Perylenetetracarboxylic acid nanosheets with internal electric fields and anisotropic charge migration for photocatalytic hydrogen evolution

Highly efficient hydrogen evolution reactions carried out via photocatalysis using solar light remain a formidable challenge. Herein, perylenetetracarboxylic acid nanosheets with a monolayer thickness of ~1.5 nm were synthesized and shown to be active hydrogen evolution photocatalysts with production rates of 118.9 mmol g−1 h−1. The carboxyl groups increased the intensity of the internal electric fields of perylenetetracarboxylic acid from the perylene center to the carboxyl border by 10.3 times to promote charge-carrier separation. The photogenerated electrons and holes migrated to the edge and plane, respectively, to weaken charge-carrier recombination. Moreover, the perylenetetracarboxylic acid reduction potential increases from −0.47 V to −1.13 V due to the decreased molecular conjugation and enhances the reduction ability. In addition, the carboxyl groups created hydrophilic sites. This work provides a strategy to engineer the molecular structures of future efficient photocatalysts.

Scan the water-absorbent sample by lowering the test vacuum (Acceleration voltage: 15kV & 2 kV).
The water content of PTA absorbed in the atomized water vapor for half an hour is about 33000 ppm. Although water causes more serious charged effect, it can still be observed that the nanosheet-like morphology of PTA has hardly changed. PTA is in a single-molecule state in alkali (Supplementary Figure 10a). In comparison, the absorption peak of PTA dispersed in water is blue-shifted. It shows that PTA is in an aggregated state in water. In-situ FT-IR data showed that the chemical structure of PTA did not change after water absorption.

Supplementary
Supplementary Figure 11. Effect of water environment on the thickness of PTA nanosheets. Small Angle X-ray Scattering (SAXS) data for PTA in water (1mg mL -1 ).
(The data is fitted using lamellar model.) The SAXS data is simulated by the lamellar model to obtain a thickness of 15.6 Å, which is highly consistent with the thickness of the dry state tested, indicating that PTA is still nanosheets in water. Compared with the XRD of dry-state, the average thickness (D) of the crystal grains perpendicular to the crystal plane calculated by the Scherer formula changes slightly when water is added (Supplementary table 1). The two main peaks change by 4% and 14%, respectively. However, Si powder as an external standard, the main peak change is greater than 11% after adding water. Therefore, it can be considered that the longrange order structure of each crystal face in water is the same as that in the dry state for PTA.

Supplementary Figure 13. Optimized PTA crystals in a water environment.
Optimized model by PM6 semi-empirical method (Black: carbon, yellow: oxygen, white: hydrogen) .
Firstly, the interaction of PTA nanosheets with H2O molecular in the water environment was explored through theoretical calculations. Under the action of the hydrogen bond between the carboxyl groups and H2O, a hydration layer is formed around the PTA nanosheet, which effectively prevents the formation of a multilayer structure. Next, explain the reasons why the morphology of the nanosheets can be maintained in water. A model in which H2O molecules exist between PTA molecules is constructed.

Supplementary
After optimization, the H2O molecules between the layers migrate to the two ends due to the hydrophilicity of carboxyl. As a result, the π-π interaction between PTA molecules is enhanced, so that the nanosheets morphology can be maintained. We further calculated that the hydrogen bonding energy between H2O and PTA molecule is 3.78 kJ/mol. However, the π-π stacking energy between the paired PTA molecules is 46.21 kJ/mol. This proves that π-π accumulation dominates in the water environment, which is responsible for the morphology of the nanosheets. We evaluated the effect of the dosage of PTA on the photocatalytic activity and explored the amount of photocatalyst to achieve saturated light absorption 1,2 . With the increase of dosage from 6 mg to 7 mg, the actual observed HER increased from 332 μmol h -1 to 345 μmol h -1 . As the dosage increased from 7 mg to 8 mg, HER decreased from 345 μmol h -1 to 310 μmol h -1 . The reason for the decreased HER rate of large-dosage PTA photocatalyst is the light scattering and the low utilization of active sites of the photocatalyst. Therefore, the optimal dosage was 7 mg for HER 3 .
Besides, it was found that the best photocatalytic performance was achieved when the mass fraction of loaded Pt was ~4.6 wt. % and the amount of catalyst was 7 mg. Too much Pt or is not conducive to the absorption of light.

Supplementary Note 1: Fluorescence quantum yields
Relative quantum yield calculations were performed using a modified procedure by Williams et al 7 . The quantum yield of a sample relative to a reference compound can be calculated using the following equation: This calculation can be done using a single concentration, or a range of concentrations.
We use the slope of the integrated emission intensity versus absorbance curve at the excitation wavelength, avoiding errors in individual measurements.
We used the reported QY = 0.94 for rhodamine 6G in ethanol as a control compound.
Typically, fluorescence and absorbance measurements were made by injecting 3. To evaluate the charge mobility, the electron mobility of PTCDA and PTA was measured by SCLC method 10 . Its J-V curve is shown in Supplementary Figure 51. The mobility values were calculated by the Mott-Gurney law for the sandwich-type device: Where J, εr, ε0 and L represent the current density, relative permittivity, the vacuum permittivity and the distance between the two electrodes, respectively. It is assumed that εr=3, and ε0 is the same for all crystals and a value of 8.85×10 -14 has been used from previous reports 11 . The electron mobility is calculated to be 1.1 and 3.2 cm 2 V -1 s -1 for PTCDA and PTA, respectively. For PTA nanosheets, the enhanced internal electric fields and the long-range ordered structure contribute to their electron mobility. The observed rate constant in PTA nanosheets is: 1/τ= k1 + k2

Supplementary
Where τ is the photogenerated electron lifetime from TAS data, 134.2 ps.
k1 is the rate of monomer decay, k1=1/τ1, τ1 is the stimulated emission (SE) delay from TAS data, 1421.9 ps. k2 is the charge transfer rate constant, it is further calculated that k2=0.006748 ps -1 .
And then computing the efficiency carrier production (Q): As shown in Figure 2c in manuscript, the apparent quantum efficiency of hydrogen production of the PTA nanosheets at 400 nm excitation was 4.8%, which is lower than the estimated carriers generation rate (90.6%). The significant difference between the apparent quantum efficiency of hydrogen production and the carrier hydrogen production rate can be explained as follows: On the one hand, the carriers produced were not fully used for hydrogen production, which may be attributed to processes such as carriers are used to generate reactive oxygen species. As shown in the electron paramagnetic resonance (EPR) measurement ( Supplementary Fig.51), where the photogenerated electrons of PTA are trapped by dissolved oxygen to form · O2in an air atmosphere. It further reacts with H + to form · OH.
On the other hand, in general, after photoexcitation of a semiconductor to generate photogenerated excitons or carriers, these photoexcited species need to migrate or diffuse to the reaction sites to drive the reaction. However, for excitons produced in organic semiconductors, the typically low exciton migration distances limit the In this study, the PTA nanosheets show an ordered crystal structure, which was confirmed by XRD as well as by electron diffraction patterns (Figure 1). According to the previous reports 12 , excitons of highly crystalline organic materials are not prone to exciton-exciton annihilation because of the large effective volume, which is expected to give favourable exponential dynamics. From the results, we fit the TAS data according to the second-order exponential and all the R 2 can reach above 0.95, and the residual analysis shows that the mathematical model has a high confidence level. When HCl is used, the PTA formed is in strips of ~2.0 μm in length. When H3PO4 was used, both short strips and sheets were available. The ~520 nm long PTA nanosheets shown in the main text were obtained using CH3COOH.

Supplementary
The photogenerated charge properties are then explored for different morphologies of PTA. Surface photovoltage is a means of characterizing the accumulation of photogenerated charge on the surface. There are no significant differences in the SPV signals of the three morphologies of PTA ( Supplementary Fig. 53d). In addition, their transient photocurrent signals are also relatively consistent ( Supplementary Fig. 53e).
This suggests that the effect of morphology on the photogenerated charge movement of PTA is minimal. Finally, the presence of ascorbic acid in the photocatalytic hydrogen production halfreaction is highly oxidizable and the reactive oxygen species generated on the surface of the PTA nanosheets would preferentially react with the thermodynamically easy ascorbic acid.

Supplementary
In conclusion, the long-range ordered structure of PTA, and the hydrogen production environment that is not conducive to the generation of reactive oxygen species ensure the stability of PTA nanosheets.

Supplementary Methods 2: Determination of internal electric fields (IEF)
The Poisson equation describes the spatial distribution of the IEF, which shows that IEF is a function of surface potential and surface charge density [13][14][15][16][17] .

Supplementary Tables
Supplementary