Why are high- Tc superconductors, HTSC, deposited by 248 nm lasers at 400 MW/cm 2?

Abstract

HTSC are commonly deposited by pulsed laser deposition. The laser ‘operating point’ is usually near the above values. It is interesting to explore the underlying physical processes which make these values near optimum. The critical point in understanding the operating point is whether the stoichiometry of the (source) target is retained upon deposition. This question usually reverts to a question of retaining CuO as a diatomic (dissociation energy D 0 = 2.8 eV) as compared to the much more stable YO ( D 0 = 7.3 eV) or BaO ( D 0 = 5.8 eV). High temperatures obviously serve to dissociate these diatomics, with CuO being the most susceptible. First, consider the (very important) irradiance level of I ≈ 400 MW/cm 2. This value is at the lower limit of intense plume ionization, above which the ablation enters the ‘plasma controlled’ situation. This latter occurs when inverse Bremsstrahlung absorption by free electrons (10 to 20 m) above the target surface control the laser transmission and ablation dynamics. While this description is only firmly established at I ≧ 1 GW/cm 2, computer modeling [2] confirms an extrapolation down to 500 MW/cm 2. The important predictions of plasma controlled etching are: The total material transfer goes as area, A 1 / 2 (and not A − 1 / 2 ). The reason is higher fluences waste energy in heating the free electron density above the target instead of heating the target surface directly. On the other hand, one wishes to work adjacent to this border to maximize the temperature, T, and (exponential in T) vapor pressure. Avoiding intense plasma heating has additional benefits for maintaining the Cu as CuO molecules. Laser-induced fluorescence measurements indicate that the reaction Cu + N 2 O → CuO + NO requires > 10 s and is not a simple collisional process, but rather an attainment [4–8] of quasi-equilibrium at T2000 K. Combining this observation with the fact that typically < 10 −6 cm 3 or typically ∼ 10 16 atoms are removed per pulse, a static gas fill of ∼ 0.2 Torr is necessary to decelerate the expanding plume. The initial (typically 3–10 eV) beam energies are transformed into a gas temperature2500 K when stopped by and mixed with the oxidizing gas. For present gas densities this temperature is near the maximum [7] for the non-dissociation of CuO (as based on thermodynamics and densities of 3 × 10 17 Cu/cm 3). T is rapidly elevated for too intense plasmas; i.e. for I > 1 GW/cm 2. Hence, increased irradiance places one in the (undesirable) region of CuO decomposition. Furthermore, large I increases the pressure impulse on the target, which may be the critical quantity leading to particulate emission. One can now visualize the advantage of 248 nm versus longer wavelengths. At longer λ the inverse Bremsstrahlung heating of free electrons is elevated by a factor ⩾ λ 2, again unfortunately elevating T plume. At shorter wavelength (193 nm), elevated photodissociation eliminates nearby all the initial CuO; i.e. an undersirable situation. In summary, one can now make quantitative statements which point to similar laser λ and I values as found empirically.