Synergy between Brønsted Acid Sites and Carbonaceous Deposits during Skeletal 1-Butene Isomerization over Ferrierite

During skeletal 1-butene isomerization over ferrierite carbonaceous deposits block 98% of the micropores within 24 h, rendering them effectively inaccessible to reactants, while the catalytic activity improves continuously for 100 h on stream. Ex-situ pyridine adsorption shows that the concentration of conventional Brønsted acid sites in the 10-R channels decreases below the detection threshold of infrared spectroscopy within 2 h. However, the operando addition of the base triethyl amine to the feed quenches the reaction, showing that mediated acidity is necessary. The larger base 2,2,6,6-tetramethyl piperidine only deactivates catalytic activity after several hours because it cannot directly bind to active sites at the sterically restricted pore mouths. The communication of internal Brønsted acid sites to the external reactants via a concerted mechanism involving protonated monoaromatic deposits trapped in the pore mouths explains the promoting effects of coke species in zeolite-catalyzed skeletal butene isomerization. This work presents a consolidated explanation of the synergy of solid acidity, structural confinement, and carbonaceous deposits in zeolites.


Introduction
Scheme S1: Interaction of butene with a Brønsted acid site via the non-selective dimerization mechanism.Adapted from 1 Scheme S2: Interaction of butene with a Brønsted acid site via the co-dimerization mechanism.Adapted from 1 Scheme S3: Interaction of butene with a Brønsted acid site via the monomolecular mechanism.Adapted from 1

27 Al MAS NMR Spectroscopy
A Bruker III 400 MHz spectrometer was utilized to collect 27 Al solid-state magic angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy data.To mitigate the influence of the quadrupolar effect for aluminum, the sample was hydrated overnight at room temperature in air saturated with water vapor.The samples were then packed in a zirconia rotor (4 mm diameter) and rotated at a spinning frequency of 10 kHz.
The pulse length was set to 3 μs and a recycle delay of 4 s.The total number of scans was set to 1024.All spectra were referenced to solid Al(NO3)3 at 0 ppm.S 3

X-ray Powder Diffraction Peak Fitting Routine
The raw diffractograms are shown in Figure S7.Peak assignment was performed as described by Arletti et al. 2 The X-ray powder diffractograms were fitted as described in Hebisch et al. 3 and in equations ( 1) to (4)   with skewed Lorentzian distribution peaks (Figure S1), to determine the exact location of the peak maxima.This was done because there exists no consensus in literature about whether FER crystalizes in the Pnnm or Immm crystal system but there is agreement that it must be an orthorhombic unit cell.
The plane spacing (Q) as function of scattering angle () and x-ray wavelength () are related to each other by Bragg's law.In the case of an orthorhombic unit cell with lattice parameters a, b, and c and Miller indices h, k, and l this equation takes the form: To phrase this as a linear least squares problem, one substitutes: To yield: Each set of Miller indices plus the exact location of the diffraction peak adds another equation with three unknow quantities.This results in an overdetermined linear system of equations: With this information, the lattice parameters a, b, and c are calculated through a linear least squares problem.
Solving this system of equations using standard linear algebra methods yields the lattice parameters.

S 4
Figure S1: Exemplary fit of XRD data of H-FER with skewed Lorentzian peaks that are used to calculate lattice parameters.
S 5

Activity of Activated Charcoal as the Catalyst
To test if activated charcoal alone can catalyze skeletal butene isomerization at the reaction temperature, 0.6 g of activated charcoal were packed similarly to the experiments with H-FER (section 2.2).The chromatographic results of feeding 12 sccm undiluted 1-butene at a temperature of 420 °C shows negligible iso-butene formation, and thermally catalyzed double-bond isomerization to 2-butene as the main reaction product (Figure S2).S 7  S 8

Overview of Acidity Results
The table below summarized the results of characterizing the acid site strength and type with the three probe molecules ammonia, pyridine, and acetonitrile.

Molecular Kinetic Diameters
The kinetic diameters (dkin) of relevant probe molecules were calculated by optimizing molecular geometries via force-field simulations using the Python package "atomic simulation environment" (ASE).
The calculated kinetic diameters of reactants and products are listed in Table S2.

Operando IR Spectroscopy
Figure S5 shows the full spectrum from 4000 cm -1 to 650 cm -1 .The time (30 min total) is represented as color from dark purple (0 min) to light yellow (30 min).The first appearance of 1-butene occurs at the 6 min mark.S 11

X-ray Powder Diffraction
The powder diffractograms of the analyzed data in Figure 1 (C) are displayed in Figure S7.S 12

N2 Physisorption Isotherms
The isotherms of the three samples analyzed in Figure 1 (D) are shown in Figure S8.

Proposed Reaction Network based on Carbocationic Intermediates
Scheme S4 depicts the resulting reaction network with a benzylic carbocation as the reactive species, as proposed by Guisnet. 8,9 he color scale indicates the stability of certain intermediates, with tertiary carbenium ions are assumed to be too stable to react further.H-shift indicates a proton transfer, while Meshift indicates rearrangement of a -CH3 group.As Guisnet proposes the reaction mechanism, the reactive species is not directly located in the pore mouths and is thus sterically unconstrained.If larger side-chains with five or six carbon atoms form, they could form a second five-or six-membered ring by fusing with the existing aromatic ring in a Friedel-Crafts-type alkylation, thereby forming condensed aromatic species which deactivate the catalyst.
Scheme S4: Reaction network showing the formation of iso-butene and various side products based on the transition stated proposed by Guisnet. 8,9

Figure
Figure S4 depicts the peak assignment and fitting results of the d3-acetonitrile adsorption.

Figure S4 :
Figure S4: Results of peak fitting and band assignment of spent H-FER (TOS= 2 h; above) and fresh H-FER (calcined at 550 °C; below).

Figure S5 :
FigureS5: Operando Infrared spectra for the first 30 min of the reaction as 1-butene is dosed to H-FER at 420 °C.1-butene first appears in the gas phase at the 6 min mark (C-H stretch region).The correlation maps were generated using data from 6-13 minutes from this dataset, corresponding to the first appearance of 1-butene.

Figure S6 :
Figure S6: 27 Al solid state Magic Angle Spinning (MAS) NMR spectrum of calcined, rehydrated H-FER before reaction (blue) and the sample after reaction for 100 h (orange).

Figure
Figure S7: X-ray powder diffractograms of the four samples analyzed in Figure 1.The time indicates the reaction time.0 h refers to the calcined H-FER prior to reaction.

Figure S8 :
Figure S8: Isotherms of the three samples shown in Figure 1 (D).

Table S1 :
Summary of characterizing the acid site strength and type with the three probe molecules ammonia, pyridine, and acetonitrile.

Table S2 :
Calculated kinetic diameters of relevant probe molecules and comparison with literature values.