In This Issue
Fall Issue of The Bridge on Space Exploration
September 1, 2021 Volume 51 Issue 3
Close collaboration between engineering and science has enabled marvels of space exploration over decades. Eight exemplary missions are described in this issue, conveying the excitement, challenges, and breakthroughs involved in efforts to better understand the wonders and mysteries of this solar system.

The James Webb Space Telescope: Where Engineering and Science Merge

Monday, September 13, 2021

Author: Kenneth Sembach and Lee D. Feinberg

JWST’s engineering advances and scientific discoveries will set the course for the future of astrophysics and space-based observatory development.

The decision to stare at a “blank” region of the sky with the Hubble Space Telescope (HST) for several weeks in 1995 forever changed perceptions of the universe. The purpose was clear—to see whether the sky really was devoid of light as it appeared to be—and motivated by an underlying question about whether distant (younger) galaxies look the same as those in the nearby (older) universe.

Sembach and Feinberg fig 1.gif
FIGURE 1 The Hubble Deep Field, containing thousands of galaxies. (Right) The Orion Nebula, a star-forming region shrouded in gas and dust. Image credits: NASA/STScI.

The Hubble Deep Field showed that the universe is dynamic, a fabric of organized chaos with galaxies colliding, merging, and evolving. Within this image—encompassing no more sky than can be seen by looking through a drinking straw—were thousands of galaxies, each composed of hundreds of billions of stars (figure 1). It was soon hypothesized that even Hubble may not be able to see all that was in the Deep Field,[1] for if the universe were expanding as predicted by the Big Bang theory, galaxies would appear redder the farther they were from Earth, until eventually the nascent galaxies would be so red that Hubble could not see them.[2]

A question naturally arose that could not be answered without technological innovation and an observatory more advanced than Hubble: When did the first ­galaxies form?


Motivated by earlier observations preceding the Deep Field, the Association of Universities for Research in Astronomy commissioned the “HST and Beyond” Committee to formulate a vision for ultraviolet-­optical-infrared space astronomy after Hubble. The com­mittee recommended that NASA develop a large space observatory optimized for imaging and spectroscopy at infrared wavelengths. Such an observatory would “be an essential tool in an ambitious program of study in many areas of astronomy” and “especially powerful in studying the origin and evolution of galaxies” (Dressler et al. 1996, pp. ix–x).

Ultraviolet and optical light can penetrate only so deeply into, or escape from, regions that are heavily obscured by cosmic dust, a fundamental component of galaxies and star-forming regions. A telescope with sensitivity at infrared wavelengths would see into, or even through, this interstellar screening material to reveal stars and planetary systems embedded therein, as well as galaxies in the distant universe.

In 2001 the National Academies’ decadal survey in astronomy and astrophysics identified a next-generation space telescope as an imperative for the decade because it would “reveal the first epoch of star formation and trace the evolution of galaxies from their birth to the present...and provide a unique window onto the birth of stars and planets in our own galaxy” (NRC 2001, p. 36). The recommended large space telescope, with 100 times the sensitivity and 10 times the image sharpness of the HST at infrared wavelengths, began development in 2002 as the James Webb Space Telescope (JWST).[3] The JWST astronomical observatory is being developed by NASA with substantial contributions from the European and Canadian Space Agencies and will be launched into space in late 2021 after nearly 2 decades of development and preparation.

The scientific impetus to understand how galaxies formed in the early universe was the seed for a partnership between engineering and science that is advancing both disciplines. Its aspirational science goals led to a notional experimental architecture with accompanying technical specifications, and then development of the requisite technologies (figure 2). Iterating with science requirements, the architecture was built and verified through rigorous testing to be capable of achieving the observatory’s science goals.

Sembach and Feinberg fig 2.gif

The engineering-science cycle comprising JWST’s engineering advances and forthcoming scientific discoveries is setting the course for the future of astro­physics and space-based observatory development.

Engineering Specifications and Technologies

Because the first galaxies to form in the universe are fainter and redder than can be observed with Hubble, a larger space telescope is needed to collect more light,[4] equipped with instrumentation sensitive to near- and mid-infrared wavelengths (1–28 mm), and cooled to cryogenic temperatures to limit thermal backgrounds to manageable levels. These requirements drive choices of telescope size, observatory architecture, orbit, power, and so forth, resulting in a slew of new technologies necessary for JWST to achieve its science goals (Stockman 1997).

Many technologies that enabled breakthrough science with Hubble are not suitable for JWST. For example, scaling Hubble’s 800 kg fused silica glass mirror up to the 6.5 m diameter specified for JWST would consume over 90 percent of the 6500 kg mass budget of the entire observatory. Furthermore, a monolithic mirror of this size would not fit in available rocket shrouds. This led to the eventual choice of a lighter, segmented, deployable primary mirror design.

To meet the stringent wavefront and temperature requirements driven by science, NASA had to develop a new isotropic form of optical-grade beryllium, called O-30, for use on JWST. Achieving a 10x lower areal mass than Hubble’s primary mirror was key to properly sizing the telescope, as was the development of a deployable secondary mirror at the end of a 25-foot lightweight tripod boom that folds for stowage in the rocket shroud until deployment in space. This engineering contributed to an overall observatory weight roughly half that of Hubble, despite JWST being a much larger observatory.[5]

Cooling the telescope is a primary driving requirement to ensure that light detected is truly astronomical in nature rather than heat from the observatory itself. This requirement ripples through specifications for ­other systems and technology choices.

For example, the telescope must be shielded from sunlight, earthlight, and moonlight, all of which would heat the observatory. JWST uses a five-layer deployable sunshield made of a mylar-like material (aluminum and doped-silicon-coated Kapton) that blocks these heat sources and reduces the heat impinging on the telescope by a factor of approximately 106, allowing the telescope to cool passively to ~40 K (−233°C) by radiating its residual heat into empty space.

All instrumentation is also thermally shielded from the telescope. Operating at a distance of about a million miles at a location in space coaligned with the Sun and Earth (the Lagrange L2 point) rather than orbiting the Earth makes shielding the telescope from heat sources easier and dramatically reduces time-variable thermal loading on the observatory (figure 3).

Sembach and Feinberg fig 3.gif
FIGURE 3 JWST field of regard and orientation relative to the Sun. Proper sunshield orientation and judicious scheduling to avoid moonlight result in an instantaneous field of regard confined between 85° and 135° of the Sun-Earth line, but all points on the sky are visible at some time during the year. Reproduced from JWST documentation online at https://jws

The need to cool JWST to cryogenic temperatures strongly constrained the materials that could be used for its mirrors and optical telescope assembly. ­Beryllium and carbon fiber, the two materials of choice for the mirrors and structure, respectively, have low ­coefficients of thermal expansion. Both can be formed into stiff lightweight structures, which is important because nearly every material property, from damping to stiffness to adhesion, changes as temperature becomes cryogenic. All material properties had to be engineered with adequate margin using new levels of integrated modeling with unprecedented mesh fidelities. The need for mechanism deployment, latching, and operation at cryogenic temperatures posed design and fabrication challenges as well.

To achieve high sensitivity to faint sources of infrared light requires infrared detectors more advanced than previous generations on Hubble and other missions. The JWST instruments incorporate two types of infrared detectors: 4-megapixel mercury-cadmium-telluride (HgCdTe) detectors in the near-infrared instruments and 1-megapixel arsenic-doped silicon (Si:As) detectors in the mid-infrared instrument. In both cases, new generations of detectors with superior noise properties and larger formats make it possible to detect faint light with greater efficiency.[6]

While passive cooling yields sufficiently low temperatures to limit thermal emission by the telescope optics and enable near-infrared detector operation, further active cooling by an advanced mechanical cryocooler is needed to cool the detector arrays in the mid-infrared (5–28 mm) instrument to an operating temperature of 6.7 K (−266°C).

Building and Testing: Challenges and Design Trade-Offs

JWST is an extraordinarily complex observatory with many unique parts and software developed by different teams in several countries. This assemblage of parts and systems has to work seamlessly in space for a decade or more, which necessitated an extensive testing program involving three cryovacuum campaigns for the instruments, a complex end-to-end optical test, and numerous integrated (spacecraft-sunshield-telescope) environmental test campaigns. Not surprisingly, with its stringent system requirements and constraints, building and testing JWST was an extensive series of engineering challenges.

Leading up to testing, trade studies to address a challenge or requirement on one subsystem had to consider the system performance ramifications for others. For example, the decision to choose a long-life ­mechanical cryocooler over a limited-life solid hydrogen sublimation dewar saved mass, which was a critical system resource early in the program. However, it introduced additional vibration that consumed image motion budget and added to spacecraft power consumption while trading away some complexities (e.g., loading and ­handling of dangerous solid hydrogen) for others (e.g., thermal rejection, vibration, and thermal isolation).

From early technology considerations through final testing, the science-driven requirements and system constraints required constant iterative attention. In the end, all system performance requirements were met with adequate margins, but it took several iterations of trade studies and design improvements to achieve this success.

An important challenge early in development was meeting the mass and volume constraints imposed on the observatory’s wavefront sensing and control system, which is needed to ensure that a single, focused image is formed when the light beams from the 18 mirror segments that constitute the primary mirror are combined coherently. NASA considered a dedicated wavefront sensor as used at the Keck Observatory in Hawaii, but the tight mass and volume requirements precluded its use.

The solution was the addition of optical elements to JWST’s near-infrared camera (NIRCam) and the development of wavefront sensing and control algorithms that enabled this architecture. At the critical final stage of this process, the algorithms are an adaptation of those developed by NASA in the early 1990s to determine the prescription of Hubble’s primary mirror aberration, modified to work with a segmented telescope using information from images obtained with NIRCam. The efficient approach to wavefront sensing and control incorporating the Hubble-derived algorithms is not only a lasting legacy of JWST but also likely to be adapted and improved to help future segmented space telescopes address new science questions.

A primary lesson learned from the Hubble spherical aberration experience was to make sure the optical system was vetted using both independent crosschecks and end-to-end testing (Feinberg and Geithner 2008). End-to-end testing of the telescope with the instruments was one of the most significant challenges in the testing program.

End-to-end testing of the telescope with the instruments was one of the most significant challenges in the testing program.

The need for a vacuum chamber capable of cooling the entire telescope to 40 K while performing an end-to-end optical test spawned numerous architectural and optical innovations. These included a new test device that used multiwavelength interferometry to measure ultraprecise relative positions of the telescope primary mirror segments and illumination source, and development of a new helium shroud in the NASA Johnson Space Center Chamber A. This shroud had to be large enough to cool and support the optical test equipment and telescope.

Basing the test architecture on previous programs would have required a large and massive stainless steel tower in the chamber between the telescope and complicated optical test equipment—a truss structure that would take weeks to cool down. Instead of being hung pointing downward as originally planned to ease particulate contamination fallout concerns, a new test architecture was adopted that positioned the telescope “cup up” so it could be rolled into or out of the chamber on tracks, thereby eliminating the need for the internal tower and its accompanying complexities. Once in place “cup up,” the telescope was hung from rods that attached above the chamber (figure 4).

Sembach and Feinberg fig 4.gifPreparing and executing the thermal vacuum test was a monumental effort to ensure that every possible contingency had been considered. Extensive planning and infrastructure improvements were essential. The team treated the test similar to a Hubble servicing mission and developed contingency flow charts for everything from the optical support equipment to the chamber systems. The addition of a large cleanroom at the entrance to Chamber A, extensive cleaning and refurbishment of existing vacuum chamber systems, and system modifications and additions added redundancy and robustness.

All of this preparation proved its worth and allowed testing to proceed when Hurricane Harvey settled on Houston for 5 days in 2017, just as the telescope reached its cold temperature 30 days into the 100-day test. Incredibly, some of the most critical tests of the primary mirror and the closed-loop guiding system occurred just as the hurricane bore down.

As development proceeded, it was necessary to ensure an adequate thermal margin to cool the detectors. This challenge took many forms, from modeling heat transfer to designing efficient radiators. Early on, detailed studies showed that the cryogenic budget for the cold region of the telescope, measured in milliwatts, did not have sufficient margin. A deployable radiator adopted for the entire instrument suite improved passive cooling and provides a better view of cold space (Yang et al. 2018). This resulted in adequate thermal margin but came at the cost of an additional deployment.

Another notable challenge was assembling, testing, and stowing the large (22 m × 10 m) sunshield, whose five layers each have to be carefully controlled during testing, launch, and deployment. The effects of gravity, criticality of single point failures, and overall size created unique challenges.

The sunshield design and fabrication were driven by temperature and straylight requirements that dictated its overall shape, deployment choreography, and tensioning requirements. All of these complexities flowed directly from the infrared nature and scientific needs of the observatory.

Sembach and Feinberg fig 5.gif
FIGURE 5 Fully integrated JWST in cleanroom at Northrop Grumman Space Park with deployed sunshield, secondary mirror support structure folded back against the primary mirror in its position stowed for launch, and six primary mirror segments (three on each side) also folded back in their launch positions. Photo credit: NASA/Chris Gunn.

Future systems, like a warm telescope optimized for detecting extrasolar planet biosignatures, will require a sunshade that blocks the Sun but will not face many of the critical constraints needed to maintain tight thermal control at the cryogenic temperatures necessary for JWST. Nonetheless, the deployment and testing methods developed for JWST have set a standard for large deployed membrane systems of this type (figure 5).

Science and the Next Big Question

Nearing the end of its development cycle, JWST is poised to revolutionize all areas of astronomical study, from faint galaxies whose light beams began their journey to Earth more than 13.5 billion years ago to giant planets and icy bodies on the fringes of the solar system (Kalirai 2018). The science community has brought forward its best ideas and most pressing questions, and the science program for the first year of observations is full of innovative uses for the observatory (O’Callaghan 2021).

This isn’t to say that science has stood still in the interim. If anything, the science potential for JWST is greater now than ever. Engineers and scientists worked together throughout JWST’s development to respond to a changing astronomical landscape while retaining the core requirements and capabilities of a telescope originally designed to explore the distant universe.

For example, upgraded instrument capabilities and operational enhancements to enable extra­solar planetary atmosphere studies were introduced well into the development cycle.[7] These late-stage additions to the ­technology-science trade space increased science potential, but have their limitations. JWST will take an unprecedented step forward in the characterization of extra­solar planetary atmospheres for Neptune and Jupiter-sized planets around stars cooler than the Sun, but is unlikely to characterize many Earth-sized planets around Sun-like stars in orbits conducive to the retention of liquid water.

Although not yet launched or specifically designed to search for evidence of life, numerous technologies developed for JWST and lessons learned about building and operating large telescopes already hold significant promise for development of a future life-finding space telescope (Feinberg et al. 2018). An ultrastable 10–15 meter segmented-mirror telescope operating at room temperature and equipped with high-performance instrumentation would be challenging to engineer and build (NASA 2019), but it is now within the realm of possibility—and the possible questions to be addressed are profound.

JWST science discoveries will help shape these and other questions, and a new engineering-science cycle aimed specifically at addressing whether there are ­habitable worlds elsewhere seems to be a natural step in the progression of space telescope investigations of the universe.


Dressler A, Brown RA, Davidsen AF, Ellis RS, Freedman WL, Green RF, Hauser MG, Kirshner RP, Kulkarni S, Lilly SJ, and 8 others. 1996. Exploration and the Search for Origins: A Vision for Ultraviolet-Optical-Infrared Space Astronomy. Report of the “HST & Beyond” Committee. Washington: Association of Universities for Research in Astronomy.

Feinberg L, Geithner P. 2008. Applying HST lessons learned to JWST. Society of Photo-Optical Instrumentation ­Engineers Conf Series 7010:0.

Feinberg L, Arenberg J, Yanatsis D, Lightsey P. 2018. Breaking the cost curve: Applying lessons learned from the James Webb Space Telescope development. Society of Photo-Optical Instrumentation Engineers Conf Series 10698:23.

Kalirai J. 2018. Scientific discovery with the James Webb Space Telescope. Contemporary Physics 59(3):251–90.

NASA [National Aeronautics and Space Administration]. 2019. LUVOIR: Large UV/Optical/Infrared Surveyor – Final Report. Washington.

NRC [National Research Council]. 2001. Astronomy and Astrophysics in the New Millennium. Washington: National Academy Press.

O’Callaghan J. 2021. The James Webb Space Telescope’s first year of extraordinary science has been revealed. Scientific American, Apr 7.

Stockman P. 1997. Next Generation Space Telescope: Visit­ing a Time When Galaxies Were Young. Washington: Association of Universities for Research in Astronomy.

Yang K, Glazer SD, Thomson SR, Feinberg LD, Burt W, Comber BJ, Ousley W, Franck R. 2018. Thermal model performance for the James Webb Space Telescope OTIS cryo-vacuum test. 48th Internatl Conf on Environmental Systems, Jul 8–12, Albuquerque (ICES-2018-35).


[1]  Hubble’s vision is limited to ultraviolet, optical, and near-infrared light in the wavelength range of 0.1–1.8 mm.

[2]  The redshift of light is given by z = (lobs − lrest) / lrest, where lobs is the observed wavelength of light and lrest is the wavelength of light emitted or absorbed by the redshifted source. Redshift and distance are related through Hubble’s Law, vr = HoD, which links the observed recession velocity (vr) of a galaxy to its distance (D), with Ho being Hubble’s constant. The relationship between redshift and recession velocity is z = square root [(1 + vr/c)/(1 − vr/c)], where c is the speed of light in vacuum.

[3]  “NASA Announces Contract for Next-Generation Space Telescope Named after Space Pioneer,” Sep 11, 2002 ( 2002-20.html?Year=2002)

[4]  Hubble can detect objects approximately 10 billion times ­fainter than observable with the human eye. The first galaxies to form are probably >10 times fainter still.

[5]  JWST’s launch mass is ~6500 kg, compared to Hubble’s launch mass of ~11,000 kg.

[6]  See html

[7]  Only a handful of extrasolar planets were confirmed at the time of the Dressler report in 1996.

About the Author:Kenneth Sembach is an astronomer and director of the Space Telescope Science Institute in Baltimore. Lee Feinberg is the senior large optical systems engineer in the Instrument Systems and Technology Division at the Goddard Space Flight Center in Greenbelt, MD.