During the alignment of JWST primary mirrors, this selfie was taken:

JWST selfie

(source: NASA)

I think we're seeing here the primary mirrors, with the secondary in the center and the reflection of the support struts in the primary mirror, i.e. we're looking in the reverse direction of the normal light path.

The NASA blog post (and also one from ESA) mentions:

This “selfie” was created using a specialized pupil imaging lens inside of the NIRCam instrument that was designed to take images of the primary mirror segments instead of images of space. This configuration is not used during scientific operations and is used strictly for engineering and alignment purposes.

How did this work? What was special about the lens in NIRCam that allowed this picture to be taken? Was it a dedicated lens or some special configuration of the standard components?

  • 2
    $\begingroup$ Any telescope can take a 'mirror selfie', you just have to defocus it far enough. $\endgroup$ Jul 17, 2022 at 6:08
  • $\begingroup$ the secondary mirror is fundamentally...... a mirror $\endgroup$ Jul 17, 2022 at 11:52

1 Answer 1


Point source viewed out of focus

when you are out of focus the image of a collimated beam of light (e.g. a star) will have the same pattern as the imaging system's aperture

All you need to do to make a pretty good image of your aperture is to:

  1. Find a collimated (parallel) bundle of rays, e.g. a point source far away, e.g. a relatively bright star on a dark background, and
  2. Move your photographic plate or CCD sensor substantially out of focus, either by moving it far forward or backward, or inserting another optical element that moves the focal point forward or backward instead. I think in the case of the JWST selfie it's the latter.

See also Astronomy SE's:

What's different about the JWST is that each element of the mirror is independently articulated. For the image in the question one mirror is pointed directly at the bright star and the others are not.

I can guess that the bright lines at the edges of some of the dark mirror segments might be edge diffraction but I don't know. I think that would be an excellent follow-up question!

unknown object

Above: from Astronomy SE's Help identify a bright and round object photographed through a telescope below: from Astronomy SE's What is that donut-shaped object I see in my telescope?

enter image description here

enter image description here

above: From Astronomy SE's Jupiter with a mobile phone and Celestron Astro FI 102mm Maksutov below: from Astronomy SE's What cluster of stars is this with a "dark donut" to one side?


It even works in reverse; an occlusion like a bit of dust on a surface near but separated from the focal plane in an area with extended illumination will cast a shadow in the shape of the aperture as well.

  • $\begingroup$ Collimated to avoid distortion? I.e. the star is used to illuminate the telescope mirror? And would that imply that they had to turn (at least) one segment towards a star for this to work? Turns out I know so little about optics... $\endgroup$
    – Ludo
    Jul 18, 2022 at 11:25
  • 2
    $\begingroup$ @Ludo Actually it doesn't have to be collimated; the light could be diverging or converging as long as it was from or towards a point. In other words a point source of light could be at finite distance, as long as it didn't have significant size. That would smear the image out. As for the star, there's gotta be light in order to see something right? If they were all pointed at absolutely black areas of the sky (in the wavelength range being observed) then they'd see nothing. I'm not sure if I've actually answered your question thought. $\endgroup$
    – uhoh
    Jul 18, 2022 at 12:09
  • 2
    $\begingroup$ you did actually, thanks! Before it just didn't connect, the images helped. $\endgroup$
    – Ludo
    Jul 18, 2022 at 20:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.