Simultaneous optical and near-infrared linear spectropolarimetry of the earthshine

Simultaneous optical and near-infrared linear spectropolarimetry of the earthshine

P. A. Miles-Páez Instituto de Astrofísica de Canarias, La Laguna, E38205 Spain
pamp@iac.es; epalle@iac.es Departamento de Astrofísica, Universidad de La Laguna, Av Astrofísico Francisco Sánchez, sn E38206–La Laguna, Spain 
   E. Pallé Instituto de Astrofísica de Canarias, La Laguna, E38205 Spain
pamp@iac.es; epalle@iac.es Departamento de Astrofísica, Universidad de La Laguna, Av Astrofísico Francisco Sánchez, sn E38206–La Laguna, Spain 
   M. R. Zapatero Osorio Centro de Astrobiología (CSIC-INTA), Carretera de Ajalvir km 4, E-28850 Torrejón de Ardoz, Madrid, Spain
mosorio@cab.inta-csic.es
Received 2013; accepted 2013
Key Words.:
polarization – Earth – Moon – scattering
Abstract

Context:

Aims:We aim to extend our current observational understanding of the integrated planet Earth spectropolarimetry from the optical to the near-infrared wavelengths. Major biomarkers like O and water vapor are strong flux absorbents in the Earth atmosphere and some linear polarization of the reflected stellar light is expected to occur at these wavelengths.

Methods:Simultaneous optical ( m) and near-infrared ( m) linear spectropolarimetric data of the earthshine were acquired by observing the nightside of the waxing Moon. The data have sufficient spectral resolution (2.51 nm in the optical, and 1.83 and 2.91 nm in the near-infrared) to resolve major molecular species present in the Earth atmosphere.

Results:We find the highest values of linear polarization () at the bluest wavelengths, which agrees with the literature. Linear polarization intensity steadily decreases towards red wavelengths reaching a nearly flat value beyond 0.8 m. In the near-infrared, we measured a polarization degree of 4.5 % for the continuum. We report the detection of molecular features due to O at m and HO at 0.653–0.725 m, 0.780–0.825 m, 0.93 and 1.12 m in the spectropolarimetric data; most of them show high linear polarimetry degrees above the continuum. In particular, the broad HO 1.12 m band displays a polarimetric intensity as high as that of the blue optical. These features may become a powerful tool to characterize Earth-like planets in polarized light.

Conclusions:

1 Introduction

The increasing progress in instrumentation and observational techniques has allowed the detection of over 1,000 planets with masses ranging from a few times the Jupiter mass to the super-Earth regime ( M). The most recent observational studies (e.g., 2012Natur.481..167C) suggest that as many as 62 of the Milky Way stars can harbor Earth-like planets. It seems probable that a significant number of Earth-sized planets located within their parent stars habitable zone will be found in the near future.

Transit spectroscopy is currently the most extended technique to characterize the planetary atmospheres of the new discoveries, and has proved successful for gaseous giant planets (e.g., 2008Natur.452..329S). The atmospheres of Earth-sized planets are too difficult to study with actual instrumentation. While biomarkers are readily identifiable in the Earth’s transmission spectrum (2009Natur.459..814P), even in the most favorable case of an Earth-like planet orbiting small stars or brown dwarfs, several transits would be necessary for a reliable detection even using large-aperture (and space-based) telescopes (2009ApJ...698..519K; 2011ApJ...728...19P). The probability of finding a transiting exo-Earth in the habitable zone of a G2V-type star is  (2012ApJS..201...15H). The direct detection of the light reflected/emitted by the exo-Earth would be necessary for its proper characterization; this is indeed a very challenging task.

Polarization may contribute to the analysis of the atmospheres of these new worlds by taking advantage of the fact that the stellar light is generally not polarized, whereas the light reflected from the planet is. Various groups (2008A&A...482..989S; 2011A&A...530A..69K) have theoretically studied the optical linear polarimetric signal of an Earth-like planet as a function of orbital phase and wavelength. Their predictions must be tested against observations of the only ‘Earth-like planet’ known so far: the Earth. This can be done by observing the earthshine, which is the sunlight scattered by the dayside Earth and reflected by the nightside of the Moon. The optical and near-infrared earthshine spectra reveal atmospheric and surface biosignatures and are sensitive to features like water clouds, oceans, deserts, and volcanic activity (e.g., 2002ApJ...574..430W; 2006ApJ...644..551T; 2006ApJ...651..544M; 2009Natur.459..814P; 2012ApJ...755..103G); many are present in the earthshine optical spectropolarimetric data (2012Natur.483...64S; 2013PASJ...65...38T).

Here, we provide the first linear spectropolarimetric measurements of the earthshine from the visible to the near-infrared (NIR) wavelengths (0.4–2.3 m). We used observations of the Moon’s dark side for a characterization of the integrated polarimetric properties of the Earth. These observations extend our current understanding of the optical linear polarimetric data of the earthshine (2013A&A...556A.117B and references therein) towards red wavelengths.

2 Observations and data analysis

Earthshine optical and NIR linear polarimetric spectra were taken using the Andalucía faint object spectrograph and camera (ALFOSC) and the long-slit intermediate resolution infrared spectrograph (LIRIS; 2004SPIE.5492.1094M) of the 2.56-m Nordic Optical Telescope (NOT) and the 4.2-m William Herschel Telescope (WHT), respectively. Observing date was 2013 May 18; the journal of the observations is provided in Table 1.

ALFOSC polarimetric mode uses a calcite plate that provides simultaneous measurements of the ordinary and the extraordinary components of two orthogonally polarized beams separated by 15″, and a half-wave plate that we rotated four times in steps of 22.5. For spectroscopy we employed the grism 4 and a slit 15 width and 15″long, which in combination with a pixel pitch of 019 (2048  2048 E2V detector) yield a wavelength coverage of m and a spectral resolution of 2.51 nm ( at 0.65 m).

LIRIS polarimetric mode uses a wedged double Wollaston device (1997A&AS..123..589O), consisting of the combination of two Wollaston prisms that deliver four simultaneous images of the polarized flux at vector angles 0 and 90, 45 and 135. An aperture mask 4′1′ in size is in the light path to prevent overlapping effects between the different polarization vector images. Two retarder plates provide polarimetric images with exchanged orthogonal vector angles, useful to minimize flat-fielding effects and provide better signal-to-noise (S/N) data (2011ApJ...740....4Z; 2013A&A...556A.125M). For spectroscopy we employed the grisms and , and a slit 075 width and 50″ long, which in combination with a pixel size of 025 (1024  1024 HgCdTe HAWAII array) yield wavelength coverages of m () and m (), and spectral resolutions of 1.83 nm ( at 1 m) and 2.91 nm ( at 1.6 and 2.2 m).

Spectropolarimetric observations were carried out under photometric sky conditions and raw seeing of 09. The angle formed by the Sun, the Moon, and the Earth was 81°, and the waxing Moon illuminated area was 59%. At the beginning of the astronomical night, the NOT and WHT were pointed to the center of the lunar disk, and an offset was applied to move both telescopes to the dark side (East) of the Moon. To track the Moon with the NOT during the observations, we manually introduced proper right ascension and declination rates111Ephemeris computations by the Jet Propulsion Laboratory HORIZONS project: http://ssd.jpl.nasa.gov/?horizons every five minutes. The location of the ALFOSC and LIRIS slits on the Moon is illustrated in Figure 1. Half of the LIRIS slit was positioned at the Moon terminator and the “dark” side while the other half registered the sky contribution simultaneously. Because of the small size of the ALFOSC slit, alternate Moon and sky observations were conducted at a separation of 600″.

One polarimetric cycle consisted of four consecutive frames with half-wave plate positions of 0°, 225, 45°, and 675 (ALFOSC), and two consecutive frames obtained with two different retarder plates (LIRIS). Typical exposure times per frame were 90 s and 400 s for the optical and NIR; one full polarimetric cycle was thus completed in about 7 min (ALFOSC) and 14 min (LIRIS) including 0.5–1-min overheads. We collected a total of 15 cycles of earthshine polarimetric data and 7 cycles of “sky” observations with ALFOSC, and 6 () and 4 () cycles with LIRIS. Before sky subtraction, all LIRIS frames were rectified using the telluric emission lines and observations of Ar and Xe arcs so that the spatial axis lies perpendicular to the spectral axis. The earthshine ALFOSC data were sky subtracted using sky frames obtained very close in time and typically within 0.1 airmass. Wavelength calibration was performed using observations of HeNe (ALFOSC) and ArXe (LIRIS) lamps, with typical uncertainty of 0.15–0.20 Å.

Figure 1: Relative positions of the ALFOSC and LIRIS slits on the Moon surface during the observations of the lunar nightside on 2013 May 18. Equatorial coordinates and correspond to 21 20 universal time. For displaying purposes, the slit widths are multiplied by a factor of four, and the slit lengths are plotted on true scale.

The normalized Stokes parameters and were computed using the flux ratio method and the equations given in 2005ApJ...621..445Z for the optical and in 2011ApJ...740....4Z for the NIR. We obtained as many and spectra as polarimetric cycles were observed at optical wavelengths. To improve the signal-to-noise (S/N) ratio of the data, we present here final spectra computed by combining all polarimetric cycles and by collapsing all pixels along the spatial direction. A total of 65 ALFOSC pixels (or 1235) per wavelength were fused together into one-dimension spectra, while LIRIS data were folded up into two regions (indicated in Figure 1): one (A, 39 pix or 975) close to the Moon terminator, and another (B, 38 pix or 95) spatially near the location of the ALFOSC slit; each section corresponds to 1/4 of the total LIRIS slit length. In the collapsing processes, we obtained the average of all pixels and rejected the maximum and minimum values per wavelength. Optical data have better S/N than NIR spectra by factors of 15 (optical versus ) and 25 (optical versus ). The linear polarization degree was derived as the quadratic sum of the normalized Stokes parameters. The debiased polarization degree, , was computed by considering the polarization uncertainty (see below) and according to the equation given in 1974ApJ...194..249W.

Together with the earthshine data, observations of polarized and non-polarized standard stars (HD 154892, HD 161056, and Elia 2-25) from the catalogs of 1992AJ....104.1563S and 1992ApJ...386..562W were acquired with the same instrumental configuration and on the same observing night as the earthsine, and were used to control the instrumental linear polarization and to check the efficiency of the ALFOSC and LIRIS optics. The same unpolarized standard was targeted at the two sites. We found an upper limit of 0.1% on the instrumental linear polarization degree across the explored wavelengths. We also found that the measured vibration angle of the linear polarization, , has to be corrected by a constant value () at the LIRIS , , , and wavelengths (which agrees with the assumption made by 2013A&A...556A.125M), while the angle correction is wavelength-dependent for ALFOSC: from at 0.4 m to at 0.9 m. We used the observations of the zero-polarized standard stars (conveniently scaled in terms of S/N) to estimate the uncertainties associated with the earthshine , , and measurements. They are as follows for : 0.3% at 0.45 m, 0.2% at 0.60 m, 0.6% at 0.85 m, 0.7% at 1.04 m, 0.6% at 1.25 m, 0.8% at 1.64 m, and 1.2% at 2.15 m.

Instrument Elapsed time Exp. time Airmass
(UT) (s) (m) (m) (%) (deg)
ALFOSC 21:20–00:32 15490 1.19–2.58 0.4–0.9 0.428
0.535
0.631
0.810
LIRIS 21:32–23:05 62400 1.22–1.57 0.9–1.5 1.035
1.250
LIRIS 23:08–00:01 42400 1.67–2.18 1.4–2.4 1.640
2.150
222 $a$$a$footnotetext: Number of polarimetric cycles as defined in Section 2, number of frames in one cycle, integration time per frame.
Table 1: Observing log (2013 May 18).

3 Results and discussion

Figure 2: Top: Spectroscopic visible (black line) and NIR (gray lines) linear polarization degree of the earthshine obtained on 2013 May 18. Region B (see text) lies spatially near the location of the optical observations; region A includes the Moon terminator. Middle: Spectroscopic visible and NIR linear polarization degree of observed polarized and non-polarized standard stars (black lines). Literature photometric polarimetry (1992AJ....104.1563S; 1992ApJ...386..562W) is plotted as white diamonds, gray horizontal error bars represent the filter widths. Bottom: Spectroscopic polarization vibration angle of the earthshine and polarized standard stars. Symbols are as in previous panels. In all panels, wavelengths strongly affected by telluric absorption are removed.

Figure 2 depicts the ALFOSC and LIRIS debiased linear polarization degree and vibration angle of polarization () spectra for both the Earth and the standard stars. Note that the optical and NIR polarized stars (middle and bottom panels) are different sources. The data of the standards agree with the literature measurements at the 1- level, thus supporting our confidence in the data analysis and instruments performance.

We show the Earth NIR spectra (top panel of Figure 2) for the two Moon sections A and B of the LIRIS slit (see Figure 1). Both Moon regions yield the same spectral shape and display similar features, but the polarization intensity of region B (more similar to the optical polarimetric signal and also spatially closer to the ALFOSC slit) is a factor of larger than that of region A. This factor remains practically constant throughout all NIR wavelengths. Since regions A and B extend over 17 km each on the Moon limb, we attribute this difference to the fact of observing distinct lunar superficial regions, e.g., maria and highlands. Our finding agrees with the recent study by 2013A&A...556A.117B; these authors measured the earthshine ( since ) by observing simultaneously highland and mare lunar areas using optical filters, and found that mare regions yield values of a factor of 1.3 times larger than highlands. There is also an offset of about 0.5–1.0 % in the linear polarimetry degree between the red wavelengths of the optical observations and the NIR B-region data, which may be also explained by the monitoring of distinct lunar areas. This highly contrasts with the excellent optical–NIR match of the polarization vibration angle measurements seen in the bottom panel of Figure 2. We provide Earth and measurements (B region) at different wavelengths of reference in the last three columns of Table1.

The most noticeable feature of the earthshine linear polarization degree is that the polarization intensity is the largest () at the bluest wavelengths and steadily decreases towards the red. Various groups (1957SAnAp...4....3D; 2005ASPC..343..211W; 2012Natur.483...64S; 2013PASJ...65...38T; 2013A&A...556A.117B) have reported a similar property. In our data, we observe that the linear polarization intensity remains nearly constant at about 5.7 % between m and m. Interestingly, the general shape of the NIR m polarimetric spectrum (except for some signatures due to water vapor and oxygen) also appears rather flat within the quoted uncertainties with a value around  % independently of wavelength. On the contrary, the behavior of the vibration angle appears slightly different. Whereas it remains constant within for all optical frequencies, the broad-shaped variations of up to 30° detected in the NIR (which cannot be attributed to instrumental or measurement errors) seem to correlate with the spectral features observed in . This may be attributed to the wavelength-dependent relative contribution of different components (atmospheric species, clouds, aerosols, and surface albedos) to the globally-integrated linear polarization of the earthshine.

Figure 3: Our visible and NIR spectropolarimetric measurements of the earthshine are compared to literature data. A 10-pixel binning was applied to the NIR spectrum of region B. The uncertainty per wavelength is plotted as vertical gray error bars. Wavelengths of strong telluric absorption have been removed. Some molecular species seen in “emission” (indicative of strong atmospheric flux absorption and less multiscattering processes occurring at those particular wavelengths) are labeled. The vertical dashed line separates the ALFOSC and LIRIS data.

We compare our measurements with data from the literature in Figure 3. To improve the quality of the NIR linear polarization degree spectrum of region B, we applied a 10-pixel binning in the spectral dimension. Overlaid in Figure 3 are the optical values obtained at a spectral resolution of 3 nm and for two separated dates by 2012Natur.483...64S, and the broadband filter measurements of Moon highlands and maria made by 2013A&A...556A.117B for a Sun-Earth-Moon phase angle similar to ours. All optical data display a qualitatively similar pattern (previously discussed), but they differ quantitatively in the amount of polarization per wavelength and the spectral slope. The spectral slope of 2012Natur.483...64S and 2013A&A...556A.117B data is steeper than the ALFOSC spectrum, while our measurements and those of 2013PASJ...65...38T display related declivity. These differences may be understood in terms of distinct lunar areas explored by the various groups and different observing dates. 2012Natur.483...64S attributed the discrepancies of their two spectra solely to the time-dependent fraction of Earth clouds, continents, and oceans contributing to the earthshine. Our simultaneous NIR spectra of regions A and B support the conclusion of 2013A&A...556A.117B that linear values may also partially depend on the exact location of the Moon observed.

Despite the featureless appearance of the polarimetric spectrum of the Earth, some signatures are still observable at the level of 3  and at the resolution and quality of our data. The most prominent molecular features have been identified as labeled in Figure 3: the optical O at m (previously reported by 2012Natur.483...64S) and HO in the intervals 0.653–0.725 m and 0.780–0.825 m, and the NIR HO at 0.93 m and m and O at 1.25 m, all the latter reported here for the first time. Linear polarization is larger inside deep absorption molecular bands because strong opacity leaves only upper atmospheric layers to contribute significantly to the observed flux, thus reducing multiple scattered with respect to single scattered photons. As discussed by 2003ApJ...585L.155S and 2008A&A...482..989S, single scattering produces more intense polarization indices than multiple scattering events. The presence of the O A-band at m and HO at 0.653–0.725 m, 0.780–0.825 m, and m in the Earth spectropolarimetry was already predicted by 2008A&A...482..989S. These authors stressed the sensitiveness of the A-band polarization index to the planet gas mixing ratio and altitude of the clouds. Interestingly, the peak of the linear polarization at the center of the 1.12-m HO band is times larger than the values of the surrounding continuum, i.e., similar in intensity to the blue optical wavelengths. Even more important should be the linear polarization signal of water bands at m and m, which unfortunately cannot be characterized from the ground due to strong telluric absorption. Spectropolarimetric models of the Earth, guided by the visible and NIR observations shown here, could provide hints about their expected polarization values.

To this point, we have presented earthshine values as measured from the light deflection on the Moon surface. However, this process introduces significant depolarization due to the back-scattering from the lunar soil (1957SAnAp...4....3D); the true linear polarization intensity of the planet Earth is actually larger. At optical wavelengths, the depolarization is estimated at a factor of ( in nm) by 1957SAnAp...4....3D, making true polarization be in the range 26–31%. We are not aware of any determination of the depolarization factor for the NIR in the literature. The extrapolation of equation 9 by 2013A&A...556A.117B towards the NIR yields corrections of for the wavelength interval m, implying that the true linear polarization intensity of the Earth may be for the NIR continuum, and at the peak of the O (1.25 m) and HO (1.12 m) bands. The extrapolation of 1957SAnAp...4....3D depolarization wavelength dependency towards the NIR would yield even higher true polarization values by a factor of 4. Further modeling efforts are needed to confirm these features, which may become a powerful tool for the search for Earth-like worlds and their characterization in polarized light.

Acknowledgements.
Based on observations made with the William Herschel and Nordic Optical Telescopes operated on the island of La Palma by the Isaac Newton Group and the Nordic Optical Telescope Scientific Association (NOTSA), respectively, in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofísica de Canarias. The instrument ALFOSC is provided by the Instituto de Astrofísica de Andalucía under a joint agreement with the University of Copenhagen and NOTSA. We thank Dr. D. Stam for helpful discussion. This work is partly financed by the Spanish Ministry of Economics and Competitiveness through projects AYA2012-39612-C03-02 and AYA2011-30147-C03-03.

References

Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
229107
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description