PROCEEDINGS OF SPIE

SPIEDigitalLibrary.org/conference-proceedings-of-spie

The use of sodium layer density

anisotropies to fully measure the
atmospheric turbulence, including tip-
tilt, focus, and higher order
aberrations

Martinez, Noelia, Rodríguez Ramos, Luis Fernando

Noelia Martinez, Luis Fernando Rodríguez Ramos, "The use of sodium layer
density anisotropies to fully measure the atmospheric turbulence, including
tip-tilt, focus, and higher order aberrations," Proc. SPIE 11448, Adaptive
Optics Systems VII, 114481V (13 December 2020); doi: 10.1117/12.2560784

Event: SPIE Astronomical Telescopes + Instrumentation, 2020, Online Only

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 The use of Sodium Layer density anisotropies to fully
measure the atmospheric turbulence, including tip-tilt, focus
and higher order aberrations
Noelia Martinezb and Luis Fernando Rodrı́guez Ramosa
a Institutode Astrofı́sica de Canarias (IAC),Vı́a Láctea s/n, San Cristobal de La Laguna,
Spain;
b Australian National University, Research School of Astronomy and Astrophysics, Mount

Stromlo Observatory, Cotter Road, Weston Creek, Australia.

ABSTRACT
It is well known that Adaptive Optics (AO) observations do require the availability of natural stars (beside laser
guide stars - LGSs) to obtain the tip-tilt and focus information of the atmospheric turbulence. Bright natural
stars are not always available imposing the ultimate limit to the AO technology due to the small technical
field planned for 40m range telescopes. The use of multiple LGSs, with their respective wavefront sensors
and tomographic computations, requires the proper reconstruction of the turbulence column adding significant
complexity and cost to the Adaptive Optics systems.
Precise knowledge of the tip-tilt information is extremely useful for the accurate pointing of lasers in ground-
to-space optical communications and in space situational awareness applications. In these contexts the tip-tilt
information cannot be obtained from natural stars.
Our group have proposed an alternative way of measuring all relevant values of the atmospheric turbulence,
specifically including tip-tilt, focus and also higher order aberrations, and tomographic information based on the
use of the foreseen density anisotropies in the sodium layer.
Results of an analysis using the available information about sodium layer profiles will be presented, showing up
to what point anisotropies with the proper spatial and time scales could be expected.
The requirements of a laser launch system capable of illuminating uniformly the metapupil at the height of the
Na layer are identified and special attention is paid to the required laser power.
The extended object generated at the Na-Layer should then be analysed with a wavefront sensor suited to that
characteristic. The plenoptic camera is the potential candidate under study in this paper.
Keywords: Laser Guide Star, Sodium layer, tip-tilt, plenoptic camera

1. INTRODUCTION
Adaptive Optics (AO) measure and correct for the negative effects the atmospheric turbulence induces on the
light. Applications involving light coming from any object on the sky (astronomical observations) or any laser
light propagated through the atmosphere are deeply affected by the turbulence. AO systems probe the atmo-
sphere and measure its negative contribution by sensing a so-called ”guide star”. This guide star could be a
natural star, nearby the object of interest inside the isoplanatic area,[1] or an artificial star (Laser Guide Star)
propagated to the location where the object of interest is. Sodium Laser Guide Stars (LGS) are artificial stars
generated on the sodium layer at around 90 km above the Earth by the excitation of the sodium atoms. Guide
stars need to produce enough photons at the collecting aperture such that the AO instrument could sense the
turbulence faster than the atmospheric temporal change rate. To compensate the lack of enough bright natural
stars, Laser Guide Stars are extensively used in astronomy to increase the sky coverage from 10% up to 85%.[2]
However, probing the atmosphere by the utilisation of artificial stars does not allow to retrieve the first modes
of the atmospheric turbulence, known as tip-tilt.
Further author information: (Send correspondence to L.R.R)
L.R.R.: E-mail: [email protected]

Adaptive Optics Systems VII, edited by Laura Schreiber, Dirk Schmidt, Elise Vernet, Proc. of SPIE
Vol. 11448, 114481V · © 2020 SPIE · CCC code: 0277-786X/20/$21 · doi: 10.1117/12.2560784

Proc. of SPIE Vol. 11448 114481V-1

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 This paper presents the tip-tilt retrieval problem when using LGS AO systems and proposes the investigation
of the sodium inhomogeneities in combination with the plenoptic wavefront sensor as a solution for the current
limitation of Adaptive Optics. This paper is a follow up of previous research.[3]

1.1 The Tip-Tilt Retrieval Problem

The atmospheric tip-tilt represents the absolute change on the wavefront slope as a result of the light wave
travelling through the atmosphere. When using Laser Guide Stars, measuring the tip-tilt has not been techni-
cally feasible until now: the guidestar laser is propagated upwards in the atmosphere and excites the sodium
atoms, whose backscattered light travels downwards through almost the same path; as a consequence, the laser
light is deflected twice cancelling out the tip-tilt effect on the wavefront sensor. Natural guide stars provide this
information for the AO systems in operation up to present.
The natural star is therefore the current limitation of the Adaptive Optics technology. Cutting-edge instruments
like HARMONI to be installed in the future European Extremely Large Telescope need to be designed around
this constraint: i.e. HARMONI aims to cover only a 10% of the sky on the South Galactic pole; to fulfil this
objective, a dedicated deformable mirror needs to work in open loop to locate natural stars with enough bright-
ness.
A mechanism enabling the tip-tilt retrieval from Laser Guide Stars would greatly influence the field of Adaptive
Optics. Measuring the atmospheric tip-tilt with an LGS would take the technology to the next level, allowing
100% sky coverage no matter the application.

1.2 The Plenoptic Camera

Considering a fixed location on a scene, all the light rays passing through that location could be defined by the
incident angles of the light (θ, φ) at a certain wavelength λ, through the fixed coordinates (Px , Py , Pz ), and at a
time instant t. The directional light distribution at all possible positions in a 3D space was defined by[4] as the
plenoptic function.The plenoptic function is the link between the objects of a scene and their images.

Figure 1. The plenoptic function, a scalar function of four dimensions, is a generalisation of an image and describes all
information available in an imaging system.[5]

It is possible to have information about the position of any point in the space when the structure generating
it in the 3D plenoptic space is known and detectable.[6] When further developing this idea, the benefit of the
plenoptic technology becomes clear: working on the plenoptic space enables the generation of images from dif-
ferent viewpoints.
By placing the plenoptic camera at the telescope focal plane, pupil images from as many different viewpoints
as pixels within the microlenses array are generated. The aperture images associated to each viewpoint are
processed and the relative displacement between them is calculated, estimating the wavefront slope.
One of the main advantages of the plenoptic cameras as a wavefront sensor is its performance with extended
objects and strong turbulence conditions[7][8][9] .

Proc. of SPIE Vol. 11448 114481V-2

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 1.3 The Sodium Layer Anisotropies
The sodium layer consists of an accumulation of free sodium atoms maintained by the deposition of meteoritic
ablation particles at altitudes between 80 km and 100 km in the upper atmosphere. The sodium abundance
varies on hourly, daily and yearly time scales with minimum and maximum reported abundance occurring in
summer and winter, respectively[10][11] .
The sodium layer has been deeply studied on its vertical distribution[10][12][13] as the layer centroid is an important
parameter for the performance of Adaptive Optics systems. To the contrary, the horizontal distribution of the
sodium atoms has not been analysed as by a dedicated experiment yet.
Figure 2 shows three different sodium profiles obtained with the Natrium experiment at Teide Observatory
(Canary Islands).[11] The temporal and spatial variability of the sodium layer is thereby represented.

Figure 2. Sodium profiles measured at Teide Observatory (Canary Islands) by the Natrium experiment[11] showing the
great non-uniformity of the layer.

2. PROPOSED SOLUTION FOR THE LGS-TT RETRIEVAL

A solution for the tip-tilt retrieval on Laser Guide Stars is proposed in this paper. By illuminating the sodium
layer with an homogeneous beam of the required width, one could observe the image generated by the anisotropies
with a wavefront sensor adequate for extended objects. The situation will then resemble the well-known use of
solar granulation in solar adaptive optics. The large illuminated area will wander on the sky as a result of the
uplink tip-tilt; the non-uniformities within that area will show a different movement: the one resulting from the
downlink, thereby isolating the atmospheric tip-tilt as if it were sensed on a natural star.
The plenoptic wavefront sensor will sense the extended illuminated area of the sodium layer and the atmospheric
turbulence including tip-tilt will be extracted by correlation-based techniques.
Extended illumination of the sodium anisotropies would not only greatly overcome the tip-tilt retrieval problem,
but could also be used to replace LGS asterisms in Laser Tomography Adaptive Optics (LTAO) systems; the
cost of LTAO instrument would be considerably reduced as there is no longer need of more than one guidestar
laser sources.
Figure 3 illustrates the concept: the whole projection of the telescope pupils on the sodium layer is illuminated
by the guidestar laser. Additionally an extra area around the telescope field of view is required to always have
an illuminated sodium patch on the wavefront sensor (the plenoptic camera).

The sodium anisotropies are required to be stable on time at the atmospheric scale. No variation could be
observed during time scales of typically 10 ms or less (atmospheric coherence time). The layer sodium distribu-
tion needs to change only as a result of the atmospheric wavefront. The non-uniformities also need to present
the appropriate spatial scales for the wavefront sensor to detect contrast.
The LGS launch system needs to illuminate the sodium layer as uniformly as possible. Adaptive Optics pre-
compensation (high order modes) of the guidestar laser wavefront on the upwards path is therefore likely to be
required. The effect of the scintillation on the LGS uplink will be further investigated in future publications.

Proc. of SPIE Vol. 11448 114481V-3

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 3. The projection of the telescope pupil needs to be fully illuminated by the guidestar laser as well as an additional
area around it determined by the field of view of interest. This total area is called metapupil.

3. ANALYSIS OF REQUIRED LASER POWER

Perhaps the most challenging part of the tip-tilt, and other Zernikes, atmospheric turbulence wavefront mea-
surement is the amount of laser power required to properly illuminate the expect inhomogeneities on the sodium
layer. In this section calculations are outlined to provide some insight on this relevant issue.
Figure 3 depicts the fraction of the Na Layer that should be illuminated: essentially the one covering the field of
view of the observation. It is basically the equivalent meta-pupil generated by the telescope pupil plus the extra
diameter depending on the field of view, computed at the height of the sodium layer.
Table 1 shows four example situations that can be considered as representative of the span required nowadays
along observatories, going from the HARMONI (E-ELT (D = 39 m) first light instrument) field of view, the
FRIDA instrument in development for GTC telescope (D = 10 m), and the sample cases of WHT telescope (D
= 4.2 m) and the ESA’s Optical Ground Station (D = 1 m). Fields of view and pupil sizes have been obtained
based on each specific case.
One observes the similarity on the area of the Na layer to be illuminated, ranging over a factor of just two for all
considered examples. This is an important result, which can lead to considering just one design of the launching
system for many possible applications.
Any wavefront sensor capable of accepting extended objects can be used for the tip-tilt and other Zernikes mea-
surement; the plenoptic camera has been proposed to this aim. In this case the wavefront sensor shall be focused
at the sodium Layer, not at infinity, and the resultant plenoptic image will look like Figure 4 in which the orange
circle represents the Sodium Layer illumination. Table 2 shows the effect of the required pupil sampling and
hence, the number of microlenses in the lenslet array for each study case.

Proc. of SPIE Vol. 11448 114481V-4

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 4. The plenoptic frame would be focused on the sodium layer at 90 km and receive photons coming from the
illuminated metapupil.
Table 1. Calculation of the metapupil area on the sodium layer for the proposed study cases.

STUDY CASE (1) (2) (3) (4)

Telescope Diameter [m] 1 4.2 10 39
Field-of-View [arcsec] 60 50 40 20
Sodium layer height [km] 90 90 90 90
2
Circular illuminated area [m ] 1135.55 965.05 944.75 2070.26

It can be concluded that even though the different telescopes examples need similar sodium layer areas to
be illuminated, the areas seen by each pixel are very different, spanning along a couple of orders of magnitude,
mainly due to the pupil sampling requirement: i.e. the case of the E-ELT will require two orders of magnitude
more power than the Optical Ground Station.
The absolute amount of power required in each case depends on the pixel noise being considered, essentially
readout and photon noise. The Na layer seen by each pixel shall generate enough photons to overcome the noise
and produce an adequate signal-to-noise ratio.

4. SIMULATIONS: FRAMEWORK AND RESULTS

The proposed tip-tilt measurement system has been simulated with the aim of verifying its feasibility for three
possible telescopes: the Optical Ground Station telescope at Teide Observatory (D = 1 m), the William Herschel
Telescope (D = 4.2 m), the GTC telescope -both at Roque de los Muchahos Observatory (D = 10 m), and the
future E-ELT (D = 39 m).
The goal of the numerical simulation is to demonstrate that the atmospheric tip-tilt information could be mea-
sured on laser guide stars by sensing the sodium anisotropies with the plenoptic wavefront sensor. Simulations
operate under the assumptions that perfect illumination of the sodium layer is achieved (no scintillation effects)
and the sodium spatial and temporal distribution is ideal to detect high contrast images on the plenoptic camera.

The system has been modelled using the Object Oriented Matlab Adaptive Optics (OOMAO) toolbox[14]
after implementing the plenoptic wavefront sensor functionality. The simulation framework is summarised in
Table 3.

Proc. of SPIE Vol. 11448 114481V-5

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Table 2. Calculation of the plenoptic wavefront sensor configuration for the proposed study cases.

STUDY CASE (1) (2) (3) (4)

Number of pixels/microlens 12 12 20 80
Number of microlenses along the axis 12 12 12 12
2
Sodium layer area/microlens [m ] 5.02 4.27 4.18 9.15
Sodium layer area/px [m2 ] 0.035 0.030 0.010 0.001

Table 3. Simulation Framework

Atmosphere r0 = 10 cm (based on OT atmospheric profile[15] )

3 layers at 0 km, 4 km and 10 km
Telescope Study Cases (1) D = 1 m and FOV = 60 arcsec
(2) D = 4.2 m and FOV = 50 arcsec
(3) D = 10 m and FOV = 40 arcsec
(4) D = 39 m and FOV = 20 arcsec
Laser Guide Star λ = 589 nm
Height = 90 km
Extensively illuminated
Wavefront Sensor Plenoptic WFS
Lenslet array = 12x12
Microlens size = 12 px (Case (1))
12 px (Case (2))
20 px (Case (3))
80 px (Case (4))
Photon noise and readout noise (0.1 e− ) included

The effect of the contrast on the sodium layer has been previously studied.[3] The high-contrast sodium
distribution in Figure 5 has been assumed as a horizontal snapshot of the layer.

Figure 6 shows an example of atmospheric phase retrieval on Study Case (1) by sensing the sodium anisotropies
with the plenoptic wavefront sensor.

Preliminary results on the method feasibility are presented; Figure 7 shows the root-mean square error
(RMSE) between the real and the reconstructed tilt in Study Case (1) with a 1-meter telescope when increasing
the Signal-to-Noise (SNR) ratio. Same procedure was repeated for all Study Cases (2) to (4) and the results
are summarised in Figure 8. As no Adaptive Optics correction of any kind is applied, the larger the telescope
collecting area, the stronger the turbulence effect will be. As a consequence, the image contrast of the sodium
anisotropies decreases (Figure 9) and the error in the TT retrieval becomes apparent.
Future simulations will introduce AO correction of high order modes on both upwards and downwards propaga-
tion path of the guidestar laser and results will be effectively compared with the ones presented in this paper.

Proc. of SPIE Vol. 11448 114481V-6

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 5. High contrast distribution assumed to be the illuminated area on the sodium layer for the plenoptic wavefront
measurements.

Figure 6. Example of phase retrieval using the sodium anisotropies and a 12x12 plenoptic wavefront sensor with 144x144
resolution. Z axis with arbitrary units.

5. CONCLUSIONS AND FUTURE WORK

The use of the sodium layer density anisotropies as the solution for the LGS-TT retrieval problem has been
validated through numerical simulations demonstrating its potential for future Adaptive Optics systems. The
plenoptic camera has proven its capabilities to fulfil the required performance in the proposed measurement
system.
Future research includes the assessment of the scintillation effect on the upwards propagated laser beam and
plan for a measurement campaign to take place at the Canary Islands Observatories (Spain).

ACKNOWLEDGMENTS
The authors would like to thank the Instituto de Astrofı́sica de Canarias (IAC) and the Australian National
University (ANU) for supporting this research.

REFERENCES
[1] Glindemann, A., “Principles of stellar interferometry,” Principles of Stellar Interferometry: , Astronomy
and Astrophysics Library. ISBN 978-3-642-15027-2. Springer-Verlag Berlin Heidelberg, 2011 (2011).

Proc. of SPIE Vol. 11448 114481V-7

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 7. Root Mean Square Error between the real tilt and the reconstructed tilt when using the plenoptic wavefront
sensor to measure the sodium anisotropies on a 1-m telescope at different SNR values.

[2] Ellerbroek, B. and Tyler, D., “Adaptive optics sky coverage calculations for the gemini-north telescope,”
Publications of the Astronomical Society of the Pacific 110(744), 165–185 (1998).
[3] Rodriguez-Ramos, L. F., Martinez, N., Alonso, A., and Castro-Almazan, J., “Sodium layer density
anisotropies as a reference for tip-tilt measurement in laser guide stars,” (2019).
[4] Adelson, E. H. and Bergen, J. R., “The plenoptic function and the elements of early vision,” in [Computa-
tional Models of Visual Processing ], 3–20, MIT Press (1991).
[5] Ng, R., Levoy, M., Brédif, M., Duval, G., Horowitz, M., and Hanrahan, P., Light field photography with a
hand-held plenoptic camera, PhD thesis, Stanford University (2005).
[6] Berent, J. and Dragotti, P. L., “Plenoptic manifolds,” IEEE Signal Processing Magazine 24(6), 34–44 (2007).
[7] Rodriguez Ramos, L. F., Utilización de la cámara plenóptica como sensor de frente de onda para Óptica
Adaptiva en Astrofı́sica, PhD thesis, ULL (2015).
[8] Martı́nez, N. and Ramos, L. F. R., “Solar mcao with a single sensor: simulating tomographic reconstruction
with the plenoptic camera,” in [Adaptive Optics Systems VI], Close, L. M., Schreiber, L., and Schmidt, D.,
eds., 10703, 1303 – 1312, International Society for Optics and Photonics, SPIE (2018).
[9] Martinez, N., Rodriguez Ramos, L. F., Montilla, I., Collados, M., and Montoya, L., “The plenoptic camera
as wavefront sensor for the vtt solar telescope,” (01 2017).
[10] Michaille, L., Clifford, J. B., Dainty, J. C., Gregory, T., Quartel, J. C., Reavell, F. C., Wilson, R. W., and
Wooder, N. J., “Characterization of the mesospheric sodium layer at la palma,” Monthly Notices of the
Royal Astronomical Society 328, 993–1000 (12 2001).

Proc. of SPIE Vol. 11448 114481V-8

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 8. Root Mean Square Error between the real tilt and the reconstructed tilt when using the plenoptic wavefront
sensor to measure the sodium anisotropies on telescopes of different diameter at a fixed SNR of 10.

[11] Castro-Almazán, J., Fuensalida, J., Alonso, A., and Montilla, I., “Natrium-teide experiment. sodium layer
observations at teide observatory (canary is.),” Journal of Physics: Conference Series 595, 012007 (04
2015).
[12] Neichel, B., D’Orgeville, C., Callingham, J., Rigaut, F., Winge, C., and Trancho, G., “Characterization
of the sodium layer at cerro pachón, and impact on laser guide star performance,” Monthly Notices of the
Royal Astronomical Society 429, 3522–3532 (01 2013).
[13] Castro-Almazan, J. Alonso, A. B. C. D. C. M. L. G. M. I. M.-T. C. R. M., “Na lgs height profiles at teide
observatory, canary islands.,” (2017).
[14] Conan, R., [Object-Oriented Matlab Adaptive Optics. User Guide ] (2013).
[15] Garcı́a-Lorenzo, B. and Fuensalida, J. J., “Statistical structure of the atmospheric optical turbulence at
teide observatory from recalibrated generalized scidar data,” Monthly Notices of the Royal Astronomical
Society 410(2), 934–945 (2011).

Proc. of SPIE Vol. 11448 114481V-9

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use
 Figure 9. Effect of the atmospheric turbulence on large telescopes. As the collecting area increases, the contrast of the
sodium inhomogeneities decreases affecting the performance of the TT retrieval method.

Proc. of SPIE Vol. 11448 114481V-10

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 17 Dec 2020
Terms of Use: https://www.spiedigitallibrary.org/terms-of-use