BTeV – A Dedicated B Physics Experiment at the Tevatron Collider
Klaus Honscheid
The Ohio State University
DPF 2002, May 24, 2001
Klaus Honscheid 1
Ohio State University
BTeV Collaboration
Belarussian State- D .Drobychev,
A. Lobko, A. Lopatrik, R. Zouversky
UC Davis - P. Yager
Univ. of Colorado-J. Cumalat
Fermi National Lab
J. Appel, E. Barsotti, C. N. Brown ,
J. Butler, H. Cheung, G. Chiodini,
D. Christian, S. Cihangir, R. Coluucia,
I. Gaines, P. Garbincius, L. Garren,
E. Gottschalk, A. Hahn, P. Kasper,
P. Kasper, R. Kutschke, S. Kwan,
P. Lebrun, P. McBride, M. Votava,
M. Wang, J. Yarba
Univ. of Florida at Gainesville
P. Avery
University of Houston
M. Bukhari, K. Lau, B. W. Mayes,
V. Rodriguez, S. Subramania
Illinois Institute of Technology
R. A. Burnstein, D. M. Kaplan,
L. M. Lederman, H. A. Rubin,
C. White
Univ. of Illinois- M. Haney,
D. Kim, M. Selen, J. Wiss
Univ. of Insubria in ComoP. Ratcliffe, M. Rovere
INFN - Frascati- M. Bertani,
D. Morozov, L. Nogach, K.
Shestermanov, L. Soloviev, A.
Uzunian, A. Vasiliev
University of Iowa
C. Newsom, & R. Braunger
University of Minnesota
L. Benussi, S. Bianco, M. Caponero,
V. V. Frolov, Y. Kubota, R. Poling,
F. Fabbri, F. Felli, M. Giardoni,
& A. Smith
A. La Monaca, E. Pace, M. Pallotta,
Nanjing Univ. (China)
A. Paolozzi
T. Y. Chen, D. Gao, S. Du, M. Qi,
INFN - Milano - G. Alimonti, P. B. P. Zhang, Z. Xi Zhang, J. W.
D’Angelo, L. Edera, S. Magni, D.
Zhao
Menasce, L. Moroni, D. Pedrini,
Ohio State University
S.Sala, L. Uplegger
K. Honscheid, & H. Kagan
INFN - Pavia - G. Boca,
Univ. of Pennsylvania
G. Liguori, & P. Torre
W. Selove
INFN - Torino –
Univ. of Puerto Rico
R. Arcidiacono, S. Argiro,
A. Lopez, & W. Xiong
S. Bagnasco, N. Cartiglia, R. Cester,
Univ. of Science & Tech. of
F. Marchetto, E. Menichetti,
China - G. Datao, L. Hao, Ge Jin,
R. Mussa, N. Pastrone
T. Yang, & X. Q. Yu
IHEP Protvino, Russia
Shandong Univ. (China)- C.
A. Derevschikov, Y. Goncharenko, F. Feng, Yu Fu, Mao He, J. Y. Li,
L. Xue, N. Zhang, & X. Y. Zhang
V. Khodyrev, V. Kravtsov, A.
Meschanin, V. Mochalov,
Southern Methodist
University - T. Coan
SUNY Albany - M. Alam
Syracuse University
M. Artuso, S. Blusk, C.
Boulahouache, O. Dorjkhaidav, K.
Khroustalev, R.Mountain, N. Raja, T.
Skwarnicki, S. Stone, J. C. Wang
Univ. of Tennessee
T. Handler, R. Mitchell
Vanderbilt University
W. Johns, P. Sheldon, K. Stenson, E.
Vaandering, & M. Webster
Univ. of Virginia:
M. Arenton, S. Conetti, B. Cox,
A. Ledovskoy
Wayne State University
G. Bonvicini, & D. Cinabro,
A. Shreiner
University of Wisconsin
M. Sheaff
York University
S. Menary
Klaus Honscheid 2
Ohio State University
CKM, CP and all that
From Hocker et al:
Current constraints on
bd CKM triangle
Magnitudes
From Peskin:
Bd, Bs Mixing Phases
Many independent
checks are needed
Klaus Honscheid 3
Ohio State University
Some crucial measurements
Physics
Quantity
sin(2)
cos(2)
sign(sin(2))
sin()
sin()
sin()
sin()
sin(2)
sin(2)
sin(2)
cos(2)
xs
 for Bs
Decay Mode
Bo
Bo
Bo & Bo
BsDs K
B+Do K
BK
BoBsKK
BsJ/ J/
BoJ/Ks
BoKs,Ks, o
BoJ/K* & BsJ/
BsDs
BsJ/ KK Ds
Vertex
Trigger











K/  det Decay Lepsep
time  ton Id




























Klaus Honscheid 4
Ohio State University
B Physics at Hadron Colliders
• The Opportunity
– Lot’s of b’s
(a few times 1011 b-pairs per
year at the Tevatron)
– “Broadband, High Luminosity
B Factory”
(Bd, Bu, Bs, b-baryon, and Bc )
– Tevatron luminosity will
increase to at least 5x1032
– Cross sections at the LHC will
be 5 times larger
– Because you are colliding
gluons, it is intrinsically
asymmetric so time evolution
studies are possible (and
integrated asymmetries are
nonzero)
• The Challenge
– Lot’s of background
(S/N 1:500 to 1:1000)
– Complicated underlying event
– No stringent kinematic
constraints that one has at an
e+e- machine
– Multiple interactions
These lead to questions about the
triggering, tagging, and
reconstruction efficiency and the
background rejection that can be
achieved at a hadron collider
Klaus Honscheid 5
Ohio State University
The Tevatron as a b & c source
property
Luminosity
b cross-section
# of b’s per 107 sec
b fraction
c cross-section
Bunch Spacing
Luminous region length
Luminous region width
Interactions/crossing
Value
2 x 1032
100 mb
4 x 1011
10-3
>500 mb
132 ns
z = 10-20 cm
x ~ y ~ 50 mm
<2.0>
Klaus Honscheid 6
Ohio State University
Characteristics of hadronic b production
The higher momentum
b’s are at larger ’s
BTeV
Central
Detectors
Mean 
~0.75
b production peaks at large
angles with large bb correlation
BTeV


Klaus Honscheid 7
Ohio State University
A simulated BTeV Event
Klaus Honscheid 8
Ohio State University
Klaus Honscheid 9
Ohio State University
The BTeV Spectrometer (revised)
• Single Arm
• Full Pixel Detector
• Full Trigger and DAQ
Reduce costs
New baseline cost ~M$100
• Use extra bandwidth
to loosen trigger
• Better lepton id
Re-gain some performance
New baseline performance
about 70% of original twoarm spectrometer
Klaus Honscheid 10
Ohio State University
Key Design Features of BTeV
A dipole magnet (1.6 T) and 2 1 spectrometer arms
A precision vertex detector based on planar pixel arrays
A detached vertex trigger at Level I
High resolution tracking system (straws and silicon strips)
Excellent particle identification based on a Ring Imaging Cerenkov
counter.
 A lead tungstate electromagnetic calorimeter for photon and 0
reconstruction
 Excellent identification of muons
 A very high capacity data acquisition system which frees us from
making excessively specific choices at the trigger level





Klaus Honscheid 11
Ohio State University
Pixel Vertex Detector
Reasons for Pixel Detector:
• Superior signal to noise
• Excellent spatial resolution -- 5-10
microns depending on angle, etc
• Very low occupancy
• Very fast
• Radiation hard
Special features:
• It is used directly in the L1 trigger
• Pulse height is measured on
every channel with a 3 bit FADC
• It is inside a dipole and gives a
crude standalone momentum
Klaus Honscheid 12
Ohio State University
Pixels – Close up of 3/31 stations
•50mm x 400mm pixels - 30 million total
•Two pixel planes per station
(supported on a single substrate)
•Detectors in vacuum
•Half planes move together when Tevatron
beams are stable.
10 cm
Klaus Honscheid 13
Ohio State University
Pixel Readout Chip
•Different problem than LHC pixels:
132 ns crossing time (vs. 25ns)  easier
Very fast readout required  harder
7.2 mm
•R&D started in 1997
Test outputs
8 mm
•Two generations of prototype chips
(FPIX0 & FPIX1) have been designed
& tested, with & without sensors,
including a beam test (1999) in
which resolution <9m was demonstrated.
•New “deep submicron” radiation
hard design (FPIX2):Three test chip
designs have been produced & tested.
Expect to submit the final design
~Dec. 2001
Readout
A pixel
FPIX1
Klaus Honscheid 14
Ohio State University
Pixel Test Beam Results
No change
after 33 Mrad
(10 years, worst
case, BTeV)
Analog output of pixel amplifier before
and after 33 Mrad irradiation.
0.25m CMOS design verified radiation
hard with both  and protons.
Track angle (mr)
Klaus Honscheid 15
Ohio State University
Pixel detectors are hybrid assemblies
•Sensors & readout “bump bonded” to one another.
•Readout chip is wire bonded
to a “high density interconnect”
HDI
which carries bias voltages,
Sensor
control signals, and output data.
Readout chip
HDI
Wire bonds
Sensor
(5 readout chips underneath)
Micrograph of FPIX1: bump bonds are visible
Klaus Honscheid 16
Ohio State University
Physics Performance of Pixel Detector
Distribution in L/ of
Reconstructed Bs
Primary-secondary vertex separation
Minus generated.  = 138m
Mean = 44
tproper (reconstructed) - t proper (generated)
= 46 fs
Klaus Honscheid 17
Ohio State University
The BTeV Level I Detached Vertex Trigger
Three Key Requirements:
– Reconstruct every beam crossing, run at 7.6 MHz
– Find primary vertex
– Find evidence for a B decaying downstream of primary vertex
Five Key Ingredients:
– A vertex detector with excellent spatial resolution, fast readout, and
low occupancy
– A heavily pipelined and parallel processing architecture well suited to
tracking and vertex fining
– Inexpensive processing nodes, optimized for specific tasks within
the overall architecture ~3000 CPUs
– Sufficient memory to buffer the vent data while calculations are
carried out ~1 Terabyte
– A switching and control network to orchestrate the data movement
through the system
Klaus Honscheid 18
Ohio State University
Information available to the L1 vertex trigger
p
p
Klaus Honscheid 19
Ohio State University
Track segments found by the L1 vertex trigger
“entering” track segment
“exiting” track segment
Klaus Honscheid 20
Ohio State University
Block Diagram of the Level 1 Vertex Trigger
Pixel
System
Pattern
Recognition
FPGA
Associative
Memory
Average
Latency:
0.4 ms
Track
Reconstruction
Vertex
Reconstruction
DSP
Track Farm
(4 DSPs per crossing)
Average
Latency:
55 ms
DSP
Vertex Farm
(1 DSP per crossing)
Average
Latency:
77 ms
Level 1 Vertex Trigger:
FPGAs: ~ 500
Track Farm DSPs: ~ 2000
Vertex Farm DSPs: ~ 500
Level 2/3 Trigger
2500 high performance
LINUX processors
Klaus Honscheid 21
Ohio State University
Pixel Trigger Performance
Select number (N) of detached tracks
typically pt> 0.5 GeV, N = 2
Select impact parameter (b) w.r.t. primary vertex
typically (b) = 6
Options include cuts on vertex, vertex mass etc.
N=1
N=1
N=2
74%
N=2
N=3
N=3
1%
N=4
N=4
Klaus Honscheid 22
Ohio State University
Trigger Simulation Results
• L1 acceptance: 1% 2%
• Profile of accepted events
• 4 % from b quarks including
50-70% of all “analyzable” b events
• 10 % from c quarks
Level 1 Efficiency on Interesting Physics States
• 40 % from s quarks
• 45 % pure fakes
• L2/L3 acceptance: 4 %
• 4000 Hz output rate
• 200 Mbytes/s
Klaus Honscheid 23
Ohio State University
BTeV Data Acquisition Overview
Trigger/DAQ
Klaus Honscheid 24
Ohio State University
Importance of Particle Identification
Cherenkov angle vs Mom. P
Gas
Klaus Honscheid 25
Ohio State University
Ring Imaging Cherenkov (RICH)
•
•
Two identical RICH detectors
Components:
–
–
–
–
–
Gaseous Radiator (C4F10 )
Liquid Radiator (C5F12 )
Spherical mirrors
Photo-detectors (PM + HPD)
Vessel to contain gas volume
Cherenkov photons from
Bo+ and rest of the beam
collision
Klaus Honscheid 26
Ohio State University
RICH Readout: Hybrid Photodiodes (HPD)
•
Commercial supplier from Holland
(DEP)
•
A number of prototypes was
successfully produced and tested
for CMS
•
New silicon diode customized for
BTeV at DEP (163 pixels per HPD)
•
Challenges:
– high voltages (20 kV!)
– Signal ~5000e- a need low
noise electronics (Viking)
– In the hottest region up to 40
channels fire per tube
– About 2000 tubes total (cost)
Klaus Honscheid 27
Ohio State University
Use of RICH to extend Lepton Identification
•
•
While the full detector aperture is 300 mr, the acceptance of the
electromagnetic calorimeter and muon detector is only 200mr
The RICH, however, has both electron-pion and muon-pion
discrimination at low momentum. The wide angle particles are mostly
at low momentum!
This gains us a
factor of 2.4 for
single lepton id
and 3.9 if we
require both
leptons to be
identified
Klaus Honscheid 28
Ohio State University
Electromagnetic Calorimeter
The main challenges include
• Can the detector survive the high radiation environment ?
• Can the detector handle the rate and occupancy ?
• Can the detector achieve adequate angle and energy resolution ?
BTeV now plans to use a PbWO4 calorimeter
• Developed by CMS for use at the LHC
• Large granularity
Block size 2.7 x 2.7 x 22 cm3 (25 Xo)
~11000 crystals
• Photomultiplier readout (no magnetic field)
• Pre-amp based on QIE chip (KTeV)
• Energy resolution
Stochastic term 1.8%
Constant term 0.33%
• Position resolution
 x  3.32 mm / E  0.28 mm
Klaus Honscheid 29
Ohio State University
PbWO4 Calorimeter Properties
Property
Density(gm/cm2)
Radiation Length(cm)
Interaction Length(cm)
Light Decay time(ns)
(3components)
Value
8.28
0.89
22.4
5(39%)
15(60%)
100(1%)
Refractive index
2.30
Max of light emission 440nm
Temperature
Coefficient (%/oC) -2
Light output/Na(Tl)(%) 1.3
Light output(pe/MeV)
into 2” PMT
10
Property
Transverse block size
Block Length
Radiation Length
Front end Electronics
Inner dimension
Energy Resolution:
Stochastic term
Constant term
Spatial Resolution:
Value
2.7cm X 2.7 cm
22 cm
25
PMT
+/-9.8cm (X,Y)
1.8%
0.33%
 x  3.32 mm / E
 0.28 mm
Outer Radius
Total Blocks/arm
140 cm--215 cm
$ driven
11,500
Klaus Honscheid 30
Ohio State University
Electromagnetic Calorimeter Tests
Block from China’s Shanghai
Institute
5X5 stack of blocks from Russia ready
for testing at Protvino
• Lead Tungstate Crystals from Shanghai, Bogoroditsk, other vendors
• Verify resolution, test radiation hardness (test beam at Protvino)
• Test uniformity
Klaus Honscheid 31
Ohio State University
Preliminary Testbeam Results
• Resolution (energy and
position) close to expectations
• Some non-uniformity in light
output
• more Radiation Hardness studies
Energy Resolution
Position Resolution
Klaus Honscheid 32
Ohio State University
EM calorimetry
*
CLEO
barrel
e=89%
radius
Klaus Honscheid 33
Ohio State University
Muon System
•
Provides Muon ID and Trigger
– Trigger for interesting physics
states
– Check/debug pixel trigger
•
fine-grained tracking + toroid
•
– Stand-alone mom./mass trig.
– Momentum “confirmation”
Basic building block: Proportional tube
“Planks”
~3 m
toroid(s) / iron
track from IP
2.4 m
half
height
Klaus Honscheid 34
Ohio State University
Physics Reach (CKM) in 107 s
Reaction
Bo+-
Parameter
Error or (Value)
14,600
3
Asymmetry
0.030
300
7500
7

8o
445
168,000
10
sin(2)
0.017
3000
59,000
3
170
1
4.5
Bs Ds KBoJ/KS
S/B
B (B)(x10-6)
J/l+ l -
Bs Ds -
# of Events
B-Do (K+-) K-
0.17
B-Do (K+K-) K-
1.1
1,000
>10
B-KS -
12.1
4,600
1
Bo K+-
18.8
62,100
20
Bo+-
28
5,400
4.1
Booo
5
780
0.3
BsJ/
330
2,800
15
BsJ/
670
9,800
30
J/m+m-
xs
(75)

13o
<4o +



theory errors
~4o
0.024
Klaus Honscheid 35
Ohio State University
Concluding Remarks
PAC Recommendation
“ … BTeV has designed and prototyped an
ambitious trigger that will use B decay
displaced vertices as its primary criterion.
This capability, together with BTeV’s excellent
electromagnetic calorimetry and particle ID
and enormous yields, will allow this
experiment to study a broad array of B and Bs
decays. BTeV has a broader physics reach
than LHCb and should provide definitive
measurements of CKM parameters and the
most sensitive tests for new physics in the
flavor sector.”
C0 Hall at the Tevatron
Klaus Honscheid 36
Ohio State University
Additional Transparencies
Klaus Honscheid 37
Ohio State University
Changes wrt Proposal in Event Yields & Sensitivities
• We lost one arm, factor = 0.5
• We gained on dileptons
– From RICH ID
– in proposal we used only m+m-, now we include e+e– factor = 2.4 (or 3.9), depending on whether analysis used single
(dilepton) id
• We gained on bandwidth, factor = 1.15
• Gained on eD2 for Bs only, factor = 1.3
Klaus Honscheid 38
Ohio State University
Examples
mode
Yield
prop
BoJ/ Ks
80,500
BsJ/()
9,940
Bs Ds K-
13,100
Yield
factor
Yield
one-arm
quan/Err
prop (eD2)
quan/Err
one arm(eD2)
0.5*3.9*
1.15=2.24
168,000
sin(2)/
0.025 (0.10)
sin(2)/
0.017 (0.10)
0.5*2.4*
1.15=1.38
12,600
sin(2)/
0.033 (0.10)
sin(2)/
0.024 (0.13)
0.5*1.15
=0.58
7,500
/
6o (0.10)
/
8o (0.13)
Klaus Honscheid 39
Ohio State University
Reconstructed Events in New Physics Modes:
Comparison of BTeV with B-factories
Mode
BTeV (107s)
Yield
Tagged
B-fact (500 fb-1)
S/B
Yield
Tagged
-
-
S/B
BsJ/()
12650
1645
>15
B-K-
11000
11000
>10
700
700
4
BoKs
2000
200
5.2
250
75
4
BoK*m+m-
2530
2530
11
~50
~50
3
Bs m+m-
6
0.7
>15
0
Bom+m-
1
0.1
>10
0
~108
~108
large
8x105
D*++Do, DoK+
8x105
large
Klaus Honscheid 40
Ohio State University
Comparisons with LHCb
• Advantages of LHCb
 b 5x larger at LHC, while t is only 1.6x larger
– The mean number of interactions per beam crossing is 3x lower at
LHC, when the FNAL bunch spacing is 132 ns
• Advantages of BTeV (machine specific)
– The 25 ns bunch spacing at LHC causes increased electronic noise.
It makes 1st level detached vertex triggering more difficult; in fact
LHCb only does vertexing in level 2
Klaus Honscheid 41
Ohio State University
Comparisons with LHCb II
– The 7x larger LHC beam energy causes problems: much
larger range of track momenta that need to be analyzed and
large increase in track multiplicity, which causes triggering
and tracking problems
– The long interaction region at FNAL, =30 cm compared
with 5 cm at LHC, somewhat compensates for the larger
number of interactions per crossing, since the interactions
are well separated
– There is an ion-trapping problem that stops them from
moving their silicon closer than ~1 cm (compared to 6 mm
for BTeV)
Klaus Honscheid 42
Ohio State University
Comparisons with LHCb III
• Advantages of BTeV (detector specific)
– BTeV has vertex detector in magnetic field which allows rejection of high
multiple scattering (low p) tracks in the trigger
– BTeV is designed around a pixel vertex detector which has much less
occupancy, and allows for a detached vertex trigger in the first trigger
level.
• Important for accumulation of large samples of rare hadronic decays
and charm physics.
• Allows BTeV to run with multiple interactions per crossing, L in
excess of 2x1032 cm-2 s-1
– BTeV will have a much better EM calorimeter
– BTeV is planning to read out 5x as many b's/second
Klaus Honscheid 43
Ohio State University
Specific Comparisons with LHC-b
Yields in important final states
Mode
BR
Bs Ds K-
3.0x10-4
Bo+-
2.8x10-5
Booo
0.5x10-5
Yield
BTeV
7530
S/B
Yield
LHC-b
S/B
7
7660
7
5400
4.1
2140
0.8
776
0.3
880
not
known
Klaus Honscheid 44
Ohio State University
Fundamentals: Decay Time Resolution
• Excellent decay time
resolution
– Reduces background
– Allows detached vertex
trigger
• The average decay distance and
the uncertainty in the average
decay distance are functions of B
momentum:
<L> = ctB
= 480 mm x pB/mB
from b
L/
CDF/D0
region
direct 
L/
LHC-b
region
Klaus Honscheid 45
Ohio State University
Trigger
Klaus Honscheid 46
Ohio State University
Klaus Honscheid 47
Ohio State University
Klaus Honscheid 48
Ohio State University