Optimal Design of Adaptive Optics Based Systems Using High Fidelity MEMS Deformable Mirror Models

1
Optimal Design of Adaptive Optics Based Systems
Using High Fidelity MEMS Deformable Mirror Models
Center for Automation Technologies and Systems
(CATS)

• Benjamin Potsaid and John T. Wen
• Center for Automation Technologies and Systems
• Rensselaer Polytechnic Institute
• Troy, New York

ISOT 07 September 9, 2007 Lausanne, Switzerland
2
Outline

• Introduction to Adaptive Optics
• Adaptive Scanning Optical Microscope (ASOM)
• Deformable Mirror Simulation Multidisciplinary
  Design Optimization Approach
• Summary

3
Deformable Mirror Technology
Boston Micromachines (3-Layer MEMS)
Xinetics Inc. (Piezo Electric)
AgilOptics Inc. OKO Tech (Membrane MEMS)
4
Adaptive Optics (AO)

• Adaptive Optics (AO) have traditionally been used
  in astronomy, vision science, directed energy
  (weaponry communications).
• AO corrects for a wavefront disturbance external
  to the optical system.

Adaptive optics telescope
5
Microscope Overview
Observation with human eye
Basic digital camera system

• Resurgence of interest in the optical microscope
• Industrial manufacturing
• Product miniaturization
• MEMS devices
• Automated medical diagnostics
• Biological research
• Microrobotics

6
Field-of-View vs. Resolution Tradeoff

• Inherent tradeoff between field-of-view and
  resolution in optical microscopy

0.13 NA 5X 2.6 micron resolution 2.4 mm field
diameter
0.075 NA 2.5X 4.5 micron resolution 4.8 mm field
diameter
0.30 NA 10X 1.1 micron resolution 1.2 mm field
diameter
This tradeoff is a significant hindrance when
using the optical microscope in practice.
7
Common Solutions to the Tradeoff
Multiple Parfocal Objectives
Multiple or Moving Stage/Microscope
Zoom Lens Design
8
Adaptive Scanning Optical Microscope

• Deformable mirror aberration correction
• Two order of magnitude larger field area
• Simplified scan lens design

• Post-objective image field scanning (not a point)
• Faster than a moving stage
• No agitation to the specimen

• Deformable mirror aberration correction
• Two order of magnitude larger field area
• Simplified scan lens design

9
Simplified Lens Design Using DM

• It is challenging and expensive to design and
  fabricate wide field, high numerical aperture,
  and flat field lens assemblies.

10
Deformable Mirror Correction
2D steering mirror angle defines observable field
location
Deformable mirror stays within 2 mm stroke
Perfect Strehl 1.0 Diffraction limited gt 0.8
11
ASOM Field of View and Scan Time
Steerable within workspace
Current design
0.21 NA 1.5mm
Commonly used today
12
CATS/Thorlabs Breadboarded ASOM System

• Close collaboration for Photonics West 2007 demo
  of breadboarded ASOM protoype
• Tech transfer from RPI to Thorlabs occurred
  through building the ASOM together

Custom optics scan lens Thorlabs
Off-the-shelf imaging optics Thorlabs
140 actuator DM Boston Micromachines
13
Thorlabs ASOM Product
Top Illumination Light Injection
X-Y Scanning Mirror
CCD Camera
Field of view 40 mm Resolution 1.5um
Scan Lens
Imaging Optics (custom)
XYZ Position Stage
Deformable Mirror
Back Illumination Light Source
14
ASOM Multiscale Live Organism Imaging

• Unprecedented resolution of freely moving
  organisms longitudinal/behavioral studies
• Simultaneous cell, organ, animal, population
  studies multiscale biological investigations

Single camera exposure per 1/2 of worm

• Unprecedented resolution of freely moving
  organisms longitudinal/behavioral studies
• Simultaneous cell, organ, animal, population
  studies multiscale biological investigations

15
Adaptive Optics Technology

• Past 10 years has seen low cost AO technology
  development.
• Low cost adaptive optics are now becoming a
  reality.
• However, there exists much confusion concerning
• Actual capabilities of deformable mirrors
• Differences between deformable mirror types
  (piezo, electrostatic, electromagnetic, membrane,
  three-layer, bimorph, etc.)
• How to effectively design optical systems to best
  use DM technology

16
Boston Micromachines DM

• Uses the Boston Micromachines (BMC) 140 actuator
  deformable mirror as a test case.

17
DM Modeling using ABAQUS FEM

• Brute force modeling of all silicon features not
  practical
• High aspect ratio features, complex geometry,
  etc.
• Modeling Approach Partition into actuator
  subsystem and membrane surface subsystem.
• Use approximating function Fm(V,d) for actuators

Membrane Surface
Actuator Subsystem
18
FEM Model Validation

• CATS FEM model verified against BMC white light
  interferometer (ZYGO NewView) data
• Subjectively, good agreement in shape and
  amplitude

3x3 actuator grid at 250V
19
Combining Optics and FEM Tools

• MATLAB Scripts communicate with ZEMAX and ABAQUS
  (Finite Element Software) for concurrent
  optimization.

20
Collaborative Optimization (CO)

• Collaborative Optimization (CO) is one method for
  Multidisciplinary Design Optimization (MDO). A
  relaxation and a coordination are introduced
  between the subsystems.

System Level Optimization GOAL minimize system
objective
sshared variables
ttargets for analysis outputs
si ti
Y3
si ti
si ti
Y1
Y2
Subsystem 1 Optimization S.T. analysis
constraints
Subsystem 2 Optimization S.T. analysis
constraints

xn
x2
x1
21
DM Shape Communications

• Represent DM shape with basis shapes from FEM
  simulation.

Collaborative Optimization
Shared variables are the basis coefficients.
22
ASOM Wavefront Error

• Consider an ASOM Design

Is this aberration correctable with a deformable
mirror?
Off axis wavefront
23
Optimize DM Geometry
Residual shape error
DM Voltages
Strehl ratio
Wavefront error
Voltage
Strehl 0.96
2.5 waves Peak-Valley
Iteration
Iteration
Strehl 0.49
5.0 waves Peak-Valley
Note Strehl of 1.0 is diffraction limit
24
Feasible Deformable Mirror Shapes

• Collaborative Optimization incorporates realistic
  DM shape constraints.
• Glass geometry designed for specific DM shape
  forming capabilities.

Note result from different DM
25
Summary

• Novel optical microscope design (ASOM) achieves a
  wide field of view while preserving resolution by
  using a deformable mirror.
• New class of systems use adaptive optics as
  integral component of system.
• Static optical components designed to the
  specific shape correcting capabilities of the DM
  for improved performance.
• Solution by combining high fidelity DM model with
  Multidisciplinary Design Optimization (MDO)
  techniques.
• Part of program to create tools for students,
  engineers, designers to start simulating adaptive
  optics in non traditional applications.

26
Optimization of Deformable Mirror (DM)

• Calculate image quality metric based on image
  sharpness (Muller and Buffington 1974)

where

• Update actuator voltages with steepest descent
  optimization algorithm

(3) Update DM Actuator voltages
(2) Perform line search

• (1) Approximate gradient by finite difference

Iterate
27
Calibration Target Results
Before optimization (85V all actuators)
After optimization