Nanocasting SiO2 into metal–organic frameworks imparts dual protection to high-loading Fe single-atom electrocatalysts

Single-atom catalysts (SACs) have sparked broad interest recently while the low metal loading poses a big challenge for further applications. Herein, a dual protection strategy has been developed to give high-content SACs by nanocasting SiO2 into porphyrinic metal–organic frameworks (MOFs). The pyrolysis of SiO2@MOF composite affords single-atom Fe implanted N-doped porous carbon (FeSA–N–C) with high Fe loading (3.46 wt%). The spatial isolation of Fe atoms centered in porphyrin linkers of MOF sets the first protective barrier to inhibit the Fe agglomeration during pyrolysis. The SiO2 in MOF provides additional protection by creating thermally stable FeN4/SiO2 interfaces. Thanks to the high-density FeSA sites, FeSA–N–C demonstrates excellent oxygen reduction performance in both alkaline and acidic medias. Meanwhile, FeSA–N–C also exhibits encouraging performance in proton exchange membrane fuel cell, demonstrating great potential for practical application. More far-reaching, this work grants a general synthetic methodology toward high-content SACs (such as FeSA, CoSA, NiSA).

Field-emission scanning electron microscopy (FE-SEM) was carried out with a field emission scanning electron microanalyzer (Zeiss Supra 40 scanning electron microscope at an acceleration voltage of 5 kV). Nitrogen sorption measurement was conducted using a Micromeritics ASAP 2020 system at 77 K. Prior to nitrogen adsorption/desorption measurement, the samples were dried overnight at 160 °C under vacuum. The Optima 7300 DV inductively coupled plasma atomic emission spectrometer (ICP-AES) was utilized for the quantification of the content of Fe. X-ray photoelectron spectroscopy (XPS) measurements were performed by an ESCALAB 250 high-performance electron spectrometer using monochromatized Al Kα (hν = 1486.7 eV) as the excitation source. The binding energy was calibrated using the C 1s from graphene oxide nanosheets located at 284.8 eV as the reference. The spectral decomposition was performed using the XPS Peak 41 program with Gaussian functions after subtraction of a Shirley background.

Supplementary Note 2 | X-ray Adsorption Spectra.
The Fe/Co/Ni K-edge X-ray absorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) experiments were carried out at the beamlines 5BM-D and 20BM-B of Advanced Photon Source (APS) at Argonne National Laboratory (ANL). The catalyst powder and reference were pressed into pellets with boron nitrides and measured in transmission mode. The incident beam was monochromatized by using a Si (111) fixed-exit, double-crystal monochromator, and a harmonic rejection mirror was applied to cut off the harmonics at higher X-ray energy. Data reduction, data analysis, and EXAFS fitting were performed with the Athena and Artemiss software packages. The energy calibration of the catalysts was conducted through a standard metal foil, which as a reference was simultaneously measured.
Data points were acquired in three separate regions (energies relative to the Fe edges): a pre-edge region (-300 to -30 eV, step size of 10 eV), the XANES region (-30 to +30 eV, step size of 0.5 eV) and the EXAFS region (to 12.5 Å -1 , step size of 0.07 Å -1 , dwell time of 1 s). The ionization chambers were optimized for the maximum current with a linear response (~1010 photons detected per second) with 10% absorption in the incident ion chamber and 70% absorption in the transmission detector.
Standard procedures for the normalization and background subtraction were performed using the Demeter 0.9.25 software package. XANES at Fe K-edge was determined from the inflection point in the leading edge, that is, the maximum in the S5 first derivative of the leading edge of the XANES spectrum. The pre-edge energies were also obtained in the first derivative using the zero-crossing point.
For EXAFS modeling, EXAFS of the Fe foil is fitted and the obtained amplitude reduction factor S0 2 value (0.81) was set in the EXAFS analysis to determine the coordination numbers (CNs) in the Fe-N scattering path in FeSA-N-C catalysts.
Following the standard EXAFS fitting procedures and model construction (the model of Fe-N4-C as is shown in Figure 3c) that have been demonstrated in the previous studies 1-2 , the EXAFS fitting was conducted through a least-squares fitting in R space for the k 2 -weighted Fourier transformed data using Artemis.
The obtained best fitting results can be found in Supplementary Table 4, all the parameters are found within the reasonable ranges, indicating a good match between the model and the experimental data. Additionally, this can be also evidenced by the good match between the k-space oscillations the experimentally data and the as-generated from the model (Supplementary Figure 14).

Supplementary Note 3 | Ligand Synthesis.
The ligand was synthesized according to the previous report with modifications. [3][4] The typical synthetic procedures are described below.

5,10,15,20-Tetrakis(4-methoxycarbonylphenyl)porphyrin (TPPCOOMe).
Typically, pyrrole (3.0 g, 0.043 mol), methyl 4-formylbenzoate (6.9 g, 0.042 mol) and propionic acid (100 mL) were added to a 500 mL three-necked flask to form mixed solution, which was then refluxed at 140 °C for 12 h in darkness. After the reaction mixture was cooled down to room temperature, purple crystals were collected by suction-filtration. Then the crystals were washed in the sequence of ethanol, ethyl acetate and THF, and the obtained crystals were dried under vacuum at 60 °C . Typically, TPP-COOMe (0.854 g, 1.0 mmol) was dissolved in DMF solution (100 mL) containing FeCl2•4H2O (2.5 g, 12.8 mmol) and the mixed solution was refluxed at 160 °C for 6 h. When the mixture was cooled down to room temperature, 150 mL of H2O was added. The resultant precipitate was filtered and washed with 50 mL of H2O for twice. The obtained solid was dissolved in CHCl3, followed by extracting three times with 1 M HCl and twice with water. The Co-TPPCOOMe and Ni-TPPCOOMe were synthesized following the same procedure as Fe-TCPPCl, except for CoCl2• 6H2O (3.1 g, 12.8 mmol) and NiCl2•6H2O (3.1 g, 12.8 mmol) were used instead of FeCl2•4H2O (2.5 g, 12.8 mmol) to synthesis Co-TPPCOOMe and Ni-TPPCOOMe.

S9
Typically, the N-C was synthesized through the similar methods to FeSA-N-C except for employing SiO2@PCN-222 as the precursor.

Supplementary Note 8 | Synthesis of CoSA-N-C and NiSA-N-C catalysts.
The CoSA-N-C and NiSA-N-C were synthesized through the similar methods to FeSA-N-C except for employing SiO2@PCN-222(Co) and SiO2@PCN-222(Ni) as precursors, respectively. S10

Supplementary Note 9 | Electrochemical measurements.
Electrochemical measurements were performed with a CHI 760E electrochemical analyzer (CH Instruments, Inc., Shanghai) and a rotating disk electrode (RDE) (Pine Instruments, Grove city, PA). All electrochemical measurements were conducted in a typical three-electrode setup with a graphite rod counter electrode and Ag/AgCl reference electrode. LSV measurements were conducted with scan rate of 5 mV/s. All potentials reported in this work were converted from vs Ag/AgCl to vs RHE by adding a value of 0.197 + 0.059 × pH. All data are presented without iR compensation.
The catalyst ink was prepared by dispersing 2 mg of catalyst into 1 mL of ethanol containing 10 μL of 5 wt% Nafion and sonicated for 30 min. Then 28 μL of the catalyst ink was loaded onto a GCE of 5 mm diameter (loading amount: ~0.28 mg cm -2 ). For comparison, Pt/C (20 wt% platinum) was conducted on the same electrochemical tests with a catalyst loading of 0.1 mg cm -2 .
For the ORR at a RDE, the electron transfer number can be calculated with Koutecky-Levich equations: where j is the measured current density; jk and jL are the kinetic and diffusion-limiting current densities, respectively; ω is the angular velocity of the disk ( = 2πN, N is the linear rotation speed); n represents the overall number of electrons S11 transferred in oxygen reduction; F is the Faraday constant (F = 96485 C mol -1 ); C0 is the bulk concentration of O2 (1.2 × 10 -6 mol cm -3 ); D0 is the diffusion coefficient of O2 in 0.1 M KOH electrolyte (1.9 × 10 -5 cm 2 s -1 ); v is the kinematics viscosity for electrolyte, and k is the electron-transferred rate constant.
Rotating ring-disk electrode (RRDE) measurements were carried out to determine the four-electron selectivity. The disk electrode was scanned at a rate of 10 mV s -1 , and the ring electrode potential was set to 1.2 V vs. RHE. The hydrogen peroxide yield (%H2O2) and the electron transfer number (n) were calculated by the following equations: where id and ir are the disk and ring currents, respectively. The N represents the ring current collection efficiency which was determined to be ~37 %.

Supplementary Note 10 | PEMFC tests.
The catalyst was mixed with Nafion® alcohol solution (5 wt%, Aldrich), isopropanol and deionized water to prepare the catalyst ink, which contained the same weight of Nafion ionomer as the catalyst. The ink was subjected to a sonication for 10 min and a stirring for 12 h. The well-dispersed ink was brushed on a piece of carbon paper (5

S24
To investigate the migration of Fe atom on the N-doped carbon surface, we considered different migration structures and obtained three stable migration structures, as shown in Figure S15. In Figure S15, the ground state energies of the corresponding systems were labelled. To compare the energy differences between different structures, the energy of FeN4 was selected as the zero of relative energy. To  images of FeSA-N-C is Fe atoms rather than Zr atoms. Amplitude reduction factor S0 2 is determined to be 0.81 through fitting the FT-EXAFS of standard Fe foil which is measured simultaneously during the experience. CN, coordination number; R, distance between absorber and backscatter atoms; σ 2 , Debye-Waller factor to describe the variance in due to disorder (both lattice and thermal); ∆E, threshold Energy Correction; R factor (%) indicates the goodness of the fit.  *∆G: free energy change at T=298 K, pH=1 and U= 0 V.