Molecular dynamics calculations have been used to explore the structure and dynamics of clusters of carbon dioxide, ranging in size up to 164 molecules. The most detailed calculations have been carried out around T=90 K with a view to interpreting the results of infrared data on clusters produced in a seeded argon beam. Under appropriate conditions, in addition to liquid‐like clusters, we have been able to produce clusters with the following structures: (i) an ordered solid—analogous to the bulk crystal, (ii) a bulk–solid core with liquid‐like outer layers, (iii) an amorphous solid, and (iv) an amorphous solid core with liquid‐like outer layers. We have calculated the dispersion of the infrared active intramolecular Q₃ mode for these clusters using the transition‐dipole transition‐dipole interaction that is usually invoked to explain the infrared spectrum of the bulk crystal. The resulting spectrum for the Q₃ mode of each type of cluster is quite distinct in appearance. This observation suggests that molecular dynamics calculations could be a valuable diagnostic tool in the study of clusters. The dynamical behavior of these clusters has also been investigated, and we find evidence for the existence of large‐amplitude, low‐frequency symmetric breathing modes with frequencies in the range 2–4 cm⁻¹. The overall shape of the experimental Fourier transform infrared spectra of clusters produced in a seeded argon jet is very similar in appearance to the calculated spectrum for a cluster of 116 molecules with an ordered solid core and liquid‐like outer layers.

Collision-induced absorption (CIA) by CO₂ is measured in the 1100–1600 cm⁻¹range using a Fourier-transform spectrometer with a resolution of 0.5 cm⁻¹. The current measurements, which agree well with previous ones but are more precise, reveal pronounced structures on top of both unresolved Fermi doublet bands consisting of P-, Q-,and R-like branches. Assignment of Q-branches at 1284.75 cm⁻¹and 1387.75 cm⁻¹to (CO₂)₂ dimers seems highly probable. The nature of other peaks observed in CIA and Raman spectra of the CO₂ Fermi doublet region is discussed.

Accurate measurements of collisionЃ]induced absorption in CO₂ are made at a number of temperatures in the range from − 40 to 60°C in the wavelength region 7–250 cm⁻¹. Direct evidence for the separation of the pure translational band from the rotational–translational band is obtained at all temperatures. This and other aspects of the band shape are discussed. Over the entire temperature range, the experimentally determined Kramers–Kronig integral is found to be in good agreement with the theoretical value, i.e., the static dielectric constant. This agreement is achieved only when the contribution of the quadrupole–quadrupole energy in the radial distribution function, of particular importance for CO₂ because of its large quadrupole moment, is calculated accurately. A value of the quadrupole moment is obtained, (4.5 ± 0.2)10⁻²⁶ esu, which is in satisfactory agreement with that obtained by the method of Buckingham and Disch, which does not depend on a knowledge of intermolecular force constants. Induction due to higher multipole moments and the overlap interaction is considered.

Collision induced microwave absorption is reported in pure CO₂, and CO₂-Ar, CO₂-CH₄ mixtures in the 70 G H z (2.3 cm^-1) region at a temperature of 22°C, using a sensitive cavity technique previously described. The results in pure CO₂ in the very low density region from 5 to 30 amagat accurately establish the dependence of the loss on the square and cube of the density, and the relaxation times are calculated. The experimental results agree well with previously reported lower frequency data at 0.3-0.8 cm^-1 which establishes the linear dependence on frequency of the absorption up to 2.3 cm^-1. There is also good agreement with an extrapolation of higher frequency infrared results of Ho et al. The relaxation times associated with the two and three body collisions are shown to be nearly equal at room temperatures with T~ = 0.84 x 10^-12 s and T~ = 1.0 x 10^-12 S. Higher order dependence on the density is observed for the CO₂-Ar and CO₂-CH₄ mixtures. The results are compared with earlier low frequency measurements at 0.8 cm^-1 and with the theory of Maryott and Kryder, taking account of correction terms in the dielectric virial coefficient according to Bose and Cole.

Molecular dynamics calculations have been used to explore the structure and dynamics of clusters of carbon dioxide, ranging in size up to 164 molecules. The most detailed calculations have been carried out around T=90 K with a view to interpreting the results of infrared data on clusters produced in a seeded argon beam. Under appropriate conditions, in addition to liquid‐like clusters, we have been able to produce clusters with the following structures: (i) an ordered solid—analogous to the bulk crystal, (ii) a bulk–solid core with liquid‐like outer layers, (iii) an amorphous solid, and (iv) an amorphous solid core with liquid‐like outer layers. We have calculated the dispersion of the infrared active intramolecular Q₃ mode for these clusters using the transition‐dipole transition‐dipole interaction that is usually invoked to explain the infrared spectrum of the bulk crystal. The resulting spectrum for the Q₃ mode of each type of cluster is quite distinct in appearance. This observation suggests that molecular dynamics calculations could be a valuable diagnostic tool in the study of clusters. The dynamical behavior of these clusters has also been investigated, and we find evidence for the existence of large‐amplitude, low‐frequency symmetric breathing modes with frequencies in the range 2–4 cm⁻¹. The overall shape of the experimental Fourier transform infrared spectra of clusters produced in a seeded argon jet is very similar in appearance to the calculated spectrum for a cluster of 116 molecules with an ordered solid core and liquid‐like outer layers.

The collision-induced absorption of gaseous CO₂is the primary source of far-infrared opacity of the atmosphere of Venus. At the temperatures and densities of the venusian atmosphere, the absorption is due mainly to binary collisions of CO₂molecules. Using a realistic anisotropic intermolecular potential and assuming the absorbing dipole to be due to the electrostatic induction and a quantum overlap, a series of molecular dynamics simulations were performed for the temperature range 200 to 800 K, and the roto-translational collision-induced absorption spectra at frequencies from 0 to 250 cm⁻¹were derived. The absorption coefficient in the submillimeter region, used in constituency retrieval studies, decreases more than 10 times in the temperature range 200 to 800 K. On the other hand, the absorption coefficient at 800 K and at the frequency range above 150 cm⁻¹was found to be almost 10 times higher than at 200 K. Earlier works relied on experimental RT CIA data at a fixed temperature of 300 K. The new, temperature-dependent absorption bands may, when included in the analysis of the atmospheric radiative transfer of the planet, help explain the observed high far-infrared opacity of the lower layers of the atmosphere. To make the results of the simulations readily available for atmospheric abundance and radiative transfer analysis, an analytic model of the roto-translational collision-induced absorption spectral profile, applicable from 200 to 800 K, is being proposed here. The FORTRAN computer code of this newly developed model is available from the authors on request.

Classical molecular dynamics simulations have been carried out for gaseous CO₂ starting from various anisotropic intermolecular potential energy surfaces. Through calculations for a large number of molecules treated as rigid rotors, the time evolution of the interaction-induced electric dipole vector is obtained and the Laplace transform of its autocorrelation function gives the collision-induced absorption rototranslational spectrum. The results are successfully compared with those of previous similar calculations before studies of the influences of the intermolecular potential and induced-dipole components are made. The calculated spectra show a significant sensitivity to anisotropic forces consistently with previous analyses limited to the spectral moments. The present results also demonstrate the importance of vibrational and back-induction contributions to the induced dipole. Comparisons between measured far infrared (0–250 cm⁻¹) spectra at different temperatures and results calculated without the use of any adjustable parameter are made. When the best and more complete input data are used, the quality of our predictions is similar to that obtained by Gruszka et al. [Mol. Phys. 93, 1007 (1998)] after the introduction of ad hoc short-range overlap contributions. Our results thus largely obviate the need for such contributions the magnitudes of which remain questioned. Nevertheless, problems remain since, whereas good agreements with measurements are obtained above 50 cm⁻¹, the calculations significantly underestimate the absorption below, a problem which is discussed in terms of various possible error sources.

The ν₁ band of gaseous carbon dioxide has been studied in pressure-induced absorption at temperatures of ~ 190, ~ 300, and ~ 470 K, over a density range from 0.5 to 300 amagat, and with path lengths from 0.007 to 56 m. The observed temperature variation of the binary absorption coefficient can be satisfactorily accounted for only by adding a quadrupole–quadrupole interaction term to the usual Lennard–Jones model for the inter-molecular potential. The band profile is in agreement with the theory of quadrupole-induced absorption. There is some increase in the intensity of the band near the critical point due to the divergence of the correlation length. A very marked increase in the intensity is possibly prevented by the "cancellation effect".

Samples of CO₂ at pressures up to 21.3 atm and with paths as long as 1067 m have been employed to measure CO₂ absorption between 1100 and 1835 cm^-1. Most of the absorption is due to the ν₁ and 2ν₂ bands. These bands are forbidden for the symmetric ¹⁶O¹²C^l6O and ¹⁶O¹³C¹⁶O molecules and, therefore, produce only pressure-induced absorption. The asymmetric ¹⁶O¹²C¹⁸O and ¹⁶O¹²C¹⁷O molecules produce bands that contain components of both intrinsic and pressure-induced absorption. The integrated-absorption coefficient for the intrinsic absorption is 1.08 × 10^-22 (±10%) molecules ^-l cm² cm^-1. The integrated pressure-induced absorption coefficient is 1.68×10^-22 (±8%) molecules^-1 cm² atm^-1 cm^-l.

Induced absorption in carbon dioxide has been studied at a frequency of 9260 Mc/sec over the temperature range 270°—500°K and to pressures as high as 95 atm. Since the dielectric loss ϵ″ is expected on the basis of the dispersion relations to be proportional to frequency throughout the microwave region the data has been fit with an expansion in the amagat density ρ of the form ϵ^′′/ = Aρ²+Bρ³, where is in wavenumbers. It is found that A=2.5(1)×10⁻⁸(T/273)^−3.0(1), and B/A=−1.05(8)×10⁻²(T/273)^−2.3(3). The molecular quadrupole moment calculated from the first coefficient of the dielectric loss A is then 6.7×10⁻²⁶ (esu).

The absorption of carbon dioxide-nitrogen mixtures at 4.40 µ was examined over a temperature range of 1200 to 2100°K for optical densities of 0.10 to 0.40 atm-cm. The test gas was elevated to desired temperatures by shock compression. Properties of the test gas were determined by measurement of initial concentrations and shock wave velocities. Absorption was observed as a diminution of source beam intensity monitored with a rapid response infrared detector and displayed on an oscilloscope. Beer's law was found to be applicable over the range of temperature and concentration studied. Absorption coefficients calculated from Beer's law plots reached a maximum of 3.0 atm^—1 cm^—1 at about 1400°K, and were independent of total pressure from 0.26 to 1.05 atm. The absorption was also measured over a wavelength interval of 4.37 to 4.6 µ. The results of this investigation are compared to the existing theoretical and experimentally determined absorption data.

The infrared transmittance of carbon dioxide has been calculated over a wide range of path lengths, pressures, and temperatures from 500 to 10,000 cm^-1. Values of the transmittance are given at intervals of 2.5 cm^-1. In addition, transmittance values are also given which have been averaged over larger intervals. All contributing spectral lines whose relative intensity is greater than 10-8 that of the strongest line in any particular band have been included in the calculation. In addition, the contributions from the eight major isotopic species have been included. The calculation of the vibrational energy levels included terms through the third power of the vibrational quantum number and also the effects of Fermi resonanse. The final transmittance tables were generated using the the quasi-random model of molecular band absorption.

A theoretical description of line mixing in the Q branch of an infrared rotation-vibration band is obtaibed in the weak coupling approximation. The description is applied to the case of the ν₂ band of CO₂ near 667 cm^-1. Theoretical absorption profiles are given for selected spectral regions corresponding to representative values of the pressure. A comparision is made between the weak coupling results and the results of a forthcoming dynamical calculation of the absorption band shape.

Two types of theoretical models for an absorption band are described and applied to infrared absorption bands of CO₂ and H₂O. For CO₂ , which has a regular fine structure, the so-called Elsasser model, which assumes equally-spaced, equally-intense lines, can be applied. For H₂O, which has a highly irregular fine structure, the so-called statistical or Goody model, which assumes random spacings and intensity distribution, is applicable. The individual lines are assumed to have the Lorentz shape in both models. The method of application of these models to the experimentally observed bands is discussed in detail.
 

The temperature dependence of the high frequency far wings of the self-broadened CO₂ lines has been investigated in the 2400–2600-cm^-1 spectral region. The temperature dependence of the corrective shape factor X(σ,T) is demonstrated for the first time.

High temperature absorption measurements were made for CO₂ gas in Local Thermodynamic Equilibrium (LTE) with a hot cell and high resolution interferometer. The experimental data were compared to band-model and line-by-line model transmittance calculations using line parameters from the HITRAN and HITEMP data bases. The line-by-line calculations using HITEMP were in excellent agreement with experimental measurements, while the model calculations using the HITRAN data underpredicted the absorption by approx. 10%.

A simple theoretical approach of Q-branch absorption is developed. It is based on the classical approximation of the rotational distribution and a Strong-Collision-type modelling of line-coupling coefficients. The Q-branch absorption is then represented by a very simple analytical expression which depends on six average parameters; pressure and wave-number dependences are included in the model so that the parameters only depend on the molecular system (active molecule + perturber), band, and temperature. Tests show that, provided effective parameters are used, our model enables very satisfactory predictions of the pressure, temperature, and wavenumber dependences of Q-branch absorption. These effective parameters can be deduced from experimental spectra and no previous knowledge of neither the spectroscopic nor collision parameters is required. This makes the present approach suitable for species (most of those of atmospheric interest) for which no alternative and more accurate approach is yet available.

The shapes of Q-branches in CO₂ spectra in the 580–850 cm⁻¹ region for pure gas and mixtures CO₂---He and CO₂---Ar have been studied at resolutions up to 0.002cm⁻¹. The branch broadening coefficients were measured in the pressure range from a few torr to 50 atm. Those for 11¹02-02²01 Π-Δ transition are different for pressures below and above 5 atm. At higher pressures the branch broadening coefficients are similar for all Q-branches. The observed Q-branch shape transformation is explained by line mixing effects.

We propose a new empirical method of band shape calculations which takes into account line mixing effects. The proposed shape is based on the strong collision model with attenuated interbranch coupling. Apart from the conventional set of spectral line parameters one needs only one additional parameter to account for line mixing effects in band profile calculations. This parameter depends on the perturbing gas type. It is the same for all molecular infrared absorption bands. The new shapes are shown to be successful in representing the measured absorption bands of CO₂ in nitrogen (v₂, v₃, v₂ + 2v₃, 2v₂ − v₂, and 2v₁ − v₂) and the 3.6 μm band of O₃ in nitrogen.

The formalism developed previously for the calculation of the far-wing line shape function and the corresponding absorption coefficient using a single-component anisotropic interaction term and the binary collision and quasistatic approximations is generalized to multicomponent anisotropic potential functions. Explicit expressions are presented for several common cases, including the long-range dipole-dipole plus dipole-quadrupole interaction and a linear molecule interacting with a perturber atom. After determining the multicomponent functional representation for the interaction between CO₂ and Ar from previous published data, we calculate the theoretical line shape function and the corresponding absorption due to the v₃ band of CO₂ in the frequency region 2400-2580 cm^-1 and compare our results with previous calculations carried out using a single-component anisotropic interaction, and with the results obtained assuming Lorentzian line shapes. The principal uncertainties in the present results, possible refinements of the theoretical formalism, and the applicability to other systems are discussed briefly.

 

A theory is presented for the modification of the line-shape functions and absorption coefficient due to the breakdown of the quasistatic approximation. This breakdown arises from the effects of molecular motion and increases the absorption in the near wings. Numerical calculations for the high-frequency wing of the ν₃ band of CO₂ broadened by Ar are reported and it is shown that these effects are significant near the bandhead. The importance of such. corrections in other spectral regions and for other systems is discussed briefly.

 

The IR absorption of the pure CO₂ gas in the region of 15 μm was examined. A special attention was given to the line mixing effect that influences the spectral shape of the vibration–rotation absorption bands in the Q-branch regions. Two methods in shape description were analyzed. The first method uses the Rosenkranz line shapes with the line mixing parameters, which are found from an empirical rotational relaxation matrix. The second method is based on the strong collision model with adjusted branch coupling (ABC-model). The merits and the demerits of these two methods are discussed, and the results of the corresponding calculations are compared to the measured shapes. It is inferred that the ABC-model for the absorption coefficient calculations can be successfully applied for solving the non-LTE radiative transfer problem in CO₂ bands in the atmospheres of Earth-like planets.

We present a high-temperature version, CDSD-1000, of the carbon dioxide spectroscopic databank. The databank contains the line parameters (positions, intensities, air- and self-broadened half-widths and coefficients of temperature dependence of air-broadened half-widths) of the four most abundant isotopic species of the carbon dioxide molecule. The reference temperature is T_ref=1000°K and the intensity cutoff is I_cut=10⁻²⁷ cm⁻¹/moleculecm⁻². More than 3 million lines covering the 260–8310, 418–2454, 394–4662, and 429–2846 cm⁻¹ spectral ranges for ¹²C¹⁶O₂, ¹³C¹⁶O₂, ¹²C¹⁶O¹⁸O, and ¹²C¹⁶O¹⁷O, respectively, are included in CDSD-1000. The databank has been generated within the framework of the method of effective operators and based on the global fittings of spectroscopic parameters (parameters of the effective Hamiltonians and effective dipole moment operators) to observed data collected from the literature. Line-by-line simulations of several low- and medium-resolution high-temperature (T=800-3000 K) spectra have been performed in order to validate the databank. Comparisons of CDSD-1000 with other high-temperature databanks HITEMP, HITELOR, and EM2C are also given. CDSD-1000 is able to reproduce observed spectra in a more satisfactory way than the high-resolution databank HITEMP for temperatures higher than 1000°K. The databank is useful for studying high-temperature radiative properties of CO₂. CDSD-1000 is freely accessible via the Internet.

Temperature (200–300 K) and pressure (70–200 atm) dependent laboratory measurements of infrared transmission by CO2–N2 mixtures have been made. From these experiments the absorption coefficient is reconstructed, over a range of several orders of magnitude, between 600 and 1000 cm-1. The elevated densities used in the experiments (up to 200 atm) magnify the contribution of the wings of the v2 band lines. In order to analyze the spectra, a theoretical model based on the energy corrected sudden approximation is proposed which accounts for line-mixing effects within the impact approximation. This approach uses the model and associated parameters built previously to model Q branches (JQSRT 1999;61:153) but extends it by nowincluding all P, Q, and R lines. No adjustable parameters are used and fundamental properties of the collisional relaxation operator are verified by using a renormalization procedure. Comparisons between measured and calculated spectra confirm that neglecting line-mixing (Lorentzian model) leads to an overestimation of absorption by up to three orders of magnitude in the far wings. On the other hand, the proposed approach leads to satisfactory results both in regions dominated by contributions of local lines and in the wing: measured spectra are correctly modeled over a range where absorption varies by more than four orders of magnitude. The largest discrepancies, which appear about 150 cm-1 from the v2 center, can be due to finite duration of collisions effects or to uncertainties in the experimental determination of very weak absorption.
 

The theoretical approach based on the Energy Corrected Sudden Approximation presented in the previous companion paper is used in order to account for line-mixing effects in infrared bands of CO₂. Its performance, which was demonstrated using laboratory spectra is confirmed here by considering atmospheric transmission in the 10–14 μm region. Comparisons are made between forward calculations of atmospheric transmission spectra and values measured using two different solar occultation experiments based on high resolution Fourier transform instruments. The results demonstrate that neglecting line-mixing and using a Voigt model can lead to a very large overestimation of absorption that may extend over more than 300 cm⁻¹ in the wing of the CO₂ ν₂ band. They also demonstrate the capability of our model to represent accurately the absorption in the entire region for a variety of atmospheric paths. Among positive consequences of the quality of the model, the possibility of retrieving amounts of (heavy) trace gases with weak and broad absorption features is demonstrated.

Precise modelling of infrared absorption by carbon dioxide is of primary importance for radiative transfer calculations in CO

-rich atmospheres like those of Venus and Mars. Despite various measurements and theoretical models dedicated to this subject, accurate data at different temperatures and pressures are still lacking in numerous spectral regions. In this work, using two Fourier Transform Spectrometers, we have measured spectra of pure CO

in a large spectral region range, from 750 to 8500 cm

at various densities (3–57 amagat) and temperatures (230–473 K). Comparisons between measured dipolar absorption bands and spectra calculated with the widely used Lorentz line shape show very large discrepancies. This result is expected since the Lorentz approach neglects line-coupling effects due to intermolecular collisions which transfer absorption from the wings to the band center. In order to account for this effect, a theoretical approach based on the impact and Energy Corrected Sudden approximations has been developed. Comparisons of this model with numerous laboratory spectra in a wide range of pressure, temperature and spectral domain show satisfactory agreements for band centers and near wing regions where the impact approximation is valid. However, as expected, due to the breakdown of the impact approximation, the model fails when considering far wing regions. In the absence of precise models accounting for line-mixing

finite collision duration (non impact) effects, empirical approximations are proposed in order to model the far wings.

An isotopic-independent, highly accurate potential energy surface (PES) has been determined for CO₂ by refining a purely ab initio PES with selected, purely experimentally determined rovibrational energy levels. The purely ab initio PES is denoted Ames-0, while the refined PES is denoted Ames-1. Detailed tests are performed to demonstrate the spectroscopic accuracy of the Ames-1 PES. It is shown that Ames-1 yields σ _rms (root-mean-squares error) = 0.0156 cm⁻¹ for 6873 J = 0–117 ¹²C¹⁶O₂ experimental energy levels, even though less than 500 ¹²C¹⁶O₂ energy levels were included in the refinement procedure. It is also demonstrated that, without any additional refinement, Ames-1 yields very good agreement for isotopologues. Specifically, for the ¹²C¹⁶O₂ and ¹³C¹⁶O₂ isotopologues, spectroscopic constants G _v computed from Ames-1 are within ±0.01 and 0.02 cm⁻¹ of reliable experimentally derived values, while for the ¹⁶O¹²C¹⁸O, ¹⁶O¹²C¹⁷O, ¹⁶O¹³C¹⁸O, ¹⁶O¹³C¹⁷O, ¹²C¹⁸O₂, ¹⁷O¹²C¹⁸O, ¹²C¹⁷O₂, ¹³C¹⁸O₂, ¹³C¹⁷O₂, ¹⁷O¹³C¹⁸O, and ¹⁴C¹⁶O₂ isotopologues, the differences are between ±0.10 and 0.15 cm⁻¹. To our knowledge, this is the first time a polyatomic PES has been refined using such high J values, and this has led to new challenges in the refinement procedure. An initial high quality, purely ab initiodipole moment surface (DMS) is constructed and used to generate a 296 K line list. For most bands, experimental IR intensities are well reproduced for ¹²C¹⁶O₂ using Ames-1 and the DMS. For more than 80% of the bands, the experimental intensities are reproduced with σ _rms(ΔI) < 20% or σ _rms(ΔI/δ _obs) < 5. A few exceptions are analyzed and discussed. Directions for future improvements are discussed, though it is concluded that the current Ames-1 and the DMS should be useful in analyzing and assigning high-resolution laboratory or astronomical spectra.

Non-intrusive spectroscopic probing of weakly bound van der Waals complexes forming in gaseous carbon dioxide is generally performed at low pressures, for instance in supersonic jets, where the low temperature favors dimers, or in few-atmosphere samples, where the signature of dimers varying as the squared gas density is entangled with the dominating collision-induced absorption. We report experimental and theoretical results on CO₂ dimers at very high pressures approaching the liquid phase. We observe that the shape of the CO₂ -dimer bands undergoes a distinctive line-mixing transformation, which reveals an unexpected stability of the dimers despite the collisions with the surrounding particles and negates the common belief that CO₂ dimers are short-lived complexes. Our results furnish a deeper insight allowing a better modeling of CO₂ -rich atmospheres and provide also a new spectroscopic tool for studying the robustness of molecular clusters.

The CO₂ absorption was measured in the 7000 and 8000 cm^–1 regions. The absorption coefficients were calculated using the asymptotic line wing shape theory. Line shape parameters were found from fitting to experimental data. The calculation results agree well with the measurement data. According to the line wing theory, the absorption in the band wings is caused by the wings of strong lines of an adjacent band. Within these assumptions, experimental and calculated data on the CO₂ absorption coefficient in the band wings in the 7000 and 8000 cm^–1 regions can provide information on the line shape at frequency detuning from several tens to several hundreds of half-widths. The results support the hypothesis that line shape parameters in the line wings related to transitions with the same initial state are close to each other. Deviations from a Lorentzian profile are found for some CO₂ bands and turn out different for the wings of different bands

Collision-induced absorption (CIA) by CO₂ is measured in the 1100–1600 cm⁻¹range using a Fourier-transform spectrometer with a resolution of 0.5 cm⁻¹. The current measurements, which agree well with previous ones but are more precise, reveal pronounced structures on top of both unresolved Fermi doublet bands consisting of P-, Q-,and R-like branches. Assignment of Q-branches at 1284.75 cm⁻¹and 1387.75 cm⁻¹to (CO₂)₂ dimers seems highly probable. The nature of other peaks observed in CIA and Raman spectra of the CO₂ Fermi doublet region is discussed.

The two components of the (ν₁, 2ν₂) Fermi doublet of gaseous CO₂ in an absorption path length of 56 m at 192 °K show a complex structure. When the allowed C¹⁶O¹⁸O absorption and the pressure-induced CO₂ absorption are removed by computational procedures, the residual spectrum consists of two similar symmetric patterns of five maxima. These are interpreted in terms of the rotation and vibration of (CO₂)₂ dimers held by quadrupole–quadrupole interaction in the locked T position at an intermolecular distance of 4.1 Å.

The temperature variations in collision-induced absorption (CIA) spectra of carbon dioxide in the region of the Fermi doublet are examined. New FTIR CIA spectra are recorded in the temperature range T=206–296°K. The spectra were subject to decomposition in order to separate true dimer contributions to the CIA profile from the base absorption caused by unbound pairs. The use of statistical physics theory allowed for quite nice reproduction of the observed temperature variations of the normalized dimer intensity.

The polarized and depolarized Raman profiles of supercritical CO₂ have been measured in the region of the ν₂ bending mode (forbidden transition at about 668 cm⁻¹) and for the Fermi dyad (1285 and 1388 cm⁻¹) along the isotherms 307, 309, 313, and 323 K in a reduced density domain 0.04<ρ^* = ρ/ρ_C<2.04 (ρ_C ∼ 467.6 kg m⁻³, ρ_C is the critical density). The spectral features associated with the ν₂ mode (degeneracy removal of the mode and Raman intensity activation) are found to be due to the formation of transient complexes. This is supported by the spectral signatures predicted for parallel slipped dimer and trimers (cyclic and noncyclic) from ab initio calculations taking into account the frequency anharmonicity. The band-shape analysis of the Fermi doublet (observed in the spectral range of 1260–1400 cm⁻¹) shows that on the subpicosecond time scale of the Raman spectroscopy, a tagged CO₂ molecule probed two kinds of environment in its first shell of neighbors independent of local density enhancement phenomenon. The first one involves interactions of CO₂ with surrounding molecules in the first shell whereas the latter is associated with a transient dimer formation. Finally, a broad band observed between the Fermi dyad (at about 1335 cm⁻¹) is assessed from symmetry considerations and from its depolarization ratio as a further evidence of transient complex formation in supercritical CO₂.

The carbon dioxide dimer spectroscopic patterns are retrieved from the analysis of collision-induced absorption (CIA) spectral bandshape at room temperature. It is shown that the use of the simplified model based on the symmetric-top approximation allows roughly consistent simulation of the observed (CO₂)₂ dimer spectrum. The rotational constants obtained can be considered as effective thermally averaged constants which characterize dimeric structure, strongly distorted from the ground state. The overall CIA bandshape and the integrated intensity of absorption are broken down into partial contributions from tightly bound and metastable dimers and free-pair states. This approach is shown to be in agreement with a wide range of independent spectroscopic and thermodynamic data.

The ν₁ band of gaseous carbon dioxide has been studied in pressure-induced absorption at temperatures of ~ 190, ~ 300, and ~ 470 K, over a density range from 0.5 to 300 amagat, and with path lengths from 0.007 to 56 m. The observed temperature variation of the binary absorption coefficient can be satisfactorily accounted for only by adding a quadrupole–quadrupole interaction term to the usual Lennard–Jones model for the inter-molecular potential. The band profile is in agreement with the theory of quadrupole-induced absorption. There is some increase in the intensity of the band near the critical point due to the divergence of the correlation length. A very marked increase in the intensity is possibly prevented by the "cancellation effect".

Samples of CO₂ at pressures up to 21.3 atm and with paths as long as 1067 m have been employed to measure CO₂ absorption between 1100 and 1835 cm^-1. Most of the absorption is due to the ν₁ and 2ν₂ bands. These bands are forbidden for the symmetric ¹⁶O¹²C^l6O and ¹⁶O¹³C¹⁶O molecules and, therefore, produce only pressure-induced absorption. The asymmetric ¹⁶O¹²C¹⁸O and ¹⁶O¹²C¹⁷O molecules produce bands that contain components of both intrinsic and pressure-induced absorption. The integrated-absorption coefficient for the intrinsic absorption is 1.08 × 10^-22 (±10%) molecules ^-l cm² cm^-1. The integrated pressure-induced absorption coefficient is 1.68×10^-22 (±8%) molecules^-1 cm² atm^-1 cm^-l.

A simple theoretical approach of Q-branch absorption is developed. It is based on the classical approximation of the rotational distribution and a Strong-Collision-type modelling of line-coupling coefficients. The Q-branch absorption is then represented by a very simple analytical expression which depends on six average parameters; pressure and wave-number dependences are included in the model so that the parameters only depend on the molecular system (active molecule + perturber), band, and temperature. Tests show that, provided effective parameters are used, our model enables very satisfactory predictions of the pressure, temperature, and wavenumber dependences of Q-branch absorption. These effective parameters can be deduced from experimental spectra and no previous knowledge of neither the spectroscopic nor collision parameters is required. This makes the present approach suitable for species (most of those of atmospheric interest) for which no alternative and more accurate approach is yet available.

An isotopic-independent, highly accurate potential energy surface (PES) has been determined for CO₂ by refining a purely ab initio PES with selected, purely experimentally determined rovibrational energy levels. The purely ab initio PES is denoted Ames-0, while the refined PES is denoted Ames-1. Detailed tests are performed to demonstrate the spectroscopic accuracy of the Ames-1 PES. It is shown that Ames-1 yields σ _rms (root-mean-squares error) = 0.0156 cm⁻¹ for 6873 J = 0–117 ¹²C¹⁶O₂ experimental energy levels, even though less than 500 ¹²C¹⁶O₂ energy levels were included in the refinement procedure. It is also demonstrated that, without any additional refinement, Ames-1 yields very good agreement for isotopologues. Specifically, for the ¹²C¹⁶O₂ and ¹³C¹⁶O₂ isotopologues, spectroscopic constants G _v computed from Ames-1 are within ±0.01 and 0.02 cm⁻¹ of reliable experimentally derived values, while for the ¹⁶O¹²C¹⁸O, ¹⁶O¹²C¹⁷O, ¹⁶O¹³C¹⁸O, ¹⁶O¹³C¹⁷O, ¹²C¹⁸O₂, ¹⁷O¹²C¹⁸O, ¹²C¹⁷O₂, ¹³C¹⁸O₂, ¹³C¹⁷O₂, ¹⁷O¹³C¹⁸O, and ¹⁴C¹⁶O₂ isotopologues, the differences are between ±0.10 and 0.15 cm⁻¹. To our knowledge, this is the first time a polyatomic PES has been refined using such high J values, and this has led to new challenges in the refinement procedure. An initial high quality, purely ab initiodipole moment surface (DMS) is constructed and used to generate a 296 K line list. For most bands, experimental IR intensities are well reproduced for ¹²C¹⁶O₂ using Ames-1 and the DMS. For more than 80% of the bands, the experimental intensities are reproduced with σ _rms(ΔI) < 20% or σ _rms(ΔI/δ _obs) < 5. A few exceptions are analyzed and discussed. Directions for future improvements are discussed, though it is concluded that the current Ames-1 and the DMS should be useful in analyzing and assigning high-resolution laboratory or astronomical spectra.

Non-intrusive spectroscopic probing of weakly bound van der Waals complexes forming in gaseous carbon dioxide is generally performed at low pressures, for instance in supersonic jets, where the low temperature favors dimers, or in few-atmosphere samples, where the signature of dimers varying as the squared gas density is entangled with the dominating collision-induced absorption. We report experimental and theoretical results on CO₂ dimers at very high pressures approaching the liquid phase. We observe that the shape of the CO₂ -dimer bands undergoes a distinctive line-mixing transformation, which reveals an unexpected stability of the dimers despite the collisions with the surrounding particles and negates the common belief that CO₂ dimers are short-lived complexes. Our results furnish a deeper insight allowing a better modeling of CO₂ -rich atmospheres and provide also a new spectroscopic tool for studying the robustness of molecular clusters.

The CO₂ absorption was measured in the 7000 and 8000 cm^–1 regions. The absorption coefficients were calculated using the asymptotic line wing shape theory. Line shape parameters were found from fitting to experimental data. The calculation results agree well with the measurement data. According to the line wing theory, the absorption in the band wings is caused by the wings of strong lines of an adjacent band. Within these assumptions, experimental and calculated data on the CO₂ absorption coefficient in the band wings in the 7000 and 8000 cm^–1 regions can provide information on the line shape at frequency detuning from several tens to several hundreds of half-widths. The results support the hypothesis that line shape parameters in the line wings related to transitions with the same initial state are close to each other. Deviations from a Lorentzian profile are found for some CO₂ bands and turn out different for the wings of different bands

Collision-induced absorption (CIA) by CO₂ is measured in the 1100–1600 cm⁻¹range using a Fourier-transform spectrometer with a resolution of 0.5 cm⁻¹. The current measurements, which agree well with previous ones but are more precise, reveal pronounced structures on top of both unresolved Fermi doublet bands consisting of P-, Q-,and R-like branches. Assignment of Q-branches at 1284.75 cm⁻¹and 1387.75 cm⁻¹to (CO₂)₂ dimers seems highly probable. The nature of other peaks observed in CIA and Raman spectra of the CO₂ Fermi doublet region is discussed.

The two components of the (ν₁, 2ν₂) Fermi doublet of gaseous CO₂ in an absorption path length of 56 m at 192 °K show a complex structure. When the allowed C¹⁶O¹⁸O absorption and the pressure-induced CO₂ absorption are removed by computational procedures, the residual spectrum consists of two similar symmetric patterns of five maxima. These are interpreted in terms of the rotation and vibration of (CO₂)₂ dimers held by quadrupole–quadrupole interaction in the locked T position at an intermolecular distance of 4.1 Å.

The carbon dioxide dimer spectroscopic patterns are retrieved from the analysis of collision-induced absorption (CIA) spectral bandshape at room temperature. It is shown that the use of the simplified model based on the symmetric-top approximation allows roughly consistent simulation of the observed (CO₂)₂ dimer spectrum. The rotational constants obtained can be considered as effective thermally averaged constants which characterize dimeric structure, strongly distorted from the ground state. The overall CIA bandshape and the integrated intensity of absorption are broken down into partial contributions from tightly bound and metastable dimers and free-pair states. This approach is shown to be in agreement with a wide range of independent spectroscopic and thermodynamic data.

The ν₁ band of gaseous carbon dioxide has been studied in pressure-induced absorption at temperatures of ~ 190, ~ 300, and ~ 470 K, over a density range from 0.5 to 300 amagat, and with path lengths from 0.007 to 56 m. The observed temperature variation of the binary absorption coefficient can be satisfactorily accounted for only by adding a quadrupole–quadrupole interaction term to the usual Lennard–Jones model for the inter-molecular potential. The band profile is in agreement with the theory of quadrupole-induced absorption. There is some increase in the intensity of the band near the critical point due to the divergence of the correlation length. A very marked increase in the intensity is possibly prevented by the "cancellation effect".

An isotopic-independent, highly accurate potential energy surface (PES) has been determined for CO₂ by refining a purely ab initio PES with selected, purely experimentally determined rovibrational energy levels. The purely ab initio PES is denoted Ames-0, while the refined PES is denoted Ames-1. Detailed tests are performed to demonstrate the spectroscopic accuracy of the Ames-1 PES. It is shown that Ames-1 yields σ _rms (root-mean-squares error) = 0.0156 cm⁻¹ for 6873 J = 0–117 ¹²C¹⁶O₂ experimental energy levels, even though less than 500 ¹²C¹⁶O₂ energy levels were included in the refinement procedure. It is also demonstrated that, without any additional refinement, Ames-1 yields very good agreement for isotopologues. Specifically, for the ¹²C¹⁶O₂ and ¹³C¹⁶O₂ isotopologues, spectroscopic constants G _v computed from Ames-1 are within ±0.01 and 0.02 cm⁻¹ of reliable experimentally derived values, while for the ¹⁶O¹²C¹⁸O, ¹⁶O¹²C¹⁷O, ¹⁶O¹³C¹⁸O, ¹⁶O¹³C¹⁷O, ¹²C¹⁸O₂, ¹⁷O¹²C¹⁸O, ¹²C¹⁷O₂, ¹³C¹⁸O₂, ¹³C¹⁷O₂, ¹⁷O¹³C¹⁸O, and ¹⁴C¹⁶O₂ isotopologues, the differences are between ±0.10 and 0.15 cm⁻¹. To our knowledge, this is the first time a polyatomic PES has been refined using such high J values, and this has led to new challenges in the refinement procedure. An initial high quality, purely ab initiodipole moment surface (DMS) is constructed and used to generate a 296 K line list. For most bands, experimental IR intensities are well reproduced for ¹²C¹⁶O₂ using Ames-1 and the DMS. For more than 80% of the bands, the experimental intensities are reproduced with σ _rms(ΔI) < 20% or σ _rms(ΔI/δ _obs) < 5. A few exceptions are analyzed and discussed. Directions for future improvements are discussed, though it is concluded that the current Ames-1 and the DMS should be useful in analyzing and assigning high-resolution laboratory or astronomical spectra.

Non-intrusive spectroscopic probing of weakly bound van der Waals complexes forming in gaseous carbon dioxide is generally performed at low pressures, for instance in supersonic jets, where the low temperature favors dimers, or in few-atmosphere samples, where the signature of dimers varying as the squared gas density is entangled with the dominating collision-induced absorption. We report experimental and theoretical results on CO₂ dimers at very high pressures approaching the liquid phase. We observe that the shape of the CO₂ -dimer bands undergoes a distinctive line-mixing transformation, which reveals an unexpected stability of the dimers despite the collisions with the surrounding particles and negates the common belief that CO₂ dimers are short-lived complexes. Our results furnish a deeper insight allowing a better modeling of CO₂ -rich atmospheres and provide also a new spectroscopic tool for studying the robustness of molecular clusters.

The CO₂ absorption was measured in the 7000 and 8000 cm^–1 regions. The absorption coefficients were calculated using the asymptotic line wing shape theory. Line shape parameters were found from fitting to experimental data. The calculation results agree well with the measurement data. According to the line wing theory, the absorption in the band wings is caused by the wings of strong lines of an adjacent band. Within these assumptions, experimental and calculated data on the CO₂ absorption coefficient in the band wings in the 7000 and 8000 cm^–1 regions can provide information on the line shape at frequency detuning from several tens to several hundreds of half-widths. The results support the hypothesis that line shape parameters in the line wings related to transitions with the same initial state are close to each other. Deviations from a Lorentzian profile are found for some CO₂ bands and turn out different for the wings of different bands

Absorption spectra of pure CO₂ have been recorded in the vicinity of the 2675 cm⁻¹ Fermi triad for temperatures between 211 K and 296 K. The 2ν₁, ν₁ + 2ν₂, 2ν₂ collision-induced components have been extracted from the measured spectra, including for the low frequency band at 2547 cm⁻¹, which is strongly masked by the ν₃ wing absorption. Dimeric features are clearly seen on top of the structureless profiles. Integrated intensities of the Fermi-triad components are determined as a function of temperature.

The temperature dependence of the high frequency far wings of the self-broadened CO₂ lines has been investigated in the 2400–2600-cm^-1 spectral region. The temperature dependence of the corrective shape factor X(σ,T) is demonstrated for the first time.

The shapes of the self- and N₂-broadened ν₃ CO₂ fundamental vibration-rotation band in the microwindows (troughs between the lines) have been measured at various temperatures. Important deviations with respect to the superposition of Lorentzian profiles are observed. These deviations are interpreted in terms of line coupling, which redistributes the intensity in the whole band. In order to take into account this line coupling, two models are considered within the frame of the impact theory. The first model uses the strong-collision approximation to describe the rotational energy transferred by collisions. It leads to a simple analytical expression for the band profile. The second model is based on the exponential-gap law. These two models account well for the frequency dependence of the measured absorption in the microwindows and for the temperature dependence in the case of the N₂-broadened CO₂ band but not in the self-broadened case. The influence of the line-coupling rotational distribution, which differs significantly in the two models, is discussed. The possible role of the finite duration of collision in rotational energy transfer is examined.

Solar occultation spectra obtained with a balloon-borne interferometer have been used to study continuous absorption by N₂ and CO₂ near 2400 cm^-1 in the lower stratosphere. Synthetic continuum transmittances, calculated from published coefficients for far-wing absorption by CO₂ lines and for pressure-induced absorption by the fundamental band of N₂, are in fair agreement with the observed stratospheric values. The continuum close to the ν₃ R-branch band head of CO₂ is sensitive to the CO₂ far-wing line shape. Therefore, given highly accurate knowledge of the N₂ continuum from laboratory data, high-resolution stratospheric
spectra provide a sensitive means for in situ testing of various  air-broadened CO₂ line shapes at low temperatures.

The absorption by pure CO₂ beyond the ν₃ bandhead has been measured with a grating spectrometer. Experiments have been made in the 0–60-bar and 291–751-K pressure and temperature ranges. Our room temperature determinations are in good agreement with previous ones and the measured temperature dependence above room temperature is consistent with recent determinations below 300 K. Lorentzian calculations, modified by the introduction of a line shape corrective factor x, are presented. Good agreement between the observed and calculated spectra is obtained when a temperature independent x factor, determined by Cousin et al. at 296 K, is used.

We present experimental results on the absorption by CO₂-N₂ mixtures in the 4.3 μm region. Measurements have been made in the 300–800 K and 0–60 bar ranges; these are in good agreement with previous determinations below 296 K. Frequency- and temperature-dependent factors χ are introduced in order to account for the subLorentzian behavior of the CO₂ far line-wings. Their dependences on temperature are deduced from experimental results beyond the 4.3μm band-head in the 193–773 K range for both CO₂-CO₂ and CO₂-N₂ and fitted by simple analytical functions. Comparisons are presented between experimental and theoretical spectra on both the low- and high-frequency sides of the band (2100–2600 cm^-1 range). It is shown that calculations using χ factors are inaccurate near line-centers at elevated pressures.

The ν₁ band of gaseous carbon dioxide has been studied in pressure-induced absorption at temperatures of ~ 190, ~ 300, and ~ 470 K, over a density range from 0.5 to 300 amagat, and with path lengths from 0.007 to 56 m. The observed temperature variation of the binary absorption coefficient can be satisfactorily accounted for only by adding a quadrupole–quadrupole interaction term to the usual Lennard–Jones model for the inter-molecular potential. The band profile is in agreement with the theory of quadrupole-induced absorption. There is some increase in the intensity of the band near the critical point due to the divergence of the correlation length. A very marked increase in the intensity is possibly prevented by the "cancellation effect".

A line shape for collision-broadened lines applicable from the resonance region to the far wings is developed. An empirical correlation function is used to represent the known short and long time behavior of the true correlation function and to interpolate the unknown intermediate time regime. The resulting spectral shape is a simple analytic function which reduces to essentially the Van Vleck-Weisskopf equation when |ω−ω₀|τ₂1, but which smoothly assumes an exponential behavior in the far wings where |ω−ω₀|τ₂1. The time parameter, τ₂, is a measure of the duration of collision and is related to the mean squared torque produced in bimolecular collisions. The effect of overlapping lines, also variations in the model correlation function on the line shape are considered. The theory is applied to the absorption in the high-frequency far wing of the 4.3υ band in CO₂.

The shapes of the extreme wings of self-broadened CO₂(lines have been investigated)in three spectral regions near 7000, 3800, and 2400 cm^-1. Absorption measurements have been made on the high-wavenumber sides of band heads where much of the absorption by samples at a few atm is due to the extreme wings of strong lines whose centers occur below the band heads. New information has been obtained about the shapes of self-broadened CO₂ lines as well as CO₂ lines broadened by N₂, O₂, Ar, He, and H₂. Beyond a few cm^-1 from the line centers, all of the lines absorb less than Lorentz-shaped lines having the same half-widths. The deviation from the Lorentz shape decreases with increasing wavenumber, from one of the three spectral regions to the next. The absorption by the wings of H₂- and He-broadened lines is particularly low, and the absorption decreases with increasing temperature at a rate faster than predicted by existing theories.

Quantitative absorpance measurements have been made in pure CO₂ and mixtures of CO₂ with N₂ and O₂ in a 10 m White Perkin-Elmer cell. With absorbing paths up to 50 m-atm, results have been obtained from the band head at 2397 cm^-1 to 2575 cm^-1. The continuous absorption, which is due to the extreme tails of the strong lines in the v₃ band centered at 2349 cm^-1, is much less than calculated with the Lorentz line shape. A good fit with the data is obtained with an empirical line-shape which retains the Lorentz pressure dependence but requires a nearly exponential modification of the frequency dependence.

The absorption coefficients of the v₃, 4.3-µ CO₂ band were measured over a temperature range of about 1200° to 2400°K at various pathlengths. The hot CO₂ sample was investigated in the exhaust jet of a small supersonic burner, while the jet composition was obtained using a radiation temperature measurement and a simple thermochemical calculation. The present experimental results are compared with existing experimental values and theoretical calculations. The results indicate that above 2000°K a weak line approximation can be used to describe the absorption or emission of the band. The apparent integrated band intensity of the CO₂ fundamental as a function of temperature is also reported and discussed in relation to other measurements.

The infrared transmittance of carbon dioxide has been calculated over a wide range of path lengths, pressures, and temperatures from 500 to 10,000 cm^-1. Values of the transmittance are given at intervals of 2.5 cm^-1. In addition, transmittance values are also given which have been averaged over larger intervals. All contributing spectral lines whose relative intensity is greater than 10-8 that of the strongest line in any particular band have been included in the calculation. In addition, the contributions from the eight major isotopic species have been included. The calculation of the vibrational energy levels included terms through the third power of the vibrational quantum number and also the effects of Fermi resonanse. The final transmittance tables were generated using the the quasi-random model of molecular band absorption.

The shapes of the CO, v₃, CO₂, and v₃ N₂O fundamental vibration-rotation bands have been studied at various temperatures and in the presence of several perturbing gases. Also the half-widths of CO vibration-rotation lines have been measured at 78 K. In the region of line wings, the measured absorption coefficients deviate from those given by the superposition of Lorentzian profiles. These deviations are explained by the collision-induced line interference that causes redistribution of absorption inside the band. A theory of line mixing is formulated which is based on Markov approximation and on the strong collision model. Simple analytical expressions are obtained for the band shape. The computed shapes are in satisfactory agreement with the experimental results. The deviations from the Lorentz absorption observed in pure CO and in CO-N₂ at low temperature are partially ascribed to the formation of van der Waals dimers.

Diode-laser measurements of absorption by CO₂-N₂ mixtures in the trough between the R₆₆ and R₆₈ lines of the ¹²C¹⁶O₂ v₃-band are presented. Experiments have been performed in the 1.0–7.0 atm pressure range at temperatures of 296 and 370 K. The measured absorption is very different from that predicted by the addition of Lorentzian contributions. Theoretical calculations accounting for line-overlapping based on a fitting law are presented. Although the effects of line-overlapping are slightly underestimated, the agreement between experiments and calculations is quite good, for both the pressure and temperature dependences.

We present experimental and theoretical results on the absorption in the Ar-broadened wing of the infrared CO₂ v₃ band. Measurements have been made in the 293–765 K temperature range for total gas pressures up to 60 bar and are in good agreement with previous determinations at room temperature. Calculations have been carried out by using a line by line coupling theory presented and tested previously [J. Chem. Phys. 89, 625 (1988) and 91, 2163 (1989)]. The theoretical approach accounts for both line mixing and the wave-number dependence of the relaxation operator. It enables correct modeling of the temperature dependence of absorption. The strong sublorentzian absorption resulting from competitions between negative and positive individual line contributions is analyzed.

Fine coupling induced by collisions leads to drastic modifications of the absorption profile in the microwindows of the self- and N₂-broadened v₃ CO₂ fundamental vibration-rotation band. Calculations of these modifications have been performed by using the energy-corrected sudden (ECS) scaling law. Linewidths have been inverted to obtain effective rotation-translation basis rate constants and to deduce R-R, P-P and R-P line couplings. The characteristics of these couplings are presented and discussed. Calculated ECS profiles in the microwindows are compared with experimental data and also with previous results based on a statistical fitting law. The ECS approach is particularly suitable for practical infrared spectroscopic applications including line-coupling effects.

Dokuchaev, A. B., Pavlov, A.Y., Stroganova, E.N., Tonkov, M.V.
Effect of gas density on the shape of Q-branches of the IR absorption bands of CO₂ in the 5 μm region
Opt. Spectrosc. 60 (1986) 585.   (English Transl.)

Measurements of the absorption by pure CO₂ in the 4.3 µm region are presented. They refer to the temperature of 291 K and densities in the 1–80 amagat range. Calculations based on the addition of Lorentzian contributions strongly overestimate the absorption since they do not account for line overlapping. The energy corrected sudden approximation (ECSA) is used to model line mixing within the impact approximation. The scaling parameters of this model are deduced from line-broadening data. This calculation strongly underestimates the absorption in the spectral region near the edges of the vibrational band, whereas it overestimates the absorption in the far wings of the lines. This is attributed to the impact approximation which does not account for the frequency dependence of the relaxation operator. This dependence is roughly determined from our results; it is in good qualitative agreement with the few corresponding available results

Line broadening coefficients of the CO₂–Ar fundamental ₃ band are obtained experimentally and theoretically for temperatures from 120 to 765 K. The experimental results are obtained by Fourier-transform infrared spectroscopy. Theoretical values are provided ia quantum-mechanical (close coupling and coupled states) approaches as well as semiclassical (improved Smith–Giraud–Cooper and Robert–Bonamy) methods. The most up-to-date CO₂–Ar ab initio intermolecular potential (J. M. Hutson, A. Ernesti, M. M. Law, C. F. Roche and R. J. Wheatley, J. Chem. Phys., 1996, 105, 9130) is employed for all calculations. For all the temperatures probed, theoretical values are found to be in a rather good agreement with experiment. In addition, the Raman Q(j) line broadening coefficients and the application of the random phase approximation are presented.