, by cooling the CCD to a low enough temperature, that the lifetime of the bound electrons in the trap becomes very long, so that the filled traps remain occupied when the next charge packet passes. Filled trap can't capture more electrons, so this trap will not lead to charge transfer inefficiency.

cnnn

kTEE

N

e trc

/)(


7

R&D Plan

• 1998-99 - first neutron damage studyIEEE Trans. Nucl. Sci. 47, 1898 (2000)

• 2003 - investigation of status of old irradiations (passive annealing)

- study of the nature of the damage in SLD VXD3 (in progress)

- exposure of test CCD ladders to moderate energy electrons (planning)


8

History of Neutron ExposuresNov 98 ~ 2 109 n/cm2 room temperature

Pu(Be)<En> 4 MeV

Dec98-Jan 99 Annealing study 100 C for 35 days Mar 99 ~ 3 109 n/cm2 room temperature

reactor neutrons<En> 1 MeV ( ~1 109 n/cm2 lower energy)

Apr 99 ~ 1.5 109 n/cm2 dry ice cooled (~190K)reactor neutrons<En> 1 MeV ( ~1 109 n/cm2 lower energy)

Total exposure ~ 6.5 109 n/cm2

mix of source and reactor

* UC Davis (G. Grim et al)


9

Images of damaged sites

Neutron Damage Study

Individual traps are identified by charge losses during readoutIEEE Trans. Nucl. Sci. 47, 1898 (2000)

T=187K, after dose of 2x109 n/cm2 T=187K, after dose of 5x109 n/cm2


10

Readout timing diagram

600 e/pxl 25 e/pxlFilling traps

Clearing pixels

Filling traps

Clearing pixels Clearing pixels


11

Image of damaged sites Image of damaged sites after flushing

Neutron Damage and Amelioration Study

Basic concept demonstrated; future work will involve charge injection to keep traps filled. IEEE Trans. Nucl. Sci. 47, 1898 (2000)


12

Pulse height distribution


13

Damage ResultsDefect Results from Exposures

# defect (> 6 e-) # defect (>20e-) 800,000 pixels 800,000 pixels

Prior to exposure 125 24

Nov 98 exposure 916 160 ~ 2 109 n/cm2

source

Mar 99 exposure 5476 442 + ~ 3 109 n/cm2

reactor

Apr 99 exposure 7036 298 + ~ 1.5 109 n/cm2

reactor * this surprising decrease is not understood

Signal Loss Results from Exposures

Exposure(109 n/cm2) ~ 2 ~ 6.5

T = 185K, cluster sum 4.05% 29.1% no flushing light

T = 185K, cluster sum 1.5% 18.0%

with flushing light

T = 178K 11.0%

Note () - flush is only partially effective due to required delay between flush and readout (1 second)In LC detector – better resultexpected


14

Measurements repeated in 2003

Damage sites remain


15

Comparison of loss 1999/2003

Electrons (1999)Electrons (1999)

Electrons (2003)

Electrons (2003)


16

Post-mortem tests of VXD3

• Have removed VXD3 from SLD and will do measurements to establish the nature of the damage


17

Summary of new studies1. Test stand for testing radiation damage of the CCDs has

been fully operational for about 2 months now.2. The 1998-1999 measurements of neutron induced

radiation damage has been confirmed.3. Some of the damage (~30%) has annealed in the 4 years,

during which the CCDs were stored at room temperature.4. We are now collecting high statistics on the damage in

these detectors, after which we plan to irradiate the same CCDs with relatively high energy (tens of Mev) electrons, to compare the characteristics of damage inflicted by electrons and neutrons.We are looking for the beam to do the electron irradiation.

5. VXD3 detector has been extracted from SLD and is waiting of disassembling.


18

Conclusion


19

Extras


20

Vertex Detectors

• Design CCD’s for– Optimal shape ~2 x 12 cm– Multiple (~20) ReadOut nodes for fast readout– Thin -≤ 100 µ– Improved radiation hardness– Low power

• Readout ASIC– No connectors, cables, output to F.O.– High reliability– Increased RO speed from SLD VXD3– Lower power than SLD VXD3


21

Vertex Detectors, continued

• Mechanical– Eliminate CCD supports, “stretch” Si.– Very thin beampipes??– Cooling

• Simulation– Quantify/justify needs

• SLD VXD3 has been removed from SLD for damage analysis of CCD’s.


22

cNNN

kTEE

N

e trc

/)(


23


24

(100o C after 2 x 109)

(days)


25

Documents

Radiation Damage Studies of Vertex Detector CCDs