Effect of SO₂ and SO₃ Exposure to Cu-CHA on Surface Nitrate and N₂O Formation for NH₃–SCR

Abstract

We report effects of SO₂ and SO₃ exposure on ammonium nitrate (AN) and N₂O formation in Cu-CHA used for NH₃–SCR. First-principles calculations and several characterizations (ICP, BET, XRD, UV–vis–DRS) were applied to characterize the Cu-CHA material and speciation of sulfur species. The first-principles calculations demonstrate that the SO₂ exposure results in both (bi)sulfite and (bi)sulfate whereas the SO₃ exposure yields only (bi)sulfate. Furthermore, SOx adsorption on framework-bound dicopper species is shown to be favored with respect to adsorption onto framework-bound monocopper species. Temperature-programmed reduction with H₂ shows two clear reduction states and larger sulfur uptake for the SO₃-exposed Cu-CHA compared to the SO₂-exposed counterpart. Temperature-programmed desorption of formed ammonium nitrate (AN) highlights a significant decrease in nitrate storage due to sulfur species interacting with copper sites in the form of ammonium/copper (bi)bisulfite/sulfate. Especially, highly stable sulfur species from SO₃ exposure influence the NO₂–SCR chemistry by decreasing the N₂O selectivity during NH₃–SCR whereas an increased N₂O selectivity was observed for the SO₂-exposed Cu-CHA sample. This study provides fundamental insights into how SO₂ and SO₃ affect the N₂O formation during ammonium nitrate decomposition in NH₃–SCR applications, which is a very important topic for practical applications.

1. Introduction

NH₃-assisted selective catalytic reduction (SCR) has been extensively applied for NOx emission control in stationary and mobile applications (e.g., power plants, automobiles, and ships). Copper-exchanged chabazite (Cu-CHA), composed of a small pore entrance (eight-membered rings, 3.8 × 3.8 Å) and double six-membered rings, has been successfully commercialized in exhaust after treatment systems (EATS) for diesel combustion thanks to its high activity and hydrothermal stability at low and high temperatures, respectively.¹⁻³ Nitrous oxide (N₂O) emissions, which may be formed via side reactions during NOx reduction, must be regulated because of their ∼300 times stronger greenhouse gas potential than carbon dioxide (CO₂) and depletion of the stratospheric ozone.^4,5 As a result, there is a great deal of interest in understanding the fundamental mechanisms of N₂O formation during NH₃–SCR of NOx to improve N₂ selectivity and minimize N₂O emissions.

Utilization of biomass in fuels is indispensable to achieve sustainability goals and reduce fossil fuel usage, resulting in lower fossil carbon dioxide (CO₂) emissions.⁶ However, combustion of petrol, diesel, and biofuels and usage of engine lubricants result in exhaust containing common poisons such as alkali metals, sulfur, phosphorus, and ash, causing contamination of the catalyst and lowering overall deNOx performance in EATS.^7,8 Among the contamination sources, sulfur has been measured to have severe effects on the activity of the SCR catalysts.⁹⁻¹⁶ Sulfur is present in the form of sulfur oxides (SOx) after combustion of the fuels and originates mainly from the fuel. SO₂-exposed Cu-CHA shows reduced deNOx performance compared to degreened Cu-CHA.^11,12 SO₃-exposed Cu-CHA, however, shows critically reduced deNOx performance and involves larger sulfur storage compared to SO₂-exposed Cu-CHA.^9,17 Thus, there is a strong demand to find strategies to reduce the effects of sulfur to maintain the catalytic activity during the lifetime of the catalyst.⁷ In modern diesel EATS, a diesel oxidation catalyst (DOC) plays an important role to increase the NO₂ content, which promotes NO_X reduction and passive regeneration of the SCR and diesel particulate filter (DPF), respectively.¹⁸ Unfortunately, SO₂ is mainly oxidized via the DOC and coated DPF, resulting in SO₃ and H₂SO₄ (formed with water via SO₃+H₂O → H₂SO₄) in the exhaust gas mixture.^18,19

Although Europe follows the fuel standard (EN590) limiting the sulfur concentration to a maximum of 10 ppm(w), a substantial amount of sulfur (around 2.5 kg of S) could reach the catalysts in the EATS for heavy-duty applications during its lifespan.^8,20 Periodic regeneration of the DPF (also referred to as active regeneration) can cause temperatures reaching over 650 °C in the filter, which means that also the downstream SCR catalyst is exposed to these high temperatures.²¹ During the high temperature, the accumulated surface sulfur species can leave the surface of the oxidation catalysts (DOC/DPF) whereupon the resulting sulfur oxides (SO₂, SO₃, and H₂SO₄) reach the SCR catalyst. According to the SOx-TPD with an EATS configuration (SCR+DOC+DPF+SCR+ASC), the DOC is the main contributor of the SO₃/H₂SO₄ generation.²² With a sulfated SCR catalyst, the adsorbed sulfur species cause CuO formation during the desulfation event. This desulfation event can cause critical loss of the SCR active copper ions to inactive Cu_xO_y species without severe zeolite structure deterioration.²³

NH₃–SCR reactions undergo different reaction schemes depending on the NO and NO₂ ratio in the exhaust gas mixture.^24,25 With only NO present, the NH₃–SCR reaction includes O₂ in the reaction stoichiometry and the reaction follows the so-called standard SCR reaction.²⁴ The SCR catalyst can be exposed to NO₂, which can be formed via an oxidation catalyst located upstream of the SCR catalyst. When using equimolar NO/NO₂ amounts, the NOx reduction rate is enhanced compared to standard SCR, in the so-called fast SCR reaction.²⁵ However, when the NO₂ content is high, the slow NO₂ SCR reaction occurs, which reduces the reaction rate even when compared to the standard SCR reaction.²⁵

Standard SCR: 4NO + 4NH₃ + O₂ → 4N₂ + 6H₂O

Fast SCR: NO + NO₂ + 2NH₃ → 2N₂ + 3H₂O

NO₂ SCR (slow SCR): 3NO₂ + 4NH₃ → 3.5N₂ + 6H₂O

A variety of Cu⁺/Cu²⁺ active sites have been identified depending on the temperature and gas composition in Cu-CHA. At a low temperature, H₂O or NH₃ solvation of the isolated copper sites causes framework-detached mobile Cu species such as Cu²⁺(H₂O)_x, Cu²⁺(OH)(H₂O)_x, Cu²⁺₂O₂(NH₃)₄, and Cu⁺(NH₃)₂.²⁶ At a high temperature, H₂O or NH₃ desorption from these mobile Cu-monomer or Cu-dimer species produces framework-bound Cu-monomer (ZCu, ZCu(OH), Z₂Cu)²⁶⁻²⁹ and oxygen- or hydrogen-containing dimer (Z₂Cu₂O, Z₂Cu₂O₂, Z₂Cu₂O_xH_y) species.^27,30−32 The Z denotes a negatively charged framework, and its subscription represents the number of the site. However, under standard SCR conditions at low temperatures (below 250 °C) in the presence of ammonia, the copper species are mobile and the NH₃-solvated copper dimers ([Cu₂(NH₃)₄O₂]²⁺) and monomers ([Cu(NH₃)₂]⁺) are considered to be key intermediates in the SCR cycle.^2,24,26,33

Bimodal N₂O release is measured with a peak at low (below 300 °C) and high (above 300 °C) temperatures during exposure of Cu-CHA in standard SCR conditions.³⁴⁻³⁶ According to the standard SCR mechanism at low temperatures, recent experimental studies have shown that N₂O is only formed in the reduction half cycle (RHC) during the catalytic turnover of the standard SCR cycle.³⁷ Ab initio calculations describing H₂NNO decomposition over [Cu₂(NH₃)₄OOH]²⁺ complexes explain not only N₂O formation during the RHC but also the trend of increasing N₂O formation as a function of Cu content.³⁶ Low NO oxidation rates compared to NH₃–SCR rates on Cu-CHA have been reported by Paolucci et al.,²⁸ suggesting that NO oxidation is irrelevant to the standard SCR at low temperatures. In addition, the N₂O formation rate is an order of magnitude higher than the NO₂ formation rate.³⁷ A pathway where N₂O is formed via ammonium nitrate (AN) is, thus, not likely to contribute to the N₂O formation during low-temperature standard SCR. In contrast to the standard SCR, the extent of N₂O formation is significantly increased when the NO₂ fraction is increased in the gas mixture⁵ and the 2+ oxidation state of Cu ions is predominantly observed from in situ XAS over Cu-CHA under fast and NO₂ SCR conditions,^38,39 suggesting that different reaction schemes are involved with respect to the fast and NO₂ SCR conditions. Kubota et al.⁴⁰ and Liu et al.⁴¹ demonstrated that AN is an intermediate species for N₂O under fast/NO₂ SCR conditions. Liu et al.⁴¹ proposed that the fast SCR follows a dual-site mechanism where derived intermediate species (HONO and H₂NNO) are decomposed at Brønsted sites in a similar manner as the standard SCR scheme suggested by Chen et al.⁴² Han et al.⁵ proposed that the CHA structure is favorable to form surface nitrates due to promotion of the NO₂ disproportionation within the CHA cage compared to the MFI and BEA structures, and thereby more AN is observed in the CHA structure, followed by MFI and BEA.

After SO₂ exposure, only small or negligible amounts of stored sulfur has been observed in H forms of zeolites upon sulfur exposure,⁴³⁻⁴⁵ indicating that SO₂ storage mainly requires the presence of Cu ions. Physical and chemical poisoning is reported as the sulfur deactivation mechanism in Cu zeolites. SO₂ and NH₃ exposure to Cu-CHA shows decreasing micropore volume with increasing S content, suggesting that pore blocking occurs by ammonia interacting with derived sulfur species.⁴⁶ In comparison, only SO₂ exposure shows no considerable micropore volume decrease with S content.⁴⁶ Temperature-programmed desorption (TPD) tests with an SO₂+SCR gas composition gives bimodal SO₂ peaks at low (∼400 °C) and high (↑600 °C) temperatures where ammonium sulfate and Cu sulfate are associated, respectively.^{10,45,47−49} After regeneration, ammonium sulfate is decomposed and thus it causes a reversible deactivation.^9,47 However, chemically interacting sulfur species at Cu sites (i.e., ZCuHSO₄) or Al sites (i.e., Al₂(SO₄)₃) are not fully regenerated and these derived sulfur species are highly stable.^9,50 Furthermore, a fairly high temperature (over 700 °C) is necessary to fully desulfate these highly stable sulfur species over Cu-CHA.^9,50 A highly thermal regeneration event can result in a hydrothermal aging (HTA) effect that converts the Cu ions to inactive Cu_XO_y and CuAlO_x species as well as CHA structure degradation.⁵¹ Therefore, high-temperature regeneration can result in irreversible thermal deactivation.

The current mechanistic understanding of the deactivation of SCR catalysts resulting from sulfur is mainly focused on SO₂ exposure of Cu-CHA. There is a strong consensus that SO₂ readily interacts with ZCuOH and that the interaction with Z₂Cu is weaker.^{45,48,52−54} In situ XAS shows that SO₂ interaction with Cu ions, especially NH₃-solvated mobile Cu dimers ([Cu₂(NH₃)₄O₂]²⁺), is particularly sensitive to SO₂ as compared to other-state Cu ions such as framework-interacting Cu⁺/Cu²⁺ (ZCu/Z₂Cu) or NH₃-solvated Cu monomers ([Cu(NH₃)₂]⁺).⁵⁵ Another XAS study shows that SO₃ exposure does not change the local structure of Cu sites within the zeolite framework (i.e., Cu–O coordination number and the Cu–Cu distance, etc.), indicating that Cu sites remain highly dispersed after sulfation as well as following desulfation.⁵⁶ First-principles calculations performed by Bjerregaard et al.⁵⁷ agree with the experimental observations and suggest that the standard SCR deactivation at low temperatures is due to hindered intercage diffusion by ammonium (bi)sulfate species. Thereby, the pairing of Cu-monomer complexes ([Cu(NH₃)₂]⁺) via O₂ activation is hindered, which is a key intermediate step of the oxidation half cycle (OHC) during standard SCR at low temperatures.³⁴ The mechanism proposed by Bjerregaard et al.⁵⁷ rationalizes the experimental findings, where the NO conversion and the N₂O formation are reduced upon SO₂ exposure of Cu-CHA during the standard SCR at low temperature.^12,58,59 Mesilov et al.¹² reported that NOx conversion and N₂O selectivity are decreased for the SO₂-exposed Cu-CHA (Si/Al ≈ 7, 3 wt % Cu) during the standard, fast, and NO₂ SCR reactions. After regeneration, the NOx conversion and N₂O selectivity were not fully regenerated for the three overall SCR reactions but the N₂O selectivity was noticeably increased for the NO₂ SCR compared to the degreened state. A comparative study between the Cu-CHA aged in EATS (270,000–710,000 miles) and HTA+SOx-treated Cu-CHA suggests that sulfur species assists CuO formation without a clear structural damage during desulfation.²³ In contrast to the SO₂ exposure, SO₃ exposure to Cu-CHA results in larger sulfur uptake and more severe irreversible deactivation.⁹

For industrial applications, understanding the fundamentals of SO_X chemistry in Cu-CHA as part of an NH₃ SCR system is critically important to establish an effective desulfation strategy that maintains catalytic activity throughout the usage lifetime of the SCR catalyst. Limited studies are reported with respect to the SO₃ exposure compared to the SO₂ exposure effects on Cu-CHA in terms of its deNOx performance and N₂O release. There are some studies that have reported the SO₂ exposure effect on the N₂O formation.^12,37,58,60 However, to the best of our knowledge, there is no comparative study of the evaluation of the AN and N₂O formation for the SO₂- and SO₃-exposed Cu-CHA, which is the objective of the current study.

2. Experimental and Computational Methods

2.1. Catalyst Synthesis and Monolith Sample Preparation

Copper-exchanged SSZ-13 (CHA) was prepared by following the hydrothermal synthesis method to acquire a Si to Al molar ratio of ≈15 and ca. 1 wt % Cu. Specifically, 1 wt % Cu was set to favor the exclusive presence of ion-exchanged Cu ions inside CHA cage and avoid possible formation of CuO particles. The acquired Cu/SSZ-13 powders were washcoated on honeycomb monolith substrates (cordierite, 400 cpsi, 15 mm × 20 mm). The specific synthesis and washcoating procedure are described in the Supporting Information (S1. Catalyst synthesis and monolith preparation). The resulting fresh Cu/SSZ-13 powder and monoliths were degreened under standard SCR conditions (400 ppm of NH₃/NO + 10% O₂ + 5% H₂O + Ar as a balance) at 750 °C for 5 h prior to powder sample characterizations and experiments in a synthetic gas bench reactor.

2.2. Synthetic Gas Bench (SGB) Reactor Setup

An SGB reactor was used to perform degreening of fresh catalyst powders, SO_x treatment, and AN temperature-programmed desorption (AN-TPD). Mass flow controllers (MFCs, Bronkhorst) and a controlled evaporator and mixing (CEM, Bronkhorst) system were equipped to provide the required gases (Ar, O₂, NH₃, NO, NO₂, N₂O, N₂, H₂O, and SO₂). The gas lines were carefully insulated and heated to 191 °C to avoid water condensation and deposition of solids such as AN. Ar was used as an inert balance for all the experiments, and the total flow was 1.2 NL·min^–1. The required gas mixture was fed to a horizontal quartz tube (inner diameter: 16 mm). The horizontal quartz tube was wrapped with a heating coil and insulated to heat up and control the reactor temperature. The prepared sample was located inside the quartz tube during the pretreatment and experiments.

2.3. SO₂ and SO₃ Exposure to Degreened Cu-CHA

The prepared monoliths were pretreated in the SGB reactor with SO₂ and a SO₂+SO₃ mixture, which will be referred to as SO₂ and SO₃ exposure treatments. First, the monolith was degreened under standard SCR conditions (400 ppm of NH₃, 400 ppm of NO, 10% O₂, and 5% H₂O) at 750 °C for 5 h, following the SO₂ or SO₃ treatment sequence, consisting of the following steps: first step: 30 ppm of SO₂ + base feed (10% O₂ + 5% H₂O + Ar) for 1 h; second step: 400 ppm of NH₃ + base feed for 2 h; third step: 30 ppm of SO₂ + base feed for 1 h. The reason for applying the ammonia in a separate step between the sulfur steps and not simultaneously with SOx is that feeding SO₃ and NH₃ simultaneously causes large ammonium sulfate deposits in the reactor, which could block the lines, even when heating the lines thoroughly. For the SO₂+SO₃ treatment, an external device connected upstream from the SGB reactor was used to generate SO₃. Pt/Al₂O₃ (7.5 wt % Pt) was used as an oxidation catalyst in the SO₃ generator. The specific procedures are described in the Supporting Information (S2. SO₃ calibration and SO_x-exposure of degreened monoliths). This provided the capability to introduce 24 ppm of SO₃ + 6 ppm of SO₂ with base feed to the monolith instead of 30 ppm of SO₂, as illustrated in Figure S2. Meanwhile, the temperature of the monolith was maintained at 400 °C during the SOx treatment to mimic the SOx exposure environment to the SCR catalyst during periodic regeneration of the DPF. Note that for simplification from here on, the SO₂ and SO₂+SO₃ treatments shall be referred to as SO₂ and SO₃ exposure, respectively.

2.4. Catalyst Characterization

Elemental analysis was performed with powder forms of the degreened and SO₂- and SO₃-exposed samples. The SOx-exposed samples’ washcoats were carefully scraped from the SO₂- and SO₃-exposed monoliths. Afterward, the SOx-exposed washcoat samples were used for the characterizations. However, for diffuse reflectance infrared Fourier transform spectroscopy (ICP) and diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) results of the sulfated samples, crushed monoliths were used.

2.4.1. Catalyst Powder Sample Characterization

Elemental compositions of the fresh catalyst powders were analyzed using inductively coupled plasma sector field mass spectrometry (ICP-SFMS) at ALS Scandinavia AB. For sulfated samples, washcoated sulfated monoliths were crushed well into a fine powder form to acquire the S/Cu molar ratio; afterward, the sulfur content of the sulfated samples was determined.

Powder X-ray diffraction (XRD) was performed with degreened and SO₂- and SO₃-exposed samples to confirm the crystalline structure of CHA. A Siemens D5000 diffractometer operating at 40 kV, using the Kα₁ radiation of a Cu anode as X-ray source (λ = 1.54060 Å), was used for measuring the X-ray diffractogram from the powder samples.

N₂ physisorption was performed to measure specific surface area using the Brunauer–Emmett–Teller (BET) method of the degreed catalyst powder. A Micromeritics TriStar 3000 instrument was used to measure the N₂ adsorption/desorption isotherms at 77 K. Prior to the measurements, the H and Cu forms of the degreened catalyst powders were degassed at 225 °C for 8 h.

UV–visible diffuse reflectance spectroscopy (UV–vis-DRS) was performed to investigate the formation of copper and ammonium sulfite/sulfate species. A Lambda 365 (PerkinElmer) double-beam UV visible spectrophotometer was used to acquire the UV spectrum. The system is composed of two different lamps (D2 and tungsten lamp) for measurement within the UV and visible-light regions. The lamp change occurred at 380 nm during the measurement. The light is split into two beams before it interacts with the sample. One of the beams is used as a reference, and the other beam reaches the sample. Two detectors (photodiodes) measured the light intensity, which came from the reference and sample beams simultaneously. Detector and reflectance sphere (sphere diameter: 60 mm, sample port aperture: 12.5 mm, sphere coating: Barium sulfate) modules were installed. Thereafter, a light trapper and a white standard were used for the calibration of UV intensity. Finally, the prepared catalyst powder was transferred to a powder cell for measuring the powder sample. The sample was scanned from 1100 to 210 at 5 nm slit width and 480 nm·min^–1 scanning speed.

2.4.2. H₂ Temperature-Programmed Reduction (H₂-TPR)

H₂-TPR was carried out with a differential scanning calorimeter (Sensys DSC calorimeter from Setaram) to investigate the reduction properties of the degreened and SO₂- and SO₃-exposed catalyst powders. The powders were carefully sieved to 180–250 μm fraction. Degreened (ca. 50 mg) or SOx-exposed (ca. 26 mg) samples were loaded into a quartz tube located inside the DSC, and 0.2% H₂/Ar was continuously fed at 20 N mL·min^–1 during the experiment. Initially, the temperature was set at 25 °C for 20 min and then was ramped up to 800 °C at 10 °C·min^–1; thereafter, 800 °C was maintained for 20 min. Note that there was no pretreatment done prior to H₂-TPR. In the meantime, the outlet gases from the DSC were monitored with mass spectrometry (Hidden HPR-20 QUI MS) for H₂ (m/e = 2), NH₃ (15), H₂O (18), Ar (20), O₂ (32), H₂S (34), SO₂ (64), SO₃ (80), and H₂SO₄ (98).

2.4.3. Ammonium Nitrate Temperature-Programmed Desorption (AN-TPD)

AN-TPD was performed to investigate the effect of SO₂ and SO₃ on AN and N₂O formation in the SGB reactor. 200 ppm of NH₃/NO₂ + 5% H₂O + Ar was fed to the monolith for 90 min at 150 °C, followed by purging with 5% H₂O+Ar for 90 min. Thereafter, the temperature was increased to 550 °C at 10 °C·min^–1 and then maintained at 550 °C for 20 min while the monolith was exposed to 5% H₂O+Ar. Note that no pretreatment was done prior to AN-TPD for the SO₂- and SO₃-exposed samples, to avoid the removal of sulfur species. However, the degreed sample was pretreated before AN-TPD. An empty-tube test was done prior to AN-TPD to check for possible AN formation. The test result showed no N₂O formation during temperature ramping, indicating that the system was sufficiently insulated to prevent any significant AN formation during the NH₃+NO₂ feed to the SGB. Meanwhile, the N₂ (m/e = 28) signal evolved while 200 ppm of NO₂/NH₃ was fed, suggesting that NO₂–SCR occurred during ionization within MS.

2.4.4. In Situ DRIFTS Measurements

In situ DRIFTS was performed to monitor surface nitrite/nitrate and NH₃ interaction with the prepared samples (i.e., degreened and SO₂- and SO₃-exposed samples). A VERTEX 70 FTIR spectrometer (Bruker) equipped with a liquid N₂-cooled mercury cadmium telluride (MCT) detector was used to acquire the IR spectra. Degreened powder samples or crushed SOx-exposed monoliths were loaded into a sample cup inside a reaction cell. A KBr bed was set on the bottom of the sample cup, and then the remaining volume of the sample cup was filled with sample. The reaction cell was covered well and mounted with a Ca₂F window. A thermocouple was installed at the center of a sample bed to monitor the sample bed temperature during the measurement. The temperature of the sample bed was controlled at 150 °C with a PID regulator (Eurotherm 2416), and MFCs (Bronkhorst Hi-Tech) were set to supply the required gas mixture to the reaction cell. 1% H₂O + Ar was fed as base feed throughout the measurement. The required gases were fed in the following steps: (1) heating to 150 °C in base feed only and (2) 400 ppm of NO₂ + base feed for 1.5 h. Meanwhile, DRIFTS spectra were measured (accumulation of 60 scans, resolution of 4 cm^–1) over time under 100 N mL·min^–1 of total flow. The outlet gas leaving the reaction cell was continuously monitored with mass spectrometry (Hidden HPR-20 QUI MS). Note that the reported temperature in the DRIFTS test was monitored by a thermocouple located at the center of the sample bed. However, the heating plate located under the sample cup was set at 242 °C. Thus, a temperature gradient resulting from thermal radiation was present and expected.

2.4.5. First-Principles Calculations

Spin-polarized density functional theory (DFT) calculations were performed using the Vienna Ab initio Simulation Package (VASP).^61,62 The Kohn–Sham orbitals were expanded with plane waves with a 480 eV cutoff value, and the interaction between the core and valence electrons was described by the projector augmented wave (PAW)^63,64 method. The valence electrons treated for each atom were Cu(11), Si(4), S(6), Al(3), O(6), and H(1). The k-point sampling was restricted to the gamma point. The Perdew–Burke–Ernzerhof (PBE)⁶⁵ functional was used to describe the exchange-correlation effects. To account for van der Waals interactions, the Grimme 3D approach⁶⁶ was applied and a Hubbard U-term was used to describe the localized Cu 3d electrons, with the U value set to 6 eV.

The convergence criteria for the self-consistent field (SCF) loop was set to 10^–5 eV, and structures were considered to be at a minimum if the norm of all forces was less than 0.02 eV/Å. The transition states were located using the Climbing Image Nudged Elastic Band (CI-NEB) approach^67,68 and was confirmed by a single imaginary frequency. Vibrational analysis was performed using the finite difference method. To sample low energy configurations, ab initio molecular dynamics (AIMD) simulations were carried out at 300 K using a Nosé–Hoover thermostat in the NVT ensemble.^69,70 Several structures along the trajectories were extracted and relaxed to determine the low-energy structures. Bader charge analysis was performed using the implementation from the Henkelman group.^71,72 To describe the CHA cage, a hexagonal unit cell containing 36 Si and 72 O was used with the experimental lattice constants (13.8026 Å, 13.8026 Å, 15.0753 Å). For each Cu ion introduced into the unit cell, one Si atom was replaced by one Al atom for charge compensation.

3. Results

3.1. First-Principles Calculation on SO_x Adsorption over Cu Sites

Two experimentally observed resting states of the Cu²⁺ ions in CHA are Z₂Cu and ZCuOH.^26,73 Depending on the sample composition and pretreatment, other Cu²⁺ sites may form and it is important to consider which Cu sites may be present under the conditions investigated. In our study, the samples have been exposed to SO_x at 400 °C in the presence of O₂ and H₂O. At 400 °C, Cu²⁺ is mainly bound to the framework and we would therefore expect to have both Z₂Cu and ZCuOH present. In addition, the presence of O₂ could lead to the formation of Cu dimers as copper ions can possibly migrate, forming a pair of Cu²⁺ ions, which can adsorb O₂ forming Z₂CuO₂Cu complexes.⁷⁴ Other similar dimer sites have been proposed such as Z₂Cu–O–Cu and Z₂CuO₂H₂Cu.^30,75,76 The Cu sites that will be investigated in the computational study are Z₂Cu, ZCuOH, and Z₂CuO₂Cu.

To explore the effect of SO₂ and SO₃, we calculated reaction pathways for the formation of CuHSO₃ and CuHSO₄ on Z₂Cu, ZCuOH, and Z₂CuO₂Cu. The reaction of SO₂ on Z₂Cu and ZCuOH has previously been studied using DFT calculations by Jangjou et al.,⁴⁵ and they found that the addition of SO₂ results in the formation of more stable CuHSO₃ complexes on ZCuOH than on Z₂Cu. In addition, they found that the formation of CuHSO₃ exhibited a lower activation barrier when it was formed on ZCuOH.⁴⁵ Furthermore, the stability of CuHSOx complexes from the reaction of SO₂ and SO₃ with ZCuOH has been considered with DFT calculations¹³ and it was found that adsorption of SO₃ gives rise to more stable complexes than SO₂ adsorption. However, the understanding of the reaction of SO₃ with ZCuOH and Z₂Cu and the role of Z₂CuO₂Cu upon sulfur exposure is currently limited. Herein, we investigate the reaction pathway for the combinations of SO₂ and SO₃ on Z₂Cu, ZCuOH, and Z₂CuO₂Cu, which are Cu sites expected to be present during the first sulfation step in our study.

The reaction of SO₂ and SO₃ with Z₂Cu is calculated in the presence of a single adsorbed H₂O molecule. If there is no adsorbed H₂O molecule on Z₂Cu, it was not possible to form any stable structures with SO₂ and SO₃, which is in agreement with a previous observation by Jangjou et al.⁴⁵ for SO₂ reacting with Z₂Cu. For the reaction of SO₂ with Z₂Cu (Figure 1a), H₂O adsorbs on the Cu ion with a Cu–O bond length of 1.95 Å and an adsorption energy of 0.87 eV (I → II). The introduction of SO₂ in the zeolite cage is preferred by 0.55 eV, compared to SO₂ in the gas phase; however, we do not find any bond to the Cu ion (II → III). Instead, the oxygen in SO₂ coordinates to Cu with a bond length of 2.49 Å. The first step in the reaction of SO₂ with Z₂Cu–H₂O results in the cleavage of the H–OH bond, forming a CuHSO₃ complex and a Brønsted acid site (BAS), which is endothermic by 0.52 eV (III → IV). From a Bader charge analysis, sulfur in CuHSO₃ is in a +4 oxidation state, as is the case for the gas-phase SO₂. In IV, sulfur binds through an OH group, which is not the preferred configuration. In structure IV, the −OH group in CuHSO₃ coordinates with the BAS with an SO-HBAS bond length of 3.73 Å. In the next step (IV → V), the CuHSO₃ complex is rotated. The barrier for the rotation (0.56 eV) is related to the change in which group is coordinating with the Brønsted site. In structure V, an oxygen from CuHSO₃ is coordinating with the Brønsted site resulting in a slight stabilization. The last step is the breakage of the SOH–Cu bond and formation of a SO–Cu bond (V → VI), which is 0.29 eV lower in energy. The final structure is 0.15 eV higher in energy compared to the case with SO₂ physisorbed in the cage (structure III). Moreover, as the formation of structure Cu is associated with a high activation barrier, we conclude that the reaction of SO₂ with Z₂Cu is not favorable.

Figure 1.

Energy landscapes for SO₂ and SO₃ adsorption onto Z₂Cu (a, b) and ZCuOH (c, d) and the subsequent formation of CuHSO_x complexes. Atomic color codes: H (white), N (blue), O (red), Al (pink), Si (dark yellow sticks), S (yellow), and Cu (bronze).

SO₃ can also react with Z₂Cu, as shown in Figure 1b. The first step is also, in this case, the adsorption of H₂O (I → II). Introducing SO₃ into the cage does not result in adsorption; instead, SO₃ is physisorbed (exothermic by 0.51 eV) with one oxygen coordinated to Cu with a bond length of 3.3 Å (II→ III). SO₃ can react with H₂O, splitting the O–H bond in H₂O and forming a CuHSO₄ complex and a Brønsted site (III → IV). From Bader charge analysis, sulfur remains in a +6 oxidation state, as is the case for the gas-phase SO₃. In structure IV, the OH group in CuHSO₄ is coordinating with HBAS. The next step (IV → V) is related to a change of coordination, resulting in the oxygen of CuHSO₄ coordinating to HBAS with a bond length of 1.64 Å. The scission of the SOH–Cu bond has a small barrier of 0.11 eV and is preferred by 0.15 eV (V → VI). The last step is the rotation of CuHSO₄, resulting in the formation of a second SO–Cu bond (VI → VII). This step has a barrier of 0.41 eV and is exothermic by 0.11 eV. The final structure (VIII) is 0.75 eV more stable compared to the case of having physisorbed SO₃ (structure III). When comparing the reaction of SO₂ and SO₃ with Z₂Cu, it becomes clear that SO₃ forms ZCuHSO₄ complexes that are more stable (0.86 eV in difference) than ZCuHSO₃ complexes formed by the reaction with SO₂. The highest activation barrier for formation of ZCuHSO₄ is 0.41 eV (VI → VII), while the effective barrier is 1.46 eV (III → V) for formation of ZCuHSO₃. This suggests that SO₃ may result in more severe deactivation than SO₂. The higher stability of sulfur species in Cu-CHA with the oxidation state of +6 is in agreement with previous DFT studies.⁷⁷

The reaction of SO_x with ZCuOH can proceed in the absence of H₂O. The introduction of SO₂ into the CHA cage is preferred by 0.25 eV (I → II, Figure 1c). Next, SO₂ can react with ZCuOH forming a CuHSO₃ complex that is preferred by 0.14 eV and associated with a small barrier of 0.2 eV (II → III). The last step is the rotation of the sulfur complex, resulting in the breakage of a SOH–Cu bond and, at the same time, the formation of a SO–Cu bond. This step has a barrier of 0.39 eV and is 0.41 eV lower in energy. The stable CuHSO₃ complex (structure IV) is 0.55 eV more stable compared to physisorbed SO₂ (structure II).

The reaction of SO₃ with ZCuOH is shown in Figure 1d. Physisorption of SO₃ is exothermic by 0.56 eV (I → II). SO₃ can react with ZCuOH forming a CuHSO₄ complex that is 0.77 eV more stable and associated with a barrier of 0.16 eV (II → III). The following steps are connected to the reorientation of the CuHSO₄ complex away from the eight-membered ring. The first step has a barrier of 0.13 eV and is endothermic by 0.08 eV (III → IV). The second step has a barrier of 0.12 eV and is exothermic by 0.33 eV (IV → V). The last step is the rotation of the CuHSO₄ complex, resulting in the breakage of a SOH–Cu bond and the formation of a SO–Cu bond (V → VI). The final CuHSO₄ complex (structure VI) is 1.3 eV more stable than physisorbed SO₃ (structure II).

For ZCuOH, the same trend as for Z₂Cu is observed, where SO₃ forms more stable sulfur complexes (−1.3 versus −0.55 eV), which is associated with lower activation barriers compared to the SO₂-derived sulfur complexes (0.22 versus 0.39 eV).

As the SOx exposure is performed in the presence of O₂ at 400 °C, a pair of Cu ions may adsorb O₂ forming framework-bound peroxo species, Z₂CuO₂Cu.⁷⁴ The reaction landscape for the reaction of SO₂ and SO₃ with Z₂CuO₂Cu is shown in Figure 2. The starting structure I could originate from a pair of Cu⁺ ions that adsorb O₂.⁷⁴ SO₂ cannot adsorb directly onto the complex (structure II). Instead, SO₂ can react (low barrier of 0.16 eV) with the adsorbed O₂ molecule forming a stable Z₂Cu(SO₄)Cu complex (II → III), which is 2.68 eV more stable. During the reaction, sulfur is oxidized from +4 to +6. The strongly exothermic nature of sulfur oxidizing for Cu-CHA has been observed previously for SO₂ oxidation over the mobile peroxo complex [Cu₂O₂(NH₃)₄]²⁺.⁵⁷ The Cu ions remain in a +2 oxidation state, supplying the two electrons needed to form the sulfate species (SO^2–₄). This complex cannot react directly with SO₂ or SO₃; however, in the presence of H₂O, it is possible to form CuHSO₃ and CuHSO₄ complexes. H₂O can adsorb onto Cu with a binding energy of 0.46 eV (III → IV). The next step is the breakage of a SO–Cu bond (IV → V). Furthermore, a bond between Cu and an oxygen in the framework (O_fw) is broken and a new Cu–O_fw bond is formed. The configuration of structure V is favored for the formation of sulfur complexes. The next step (V → VI) is endothermic by 0.26 eV and is the result of the coordination of hydrogen from H₂O, which changes from one oxygen to another that is bonded to sulfur. The next steps follow two different paths, which involves the reaction with either SO₂ or SO₃ indicated in black or red with a star (*), respectively. SO₂ (black) can react with H₂O, splitting the O–H bond of H₂O and forming a CuHSO₃ and CuHSO₄ complex (VII → VIII), which is endothermic with 0.35 eV. The next step (VIII → IX) is the change of the coordination of an OH group bound to sulfur, which is turned away. The last step is the rotation of CuHSO₃ that leads to the breakage of a SOH–Cu bond and formation of a SO–Cu bond (IX → X). This step has a barrier of 0.56 eV and is exothermic by 0.21 eV. From structure VI, it is also possible to follow the structure denoted with an *, which is the reaction with SO₃. Here, SO₃ can react with H₂O forming two CuHSO₄ complexes (VII* → VIII*). The formation of CuHSO₄ is associated with a small barrier of 0.11 eV and is exothermic by 0.38 eV. As in the previous case, the next step involves the rotation of HSO₄, so the −SOH group is no longer bonded to Cu (VIII* → IX*). The rotation has a barrier of 0.4 eV and is exothermic by 0.4 eV.

Figure 2.

Energy landscape for the reaction of SO₃ and SO₂ with Z₂CuO₂Cu and the subsequent formation of ZCuHSO_x complexes. Atomic color codes are as in Figure 1.

From the reaction landscapes, we conclude that more stable sulfur complexes are formed when reacting Z₂CuO₂Cu with SO₃ compared to SO₂. This is similar to the reactions on Z₂Cu and ZCuOH. In addition, the activation barriers associated with the reaction with SO₃ are lower than for SO₂. Compared to Z₂Cu and ZCuOH, the first reaction with SO₂ (II → III) forms more stable complexes suggesting that Z₂CuO₂Cu causes more severe deactivation. To form Z₂CuO₂Cu, the catalyst must be exposed to O₂. When SO₂ is cofed with O₂, severe deactivation has been experimentally observed,⁴⁹ which is consistent with our calculations. Furthermore, it is possible to oxidize SO₂ to SO^2–₄ over the Z₂CuO₂Cu and the presence of sulfur in +6 oxidation states has been observed using X-ray absorption spectroscopy.⁷⁸ The sulfur-derived complexes formed when reacting SOx with Z₂CuO₂Cu are the same as those for the cases with Z₂Cu and ZCuOH. This indicates that even though more sulfur complexes may form on Z₂CuO₂Cu, it is difficult to unravel the origin of the species. However, the proximity of two CuHSO₄ ions could potentially stabilize the complexes. The observation that SO₃ forms more stable sulfur complexes that are associated with lower barriers compared to SO₂ for all Cu sites is consistent with the observation of more SOx release in TPD experiments.⁹ In addition, it is observed that SO₃ in general causes more deactivation of the low-temperature NH₃–SCR.⁹

3.2. Elemental Analysis, N₂ Physisorption, and XRD Analysis of Powder Cu/SSZ-13

Elemental analysis and N₂ physisorption were conducted for the prepared copper-exchanged SSZ-13 powders, and the results for the degreened samples are shown in Table S1. SAR (Si/Al molar ratio) was around 14, and the Cu content was 0.83 wt %. The BET surface area decreased (from 711 to 651 m²·g^–1) after impregnation of Cu ions. With respect to sulfated samples, 33 and 159 μmol·g^–1_washcoat of sulfur content were obtained for SO₂- and SO₃-exposed samples, respectively. In Figure S3, a comparison of X-ray diffractograms for the degreened H and Cu forms of SSZ-13 shows that the CHA framework structure was well maintained without noticeable diffraction peaks of Cu_xO_y crystalline particles (36.30 and 38.72° 2θ)⁵ or Cu(OH)₂ precipitates (12.83 and 25.84° 2θ)³⁷ in the sample powders. Note that the fresh sample was degreened in standard SCR conditions at 750 °C for 5 h and a low Cu loading amount was set (ca. 1 wt % Cu) to avoid possible formation of copper-containing particles. This suggests that no or minor amounts of Cu_xO_y particles were formed and if any was formed; they were below the detection limit of the XRD. It is, thus, assumed that most copper is present as isolated Cu ions interacting with 1 or 2 Al_f over the CHA structure of the degreened powder.

The effect of SO₂ and SO₃ exposure to Cu-CHA was examined using SEM and XRD with the following sulfation sequence: (i) SO₂/SO₃+O₂+H₂O; (ii) NH₃ + O₂+H₂O and (iii) SO₂/SO₃+O₂+H₂O. We confirmed that no clear changes in crystal morphology from SO₂- and SO₃-exposed samples using SEM (see Figure S4) were observed compared to the degreened sample. XRD data are shown in Figure 3, and for the degreened sample, catalyst powder was used while for the SO₂ and SO₃-exposed samples, washcoat was scraped off from the monoliths. In addition, XRD of the cordierite is shown and it is clear that the cordierite present in the samples that were scraped off the monoliths can be neglected. The XRD results show shifted diffraction peaks (Figure 3). The inset (dashed purple) in Figure 3 represents the region of crystallographic planes containing [1 0 0] (ca. 9.5°), [−1 1 0] (ca. 13°), and [1 1 0] (14°).⁴⁶ Both SO₂- and SO₃-exposed samples caused peak shifting toward either lower or higher angles compared to the degreened sample, respectively. This suggests that SO₂ and SO₃ exposure at 400 °C results in changes of the CHA unit cell size but no damaged framework structure according to Bragg’s law.⁷⁹ Decreased peak intensity for SO₂- and SO₃-exposed samples indicates that sulfur species were successfully introduced into the CHA unit cell. However, questions remain as to why SO₂ and SO₃ caused different trends in the shift of the peak toward lower and higher angles, respectively. This different peak shifting trend implies that different lattice distances resulted from SO₂ and SO₃ exposure with NH₃ oxidation as the middle step.

Figure 3.

X-ray diffractogram for degreened and SO₂- and SO₃-exposed samples. The SO₂- and SO₃- exposed samples were scraped off the washcoat from monoliths. In addition, XRD from pure cordierite is shown. The inset illustrates X-ray diffraction peak shifting in the position of crystallographic planes with indices of [1 0 0] (ca. 9.5°), [−1 1 0] (ca. 13°), and [1 1 0] (ca. 14°).

UV–visible diffuse reflectance spectroscopy (UV–vis-DRS) was performed to investigate possible formation of different sulfur species. Reference UV spectra were measured for copper compounds (i.e., CuSO₄, Cu(OH)₂, CuO) and various ammonium (bi)sulfite/sulfate in hydrated and ambient conditions, as shown in Figure S5. Figure 4 presents the resulting UV spectra as a function of wavelength for degreened and SO₂- and SO₃-exposed samples. Figure 4a shows for all samples two typical UV absorption features corresponding to the d–d transition of hydrated Cu²⁺ ions (ca. at 800 nm) and the ligand to metal charge transfer (LMCT) transition caused by the isolated Cu²⁺ ions (lattice O → Cu²⁺, at ca. 220 nm) of Cu/SSZ-13, as reported in literature.^49,80,81 Moreover, UV absorption was observed for both SO₂- and SO₃-exposed samples over the entire measured wavelength, which was not the case for the degreened sample. This suggests the presence of ammonium (bi)sulfite/sulfate species since these species display considerable UV absorbance over the entire wavelength, as shown in Figure S5b. Both SO₂- and SO₃-exposed samples show in addition a broader UV absorption band at 250–300 nm compared to the degreened sample. The first derivative of Figure 4a was applied to qualitatively examine the effect of SO₂ and SO₃ with NH₃ oxidation as a middle step, as shown in Figure 4b. The result shows different features of the SO₂- and SO₃-exposed samples compared to the degreened sample, especially at 220–400 nm. Dissimilar UV absorption features between the SO₂- and SO₃-exposed samples as in Figure 4c indicates the possible formation of different types of sulfur species interacting with NH₃.

Figure 4.

UV–visible diffuse reflectance spectra for degreened and SO₂- and SO₃-exposed powder samples in terms of (a) absorbance, (b) first derivative of absorbance, and (c) first derivative of absorbance with the degreened sample subtracted.

In comparison with reference bulk materials in Figure S5, CuSO₄ (Figure S5a) and various ammonium (bi)sulfites/sulfates (Figure S5b) were considered as potentially formed sulfur species arising during SO₂ and SO₃ exposure. In the region of 220–400 nm, the broad UV absorbance for CuSO₄ and ammonium (bi)sulfite/sulfate indicates that copper sulfate can be potential sulfur species interacting with Cu²⁺ ions, and the observed additional absorbance (shoulder at 250–300 nm in Figure 4a) is likely to originate from ammonium sulfate ((NH₄)₂SO₄) or ammonium sulfite ((NH₄)₂SO₃). To exclude the background resulting from Cu-CHA with respect to the SO₂- and SO₃-exposed samples, the UV spectrum of the degreened sample was subtracted from the SO₂- and SO₃-exposed samples. In Figure 4c, the resulting first derivative of UV absorbance shows the comparison of the derived sulfur species during the SO₂ and SO₃ exposure with NH₃ saturation. By comparing the reference samples (i.e., bulk ammonium (bi)sulfite/sulfate) in Figure S5d, the result implies that the SO₂ and SO₃ exposures, with NH₃ oxidation as the middle step, are associated with the formation of ammonium sulfite ((NH₄)₂SO₃), ammonium sulfate ((NH₄)₂SO₄), or (bi)sulfate ((NH₄)HSO₄).

3.3. H₂ TPR in Degreened and SO₂- and SO₃-Exposed Cu/SSZ-13

XRD and UV–vis-DRS results suggest that different sulfated species are interacting with NH₃ and copper ions inside the CHA unit cell, resulting from SO₂ and SO₃ exposure of Cu-CHA (with ammonia oxidation as a middle step). This implies that the formed sulfated species can influence the reduction properties of isolated Cu²⁺ ions. Bergman et al.⁶⁰ showed that SO₂ exposure influences the reduction property of Cu²⁺ in Cu-CHA using in situ XAS. Thus, H₂ temperature-programmed reduction (H₂-TPR) was carried out to investigate the reduction property of the SO₂- and SO₃-exposed samples.

Figure 5a shows H₂ consumption as a function of temperature for degreened and SO₂- and SO₃-exposed samples. The degreened sample shows H₂ consumption in the temperature range of 200–650 °C, resulting from Cu²⁺ species reduction by H₂. In general, ZCuOH and Z₂Cu are regarded as primarily isolated Cu²⁺ species over Cu-CHA and the observed H₂ consumption peaks are commonly assigned to ZCuOH and Z₂Cu at low and high temperatures, respectively, due to different binding strengths of Cu²⁺ species interacting with the CHA framework.^{5,48,73,82,83} In addition, in our recent study, combining ab initio calculations with experimental results, we also found the importance Z₂CuOCu, Z₂CuHOOHCu, and Z₂CuOOCu to describe the reduction process.⁸⁴

Figure 5.

H₂ temperature-programmed reduction for degreened and SO₂- and SO₃-exposed samples showing H₂ consumption (a), desorption of SO₂ (b), and NH₃ release (c), and molar ratios of consumed H₂ and desorbed SO₂ per Cu content of samples (hydrated condition, gas feed: 0.2% H₂/Ar, heating rate: 10 °C·min^–1, total flow rate: 20 N mL·min^–1).

For the degreened sample in Figure 5a, the main reduction peak at 430 °C was impeded due to sulfur species, implying an interaction between sulfur and copper species. In contrast to the degreened sample, the sulfated samples show larger H₂ consumption. SO₂ and NH₃ signals were detected as shown in Figure 5b,c, suggesting that SO₂ and NH₃ originated from decomposition of the remaining ammonium (bi)sulfite/sulfate species. The SO₂-exposed sample shows two distinct reduction features during H₂ TPR at ca. 293 and 652 °C (Figure 5a). For the SO₃-exposed sample, relatively intense reduction peaks were observed at 657 °C with a shoulder (604 °C) and at 293 °C. The SO₃-exposed sample showed three distinct H₂ consumption peaks at 293, 604, and 657 °C, but only two reduction peaks corresponding to ca. 293 and 652 °C were observed for the SO₂-exposed sample. Thus, it is likely that the peak at 293 °C resulted from hydrogen reacting with ammonium sulfite/sulfate species and the peaks at 604 and 657 °C originated from sulfur species interacting with Cu²⁺ ions since SO₂ started to be observed above 400 °C with a peak at 577 °C in Figure 5b. In addition, an H₂S (mass 34) was measured and a weak signal was detected above 650 °C for the SO₃-exposed sample in Figure S6. Based on this, we suggest that the main hydrogen consumption peak at 657 °C might be associated with H₂S formation since it requires a two-electron transfer to form H₂S as well as cleavage of S–O/S=O bonds to become the sulfide (S^2–) of H₂S. It should be noted that the detection of H₂S in the MS is rather difficult, and we suggest that this is the reason for the significantly smaller and noisier H₂S peak compared to the hydrogen consumption peak. In contrast. no H₂S signal was observed for the SO₂-exposed sample. This is probably because of the lower amount of H₂S formation as a result of the relatively lower sulfur content in the SO₂-exposed sample compared to the SO₃-exposed sample according to the ICP results in Section S3, SI. Both SO₂- and SO₃-exposed samples showed an SO₂ desorption, with a single peak in the same temperature region (peak at 577 °C) (see Figure 5b). However, the sulfur release from the SO₃-exposed sample was significantly larger. Note that H₂SO₄ and SO₃ signals were not detected during H₂-TPR but these compounds would be challenging to detect in the calorimeter setup if they are in low concentrations. In addition, ammonia desorption was clear from both SO₂- and SO₃-exposed samples, with two desorption peaks Figure 5c.

The consumed H₂ per Cu and SO₂ release per Cu from the H₂ TPR are shown in Figure 5d. For the degreened sample without sulfur added, the H₂/Cu ratio is 0.55, indicating that a one-step reduction of copper occurred. This is common at these temperatures, since a full reduction to metallic copper usually requires higher temperatures. Larger H₂ consumption is found for the SO₂-exposed sample, which is significantly increased for the SO₃-exposed sample, simultaneously with the SO₂ desorption increasing. Larger H₂/Cu and SO₂/Cu ratios are expected if the SOx-exposed sample involves a larger sulfur content. Figure 5d suggests that SO₃ exposure results in a much larger sulfur uptake compared to SO₂ exposure. This is in line with the ICP analysis results regarding the sulfur content (Section S3, SI). Considering the observed NH₃ release and SO₂ desorption features in Figure 5b,c, it can be concluded that NH₃-sulfur interacting species remained following the sulfation and ammonia exposure step at 400 °C. Different sulfur speciation is expected to result from the SO₂ and SO₃ exposure, with NH₃ as a middle step, forming ammonium sulfite ((NH₄)₂SO₃) and ammonium sulfate ((NH₄)₂SO₄) or (bi)sulfate ((NH₄)HSO₄) species according to the UV–vis DRS results.

3.4. AN-TPD with Degreened and SO₂- and SO₃-Exposed Cu/SSZ-13

AN can easily be formed from NO₂ and NH₃, and during its decomposition, it is well known that N₂O is formed. Normally, a larger amount of N₂O is reported as the NO₂ concentration is increased during SCR reaction.³⁵ We have reported increased N₂O formation for Cu-CHA compared to H-CHA, indicating that AN interacts with copper species.⁵ It is, therefore, expected that SO₂ and SO₃ influence the AN and N₂O formation.

Temperature-programmed desorption with AN (AN-TPD) was carried out for degreened and SO₂- and SO₃-exposed samples to investigate the effects of SO₂ and SO₃ on AN and N₂O formation. The mass 28 (N₂) signal was detected, while 200 ppm of NH₃/NO₂ was fed for 90 min in the empty-tube test (see Figure S7). This suggests that the NO₂ SCR occurred during the ionization process in the mass spectrometer. In addition, no N₂O signal was seen during heating when measuring with the FTIR as shown in Figure S7, suggesting that no AN was formed in the reactor system. The results for AN-TPD over Cu/SSZ-13 are presented in Figure 6. The NH₃ concentrations during heating are presented in Figure 6a, where the NH₃ release was increased for SO₂- and SO₃-exposed samples compared to the degreened samples. An increased NH₃ release is expected due to residual sulfur species, since ammonium (bi)sulfite/sulfate species could form during the NH₃+NO₂ feed period of the AN-TPD test. These results are in line with the studies in the literature,^43,58,85,86 where ammonia was suggested to interact with sulfate species resulting in larger ammonia storage. Notably, the SO₃-exposed sample shows the largest NH₃ release; furthermore, its NH₃ release peak is also shifted from lower to higher temperatures compared to that for the SO₂-exposed and degreened samples.

Figure 6.

NH₃ (a), NO (b), NO₂ (c), and N₂O (d) concentrations during ammonium nitrate temperature-programmed desorption for degreened and SO₂- and SO₃-exposed samples (gas feed: 200 ppm of NH₃/NO₂ + 5% H₂O + Ar at 150 °C during adsorption, heating rate: 10 °C·min^–1, total flow rate: 1200 N mL·min^–1). Note that the NO concentration is during the adsorption phase.

Figure 6b shows the NO formation curves during NH₃+NO₂ adsorption for degreened and SO₂- and SO₃- exposed samples. It has been reported that NO formation originates from NO₂ disproportionation and subsequent surface nitrite oxidation by NO₂, leading to surface nitrate formation with gas-phase NO release.⁸⁷⁻⁸⁹ The ratios for NO formed divided by removed NO₂ during the adsorption phase were 0.03, 0.03, and 0.01 for the degreened, SO₂-exposed, and SO₃-exposed samples, respectively. A ratio of 0.33 is expected due to the disproportionation reaction where nitrates are formed.⁸⁹ Thus, these values are significantly lower and the reason for this is likely that we also have slow NO₂ SCR occurring at this temperature, since we use a quite high temperature (150 °C) and a large amount of the NO₂ that is consumed is due to the SCR reaction. NO formation was decreased for both SO₂-exposed and, in particular, SO₃-exposed samples compared to the degreened sample. This suggests that surface nitrate storage was decreased due to sulfur species interacting with copper species since sulfur species do not show noticeable interaction with Brønsted sites.⁴⁵ In this respect, it is postulated that most surface nitrate was formed on copper species under these conditions (presence of water, for example), since significantly reduced NO formation was observed for the SO₃-exposed sample containing the largest sulfur content. Indeed, lower NO₂ formation during the heating is found (Figure 6c), which is in line with the decreased surface nitrate storage, as NO₂ formation is expected from the decomposition of the formed surface nitrate species during the NH₃+NO₂ adsorption.

Figure 6d shows N₂O formation during heating for the degreened and SO₂- and SO₃-exposed samples. N₂O formation due to thermal AN decomposition is commonly observed.⁹⁰ The N₂O curve from the SO₂-exposed sample shifted toward slightly lower temperatures compared to the degreened sample. Moreover, a similar amount of N₂O was released from the SO₂-exposed sample (31.9 μmol·g^–1_washcoat) as for the degreened sample (31.3 μmol·g^–1_washcoat). It is slightly more N₂O amount for the SO₂-exposed sample (↑ 1.9%), but the differences could be within the accuracy of the experiments. Despite similar N₂O formation, the NO₂ release was clearly decreased for the SO₂-exposed sample (Figure 6c) as compared to that of the degreened sample, which we interpret as a result of nitrate decomposition. Thus, these results overall indicate that AN formation also decreased for the SO₂-exposed sample. However, it should be noted that some of the released NO₂ could react in the slow NO₂ SCR reaction. In contrast to the SO₂-exposed sample, the SO₃-exposed sample showed a remarkably decreased N₂O release (14.6 μmol·g^–1_washcoat) corresponding to a ca. 53% reduction in N₂O release compared to the degreened sample. The N₂O peak was also shifted toward lower temperatures compared to the SO₂-exposed and degreened samples. This suggests that SO₃ exposure has a stronger influence on N₂O formation compared to SO₂ exposure.

It should be mentioned that sulfur species such as SO₂, SO₃, H₂SO₄, and H₂S were measured by FTIR during AN-TPD. Among the sulfur species, SO₂ and H₂SO₄ were observed but the SO₃ and H₂S were below the detection limit. However, a slightly negative SO₂ signal (below 0 ppm) was observed from the SO₂- and SO₃-exposed samples that was symmetrical with the positive H₂SO₄ curve, as shown in Figure S8. Thus, it is likely that there were overlapping bands in the FTIR that were slightly overcompensated by the software for SO₂ and H₂SO₄ according to Figure S9. To ease interpretation, the H₂SO₄ and SO₂ signals were merged as SO₂+H₂SO₄ (blue curve in Figure S8), and sulfur species are denoted in this paragraph. The desorption of these sulfur species features was still maintained, showing three desorption peaks at 250, 368, and 537 °C for the SO₂-exposed sample and a monotonically increasing SO₂+H₂SO₄ curve with a peak at 533 °C for the SO₃-exposed sample. The marked different desorption features for these sulfur species suggests that SO₃ exposure causes more stable sulfur species compared to SO₂ exposure of Cu-CHA. For the SO₂-exposed sample, the H₂SO₄ peaks at 250 and 368 °C were located at same positions with two NH₃ release peaks at 180–215 min, as shown in Figure 6a. On the other hand, the SO₃-exposed sample showed an even larger NH₃ release; however, there was no significant desorption of sulfur species during NH₃ release as compared to the SO₂-exposed sample.

3.5. In Situ DRIFTS Measurement with Degreened and SO₂- and SO₃-Exposed Cu-CHA

AN-TPD results showed decreased NO formation during the NH₃+NO₂ feed for the SO₂- and SO₃-exposed samples, implying that surface nitrite oxidation with NO₂ was severely limited. Especially, the SO₃-exposed sample led to a considerably lower N₂O formation. However, the SO₂-exposed sample gave comparable N₂O formation compared to the degreened sample, although less NO₂ formation was observed, consistent with that less AN was formed, as evident from the NO formation in the adsorption phase. To corroborate on these results, surface nitrite/nitrate species were investigated with DRIFTS in the presence of 400 ppm of NO₂ + 1% H₂O + Ar at 150 °C. It should be mentioned that a relatively higher concentration of H₂O is usually introduced to mimic the real exhaust environment. However, 1% H₂O was set as the base feed here along with the background feed (Ar) since this was the maximum water feed in our DRIFTS setup. In addition, note that the powder sample was used for the degreened case, while crushed monoliths were used for the SOx-exposed monoliths. Thus, the sulfur exposed samples contain a large amount of cordierite and the spectra were therefore normalized with the amount of copper they contained.

Figure 7 shows the NO₂ adsorption phase for the degreened AND SO₂- and SO₃-exposed samples at 150 °C. The resulting spectral evolutions are illustrated in the region 2000–1000 cm^–1 to investigate the surface adsorption species. The surface nitrate region (1800–1500 cm^–1) is magnified in the right panel from the original spectrum marked with the dashed line box. The addition of 400 ppm of the NO₂ + base feed mixture (Figure 7a) results in peaks related to the formation of chelating bidentate ZCu²⁺-(NO₃) structures (1750–1500 cm^–1).⁹¹ The four bands at 1624, 1608, 1597, and 1575 cm^–1 are associated with υ(NO) vibrations of different types of chelating bidentate nitrates.^91,92 The peak at 1608 cm^–1 reached the highest intensity, when NO₂ was saturated (orange curve). Typical NO⁺ formation (∼2155 cm^–1)⁹³ was not observed here under an excess of water, due to H₂O interaction with Brønsted sites Si(OH)Al. The negative band (∼1453 cm^–1) is associated with H₂O removal.⁹⁴ The band 1898 cm^–1 was assigned as NO interaction with Cu²⁺ (i.e., Cu²⁺-NO).⁹⁵ Note that the negative peak (∼1453 cm^–1) may affect the neighboring nitrate feature.

Figure 7.

Spectrum evolution during NO₂ adsorption in DRIFTS for the (a) degreened, (b) SO₂-exposed, and (c) SO₃-exposed samples with (d) comparison of the NO₂ spectra of panels (a), (b), and (c) after 2 min of NO₂+H₂O exposure. Spectral evolution with gray represents intermediate spectra prior to NO₂ saturation (orange). Spectra in all panels were normalized by the number of moles of copper within the sample cup (gas feed: 400 ppm of NO₂ + 1% H₂O + Ar at 150 °C, total flow rate: 100 N mL·min^–1, sample bed temp.: 150 °C).

For the SO₂-exposed sample in Figure 7b, surface nitrate features were observed over the region 2000–1000 cm^–1. The four bands associated with nitrate species (1629, 1611, 1597, and 1576 cm^–1) were developed within 5 min and then consumed. Moreover, bands 1887 (positive band) and 1441 cm^–1 (negative band) were observed as well as the degreened sample. Surprisingly, at 10 min (blue curve), a decrease in the surface nitrate species peaks was observed, which were further decreased until NO₂ saturation (orange curve). During NO₂ adsorption, a decrease in the nitrates is not expected. The negative band at 1441 cm^–1, due to water removal,⁹⁴ is quite large. It is possible that there is an overlap between the 1441 cm^–1 negative band and the nitrate bands; thus, when the water removal increases, it decreases the bands in the nitrate region. This is not observed for the degreened sample, and the reason for that could be that the nitrate peaks are significantly larger; thus, the interference with the band at 1441 cm^–1 might be less. For the SO₃-exposed sample in Figure 7c, a similar trend was observed as the SO₂-exposed sample. For example, the surface nitrate region increased at 5 min (green curve), followed by decreased intensities until NO₂ saturation (orange curve). Likewise, for the degreened sample, the positive band at 1885 and negative band at 1439 cm^–1 were observed. In Figure 7d, a comparison is done between the samples after only 2 min of NO₂+H₂O exposure, where the water peaks for the SO₂- and SO₃-exposed samples are still quite small. From these results, it is clear that the nitrate bands are decreased for SO₂- and SO₃-exposed samples compared to the degreened sample.

To summarize, the DRIFT results show that there are fewer nitrates on the SO₂- and SO₃-exposed samples, as would be expected from the flow reactor experiments. However, it should be mentioned that there could be an overlap with the band for water removal, which could decrease the peak for nitrates.

4. Discussion

At low temperature, surface nitrate formation is observed in the presence of either NO+O₂ or NO₂ in Cu-CHA.^89,91,92,96 An in situ DRIFTS measurement shows that NO₂ kinetically promotes surface nitrate formation in Cu-CHA compared to NO+O₂ presence in dry conditions.⁸⁹ In wet conditions, the nitrate species formation is significantly hindered due to water interactions.⁸⁹ NO release is typically seen during isothermal NO₂ adsorption in Cu-CHA, as shown in Figure 6b as well as in the literature,^87,88,97,98 resulting from surface nitrite oxidation by NO₂, and thereby, NO is formed through the following reactions (R.1 and R.2) below. According to Ruggeri et al.,⁸⁹ when H₂O is present, Cu dimers are likely dissociated into two ZCuOH via R.3.

R.1

R.2

R.3

Ruggeri et al.⁸⁹ derived R.1–R.3 by NO or NO₂ exposure to Cu-CHA at 120 or 150 °C in dry or wet conditions using in situ DRIFTS. Cu dimers in R.1 are critically important as a surface oxygen provider, initiating surface nitrite formation and leading to the additional nitrate in the form of copper nitrate and NO via R.2. A recent study performed by Negri et al.⁹⁶ reported that possible structures of nitrate species and their local environment show that only ZCuOH in the 8-MR within the CHA cage is the active copper species forming the framework-interacting chelating bidentate nitrate (i.e., Z[Cu²⁺(NO₃)]).⁹⁶ Other nitrate candidates such as bridging bidentate or monodentate nitrate configurations were difficult to reconcile the experimental features under exposure of Cu-CHA to an NO+O₂ mixture.^91,96

The SO₂ and SO₃ exposure to Cu-CHA significantly influences the surface nitrate formation. In the AN TPD (Figure 6b), the NO release during isothermal NH₃+NO₂ adsorption is significantly decreased. Since NO production during NO₂ exposure is associated with nitrate formation, according to reactions R.1 and R.2,⁸⁹ these results suggest that the surface nitrate formation is hindered due to resulting sulfur species after SO₂ and SO₃ treatment of Cu-CHA. The lowest NO release was observed for the SO₃-exposed sample (see Figure 6b), suggesting that this sample exhibited the lowest amount of nitrates, which is consistent with the fact that it contained the largest amount of sulfur, as seen in the ICP measurements (Section S3, SI). These results are also consistent with the in situ DRIFT experiments (see Figure 7d). Our first-principles calculations shown in Figure 1b,d describe that the SO₃ interaction with both Z₂Cu and ZCuOH is possible. These findings are in agreement with ICP analysis showing an S/Cu molar ratio of 1.2, which indicates that all copper sites were occupied by sulfur species. In contrast, SO₂ primarily interacts with ZCuOH but not Z₂Cu as suggested by first-principles calculations (Figure 1a,c). Since ICP results show low sulfur storage for the SO₂-exposed sample, it suggests that there is a low amount of ZCuOH in our sample. These results are consistent with our H₂ TPR (Figure 5a) where the lower reduction peak, usually assigned to ZCuOH, is small for the degreened sample. A deconvolution of hydrogen TPR suggests only 15% ZCuOH. This value is slightly lower than the theoretical value according to Paolucci et al.,²⁶ which is 35%. The reason for this could be that in our synthesis of SSZ-13, we use Na cations and, according to Lee et al.,⁹⁹ SSZ-13 synthesis with the Na cation can increase 2Al sites over CHA. In our study, SSZ-13 was synthesized using Na and TMAda⁺, and this could result in a larger amount of 2Al, which increases Z2Cu and thereby decreases ZCuOH.

In terms of speciation of sulfur according to the first-principles calculation in Figures 1 and 2, SO₂ interaction with ZCuOH results in (bi)sulfite (ZCuHSO₃) and SO₂ interactions with dicopper cause both (bi)sulfite and (bi)sulfate surface species. Meanwhile, the interaction of SO₃ with Z₂Cu and ZCuOH leads to only (bi)sulfate (ZCuHSO₄). The (bi)sulfate resulting from the SO₃ exposure shows a lower energy compared to the other sulfur species (i.e., ZCuHSO₃) derived from the SO₂ exposure. Importantly, SO₂ and SO₃ interaction with Cu dimers shows a further decrease in energy compared to the SO₂ and SO₃ adsorption with monomeric copper species (i.e., Z₂Cu and ZCuOH). The difference between the interaction of SO₂ and SO₃ with Cu dimers is that SO₂ adsorption causes both ZCuHSO₃ and ZCuHSO₄. The SO₃ adsorption, however, results in only two ZCuHSO₄. In this respect, two H₂ consumption peaks at 604 and 657 °C can be associated with (bi)sulfate interacting with copper monomers and dimers, respectively, in Figure 5a.

As mentioned above and described in reaction R.3, in the presence of water, ZCuOH is the active species for the surface nitrate formation in the form of copper nitrate (ZCuNO₃) by following reactions below (R.4–R.6), as described in the refs (89,100). In the feed of only NO₂, NO⁺ on the copper site results from NO₂ disproportionation with forming nitric acid (HNO₃) via R.4. In R.6, the resulting copper-nitrite oxidation by NO₂ leads to copper nitrate and NO. For the SO₃-exposed sample, most copper sites are occupied by (bi)sulfate in the form of copper-(bi)sulfate (ZCuHSO₄). ZCuHSO₄, therefore, deactivates the surface nitrate formation by blocking the copper site from the access of the NO⁺ or HNO₃, resulting in the lowest NO release during the adsorption phase in AN-TPD. In the same manner, decreased NO release was seen for the SO₂-exposed sample compared with the degreened sample. However, more NO release was observed due to less sulfur content compared to the SO₃-exposed sample.

R.4

R.5

R.6

In AN-TPD, stored surface nitrate species are typically desorbed in the form of NO₂ for the degreened sample, as can be seen in Figure 6c. For the SO₂- and SO₃-exposed samples, decreased NO₂ release was seen because of decreased nitrate storage resulting from sulfur species, especially the SO₃-exposed sample. With respect to N₂O, the SO₃-exposed sample released the lowest amount of N₂O (14.6 μmol·g^–1_washcoat) compared to the SO₂-exposed sample (31.9 μmol·g^–1_washcoat).

Surprisingly, a comparable amount of N₂O was released for the SO₂-exposed sample as degreened sample (31.3 μmol·g^–1_washcoat), although decreased nitrate storage was clearly observed via NO formation in Figure 6b. This suggests that sulfur may increase N₂O selectivity during NO₂ SCR. Sulfur enhances ammonia storage on the catalyst. We hypothesize that the increased availability of ammonia on the catalyst for the SO₂ poisoned sample could result in that more N₂O is formed compared to NO₂ release. According to first-principle calculations performed by Feng et al.,³⁶ the decomposition of the reaction intermediate (H₂NNO) over the Cu site leads to N₂O and H₂O. Their suggested specific copper-configuration is NH₃-solvated Cu dimers. In Figure S8, the observed SO₂+H₂SO₄ release and N₂O production (Figure 6d) show that for the SO₂ poisoned sample, there is sulfur release already from 200 °C. The main N₂O desorption occurs at 300 °C, suggesting that some of the copper sites were regenerated for the SO₂-exposed sample; thus, resulting H₂NNO intermediates can possibly lead to N₂O through the regenerated copper sites. This combination with the increased availability of ammonia could be one possible explanation. However, for the SO₃-exposed sample, there was significantly more sulfur stored on the sample as seen in ICP experiments. In addition, the sulfate species were very stable, which is seen by the fact that the SO₂+H₂SO₄ release occurred at significantly higher temperature and was not detected during N₂O formation. These results suggest that more copper sites were occupied by sulfur species when N₂O was formed and in addition significantly less nitrates were formed in the adsorption phase, which could explain the low N₂O formation for the SO₃-exposed sample.

5. Conclusions

NO₂+NH₃ titration to form AN and subsequent temperature-programmed desorption was applied for the degreened and SO₂- and SO₃-exposed Cu-CHA to investigate the effect of SO₂ and SO₃ on surface nitrate and N₂O formation. Sulfation was carried out at 400 °C to mimic the scenario of SO₂/SO₃ contamination of the SCR catalyst resulting from SO₂ oxidation to SO₃/H₂SO₄ over the DOC or coated filter. H₂-TPR and in situ DRIFTS were applied to investigate the reduction properties and surface species response in the presence of sulfur species derived by SO₂- and SO₃-exposure treatments. Critical effects of SO₂ and SO₃ on Cu-CHA, its surface nitrate speciation, and N₂O formation were found. The following major findings were established:

• 1.First-principles calculations demonstrate that SO₂ adsorption on monocopper sites results in (bi)sulfite and dicopper causes both (bi)sulfite and (bi)sulfate surface species. SO₃ adsorption on mono- and dicopper results in only (bi)sulfate. Sulfur species derived from the SO₃ exposure are lower in energy compared to sulfur species derived from the SO₂ exposure. In addition, SOx adsorption on the dicopper exhibits higher thermodynamic stability as compared to SOx adsorption on the monocopper in Cu-CHA, indicating that speciation of copper has a significant effect on the speciation and thermal stability of sulfur species over Cu-CHA.
• 2.SO₃ exposure causes a more detrimental contamination of Cu-CHA compared to SO₂ exposure, according to the ICP and H₂-TPR results, in line with an earlier reported SO₂/SO₃-exposure study⁹ as well as showing good agreement with our theoretical calculations. The SO₂ exposure mainly shows sulfur contamination with ZCuOH, but the SO₃ exposure results in the sulfation not only with ZCuOH but also with Z₂Cu. Therefore, the SO₃-exposed Cu-CHA shows a larger sulfur uptake compared to the SO₂-exposed Cu-CHA.
• 3.In situ DRIFTS demonstrated that similar surface species result from reactions between NO₂ and sulfur species derived from the SO₂ and SO₃ exposures. After 2 min of NO₂ exposure, significantly lower bands were observed in the nitrate region for the SO₂- and SO₃-exposed samples compared to the degreened sample. The results from DRIFT could explain why NO release was decreased for both SO₂- and SO₃-exposed samples during NH₃+NO₂ adsorption in AN-TPD. Especially, the SO₃-exposed sample showed the lowest NO formation, suggesting that surface nitrate formation was significantly hindered due to (bi)sulfate occupying the copper sites in the form of copper-(bi)sulfate (ZCuHSO₄).
• 4.AN-TPD showed similar N₂O release for the SO₂-exposed Cu-CHA compared to the degreened Cu-CHA, while a lower NO₂ release for SO₂-exposed Cu-CHA. These results suggest that SO₂ exposure can increase the N₂O selectivity compared to the degree end Cu-CHA during AN decomposition. In contrast, SO₃-exposed Cu-CHA shows significantly lower N₂O formation compared to the degreened and SO₂-exposed Cu-CHA, due to its highly stable sulfur species (i.e., Cu-(bi)sulfate). SO₃ exposure caused more stable sulfur species compared to SO₂ exposure. The stability and amount of the sulfur species are likely important factors that influence N₂O and N₂ selectivity since SO₂-exposed and SO₃-exposed samples shows totally different N₂O formation properties.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsengineeringau.4c00004.

• Hydrothermal synthesis of Cu-CHA and washcoating on the monolith substrate, SO₂ and SO₃ exposure to the Cu-CHA procedure with SO₃ calibration, elemental analysis results from ICP, N₂ physisorption and XRD, SEM images of zeolite crystal morphology, UV–vis DRS measurement for bulk sulfur-containing compounds, H₂S detection during H₂-TPR, DRIFTS, empty-tube test during the adsorption step, and SO₂/H₂SO₄/N₂ release during heating in AN-TPD (PDF)

Author Contributions

CRediT: Joonsoo Han conceptualization, data curation, formal analysis, investigation, methodology, writing-original draft; Joachim Dithmer Bjerregaard conceptualization, data curation, formal analysis, investigation, writing-original draft; Henrik Grönbeck conceptualization, funding acquisition, methodology, supervision, writing-review & editing; Derek Creaser conceptualization, methodology, supervision, writing-review & editing; Louise Olsson conceptualization, funding acquisition, methodology, supervision, writing-review & editing.

The authors declare no competing financial interest.