The SPD geometry in AliRoot

1
The SPD geometry in AliRoot

• Alberto Pulvirenti
• University INFN Catania
• In collaboration with
• Domenico Elia (INFN Bari)

3 Convegno Nazionale sulla Fisica di
ALICE Frascati LNF, 14 November 2007

• Outline
• Implementation
• Volumes Displacement
• Material budget estimates
• Cables and sevices on the cones
• Outlook

2
Advantages of ROOT geometry modeling

• Modeler independent from transport code (GEANT3,
  FLUKA, )
• geometry is implemented once for all transport
  engines
• easy to be interfaced with the virtual generic
  simulation engine (TVirtualMC)
• easy to swtich among different transport codes
• Geometry built with ROOT classes
• reusability for reconstruction
• easy to implement (mis)alignment of modules

3
Components of a TGeo geometry

• TGeoMedium
• a tracking medium (material physical status)
• TGeoVolume
• a block of material, which represents a part of
  the detector
• a box which contains several sub-volumes, in
  order to be able to replicate a composite
  structure made of several parts
• TGeoVolumeAssembly
• a virtual space with several volumes inside
• useful to manage situations where a group of
  volumes can superimpose on another group of
  volumes

4
Implementation philosophy

• Multi-level implementation
• all groups of volumes which are replicated many
  times in the whole detector are inserted into an
  upper level container
• which will be a TGeoVolume or TGeoVolumeAssembly
  depending on how its components are displaced in
  space
• Advantages
• reduces the elements to be checked in case that
  corrections are needed
• logic of the implementation is more easily
  readable and followable

5
Half-stave architecture
Aluminum-polyimide multi-layer bus to connect the
MCM and FE chip
Aluminum-polyimide grounding foil (25 50 µm
thick) with 11 windows to improve the thermal
coupling

• 2 Ladders consisting of
• pn silicon sensor matrix 200 µm thick with 40960
  pixels arranged in 256 rows and 160 columns
• 5 FE chip Flip-chip bonded to the sensor through
  Sn-Pb bumps single cell
  dimensions 50 µm (r?) x 425 µm (z)

Multi-chip module (MCM) to configure and read-out
the half-stave
6
Implementation levels
Sensor
Chips
Bumps
Ladder
Alignable Volumes
BASE
Base
Resistors
Pixel Bus
Box volume containing the grounding foil and
the ladders.
Pt1000
Capacitors
Kapton
Glues
HALF STAVE
Grounding Foil
Al
Grease
MCM base
Uppermost level Implemented as an assembly, to
avoid some overlaps on the sector.
MCM
MCM Cover
Chips inside MCM
Pixel Bus Extender
Extenders
MCM Extender
Thin cables which go from inside to outside the
sensitive area of the SPD.
7
Stave architecture

• 1 Stave 1 left half-stave 1 right
  half-stave

LEFT-type half-stave
RIGHT-type half-stave
C - side
A - side
Example couple of half-staves on the outer layer
z
x
8
Ladder

• 1 sensor 5 chips 32 bump-bondings ? one
  single container

• bump bondings implemented in stripes (1 x
  column)? of
• 0.042 mm width
• 0.013 mm thickness

guard ring around the sensor
9
Grounding foil

• Complicated shape with holes inside
• holes are filled with thermal grease
• Cosisting in two layers (kapton aluminum)
• Small differences in size

10
Pixel bus
Pt1000s (one per chip)
Big resistors and capacitors in correspondence of
the end of each ladder
11
Half-stave assembly
Needed some room for movement of ladders and
half-staves to implement misalignment this could
cause an overlap of volumes.
SOLUTION reduce glue layer thickness to leave
some free space around the ladder and between
GF and support, without changing the spacing
between components
Pixel Bus
Glue
Glue
Ladder
Glue
Glue
Grounding Foil
Glue
Glue
CARBON FIBER SUPPORT
12
Pixel bus extenders (by R. Vernet)

• Implemented as folded foils
• Volumes must intrude in each other ?
  TGeoVolumeAssembly

13
MCM

• Thin integrated circuit Chips Thick cover

14
Clips

• Component on 3 over the 4 staves lying on layer 2

15
Placement on sector

• Use reference points in the support placement
  planes

16
Final appearance
17
Tests (1)what is done on the way along
implementation

• Fix coding conventions
• usually done before committing on CVS (by me or
  Massimo Masera)
• Remove overlaps
• the volumes must not overlap with each other,
  because this can cause the transport of particles
  to get confused and return meaningless data
• a ROOT facility allows to check overlaps by
  sampling
• points are generated randomly in the volume of
  the complete geometry
• for each point it is checked if it belongs to
  more than 1 volume
• an alert is raised when this happens ? an overlap
  is present somewhere
• Event generation in AliRoot with new geometry
• check execution CPU time to detect anomalous
  increases due to slow geometry creation (e.g. due
  to a too large amount of volumes)
• make sure that no run-time errors are raised

18
Tests (2)what will be done with dedicated tests

• Check materials used for implementation
• for objects present in the old geometry
• translate their definition in TGeo language (done
  by Ludovic Gaudichet)
• for new objects only present in new geometry
• when possible, reuse old definition (chips,
  silicon, )
• when not possible, a dedicated study is required
  to define new materials
• Radiation Length maps
• comparison between old and new geometries
• comparison with computations from technical
  details

19
Calculated material budget (as implemented)
1.090
INNER LAYER
OUTER LAYER
1.197
0.530
TH. SHIELD
20
Cables and services on the cones what is there
21
Summary outlook

• The new TGeo package allows a definition of a
  detector geometry decoupled from transport code
• ease switching among different transport codes
• ease interacting with geometry also in
  reconstruction
• Implementation of SPD has started since several
  months
• Implemented part is almost equivalent to the
  actual geometry
• work started for implementation of other
  components on cones
• Testing of new geometry on the way
• Preliminary tests being done for radiation length
  maps and event generation
• Test on materials is going to start

22
(No Transcript)
23
X0 map comparison with old geometry
(very preliminary result) with geantinos
New
Layer 2 R 6.5 ? 7.5 cm
Z (cm)
F (deg)
Old
Z (cm)
F (deg)
24
X0 map comparison with old geometry
(very preliminary result) with geantinosdiffer
ence
Layer 2 R 6.5 ? 7.5 cm
Z (cm)
F (deg)
25
Cables and services on the cones some estimates
1. Extenders
12 per each (half-)sector - 6 x
pixel-bus - 6 x MCM
X/Xo 6(0.11/285.70.14/14.3) 6
X/Xo 6(0.10/285.70.10/14.3) 4.4
162o
18o
126o
54o
90o
26
Cables and services on the cones some estimates
2. Optical patch-panels
X/Xo 4/27.0 14.8
10 in total, 1 per each (half-)sector
90o
62o
115o
162o
18o
100mm
yz
50mm
Aluminium
50mm
4mm
xy
50mm
27
Cables and services on the cones some estimates
3. Plates holding the extenders
X/Xo 5/223.5 2.2
10 in total, 1 per each (half-)sector at ?
middle of the cone
200mm
50mm
30mm
50mm
2mm
5mm
Carbon fiber
xy
xz
28
Cables and services on the cones some estimates
4. Tubes for detector cooling
X/Xo 2/17.2 11.6
Flexible parts
1mm
Inox
6mm
Central (rigid) part may be assumed with same
diameter but thinner walls (0.3mm)
29
Cables and services on the cones some estimates
5. Other materials
4 of these capillars on the cone
PHYNOX ducts 2.6mm xternal diameter 0.040mm thick
walls X/X0 0.080/16.1 0.5
CARBON FIBER holding plate 300mm x 30mm 0.3mm
thick X/X0 0.3/223.5 0.1
Cu/Ni (30/70) ducts 1.85mm external diameter
0.35mm thick walls 200mm length X/X0 0.7/14.3
4.9
Optical fibers (quartz) 18 per
halh-sector 9/125/900mm buffered fibers 1 fiber
X/X0 0.9/100 1.6 18 fiber pockets ?
X/Xo 4.5/100 4.5
4.5mm