The Project
Current capabilities
Future capabilities
The Team

Welcome to the SCExAO project webpage!

Pulpit rock Pulpit rock Pulpit rock

Image of various components in SCExAO: (Left) 2000 actuator deformable mirror (middle) IR/visible dichroic splitter, fixed pupil
mask and delivery fiber for internal calibration source (foreground) (Right) Starbowl wheel.


The SCExAO acronym stands for Subaru Telescope Extreme Adaptive Optics project. It is an active, ongoing effort to equip the Subaru Telescope with a high performance coronagraph and a series of wavefront control solutions that make an optimal use of the angular resolution that an 8-meter telescope has to offer. The ultimate science goal of SCExAO is the direct imaging of extrasolar planets around stars at a separation corresponding to the diffraction limit of the telescope in the near infrared, more specifically in the y-K-bands, at 0.95-2.4 μm. Two other large scale projects with comparable science goals are currently being assembled: The Gemini Planet Imager (GPI) that is currently being commissioned at the Gemini South Telescope and the SPHERE for the VLT. Instead of competing with these two heavyweight projects, SCExAO is trying to offer complementary capabilities, by focusing on very small angular separation: 40 - 500 milli-arcseconds.

SCExAO can do this because it implements several a high efficiency, low inner working angle coronagraphs. The workhorse of these units is the PIAA (Phase Induced Amplitude Apodization), invented by Olivier Guyon (SCExAO project PI), that exhibits an inner working angle that is as close to the diffraction limit as you can get (1 λ/D). The PIAA used in SCExAO was designed to provide a raw-contrast of 106 at 1.5 λ/D. In addition SCExAO offers other low inner working angle coronagraphs such as the Vortex, Four-Quadrant Phase Mask and 8-Octant Phase Mask versions. In addition it also offers a shaped pupil coronagraph for high contrast work where the inner working angle can be relaxed.

While reaching such a high level of contrast has already been demonstrated in the well controlled environment of the laboratory, it is unlikely to be achieved at the telescope, as wavefront aberrations induced by the atmosphere and the optics of the telescope quickly degrade image quality. To remedy this, SCExAO is/has developed a series of appropriate measures, for wavefront sensing and control. SCExAO uses:

  • The Subaru Telescope facility AO system called AO188, operating at visible wavelengths, to provide the initial wavefront correction, which considerably reduces the strain on the SCExAO components.

  • A Coronagraphic Low-Order Wavefront Sensor (CLOWFS) for ultra-fine pointing control, operating in the IR and using the light rejected by the focal plane mask of the coronagraph to fine tune the tip/tilt.

  • A visible pyramid-based High Order Wavefront sensor (only available in the second phase of the project) to cancel out the dynamic atmospherically induced aberrations responsible for the presence of fast speckles.

  • Speckle probing techniques, that use commands sent to a deformable mirror to "probe" the speckles the coherence of the speckles in the field, and eventually erase them. This also offers the ability for precise astrometry.

SCExAO has been undergoing commissioning of various modules since 2011. Most runs have been hampered by weather but the few that haven't, have been very successful. The modules/features that have been commissioned and can now be offered for open use are listed in the following section. Beneath this is a list of the modules we are currently engineering and hope to offer in the not to distant future.

Pulpit rock

Image of SCExAO mounted at the Nasmyth IR platform. AO188 can be seen to the left (large black box), SCExAO is the
dual level instrument in the middle (black and white panels) and HiCIAO is to the right of the image (Blue instrument).

Science ready Capabilities (Can be used from S14B onwards):

The SCExAO instrument is continuously growing and is therefore being commissioned in phases. As of January 2014 only phase I engineering has been completed and therefore only the modes tested within that phase can be used for science. These include the following:

  • Coronagraphs: The SCExAO instrument is equipped with several coronagraphs, which operate across the near-IR (y-K bands). Those that can be used include the Phase induced amplitude apodisation (PIAA), Vector Vortex, Four-Quadrant Phase Mask and Shaped pupil versions.

  • Tip/tilt (low order) control: The low order wavefront sensor (LOWFS) uses the rejected light from the coronagraphs in the near-IR to fine-tune the tip/tilt for optimum starlight rejection.

  • Coherent differential imaging: The speckle nulling routine allows for quasi-static speckles due to aberrations in the optics, or diffraction from the spiders and obstructions to be systematically beaten down. The performance of this mode is limited by the current generation of cameras (See below).

  • High frame rate imaging: The internal near-IR camera in SCExAO can run at up to 346 Hz offering the possibility to collect fast frames for lucky imaging or advanced PSF calibration of long exposures taken with HiCIAO.

  • Deep exposure imaging: HiCIAO is used to take longer exposures of a given target once it has passed through SCExAO. It can be used in conjunction with all the other modes described above or separate to them.

Please note SCExAO is a booster stage for the initial AO correction provided by AO188. In good seeing AO188 offers 20-40% Strehl ratios in the H-band. Please use this value when preparing proposals.

SCExAO Phase I capabilities and performance summary table

Target acquisition and Image quality
Strehl Ratio, FWHM See AO188 performance SCExAO phase I provides no significant
improvement in image quality over AO188
Target Acquisition and
setup overhead
AO188 target acquisition time + 10 mn Does not include fine tuning of speckle control loop
Seeing limit ~1.2" See AO188 performance
Faint limit R=8, H=5 Performance improves steeply with brighter targets
Bright limit N/A Neutral Density filters can be placed in front of wavefront
sensor and HiCIAO when observing bright targets
Cameras: HiCIAO
Detector Hawaii 2RG See HiCIAO instrument description
Wavelength coverage See HiCIAO instrument description SCExAO feeds HiCIAO with wavelengths longer than 950nm
Field-of-view, plate scale nominal HiCIAO values SCExAO preserves HiCIAO's plate scale and FOV
SCExAO throughput to HiCIAO 35 % Assumes 50%/50% beam splitter between HiCIAO and
SCExAO science camera, does not include PIAA lenses
Coronagraph modes PIAA, Vortex, 4QPM,
Shaped Pupil
Angular Differential Imaging YES SCExAO operates in fixed pupil mode
Active Speckle control YES Speckle control performed using SCExAO
internal science camera
Tip-Tilt control YES Tip-Tilt control performed using SCExAO
Low Order Wavefront Sensor
Spectral Differential Imaging NO Will be offered at a later time
Polarimetric imaging NO Will be offered at a later time
Cameras: Internal SCExAO science camera
Detector InGaAs, 320x256 Axiom Optics OWL SW1.7HS
Wavelength coverage J, H band 0.9-1.7 μm
Field-of-View 3.4"x 2.7" 10.76 mas/pixel
Frame rate 346 Hz max adjustable
Readout Noise 115 e-
Phase I High Contrast Imaging Performance
Inner Working Angle 1 to 3 l/D Function of coronagraph configuration
PIAA Coronagraph optics throughput 52%
Detection contrast threshold:
1-4 λ/D (40-160 mas)
1e-4 H band, Approximate value, assumes bright star
and PSF post-calibration
Detection contrast threshold:
4-20 λ/D (160-800 mas)
1e-5 H band, Approximate value, assumes bright star
and PSF post-calibration
Phase II High Order Wavefront Correction
Extreme AO HOWFS development is progressing well and we should
be able to offer additional wavefront correction on top of
AO188 in S14B. However, at this point we can not commit to a Strehl ratio.


Coronagraph type PIAA Vortex 4QPM Shaped pupil
Inner working angle (λ/D) 1.5 1 1 3
Waveband(s) y-K H H y-K


Speckle nulling is performed with the camera used for high frame rate imaging. This camera is the same as the one used for the LOWFS. They are both Axiom Optics OWL SW1.7HS units, which have an InGaAs CMOS array with 320x256 pixels that are 30x30 μm in size. These cameras can run full frame at 346 Hz and have a read noise of 115 and 140 e- respectively. The dark current is such that the maximum exposure is ~5-10 s. Please note that the speckle nulling routine operates on the speckles which are ~1000 x fainter than the central PSF and hence this must be taken into consideration when selecting your target, in light of the noise characteristics. The high frame rate/speckle nulling camera has a plate scale of 10.76 mas/pixel and a maximum field-of-view of 3.4"x 2.7". Please refer to the HiCIAO page for information on that camera. The plate scale of HiCIAO when used in conjunction with SCExAO is preserved. The high frame rate/speckle nulling camera can be used with a series of bandpass filters which include: y-band, J-band, H-band and 50 nm bandwidth centered at 1650 nm.

Pulpit rock Pulpit rock Pulpit rock

(Left) High frame rate science camera/speckle nulling camera. Mounted on a long translation stage to move between
the focal and pupil planes. (Middle) LOWFS camera. (Right) HiCIAO camera.

Calibration source

The internal calibration source consists of a Fianium super continuum source and 2 laser diodes centered at 650 and 1550 nm. The light from the source is delivered via an endlessely single-mode fiber to the bench ensuring a diffraction-limited calibration PSF from the visible to K-band. The brightness and spectral bands of the injected light can be easily controlled. In addition there is a turbulence simulator that can be used for realistic simulations of on-sky performance in the laboratory.

Pulpit rock Pulpit rock Pulpit rock

(Left) Internal calibration source unit (a.k.a rainbow maker). (Middle) Super continuum laser source attached to source box. (Right) Internals of
calibration source box. There is a shutter to turn the light on/off, and three wheels to control the ND in the vis/IR and spectral content.

Hot News:

April 2014:
The April observing run was a great success. We revalidated speckle nulling and the LOWFS on-sky with convincing results and VAMPIRES collected a ton of data following its upgrades as well. The icing on the cake was the demonstration of closed loop operation of the high order wavefront sensor (non-modulated pyramid wavefront sensor). With the implementation of 5 GPU's since the December run, we managed to drive the PyWFS at 800 Hz on 200 modes on-sky. We tested the loops performance to cancel aberrations by adding varying static wavefront errors via our deformable mirror on top of the pre-existing aberrations and then watched the restoration of the PSF and the applied volt map post correction. Next steps include adding regularization to prevent actuators in and around the pupil edges from saturating, eliminating existing delays in the loop and applying off-sky response matrices to attempt to get extreme AO correction on the coming June run. Stay tuned! In addition we observed significant improvements in PSF stability as a result of the HiCIAO vibration isolation upgrades. Great success!

March 2014:
The HiCIAO camera was overhauled to include a dual bellows vibration isolation unit. More information can be found on the vibrations page.

Pulpit rock

(Left) The new dual bellows vibration isolation system installed on HiCIAO.

January 2014:
We received funding to purchase two state-of-the-art cameras, namely OCAM 2K and MKIDS. OCAM 2K (FirstLight) is one of the most advanced visible cameras (frame rate of 3.7 kHz with binning, 0.3e- read noise, <0.01e- dark current and very high quantum efficiency between 500 and 900 nm). This camera will be used to replace the Andor Zyla currently used for the non-modulated pyramid wavefront sensor and boost performance significantly. This camera is expected to be delivered in late May 2014 and installed and running a few months after this date. MKIDS stands for Microwave Kinetic Inductance Detector and is a photon counting, energy discriminating superconducting based technology. It is being developed by Prof. Ben Mazin's group at UC Santa Barabara and we are fortunate enough to work with them developing a specific unit for SCExAO. This camera will replace the current high frame rate/speckle nulling camera. With the boost in sensitivity and speed, we will be able to do fast wavefront control and calibration and push speckle nulling and even focal plane wavefront sensing to the limit. MKIDS will take several years to build and test.

Pulpit rock Pulpit rock Pulpit rock

(Left) OCAM 2k (Image taken by FirstLight Advanced Imaging). (Middle) MKIDS detector.
(Right) Prof. Ben Mazin with the entire MKIDS camera assembly (Images taken by Mazin's group at UCSB).