STScI Logo

James Webb Space Telescope
Wavefront Sensing and Control

The optical telescope element (OTE) of the James Webb Space Telescope deploys some four days after launch. Then a 1-2 month commissioning process to optimize the telescope image quality begins. The goal of this process is to align the primary segments into a co-phased state, and to align the primary and secondary mirrors with the aft optics system, which is fixed and houses the tertiary mirror, the fine-steering mirror, and a central baffle.

To align the telescope, each primary mirror segment is mounted on three bipods (a hexapod) to form a kinematic attachment to the backplane. Six actuators with both coarse and fine-positioning capability adjust the lengths of the hexapod’s legs. This arrangement permits six degrees of freedom to be adjusted independently: x- and y- position, piston, tip, tilt, and clocking. The secondary mirror also has six actuators, permitting the same kind of control. For each of the primary mirror segments, an additional (seventh) actuator controls the radius of curvature.


Deployment sequence for the primary and secondary mirrors of the Webb telescope.
The aft optics system is housed in the primary mirror gap, and includes the central baffle.
Following deployment, a sequence of wavefront sensing and control activities
bring the telescope into focus.

In total, there are 132 controllable degrees of freedom in aligning the OTE. The process by which these degrees are used to optimize the Webb image quality is called wavefront sensing and control (WFS&C). Unlike the case for ground-based adaptive optics systems, WFS&C is not a real-time process that runs autonomously on the telescope. For ground-based observations, the short characteristic timescales of atmospheric turbulence dictate rapid sensing and correction. By contrast, Webb operates at the second Sun-Earth Lagrange point, L2, in an environment with no atmosphere and little gravity. The relevant timescales for WFS&C are dictated by slow thermal drifts, which occur on day-to-week timescales. Therefore, the cheapest and most reliable approach is to obtain data through the regular, pre-planned observing methods, downlink of the data, analysis on the ground, and uplink to the telescope of any commands to move actuators.

Immediately after OTE deployment, the telescope point-spread function (PSF) consists of 18 separate, out-of-focus images, one from each primary mirror segment. The first OTE commissioning steps use a bright star to determine and calibrate the telescope line of sight, and to locate and identify the image of each individual segment. The mirror segments can then be tilted to align the 18 separate images into a single image. The wavefront superposition thus obtained is incoherent. As a result, the PSF at this stage has a width typical of a single (1.32 m flat-to-flat) mirror segment.

To obtain the 6.5m diffraction-limited performance, it is necessary to co-phase the segments. This involves two main steps, each of which uses dedicated hardware within the Near-Infrared Camera (NIRCam). These steps are called coarse phasing and fine phasing.

Coarse phasing uses dispersed Hartmann sensing. This process uses grisms to create wavelength-dependent interferograms based on the light reflected by adjacent mirror segments. This allows identification of large piston errors between segments, which would then be corrected.

Fine phasing is done using a set of weak lenses, which can be chosen to yield images with five defocus settings: –4, –8, 4, 8 and 12 waves at 2.12 μm. Images thus obtained can be analyzed with focus-diverse phase retrieval algorithms, which yield a map of the optical path difference (OPD) over the telescope pupil. The specific algorithm used for Webb is the “hybrid diversity algorithm” (HDA), which is a variation of the well-known Gerchberg-Saxton scheme, with an additional outer loop in the process to provide phase-unwrapping and robustness even with large phase errors.

The phase-retrieval process requires amplitude information for the pupil, which is also obtained using dedicated NIRCam hardware. A Pupil Imaging Lens (PIL) provides the required imaging. The resulting data can also reveal insight into vignetting or other alignment problems.

Once an OPD map has been determined from NIRCam imagery, its low-spatial frequency content is corrected by moving actuators appropriately to reposition the segments. Mid- and high-spatial-frequency figure errors, which may be caused by manufacturing errors within a segment, cannot be corrected by this process. Nevertheless, these errors are bounded by robust error budgets in the manufacturing process, which ensures diffraction-limit performance. Furthermore, these uncorrectable errors are stable and can be accurately characterized during telescope integration and testing (I&T). This enables accurate prediction and characterization of the post-commissioning PSF.

Most WFS&C steps are performed at a single field point in NIRCam. During the initial alignment, however, multi-field, multi-instrument wavefront sensing is also performed in order to remove degeneracies ambiguities that might arise when using NIRCam alone. In principle, the WFS&C steps can be repeated as necessary during the Webb mission.

After OTE commissioning has been completed, a maintenance process of WFS&C executes for the duration of the mission. In its cooled, equilibrium state, the primary mirror experiences a large temperature gradient (some 20–30 degrees over the 6.5 m aperture) in the direction away from the sunshield. This gradient is stable, and any distortion that it induces can be corrected during OTE commissioning. However, as the telescope is operated and pointed at different positions within the field of regard, small temperature changes can occur at the level of tenths of a degree. These changes induce changes in the mirror figures, which must be controlled to maintain image quality.

During regular operations, NIRCam observations for WFS&C are executed every two days to monitor the Webb image quality. The weak lenses are used to image a bright target star using several defocus settings. The implied wavefront errors are corrected if they are deemed unacceptable, which is not expected to occur more frequently than every two weeks.

The software modules to analyze WFS data and control the mirror actuators are being developed by Ball Aerospace. The Institute is incorporating these modules into the WFS&C executive software, which handles the interfaces with other aspects of the Webb Science and Operations Center (such as the archive, proposal planning, and flight operations). The entire WFS&C software subsystem (WSS) has passed its critical design audit.

The functionality of the WFS&C algorithms has been verified on the test bed telescope (TBT), a 1/6 scale model of the Webb OTE, at Ball Aerospace. Additional verification and validation activities continue in the coming years. The NIRCam WFS&C hardware has recently been tested with good results in a cryogenic and vacuum environment (cryo-vac), using the NIRCam engineering test unit. Later tests include cryo-vac testing at Johnson Space Center of the integrated OTE and ISIM.

OPD maps inferred from WFS data will be delivered to the Webb archive and available to astronomers. Software that calculates PSFs as a function of instrumental configuration, field position, and time by combining OPD data with optics models and I&T data is under development at the Institute. This will be akin to Hubble’s TinyTim software, but will deal with optical details specific to Webb such as the time evolution of the primary mirror. A limited amount of this functionality is already available in the WebbPSF program, which produces calculations based on realizations of the wavefront error budget, rather than hardware characterizations. WebbPSF will improve in fidelity as I&T data become available. However, it should already be useful for prospective observers to assess the expected image quality as function of instrument, detector, wavelength, and filter.