Direct Evidence for Excitation Energy Transfer Limitations Imposed by Low-Energy Chlorophylls in Photosystem I–Light Harvesting Complex I of Land Plants

The overall efficiency of photosynthetic energy conversion depends both on photochemical and excitation energy transfer processes from extended light-harvesting antenna networks. Understanding the trade-offs between increase in the antenna cross section and bandwidth and photochemical conversion efficiency is of central importance both from a biological perspective and for the design of biomimetic artificial photosynthetic complexes. Here, we employ two-dimensional electronic spectroscopy to spectrally resolve the excitation energy transfer dynamics and directly correlate them with the initial site of excitation in photosystem I–light harvesting complex I (PSI-LHCI) supercomplex of land plants, which has both a large antenna dimension and a wide optical bandwidth extending to energies lower than the peak of the reaction center chlorophylls. Upon preferential excitation of the low-energy chlorophylls (red forms), the average relaxation time in the bulk supercomplex increases by a factor of 2–3 with respect to unselective excitation at higher photon energies. This slowdown is interpreted in terms of an excitation energy transfer limitation from low-energy chlorophyll forms in the PSI-LHCI. These results aid in defining the optimum balance between the extension of the antenna bandwidth to the near-infrared region, which increases light-harvesting capacity, and high photoconversion quantum efficiency.

The expanded 2DES maps ( Figure S3a) show the persistence of dynamics along the diagonal pump/probe axis, whereas most of the evolution results in an offdiagonal cross peak(s), centered, initially, at 680 and then further shifting to 700 nm for pumping in the bulk absorptions (λ pump <700 nm). Figure S3b shows an expanded view of the 2D-DAS (for open centre conditions) in the red forms spectral window. It is noticeable that in the fastest DAS (0.3 ps) there is an inversion of sign, due to 4 prompt depopulation of the red forms by energetically uphill energy transfer to the bulk, whereas a downhill energy transfer to these states is observed for λ pump <700 nm.
The inversion of the amplitude sign is also present in the 3 ps 2D-DAS, indicating a slower kinetic component reflecting the pre-equilibration of the bulk absorption from the red forms. Furthermore, the persistence of depopulation dynamics along the pump/probe diagonal is also observed in this spectral window in 2D-DAS associated with the 12 ps, 56 ps and, even more clearly, in the >160 ps lifetimes. This further spectral-temporal evolution corresponds to the depopulation of red-forms having absorption maxima at ~705 and ~710 nm respectively. The progressive slowdown of de-excitation correlating with the red-shift of different long-wavelength absorption forms is in excellent agreement with previous excited state relaxation dynamics studies 2-6 .

Decay associated spectra (DAS) at selected pump-wavelengths
To facilitate the comparison of the DAS obtained from the analysis of 2DES data with those obtained previously from classical TA techniques with preferential excitation at defined wavelengths, slices of the 2D-DAS map ( Figure 2  The mean decay lifetime ( ) is defined as the first moment of the normalised m  decay distribution. When the decay function is a linear combination of exponentials, the moment has an analytical solution, given by:

Section V: Reproducibility of the experimental results and analysis.
In the following paragraph data are presented for an independent set of measurements performed on the same PSI-LHCI preparation. The measurement conditions were the same as in the data presented in the main text, but the pulse spectrum was further shifted to the blue (down to 550 nm) and, as a consequence, had lower intensity in the near infrared above 700 nm. Moreover, t 2 kinetics were acquired only up to 40 ps. Therefore, slowly decaying component are less dynamically The 2D-DAS and the associated lifetimes obtained in an independent measurement set and employing slightly different acquisition conditions, closely resemble those reported in Figure 3 of the main text. The main difference relates to the long-lived components which are not fully resolved in Figure S9 (with respect to Figure 3) because of the shorter t 2 scale employed in these measurements.
It can nonetheless be appreciated that there are specific spectral features characterising the long-lived component (>40 ps) that are specific for open and closed centre conditions. In the first case, a broader GSB contribution is retrieved having maximum intensity at ~700 nm. The corresponding 2D-DAS spectral feature narrows and red-shifts upon centers closure. This difference is interpretable in terms of the disappearance of the long-lived signal due to P 700 + under closed centers, when only the slowly decaying excited relaxation of the low energy state is observed instead.
In the 2D-DAS reported in Figure 3 the contributions of EET from the redforms and the long-lived (non-decaying) contribution from P 700 + are better, but not fully, separated. This is because in PSI-LHCI supercomplex of higher plants the relaxation of the low energy state has been shown to be polyphasic, often characterised by lifetimes of ~30-40 ps and 60-80 ps, respectively. The latter lifetime is clearly resolved under closed centers conditions ( Figure 3 and Figure S3), but it is cumbersome to fully deconvolute it from the non-relaxing P 700 + , which is known to decay on a millisecond time scale in the isolated photosystem.