Free Access
Issue
A&A
Volume 655, November 2021
Article Number L3
Number of page(s) 5
Section Letters to the Editor
DOI https://doi.org/10.1051/0004-6361/202142470
Published online 25 November 2021

© ESO 2021

1. Introduction

Brown dwarfs (BDs) do not stabilize on the hydrogen (H) burning main sequence, and hence they keep cooling down to very low temperatures. The coolest BDs have been spectroscopically classified as T dwarfs and Y dwarfs, based on the presence of methane and ammonia in near-infrared spectra, respectively (Burgasser et al. 2002; Delorme et al. 2008; Cushing et al. 2011), although the presence of ammonia has been established not only in Y dwarfs but also in late-T dwarfs (Bochanski et al. 2011; Canty et al. 2015).

The Wide Field Infrared Survey Explorer (WISE; Wright et al. 2010) has been very effective in discovering Y dwarfs and low-resolution spectroscopic follow-up has been carried out from the ground and from space. In particular, homogeneous sets of low-resolution near-infrared spectra have been collected with the Keck Near Infrared Spectrometer (NIRSPEC) and with the Hubble Space Telescope (HST) for late-T and Y dwarfs down to spectral class Y2 (McLean et al. 2003; Schneider et al. 2015; Cushing et al. 2021).

Estimates of the effective temperature (Teff) of Y dwarfs are difficult because theoretical fits to their spectral energy distribution are of modest quality, which indicates that the chemistry calculations and/or the sources of opacity may be incomplete. Furthermore, values estimated from their luminosity and assumed radii using the Stefan-Boltzmann law are hampered by uncertainties due to possible effects of unresolved binarity (Leggett et al. 2013; Luhman & Esplin 2016).

The appearance of ammonia in the near-infrared spectra of ultra-cool dwarfs has been discussed by several authors (Saumon et al. 2000; Leggett et al. 2007; Warren et al. 2007; Bochanski et al. 2011; Burgasser et al. 2012; Canty et al. 2015). A spectral index named NH3-H was defined by Delorme et al. (2008) to quantify the effects of ammonia absorption from 1.53 to 1.56 microns. Their NH3-H index has been adopted for spectral type quantification for dwarfs of classes T5 to Y2 (Cushing et al. 2021).

The upcoming Euclid1 space mission is expected to provide the next order of magnitude leap in the numbers of detected ultra-cool dwarfs. By the end of its six-year nominal mission, the Euclid wide survey is expected to reach 15 000 square degrees of the extragalactic sky observed at a single epoch, and the Euclid deep surveys are expected to cover between 40 and 50 square degrees at multiple epochs. The Near Infrared Spectrometer and Photometer (NISP) is a two-channel instrument for the Euclid space mission. The photometric channel is equipped with three broadband filters, and the spectroscopic channel is equipped with three identical low-resolution grisms that cover the spectral range from 1250 nm to 1850 nm with a spectral resolving power of R = 380 as well as one grism that covers the spectral range from 920 nm to 1250 nm. This last grism is expected to be used only for the Euclid deep fields and not for the wide survey. The spectroscopic observations will be made in survey mode without any slit. The NISP field of view is 0.53 square degrees covered with a matrix of 4 × 4 Teledyne detectors, each of them with 2040 × 2040 18-micron pixels with a scale of 0.3 arcsec per pixel.

Updated simulations of the number counts of ultra-cool dwarfs in the Euclid wide survey using the most recent information on the photometric passbands were provided in Solano et al. (2021). In this Letter we provide updated simulations for the number of T and Y dwarfs expected to be detected with the NISP spectroscopic mode in the Euclid wide survey. We show that Euclid will provide useful low-resolution H-band spectra for an unprecedentedly large number of T dwarfs (≥103 for T5–T8 subclasses). Consequently, it is important to investigate efficient methods for deriving Teff values for T dwarfs using NISP spectra. Our first goal is to confirm that the NH3-H spectral index is really an indicator of the absorption of ammonia in late-T and Y dwarfs, and to adjust its integration limits using our laboratory spectra. Our second goal is the determination of the ammonia-to-methane ratio as a probe of the atmospheres of ultra-cool dwarfs using low-resolution H-band spectra.

This Letter presents laboratory spectra of gas cells filled with ammonia and methane, which were made with the aim of helping with the identification of these molecules in the atmospheres of ultra-cool dwarfs. The rest of the paper is organized as follows: Sect. 2 describes the experimental setup for the laboratory measurements. Section 3 compares the laboratory spectra with simulated spectra using the high-resolution transmission molecular absorption database HITRAN and redefines the NH3-H index. Section 4 compares the laboratory spectra with observed near-infrared spectra from the Infrared Telescope Facility (IRTF), Hubble, and Keck for dwarfs with spectral types from T5 to Y2 as well as for Jupiter and Saturn. Ammonia-to-methane ratios are derived using a best matching algorithm on the laboratory and the observed spectra. Section 5 summarizes our results and their implications.

2. Low-resolution laboratory spectra

Custom-made gas cells were created to fit into the spectrophotometer Cary 5000 located at the optical laboratory of the Instituto de Astrofísica de Canarias (IAC). They are G2 Infrasil cells from International Crystal Lab, 13 cm long and with a window size 38 mm in diameter and 6 mm thick.

The cells were filled with different molecular gases using a vacuum pump system in our laboratory located in the chemistry department of the Universidad de La Laguna. The procedure for filling the cells with gas is described in more detail in Valdivielso et al. (2010). The gas cells with ammonia (NH3), argon (Ar), and methane (CH4) were completely filled with each of those gases at the atmospheric pressure.

The gas cells were taken to the IAC and were left to settle to the ambient conditions inside the Cary 5000 for at least five hours before the measurements were taken. The Cary 5000 has two channels, which allows for two gas cells to be measured simultaneously. In one channel we placed a gas cell filled with ammonia or methane, and in the other channel we placed an identical gas cell filled with Ar so that the instrumental response would be corrected by the instrument’s software.

The laboratory spectra of all the gas cells were taken on July 8 and 9, 2021, with an integration time of 0.2 s for each step of 0.5 nm. The total time to measure a spectrum from 800 nm to 3000 nm was 6.3 h. The ambient conditions in the laboratory were a temperature of 298 K and a humidity level of 57%.

3. Comparison between laboratory and simulated spectra

Our laboratory spectra for ammonia and methane were compared with simulated spectra for the same molecules using the HITRAN database (Gordon et al. 2017) for temperatures of 300 K and 500 K. The comparisons in the full spectral range from 1.0 to 2.5 microns are shown in Fig. 1. The agreement between 2.0 and 2.5 microns is quite good, but the differences increase toward shorter wavelengths. This is not a surprise because it had previously been reported that HITRAN is very incomplete for hot bands and high-J molecular transitions (Bailey & Kedziora-Chudczer 2012; Wong et al. 2019).

thumbnail Fig. 1.

Our laboratory spectrum for ammonia and methane compared to simulated spectra of the same molecules using HITRAN for temperatures of 500 K and 300 K. The wavelength regions proposed for the new NH3-H index are marked with color bands: green for the numerator, red for the denominator.

We used our laboratory spectrum of ammonia to propose a redefinition of the integration limits for the spectral index NH3-H, which was originally defined using HITRAN opacity data (Delorme et al. 2008). These authors used as numerator the spectral region from 1.53 to 1.56 microns, and as denominator the region from 1.57 to 1.60 microns. Based on the appearance of the ammonia band in our laboratory spectra, we propose redefining the NH3-H index using as numerator the region from 1.50 to 1.57 microns, and as denominator the region from 1.57 to 1.61 microns. These integration limits are shown in the bottom part of Fig. 1. They sample a region dominated by the ammonia absorption and not by methane.

The increased wavelength range chosen for our proposed modification of the NH3-H spectral index was motivated by a desire to better understand the ammonia absorption in this region, and it has the advantage of increasing the flux collected in each passband for the classification of ultra-faint objects of T and Y type detectable with NISP. The comparison between the previous NH3-H index and ours of dwarfs from spectral type T5 to Y2 is shown in Fig. 2.

thumbnail Fig. 2.

Redefined NH3-H index using our laboratory spectra. It has a higher sensitivity to spectral class (larger amplitude of variation) than the original definition by Delorme et al. (2008) based on HITRAN data. Third-order polynomial fits to the relationship between the indices and the spectral types are denoted with dashed lines of different colors, as labeled.

4. Comparison between laboratory and HST spectra

Different combinations of the relative strengths of ammonia and methane were computed using our laboratory spectra. An example of the typical combination that provides the best fit is shown in Fig. 3 together with HST spectra of a T8 dwarf and two Y dwarfs (Schneider et al. 2016; Cushing et al. 2014). This combination, denoted with a blue line, corresponds to the case where we scaled our laboratory spectrum of ammonia by a factor of 0.33 and added it to our laboratory spectrum of methane.

thumbnail Fig. 3.

Our laboratory spectra for a combination of ammonia and methane (blue) and for pure methane (red), compared to HST spectra of a T8 dwarf and two Y dwarfs, and IRTF spectra of two giant gas planets.

The spectral region from 1.5 to 1.7 microns is particularly sensitive to the effects of changing the relative contribution of ammonia with respect to methane. Furthermore, this spectral region is thought to be free from a significant contribution of other molecules, such as water vapor (Delorme et al. 2008; Canty et al. 2015).

A least square fit and a confidence interval estimate were computed using the lsqcurvefit and nlparci routines available in the Optimization Toolbox and the Statistics and Machine Learning Toolbox in the MATrix LABoratory (MATLAB). We found that the best fit to the HST and NIRSPEC spectra of late-T and Y dwarfs in the spectral window from 1.5 to 1.7 microns was achieved with the ratios of ammonia to methane shown in Fig. 4.

thumbnail Fig. 4.

Ammonia-to-methane ratios derived for late-T dwarfs and Y dwarfs with a 95% confidence level from NIRSPEC data (blue) and HST data (red) compared with those of Jupiter and Saturn from IRTF data (red) as functions of spectral type (upper panel) and Teff (lower panel). The Teff of T dwarfs, Y dwarfs, Jupiter, and Saturn adopted in this work are listed in Table 1. The dashed pink curves are the third-degree polynomial fittings for all objects (except Saturn) against spectral type and Teff, and the dashed black line is a linear fitting just for objects from T8 to Y2. In the lower panel, the results are compared with the scaled theoretical ratios of molecular column densities estimated from Zahnle & Marley (2014), which are shown in light blue.

Table 1.

Adopted Teff values and uncertainties for the different spectral types and giant gas planets considered in this work.

Using the same method as described above, we also derived the ammonia-to-methane ratio in the spectra of Jupiter and Saturn that we downloaded from the NASA IRTF archive (Vacca et al. 2003). Our ratio for Jupiter is plotted in Fig. 4, and it is consistent (within the 1 sigma uncertainty) with that found in the literature using models and observations of the Galileo spacecraft (0.26; Taylor et al. 2007). Our ratio for Saturn agrees with that of Owen et al. (1977), who reported on the non-detection of ammonia in high-resolution spectra of Saturn obtained at Palomar Observatory in the wavelength region relevant for this work.

In Fig. 4 (lower panel) we show the comparison of our ammonia-to-methane ratios with the theoretical predictions for high-gravity objects (g = 105 cm s−2) from Zahnle & Marley (2014). To compare the behavior of the models with other data, we used the ratios of theoretical column densities scaled by a constant factor to produce two values that reach an agreement at 600 K. There is a good qualitative agreement in the sense that the models predict a ratio that has a minimum around 800 K and increases toward cooler and hotter temperatures.

5. Discussion and final remarks

Our laboratory spectra of ammonia and methane contain absorption features that are too weak or not present in the HITRAN database. We used our laboratory spectrum of ammonia to redefine the spectral index NH3-H in late-T and Y dwarfs. This index is a sensitive indicator of the spectral type of ultra-cool BDs, and it can be used to calibrate the data provided by the slitless spectroscopic mode of the NISP instrument on Euclid. Our new defined NH3-H index has broader integration limits and an amplitude of variation with spectral type that is almost twice as large as the one based on HITRAN and originally proposed by Delorme et al. (2008).

There is a smooth relation between the ammonia-to-methane ratio derived from our best fits to the observed spectral region from 1.5 to 1.7 microns and the Teff of the objects considered, which is qualitatively in agreement with theoretical predictions taken from the ratio of the theoretical column densities of ammonia and methane under a high surface gravity environment (g = 105 cm s−2) (Zahnle & Marley 2014). All the late-T and Y dwarfs considered in this work have ammonia-to-methane ratios consistent with that of Jupiter within the uncertainties, suggesting that they have similar chemical abundances.

Using the same assumptions as in Solano et al. (2021) and the densities of objects of different spectral types in the solar neighborhood (Kirkpatrick et al. 2021), we provide here estimates of the numbers of dwarfs with spectral types from T0 to Y2 that could be detected by NISP spectroscopy with S/N > 10 in the Euclid wide survey (Fig. 5). We note that the number of T5 to T8 dwarfs spectroscopically characterized by NISP is expected to be ≥103, and this number rapidly decreases in the Y dwarf domain due to the intrinsic faintness at near-infrared wavelengths of those sources. The results of this work indicate that ammonia-to-methane ratios can be measured in late-T and Y dwarfs using low-resolution near-infrared spectra such as those that will be provided by Euclid/NISP. These ratios measured in a large sample with NISP will be useful for exploring the diversity of chemical and physical properties that are likely to exist among the coolest objects in the solar vicinity.

thumbnail Fig. 5.

Simulated number counts of T and Y dwarfs detectable (S/N ≥ 10) in the full Euclid wide survey with the spectroscopic mode of NISP for different values of the Galactic scale height of the BD population in the thin disk.


Acknowledgments

ELM acknowledges support from the Agencia Estatal de Investigación del Ministerio de Ciencia e Innovación (AEI-MCINN) under grant PID2019-109522GB-C53. J.-Y. Zhang acknowledges a summer grant from the Instituto de Astrofisica de Canarias. ELM thanks Mike Cushing for sending the HST infrared spectrum published in Cushing et al. (2021). This research has made use of the Simbad database, operated at the centre de Données Astronomiques de Strasbourg (CDS), and of NASA’s Astrophysics Data System Bibliographic Services (ADS).

References

  1. Aumann, H., Gillespie, C., Jr., & Low, F. 1969, ApJ, 157, L69 [Google Scholar]
  2. Bailey, J., & Kedziora-Chudczer, L. 2012, MNRAS, 419, 1913 [Google Scholar]
  3. Bochanski, J. J., Burgasser, A. J., Simcoe, R. A., & West, A. A. 2011, AJ, 142, 169 [Google Scholar]
  4. Burgasser, A. J., Kirkpatrick, J. D., Brown, M. E., et al. 2002, ApJ, 564, 421 [Google Scholar]
  5. Burgasser, A. J., Gelino, C. R., Cushing, M. C., & Kirkpatrick, J. D. 2012, ApJ, 745, 26 [Google Scholar]
  6. Canty, J. I., Lucas, P. W., Yurchenko, S. N., et al. 2015, MNRAS, 450, 454 [Google Scholar]
  7. Cushing, M. C., Kirkpatrick, J. D., Gelino, C. R., et al. 2011, ApJ, 743, 50 [Google Scholar]
  8. Cushing, M. C., Kirkpatrick, J. D., Gelino, C. R., et al. 2014, AJ, 147, 113 [Google Scholar]
  9. Cushing, M. C., Schneider, A. C., Kirkpatrick, J. D., et al. 2021, ApJ, 920, 20 [Google Scholar]
  10. Delorme, P., Delfosse, X., Albert, L., et al. 2008, A&A, 482, 961 [Google Scholar]
  11. Dupuy, T. J., & Liu, M. C. 2017, ApJS, 231, 15 [Google Scholar]
  12. Gordon, I. E., Rothman, L. S., Hill, C., et al. 2017, J. Quant. Spectr. Rad. Transf., 203, 3 [Google Scholar]
  13. Kirkpatrick, J. D., Marocco, F., Caselden, D., et al. 2021, ApJ, 915, L6 [Google Scholar]
  14. Leggett, S. K., Saumon, D., Marley, M. S., et al. 2007, ApJ, 655, 1079 [Google Scholar]
  15. Leggett, S. K., Morley, C. V., Marley, M. S., et al. 2013, ApJ, 763, 130 [Google Scholar]
  16. Luhman, K. L., & Esplin, T. L. 2016, AJ, 152, 78 [Google Scholar]
  17. McLean, I. S., McGovern, M. R., Burgasser, A. J., et al. 2003, ApJ, 596, 561 [Google Scholar]
  18. Owen, T., McKellar, A., Encrenaz, T., et al. 1977, A&A, 54, 291 [Google Scholar]
  19. Saumon, D., Geballe, T. R., Leggett, S. K., et al. 2000, ApJ, 541, 374 [Google Scholar]
  20. Schneider, A. C., Cushing, M. C., Kirkpatrick, J. D., et al. 2015, ApJ, 804, 92 [Google Scholar]
  21. Schneider, A. C., Greco, J., Cushing, M. C., et al. 2016, ApJ, 817, 112 [Google Scholar]
  22. Solano, E., Gálvez-Ortiz, M. C., Martín, E. L., et al. 2021, MNRAS, 501, 281 [Google Scholar]
  23. Taylor, F. W., Atreya, S. K., Encrenaz, T., et al. 2007, The Composition of the Atmosphere of Jupiter, 59 [Google Scholar]
  24. Vacca, W. D., Cushing, M. C., & Rayner, J. T. 2003, PASP, 115, 389 [Google Scholar]
  25. Valdivielso, L., Esparza, P., Martín, E. L., Maukonen, D., & Peale, R. E. 2010, ApJ, 715, 1366 [Google Scholar]
  26. Warren, S. J., Mortlock, D. J., Leggett, S. K., et al. 2007, MNRAS, 375, 213 [Google Scholar]
  27. Wong, A., Bernath, P. F., Rey, M., Nikitin, A. V., & Tyuterev, V. G. 2019, ApJS, 240, 4 [Google Scholar]
  28. Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868 [Google Scholar]
  29. Zahnle, K. J., & Marley, M. S. 2014, ApJ, 797, 41 [Google Scholar]

All Tables

Table 1.

Adopted Teff values and uncertainties for the different spectral types and giant gas planets considered in this work.

All Figures

thumbnail Fig. 1.

Our laboratory spectrum for ammonia and methane compared to simulated spectra of the same molecules using HITRAN for temperatures of 500 K and 300 K. The wavelength regions proposed for the new NH3-H index are marked with color bands: green for the numerator, red for the denominator.

In the text
thumbnail Fig. 2.

Redefined NH3-H index using our laboratory spectra. It has a higher sensitivity to spectral class (larger amplitude of variation) than the original definition by Delorme et al. (2008) based on HITRAN data. Third-order polynomial fits to the relationship between the indices and the spectral types are denoted with dashed lines of different colors, as labeled.

In the text
thumbnail Fig. 3.

Our laboratory spectra for a combination of ammonia and methane (blue) and for pure methane (red), compared to HST spectra of a T8 dwarf and two Y dwarfs, and IRTF spectra of two giant gas planets.

In the text
thumbnail Fig. 4.

Ammonia-to-methane ratios derived for late-T dwarfs and Y dwarfs with a 95% confidence level from NIRSPEC data (blue) and HST data (red) compared with those of Jupiter and Saturn from IRTF data (red) as functions of spectral type (upper panel) and Teff (lower panel). The Teff of T dwarfs, Y dwarfs, Jupiter, and Saturn adopted in this work are listed in Table 1. The dashed pink curves are the third-degree polynomial fittings for all objects (except Saturn) against spectral type and Teff, and the dashed black line is a linear fitting just for objects from T8 to Y2. In the lower panel, the results are compared with the scaled theoretical ratios of molecular column densities estimated from Zahnle & Marley (2014), which are shown in light blue.

In the text
thumbnail Fig. 5.

Simulated number counts of T and Y dwarfs detectable (S/N ≥ 10) in the full Euclid wide survey with the spectroscopic mode of NISP for different values of the Galactic scale height of the BD population in the thin disk.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.