The Dynamic Nature of Brønsted Acid Sites in Cu-Zeolites During Nox Selective Catalytic

The Dynamic Nature of Brønsted Acid Sites in Cu-Zeolites During NOx Selective Catalytic Reduction: Quantification by Gas-Phase Ammonia Titration

John R. Di Iorio, Shane A. Bates, Anuj A. Verma, W. Nicholas Delgass, Fabio H. Ribeiro, Jeffrey T. Miller, Rajamani Gounder*

SUPPORTING INFORMATION

Section S.1. Atomic Absorption Spectrocopy of Catalyst Samples.

The total Si, Al, and Cu contents of each sample were determined usingatomic absorption spectroscopy (AAS) on a Perkin Elmer AAnalyst 300 Atomic Absorption Spectrometer. Typically,0.02 g of zeolite were dissolved in 2 g of hydrofluoric acid (48 wt%, Alfa Aesar) and the solution was allowed to sit overnight, prior to dilution with 60 g of deionized water. Absorbances were measured using radiation sources at wavelengths of 251.6 nm, 309.3 nm, and 324.8 nm for Si, Al, and Cu, respectively, in an acetylene/nitrous oxide flame. Elemental compositions were determined from calibration curves derived from known standard solutions, and are listed for each sample in Table S.1.

Section S.2. X-Ray Diffraction Characterization of Catalyst Samples.

Zeolite framework topologies were determinedfrompowder X-ray diffraction (XRD) patterns measured on a Rigaku SmartLab X-ray diffractometer with a Cu Kα radiation source operated at 1.76 kW. Typically, 0.01g of zeolite powder were loaded onto a zero background, low dead volume sample holder (Rigaku) and the diffraction pattern was measured from 4–40° 2θ at a scan rate of 2.4° min-1. The resulting powder XRD patterns are shown in Figure S.1, and were compared to a known reference material and to diffraction patterns for MFI (H-ZSM-5) and CHA (H-SSZ-13) reported in the International Zeolite Association (IZA) structure database [1].

Table S.1. Elemental composition, and N2-derived BET surface areas and micropore volumes, for all zeolite samples in this study [2].

a Measured Si/Altot and Cu/Altot ratios were determined from atomic absorption spectroscopy.

b The 90% confidence intervals for surface area and micropore volumes are ±51 m2 g-1 and ±0.01 cm3 g-1, respectively.

Figure S1. X-Ray diffraction patterns for (a) H-SSZ-13 (Si/Altot = 4.5) and H-ZSM-5 with Si/Altot of (b) 12.5, (c) 17, (d) 30, (e) 43, and (f) 89.

Section S.3. N2 Adsorption Isotherms of Catalyst Samples.

Micropore volumes were determined from N2 adsorption isotherms measured at 77 K on a Micromeritics ASAP 2020 Surface Area and Porosity Analyzer. Samples were first pelleted and sieved to retain particles between 180-250 μm in diameter, degassed by heating 0.03–0.05 g of sample to 393 K (0.167 K s-1) under high vacuum (~5 μmHg) for 2 h, and then further heated to 623 K (0.167 K s-1) under high vacuum (~5 μmHg) for 8 h. Extrapolation of the linear volumetric uptake during mesopore filling (~0.08–0.30 P/P0) to zero relative pressure gave an estimate for the volume of adsorbed gas in micropores (cm3 g cat-1 at STP). These estimates agreed with the micropore volume determined by analyzing the semi-log derivative plot of the adsorption isotherm, ∂(Vads)/∂(ln(P/P0)) vs. ln(P/P0). This analysis requires identifying the first maximum of ∂(Vads)/∂(ln(P/P0)), which corresponds to the relative pressure at which the micropore filling transition occurs, and then identifying the subsequent minimum that corresponds to the end of micropore filling [3,4]. Micropore volumes were obtained by converting adsorbed gas volumes (at STP) to liquid volumes assuming the liquid density of N2 at 77 K, and are listed in Table S.1 along withBET surface areas. The H-ZSM-5 samples with Si/Altot = 12.5, 17, 30, and 43 had micropore volumes between 0.13–0.17 cm3 g-1. The H-ZSM-5 with Si/Altot = 89, however, exhibited an abnormally low micropore volume of 0.06 cm3 g-1, despite the similar XRD patterns among all ZSM-5 samples. Cu-ZSM-5 samples (Si/Altot = 17, Cu/Altot = 0–0.27) gave micropore volumes between 0.14–0.15 cm3 g-1. The H-SSZ-13 (Si/Altot = 4.5) sample had a micropore volume of 0.26 cm3 g-1 and the Cu-SSZ-13 (Si/Altot = 4.5, Cu/Altot = 0–0.35) samples gave micropore volumes between 0.20–0.24 cm3 g-1.

Section S.4. 27Al Magic Angle Spinning Nuclear Magnetic Resonance (27Al MAS NMR) on H-ZSM-5 and H-SSZ-13 zeolites.

27Al MAS NMR spectra were recorded on NH4-ZSM-5, H-ZSM-5, NH4-SSZ-13, and H-SSZ-13 to estimate the fraction of framework (Alf) and extraframework Al (Alex) species present in each material. Spectra of H-ZSM-5 samples were collected using a Bruker Avance 500 MHz spectrometer in a wide-bore 11.7 Tesla magnet (Caltech Solid-State NMR Facility), and of H-SSZ-13 were collected using a Chemagnetics CMX400 400 MHz spectrometer in a wide-bore 9.4 Tesla magnet (Purdue Interdepartmental NMR Facility), as described elsewhere [2]. Briefly, spectra were acquired at ambient conditions from 456 scans with 12.5 μs pulses and a 2 s delay and were measured at 104.24 MHz and a MAS rate of 5 kHz. Zeolite samples were hydrated by holding for >48 h in a desiccator containing a saturated potassium chloride (KCl) solution prior to packing in a 4mm ZrO2 rotor. All 27Al MAS NMR spectra are referenced to an aqueous 1.0MAl(NO3)3 solution.

The 27Al MAS NMR spectra for the ZSM-5 samples are shown in Figure S.2 and the spectra for SSZ-13 are shown in Figure S.3. Tetrahedrally-coordinated Al atoms were characterized by a lins centered around 55 ppm (52 ppm for ZSM-5 and 58 ppm for SSZ-13), with a small shoulder representative of pentacoordinated Al [5-7]. The relative intensity of the resonance at 52 ppm in H-ZSM-5 increased with increasing Al content (Fig. S.2). Octahedrally-coordinated Al atoms were characterized by a line centered at 0 ppm. The sum of the tertrahedral and pentacoordinated Al were taken to reflect the total amount of Alfspecies present, while the octahedrally-coordinated Al atoms were taken to reflect Alex species. The fraction of Al present in tetrahedral coordination was estimated from the ratio of integrated areas for the 55 ppm line to the total Al area, and is listed for each sample in Table S.2.

Small amounts of Alex species were observed in NMR spectra for H-ZSM-5 with Si/Altot = 43 and 89, 6% and 7%, respectively, and the H-ZSM-5 with Si/Altot = 17, and 30 showed Alex content of less than 2% (Fig. S.2). The NMR spectra of H-SSZ-13 showed a larger Alex contribution of 15% (Fig. S.3). All NH4-ZSM-5 and NH4-SSZ-13 showed Alex content of less than 2%. A small increase in the amount of Alex was observed in the H-form relative to the NH4-form zeolite, which can be attributed to either irreversible removal of Alf species from the lattice to form Alex or to the reversible change between tetrahedral (Alf) and octahedral (Alex) coordination depending on the surrounding conditions [7-9].

Figure S.2. 27Al MAS NMR spectra for all H-ZSM-5 (dashed line) and NH4-ZSM-5 (solid line) samples with Si/Altot = (a) 17, (b) 30, (c) 43, and (d) 89[2].

Figure S.3. 27Al MAS NMR spectra for H-SSZ-13 (dashed line) and NH4-ZSM-5 (solid line) samples with Si/Altot = 4.5[2].

Table S.2. 27Al MAS NMR quantification results for H and NH4–form of the ZSM-5 (Si/Altot = 17, 30, 43, and 89) and SSZ-13 (Si/Altot = 4.5) samples[2].

a Measured Si/Altot ratios were determined from atomic absorption spectroscopy.

b Alfr contains the sum of areas for the tetrahedral and pentacoordinated Al species.

Section S.5. Comparison of NH3 and n-propylamine titrations on ZSM-5 and SSZ-13

The number of H+ sites measured by NH3 titrations at 433 K, followed be purging in flowing helium, and measured by n-propylamine (Table S.4, Section S.7) are listed in Table S.3, and are plotted in Figure 3 of the manuscript.

Table S.3. Comparison of the H+/Altotratios for zeolite samples measured by NH3 and n-propylamine (NPA).

a NH3 saturation performed at 433 K with He purge.

bn-Propylamine H+/Altot is the average of the H+/Altot determined from the decomposition products (NH3 and C3H6) [2].

Section S.6. NH3 TPD profiles after saturation at 323 K

The NH3 TPD profiles for H-ZSM-5 (Si/Altot = 17) and H-SSZ-13 (Si/Altot = 4.5) are shown in Figure S.4, after the samples were saturated with gas-phase NH3 at 323 K. The low temperature peak centered at ~433 K corresponded to weakly-bound and Lewis acid-bound NH3, while the high temperature peak at ~600 K corresponded to NH4+ bound to H+ sites[2].

Figure S.4. NH3 desorption rates during TPD of (a) H-ZSM-5 (Si/Altot = 43) and (b) H-SSZ-13 (Si/Altot = 4.5) after gas phase saturation with NH3 at 323 K[2].

Section S.7. n-Propylamine titrations of ZSM-5 and SSZ-13 zeolites

Alkylamine titrants of varying alkyl chain length (C1-C4) are often chosen to count H+ sites on solid acids, because althoughthey adsorb onto both Brønsted and Lewis acid sites, they decompose via Hoffman-type elimination reactions only on H+ sites to form NH3 and the corresponding alkene (Scheme S.1) [10-13]. The adsorption of n-propylamine (~0.53 nm in diameter [2]) on medium-pore H-ZSM-5 and Cu-ZSM-5 zeolites (~0.51 nm in diameter [14]) ledto the evolution of equimolar amounts of NH3 and C3H6 during a subsequent TPD (Figure S.5, Table S.4), as expected from the reactive decomposition of n-propylammonium ions at H+ sites. The numbers of NH3 and C3H6 species (per Altot) evolved via reactive n-propylamine decomposition on H-ZSM-5 zeolites of varying Al content (Si/Altot = 17-89), and on Cu-ZSM-5 zeolites of varying Cu content (Si/Altot=17 Cu/Altot=0-0.27) are listed in Table S.4[2]. For the samples with Cu/Altot = 0.15 and 0.20, reactive n-propylamine decomposition detects two fewer residual H+ sites per Cu (Table S.4), consistent with the exchange of each Cu2+ cation for two H+ sites. The number of residual H+ sites is similar for the samples with Cu/Altot = 0.20 and 0.27 (Table S.4), indicating that different Cu species may be present on the Cu/Altot = 0.27 sample.

Scheme S.1.n-Propylamine decomposes only on Brønsted acid sites to form NH3 and C3H6 [13].

Figure S.5. Desorption rates of reactive n-propylamine decomposition products NH3 (dark trace) and C3H6 (light trace) during TPD of H-ZSM-5 (Si/Altot = 17) [2].

Table S.4. Number of H+ sites measured from the NH3 and C3H6 desorbed after n-propylamine decomposition on H- and Cu-form zeolites. Data originally reported by Bates et al. [2].

a Alf/Altot values determined from 27Al MAS NMR and details can be found in Table S2 (Supporting Information).

b H+/Alf is the average of the H+/Alf counts from NH3 and C3H6 desorption.

Section S.8. Ambient UV-Visible Spectroscopy of Cu-SSZ-13

Ambient UV-Visible-Near Infrared (UV–Vis–NIR) spectra measuredusing a diffuse reflectance technique were collected on a Varian UV–Vis–NIR spectrophotometer (Cary 5000) equipped with a Harrick-Scientific Praying-Mantis diffuse reflectance accessory (DRA). Barium sulfate (BaSO4, 98 wt%, SigmaAldrich) was used as a reference material when collecting background scans. This material was sieved to a particle size of 125–250 μm before taking the background scan. UV-Vis spectra were collected from 7000 cm-1 to 50,000 cm-1 at a scan speed of 2000 cm-1 min-1. Cu-SSZ-13 samples were pretreated in 10% O2/N2 at 798 K elsewhere and subsequently exposed to ambient conditions. This was done to ensure that these samples were hydrated. Approximately 0.1 g of each sample was placed in the sample cup and spectral measurements were taken.The peakobserved at 12,500 cm-1 is the contribution from the d–d transition of hydrated isolated Cu2+ ions [15,16]. The peaks between 30,000 cm-1 and 50,000 cm-1 have been assigned to a convolution of different oxygen to Cu2+ charge transfer [15,16] and bare zeolite absorption edge. Theintegrated spectra of the band at 12,500 cm-1can be found below in Fig. S.6.A linear increase in the integrated absorbance band area centered around 12,500 cm-1 was observed for increasing total Cu content (Cu/Altot) and is demonstrated by a best fit trendline passing through the origin.

Figure S.6. Integrated area of the hydrated Cu(II) d-d transition absorbance band (12,500 cm-1) measured by UV-Vis spectroscopy as a function of total Cu content (Cu/Altot). The dashed line represents the best fit for a linear trendline through the origin.

Section S.9. References

1. International Zeolite Association Database of Zeolite Structures.

2. Bates SA, Delgass WN, Ribeiro FH, Miller JT, Gounder R (2014) J. Catal. 312:26-36

3. Borghard WS, Reischman PT, Sheppard EW (1993) J. Catal. 139:19-23

4. Borghard WS, Sheppard EW, Schoennagel HJ (1991) Rev. Sci. Instrum. 62:2801-2809

5. Kraus H, Müller M, Prins R, Kentgens APM (1998) J. Phys. Chem. B 102:3862-3865

6. Luz Z, Vega AJ (1987) J Phys. Chem. 91:374-382

7. Gounder R, Jones AJ, Carr RT, Iglesia E (2012) J. Catal. 286:214-223

8. Omegna A, van Bokhoven JA, Prins R (2003) J. Phys. Chem. B 107:8854-8860

9. Omegna A, Prins R, van Bokhoven JA (2005) J. Phys. Chem. B 109:9280-9283

10. Kresnawahjuesa O, Gorte RJ, Oliveira Dd, Lau LY (2002) Catal. Lett. 82:155-160

11. Kresnawahjuesa O, Heussner R, Lee C-C, Kuehl G, Gorte RJ (2000) Appl. Catal., A 199:53-60

12. Farneth WE, Gorte RJ (1995) Chem. Rev. 95:615-635

13. Parrillo DJ, Adamo AT, Kokotailo GT, Gorte RJ (1990) Appl. Catal. 67:107-118

14. Olson DH, Kokotailo GT, Lawton SL, Meier WM (1981) J. Phys. Chem. 85:2238-2243

15. Giordanino F, Vennestrom PNR, Lundegaard LF, Stappen FN, Mossin S, Beato P, Bordiga S, Lamberti C (2013) Dalton T. 42:12741-12761

16. Verma AA, Bates SA, Anggara T, Paolucci C, Parekh AA, Kamasamudram K, Yezerets A, Miller JT, Delgass WN, Schneider WF, Ribeiro FH (2014) J. Catal. 312:179-190