Search for direct CP-violation in K ± ± + – decays by NA48/2 Ivan Mikulec HEPHY Vienna, Austria On behalf of the NA48/2 collaboration: Cambridge,

Search for direct CP-violationin K± ±+– decays by

NA48/2

Ivan MikulecHEPHY Vienna, Austria

On behalf of the NA48/2 collaboration:Cambridge, CERN, Chicago, Dubna, Edinburgh, Ferrara,

Firenze, Mainz, Northwestern, Perugia, Pisa, Saclay, Siegen, Torino, Vienna

XXXXth Rencontres de Moriond EWI&UTLa Thuile March 5-12 2005

I. Mikulec: Search for direct CP-violation in K± by NA48 2

Overview

Direct CP violation in K±→3 decays

NA48/2 experimental setup

Measurement principle

Systematic effects

Preliminary result in K± ±+– decay mode

Outlook for K±±00 analysis

Conclusions


Brief history of CP violation

1964 - CP violation in K0 (Cronin, Christenson, Fitch, Turlay)

1993-99 - Direct CP violation in K0 (NA31, NA48, KTeV)

2001 - CP violation in B0 (Babar, Belle)

2004 - Direct CP violation in B0 (Belle, Babar)

CP violation and especially direct:

A window to physics beyond SM

Complementary observables in Kaons: ’/↔Ag↔rare decays

Look for direct CP violation in K± !(only direct CPV in K± possible – no mixing)


2 21 gu v uM( , ) h ku v K±→3matrix element:

Direct CP-violation observable Ag

2 1

03 s

s

s

m

s

m

u

v

2

10 3

i K i

i

s (P p )

s s

0gA

K+-K- asymmetry in g:

g

g g

g gA

Direct CP violation

Dalitz variables

i=3 odd pion

Kπ+π-π±: g = -0.2154Kπ0π0π±: g = 0.652

|h|, |k| << |g|2even

1even

3oddK±

Kπ+π-π±

if


Experimental and theoretical status

10-6

|Ag|

Experimental results

10-5

10-4

10-3

10-2

Ford et al. (1970)

HyperCP prelim. (2000)

TNF prelim. (2002)“neutral” mode

NA48/2proposal

SM estimates of Ag vary within an order of

magnitude (few 10-6…8x10-5).

“charged”“neutral”

Theory

Asymmetry in decay widths A

expected to be smaller than in Dalitz-plot slopes Ag

(SM: ~10-7…10-6).

SMSM SUSYSUSY NewNewphysicsphysics

Models beyond SM predict substantial enhancements partially

within the reach of NA48/2.(theoretical analyses are by far not

exhaustive by now)


Goals and method

Primary NA48/2 goals:– Measure slope asymmetries in “charged” and “neutral”

modes with precisions δAg<2.2x10-4, and δAg0<3.5x10-4,

respectively.

– Statistics required for this measurement: >2x109 in “charged” mode and >108 in “neutral” mode.

NA48/2 method:– Two simultaneous K+ and K– beams, superimposed in

space, with narrow momentum spectra;

– Detect asymmetry exclusively considering slopes of ratios of normalized U distributions;

– Equalise K+ and K– acceptances by frequently alternating polarities of relevant magnets.


Experimental setupPK spectra, 603 GeV/c

54 60 66

1cm

50 100

10 cm

200 250 m

He tank+ spectrometer

Front-end achromat

• Momentum selection

Quadrupole quadruplet

• Focusing• sweeping

Second achromat

• Cleaning• Beam spectrometer

~71011

ppp

K+

K

Beams coincide within ~1mm

all along 114m decay volume

focusing beamsfocusing beams

Analysingmagnet

vacuum tank

not to scale

K+

K

beam pipebeam pipe


Data taking

2003 run: ~ 50 days2004 run: ~ 60 days

Total statistics in 2 years:

K +: ~4x109

K 00: ~2x108

~ 200 TB of data recorded

First result based on 2003 K± ±+– sample will be presented here


Accepted statistics

U

|V|

even pionin beam pipe

Data-taking 2003: 1.61x101.61x1099 KK±± ±±++–– events

KK+ + : 1.03 x10: 1.03 x109 9 eventseventsKK : 0.58 x10: 0.58 x109 9 eventsevents

KK++/K/K–– ≈ 1.8≈ 1.8

odd pionin beam pipe

M=1.7 MeV/c2

Events

Invariant mass


1

11N ( )

N ( )

g uu

u ugR( ) R R( )u gu

Principle of the experiment

2g

g gA

• Build u-distributions of K+ and K-: N+(u),N-(u)• Make a ratio of these distributions: R(u)• Fit a linear function to this ratio: normalised slope ≈ g

e.g. uncertaintyδAg<2.2∙10-4

corresponds to δg<0.9∙10-4

But, achromat and spectrometer magnet unavoidable sources of apparatus asymmetry!


Four ratios

ZZ

XXYY

JuraJura(left)(left)

SaleveSaleve(right)(right)

Achromats: KAchromats: K+ + UpUp

Achromats: KAchromats: K++ DownDown

B+

B

Indeces of R’s’s correspond to• beamline polarity (UU/DD)• kaon deviation in spectrometer mag. field (SS/JJ).

B K

BJ K

AD A B K

A B

N( )N( )

N( )N

DJ

A B KUS A

A KU A B

B KS

KB

( )

N( )N( )

N( )N( )A K

R

R

R

R

Supersample data taking strategy:– achromat polarity (A) was reversed on weekly basis;– spectrometer magnet polarity (B) was reversed on daily basis.

1 Supersample ~ 2 weeks 2003 data 4 supersamples


Quadruple ratio

R = RUSRUJRDSRDJ ~ 1+4g·u

3-fold cancellation of systematic biases:1) Global time-variable biases (K+,K- simultaneously recorded)2) Beam line biases (K+ beam up / K- beam up etc.)3) Detector asymmetries (K+ toward Saleve / K- toward Saleve etc.)

In addition, acceptance is defined respecting azimuthal symmetry4) Effects of permanent stray fields (earth, vacuum tank magnetisation) cancel

The result is sensitive onlyonly totime variationtime variation of asymmetriesasymmetries

in experimental conditionswith a characteristic time smaller than

corresponding field-alternation period (beam-week, detector-day)


Monte Carlo simulation

Still MC is used to study systematics. MC features:– Based on GEANT;– Full detector geometry and

material description;– Local DCH inefficiencies

simulated;– Variations of beam geometry

and DCH alignment are followed;– Simulated statistics similar to

experimental one.

Example of data/MC agreement:mean beam positions @DCH1

K+ dataK dataK+ MC K MC

K+

K

Due to acceptance cancellations, the analysis does not rely on Monte-

Carloto calculate acceptance


Beam systematics

Time variations of beam geometry

Acceptance largely defined by central beam hole edge (~10 cm radius)

Acceptance cut defined by (larger) “virtual pipe” centered on averaged beam positions as a function of charge, time and K momentum

Y,

cmSample beam profile at

DCH1

0 0.4 0.8-0.4-0.8

0

0.4

0.8

-0.4

-0.8

X, cmX, cm

Beam movements: ~ 2 mm

Y,

cm

A-K+A+K-

A-K-A+K+

5565

55

65

x, cm

y, cm DCH1

A-K+A+K-

A-K-A+K+

5565

55

65

x, cm

y, cm DCH1

x,y-beam vs. K momentum

Beam widths: ~ 5 mm

5mm

2mm


Spectrometer systematics

Time variations of spectrometer geometry - Alignment is fine tuned by forcing mean reconstructed invariant masses to be equal for K+ and K-

Maximumequivalent

horizontal shift:

~200m @DCH1or

~120m @DCH2or

~280m @DCH4

E.g. sensitivity to DCH4 horizontal shift: M/x 1.5 keV/m

Momentum scale variation due to limited control of spectrometer magnet current (10-3) cancels due to simultaneous beams

In addition, it is adjusted by forcing mean reconstructed invariant masses to PDG value of MK+

MM


Trigger systematicsInefficiencies measured using control data from low bias triggers.Rate-dependent parts of trigger inefficiencies assumed to be symmetric.

L2 inefficiency

L2 trigger (online vertex reconstruction): time-varying inefficiency (local DCH inefficiencies) 1-≈ 0.2% to 1.8% flat in u within measurement precision

u-dependent correction applied

3x10-3

cut cut

statistical uncertainty from control sample

L1 trigger (2 hodoscope hits): stable and small inefficiency 1-≈ 0.7·10-3

(no correction)

L2correction Δgx104

SS0 0.5±1.8

SS1 1.4±1.0

SS2 -0.2±1.2

SS3 -4.5±1.9

Possibility to use MC studied


Other systematics

No magnetic field correction

Magnetic fieldcorrected for

• Bias due to resolution in u calculation• Sensitivity to fitting interval and method• Effects connected to → decay • Effects due to event pile-up• +/- interactions in material• Track charge misidentification

Further systematic effects studied:

Residual effects of stray magnetic fields (magnetised vacuum tank, earth field)

minimised by explicit field map correction

Field map in decay volume: Y projection

Decay volume: Z coordinate


Fit linearity – four supersamples

22=39.7/38=39.7/38SS0: SS0: g=(0.6±2.4)x10g=(0.6±2.4)x10-4-4

22=38.1/38=38.1/38SS1: SS1: g=(2.3±2.2)x10g=(2.3±2.2)x10-4-4

22=29.5/38=29.5/38SS2: SS2: g=(-g=(-3.1±2.5)x103.1±2.5)x10-4-4

22=32.9/38=32.9/38SS3: SS3: g=(-g=(-2.9±3.9)x102.9±3.9)x10-4-4

U

U

U

U


Time stability

Quadruple ratio with

K(right)/K(left) K(up)/K(down)K(+)/K(-)

4 supersamplesgive consistent

results

controlof detectorasymmetry

controlof beam lineasymmetry

By regrouping the components in quadruple ratio – check residual detector and beam line asymmetries (~few 10-4) – they cancel safely in g fits

g

MC can reproduce these apparatus

asymmetries


Systematics summary and result

Preliminary estimates of systematic uncertainties

Effect on

Δgx104

Acceptance and beam geometry

0.5

Spectrometer alignment 0.1

Analyzing magnet field 0.1

± decay 0.4

U calculation and fitting 0.5

Pile-up 0.3

Syst. errors of statistical nature

Trigger efficiency: L2 0.8

Trigger efficiency: L1 0.4

Total systematic error 1.3

Raw Corrected for L2 eff

SS0 0.0±1.5 0.5±2.4

SS1 0.9±2.0 2.2±2.2

SS2 -2.8±2.2 -3.0±2.5

SS3 2.0±3.4 -2.6±3.9

Total

-0.2±1.0

-0.2±1.3

2 2.2/3 3.2/3

Combined preliminary result: in Δgx104 units

(3 independent analyses) Including L2 trigger

correction


Preliminary result K±±+- (2003 data)

Δg =(-0.2±1.0stat.±0.9stat.(trig.)±0.9syst.)x10-4

Δg =(-0.2±1.7)x10-4

Ag = (0.5±2.4stat.±2.1stat.(trig.)

±2.1syst.)x10-4 Ag = (0.5±3.8)x10-4

• This is a preliminary result with conservative estimate of systematic uncertainties• Extrapolated statistical uncertainty 2003+2004:

Ag=1.6x10-4

• Expect smaller systematic effects in 2004 data (due to more frequent polarity alternation, better L2 performance).

Converted to asymmetry:


Comparison K±±+-

NA48/2 prelim.: 2003 data

10-6

|Ag|

10-5

10-4

10-3

10-2

SMSM SUSYSUSY

Ford et al. (1970)

HyperCP prelim. (2000)

NA48/2 goal:2003-04 data

NewNewphysicsphysics

This preliminary

result is already an

order of magnitude better than previous

experiments


K±±00 analysis

∙ u can be reconstructed with LKr calorimeter only∙ Statistics analyzed: 28 × 106 events (1 month of 2003)∙ Statistical error with analyzed data: Ag(stat)= 2.2 × 10−4

∙ Extrapolation to 2003+2004 data: Ag(stat)= 1.3 × 10−4

∙ Similar statistical precision as in “charged” mode∙ Possibly larger systematic errors

v

-1.5 -1.0 -0.5 0 0.5 1.0 u

0

-1

-2

1

2

3

Dalitz-plot

Statistical precision in Ag0

similar to “charged” mode: - Ratio of “neutral” to “charged” statistics: N0/N±~1/20 (√=1/4.5) - Ratio of slopes: |g0/g±|3 - More favourable Dalitz-plot distribution (gain factor f~1.5)


Observation of scattering effect in K→3 decays

1 bin = 0.00015 GeV2

MC (no rescattering)

Data

K±±00

M(00) GeV/c 2

4mπ+2

Charge exchange process +00 not negligible under 2m threshold,destructive interference generates a cusp in the Dalitz plot,

not seen earlier by lower precision experiments

30M events

4mπ+2

Great potential for a new accurate measurement of ππ scattering lengths

from this data

New preliminary result at Moriond QCD!


Conclusions

Preliminary NA48/2 result (2003 data) on direct CP−violating charge asymmetry in K± ±+– decays is Ag = (0.5 ± 2.4stat. ± 2.1stat.(trig.) ± 2.1syst.)×10-

4

>10 times better precision than previous measurements Further room to decrease systematic uncertainties 2004 data contain another 2×109 K± ±+– events, with

higher quality

Neutral mode asymmetry: complementary, comparable sensitivity

Design goal is within reach in both decay modes

Other interesting results will follow ( scattering lengths, other CP asymmetries, rare decays)


Spare slides


Theoretical predictions of Ag

StandardModel

L.Maiani, N.Paver ’95 (2.3±0.6)x10-6

A. Bel’kov ’95 <4x10-4

G.D’Ambrosio, G.Isidori ’98 <10-5

E.Shabalin ’01 <3x10-5

E.Gamiz, J.Prades, I.Scimemi ’03

(-2.4±1.2)x10-

5

E.Shabalin ’05 (La Thuile’05) <8x10-5

SUSY G.D’Ambrosio, G.Isidori, G.Martinelli

~10-4

Newphysics

E.Shabalin ’98 [Weinberg model of extended Higgs doublet]

~4x10-4

I.Scimemi ’04 >3x10-5


Experimental results so far

””Charged” mode KCharged” mode K±±33ππ±± : :Ford et al. (1970) at BNL: Ag=(-7.0±5.3)∙10-3; Statistics: 3.2M KStatistics: 3.2M K±±;;

HyperCP prelim. (2000) at FNAL: Ag=(2.2±1.5±3.7)∙10-3; Statistics: 41.8M KStatistics: 41.8M K++, 12.4M K, 12.4M K;; Systematics due to knowledge of magnetic fields;Systematics due to knowledge of magnetic fields; Published as PhD thesis W.-S.Choong LBNL-47014 Berkeley Published as PhD thesis W.-S.Choong LBNL-47014 Berkeley 2000;2000;

””Neutral” mode KNeutral” mode K±±ππ±±ππ00ππ00 : :Smith et al. (1975) at CERN-PS: Ag

0=(1.9±12.3)∙10-3; Statistics: 28000 KStatistics: 28000 K±±;;

TNF (2004) at IHEP Protvino: Ag0=(0.2±1.9)∙10-3;

Statistics: 0.52M KStatistics: 0.52M K±±..


NA48 detector

Main detector components:

• Magnetic spectrometer (4 DCHs): redundancy redundancy high efficiency; high efficiency; ΔΔp/p = 1.0% + 0.044%*p [GeV/c]p/p = 1.0% + 0.044%*p [GeV/c]

• Hodoscope fast trigger;fast trigger; precise time measurement (150ps).precise time measurement (150ps). • Liquid Krypton EM calorimeter (LKr) High granularity, quasi-High granularity, quasi-homogenious;homogenious; ΔΔE/E = 3.2%/√E + 9%/E + 0.42% E/E = 3.2%/√E + 9%/E + 0.42% [GeV].[GeV].

• Hadron calorimeter, muon veto counters, photon vetoes.

Beam pipe


KAon Beam Spectrometer (KABES)

3 MICROMEGA gas chamber stations

measure beam particle• charge (prob. mis-ID~10-2);• momentum (p/p=0.7%);• position in the 2nd achromat (x,y100m).

Not used yet forK±3± analysis

Measurement of kaon momentum:

• Reconstruct K3π with a lost pion;• Redundancy in K3π analysis;• Resolve Ke4 reconstruction ambiguity.


Break down of K±±+- statistics

Dates Sub-sample

s

Achromat A+ Achromat A–

KK++ KK– KK++ KK–

0 22.06-25.07

26 229.6229.6 125.9125.9 201.0201.0 114.0114.0

1 6.08-20.08 12 122.5122.5 68.168.1 135.1135.1 75.475.4

2 20.08-3.09 12 147.2147.2 81.881.8 105.5105.5 58.958.9

3 3.09-7.09 4 40.640.6 22.622.6 54.554.5 30.430.4

Total 54 Total events selected 1613.2

Statistics selected for Ag measurement, events x106


Cancellation of beam spectra

Achromat reversalAchromat reversalreverses reverses KK++ and and KK––

beam spectrabeam spectra

Systematic differencesSystematic differencesof of KK++ and and KK–– acceptanceacceptance

due to beam spectradue to beam spectramostly cancelmostly cancel in in RRUU**RRDD

Supersample 1Supersample 1 Supersample 2Supersample 2 SS 3SS 3

Systematic check:Systematic check:Reweighting KReweighting K++ eventseventsso as to equalise so as to equalise momentum spectramomentum spectraleads to negligible effectleads to negligible effectΔΔg=0.03x10g=0.03x10-4-4

KK++

KK––


Fits to the “cusp” effect in K±±00

data

fit - data

M(00), GeV/c2

Standard Dalitz-plot parametrization

2 = 133/139for M(00)>80

MeV/c2

One-loop exchange: 2 = 463/149

One and two loops: 2 = 159/147

Incl. + atoms: 2 = 144/146

M(00), GeV/c2

Cabibbo: PRL 93 (2004) 121801

Cabibbo, Isidori: hep-ph/0502130

Documents

Search for direct CP-violation in K ± ± + – decays by NA48/2 Ivan Mikulec HEPHY Vienna, Austria On behalf of the NA48/2 collaboration: Cambridge,