Search

We shall call Far-Infrared (FIR), the part of the electromagnetic (e.m.) spectrum ranging from 30 μm to 3000 μm. A short history of 50 years of research in the Far Infrared: 1952-2002. is worthwhile to consider now that the field has so much progressed. It is possible to write this story from published documents and from our own memory of the major events. It has to be supplemented by our colleague's own souvenirs.

New experiments show that the solution of the complicated problem of Far Infrared (FIR) absorption by crystals of very low symmetry, illuminated at quasi-normal incidence, might have a simple solution, which has been suggested now that a number of spectroscopic studies on monoclinic Triglycine Sulphate (TGS) are available: i) in a first classical approximation where only the strongest bands are considered, the absorption spectrum depends only on the direction of the electric field Eof the IR radiation, and Transverse Optic (TO) vibrations are directly excited with electric dipole variations dpparallel to E. ii) in a second approximation looking only at the weakest absorption bands, it depends only on the direction of the wave vector kand Longitudinal Optic (LO) vibrations are excited with dpparallel to kand extinction coefficients 3 orders of magnitude smaller.

In the parts of the Far IR (FIR) spectrum where Ferroelectric Triglycine Sulphate (TGS) is transparent for a given direction of the electric vector [image omitted], the forced transverse vibrations that can be excited in that direction are shown to couple slightly with the Longitudinal Optical (LO) vibrations propagating along the wave vector [image omitted] (wave normal) of the FIR wave. Very clear LO phonons FIR transmission spectra at quasi-normal incidence with zero transmission sharp lines can thus be obtained at low temperature with thick samples, while classical spectra of Transversal Optical (TO) phonons are obtained with thin crystal plates. This coupling seems a second order effect, quite new, applicable to all low frequencies LO phonons of TGS and probably other crystals of low symetry. On the opposite, no coupling has been detected with an other direction than the [image omitted] vector.

The modification of domain texture in triglycine sulphate (TGS) is observed during different ferroelectric loops. Nucleation and domain wall motions are easily observed and they generally involve a small number of domains except at high frequencies. The dielectric constant shows important variations along the loop from ε = 28 to ε = 55 at room temperature and at low frequency (0.005 Hz). The smallest value corresponds to one single domain and a high dc field (2.8 MV/m), the highest one to a polydomain crystal. At higher frequencies the number of domains and the maximum of the dielectric constant are much higher (i.e. ε = 260 at 1600 Hz).

A new cryostat, a Fourier transform grid-spectrometer, and a new interferometric method have led to more accuracy for transmission spectra of YBaCuO films carefully deposited on MgO plates, than reported before in the 10–40 cm−1very far infrared (VFIR) range.

The Miller-Savage law is verified for low intensity electric fields and the power law for high fields, in the case of thick samples. For very thin crystal plates (e < 10 μm), the Miller-Savage law must be replaced by an exponential law. In all cases the different parameters can be kept constant in a small range of thickness by using an "effective electric field."

We have studied the Far Infrared transmission spectra of an MgO plate covered with an YBaCuO film of thickness t=200 Å, at different temperatures. By cooling from room temperature to temperatures Θ < 140 K, the Far Infrared interference fringes are displaced by half a period. Explanation in terms of different adaptations of the YBaCuO sheet resistance to the MgO substrate impedance during cooling.

In a very rough experiment using two MgO plates, coated with a 200 Å thick YBaCuO film to increase their reflectivity, and facing each other their YBaCuO films to make a Perot-Fabry (PF) interferometer, we have observed a 43% maximum transmission and a finesse around 33 at v=29 cm−1, in good accordance with a new phenomenological model for YBaCuO. Much higher performances can be expected by optimizing the YBaCuO film thickness.

Far IR transmission spectra of very thin YBaCuO films in the 10–40 cm−1 range have now sufficient accuracy and reproducibility to show that the number of quasiparticles at temperature Θ<Θc, Θc being the transition temperature, do not follow a Θ4 law, but rather a Θ1.5 one.

Conclusion We have presented the possibility to make optical coatings on quartz in the FIR and we have got a transmission coefficient as high as 97% at room temperature. It will rise probably close to 100% at liquid helium temperature where the quartz two-phonon difference processes are frozen. The use of polyethylene or TPX is also possible to make plates and lenses. The reflectivity losses are smaller because the refractive index is much smaller, but they cannot be further reduced by any coating because there is no transparent material with n≈1.2 in the FIR. These lenses are very useful in the FIR (except around 60 cm−1 if they are made of polyethylene). TPX is quite transparent over the whole FIR range, and also in the visible with nearly the same refractive index as in the FIR. Alignments can be made with visible light, and still work in the FIR.

The dielectric properties of MgO single crystals are carefully measured in the microwaves and compared with the predictions given by a dynamical model made of 4 Lorentz oscillators with parameters fitting quite well the IR spectra. The microwave losses can be made very low (e.g. 4 × 10−7 for T=80 K and f=8 GHz), with the same order of magnitude expected from the model. It is a hint that Debye relaxation of small amounts of impurities (e.g. Mn which is sometimes detected by ESR analysis), bring a negligible contribution, and MgO crystal plates are indeed a good support to grow High Tc superconducting films.

Finesse as high as 260 and peak transmission close to 98% at frequencies around 30 cm−1 are predicted at 10 K by coating the Perot-Fabry (PF) mirrors with a suitable thickness of High Tc (HTC) superconductor YBaCuO. In fact, Finesse will be limited by the lack of parallelism and the lack of perfect flatness of the mirrors to values around 100. These predictions are based upon phenomenological models for both the YBaCuO coating and the MgO support, i.e. from direct transmission measurements of MgO and YBaCuO in the Far IR, which had to be fitted by the models.

Abstract: The different technical approaches are considered mostly to compare the accuracy attainable in microwaves to Far IR where new resonant tchniques have been developped. It is possible to use the fringes of the observed spectra to get directly the London penetration length. However in the present state of the Art there is more accuracy to use a two-fluid model with at least one Mid-IR oscillator, and a strategy is possible to fit separately the different parameters. The different results are reviewed and compared to the microwaves data. There is a good accordance, except perhaps for the power law-giving the concentration of quasiparticles in the superconductive phase. For a thin film of YBCO devoid of grain-boundaries (no twins), a θ4 law is found instead of a θ1.5 law for an ordinary film. It should be highly valuable that both microwaves and Far-IR data should be taken on the same sample.

A number of far infrared absorption bands are found in molecular and hydrogen-bonded crystals when there are several formula units in the primitive cell. They are very sensitive to the crystal structure and constitute a convenient probe for looking at small changes in the structure, especially at low temperatures where there is a lack of X-Ray data. As well known examples, the case of LTT (Lithium Thallium Tartrate) where the absorption band observed at 21 cm-1 at room temperature softens down to 8 cm-1 at the Curie temperature (Tc = 12 K), and LiNH4SO4 which has a phase transition at 26 K, will be considered.It is less known that some ferroelectric crystals can be studied as very thick single crystal plates (thickness t up to 10 mm) at 4 K, and they show in some cases a far IR transmission that is much higher than expected from the study of thin plates (i.e. t = 20 μm). In fact the transmission does not decrease very much when thickness is higher than some specific value t0/2. The crystal is not homogeneous at 4 K. A model with a far IR absorbing surface layer (thickness t0/2) and a transparent bulk is a good first approximation.At some temperature T0 located between liquid helium and liquid nitrogen temperatures, the center of the bulk undergoes a phase transition from the absorbing phase into the transparent one. When temperature is still lowered, the transparent phase is extended towards the surface with an incommensurate phase between the surface and the bulk.

Far infrared spectra of TGS from room temperature to liquid helium temperature in the spectral range extending from 3 to 400 cm-1, and description of a triglycine sulphate pyroelectric detector working at frequencies as high as 1 Megacycle per second.

Absorption spectra at liquid helium temperature in the far infrared are given for thick plates of CsI, CsBr, TlBr, TlCl, KI, NaCl, InSb, KBr, CuCl, Al_2O_3, SiO_2, MgO, and LaCl_3. For a thickness of about 3 mm all these crystals, completely opaque at room temperature, become quasitransparent at very low temperature. This supertransparency of crystals at low temperature gives several far infrared filters with remarkable cutoffs and a lot of available materials to the spectroscopist in a part of the spectrum where only quartz and polyethene were usable up to now. This is also evidence that the main part of the absorption at room temperature comes from the lattice phonons. It decreases with their number when the crystal is cooled (2 phonons difference process). The small absorption which remains in a few cases—KBr, —NaCl, KI—could arise from 2 phonons addition process nearly independent of temperature and much less efficient to give absorption. In the case of glasses the absorption coefficient is higher by 2 or 3 orders of magnitude and keeps constant when the glass is cooled. For quartz glass, the index of refraction also seemed invariant. An explanation is proposed in terms of one single phonon process, independent of temperature and allowed by the disorder characteristic of glassy states.

When a laser spot is focused on a Triglycine Sulphate plate used as a pyroelectric detector, a non-linear response is observed when the illumination is above a threshold of about 40W.cm-2. It appears that the items of primary significance are:(a) weak laser pulses can cause nucleation of domains whose polarization are opposed to the bulk polarization of the crystal.(b) this new process is elastic, polarization recovers at termination of irradition.(c) it can be inhibited by large electric fields.(d) it is time dependent in accordance with heat propagation to a nucleation site and nucleation growth processes of a domain.Application to pyroelectric detectors are considered.

The reflection spectra of AgCl, AgBr, and AgI are measured at liquid hydrogen, liquid nitrogen, and room temperatures. For AgCl, they show distinctly two components. For AgBr and AgI, the structure, though complex, is not resolved. Nevertheless, in all cases, except for AgI, at low temperature, the Lorentz analysis needs two oscillators. The stronger one ν_1 corresponds to the TO transition at point Γ, the second one ν_2, about an order of magnitude smaller, corresponds to a TA + TO addition band for the NaCl type structure, and to a TA + LA for the zinc blende or wurtzite structure, confirming preliminary measurements made by a completely different experimental approach (transmission of thin evaporated films instead of reflection on bulk crystals). They are, however, slightly distorted towards high frequencies. The proposed frequencies for the TO branch at point Γ are, for AgCl: 105 cm^−1 (290°K), 114 cm^−1 (80°K), 119 cm^−1 (25°K); for AgBr: 79 cm^−1 (290°K), 84 cm^−1 (80°K), 85 cm^−1 (25°K), and for AgI: 103 cm^−1 (290°K), 106 cm^−1 (80°K), 106 cm^−1 (25°K). The oscillator strengths and the damping of both oscillators are studied vs temperature. A large decrease of the damping is still observed between liquid nitrogen and liquid hydrogen temperatures. The refractive index and the absorption index are determined in the whole ir for the first time at liquid hydrogen, liquid nitrogen, and room temperatures, showing large variations, particularly a high transparency in the very far ir at liquid hydrogen temperature. The maximum values of n and k range from 2 to 10. The real and the imaginary parts of the dielectric constants are also computed for the whole ir range at room and lower temperatures down to liquid hydrogen temperature.

A special dichroism is observed in the far-infrared spectra of triglycine sulfate crystals. The absorption of a few peculiar lines is dependent on the orientation of the sample surface while, indeed, the infrared electric vector is kept in a constant direction. A significant broadening of some lines in a plate cut perpendicular to the axis is explained by a heterogeneous electric field in the surface layer. New lines appearing only in a plate cut parallel may be interpreted either by a loss of symmetry in the surface parallel to the binary C_2 axis or by more classic arguments derived from the complexity introduced by a high absorption index in a very anisotropic crystal. This dichroism is observed in reflectivity and, when transmission is considered, for some lines the dichroism is limited to very thin samples in good accordance with a surface layer explanation, while for other lines it is still observed for thick samples.

High-resolution spectra from 20 to 200 cm^−1 for temperatures from 10 to 300 K give interference fringes that cannot be fitted by previous analysis in terms of four Lorentz oscillators. Parameters in the theory are revised to give a better fit both to the far IR and to the reststrahlen spectra.

Abstract: In an Alanine doped triglycine sulphate ferroelectric crystal, domain wall motions show roughly the same velocity vs. electric field as in pure TGS (after the specific bias field of the observed region has been annealed by an opposite external field). The different steps of the hysteresis loop can be brought together with domain switchings which have been made visible by the pyroelectric probe technique.