Minutes of the
"4th Meeting on planning of the work on oxygenated silicon"

======================================================
The Meeting was held at CERN 15.6.2000, 14:00-17:30, Room:  40-R-C10
======================================================


====================
Contents
====================


====================
Participants  (Contents)
====================

 

Laci Andricek ATLAS-SCT, MPI Muenchen
Gennaro Ruggiers ATLAS-SCT, CERN
Olaf Krasel ATLAS-Pixel, Dortmund
Jens Wuestenfeld ATLAS-Pixel, Dortmund
Andreas Furtjes CMS, CERN
Kerstin Hoepfner CMS, Berlin
Antti Honkanen CMS, Helsinki
Mika Huhtinen CMS, CERN
Alexandre Kaminsky CMS, Padova
Kati Lassila-Perini CMS, Helsinki
Ernesto Migliore CMS, CERN
Eugene Grigoriev CMS-Pixel, Geneva
Roland Horisberger CMS-Pixel, PSI
Petra Riedler CMS-Pixel, PSI
Paula Collins LHCb, CERN
Frederic Teubert LHCb, CERN
Maurice Glaser RD48, CERN
Mike Letheren RD48/MIC, CERN

Michael Moll  

RD48, CERN

Angela Vasilescu RD48, Bucharest
Steve Watts RD48, Brunel
Karl Zankel RD48, CERN

=======================
Transparencies    (Contents)
=======================


Copies of the transparencies have been send to all participants. Those who have not attended the meeting are invited to ask for a copy (michael.moll@cern.ch ). Some transparencies are available in electronic format:


=================================
Presentation: Michael Moll - RD48    (Contents)
"Studies on the radiation hardness of oxygen
enriched silicon detectors"
=================================


I gave an overview of the results of the ROSE collaboration on oxygenated silicon. A detailed description of most of the presented data can be found in the  3rd RD48 STATUS REPORT and in the ROSE Technical Note TN/2000-03. Furthermore, some of the data can be found in the Proceedings of the 5th ROSE Workshop.

The report, the technical note and an online request form for the proceedings are available on the Web: http://cern.ch/rd48.

Summary: Key scientific results

--- macroscopic ---

* The leakage current damage parameter is material independent (no impurity, resistivity or conduction type dependence). It has been linked to defect clusters which are not affected by the material. Annealing of the leakage current is also material independent. The damage parameter and its annealing has been shown to scale with NIEL (non ionising energy loss), i.e. with-out any remaining particle or energy dependence.

* Effective doping changes can be improved by oxygenation of the material (factor 3 in the stable damage component). Such improvement is only observed when the radiation environment contains a significant charged particle component. This has been understood in terms of the production of larger numbers of isolated vacancy/interstitial pairs during charged particle irradiation.

* Lower resistivity oxygenated or standard material is beneficial for detectors that operate in a radiation environment dominated by reactor energy neutrons.

* Reverse annealing has been linked to defect clusters. After proton irradiation, this process is found to saturate at high fluence (>2 1014p.cm-2 ) for oxygenated silicon. This provides a significant safety margin. In addition, the time constant for the process is found to be a factor of about 2 larger in oxygenated material (see presentation of A.Vasilescu). This would allow detectors to remain at room temperature for longer periods during maintenance periods and thus offers a substantial safety margin.

* A macroscopic damage parameter model has been developed which can be used to predict detector parameters in a given radiation environment. This model has been used already in operational projections for major LHC experiments.

* A set of silicon diodes produced by ST Microelectronics on standard Wacker silicon with different resistivity (1 to 15Kohmcm) and different orientation (<111> and <100>) was irradiated with 24 GeV/c protons. Some of the diodes were less radiation hard than expected from previous experiments on standard silicon and some were radiation harder (almost as good as oxygenated silicon diodes). This fluctuation of the beta factor was found to be independent of the resistivity and the crystal orientation. Up to now it is not clear which material property is responsible for this behavior.
For the oxygenated silicon such a variation was so far not observed. This leads to the preliminary conclusion that the oxygenation process gives a reproducible good result with respect to the radiation hardness of the diodes while for material bought as "standard silicon" a wide variation can be possible. Even a less radiation hard behavior than expected from our previous experiments on standard silicon and published in the STATUS REPORT is possible!

-- microscopic --

* Detailed correlations have been found between microscopic defect formation and macroscopic damage parameters. 

* Charged particle irradiation produces more point defects than irradiation with reactor energy neutrons.

* Defect kinetics models and device models can predict macroscopic behavior, even for hadron irradiation. However, these models are still under discussion.

Summary: Key technological results:

* Two methods were found to highly oxygenate silicon. Firstly, at the ingot growing stage. Secondly by diffusion of oxygen into ANY wafer using a high temperature drive-in (a minimum of 16 hours at 1150C seems to be sufficient).
* This technology has been successfully transferred to several silicon detector manufacturers (SINTEF, Micron, ST-Microelectronics, CIS, IRST) and full-scale microstrip and pixel detectors have been produced.
* DOFZ wafers produce detectors which prior to irradiation are no different to those produced on standard material.
* Irradiated standard and oxygenated test structures show the same increase in interface generation current and oxide charge.

Summary: Open questions/further work

* The optimal diffusion time required to give radiation hardening needs further study.

* How reliable is the depth measurement of SIMS analysis ?

* Why are the oxygen depth profiles of some devices produced by different producers different although almost the "same" diffusion time and temperature have been used?

* The beneficial effect that oxygen has on the reverse annealing process needs more work. As this effect is crucial to the maximum maintenance period that can be used by the experiments, it needs further investigation. This work is extremely time consuming.

* The physics of bulk damage should be the same in full-scale detectors as in simple diodes. Nevertheless, bulk damage parameters should be extracted from irradiated strip detectors and compared to the well-measured parameters obtained with diodes.

* The proton-neutron puzzle: Violation of NIEL ?
The violation of NIEL by charged hadrons in oxygenated material needs further study. Testing with radiation sources that better represent the environment in the LHC experiments needs to be performed. Further experiments also on the microscopic scale are needed to support the simulations of the primary damage produced by different particles (see also presentation of Mika Huhtinen). Presently further irradiations with pions and low energetic protons are planned.

* Variation of standard material needs to be investigated. How could a radiation hardness quality assurance test look like?

* So far Hamamatsu detectors (and silicon used by Hamamatsu) have not been investigated by the ROSE collaboration.

=======================================
Presentation: Angela Vasilescu - RD48    (Contents)
"Long-Term Annealing Behaviour of Oxygenated and
Standard Diodes after 24 GeV/c Proton Irradiation"
=======================================


Angela presented a systematic investigation on the annealing behaviour of  24 GeV/c proton irradiated Sintef test structures. Several sets of oxygenated and standard devices (irradiated at the same time to a fluence of  3.2e14p/cm2)  were annealed at different temperatures (60C,70C,80C,90C,100C +50C under way).

A speciality of the oxy and std devices produced by Sintef was that in a previous experiment almost no difference in the depletion voltage was observed after irradiation with 24 GeV/c protons (measured in the minimum of the annealing curve). However, SIMS measurements clearly revealed that the oxygen concentration was about a factor of 10 higher in the oxygenated material thus ruling out the possibility of a mix-up of the materials.

Material:  Topsil <100> about 1Kohmcm
a) standard
b) oxygenated (80h - 1150C)

Devices: Sintef, ROSE-Test structure

Measurements: CV, IV - isothermal annealing at 60,70,80,90,100C

The annealing experiments exhibited differences for the oxy and std devices:
(preliminary data - experiment is still running)

- Suppressed beneficial annealing amplitude
  Na(oxy) = 1.4e12cm-3   Na(std) = 2.5e12cm-3
- Suppressed reverse annealing amplitude:
   Ny (oxy) = 8.2e12 cm-3    Ny(std) = 1.0e13cm-3
- The time constant for the reverse annealing is reduced in oxy silicon
   e.g. 60C : ty(oxy)= 2900min and ty(std) = 4800min

The difference in the time dependence of the reverse annealing effect in oxy and std silicon is found in the frequency factor rather than in the activation energy of the underlying microscopic process. For the damage projection for oxygenated silicon this implies that the reverse annealing time dependence can simply be scaled by a factor between oxy and std silicon (in other words: If the difference of the reverse annealing process is found in the frequency factor and the activation energy is the same, then for every temperature the factor between the reverse-annealing time constants of oxy and std material is the same).

=======================================
Presentation: Mika Huhtinen - CMS     (Contents)
"Simulations of NIEL for oxygenated
and standard Si"
=======================================

           Transparencies:  (ps-file, 317 kB)

Mika presented his latest results of his simulations of the defect formation during the irradiation of oxygenated and standard silicon with different particles of different energy:

1st step - Simualtion of hadronic interactions
The neutron interactions are simulated by the FLUKA event generators above 20 MeV and taken from the ENDFB-VI data for neutrons with energies below 20 MeV. For protons the elastic scattering is simulated by the optical Glauber formalism while for the inelastic reactions the FLUKA event generators are used.

2nd step - Transport of recoils
Mika uses a modified version of the TRIM85 code to generate all atomic recoils down to the threshold of 20eV. This results in a three dimensional distribution of all vacancies (V) and interstitials (I) produced by the recoils.

3rd step - Simulation of defect production within clusters
The defect migration is simulated by a random walk of all defects (interstitials diffuse 10000 times faster than vacancies). Recombination (V+I => perfect lattice structure) or VV formation (V+V => VV) can occur when defects get closer than 3 lattice constants (The recombination probability is obtained from fits to DLTS data of defect concentrations).

4th step - Simulation of defect production outside clusters
When the defects migrate out of the primary damage region (1000 A) and nearest neighbors are further away than 200A the "Davies model" is applied for simulation of the quasichemical reactions of the primary defects with the impurities in the silicon. (The simulations in step 3 and 4 include many defects (up to V6, V4O etc and I2, VP, CC, CCI, CO,...) ).

Results  (1, 10 MeV neutrons, 24 GeV/c protons):

* VV production scales with NIEL  in oxy and std silicon
*V2O production does not scale with NIEL for std silicon (5e15 O/cm3)
*V2O production scales with NIEL for oxy silicon (5e17 O/cm3)
==> V2O can not explain the observed differences in Neff of oxy and std silicon
==> For standard silicon VV seems to be the only candidate that could explain the NIEL scaling !
==> Suggestion of a new model: The macroscopic effect is produced by both the VV and the V2O
==>The occupation of the VV depends on the size of the clusters (particle dependence)
==>The occupation of the V2O accounts for the oxygen dependence of the damage
     (many free parameters but macroscopic results can be qualitatively predicted with such an approach)

Conclusions (preliminary):
-oxygen effect can be at least qualitatively explained
- NIEL scaling for alpha must be slightly violated
- there is no single microscopic reason for NIEL scaling et all

Predictions of the simulations:
- deviation from NIEL scaling for all hadrons should become very clear at fluences above 1e15
- the oxygen effect should become visible for neutrons at fluences well above 1e14
- the oxygen effect should be very pronounced for 10 MeV protons
- the alpha for 10 MeV protons should have significant deviation from NIEL scaling

=======================================
Presentation: Laci Andricek - ATLAS-SCT     (Contents)
"Suppression of Reverse Annealing in Oxygen
enriched Si-Strip Detectors"
=======================================

Laci presented an investigation on 24 GeV/c proton irradiated (3e14p/cm2) oxygenated and standard ATLAS W12 detectors produced by CIS ( 2 oxygenated (280mum, 24h 1150C), 1 standard (280mum) and 2 thin standard (260mum) detectors):

* before irradiation: no difference between standard and oxygenated detector

* after irradiation:
- Oxy/Std: Currents after irradiation are very similar and at the level expected from other irradiations
- CV measurements (7d at 25C):  30V to 40V lower depletion voltage for the oxygenated detectors.
- CV measurements (32d at 25C):  about 80V lower depletion voltage for the oxygenated detectors.

* Comparison with irradiation tests performed previously (9/98 and 11/98)  on CIS (std) detectors shows that for the standard material investigated in this experiment the depletion voltage is significantly (about 100V) lower ! (Variation of the material properties ?)

* Source measurement after 7d at 25C with fast analogue readout (SCT128a):
- for oxygenated detectors CCE plateau around 300V, about 60 V lower than for the standard detector
- oxygenated and standard detector give 92% of the signal that was measured on the unirradiated detectors

Conclusions:
* Even in the annealing state at the minimum of the depletion voltage oxygen enriched detectors give the same signal at 60 V lower voltage
* Using thin detectors, one has to got as low as 250 microns to get the same performance as an oxygen enriched detector (in this radiation field and after annealing of 7d at 25C).
* The reverse annealing is suppressed in the oxygenated detectors: The increase of Vdep between 7d and 32d at 25C is only 21V vor the oxygenated but about 65V for the standard detector.

==========================================
Presentation: Gianluigi Casse
(presented by M.Moll)    (Contents)
"Charge trapping in oxygenated and
control irradiated silicon detectors"
==========================================

Transparencies: (ps-file, 1360 kB)

Unfortunately Gianluigi could not come to the meeting. However, he sent me his transparencies and I presented them:

Experiment:
Four miniature microstrip detectors (1x1 cm2, 128 strips) produced by SINTEF were irradiated unbiased at room temperature with 24 GeV/c protons. Two pairs of non-oxygenated and oxygenated devices were irradiated to 3× 1014 cm-2 and 4× 1014 cm-2. A non irradiated non-oxygenated detector with identical characteristics has been used for reference.

Measurement:
Their 128 strips were bonded to a wide bandwidth current amplifier. The CCE was obtained by the integral of the averaged signal (from a 106Ru source) on the oscilloscope.

Results/Conclusions:
The oxygen enriched material gives a certain improvement in term of bias required for complete charge collection of heavily irradiated silicon detectors. For this experiment the improvement was ~40 volts for the detectors irradiated to 4× 1014cm-2, while no difference has been observed for the detectors irradiated to 3× 1014 cm-2.
The level of the plateau of the charge collection is the same for the oxygenated and non-oxygenated 300 micron thick detectors after the two different doses of irradiation, implying no relevant differences in the permanent charge trapping between these substrates. The improvement of the charge collection in heavily irradiated oxygenated detectors is probably mainly due to the reduction of the space charge concentration (and therefore of VFD) in respect of non-oxygenated devices.

The effect of the trapping was estimated by the ratio of the charge collection at VFD to the maximum charge. It is found that only ~72% of the charge is collected at full depletion in heavily irradiated detector.

The charge deficit after irradiation was evaluated for the two given doses by comparison to the pre-irradiation value and is ~10% after 3× 1014 cm-2 and ~23% after 4× 1014 cm-2.

=============================================
Presentation: Jens Wuestenfeld - ATLAS-Pixel (Contents)
"ATLAS Pixel Detector and oxygenated silicon"
=============================================

* Requirements:
- 3 barrel layers and 5 discs have to be equipped with 5.2m2 of silicon detectors
- The barrel layers are located at radii of 4.3cm (b-layer) 10cm (layer1) and 12cm (layer2)
- The devices will face high fluences of up to about 1e15 1/cm2 (1MeV n - equivalent)
- The damage is dominated by charged hadrons: b-layer : 82% charged hadrons/ layer1 : 70% charged hadrons

* According to these requirements oxygenated silicon is the perfect choice.
  - oxygenated silicon will be used for the detectors

* The ATLAS-Pixel Collaboration will need the ROSE collaboration in the future to solve/ work on the following topics:
- further optimization of the oxygenation process
- better understanding of the oxygenation effect
- deeper understanding of the field distribution in the detectors
- Help for the quality assurance:
  monitoring of radiation hardness, how to measure the oxygen content, development of performance tests

* The b-layer will be ordered in two years:
  - new results can still be considered !

=============================================
Presentation: Olaf Krasel - ATLAS-Pixel (Contents)
"Damage Projections for Pixel Sensors"
=============================================

        Transparencies: (pdf-file, 34 kB)

* Olaf presented a new "standard scenario" for the operation of the ATLAS-Pixel detectors:
  - 100 days beam at 0C
  -  30 days warm-up at 20C
  - 235 days storage at -10C
* Based on this scenario projections for the depletion voltage, the leakage current and the CCE were given (with damage parameters presented in the last ROSE status report).
* Furthermore, he presented projections for various operation temperatures, warm-up temperatures and warm-up times.
*The main conclusions are (taken from the transparencies):
  - ATLAS Pixel "Layer 1": Vdep and CCE allow operation at 0C
  - ATLAS Pixel "B-Layer": 5 years at nominal fluence with 0C operation temperature, 30 days warm-up to 20C, 200mum, 600V bias will result in 10000 electrons.
  - leakage current will stay within limits (<50 nA per pixel)
  - higher power consumption due to higher leakage current (about a factor 2 per 8C higher temperature)


=============================================
Comments: Roland Horisberger - CMS-Pixel (Contents)
=============================================

* CSEM has produced 24 CMS-Pixel detectors: 8 high resistivity and 8 low resistivity detectors on standard silicon and 8 oxygenated detectors. However, the production of the oxygenated detectors failed.
* Several pixel detectors with new p-stop design were irradiated at the CERN PS at the end of April and are now being tested.
* More detectors will be irradiated at the PSI irradiation in July.
* Oxygenated pixel detectors produced by Sintef have been produced for the forward pixels. Teststructures could be included in the irradiations.

=============================================
Comments: Paula Collins/Frederic Teubert - LHCb (Contents)
=============================================

- The TDR has to be written until April 2001.
- The LHCb collaboration is interested in a close collaboration with the ROSE collaboration.
-LHCb still has as a priority thin detectors, but the option to use oxygenated detectors is not excluded. Several detectors from Micron have been ordered (p on n detectors of different thickness 300 um, 220 um and 140 um), some of them oxygenated. Probably some oxygenated detectors from Hamamatsu (300 um) maybe be available and tested.
LHCb expects to reach a conclusion on the technology by the end of 2000.
Some of the results shown by ROSE, ATLAS and CMS need to be adapted to the special conditions of LHCb (mainly very non-uniform radiation). LHCb needs also to look at what is the effect on the resolution, and not only improvements on Vdep or CCE.


=============================================
Discussion/Comments/Questions  (Contents)
=============================================

* Mika: It would be very interesting to measure the radiation damage at low temperatures. Some of the annealing processes are frozen and thus it might be easier to compare the experimental results with the simulated results. Furthermore, such an experiment could give some indication whether the beneficial oxygen effect arises during the irradiation or during the following annealing.

* Roland: Is it feasible to use low energy protons for radiation tests with the purpose of a radiation quality assurance test ?

*The GDS-file of the ROSE-teststructure should be distributed (made easily available) as fast as possible to everybody producing new masks in order to have a standardized structure for radiation tests.

* Dortmund: new (simpler) gate controled diode design is available as GDS

* A standard setup and a standard measurement procedure (annealing cycle!!) for irradiation tests should be defined in order to make experimental results comparable. It would be even better to have a central place where such measurements could be performed regularely by one person.

* The discrepancies between the results obtained on full size detectors and test structures have to be understood. Irradiation/Annealing experiments on full size detectors and test structures originating from the same wafer are necessary.

* Hamamatsu detectors (and silicon bought by Hamamatsu) are widely used now in the experiments. Therefore, it is agreed that it is important that ROSE also investigates Silicon Detectors from Hamamatsu.

=============================================
Next Meeting (Contents)
=============================================

- Will be held at CERN at the end of this year.

- At the end of this year also the "6th ROSE Workshop on Radiation Hardening of Silicon Detectors" will also take place at CERN.

- The announcements for both meetings will be distributed to all persons on the mailing list in time.

The END             Top of page             Back to main page