At McGill: scrambling for tenure precision

Realistic goal: room-temperature squeezing

Goals:

• Squeeze all relevant frequencies for LIGO?
• Tunable quadrature frequency dependence?
• Purely quantum bath physics, transduction, ...

"Realistic" calculation

for our systems

​

~50 micron fiber cavity

~1 W circulating

Room temperature

Displacement noise (m/Hz^1/2)

Our 2D SiN design (simulated structure)

Parameters

• 100 nm thick
• ~10’s-100’s microns pad diameters
• Square or hex lattice behave similarly

Advantages

• Stronger localization than 1D
• More unit cells / larger effect
• Technically easier

High-contrast devices

Example large-pad device

• 128 um pad
• 350 um tether length
• ~1 um tether width
• window size 4.5 mm (90 unit cells)

Bandgap

~ 1.3 x Band Edge

Brownian

motion of central pad

Optomechanics at McGill

1. Optomechanics boot camp
2. Trampolines and cavities
3. Optically defined mechanical geometry

Common mechanical sensors

Micro-mechanics on-chip

• Oscillators
• Electronic filters
• Accelerometers

Sensors for fundamental science

Spacetime strain sensors

LIGO

Light measures and pushes things

Each photon carries momentum

​

Particles of light (photons) probe information about position and velocity:

​

bonk!

recoil

Sense of scale: momentum of a falling snowflake

= 100,000,000,000,000,000,000's of these green photons

= about a day of light from this laser pointer (battery?)

= column of this light 100's of times the distance to the sun

= 10's of calories (enough to heat your finger tip by 10's of degrees)

​

You (personally) would feel the heat before you notice the force!

information

Light measures and pushes things

Each photon carries momentum

​

Particles of light (photons) probe information about position and velocity:

​

bonk!

recoil

Sense of scale: momentum of a falling snowflake

= 100,000,000,000,000,000,000's of these green photons

= about a day of light from this laser pointer (battery?)

= column of this light 100's of times the distance to the sun

= 10's of calories (enough to heat your finger tip by 10's of degrees)

​

You (personally) would feel the heat before you notice the force!

information

Optomechanics: control with radiation forces

Optomechanics

• New knobs to turn
• All sizes and shapes
• Mostly described by this simple toy model:

Same physics: driven slide guitar

Optical resonances

Energy in Cavity

​

Radiation Force

Mirror Position

few milliwatt input

(weak laser pointer)

force from 1,000,000,000,000,000,000,000,000 photons / second

= 1 micronewton of radiation force

= weight of about 20 grains of salt

= push a paper clip 1 cm in 2 seconds (in space)

...

...

Driving near a resonance: optical spring

Radiation Force

Mirror Position

Surprisingly stiff photons

Mechanical Response of Gram-Scale Mirror

low power

high power

Drive Frequency (Hz)

• Optical cavity with one “floppy” gram-scale mirror
• Optical spring stiffens gram-scale mirror resonance from 170 Hz to 5,000 Hz
• Column of light is stiffer than diamond! (however brittle, “windy”, unstable)

Corbitt et. al. PRL 98, 150802 (2007)

Optical spring instability

lag in the cavity response

(ring-up and ring-down time)

Radiation Force

Mirror Position

cavity light pumps mirror motion:

“antidamping”

Laser cooling

Energy in Cavity

Mirror Position

mirror motion pumps cavity light:

“laser cooling”

mechanical temperature can be made very close to absolute zero!

Optomechanics: a powerful knob

Quantum sensors?

Goals

• Coherent interface between dramatically different quantum systems
• Arbitrary wavelength conversion

light

“superposition”

"Standard" quantum limit for optical readout

Fundamental noise source: "shot noise"

• Photons from "ideal" lasers arrive randomly, producing the "quantum hiss" (like rain on a tin roof)

​

bonk!

noisier recoil

more information

bonk!

bonk!

bonk!

bonk!

more photons

Standard quantum limit applied to cavity system

Increasing probe power

• Improves readout precision
• Introduces force noise (quantum hiss)

Laser Power

Transmitted Power

Frequency

laser frequency

average power out measures mirror position

Measurement Noise Floor

Consequence: squeezing (example)

Squeezing the quantum noise:

Initial laser noise

-> force noise

-> motional noise

-> cavity resonance noise

-> correlated transmitted power noise

that can cancel original noise (one quadrature)

Frequency

Transmitted Power

output

quantum noise

squeezed

Recent awesomeness from MIT LIGO

Advantages of optomechanical nonlinearity over material nonlinearities

• Wavelength agnostic
• Tunable frequency-dependence of squeezed quadrature
• Miniaturization

N. Aggarwal, et al, arXiv:1812.10224 (Dec 2018)

At McGill: scrambling for tenure precision

At McGill: scrambling for tenure precision

High-quality "trampolines" for cavity optomechanics

Process flow

1. Buy nitride-coated silicon wafer
2. Photolithographically define trampoline shape
3. Dunk in KOH to remove silicon

SiN on Si

SiN on Si

released

nitride

photoresist on SiN

[Viewpoint, Physics 9,40] C. Reinhardt, T. Müller, A. Bourassa, J. C. Sankey, Ultralow-Noise SiN Trampoline MEMS for Sensing and Optomechanics, Phys. Rev. X 6, 021001 | arXiv:1511.01769 (2016)

​

Gory details

(or ask!)

Most promising devices

Identifying resonances

• Measured chip and trampoline modes
• Simulated modes agree within ~2%

Trampoline mechanical performance

• Mechanical Q ~ 50-100(+) million
• Sensitivities approaching nanowires, nanotubes, single graphene sheets
• Cavity compatible!

[Viewpoint, Physics 9,40] C. Reinhardt, T. Müller, A. Bourassa, J. C. Sankey, Ultralow-Noise SiN Trampoline MEMS for Sensing and Optomechanics, Phys. Rev. X 6, 021001 | arXiv:1511.01769 (2016)

​

​

Optical performance

• Sweep through resonance quickly, measure optical “ringing”
• Estimate cavity hold time, finesse, absorption, scattering, etc
• No sign of intrinsic optical loss from processed nitride (finesse 40,000)

naive expectation

[Viewpoint, Physics 9,40] C. Reinhardt, T. Müller, A. Bourassa, J. C. Sankey, Ultralow-Noise SiN Trampoline MEMS for Sensing and Optomechanics, Phys. Rev. X 6, 021001 | arXiv:1511.01769 (2016)

​

​

In a cavity: interesting parameter regime

High “single-photon cooperativity”

a.k.a. "cooling efficiency (resolved sideband)"

​

​

​

​

​

​

Notes & open questions

• Cavity occupancy of 1 can apply damping larger than the environment
• Quantized nature of light?

Time

Vibrational Amplitude

without light

with cooling

with antidamping

Way too sensitive!

Traditional lock fails!

Realistic goal: room-temperature squeezing

Goals:

• Squeeze all relevant frequencies for LIGO?
• Tunable quadrature frequency dependence?
• Purely quantum bath physics, transduction, ...

"Realistic" calculation

for our systems

​

~50 micron fiber cavity

~1 W circulating

Room temperature

Displacement noise (m/Hz^1/2)

Goal: "new" levels of control

Optomechanics: control with radiation forces

Traditional optomechanics

Strong optical control over dissipation and spring constant

​

What about this?

Phononic crystals + optomechanics?

A. Z. Barasheed, T. Müller, J. C. Sankey, Optically Defined Mechanical Geometry, Phys. Rev. A 93, 053811 | arXiv:1511.06193 (2016)

Electromagnetic energy coupled to localization length

Weird functionality (tunable defect)

• Light coupled to mechanical localization length, effective mass
• Infinitesimal light => finite localization

Our 2D SiN design (simulated structure)

Parameters

• 100 nm thick
• ~10’s-100’s microns pad diameters
• Square or hex lattice behave similarly

Advantages

• Stronger localization than 1D
• More unit cells / larger effect
• Technically easier

Single-photon localization length

Plugging in realistic numbers:

​

​

​

​

​

​

​

~ 1 cavity photon (average)

Not absurd!

Better in 2D!

~ 50 micron unit cell

~ 10 micron cavity

Increased response in larger structures?

One figure of merit: mode amplitude difference due to light (fixed mechanical energy)

​

​

​

​

​

​

​

​

Larger response from larger, more massive structures?

Single-photon => 10% for mm-scale structures?

A. Z. Barasheed, T. Müller, J. C. Sankey, Optically Defined Mechanical Geometry, Phys. Rev. A 93, 053811 | arXiv:1511.06193 (2016)

Realistic ideas

Goals

• Demonstrate control: frequency, amplitude, mass, shape
• Learn about dissipation
• Tunable mechanical coupling to other systems
• Landau-Zener dynamics with a continuum?
• State transfer through defect lines / loops?

Ideas welcome!

Half-baked idea

Dough.

Progress in fabrication: old (clean) method

H2 bubbles => low yield!

Driven Mechanical Motion

Being more careful (up to ~4mm, but low yield)

Improving the yield

Goal: protect the front side

• Partial etching the nitride (fail)
• Wafer chuck (haven't tried)
• SiO2+Protek (works surprisingly well)

Be more careful (also works)

2.5 mm

(up to 5 mm)

High-contrast devices

Example large-pad device

• 128 um pad
• 350 um tether length
• ~1 um tether width
• window size 4.5 mm (90 unit cells)

Bandgap

~ 1.3 x Band Edge

Brownian

motion of central pad

Small-cell, higher-frequency devices

• 16 um pad,
• 70 um tether length,
• 1 um tether width,
• Window size 4.5 mm (2538 cells)
• Factor of 2500 mass tuning possible!

In progress

Fiber Mirror Etching /

Finesse-Tuning

Vibration-isolated, monolithic UHV system

Quantum-limited laser

Optomechanics at McGill

Devices discussed here

• Ultralow-noise “trampolines"
• Large phononic crystal membranes

Current Goals

• Broadband squeezing & SQL
• Laser-controlled geometry

​

​

​

​

​

​

Other Projects

• Low-photon optomechanics
• Torsional levitation
• Dissipative optomechanical coupling
• Quantum nondemolition readout
• Spin transfer and quantum emitters in diamond

Abeer Barasheed

Steffi Ruf (visiting)

Vincent Dumont

Max

Ruf

Simon

Bernard

Christoph Reinhardt

Tina

Müller

Wife

Self

Cal

Wife

Cal

Yannik Fontana

Luke Hacquebard

Adrian Solyom

Mark Dimmock

Erika Janitz

Zack Flansberry

Tommy

Clark

Raphael St-Gelais

Spin transfer controlled magnetic circuits & NV's

NV center as local magnetic probe

• ~noninvasive readout of stray fields
• nm-scale resolution (in principle)
• protected circuit (doesn't oxidize)
• learn about spin Hall, image modes

Magnetic circuits for NV spin control?

• Spin-transfer control of NV states
• Controllably couple remote spins?

Dip when driving spin flips (more time in a less- fluorescent state)

(Loss Group)

Collaboration with Childress Group (McGill)

& Pioro-Ladriere Group (Sherbrooke)

Related geometries

Hammel Group

Spin transfer?

Advertisement: spin transfer + quantum spins

Goal:

• Sense spin-transfer-controlled nanoscale magnetization dynamics with NV centers in diamond
• Answer questions not answerable through transport alone
• Feasibility of hybrid systems?

Part II: Spin transfer and NV's

Efficient control over nanoscale magnets

• Electrons repolarize at interface
• Transverse spin component -> ferromagnet

Can electronically drive (and measure)

• Ferromagnetic resonance
• Magnetic antidamping / damping (cooling)
• Parametric oscillations

Source of spins: Spin Hall torque

A source of spin current: Spin-orbit coupling

• Spin-orbit coupling produces spin current
• Spins apply torque to magnetic layer

Applications

• Hard drive read heads / magnetic sensors
• magnetic RAM
• magnetic neural networks (?!)

​

​

Preliminary: spin transfer ferromagnetic resonance

GHz in, DC out

• Lineshape provides torque magnitude and direction
• Spin transfer torque ~ Oerstead field torque
• Sensible frequency shift with field

B-Field 35^o out of plane

Preliminary: implanted NV electron spin resonance

ESR sensitivity to stray fields

• Frequency: DC field
• Linewidth & contrast: RF field strength & noise
• Ask wife.

Field Stripline

Parametric driving and magnetic cooling

​

Parametric drive

• Drive magnetics at 4 GHz, detect NV photoluminescence at 2 GHz
• Threshold behavior while antidamping
• Strong spin-transfer control of NV

​

​

​

Cooling

• NV spin relaxation rate measures local field noise => magnon temperature
• DC current cools magnons to ~150 K
• NV sensitive enough to detect magnon zero-point fluctuations?

Fabricating nanowires on diamond

electron current

single-crystal

diamond

Implanted NV Centres

• ~100 nm deep (high-quality bulk properties)
• ~1 NV per confocal volume

Nanowire stacks

• 3-10 nm Py / 3-10 nm Pt (no sticking layer)
• ~50-500 nm-wide
• 1-20 microns long
• Curren in nanowire stack:
• Spin Hall torque in plane
• Oerstead field torque out of plane

Other structures: kinks, zags, vortex cores

RF field stripline

Spin Hall

torque

Abeer Barasheed

Steffi Ruf (visiting)

Vincent Dumont

Max

Ruf

Simon

Bernard

Christoph Reinhardt

Tina

Müller

Wife

Self

Spawn

Wife

Spawn

Yannik Fontana

Luke Hacquebard

Adrian Solyom

Mark Dimmock

Erika Janitz

Zack Flansberry

Preliminary: resistive readout

Anisotropic magnetoresistance (AMR)

• ~25% of current flows through Py layer
• Follows expected cos²(θ) dependence
• Provides estimate of magnetization angle

Technologies: Ultrashort optical cavities

Fiber-based microcavities

• Shorter cavity = more bounces per second = larger per-photon force
• Confined optical mode = suited to smaller, lighter MEMS
• Compact, monolithic, fiber-coupled package (ideal for cryostat)

Technology: Ultrashort optical cavities

E. Janitz, M. Ruf, M. Dimock, A. Bourassa, J. C. Sankey, L. Childress, A Fabry-Perot Microcavity for Diamond-Based Photonics, Phys. Rev. A 92, 043844 | arxiv:1508.06588 (2015)

Increasing reflectivity by removing material

Predictions (MEEP)

• Resonant reflection of one wavelength
• Subsequent HF (acid) dips thin membrane, widen holes
• Controlled tuning of resonance

increased reflectivity by removing material

(Solgaard group, Stanford)

Collaboration with Yves-Alain Peter (Polytechnique)

Resonance tuning with acid

"Slightly" modified fabrication

Thin SiN crystals

• Electron beam instead of photolith
• No surprise: new surprises
• Resonance fine-tuned with HF dips

Why? Why?!

Tuning the resonance post fabrication (~60 nm thick)

S. Bernard, C. Reinhardt, V. Dumont, Y.-A. Peter, J. C. Sankey, Precision Resonance Tuning and Design of Photonic Crystal Reflectors, Optics Letters 41, 24, 5624-5627 | arXiv:1609.00858 (2016)

Problem: collimation = plane waves at all angles

Collimation broadening (simulations)

Simulations (plane waves only)

• "Parasitic" mode couples to angled plane waves, shifts main resonance

​

• Really messes with main resonance when degenerate

​

• Main resonance also depends on angle

​

=> Collimated beam less perfectly reflected

• Especially in thin structures
• Figure of merit and optimization

S. Bernard, C. Reinhardt, V. Dumont, Y.-A. Peter, J. C. Sankey, Precision Resonance Tuning and Design of Photonic Crystal Reflectors, Optics Letters 41, 24, 5624-5627 | arXiv:1609.00858 (2016)

R&D: Magnetic nanocircuits and NV centers

N-V magnetometry of magnetic nanocircuits

• All-electrical FMR on protected magnetic nanocircuit
• Imaging of dynamics with N-V defects in diamond
• Nanoscale magnetic elements, sub-wavelength techniques
• Coherent coupling between N-V spins and magnons?

collaboration with Wife (McGill) & Michel Pioro-Ladrière (Sherbrooke)

Simulated Dynamics

Fabricated

Contacts

Preliminary: first (weak) check for interactions

DC current through device

• Two NV's in spectrum (one very dim)
• Asymmetric bending of the resonances
• Magnetization responding to current?

Other fun: stable optical spring

Stable optical spring!

​

​

Can be enhanced by orders of magnitude

Optical trapping

Oscillator Position

Time

Enhanced MEMS

• faster response
• less dissipation
• less force noise

Other fun: optical trapping to increase Q

With trampolines

• Only milliwatts required
• Levitated Q > 10⁸ limited by tether / sheet modes
• (not optimal)

Simulation with optical spring:

Solution: eliminate tether losses with torsional symmetry

Torsional mode levitation

• Does not mix with violin modes (symmetry)
• First hybridization with “flappy” disc mode
• 1500 x Q enhancement (hours-long ringdown? effectively levitated?)

Simulations of optically trapped mechanical modes

Q-Enhancement

Trap Strength (MHz)

Trap Strength (kHz)

Q-Enhancement Frequency (kHz)

increased laser spot size

T. Müller, C. Reinhardt, J. C. Sankey, Enhanced Optomechanical Levitation of Minimally Supported Dielectrics, Phys. Rev. A 91 153849 | arXiv:1412.7733 (2015)

Closed-loop gain

• Data explained with measured cavity, amplifier, and delay
• Compatible with noise squashing ~10⁷ at 1 kHz? (In progress!)
• Details, part numbers, pedagogical introduction
• Up next: lock, demo, squeeze (?)

C. Reinhardt, T. Müller, J. C. Sankey, Simple High-Bandwidth Sideband Locking with Heterodyne Readout Opt. Express 25, 1582-1597 | arXiv:1610.01631 (2017)

​

​

In progress: 10 mK trampolines

At Low Temperature

• Reduced force noise, increased bare Q:
• 100 x higher?
• Zeptonewtons?
• Quantum motion experiments
• Vary temperature & optical spring, study dissipation
• Nanoscale MRI?

Now at: 7 mK

3K - 300K

Mechanical Feedback of Fiber Cavities (0-40 kHz)

• ​

E. Janitz, M. Ruf, Y. Fontana, J. Sankey, L. Childress, High Mechanical Bandwidth Fiber-Coupled Fabry-Perot Cavity, Opt. Express 25, 20932-20943 | arXiv:1706.09843 (2017)

​

​

Immediate problem: laser locking

Sideband locking with (single laser) heterodyne readout

• Feedback to sideband frequency
• Carrier acts as heterodyne local oscillator
• Ideal for low cavity powers
• Limited only by signal delay

C. Reinhardt, T. Müller, J. C. Sankey, Simple High-Bandwidth Sideband Locking with Heterodyne Readout Opt. Express 25, 1582-1597 | arXiv:1610.01631 (2017)

​

​

Only

$44.95

Act now!!