Docstoc

PAVI06 HC Asymmetries in a Polarized e- Beam

Document Sample
PAVI06 HC Asymmetries in a Polarized e- Beam Powered By Docstoc
					        Controlling Helicity-Correlated
         Asymmetries in a Polarized
                 Electron Beam
                           Kent Paschke
              University of Massachusetts, Amherst




                                                        Some slides adapted from
                                                        G. Cates, PAVI ‘04



PAVI ’06 Milos May 20, 2006       Kent Paschke – University of Massachusetts
               World Data near Q2 ~0.1 GeV2

                                          ~3% +/- 2.3% of
                                          proton magnetic moment
                                          ~0.2 +/- 0.5% of
                                          proton electric FF
                                          ~20 +/- 15% of
                                          isoscaler magnetic FF
                                        Caution: the combined fit is
                                        approximate. Correlated errors and
                                        assumptions not taken into account


                                           HAPPEX-only fit suggests
                                           something even smaller:
                                           GMs = 0.12 +/- 0.24

           Preliminary                     GEs = -0.002 +/- 0.017




PAVI ’06 Milos May 20, 2006   Kent Paschke – University of Massachusetts
     Precision goals for PVeS experiments
    JLAB
Generation
       1      • HAPPEX:                 dA ~ 1 ppm

              •   A4:                   dA ~ 300 ppb
              •   G0:                   dA ~ 300 ppb
       2      •   HAPPEX-II He:         dA ~ 250 ppb
              •   HAPPEX-II H:          dA ~ 100 ppb          208Pb




              • SLAC E158:              dA ~ 15 ppb
       3      • PREx:                   dA ~ 15 ppb
              • QWeak :                 dA ~ 5 ppb

       4      • Moller at 12GeV (e2e)

PAVI ’06 Milos May 20, 2006       Kent Paschke – University of Massachusetts
    Helicity-Correlated Beam Asymmetries
         • Helicity-correlated intensity (charge) asymmetries
         • Helicity-correlated position differences

         This is what has been considered for 2nd generation
            (precision goals at the 10-7 level)
             See G.D. Cates, PAVI ’04 presentation.

         • Helicity-correlated beam spot size asymmetries,
           x/x’ correlations… in general, higher-order helicity-
           correlated effects…

         If detector is sufficiently symmetric, higher-order effects
            will be dominant!

         One needs to be careful to focus on the largest problems and
           develop systems for measuring, removing, and/or
           estimating corrections for higher order helicity-
           correlated beam parameters. (Not in this talk.)



PAVI ’06 Milos May 20, 2006                  Kent Paschke – University of Massachusetts
                 The Polarized e- Source
                                Optical Pumping:
                                 … of strained GaAs cathode
                                 produces highly-polarized e- beam.




                                HV Extraction and Injection
  Preparation of
  Circularly-polarized light
          Pockels Cell:
          Rapid Helicity Flip
HC beam asymmetries
correspond to differences in
preparation of circularly
polarized laser light*.
PAVI ’06 Milos May 20, 2006         Kent Paschke – University of Massachusetts
Feedback is effective… as far as that goes
       charge                             position
                                                                   Charge and position
                                                                   feedback successful
                                                                   for G0 forward-angle
                                                                   Figures from K.Nakahara



   Helicity-correlated cut         Helicity-correlated deflection by
   laser power to zero             a piezoelectric-controlled mirror
   charge asymmetry

This works, but these are heavy hammers for a subtle problem.
         Does nothing to fix higher-order problems, may even create them.
Preferred strategy: configure system with care to minimize effects.
                  If you do it right, all problems get small together*!
If you do your best there, you can use feedback to go the last mile (or nanometer).


                                                                       *At least, so you hope.
PAVI ’06 Milos May 20, 2006                    Kent Paschke – University of Massachusetts
Fine Control of Beam Asymmetries in Laser Optics


                 Recent work:
 Lisa Kaufman, Ryan Snyder, Kent Paschke,
         T.B. Humensky, G.D. Cates



 Cates et al., NIM A vol. 278, p. 293 (1989)
 T.B. Humensky et. al., NIM A 521, 261 (2004)
 G.D. Cates, Proceedings from PAVI ’04




            Close Collaboration with the JLab Electron Gun Group in
                   analyzing causes and developing solutions


PAVI ’06 Milos May 20, 2006                     Kent Paschke – University of Massachusetts
   Various Causes of Helicity-correlated Beam Changes

    • Steering effects – Pockels cell
    • Imperfect circularly polarized light
        –   Intrinsic birefringence of the Pockels cell
        –   Other birefringent beamline elements (vacuum window)
        –   Phase gradient in beam before Pockels Cell
        –   Laser divergence in the Pockels cell
        –   Quantum Efficiency Anisotropy Gradient
    • Beam element/helicity electronics pickup
    • Quantum Efficiency Variation (“QE holes”)
    • Cross-talk between different beams: cathode
      effects or cross-talk in electron-beam transport


                              (partial list)
PAVI ’06 Milos May 20, 2006               Kent Paschke – University of Massachusetts
                      Piezoelectric Steering
      The piezoelectric Pockels                                                             Red, IHWP Out




                                         X position diff. (um)
      Cell acts as “active” lens                                                            Blue, IHWP IN




                                         Y position diff. (um)
Signature of steering:
   • scales with lever arm                                               Translation (inches)
    • not related to beam polarization
    • does cancel on slow reversal
 PAVI ’06 Milos May 20, 2006                                     Kent Paschke – University of Massachusetts
   The simplest consequences of imperfectly
           circularly polarized light
             Polarization Induced Transport Asymmetry (“PITA” effect)
                      creating intensity (charge) asymmetry AQ

     Perfect ±l/4 retardation
     leads to perfect D.o.C.P.              Now L/R states have opposite sign
                                            linear components.

                                            This couples to “asymmetric
                                            transport” in the optics
                                            system to produce an
                                            intensity asymmetry.
A common retardation offset over-rotates
   one state, under-rotates the other

  Right helicity   Left helicity
                                              Significant DoLP with small change in DoCP

                                               ( DoLP )  1  ( DoCP )
                                                           2                      2


      This is the D phase           In the photocathode, there is a preferred axis:
                                     Quantum Efficiency is higher for light that is
  PAVI ’06 Milos May 20, 2006                  polarized – University of
                                              Kent Paschkealong that axisMassachusetts
                                              Measuring analyzing power
                                                      Perfect
                                                      DoCP
                                                                            Scanning the Pockels Cell voltage
                                                                            = scanning the retardation phase
Intensity Asymmetry (ppm)




                                                                            = scanning residual DoLP
                                                                                   A simplified picture:
                                                                               asymmetry=0 corresponds to
                                                                                minimized DoLP at analyzer
                                   Pockels cell voltage D offset (V)   Voltage change of 58 Volts, added to
                                                                       both the + and - voltages, would zero
                                                                       the asymmetry.



                              A rotatable l/2 waveplate
                              downstream of the P.C.
                              allows arbitrary
                              orientation of DoLP


                            PAVI ’06 Milos May 20, 2006                        Kent Paschke – University of Massachusetts
                          Intensity Asymmetry using RHWP
minimum                                                                   maximum
analyzing                                                                 analyzing
power                                                                     power
Electron beam intensity
   asymmetry (ppm)




                                 Rotating waveplate angle




                                                                     4q term measures
                                                                     analyzing power*DoLP
                                                                     (from Pockels cell)
PAVI ’06 Milos May 20, 2006                                 Kent Paschke – University of Massachusetts
       What happens if there are phase gradients
               across the laser beam?
A gradient in the phase results in a DoLP gradient across the beamspot.

                                  Big charge
      Large D                     asymmetry        Gradient in charge
                                                   asymmetry creates a beam
   Medium D
                                   Medium charge   profiles with helicity-
                                   asymmetry       dependent centroid.
                                  Small charge
       Small D                    asymmetry



     Same effect (Charge asymmetry
     gradient -> position difference)
     can be created by constant linear
     polarization but gradient in
     Cathode Analyzing Power

  PAVI ’06 Milos May 20, 2006             Kent Paschke – University of Massachusetts
Evaluating phase gradients and their effects
 Optics-table data looking at asymmetries while translating Pockels cell


        Intensity asymmetry
        is proportional to the
        phase D.


        Position difference is
        roughly proportional to
        the derivative of the
        intensity asymmetry.


        Spot size difference is
        roughly proportional to
        the derivative of the
        position difference.


 PAVI ’06 Milos May 20, 2006           Kent Paschke – University of Massachusetts
         Electron beam position
                                    Position Differences using RHWP

                                                                               Position differences also
           difference (micron)




                                                                                  follow “2q/4q” fit.

                                                                       4q term measures:
                                                                       analyzing power*(gradient in DoLP)
                                                                                           +
                                           Rotating waveplate angle
                                                                        (gradient in analyzing power)*DoLP

                                  With Large D voltage offset
Electron beam position
  difference (micron)




                                                                      Large DoLP -> large position difference
                                                                        -> Gradient in cathode analyzing power



                                          Rotating waveplate angle

                To minimize all effects, keep DoLP small and stay at small effective analyzing power

             PAVI ’06 Milos May 20, 2006                              Kent Paschke – University of Massachusetts
 Beam Divergence and Fine Alignment of Cell
                                                             New!
• Off-axis beam mixes index of refraction between
optic and extraordinary axes
• Divergent beam couples D-phase to divergence angle
• Beam divergence couples angle to position, resulting
in a position-sensitive D-phase

                                        Laser spot centroid difference, after linear
                                           polarizer (maximum “analyzing power”)
    Vertical position difference (mm)




                                                                                         Simultaneous zero position
                                                                          IHWP IN        differences for pitch and yaw
                                                                                         angles (same for both waveplate
                                                                                         states) can be found,
                                                                                         representing best average
                                                                                         alignment along optic axis.
                                                                          IHWP OUT

                                                                                        Higher order: when alignment is complete,
                                                                                        this effect will lead to “quadrapole”
                                                       Yaw Angle (mrad)                 breathing mode of beam spot.

  PAVI ’06 Milos May 20, 2006                                                       Kent Paschke – University of Massachusetts
                   Strategy for success
  • Well chosen Pockels cells and careful alignment
    minimize effects.
  • Adjust RHWP to get small analyzing power.
      – Not large, but not zero. You want to be able to tune DoCP on
        cathode to counteract vacuum window effect
  • Adjust voltage to maximize DoCP on cathode
  • Use feedback on PC voltage to reduce charge
    asymmetry.
      – Pockels cell voltage feedback maximizes circular polarization,
        which is good for both “zeroth” AND higher orders

     This technique is robust. <300 nm position differences in injector for
      HAPPEX-H setup, and for G0 back-angle setup (same algorithm for
                     optics alignment, different personnel)!

                         If you still care about the remaining position differences:
                          use position feedback, keeping in mind you may just be
                              pushing your problem to the next highest order.
PAVI ’06 Milos May 20, 2006                 Kent Paschke – University of Massachusetts
         Beam Position Differences, Helium 2005
 HC beam asymmetries                           *unless you decide to add helicity
 correspond to differences in                  information to the electron beam after
 preparation of circularly                     it is generated from the cathode
 polarized laser light*.
                                                                    • Problem clearly identified
             X Angle BPM AT 3 GeV!                                  as beam steering from
                                                                    electronic cross-talk
                                                                    • Tests verify no helicity-
                                                                    correlated electronics noise
                                                                    in Hall DAQ at sub ppb level
    micron




                                                                    • Large position differences
                                                                    mostly cancel in average over
                                                                    both detectors, cancels well
                                                                    with slow reversal

                                                                    Raw ALL Asymetry




                                                              ppm
   Helicity signal to           Helicity signal to
   driver reversed              driver removed

          Problem: Helicity signal deflecting the beam through electronics “pickup”
PAVI ’06 Milos May 20, 2006                        Kent Paschke – University of Massachusetts
                   Large beam deflections even when Pockels cell is off
                                   Injector Position Differences for
                                           2005 HAPPEX-H
  After configuration:
  position differences in injector had maximum around 200 nanometers
    position difference (micron)




                                                                                                            200 nm
      Electron beam vertical




                                                                                                            -200 nm


                                                          Location in Injector

                                   Additional suppression from slow reversal and adiabatic damping



PAVI ’06 Milos May 20, 2006                                               Kent Paschke – University of Massachusetts
                        Adiabatic Damping
 Area of beam distribution in the
phase space (emittence) is inversely
    proportional to momentum.

         From 100 keV injection
         energy to 3 GeV at target,
                                            3 GeV
         one expects helicity-                      95
         correlated position               335 keV
         differences to get smaller


                                                                            Design
   The critical parameter in position difference                            transport

                isn’t sqrt(emittence)
                                                                                  Bad match
   The projection along each axis is sensitive to          X’                     to design
                                                                                  transport
                     coupling.
    If the coupling develops, it is difficult to remove…
  To take advantage of adiabatic damping, keep machine
    close to design to minimize undesired correlations.                 X
PAVI ’06 Milos May 20, 2006                    Kent Paschke – University of Massachusetts
           Taking Advantage of Phase Space Reduction
  Major work invested to controlling beam transport as designed (Yu-Chiu Chao)

  • Transport matching design (linacs & arcs) now routine.             Factor between
  • Improvements in the 5MeV injector major step forward               5-30 observed
  • Configuration very stable over 2+ months                          during HAPPEX-H
  • Next battle: 100 keV injector
                  X-BPM (mm)                           Y-BPM (mm)
                                       without


X-PZT
(Source)
                                         1C-Line
                                                                               1C-Line
                                with

                  X-BPM (mm)                           Y-BPM (mm)



Y-PZT
(Source)
                                         1C-Line                               1C-Line


  PAVI ’06 Milos May 20, 2006                    Kent Paschke – University of Massachusetts
         Hydrogen 2005 position differences
                  Dx                                                      5*Dx’
micron




                                           micron
                  5*Dy’                                          Dy
micron




                                           micron                  4*DE/E
Run Averaged:
Energy: -0.25 ppb
                                           4*ppm


X Target: 1 nm
X Angle: 2 nm
Y Target : 1 nm • Degradation of source
Y Angle: <1 nm   setup at end of run but
                 good adiabatic damping
PAVI ’06 Milos May 20, 2006                         Kent Paschke – University of Massachusetts
               What is needed for the future
       • The next generation experiments at JLab (QWeak and
         PREx) will increase demand to understand and control
         higher order effects.

       • Significant progress has been made by thoroughly
         understanding the origins of the effects.
       • Continued empirical work is critical.
       • Need to focus on passive suppression, while exploring
         what might be gained through (the right kind of)
         feedback.

       • Improvements in beam diagnostics and sensitivity
         measurements will be required. This may involve new
         hardware... and new thinking.



PAVI ’06 Milos May 20, 2006           Kent Paschke – University of Massachusetts
                              Backup




PAVI ’06 Milos May 20, 2006       Kent Paschke – University of Massachusetts
                  Non-linearity in Beam Corrections
                                                                      Based on slides by Dave Mack

                                          Magnitude depends on product of:
                                             Nonlinear response in apparatus,
                                                              and
                                             Size modulation at frev in beam <xi 2>
No significant JLab bounds on Δ<I2>, Δ<E2>, Δ<X2>, Δ<X’2>, Δ<Y2>, or Δ<Y’2>.
(“Significant” means they haven’t been proven to be smaller than 1 ppm.)
      (also, no significant JLab bounds on the         How big are these terms?
      15 unique Δ<xixj>.)
                                                       Do these effects cancel under half-
Examples of currently invisible <xi2>:                 wave plate reversal?
        Simple breathing .       Interaction between scraping and   Differential intensity bounce.
                                 intensity feedback.                Same <x>, <I>,
        Same <x>, <I>,
                                 Same <x>, <I>,                     Different <I 2>
        Different <x2>
                                 Different <x2>




  PAVI ’06 Milos May 20, 2006                           Kent Paschke – University of Massachusetts
                Xi                                Xi                          time
                                Phase Trombone
•Goal: vary beta phase
    • implemented with eight existing quads at the beginning of the Hall A arc
    • Allows for independent beta fcn phase control in horizontal and vertical
    planes
• Uses:
    • Allows one to trade off position and angle differences (10:1 scale
    between size in accelerator and senstivity for experiment)
    • Periodic phase changes can be used to randomize or reverse the
      sign of position differences




                                                    Constraints:
                                                        • Preserve beam size at the
                                                        location of the Compton
                                                        polarimeter
                 horizontal phase advanced by 60o       • Preserve large dispersion at
                   while vertical stays fixed
                                                        center of arc
                                                        • Preserve ability to independently
                                                        vary spot size at target
PAVI ’06 Milos May 20, 2006                         Kent Paschke – University of Massachusetts
                                                      Figures from Beck, PAVI’04
    PhaseTrombone, Results from First Test in Hall A
   Data from 2004 (Bogacz and Paschke):
      Phase Trombone          Dx (mm)   Dy (mm)      Dqx(mrad)     Dqy (mrad)
      Setpoint (Dqx , Dqy)    0.3 mm   0.3 mm     0.01 mrad     0.02 mrad
            (0o,0o)             2.9       2.0          -0.08           -0.19
           (30o,0o)             2.7       1.2          -0.07          -0.22
           (-30o,0o)            2.8       3.2          -0.07           -0.16
           (30o,30o)            1.0       1.2          -0.12           -0.21


         Promising approach, but not applied in 2005
  • “Local” phase trombone undone by over contraints (too few independent
  quads)
  • “Linac” phase trombone promising, but brief test was ambiguous.
  Diagnositics are probably insufficient.
  • Electronics pickup made tests uninterpretable



PAVI ’06 Milos May 20, 2006                     Kent Paschke – University of Massachusetts
                 Polarized beam without PC
                                                     Slide from M.Poelker

    60 degree
 optical delay line         s-polarized
                                                          steering
                                           atten           mirror

Fiber-based      l /2                                             l /4
   laser


                                          atten
       Fast RF                                                        s and p
        phase               p-polarized                              polarized
                                                           l /2
       shifter



Fast phase shifter moves beam IN/OUT of slit;
  Downside: extract
 PAVI ’06 Milos May 20, 2006 2x required beam current
                                         Kent Paschke – University of Massachusetts
Position differences at the end of H-2005




PAVI ’06 Milos May 20, 2006   Kent Paschke – University of Massachusetts
     Improved measurement of high-frequency
                beam parameters


  Calculated response of slow
 beam monitors with fast raster

                                        What are the
                                        implications of such
                                        non-linear response for
                                        false asymmetries, or
                                        normalization?


PAVI ’06 Milos May 20, 2006       Kent Paschke – University of Massachusetts

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:4
posted:8/13/2012
language:
pages:30