Photonic Ring Resonators for Near-Infrared CosmologyRavi Gupta (LBNL)
S. Kuhlmann, H. Spinka, D. Underwood (Argonne National Laboratory), Simon Ellis, Kyler Kuehn (Australian Astronomical Observatory),
Pufan Liu, Guohua Wei, Nathaniel Stern (Northwestern University Nanophotonics Group),Leonidas Ocola, Dave Czaplewski, Ralu Divan, Suzanne Miller (ANL Center for Nanoscale Materials)
1 µm
Outline
• Ring Resonator Technology as Notch Filters
• Near-IR Sky Background & How Ring Resonators Can Help
• Applications for Dark Energy Science: SNe & Galaxies
• Fabrication: Progress, Challenges, & Plans
2
3
Optical/Infrared Silicon Waveguides (<1µm wide/tall)
Difference in index of refraction of silicon with surrounding material
Design file with 32 independent devices (300µm separation)
10µm ring in middle (SEM pic)
Our test-stand, optical fiber input and output, tunable and fixed red and IR lasers
4
How Do Ring Resonators Work?
Silicon ring and waveguides fabricated at Argonne
25µm single SiN ring with different modes (7nm gaps)
Free Spectral Range (gap between modes):
Fig. 2. Top panel: schematic diagram of a simple ring resonator showing the input, throughand drop ports and a sketch of the spectrum at each port. Bottom panel: SEM imageof one of our prototype silicon-based ring resonators with a through port and drop port,manufactured at the Center for Nanoscale Materials at Argonne National Laboratory.
remains. The condition for resonance is therefore
mλ = neL, (1)
where m is an integer, λ is the wavelength, ne is the effective index and L is the circumference ofthe ring. The resonant light couples back into the input waveguide and destructively interfereswith the input light. Thus, a series of ring resonators, each tuned to the wavelength of a differentOH night sky line, could provide a means of OH suppression [25].Ring resonators have been developed for applications in telecommunications, industry, and
photonics research as filters, add/drop multiplexers, delay lines, modulators, sensors, lasergeneration, tuneable dispersion compensators, all-optical wavelength converters, frequencycombs, and tuneable cross-connects [26–28]. The motivation to explore the use of ringresonators for OH suppression is primarily due to their method of manufacture. Ring resonators
Vol. 25, No. 14 | 10 Jul 2017 | OPTICS EXPRESS 15873
SEM imageFig. 2. Top panel: schematic diagram of a simple ring resonator showing the input, throughand drop ports and a sketch of the spectrum at each port. Bottom panel: SEM imageof one of our prototype silicon-based ring resonators with a through port and drop port,manufactured at the Center for Nanoscale Materials at Argonne National Laboratory.
remains. The condition for resonance is therefore
mλ = neL, (1)
where m is an integer, λ is the wavelength, ne is the effective index and L is the circumference ofthe ring. The resonant light couples back into the input waveguide and destructively interfereswith the input light. Thus, a series of ring resonators, each tuned to the wavelength of a differentOH night sky line, could provide a means of OH suppression [25].Ring resonators have been developed for applications in telecommunications, industry, and
photonics research as filters, add/drop multiplexers, delay lines, modulators, sensors, lasergeneration, tuneable dispersion compensators, all-optical wavelength converters, frequencycombs, and tuneable cross-connects [26–28]. The motivation to explore the use of ringresonators for OH suppression is primarily due to their method of manufacture. Ring resonators
Vol. 25, No. 14 | 10 Jul 2017 | OPTICS EXPRESS 15873
5
Single wavelength, on-resonance animation (2ps total)
output
input
Simulation of electric field inside ring at resonance
Waveguide Input
How Do Ring Resonators Work?
Silicon on Insulator (SOI) wafers
6
7
Nanofabrication Steps With Electron Beam Lithography
Entire process ~6 hours total, <$200/wafer (ANL/CNM is a DOE user facility)
8
Source of Sky Background in Near-Infrared is
Atmospheric OH molecules
9
Proven OH Wavelength Suppression Techniques GNOSIS Instrument using Fiber Bragg Gratings
Ring Resonators (this talk)
Ring Resonators as Multi-notch Filters
Revolutionize Infrared Astrophysics by removing sky background spikes
10
Need a series of rings on a single waveguide, each with a different radius, tuned to the sky background wavelengths
Even with only 6 notches over 60nm window, can improve SNR from 1 to 5
Fig. 4. Sketch of a ring resonator based OH suppression instrument.
where S is the integrated signal, and B is the integrated background, over the passband ofinterest. For the purposes of these calculations we assume idealised notches with a perfectlyrectangular profile, and that all notches have the same depth and width (except where notchesoverlap). Lines are selected to be suppressed in order of their average brightness. Further weassume that the signal spectrum is flat; for many science cases there may be a particular featurewhich is important, rather than the whole spectrum. Neverthless, these calculations give an ideaof the general requirements for OH suppression. We assume the sky-background to be as givenin Ellis & Bland-Hawthorn [3], except with the interline continuum as measured by Maihara etal. [34].Figure 5 shows the improvement in signal to noise as a function of the number of notches, for
notch widths of 100, 150 and 200 pm and for notch depths of 10, 20, 30 and 40 dB over the J andH bands. The signal-to-noise improves with increasing notch depth, but the difference between30 dB and 40 dB notches is very slight. In the J band, the notch width is relatively unimportantbetween 100 and 200 pm. In the H band, 200 pm notches are better than narrower notches inall cases, except for the 10 dB notches. The optimal number of notches for each combinationof notch width and depth are given in Table 1. In both the J and H band these correspondto approximately 1 notch for every 2 nm of passband, which is a useful approximation whenconsidering shorter passbands.The improvement in signal to noise will be compromised by the total throughput of the
system. However, because S/N ∝ √η, where η is the throughput, the total end-to-end throughputof the system need only be > 4 % if the OH suppression increases the S/N by a factor 5. To becompetitive with FBG OH suppression the total throughput of the OH suppression system itself
Vol. 25, No. 14 | 10 Jul 2017 | OPTICS EXPRESS 15876
11
Ultimate Goal: Integrating Ring Resonators into Astronomical Instruments
12
• 27 optical/NIR telescopes with >4m mirror, operating costs >$3K/hour
• Oversubscribed: 2-4× more proposal time than available
• OH suppression gives 10× better SNR for same exposure time
• Pays for itself in a few weeks with operating costs
Many Potential Customers World-Wide
13
• Redshift >1 SNe: smothered by OH lines
• Low Redshift: best standard candle with one H-band epoch
• All redshift SNe: NIR insensitive to dust (which is the likely dominant LSST systematic uncertainty w/o NIR measurements)
• OH suppression only hope for 100% NIR single epoch H-band measurement
Motivations for Supernova Science
Supernovae in the Near-IR are Better Standard Candles
“NIR observations are expensive to take from the ground as a result of the
significant emission and absorption from the atmosphere…”
“The optical light curve will give us the phase and we will measure the brightness in the near infrared.”
“… significant potential in supplementing future large ground-
based surveys such as LSST…”
(LSST Supernova Group Leaders)
14
15
• Ring resonators could provide a “clean” segment of new Southern Spectroscopic Survey Instrument (i.e., some fraction of fibers with OH suppression)
• OH lines increase costs, limit wavelength ranges, and degrade performance
• All SSSI proposals push farther into NIR
Motivations for Galaxy Science
Detailed Specifications and Early Test Results in Recent Publication
16
ring resonators on chip
17
Recent Ring Resonator Milestones
• Feb 2016: First working devices in red wavelengths
• June 2016: First working devices (and test-stand) in infrared
• January 2017: Northwestern Nanophotonics graduate student Pufan Liu starts (One specification met and 16 working IR devices at that time, 10% silicon yield, much better yield in 3× thicker SiN)
• September 2017: Five of our six specifications satisfied, 132 working devices, 90% silicon yield
1. Free spectral range > 30nm ✓
2. Notch width < 0.4nm ✓
3. Polarization-independent enough to not ruin other specs ✓
4. Notch depth > 20dB ✓
5. >5 notches per chip at correct wavelengths ✓
6. Transmission enough to significantly improve signal-to-noise (>50%)
Specs Needed For an On-sky Engineering Test and Science
✓ = Specs currently satisfied
18
Summary Plot of Free Spectral Range Measurements 132 devices have been fabricated and demonstrated suppression
Spec
19
20
Notch Width Lumerical 3D Finite Difference Time Domain Simulation Results
(confirmed with data)
5 µm radius SiN 1.5 µm radius Si
21
Notch Width OH lines are actually narrow doublets
H BandWith larger OH doublet spacing,
2 rings are better than 1: Optimization study continuing
3µm Si ring
22
Notch Depth Data with 3µm Si Rings
(Australian student simulations show ~3dB reduction in depth due to laser line width)
One advantage of ring resonators is that multiple rings can easily be
applied to the strongest OH lines for more suppression
23
Post-fabrication Wavelength Re-Tuning
• Even with the precise electron beam lithography, not able to fabricate resonance locations to within our specification of 0.2nm.
• Several papers address this by modification of the SiO2 cladding index of refraction, effectively changing ring resonance.
• None had accuracy and long-term stability needed for this project.
• Pufan Liu tried various modifications of cladding exposure techniques and annealing for stability, and found an iterative 2-4 step solution.
• Stability has been shown over ~1 month, and will continue to be monitored in the future.
5 Triple-ring System, Each Targeting 3 OH lines Initial fabrication brings within 2nm, iterative tuning and
re-testing brings wavelengths within 0.2nm spec
OH Line (nm) 1518.71 1524.09 1528.78
Waveguide #25 – OH Line (nm) -0.16 -0.2 0.09
Waveguide #27 – OH Line (nm) 0.04 0.04 0.01
Waveguide #28 – OH Line (nm) -0.18 0 0.15
Waveguide #29 – OH Line (nm) -0.05 -0.07 -0.06
Waveguide #30 – OH Line (nm) -0.04 0.21 -0.06
3 rings, radii differ by 20nm
OH lines
24
25
Coupling Light from Fiber to Silicon
Many papers achieve <2dB “insertion loss.” A month ago our best was 15dB loss, now 10dB.
Would like to achieve 2dB before a “publishable” on-sky test, but obviously many other benefits of an engineering test
Not to scale: fiber core 80× larger than tapered waveguide
26
Coupling Light from Fiber to Silicon
Many papers have solved this with better than 90% coupling, starting in 2003. Google search gives 506,000 results.
We are starting to implement those ideas
27
Simulated Insertion Loss vs. Angle (current test-stand limit 2-3 degrees, ordered new tilt equipment to improve)
28
Fiber-Chip Coupling and Packaging Investigating both industry solutions and custom on-chip V-groove solutions.
This is now the focus of the project.
SEM
29
Ring Resonator Technical + Simulation Plans Now
AAT hole-in-wall where fiber points
toward sky
• Working with LSST SN group on optical+NIR simulations. What is the best use of NIR information with OH suppression?
• Next month, 1st “dry run” 2-night on-sky test at AAT, using location and equipment used for previous FBG test
IRIS2 spectrograph
30
Ring Resonator Technical + Simulation Plans Now
• Repeat all tests with 1.7um Silicon rings (63nm FSR)
• Continue work on fiber-to-chip coupling efficiency
• Work with industrial partners on packaging solutions, which may solve coupling problem at the same time
31
Ring Resonator Technical + Simulation Plans
By end of 2018
• On-sky “fixed fiber” test with chip meeting all specs
• Finish LSST SN NIR simulations and optimization
After that…
• On-sky Supernova/Transient Demonstrator with Fiber Positioner and Final Chip Packaging
• Pursue published advanced ring resonator improvements such as FSR>150nm tricks, multi-mode rings, …
32
Summary
• Silicon Photonics an emerging technology, new to US HEP
• High-impact applications in Cosmology
• Notch filters for NIR background removal is advancing quickly, with planning underway for an on-sky test
33
Backup Slides
34
Wavelength Suppression Techniques
• Major effort for Australian astronomy for almost a decade (Bland-Hawthorn and Ellis)
• Current FBG techniques not easily scalable
• Silicon Photonics like ring resonators could provide scalable solutions
GNOSIS Instrument and first results using Fiber Bragg Gratings (S. Ellis et al. 2012)
2012 publication
35
Wavelength Suppression Techniques
b1 = 0 for α = t, critical coupling
R=25µm critical coupling
36
37
“n” is not so simple
OH lines in 60nm slices of H band
38
Why not just use a high resolution spectrometer?
The resolution tails from the NIR spikes cause a dominant background
39
Note: DESI wavelength range up to 980nm, resolution 0.2nm, not an issue until NIR region
SOI8E, WG9, 3um ring, 5 different polarizations (0,22,45,67,90 deg) 40
Polarization Effects
