OPN Astrophotonics Revised
OPN Astrophotonics Revised
OPN Astrophotonics Revised
Astrophotonics
The Rise of Integrated
Photonics in Astronomy
Photonic components
are fueling impressive
gains by Earth-based
astrophysicists—with
more to come in the
soon-to-be-initiated
age of extremely
large telescopes.
Single-mode waveguides transport Y-splitters/junctions can split individ- Directional couplers combine beams
light between components with ual telescopes/sub-pupils for later with two interference outputs (dark
precisely matched optical path dif- recombination, or combine beams and light), maintaining all flux.
ference, maintaining coherence and with a single interference output.
performing spatial filtering.
SMF cores
AO systems are now fully integrated into the design Transition
of telescopes, and can compensate for the bulk of the dis- MMF core
tortion induced by the turbulent atmosphere. But there
are always residual, uncompensated effects that must
be handled downstream within the optical instrument
located at the telescope focus. The versatility of photon-
ics gives us many ways to manipulate light with a view The photonic lantern
to forming an undistorted image. A multimode fiber device consisting of an array of identical
single-mode fiber cores, the photonic lantern allows the light
Coupling light signal captured by a telescope’s multimode fiber to be chan-
neled into an array of output single-mode waveguides.
Photonics allows for some remarkably simple but effec-
Adapted from S.G. Leon-Saval et al., Opt. Lett. 30, 2545 (2005)
tive operations that have no other counterpart in optics.
For example, if we place a single-mode fiber (or a wave-
guide within an integrated-photonic chip) at the telescope another solution to this quandary: the photonic lantern, a
focus, we can couple some light into it, albeit inefficiently. multimode fiber device consisting of an array of identical
For light to couple into the fiber, it must have a wavefront single-mode fiber cores.
whose wave vector is perpendicular to the fiber’s front face. The lantern allows telescopes to use a large-diameter
By definition, the fiber “cleans” the input beam, allow- multimode fiber at the telescope focus that receives essen-
ing only coherent light to couple into the fiber. Since the tially all of the focused light. If N spatial (unpolarized)
fiber supports only the fundamental mode, higher-order modes couple into the photonic lantern, these can be cou-
spatial information is filtered out (at the cost of injection pled efficiently into N output single-mode waveguides—a
efficiency). The fiber thus acts as the perfect spatial filter, an device referred to as a lantern waveguide. Generally, there
invaluable property for interferometry and spectroscopy. is no one-to-one mapping between a specific input mode
Ideally, the beam would have a Gaussian cross-section and output waveguide; the information is shared across
with a diameter matched to that of the fiber core. AO all outputs. Thus beam-smeared incoherent light can
systems come close to producing a diffraction-limited cir- be converted to single-mode propagation along a set of
cular beam for optimal injection, but the results are never waveguides that can be used as inputs to other photonic
perfect. If the corrected beam is smeared spatially, the actions—for example, array waveguide gratings (AWGs)
coupling will be poor, leading to low overall throughput or Bragg gratings. All of this functionality can now be
in the astronomical instrument. Astrophotonics provides integrated into a 3-D photonic device.
Tri-couplers and other multi-input, Multimode interference (MMI) couplers Pupil-remappers have a 3-D arrange-
multi-output couplers allow more use a multimode propagation region, ment of waveguides to convert sub-
complex beam combinations to be rather than evanescent coupling, pupils from a single telescope pupil
performed. to perform complex input-output into a linear array for combination,
arrangements. while maintaining optical path length.
fainter than its parent star. The star’s photon noise thus
completely obliterates the faint signal from the planet. resulting from the single-mode waveguides allows the
To address this, an interferometric instrument can be darkest possible null (destructive interference) to be pro-
designed so that the light from the parent star is destruc- duced, maximally removing the contaminating starlight
tively interfered with itself, or “nulled,” and only the and revealing even the faintest planetary companions.
uncontaminated light from the planet emerges from the Using combinations of these techniques, interferom-
outputs of interest. Making this happen requires pre- eters have recently tracked stars close to the black hole
cise control of optical delays, splitting ratios and other Sgr A* that lies at the center of the Milky Way, probed
parameters—ideally suited to the precision of a photonic the dusty chaos surrounding distant stars and caught
chip. This novel technology is being deployed on-sky, the dance of planet formation in-the-act in far-away
as in the GLINT instrument at the Subaru telescope. solar systems.
Throughout all of these interferometric applica-
tions, the intrinsic spatial-filtering effect of single-mode Microspectrographs
waveguides plays a key role. The process of making The primary instrument at most astronomical obser-
accurate measurements or reconstructing an image vatories is the spectrograph. And this instrument, too,
from interferometric data relies on having a clean, is an area that is benefiting from a breakthrough inte-
high-contrast interference pattern, such as that formed grated-photonic application developed at the University
by a set of arbitrarily small slits. In other words, while of Sydney: the photonic integrated multimode micro-
robust against atmosphere-induced wavefront error spectrograph (PIMMS).
occurring between telescopes or sub-apertures, it is not For a given spectroscopic resolution of an instrument
so robust against wavefront error occurring within a that’s not diffraction limited, the instrument’s size scales
given sub-aperture. with the slit width, which in turn scales with the diam-
The use of single-mode waveguides solves this, eter of the telescope mirror. Modern spectrographs can
removing higher order spatial structure within each be up to 6 m long, and will only grow larger with the
sub-element and yielding vastly improved precision. In giant mirrors of the next-generation ELTs. The PIMMS
the case of nulling interferometry, the ideal wavefront sidesteps the need for these huge optical instruments.
Using the photonic lantern, the instrument accepts at the output face as an undulating pattern of speckles,
multimode (incoherent) light from the telescope focus, varying in time as the atmosphere-induced wavefront
which is converted into an array of single-mode outputs. error changes. Since this pattern is at the spectrograph’s
Since the light from a single-mode fiber, by definition, input slit, its spatial structure is convolved with the
provides a diffraction-limited source, the single-mode spectral point-spread function of the spectrograph.
outputs can be lined up into a slit arrangement and This unpredictable modal noise corrupts the shape of
fed to a diffraction-limited spectrograph. The spec- a spectral element on the detector, resulting in a loss of
trograph can thus be in its most compact form (of spectral precision. The single mode fibers of the pho-
order 10 cm), regardless of the properties of the input tonic spectrograph do not have this problem, as each
aperture. That is the secret to achieving inexpensive yields a perfect Gaussian beam thanks to the spatial-
compact instruments on next-generation ELTs, planet filtering effect.
rovers and space telescopes. Such instruments have been demonstrated in a
Microspectrographs have extraordinary potential. A University of Sydney CubeSat and on telescopes in
diffraction-limited spectrograph’s resolving power (its Australia, Chile and Hawaii. They work with high
periodic spacing divided by its line thickness) is given efficiency and high reliability—so much so that other
by R = λ/δλ = mN where m is the order of interference functions, like complex filtering and ring resonators,
and N is the interference grating’s finesse—essentially are now being integrated into them. For example, sci-
the number of combining beams produced in the spec- entists at the Joint Space Science Institute, University of
trograph, or the number of tracks in the PIMMS array Maryland, USA, have recently demonstrated a complex
waveguide grating. A well-optimized PIMMS can be fiber Bragg grating embedded into a 2-D waveguide.
made very small indeed; a grating with 2,000 lines per
millimeter, illuminated over an area diameter of 10 mm, Future prospects
can achieve R = 20,000 in first order, in an instrument Integrated photonics’ impact on astronomy is just begin-
only 10 cm long. ning. Some photonic technologies are only now proving
A single-mode photonic spectrograph inherently their mettle in astronomical applications. One example
offers improved precision over traditional fiber-fed lies in the production of accurate optical frequency combs
spectrographs. In those instruments, used since the for precision spectroscopic applications—specifically,
1970s, the light from the star or galaxy of interest is the detection of extrasolar planets via the radial veloc-
injected into a large multimode fiber and transported ity shift that they impose on their parent star.
to the spectrograph, where the end of the fiber forms As a planet orbits a distant star, its mass tugs its par-
part of a pseudo-slit. But the injected starlight excites ent star backward and forward by a minuscule amount,
a large combination of modes in the fiber, appearing producing line-of-sight velocity variations of meters or
even centimeters per second. Detecting these velocities signals, previously inaccessible to astronomers. One
via the Doppler shift of the star’s spectral lines requires example is that of photon orbital angular momentum
an extraordinarily precise spectrograph. Large laser (OAM), which measures the field spatial distribution of
frequency combs can produce a precise set of lines the beam distinct from its polarization state. Photonic
with which to calibrate the spectrograph in real time, devices are now being tested to detect this information,
but their huge expense and complexity limits their use although whether the OAM signal survives astronomi-
to only the largest observatories. Cheap, reproducible cal distances is controversial.
photonics, however, offer alternative solutions. One Finally, the many individual photonic devices
is via a fiber etalon, in which the etalon cavity is con- discussed here not only offer important benefits for
structed within a single-mode fiber or waveguide. The astronomy, but can be easily combined and reproduced
small, monolithic design of such an etalon means it can in complex but stable configurations. One could imag-
be easily tuned in a closed loop via simple integrated ine a slew of photonic lanterns, fiber Bragg gratings and
heating elements, and is small and robust enough to be spectrographic elements (such as AWGs) built onto a
deployed at telescopes worldwide. Another approach single chip—and the same applies to interferometry.
is to use a compact ring resonator on a photonic chip This is the future of astronomical instrumentation
to produce such a frequency comb. once all telescopes achieve the goal of being truly
Another emerging application is the use of fiber Bragg diffraction-limited.
gratings to filter out problematic atmospheric spectral We predict that more and more bulk optical systems
lines. When observing the spectra of faint, distant galax- will be replaced by on-chip components, including
ies, signals can be badly contaminated by the relatively variable delay lines, spectral dispersion and even the
bright emission lines (such as OH lines) from our own photodetectors themselves. Devices such as supercon-
atmosphere. To mitigate this, extremely accurate optical ducting, photon-counting, energy-resolving microwave
filters can be constructed by laser-writing a pattern of kinetic inductance detectors (MKIDS) may be integrated
refractive-index modulations into a single-mode fiber with the photonic-chip package, with the whole assem-
core. Fresnel reflections occur at each modulation, and bly operating cryogenically. This will allow ultra-stable,
tuning their spacing and depth thus allows a precisely highly sensitive and easily replicable interferometric
chosen set of narrow wavelengths to be reflected rather instruments to be deployed across the world—enabling
than transmitted through the fiber. This, in turn, permits the direct imaging of planets around distant stars and
the known set of troublesome OH emission lines, for the explosive environments of supermassive black holes
example, to be filtered out before reaching the spectro- at the heart of distant galaxies. OPN
graph—as demonstrated in the GNOSIS and PRAXIS
Barnaby Norris and Joss Bland-Hawthorn (jbh@physics.
instruments at the Anglo-Australian Telescope, Siding
usyd.edu.au) are with Sydney Astrophotonic Instrumentation
Spring Observatory, Australia. Labs, University of Sydney, Australia.
Although not yet in wide use, photonic technologies
can accurately analyze the polarization state of light.
Polarimetry is an important tool in astronomy, as light References and Resources
becomes polarized by gas and dust around stars and as c J. Bland-Hawthorn and P. Kern. “Astrophotonics: A
new era for astronomical instruments,” Opt. Express
it travels through the interstellar medium. Waveguides 17, 1880 (2009).
and couplers can be designed to selectively transmit c R.M. Roettenbacher et al. “No Sun-like dynamo on the
and select specific polarization states, and so photonic- active star ζ Andromedae from starspot asymmetry,”
Nature 533, 217 (2016).
polarimetric instruments are not far off. c J. Bland-Hawthorn and S.G. Leon-Saval. “Astropho-
Moreover, photonic technologies may allow the tonics: Molding the flow of light in astronomical instru-
measurement of completely new types of astronomical ments,” Opt. Express 25, 15549 (2017).