BLS1

1
LISA optical bench design-change
propositions --- Revised from recent
presentations to the LISA International Science
Team --- 08January 2004 Bonny L. Schumaker,
Charles C. Harb Jet Propulsion Laboratory,
California Institute of Technology Bonny.Schumaker
_at_jpl.nasa.gov 818-354-4169
2
Several months ago, we circulated to the LISA
Project a collection of viewgraphs summarizing
our concerns and suggested possible changes to
the baseline design for the LISA optical bench,
titled Optical-bench design propositions --
Ideas proposed for scrutiny and presented
graphically 22 September 2003 Bonny
Schumaker, Charles Harb Jet Propulsion
Laboratory, California Institute of
Technology Bonny.Schumaker_at_jpl.nasa.gov 818-354-41
69 Filename OB-DRNov2003 A subset of those
concerns and proposed changes, which we
currently view as compelling, is presented here
for discussion.
3
Top Three (plus four) optical bench design-change
propositions
1. Use linear (vs. circular) polarization for
science and inter-bench beams. 2. Change
polarization prior to local PM in inter-bench
phase-locking/TDI link, to prevent destructive
interference of signal, reduce noise, and reduce
PM disturbance. 3. Move troublesome (TBD)
heat-producing elements off optical
bench.
Additional possible
propositions, priority TBD 4. Use free-space (vs.
fiber) for inter-bench link. 5. Make independent
backside measurements of relative motion between
proof mass and its optical bench along sensitive
axis. 6. Make optical benches interchangeable
(to ensure measurement capability for a
particular arm). 7. Remove all transmissive
optics (QWPs) from science path, to suppress
thermal effects and to symmetrize optical paths
for the local science beams (incoming LO).
4
The baseline optical bench configuration
Excerpted from LISA System Technology Study
Report, July 2000
(PM)
Circular pol (L, R)
Circular pol (R, L)
5
The baseline optical bench configuration (minus
QWP at rear)
6

• Useful primer for optical diagrams (conventions
  used here)
• Vertical (V, aka S) polarization V (S) ?
  R (QWP)
• denoted by solid lines _______
• is changed to Right Circularly Polarized (R)
  upon passage through a quarter-wave plate (QWP),
  unless noted otherwise (e.g., unless the QWP is
  deliberately angled to make V become L)
• is unchanged upon reflection.
• Horizontal (H, aka P) polarization H (P) ?
  L (QWP)
• denoted by dotted lines .................
• is changed to L upon passage through a QWP,
  unless noted otherwise
• is unchanged upon reflection
• R (L) polarizations R ? L (reflection)
• denoted by dot-dash lines . __ . __ . __
• is changed to V (H) upon passage through a QWP,
  unless noted otherwise
• is changed to L (R) upon reflection

7
Note non-interchangeability of local
benches The incoming (black), not the outgoing
(red), science beam reflects off the local PM.
Accomplish this by making the outgoing and
incoming beams oppositely polarized. Thus, the
two local benches transmit oppositely-polarized
science beams. This makes the
local benches non-identical, hence
non-interchangeable.
--- Baseline bench design (minus QWP at rear) ---
8
The baseline uses exit QWPs oriented 45? to each
other in the outgoing science beams, to transmit
RCP from one bench and LCP from the other.

9
Concern Out-going light suffers back
reflections, e.g., from secondary.
(red) (blue)
Back-reflected circularly-polarized light
interferes at the science PD.
Back-reflected linearly polarized light would
simply return toward the laser.
10

• Proposition 1.
• Use linear (vs. circular) polarization for
  science and inter-bench beams.
• ? Accomplish by replacing exit QWPs with HWPs for
  science beams, and removing post-PM exit QWPs for
  inter-bench beams.
• Gets rid of undesired backreflection into the
  science PD.
• Simplifies design and control of antireflection
  coatings.
• Enables replacement of transmissive optics
  (QWPs) with reflective optics in science paths,
  if dn/dT is so large that multiple passes through
  transmissive optics contributes too much phase
  noise. (See proposed change 7 below.)
• Recommended in the 2000 LISA Final Technical
  Report (Appendix B therein, excerpted below).
• n.b. Benches remain non-interchangeable unless
  HWPs are movable, since the requirement that the
  benches transmit oppositely polarized science
  beams still requires a 45? orientation or no HWP
  on one bench

11
Proposition 1 Use linear (vs. circular)
polarizations for science and inter-bench links
12
Modified baseline optical bench with linear
polarizations front and back (BLS CH, Nov 03)
13
Advantages of using linear polarization,
excerpted from 2000 LISA FTR
14

• Concerns with baseline design problems for
  inter-bench link
• Unnecessarily large disturbance to PM ? 0.5-1
  mW reflects off the PM, while the weaker
• (? 1-mW) leakage beam serves as the LO.
• (0.01 power fluctuations ? 2?100nW/c ?
  10-15 m/s2?Hz acceleration disturbance to the PM)
• Opposite polarizations through fiber produce
  non-common-mode noise.
• Destructive interference between LO and return
  leakage beam from other bench.

15

• Baseline inter-bench link (effect of fiber twist
  symbolized with HWP here)
• Destructive interference between LO and coherent
  return leakage beams (both ? 1mW)
• (red/pink and blue/turquoise).

16

• Proposition 2. Change polarization prior to
  reflection off local PM for inter-bench link.
• Implementation 2a leave QWP behind the PM, send
  twice-leaked beam, down by 10-6, to other bench.
• Disturbance to PM is down by 1000.
• Common polarizations in inter-bench link no
  fiber twist or HWP, polarization noise
  common-mode.
• Phase measurement compares 1-nW with 1-mW.
• Coherent leakage beam (H pol ) returns as a
  negligible ?1 fW compared to 1 mW

17

• Implementation 2b Remove QWP behind the PM, send
  once-leaked beam, down by 10-3, to other bench.
• Disturbance to PM is down by 1000.
• Opposite polarizations in back link, require
  fiber twist or HWP, polarization noise is
  non-common.
• Phase measurement compares 1-mW with 1-mW.
• Coherent leakage beam (H pol ) returns as a
  negligible ?1 fW compared to 1 mW.

18

• Proposition 3.
• Move troublesome (TBD) heat-producing elements
  off optical bench.
• This could include PDs, QPDs, CCDs, AOMs, etc.
  
• Only the science PD (PD1) must be on the OB,
  since it is small,
• wideband, and sensitive to alignment).
• The laser reference cavity might also be
  isolated, e.g., for better
• thermal control.
• TBD
• a) Does this ease manufacture or integration?
• b) Do expected thermal gradient fluctuations
  require it?

19
Proposition 3. Move troublesome (TBD)
heat-producing elements off optical bench.
20

• Proposition 4.
• Use free-space (vs. fiber) for inter-bench link.
  
• The baseline design uses a polarizing fiber for
  the inter-bench link, and it twists the fiber to
  change each beams polarization before it
  arrives at the other bench, in order to ensure
  that it does not reflect off the other benchs
  PM.
• The fiber itself adds noise caused by coherent
  back-scattering and imperfect mode-matching,
  exacerbated by the fact that the baseline design
  has the two beams propagating with opposite
  polarizations.
• ? Use of a free-space link is proposed.
• TBD Verify stability and noise of actuating
  mirrors.

21
Proposition 4. Avoid fiber-induced noise with a
free-space link between benches, controlled with
actuating mirrors. (Note this drawing does not
incorporate Proposition 2!)
22

• Proposition 5.
• Make independent measurements of the relative
  motion between each PM and its optical bench
  along the appropriate sensitive measurement
  axis.
• This option would require fixed reference
  mirrors for the PM-bench measurement.
• Inter-bench phase-locking could use a different
  set of fixed reference mirrors operated as
  Mach-Zehnder interferometers and thus never touch
  the PM at all!
• ? This approach symmetrizes the optics and
  polarizations for the inter-bench link, thereby
  offering all the advantages of Proposition 1.
• Pick-offs from the two bench lasers could be
  used for both PM-bench measurements, to provide
  10-kHz beatnotes. (n.b. ltlt 1mw is needed for a
  5-pm/rtH measurement.)
• These measurements could be made with a QPD (not
  just a PD) in order to see PM rotation, thus
  providing unambiguous SC pointing info from the
  science PD.
• TDI variables require only the difference D1-D2
  in relative PM-bench motions D for the two
  benches, but could independent measurements
  provide useful addl info, or reduce risk from
  failure, or reduce complexity (e.g., in data
  extraction)?

23
Proposition 5. Independent measurements of
PM-bench relative motion
24
Proposition 5. Independent measurements of
PM-bench relative motion
25

• Proposition 6. Make optical benches
  interchangeable.
• (This assures Michelson-type measurement
  capability in the event that two benches on
  different SC become compromised, or to ensure
  measurement capability for one arm.)
• Interchangeability requires either
• a) rolling the SC if benches are made
  identical by mounting them ?45 to each other (in
  which case the SC solar illumination is changed)
  
• or
• b) movable HWPs or QWPs and yawing of the SC.
• TDI variables may be recoverable from other
  measurement combinations, so is
  interchangeability of the benches necessary?

26
? How to replace one bench with its partner
bench? Two possible solutions 1) Orient the
OBs 90? to each other and roll the SC to
interchange benches ?Probably
impractical SC solar cell coverage not
symmetric. (This solution was
considered in the FTR, however see next slide.)
27
From Appendix B of the 2000 Astrium LISA Final
Technical Report
28
? How to replace one bench with its partner
bench? Second possible solution 2) Allow for
rotation of or insertion of different HWP.
(Trade-offs TBD!)
(shown here with a free-space backside link)
29

• Proposition 7.
• Remove all transmissive optics (QWPs) from
  science path, to suppress thermal effects and to
  symmetrize optical paths for the local science
  beams (incoming LO).
• For dn/dT?10-5/?K, dT?10-5?K/?Hz, d?5mm ?
  0.5pm/ ?Hz noise from each pass,
• ? 8 4 pm/?Hz addl noise from double passes
  four QWPs between the two SC!
• QWPs are of different material from the bench,
  so bonding also adds phase noise.
• No wave-plates in the science path eases
  requirements on telescope dimensional stability.
• A final HWP would be used to match
  polarizations of the incoming science and local
  LO beams, but this optic is encountered after
  combination of the local and distant science
  beams, so its thermal noise is common to the two
  combining beams.
• There are only two optics not common between the
  incoming science and local LO beams, but these
  are in different locations, and their noise
  contributions would differ only slightly as
  determined by thermal gradient fluctuations.

30
Proposition 7. Remove all transmissive optics
from science path, to suppress thermal effects
and to symmetrize optical paths for the local
science beams (incoming LO).
31
Extra slides follow ----------
32
--- Baseline bench design (minus QWP at rear) ---
The incoming (blue), not the outgoing (red),
science beam reflects off the local PM.
Accomplish this by making the outgoing and
incoming beams oppositely polarized. Thus, the
two local benches transmit oppositely-polarized
science beams. This makes the
local benches non-identical, hence
non-interchangeable.