Tracking Charge Transfer to Residual Metal Clusters in Conjugated Polymers for Photocatalytic Hydrogen Evolution

Semiconducting polymers are versatile materials for solar energy conversion and have gained popularity as photocatalysts for sunlight-driven hydrogen production. Organic polymers often contain residual metal impurities such as palladium (Pd) clusters that are formed during the polymerization reaction, and there is increasing evidence for a catalytic role of such metal clusters in polymer photocatalysts. Using transient and operando optical spectroscopy on nanoparticles of F8BT, P3HT, and the dibenzo[b,d]thiophene sulfone homopolymer P10, we demonstrate how differences in the time scale of electron transfer to Pd clusters translate into hydrogen evolution activity optima at different residual Pd concentrations. For F8BT nanoparticles with common Pd concentrations of >1000 ppm (>0.1 wt %), we find that residual Pd clusters quench photogenerated excitons via energy and electron transfer on the femto-nanosecond time scale, thus outcompeting reductive quenching. We spectroscopically identify reduced Pd clusters in our F8BT nanoparticles from the microsecond time scale onward and show that the predominant location of long-lived electrons gradually shifts to the F8BT polymer when the Pd content is lowered. While a low yield of long-lived electrons limits the hydrogen evolution activity of F8BT, P10 exhibits a substantially higher hydrogen evolution activity, which we demonstrate results from higher yields of long-lived electrons due to more efficient reductive quenching. Surprisingly, and despite the higher performance of P10, long-lived electrons reside on the P10 polymer rather than on the Pd clusters in P10 particles, even at very high Pd concentrations of 27000 ppm (2.7 wt %). In contrast, long-lived electrons in F8BT already reside on Pd clusters before the typical time scale of hydrogen evolution. This comparison shows that P10 exhibits efficient reductive quenching but slow electron transfer to residual Pd clusters, whereas the opposite is the case for F8BT. These findings suggest that the development of even more efficient polymer photocatalysts must target materials that combine both rapid reductive quenching and rapid charge transfer to a metal-based cocatalyst.

. Hydrogen evolution experiment using P3HT nanoparticles with a 40 nm diameter and 5 wt. % photodeposited Pt in a 1:1:1 H2O/methanol/triethylamine mixture, irradiated by the output of a 300 W xenon lamp (385 -700 nm). Note that the hydrogen evolution activity of P3HT is very low and in fact close to the resolution limit of our instrument, therefore there is visible noise in this P3HT HER data.

Calculation of exciton diffusion lengths via singlet exciton−singlet exciton annihilation (EEA)
Femtosecond-transient absorption data obtained from F8BT nanoparticles (<1 ppm Pd, ~62 nm diameter) upon excitation with various fluences (1 -150 μJ cm -2 ), and was used to calculate the diffusion length of F8BT excitons in the absence of Pd. The model used assumes that exciton decay via radiative and non-radiative deactivation with intrinsic exciton decay rate constant (k) and via bimolecular EEA 1-3 with a bimolecular decay rate coefficient (γ), as shown in Equation S1, where n(t), in cm -3 is the exciton density at a given delay time t after laser excitation, (Eq. S2) A useful indication of the rate of a first-order chemical reaction is the half-life time t1/2 of photo-generated excitons, i.e. the time after which the exciton population has decayed to half of its initial value.
The time for n(t) to decrease from n0 to 1 2 n0 in a first-order reaction is given by
We can then rearrange this Eq. S2 using the half-life time of the exciton decay, It follows that The modified EEA bimolecular decay rate coefficient was obtained from Eq. S6 and was calculated to be to be γ = 7.93×10 -9 cm 3 s -1 at a fluence of 2.5 µJ/cm 2 .
= 4 (Eq. S8) We used an annihilation radius Ra = 2.88 nm corresponding to the distance at which singlet-singlet exciton annihilation is faster than exciton diffusion. D = 2.29×10 -3 cm 2 s -1 was obtained by using Equation S8. [1][2][3] The exciton diffusion length (Ld) of 9.2 nm can be calculated using, The transient absorption spectra probed before 1 ns shown in Figure S3a exhibit a negative ΔA signal below 610 nm, which can be assigned to a mixture of ground state bleach and stimulated emission. The positive signal above 610 nm is assigned to excited state absorption of photogenerated excitons, and this signal can therefore be used to track the exciton population. Figure S3a shows the decay kinetics probed at 700 nm for aqueous nanoparticle dispersions with different Pd contents. We estimate that 10% of photogenerated excitons are quenched by Pd before 100 fs (the time resolution of the instrument) in 1170 ppm F8BT compared to <1 ppm Pd F8BT based on the initial signal amplitude. were counted over a time period of 10 s. We note that an increase in the initial signal amplitude is observed for a-c in the sample with DEA, which we attribute to the evaporation of some of the relatively volatile DEA during the argon purge before the measurement. The full lines represent fits to the data obtained using iterative reconvolution of the measured instrument response and a stretched exponential function y ∝ exp(−kt b ) with a stretching exponent b = 0.87. Figure S5. Photoluminescence decays of unpurified F8BT nanoparticles of different sizes suspended in H2O, probed at 545 nm upon 467 nm excitation. All suspensions were prepared with an absorbance of 0.05 at the excitation wavelength to ensure an equal number of absorbed photons. The full lines represent fits to the data obtained using iterative reconvolution of the measured instrument response and a stretched exponential function y ∝ exp(−kt b ) with a stretching exponent b = 0.87. We note that the decays in the presence of DEA follow the same trend, further demonstrating that quenching via Pd dominates quenching via DEA in unpurified F8BT (as seen in Figure S4).  7. X-Ray absorption spectroscopy (XAS) experiments X-Ray absorption spectroscopy (XAS) was used to further characterise the metal clusters in P10 synthesized via a Suzuki-Miyaura polycondensation reaction and via Yamamoto coupling (P10Y). We studied as-synthesized P10 containing 8000 ppm Pd (similar to the 7000 ppm Pd sample discussed in the main manuscript), P10 containing 10,000 ppm Pd following photodeposition (derived from assynthesized P10 with 2000 ppm Pd), and P10Y.

TEM images
In the XAS data processing, the photon energy of the collected XAS spectra was calibrated at the first inflection point of the derivative XAS spectrum of Pd or Ni foil, and then normalized by Athena (version of 0.9.26) software. The backgrounds in the regions of pre-edge and polynomial post-edge in these sample XAS spectra were normalized by subtracting from the raw If/I0 data (If: Intensity of reflected light projected on the filter in fluorescence mode; I0: Intensity of initial light from light source). In addition, the extended X-ray absorption fine structure (EXAFS) spectra of these samples were mathematically converted to k 2 -weighted in the region of 4.1-12.1 Å -1 . Subsequently, these k 2 -weighted EXAFS spectra were Fourier-transformed (FT) to FT-EXAFS spectra to analyze the coordination configuration, including coordination number, bond distance, and Debye-Waller factor of Pd and Ni in samples. For the model fitting in Artemis software (version of 0.9.26), the crystal structure parameters of aforementioned palladium foil and several nickel standards for model fitting were downloaded from the online Inorganic Crystal Structure Database (ICSD) (https://icsd.fiz-karlsruhe.de/search/basic.xhtml).
With the aid of XAS spectra, the valence of Pd and Ni in samples can be identified by comparing their XAS spectra with that of other metal references. As shown in Figure S9a, the XAS spectra are composed of three parts, including pre-edge (-150 to -50 eV), edge/white line (-50 to 50 eV), and postedge (50 to 250 eV) regions (The inflection point in edge region is set as the datum, 0 eV). The signal of edge/white line region in these XAS spectra is contributed to the electron excitation from 1s to 4p orbital which can be adopted for the valence identification for elements. However, the inflection point shift in these samples are unobvious in our case that can only identified with the feature variation of white line region, as shown in the inserted figure in Figure S9a. The sequence of white line intensity is P10-photodeposited>P10-synthesis proportional to their Pd amounts. Apart from Pd amount, the white line intensity is also proportional to the d-orbital empty and coordination number of Pd in samples. Therefore, the most Pd atoms in these P10 samples existed as the valence of Pd(0) and have the relatively lower coordination number compared with Pd foil, revealing that Pd atom is well-dispersed by P10 polymer. In addition, the higher absorption peak intensity at lower photon energy in the white line region also refers to the existence of oxidized Pd atoms. 4,5 That is, a few Pd atoms were bound with the oxygen atoms of P10 polymer, forming PdOx-based clusters with a higher Pd valence. In order to obviously depict the photon energy position of inflection points in each of the samples for Pd valence identification, their first derivative normalized XANES spectra are shown with that of Pd foil in Figure  S9b. The inflection point position of samples is basically located at 24350 eV which is same as that of Pd foil, implying that the average Pd valence in various samples is Pd(0). Furthermore, the coordination environment of Pd in these P10 samples and Pd foil were compared in forms of k 2 -weighted XANES spectra, as displayed in Figure S9c. In the region of wavenumber higher than 2.5 Å -1 , the oscillation frequency of samples was same with that of Pd foil, suggesting that their Pd coordination environment was metal-like. Most particularly, the oscillation frequency of the P10 samples are higher than that of Pd foil in the region of wavenumber lower than 2.5 Å -1 . This oscillation frequency difference is possibly contributed by the intermolecular correlation between Pd and surrounding P10 polymer. Figure S10 exhibits the bond distance of first coordination shell, Pd-Pd in the studied P10 samples and in Pd foil. In comparison with Pd foil, the Pd-Pd intensity of the P10 samples were relatively lower (0.5-1.0 Å -3 ) than that of Pd foil (~3.0 Å -3 ) because of their lower Pd coordination number which can be ascribed to a P10 polymer dispersion effect. On the basis of this principle, the Pd coordination number of P10-Photodeposited was the highest among these samples. The coordination number and bond distance of Pd in samples were derived from the model-fitted FT-EXAFS spectra which show good agreement between fitting curve and FT-EXAFS spectra for the P10 samples and Pd foil, especially at the first shell, Pd-Pd. The model-fitting results of FT-EXAFS spectra for samples and Pd foil are summarized in Table S2. In comparison to Pd foil, the coordination number of Pd in the P10 samples were lower, indicating that the Pd atoms were well-dispersed by P10 polymer. With a lower Pd amount and a specific synthesis route, the P10 polymer can more effectively separate the Pd atoms, resulting to a lower coordination number with a longer Pd-Pd bond distance.  In the case of P10Y a similar 3-D structure is observed, with Ni central atoms surrounding P10 polymer. Figure S11a displays the feature and the shift of inflection point in edge-jump region (inserted figure) position of P10Y as compared to those of Ni(II) references. The accurate inflection points of P10Y and nickel references are shown in the first derivative normalized XANES spectra in Figure S11b. It is apparent that the inflection point of P10Y was 8343.8 eV which was between that of Ni foil (8333.0 eV) and NiO (8345.5 eV). To precisely obtain the Ni valence in P10-Yamomoto, the linear regression line was derived from the inflection points of various nickel references with high R 2 value (> 0.980), as displayed in Figure S11c. The Ni valence in P10-Yamamoto was a lower mixed-valence, 1.906, suggesting that the existences of Ni(0) (4.7%) and Ni(II) (95.3%). Additionally, the coordination environment of Ni in P10-Yamomoto was compared with that of other nickel references in forms of k 2weighted XANES spectra, as exhibited in Figure S11d. The oscillation frequency, amplitude, and feature of P10-Yamamoto were similar with those of Ni(II) species, especially Ni(OH)2, implying a similar coordination environment.  Figure S12 shows the coordination shells of P10Y and other nickel references. Among these nickel references, the feature of Ni(OH)2 was near that of P10Y, especially in the first shell, Ni-O at 1.60 Å. The existence of Ni-O bond in first shell indicates the occurrence of interatomic bond formation between Ni and O atoms in P10 polymer. These Ni atoms and P10 polymers arranged into an six-coordinated octahedral structure similar with Ni(OH)2. Regarding to the second shell, Ni-Ni at 2.27 Å, was approaching the bond distance of Ni-Ni in Ni foil, 2.18 Å. This may be attributed to the existence of a few Ni(0) atoms, which was consistent with the aforementioned results from XANES spectra.   Figure 1, thus exhibiting more scattering); (b) Fluorescence lifetime decays in the solid state, recorded with a fixed peak count of 10,000 counts. (c) transient absorption kinetics in H2O/MeOH/TEA probed at 630 nm on the fs -early ns timescale following 355 nm excitation using a fluence of 0.08 mJ cm -2 ; (d) transient absorption spectra in H2O/MeOH/TEA probed at 10 µs following nm following 355 nm excitation using a fluence of 0.32 mJ cm -2 . All samples were prepared with a polymer concentration of 0.24 g L -1 . Note that the charge yield on the ps -ns timescale largely reflects the absorbance differences at the excitation wavelength between the different samples and Figure S15. Photocatalytic testing of P10 samples with different Pd content. P10 (25 mg) was suspended in H2O/MeOH/TEA (1:1:1, 25 mL) and sonicated for 10 minutes before degassing by N2 bubbling. Samples were irradiated with a 300 W Xe lamp fitted with a λ > 420 nm filter. The samples with added Pd (orange and red open circles) were prepared via the same method but with the addition of 2 wt. % Pd to the sample, in the form of Pd(NH4)2Cl4, before sonication.