Kaila Yallum1,Natalie Banerji1,Julien Réhault1
Universität Bern1
Kaila Yallum1,Natalie Banerji1,Julien Réhault1
Universität Bern1
Femtosecond transient absorption (TA) allows us to observe photophysical processes on the pico-, nano-, and micro-second timescales with femtosecond time resolution. Broadband white-light generation and detection determine the scope and limits of a TA setup. With the goals of 1) extending the white light into the NIR, and 2) removing the gap in the 750-850 nm range, a different white light generation scheme was employed consisting of a 800 nm-pumped optical parametric amplifier (OPA) to generate an idler around 2000 nm. The 2000 nm idler is used as the driving beam in a Yttrium Aluminum Garnet (YAG) crystal to generate a white light supercontinuum that spans a range starting at 600 nm up to 1700 nm. Because the white light generation relies on the OPA, additional shot-to-shot noise is introduced yielding shot to shot noise of up to 15%. Such significant noise contributions prevent this generation mechanism from being useful for TA experiments with a classic ratiometric referencing scheme. Therefore, an improved referencing scheme was employed in the signal processing to recover the signal to noise ratio<sup>1</sup>.<br/><br/>In short, the proposed B-matrix referencing scheme relies on a cross covariance matrix, <b>B</b>, to predict the change in intensity between two pumped shots, ��<i>I<sub>r</sub><sup>*</sup></i>, for each wavelength. The difference between the measured change in intensity, ��<i>I<sub>m</sub><sup>*</sup></i>, and the reconstructed ��<i>I<sub>r</sub><sup>*</sup></i>, is taken to isolate the sample response, <i>S</i> = ��<i>I<sub>m</sub><sup>*</sup></i>- ��<i>I<sub>r</sub><sup>*</sup></i>. Calculations to determine <b>B</b> can be carried out by a series of unpumped shots measured before the experiment, therefore removing the need for a chopper, or it can be simply calculated by the unpumped shots measured when the chopper is in the closed position. In practice, the B-matrix referencing method reduced the root mean square error in detection by a factor of 40 when compared to the ratiometric referencing scheme.<br/><br/>The case study presented herein focuses on the benefits of the white-light generation and noise reduction methods employed on a spectroscopic study of the organic photovoltaic (OPV) blend PM6:Y6. Newer generation organic semiconducting materials for photovoltaic applications have been chemically modified to absorb closer to the red edge of the visible light spectrum<sup>2</sup>. Because of this, the spectra of triplet and charged states of these materials are often observable in the NIR. Furthermore, the ground-state bleach (GSB) associated with these materials falls into the detection gap brought about by the optical filters needed to remove the 800 light. Furthermore, the overlapping GSB and excited-state absorption result in a differential absorption of the exciton in the linear regime of often less than 1 mOD, sometimes as low as 50 ��OD. The referencing scheme employed in this unique TA setup successfully represses the noise in order to recover such small signals.<br/><br/>1. Y. Feng, I. Vinogradov, and N-H. Ge, <i>Optics Express</i> (<b>2017</b>) Vol. 25, No. 21, 26263<br/>2. J. Hou, O. Inganäs, R. H. Friend, F. Gao, <i>Nature MaterialsI </i>(<b>2018</b>) Vol. 17, 119-128.