Stability Issues for NSLS-II Infrared

							Stability Issues for NSLS-II Infrared




               L. Carr
17-April-07 NSLS-II Stability Workshop




                                         BROOKHAVEN SCIENCE ASSOCIATES
                              Outline
• Overview: Why is IR so sensitive?

• Frequency range

• Position Specification (tighter in vertical than horizontal)

• Angular Specification – generally less restrictive

• Summary




                                                        BROOKHAVEN SCIENCE ASSOCIATES
    Infrared Performance Issues: Why is IR so demanding?
• InfraRed Synchrotron Radiation (IRSR) is used for low-throughput
  techniques such as microspectroscopy.
    – Spectrometer endstations are based on highly evolved commercial instruments.
         • Instruments already optimized for highest S/N
    – Fourier Transform (FT-IR) interferometers (modulate spectral intensity into AC signal).
    – Detectors achieve “background limited infrared performance” (BLIP)
      => Photon Noise is often limiting factor.
    – Spectral range: n (1/l) from 1 cm-1 up to 10000 cm-1 (1 cm < l < 1 mm)

• IRSR is typically 1000x brighter than standard laboratory IR source … a
  thermal (blackbody) radiator at ~1200K.
    – background photon noise from 1200K source only factor of ~2x above 300K background.
• In order to benefit from the brightness advantage, IRSR source noise
  should be no more than 10X the thermal source noise. Ideally 1X.
    => thermal source noise serves as reference point.


                                                                            BROOKHAVEN SCIENCE ASSOCIATES
         Fourier Transform Infrared (FTIR) Spectrometer
• Michelson-type interferometers, typically operating in “continuous” or
  “rapid-scan” mode.
                  fixed mirror

   source
                                            moving mirror (scans back and forth)
                                            typical velocity of ~1 cm/s




                                 to experiment endstation and detector

• Each spectral component receives sinusoidal modulation:
  FT gets you spectrum: frequency (in Hz) ~ n (in cm-1).


                                                                  BROOKHAVEN SCIENCE ASSOCIATES
                      Frequency Range Requirement
• Determined by several factors
   – Desired spectral range (usually spans from 1 to 2 full decades)
   – Mechanical movement of FTIR scanner (available velocities)
   – Digitizing rate capability
   – Detector and amplifier response time
   => Modulated frequencies span 1 to 2 decades in a single measurement.
   – One more time-scale: sample/reference measurement.
      Can be < 1 minute or several hours.


• Result:
   –   ~1 Hz up to 1 kHz for far-infrared and THz spectral range.
   –   100 Hz up to 10 kHz for mid-infrared spectral range.
   –   sub 1 Hz for all measurements (sample-in / sample-out).
   –   SUM: need stable up to ~10kHz


                                                                     BROOKHAVEN SCIENCE ASSOCIATES
           Noise Sources in Required Frequency Range
Typical SR Noise Sources
   – Mechanical motion (drift, vibrations). < 200 Hz

   – 60 Hz related (electrical pickup). 60 Hz and multiples
       • up to ~720 Hz. Multiples of 720 Hz in RF sidebands.

       • Note: some low frequency noise can be compensated using dynamic beam steering
         mirrors with feedback.

   – RF (100s of Hz to > 10 kHz)
       • too fast to correct using optomechanics.




                                                                         BROOKHAVEN SCIENCE ASSOCIATES
                          Position Stability Requirements
What motion magnitude can be tolerated?

• The good news: the effective or apparent source size will always be diffraction-limited.

          sdiffraction ~ l2/3 r1/3

• At short wavelength of 2 mm (2x10-4 cm) and r = 2500 cm (NSLS-II), smallest effective
  beam size is about 500 microns.

• Model: use Gaussian beam and “aperture” to determine signal fluctuation as function of
  motion: defines allowable movement.

• Assume upper noise limit of 1%, set requirement at 0.3% under “worst case scenario”.
     – achieving below 0.1% is still beneficial.




                                                                           BROOKHAVEN SCIENCE ASSOCIATES
     Position Stability Requirements
Optimal case: beam perfectly centered on all apertures




                                                   BROOKHAVEN SCIENCE ASSOCIATES
     Position Stability Requirements
Optimal case: beam perfectly centered on all apertures
10 mm movement cause 0.3% change.




                                                   BROOKHAVEN SCIENCE ASSOCIATES
    Position Stability Requirements
Worst case: Sample with sharp edge centered on beam
1 mm movement yields 0.3% change




                                                BROOKHAVEN SCIENCE ASSOCIATES
                    Position Stability Requirements
Result:
• For a symmetrically aligned aperture, beam motion must be kept to below 5% of the
  effective bunch size for 1% noise. This sets an upper limit of 25 microns and goal
  requirement of 10 microns.
• If the aperture can not be symmetrically aligned or experiment can not use a symmetric
  beam profile, then the constraint becomes 10 times more severe (1 micron for 0.3%
  noise).
• This does not reduce noise to background level. It makes it quite tolerable (moderate
  improvement relative to NSLS). Another factor of 10x smaller would improve S/N for
  many measurements.
• Source is always diffraction-limited in vertical, but becomes extended source
  horizontally. Can tolerate more horizontal movement (at least 3X).

• SUM: limit beam motion to 1 mm in vertical, ~ 3 mm in horizontal (more forgiving).




                                                                         BROOKHAVEN SCIENCE ASSOCIATES
              Noise Example: NSLS Infrared beamline U10
Mechanical & electrical noise below 500 Hz (could be reduced by optical stabilization)
“Other” noise (RF?) at higher frequencies
“Noise Floor” => intrinsic noise at detector (baseline with no beam)




                                                                       BROOKHAVEN SCIENCE ASSOCIATES
         Noise Example: NSLS Infrared beamline U10
Example S/N (red) and Ideal (blue)   Loss in S/N due to Beam-related Noise




                                            > 10X loss
                                            below 1.5kHz




                                                           BROOKHAVEN SCIENCE ASSOCIATES
Apparent Beam Motion Example: NSLS Infrared beamline U10
         Position Sensitive Photodetector at endstation, ~ 300 Hz BW




                                                              Equivalent
                                                              NSLS-II goal




                                                             BROOKHAVEN SCIENCE ASSOCIATES
                  Angular Stability Requirements
•   Main issue is beam spillage at an aperture.
•   Exit (collecting) aperture at dipole extraction serves as limit.
•   Typical mid-IR collection of ~ 10 mrad, assume 0.1% tolerance.
•   Vertical may be more sensitive (downstream aperture).

SUM: 10 mrad sufficient for horizontal, suggest 3 mrad for vertical.




                                                              BROOKHAVEN SCIENCE ASSOCIATES
                                  Infrared Noise Summary
• Noise from Source & Beamline components limits performance at most IR beamlines.
    – Mechanical movement (mostly below 100 Hz)
    – Electrical (60 Hz multiples)
    – RF (multiple lines, above 500 Hz)

• Nothing is “magic” about IR requirements: The “competition’s” noise is very low!
    –   1 micron position stability would achieve 300:1 S/N for “worst case".
    –   more forgiving in horizontal than vertical (extended horizontal source, assume 3X more tolerance = 3 mm).
    –   angular position less critical (several mrads is fine, plan to under-fill optics & avoid beam “spillage”).
    –   frequencies to at least 10 kHz.

• Existing NSLS VUV/IR Mid-IR: effectively lose ~10X of potential S/N benefit.
    – all types of noise, RF sidebands difficult to avoid (occur at many different frequencies).
    – optimal alignment helps, but optimization lost when beam position drifts.

• Existing NSLS VUV/IR Far-IR: effectively lose up to ~100X
    – noise is both mechanical and electrical.
    – beam stabilization expected to yield significant improvement.

• Users do not always recognize unusual noise.
    – need independent diagnostic beamport for constant monitoring.



                                                                                              BROOKHAVEN SCIENCE ASSOCIATES