HARPS-S

Laser comb being tested on the "old" HARPS in Chile:

http://www.space.com/15927-earth-alien-planets-laser-comb.html
http://www.nature.com/nature/journal/v485/n7400/full/nature11092.html?WT.ec_id=NATURE-20120531

As HARPS-N is being commissioned on the northern hemisphere, we should call a future laser comb equipped HARPS as its southern counterpart - HARPS-S.
Let the exoplanet competition began..

Achieving a few cm/sec calibration repeatability for high resolution spectrographs: the laser frequency comb on HARPS

http://proceedings.spiedigitallibrary.org/proceeding.aspx?articleid=1362421

Quote :

The laser frequency comb, with its extreme precision, opens a new window for high precision spectroscopy for current facilities, as well as for the ELT's. We report on the latest performance of the laser frequency comb obtained in combination with the HARPS spectrograph, which allowed calibration with cm/sec repeatability. The laser frequency comb system developed is described. Details of its laboratory set-up, characterization and integration with HARPS are shown. The results of the recent test campaigns are presented, showing excellent performance in terms of repeatability as well as wavelength coverage. Preliminary on sky data and next activities to integrate such a system in HARPS are presented.

The article is beyond a pay-wall, but results from previous tests are available and were quite promising:

http://obswww.unige.ch/~wildif/publications/2010_7735-181.pdf

Quote :

[...] the Observatory of Geneva has designed, built and tested in collaboration with ESO a calibrator system based on a Fabry-Perot interferometer to explore its potential to improve the wavelength calibration of RV spectrographs. Unlike the Th-Ar lamps used today [3], this device allows the production of optimally and regularly spaced calibration lines covering all orders of the spectrograph. The stability of this system has already shown to be on par or better than the Th-Ar lamps.

http://www.eso.org/sci/libraries/SPIE2010/7735-29.pdf

Quote :

In summary, we have set up a frequency comb system with a mode separation and spectral bandwidth adequate for calibrating 2 echelle orders of the HARPS spectrograph. We could demonstrate a repeatability of 9 cm s-1, compatible with the computed photon noise and already better than a Th-Ar calibration, which needs to use all 72 orders. By using the LFC we could measure the contribution of the CCD pixel irregularities to the calibration curve for the first time. This leads to an unprecedented low scatter of the residuals. Applied to the measurement of absolute Th-Ar frequencies we could exceed the accuracy of the published values. After spectrally broadening our LFC in a highly nonlinear fiber we hope to cover the whole visible range in future. By using an adjustable spectral filter the comb modes will also have about equal intensity. This system will reduce the photon noise by more than an order of magnitude which is expected to enable Doppler shift measurements at the cm s-1 level.

Have to say I really like the frequency comb idea, particularly as it also has applications to infrared radial velocities which are an interesting area for several reasons.

If all goes well, the hunt of truly earth like planets is near to begin. It' s very exciting.

Is there an entry-level explanation of this technique? I'm intrigued.

ciceron wrote:

Is there an entry-level explanation of this technique? I'm intrigued.

Very good question. I still don't understand laser combs either.

ciceron wrote:

Is there an entry-level explanation of this technique? I'm intrigued.

I'm not an expert of the field, I'll try to explain what I think to know on the affair.

The RV technique measures the changes of radial velocity of the star over time (due to the tugs of orbiting exoplanets) by measuring the displacements of the spectral lines of the star caused by the Doppler effect; this is done through a wavelength reference, i.e. a stable "ruler" by which the star spectra is compared.
Modern spectrographs employ as rulers the spectra of Th-Ar lamps (Harps) or Iodine cells (Hires-PFS), which provides a certain number of weighing reference lines which statistically allows precisions up to 1 m/s. As a ruler, a laser comb provides a far bigger number of useful calibration lines, on a larger spectral coverage, using regularly spaced pulsed laser as a reference. The increased number of reference lines statistically allows the detection of far subtler RV displacements than previously possible.
In a few words, you have a finer and more precise ruler by which weight RV variations.

A more technical explanation could be found here:

http://tf.nist.gov/timefreq/general/pdf/2255.pdf

Besides increased accuracy, there are additional benefits: the reference lamps nowadays used as calibrators age with time, causing changes in their spectra (i.e. the ruler vary with time, with the need of corrections for long term stability) and since there are no identical calibration lamps, when merging observations of different instruments you could be misled by offsets and small undetected systematics due to different calibrations properties between the spectrographs. With a standard laser ruler calibrated with an atomic clock, these two problems disappear altogether.
In addition, as correctly pointed out by Lazarus, no reference lamps today allows high precision RV in infrared; with a laser comb, the reference lines could be shifted from optical to infrared with a cascade of commercially available electronic components. Since 80% of stars are M dwarfs whose emissions peaks in the infrared, this would permit to obtain high precision RV of stars being never or poorly studied before due to their faintness in the optical, including many nearby stars.

This is my roughly understanding of the affair, if someone can be more precise and/or provide corrections I would welcome that.

So , if i got this right , it's a better "test pattern" to calibrate the sensors , guaranteed to be identical in all the installations and with no distorsions due to material fatigue. That's it?

With a better test pattern , you can write a transfer function to allow for defects in the sensors and/or ambient effects.

ciceron wrote:

So , if i got this right , it's a better "test pattern" to calibrate the sensors , guaranteed to be identical in all the installations and with no distorsions due to material fatigue. That's it?

It's a more precise calibration system of stellar spectra, and more stable (atomic clocks has negligible errors over time). Once the setup of the laser combs in different spectrographs is known, I think that they can correlate RV measurements of separate facilities without the troubles arising today with spectrographs each operating with their own particular calibration lamp and correlated issues (aging, or material fatigue, as you said).

ciceron wrote:

With a better test pattern , you can write a transfer function to allow for defects in the sensors and/or ambient effects.

In my yesterday post I noted that previously undetected systematics were characterized thanks to the improved accuracy.

Quote :

In summary, we have set up a frequency comb system with a mode separation and spectral bandwidth adequate for calibrating 2 echelle orders of the HARPS spectrograph. We could demonstrate a repeatability of 9 cm s-1, compatible with the computed photon noise and already better than a Th-Ar calibration, which needs to use all 72 orders. By using the LFC we could measure the contribution of the CCD pixel irregularities to the calibration curve for the first time. This leads to an unprecedented low scatter of the residuals. Applied to the measurement of absolute Th-Ar frequencies we could exceed the accuracy of the published values. After spectrally broadening our LFC in a highly nonlinear fiber we hope to cover the whole visible range in future. By using an adjustable spectral filter the comb modes will also have about equal intensity. This system will reduce the photon noise by more than an order of magnitude which is expected to enable Doppler shift measurements at the cm s-1 level.

Very informative. It's greatly appreciated.

We wait for precise measure gj 581 and solve problems with dodgy planets there

The latest edition of ESO’s quarterly journal, The Messenger, is now available online. Find out the latest news from ESO on topics ranging from new instruments to the latest science disconveniez
http://www.eso.org/public/announcements/ann12077/

Include :
- A report on the calibration of HARPS at the exo-Earth detection limit (page 2 to 6)

- An update of the progress made in the construction of the second generation VLT instrument SPHERE (the Spectro-Polarimetric High-contrast Exoplanet REsearch instrument). (page 17 to 21)

Very good. This thread may be joined with the HARPS-S thread.

Agreed, and done.

http://www.eso.org/public/usa/announcements/ann15037/

HARPS Laser Frequency Comb Commissioned

« …A precision in measuring stellar velocities of better than 0.2 kilometres/hour (5 centimetres/second) on a single exposure was obtained. This improved to better than 0.07 kilometres/hour (2 centimetres/second) when subsequent exposures were combined [1]. By comparison, the amplitude of the radial velocity induced in the Sun by the Earth as it orbits is 0.3 kilometres/hour (9 centimetres/second), so the new laser frequency combs in principle enable, for the first time, the detection of Earth-like planets in Earth-like orbits… »

http://www.eso.org/public/announcements/ann17056/

A powerful new instrument called NIRPS (Near Infra Red Planet Searcher) is to be installed on the ESO 3.6-metre telescope at the La Silla Observatory in Chile
(…)
NIRPS will complement the HARPS (High Accuracy Radial velocity Planet Searcher) instrument currently attached to the ESO 3.6-metre telescope
(…)
NIRPS will become the “red arm” of HARPS, extending the telescope’s capability into the infrared and allowing astronomers to characterise planetary systems. ESO will allocate 725 observing nights over a five-year period to the NIRPS team.
The main goal of NIRPS is to use the radial velocity method to detect and characterise planets orbiting cool, red, low-mass M-type stars. In particular, NIRPS aims to find Earth-like rocky planets that could potentially be habitable.
(…)
NIRPS is expected to see first light in the last quarter of 2019