The spectrum of ammonia near 0.793 $\mathrm{\mu }$m

Highlights

• Two ammonia spectra are recorded in the 11000 – 14500 cm-1 region.

• 278 lines are assigned using combination differences and variational nuclear motion calculations.

• 119 new energy levels belonging to 16 vibrational states are determined.

Abstract

Two sets of NH${}_3$ absorption spectra covering the 0.793 $\mathrm{\mu }$m region are recorded using two Bruker IFS 125 HR Fourier transform spectrometers. Three unapodized absorption spectra are recorded in Brussels over the range $11000-14500$ cm${}^{-1}$ and the positions and intensities of 1114 ammonia lines observed in the $12491-12810$ cm${}^{-1}$ region are measured using a multi-spectrum least squares fitting algorithm. 367 additional lines are identified in an ammonia absorption spectrum recorded in two steps at the J. Heyrovsky Institute of Physical Chemistry in Prague, using two different interference filters covering the $12000-12500$ and $12400-13000$ cm${}^{-1}$ ranges. The 1481 measured ammonia lines are analyzed using an empirical line list computed using variational nuclear motion calculations and ground state combination differences. Transitions are assigned to vibrational states with $4{\nu }_{NH}$ stretching excitation $(v_1+v_3=4)$. 278 out of the 1481 measured lines are assigned to 300 transitions and 119 upper state energy levels are derived from the frequencies of the assigned transitions.

Introduction

Complete characterization of ammonia molecular spectra in the microwave, infrared and optical ranges represents one of the major aims of both fundamental as well as applied high resolution molecular spectroscopy. In this field, the ammonia molecule has become a totemic system. Serendipitous observation of ammonia inversion-rotation can be found at the very beginning of this discipline [1]. The umbrella motion of ammonia is a textbook example of large amplitude motion in a molecule.

Ammonia is hazardous chemical, highly toxic for aquatic organisms, and its ever-increasing release into Earth’s atmosphere has undesirable consequences [2]; monitoring its presence in the atmosphere and a detailed understanding of the nitrogen cycle is therefore a particularly important scientific objective. Remote sensing of spatially resolved atmospheric concentrations of ammonia requires reliable and extensive spectroscopic datasets and their deficiencies remain a significant source of error [3]. Ammonia is the second biggest synthetic chemical product [4], [5]. NH${}_3$ may be a biosignature gas in H${}_2$-dominated exoplanetary atmospheres [6] or even a candidate for evidence of industrial civilization due to its relation with industry and agriculture.

Many areas in astronomy require spectroscopic data: ammonia is thought to be the key spectroscopic signature of the coldest failed stars, so-called brown dwarfs [7], [8], and is probably also prominent in the atmospheres of gas giant planets both in the solar system [9] and those of other stars [10]. Indeed, Fortney et al. [11] recently suggested that hotter gas giants should show pronounced ammonia features, emphasizing the importance of characterizing highly excited vibrational states of ammonia. All these applications require accurate spectroscopic data over extended frequency and temperatures ranges. This information is also required for the analysis and assignment of hot laboratory spectra [12], [13], [14], [15], [16].

However, even in the relatively recent editions of HITRAN, HITRAN2008 [17], ammonia data were presented only over the limited wavenumber range below 5000 cm${}^{-1}$. The reason for this limitation was mostly connected with the theoretical problem of accurately modelling the near IR and visible spectrum of NH${}_3$ which meant that spectra in these regions remained unassigned and often unanalyzed. Indeed, observations of ammonia spectra in the region around 1 $\mathrm{\mu }$m were made at Kitt Peak in the early 1980s (C. de Bergh, unpublished). Analysis of these spectra had to wait for improvements in theory, and in particular the generation of variational line lists. These spectra have only recently been assigned [18], [19], [20], [21]. The analogous situation arose with the spectrum of water in sunspots where detailed spectra were recorded [22] but spectral analysis [23], [24], [25] had to await variational calculations.

Returning to ammonia, recent experimental studies of the infrared spectrum include ones performed at JPL [26], [27], Laboratoire de Physique Moléculaire [28], Lille-Bratislava [29], Bruxelles [30], [31] as well as on Kitt Peak itself [32]. The latest edition of HITRAN, HITRAN2016 [33] includes ammonia spectra up to 10 000 cm${}^{-1}$. The need to improve the representation of the data in ammonia is illustrated by a recent study of the visible spectrum of ammonia on Jupiter [9] which showed that the CoYuTe variational line list [34] gave a good representation of the shape of the observed features but showed a shift in wavelength which can be attributed to the lack of experimental energy levels to which the line list could be tuned. This behavior is also found in our analysis below. A review of experimental spectroscopic studies on ${}^{14}$NH${}_3$ up to 7500 cm${}^{-1}$ is given as part of MARVEL (Measured Active Rotation-Vibrational Energy Levels) studies of ammonia energy levels [35], [36]. At higher wavenumbers double resonance studies performed in the 1980s by Coy and Lehmann [37], [38], [39] probed ammonia levels in the 15 000 – 18 000 cm${}^{-1}$ region as did dye laser experiments by Kuga et al. [40], and the work of Giver et al. [41]. Recent work has actually managed to assign portions of these spectra up to 18 000 cm${}^{-1}$ [20]. However, there remains a conspicuous gap in the region around 12 000 cm${}^{-1}$ due, in this case, to the absence of high resolution laboratory measurements, not theory. In this paper, we report new spectra which span this gap.

This paper is organized as follows. Section 2 gives experimental details for both Brussels and Prague observations of ammonia in the 12 000 cm${}^{-1}$ spectral region. Section 3 presents the spectral analysis, the results of line assignment and the derived energy. Section 4 concludes this paper.