the mulan experiment - 2011. 7. 24.آ  boston university robert carey mulan collaboration boston...

Download The MuLan Experiment - 2011. 7. 24.آ  BOSTON UNIVERSITY Robert Carey MuLan Collaboration Boston University

Post on 09-Oct-2020

0 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • BOSTON UNIVERSITY

    Robert Carey MuLan Collaboration Boston University

    The MuLan Experiment A New Measurement of the Fermi Constant

    Berkeley, Boston University, University of Illinois, James Madison, University of Kentucky, KVI, PSI

    Outline: ●Motivate the measurement ●Describe the experiment ●Final results

  • Precision electroweak predictions rest on three parameters

    Fermi Constant Giovanetti et al

    1984 ±GF GF

    ¼ 9 ppm

    Mass of the neutral weak boson LEP EWWG

    2005 ±MZ0

    MZ0 ¼ 23 ppm

    Fine Structure Constant Gabrielse et al

    2008 ±®em ®em

    ¼ 0:37ppb0.37 ppb

  • RMC, PANIC, July 2011

    The V-A theory factorizes into a pure weak contribution, and non-weak corrections, essentially uncontaminated by hadronic uncertainties.

    Muon decay gives us unique access to the electroweak scale

    The muon only decays via the weak interaction, which gives it a very long lifetime.

    All relevant weak interaction physics confined to one easily(!) measured parameter with a clean theoretical interpretation.

  • RMC, PANIC, July 2011

    The Fermi constant is an implicit input to all precision electroweak studies

    Contains all weak interaction loop corrections.

  • RMC, PANIC, July 2011

    The Fermi constant is an implicit input to all precision electroweak studies

    Plot clipped from LEP Electroweak Working Group publications

    Contains all weak interaction loop corrections.

    Top quark mass prediction:

    +

    e 

    W t

    b

    W e

    +

    +

    +

  • RMC, PANIC, July 2011

    ±GF GF

    = 1

    2

    sµ ±¿¹ ¿¹

    ¶2 +

    µ 5 ±m¹ m¹

    ¶2 +

    µ ±theory

    theory

    ¶2

    18 ppm 90 ppb 30 ppmMid 90s: 17 ppm 90 ppb

    Standard Model Fermi Constant extraction used to be theory-limited

    T. van Ritbergen and R. G. Stuart, Nucl. Phys. B564, 343 (2000)

  • RMC, PANIC, July 2011

    ±GF GF

    = 1

    2

    sµ ±¿¹ ¿¹

    ¶2 +

    µ 5 ±m¹ m¹

    ¶2 +

    µ ±theory

    theory

    ¶2

    van Ritbergen and Stuart: 2- loop QED corrections (massless electrons)

    18 ppm 90 ppb < 0.3 ppm1999: 9 ppm 90 ppb Lifetime error now limits the Fermi constant

    Standard Model Fermi Constant extraction used to be theory-limited

    T. van Ritbergen and R. G. Stuart, Nucl. Phys. B564, 343 (2000)

  • RMC, PANIC, July 2011

    ±GF GF

    = 1

    2

    sµ ±¿¹ ¿¹

    ¶2 +

    µ 5 ±m¹ m¹

    ¶2 +

    µ ±theory

    theory

    ¶2

    1 ppm < 0.3 ppmGoal: 0.5 ppm Lifetime error now limits the Fermi constant

    Standard Model Fermi Constant extraction used to be theory-limited

    T. van Ritbergen and R. G. Stuart, Nucl. Phys. B564, 343 (2000)

  • RMC, PANIC, July 2011

    Also .... Two muon capture experiments at PSI

    Protium: MuCap Deuterium: MuSun ... (where the capture rate is inferred from the difference between the positive and negative muon lifetimes) require measurements of the positive muon lifetime to 10 ppm or better.

  • For 1ppm, need more than 2 trillion muons ...

    πE3 Beamline, Paul Scherrer Institut, Villigen, Switzerland

  • For 1ppm, need more than 2 trillion muons ...

    πE3 Beamline, Paul Scherrer Institut, Villigen, Switzerland

  • For 1ppm, need more than 2 trillion muons ...

    πE3 Beamline, Paul Scherrer Institut, Villigen, Switzerland

  • +12.5 kV

    -12.5 kV

    … with a pulsed time structure

    SlitTarget Kicker

    M uo

    ns in

    ta rg

    et

  • +12.5 kV

    -12.5 kV

    … with a pulsed time structure

    SlitTarget Kicker

    M uo

    ns in

    ta rg

    et

    5¹s

    On

    Nin(t) = R¹¿ ³ 1¡ e¡t=¿

    ´

  • +12.5 kV

    -12.5 kV

    … with a pulsed time structure

    SlitTarget Kicker

    M uo

    ns in

    ta rg

    et

    22¹s5¹s

    On Off

    Nin(t) = R¹¿ ³ 1¡ e¡t=¿

    ´

    Nout(t) = Nin(tc)e ¡t=¿

  • +12.5 kV

    -12.5 kV

    … with a pulsed time structure

    SlitTarget Kicker

    M uo

    ns in

    ta rg

    et

    22¹s5¹s

    On Off

    Nin(t) = R¹¿ ³ 1¡ e¡t=¿

    ´

    Nout(t) = Nin(tc)e ¡t=¿

    Modulators designed, built at TRIUMF

    Plates, vacuum chamber built by PSI

  • RMC, PANIC, July 2011

    Symmetric, highly segmented detector

    Thin stopping

    target

    We have reached our goal by running about 20 muon decay experiments simultaneously

    Polarized surface muon beam

  • RMC, PANIC, July 2011

    Symmetric, highly segmented detector

    Thin stopping

    target

    We have reached our goal by running about 20 muon decay experiments simultaneously

    +12.5 kV

    -12.5 kV

    Polarized surface muon beam

    Electrostatic beam kicker

  • RMC, PANIC, July 2011

    Symmetric, highly segmented detector

    Thin stopping

    target

    We have reached our goal by running about 20 muon decay experiments simultaneously

    +12.5 kV

    -12.5 kV

    Polarized surface muon beam

    Electrostatic beam kicker

    Inner/Outer tile pair

  • RMC, PANIC, July 2011

    Symmetric, highly segmented detector

    Thin stopping

    target

    We have reached our goal by running about 20 muon decay experiments simultaneously

    +12.5 kV

    -12.5 kV

    Polarized surface muon beam

    Electrostatic beam kicker

    Inner/Outer tile pair

    500 Mhz waveform digitization

    MHTDC (2004)

  • RMC, PANIC, July 2011

    Symmetric, highly segmented detector

    Thin stopping

    target

    We have reached our goal by running about 20 muon decay experiments simultaneously

    +12.5 kV

    -12.5 kV

    Polarized surface muon beam

    Electrostatic beam kicker

    Inner/Outer tile pair

    500 Mhz waveform digitization

    N(t) = N0e ¡t=¿ +B

    MHTDC (2004)

  • RMC, PANIC, July 2011

    Data Production in 2006-7

    2006: 1.16 ppm (stat.) 2007: 1.7 ppm (stat.)

    We collected roughly a trillion muons each year

  • 23

    Time-dependent systematic errors are the principal analysis issue

    Early-to-late changes:

    Time in fill

    lo g(

    co un

    ts )

  • 24

    Time-dependent systematic errors are the principal analysis issue

    Early-to-late changes:

    Instrumental issues PMT time pickoff, gain Pulse shape variations Kicker voltage sag and beam dynamics effects Pileup : does detector

    system performance change when pulses get

    close together?

    Time in fill

    lo g(

    co un

    ts )

  • 25

    Time-dependent systematic errors are the principal analysis issue

    Early-to-late changes:

    Instrumental issues PMT time pickoff, gain Pulse shape variations Kicker voltage sag and beam dynamics effects Pileup : does detector

    system performance change when pulses get

    close together? Physics issues

    Spin (de)polarization and precession Non-flat background

    Time in fill

    lo g(

    co un

    ts )

  • 26

    Time-dependent systematic errors are the principal analysis issue

    Ee

    threshold

    time

    co un

    ts i

    Early-to-late changes:

    Instrumental issues PMT time pickoff, gain Pulse shape variations Kicker voltage sag and beam dynamics effects Pileup : does detector

    system performance change when pulses get

    close together? Physics issues

    Spin (de)polarization and precession Non-flat background

    Time in fill

    lo g(

    co un

    ts )

    1 cut data selection

  • Pileup

    time

    A MuLan Detector

    Tile

    Hidden pulses distort lifetime ~ 100 ppm, if uncorrected

    Pulse Resolving

    Time

  • Pileup

    time

    A MuLan Detector

    Tile

    Hidden pulses distort lifetime ~ 100 ppm, if uncorrected

    time

    co un

    ts

    Ppileup /

    Z tr 0

    P (t)P (t+ t0)dt0

    / e¡2t=¿

    Pulse Resolving

    Time

    Pileup goes as Rate2

  • Don't fit for pileup losses (big statistical cost) => reconstruct!

    A MuLan Detector

    Tile

  • Don't fit for pileup losses (big statistical cost) => reconstruct!

    A MuLan Detector

    Tile Artificial

    Resolving Time

    time

    Fill n

    Artificial Resolving

    Time

    Fill n+1

  • Don't fit for pileup losses (big statistical cost) => reconstruct!

    Pileup Time Distribution

    Normal Time Distribution

    A MuLan Detector

    Tile Artific