Dynamic Active Site Evolution in Lanthanum‐Based Catalysts Dictates Ethane Chlorination Pathways

Abstract

Radical‐mediated chlorination of ethane presents a low‐carbon alternative for polyvinyl chloride (PVC) synthesis, yet selectivity toward 1,2‐dichloroethane remains challenged by uncontrolled over‐chlorination. Lanthanum oxychloride (LaOCl) has emerged as a promising catalyst, but its structural dynamics under Cl₂‐rich conditions and the origin of selectivity loss remain elusive. Here, we integrate advanced spectroscopic techniques with theoretical calculations to address this knowledge gap. Our findings unveil a sequential LaOCl → LaCl₃ transformation that dictates product distribution shifting from 1,2‐dichloroethane to trichloroethane. Mechanistic insights reveal that surface hydroxyl groups, generated during catalyst chlorination, promote bidentate adsorption of 1,2‐dichloroethane via hydrogen‐bond networks, thereby activating C─Cl over‐chlorination. Additionally, by employing Al₂O₃‐supported LaCl₃ model catalysts, the size‐dependent chlorophilicity of the LaCl₃ species is demonstrated. The bonding of interfacial oxygen with monolayer‐dispersed LaCl₃ species generates empty 4f‐states above the Fermi level, creating strong Lewis acid sites that stabilize Cl radicals and selectively convert chloroethane to 1,2‐dichloroethane. In contrast, aggregated nanoparticles are inactive due to their inability to stabilize chlorine radical. Our findings establish important structure sensitivity in lanthanum‐catalyzed chlorination and provide guiding principles for catalyst design, highlighting the importance of stabilizing metastable LaOCl _x species and modulating surface hydroxyl chemistry to overcome selectivity limitations.

LaOCl‐catalyzed C₂H₆ chlorination offers a low‐carbon route for polyvinyl chloride synthesis, yet structural evolution occurs under Cl₂ conditions. We establish structural sensitivity in La‐catalyzed chlorination, and elucidate the size‐dependent chlorophilicity of LaCl₃ and the role of hydroxyl groups in overchlorination, highlighting stabilization of LaOCl and hydroxyl chemistry modulation as key design principles to break selectivity limitations.

Introduction

Polyvinyl chloride (PVC) is the third most manufactured polymer globally, with annual production exceeding 50 million tons.^{[ 1} , ^{2 ]} Such widespread application has driven significant industrial interest in developing more efficient PVC production methods.^{[ 3} , ⁴ , ⁵ , ^{6 ]} PVC can be readily synthesized from vinyl chloride monomer (VCM) through a radical‐based polymerization process.^{[ 7} , ^{8 ]} However, the current coal‐to‐acetylene and petroleum‐to‐ethylene routes for VCM synthesis face critical sustainability limitations.^{[ 9} , ¹⁰ , ¹¹ , ¹² , ¹³ , ^{14 ]} The coal‐to‐acetylene route requires energy‐demanding calcium carbide production at >2000 °C,^{[ 15 ]} resulting in 8–12 tons of CO₂ emissions per ton of VCM.^{[ 16 ]} Similarly, petroleum‐derived routes depend on high‐temperature steam‐cracking operations (>500 °C) for ethylene (C₂H₄) generation,^{[ 17 ]} emitting 3–4 tons of CO₂ emissions per VCM ton.^{[ 18} , ^{19 ]} These carbon‐intensive methodologies collectively represent a notable portion of global chemical industry emissions, highlighting the urgent need for developing alternative feedstocks for sustainable VCM synthesis.

Natural gas‐derived ethane (C₂H₆) is a promising candidate for VCM production due to its abundance and low carbon footprint.^{[ 20} , ²¹ , ²² , ^{23 ]} Over the past few decades, the C₂H₆ oxychlorination process, which involves co‐feeding HCl, O₂, and Cl₂ in a single reactor, has been extensively researched.^{[ 24} , ²⁵ , ^{26 ]} However, the C₂H₆ oxychlorination process requires high temperatures (450–550 °C) to activate the relatively unreactive C₂H₆.^{[ 27} , ^{28 ]} The limited selectivity for VCM, due to C₂H₄ formation, along with considerable carbon loss in the form of carbon oxides due to the use of O₂, poses significant challenges.^{[ 29 ]} Compared with the oxychlorination process, reacting C₂H₆ with Cl₂ is an excellent alternative strategy for VCM production.^{[ 30 ]} Previous work by Olsbye et al. indicated that thermal‐induced dissociation of Cl₂ produces chlorine radicals (Cl·), which activates C₂H₆ at mild temperatures (<300 °C).^{[ 31 ]} Notably, the radical‐driven chlorination of C₂H₆ yields almost exclusively ethyl chloride (C₂H₅Cl) even at high conversion of up to 50%.^{[ 32 ]} A catalyst can be designed to target the chlorination of C₂H₅Cl into 1,2‐dichloroethane (1,2‐C₂H₄Cl₂), which can then be thermally cracked into VCM. However, despite showing promise, chlorination of C₂H₆ directly to 1,2‐C₂H₄Cl₂ has yet to be implemented due to inherent challenges in achieving acceptable product selectivity. Gas‐phase Cl radical‐mediated hydrogen abstraction favors the formation of the 1‐chloroethyl radical (15 kJ mol⁻¹ lower in energy than the 2‐chloroethyl isomer), leading to the undesired production of 1,1‐C₂H₄Cl₂.^{[ 31 ]} The low activation barriers for sequential H‐abstraction and Cl· transfer further exacerbate competing pathways, including over‐chlorination and thermal dehydrochlorination.

Recent studies by Zichittella and Pérez‐Ramírez have demonstrated that these challenges can be addressed by activating the C₂H₅Cl intermediate on rare‐earth oxychloride catalysts.^{[ 30 ]} Among these catalysts, LaOCl exhibits exceptional promise due to its unique chlorophilic properties, which facilitate efficient Cl₂ dissociation. This enables the selective conversion of C₂H₅Cl to 1,2‐C₂H₄Cl₂, achieving 80% selectivity at 20% C₂H₆ conversion under a C₂H₆/Cl₂ ratio of 6:3 (v/v). Nevertheless, a critical 1,2‐C₂H₄Cl₂ selectivity limitation emerges at elevated C₂H₆ conversion. Notably, LaOCl is prone to progressive surface chlorination, leading to the formation of LaCl₃, under Cl₂‐rich conditions.^{[ 33} , ³⁴ , ^{35 ]} While the structural sensitivity of C₂H₆ chlorination to La sites can be mitigated by lowering Cl₂ feed concentrations, the impact of these dynamic active sites under industrially relevant conditions remains unresolved. Addressing the mechanistic gap between the electronic states of La sites and their chlorination behavior is imperative to reduce selectivity constraints inherent with La‐based catalysts.

In this work, we investigate the structural evolution of La₂O₃ catalysts under Cl₂‐rich conditions (C₂H₆:Cl₂ = 4:9, v/v), aiming to shed light on the complex C₂H₆ chlorination reaction network on oxychloride‐based materials. Our in‐depth characterization reveals a sequential La₂O₃ → LaOCl → LaCl₃ transformation, driven by progressive chlorination. This structural evolution directly governs product distribution, with chlorination selectivity shifting from C₂H₅Cl to 1,2‐C₂H₄Cl₂, and ultimately to trichloroethane (C₂H₃Cl₃). By combining Al₂O₃‐supported LaCl₃ model catalysts and theoretical calculations, we elucidate the size‐dependent chlorophilicity of LaCl₃ and the role of hydroxyl groups in promoting over‐chlorination. These findings establish chlorination‐driven phase dynamics as the central determinant of product selectivity, offering a blueprint for designing selective catalysts via precise control of metal speciation and surface microenvironment.

Results and Discussion

Structural Evolution of the La₂O₃ Under Cl₂‐Rich Conditions

The La₂O₃ catalyst was synthesized via an ammonia precipitation method followed by high‐temperature calcination at 800 °C. To investigate structural evolution under industrially relevant conditions (C₂H₆:Cl₂ = 4:9, v/v), we conducted time‐resolved characterization combining high‐resolution transmission electron microscopy (HRTEM) and energy‐dispersive spectroscopy (EDS). These complementary techniques provide critical insights into morphological transformations and chlorination dynamics during the catalytic process (Figure S1). Initial HRTEM characterization (Figure 1a–c) reveals that the pristine catalyst exhibited well‐defined lattice spacings of 0.30 and 0.23 nm, corresponding to the (011) and (012) planes of hexagonal La₂O₃, respectively.^{[ 36} , ^{37 ]} Upon exposure to the Cl₂‐rich conditions, time‐dependent structural degradation was observed (Figure 1d). When the reaction proceeded to 55 min, the emergence of 0.26 nm lattice fringes (Figure 1e) are observable, matching the (102) planes of tetragonal LaOCl.^{[ 38 ]} The intermediate chlorination state of LaOCl is further confirmed by Figure 1f which displays characteristic spacings of 0.29 and 0.35 nm, corresponding to Miller indices (hkl) of 110 and 101, respectively.^{[ 39 ]} The structural evolution progresses more extensively at extended reaction times (Figure 1g). After 770 min, HRTEM images (Figure 1h,i) reveal 0.65 nm lattice spacing corresponding to the (100) planes of hexagonal LaCl₃, demonstrating further chlorination of the LaOCl phase. However, residual 0.35 nm spacings corresponding to LaOCl (101) planes persist in localized regions (Figure 1i), revealing incomplete conversion of the intermediate phase. This behavior may be attributed to the higher activation energy required for complete oxygen substitution in the LaOCl lattice compared to initial La₂O₃ chlorination. Complementary EDS analysis (Figure S2) quantitatively tracked the chlorination progression, showing a monotonic increase in Cl/La atomic ratios from 0 (initial) to 1 ± 0.04 (770 min).

Figure 1.

Time‐resolved HAADF‐STEM analysis of La₂O₃ structural evolution under Cl₂‐rich conditions. a)–c) Pristine La₂O₃ catalyst. d)–f) Intermediate stage after 55 min reaction. g)–i) Final stage after 770 min reaction.

The phase evolution of La₂O₃ during chlorination was further investigated through analysis of synergistic X‐ray diffraction (XRD) and Raman spectroscopy. Figure 2a presents time‐resolved XRD patterns tracking crystalline phase evolution. Initially, distinct diffraction peaks corresponding to hexagonal La₂O₃ dominate the profile.^{[ 34 ]} At 55 min, reflections assignable to tetragonal LaOCl emerge,^{[ 30 ]} indicating partial oxychlorination of La₂O₃, in agreement with HRTEM observations of LaOCl lattice formation. After 360 min, the XRD pattern transitions to a LaCl₃·3H₂O dominated profile, demonstrating progressive chlorine substitution at oxygen sites. Notably, the disappearance of LaOCl signatures at this stage suggests thermodynamic instability of the intermediate phase under sustained Cl₂ exposure, driving structural degradation and conversion to LaCl₃.

Figure 2.

Structural changes of the La₂O₃ catalyst under reaction gas atmosphere at 260 °C. Time‐resolved XRD patterns a) and Raman spectra b). The numbers in the figures represent the reaction time of the catalyst in minutes. c) Time‐resolved XPS Cl 2p and La 4p_2/3 spectra. d) Changes in surface atomic contents. e) Conversion and selectivity of C₂H₆ chlorination over the La₂O₃ catalyst. Reaction conditions: C₂H₆/Cl₂/N₂ = 4:9:87, 260 °C, WHSV = 2000 mL h⁻¹ g⁻¹. f) Gibbs free energy change during the surface chlorination of La₂O₃.

Raman spectroscopic analysis (Figure 2b) provides insight into atomic‐scale structural dynamics. The pristine La₂O₃ spectrum exhibits characteristic vibrational modes at 104.5 cm⁻¹ (A_1g symmetry, La─O symmetric stretching), 189.7 cm⁻¹ (E _g symmetry, O─La─O bending), and 405.8 cm⁻¹ (multiphonon coupling), indicative of its hexagonal lattice.^{[ 40 ]} After 55 min of reaction, new bands emerge at 184 cm⁻¹ (Cl─La─Cl bending) and 207 cm⁻¹ (La─Cl stretching), accompanied by oxygen‐associated vibrations at 330 cm⁻¹ (La─O─La bridging) and 432 cm⁻¹ (terminal La─O stretching), confirming LaOCl formation.^{[ 41 ]} The spectral evolution is consistent with the XRD‐detected phase progression, validating La₂O₃→LaOCl transformation. The spectrum collected after 770 min of reaction shows two dominant bands at 184 and 206 cm⁻¹, indicative of the formation of LaCl₃ species.^{[ 33 ]} Residual weak signals at 330/432 cm⁻¹ indicate incomplete LaOCl conversion, highlighting kinetic limitations in lattice oxygen substitution.

To elucidate the evolution of surface electronic states during the chlorination reaction, X‐ray photoelectron spectroscopy (XPS) analysis was conducted. The Cl 2p region shown in Figure 2c can be deconvoluted into two peaks at ca. 198.8 and 200.4 eV, which are attributed to spin‐orbit splitting at the 2p_3/2 and 2p_1/2, respectively. Since the binding energy of the La 4p_3/2 electrons overlaps with that of the Cl 2p regions, two additional peaks corresponding to the La 4p_3/2 (195.1 eV) and a satellite peak (196.5 eV) were also fitted.^{[ 42 ]} The peak of Cl shifts to a lower binding energy, while the peak of La 4p_3/2 shifts to a higher binding energy after the reaction. Meanwhile, the peak of La 3d also shifted to higher binding energy (Figure S3). This implies that La atoms donate electrons to Cl atoms. Quantitative analysis of elemental surface composition (Figure 2d) demonstrates a monotonic increase in Cl/La atomic ratio (derived from peak area integration) accompanied by oxygen depletion, confirming progressive replacement of surface oxygen with chlorine adatoms.

Density functional theory (DFT) calculations were performed to map the thermodynamic landscape of surface chlorination (Figure 2f). The Gibbs free energy profile reveals a stepwise chlorination mechanism. Initial exothermic adsorption of HCl on La₂O₃ (ΔG = −2.08 eV per HCl molecule) precedes saturation‐induced phase transformation to LaOCl, followed by metastable LaOCl decomposition into LaCl₃ under continuous HCl exposure. Notably, the calculated energy barrier for LaOCl → LaCl₃ conversion (ΔG = −1.82 eV per HCl molecule) corroborates its observed instability in HCl‐rich environments. This synergy between theory and experiment conclusively establishes HCl‐mediated phase evolution as a thermodynamically favored process, governed by sequential surface chlorination.

The catalytic activity of the as‐synthesized La₂O₃ was measured through the performance of long‐term experiments under C₂H₆ chlorination conditions. As illustrated in Figure 2e, the activity of the La₂O₃ catalyst displays a dynamic trend throughout the reaction. In the initial reaction stage, the selectivity for 1,2‐C₂H₄Cl₂ is relatively low, which can be ascribed to the relatively inert surface of La₂O₃ that had undergone high‐temperature calcination. With the progression of the reaction, the selectivity for 1,2‐C₂H₄Cl₂ exhibits a continuous increase, as La₂O₃ is chlorinated to form LaOCl; which is evidenced by the above time‐resolved spectroscopic and structural characterization. It is well established that LaOCl serves as the active phase for the catalytic chlorination of C₂H₆ to 1,2‐C₂H₄Cl₂.^{[ 30 ]} Nevertheless, as the reaction proceeded further, the structural degradation of LaOCl triggers the over‐chlorination of 1,2‐C₂H₄Cl₂, consequently causing the main product to gradually shift from 1,2‐C₂H₄Cl₂ to C₂H₃Cl₃. Notably, the occurrence of the side reaction of over‐chlorination consumes the Cl₂ that is fed at stoichiometric ratio, resulting in insufficient Cl₂ to drive the conversion of C₂H₆. Therefore, an inverse correlation between over‐chlorination selectivity and C₂H₆ conversion is observed. This shift in product selectivity may be attributed to the structural evolution to LaCl₃ species and the presence of hydroxyl species on the catalyst surface, both of which can dynamically form during the reaction and influence the reaction pathway. We believe that such structural degradation of LaOCl under Cl₂/HCl‐rich conditions is a common phenomenon for other rare‐earth oxide catalysts, highlighting the need to identify the primary driving forces behind the over‐chlorination of 1,2‐C₂H₄Cl₂.

The Role of LaCl₃ Species on Ethane Chlorination

As LaCl₃ is the primary phase present after reaction, we first investigated the mechanistic role of this species in C₂H₆ chlorination. To systematically explore this, we employed a model catalyst approach by synthesizing a series of LaCl₃/θ‐Al₂O₃ catalysts with precisely controlled LaCl₃ loadings (denoted as x% LaCl₃/Al₂O₃, where x indicates La wt.% loading quantified by ICP‐OES, Table S1). θ‐Al₂O₃, which is chemically inert under reaction conditions (as evidenced by control experiments showing comparable activity between θ‐Al₂O₃ and the empty reactor, Figure 3a), serves as a stable support for LaCl₃. XRD characterization (Figure S4) confirmed the evolution of LaCl₃·3H₂O crystallinity with increasing loadings. At La loadings below 7.4 wt.%, no discernible diffraction peaks associated with LaCl₃ phases are observed, indicating atomic‐level dispersion of La species on the θ‐Al₂O₃ surface. This conclusion was further corroborated by atomic‐resolution scanning transmission electron microscopy (STEM), which confirmed the absence of La‐containing nanoparticles and demonstrated uniform atomic dispersion of La across the support (Figures S5 and S6). However, as the LaCl₃ loading increases further, diffraction peaks for LaCl₃·3H₂O emerges at 15° in the XRD patterns, indicating that LaCl₃ species start to aggregate and form nanoparticles.

Figure 3.

a) Activity tests over empty tube and the LaCl₃/Al₂O₃ catalysts. Reaction conditions: C₂H₆/Cl₂/N₂ = 4:9:87, 260 °C, WHSV = 2000 mL h⁻¹ g⁻¹. b) Conversion as a function of temperature in the chlorination of C₂H₅Cl over the LaCl₃/Al₂O₃ catalysts. Reaction conditions: C₂H₅Cl/Cl₂/N₂ = 3.5:5:91.5, 150–260 °C, WHSV = 5500 mL h⁻¹ g⁻¹. Reaction orders of C₂H₅Cl c) and Cl₂ d). e) Apparent activation energy of C₂H₅Cl chlorination. f) Reaction order of Cl₂ as a function of the apparent activation energy in C₂H₅Cl chlorination over the LaCl₃/Al₂O₃ catalysts. Reaction conditions: C₂H₅Cl:Cl₂ = 1.5–4.5:2–5, 200–230 °C, WHSV = 6000–8000 mL h⁻¹ g⁻¹. g) Adsorption energies of surface species (C₂H₅Cl, Cl₂, and Cl·) on the sites of LaCl₃/Al₂O₃ and LaCl₃(100). PDOS analysis of Cl· adsorbed on the surface of LaCl₃/Al₂O₃ h) and LaCl₃(100) i).

The structure–activity relationship for the range of model catalysts was elucidated subsequently. Figure 3a shows the conversion of C₂H₆ and the product distribution on the LaCl₃/Al₂O₃ catalysts. By analyzing the activity of the catalysts as well as of θ‐Al₂O₃, it is noted that over all the catalysts, the C₂H₆ conversion is across a rather narrow range from 60% to 66%, indicating virtually no activity difference between the catalysts and θ‐Al₂O₃. This is in line with the studies of Olsbye et al., which have shown that C₂H₆ chlorination can be feasibly initiated by Cl· via a gas‐phase pathway, which drives virtually exclusive formation of C₂H₅Cl (>83.7%) up to 66.6% C₂H₆ conversion at 260 °C (Figure S7). Notably, increasing the reaction temperature did not favor the transformation of C₂H₅Cl into 1,2‐C₂H₄Cl₂, but instead the undesired 1,1‐C₂H₄Cl₂. Indeed, the gas‐phase Cl· driven chlorination of C₂H₅Cl is a major side reaction because chlorination occurs preferentially at the chlorine‐functionalized carbon, hindering the generation of the desired isomer. Interestingly, the incorporation of LaCl₃ into θ‐Al₂O₃ significantly enhances the formation of 1,2‐C₂H₄Cl₂ and exhibits remarkable structural sensitivity. The 1,2‐C₂H₄Cl₂ selectivity shows a marked increase from 32% to 62% with La loading increasing from 1.6% to 7.4%, followed by subsequent inhibition at higher loadings. Mechanistic investigations via the temperature‐programmed surface reaction (TPSR) demonstrate the chemical inertness of lattice chlorine in LaCl₃ (Figure S8), as evidenced by the absence of 1,2‐C₂H₄Cl₂ formation, which hinders its participation as a Cl donor in the chlorination of C₂H₅Cl.

To rationalize the differences observed in catalytic performance, extensive kinetic tests were performed for the reaction of C₂H₅Cl with Cl₂ (Table S2). The light‐off curves of C₂H₅Cl chlorination reveal the limitations of θ‐Al₂O₃ in catalyzing the chlorination of C₂H₅Cl, requiring higher temperatures to initiate the reaction (Figures 3b and S9). However, increasing the temperature mainly induced the formation of 1,1‐C₂H₄Cl₂, indicating that the chlorination of C₂H₅Cl on the θ‐Al₂O₃ surface is still controlled by the gas‐phase radical pathway (Figure S9a). Interestingly, the loading of LaCl₃ significantly shifts the light‐off curves to lower temperatures, the extent of which is related to the LaCl₃ loading, indicating a catalytic effect on C₂H₅Cl chlorination. More importantly, the chlorination of C₂H₅Cl over all the LaCl₃/Al₂O₃ catalysts predominantly resulted in the formation of 1,2‐C₂H₄Cl₂ (60–100% selectivity). These observations indicate a cascade mechanism for C₂H₆ chlorination, where the transformation of C₂H₆ into C₂H₅Cl is likely dominated by gas‐phase radical pathways, while the role of the catalyst is to activate Cl₂ and intermediate C₂H₅Cl to facilitate the formation of 1,2‐C₂H₄Cl₂.

Temperature‐programmed desorption (TPD) experiments show negligible adsorption for C₂H₅Cl for all catalysts (Figure S10), indicating that C₂H₅Cl only weakly adsorbs onto the catalyst surface and that the presence of LaCl₃ does not significantly enhance this adsorption. This suggests that the mechanism by which LaCl₃ enhances 1,2‐C₂H₄Cl₂ selectivity involves alternative pathways that do not significantly rely on the adsorption and activation of C₂H₅Cl on the catalyst surface. This inference is consistent with kinetic findings, which reveals near‐unity reaction orders for C₂H₅Cl independent of LaCl₃ loadings (Figure 3c). In contrast, the reaction order of Cl₂ exhibits a tendency to decrease first and then increase as the loading of LaCl₃ increases (Figure 3d). Notably, the reaction order of Cl₂ decreases considerably over the 7.4% LaCl₃/Al₂O₃ catalyst, with the reaction order being only 0.3. Interestingly, a negative correlation is found between the reaction order of Cl₂ and the 1,2‐C₂H₄Cl₂ selectivity (Figure S11). These findings clearly demonstrate that the rate‐determining step involving the Cl radical is significantly structure‐sensitive, with highly dispersed LaCl₃ playing a promotional role, and this effect is quenched when LaCl₃ agglomerates. This can be rationalized by the relationship in Figure 3f, which evidences a linear correlation between the apparent activation energy and the reaction order of Cl₂. Arrhenius analysis reveals a loading‐dependent activation energy of C₂H₅Cl chlorination in the range of 64.0–117.4 kJ mol⁻¹ (Figure 3e). The activation energy is significantly lower for the highly dispersed LaCl₃ species, while the aggregated LaCl₃ particles formed at higher loadings result in a higher activation energy.

To gain atomic‐level insights into the structural sensitivity of C₂H₅Cl chlorination at LaCl₃ sites, DFT calculations were performed using two model systems: (i) a (LaCl₃)₂ dimer supported on θ‐Al₂O₃ (110) to mimic highly dispersed LaCl₃ (denoted as LaCl₃/Al₂O₃), and (ii) chlorine‐terminated LaCl₃(100) to represent aggregated configurations. Notably, the interaction between La atoms in the (LaCl₃)₂ dimer and surface oxygen atoms significantly weakens the La─Cl bond. Upon structural optimization, one La─Cl bond within the dimer undergoes cleavage, resulting in a free Cl ion that is captured by an adjacent Al³⁺ site. The resulting dimer of La ions is subsequently anchored on the Al₂O₃ surface by hybrid O/Cl coordination. This is consistent with the study by Chen et al., which has shown that CuCl₂ undergoes dissociative adsorption on an γ‐Al₂O₃ (110) surface, where only one chloride ion binds to copper, and the other binds to the Al₂O₃ surface.^{[ 43 ]}

Typically, interactions between rare‐earth metal sites and molecules primarily involve electron orbitals below the Fermi level, as these electrons dominate bonding and antibonding interactions at the metal surface.^{[ 44} , ^{45 ]} Projected density of states (PDOS) analysis demonstrates negligible perturbation to La 4f and 5d orbitals below the Fermi level in the (LaCl₃)₂ dimer (Figure S12). This finding explains the structural insensitivity observed in the adsorption of C₂H₅Cl and Cl₂ at La sites, with both exhibiting weak adsorption energies (Figure 3g). These findings align with kinetic data showing C₂H₅Cl reaction orders (0.94–1.1) independent of LaCl₃ dispersion. Furthermore, it is important to note that the activation of Cl₂ is not a critical factor in C₂H₆ chlorination, as the thermal scission of Cl₂ into two Cl radicals occurs readily even at 200 °C (Figure S7).

Interestingly, the bonding of the (LaCl₃)₂ dimer with interface oxygen significantly enhances the spin degenerate empty states above the Fermi level, which involve the atomic orbitals of La 4f and O 2p (Figure S12). These vacant orbitals can act as Lewis acid sites, facilitating electron donation from Cl· via frontier orbital interactions.^{[ 46 ]} The enhanced electron‐accepting capacity of La centers strengthens their interaction with Cl· through charge transfer. Notably, Cl· retains a significant amount of electron density at the Fermi level when interacting with the (LaCl₃)₂ dimer (Figure 3h). These highly active electrons potentially lower the activation energy for subsequent chlorination reactions. In contrast, on the LaCl₃(100) surface, Cl· exhibits weak interactions, resulting in Cl 2p electrons occupying states above the Fermi level (Figure 3i). This conclusion is further confirmed by the calculated adsorption energies (ΔE; the more negative the value, the stronger the binding), where the value for Cl· adsorption on the (LaCl₃)₂ dimer (−0.98 eV) is much less than that on LaCl₃(100) (1.26 eV) (Figure 3g). The repulsion of the Cl· on the LaCl₃(100) surface hinders further chlorination of C₂H₅Cl, making the product of C₂H₆ chlorination primarily C₂H₅Cl. The observed performance degradation at aggregated LaCl₃ highlights the necessity of controlling the degree of chlorination of the catalyst to prevent crystalline phase segregation.

To elucidate the atomic‐level mechanism by which phase transformation governs over‐chlorination, we simulated the adsorption of 1,2‐C₂H₄Cl₂ on the LaOCl (100) surface (Figure 4a). To model realistic reaction conditions, unsaturated La sites on the LaOCl surface were saturated with Cl radicals, as thermal dissociation of Cl₂ at 260 °C readily generates gas‐phase Cl radicals (Figure S13). The calculations reveal weak adsorption of 1,2‐C₂H₄Cl₂ on the Cl‐saturated LaOCl surface, with an adsorption energy of −0.65 eV. Crucially, the energy barrier for Cl radical‐mediated H abstraction from adsorbed 1,2‐C₂H₄Cl₂ to form the 1,2‐C₂H₃Cl₂⋅ radical (1.06 eV) remains significantly higher than the adsorption energy. This thermodynamic preference for desorption over further chlorination aligns with experimental observations of high 1,2‐C₂H₄Cl₂ selectivity on LaOCl. To probe the role of LaCl₃ in over‐chlorination, we subsequently examined 1,2‐C₂H₄Cl₂ adsorption on the LaCl₃(100) surface (Figure 4a). Similarly, weak linear adsorption (−0.5 eV) was observed, with the adsorption energy being significantly lower than the energy barrier (0.97 eV) for subsequent chlorination. This confirms that LaCl₃ formation alone does not drive over‐chlorination.

Figure 4.

Determining the role of hydroxyl groups in C₂H₆ chlorination. a) Energy profile for 1,2‐C₂H₄Cl₂ chlorination on the surfaces of LaCl₃, LaOCl, and LaOCl‐OH. The numbers above the models indicate the adsorption energies of different adsorption states. b) O 1s XPS spectra of La₂O₃ recorded at different reaction times (unit: minute). In situ DRIFTS of 1,2‐C₂H₄Cl₂ adsorption on La catalysts sampled at 55 min c) and 770 min d). e) Activity profile of the La₂O₃ catalyst at 260 °C. The catalyst underwent periodic Ar treatment at 350 °C for 5 h after 15 h of reaction, followed by re‐exposure to reaction conditions (C₂H₆/Cl₂/N₂ = 4:9:87, 260 °C, WHSV = 2000 mL h⁻¹ g⁻¹).

Hydroxyl‐Mediated Over‐Chlorination

As illustrated by the DFT calculations in Figure 2f, the exothermic adsorption of HCl on LaOCl surfaces drives a thermodynamically favorable chlorination process, ultimately saturating La sites with Cl atoms and hydroxylating surface oxygen species. Time‐resolved O 1s XPS spectra provide direct evidence of this dynamic surface hydroxylation (Figure 4b). After treatment at 800 °C, hydroxyl groups on La₂O₃ surfaces are completely removed, leaving only lattice oxygen and a minor concentration of oxygen vacancies, which appear at 531.2 and 528.6 eV, respectively.^{[ 47 ]} However, under Cl₂/HCl atmospheres, the La₂O₃ surface is rapidly etched, leading to a significant increase in the number of oxygen vacancy species, accompanied by the generation of hydroxyl species, identified by the peak at 532.7 eV.^{[ 48 ]} As the reaction progresses, water produced during chlorination progressively quenches oxygen vacancies, while a substantial amount of hydroxyl groups is generated. Notably, due to the kinetic limitations in the complete chlorination of LaOCl to LaCl₃, a minor concentration of oxygen vacancies remains on the catalyst surface after 770 min of reaction. Fourier transform infrared (FTIR) spectroscopy experiments further indicate the hydroxyl groups accumulation, with a distinct peak corresponding to hydroxyl groups observed at 3385 cm⁻¹ after the reaction (Figure S14).^{[ 49 ]}

Subsequently, we focus on the role of surface hydroxyls in the adsorption of 1,2‐C₂H₄Cl₂. To obtain direct evidence of hydroxyl group involvement in C─Cl bond activation, we performed in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) studies to investigate the adsorption of 1,2‐C₂H₄Cl₂ on La catalysts sampled at 55 min (LaOCl‐dominated phase) and 770 min (denoted as LaOCl‐OH) under reaction conditions. Dynamic analysis of surface‐adsorbed species demonstrates well‐defined chemisorption states of 1,2‐C₂H₄Cl₂ on both catalysts at room temperature. As illustrated in Figure 4c, the infrared spectra display distinct vibrational bands at 2971 cm⁻¹ (ν _as(CH₂) asymmetric stretching), 2890 cm⁻¹ (ν _s(CH₂) symmetric stretching), 1455 cm⁻¹ (δ(CH₂) scissoring), 1289 cm⁻¹ (ω(CCH) wagging), 1232 cm⁻¹ (ρ(HCCl) rocking), and 725 cm⁻¹ (ν(C─Cl) stretching).^{[ 50} , ^{51 ]} Notably, these adsorbed species exhibit systematic blueshifts compared to the intrinsic vibrational frequencies of gaseous 1,2‐C₂H₄Cl₂. Furthermore, weak splitting of the δ(CH₂) mode at 1453 cm⁻¹ and the ω(CCH) mode at 1291 cm⁻¹ are observed in the adsorbed state, indicative of symmetry disruption and electron density redistribution during chemisorption.

Intriguingly, on the LaOCl‐OH surface (Figure 4d), the δ(CH₂) and ω(CCH) peaks exhibit enhanced splitting, accompanied by the emergence of a new C─H bending vibration at 1394 cm⁻¹. These spectral changes suggest a transition from linear to bridged adsorption geometry for 1,2‐C₂H₄Cl₂, further perturbing molecular symmetry and amplifying vibrational coupling of C─H modes.^{[ 52 ]} This conclusion is strongly supported by a broadened absorption band at 3224 cm⁻¹, assigned to O─H⋯Cl hydrogen bonding interactions between surface μ‐OH groups and adsorbed molecules.^{[ 51 ]} These findings provide unequivocal evidence for hydroxyl‐mediated adsorption of 1,2‐C₂H₄Cl₂ on LaOCl surfaces, as corroborated by PDOS analysis, which reveals hydrogen‐bonding interactions between surface hydroxyls and the Cl atoms of 1,2‐C₂H₄Cl₂ (Figure S15). We thus constructed a hydroxyl‐covered LaOCl model (LaOCl‐OH). Intriguingly, 1,2‐C₂H₄Cl₂ adopts a bidentate adsorption mode via hydrogen bonding with surface hydroxyls, dramatically increasing the adsorption energy to −1.35 eV. The enhanced adsorption enables 1,2‐C₂H₄Cl₂ to overcome the chlorination barrier (0.96 eV), identifying hydroxyl‐induced adsorption as the primary driving force of over‐chlorination.

To further validate the hydrogen bonding as the driving force behind the over‐chlorination of 1,2‐C₂H₄Cl₂, we subjected the spent catalyst to treatment with Ar for 5 h (Figure 4e). FTIR experiments indicate that treatment at 350 °C effectively reduces the number of hydroxyl groups on the catalyst surface (Figure S14). Interestingly, after Ar treatment, the selectivity of the catalyst for 1,2‐C₂H₄Cl₂ recovers from 18.3% to 61.7% (Figure 4e), while the selectivity for C₂H₃Cl₃ is suppressed from 73.2% to 35.8%. Notably, as the reaction continues, surface hydroxyls reaccumulate, leading to a shift in chlorination products from 1,2‐C₂H₄Cl₂ to C₂H₃Cl₃. It needs to be emphasized that repeated Ar treatments only induce the dynamic reversible changes of surface hydroxyl sites, with almost no impact on La sites. The above results fully demonstrate that the hydroxyl groups on the catalyst surface induce the over‐adsorption of 1,2‐C₂H₄Cl₂, which is the intrinsic driving force for over‐chlorination.

Conclusions

In this study, we provide a comprehensive understanding of the chlorination of C₂H₆ over lanthanum‐based catalysts. By employing advanced spectroscopic techniques and DFT calculations, we reveal the structural evolution from La₂O₃ to LaOCl and subsequently to LaCl₃ under HCl/Cl₂‐rich conditions. This transformation significantly influences the product distribution, shifting selectivity from C₂H₅Cl to 1,2‐C₂H₄Cl₂, and ultimately to C₂H₃Cl₃. Our research elucidates the pivotal roles played by LaCl₃ species and surface hydroxyl groups in the chlorination process. Monolayer‐dispersed LaCl₃ clusters, stabilized by La─O interactions with the support, effectively stabilize Cl·, thereby enhancing the selectivity toward 1,2‐C₂H₄Cl₂. In contrast, the aggregated LaCl₃ nanoparticles repulse Cl· access. Furthermore, we have identified that hydroxyl groups formed during surface chlorination crucially drive the over‐chlorination of 1,2‐C₂H₄Cl₂ through promoting hydrogen‐bond‐assisted bidentate adsorption. The findings of this study are significant as they confirm the structure sensitivity of lanthanum‐catalyzed chlorination and highlight the importance of maintaining metastable LaOCl species and suppressing hydroxyl accumulation to enhance selectivity. Importantly, the unique role of highly dispersed LaCl₃ in stabilizing Cl· highlights the potential application of lanthanum‐based single‐atom catalysts in C₂H₆ chlorination. These insights pave the way for the development of more efficient and selective catalysts for C₂H₆ chlorination, potentially leading to breakthroughs in the sustainable production of PVC and other chlorinated hydrocarbons from new renewable resources.

Experimental Section

For the experimental details, see Supporting Information.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Li Y., Qi H., Zhu Z., Wu X., Dummer N. F., Taylor S. H., Ma L., Yang X., Liu Q., Hutchings G. J., Huang Y., Angew. Chem. Int. Ed.. 2025, 64, e202505846. 10.1002/anie.202505846,

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.