980 GeV/c2 p

Precision measurements of CP violation and

D0-D0bar mixing at CDF

Mark MattsonWayne State University

ICHEP 23 Jul 2010

CDF

Fermilab, IL, USA

980 GeV/c2 p

Overview

• Why is charm mixing and CPV interesting?

• Quick overview of the CDF II detector

• Charm mixing with D0 -> K+π- or π+K-

• Previous result, and prospects with current data

• CP violation measurements with D0 -> h+h-

• Previous result, and prospects with current data

2

Charm Mixing• Neutral K/D/B mesons

can oscillate between matter to anti-matter

• For no CPV,

• Charm mixing is small

• x, y ~ O(1 %)

• kaon mixing seen 1962

• beauty mixing seen 1987

• first evidence of charm mixing was in 2007

3

|D1,2(t)� = |D1,2� e−�

Γ1,22 +iM1,2

�t

x ≡ M2 −M1

Γ

y ≡ Γ2 − Γ1

2Γ

xy

Bs 0 Bd 0 K 0 D 0

|D1,2� = p��D0

�± q

��D̄0�

|q/p| = 1

• No mixing point (0,0)

• excluded at 10.2σ

• Many results combined; no single measurement has reached 5σ significance

• No CPV (|q/p|, φ) = (1,0)

• Experiments are currently consistent with CP conservation

• Experiment and theory need more precision to test for new physics

4

World Average

CDF II Detector

• Located at point B0 on the Fermilab Tevatron

• Looking at fully reconstructed D0 decaying to charged K and π

• silicon vertex detector surrounded by wire drift chamber (COT) in 1.4T solenoid (central tracking)

• These analyses do not use the (EM, hadronic, muon) calorimeters

• Particle identification using energy loss (dE/dX) in the COT

5

Displaced Track Trigger• Using events from trigger

that selects two oppositely charged tracks that are consistent with a detached vertex• Track momentum transverse to

the beam pT > 2.0 GeV

• Track impact parameter > 100 µm

• Initially optimized for B decays, but also good acceptance for charm

6

D0→K+π-

D0→π+π-

D0→K+K-

Untagged D0N(D0→π+π-) ≈ 1.7×106 N(D0→K+K-) ≈ 4.7×106 N(D0→K-π+) ≈ 47×106 D* tagged

Note on Charm Yield & Luminosity

• Number of D* does not scale linearly with integrated luminosity

• Charm trigger is prescaled at higher beam luminosites

• Earlier data had higher yield of charm per 1/fb

7

Jan 2007 1.5/fb good for hadronic charm analysisJun 2009 5 / fb good for hadronic charm analysis

Charm Mixing with Kπ• Use D* to tag initial production of meson

• “right-sign” (RS) - Cabibbo favored decay

• “wrong-sign” (WS) - doubly Cabibbo suppressed decay, or mixing followed by a CF decay

• In the limit of |x|, |y| << 1 and no CPV, ratio of WS to RS versus decay time is

• Cannot measure mixing parameters directly, but still put limits on amplitude

8

r(t) ∝ e−Γt

[

RD +√

RD y′(Γt) +x′2 + y′2

4(Γt)2

]

D∗+→ D0π+ , D0

→ K−π+

D∗+→ D0π+ , D0

→ K+π− D∗+→ D0π+ , D0

↔ D̄0→ K+π−

δKπ is the strong phase differencebetween the DCS and CF amplitudes

y� = y cos δKπ − x sin δKπx� = x cos δKπ + y sin δKπ

!t/0 2 4 6 8 10

R

0.002

0.004

0.006

0.008

0.01

Previous CDF Result

• PRL 100, 121802 (2008)

• “Evidence for D0-D0bar mixing using the CDF II Detector”

• Difference in chi2 between mixing fit (red dashed) and the no-mixing fit (blue bots) is 17.6• Equivalent to 3.8σ significance

• Results were competitive with the best experimental results at the time

9

)-3 (102x'-0.5 0 0.5

)-3

y' (1

0

-10

0

10

20

No mixing point (x’2,y’=0)Best fit point Best fit point with x’2=0

10

1.5/fb: 3.03 million RS D*5.2/fb: ~6.3 million RS D*

1.5/fb: 12.8 thousand WS D*5.2/fb: ~26 thousand WS D*

Work in Progress

• Time-integrated D*• Light green is the published result with 1.5/fb• Blue is current, with 5.2/fb

• Peak at 5.9 MeV/c2 is D*• Rest of the distribution is D0 + random track

from the primary vertex

• Working to improve systematic uncertainties on the analysis (ex. removing D* from B decays, dE/dX variation over time, etc.)

)2) (MeV/c! - M!K - M!!K

WS (M0 10 20 30

2 p

er 0

.5 M

eV/c

0W

S D

0

5000

10000

15000

20000

)2) (MeV/c! - M!K - M!!K

WS (M0 10 20 30

2 p

er 0

.5 M

eV/c

0W

S D

0

5000

10000

15000

20000

)-1CDF II Preliminary (5.2 fb

)2) (MeV/c! - M!K - M!!K

RS (M0 10 20 30

2 p

er 0

.5 M

eV/c

0R

S D

0

500

1000

1500

310"

)2) (MeV/c! - M!K - M!!K

RS (M0 10 20 30

2 p

er 0

.5 M

eV/c

0R

S D

0

500

1000

1500

310" )-1CDF II Preliminary (5.2 fb

• In addition to direct CP violation, D0 oscillations can generate time dependent CP asymmetries that survive integrating over time

• To make precision measurement, need to correct for detector systematics that can bias the asymmetry

CP Violation with

€

ACPππ =

Γ(D0 →π−π +) −Γ(D 0 →π +π−)Γ(D0 →π−π +) + Γ(D 0 →π +π−)

€

Γ(D*− → D 0π s− → [h+h−]π s

−)Γ(D*+ → D0π s

+ → [h+h−]π s+)

=Nh + h −π s

−

Nh + h −π s+

⋅εh + h −

εh + h −

⋅επ s

+

επ −

€

Γ(D 0 →K +π−)Γ(D0 →K−π +)

=NK +π −

NK −π +

⋅εK −π +

εK +π −

D0 → h+h−

Previous CDF Result• PRL 94, 122001 (2005)

• “Relative Branching Fractions and Search for CP Asymmetry for D0 -> Kπ/KK/ππ

• L = 0.123 / fb• (first 15 months of data taking)• D0 -> KK 8.2 thousand• D0 -> ππ 3.7 thousand• D0 -> Kπ 88.3 thousand

12]2KK Mass [GeV/c1.75 1.8 1.85 1.9 1.95

2En

trie

s/3

MeV

/c

0

1000

2000

3000

4000

5000

+]+K-[K+0D*+D+ charge conjugate

CDF II

]2 Mass [GeV/c1.78 1.8 1.82 1.84 1.86 1.88 1.9

2En

trie

s/4

MeV

/c

0

1000

2000

3000

4000

5000

+]+-[+0 D*+D+ charge conjugate

CDF II

]2) [GeV/c+-) - M(K++-M(K0.14 0.145 0.15 0.155 0.16

2En

trie

s/0.

2 M

eV/c

0

2000

4000

6000

8000

10000

12000

+]+-[K+0D*+D+ charge conjugate

CDF II

]2 Mass [GeV/cK1.82 1.84 1.86 1.88 1.9

2En

trie

s/1

MeV

/c

0

1000

2000

3000

4000

5000

6000

7000

8000+]+-[K+0D*+D

CDF II

+ charge conjugate

• These plots are to get an estimate for current signal size

• a lot more than previous published result)

• Different method

• data-driven technique to measure detector efficiencies

• previous used MC

13

133k D*+→D0π+→[π-π+]π+ 140k D*-→antiD0π−→[π+π−]π−D*+→D0π+ →[h−h+] π+ (now)

380k D*+→D0π+→[K−K+]π+ 401k D*−→antiD0π−→[K+K−]π−

Work in Progress

14

Assuming:

€

σN ≅σN ≅1/ N ⇒σACP=1/ N + N

Systematic uncertainty is expected to be O(0.1%), comparable to statistical uncertainty.

Experiment N (D0→π+π-) ACP(D0→π+π-) (%)CDF(0.123/fb) 7.3K 1.0 ± 1.3(stat) ± 0.6 (syst)CDF(4.8/fb) 273K xxx ± 0.19(stat) ± xxx (syst)

Babar (386/fb) 64K −0.24 ± 0.52(stat) ± 0.22(syst) Belle(540/fb) 51K +0.43 ± 0.52(stat) ± 0.12 (syst)

Experiment N (D0→K+K-) ACP(D0→K+K-) (%)CDF(0.123/fb) 7.3K 1.0 ± 1.3(stat) ± 0.6 (syst)CDF(4.8/fb) 781K xxx ± 0.11(stat) ± xxx (syst)

Babar (386/fb) 129K 0. ± 0.34(stat) ± 0.13(syst)Belle(540/fb) 120K -0.43 ± 0.30(stat) ± 0.11 (syst)

Conclusion

15

• Working to update two previous results (charm mixing, charm direct CPV) with more data

• Substantial charm samples

• Mature detector, understood systematic effects

• Working to improve precision of the syst. errors

• Stay tuned!

16

Backup Slides

17

Displaced Track Trigger

18

07/15/10

M.J. Morello

Run I collected O(1) Bs--> Dsπ (all Ds modes)

Run II collected ~2000 Bs--> Dsπ (Ds--> φ[-->K+K-] π)

Compare with only 10x integrated luminosity!

The trigger had a much bigger impact than Tevatron upgrade!!!

Without SVT With SVT

Charm Mix plots

19

No mixing point (x’2,y’=0)Best fit point Best fit point with x’2=0