How big of a space telescope would we need to see this planet in any actual detail?
One of my sci-fi fantasies is to take a photo of an extrasolar planet and see someone else's city lights. :) Of course if we could see that we could also probably detect their radio emissions, but seeing someone else's lights would somehow be cooler.
The most promising technique at this point is to use optical interferometry to resolve the surface of the planet. In this case, you will need two telescopes (likely in space) separated by a baseline distance, "d". The two telescopes will require extremely precise synchronization in both spatial position and timing so that the light they collect will interfere at precisely the right phase, but if this can be achieved, the angular resolution of such a telescope in radians is wavelength/d. An earth sized planet has a diameter of ~13,000 km. To "resolve the planet", we would need to be able to distinguish one half of the planet from the other, meaning we need to see at a resolution of ~6000 km on the surface. Proxima Centauri is ~4 light years away, meaning that this requires an angular resolution of 1.5e-10 radians or 30 microarcseconds. That's over 30,000 times smaller than the angular size of a human hair held at arms length!) To achieve 30 microarcsecond resolution with our optical interferometer operating at 600 nm (visible light), we need the two telescopes to be 4 km apart, which practically doesn't sound unfeasible.
Alternatively, you could construct a 4 km telescope, but that's far bigger than any optical telescope we have now or in the near future (the biggest telescope in the next 20 years will be 40 meter in diameter).
NASA actually did do a feasibility study on the Terrestrial Planet Finder (https://en.wikipedia.org/wiki/Terrestrial_Planet_Finder) as a space-based telescope to find terrestial planets orbiting other stars, and one of the versions is an infrared interferometer flying in formation in space. Unfortunately, they never followed through with it (I'm guessing due to technical challenges).
The planet is about 0.05 AU from Proxima Centauri, meaning we need an angular resolution of about 1.9e-7 radians to even distinguish it from its host star. Is that realistic?
In theory, an orbiting space telescope has a diffraction-limited resolution of approximately 1.22λ/D (λ = wavelength, D = aperture size). Modern image processing techniques can improve on this somewhat, but it makes a good order-of-magnitude estimate. Anyway, this formula tells us a 4-meter telescope has a maximum angular resolution of about 1.8e-7 radians at a typical visible light wavelength of 600 nm. That would be just good enough... except that we don't actually have a 4 meter orbiting space telescope. Resolving even large features on the planet would require a much larger telescope, probably kilometers or more.
For ground based telescopes, the situation is even worse because of atmospheric effects. Despite being 10 meters in aperture, the Keck telescopes in Hawaii are limited to an angular resolution of about 2e-7 radians because of the atmosphere. However, there is reason to hope that the even larger European Extremely Large Telescope will have enough resolution (about 5e-8 radians? hard to tell from their official publications) to image Proxima b directly. https://www.eso.org/sci/meetings/2011/VLTI2011/presentations... Again, this is still not enough to resolve surface features.
So, long story short the answer is unfortunately no at present. Maybe space-based manufacturing will let us build a big enough telescope someday?
The planet is about 0.05 AU from Proxima Centauri, meaning
we need an angular resolution of about 1.9e-7 radians to
even distinguish it from its host star. Is that realistic?
Much more than that; that's the angle for HALF-maximum brightness, but since the star's many orders of magnitude brighter than the planet, you'd need a much larger reduction than 1/2. Unfortunately, the diffraction-limited pattern [0] has fat tails -- it's not Gaussian, the brightness is slow to drop off away from the center (polynomially slow? [1]). I understand you'd need >100 times the FWHM angle in practice, on the order of 1" for JWST for instance [2]
Technically, 1.22λ/D is the angle for the first dark circular ring of the Airy disc (first zero of the relevant first-order Bessel function [0]). But you are still right that the host star needs to be blocked out in some way to produce a useful image. I think NASA is working on some ways to do this, see [1].
Can we directly image the planet from earth?
1. "The planet/star contrast is 10^-7 " This basically means
for every 10,000,000 photons from the star, we would measure
~ one from the planet.
2. "Current instrumentation using adaptive optics and
coronography on 10 m class telescopes (like Sphere on VLT or
Gemini Planetary Imager) aims at achieving a contrast of
10^-6 to 10^-7 at an angular resolution of 100-200 mas"
3. "The planet has a separation of 38 mas".
4. Therefore with the best planet imagers we cannot
currently directly image the planet. Our best hope is the
E-ELT which should have first light in 2024.
To add on to this, adaptive optics systems can be used to correct for atmospheric turbulence. The latest adaptive optics instruments have been able to achieve diffraction-limited imaging (i.e. comparable to if the telescope was in space) with 8 meter telescopes in the near-infrared (1-2 microns), resulting in ~2e-7 radian resolution. Pushing this to larger telescopes and shorter wavelengths will improve this resolution.
The bigger problem is actually blocking out the glare of the host star. Especially for a planet this close in angular separation to its host star, blocking out the light from the host star is challenging and requires sophisticated instrumentation (e.g. coronagraphs). The problem with the more sophisticated instrumentation is that you also lose throughput (when trying to block out stellar light, you also end up blocking a lot of light from the planet), meaning we will need a lot of telescope time in addition to sophisticated instrumentation to eventually image this planet.
However, there is reason to hope that the even larger
European Extremely Large Telescope will have enough
resolution (about 5e-8 radians? hard to tell from their
official publications)
6-12 mas is the advertised figure (0.006" = 3e-8 rad). That's the FWHM for its adaptive-optics imaging camera [0]. If you look at the details [1], it achieves the best resolution (6 mas) in the near-infrared J band, and for Nyquist-sampling reasons the pixel scale is half that (3 mas).
To put things in perspective, what was the angular resolution of Hubble's Ultra Deep Field measurements/photos? (keeping in mind that angular resolution isn't the only challenge here)
Hubble's angular resolution is about 0.05 arcsec = 2.4e-7 radians. And you're absolutely right that angular resolution isn't the only challenge here - some way of filtering out the extremely bright (relatively) light from the host star is needed as well, for starters.
From the link 'evilduck posted, it seems that's big indeed, but possibly big.
240km is a big thing, but within the realm of possibility; also I'm personally hoping someone will invent some magic mathematical trick around phased arrays or whatever, and cut that size by an order of magnitude or two :).
240km is also what's required to resolve a feature no smaller than 100km across. I'm no mathematician and I am not equipped at the moment to pretend to be one, but to resolve a feature an order of magnitude smaller, I suspect you'd need a much larger instrument.
If I did the math correctly, it turns out to be ~2400km for 10km across (changing 100 to 10 in the eq). This could very easily be wrong as I only glanced through the comment. Think of it as Kirkian mathematics.
It's not a coincidence that we can see in the 390 to 700 nm part of the spectrum. These are the wavelengths that pass through water most easily[1]. Any alien lifeforms which started their evolution in (illuminated) water could be expected to be sensitive to the same slice of the spectrum.
At ~500 light years away, even if we did see lights, who would we be communicating with _now_? And would it be disappointing if we saw a technologically-similar civilization that hadn't contacted us yet, being 500 years ahead of us?
Sometimes it's romantic to look up at stars that probably died a long time ago. Sometimes I hate that we can't see what's happening now.
Why would a civilization be exactly 500 years ahead of us? While evolution is a powerful trend, I think there's so much randomness about... when their planet cooled off enough (and had enough alkaline vents) to support abiogenesis, when bacteria form eukaryotes, when eukaryotes become multi-cellular, when life moves out of the oceans, when plant-like organisms start producing fruit leading to animals that take advantage of the high-energy food source and are able to develop smart brains---
all of these events could happen +/- a million years from eachother. And that's assuming that their evolution of life followed a similar path! So worrying about a race having 500 years on us is just... a worry based on so many wrong assumptions.
Why are people in this thread constantly citing 500 light-years? Proxima Centauri's 4 ly away, which, granted, isn't a walk in the park, but it's remarkably near to us.
One of my sci-fi fantasies is to take a photo of an extrasolar planet and see someone else's city lights. :) Of course if we could see that we could also probably detect their radio emissions, but seeing someone else's lights would somehow be cooler.