JWST wavelength sensing starts at 600nm, corresponding to red/orange visible light, and then into the IR. It seems there is no sensor or filter included below 600nm meaning that green and blue light cannot be imaged at all, meaning all images must have to be displayed as false color composites. The mirror surface is gold which reflects well (>90%) above 600nm and falls off to around 35% - 40% at 400nm. This seems to be the justification for not including any blue or green sensing, even though gold still reflects 35% of 450nm blue light and 70% of 540nm green light (ref https://www.inradoptics.com/capabilities/coating/spectral-reflectivity-curves).

So a blue and green filter could easily have been fitted - their image gains increased electronically to compensate for the gold spectral response (the lower SNR would not be a problem for the very bright outer solar system planets). This means, very disappointingly, that when the outer solar system planets are imaged, we will not get real color images but false color instead. Very disappointing and a major oversight in my view. When the next event happens on Jupiter or Saturn we cannot get real color images of it from JWST, for the sake of two cheap and lightweight dichroic filters.

  • 37
    $\begingroup$ It might be disappointing to you, but the purpose of JWST isn't really to appease normies like us I'm afraid. And to suggest it's an oversight is a bit weird - don't you maybe think the team designing JWST thought of this and had a very good reason not to do it? I personally wouldn't assume I had thought of something which eluded NASA :) $\endgroup$
    – user438383
    Commented Jul 15, 2022 at 11:52
  • 2
    $\begingroup$ @user438383 it's not the purpose, no, but it's also not completely extraneous for a project that ate up as much taxpayers' money as JWST to produce results that are viscerally impressive, as well as scientifically valuable. Fortunately, the false-colour images seem to do that job just fine. $\endgroup$ Commented Jul 15, 2022 at 13:32
  • 25
    $\begingroup$ There is no such thing as "true color". That's just a narrow range of electromagnetic wavelengths the human eye is able to detect and interpret. The interesting stuff for scientists isn't in that range. The JWST is specialized to explore the longer wavelengths that thus far we have not been able to, because that's where we're going to learn new things. We already have telescopes that can see Jupiter and Saturn just fine. $\endgroup$
    – Seth R
    Commented Jul 15, 2022 at 15:28
  • 5
    $\begingroup$ One of the reasons we don't need JWST to look at visible light is that the new generation of ground-based telescopes with adaptive optics can do just as well. We needed Webb in space because thermal noise in the aatmosphere would not allow looking at high redshift galaxies from Earth, and spectrographic analysis of elements would also be fery hard. But for visible light, the largest earth-bound observatories are very good and already can exceed Hubble resolution. The next generation will be even better. $\endgroup$
    – Dan Hanson
    Commented Jul 16, 2022 at 19:41
  • 4
    $\begingroup$ Isn't the reason they constructed JWST to see things we couldn't see very well with visible light. In that case if they viewed a very distant or faint object on infrared, they might not even be able to pick up the object on the visible spectrum. In those cases, having the visible light sensors wouldn't add anything. $\endgroup$
    – user4574
    Commented Jul 17, 2022 at 15:53

8 Answers 8


JWST's optical path includes more than just the large primary mirror. In the OTE (the telescope proper, rather than the instrumentation), there are three additional mirrors:

  • A secondary mirror, located some distance forward of the concave primary mirror.
  • A tertiary mirror, located aft.
  • A flat fine steering mirror, also located aft.

All these mirrors are coated with gold (link.)

JWST mirrors

The instruments add still more elements to the path. For example, in each of the two NIRCam modules, for the short-wavelength sensors, the path has ("NIRCam optical analysis", Mao, Huff and Granger, Proc. SPIE, 5904 (Aug. 2005)):

  • A pick-off mirror.
  • A coronagraphic mask.
  • The first fold mirror.
  • Three lenses, the first made out of zinc selenide, the second barium fluoride, and the third lithium fluoride.
  • A wavelength-sensitive dichroic mirror to split shortwave from longwave light.
  • A rotatable pupil wheel.
  • A rotatable filter wheel.
  • Another group of three lenses (again LiF, BaF2 and ZnSe.)
  • A flat fold mirror.
  • And finally another mirror to reflect the light into the detector array.

So, although you could place a short-wavelength (≤ 500 nm) filter in the filter wheel, you would not end up with anywhere near 40% transmission at these wavelengths. Considering only the first four mirrors, you would end up with at best $(40\%)^4 = 2.56\%$ of the light. (The other elements may also make things worse.) This pushes the tradeoff against including such a filter.


It all comes down to cost vs benefit

If you shoot something into space every gram of mass and every little bit of space counts. Also every bit of energy an instrument uses and heat it produces must be carefully considered.

So the scientists and engineers sit together and fight over every bit like:"what does it cost (in terms of weight, space and energy) and what's the scientific value". In the end, there is a list, sorted by cost/benefit and at some point there has to be a cut made (sure, a lot more factors will play into it like:"can it be integrated without interfering with other parts of the spacecraft" or "is there another telescope that can fulfill this role better")

So, they just came to the conclusion that the benefit of integrating sensors for shorter wavelengths was not worth the cost. Maybe, because there are already other instruments that can fulfill that task (visible light astronomy can be done with Earth based observatories while the infrared spectrum is blocked by the atmosphere) or other instruments, that will do this task better are already planned.

Also, telescopes are not created for the pretty pictures but for the scientific value. So "true color" is no concern at all for most scientists. And what's "true color" anyway if you're looking at heavily red-shifted objects? Is "true color" how they would look to an observer that is stationary relative to the object? Then infrared is the way to go!

  • 4
    $\begingroup$ @RogerWood ythe better question is: why should ther be? As already said: it's an infrared telescope. There are telescopes better suited for visible light... $\endgroup$
    – TrySCE2AUX
    Commented Jul 15, 2022 at 8:51
  • 5
    $\begingroup$ Adding to @TrySCE2AUX's point, why should there be visible frequency sensors on a telescope that by design is highly optimized for collecting infrared, and on a telescope that was already years late in delivery and billions of dollars over budget by the time it did finally launch? $\endgroup$ Commented Jul 15, 2022 at 12:41
  • 6
    $\begingroup$ @RogerWood A "no filter" option would result in a grayscale image. This would not satisfy the OP's desire for "true color". While Ansel Adam's black and white (another name for grayscale) photos of the American West are absolutely beautiful, they are not color photos. $\endgroup$ Commented Jul 15, 2022 at 12:46
  • 2
    $\begingroup$ No filter may also mean having to design the detector to handle more photons before saturating the detector. That may then mean the camera is less sensitive to faint objects. Again trade offs had to be made to optimizes for the different sciences cases. $\endgroup$
    – Rob
    Commented Jul 15, 2022 at 15:13
  • 4
    $\begingroup$ @RogerWood NIRcam is actually less sensitive at blue wavelengths than red; see NIRcam detector performance which also links to the filter set. It's not uncommon for astronomical instruments to not support no filter since it often requires a refocus of the telescope as there is now no longer the several mm of filter in the beam. $\endgroup$ Commented Jul 15, 2022 at 16:48

The mirror is made out of gold. Gold is an amazing reflector of anything past about 600 nm, some between 500 and 600 nm, but very little smaller than that. It works perfectly for the infrared nature of the telescope. Having a camera with wavelengths less than 600 nm doesn't make sense because of the gold mirror, and the gold mirror is optimized for infrared.

JWST is designed to see through dust clouds and back to the early days of the universe, both of which work better in infrared.

Bottom line, it wouldn't work well with the mirror, and would detract from the main mission.

  • 5
    $\begingroup$ But as the question already remarked, gold does also reflect green and blue light, and not even that badly. JWST would still outperform Hubble in the visible range if it had suitable instruments: the larger aperture would easily make up for the gain loss due to the gold surface. It's just that the visible range is covered pretty well already by both Hubble and ground-based telescopes, whereas in the infrared JWST outperforms everything else by a huge margin, so it would pretty much a waste of observation time to use it for visible-light tasks. $\endgroup$ Commented Jul 15, 2022 at 13:42
  • 4
    $\begingroup$ It's more of a pain to make an imaging sensor when there is a non-uniform gain. It's doable, sure, but would involve a lot more careful calibration, and it wasn't really the main point. $\endgroup$
    – PearsonArtPhoto
    Commented Jul 15, 2022 at 14:15
  • $\begingroup$ There's always non-uniform gain which needs to be accounted for. With aluminium you need to account for the dip to 85% at 800 nm, with gold for the dips to 40% at 350 and 450 nm. Really, in both cases for the whole curve needs to be calibrated for. There's no difference at all in how carefully this needs to be done, the only difference is what factors you multiply with. And again, sure, of course gold is less favourable than aluminium for visible light and you wouldn't use it for a visible light telescope – but it would totally still work ok if you had to use it. $\endgroup$ Commented Jul 15, 2022 at 15:57
  • 7
    $\begingroup$ "... if you had to use it." - the entire point being, JWST does not need to, because that's not within mission scope. It's up there to see what can't be done from Earth, so anything can be done on Earth, isn't necessary. $\endgroup$
    – Nij
    Commented Jul 15, 2022 at 22:58

It's not just a filter

The mission of the JWST is to image heavily redshifted light from objects that are moving quickly away from the telescope. Far away means long in the past, in this case, so we can study the early universe. The entire telescope design is tightly centered around that goal, visible most prominently in the sunshade preventing IR emissions from Sol from washing out the signal received on the primary mirror. There are not a bunch of add-ons to look at "other cool stuff."

There is no non-navigational sensor on board that can see blue and green light. "True color" imaging capability would mean switching out one of the HgCdTe detectors (which are only sensitive to light in the IR spectrum) for a different detector that is sensitive to blue and green light. This isn't a matter of just needing a different filter to swap in, you're talking about an entire additional "camera." Because of the architecture of the system, that would be one fewer camera dedicated to the actual science mission.

Fundamentally, the JWST is not just a "bigger and better" Hubble, though that's often how it's portrayed in the media. It's dedicated to an entirely different set of missions, and the capability "left out" is orthogonal to the mission. It won't be looking at Jupiter or Saturn, but at faraway galaxies. There were a lot of other things also left out, science instruments that many astronomers want, as there is finite and very precious SWAP (Space, Weight, And Power) budget available on the telescope.

For what it's worth, I'd like a bigger and more modern telescope with different science instruments on board, but that telescope is not this telescope. Luckily, there are several other instruments in various stages of planning and development. This telescope cost everyone in the US and EU an average of ~$15 ea, and took nearly two decades to design, build, test, and launch.

You can read about the HgCdTe sensors here: https://webb.nasa.gov/content/about/innovations/infrared.html

Source for the rest: I helped build the telescope.

  • 1
    $\begingroup$ Great to have someone with direct knowledge. Welcome to the forum! Table 1 here jwst-docs.stsci.edu/jwst-near-infrared-camera/… says the range for the actual sensor is 0.4-2.5 um with a footnote saying it's limited by filters and dichroics. $\endgroup$
    – Roger Wood
    Commented Jul 16, 2022 at 4:17
  • $\begingroup$ is ea a typo for each, or does it mean something else? $\endgroup$ Commented Jul 16, 2022 at 16:36
  • 2
    $\begingroup$ @Richard, not a typo; a standard abbreviation. $\endgroup$
    – prl
    Commented Jul 17, 2022 at 10:52

I see a number of good reasons in the other answers, but I want to add this information about the use of dual detectors and the dichroic beam-splitter. NIRCAM has two separate electronic detectors that can be used simultaneously. Preceding these detectors is a beam-splitter (dichroic mirror) that directs short-wavelengths to one detector and long-wavelengths to the other detector. There is a relatively sharp crossover between the two regimes and a wide span of wavelengths that are covered on either side. In this way the images include almost all the received photons automatically separated into two 'colors'. Further refinements in wavelength (color) can be done by inserting various filters but at the expense of longer exposures.

This shows the response of the dichroic beam-splitter:

JWST NIRCAM dichroic mirror

The response below 0.6 microns is quite messy and the response also rolls off above 5 microns. There is a big advantage in using this dual-sensor dichroic beam-splitter architecture, but I'm guessing it is also a primary factor limiting the short-wavelength (visible) response. I imagine there is a trade-off between achieving very wide 'bandwidth' on the beam-splitter and maintaining its efficiency and flatness over the bulk of the spectrum.


The main reason is that, due to the nature of the mission, they are not needed (in most cases are actually useless), and adding them will just add unnecessary overhead to the mission in all aspects and shorten its lifetime. JWST will be mainly used to image and study the galaxies further away from us, study the composition of the atmospheres of exoplanets, and study star and planetary formation processes and regions (plus anything I might have forgotten, and other crazy application they might come up with later). This poses some particular challenges.

One is that the further away a galaxy is from us the more redshifted its light is, to the point a lot of them might only be visible in the infrared.

Another issue is that a lot of the JWST targets will be blocked to some degree by dust clouds, which are opaque to the shorter wavelengths (UV, Visible), which means there is no way to see through them except for using IR and NIR. (Roughly speaking, the light is blocked by dust or particles when its wavelength is smaller or very similar in size to the dust/particles. Red and infrared wavelengths are several times larger, so for practical purposes, the dust becomes "transparent" for them.)

Simplifying the technical details a lot, that's why all the instruments on board the JWST are built to detect in the range of infrared and near-infrared. Adding the hardware necessary for the blue/green color or any other wavelength shorter than what the telescope can currently detect won't add much value to the mission (if any), but would make it more expensive, and add additional mass that would need to be removed from somewhere else, in this case probably less fuel, which would reduce the amount of time the orbit and attitude can be maintained and shorten the mission.


What’s disappointing, and what’s a major oversight?

Signal-to-noise ratio (SNR): the ratio of desired information (signal) to undesired (background) detections

As part of scientific observation, targets must be discerned from background. This includes:

-faint objects imaged against space that isn’t literally, completely black

-pixels cleanly registered, against spillover light from other pixels

-pixel brightness levels cleanly registered versus zero or saturation or whatever

-spectral features detected against other spectral features

-and I’m sure other forms of measurement I can’t articulate off the top of the head.

These measurements are done by excluding stray light (by various definitions depending on the situation), not being a garbage disposal for photons. One of the more elementary means of stray light exclusion is band limits: defining long (more red) and short (more blue) cutoffs for the light path. Each undesired photon excluded will reduce the risk of light contamination and noise contributions, if nothing else by reducing absorption and its resulting heat. Shorter wavelengths are better at generating waste heat, per Planck’s Law. If those wavelengths were never meant to be measurements in the first place, then why should they be there? There are other materials in the light paths, besides the primary and secondary mirror coatings. The aft optics and instrument paths would boggle you.

These measurements are done to reach defined, targeted program objectives. Principal investigators are under extreme competition to get access to telescope time in general, and JW time certainly. Each and every principal investigator is then forced to submit program proposals that show defined objectives and pathways, for answering the question under target. One metric for most (if not all) proposals is showing the signal-to-noise ratio is past the threshold implied to do the measurement and answer the given question. If the SNR is low, observing time must be increased, to capture more signal photons. If the SNR is palpably low, the proposed observation is a waste of the telescope’s time.

These measurements are the ACTUAL goal of JWST, or any other research-grade, funding-limited, time-oversubscribed instrument. Disappointing or not disappointing you is not an observing objective, much less a mission criterion. Passing your oversight is not a defined metric. We gather signal photons from astronomical objects. We exclude noise from undesired sources.


The biggest reason is that we don't need a really good space based visible telescope. There are a number of much bigger visible light telescopes (on earth) that are much more effective at observing the visible spectrum (https://en.wikipedia.org/wiki/Extremely_large_telescope). These telescopes have a much bigger aperture, have more flexible optics, and have costs in the range of $100M to $1B instead of $10B. Adding weight and complexity to the JWST to make a far worse visible wavelength telescope just isn't interesting from a scientific perspective. You only bother putting a telescope in space if it gives you new capability. JWST is incomparably better than any ground based IR telescope, but would be a mediocre to bad visible light telescope.


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.