Nonlinear materials for optical power limiting : characterization and modelling

Sammanfattning: High power laser pulses can be a threat to optical sensors, including the human eye. Traditionally this threat has been alleviated by colour filters that block radiation in chosen wavelength ranges. Colour filters’ main drawback is that they block radiation regardless of it being useful or damaging, information is lost for wavelengths at which the filter is active. Protecting the entire wavelength range of a sensor would block or strongly attenuate the radiation needed for the operation of the sensor.Sol-gel glasses highly doped with optically non-linear chromophores have previously shown high optical quality in combination with efficient optical power limiting (OPL) through reverse saturable absorption (RSA). These filters transmit visible light unless the light fluence is above a certain threshold. A key design consideration of laser protection filters is linear absorption in relation to the threshold level. A high linear absorption means that the user’s view is degraded by the filter.To model the photokinetics of RSA chromophores, the five-level population model is widely used. It consists of three singlet and two triplet levels. Model parameters relevant for OPL performance include linear absorption cross-sections, two-photon absorption (2PA) cross-sections, lifetimes, quantum yields and inter-system-crossing (ISC) times. The dominant design paradigm is to have a highly absorbing and long-lived triplet state that is quickly populated by ISC during the beginning of a laser pulse.To simultaneously achieve a lower threshold and linear absorption a vast number of materials for self-activated filters were evaluated, either as bulk glasses or solutions. An f/5 setup was used to evaluate their OPL performance while several photophysical measurements were performed to gain an understanding of system behaviour. The first three series of methyltriethoxysilane (MTEOS) Sol-Gel glasses were doped with gold nanoparticles either solely, or with one of two Pt-acetylide chromophores. One with shorter conjugated ligands, the second with similar but longer conjugated ligands. Finally, a series of multi-branched fluorene chromophores were evaluated in solution. Their central moiety was either an organic benzene unit or an ISC promoter in the form of para-dibromobenzene or a platinum(II)-alkynyl unit.For the gold nanoparticle doped glasses, the lower performance Pt-acetylide with short ligands had its OPL threshold lowered at 600nm while the glasses doped with only gold nanoparticles showed no OPL at all. Secondly, the enhancement was most pronounced for very low gold nanoparticle concentrations. While gold nanoparticles alone showed good OPL performance at 532 nm, at this wavelength neither Pt-acetylide showed an obvious OPL enhancement beyond linear absorption losses from codoping with gold nanoparticles.The improved OPL performance at 600 nm was attributed to stronger 2PA, by electric field enhancement from the gold nanoparticles. The lack of detectable OPL improvement for 532 nm and for the higher performance Pt-Acetylide chromophore with long ligands were qualitatively explained by a lower sensitivity to 2PA on system performance. A degraded performance from linear absorption by excess nanoparticles in front of the focus explained the weakening of the enhancement at higher gold nanoparticle concentrations.All three fluorene chromophores, including the chromophore without a central ISC promoter, showed broadband OPL through the visible spectrum. The OPL performance of the two chromophores with ISC promoters was expected considering their transient absorption at microsecond time-scales. For the fluorene chromophore without an ISC-promoter, ultra-fast transient absorption was used to identify singlet excited state absorption as the source of the OPL performance.Both of these series of experiments demonstrate how a simplistic view of simply increasing desired photophysical parameters, e.g. effective 2PA cross-section or ISC quantum yield, do not always result in a noticeable increase in system performance. By employing numerical population models it was possible to identify which parameters had the highest impact on OPL performance.