Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/iop-publishing/the-science-performance-of-jwst-as-characterized-in-commissioning-37o2D4uvty
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- IOP Publishing
- Copyright
- © 2023. The Author(s). Published by IOP Publishing Ltd on behalf of the Astronomical Society of the Pacific (ASP). All rights reserved
- ISSN
- 0004-6280
- eISSN
- 1538-3873
- DOI
- 10.1088/1538-3873/acb293
- Publisher site
- See Article on Publisher Site
Abstract
Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April https://doi.org/10.1088/1538-3873/acb293 © 2023. The Author(s). Published by IOP Publishing Ltd on behalf of the Astronomical Society of the Pacific (ASP). All rights reserved 1,46 2 1 1 2 2 3 Jane Rigby , Marshall Perrin , Michael McElwain , Randy Kimble , Scott Friedman , Matt Lallo , René Doyon , 1 4 5 6 6 5 7 Lee Feinberg , Pierre Ferruit , Alistair Glasse , Marcia Rieke , George Rieke , Gillian Wright , Chris Willott , 1 1 1 1 2 1 2 2 Knicole Colon , Stefanie Milam , Susan Neff , Christopher Stark , Jeff Valenti , Jim Abell , Faith Abney , Yasin Abul-Huda , 8 2 2 2 2 3 6 2 D. Scott Acton , Evan Adams , David Adler , Jonathan Aguilar , Nasif Ahmed , Loïc Albert , Stacey Alberts , David Aldridge , 2 10 11 4 1 2 Marsha Allen , Martin Altenburg , Javier Álvarez-Márquez , Catarina Alves de Oliveira , Greg Andersen , Harry Anderson , 2 12 2 11 3 2 13 Sara Anderson , Ioannis Argyriou , Amber Armstrong , Santiago Arribas , Etienne Artigau , Amanda Arvai , Charles Atkinson , 2 2 1 2 2 9 2 Gregory Bacon , Thomas Bair , Kimberly Banks , Jaclyn Barrientes , Bruce Barringer , Peter Bartosik , William Bast , 14 6 2 2 2 1 4 Pierre Baudoz , Thomas Beatty , Katie Bechtold , Tracy Beck , Eddie Bergeron , Matthew Bergkoetter , Rachana Bhatawdekar , 4 2 2 14 4 2 15 Stephan Birkmann , Ronald Blazek , Claire Blome , Anthony Boccaletti , Torsten Böker , John Boia , Nina Bonaventura , 1 2 1 2 16 2 1 Nicholas Bond , Kari Bosley , Ray Boucarut , Matthew Bourque , Jeroen Bouwman , Gary Bower , Charles Bowers , 2 2 2 2 17 2 2 Martha Boyer , Larry Bradley , Greg Brady , Hannah Braun , David Breda , Pamela Bresnahan , Stacey Bright , 2 2 2 2 8 2 2 Christopher Britt , Asa Bromenschenkel , Brian Brooks , Keira Brooks , Bob Brown , Matthew Brown , Patricia Brown , 18 2 2 1 18 9 2 Andy Bunker , Matthew Burger , Howard Bushouse , Steven Cale , Alex Cameron , Peter Cameron , Alicia Canipe , 2 2 2 8 19 8 2 James Caplinger , Francis Caputo , Mihai Cara , Larkin Carey , Stefano Carniani , Maria Carrasquilla , Margaret Carruthers , 2 2 2 2 20 2 2 2 Michael Case , Riggs Catherine , Don Chance , George Chapman , Stéphane Charlot , Brian Charlow , Pierre Chayer , Bin Chen , 2 2 2 8 1 1 2 2 Brian Cherinka , Sarah Chichester , Zack Chilton , Taylor Chonis , Mark Clampin , Charles Clark , Kerry Clark , Dan Coe , 2 1 2 1 1 2 8 Benee Coleman , Brian Comber , Tom Comeau , Dennis Connolly , James Cooper , Rachel Cooper , Eric Coppock , 2 21 14 8 2 22 9 Matteo Correnti , Christophe Cossou , Alain Coulais , Laura Coyle , Misty Cracraft , Mirko Curti , Steven Cuturic , 2 1 1 1 12 2 2 Katherine Davis , Michael Davis , Bruce Dean , Amy DeLisa , Wim deMeester , Nadia Dencheva , Nadezhda Dencheva , 2 2 16 2 5 2 2 Joseph DePasquale , Jeremy Deschenes , Örs Hunor Detre , Rosa Diaz , Dan Dicken , Audrey DiFelice , Matthew Dillman , 2 2 2 2 2 23 1 William Dixon , Jesse Doggett , Tom Donaldson , Rob Douglas , Kimberly DuPrie , Jean Dupuis , John Durning , 2 2 2 6 10 2 24 Nilufar Easmin , Weston Eck , Chinwe Edeani , Eiichi Egami , Ralf Ehrenwinkler , Jonathan Eisenhamer , Michael Eisenhower , 2 2 2 2 2 2 2 25 Michelle Elie , James Elliott , Kyle Elliott , Tracy Ellis , Michael Engesser , Nestor Espinoza , Odessa Etienne , Mireya Etxaluze , 2 2 26 2 2 2 2 Patrick Falini , Matthew Feeney , Malcolm Ferry , Joseph Filippazzo , Brian Fincham , Mees Fix , Nicolas Flagey , 6 13 2 1 2 2 1 13 Michael Florian , Jim Flynn , Erin Fontanella , Terrance Ford , Peter Forshay , Ori Fox , David Franz , Henry Fu , 2 2 1 4 1 2 Alexander Fullerton , Sergey Galkin , Anthony Galyer , Macarena García Marín , Jonathan P. Gardner , Lisa Gardner , 2 2 12 6 23 2 5 Dennis Garland , Bruce Garrett , Danny Gasman , Andras Gaspar , Daniel Gaudreau , Peter Gauthier , Vincent Geers , 1 2 27 2 2 2 28 Paul Geithner , Mario Gennaro , Giovanna Giardino , Julien Girard , Mark Giuliano , Kirk Glassmire , Adrian Glauser , 1 2 2 2 9 2 8 Stuart Glazer , John Godfrey , David Golimowski , David Gollnitz , Fan Gong , Shireen Gonzaga , Michael Gordon , 2 2 29 1 8 2 Karl Gordon , Paul Goudfrooij , Thomas Greene , Matthew Greenhouse , Stefano Grimaldi , Andrew Groebner , 25 20 2 1 17 13 6 Timothy Grundy , Pierre Guillard , Irvin Gutman , Kong Q. Ha , Peter Haderlein , Andria Hagedorn , Kevin Hainline , 9 2 2 30 2 13 2 2 Craig Haley , Maryam Hami , Forrest Hamilton , Heidi Hammel , Carl Hansen , Tom Harkins , Michael Harr , Jessica Hart , 2 2 13 1 2 1 2 Quyen Hart , George Hartig , Ryan Hashimoto , Sujee Haskins , William Hathaway , Keith Havey , Brian Hayden , 2 2 2 2 10 2 2 Karen Hecht , Chris Heller-Boyer , Caroline Henriques , Alaina Henry , Karl Hermann , Scarlin Hernandez , Brigette Hesman , 8 2 2 2 2 2 2 13 Brian Hicks , Bryan Hilbert , Dean Hines , Melissa Hoffman , Sherie Holfeltz , Bryan J. Holler , Jennifer Hoppa , Kyle Hott , 1 31 2 2 2 16 2 Joseph M. Howard , Rick Howard , Alexander Hunter , David Hunter , Brendan Hurst , Bernd Husemann , Leah Hustak , 23 32 1 13 1 33 2 Luminita Ilinca Ignat , Garth Illingworth , Sandra Irish , Wallace Jackson , Amir Jahromi , Peter Jakobsen , LeAndrea James , 1 2 2 2 2 2 2 Bryan James , William Januszewski , Ann Jenkins , Hussein Jirdeh , Phillip Johnson , Timothy Johnson , Vicki Jones , 1 2 5 2 2 8 1 2 Ron Jones , Danny Jones , Olivia Jones , Ian Jordan , Margaret Jordan , Sarah Jurczyk , Alden Jurling , Catherine Kaleida , 2 2 2 2 2 34 6 Phillip Kalmanson , Jens Kammerer , Huijo Kang , Shaw-Hong Kao , Diane Karakla , Patrick Kavanagh , Doug Kelly , 4 2 2 1 2 2 2 Sarah Kendrew , Herbert Kennedy , Deborah Kenny , Ritva Keski-kuha , Charles Keyes , Richard Kidwell , Wayne Kinzel , 1 13 2 5 2 8 13 Jeff Kirk , Mark Kirkpatrick , Danielle Kirshenblat , Pamela Klaassen , Bryan Knapp , J. Scott Knight , Perry Knollenberg , 2 2 2 2 35 2 2 Robert Koehler , Anton Koekemoer , Aiden Kovacs , Trey Kulp , Nimisha Kumari , Mark Kyprianou , Stephanie La Massa , 2 11,36 21 2 2 2 Aurora Labador , Alvaro Labiano , Pierre-Olivier Lagage , Charles-Philippe Lajoie , Matthew Lallo , May Lam , 2 1 2 2 35 2 1 5 Tracy Lamb , Scott Lambros , Richard Lampenfield , James Langston , Kirsten Larson , David Law , Jon Lawrence , David Lee , 6 2 2 2 17 2 26 2 Jarron Leisenring , Kelly Lepo , Michael Leveille , Nancy Levenson , Marie Levine , Zena Levy , Dan Lewis , Hannah Lewis , 35 8 2 2 13 2 2 2 Mattia Libralato , Paul Lightsey , Miranda Link , Lily Liu , Amy Lo , Alexandra Lockwood , Ryan Logue , Chris Long , 2 2 37 4 13 2 Douglas Long , Charles Loomis , Marcos Lopez-Caniego , Jose Lorenzo Alvarez , Jennifer Love-Pruitt , Adrian Lucy , 1 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. 4 1 22 2 2 1 35 Nora Luetzgendorf , Peiman Maghami , Roberto Maiolino , Melissa Major , Sunita Malla , Eliot Malumuth , Elena Manjavacas , 2 2 4 2 10 2 Crystal Mannfolk , Amanda Marrione , Anthony Marston , André Martel , Marc Maschmann , Gregory Masci , 8 23 1 1 2 2 Michaela Masciarelli , Michael Maszkiewicz , John Mather , Kenny McKenzie , Brian McLean , Matthew McMaster , 8 2 1 2 2 13 1 Katie Melbourne , Marcio Meléndez , Michael Menzel , Kaiya Merz , Michele Meyett , Luis Meza , Cherie Miskey , 6 1 6 2 38 1 30 Karl Misselt , Christopher Moller , Jane Morrison , Ernie Morse , Harvey Moseley , Gary Mosier , Matt Mountain , 8 39 2 8 2 2 40 Julio Mueckay , Michael Mueller , Susan Mullally , Jess Murphy , Katherine Murray , Claire Murray , David Mustelier , 2 17 13 2 13 2 James Muzerolle , Matthew Mycroft , Richard Myers , Kaila Myrick , Shashvat Nanavati , Elizabeth Nance , 2 17 2 2 1 2 Omnarayani Nayak , Bret Naylor , Edmund Nelan , Bryony Nickson , Alethea Nielson , Maria Nieto-Santisteban , 2 2 2 4 1 2 Nikolay Nikolov , Alberto Noriega-Crespo , Brian O’Shaughnessy , Brian O’Sullivan , William Ochs , Patrick Ogle , 2 2 2 1 2 2 2 Brenda Oleszczuk , Joseph Olmsted , Shannon Osborne , Richard Ottens , Beverly Owens , Camilla Pacifici , Alyssa Pagan , 2 24 1 28 2 2 2 2 James Page , Sang Park , Keith Parrish , Polychronis Patapis , Lee Paul , Tyler Pauly , Cheryl Pavlovsky , Andrew Pedder , 2 2 17 2 11 2 1 Matthew Peek , Maria Pena-Guerrero , Konstantin Penanen , Yesenia Perez , Michele Perna , Beth Perriello , Kevin Phillips , 13 13 35 41 9 2 2 Martin Pietraszkiewicz , Jean-Paul Pinaud , Norbert Pirzkal , Joseph Pitman , Aidan Piwowar , Vera Platais , Danielle Player , 2 2 2 2 2 2 2 Rachel Plesha , Joe Pollizi , Ethan Polster , Klaus Pontoppidan , Blair Porterfield , Charles Proffitt , Laurent Pueyo , 2 2 2 2 2 8 1 Christine Pulliam , Brian Quirt , Irma Quispe Neira , Rafael Ramos Alarcon , Leah Ramsay , Greg Rapp , Robert Rapp , 1 2 4 2 1 42 Bernard Rauscher , Swara Ravindranath , Timothy Rawle , Michael Regan , Timothy A. Reichard , Carl Reis , 17 2 13 2 1 4 2 Michael E. Ressler , Armin Rest , Paul Reynolds , Timothy Rhue , Karen Richon , Emily Rickman , Michael Ridgaway , 2 16 2 31 2 2 Christine Ritchie , Hans-Walter Rix , Massimo Robberto , Gregory Robinson , Michael Robinson , Orion Robinson , 2 2 11 29 1 2 Frank Rock , David Rodriguez , Bruno Rodriguez Del Pino , Thomas Roellig , Scott Rohrbach , Anthony Roman , 2 2 13 13 2 9 2 Fred Romelfanger , Perry Rose , Anthony Roteliuk , Marc Roth , Braden Rothwell , Neil Rowlands , Arpita Roy , 12 2 13 4 8 2 43 13 Pierre Royer , Patricia Royle , Chunlei Rui , Peter Rumler , Joel Runnels , Melissa Russ , Zafar Rustamkulov , Grant Ryden , 2 2 8 2 13 2 2 Holly Ryer , Modhumita Sabata , Derek Sabatke , Elena Sabbi , Bridget Samuelson , Benjamin Sapp , Bradley Sappington , 2,43 10 16 6 2 2 17 B. Sargent , Arne Sauer , Silvia Scheithauer , Everett Schlawin , Joseph Schlitz , Tyler Schmitz , Analyn Schneider , 16 2 2 2 2 2 25 Jürgen Schreiber , Vonessa Schulze , Ryan Schwab , John Scott , Kenneth Sembach , Clare Shanahan , Bryan Shaughnessy , 2 13 2 1 2 1 2 Richard Shaw , Nanci Shawger , Christopher Shay , Evan Sheehan , Sharon Shen , Allan Sherman , Bernard Shiao , 2 6 2 43 4 2 2 Hsin-Yi Shih , Irene Shivaei , Matthew Sienkiewicz , David Sing , Marco Sirianni , Anand Sivaramakrishnan , Joy Skipper , 2 2 2 1 31 2 1 2 G. C. Sloan , Christine Slocum , Steven Slowinski , Erin Smith , Eric Smith , Denise Smith , Corbett Smith , Gregory Snyder , 9 2 2 2 2 2 1 13 Warren Soh , Sangmo Tony Sohn , Christian Soto , Richard Spencer , Scott Stallcup , John Stansberry , Carl Starr , Elysia Starr , 1 2 1 2 2 2 6 Alphonso Stewart , Massimo Stiavelli , Amber Straughn , David Strickland , Jeff Stys , Francis Summers , Fengwu Sun , 2 2 2 2 13 2 2 Ben Sunnquist , Daryl Swade , Michael Swam , Robert Swaters , Robby Swoish , Joanna M. Taylor , Rolanda Taylor , 4 2 2 2 44 2 2 Maurice Te Plate , Mason Tea , Kelly Teague , Randal Telfer , Tea Temim , Deepashri Thatte , Christopher Thompson , 2 1 45 2 2 2 2 Linda Thompson , Shaun Thomson , Tuomo Tikkanen , William Tippet , Connor Todd , Sharon Toolan , Hien Tran , 2 2 13 25 2 2 2 Edwin Trejo , Justin Truong , Chris Tsukamoto , Samuel Tustain , Harrison Tyra , Leonardo Ubeda , Kelli Underwood , 2 1 3 12 1 2 2 Michael Uzzo , Julie Van Campen , Thomas Vandal , Bart Vandenbussche , Begoña Vila , Kevin Volk , Glenn Wahlgren , 1 8 23 2 9 1 2 Mark Waldman , Chanda Walker , Michel Wander , Christine Warfield , Gerald Warner , Matthew Wasiak , Mitchell Watkins , 1 17 8 13 2 2 8 1 Andrew Weaver , Mark Weilert , Nick Weiser , Ben Weiss , Sarah Weissman , Alan Welty , Garrett West , Lauren Wheate , 2 2 2 2 1 2 Elizabeth Wheatley , Thomas Wheeler , Rick White , Kevin Whiteaker , Paul Whitehouse , Jennifer Whiteleather , 2 6 6 13 9 8 William Whitman , Christina Williams , Christopher Willmer , Scott Willoughby , Andrew Wilson , Gregory Wirth , 2 8 2 6 2 8 1 2 2 Emily Wislowski , Erin Wolf , David Wolfe , Schuyler Wolff , Bill Workman , Ray Wright , Carl Wu , Rai Wu , Kristen Wymer , 2 2 9 2 2 4 2 2 Kayla Yates , Christopher Yeager , Jared Yeates , Ethan Yerger , Jinmi Yoon , Alice Young , Susan Yu , Dean Zak , 35 9 1 1 2 Peter Zeidler , Julia Zhou , Thomas Zielinski , Cristian Zincke , and Stephanie Zonak NASA’s Goddard Space Flight Center, USA; [email protected] Space Telescope Science Institute, USA Université de Montréal, Canada European Space Agency UK Astronomy Technology Centre, UK University of Arizona, USA National Research Council Canada (Herzberg), Canada Ball Aerospace, USA Honeywell International, USA Airbus Defence and Space GmbH, Germany Centro de Astrobiología (INTA/CSIC), Spain Katholieke Universiteit Leuven, Belgium Northrop Grumman Space Systems, USA Observatoire de Paris, France Niels Bohr Institute, Denmark Max-Planck-Institut für Astronomie, Germany 2 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Jet Propulsion Laboratory, California Institute of Technology, USA University of Oxford, UK Scuola Normale Superiore di Pisa, Italy Institut d’Astrophysique de Paris, France Commissariat à l’énergie atomique et aux énergies alternatives, France University of Cambridge, UK Canadian Space Agency, Canada Center for Astrophysics | Harvard & Smithsonian, USA Rutherford Appleton Laboratory, UK Lockheed Martin Advanced Technology Center, USA ATG Europe for the European Space Agency, The Netherlands Eidgenössische Technische Hochschule, Switzerland NASA’s Ames Research Center, USA Association of Universities for Research in Astronomy, USA National Aeronautics and Space Administration, USA University of California Santa Cruz, USA University of Copenhagen, Denmark Dublin Institute for Advanced Studies, Ireland AURA for the European Space Agency, USA Telespazio UK for the European Space Agency, UK Aurora Technology for the European Space Agency, The Netherlands Quantum Circuits, USA University of Groningen, The Netherlands ATA Aerospace, USA Heliospace, USA NASA’s Johnson Space Center, USA John Hopkins University, USA Princeton University, USA University of Leicester, UK Received 2022 November 16; accepted 2023 January 11; published 2023 April 3 Abstract This paper characterizes the actual science performance of the James Webb Space Telescope (JWST),as determined from the six month commissioning period. We summarize the performance of the spacecraft, telescope, science instruments, and ground system, with an emphasis on differences from pre-launch expectations. Commissioning has made clear that JWST is fully capable of achieving the discoveries for which it was built. Moreover, almost across the board, the science performance of JWST is better than expected; in most cases, JWST will go deeper faster than expected. The telescope and instrument suite have demonstrated the sensitivity, stability, image quality, and spectral range that are necessary to transform our understanding of the cosmos through observations spanning from near-earth asteroids to the most distant galaxies. Key words: Observatories – Infrared astronomy – Astronomical instrumentation 1. Introduction (ESA), and the Canadian Space Agency (CSA). The review article Gardner et al. (2006) described JWST’s design and This paper characterizes the delivered science performance science goals, which are divided into four themes: “End of the of JWST, as determined through the six month commissioning Dark Ages: First Light and Reionization;”“Assembly of period that ended on 2022 July 12, and describes how the Galaxies;”“Birth of Stars and Protoplanetary Systems;” and actual performance differs from pre-launch expectations. “Planetary Systems and the Origins of Life.” The design and JWST is a large (6.6m), cold (<50 K), infrared-optimized architecture of JWST are described in Menzel et al. (2022)(this observatory (Gardner et al. 2022, this issue) with a segmented issue). mirror design, that was launched on 2021 December 25 and is During the six-month commissioning period of JWST, the now in science operations. The project is an international mission team worked with dedication and focus to prepare the collaboration among NASA, the European Space Agency observatory for science operations. A key part of commission- Author to whom any correspondence should be addressed. ing activities was characterizing the on-orbit performance of the observatory, including the performance of the spacecraft, Original content from this work may be used under the terms telescope, science instruments, and ground system. This paper of the Creative Commons Attribution 3.0 licence. Any further summarizes those results, as drawn from many activities and distribution of this work must maintain attribution to the author(s) and the title analyses from the commissioning period. of the work, journal citation and DOI. 3 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. The expected pre-launch performance was incorporated into the sunshield and radiators, and MIRI’s active cryocooler the pre-launch versions of the exposure time calculator (ETC) recycles its helium. The only onboard consumables are and the JWST backgrounds tool. These tools have now been propellant: fuel and oxidizer. Before launch, JWST was revised to reflect actual performance, to support the Cycle 2 required to carry propellant for at least 10.5 yr of mission Call for Proposals. lifetime. Now that JWST is in orbit around L2, it is clear that To inform the scientific community, an early version of this the remaining propellant will last for more than 20 yr of paper was posted online (arXiv:2207.05632) at the end of mission lifetime. This fortunate surfeit has multiple causes: an commissioning on 2022 July 12. For the PASP special issue on accurate launch; launch on a day that required relatively less JWST, some description of science instrument performance has energy to get to L2 than most other possible launch dates; three been moved to companion papers: Böker et al. (2023) for timely and accurate mid-course corrections that sent JWST to NIRSpec; Doyon et al. (2022) for NIRISS/FGS, Rieke et al. L2 with the minimum possible propellant usage; and finally, (2023) for NIRCam, and Wright et al. (2023) for MIRI. Some careful stewardship of mass margins by the engineering team description of the JWST background levels has been moved to over the years, such that the remaining mass margin was used a different PASP paper (Rigby et al. 2023). The companion to add more propellant than required, until the tanks were full. PASP paper by McElwain et al. (2023) describes more fully the For the remainder of the mission, propellant will be used for two telescope element of JWST. purposes: stationkeeping burns (using fuel and oxidizer) to The transformative scientific performance of JWST is the maintain the orbit around L2, and momentum dumps (using only result of the collective effort, spanning decades, of thousands of fuel) to remove momentum from the reaction wheels. Momentum individuals from multiple institutions. The authors acknowl- accumulates as solar photons hit the sunshield and impart a net edge the tremendous amount of work by the entire international torque, which the reaction wheels resist by spinning up. The actual team to bring JWST through commissioning into science rate at which the observatory builds up momentum is within operations. specifications and is well below worst-case allocations, which further contributes to propellant lifetime. While the detailed 2. Spacecraft propellant usage depends on orientation, which is set by the 2.1. Orbit observing schedule, the big picture is that JWST has sufficient propellant onboard to support science operations for more than The Ariane 5 rocket that launched JWST on 2021 December 20 yr. 25 UT injected it into an orbit that was well within specification, with a semimajor axis approximately 0.5σ larger than nominal. This very slightly “hot” injection state had a 2.3. Projected Observatory Lifetime semimajor axis of 542,120.1 km versus the nominal of 536,533.8 km, and a delta-v at launch vehicle separation within At this point, it is not clear what will determine the duration of −1 −1 3ms of the target velocity of 10,089 m s . This accurate JWST’s mission. The mirrors and sunshield are expected to launch, in combination with three on-time, nominal mid-course slowly degrade from micrometeoroid impacts; the detectors are corrections, minimized propellant consumption and delivered expected to experience cumulative slow damage from charged JWST to a nominal orbit around the second Earth-Sun particles; the sunshield and multilayer insulation will degrade Lagrange point (known as L2). This orbit fully complies with from space weathering; the spacecraft was designed for a five year all geometry requirements, and supports communications with mission (as is standard for NASA science missions);and the the Science & Operations Center using the Deep Space science instruments include many moving parts at cryogenic Network. temperatures. These sources of degradation were all taken into Orbit around L2 is maintained through regular station- account in the design of JWST, with performance margins set so keeping burns, which are scheduled every three weeks. As of that JWST will still perform after many years of operation. At 2022 July 12, there have been four station-keeping burns, with present, the largest source of uncertainty is long term effects of typical durations of tens of seconds. During commissioning, micrometeoroid impacts that slowly degrade the primary mirror. three station-keeping burns were skipped because the computed As discussed in Section 4.7, the single micrometeorite impact that correction was negligibly small. occurred between 2022 May 22 and 24 UT exceeded prelaunch expectations of damage for a single micrometeoroid, triggering 2.2. Predicted Lifetime of Consumables further investigation and modeling by the JWST Project. The There are no consumable cryogens onboard JWST; the Pre-launch projections, informed by micrometeoroid population models and telescope and the science instruments are passively cooled by experimental studies and numerical simulations of impacts to beryllium mirrors, predicted that on average each segment would receive a cumulative https://jwst.etc.stsci.edu/ total of 16 nm added WFE over six years. The May impact resulted in one https://jwst-docs.stsci.edu/jwst-other-tools/jwst-backgrounds-tool segment receiving more than 10 times that average in a single event. 4 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Project is actively working this issue to ensure a long, productive structures, the stability is within the 40 mK noise of the science mission with JWST. temperature sensors on those components. The science instrument temperatures vary based on their activity, but they are also within the 10 mK noise for the instrument sensors. Any 2.4. Slew Speed and Settle Times resulting temperature change impact can only be identified The slew speed and settle time are important drivers for optically (see Section 4.5). No long term temperature drift has efficient operations as the observatory changes pointing to look been detected since achieving final cooldown conditions. at different targets on sky and carry out orbital stationkeeping Temperatures of detector focal plane arrays are actively and momentum unloading maneuvers. Slews use all six stabilized to a precision of a few mK. reaction wheels on the spacecraft to repoint the observatory, The observatory was designed to minimize ice deposition on although it is possible to control with five reaction wheels. The the optical surfaces. Sensitive components of the science control system was designed to slew 90° in less than 60 instruments and the telescope’s fine steering mirror were heated minutes, which has been demonstrated during commissioning. through the cooldown to prevent ice accumulation. Throughput Achieved slew speeds and durations to repoint to new targets measurements detect no spectral signatures of ice, which at the start of each visit are broadly consistent with pre-flight constrains any ice deposition that may have occurred to be far expectations, such as the timing model encoded in the better (less ice) than the requirements. Astronomer’s Proposal Tool (APT), plus typically 2 minutes Components within the spacecraft bus (computer, commu- duration for control overheads and settling. nications, cryocooler compressors and electronics, attitude At the end of a slew, pointing transients are observed while control and propulsion systems, etc.) are comfortably within the observatory settles. The pointing settles in ∼30 s with required operating temperature ranges regardless of pointing damping from isolators between the telescope and spacecraft direction or telescope activities. Instrument electronics housed bus. Re-pointing maneuvers can generate fuel slosh, which is at on the cold side of the Observatory are also within required ∼0.045 Hz and not compensated by the fine guidance control operating range and are under tight temperature control to system. The slew rate profile has been tuned to reduce the minimize any temperature-induced distortions (see excitation of the fuel slosh mode. Measurements of line-of- Section 4.5.3). All heaters on the observatory are functional sight pointing performance (Section 3.3 below) confirm that the on prime circuits and demonstrating expected margins. resulting effect of fuel slosh on pointing is <0.5 milliarcse- conds (mas). 2.6. Sunshield Performance After slewing to a new target, the pointing stabilizes quickly in less time than it takes the FGS guide star acquisition process The shape of the deployed sunshield affects the temperature to complete. This was not initially the case; in the first months and thermal stability of JWST, the amount of scattered light of commissioning, long slew settling times and high image from the Earth, Moon, and stars, and the background levels at motion impacted many observations, and required efforts to the longer wavelengths. Telemetry (microswitch, motor, investigate, diagnose, and mitigate. Adjustments to Attitude thermal, power, and inertial reference unit) indicated a Control System parameters and several software patches successful deployment ending with a nominal deployed shape. dramatically improved slew settling performance and resolved There are no subsequent indications of any issue with the this issue. Users examining very early commissioning data deployed shape, from the many thermal sensors onboard or (prior to mid 2022 April), particularly images taken in coarse from the background levels seen in the science instruments. point mode without fine guiding, should be aware of this The shape of the deployed sunshield affects how quickly the caveat. observatory builds up angular momentum from solar photons, which is then stored in the reaction wheels and must be dumped periodically using thrusters. The sunshield geometry was 2.5. Thermal designed to minimize momentum accumulation. The measured The cooldown of the telescope and science instruments was torque table on-orbit is consistent with the pre-launch model nominal and closely matched predictions. As predicted, the within the allocated uncertainties. As noted above, this implies primary mirror segments have cooled to temperatures of 35–55 a lower rate of fuel use for momentum management. K, with the hotter segments those closest to the sunshield. The secondary mirror has cooled to 29.3 K, the near-IR instruments 2.7. Other Spacecraft Performance to 35–39 K, and MIRI to 6.4 K. The MIRI cooler achieves this temperature at nominal, pre-flight predicted performance levels After optimization of the solar array regulator settings, and has no perceptible effect on pointing stability (e.g., jitter) or JWST is now generating 1.5 kW to match the power load, with on the performance of the other instruments. Since cooling to a capability of >2kW—as such, the power margins are operational temperatures, these temperatures have remained comfortable. JWST is projected to have a tight (11%) margin extremely stable with time. For the telescope mirrors and on data downlink during Cycle 1. The project is working 5 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. closely with NASA’s Deep Space Network (DSN) to resolve Line-of-sight pointing control with JWST requires the inter- all issues and ensure DSN can adequately support all Cycle 1 connected operation of star trackers, inertial reference units science and beyond. The JWST spacecraft has all the (gyroscopes), fine Sun sensors, reaction wheels, the telescope’s redundancy it had at launch. Of the 344 single point failures fine steering mirror, the fine guidance sensor, various target that were present at launch—almost all of them related to acquisition modes in science instruments, flight software in deployments—only 49 remain; these are common to most both the spacecraft and science instrument computers, and science missions (for example, only one set of propellant tanks, ground software systems for guide star selection and observa- only one high gain antenna). Fifteen of the remaining single tion planning, as well as the structural dynamics of the point failure items are associated with the science instruments deployed observatory. —any future failure with these items would degrade science It is inherently not possible to test those systems together in performance but would not end the mission. an end-to-end fashion on the ground. As such, on-orbit performance offers the first chance to characterize the system 2.8. Fault Management as a whole. As commissioning concludes, the pointing performance of the observatory meets or exceeds expectations. JWST has a robust fault management system that makes use of its redundant components to ensure the observatory remains 3.1. Pointing Accuracy after Guide Star Acquisition safe. With any new spacecraft/observatory, the first couple Absolute pointing accuracy after guide star acquisition is months after launch provide the operations team an opportunity excellent. Observed pointing offsets are generally below 0 19 to learn how the vehicle performs on-orbit and adjust the fault (1σ, radial). When systematic offsets are removed by improved management and system parameters to fine tune the system behavior. During commissioning, JWST experienced seven astrometric calibration between the Guiders and the other safe mode entries (six safe haven and one inertial point mode). science instruments (SIs), the residual scatter around the Early science operations (2022 July -December) experienced desired target position is generally below 0 10 (1σ, radial) , one safe haven entry and four entries into inertial point mode. better than the prelaunch predicted value of 0 14, which is in The majority of these safe mode entries were a result of on- turn significantly better than the required value of 1 0. Updates orbit learning of the nuances of the control system behavior and for such systematic offsets are in progress. This confirms system level interactions. All of the safe mode entries’ expectations that many integral field spectroscopy (IFS) underlying causes are understood and several updates to the observations may omit target acquisition and rely solely on fault management and control system set items have been guide star acquisition to achieve the necessary science pointing. loaded to flight software to prevent the issues from recurring. The excellent pointing performance can be credited to JWST’s Likewise, flight software has placed individual instruments own systems, as well as the high accuracy and precision of the into safe modes on several occasions in responses to guide star catalog enabled by the Gaia mission. unexpected telemetry values or conflicting commands. In all Early in commissioning, some observations had significantly cases the underlying issues were quickly understood, and larger pointing offsets, of order ∼1 5, due to catalog cross- appropriate steps were taken to bring the instrument back into matching issues that occurred when data from Gaia and other operation and prevent recurrence of the fault. JWST’s onboard catalogs such as 2MASS were combined to produce the current event-driven operations system works as intended to allow the JWST guide star catalog. Updated guidelines for selecting observatory to flexibly continue observations when an guide/reference stars was implemented 2022 May 5, and non- observation cannot execute due to an instrument fault or a stars were disallowed as reference objects 2022 June 24; both greatly reduced mis-identification. A future version of the guide star acquisition error; in such cases, JWST will catalog is planned for 2023 which is expected to provide automatically move on to the next observation in the onboard further improvements. observation plan. Pointing repeatability is likewise excellent: independent separate observations returning to the same target and at the 3. Pointing and Guiding same position angle generally result in identical target Observatory attitude control and line-of-sight stabilization to pointings on a detector to within <0 1. achieve JWST’s stringent requirements present a complex engineering challenge, involving a sophisticated interplay 3.2. Pointing Accuracy After Target Acquisition between hardware systems across the entire observatory and arguably the most complex set of flight software components. For observing scenarios requiring better final pointing accuracy than provided by the guide star catalog alone, the The NASA Goddard GOLD rules, https://standards.nasa.gov/standard/ instruments offer onboard Target Acquisition procedures to GSFC/GSFC-STD-1000,define a failure in this context as preventing the place targets where desired within a science instrument field, mission from fully meeting level 1 requirements; this is a stricter definition than failure meaning loss of the mission. e.g., centered on a coronagraphic spot or null, or in a repeatable 6 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. position for NIRISS Aperture Masking Interferometry or Single achieved NEA can vary by ∼0.3 mas between dithers on the Object Slitless Spectroscopy (SOSS). There are several distinct same source. For moving targets, the guide star faint limits are versions of target acquisition for the various instrument modes. somewhat reduced to 16.5 in Guider 1 and 17.0 in Guider 2. These onboard processes have been confirmed to meet their Guiding performance can vary under some circumstances. requirements, typically yielding target positions within a few Reasons that can cause a brighter star to appear to register mas rms per axis in the near-IR instruments, and slightly larger lower counts than expected (and give an increased NEA) in MIRI with its broader longer-wavelength PSFs. NIRCam include: unflagged bad pixels in the guider box, (flagged) bad coronagraphy, the final mode to have its Target Acquisition pixels exactly coincident with the guider star’s position, process evaluated, currently experiences larger offsets, which guiding on an extended galaxy instead of a star, quantum will be alleviated by future calibrations—even so, the Small efficiency variations due to cross hatching on the detector Grid Dithers recommended for that mode cover the range of surface, or a guidestar falling at the edge of the centroiding uncertainty and yield excellent coronagraphic contrast. box. Work is in progress to reduce these variations. In most The algorithm for NIRSpec Multi-Object Spectroscopy is the cases, observers should expect excellent stability in their most complex, because it derives a roll correction for the observations. observatory position angle as well as the usual x, y positional Very precise flight jitter measurements have yielded detailed offsets. That target acquisition required great care in imple- insights into observatory dynamics, and confirmed many mentation, and was the most challenging type of target aspects of preflight integrated modeling. These data show that acquisition to get working during commissioning; it is now neither the spacecraft’s reaction wheels nor the MIRI working within requirements when the requisite attention is cryocooler produce measurable contributions to the line-of- paid to the accuracy of the input reference star information. sight jitter. Instead there are two signatures, detectable but at low level, that come from a 0.3 Hz oscillation mode of a vibration damper between the telescope and spacecraft bus, and 3.3. Guiding Precision and Line of Sight Pointing a ∼0.045 Hz oscillation understood to be due to fuel slosh, both Stability consistent with models. The amplitudes of these modes range The pointing stability of the line of sight under fine guidance up to ∼0.3 mas each, with variation over time that does not control is superb, several times better than requirements. The appear correlated with pointing history such as slew distances. FGS sensing precision, parameterized as the Noise Equivalent Momentum/reaction wheel operations or unplanned High Angle (NEA, an equivalent jitter angle calculated based on Gain Antenna motion may occasionally disturb observatory centroid precision and S/N) is usually ∼1 mas (1σ per axis),as dynamics and degrade pointing stability; these are generally compared to the requirement of 4 mas; the NEA is usually short-lived and fine guiding is expected to be maintained. High symmetric in x and y. Gain Antenna moves are not planned during any science The achieved line-of-sight jitter as seen in the science observations except for long duration time series observations. instruments is similarly good, even when measured at higher frequency than the 15.6 Hz cadence of the FGS measurements used in calculating the Guider NEA. Jitter measured via high- 3.4. Precision of Dithering frequency sampling (every 2.2 ms) using a small NIRCam JWST has three methods of performing a small repointing, subarray has also been ∼1 mas (1σ per axis). For long commonly called a dither. Dithers less than 60 mas are observations, drift in observatory roll could generate systematic executed with the fine steering mirror (FSM) while maintaining motion of sources in an SI field, as the FGS can not sense or closed loop on the guide star. Dithers between 60 mas and 25″ correct roll with the single guide star used in the Fine Guide are executed by dropping closed-loop guiding, slewing, and re- control loop. However, measurements during a commissioning entering guiding at the track mode stage. Dithers larger than thermal characterization test indicated that the roll drifts are 25″ are executed by dropping closed-loop guiding, slewing, extremely small, far below requirements, even after a worst- and re-entering guiding at the guide star acquisition stage. case hot-to-cold slew, and contribute negligibly to the total All three methods of dithers result in an offset precision as pointing error. In comparison, pre-flight predictions for measured by the Guider of 1–2 mas. On the sky, the accuracy pointing stability were in the range 6–7 mas (1σ per axis). of the dithers is typically 2–4 mas rms per axis, due to residuals The FGS guide star magnitude scale matches the 2MASS J- in the Guider’s astrometric calibration or small systematic band, Vega scale, with guide star brightnesses of 12.5 < J < 18 offsets in the Guider’s calculated 3 × 3 pixel centroids. As allowed for fixed targets. Dimmer stars will typically have a reported by the observing SI, offsets for large dithers may differ larger NEA, due to the lower signal to noise per time step, but with NEA that is still well within the requirement. The guide by another few mas, due to residuals in the astrometric star selection system preferentially selects the brightest calibration of the SI. As SI astrometric calibrations continue to available guide star for a given pointing. In practice the improve, these residuals may decrease. 7 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Table 1 Moving Targets Tested during Commissioning and Early Science Operations Apparent Rate of Motion −1 Moving target (type) (mas s ) Program ID Instrument/Mode Jupiter (planet) 3.3 1022 NIRCam Imaging, NIRISS Imaging, NIRSpec fixed slits and IFS, MIRI MRS and imaging 2516 Roman (MBA) 4.7 1449 MIRI/Imaging 118 Peitho (MBA) 4.9 1449 MIRI/ LRS and MRS 6481 Tenzing (MBA) 5 1021 NIRCam/Imaging 1773 Rumpelstilz (MBA) 6.6 1021 NIRISS/AMI 216 Kleopatra (MBA) 11 1444 NIRSpec/ IFS and MOS longslit 2035 Stearns (Mars- crossing asteroid) 24 1021 NIRCam/Imaging 4015 Wilson-Harrington (Apollo, 40 1021 NIRCam/Imaging NEO, PHA) 464798 (2004 JX20)(Aten, NEO) 67 1021 NIRCam/Imaging 411165 (2010 DF1)(Apollo, NEO, PHA) 90, 110 2744 NIRCam/Imaging Note. Targets are sorted by apparent rate of motion. The last target on the list was tested during early science operations; the others were tested during commissioning. Target type is listed in parenthesis after the target name, where MBA = main belt asteroid, NEO = near-Earth object, and PHA = potentially hazardous asteroid. Aten and Apollo asteroids have orbits that cross the Earth’s orbit around the Sun. 3.5. Performance for Tracking Moving Targets object’s orbit for the purpose of planetary defense (Thomas, C. et al. in preparation.) JWST has a Level 1 requirement to track objects within the While tracking at super-fast rates has now been demon- −1 solar system at speeds up to 30 mas per second (mas s ).In strated, the observatory efficiency was poor due to guide star commissioning, tracking was tested at rates from 5 to −1 availability and acquisition at these rates. Thus, the new 67 mas s . These tests verified tracking and science instrument −1 maximum rate of motion being offered is 75 mas s without performance for moving targets, including dithering and −1 limitations, but special permission for rates up to 100 mas s mosaicking. may be requested and subjected to approval for future cycles. All tests of moving targets during commissioning were Observing a bright planet and its satellites and rings was successful. Centroids showed sub-pixel scatter in all instru- expected to be challenging, due to scattered light that may ments. No test showed elongation of moving targets, as would affect the science instrument employed, but also the fine be expected in the case of poor tracking. rms jitter in the guider guidance sensor must track guide stars near the bright planet. was typically <2 mas (1σ, radial), comparable to that seen for Therefore, commissioning included tests of moving target fixed targets. Image quality as determined by FWHM tracking with NIRCam, where Jupiter was incrementally measurements of point spread functions was also comparable moved from the NIRCam field of view (FOV) to the FGS-2 to that of fixed targets. Table 1 summarizes the moving targets FOV. See Figure 1. These observations verified the expectation observations during commissioning. −1 that guide star acquisition works successfully as long as Jupiter Tracking at faster-than-requirements rates of 30–67 mas s is at least 140″ away from the FGS, consistent with pre-flight showed accuracies similar to tracking of slower-moving modeling. targets; this potentially opens up science for near-Earth The other SIs were also tested for efficiency with nearby asteroids (NEAs), comets closer to perihelion, and interstellar scattered light, also using the Jupiter system. Preliminary objects. This capability was pushed further early in science results have measured scattered light contamination on the operations, when JWST successfully tested rates up to −1 −1 detectors for all instruments when the planet was not in the 110 mas s (396″ hr ) on near-Earth asteroid 2010 DF1, to primary FOV, which will need to be considered for planning observe and track the Double Asteroid Redirection Test nearby satellite observations. The most notable impact for (DART) mission target (65803) Didymos at the time when scattered light is in NIRSpec IFS mode—if a bright planet is on the spacecraft impacted the asteroid moonlet on 2022 the microshutter array, then light seepage becomes significant. September 26, to demonstrate the capability of modifying an 8 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 1. NIRCam narrow-band imaging of Jupiter, moons, and ring. PID 1022 demonstrated that JWST can track moving targets even when there is scattered light from a bright Jovian planet. At left is a NIRCam short-wavelength image in filter F212N (2.12 μm); at right is a NIRCam long-wavelength image in filter F323N (3.23 μm). The exposure time was 75 s. The Jovian moons Europa, Thebe, and Metis are labeled. The shadow of Europa is also visible, just to the left of the Great Red Spot. The stretch is fairly harsh to bring out the faint moons as well as Jupiter’s ring. Each image is from one NIRCam short wavelength detector, which spans 63″. 3. Guide star coordinates, or occasionally brightnesses, in 3.6. Success Rate for Closed-loop Guiding the catalog are incorrect. The catalog is corrected when The closed-loop fine guidance system is one of the most such errors are found. complex aspects of JWST operations, requiring close coordina- 4. The guide star catalog contained duplicate entries for some guide stars. This duplication was reduced greatly tion between several subsystems and multiple control loops with new rules for selecting guide stars, implemented running in real time. Activating and tuning the fine guidance 2022 May 5. A new version of the guide star catalog system was a significant focus during commissioning. The expected in 2023 should largely eliminate this issue. robustness of the system has steadily improved and continues 5. The guide star may be placed on a bad pixel. Bad pixels to do so as reasons for failure are identified and mitigated. are flagged once identified. As a snapshot of performance during the later period of 6. The pointing may not have stabilized sufficiently at the commissioning, from May 25 through 2022 June 16, guiding start of the Guide Star ID process. worked successfully ∼93% of the time: guiding was successful on the first try ∼81% of the time, and 12% of the time In Cycle 1, users should expect a closed-loop guiding succeeded on the second or third guide star candidate success rate close to the 93% value that was achieved late in attempted. (Up to three guide star candidates can be tried in commissioning. That value is, largely by coincidence, very a visit.) In the same time period, guiding failed or skipped 7% close to the success rate for Hubble guide star acquisitions in recent cycles. Efforts continue to optimize guiding success of the time. rates. Several reasons for guiding failure were identified, with steps taken to mitigate them: 4. Optical Performance 1. The “guidestar” is actually a galaxy, but was classified as a star in the guide star catalog. These are flagged in the The image quality achieved by JWST exceeds performance guide star catalog once found. requirements and expectations, having diffraction-limited 2. Attempts to guide on known galaxies frequently failed. image quality at wavelengths much lower than requirements, Using a known galaxy as a reference source for ID was very good stability, and superb throughput. There is not one disallowed starting 2022 June 24. single factor to credit for the high performance; rather it is the 9 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 2. Measured Point Spread Functions spanning the full wavelength range of NIRCam.The filters shown are the shortest and longest wide filters in each of the NIRCam short wavelength and long wavelength channels. The left two panels show individual PSFs (single exposure each), on two different logarithmic scales for higher dynamic range. The right two panels show 4× subsampled effective PSFs (ePSFs), generated following the method of Anderson & King (2000) using dithered PSF measurements of many stars; these are shown zoomed in by 2× compared to the left panels. The second column from right, shown on a linear scale, highlights the compact PSF core, while the log display in the other columns emphasizes the diffraction features from the primary mirror geometry. The PSF core is sharp even at the shortest wavelengths in F070W. Data from PIDs 1067 and 1072. 10 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. accumulation of performance margins throughout the observa- Table 2 Measured Static Wave front Errors After Multi-instrument Alignment tory, and the result of many careful and precise efforts throughout the design, assembly, and alignment of the Instrument Measured Observatory Static WFE (nm rms) telescope and instruments. NIRCam, short-wavelength 61 ± 8 (module A),69 ± 11 (module B) JWST’s top-level image quality requirements were defined NIRCam, long wavelength 134 ± 38 (module A) 134 ± 39 (module B) to achieve Strehl ratio greater than 0.8 at a given wavelength; this can be converted to an equivalent condition of having rms NIRISS 68 ± 12 wave front error less than λ/14, using the optical Marechal NIRSpec 110 ± 20 approximation (Marechal 1947). These conditions are equiva- MIRI 99 ± 28 lent; in practice, the wave front error formulation was used for JWST’s optical budgets and in-flight wave front sensing FGS 77 ± 15 (FGS1),69 ± 8 (FGS2) measurements. JWST’s requirement was to be diffraction limited at Note. Quoted values are the average over multiple field points within each instrument, as measured from multi-instrument multi field sensing during λ = 2 μm wavelength, which corresponds to 150 nm wave commissioning in 2022 May, and updated for the final focus adjustment at the front error by the λ/14 criterion. The achieved wave front error end of that process. The plus and minus ranges reflect the measured variation in is routinely between 65 and 70 nm in the NIRCam shortwave wave front error at different field points within each instrument’s field of view. channel. During early science operations, wave front control These values represent just the static (non-time-varying) portion of the updates have been scheduled when the wave front error wavefronts, and include the sum of telescope and instrument WFE together. Additional time-dependent terms sum on top of these at any given time. Data exceeds 80 nm (McElwain et al. 2023). Therefore, operation- from PID 1465. ally, the observatory is diffraction limited at λ = 1.1 μm. McElwain et al. (2023)(this issue) describes in more detail the telescope deployment and alignment processes as carried such that the NIRCam short wavelength channel, with typical out in flight. Here we summarize the achieved optical performance, drawing on extensive telescope wave front WFE ∼80 nm, in fact achieves λ/14 at λ = 1.1 μm. sensing and image characterizations throughout commission- That wave front quality yields optical point spread functions ing, including a dedicated thermal stability characterization (PSFs) with angular resolutions set by the diffraction limit (i.e., exercise. PSF full width at half maximum ∼λ/D) across the full range of available wavelengths. See Figure 2. In particular, even in its shortest filter F070W, the NIRCam short wavelength channel 4.1. Wave front Error and Angular Resolution achieves a Strehl ratio of ∼0.6. Though ∼40%–50% of the The achieved telescope wave front error (WFE), measured at light is in a diffuse speckle halo at that wavelength, the PSF the primary wave front sensing field point in NIRCam module prior to detector sampling still has a tight core with angular A, is generally in the range 60–80 nanometers rms; it varies on resolution ∼λ/D. (In that sense, JWST’s PSF quality at 0.7 μm multiple timescales as described below. That WFE contrib- is similar to PSFs achieved at 2 μm by adaptive optics systems ution sums with field-dependence of the telescope WFE and on 8–10 m ground-based telescopes in good conditions.) In instrument internal WFE to yield total observatory WFE values practice, angular resolution at the shortest wavelengths (<2 μm which are modestly higher: 75–130 nm depending on instru- for NIRCam short wavelength, <4 μm for NIRCam long ment, observing mode, and field position. See Table 2. Motions wavelength or NIRISS) is limited more significantly by of mirror segments over time can lead occasionally to wave detector pixel Nyquist sampling than by optical performance; front error levels higher than those values, which are corrected subpixel dithering and image reconstruction (“drizzling”) will through the routine wave front sensing and control process. be required to make use of the full resolution at these JWST has exquisite image quality across the entire telescope wavelengths. field of view and at all available wavelengths. Expressed Wave front sensing confirms the surface quality of the relative to wavelength λ, JWST ranges from ∼λ/10 for individual mirror segments in space matches closely the mirror NIRCam F070W to better than λ/100 for MIRI F1000W and surface maps measured during cryogenic testing on the ground. longer. JWST aimed to achieve at λ = 2 μm a Strehl ratio of See Figure 3. In other words, after launch into space, and 0.8 (corresponding to wave front rms λ/14 or better), which is significant thermal contraction and deformation while cooling considered having diffraction-limited image quality for space- based telescopes. This is achieved with substantial margin, For example, in NIRCam’s shortest filter F070W, the optical PSF pre- The best achieved telescope wavefronts at the completion of alignment were sampling is roughly 28 mas ∼ 0.9 pixels in the NIRCam short wavelength as low as 50 nm rms; the May 2022 micrometeoroid impact on segment C3 channel, but after sampling onto the detector pixel scale of 32 mas, the apparent subsequently raised the high-order uncorrectable WFE term enough that the PSF resolution becomes 40 mas ∼ 1.25 pixels in the NIRCam short wavelength floor is now 59 nm rms. channel. 11 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 3. Wave front Sensing Measurements showing the quality of achieved mirror alignment on orbit. The telescope wave front error achieved in flight, shown in the right panel, closely tracks the as-polished surface figures of the individual segments, as measured during ground testing, shown in the left panel. JWST’s wave front sensing and mirror control systems are working as intended, achieving optimal alignments within the ∼10 nm resolution of the sensing and control system and correcting as necessary to maintain that alignment. Data from PID 1163. from room temperature to ∼45 K, the achieved wavefronts on the sole exception of the high frequency WFE which was each segment were just as expected. increased by the 2022 May micrometeoroid impact. Terms with particularly notable performance relative to requirements include the telescope stability (see Section 4.5), the line of 4.2. Comparison to Optical Budget for the Telescope and sight pointing (Section 3.3), and the wave front quality of the Science Instruments science instruments (which are all significantly better than their JWST has top-level science requirements to achieve image requirements). quality with Strehl ratio greater than 0.8 in NIRCam at a An initial assessment, which combined the measured wavelength of 2 μm (equivalent to wave front error of 150 nm beginning-of-life performance with model predictions for rms) and MIRI at a wavelength of 5.6 μm (equivalent to observatory aging in the L2 environment, predicts that JWST 420 nm rms). These and other top-level requirements flowed should meet its optical performance requirements for many into detailed optical performance budgets and lower-level years. The current largest source of uncertainty in models is the requirements, such as a required telescope wave front error rate of mirror surface degradation from micrometeoroids, „131 nm rms over the fields of view of the science discussed below. instruments, a required telescope stability of „54 nm rms over two weeks, and so on. Though line-of-sight image jitter is distinct from WFE, it is also tracked within the same budget via 4.3. Shape of the Point-spread Function a computed equivalent WFE. JWST’s hexagonal aperture creates a characteristic diffrac- Figure 4 presents an abbreviated summary comparing tion pattern in its point spread functions, with six stronger measured performance in flight to those budgets at high level. diffraction spikes at 60° intervals created by the segment and Note that the measured values shown here are from aperture edges, plus two fainter horizontal spikes created by the commissioning at observatory “beginning of life,” while the vertical secondary mirror support. While these diffraction required and predicted values were derived for “end of life” spikes can be visually dramatic in images which are deeply after a notional 5 yr mission. At the observatory top level, the exposed or are plotted with log stretches, it is the case that the achieved WFE is well below predictions by ∼30%, and the majority of light is focused into the PSF core (typically ∼66% remaining performance margin relative to requirements is large. All lower-level terms are at or below requirements values, with within the first Airy ring). 12 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 4. High-level summary representation of JWST optical performance for 2.0 and 5.6 μm. For each contribution to wave front error, triangles represent the required level, Xs mark the pre-launch optical budget predicted levels, and bars indicate the measured performance. All Optical Telescope Element (OTE) values shown are from on-orbit measurements. Science instrument WFE values shown are for typical field points (median SI image quality) in NIRCam and MIRI, as measured on the ground during ISIM CV3 testing. The colored lines depict which terms sum together; for instance the OTE total WFE is the RSS sum of the OTE static WFE and OTE time-varying WFE. Shaded portions of the bars indicate the delta in performance from the 2022 May micrometeoroid strike on segment C3. Compared to the diffraction patterns from circular apertures outline of the larger hexagon which is created is rotated 30 deg which many astronomers are more familiar with, such as relative to the smaller ones. Hubble, the hexagonal geometry of JWST concentrates wide- Additional PSF details per instrument: angle diffracted light more strongly into the diffraction spikes, 1. MIRI imager and MRS PSFs, particularly at wavelengths while the areas between those spikes are relatively darker. For „10 μm, show additional diffraction spikes in vertical bright sources the diffraction spikes can be seen to a separation and horizontal directions in the instrument coordinate −1.5 of many arc minutes, falling off as ∼R , including frame, called the “cruciform” or “cross artifact.” These diffraction spikes from sources outside of an instrument’s field arise due to diffraction that occurs internal to the detector of view. at the detector electrical contacts. This effect was seen in Note that the position angles of the six bright diffraction Spitzer’s similar Si:As detectors, and was modeled and features are different between the PSF core region (2–5 λ/D, expected for MIRI (Gaspar et al. 2021). Improved models dominated by diffraction from the overall outer hexagonal based on flight data are able to reproduce this artifact in outline) and the outer wings (>5 λ/D, dominated by detail including the field dependence. Pipeline steps are diffraction from the individual segments). See Figure 2. This being developed to compensate for its impact on MIRI can be understood intuitively from the pupil geometry: when MRS data cubes. assembling a larger hexagon from smaller hexagons, the 13 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. 2. NIRISS imaging PSFs at wavelengths …2.7 μm have close to predicted values, however some timescales are faster stronger diffraction spikes due to additional pupil than expected. These variations have small but measurable obscuration from the CLEARP pupil wheel position effects on PSF properties, which should be taken into account which must be used with those filters. for measurements at the highest precision, but will be 3. NIRISS imager PSFs at wavelengths „2.0 μm show negligible for many science cases (similar to the effect of the additional anomalous or “extra” diffraction spikes, which “breathing” variations seen in Hubble PSFs on orbital are more diffuse than the ordinary spikes and have a timescales). position angle that varies strongly with field position. The Overall stability: During the last month of commissioning, intensity of the anomalous spikes decreases with 2022 June, the change in wave front measured between wavelength, in F090W containing ∼70% of the flux of successive observations roughly 2 days apart was typically the normal vertical diffraction spike, becoming barely „25 nm rms, and frequently less than half that (range 8–50 nm detectable by F200W. These spikes are now understood rms per 2 days). During this time period, the observatory was as due to diamond-turning tool marks on off-axis mirrors conducting a wide range of commissioning and Early Release within NIRISS. Retroactive examination of some ground Observations typical of science activities, so this level of test data also shows their presence at lower S/N. A stability should be representative of what can be expected similar effect is also seen in FGS PSFs. during Cycle 1. Indeed, during the first half of Cycle 1 the 4. The rotation of NIRSpec in the focal plane means that, in median change in wavefront between successive observations the detector coordinate system, NIRSpec PSFs are rotated was only 10 nm rms, just slightly above the typical wavefront by 139 deg relative to the other instruments. MIRI is sensing measurement uncertainty of 7 nm rms. similarly rotated 5 deg. Though we provide details below on observed variations 5. Instrument modes with specialized optical components, over time, we wish to emphasize that these are small variations. such as NIRCam coronagraphy, MIRI coronagraphy, NIRISS aperture masking interferometry, and NIRISS The absolute amplitude of drifts seen between successive wave SOSS, each have their own PSF properties, as described front monitoring visits is comparable to the Hubble Space in Girard et al. (2022), Kammerer et al. (2022), Telescope’s ∼18 nm rms of focus variation typical on orbital Kammerer et al. (2023), Boccaletti et al. (2022). These timescales (see e.g., Lallo 2012). Yet for JWST the variations observed PSFs are in general highly consistent with are observed to mostly occur on significantly longer timescales, model predictions. and the effect on images at longer wavelengths is correspond- ingly reduced. (Hubble is typically stable to λ/30 at λ = 0.5 μm over 90 minutes; JWST is often stable to λ/100 4.4. Transmission and Contamination at λ = 2 μm over 2 days.) Also similar to Hubble, the The telescope’s effective area, the product of the telescope amplitude of wave front variations over time is comparable to area and its transmission, is a key optical performance and generally less than the field-dependent variations within a parameter that was tracked through development and char- given instrument. acterized on orbit. The JWST unobscured telescope area was 2 Factors contributing to stability levels:The telescope is required to be >25 m . The measured value using the NIRCam 2 very thermally stable, but small changes in equilibrium pupil imaging lens was 25.44 m . The telescope’s wavelength- temperature (<0.1 K, within the temperature sensor noise) dependent transmission ranged from 0.786 at 0.8 μm to 0.933 can still occur in response to changes in attitude with respect at 28 μm, better than requirements at each wavelength. Even to the Sun. Variations from instrument heat sources can also though the JWST optics spent significantly more time in affect telescope structures. A thermal stability exercise ground facilities than originally anticipated, these high measured these effects by moving between the extremes of transmissions were maintained with careful control of con- JWST’s field of view, with a week-long soak at “cold” tamination throughout the integration and test phases and during preparation for launch. Brush cleaning was also carried attitude (−40° pitchrelativetothe Sun) sandwiched between out on the telescope primary mirror segments and secondary an equalamountoftimeatthe “hot” attitude (0° pitch);this mirror to remove particulates. test is intentionally more stressing than typical attitude profiles during science operations. The thermal slew exercise 4.5. Optical Stability on Different Timescales (PIDs 1445, 1446) included a wide range of wave front sensing, imaging, and sensor telemetry investigations to Commissioning observations characterized the telescope’s characterize many aspects of JWST’s performance on optical stability and variations on multiple timescales. Highly precise wave front sensing allows measurement of very small multiple timescales. Modes of variation observed with changes (<10 nm). The overall amplitudes of variation are JWST include the following. 14 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. 4.5.1. Telescope Backplane Thermal Distortion with varying amplitudes of a few nanometers to tens of nanometers. See Figure 23 of McElwain et al. (2023)(this issue). After backing out the contributions from tilt events (see These are hypothesized to be due to structural microdynamics below), the observed drift in wave front during the thermal within the telescope (e.g., localized relaxations of stiction or slew test could be well fit as a double exponential. The largest microstresses within the backplane structure). This hypothesis is and slowest term arises from thermal deformation of the supported by a few cases in which measurable stress relief was observatory backplane in response to the small change in detected by telescope wing latch strain sensors at the time of tilt equilibrium temperature. In the thermal slew test this was events. Fast wave front sensing measurements constrain the measured to be ∼18 nm with a time constant of 1.5–2 days. In timescale for some tilt events to be <10 s; these are effectively comparison the preflight prediction from integrated modeling instantaneous step functions in mirror position. was 14.6 nm with a time constant of 5–6 days predicted at Since a few such events were first seen during the telescope observatory beginning of life, versus requirements of 54 nm cryovacuum testing on the ground, it was expected these might rms. The dominant mode in this drift is zero-degree be seen in flight, particularly initially after cooldown. Tilt astigmatism, essentially a bending mode of the primary outer events proved to be not infrequent during both commissioning segment “wings” relative to the center section, in accordance and early Cycle 1, though the rate has decreased overtime.The with predictions. rate of tilt events continued to decrease through the first half of Cycle 1, and as of this writing in 2023 February there have 4.5.2. Telescope Soft Structure Thermal Distortion been no additional larger tilt events since 2022 October.Dur- The second term in the double exponential fit represents ing 2022 April–May there were periods in which small tilt wave front drift which arises from soft structures. The black events occurred as frequently as two to three per day, but by Kapton stray-light-blocking “frill” around the outside of the later 2022 June there were weeks with few or no tilt primary can place mechanical stress on its points of attachment events.Figure 23 of McElwain et al. (2023)(this issue) shows when it thermally expands and contracts. The lower thermal that tilt events became increasingly rare during the first five mass of the frill causes this to have a shorter characteristic months of science operations (2022 July through November). timescale. The observed effect in the thermal slew test was an The expectation that tilt events would be small and would exponential with 4.45 nm amplitude and time constant of 0.77 stabilize as the microstresses are relieved seems to be in line hr. mac-fn7 In comparison, the preflight prediction was with the reduction we are seeing and the fact that they are 8.6 nm with a time constant of 8–10 hr. small. Tilt events may cause small but observable changes in PSFs 4.5.3. Fast Oscillations from Heaters during some science observations. The resulting change to encircled energy is very small (small fractions of a percent) and On short timescales (2–4 minutes), the thermal cycling of much smaller than encircled energy stability requirements. heaters in the ISIM Electronics Compartment (IEC) induces However it can be significant for observations at very high small forces on the telescope backplane structure. The observed precision. For instance tilt events were observed in the result is a semi-periodic variation primarily in astigmatism with NIRCam and NIRISS time series commissioning tests, and amplitude of about 2.5 nm, which is closely consistent with caused abrupt small step-function offsets in the initial measured preflight predictions. The exact amplitude and timescale vary as flux analysis (for example, see Section 6.1.2). Flux jumps the cycles of several different heaters beat together. This caused by occasional tilt events can be calibrated as a function oscillation is sufficiently small and rapid that it has thus far of aperture size or wavelength and/or be included as a only been measured in special high-cadence differential wave parameter in the transit fit model. The time sequence of FGS front sensing measurements which achieve sub-nanometer image data and centroids obtained at 15.6 Hz during all science precision. This effect is expected to be negligible for the vast observations may also be a useful diagnostic channel for some majority of science observations. tilt events. Not all changes in segment position occur on the rapid 4.5.4. Segment Tilt Events and other Drifts timescale of tilt events; at times, slow drifts of segment “Tilt events” are occasional abrupt changes in position of an positions over hours or days have been observed as well. individual segment which are seen to occur from time to time, Understanding of the structural dynamics of this segmented telescope in space will continue to improve with time. While one might assume the amplitude of this effect to be independent of slew direction, this was not observed to be the case: after the reverse slew from cold to hot attitude, a similar exponential drift was not seen, and instead the 4.6. Routine Wave front Sensing and Control wave front was stable to within 3 nm over the 6 hr measurement period. Due to test duration it was not possible to repeat these measurements for further Since the completion of telescope alignment, regular wave investigation of under what circumstances the frill term does or does not manifest. front sensing and control has monitored and maintained 15 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. telescope alignment, as will continue throughout the mission. shortwave channel has the least wavefront error of all the During science operations, wave front sensing measurements modes (see Table 2 of McElwain et al. (2023), this issue), the are conducted every twodays (roughly, with flexible cadence increase in wavefront error for all other instruments and modes around science observations). The resulting measurements of from the C3 strike should be comparable to or less than mirror alignment state versus time are automatically made NIRCam. available in the MAST archive; software for making use of Images of the telescope optics using the NIRCam pupil these measurements to produce time-dependent PSF models is imaging lens can reveal smaller impacts below the threshold now included in the WebbPSF package. detectable by wave front sensing. Comparison of pupil images Wavefront control is nominally applied whenever the total taken 23 February and 2022 May 26 show evidence for 19 such wavefront error (telescope + NIRCam, short wave channel, at minor strikes over that 92 days period. Regular monitoring of the reference field point) reaches 80 nm rms due to drift or any the pupil may help constrain the micrometeoroid hit rate and other instability. Over the period 2022 November through 2023 power spectrum. Micrometeoroid impacts should also slightly January, this total has varied over a narrow range between lower the telescope throughput; this effect is not yet about 65–77 nm rms, with only two wavefront corrections measurable. needed. It appears the 2022 May hit to segment C3 was a fairly rare Because of the every-two-days cadence of wave front event. Still an open question is how rare — every year versus sensing measurements, wave front states at times between every few years. Ten such large hits would degrade the mirror those measurements are not directly measured (e.g., if a tilt such that it is no longer diffraction limited at lambda < 2micron. event or micrometeoroid impact occurs within some 2 days, its Therefore, after further investigation of the micrometeoroid exact time of occurrence is not measured). Some science modes population (see Section 6.2 of McElwain et al. (2023),this do allow for more frequent wave front measurements, in issue), the JWST Project has decided, starting in Cycle 2, to limit particular NIRCam time series using the weak lens in the short the amount of time the telescope spends pointed toward the wave channel, and to some extent the NIRISS SOSS mode. direction of orbital motion, as those directions statistically have higher micrometeoroid rates and energies. 4.7. Micrometeoroids 5. Backgrounds, Stray Light, and Scattered Light Inevitably, any spacecraft will encounter micrometeoroids. 5.1. Backgrounds The part of JWST that is most vulnerable is the primary mirror. Some of the resulting wavefront degradation from microme- The level of background emission is critical to the depth of teoroid strikes is correctable through regular wavefront control, many JWST imaging and low-spectral-resolution spectroscopic while some of it comprises high frequency terms that cannot be observations. In addition to the unavoidable in-field back- corrected; this latter, cumulative damage was incorporated into grounds from our solar system and our galaxy, as an unbaffled the prelaunch JWST wavefront error budget (Feinberg et al. telescope with a non-zero temperature, JWST sees two 2022). Over the period 2022 March11 to 2023 January12, additional sources of background: (a) the astronomical stray wavefront sensing recorded a total of 25 localized surface light, mainly affecting the near-infrared wavelengths, in which deformations on the primary mirror that are attributed to impact light is scattered into the field of view; and (b) thermal self- by micrometeoroids (a hit rate of 2.5 per month.) With one emission from the glowing mirrors, sunshield, and other major exception, after correction these micrometeoroid strikes observatory components, which affects the mid-infrared. have had no measurable impact on the overall wavefront error. Before launch, the predicted levels of these components carried The outlier is the micrometeoroid hit to segment C3 in the uncertainties of ∼20% and ∼50%, respectively. These back- period 2022 May 22–24, which caused significant uncorrect- grounds, now measured from commissioning and early science able change in the overall figure of that segment. However, the data, are described in a companion PASP paper (Rigby et al. effect was small at the full telescope level because only a small 2023). Here, we briefly summarize the key results. portion of the telescope area was affected. After two First, the astronomical stray light is observed at levels subsequent realignment steps, the telescope was aligned to a considerably below the requirements, and 20% lower than the minimum of 59 nm rms, which is about 9 nm rms above the pre-launch predictions. As a result, deep fields at high ecliptic previous best wave front error rms values. Since the NIRCam latitude will go deeper faster than expected for wave- lengths <5 μm. Available at https://www.stsci.edu/jwst/science-planning/proposal- In the mid-infrared, the main JWST background is a planning-toolbox/psf-simulation-tool. combination of thermal emission from the primary mirror and The impact raised the wave front error of segment C3 from 56 to 280 nm rms. Mirror commanding to adjust segment position and curvature reduced this scattered thermal emission from the sunshield and other parts error to 178 nm rms. This, after dividing by area and adding in quadrature to of the spacecraft. The measured background spectrum is very the other sources of WFE in the telescope, results in ∼9 nm rms increase to the total telescope wave front error. close to the predicted thermal spectrum that was incorporated 16 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. into the exposure time calculator before launch, with somewhat higher backgrounds at at 5–15 μm. The measured thermal background at 10 and 20 μm is close to the requirements values. The mid-infrared backgrounds in the F2550W filter are variable at the level of 7% over timescales of weeks to months, which is less variability than expected, due to greater than expected thermal stability of the primary mirrors. Given these measurements, JWST is indeed, as it was designed to be, limited by the irreducible astronomical background emission, not by stray light or its own self- emission, for all wavelengths <12.5 μm. The control of stray light and thermal emission is an engineering triumph that will translate into substantially better than expected scientific performance for many applications. 5.2. Scattered Light Features During commissioning, a few unexpected stray light features were discovered, characterized and understood, and mitigation Figure 5. The NIRCam scattered light features “the claws” and “the wisps.” The plans were developed. Early science operations have shown the claws are marked by green triangles, and the wisps by cyan triangles. Roughly success of these mitigations. 4% of the pixels in the detector B4 are involved in a claw, and 1.5% in a wisp. There exists a stray light path, termed the “rogue path,” that Data are from detector B4 of the NIRCam short wavelength channel. Of the 8 bypasses the primary and secondary mirrors, directly passes short-wavelength detectors in NIRCam, detector B4 is the one most affected by wisps, and one of the two most affected by claws. Data are from program PID through the aft optics system (AOS) aperture just over the back 01063, in which bright star X Cancri (K = 0.25 mag) was inside the claws of the fine steering mirror, and reaches the science instrument susceptibility zone. Image used is jw01063142001_02101_00001_nrcb4_ra- pick off mirrors (Lightsey et al. 2014). While the rogue path teints.fits. The detector covers 63″ on a side. was anticipated, and steps taken to block it to the extent possible, the NIRCam and NIRSS instruments at times show some unexpected scattered light features which have been coordinates as the telescope dithers. See Figure 5. Claws are identified as being due to this path. The resulting features are rare, but when present, have a brightness of about 10% of the relatively faint, but for some observations such as deep fields zodiacal background. When present, claws occur primarily in these could be significant noise terms if not mitigated. More the NRCA1 or NRCB4 detectors, affecting roughly 5% of the information may be found at the JDox Data Features and Image pixels on those detectors. Since claws move in detector Artifacts page. Here we summarize these features and their coordinates, they are more difficult to subtract off than wisps. mitigations. Claws arise when a bright star is located in the rogue path susceptibility region. 5.2.1 Wisps A few percent of the pixels in four of the eight NIRCam 5.2.3 The Lightsaber short wavelength detectors show small, faint, diffuse features Some NIRISS images show a narrow band of excess stray that are termed “wisps.” See Figure 5. In the B4 detector wisps light running almost horizontally across the detector. Dubbed are always present, with variable brightness that is typically “the lightsaber,” this light originates from the rogue path (see about 10% of the zodiacal background. Fainter wisps have also NIRCam claws above), grazes off the NIRISS entrance housing been seen in detectors A3, A4, and B3. Wisps occur at fixed wall and then experiences double reflections off two mirrors detector positions. The origin of wisps has been traced to inside NIRISS. See Figure 6. The light comes from a reflections from the upper strut that supports the secondary susceptibility region mapped and modeled to be far away from mirror. Wisps have not been seen in the NIRCam long the NIRISS field-of-view (+2°.0 < V2<+5°.0, +12°.4 < wavelength channel. V3<+12°.8). The zodiacal light and any stars in this region contribute to the intensity. When the light is dominated by the 5.2.2 Claws zodiacal light the observed brightness is typically up to 1.5% A minority of NIRCam short wavelength channel images per pixel of the in-field background. When bright (H ∼ 0) Vega from commissioning showed “claws”—a faint diffuse pattern stars are in that region, the observed brightness is up to 10% of scattered light features that moves across the detector per pixel. The lightsaber from bright stars is somewhat 17 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. subtracted from images or low-resolution grism spectra. More information on this and other NIRISS features may be found at the Data Features and Image Artifacts JDox page. The “wisp” features in NIRCam arise from a different optical path: stray light that reflects from the upper strut which supports the secondary mirror. These are fixed in position; the NIRCam team has demonstrated that they subtract out well (see Rieke et al. 2023, this issue). 6. Science Instrument Performance JWST has 17 science instrument modes. Near the end of commissioning, the science performance of each mode in turn was reviewed against criteria developed pre-launch for such parameters as sensitivity, image quality, wavelength calibra- tion, astrometric calibration, ghosts, stability, and so on. As of 2022 July 10, all 17 of the 17 modes have been approved to begin science operations. With the performance of the instruments being uniformly excellent versus requirements and typically also better than pre-flight estimates, the assess- ments of modes as being ready for Cycle 1 science have been quite straightforward. Below we summarize the performance Figure 6. The NIRISS lightsaber. This image in the F150W filter shows a prominent lightsaber stray light feature running almost horizontally across the and any known issues for each instrument mode. lower part of the image. For this pointing there is a bright star in the rogue path A key result of science instrument commissioning is that susceptibility region. The field of view of the detector is 2 2 on a side. Data overall, the JWST science instruments have substantially better from PID 1063. sensitivity than was predicted pre-launch. This result is due to higher science instrument throughput, sharper point spread functions, cleaner mirrors, and lower levels of near-infrared stray light background compared to pre-launch expectations. narrower and shifts position within the broader band caused by The exposure time calculator (ETC) and its underlying Pandeia the zodiacal light. engine were overhauled in v2.0 (released 2022 December) to reflect on-orbit performance, timed to support the Cycle 2 Call 5.2.4 Mitigation of the Stray Light Features for Proposals. The JDox documentation has been similarly Stray light modeling associated the claws and lightsaber with updated. The ETC and its Pandeia engine are the definitive the presence of bright (0–1st Vega mag) stars in the “rogue reference as to sensitivity. For now, we quote several path” susceptibility zone: a region of the field of view located representative measurements and calculations that were several degrees off the telescope boresight (10° for NIRCam, determined during commissioning. 13° for NIRISS). Light from bright stars in this zone can In Figures 7 and 8, we summarize the sensitivity of JWST in bypass the telescope optical train, enter the SIs through the SI two common modes: imaging and emission line spectroscopy, pickoff mirrors, then bounce from non-optical surfaces within and compare to previous and current observatories.We convert the science instruments to the detectors. During commission- from the continuum sensitivity calculated by Pandeia to the ing, we directly mapped the extent of the rogue path more intuitive limiting emission line flux, using the following susceptibility zones for both NIRCam and NIRISS by moving equation: a bright star through the susceptibility zone. pix -23 The claws and lightsaber may be avoided simply by ensuring ff = 10 wl , line cont observations do not place very bright stars in the susceptibility −1 −2 zone. In early science observations, manual checks of where f is the limiting emission line flux in erg s cm , line pix scheduled observations by STScI staff have largely prevented f is the per-pixel continuum sensitivity in Janskies, w is the cont recurrence. STScI plans to update the Astronomers’ Proposal pixel width in Hz, and l is the line width in pixels. Tool to alert users when planned observations would place Across all instruments, JWST has multiple time series bright stars in the susceptibility zone, to give users advance observation (TSO) modes to support observations of transiting predictions of the artifacts and allow replanning observations to exoplanets. During commissioning, observations of the exo- avoid them. The much fainter but ever-present lightsaber due to planet HAT-P-14-b were obtained with three NIR spectro- zodiacal light has been modeled, and can be scaled and scopic modes—the NIRCam grism time series mode, the 18 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 7. Imaging sensitivity for JWST. The Y-axis shows limiting flux density: the faintest point source that can be detected at S/N = 10in an integration time of ten thousand seconds, in units of Janskies. The X-axis is wavelength in units of microns. JWST instruments are shown by the points connected by solid lines, color-coded by instrument as NIRISS (red), NIRCam (blue), MIRI (black). JWST brings two orders of magnitude improvement in imaging sensitivity at 2–3 μm. Actual JWST sensitivity was calculated using Pandeia v2.0, which is the version of the exposure time calculator engine that was released to support the Cycle 2 call for proposals; calculated sensitivities in this released version reflect on-orbit performance as characterized during commissioning. For comparison, comparable sensitivities are shown for other observatories, with points connected by dashed lines: Hubble (WFC3, ACS, and NICMOS instruments); Gemini (GMOS and NIRI instruments); and Spitzer (IRAC and MIPS instruments). Plotted sensitivities for JWST and the comparison observatories are included as supplementary data tables, which also describe computation methods for the sensitivity of these comparison instruments are given in the supplementary information. NIRISS SOSS mode, and the NIRSpec bright object time series programs (or that are being addressed by the operations team in mode—to enable cross-comparison between instruments and the near term). Liens have not prevented any Cycle 1 observations from executing, nor have any observing windows assessment of astrophysical versus instrumental systematics. been missed due to a lien. Liens have caused delays in This target was chosen because, as a massive planet with a high scheduling some programs, and some have required work- surface gravity, it is expected to have a flat transmission arounds to make some programs executable before a permanent spectrum. It also has a relatively bright host star (K = 8.9). s,Vega fix for the lien was ready. In addition the transiting exoplanet L 168-9 b was observed Observations not affected by liens are ready for immediate with the MIRI LRS time series mode. Results are given in the insertion into the observing plan. Following the first several subsections below for each instrument; the overall picture is mode readiness reviews, cycle 1 observations of early release that JWST is returning precise transit spectra with minimal science targets began on 2022 June 20. General Observer and processing. Guaranteed Time Observer cycle 1 programs followed. Mode readiness reviews also included the documentation of any remaining work needed to enable particular use cases (e.g., 6.1. NIRCam Performance nudging a subarray location slightly to not clip a spectrum; The Near Infrared Camera (NIRCam) instrument is updating an astrometric file needed for target acquisition) or the described in the companion PASP special issue paper Rieke identification of performance features that observers should be et al. (2023). Here, we summarize the science performance of made aware of before final planning of their observations (e.g., NIRCam as characterized during commissioning. alerts regarding faster-than-expected saturation due to higher- than-expected throughput, or warnings regarding stray light 6.1.1. NIRCam Imaging features such as the “claws” and “lightsaber,” with tips for mitigation). These issues were captured as “liens” on the mode The throughput of NIRCam meets or slightly exceeds pre- readiness that must be addressed for some specific observing launch expectations for all but a few of the filters in the short 19 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 8. Spectroscopic sensitivity for JWST. The Y-axis shows the limiting line flux, which is the flux of the faintest narrow (spectrally unresolved) emission line in a point source that can be detected at S/N = 10 in an integration time of ten thousands seconds. The JWST instrument sensitivities are plotted in bold lines and labeled in boldface. JWST brings dramatic improvement in spectroscopic sensitivity and spectral resolution compared to previous observatories. As in Figure 7, calculations for the JWST instruments were done using Pandeia v2.0.For comparison, sensitivities are plotted (thin dashed lines) for other observatories: SOFIA (FLITECAM; sensitivity scaled from their exposure time calculator), Gemini (NIRI instrument; sensitivity from instrument website), VLT (ISAAC instrument, sensitivity from their ETC), Keck (MOSFIRE instrument; sensitivity from Wirth et al. 2015), Spitzer (the IRS instrument, for the “L” or low-resolution (R = 60–120) gratings and the “H” high (R = 600) resolution gratings; sensitivity from the SPEC-PET calculator). Plotted sensitivities for JWST and the comparison observatories are included as supplementary data tables, which include more information on how the comparison sensitivites were calculated. Table 3 NIRCam Limiting Point Source Sensitivity Wavelength (μm) 2 3.5 filter F200W F356W Requirement (nJy) 11.4 13.8 ETC prediction (nJy) 10 14.1 Actual (nJy) 6.2 8.9 Note. What is quoted is the faintest flux density (in nanojanskies) that can be detected at S/N = 10 in 10,000 s, for imaging in broad-band filters, assuming a background that is 1.2 times the minimum zodiacal light level. Smaller numbers are better. The equation to convert from flux density f in nJy to AB Figure 9. NIRCam imaging throughput compared to what was assumed in the −32 magnitudes is: mAB = −2.5 log [f (nJy) × 10 ] −48.57. Actual sensitivities pre-launch ETC. For most filters, the observed throughput is higher than the 10 are from Table 2 of Rieke et al. (2023), this issue). pre-launch expectations. Data are from spectrophotometric standard star P330- E observed in program PID 1074. The point-spread function is better than expected, as parameterized by encircled energy or full width at half wavelength channel. In the long wavelength channel, the maximum. The photometric stability is stable to at least 4%, throughput is systematically 20% higher than expected for most and is likely much better. The residual astrometric errors are filters. Figure 9 shows the throughput of NIRCam compared to 2–4 mas per filter across a detector. Optical ghosts are what was assumed in the pre-launch version of the ETC. consistent with expectations from ground tests. Some scattered 20 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 10. The summed broadband (2.4–4.0 μm) grism light curve from the HAT-P-14 observation. The lightcurve from the NIRCam long wavelength channel shows excellent precision (195 ppm standard deviation out of transit) and no significant ramp systematics from charge trapping. For a smaller aperture size of 8 pixels (blue points), a larger jump (∼600 ppm) was observed in the middle of the primary transit, as well as a smaller one (∼100 ppm) toward egress, both due to tilts in a primary mirror segment which occurred during the observation (see Section 4.5.4 and Figure 11). The magnitude of the jumps can be reduced by increasing the extraction aperture size (orange points). Data are from PID 1442. Figure 11. Weak lens data acquired at the same time as the light curve shown in Figure 10. The weak lens image at the start of the time series is shown along with ratios of the images leading up to the tilt event. The tilted segment is very apparent, and the time of the appearance of the tilt coincides with the jump in the white-light curve. light features are seen (the “claws” and “wisps,” described in NIRCam at 3.5 μm, it is 68 times better than IRAC on Spitzer. Section 5.2), which may have some impact on deep imaging. Again, the integration time required to achieve a given limiting Table 3 compares the required, predicted, and on-orbit sensitivity, relative to these previous instruments, scales limiting point-source sensitivity of NIRCam imaging, using the roughly as the square of these advantage factors for back- flux calibration of 2022∼October. The limiting sensitivity is ground-limited broadband imaging. Clearly, NIRCam imaging the flux density of thefaintest point source that can be detected should detect faint objects substantially faster than pre-launch at signal to noise ratio S/N = 10 in an integration time of expectations. 10,000 s. For a representative wavelength for each of the short and long-wavelength channels, the table quotes the require- 6.1.2. NIRCam Grism Time-series ments values, the pre-launch predictions from the exposure time calculator, and the predicted performance assuming the Observations of exoplanet HAT-P-14 b taken during measured on-orbit throughput, PSF, and detector noise levels. commissioning (PID 1442) demonstrated that NIRCam grism Table 3 shows that NIRCam imaging is substantially more time series spectroscopy is working well, meeting performance sensitive than pre-launch expectations. requirements (Schlawin et al. 2023, this issue) after only simple For the common case of background–limited broadband removal of systematic instrumental noise, known as detrending, imaging of a point source, the integration time required (to by fitting an astrophysical lightcurve model times a polynomial reach a given S/N on a target of a given brightness) scales as and exponential model that are both functions of time. The the square of the sensitivity. As such, given Table 3, NIRCam noise for this mode is within 150% of the theoretical photon deep imaging should proceed 3.3 and 2.4 times faster at 2.0 and noise. Additional detrending and analysis will presumably 3.5 μm, respectively, than a system that just met requirements. move even closer to the photon noise limit. We quickly compare this limiting point source sensitivity to With simple detrending, the standard deviation of the transit Hubble and Spitzer, again considering the faintest point source spectrum (as fit by a limb darkened transit model) was 91 ppm detectable at S/N = 10 in 10,000 s. For NIRCam at 1.5 μm, the at R = 100, compared to the expected photon noise of 55 ppm. sensitivity is 9 times better than WFC3-IR on Hubble. For The settling time was observed to be 5–15 minutes. The throughput for NIRCam grism spectroscopy is 20%–40% Comparing JWST/NIRCam F150W to HST/WFC3-IR F160W for a flat- spectrum source. higher than expected for most wavelengths. This may make 21 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 12. Spectrum of a z = 4.39 emission-line galaxy. This spectrum was detected serendipitously in 386 s of exposure time, in NIRCam wide field slitless spectroscopy mode data that targeted standard star P330-E for flux calibration, in PID 1076. Forbidden [O III] 5007 and H alpha are clearly detected. saturation more likely for approved programs, particularly at 6.1.5. NIRCam Coronagraphy long wavelengths. As described in Girard et al. (2022), the performance of the Figures 10 and 11 illustrate the utility of monitoring the star NIRCam coronagraphs exceed expectations. JWST commis- in NIRCam’s short wavelength channel with the weak lens, sioning demonstrated the 5σ PSF-subtracted contrast of the during such grism transit observations in the long wavelength 335R mask at 1″ to be roughly ten times better than channel. In this example, such monitoring with the weak lens −5 requirements, achieving contrasts of ∼4 × 10 to captured two distinct tilts of primary mirror segments (see −6 ∼4 × 10 . Coronagraphic target acquisition was demonstrated Section 4.5.4), which caused jumps in the grism data of to be on par with requirements. 200–1000 ppm, depending on wavelength and extraction Stray light features have not been observed in the NIRCam aperture size. coronagraph fields of view. At this time there is no evidence that stray light will impact coronagraphy or that mitigation strategies need to be taken. 6.1.3. NIRCam Photometric Time-series This mode is not heavily used in Cycle 1. In commissioning, 6.2. NIRISS Performance the mode was checked out using a J = 14 star with a short Vega The Near Infrared Imager and Slitless Spectrograph (NIRISS) 500 s observation, in PID 1068. The performance was nominal: instrument is described in the companion PASP special issue the standard deviation in the normalized flux was measured to paper Doyon et al. (2022). Here, we summarize the science be 0.62% (long wavelength channel) and 1% (short wavelength performance of NIRISS as characterized during commissioning. channel), both close to theoretical expectations. 6.2.1. NIRISS Imaging (Parallel Only) 6.1.4. NIRCam Wide Field Slitless Spectroscopy The imaging performance of NIRISS matches or exceeds NIRCam’s Wide Field Slitless Spectroscopy (WFSS) has expectations. Shortward of 2 μm, the measured throughput is shown excellent performance through commissioning observa- about 20% higher than the instrument team’s expectations, and tions (see Rieke et al., this issue). With the F322W2 filter at about 25%–30% higher than the ETC’s predictions. Longward 3.5 μm, in an integration of 10 s, the continuum sensitivity (S/ of 2 μm, the throughput is about 5%–10% better than the N=10 per resolution element) is 4 microJy, and the emission instrument team’s expectations, and 25% better than thepre- −18 −1 line sensitivity (line flux S/N = 10) is 2.8 × 10 erg s launchETC predicted. Detector noise properties are similar to −2 cm . Figure 12 illustrates the power of the NIRCam/WFSS as measured on the ground, with a slightly higher “dark mode, by showing a serendipitous detection of a line-emitting current” and total noise, likely due to residuals from galaxy at z = 4.39 in the commissioning data. Note the strong uncorrected cosmic ray events. The sky background is lower detections of the [O III] 5007 Å and Hα lines in the spectrum, than predicted before flight. which securely identify the redshift of this galaxy. A similar Table 4 compares the ETC-predicted and actual limiting serendipitous detection of a z = 6.11 galaxy was reported by point-source sensitivity of NIRISS imaging. The limiting Sun et al. (2022). sensitivity is the faintest point source that can be detected at 22 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Table 4 diffuse, mostly circular, events that usually saturate in their NIRISS Limiting Point Source Sensitivity centers. See Section 6.6. Wavelength (μm) 1.15 2 3.5 4.4 6.2.2. NIRISS Single Object Slitless Spectroscopy ETC prediction (nJy) 13 10.2 14.5 22.8 Time series observations of a spectrophotometric standard Actual (nJy) 10.0 8.4 11.8 17.9 A-star (BD+601753, K = 9.6, PID 1091, duration 5 hr) s,Vega Note. What is quoted is the faintest flux density that can be detectedfor a point were used for flux calibration during commissioning. The source at S/N = 10 in 10,000 s. Values are for wide-band filters. Smaller median precision obtained in order 1 was 147 ppm and order 2 numbers are better. The requirement level was set at 13 nJy for the 3.5 μm was 206 ppm, binned to a 22 s integration time. This provides filter. an independent test of the stability and precision achieved for a non-variable star, with only a low-level polynomial trend observed in spectral order 1 and no significant trends observed signal to noise ratio S/N = 10 in an integration time of ten in orders 2 and 3. This observation was affected by a tilt event thousand seconds. The table quotes the pre-launch predictions approximately in the middle of the time series resulting in a from the exposure time calculator, and the predicted perfor- flux jump of a few 100 s ppm. The tilt event was easily detected mance assuming the measured on-orbit throughput, PSF, and by monitoring the PSF shape along the spatial direction, more detector noise levels, and pre-flight background. This is specifically by measuring the second derivative of the PSF predicted performance, not measured performance, as com- which is a good proxy of the FWHM. The flux jump was missioning in general did not involve long integrations. Table 4 demonstrated to be achromatic. predicts that NIRISS parallel imaging should detect faint Observations of the exoplanet HAT-P-14 b (PID 1541, objects substantially faster than pre-launch expectations. The duration 6 hr) returned a point-per-point median precision in photometric stability is better than 1%, based on two the transit depth of 85 ppm at R = 100 for order 1, and 90 ppm measurements of the same standard star made 16 days apart. at R = 100 for order 2. The weighted scatter of the spectrum The field distortion of NIRISS was calibrated using itself was 92 ppm for order 1 and 85 ppm for order 2. Errors in thousands of stars in the LMC astrometric field. Residual the transit depth are within <10%–20% from expectations at astrometric errors with respect to the catalog are 3 mas per axis. this resolution. A tilt event was also noted early in the sequence This is better than the requirement for accurate targeting of well before ingress. The NIRISS SOSS mode readily meets NIRISS sources with NIRSpec multi-object spectroscopy. performance requirements. The NIRISS PSF is better or similar to pre-flight WebbPSF Throughput in this mode is 25% better for order 1 near the predictions as defined by encircled energy, FWHM, and blaze wavelength at 1.3 μm, and ∼50% better for order 2. ellipticity. There is very little field-dependence of the PSF as There appears to be no significant noise penalty to operate at measured by these parameters. NIRISS imaging of very bright 70%–75% of the saturation level. Observations should stars shows an extra diffraction spike offset from the vertical by generally not exceed 35,000 ADUs (56,000 e-) to avoid extra between −10° and +20° with an angle that rotates smoothly as noise. There is also a trade-off with observing efficiency; users one moves from the right to the left edge of the detector. This may opt to saturate part of the spectrum to improve the duty cycle. spike is stronger at shorter wavelengths with an integrated intensity ∼70% of the vertical diffraction spike for the shortest 6.2.3. NIRISS Wide Field Slitless Spectroscopy wavelength filter, F090W. The cause of this spike has been traced to diamond turning residual wave front errors in some of The throughput for NIRISS wide field slitless spectroscopy the NIRISS off-axis mirrors. (WFSS, Willott et al. 2022; Figure 13) is generally 30% better NIRISS images show a narrow band of excess stray light than expected from the prelaunch ETC. As a result, the running almost horizontally across the detector, dubbed “the predicted on-sky sensitivities (assuming the pre-flight back- lightsaber.” See Section 5.2. ground model) are better than the prelaunch ETC predictions Imaging ghosts were seen in ground testing of NIRISS and by 7%–20%, similar to the results in imaging mode at these similar behavior is seen in flight. They are the result of internal wavelengths. The GR150C filter has higher throughput than reflections in the optical system. Ghost positions are predictable GR150R at wavelengths below 1.2 μm, likely due to a different for each filter as they form at a position that is symmetric anti-reflective coating, so is preferred for F090W and F115W if around the ghost axis point (GAP). The GAPs for each filter only one grism is used (normally both are used to mitigate were measured during commissioning and the intensity of the contamination). ghosts was found to be ∼1% of the original source intensity. Trace positions, curvature, dispersion and spectral resolution The cosmic ray rate at L2 is similar to predictions. However for WFSS are close to those measured on the ground. In there is a much higher rate of large “snowballs” that appear as addition to dispersed versions of imaging ghosts, additional 23 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. enables sub λ/D imaging but with better throughput at the price of a lower contrast at small separation. Observations were performed on 4 targets, all presumed to be single stars, but one was found to have a companion at 0 15 with a contrast of 1.7 mag. Interferometry with JWST has shown unsurpassed fringe amplitude stability, promising valuable complementarity to ground-based interferometry’s significantly higher resolution. 6.3. NIRSpec Performance A detailed description of the Near-Infrared Spectrograph (NIRSpec) instrument and of its pre-launch performance can be found in a series of four papers: overview (Jakobsen et al. 2022); multi-object spectroscopy (MOS) mode (Ferruit et al. 2022); integral field spectroscopy (IFS) mode (Böker et al. 2022);and exoplanet time series (Birkmann et al. 2022). The on-orbit science performance of NIRSpec is described in the companion PASP special issue paper Böker et al. (2023).Herewebrieflysummarize those results. All modes of NIRSpec are working well, in general better than pre-launch expectations. Both of NIRSpec’s two detectors Figure 13. NIRISS WFSS. Simultaneous spectroscopy of thousands of stars in show noise levels, in the actual cosmic ray environment of L2, the NIRISS focus field of the LMC. Configuration is grism GR150C and filter similar to or lower than ETC predictions. The excellent optical F115W. The field of view of the detector is 2 2 on a side. Data from PID 1085. quality of the telescope translates to lower-than-expected slit losses in the multi-object and fixed slit modes of NIRSpec. For the 200 mas wide microshutters, this translates to increased ghosts due to the grisms are seen, but at a relatively low photon conversion efficiency of 2.5% at 5 μm, >7.5% below intensity level. The lightsaber scattered light feature is also 3 μm, and >10% below 1 μm. NIRSpec bright object time apparent for WFSS. For the typical intensity where the series mode has demonstrated precision in the transit depth of lightsaber is dominated by zodiacal light, this emission is 50–60 ppm per point (Espinoza et al. 2023). Both target included in the WFSS background reference files, so will be acquisition methods that are specific to NIRSpec, wide-aperture subtracted off in the pipeline. target acquisition (WATA) and microshutter assembly target acquisition (MSATA), are working well. 6.2.4. NIRISS Aperture Masking Interferometry Figure 4 of Böker et al. (2022) shows the measured on-orbit The AMI mode (Sivaramakrishnan et al. 2022) features a sensitivity of NIRSpec for both multi-object spectroscopy and seven-hole non-redundant mask that enables high-contrast integral field spectroscopy. Across the board, the sensitivity is −3 −4 (10 –10 ) imaging at sub λ/D angular separations better than pre-launch predictions. (0 1–0 5) over three medium-band filters (F380M, F430M, The operability rate of the un-vignetted microshutters is 82.5% (Rawle et al. 2022), with electrical short masking now F480M). This mode was successfully demonstrated through the easy detection of AB Dor C, a companion with a separation of the primary cause of non-operable shutters. The resulting ∼0 3 with a contrast ratio of 4.5 mag. The noise floor of this multiplexing levels are within 10% of the pre-launch levels and data set is 6.5–7.0 mag (3σ), very close to the photon noise are still excellent, with the possibility, for high target densities, floor limit of ∼7.5 mag. This is the first space-based to observe more than 200 scientific targets at a time at low demonstration of both infrared interferometry and non- spectral resolution or close to 60 at medium / high spectral redundant aperture masking. The nominal operational concept resolution (see Figure 14). for AMI requires staring (rather than dithering) on the science target, followed by a similar observation on an isolated 6.4. MIRI Performance reference star. Target acquisition (TA) places targets at the same detector location, and TA accuracy well within 0.1 pixel The Mid-Infrared Instrument (MIRI; Wright et al. 2023, this was demonstrated. Kernel Phase Interferometry (KPI) is the full issue) is the only instrument on JWST that operates beyond pupil generalization of AMI used without the NRM mask but 5 μm; as such it supports a broad range of measurement types: using a similar Fourier-based removal of instrument effects (1) standard imaging; (2) high contrast (coronagraphic) from the data (Kammerer et al. 2023). Like AMI, KPI also imaging; (3) low resolution spectroscopy (LRS) with and 24 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 14. Multi-object spectroscopy with the NIRSpec microshutter array. The low-spectral resolution prism mode was used, with a total of 235 microshutter slitlets opened, to capture spectral features from the diffuse interstellar medium in a region close to the Galactic Center. Extracted example spectra (not flux calibrated) show multiple emission and absorption line features. Data from PID 1448. without a slit; and (4) medium resolution integral field unit 23 μm, and providing raw contrasts of ∼10 (at 6λ/D). The spectroscopy (MRS). MIRI coronagraphs are performing significantly better than The MIRI imager uses most of a 1024 × 1024 pixel detector anticipated (Boccaletti et al. 2022), in part due to the excellent array to provide a field of view of 74″ × 113″ with eight image quality of the telescope. Other key factors are the broadband filters, providing bands starting at 5.6 μm and achieved precision of alignment of MIRI and the ISIM to the spaced every factor of ∼1.2–1.3 to 25.5 μm. An additional 6% telescope in pupil shear and focus, both of which are better than bandwidth filter is centered on the aromatic feature at 11.3 μm. the budget allocations. In common with other coronagraphs on The images are all diffraction limited, and they are Nyquist JWST, subtraction of a PSF reference star is necessary to sampled for wavelengths longer than 6.25 μm. The overall achieve the best performance. This technique is very sensitive throughput of the OTE and MIRI imager exceeds the prelaunch to high order wave front error (i.e., the shape of the PSF wings) expectations, particularly for wavelengths at 10 μm and longer. and thus scheduling needs to take account of tilt events and The result is that the imager sensitivity is improved over pre- routine telescope mirror alignments. launch predictions. The MIRI low resolution spectrometer(LRS)also lies to one JWST provides subarcsec imaging in the mid-infrared, with side of the imager. It has very high throughput (∼80%) and a beam 50 times smaller in area than that of the Spitzer Space nominal resolution of λ/Δλ = 100, designed for spectroscopy Telescope, the previous most capable space infrared telescope. of very faint objects. Its performance is optimized for 5–10 μm, Figure 15 illustrates this capability, which opens an entirely but it is usable to 14 μm. There is also a slitless mode of LRS new realm of study of the structure of mid-infrared sources. for exoplanet transits. During commissioning, a transit Where MIRI is limited by natural backgrounds (for wave- spectrum of the exoplanet L168-9b was obtained, demonstrat- lengths of 5 to at least 12.5 μm, Rigby et al. 2023, this issue), ing calibration to ∼25 ppm with a spectral resolution of ∼50 at its sensitivity for point sources is about 50 times better than that 7.5 μm (Bouwman et al. 2023). LRS transit spectroscopy will of the Spitzer Space Telescope. This gain is reduced at longer be used to study organics and water in the atmospheres of wavelengths due to telescope emission, but in deep exposures exoplanets, helping determine abundances of these molecules the lack of confusion noise with the small beam of MIRI still and whether non-equilibrium chemistry is at work influen- provides significant gains over Spitzer. cing them. Four coronagraphs lie along a side of the imaging field of The MIRI medium resolution spectrometer (MRS) covers by view, optimized for wavelengths of 10.58, 11.30, 15.50, and design the full 5–28.3 μm range (currently calibrated to 25 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 15. NGC 7320, the foreground galaxy in Stephan’s Quintet, as seen in MIRI imaging. The major axis of the galaxy is about 2′ long. The MIRI image is in three bands: F770W, F1000W, and F1500W, respectively shown in false color as blue, green, and red. In the image, red denotes dusty, star-forming regions, and blue can show either stars, or the strong and prominent 7.7 μm aromatic band (which dominates the image). What appears in the visible to be a typical dusty spiral galaxy is lit up in this MIRI image, with complex structure tracing where aromatic molecules are heated by hot stars. Data from PID 2732 (Pontoppidan et al. 2022). Credit: NASA/ESA/CSA/STScI. 27.9 μm). It uses integral field units as its inputs, with fields before launch, the commissioning data confirm the design expectations. increasing with increasing wavelength from 3 2 × 3 7to 6 6 × 7 7. The spectra and spatial information are arranged 6.5. Using Two Science Instruments in Parallel over two 1024 × 1024 pixel detector arrays to provide spectral resolution, λ/Δλ,of 3000 for wavelengths shorter than JWST supports parallel observing, in which two science ∼11.7 μm, and 1500 out to 28.3 μm, along with near- instruments are used simultaneously. There are two types of diffraction-limited imaging. The high spectral resolution, parallel: coordinated parallels and pure parallels. sensitivity, and imaging capability of the MRS together provide breakthrough capabilities for mid-infrared spectrosc- 6.5.1. Coordinated Parallels opy. For example, MRS gives access to many molecular Coordinated parallels are when two science instruments are transitions (e.g., organic molecules, H2) and to the full suite of used together for the same observing program. A total of 170 neon fine structure lines: [Ne II] 12.81 μm, [Ne III] 15.56 μm, coordinated parallel visits were successfully executed during [Ne V] 14.32 μm, and [Ne VI] 7.64 μm, which are very science instrument commissioning (Figure 17), and worked powerful for determining the excitation mechanism of emission well with only one issue (discussed just below). Parallel line objects. Figure 16 illustrates the use of these capabilities to operations are designed such that mechanism movements of dissect the planetary nebula NGC 6543. These data also each instrument do not disturb the operation of the other; this illustrate well the quality of the calibration of the MRS coordination is working as designed. distortion, fields of view and astrometry achieved during While there was no requirement that all of the eleven commissioning. While the point source illumination used coordinated templates be exercised during commissioning, in during ISIM testing on the ground was too faint at mid-IR the end seven of the templates were. The four templates that wavelengths to characterize well these aspects of the MRS were not exercised are: NIRCam Imaging + NIRISS WFSS; 26 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 16. NGC 6543, the Catseye planetary nebula, imaged at 5.6 μm (left panel), and dissected by the MIRI MRS in mosaics using its integral field unit (right two panels). The maximum extent of the nebula in the image is ∼25.” The two layers in the spectral/spatial cube show the nebula at sub-arsecond resolution in Humphreys α 12.37 μm and the fine structure line of [Ne II] 12.8 μm. MRS has four integral field units; the field of view increases with wavelength. Data are from programs PID 1023, 1031, and 1047. Credit: B. Vandenbussche Figure 17. Observations of the JWST astrometric calibration field and surroundings in the Large Magellanic Cloud. These observations from PID 1473, previously released, were taken just after the completion of multi-instrument optical alignment to assess and demonstrate image quality in all instruments. This also served as an engineering test of coordinated parallel operations: this included the first uses of NIRCam+MIRI imaging, MIRI+NIRCam imaging, and NIRSpec MOS spectroscopy +NIRCam imaging, as well as NIRCam+NIRISS imaging and NIRCam+FGS imaging, with many combinations of filters and parallel-optimized dither patterns. The pointings on sky are all shown in the correct relative orientations and scales relative to one another; the figure is approximately 18′ horizontal by 9′ tall. 27 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. NIRCam WFSS + MIRI Imaging; NIRCam WFSS + NIRISS 7. Science Operations Status Imaging; and MIRI Imaging + NIRISS WFSS. We see no Science operations includes managing the proposal process, reason why these templates should not work, as these WFSS preparing visits for execution, processing data from the templates are identical to the imaging templates except for observatory, and making data products available via the MAST putting a grism in the beam instead of a filter. archive. JWST science operations at STScI provided out- The one issue with parallel observing that was identified was standing service during commissioning and the transition to this: when all 10 NIRCam detectors were used in parallel with Cycle 1. Complex software and processes worked thanks to another instrument, data from one instrument could under some many tests and rehearsals prior to launch. Planned updates and circumstances be partially overwritten by data from the other unexpected issues were handled via operational work-arounds instrument. Flight software was patched on 2022 June 24 to fix and software updates. Nevertheless, some significant issues this issue. Data taken earlier may be affected by partial data remain and warrant attention, as discussed below. overwrite. The Operations Scripts Subsystem (OSS) carries out the science observations by executing the Observation Plan (OP), which is uploaded weekly and executed by OSS autonomously, 6.5.2. Pure Parallels often while JWST is not in contact with the Mission Operations Pure parallels are when an observing program makes use of Center. OSS was used to conduct most of the telescope and parallel observing opportunities from other accepted proposals. science instrument commissioning activities, and has been Thus, in pure parallel mode two science programs execute at patched several times to resolve issues identified during once. Pure parallel mode was not exercised in commissioning. commissioning. Pure parallel programs were not scheduled for the first four During commissioning, the JWST Exposure Time Calculator months of Cycle 1; they were enabled in 2022 November. The (ETC) reflected pre-launch expectations. As described above, concern had been the tight (11%) margin on data downlink in end-to-end flight performance is typically better, in some cases Cycle 1. Data volume is used as a constraint when scheduling by a significant amount. The Cycle 2 Call for Proposals, JWST observations; the schedule is built to keep the onboard released on 2022 November 15, uses sensitivities measured solid state recorder from filling, since if the recorder fills JWST with in-flight data. In the beginning of Cycle 1, observers were would halt observing and sit idle until the next ground contact. able to use performance information above and in JDox to scale The project will continue to assess how well data volume is ETC results. For the typical case of higher than expected managed through scheduling and downlink performance during throughput, the main concern for Cycle 1 programs was science operations. unexpected saturation during target acquisition or early in an integration. Saturation later in an integration will generally yield S/N comparable to or better than predicted by the ETC. 6.6. Cross-instrument Detector Topic: Cosmic Rays The Astronomer’s Proposal Tool (APT) and the Micro- Observed cosmic ray rates and properties are largely in line shutter Array Planning Tool now reflect flight measurements of with expectations. The vast majority of cosmic ray impacts aperture placement in the focal plane and distortion across directly affect only one or 2 pixels, but there are also instrument apertures. Additional refinements are expected, but uncommon events that affect hundreds of pixels. These large the impact on observers should be negligible. Functionality will events are colloquially termed snowballs. There are also large also evolve slightly to accommodate insights from radiation events that affect the MIRI detectors, which are called commissioning. “shower” events. Long range plan windows have been published for most The current JWST data reduction pipeline handles the first Cycle 1 visits. Investigators can search for program information order effects from cosmic ray events but the large number of and click the Visit Status Information link for more informa- electrons that result from a snowball event have secondary tion. Whenever possible, visits are scheduled and executed effects that are not currently corrected in the pipeline. Residuals within their assigned plan window, but operational issues may have a circular appearance with alternating light and dark bands cause plan windows to change, visits not to schedule, or a few tens of pixels across. Dithering exposures is the current scheduled visits not to execute. This is particularly true early in recommended mitigation strategy. A four-point or larger dither Cycle 1. Visits that do not execute are usually rescheduled. pattern will allow the pipeline outlier detection routine to JWST uses the Guide Star Catalog (GSC) to specify significantly improve the final combined image. Work is in astrometry and photometry of guide stars and reference stars progress to improve the calibration pipeline’s detection and for guide star acquisition. During commissioning, a few guide handling of regions affected by snowballs. star acquisitions failed because the guide “star” was actually a galaxy resolved by FGS or because coordinates in the GSC https://jwst-docs.stsci.edu/jwst-mid-infrared-instrument/miri-features- and-caveats were wrong by an arcsec or more. The same will happen during 28 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. Figure 18. Multi-wavelength image mosaic of the Fine Phasing field, demonstrating the efficacy of automated data processing. This is a color version of the image that was released in 2022 Marchjust after the completion of telescope fine phasing; the blue, green and red channels show NIRCam F115W, F200W, and F356W. It was produced simply by retrieving the automatically produced Level 3 mosaic data products from MAST, and opening them in ds9 to make an RGB image. The only manual step applied was a simple median subtraction of background level, used primarily to remove a temporarily high background in the 3.5 μm channel in these early data, which appeared because the instruments were not yet fully cooled. The automated pipeline generates a nearly science-ready product with all mosaic tiles stitched and the several filters registered together. The field of view is 6′ by 2 6. Data from PID 1160 observation 22. Cycle 1, but the frequency of failures will decrease with transform the WCS onto the Gaia frame. Photometric improvements to the GSC and operational procedures. calibration will reflect pre-launch expectations until calibration After processing all data taken during commissioning, the reference data are updated. JWST data management subsystem generated 1.5M files, Users should note the calibration pedigree of their data files requiring 55 TB to store a single copy. Users should be and of data files used in publications. Consult JWST documentation for a description of known calibration issues selective about which types of data products they download to and their resolution in successive versions of calibration conserve network bandwidth and storage. Users can download reference data and software. and rerun the JWST calibration pipeline with custom Data processing typically takes about a day, but can take parameters tailored to their needs, or even modify or add steps longer if downlink, data transfer, or processing issues arise. For to the calibration pipeline. complicated modes (e.g., multi-object spectroscopy) or asso- JWST data are now available in the MAST archive (see ciations containing many exposures, the calibration pipeline Accessing JWST Data), as well as in the JWST archives at can take several hours to run or even days in extreme cases. ESA and CSA. See Figure 18. Performance will improve over time, but for now the main Because Cycle 1 observing began shortly after observatory focus is functionality. commissioning, data products at first had relatively poor calibration accuracy and contained calibration artifacts. As instrument teams generate new calibration reference files 7.1. JWST User Documentation System (JDox) Updates throughout Cycle 1, the calibration accuracy will improve. Science operations will reprocess old data and make improved The JWST User Documentation system (JDox) provides data products available in MAST, as needed when new comprehensive information about the JWST Proposing pro- calibration files become available. cess, the Observatory and science instruments, science data Users should be especially vigilant about astrometric and characteristics and data access, data pipeline processing and photometric calibration errors in data products downloaded calibration, as well as introductions to post pipeline data early in Cycle 1. Errors in guide star coordinates or focal plane analysis tools and training materials. geometry will propagate to the world coordinate system (WCS) JDox was updated on 2022 July 12 to describe data features for exposure-level data products. Higher-level products (e.g., and image artifacts seen in flight data, known shortcomings in mosaics) may use tweakreg and Gaia sources in each image to current data pipeline products, and other articles to help users 29 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. understand the status of the Observatory and science instru- instruments have demonstrated the ability to precisely capture ments. The next major release of JDox coincided with the spectra of transiting exoplanets with initial precision better than release of the Cycle 2 Call for Proposals on 2022 Novem- 100 ppm per measurement point, with limiting performance ber 15. expected to be well below that level. JWST has tracked solar −1 Users can refer to the Latest Update information box at the system objects at speeds up to 67 mas s , more than twice as bottom of each JDox article to see when that article was fast as the requirement. JWST has detected faint galaxies with updated. JDox also maintains a summary page showing fluxes of several nano-Jansky, and observed targets as bright as recently changed articles by titles and when they were updated. Jupiter. JWST has obtained infrared spectra of hundreds of stars simultaneously in a dense starfield toward the Galactic center, as well as integral field spectroscopy of planetary 8. Conclusions nebulae and a Seyfert nucleus at unprecedented sensitivity. This article summarizes the science performance of JWST as Data from these and all other commissioning activities are characterized by the six month commissioning period. Almost available to the scientific community via the MAST archive. across the board, the science performance of JWST is better As each of JWST’s seventeen science instrument modes than expected. The optics are better aligned, the point-spread finished its commissioning activities, it was reviewed against function is sharper with higher encircled energy, and the optical mode-specific readiness criteria for science instrument perfor- performance is more time-stable than requirements. The fine mance. All JWST observing modes have been reviewed and guidance system points the observatory several times more confirmed to be ready for science use. In most cases, the modes accurately and precisely than required. The mirrors are cleaner surpass performance requirements. than requirements, which translates into lower-than-expected Continued analyses as well as Cycle 1 calibrations will levels of near-infrared stray light, meaning that the <5 μm sky further improve the characterization of the science instruments. background will be darker for JWST than expected.As it was Updated knowledge has been reflected in updates to data designed to be, JWST is indeed limited by irreducible pipeline reference files and algorithms, as well as thorough astronomical backgrounds, not by stray light or its own self- revisions to the JDox documentation and updates to the JWST emission, for all wavelengths <12.5 micron. The science proposal tools that were timed to support the Cycle 2 Call for instruments have generally higher total system throughput than Proposals. pre-launch expectations. Detector noise properties are similar to JWST is the product of the efforts of approximately 20,000 ground tests, albeit with higher rates of cosmic rays, as people in an international team. Commissioning JWST and expected in deep space. Collectively, these factors translate into characterizing science performance is the result of tremendous substantially better sensitivity for most instrument modes than effort by the JWST commissioning team over six months. The was assumed in the exposure time calculator for Cycle 1 achieved performance is the result of efforts over the many observation planning, in many cases by tens of percent. In most years leading to launch by team members across much of the cases, JWST will go deeper faster than expected. As a key globe. Given the measured performance described in this example, NIRCam deep imaging of point sourcesto a given document, the JWST mission entered Cycle 1 having depthshould proceed 3.3 and 2.4 times faster at 2.0 and demonstrated that the observatory exceeds its demanding pre- 3.5 μm, respectively, than a system that just met requirements. launch performance expectations. With revolutionary capabil- In addition, JWST has enough propellant onboard to last at ities, JWST has begun the first of many years of scientific least 20 yr. discovery. This characterization of science performance undergirds the key conclusion of commissioning: that JWST is fully capable of achieving the discoveries for which it was built. JWST was This work is based on observations made with the NASA/ envisioned “to enable fundamental breakthroughs in our ESA/CSA James Webb Space Telescope. The data were understanding of the formation and evolution of galaxies, obtained from the Mikulski Archive for Space Telescopes at stars, and planetary systems” (Gardner et al. 2006)—we now the Space Telescope Science Institute, which is operated by the know with certainty that it will. The telescope and instrument Association of Universities for Research in Astronomy, Inc., suite have demonstrated the sensitivity, stability, image quality, under NASA contract NAS 5-03127 for JWST. These and spectral range that are necessary to transform our observations are associated with many programs from the understanding of the cosmos through observations spanning commissioning period. Technical contributions were carried from near-earth asteroids to the most distant galaxies. Commissioning proved the observatory’s capabilities That time series precision of < 100 ppm is measured at spectral resolutions of R = 100 at wavelengths below 5 μm and R = 50 at longer wavelengths, with through approximately 2300 visits of commissioning observa- minimal detrending. tions, which exercised the same science instrument modes that The Cycle 2 Call for Proposals was released on 2022 November 15, with will be used in normal science operations. All four science proposals due 2023 January 27. 30 Publications of the Astronomical Society of the Pacific, 135:048001 (31pp), 2023 April Rigby et al. out at the Jet Propulsion Laboratory, California Institute of Ferruit, P., Jakobsen, P., Giardino, G., et al. 2022, A&A, 661, A81 Gardner, J. P., Mather, J. C., Clampin, M., et al. 2006, SSR, 123, 485 Technology, under a contract with the National Aeronautics Gaspar, A., Rieke, G. H., Guillard, P., et al. 2021, PASP, 133, 4504 and Space Administration. Girard, J. H., Leisenring, J., Kammerer, J., et al. 2022, Proc. SPIE, Facilities: JWST (NIRCam, NIRISS, NIRSpec, MIRI). 12180, 121803Q Jakobsen, P., Ferruit, P., Alves de Oliveira, C., et al. 2022, A&A, 661, A80 Software: Pandeia (Pontoppidan et al. 2016). Kammerer, J., Girard, J., Carter, A., et al. 2022, Proc. SPIE, 12180, 121803N Kammerer, J., Cooper, R. A., Vamdal, T., et al. 2023, PASP, 135, 1043 ORCID iDs Lallo, M. D. 2012, Proc. SPIE, 51, 011011 Lightsey, P., Knight, J. S., Golnick, G., et al. 2014, Proc. SPIE, 9143, 12 Jane Rigby https://orcid.org/0000-0002-7627-6551 Maréchal, A. 1947, Rev. d’Opt, 26, 257 Michael McElwain https://orcid.org/0000-0003-0241-8956 McElwain, M. W., Feinberg, L. D., Perrin, M. D., et al. 2023, PASP, submitted Menzel, M., Davis, M., Parrish, K., et al. 2023, PASP, submitted Pierre Ferruit https://orcid.org/0000-0001-8895-0606 Pontoppidan, K. M., Pickering, T. E., Laidler, V. G., et al. 2016, Proc. SPIE, Marcia Rieke https://orcid.org/0000-0002-7893-6170 9910, 991016 Chris Willott https://orcid.org/0000-0002-4201-7367 Pontoppidan, K., Barrientes, J., Blome, C., et al. 2022, ApJL, 936, 14 Rawle, T., Giardino, G, Franz, D. E., et al. 2022, Proc. SPIE, 12180 121803R References Rieke, M., Kelly, D. M., Misselt, K., et al. 2023, PASP, submitted Rigby, J., Lightsey, P., García Marín, M., et al. 2023, PASP, submitted Anderson, J., & King, I. R. 2000, PASP, 112, 1360 Schlawin, E., Beatty, T., Brooks, B., et al. 2023, PASP, submitted Birkmann, S. M., Ferruit, P., Giardino, G., et al. 2022, A&A, 661, A83 Sivaramakrishnan, A., Tuthill, P., Lloyd, J. P., et al. 2022, PASP, 135 Boccaletti, A., Cossou, C., Baudoz, P., et al. 2022, A&A, 667, 165 Böker, T., Beck, T. L., Birkmann, S. M., et al. 2023, PASP, 135, 038001 Sun, F., Egami, E., Pirzkal, N., et al. 2022, ApJL, 936, 8 Böker, T., Arribas, S., Lützgendorf, N., et al. 2022, A&A, 661, A82 Willott, C., Doyon, R., Albert, L., et al. 2022, PASP, 134, 5002 Bouwman, J., Kendrew, S., Greene, T., et al. 2023, arXiv:2211.16123 Wirth, G., Trump, J. R., Buillermo, B., et al. 2015, AJ, 150, 153 Doyon, R., Willott, C., Hutchings, J., et al. 2023, PASP, submitted Wright, G., Rieke, G. H., Glasse, A., et al. 2023, PASP, submitted Espinoza, N., Úbeda, L., Birkmann, S. M., et al. 2023, PASP, 135, 1043
Journal
Publications of the Astronomical Society of the Pacific – IOP Publishing
Published: Apr 1, 2023
Keywords: Observatories; Infrared astronomy; Astronomical instrumentation