Observation of Bs Mixing with CDF II

Alessandro CerriCERN (LBNL)

Dec 11th 2006

2

Synopsis

• Introduction– Bs mixing in the last 12 months

– Flavor physics– How the Tevatron contributes

• Detector• Benchmarks

• Bs Mixing Observation

• Is BSM physics dead?• Conclusions

3

What happened in the last year…

• Mar 2006: D0 came out with a result based on 1fb-1 (Moriond)

• Jun 2006: CDF releases a preliminary result (but not the last word) on 1fb-1

– Signal search start showing evidence– Not enough statistical power for ‘observation’ (5)– If signal is theremeasure

• Sep 2006: CDF went back and improved the analysis– Signal shows up at >5– We therefore claim observation and measurement

of

133.018.0 )(07.0)(31.17 −+

− ±=Δ pssyststatms

1)(07.0)(10.077.17 −±±=Δ pssyststatms

This last analysis will be today’s subject!

€

Δms ∈ [17,21] ps−1 @ 90% CL

4

The Flavor Sector: CKM Matrix

Quarks couple to W through VCKM: rotation in flavor space!

VCKM is Unitary

u

Wd’

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛=

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛

′′′

b

s

d

V

b

s

d

CKM

5

Last year…

TeVatron contribution is critical!

6

The Tevatron as a b factory (4s) B factories program extensive and very

successful BUT limited to Bu,Bd

• Tevatron experiments can produce all b species: Bu,Bd,Bs,Bc, B**, b, b

61@53.042.051.3 ><±±= tBpybμ o

Compare to:(4S) 1 nb (only B0, B+)–Z0 7 nb

Unfortunately–pp 100 mb

b production in pp collisions is so large (~300 Hz @ 1032 cm-2 Hz) that we could not even cope with writing it to tape!

See PRD 71, 032001 2005

7

Detector & Techniques

8

CDF and the TeVatron

Renewed detector & Accelerator chain:•Higher Luminosity higher event rateDetector changes/improvements:DAQ redesignImproved performance:Detector CoverageTracking QualityNew Trigger strategies for heavy flavors:

displaced vertex trigger

COT

Si Detector: L00,SVX II, ISL

Delivered: 1.6 fb-1

On tape : 1.4 fb-1

Delivered: 1.6 fb-1

On tape : 1.4 fb-1

Good w/o Si: 1.2 fb-1

Good w Si: 1.0 fb-1

Good w/o Si: 1.2 fb-1

Good w Si: 1.0 fb-1

TOF

9

SVT: a specialized B physics trigger

•Good IP resolution

SVT

Detector

Raw Data

Level 1

storage

pipeline:

42 clock

cycles

Level 1Trigger

L1Accept

Level 2Trigger

Level 2 buffer: 4 events

L2Accept

DAQ buffers

L3 Farm

Level 1•2.7 MHz Synch. Pipeline•5544 ns Latency•~20 KHz accept rate

Level 2• Asynch. 2 Stage Pipeline•~20 μs Latency•250 Hz accept rate

Mass Storage (30-50 Hz)

~2.7 MHz Crossing rate

396 ns clock

The CDF Trigger requirements

•As fast as possible

Customized Hardware

10

…and a successful endeavor!

•SVT is capable of digesting >20000 evts/second, identifying tracks in the silicon

~ 48 μm

Single Hit

Superstrip

Road

Dete

ctor

Laye

rs

•CDFII has been running it since day -1

•The recipe: specialized hardware

1)Clustering Find clusters (hits) from detector ‘strips’ at full detector resolution

2)Template matching Identify roads: pre-defined track templates with coarser detector bins (superstrips)

3)Linearized track fitting Fit tracks, with combinatorial limited to clusters within roads

SVT is the reason of the success and variety of B physics in CDF run II

11

Benchmarks

12

Knowledge of non-(4s)-produced b (PDG’04)

13

Measure: Branching Ratios

Hep-ex/0508014

Hep-ex/0601003

http://www-cdf.fnal.gov/physics/new/bottom/050310.blessed-dsd/

http://www-cdf.fnal.gov/physics/new/bottom/050407.blessed-lbbr/lbrBR_cdfpublic.ps

http://www-cdf.fnal.gov/physics/new/bottom/050310.blessed-dsd/

First-time measurement of many Bs and b Branching Fractions

Hep-ex/0502044

14

Lifetimes: fully reconstructed hadronic modes

(B+) = 1.661±0.027±0.013 ps (B0) = 1.511±0.023±0.013 ps(Bs) = 1.598±0.097±0.017 ps

•Testbed for our ability to understand trigger biases

•Large, clean samples with understood backgrounds

•Excellent mass and vertex resolution

•Prerequisite for mixing fits!

http://www-cdf.fnal.gov/physics/new/bottom/050303.blessed-bhadlife/

KK

Systematics (μm)

Effi

cien

cy

(AU

)

Proper decay length (mm)

43210

15

Hadronic Lifetime Results

ModeLifetime [ps](stat. only)

B-D0 - 1.638 0.017

B0D- + 1.508 0.017

BsDs () 1.538 0.040 World Average:

B+ 1.653 0.014 ps-1B0 1.534 0.013 ps-1

Bs 1.469 0.059 ps-1

Excellent agreement!

~3000 candidates

t

Bxyxy

P

mLLct ==

βγ t

P

xy

Lct

PLcttxy

⊕=

16

lDs Lifetime Results

Lifetime (ps)

Bs:Ds 1.51±0.04 stat. only

Bs:Ds K*K 1.38±0.07 stat. only

Bs:Ds 1.40±0.09 stat. only

Bs combined 1.48±0.03 stat. only

MCT

VISt

VISt

Bxy

P

P

P

mLct =

• lifetimes measured on first 355 pb-1 • compare to World Average: Bs: (1.469±0.059) ps

KPLctK

t

P

xy

Lct txy ⊗⊕=

K K

ll

Ds

Bs

17

Bs Mixing

18

Why so much fuss around Δms?

γ β

( )ηρ ,∝

tdV∝ubV∝

cbV∝•Vtd is derived from mixing effects

•QCD uncertainty is factored out in this case resorting to the relative Bs/Bd mixing rate (Vtd/Vts)

•Beyond the SM physics could enter in loops!

19

B production at the TeVatron

•Production: ggbb

•NO QM coherence, unlike B factories

•Opposite flavor at productionone of the b quarks can be used to tell the flavor of the other at production

•Fragmentation products have some memory of b flavor as well

20Δms [ps-

1]

A

Bs Mixing 101

Δms>>Δmd

•Different oscillation regime Amplitude Scan

B lifetimePerform a ‘fourier

transform’ rather than fit for frequency

Bs vs Bd oscillation

Nunmix-NmixNunmix+Nmix

A= cos(Δm t)

21

Amplitude Scan: introduction

Just an example: Not based on real data!

•Mixing amplitude fitted for each (fixed) value of Δm

•On average every Δm value (except the true Δm) will be 0

•“sensitivity” defined for the average experiment [mean 0]

•The actual experiment will have statistical fluctuations

•Actual limit for the actual experiment defined by the systematic band centered at the measured asymmetry

•Combining experiments as easy as averaging points!

Is this an effective tool to search for a signal?

22

Bs Mixing Ingredients

( )222

2

s tmS D SSignificance e

S B

σε Δ−

=+

Event yield

Flavor taggingSignal-to-noise

Proper time resolution

23

Flavor Taggers:•Opposite Side

•Lepton (e,μ)

•Average charge

•Kaon (bcs)

•Same Side:

•Kaon (hadronization)

Flavor Tagging

Several methods, none is perfect !!!

Fragmentation

product

B meson

Reconstructed decay“Same Side”

( )222

2

s tmS D SSignificance e

S B

σε Δ−

=+

Nright-Nwrong

Nright+Nwrong

D=AmplitudeDAmplitude

24

Bs Mixing: tagging performanceM

easu

red f

rom

Bd/B

s data

~5% of the Events are effectively used!

DD2 Semileptonic

(%)

Muon 4.60.0 34.70.5 0.580.02

Electron 3.20.0 30.30.7 0.290.01

JQT 95.50.1 9.70.2 0.900.03

Kaon 18.10.1 11.10.9 0.230.02

Total OST 95.80.1 12.70.2 1.540.04

Total OST (NN) ~96% ~13.7% 1.820.04

SSKT ~67% 26.76.04.81.2

(3.50.8 in hadr.)

25

Bs Mixing Ingredients: ct

( )222

2

s tmS D SSignificance e

S B

σε Δ−

=+

Proper time resolution

26

Proper time resolution

( ) T

xy

pBct

xy xy B

TL

T T

L L mct

m

p ppct

σσ σ

βγ

⎛ ⎞= ⊕ ⎜ ⎟

⎝=

⎠=

~0.5%

BsDs

KK

Ds

Bs

Semileptonic modes: momentum uncertainty

Fully reconstructed: Lxy uncertainty

BsllDs K K

ll

Ds

Bs

mcst

st

st

xyB

BP

lDP

lDP

Lmct

)(

)(

)(⋅=

~15%

Kt

PL

t

Bct

Pct

Pm t

xy

27

Mixing in the real world

Proper time resolutionFlavor tagging power

( )

BS

Se

DS ctsm

+=

Δ−

22

2

2

1 σε

σ

28

Bs Mixing: CDF semileptonic

Δms>16.5 @ 95% CL

Sensitivity: 19.3 ps-1

Reach at large Δms limited by incomplete reconstruction (ct)!

BsDsl Yield s/b

Ds ~29000 ~2:1

DsK*K ~22000 ~1:2

Ds ~10000 ~1:5

Hep-ex/0609040

29

Bs Mixing: CDF hadronic

Δms>17.1 @ 95% CL

Sensitivity: 30.7 ps-1

This looks a lot like a signal!

BsDs Yield s/b

Ds 2000

~11:1

Ds (partial)

3100

3:1

DsK*K 1400

~2:1

Ds 700 ~2:1

Hep-ex/0609040

BsDs Yield s/b

Ds 700 ~3:1

DsK*K 600 ~1:1

Ds 200 ~3:1

Total: ~8700 events!

30

Bs Mixing: combined CDF result

05.6)75.17(/ 1 ==Δ −psmA s

Δms> 17.2 ps-1 @ 95% CL

Sensitivity: 31.3 ps-1

Minimum: -17.26What is the probability

for background to mimic this?

( )⎥⎦

⎤⎢⎣

⎡==

)0(

1log

AL

AL

Hep-ex/0609040

Develop a sound statistical approach -prior to opening the box-to assess statistical significance!

31

Likelihood Ratio

gaussian 4.5108350000

28 8 ⇒×≈≈ −p

Hep-ex/0609040

32

Likelihood Ratio

Δms = 17.77 ± 0.10(stat) ± 0.07 (syst) ps-1

combined likelihoods from hadronic and semileptonic channels

•Combined hadronic+semileptonic

likelihoods gives 5 significance•Parabolic fit to minimum yields:

•the measurement is very precise!

(~2.5%)

Hep-ex/0609040

33

Systematic Uncertainties I

• Mostly related to absolute value of amplitude, relevant only when setting limits – cancel in A/ A, folded in confidence calculation for observation

– systematic uncertainties are very small compared to statistical

Hadronic Semileptonic

34

Systematic Uncertainties II: Δms

• systematic uncertainties from fit model evaluated on toy Monte Carlo

• have negligible impact

• relevant systematic unc. from lifetime scale

Syst. Unc

Fitting Model < 0.01ps-1

SVX Alignment 0.04 ps-1

Track Fit Bias 0.05 ps-1

PV bias from tagging

0.02 ps-1

Total 0.07 ps-1

All relevant systematic uncertainties are common between hadronic and semileptonic samples

35

Δms and Vtd

• compare to Belle bsγ (hep-ex/050679):|Vtd| / |Vts| = 0.199 +0.026 (stat) +0.018 (syst)

inputs: m(B0)/m(Bs) = 0.9830 (PDG 2006) = 1.21 +0.047 (M. Okamoto, hep-lat/0510113) Δ md = 0.507 ± 0.005 (PDG 2006)

-0.035

|Vtd| / |Vts| = 0.2060 0.0007(exp) +0.0081 (theo)

-0.025 -0.015

-0.0060

36

Δms & CKM (CKMFitter)

37

Δms from Tevatron & BSM Limits

( )SisSMSM ehAA σ+→ 1

Pro

bab

ilit

y

Hep-ph/0509117 Agashe/Papucci/Perez/Pirjol

38

ConclusionsExciting times:• 1987 B0 mixing (UA1, Argus)• 1989 CLEO confirms B0 mixing• 1990s LEP B0 Mixing• 1993

– First time dependent Δmd meas. (Aleph)– First lower limit on Δms

• 1999 CDF Run I lower limit (Δms>5.8 ps-1)• 2005

– D0: Δms>5.0 ps-1

– CDF: Δms>7.9 ps-1

• 2006– D0: Δms [17,21] ps-1 @ 90% CL– CDF: Δms=17.31+0.330.07 ps-1 – CDF: 5 observation, Δms=17.77+0.100.07 ps-1

-0.18

PLB 186 (1987) 247, PLB 192 (1987) 245

PRL 62 (1989) 2233

PLB 313 (1993) 498

PLB 322 (1994) 441

PRL 82 (1999) 3576

PRL 97 (2006) 021802

PRL 97 (2006) 0062003