Ligand-Aided Glycolysis of PET Using Functionalized Silica-Supported Fe2O3 Nanoparticles

The development of efficient catalysts for the chemical recycling of poly(ethylene terephthalate) (PET) is essential to tackling the global issue of plastic waste. There has been intense interest in heterogeneous catalysts as a sustainable catalyst system for PET depolymerization, having the advantage of easy separation and reuse after the reaction. In this work, we explore heterogeneous catalyst design by comparing metal-ion (Fe3+) and metal-oxide nanoparticle (Fe2O3 NP) catalysts immobilized on mesoporous silica (SiO2) functionalized with different N-containing amine ligands. Quantitative solid-state nuclear magnetic resonance (NMR) spectroscopy confirms successful grafting and elucidates the bonding mode of the organic ligands on the SiO2 surface. The surface amine ligands act as organocatalysts, enhancing the catalytic activity of the active metal species. The Fe2O3 NP catalysts in the presence of organic ligands outperform bare Fe2O3 NPs, Fe3+-ion-immobilized catalysts and homogeneous FeCl3 salts, with equivalent Fe loading. X-ray photoelectron spectroscopy analysis indicates charge transfer between the amine ligands and Fe2O3 NPs and the electron-donating ability of the N groups and hydrogen bonding may also play a role in the higher performance of the amine-ligand-assisted Fe2O3 NP catalysts. Density functional theory (DFT) calculations also reveal that the reactivity of the ion-immobilized catalysts is strongly correlated to the ligand–metal binding energy and that the products in the glycolysis reaction catalyzed by the NP catalysts are stabilized, showing a significant exergonic character compared to single ion-immobilized Fe3+ ions.


Table S1
Heterogeneous catalyst reported in the literature S3 Figure S1 1 H liquid NMR spectrum of the BHET product formed from the glycolysis of PET.

S10
Figure S2 XPS analysis of the Si 2p core level for bare SiO 2 and the functionalized SiO 2

S12
Table S2 ICP-MS for Fe loading for each ion and NP catalyst S13

S15
Table S3 Calculated relative energies (in kcal•mol -1 ) for the three different spin multiplicities, doublet (D), quadruplet (Q) and sextet (S), for the Fe +3 ion using the SiO 2 -NH 2 system as a test case.

S15
Figure S8 Mechanism for the PET glycolysis reaction.S15

Fe ion immobilization onto ligand modified SiO 2
Fe species were immobilized onto each of the functionalized SiO 2 supports by stirring 2 g FeCl 3 .4H 2 O and 1 g of the ligand modified SiO 2 with 50 mL H 2 O for 1 h.Once completed, 1 mL ammonium hydroxide solution was added, and pH of the solution adjusted from 2 to 12.The product was filtered off and washed with copious amounts of H 2 O.

Synthesis of Fe 2 O 3 NPs on SiO 2
The Fe ion-based catalysts were converted into iron oxide nanoparticle catalysts by calcination or chemical reduction.Catalysts were annealed in a tube furnace at 450 °C for 3 h in air.For chemical reduction method, 0.5 g of the Fe-ion immobilized SiO 2 were placed in a 50 mL beaker.A freshly prepared solution of NaBH 4 (0.095 g, 2.5 mmol) and 2 mL H 2 O was rapidly added under vigorous stirring, turning the catalyst from orange to brown-black.The solution was stirred for 1 h before filtering and washing with copious amounts of H 2 O.

Glycolysis reaction of PET to form BHET
In a typical procedure, 1 g of PET, 0.1 g catalyst and 5 mL of ethylene glycol were refluxed at 190 °C for 3 h.Once reaction was complete, 5 mL of ice-cold H 2 O was added to the reaction filtered off and dried.The catalyst was separated by filtration.The filtrate was placed in the fridge overnight after which the BHET crystallized out the solution.The BHET crystals were collected be filtration, washed with water and dried overnight before being weighed.The percentage PET conversion and isolated BHET yield calculated be the equation 1 and equation 2, respectively.

% Conversion of PET:
Eqn. 1 Where W 0 is the initial weight of the PET used and W 1 is the weight of the unreacted PET after the reaction.
% Yield of BHET: Eqn. 2 Where, W BHET is the weight of the BHET product after recrystallization, M BHET is the molecular weight of BHET, W 0 is the initial weight of PET used and M PET is the molecular weight of the repeat unit of PET.
NMR analysis confirmed that no formation of dimers or oligomers occurred and that BHET was the only product formed in the reaction.The 1 H liquid-state NMR spectrum of BHET can be seen in the supplementary information Figure S1, where the four main peaks were a singlet at 8.12 ppm corresponding to the aromatic ring protons, an OH triplet at

Materials Characterization
X-ray Photoelectron Spectroscopy (XPS) was acquired using a KRATOS AXIS 165 monochromatized Xray photoelectron spectrometer equipped with an Al Kα (hv = 1486.6eV) X-ray source.Spectra were collected at a take-off angle of 90 and all spectra were referenced to the C 1s peak at 284.8 eV.
Transmission electron microscopy (TEM) analysis was performed using a FEI Titan TEM, at an operating voltage of 300 kV.The iron loading in the catalyst was determined by inductively coupled plasma mass spectroscopy (ICP-MS) using an Agilent 7700 ICP-MS equipped with nickel-tip with copper base sampler cones.Autosamples (Agilent Technology, Japan) were used for all measurements.All measurements were performed in three replicates from each vial.In brief, the ICP-MS was operated in full quantitative mode with a Ni sampler and skimmer cones, MicroMist glass concentric nebulizer and quartz Scott-type spray chamber.Samples were quantified based on external calibrations constructed using standard solutions prepared on the day of analysis.Standard solutions were prepared from TraceCERT®, 1 mg/L Fe in nitric acid (Sigma-Aldrich, UK).Method blanks and certified references materials (Environmental Spike Mix, Agilent) were used to monitor the system performance and instrumental drift.Full data were recorder and analyzed with Agilent Mass Hunter Data Software (version 4.6 C.01.08).
1 H liquid-state NMR spectra were acquired on a Bruker Avance III NMR spectrometer operating at B 0 = 7.05 T ( 0 = 300.13MHz for 1 H), in proton-coupled mode.Samples were dissolved in deuterated chloroform (CDCl 3 ) or deuterated d 6 -DMSO and tetramethysilane (TMS) was used as the internal standard.
Solid-state NMR experiments were carried out on numerous instruments. 29Si NMR experiments were recorded at B 0 = 7.05 T ( 0 = 300.13MHz for 1 H) using a Varian VNMRS spectrometer, equipped with a 9.5 mm HX probes employing a MAS rate  R = 4 kHz. 29Si one pulse excitation (SPE) experiments were acquired averaging 128 transients with a recycle delay of 1800 s, employing an rf-field strength of  1  42 kHz.The integrals for the Q 4 sites were corrected for the extremely T 1 extracted from saturationrecovery curves. 1 H 29 Si CPMAS CPMG experiments were recorded with a CP contact time of 2 ms, a recycle delay of 5 s, and 512 transients.For the CPMG train, each full echo was 6 ms in duration and a total of 12 echoes were collected. 29Si chemical shifts were referenced using a solid sample of zeolite 4A ( iso = 89.70ppm).
out on a Bruker Avance NEO 600 MHz spectrometer (B 0 = 14.1 T) employing a Varian 3.2 mm HXY probe at  R = 14 kHz.The initial π/2 pulse was 2.17 μs corresponding to an rf-field of  115 KHz and was followed by a CP contact time of 1 ms.The1 H→ 13 C CPMAS spectra were recorded using recycle intervals of 8 s (Ligand 1) and 10 s (Ligand 2) and were the result of averaging 1024 transients.SPINAL 1 H decoupling 24 (∼50 kHz) was applied during acquisition.The Hartmann-Hahn condition [25][26] was calibrated using a powdered sample of adamantane. 13C chemical shifts were referenced using a solid sample of adamantane ( iso = 29.47 and 38.52 ppm).to an rf-field of  98 kHz.The VACP contact time was 1 ms (for ligands 3 and 4) and 10 ms (for ligand 5) using an 80-100 % ramp shape at the 1 H channel, while the RF nutation frequency on the 13 C channel was 65 kHz.The spectra were recorded using recycle intervals of 3.5 s (Ligand 3), 2.5 s (Ligand 4) and 7 s (Ligand 5) and were the result of averaging 2048 (Ligands 3 and 4) and 10240 (Ligand 5) transients.
SPINAL 1 H decoupling 24 (∼80 kHz) was applied during acquisition.The Hartmann-Hahn condition [25][26] was calibrated using a powdered sample of glycine.For Ligand 5; the 1 H 13 C CPMAS spectrum was acquired at 40C to decrease the dynamics in the sample which reduced the CP efficiency.For all solid-state NMR experiments, rotors were packed in air.Data processing was carried out using Bruker TopSpin (version 4.1.3)and ssNake (version 1.3). 27e relative surface coverage obtained from the quantitative 29 Si MAS NMR spectra was estimated using equation 3 and expressed with respect to the bare SiO 2 support.The degree of condensation of the Si atoms was estimated using equation: 28 Eqn. 3 100 -(

Density functional theory (DFT) calculations
Density functional theory (DFT) simulations were carried out using the CP2K package. 29The mesoporous SiO 2 structure studied here has been computationally represented as a periodic amorphous silica (SiO 2 ) surface.The surface selected has a silanol density of 7.2 OH/nm 2 assuring the maximum number of Si atoms available to be functionalized in the unit cell. 30For the initial SiO 2 surface, both internal atomic positions and cell parameters were optimized (resulting with dimensions of 13.34 x 13.73 x 49.21 Å 3 .The following optimizations were performed only relaxing the geometry parameters.Optimizations were carried out using the semi-local Perdew-Burke-Ernzerhof (PBEsol) functional, 31 combined with a double-ζ basis set (DZVP-MOLOPT-SR-GTH Gaussian basis set) for all the atom types, together with the Grimme D3BJ correction term to the electronic energy, 32 and a cutoff set at 500 Ry for the plane wave auxiliary basis set.Core electrons were described with the Goedecker-Teter-Hutter pseudopotentials 33 and valence ones with a mixed Gaussian and plane-wave (GPW) approach. 34Solvation effects were considered by performing a single point energy calculations on the optimized systems adopting the self-consistent continuum solvation (SCCS) model as implemented in CP2K, aiming to reproduce the experimental conditions of water and ethylene glycol as solvents. 35Binding energies (BE) have been calculated by applying the counterpoise correction in order to avoid basis-set superimposition errors (BSSE).In CP2K, the interaction energy E int of the metal on the SiO 2 structure can be calculated defining 2 fragments A and B, as shown in equation 5.

Eqn. 5 𝑬
where E AB is the absolute potential energy of the SiO 2 /metal system, E A corresponds to the absolute potential energy of the bare silica and E B is the absolute potential energy of the naked metal.

S10
Finally, to determine the nature of the stationary points of the potential energy surfaces (i.e., local minima and saddle points) the corresponding vibrational harmonic frequencies were calculated at the PBE-D3BJ/DZVP level using the finite differences method.A partial Hessian approach was used to reduce the computational cost of the calculations and the vibrational frequencies were calculated on the optimized geometries only for a fragment of the entire system, which included the metal centre and the organic ligands for the single atom Fe 3+ structures and, the metal-nanocluster and the adsorbates for the FeO nanoparticle structures.Table S3.Calculated relative energies (in kcal•mol -1 ) for the three different spin multiplicities, doublet (D), quadruplet (Q) and sextet (S), for the Fe +3 ion using the SiO 2 -NH 2 system as a test case.

System
Rel

Synthesis of SiO 2 -
NH 2 -SB (Ligand 2): 1 g of the SiO 2 -NH 2 prepared in the previous step, was placed in a round bottom flask along with 1.2 mL 2-pyridinecarboxaldehyde and 50 mL ethanol.The reaction mixture was refluxed at 60 °C for 24 h.Once completed, the solid product (SiO 2 -NH-SB) was filtered and washed with ethanol before being dried at 60 °C.Synthesis of SiO2 -NHNH 2 (Ligand 3): 2 g of MCF SiO 2 was placed in a round bottom flask along with 4 mL of AEAPTMS and 75 mL toluene.Mixture was refluxed at 120 °C for 12 h under argon.Once reaction had completed, the solid product (SiO2-NHNH 2 ) was vacuum filtered and washed with ethanol before dried at 60 °C.Synthesis of SiO 2 -NHNH 2 SB (Ligand 4): 1 g SiO 2 -NHNH 2 prepared in the previous step, was placed in a round bottom flask along with 1.2 mL 2-pyridinecarboxaldehyde and 50 mL ethanol.The reaction mixture was refluxed at 60 °C for 24 h.Once completed, solid product (SiO 2 -NHNH 2 SB) was filtered and washed with ethanol before being dried at 60 °C.Synthesis of SiO 2 -Pincer (Ligand 5): 1.4 g SiO 2 -NH 2 , 1.8 g cyanuric chloride (CNC), 1.8 mL DIEA and 40 mL THF were added to a round button flask.The reaction mixture was left to stir in an ice bath under argon for 8 h.The product SiO 2 -CNC was filtered off and washed with THF (5 x 20 mL) and dried overnight.In a three-neck round bottom flask, 1.5 g of the SiO 2 -CNC was dispersed in 30 mL dry acetonitrile, 2 g 2-aminopyridine and 2 mL DIEA were added and refluxed for 48 h.After completion, the solid product (SiO 2 -Pincer) was filtered off and washed with methanol (5 x 30 mL) and dried at 50 °C overnight.

Figure S1: 1 H
Figure S1: 1 H liquid NMR spectrum of the BHET product formed from the glycolysis of PET.

Figure S8 :
Figure S8: Mechanism for the PET glycolysis reaction.

Figure S9 :
Figure S9: Catalyst recyclability of the SiO 2 -NP pincer catalyst over five reaction cycles.

Table S1 :
Summary of literature reports using heterogeneous catalysts for PET glycolysis

1.2. Synthesis of surface modified mesoporous cellular foam SiO 2 Synthesis of SiO 2 -NH 2 (Ligand 1): 2 g of MCF SiO 2 was placed in
2301 ppm, a CH 2 triplet at 4.32 ppm and a quartet at 3.77 ppm for the CH 2 next to the OH group.Residual solvent d 6 -DMSO was located at 2.51 ppm.The results are in good agreement with literature values of BHET.23

Table S2 :
ICP-MS for Fe loading for each ion and NP catalyst