Near-infrared wavefront sensingauthors.library.caltech.edu/72543/1/990915.pdf · wavefront sensing in the near-IR, where the image is partially AO-corrected, offers a potentially

Near-Infrared Wavefront Sensing Peter Wizinowich1,*, Mark Chun2, Dimitri Mawet3, Guido Agapito4, Richard Dekany3, Simone Esposito4, Thierry Fusco5,6, Olivier Guyon7, Donald Hall2, Cedric Plantet5,8, Francois Rigaut9

1. W. M. Keck Observatory, 65-1120 Mamalahoa Hwy., Kamuela, HI, USA 967432. Institute for Astronomy, University of Hawaii, 640 N. A’ohoku Place, Hilo, HI, USA 96720

3. Astronomy Dept., Caltech, 1200 E. California Blvd., Pasadena, CA, USA 911254. Arcetri Astrophysical Observatory, Largo E. Fermi 5, 50125-Firenze, Italy

5. Aix Marseille Univ., CNRS, Laboratoire d’Astrophysique de Marseille UMR 7326, 13388 Marseille, France6. Onera, the French Aerospace Lab, 92322 Chatillon, France

7. Subaru Telescope, National Astronomical Observatory of Japan, 650 N, A’ohoku Place, Hilo, HI, USA 967208. Universite Grenoble Alpes, CNRS, IPAG, F-38000, Grenoble, France

9. Research School of Astronomy & Astrophysics, Australian National University, Canberra, Australia

ABSTRACT

We discuss the advantages of wavefront sensing at near-infrared (IR) wavelengths with low-noise detector technologies that have recently become available. In this paper, we consider low order sensing with laser guide star (LGS) adaptive optics (AO) and high order sensing with natural guide star (NGS) AO. We then turn to the application of near-IR sensing with the W. M. Keck Observatory (WMKO) AO systems for science and as a demonstrator for similar systems on extremely large telescopes (ELTs). These demonstrations are based upon an LGS AO near-IR tip-tilt-focus sensor and our collaboration to implement a near-IR pyramid wavefront sensor (PWFS) for a NGS AO L-band coronagraphic imaging survey to identify exoplanet candidates.

Keywords: adaptive optics, wavefront sensing, tip-tilt sensing, near infrared, pyramid, Keck, LIFT

1. INTRODUCTIONNear-infrared wavefront sensing is a critical technology for science with AO on current and future telescopes. It enables high contrast science of exoplanets and dust obscured regions, and high sky coverage for extragalactic science. It can be used to extend the performance of NGS AO to fainter (redder) targets and to increase the sky coverage of laser guide star (LGS) AO. Furthermore, it allows the application of optimal wavefront sensing approaches (e.g. pyramid and Zernike wavefront sensing) due to the AO correction at near-IR wavelengths. All of the extremely large telescopes (ELTs) are planning to use near-IR wavefront sensing as part of their AO facilities.

Infrared cameras have been used as phase and angle sensors on astronomical interferometers for many years (e.g. [1]). The VLT NACO Shack-Hartmann camera is the only high-order AO near-IR wavefront sensor in operational use [2] and a near-IR PWFS was demonstrated on-sky with the Calar Alto AO system [3]. The limited use of near-IR wavefront sensing has been due to the high read noise, and high cost, of these detectors which limited them to bright targets. The high read noise limitation has recently been removed with the introduction of new approaches and technologies.

In terms of new approaches to reduce read noise, a near-IR tip-tilt sensor, based on multiple reads of small windows on a Teledyne H2RG detector, has been implemented with the Keck I LGS AO system [4]; however there is still a high procurement cost associated with this detector. In terms of new technologies for low read noise and lower cost, both CEA-Leti/Sofradir and Selex have demonstrated the production of electron initiated avalanche photodiode arrays (APDs) [5]. A CEA-Leti/Sofradir RAPID detector is in operational use as part of the VLTI PIONIER instrument [6]. A Selex SAPHIRA array has been demonstrated for lucky imaging on the IRTF [7] and for tip and tilt with the Robo-AO system [8], and for wavefront sensing and fringe tracking on GRAVITY [9].

In this paper we report on the status of lab, on-sky and analytical results for near-IR sensors used for tip-tilt and low order modes for LGS AO, and high order pyramid wavefront sensing for NGS AO (section 2). We also report on the conceptual design of a combined near-IR tip-tilt and PWFS sensor for the Keck II AO system (section 3).

Adaptive Optics Systems V, edited by Enrico Marchetti, Laird M. Close, Jean-Pierre Véran, Proc. of SPIE Vol. 9909, 990915© 2016 SPIE · CCC code: 0277-786X/16/$18 · doi: 10.1117/12.2233035

Proc. of SPIE Vol. 9909 990915-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 12/02/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

2. ANALYSIS AND RESULTS2.1 Selex Detector Performance The initial Selex 320 x 256 at 24 µm pitch e-APD SAPHIRA arrays utilized Liquid Phase Epitaxy (LPE) HgCdTe. A development program initiated by ESO and carried on by UH switched to Metal Organic Vapor Phase Epitaxy (MOVPE) HgCdTe. This technology, which enables full band gap engineering of the APD, has allowed dramatic improvements in the quality and performance of the SAPHIRA arrays [10]. The most recent models have achieved superb cosmetic quality at avalanche gains as high as 600 (at 20 volts bias!). These arrays have 32 parallel video outputs organized as 32 adjacent pixels in a row and, for windowing, the readout allows addressing of any modulo 32 set in the array. The readout is inherently capable of pixel rates well in excess of 10 MHz although the highest rate achieved to date are 5 MHz with the ESO controller. This corresponds to a full frame rate of 2 kHz and 10 KHz for a 128 x 128 sub-array with deep sub-electron read noise. The quantum efficiency is 70% and the mean read noise at 2000 frames per second and a gain of 60 is < 1 e-.

The technology is readily scalable to larger format. Selex and its partners are actively working towards ¼ to one megapixel class formats [11].

2.2 Tip-tilt Sensing for LGS AO The performance of the Keck I near-IR tip-tilt sensor with LGS AO has been measured [12] and some sample measurements are shown in Table 1. The near-IR tip-tilt sensor manages to maintain better tip-tilt performance than the visible tip-tilt sensor as the NGS gets fainter. Both sensors can use stars up to ~ 50″ off-axis. The near-IR sensor uses a 2048x2048 pixel detector and off-axis NGS are acquired by reading out the appropriate small (e.g. 4x4 pixel) region on the detector; while the visible sensor must be moved off-axis to acquire the NGS. A potential advantage of this near-IR sensor is that multiple NGS can be used by reading out multiple regions in order to reduce tip-tilt anisoplanatism (this mode is implemented but needs on-sky testing).

Table 1: Keck LGS AO H-band performance with the visible and near-IR tip-tilt sensors for three different NGS. For these measurements the LGS was pointed at the tip-tilt NGS, while the performance was measured on a nearby star.

Figure 1: Left: A Ks-band image of the Galactic center taken with the Keck I near-IR tip-tilt sensor. Right: A 4x4 pixel (0.1″x0.1″) region of interest anywhere on the detector can be read out non-destructively multiple times to reduce the measurement error on the tip-tilt measurement.

In order to determine the relative merits of the H2RG and e-APDs we can compare their signal-to-noise ratios (SNR).

Strehl FWHM (mas) Strehl FWHM (mas) Strehl FWHM (mas)Visible 0.32 45 0.22 48 0.20 56

Near-IR 4x4 0.35 41 0.29 40 0.36 43

R=12.0 & K=11.1 8.6" off-axis

Tip-tilt Sensor

R=15.5 & K=13.8on-axis

R=13.0 & K=11.56.4" off-axis

102″



= ηSη S + S + S + n σGwhere η is the detector quantum efficiency; S, SB and SD are the signal due to the star, background (sky and thermal) and dark current on the pixels used for the measurement; F is the excess noise caused by the amplification; n is the number of pixels; σR is the read-noise per pixel; and G is the e-APD gain. Note that the calculation of the stellar signal in the n pixels includes a multiplication by the Strehl ratio. The effect of the e-APD gain is to effectively reduce the read-noise at the expense of some excess noise in the detected photons.

A plot of H-band SNR versus magnitude is shown in Figure 2. The e-APD outperforms the H2RG for the 4x4 pixel case. The higher SNR performance for the e-APD system allows an e-APD system to be operated at higher bandwidth than an H2RG system while maintaining the H2RG SNR level. The APD’s performance advantage over the H2RG increases with the number of pixels (e.g. the APD’s can be used for a high order wavefront sensor). There is little performance difference between the two detectors in the 2x2 pixel case.

Figure 2: H-band SNR versus magnitude. Additional assumptions: 49% throughput and Q = 70% for Selex, 45% throughput and Q = and 85% for H2RG, and F =1.2.

2.3 Low Order Focal Plane Wavefront Sensing for LGS AO The LGS wavefront sensor measurements include focus changes due to variations in the altitude of the sodium layer and subaperture tilt changes due to variations in the structure of the sodium layer and the telescope pupil orientation. The Keck LGS AO system [13] uses a visible low bandwidth wavefront sensor (LBWFS) to measure a NGS (i.e. truth) and hence provide focus and centroid corrections to the LGS wavefront sensor. The LBWFS can limit LGS AO performance and observing efficiency especially on faint NGS where 1 to 2 minute exposures can be required.

The errors measured by the LBWFS are low order aberrations (e.g. astigmatism, coma and spherical). Focal plane wavefront sensing in the near-IR, where the image is partially AO-corrected, offers a potentially more efficient alternative to the LBWFS, especially if this could be done with the data from the near-IR tip-tilt sensor.

The focal plane wavefront sensor algorithm LIFT [14] uses a single astigmatic image to estimate tip/tilt, focus and other low order aberrations. LIFT has been tested with calibration sources on the Keck I near-IR tip-tilt sensor and the Keck II science camera NIRC2 [15]. The near-IR tip-tilt sensor’s 50 mas pixels are under-sampled (0.45 Nyquist sampling at K), while the NIRC2 10 mas pixels provide well sampled images. The AO loop was closed on a calibration source and Zernike aberrations modes, in addition to the astigmatism offset needed by LIFT, were inserted by changing the reference slopes on the wavefront sensor. Tip-tilt was successfully retrieved on both instruments (Figure 3a and Figure 4a). Some non-linearities were observed on TRICK for the estimation of focus and 0° astigmatism (Figure 3b and d), but most amplitudes are well estimated. The 45° astigmatism is well estimated in positive values (Figure 3c). The inflection point is expected and corresponds to the opposite of the inserted astigmatism offset. These results can be improved with better knowledge of the imaging model, which is fundamental at such a sampling.

0

5

10

15

20

10 12 14 16

H-ba

nd S

NR

H-band MagnitudeCase: 4x4 pixels, Strehl = 0.06, G = 60

e-APD at 1000 Hze-APD at 300 HzH2RG at 1000 HzH2RG at 300 Hz

For the e-APDs the



Iao -200 -100

0° Ast

-200 -100

:ocas

inserle0 (nm)700

tigmatism

100Inserted (nm)

8 9 R

(sow) mow,s3g g 8

Tilt

7-20 0

Inserted (mos)

45° Astigmz

-100 000Inserted (nm)

40

atism

100 200

$..i...i...i...i...i..

iI

i,,,,

.

Est

imat

ed (

nm)

Figure 3: Esblue plot in

Figure 4: EsThe blue plo

stimation of ra(a) is a center

stimation of rot in (a) is a ce

amps of pure r of gravity est

ramps of pureenter of gravit

Zernike modetimation.

e Zernike modty estimation.

es by LIFT on

des by LIFT

n TRICK. Sam

with NIRC2.

mpling: 0.45 N

Sampling: 1.9

Nyquist sampli

96 Nyquist sa

ing. The

ampling.



1H:fraction-limit

11 12

Centro Id

LIFT

13

Star Magnitude

14 15

The good NIwell retrievedtilt. The coupimpact on skamplitude. TZernike modpupil shapesresults. LIFTmisalignmen

An analysis oa new near-Isection. The integrated tofocus errors w

Figure 5: Wassuming an

As shown insimple centro

Figure 6: NGare obtained

0

50

100

150

200

250

300

350

400

10

TT e

rror

(nm

rms)

IRC2 samplingd within the linpling with tip-t

ky, and thus haThis may also bdes on the defos (using masksT appears to nt and a comple

of tip-tilt and fIR tip-tilt sens

results for a average out swhile detector

Wavefront erron A0 star and

n Figure 6, LIFoiding algorith

GS FWHM ond with the Ke

0 12Star m

g allowed the enearity range (Ftilt is due to a

as not been corbe due to the pormable mirrors in NIRC2 inbe robust to

ex pupil shape)

focus errors versor for Keck I300 Hz sampleeing before bnoise dominat

or as a functioKs-band Stre

FT can also prhm used with th

n the science ceck I near-IR

14magnitude

estimation of loFigure 4), withmisalignment

rrected). Also, pupil misalignr, are being inncluding the K

errors linked ) present in the

rsus star magnII. The analysile rate are sho

being applied ates for faint NG

on of pixel sizehl ratio of 0.2

ovide a signifihe Keck I near-

camera for a ctip-tilt sensor

16

50 ma

32 ma

17 ma

50 ma

32 ma

17 ma

ow orders up toh the other modof the pupil wa scaling facto

nment, but othenvestigated. ThKeck primary

to complex e Keck’s AO sy

nitude has beenis used a similown in Figureas corrections. GS. The aliasin

e (mas) and m5. Left: tip-til

icant tip-tilt pe-IR tip-tilt sens

centroid algorr operating in

as (H)

as (H)

as (H)

as (Ks)

as (Ks)

as (Ks)

0

50

100

150

200

250

300

350

400

10

Focu

s err

or (n

m rm

s)

o 10 modes. Thdes being cons

with the deformor of ~0.8 to 0.er error sourcehese tests havemask) and at imaging mod

ystems.

n performed to lar approach t 5; in practiceFor bright NG

ng is a function

magnitude, in t error. Right

erformance impsor.

ithm and for Ln H-band, wh

0 12Star m

he focus and bstant and close

mable mirror (w9 has to be app

es, such as an e also been per

different wavels (e.g. very

understand theto that of [14] e the focus meGS aliasing domn of Strehl ratio

H or Ks-bandt: focus error.

mprovement for

LIFT. The Kshile the H-ban

14magnitude

both astigmatisto zero, excep

which does not plied to get theincorrect projerformed with d

velengths, withy low samplin

e optimal pixeldescribed in

easurements wminates the tipo.

d, at 300 Hz sa

r faint stars, ve

s-band sciencend science ima

16

50 ma

32 ma

17 ma

50 ma

32 ma

17 ma

ms were pt for tip-

have an e correct ection of different

h similar ng, pupil

l size for the next

would be p-tilt and

ampling

ersus the

e images ages are

as (H)

as (H)

as (H)

as (Ks)

as (Ks)

as (Ks)

at 15th and obtained witth Ks-band ti

116th magnitudp-ee. The Ks-ba

tilt sensing..nnd Strehl ratioo The sampling rate was redd

was 0.25 andduced from 800 Hz at 10th

an A0 star wwas assumed. magnitude too 100 Hz



1000

100

10

1

1

-N___\___

\--.

TurbIl

VI:

10mode number

1 2

time [s]

IR

VIS

3

RS

100

2.4 PyramiOur case forobscured andThe abundandisks at younoptical wavewith current planet to starbeen implemwavefront seFigure 7.

Figure 7: Cmedian conpredicted Nguide stars, performanc

Figure 8: Sierror versusnear-IR (red

0.0

0.2

0.4

0.6

0.8

1.0

10

L-ba

nd S

treh

l Rat

ioid Wavefronr a near-IR pyd red objects, nce of M-type ng ages imply lengths has maAO systems. Tr contrast ratio

mented with Nensing and the

Current and pditions. The cGS AO perforespectively.

e model for th

imulated Kecks mode numbed curve) and v

0 12

R

Current PWFS PrPWFS PrCurrent

nt Sensing foryramid wavefrsuch as youngstars, their lowthat they are c

ade them difficThese stars areo is particularl

NIRC2 and the e resultant pro

predicted L-bacurrent Keckrmance is forRight: Surve

he Keck L-ban

k AO performer for Mauna visible (green c

14 16

R-band Magnitude

Keck NGS AOrediction for M0V Srediction for M6V SKeck LGS AO

r NGS AO ront sensor (PWg stars in distanw close binary common sites cult to observe e however suffly favorable in

Keck II AO bability of pla

and Strehl rak NGS AO syr an efficient ley completenend vortex coro

mance versus Kea turbulen

curve) PWFS

18 2

e

StarStar

WFS) is drivent star formingfraction, and tof planet format the required

fficiently brightn L-band and asystem [16]. T

anet detection

atios versus thstem uses a vow-noise infrass, assuming

onagraph. The

PWFS wavelence (black cur. The plot at r

20

Mas

s (M

Jup)

en by the desirg regions, and the ubiquitous

mation. Unfortud contrast and st in the near-IRa suitable L-baThe limiting mwith the L-ba

he R-magnituvisible Shack-ared PWFS uthe AO perfoe red dot is wh

ength. The plorve) and the Kright shows th

Semi-

re to study plalate-type starspresence of m

unately the fainspatial resolutioR to be used aand vortex cormagnitude gainand vortex cor

de of an on-a-Hartmann w

using M0V andormance showhere Jupiter w

ot at left showKeck AO-correhe K-band Stre

-major Axis (

anet formations, including M

massive proto-pntness of theseon to detect exas AO guide stronagraph has n by going to ronagraph is s

axis guide stawavefront sens

d M6V spectrwn on the left would sit in thi

ws the rms waected residualehl ratio for th

AU)

n around M dwarfs. planetary e stars at xoplanets tars. The recently near-IR

hown in

r under sor. The ral types and the is plot.

avefront l using a he near-

assumed to IR and visibb

be 1e- and a corrected m

le PWFS verssus time. The

odes (250) is llimited by thebright NGS ((

read-noise of400 photons//ms/subapert

existing Keckk AO deform

both the visible (750 nm) uure) was assuaable mirror.

aand near-IR (m ed in both

11.6 µm) detecccases. The nu

ttors was mmber of



400

300hEL

g 200wLL

100

Given the hiavailable. Inivisible PWFSare plotted invisible PWFSerror is 237 advantage wCCD read-no

The predictedA bandwidthvariance of th

where ispropagation assuming a approach of radii of modcalculated vaand β = 7536noise factor = 0.7 were computed as

=where σron = noise and daavalanche gaband therma, = 1photons/s/pix

A demonstraperformed wtesting (see Fat 100 Hz wh

Figure 10: N

igh performancitial PWFS simS for the Keckn Figure 8 aloS even for this

nm with the ill increase whoise is consider

d wavefront erh error of 100 nhe wavefront e

s the number ocoefficients, α

diffraction-lim[17]. These co

dulation of thealues for a mod670. For the Sof F = 1.25 anused. The e

24 e- and ark current foain and is al and sky bac15 photons/s/pxel, respectivel

ation of a near-with the SubaruFigure 10). Hohich severely h

Near-IR pyramSCE

ce of visible Pmulations havek AO system gong with the re

case which asnear-IR senso

hen using typicred.

rror, including nm rms was asestimation error

of photons incand β, were c

mited spot aoefficients ince spot on the dulation of 3λ/

SAPHIRA detend a quantum eequivalent dete

= 504 e-/s aror unit gain, Gthe integration

ckground flux pixel and ly.

-IR (H-band) pu SCExAO sywever, becaus

hampered the o

mid wavefronExAO system w

PWFS it is natue been performiven Mauna Kesultant K-bansumes the samor versus 263cal stars which

measurement sumed at 100 Hr in rad2 is give=

cident on the pomputed for a

and using thecrease with the

pyramid. The/D are α = 536ector an excessefficiency of ηector noise is

, , ,

re the read-outG = 30 is then time. The H-were taken as, = 261

pyramid wavefrystem in Octobe of the limited

on-sky perform

nt sensor laborwith modulati

ural to considemed to understaKea turbulence.nd Strehl ratio

me signal and re nm with a v

h have higher f

and bandwidthHz and was scen by [14] as + pyramid during

maximum likee e e 6 s η s

t e -s

front sensor basber 2015. The d frame rate of

mance.

ratory images ion (left) and w

Figure 9: Cprediction f

er a near-IR PWand the relative The modal tuversus time.

ead-noise; the Avisible sensor. flux in the nea

h errors, is plotaled as Tint5/6 v

,

g an integratioelihood estima

sed on a SAPHsensor perform

f the readout e

of the four ovwithout modu

Combined mfor a Keck PW

WFS now thate merits of a n

urbulence and AThe near-IR PAO-corrected r

The near-IR ar-IR and when

tted in Figure 9versus bandwid

on. For the neaation of 200 Ka

HIRA detector med well in inlectronics the l

verlapping puulation (right).

easurement aWFS with 3λ/D

t the e-APD arnear-IR PWFS AO-corrected rPWFS outperforms residual wsensor’s perfo

n more realisti

9, for a near-IRdth. The total p

ar-IR PWFS, tharhunen-Loeve

and a UH camnitial unmodulloop could onl

pils from the .

and bandwidtD radius modu

rrays are versus a

residuals orms the

wavefront formance c visible

R PWFS. predicted

he noise e modes,

mera was lated lab ly be run

Subaru

th errorulation.



,I,\`

In this sectiowavefront seminimum prLIFT algorith

A dichroic band transmit

3.1 PyramiA conceptuathe f/15 AO distance f1 afdynamic rangbeams each owith each pixcm as project

Figur

According totimes the difffaint guide stdiffraction-liModulation comparator m

It will be impbe a science keep the starand PWFS wlimited diamAn x and y d

Pupil stabilitymodes. The APWFS desigposition at 1

3.2 CombinA conceptuadriven by themm for L1. Ithe results frotilt relay, from

3. A NEon we describensor is being oviding high bhm. In NGS A

beamsplitter juthe light to NI

id Wavefronl view of the poutput to the dfter lens L1, wge of the sensoof which is thexel correspondted on the teles

re 11: Left: Co

o the Subaru vffraction-limitetars. From the mited image dis achieved by

monitoring one

portant to accumode that usesr centered on thwavelengths du

meter should resdc offset added

y is critical andAO rotator can

gn includes capHz rates, suffi

ned Low Ordal layout for a e pupil size on If the SAPHIRom Figure 5 a m AO focus to

EAR-INFRAbe the conceptdesigned to subandwidth tip-O it would be

st before the NIRC2.

nt Sensor Conplanned Keck Pdesired focal ra

where a mirror or. The pyramen imaged on

ding to the 7 mmscope primary

onceptual layo

visible PWFS ed image diameLarge Binocu

diameter [18]. y using two sie of the sine wa

urately positions the PWFS. Thhe pyramid anuring an expossult in no loss oto the two mo

d should be hen be used to mpability to concient to compe

der and Pyracombined NIRthe modulator

RA detector we1-to-1 relay w

o detector, wou

ARED WAVtual design forupport both NG-tilt sensing, wthe high order

NIRC2 science

nceptual LayPWFS layout iatio on the pyrwill be placed

midal prism is lthe detector atm physical defmirror.

out of the PWF

experience, moeter (0.1" for Klar Telescope ( A Physique Iine waves for aves to trigger

n and maintain he modulator m

nd to provide dsure. Offsets oof performancedulating sine w

ld to a fractionmaintain a fixed

ntrol the collimensate for therm

amid WavefrR tip-tilt and Pr and the requirre placed direc

would work weluld be 4f1 = 300

VEFRONTr a near-IR wGS and LGS A

with the intentisensor.

e camera woul

yout is shown in Firamid. The relad to modulate tocated in a foct a pupil planeformable mirro

FS optics. Rig

odulation in a Keck at H-band(LBT) visible Instrumente 33x and y mod

the camera rea

the science tarmirror will therdifferential atmof the star frome and are suffi

waves will be u

n of a pixel on d pupil orientatmating lens cemal variations

ront SensorPWFS is shownred plate scale ctly at the AO ll for the tip-til0 mm.

SENSOR Favefront senso

AO. In LGS AOon to also pro

ld reflect the li

gure 11. Lenseay produces anthe beam on thcal plane and de. Each pupil ior actuator spac

ght: The four p

circular path, wd), will be requPWFS experie30.30 tip-tilt s

dulation. Synchadout at a partic

rget since the Lrefore also be u

mospheric refram the tip of thcient to addres

used to offset th

the camera to ption which wilentering (Lcol and the small n

n in Figure 12on the detectofocus the platelt sensing (f2tt =

FOR KECKor for the KecO it would be

ovide low band

ight to the nea

es L1 and L2 wn intermediate he tip of the pdivides the incmage will be ~cing, which in

pupil images o

with a radius ouired for the PWence the range stage will be uhronization wicular voltage.

L-band vortex cused for small

action correctiohe pyramid by ss differential ahe star from th

prevent instabill help with puin Figure 11) nutations of the

. The tip-tilt seor. For a 5 mm e scale would b= f1) and the ov

K AO ck II AO syste

a low order sedwidth focus u

ar-IR wavefron

will be used topupil image, lyramid to incr

coming beam i~20 pixels in dturn correspon

on the detecto

of approximateWFS to work wis two to six tused for this pill be achieved

coronagraph mpointing corre

on between theup to one diff

atmospheric rehe tip of the pyr

ility in the highupil stability. T

to maintain the telescope pup

ensor optical dpupil size then

be 33 mas/pixeverall length of

em. The ensor, at

using the

nt sensor

o convertlocated a rease the into four diameter nds to 56

r.

ely three well with times the purpose. d with a

mask will ctions to

e science ffraction-efraction. ramid.

her order The Keck

he pupil pil.

design is n f1 = 75 el. Given f the tip-



= Pupil Plane

= Focal Plane

1

M1

Pyrarr

M2

L2p

I

odulator

lid

Lc

M3

L3p

AddOptPyr

M

, L4pM4

L2tt

litional:ics foramid

lode

- Detecto

Figure 12: Ctip-tilt sensosecond lens blue rectangthe beam whproduces a p

Figure 13: KThe sensor wscience instraxis parabol

The L2p focfocal length i

Conceptual opor. A lens L1 cL2tt focuses t

gle. The L2p thile removingpupil image. L

Keck II AO bewould be fed rument. The Nla (3), the defo

al length is chis chosen for th

ptical layout focollimates the the light ontothrough L3p

g L2tt. L2p foL3p and L4p r

ench layout shby a dichroic

NIRC2 scienceormable mirro

hosen for the rhe required pup

for a combinedlight and reim

o the detector.components c

ocuses the lighrelay the pupil

howing the locc beamsplittere path includeor (4), a secon

equired f/36 opil size on the

d near-IR tip-mages the tele. The additioncould be fixedht onto the tipl image onto t

cation of the nr located at thes an imager r

nd off-axis par

n the tip of thdetector; for th

-tilt sensor anescope pupil onnal optics for d and a stage p of a pyramidhe detector.

new combinedhe edge of therotator (item rabola (6) and

he prism (f2p = he case of one

nd PWFS. Then a modulatorthe PWFS mcould insert Md while Lc col

near-IR tip-te AO bench ju1), a fast tip-tan IR transm

180 mm). Thesubaperture (0

e bottom imagr (a tip-tilt miode are show

M1, M4 and Lllimates the li

tilt sensor andust before thetilt mirror (2)

missive dichroi

e collimation l.5625 m) per p

ge is the rror). A

wn in the L4p into ight and

d PWFS. e NIRC2 , an off-ic (7).

lens, Lc, pixel this



requires a 10projection ofthe camera w

The positionwhile providmm from themm. For this

The requiredFigure 13. Acombined nereflection byThis is followbench and tofield dichroiclight transmifeed the NIR

Figure 14: dichroic just

3.3 Camera

Figure 15: Lfrom the fro

0 m / 0.5625 = f the pupil ontowindow.

ns of mirrors Mding the necesse detector. Thes case the pyram

d location for thA dichroic beamear-IR sensor.

y the dichroic, awed by anothe

oward the NIRc to the Selex tted by the nea

RSPEC spectrog

Preliminary Zt before NIRC

a

Left: Universitont mounting

17.78 pixel puo the detector w

M1 to M4 wersary extra pathe M1 to M4 (anmid is 49.7 mm

he combined nmsplitter locate

An initial Zea fold mirror, ler dichroic, jus

R WFS (the trandetector would

ar-IR WFS fielgraph.

Zemax opticaC2 to the Selex

ty of Hawaii csurface at left

upil (fc = 15.36with a back foc

re chosen to prh length. In thend M2 to M3) m after M3 (as

near-IR tip-tilt ed just in fronemax optical dlarge enough tost large enoughnsmitted light d be on a xy-tld dichroic wil

al design for x detector.

camera dewart. The overall

6 mm). L3p anal distance of 6

rovide a reasoe selected layoudistance is theopposed to wh

sensor and PWnt of the NIRCdesign of the o cover the fulh for the NIR will go to a fi

translation stagl go to a single

the NIR tip-

r (orange) from

package lengt

nd L4p were ch60 mm since th

onable separatiut M1 is 30 m

erefore 40 mm,hat is shown in

WFS on the exiC2 science instr

tip-tilt sensor ll field, folds thWFS field, tha

iber injection uge to access thee mode fiber in

-tilt sensor sh

m GL Scientifth is 457 mm

hosen to providhe detector is lo

ion from the mmm after the m

, and the M1 toFigure 12).

isting Keck II rument would path is show

he beam down at folds the beunit). Everythine full 120 arcsnjection unit (P

howing the op

fic. The detectincluding the

de a unit magnocated ~40 mm

modulator and odulator and Mo M2 distance

AO bench is sreflect the fiel

wn in Figure 1toward the AO

eam parallel tong from the Nsec diameter fiPI: D. Mawet)

ptical path fr

or is located 5cryocooler ex

nification m behind

detector M4 is 80

is 240.4

shown in ld to the 4. After

O bench. o the AO

NIR WFS ield. The that will

rom the

50.5 mm xtending

to the right. Right: First Light Inc.’s C-Red One camera head is 341 mm long x 250 mm high x 215 mm wide; the entrance window is 65 mm from the bottom of the mounting surface.

Proc. of SPIE Vol. 9909 990915-10


Other Systems

Telescopecontrol

Laser

control

ScienceInstruments

Rest of AO(OperationsSoftware,

Supervisory& OpticsBench

Control,LTCS, TBAD,

SATS...)

TelemetryRecording System

AL

Keck RTC

VxWorksMicrogateReal -Time

Controller

Linux

EPICS

I/F

MGP Driver

will continue to be determined by the operational system and will be passed as a calibration file to the PWFS RTC; and (3) pointing offsets to register the science target on the vortex coronagraph and compensate for differential atmospheric refraction will continue to be determined by the operational system and will be passed to the modulator driver.

4. CONCLUSIONS Wavefront sensing with the low noise near-IR detectors that have recently become available offers considerable sky coverage improvements for both LGS and NGS AO. In LGS mode these sensor can be used for tip-tilt as well as for measuring the low order aberrations introduced by the LGS. For high order NGS AO the low noise and flexible readout of the Selex SAPHIRA e-APD arrays is particularly optimal and supports improved AO correction.

A combined near-IR tip-tilt and pyramid wavefront sensor design is described which would support Keck II LGS and NGS AO with the NIRC2 and NIRSPEC science instruments. The tip-tilt sensor mode is being designed to support the determination of low order modes with the LIFT algorithm that has shown promising laboratory results.

ACKNOWLEDGEMENTS

The W. M. Keck Observatory is operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W. M. Keck Foundation. The Keck I near-IR tip-tilt sensor was supported by the National Science Foundation under Grant No. AST-1007058 and by the Gordon and Betty Moore Foundation under Grant No. 4046.

The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of Maunakea has always had within the indigenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain.

REFERENCES

[1] Colavita, M., Wizinowich, P., Akeson, R., Ragland, S., Woillez, J., Millan-Gabet, R., Serabyn, E., et al., “The Keck Interferometer,” PASP 125, 1226-1264 (2013).

[2] Gendron, E., Lacombe, F., Rouan, D., Charton, J., Collin, C., Lefort, B., Marlot, C., Michet, G., Nicol, G., Pau, S., Phan, V., Talureau, B., Lizon, J.-L., Hubin, N., “NAOS infrared wavefront sensor design and performance,” Proc. SPIE 4839, 195 (2003).

[3] Peter, D., Feldt, M., Henning, T., Hippler, S., Aceituno, J., Montoya, L., Costa, J., Dorner, B., “PYRAMIR: Exploring the On-Sky Performance of the World’s First Near-Infrared Pyramid Wavefront Sensor,” PASP 122, 63 (2010).

[4] Femenia Castella, B., Wizinowich, P., Rampy, R., Cetre, S., Lilley, S., Lyke, J., Ragland, S., Stomski, P., van Dam, M., “Status and new developments with the Keck I near-infrared tip-tilt sensor,” Proc. SPIE 9909-79 (2016).

[5] Feautrier, P., Gach, J.-L., Wizinowich, P., “State of the art IR cameras for wavefront sensing using e-APD MCT arrays,” Adaptive Optics for Extremely Large Telescopes Conference 4 (Los Angeles, 2015).

[6] Guieu, S., Feautrier, P., Zins, G., Le Bouquin, J. -B., Stadler, E., Kern, P., Rothman, J., Tauvy, M., Coussement, J., de Borniol, E., Gach, J. -L., Jacquard, M., Moulin, T., Rochat, S., Delboulb, A., Derelle, S., Robert, C., Vuillermet, M., Mérand, A., Bourget, P., “RAPID, a revolutionary fast optical to NIR camera applied to interferometry,” Proc. SPIE 9146, 91461N (2014).

[7] Atkinson, D., Hall, D., Baranec, C., Baker, I., Jacobson, S., Riddle, R., “Observatory Deployment and Characterization of SAPHIRA HgCdTe APD Arrays,” Proc. SPIE 9154, 914519 (2014).

[8] Baranec, C., Atkinson, D., Riddle, R., Hall, D., Jacobson, S., Law, N., Chun, M., “High-speed Imaging and Wavefront Sensing with an Infrared Avalanche Photodiode Array,” ApJ 809, 70 (2015).

[9] Finger, G., Baker, I., Alvarez, D., Ives, D., Mehrgan, L., Meyer, M., Stemeier, J., Weller, H., “SAPHIRA detector for infrared wavefront sensing,” Proc. SPIE 9148, 914817 (2014).

[10] Atkinson, D. et al., “Next-generation performance of SAPHIRA HgCdTe APDs,” Proc. SPIE 9915-22 (2016).

Proc. of SPIE Vol. 9909 990915-12


[11] Hall, D. et al., “Towards the next generation of L-APD MOVPE HgCdTe arrays: beyond the SAPHIRA 320x256,”Proc. SPIE 9915-21 (2016).

[12] Rampy, R., Femenia, B., Lyke, J., Wizinowich, P., Cetre, S., Ragland, S., Stomski, P., “Near-infrared tip-tiltsensing at Keck: System architecture and on-sky performance,” Adaptive Optics for Extremely Large TelescopesConference 4 (Los Angeles, 2015).

[13] Wizinowich, P., Le Mignant, D., Bouchez, A., Campbell, R., Chin, J., Contos, A., van Dam, M., Hartman, S.,Johansson, E., Lafon, R., Lewis, H., Stomski, P., Summers, D., Brown, C., Danforth, P., Max, Pennington, D.,“The W. M. Keck Observatory Laser Guide Star Adaptive Optics System: Overview,” PASP 118, 310 (2006).

[14] Meimon, S., Fusco, T., Mugnier, L., “LIFT: a focal-plane wavefront sensor for real-time low-order sensing on faintsources,” Optics Letters 35, 3036 (2010).

[15] Plantet, C., Wizinowich, P., Fusco, T., Neichel, B, Ragland, S., Rampy, R., “Lift on Keck: analysis of performanceand first experimental results,” Proc. SPIE 9909-233 (2016).

[16] Femenia Castella, B. et al., “Commissioning and first light results of an L’-band coronagraph with the Keck IIadaptive optics NIRC2 science instrument,” Proc. SPIE 9909-76 (2016b).

[17] Plantet. C., Meimon, S., Conan, J.-M., Fusco, T., “Revisiting the comparison between the Shack-Hartmann and thepyramid wavefront sensors via the Fisher information matrix,” Optics Express 23, 28619 (2015).

[18] Esposito, S., Riccardi, A., Pinna, E., Puglisi, A., Quiros-Pacheco, F., Arcidiacono, C., Xompero, M., Briguglio, R.,Agapito, G., Busoni, L., Fini, L., Argomedo, J., Gherardi, A., Brusa, G., Miller, D., Guerra, J., Stefanini, P.,Salinari, P., “Large Binocular Telescope Adaptive Optics System: New achievements and perspectives in adaptiveoptics,” Proc. SPIE 8149, 814902 (2011).

[19] Gach, J.-L., Feautrier, P., Stadler, E., Greffe, T., Clop, F., Lemarchand, S., Carmignani, T., Boutolleau, D., Baker,I., “C-RED one: ultra-high speed wavefront sensing in the infrared made possible,” Proc. SPIE 9909-41 (2016).

[20] Lockhart, C. et al., “The PizzaBox-1 array controller: a high speed low noise array controller for Teledyne H2RG,H4RG, and Selex SAPHIRA devices,” Proc. SPIE 9915-104 (2016).

[21] Johansson, E., van Dam, M., Stomski, P., Bell, J., Chin, J., Sumner, R., Wizinowich, P., Biasi, R., Andrighettoni,M., & Pescoller, D., “Upgrading the Keck AO wavefront controllers,” Proc. SPIE 7015, 7015-121 (2008).

Proc. of SPIE Vol. 9909 990915-13


Documents

Near-infrared wavefront sensingauthors.library.caltech.edu/72543/1/990915.pdf · wavefront sensing in the near-IR, where the image is partially AO-corrected, offers a potentially