Abstract
Miniaturizing and integrating atomic vapor cells is widely investigated for the purposes of fundamental measurements and technological applications such as quantum sensing. Extending such platforms to the realm of molecular physics is a fascinating prospect that paves the way for compact frequency metrology as well as for exploring light-matter interactions with complex quantum objects. Here, we perform molecular rovibrational spectroscopy in a thin-cell of micrometric thickness, comparable to excitation wavelengths. We operate the cell in two distinct regions of the electromagnetic spectrum, probing ν1 + ν3 resonances of acetylene at 1.530 µm, within the telecommunications wavelength range, as well as the ν3 and ν2 resonances of SF6 and NH3 respectively, in the mid-infrared fingerprint region around 10.55 µm. Thin-cell confinement allows linear sub-Doppler transmission spectroscopy due to the coherent Dicke narrowing effect, here demonstrated for molecular rovibrations. Our experiment can find applications extending to the fields of compact molecular frequency references, atmospheric physics or fundamental precision measurements.
Similar content being viewed by others
Introduction
Interfacing atomic and molecular gases with compact photonic platforms is a long-standing goal promising for quantum sensing1,2,3,4,5, optical image processing6, gas lasers and super-continuum sources7,8, as well as time keeping1 and frequency referencing applications9. Beyond their technological interest, compact hybrid platforms have an impact in fundamental studies of quantum electrodynamics in the presence of surfaces10,11,12.
In the realm of atomic gases compact integrated atomic cells of mesoscopic size have been realized by microfabrication techniques13 that find applications as portable atomic clocks1 or magnetometers2. Beyond these works, subwavelength confinement in the micro and nanoregime has been studied in thin-cells10,14,15, initially explored by Romer and Dicke16 in the microwave regime. Such cells are now used for studying Casimir-Polder interactions10,17, as well as cooperative18 and collective quantum effects19, paving the way for a new generation of room temperature quantum technology devices. Finally, extreme 3D subwavelength miniaturization has been recently achieved by confining atoms in opals20 or in nanostructured vapor cells21.
In contrast to atoms, molecules have a plethora of rovibrations throughout the spectrum offering the possibility of ultra-narrow spectroscopy, with a linewidth essentially proportional to gas pressure. For this reason, molecular cells based on photonic platforms are expected to find important applications in frequency referencing, especially in the telecommunications spectral window, via nonlinear spectroscopy of weak, hard-to-saturate, acetylene inter-combination lines22,23,24. Hollow core fibers filled with molecular gas9,23,24,25, tapered fiber devices22 or even chip-based waveguides ‘cladded’ with a molecular gas26 (inspired by their atomic analogues27) have been proposed for this purpose. These devices confine molecules in two dimensions, but have a macroscopic interaction length.
Extending the sub-wavelength confinement studies performed on atomic vapors over the last 20 years to molecular gases has remained up to now very challenging. This is mainly because molecular rovibrations have small transition probabilities. Moreover, the strongest rovibrations fall in the mid-infrared, the molecular fingerprint region, where easy-to-use laser sources were not accessible before the advent of Quantum Cascade Lasers (QCL). Experiments have probed confined molecular gases28,29 but this was achieved for high molecular densities compromising the frequency resolution28,29. So far, the only high-resolution (sub-Doppler) probing of molecular gases close to surfaces was achieved by selective reflection spectroscopy, where the effective optical confinement depends on the excitation wavelength30. These experiments could push Casimir-Polder studies further, unraveling fundamental effects linked to the complex molecular geometry31,32 that are inaccessible with atoms.
Here, we probe low-pressure molecular gases confined in a thin-cell of micrometric thickness comparable to the excitation wavelength. We perform rovibrational spectroscopy of acetylene at 1.53 µm, as well as of SF6 and NH3 at 10.55 µm. For these wavelengths, the linear, low-power transmission through our cell presents sub-Doppler features, due to the coherent Dicke narrowing14,16,33, with observed spectra that are very well interpreted by our theoretical models. The combination of linearity and high-resolution offers an original method of sub-Doppler molecular spectroscopy with significant advantages for enriching molecular databases and makes thin-cells attractive candidates for the fabrication of compact frequency references. We also discuss the potential of thin-cells for fundamental physics experiments with molecules.
Results
Coherent Dicke narrowing and cell thickness
Thin-cell spectroscopy under normal incidence has been studied extensively with dilute atomic vapors14,15,33,34,35,36,37,38. In thin-cells with thickness smaller than half the excitation wavelength (λ/2) the contribution of fast molecules (in the direction of the beam) is strongly inhibited due to transient broadening resulting from hard collisions with the cell walls. This leads to spectroscopic signals narrower than the Doppler profile. Nevertheless, linear thin-cell transmission spectroscopy also presents sub-Doppler features for larger cell thicknesses (L)14,33. These features, more pronounced when L is an odd multiple of λ/2, are due to the additive contribution of all molecular velocities at δ = 0, where δ is the laser frequency detuning14,34,38. The collapse and revival of the coherent Dicke narrowing, experimentally demonstrated in14, is illustrated in Fig. 1a where the predicted linear thin-cell transmission spectra (T) is shown for different cell thicknesses. In Fig. 1b we show the frequency modulated (FM) thin-cell transmission signals for L = λ/2 and 7λ/2, relevant for this work, for which the frequency resolution is mostly governed by the homogeneous transition linewidth (Γ).
Thin-cell fabrication
The thin-cell fabricated for this experiment is shown in Fig. 2. The cell consists of two ZnSe windows separated by an annular spacer, made from commercial gold-foil of nominal 5 µm thickness. The windows are held together with screws (mechanical pressure). The gold spacer determines the cell thickness but also effectively seals the cell, acting like a vacuum flange due to strain hardening. A ~ 5 mm diameter hole is drilled into one of the windows allowing connection to a vacuum system via a metallic tube and a KF flange (Fig. 2). The above fabrication method does not require processing or optical contact of the windows, therefore providing great flexibility in the choice of dielectric. Here, we fabricated a uniform thickness ZnSe cell, which is a fragile material but has the advantage of being transparent throughout the near and mid-infrared.
After pumping, the cell thickness is measured by Fabry-Perot interferometry (see Methods) and ranges between 5.27–5.5 µm (Fig. 2c). In the most used area of the cell, the thickness is 5.35 µm ± 0.02 µm, a value that is close to λ/2 for the mid-infrared strong rovibrations of SF6 and NH3 molecules at 10.55 µm and to 7λ/2 for the C2H2 overtones between 1.51–1.54 µm.
Acetylene spectroscopy at telecommunication wavelengths
To demonstrate the potential of our device, we first probed the ν1 + ν3 P(9) transition of acetylene at 1.530 µm, which is the strongest acetylene rovibration in this wavelength range accessible with our laser system. Our thin-cell is typically filled with acetylene pressures ranging from 0.65–3 Torr, for which the homogeneous linewidth (Γ), determined by molecular collisions (see Table 1) remains significantly smaller than the Doppler width (ΔD ~ 285 MHz).
For our experiment, we use an extended cavity diode whose frequency is scanned by applying a voltage to the piezoelectric actuator of the grating (see Methods). The FM amplitude and frequency are M = 5 MHz and fFM = 1 kHz respectively. One part of the laser beam, with a power of ~2-3 mW, simply crosses the thin-cell at normal incidence and is subsequently detected by a Ge photodiode. We refer to the demodulated signal after lock-in detection as TFM. An auxiliary saturated absorption experiment in a macroscopic (about 0.5 m length) cell provides a conventional narrow frequency reference in the volume. For this purpose, the laser beam is amplified by an Erbium Doped Fibre Amplifier (EDFA), providing ~200 mW with a beam size of ~1 mm, to saturate the weak acetylene rovibrations (see Methods).
Figure 3a shows thin-cell FM-transmission (TFM) spectra for three acetylene pressures normalized by the transmission signal through the empty cell (To). A sub-Doppler structure remains visible for all curves. The spectra presented in Fig. 3, result from an overall averaging of ~200 individual ~3 min scans. We have succeeded in eliminating a parasitic signal baseline below the ~10−8 level (smaller than statistical uncertainties) by subtracting consecutive scans with and without molecules (molecular pressure on/off modulation technique)30. For this, we used a system of electronically controlled valves that allow pumping and refilling of the cell at time scales on the order of 10–20 s (see Methods). The statistical noise level of this experiment is on the order of ~2 × 10−8, dominated by electronic detection noise. A larger signal-to-noise ratio could be achieved by using more laser power (while avoiding saturation effects) and by improvements on the set-up.
In Fig. 3a we plot the theoretically predicted spectra, fitted to the experimental data using the pressure broadened linewidth as a free parameter. Additionally, the theoretical curves are adjusted for amplitude and a small offset. Our fits allow us to extract the pressure broadening of the P(9) acetylene line which is measured to be 10.7 ± 1.8 MHz/Torr. This is consistent with values reported in the literature22,23,39,40 (see also Methods) that can nevertheless vary between 8–13 MHz/Torr depending on the experiment.
The theoretical model assumes that collisions with the surface destroy laser-molecule coherence and the velocity of desorbed molecules follow a Maxwell-Boltzmann distribution. This allows us to calculate the transient molecular response for ballistic molecular trajectory from one wall to the other, with intermolecular collisions described using a pressure-dependent homogeneous linewidth. In addition to transient effects theoretical calculations account for Fabry-Perot cavity effects41 and distortions due to FM modulation, calculated without any simplifying approximations30,42. The effects of laser linewidth (below 1 MHz) and of molecule-surface interactions can here be safely ignored30. Our models accurately reproduce the experimental spectra at the line center (sub-Doppler contribution) but also at the Doppler broadened wings.
Thin-cell theory is further tested in Fig. 3b, where the signal transmission, measured at different thicknesses (see Fig. 2c) for a pressure of 1.4 Torr, is almost perfectly superposed to the theoretically predicted spectra. The asymmetric lineshapes observed when the cell thickness deviates from 7λ/2 are due to the mixing of the symmetric forward molecular response (transmission) with the asymmetric backward response (reflection) in the Fabry-Perot cavity of the cell. In Fig. 3b both linewidth and amplitude of the theoretical curves are fixed to the values extracted from Fig. 3a (1.4 Torr red curve) and only a small offset (~10−8) is independently adjusted. The good parallelism between the cell windows, allows the beam to explore a uniform thickness and contributes to the remarkable agreement between experiment and theory.
Mid-infrared rovibrational spectroscopy
We also studied the mid-infrared region of the spectrum where most molecular rovibrations can be found. In particular, the fingerprint region around 10 µm is of interest for precision measurements43,44, and metrology45, but also in the field of atmospheric physics and gas tracing46,47,48. We perform spectroscopy of v3 and v2 rovibrations of SF6 and NH3 molecules respectively (see Table 1), at 10.55 µm, both accessible with our laser set-up described in30.
The QCL source used for this experiments emits ~5 mW of optical power at its output after optical isolation. The principles of our spectroscopic detection remain the same as those described for the C2H2 experiments. Here the FM modulation is applied on the laser current with a frequency fFM = 10 kHz and an amplitude of M = 0.33 MHz. A saturated absorption spectrum is also simultaneously recorded for frequency calibration. QCL technology has not yet attained the maturity of telecom wavelength lasers. We have, therefore, developed a number of techniques to improve the QCL frequency scan and render it compatible with the needs of high-resolution spectroscopy30.
Sulfur hexafluoride (SF6), is a heavy spherical top molecule of octahedral geometry that presents a complex spectroscopic landscape. SF6 is in the list of greenhouse gases and determination of both frequency and amplitude of its absorptions is required in order to measure the evolution of its atmospheric concentration48. The large transition probability of the v2 rovibrations of SF6, allows us to use lower pressures, achieving higher frequency resolution compared to acetylene spectroscopy. Ultimately the resolution of our mid-infrared experiments is limited by the QCL linewidth (~0.7 MHz FWHM).
Figure 4a shows the SF6 thin-cell FM-transmission spectrum centered on the strong Q62E, Q62A2 and Q62F2 degenerate triplet. The adjacent, smaller transitions, probably corresponding to SF6 hot-bands, are not reported in the HITRAN database49 but their frequency positions have been pinpointed in previous experiments using saturated absorption spectroscopy and a high precision wavemeter30. The curve of Fig. 4a results from averaging about 50 individual ~2 min scans, using the aforementioned molecular pressure on/off modulation technique. The noise level is here ~4 × 10−8, slightly higher than the shot noise limit of ~2 × 10−8. The theoretical model, used to interpret the mid-infrared spectra also accounts for the influence of laser linewidth, which is comparable to the FM amplitude and the homogeneous pressure broadened linewidth. The fit allows us to extract the collisional broadening (Γ) and shift, as well as the relative transition amplitudes of the small and unidentified lines compared to the main strong triplet. Experiments at a pressure range of 0.04–0.66mTorr give us a collisional broadening of 4.4 ± 0.4 MHz/Torr, while collisional shifts remain negligible.
Although these transition amplitudes were measured by selective reflection30, thin-cells (here operated at λ/2 thickness) are simpler platforms for molecular spectroscopy providing a larger signal-to-noise ratio (see Methods). This experiment, allows us to measure amplitudes with an uncertainty below 5% with a better signal-to-noise ratio compared to30 by roughly an order of magnitude. Our set-up could be used to access a wide range of v3 rovibrations of SF6 between 28.35-28.5 THz thus providing spectroscopic information that could be used to enrich molecular databases.
The ammonia (NH3) molecule, a symmetric top of tetrahedral geometry, has a simple rovibrational spectrum with well-isolated transitions. For this reason, it is better suited for metrology or fundamental physics experiments. NH3 has nevertheless a hyperfine structure with splittings as large as ~1 MHz (depending on the rovibrational level), that need to be taken into consideration when interpreting spectral lineshapes. Here, we probe the saP(1,0) transition, from the ground state to the first v2 vibration, whose hyperfine structure is shown in the inset of Fig. 4b.
The thin-cell FM-transmission spectrum of NH3 is shown in Fig. 4b for pressures of 31mTorr and 94mTorr. Further experiments (not shown here) were performed for pressures up to 300mTorr providing the pressure broadening of 18 ± 1.7 MHz/Torr (Table 1). The scans are recorded using the same parameters and conditions as in the case of SF6 (Fig. 4a). As pressure decreases, the hyperfine structure of the saP(1,0) rovibration becomes more resolved and apparent on the spectra. Reducing the pressure below 30mTorr does not significantly improve the resolution (limited by laser linewidth) but simply reduces the signal amplitude. The solid lines represent the theoretical fits, where both positions and relative amplitudes of the hyperfine transitions are fixed to their theoretically expected values50. Lineshape analysis allows us to measure a collisional shift of 3.7 ± 1 MHz/Torr for the saP(1,0) ν2 transition of NH3. We observe no differential pressure shift between the hyperfine components of the examined transition, while at low pressures the frequency spacing between the hyperfine components is consistent, to within 0.1 MHz, with previously reported values (shown on the inset of Fig. 4b).
Discussion
We probed molecular gases in a thin cell of micrometric thickness corresponding to λ/2 for SF6 and NH3 rovibrations at 10.55 µm and 7λ/2 for acetylene transitions at 1.53 µm. The coherent Dicke narrowing16, demonstrated for rovibrational transitions, allows us to obtain high-resolution, sub-Doppler transmission signals without resorting to nonlinear spectroscopic schemes. The thin-cell platforms presented here can be used to probe many molecular transitions spanning through a wide range of the electromagnetic spectrum.
One of our major results is the measurement of transmission signal amplitudes and lineshapes of subwavelength confined molecules that are very well reproduced by theory. This remarkable observation suggests that our experiment is compatible with the assumptions included in the models. The most important question concerns molecule-surface collisions and in particular their effect on the molecule-laser interaction28,51 as well as the velocity distribution52,53,54 and the partition function (redistribution in rotational states) of the molecules departing from the surface (desorbed molecules)55. The common assumption that collisions with the surface destroy the laser-particle coherence is justified28, but the case of rovibrational spectroscopy needs further investigations since molecule–surface potentials can be very similar between the two probed states51. In contrast to atoms, molecules offer the possibility of ultra-high resolution spectroscopy thus allowing for more strenuous tests of the Maxwell-Boltzmann distribution. Additionally, studies of the signal amplitude of different rovibrations inside thin-cells can offer information on the molecular partition function of confined gases. Molecular thin-cells are therefore excellent testbeds for exploring the thermodynamics of confined gases.
A fascinating perspective is to strongly confine molecular gases in the nanometric regime10 and perform spectroscopy of the Casimir-Polder interaction with molecules. Molecule-surface interactions present a multidisciplinary interest (from QED theory to physical chemistry applications) but experimental investigations remain scarce56,57. The fabrication process presented here is particularly suited for Casimir-Polder studies, as it offers high flexibility in the use of dielectric windows. In contrast to atoms, molecules could be probed in thin-cells operated below room temperatures, which can be of interest for studies of thermal Casimir-Polder effects55,58,59. Molecular cells can also be operated under extreme molecular densities, which could prove an advantage for probing cooperative18 or collective effects such as super-radiance60.
Our experiment provides an alternative technique for rovibrational molecular spectroscopy that combines linearity and crossover-transition-free sub-Doppler resolution. This allows simultaneous measurement of both transition amplitudes and frequencies with a significantly better signal amplitude than selective reflection. Thin-cell spectroscopy can therefore contribute to the compilation of molecular databases, essential for retrieving atmospheric concentrations and interpreting astrophysical spectra, complementing traditional methods such as Fourier transform infrared (FTIR) spectroscopy or Doppler broadened absorption spectroscopy that suffers from low-frequency resolution, or saturated absorption whose inherent nonlinearity makes difficult the extraction of transition amplitudes. The advantages of using thin-cells are more evident for heavy atmospheric species such as SF6 with low-lying vibrational modes, which exhibit dense rotational structures and many hot-bands, impossible to resolve by FTIR or Doppler spectroscopy, and for which the molecular databases remain largely incomplete61.
The subject of SF6 hot-bands is an open question in molecular spectroscopy, many of which still remain largely neglected in the HITRAN database even after a recent update49. Including thin-cell spectroscopic data in a global analysis of SF6 spectra (such as the one performed in61) could potentially offer unprecedented information on hot-band transition amplitudes and positions for SF6 and other heavy atmospheric molecules (ClONO2, CF4) in the future. Additionally, thin-cell spectroscopy can constitute a stringent test for validating the most advanced ab initio quantum mechanical calculations of the complete band structure of SF662,63 and other complex molecules.
Dicke-narrowed molecular thin-cell spectroscopy is an important development in the field of compact frequency references at telecommunication wavelengths allowing high frequency-resolution with a simple, one-beam, versatile, low-power set-up, in contrast to saturation spectroscopy. Gaining a factor of 100 in signal amplitude could allow stabilization of a telecommunications laser on the thin-cell transmission, whose stability can subsequently be compared against a voluminous frequency reference obtained in a macroscopic cell64. Realizing a resonant high-finesse Fabry-Perot cavity could be a viable route towards signal amplification by a factor roughly proportional to the finesse (see Methods). This could be achieved by depositing dielectric mirrors of modest reflectivity (~98% for an amplification of 100, with a cavity finesse on the order of 150) on the cell windows. Alternatively, fabricating a stack of cells (successively piling windows separated by spacers) or using a multi-pass technique could also be explored. Finally, micro-fabricated ultra-compact thin-cells, as reported in13,65 for atomic vapor cells, could also be an interesting prospect for acetylene spectroscopy.
Methods
Gas cells, thickness and pressure measurements
The micrometric thin-cell consists of 2 ZnSe windows, with a wedge of about 2o, separated by an annular spacer. For the purposes of this experiment, we used a Goodfellow gold-foil as a spacer with nominal thickness of 5 µm but we have also tested sputter coating and evaporation techniques (using a mask) that allow us to achieve a very good control of the spacer thickness. When applying mechanical pressure to seal the cell, the two external interfaces are aligned to be parallel to each other thus avoiding stress that could fracture the windows. The internal faces of the windows (the actual thin-cell walls) are uncoated, while the external interfaces are anti-reflection (AR) coated to avoid, as much as possible, parasitic reflections. The coating is optimal at 10.6 µm (NH3 and SF6 experiments) giving a reflectivity of about 1%, but at 1.53 µm (C2H2 experiments) its reflectivity is about 10%. The reflectivity of an uncoated ZnSe surface is ~17% and 18% for 10.6 µm and 1.53 µm respectively, creating a small finesse optical cavity. The curvature of the internal window interfaces (and possible deposited contaminants) can introduce variations (and uncertainties) to the cell thickness.
The thickness of the thin cells has been mapped by Fabry-Perot interferometry with an uncertainty of a few nm using three different lasers: a He-Ne at 633 nm, a diode laser at 852 nm, and the extended cavity laser emitting from 1519 nm to 1540 nm10. The three beams are aligned and are used to measure the thin cell reflectivity at different wavelengths and different spots of the cell. The thickness gradient is here small, allowing us to use a relatively big beam size of 1 mm diameter.
The macroscopic cells used to obtain an auxiliary saturated absorption reference are metal tubes sealed by windows glued at the ends. For the C2H2 experiment, we use a 45 cm long cell sealed by glass windows whereas for the SF6 and NH3 experiments we use a 15 cm long cell sealed by ZnSe windows (AR coated at both sides).
The thin-cell is connected to two motorized valves and a pressure gauge. We use a valve “Pfeiffer Vacuum GmbH RME series” to control the gas flow into the cell, and a valve “Pfeiffer Vacuum GmbH EVR 116” to control the connection to the pumping station. Both valves are actuated with an external computer-controlled voltage.
To determine the pressure inside the cells we use dual gauges (Pirani and cold cathode gauge) with a range extending from 10−6 mTorr up to almost atmospheric pressure. The Pirani gauge, which operates in the pressure range where our experiments are performed (≈20-4000 mTorr) for C2H2, SF6, or NH3 gases, was later calibrated by a capacitive gauge (readings independent of gas type). This eliminates an important source of systematic errors in the pressure measurement and gives us more confidence in the reported pressure broadenings and shifts. Gauge calibration can be a source of discrepancies between our measurements and the values of pressure broadening reported in the literature.
Frequency scan of the lasers
The extended cavity laser emitting around 1530 nm, used for C2H2 spectroscopy, is scanned by applying a voltage to the piezoelectric actuator of the grating. The laser detuning is calibrated with an uncertainty of ~1%, by using a BRISTOL 771B-MIR (1–12 µm) wavemeter and this measurement is subsequently corroborated by fitting the linear Doppler absorption spectrum of acetylene at room temperature with a Voigt profile. Saturated absorption in the volume gives an absolute molecular reference with which the frequency scans are referenced with an accuracy better than the MHz.
The frequency drift of the QCL laser emitting around 10.55 µm is significantly more important and the frequency scans are achieved by loosely frequency stabilizing the laser to a linear absorption profile while applying an offset voltage to the error signal. The technique is described in detail in previous works30.
A modulation on the laser frequency (\(f\)), is applied to both lasers, \(f(t)={f}_{{{{{{\rm{o}}}}}}}+M\sin (2\pi {f}_{{{{{{\rm{FM}}}}}}}t)\), where \({f}_{{{{{{\rm{o}}}}}}}\) is the central frequency, \(M\) is the modulation amplitude and \({f}_{{{{{{\rm{FM}}}}}}}\) is the modulation frequency. The frequency modulation (FM) is applied on the current of the 10.55 µm emitting QCL and on the piezoelectric actuator of the 1530 nm laser.
Saturated absorption spectroscopy
The C2H2 inter-combination lines in the telecommunications windows are relatively weak, with a dipole moment matrix element of µ = 3.6 × 10−32Cm22, in contrast to the much stronger NH3 and SF6 rovibrations with µ = 1.4 × 10−30Cm66 and µ = 0.8 × 10−30Cm67 respectively. Therefore, the saturation intensity Isat (Isat ∝ Γ2/ \({\mu }^{2}\), where Γ is the homogeneous linewidth) is significantly higher for C2H2 (Isat ~ 50 W/cm2) than for NH3 (Isat ~ 30 mW/cm2) or SF6 (Isat ~ 80 mW/cm2) rovibrations. Here the values of saturation intensity are indicatively given for a linewidth of 1 MHz.
In order to saturate the acetylene transitions, the laser beam is amplified with an EDFA which provides 23.5 dBm (~200 mW) at its output. In the experimental set-up we use one single beam (pump beam) which is retro-reflected and attenuated, before passing through the cell, to serve as the probe beam. The pump beam size is about 1 mm (pump intensity ~20 W/cm2).
In the case of NH3 and SF6, saturated absorption spectroscopy is performed on a 15 cm cell using a similar set-up (here the source is a QCL source). Here, the amplification step is not necessary and saturation can be observed with a maximum optical power of ~1.5 mW and a beam diameter of about 2 mm.
The saturated absorption spectra can also be demodulated at the FM frequency allowing us to observe the derivative of the saturated absorption signal and obtain a frequency reference.
Background elimination: molecular pressure on/off modulation technique
The molecular thin-cell transmission spectra are always ‘contaminated’ by a parasitic background resulting probably from the interference between the transmitted laser beam and parasitic reflections. Due to slow thermal fluctuations or small mechanical movements, the interferometric background slowly drifts, making its shape difficult to predict. To eliminate this parasitic signal, we have implemented the on/off pressure modulation technique mentioned in the main text and explained in30. Here, we recall the main steps of this technique.
-
1.
Pumping the cell: First, the thin-cell is pumped by opening the aperture of the motorized valve which connects it to the vacuum pump. The valve is closed after reaching a pressure Pvaccum of around 0.1mTorr and a frequency scan A (without molecules) is recorded.
-
2.
Filling the cell with molecular gas: By opening the aperture of the solenoid valve, the cell is filled with molecular gas up to a pressure Pgas. After reaching the desired pressure (Pgas) the valve is closed and a frequency scan B (with molecules) is recorded.
-
3.
Pumping the cell: The cell is pumped again and a frequency scan C (without molecules) is recorded.
The above steps are fully automated using computer control. Pumping and filling the cells takes 10–20 seconds, while scans A, B, and C are recorded within 120–180 seconds time depending on the required frequency resolution. After detection, the frequency scale of all scans is calibrated using the saturated absorption reference. The difference between scan B and the average of scans A and C is the ‘pure’ molecular thin-cell transmission. The cycle is then repeated N times (typically 20–200 times), allowing us to reduce the statistical noise typically below 10−7 (depending on the experiments and number of scans) and the residual interferometric background down to the 10−8 level for C2H2 experiments and down to the 10−7 level for NH3 and SF6 experiments.
Fitting
The theoretical thin-cell transmission spectra are calculated using the theory developed in41 ignoring Casimir-Polder interactions and assuming a Maxwell-Boltzmann distribution of molecular velocities inside the cell. After the theoretical ‘direct’ transmission curves have been calculated, we account for FM demodulation and the effects of the finite laser linewidth by convolving with an assumed Gaussian frequency distribution. The FWHM of the Gaussian function, that represents the laser linewidth, is fixed by independent measurements. For the mid-infrared experiments, the laser linewidth was experimentally estimated to be ~0.7 MHz (this is slightly higher than the value 0.6 MHz reported in ref. 30, probably because of small differences in the laser frequency stabilization loop). For the near-infrared experiments we have experimentally verified that the laser linewidth is well below 0.5 MHz and it can therefore safely be ignored in the fits.
To fit the spectra, we first produce a theoretical spectrum using an initial estimate of the value of Γ (collisional broadening). The curve is then adjusted for amplitude, offset and a small frequency shift to fit the experimental spectrum. We iterate the process for different values of Γ, until the best fit (the value that minimizes the least square difference) is identified. This gives us the value of collisional broadening, transition amplitude, and collisional shift.
Thin-cell transmission and selective reflection spectroscopy
For simplicity, we discuss a symmetric cell made from two interfaces with a reflection coefficient \(r\). In the infinite Doppler approximation \({\Delta }_{{{{{{\rm{D}}}}}}}\,\gg \,\Gamma\) and assuming that \(M\,\ll \,\Gamma\), the normalized FM transmission \({T}_{{{{{{\rm{FM}}}}}}}/{T}_{o}\) through a cell of \(\frac{\lambda }{2}\) thickness is given by41:
where \(N\) is the molecular density, \(\mu\) the transition dipole moment matrix element, \({u}_{p}\) is the most probable velocity and \(A\) a constant.
The normalized FM selective reflection signal (frequency modulated resonant reflection from an infinitely long cell at normal incidence assuming an infinitely small but not strictly null absorption), calculated with the same formalism, is:
where \({S}_{{{{{{\rm{R}}}}}}}^{{{{{{\rm{o}}}}}}}={r}^{2}\) is the reflected signal in the absence of molecules.
To compare the normalized FM thin-cell transmission signal reported here (for a thin cell of λ/2) and the normalized FM selective reflection signal of 30, we use the reflection coefficient (\(r=0.42\)) of our ZnSe windows and account for the different modulation amplitudes: 0.33 MHz and 0.25 MHz for thin-cell and selective reflection spectroscopy respectively. We find that thin cell spectroscopy is expected to provide a larger amplitude by a factor of ~6.6 and a larger signal-to-noise ratio by ~11.9 (assuming the same incident power and a shot-noise limited experiment) compared to selective reflection. In our experiments, the observed amplitude ratio is slightly larger, roughly between 7 and 9, depending on the type of gas (SF6 or NH3) and the pressure. This small difference is probably due to the fact that in ref. 30 some mirrors were vibrated to reduce the parasitic background. This technique was later found to slightly degrade the amplitude and noise of our signals and was not used for thin cell spectroscopy.
Pressure broadening and shift
Following the above process for thin-cell experiments we have found that the collisional broadening is 10.7 ± 1.8 MHz/Torr, 4.4 ± 0.4 MHz/Torr and 18 ± 1.7 MHz/Torr and for the C2H2, SF6 and NH3 rovibrations respectively. The collisional shift (red line in Fig. 5c) is measured to be +3.7 ± 1 MHz/Torr in the case of NH3 but stays negligible in the case of SF6 and C2H2 ( < 0.2 MHz for the explored pressures). In Fig. 5 we show the homogeneous linewidth (Γ) and shift (δ), extracted from the fits of the thin-cell transmission lineshapes as a function of pressure. The error bars on Γ and δ are conservatively extracted from our lineshape analysis. The pressure was measured with Pirani gauges that were subsequently calibrated with a capacitive gauge, whose readings are independent of the nature of the gas. The error bars on the pressure readings are due to statistical fluctuations of ~5-10mTorr and systematic errors of ~3% due to the calibration process.
In these experiments, the pressure gauge is naturally placed outside the thin-cell (see Fig. 2a of the main text). Thus a pressure differential cannot be excluded (cell thickness is ~5 µm), contrary to selective reflection measurements30, performed with the same pressure gauge (and the same set-up) but in a macroscopic cell. Nevertheless, both techniques yield comparable pressure broadenings suggesting that gas pressure readings outside the cell also reflect the pressure inside the cell. Using selective reflection, we found 4 MHz/Torr and 21 MHz/Torr broadening for the same SF6 and NH3 rovibrations respectively with an error bar on the order of 10% (here we have corrected the pressure readings reported in30 assuming the same calibration factors).
In the case of C2H2 (no selective reflection measurements available) we have performed independent saturated absorption experiments for different gas pressures ranging from 20mTorr up to 500mTorr. For these dedicated measurements, we have separated the pump (~200 mW after amplification) and probe (~3 mW) beams and have applied an amplitude modulation on the pump, which allows us to eliminate the Doppler broadened background. At low pressures (below 100mTorr) the saturated absorption profile is well fitted by a Lorentzian function. The minimum achievable linewidth is ~2.5 MHz, probably limited by transit time broadening, small angle between pump and probe beams or laser linewidth. At higher pressures (above 100mTorr) the narrow saturated absorption (still interpreted by a Lorentzian profile) sits on a broader pedestal, probably due to velocity-changing collisions, that becomes more prominent as pressure increases. Eventually the narrow saturated absorption disappears for pressures above 500mTorr. The pressure broadening, extracted purely by analysis of the narrow component, is measured to be ~15 MHz/Torr slightly higher than the values obtained in the thin cell. The discrepancy between the two experiments can be due to velocity-changing collisions that can give rise to a nonlinear dependence of linewidth on pressure (see68 and references therein) and whose effects could be different between linear (thin-cell) and nonlinear (saturated absorption) spectroscopic schemes.
Data availability
The experimental data generated in this study are deposited in the zenodo database and are available under the accession code https://zenodo.org/records/10499906. Further data can be provided by the corresponding author upon request.
References
Knappe, S. et al. A microfabricated atomic clock. Appl. Phys. Lett. 85, 1460–1462 (2004).
Shah, V., Knappe, S., Schwindt, P. D. D. & Kitching, J. Subpicotesla atomic magnetometry with a microfabricated vapour cell. Nat. Photonics 1, 649–652 (2007).
Downes, L. A. et al. Full-Field Terahertz Imaging at Kilohertz Frame Rates Using Atomic Vapor. Phys. Rev. X 10, 011027 (2020).
Fan, H. et al. Effect of vapor-cell geometry on rydberg-atom-based measurements of radio-frequency electric fields. Phys. Rev. Appl. 4, 044015 (2015).
Sedlacek, J. A. et al. Microwave electrometry with Rydberg atoms in a vapour cell using bright atomic resonances. Nat. Phys. 8, 819–824 (2012).
Biaggio, I., Partanen, J. P., Ai, B., Knize, R. J. & Hellwarth, R. W. Optical image processing by an atomic vapour. Nature 371, 318–320 (1994).
Zhou, Z. et al. High-power tunable mid-infrared fiber gas laser source by acetylene-filled hollow-core fibers. Opt. Express 26, 19144 (2018).
Sollapur, R. et al. Resonance-enhanced multi-octave supercontinuum generation in antiresonant hollow-core fibers. Light Sci. Appl. 6, e17124–e17124 (2017).
Benabid, F., Couny, F., Knight, J. C., Birks, T. A. & Russell, P. S. J. Compact, stable and efficient all-fibre gas cells using hollow-core photonic crystal fibres. Nature 434, 488–491 (2005).
M. Fichet et al. Exploring the van der Waals atom-surface attraction in the nanometric range, Europhys. Lett. EPL, vol. 77, Art. no 5, mars https://doi.org/10.1209/0295-5075/77/54001 (2007).
D. Bloch & M. Ducloy, Atom-wall interaction, in Advances In Atomic, Molecular, and Optical Physics, Elsevier, 91‑154, https://doi.org/10.1016/S1049-250X(05)80008-4 (2005).
Kübler, H., Shaffer, J. P., Baluktsian, T., Löw, R. & Pfau, T. Coherent excitation of Rydberg atoms in micrometre-sized atomic vapour cells. Nat. Photonics 4, 112–116 (2010).
Liew, L.-A. et al. Microfabricated alkali atom vapor cells. Appl. Phys. Lett. 84, 2694–2696 (2004).
Dutier, G. et al. Collapse and revival of a Dicke-type coherent narrowing in a sub-micron thick vapor cell transmission spectroscopy. Europhys. Lett. EPL 63, 35–41 (2003).
Sarkisyan, D., Bloch, D., Papoyan, A. & Ducloy, M. Sub-doppler spectroscopy by sub-micron thin Cs vapour layer. Opt. Commun. 200, 201–208 (2001).
Romer, R. H. & Dicke, R. H. New technique for high-resolution microwave spectroscopy. Phys. Rev. 99, 532–536 (1955).
Peyrot, T. et al. Measurement of the atom-surface van der Waals interaction by transmission spectroscopy in a wedged nanocell. Phys. Rev. A 100, 022503 (2019).
Peyrot, T. et al. Collective lamb shift of a nanoscale atomic vapor layer within a sapphire cavity. Phys. Rev. Lett. 120, 243401 (2018).
Ripka, F., Kübler, H., Löw, R. & Pfau, T. A room-temperature single-photon source based on strongly interacting Rydberg atoms. Science 362, 446–449 (2018).
Ballin, P., Moufarej, E., Maurin, I., Laliotis, A. & Bloch, D. Three-dimensional confinement of vapor in nanostructures for sub-Doppler optical resolution. Appl. Phys. Lett. 102, 231115 (2013).
Cutler, T. F. et al. Nanostructured alkali-metal vapor cells. Phys. Rev. Appl. 14, 034054 (2020).
Takiguchi, M., Yoshikawa, Y., Yamamoto, T., Nakayama, K. & Kuga, T. Saturated absorption spectroscopy of acetylene molecules with an optical nanofiber. Opt. Lett. 36, 1254 (2011).
Knabe, K. et al. 10 kHz accuracy of an optical frequency reference based on C2H2-filled large-core kagome photonic crystal fibers. Opt. Express 17, 16017 (2009).
Triches, M. et al. Optical frequency standard using acetylene-filled hollow-core photonic crystal fibers. Opt. Express 23, 11227 (2015).
Billotte, T. et al. Contaminant-free end-capped and single-mode acetylene photonic microcell for sub-Doppler spectroscopy. Opt. Lett. 46, 456 (2021).
Zektzer, R. et al. A chip‐scale optical frequency reference for the telecommunication band based on acetylene. Laser Photonics Rev. 14, 1900414 (2020).
Yang, W. et al. Atomic spectroscopy on a chip. Nat. Photonics 1, 331–335 (2007).
Hartmann, J. M. et al. Infrared look at the spectral effects of submicron confinements of CO2 gas. Phys. Rev. A 93, 012516 (2016).
Svensson, T. & Shen, Z. Laser spectroscopy of gas confined in nanoporous materials. Appl. Phys. Lett. 96, 021107 (2010).
Lukusa Mudiayi, J. et al. Linear Probing of Molecules at Micrometric Distances from a Surface with Sub-Doppler Frequency Resolution. Phys. Rev. Lett. 127, 043201 (2021).
Butcher, D. T., Buhmann, S. Y. & Scheel, S. Casimir–Polder forces between chiral objects. New J. Phys. 14, 113013 (2012).
Thiyam, P. et al. Anisotropic contribution to the van der Waals and the Casimir-Polder energies for CO2 and CH4 molecules near surfaces and thin films. Phys. Rev. A 92, 052704 (2015).
Briaudeau, S., Saltiel, S., Nienhuis, G., Bloch, D. & Ducloy, M. Coherent Doppler narrowing in a thin vapor cell: Observation of the Dicke regime in the optical domain. Phys. Rev. A 57, R3169–R3172 (1998).
Sarkisyan, D. et al. Spectroscopy in an extremely thin vapor cell: Comparing the cell-length dependence in fluorescence and in absorption techniques. Phys. Rev. A 69, 065802 (2004).
Briaudeau, S., Bloch, D. & Ducloy, M. Detection of slow atoms in laser spectroscopy of a thin vapor film. Europhys. Lett. EPL 35, 337–342 (1996).
Ai, B., Glassner, D. S. & Knize, R. J. Enhancement of degenerate four-wave mixing by atom-wall collisions in atomic vapors. Phys. Rev. A 50, 3345–3348 (1994).
Vartanyan, T. A. & Lin, D. L. Enhanced selective reflection from a thin layer of a dilute gaseous medium. Phys. Rev. A 51, 1959–1964 (1995).
Hamdi, I. et al. Laser spectroscopy with nanometric gas cells: distance dependence of atom-surface interaction and collisions under confinement. Laser Phys. 15, 987–996 (2005).
Swann, W. C. & Gilbert, S. L. Pressure-induced shift and broadening of 1510–1540-nm acetylene wavelength calibration lines. J. Opt. Soc. Am. B 17, 1263 (2000).
Castrillo, A. et al. Optical determination of thermodynamic temperatures from a c2h2 line-doublet in the near infrared. Phys. Rev. Appl. 11, 064060 (2019).
Dutier, G., Saltiel, S., Bloch, D. & Ducloy, M. Revisiting optical spectroscopy in a thin vapor cell: mixing of reflection and transmission as a Fabry–Perot microcavity effect. J. Opt. Soc. Am. B 20, 793 (2003).
Ducloy, M. & Fichet, M. General theory of frequency modulated selective reflection. Influence of atom surface interactions. J. Phys. II 1, 1429–1446 (1991).
Ch. Daussy et al. Limit on the parity nonconserving energy difference between the enantiomers of a chiral molecule by laser spectroscopy. Phys. Rev. Lett. 83, 1554–1557 (1999).
Fiechter, M. R. et al. Toward detection of the molecular parity violation in chiral Ru(acac) 3 and Os(acac)3. J. Phys. Chem. Lett. 13, 10011–10017 (2022).
Mejri, S. et al. Measuring the Boltzmann constant by mid-infrared laser spectroscopy of ammonia. Metrologia 52, S314–S323 (2015).
Gabard, T. et al. Spectroscopic tools for remote sensing of greenhouse gases CH4, CF4 and SF6. Environ. Chem. Lett. 1, 86–91 (2003).
Zhou, M. et al. Ground-based FTIR retrievals of SF6 on Reunion Island. Atmospheric Meas. Tech. 11, 651–662 (2018).
Khalil, M. A. K. Non-CO2 greenhouse gases in the atmosphere. Annu. Rev. Energy Environ. 24, 645–661 (1999).
Gordon, I. E. et al. The HITRAN2020 molecular spectroscopic database. J. Quant. Spectrosc. Radiat. Transf. 277, 107949 (2022).
Urban, Š., Herlemont, F., Khelkhal, M., Fichoux, H. & Legrand, J. High-accuracy determination of the frequency and quadrupole structure of the sp(1, 0) and ar(0, 0) transitions to the ν2 state of 14NH3. J. Mol. Spectrosc. 200, 280–282 (2000).
Ducloy, M. Influence of atom-surface collisional processes in FM selective reflection spectroscopy. Opt. Commun. 99, 336–339 (1993).
Todorov, P. & Bloch, D. Testing the limits of the Maxwell distribution of velocities for atoms flying nearly parallel to the walls of a thin cell. J. Chem. Phys. 147, 194202 (2017).
Rabi, O. A., Amy-Klein, A., Saltiel, S. & Ducloy, M. Three-level non-linear selective reflection at a dielectric/cs vapour interface. , EPL 25, 579–585 (1994).
Christaller, F. et al. Transient density-induced dipolar interactions in a thin vapor cell. Phys. Rev. Lett. 128, 173401 (2022).
Laliotis, A., de Silans, T. P., Maurin, I., Ducloy, M. & Bloch, D. Casimir–Polder interactions in the presence of thermally excited surface modes. Nat. Commun. 5, 4364 (2014).
Brand et al. A Green’s function approach to modeling molecular diffraction in the limit of ultra‐thin gratings. Annalen der Physik 527, 580–591 (2021).
Shih, A., Raskin, D. & Kusch, P. Investigation of the interaction potential between a neutral molecule and a conducting surface. Phys. Rev. A 9, 652–662 (1974).
De Aquino Carvalho, J. C. et al. Spectrally sharp near-field thermal emission: revealing some disagreements between a casimir-polder sensor and predictions from far-field emittance. Phys. Rev. Lett. 131, 143801 (2023).
Obrecht, J. M. et al. Measurement of the temperature dependence of the casimir-polder force. Phys. Rev. Lett. 98, 063201 (2007).
Skribanowitz, N., Herman, I. P., MacGillivray, J. C. & Feld, M. S. Observation of Dicke superradiance in optically pumped HF gas. Phys. Rev. Lett. 30, 309 (1973).
Faye, M., Boudon, V., Loëte, M., Roy, P. & Manceron, L. The high overtone and combination levels of SF6 revisited at Doppler-limited resolution: A global effective rovibrational model for highly excited vibrational states. J. Quant. Spectrosc. Radiat. Transf. 190, 38–47 (2017).
Nikitin, A. V., Rey, M., Chizhmakova, I. S. & Tyuterev, V. G. First full-dimensional potential energy and dipole moment surfaces of SF6. J. Phys. Chem. A 124, 7014–7023 (2020).
Rey, M., Chizhmakova, I. S., Nikitin, A. V. & Tyuterev, V. G. Towards a complete elucidation of the ro-vibrational band structure in the SF 6 infrared spectrum from full quantum-mechanical calculations. Phys. Chem. Chem. Phys. 23, 12115–12126 (2021).
Manamanni, K. et al. Limitations due to residual interference in a fiber-based optical frequency reference at 1.55 µm. J. Opt. Soc. Am. B 39, 438 (2022).
Hasegawa, M. et al. Microfabrication of cesium vapor cells with buffer gas for MEMS atomic clocks. , Sens. Actuators Phys. 167, 594–601 (2011).
Pracna, P., Špirko, V. & Kraemer, W. P. Electric dipole moment function of ammonia. J. Mol. Spectrosc. 136, 317–332 (1989).
Kim, K., McDowell, R. S. & King, W. T. Integrated infrared intensities and transition moments in SF 6. J. Chem. Phys. 73, 36–41 (1980).
Twagirayezu, S., Hall, G. E. & Sears, T. J. Frequency measurements and self-broadening of sub-Doppler transitions in the v1 + v3 band of C2 H2. J. Chem. Phys. 149, 154308 (2018).
Zambon, B. & Nienhuis, G. Reflection and transmission of light by thin vapor layers. Opt. Commun. 143, 308–314 (1997).
Acknowledgements
We acknowledge financial support from the ANR project SQUAT (Grant No. ANR-20-CE92-0006-01), the LABEX Cluster of Excellence FIRST-TF (ANR-10-LABX-48-01) and the PN-GRAM. We thank Daniel Bloch for discussions throughout the duration of the project and comments on the text. We also thank Jean-Michel Hartmann, Martial Ducloy, Livio Gianfrani for interesting exchanges and Yannis Pargoire for his contribution in the automation of the molecular pressure on/off modulation technique.
Author information
Authors and Affiliations
Contributions
The experiment was built by G.G.A., J.C.A.C., I.M., E.B., A.L. The data was taken by G.G.A., J.C.A.C., E.B., H.M. The fits were performed by G.G.A., I.M., H.M., A.L. The cell was fabricated by T.B., I.M., A.L. Data acquisition and automation via computer was managed by I.M with the participation of H.M. B.D. has provided expertise in molecular spectroscopy and participated in data interpretation in discussions with A.L. F.dB. has provided expertise in frequency metrology and participated in data interpretation in discussions with A.L. The manuscript and the reply to referee’s comments were written by A.L. with the participation of B.D. after discussions with F.dB., I.M., G.G.A., J.C.A.C. The project was coordinated by A.L.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Arellano, G.G., de Aquino Carvalho, J.C., Mouhanna, H. et al. Probing molecules in gas cells of subwavelength thickness with high frequency resolution. Nat Commun 15, 1862 (2024). https://doi.org/10.1038/s41467-024-45830-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-024-45830-x
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.