The Search for CP Violation in Hyperon Decays

1
The Search for CP Violation in Hyperon Decays

• E. Craig Dukes
• University of Virginia
• HyperCP Experiment
• 6 July 2006
• BEACH 06
• University of Lancaster

2
Why Search for CP Violation in Hyperon Decays?

• After 40 years of intense effort and many
  beautiful experiments we still no little about
  CP violation the origin of CP violation remains
  unknown and there is little hard evidence that it
  is explained by the Standard Model
• The importance of CP violation to our
  understanding of particle physics, indeed the
  universe, cannot be overstated
• The asymmetry in hyperon decays can be relatively
  large up to O(10-2)
• The price is modest
• No new accelerators needed
• Apparatus is modest in scope and cost
• Hyperons are sensitive to sources of CP violation
  that, for example, kaons are not
• Almost all scenarios for New Physics produce
  large CP asymmetries

We are willing to stake our reputation on the
prediction that dedicated and comprehensive
studies of CP violation will reveal the presence
of New Physics Big and Sanda, CP Violation
3
Short Primer on Non-leptonic Hyperon Decays

• Predominantly two-body non-leptonic
• Decay violates parity ? daughter baryon angular
  distribution not isotropic
• The magnitude of the parity violation is given by
  
• The parity violation is generally large
• Slope of the daughter baryon cosq distribution is
  given by

4
Short Primer on Non-leptonic Hyperon Decays
? If parent X hyperon is polarized
Daughter L baryon is polarized
? If parent X hyperon is unpolarized
5
How to Search for CP Violation in L Decays

• Due to parity violation the proton likes to go in
  the direction of the L spin

Under CP violation that antiproton prefers the go
opposite to the direction of the anti-L spin
6
Other CP Signatures in Hyperon Decays
Most precisely known bX -0.0370.015 (HyperCP)

• Expected to be smaller than a asymmetry
• Difficult experimentally to measure small BR
  differences

• Expected to be larger than a asymmetry
• Difficult experimentally to measure
• Need to measure polarization of daughter from
  polarized parent
• b is very small

7
Theory
What have theorists contributed in the course of
25 years toward a quantitative understanding of
CP violation? A. Pais Blois 1989
Nothing.
8
Phenomenology of CP Violation in Hyperon Decays

• CP violation is manifestly direct with DS 1
• Three ingredients needed to get a non-zero
  asymmetry
• At least two channels in the final state S- and
  P-wave amplitudes
• The CP-violating weak phases must be different
  for the two channels
• There must be unequal final-state strong phase
  shifts
• Asymmetry greatly reduced by small strong phase
  shifts
• the pp phase shifts have been measured to about
  1
• the Lp phase shifts cannot be directly measured
  theory predictions disagree

strong phases
weak phases
9
Measuring the L-p Phase Shift

• Done by measuring the transverse component (along
  the b axis) of the L polarization from polarized
  X- decays
• Daughter L polarization given by
• In the absence of CP violation
• Difficult measurement to make! From 144 million
  polarized X- decays we find

Confirms expected small size from recent cPT
calculations
10
Comparison of AX, AL with e?/e
AX, AL
e?/e

• Thought to be due to Penguin diagram in Standard
  Model
• Expressed through a different CP-violating phase
  in S- and P-wave amplitudes
• Probes parity-violating and parity-conserving
  amplitudes

• Thought to be due to Penguin diagram in Standard
  Model
• Expressed through a different CP-violating phase
  in I0 and I2 amplitudes
• Probes parity-violating amplitudes

Our results suggest that this measurement is
complementary to the measurement of e?/e, in that
it probes potential sources of CP violation at a
level that has not been probed by the kaon
experiments. He and Valencia, PRD 52 (1995),
5257.
11
Bad News SM Theoretical Predictions Small

• Much enthusiasm a decade ago as theory
  predictions were relatively large and
  experimentally accessible
• Standard Model predictions have slowly fallen
  since then to

• At same time there was concern that accidental
  cancellation would cause e?/e ? 0
• The expect SM asymmetry is out of reach of any
  experiment, planned or otherwise

Pachos (1991)
Valencia (1991)
(Tandean and Valencia, 2003)
Note no unambiguous connection between dCKM ?
AX, AL
12
Good News SM Theory Predictions are Small

• Most beyond-the-standard-model theories predict
  new ard large CP-violating phases
• These predictions are often not well constrained
  by kaon CP measurements as hyperon CP violation
  probes both parity conserving and partiy
  violating amplitudes
• A recent paper by Tandean (2004) shows that the
  upper bound on AXAL from e?/e and e measurements
  is O(10-2).
• For example, some supersymmetric models that do
  not generate e?/e, can lead to AL of O(10-3).
• Other BSM theories, such as Left-Right mixing
  models (Chang, He, Pakvasa, 1994), also have
  enhanced asymmetries.

It is clear that hyperon decays are much more
sensitive to new physics than e?/e. Sandip
Pakvasa
He, Pakvasa, Valencia
Any CP-violation signal will almost certainly
indicate New Physics
13
Three Ingredients Needed to have an Experiment

1. A way of producing hyperons and antihyperons with
   either
2. exactly known polarizations
3. polarizations known to be equal
4. Control systematics to the level of the expected
   asymmetry lt10-4 level or better
5. A lot of events with as large as polarization as
   possible!

14
Early Experimental Measurements

• None of these pioneering attempts were in the
  realm of testing theory
• New ideas needed to get several orders of
  magnitude improvement in sensitivity

15
Prospects at ee- Colliders

• polarized beam is needed to separately extract aL
  and anti-aL
• exacts large penalty in luminosity

• can be analyzed using the HyperCP technique
  described later
• small BR ? lots of J/ys needed

16
Prospects at pbar-p Colliders

• Pioneered by PS185 at LEAR
• L and anti-L produced polarized average is 0.27
• Polarizaton of L and anti-L rigorously equal by C
  conservaton
• Gas jet used to provide a point target
• Problem s(pp?LL) small
• 65 mb _at_ 1.65 GeV/c
• Total cross section 100 mb
• Much interest at CERN in early 1990s in an
  upgraded LEAR
• target r 1014 atoms/cm2
• luminosity 6.1x1031 cm-2s-1
• events 1.5x109
• These rates can be achieved at Fermilab and there
  is a group exploring making such a measurement
  after the end of the collider era
• PANDA at GSI

17
A New Way of Producing Polarized Ls and anti-Ls

• This technique, pioneered by HyperCP, produces Ls
  and anti-Ls of known polarization through X
  decays

If the X is produced unpolarized which can
simply be done by targetting at 0 then the L
is produced in a helicity state.
If CP is good the slopes of the proton and
antiproton cosq distributions are identical.
18
Sensitive to CP in both X and L Decays
where

• What is experimentally measured is the slope of
  the proton (antiproton) cosq distribution in the
  rest frame of the L (L).
• This is done in a special L rest frame called the
  Lambda Helicity Frame in which the L direction in
  the X rest frame defines the polar axis.

19
Technique Tested on E756 Data

• E756 W- magnetic moment measurement in mid-1980s
• Not optimized at all for CP measurement all
  data polarized!
• Nevertheless, they managed to surpass the PS185
  sensitivity

20
HyperCP Dedicated Hyperon CP Search

• Charged Secondary Beam
• 800 GeV protons on 2x2mm2 target
• mean momentum 167 GeV/c
• rate 10-15 MHz
• alternate /- beam polarity

• High-Rate Magnetic Spectrometer
• 8 high-rate, narrow-pitch MWPCs
• simple yet selective hyperon decay trigger
• very high-rate DAQ 100,000 evts/s to tape

21
HyperCP Yields

• In 12 months of data taking HyperCP recorded one
  of the largest event sample ever
• 231 billion events
• 29,401 tapes
• 120 TB
• Entire WWW as of end of data taking 5 TB

22
Extracting the CP Asymmetry

• If CP is good then the proton and antiproton cosq
  distributions are identical

• Take the ratio of the two distributions if not
  flat CP is violated

• Fit proton cosq ratios to
• to extract asymmetry d

Note No Monte Carlo used in measurement!
23
Difference Between HyperCP and e?/e Experiments
AX, AL
e?/e

• HyperCP measures alternatively 2 identical decays
  with identical apparatus
• Branching ratios 100
• No acceptance corrections

• Kaon experiments measure simultaneously 4
  different decays modes with 2 different
  spectrometers
• The KL?2p, pp- branching ratios are small
• Large acceptance corrections

24
Equalize X- and X Acceptances by Weighting

• Problem acceptances for X- and Xdecays not the
  same due to different production dynamics
• Solution weight the X- and X momentum
  distributions to force them to be identical
• only 3 momentum dependent parameters weighted
• 100x100x100 106 bins

25
Proton, L-pion, X-pion before/after Weighting
26
Monte Carlo Tests

• Monte Carlo only used to
• verify algorithm and implementation
• check that weighting procedure doesnt wash out
  asymmetry
• study a few systematic errors
• Problem how to generate 1 billion MC events?
• Solution Hybrid Monte Carlo
• We get the input asymmetry back

Important Final result has no Monte Carlo
dependence!
27
The Raw CP Asymmetry

• Data broken up into 18 Analysis Sets of roughly
  equal size, each with and - polarity data
• 10 of data sample 119 million X-, 42 million
  X
• No acceptance corrections
• No efficiency corrections
• No background subtraction

Raw (non-background subtracted) CP asymmetry AXL
from all 18 Analysis Sets
Weighted average of all 18 Analysis Sets
28
Controlling Biases to the 10-4 Level

• Targets changed to equalize secondary beam rates
• polarity 2 mm Cu
• - polarity 6 mm Cu

• Little difference in PWC effieciencies between
  and polarity running
• - data solid line
• data dashed line

• Two important features of HyperCP allow biases to
  be controlled to 10-4 level
• Fact that the same spectrometer is used for both
  X- and X proton/antiproton cosq measurements
• Need to make sure that magnetic fields were
  exactly reversed
• Need to make sure that there was no temporal
  dependence of spectrometer efficiencies
• Measuring the proton/antiproton cosq slope in the
  Lambda Helicity Frame
• Localized acceptance differences to not map to
  any part of the cosq plot

• When flipping polarity field magnitude kept to
  within 2x10-4
• This corresponds to a 0.3 mm deflection at 10 m
  for the lowest momentum (10 GeV/c) pions

Important overall acceptance differences do not
cause any bias!
29
Systematic Uncertainties

• Most estimated from data, a few from Monte Carlo
• Most systematic uncertainties can be reduced in
  analysis of full data set

30
Background Subtracted Asymmetry

• Background subtraction
• No efficiency or acceptance corrections.
• Factor of 20 improvement in sensitivity over
  previous limit
• Null result constraining allowed SUSY effects

Expect to have full data set analyzed in about a
year with factor of 3 improvement in sensitivity
31
CP Violation Search in W?LK Decays

• The only hyperons in which parity violation has
  not yet been observed are the W- and W
• HyperCP has 4.5 million W-?LK- ?pK-p- and 1.5
  million W?LK ?pKp decays
• Measure product aWaL in W-?LK-?pK-p- in the
  Lambda Helicity Frame using Hybrid MC method
• P violation observed, no CP violation still
  several orders of magnitude away from expected
  value
• Most precisely known alpha parameter

32
Conclusions

• Hyperon CP violation searches are probing limits
  not constrained by Kaon, B, or EDM measurements
• HyperCP, the first dedicated hyperon CP-violation
  experiment, is probing down to AXL ? 2?10-4, the
  regime where SUSY models allow an effect
• The HyperCP technique could easily be pushed at
  least an order of magnitude, if not more, in
  statistics, to probe the asymmetry at the SM
  level.
• Unfortunately, the Tevatron no longer available
  for fixed-target physics at Fermilab
• It is unlikely that higher-luminosity ee- and
  pbar-p colliders can push beyond the HyperCP
  limit, only perhaps the SPS, with an improved
  HyperCP-like experiment, can go to where the SM
  predicts an effect

33
Backup Slides