Progress in deep broadband interferometric nulling with the adaptive nuller

Robert D. PetersOliver P. Lay, Muthu Jeganathan

Jet Propulsion LaboratoryCalifornia Institute of Technology

June 26, 2008

Outline

• Background

• How it works

• Development activities

• Results

• Summary

Nulling in TPF-I

• For deep null require electric fields with– equal amplitudes

– opposite phases

• simultaneously at each wavelength and polarization

• Single-mode filter makes it easier – Removes all spatial effects

Horizontalpolarization

Verticalpolarization

Single mode filter

A(,pol) (,pol)

Why Do Adaptive Nulling

Include a compensator to actively control amplitude and phase for each polarization and

wavelength at low bandwidth

Perfectly matched

beamtrain optics

Perfectly matched

beamtrain optics

Fully symmetric, high performance beam combiner

Single-modespatial filter

Realistic beamtrain optics

Realistic beamtrain optics

Simple asymmetric

beam combiner

Single-modespatial filter

Compensator

Compensator

Parallel high-order compensator design

Parabolic mirror~ 10 x 14 cm

Uncompensatedbeam in (~4 cm)

Dispersive elementsplits wavelengths

Birefringent elementsplits polarizationsPupil

Stop

Compensatedbeam out (~4 cm)

PupilStop Birefringent element

re-combines polarizations

Dispersive elementre-combines wavelengths

S-polarization

Deformablemirror

P-polarization

Phase and Amplitude Control

• Deformable mirror allows independent control of piston and tilt at each wavelength and polarization

Phase control with piston*: Amplitude control with tilt*:

* Side view, shown for single wavelength & polarization

Outline

• Background

• How it works

• Development activities

• Results

• Summary

Development activates• Proof-of-concept experiment ( = 0.8 to 0.9 µm)

– Less expensive optics and detectors– Relaxed/scaled requirements– Demonstrate feasibility of the design– Gain experience with the control system– Status: Complete – all requirements met/exceeded.

• Results presented at SPIE 05 Annual Meeting, San Diego.

• Mid-IR experiment ( ≈ 8 to 12 µm)– Demonstration of phase and intensity control only– Requirements traceable to flight needs for phase and intensity control– Demonstrate to functionality of the design– Status: Complete – all requirements met.

• TPF Milestone 1 Report(http://planetquest.jpl.nasa.gov/TPF-I/adaptiveNullerTestbed.cfm)

• Applied Optics Paper http://ao.osa.org/upcoming_pdf.cfm?id=94240

• Mid-IR Nulling experiment ( ≈ 8 to 12 µm)– Demonstrate nulling at the required TPF level of 100,000:1– 3 runs of > 6 hours with and RMS null of 100,000:1 at least 3 days apart– Status: In progress – complete by end of September, 2008– Requirements detailed on PlanetQuest site in TPF-I Milestone #3 Whitepaper.

What goes into a 100,000:1 null?

Nuller

Intensity Balance < 0.12% OPD < 8nm Phase Dispersion < 5nm

100,000:1

Polarization Single Mode Fiber Leakage . . . Other effects

Schematic Experimental Setup

• Mach-Zehnder type interferometer• Reference is a “fixed” version of the adaptive nuller.• Laser metrology to maintain a stable difference between the two arms.• Bandwidth = 3.2µm FWHM.

Single PixelDetector

FlipperMirror

Outline

• Background

• How it works

• Development activities

• Results

• Summary

Recent Results

0 1000 2000 3000 4000 5000 6000 7000 800010

-7

10-6

10-5

10-4

10-3

10-2

10-1

100

101

Nulling Data 20080719

Time (s)

Nor

mal

ized

Nul

l Dep

th

Data

RMS = 93,000:1

~2 Hours

Metrology off

ShuttersClosed

NoiseFloor

Ongoing Work

• Some more improvements to the testbed– Full enclosure

– Fully fix the metrology system to increase bandwidth

– Interferometer “EKG” system installed• Measure environmental parameters to see if we are taking

“good” data.

– More photons!

• If we still can’t get to 100,000:1– Filters to narrow the bandwidth

– Wire-grid polarizers

– Pinhole on fiber output

– ???

Summary

• Mid-IR experiments in progress– Metrology problems that have slowed us are

understood.– At least one null ~90,000:1 has been achieved and was

shown to be stable over 2 hours– Identified key upgrades to be done within the next

month.– Identified potential “fixes” if it still doesn’t work.

• Many more measurements to take before we can declare victory– Null ~80,000:1 many months ago (Applied Optics)– Null ~70,000:1 month ago (w/ flaky metrology)– Null ~90,000:1 week ago

Thank you for your time.

This work was performed by the Jet Propulsion Laboratory, California Institute of Technology under contract with the National Aeronautics

and Space administration.

Thanks to Peter Lawson!

BACKUP

SLIDES

Deformable Mirror

760 780 800 820 840 860 880 900 920 940-30

-25

-20

-15

-10

-5

0

5

10

Inte

nsity D

iffe

ren

ce

(P

erc

en

t)

Wavelength (nm)

Actuator #1 2 3 4 5 6 7 8

760 780 800 820 840 860 880 900 920 940-30

-25

-20

-15

-10

-5

0

5

10

Inte

nsity D

iffe

ren

ce

(P

erc

en

t)

Wavelength (nm)

Actuator #1 2 3 4 5 6 7 8 • Tilting every other

actuator• 8 actuators to

control the spectrum

• MEMS deformable mirror from Boston Micromachines.

• 140 actuators (12x12 – 4 corners)• 3mm square continuous

membrane• ~1.8µm travel per actuator

Mid-IR components

• Using ceramic heater as source

• Transmissive optics replaced with appropriate materials (ZnSe)

• Wollaston prism tested separately.– CdSe is the only birefringent material

– Manufactured by Cleveland Crystals

– Cost ~ $27k each.

– Test sample had good agreement with our Zemax model.

• Mid-IR spectrometer– Grating and single pixel mercury cadmium telluride

detector on a computer controlled translation stage

Lab Setup

Source

ReferenceArm

AdaptiveNuller Arm

DM

ShutterDispersingWedges

FlatMirror