ALBERT

Notes: This paper addresses the separation of the contributions to the visible refractive index of colorless liquids from electronic (ultraviolet) and vibrational (infrared) absorption. The goal is to find the most accurate infrared values of nel(ν˜), the refractive index that results solely from electronic absorption, by fitting and extrapolating currently available visible refractive index data. These values are needed, interalia, to improve the accuracy of infrared real refractive index spectra calculated by the Kramers–Kronig transform of infrared imaginary refractive-index spectra. The electronic molar polarizability αel(ν˜) is calculated from the values of nel(ν˜) at wave numbers between 20 500 and 0 cm−1. The methods are applied to ten liquids: H2O, D2O, CH3OH, CH3COOH, CH3CN (CH3)2CO, CH2Cl2, C6H6, C6H5Cl, and C6H5CH3. The visible refractive indices are expressed as power series in wave number, by expansion of the Kramers–Kronig integral. Terms in ν˜+2m, m=1,2, are due to the electronic contribution and terms in ν˜−2m are due to the vibrational contribution.The vibrational contribution to the visible refractive index is also calculated from experiment by Kramers–Kronig transformation of the known infrared imaginary refractive index spectrum of the liquid. It is shown that the vibrational absorption contributes ≥0.001 to the visible refractive index only for the four hydrogen-bonded liquids, and that, for all ten liquids, at least 25% of the vibrational contribution arises from absorption below 2000 cm−1. If the vibrational intensities are not known, the available visible refractive indices yield the most accurate infrared values of nel for all liquids except H2O if they are fitted to the equation n=a0+a2ν˜2+a4ν˜4. A similar equation, with the additional term a2ν˜−2, is theoretically superior because the latter term adequately describes the vibrational contribution to the visible refractive indices, but only for H2O are the currently available visible refractive indices sufficiently accurate and sufficiently extensive to allow the four coefficients in the equation to be determined with useful accuracy. For H2O, D2O, CH3OH, CH2Cl2, C6H6, C6H5Cl, and C6H5CH3, corrections are given to slightly improve the accuracy of the previously published infrared real refractive-index spectra. © 1995 American Institute of Physics.

Notes: A new and simple procedure is presented for the calculation of the infrared real, n, and imaginary, k, refractive index spectra from s-polarized attenuated total reflection (ATR) spectra by a modified Kramers–Kronig transform of the reflectance to the phase shift on reflection. The procedure consists of two parts, first a new modified Kramers–Kronig (KK) transform, and second a new, wave number-dependent, correction to the phase shift. The procedure was tested with ATR spectra which were calculated from refractive index spectra that were synthesized under the classical damped harmonic oscillator model. The procedure is far more accurate than previous procedures for the real case of a wave number-dependent refractive index of the incident medium, and yields n and k values that are accurate to ≤0.1% provided that no errors are introduced by the omission of significant reflection bands. This new procedure can be used to obtain optical constants from any ATR experiment that yields the spectrum of Rs, the reflectance polarized perpendicular to the plane of incidence. In this laboratory Rs spectra are obtained from samples held in the Spectra-Tech CIRCLE cell in a Bruker IFS 113 V spectrometer. Accordingly the ATR spectra used to test the new procedure were calculated for the optical configuration of this system, which is m reflections at 45° incidence with equal intensities of s- and p-polarized light and retention of polarization between reflections. For the previously studied [J. S. Plaskett and P. N. Schatz, J. Chem. Phys. 38, 612 (1963); J. A. Bardwell and M. J. Dignan, ibid. 83, 5468 (1985)], but unreal, case of constant refractive index of the incident medium, n0, the new transform gave better results than either of two previously studied procedures. In this case the phase shift at each wave number was corrected by a constant which ensured that the correct phase shift was obtained at the highest wave number in the transform, 7800 or 8000 cm−1. In contrast to a previous study [J. Chem. Phys. 83, 5468 (1985)] it was found that the normal KK transform is inferior for this case to a previous modified KK transform [J. Chem. Phys. 38, 612 (1963)], and it is also inferior to the new modified KK transform. Further, the new transform has only the usual singularity of a KK transform, and this makes it numerically superior to the previous modified KK transform which has an additional singularity at 0 cm−1. For the real case, in which the refractive index of the incident medium changes with wave number, the new transform was used with a new simple wave number-dependent additive correction to the phase shift. This new correction is calculated with the actual value of n0 at each wave number. For molecular liquids such as methanol and benzene the new transform is markedly superior to the previous two transforms. It yields real and imaginary refractive index values that are accurate to better than 0.1% provided the reflection spectrum is known down to 2 cm−1. The latter condition is rarely fulfilled, and the effect of the omission of low wave number bands is illustrated. A method to reduce the impact of missing low-wave number parts of the reflectance spectrum is described, and its effectiveness is illustrated. © 1996 American Institute of Physics.

Notes: In this paper a multireference constant denominator perturbation theory (CDPT) is developed to reduce incomplete basis set errors arising when solving the Schrödinger equation with a finite basis set. The advantage of this method is that very few basis functions are needed, and all calculations if carried out to high enough order in the perturbation treatment effectively use a complete basis set. As a first step the theory has been restricted to one-particle Hamiltonians and applied to the anharmonic oscillator to study the convergence properties. For perturbation calculations carried out to fifth order, results from Pade approximates show an improvement in accuracy of between one and three orders of magnitude.