Probing Ultrafast Dynamics of Bacterial Reaction Centers Using Two-Dimensional Electronic Spectroscopy

Abstract

In the initial steps of photosynthesis, solar energy is converted to stable charge separated states with high efficiency. Understanding the relationship between structure and function in the photosynthetic reaction centers where these conversion steps take place could guide the development of more efficient artificial light harvesting systems. Reaction centers are complicated pigment-protein complexes with multiple spectrally overlapped absorption bands, making interpretation of spectroscopic data challenging. The sub-picosecond time scales involved in the energy transfer and charge separation processes present another challenge. Two-dimensional electronic spectroscopy (2DES) has proven to be a powerful tool for disentangling features in spectrally congested systems like reaction centers by resolving the optical response with respect to excitation and detection frequencies. 2DES also obtains the excitation frequency dependence without sacrificing time resolution, which is necessary to resolve energy transfer processes in reaction centers occurring on time scales faster than 100fs.

We perform 2DES on bacterial reaction centers (BRCs) from the purple bacterium Rhodobacter capsulatus, using a degenerate optical parametric amplifier producing 12fs pulses with bandwidth spanning the broad near-IR absorption bands of the BRC. The 2D spectra are analyzed using several global analysis methods to extract the underlying energy transfer and charge separation kinetics, and we compare the results to published transient absorption studies on BRCs. Commonly used 2DES global analysis techniques proved inadequate for resolving specific branched and parallel reaction mechanisms. We developed an improved 2D kinetic fitting approach which employs a common set of basis spectra for all excitation frequencies, and uses information from the linear absorption spectrum and BRC structure to model the excitation frequency dependence of the 2D spectrum. Using the improved fitting method, we show that the entire time-dependent 2D spectrum is well-represented by a sequential reaction scheme with a single charge-separation pathway. We tested several proposed alternative reaction schemes involving branched charge separation pathways, and did not find compelling evidence from our data that favors a particular branched model. Based on this analysis, we conclude that our data supports the simpler, single pathway charge separation model.