Boosting Hydrogen Production from Hydrogen Iodide Decomposition over Activated Carbon by Targeted Removal of Oxygen Functional Groups: Evidence from Experiments and DFT Calculations

Abstract

In the thermochemical sulfur–iodine water splitting cycle for hydrogen production, the hydrogen iodide (HI) decomposition reaction serves as the rate-determining step, and its high efficiency relies on the precise design of active sites on the catalyst. This paper combines experimental characterization with density functional theory (DFT) calculations, focusing on activated carbon catalysts. By regulating the types and contents of oxygen-containing functional groups through H₂ reduction treatment at different temperatures, the influence of oxygen-containing functional groups on HI decomposition was investigated. The results show that H₂ reduction treatment can gradually remove oxygen-containing functional groups such as carboxyl, hydroxyl, and carbonyl groups on the surface of activated carbon without significantly affecting the pore structure. Catalytic activity tests conducted under the typical reaction temperature of 500 °C confirmed that as the content of oxygen-containing functional groups decreases, the HI decomposition efficiency increases. DFT calculations further revealed the role of oxygen-containing functional groups: they inhibit the chemisorption of reactant HI on unsaturated carbon atoms and alter the desorption activation energy of product H₂, thereby affecting the overall reaction process. This study provides important theoretical guidance and experimental basis for designing efficient HI decomposition catalysts.

1. Introduction

With the proposal of carbon peak and carbon neutrality goals, the value of new energy in the global energy low-carbon transition and green development has gradually become prominent [1]. Hydrogen energy, characterized by high efficiency, cleanliness, and sustainability, is regarded as a clean energy carrier that can replace fossil fuels to address energy crises and alleviate increasingly severe environmental problems [2,3]. Among numerous hydrogen production methods, the thermochemical sulfur–iodine water splitting cycle has attracted significant attention due to its advantages of low cost, environmental friendliness, and large-scale applicability [4,5,6]. It decomposes water into multiple elementary reaction steps (Bunsen reaction, sulfuric acid decomposition reaction, and hydrogen iodide (HI) decomposition reaction) by introducing reaction intermediates [7]. In the actual process of cyclic hydrogen production, the HI decomposition reaction, as the rate-determining step, exhibits low efficiency. To improve decomposition efficiency, various supported metal catalysts have been developed, with active components including single metals such as Pt and Ir, as well as bimetallic Pt-Ir catalysts, which exhibit excellent catalytic activity and stability [8,9,10,11]. For instance, the HI decomposition rate of the Pt-Ir/C catalyst at 500 °C is 20% [8]. For support materials, carbon-based materials are the most commonly used, particularly activated carbon, which can also serve as a catalyst with activity comparable to that of supported metal catalysts. Therefore, carbon-based materials, due to their low cost, wide availability, and tunable physicochemical properties, are an important choice for high-performance HI decomposition catalysts. Studies have shown that the HI catalytic activity of activated carbon and carbon molecular sieves is higher than that of other carbon-based materials such as carbon nanotubes, graphite, and carbon black [12]. At 500 °C, the decomposition efficiency of hydrogen iodide can reach 23%, approaching its equilibrium decomposition rate.

Due to the multi-scale structural characteristics of carbon-based materials, their catalytic performance is determined by various physicochemical structures. Wang and Petkovic, et al. compared the catalytic performance of activated carbons with different raw materials, activation methods, and modification treatments [13,14,15]. They found that unsaturated carbon atoms are the active sites for HI decomposition. Meanwhile, a relatively high specific surface area and pore volume are conducive to HI decomposition. Rong et al. further prepared activated carbons with different specific surface areas and pore structures, confirming that hierarchical porous activated carbons with larger specific surface areas exhibit higher HI decomposition efficiency [16]. However, they failed to rule out the influence of simultaneous changes in the surface chemical structure. In addition to the pore structure, the regulation of chemical structures such as nitrogen-containing and oxygen-containing functional groups also profoundly affects the HI decomposition efficiency. The preliminary research results of our research group show that nitrogen doping can reshape the local electron density and charge distribution on the carbon surface, and the directional introduction of pyridine nitrogen proves that it can significantly reduce the activation energy of HI decomposition [17]. For oxygen-containing functional groups, Lin et al. performed acid treatment on coal-based activated carbons and found that fewer surface oxidation groups are beneficial to the catalytic activity of activated carbons [15]. Nevertheless, they overlooked an in-depth analysis of the mechanistic differences among various types of oxygen-containing functional groups. The limitations of the above studies are due to the strong interdependence of multi-scale structures in activated carbon often leads to simultaneous changes in both pore structure and chemical properties, making it difficult to accurately identify the decisive factors affecting catalytic activity. Such coupling effects of multiple variables significantly hinder the precise elucidation of the true active origins. Therefore, it is imperative to develop precise structural control strategies to independently investigate the role of specific oxygen-containing functional groups, thereby revealing the detailed mechanisms of different oxygen-containing functional groups in the HI decomposition process.

The oxygen-containing functional groups on the surface of carbon-based materials include hydroxyl, carboxyl, lactone, phenol, carbonyl, quinone, and ether bonds, which can be classified into acidic, basic, and neutral types according to their chemical properties [18,19]. Oxygen-containing functional groups are derived from the oxidation of the carbon surface by gas-phase oxidants such as oxygen, air, and carbon dioxide, as well as liquid-phase oxidants such as nitric acid and sulfuric acid. Generally, gas-phase oxidants can increase the content of hydroxyl and carbonyl groups, while liquid-phase oxidants can increase the content of carboxyl groups. Meanwhile, compared with gas-phase activation, liquid-phase activation can achieve the introduction of more oxygen-containing functional groups at low temperatures. Due to the differences in thermal stability of different oxygen-containing functional groups, oxygen-containing functional groups can be selectively removed by heat treatment in an inert atmosphere [20,21]. It has also been found that high-temperature hydrogen treatment can reduce the content of oxygen-containing functional groups on the surface of carbon materials while maintaining the pore structure of carbon-based materials. Therefore, H₂ reduction treatment at different temperatures can achieve gradient regulation of oxygen-containing functional groups.

Therefore, this paper selects three commercial activated carbons for physicochemical structure characterization and HI catalytic activity testing. Then, a suitable one is chosen for H₂ reduction treatment at different temperatures to obtain a model catalyst system with gradient changes in the types and contents of oxygen-containing functional groups. Through corresponding structural characterization and HI catalytic activity testing, the influence of oxygen-containing functional groups on HI decomposition is clarified, and the influence mechanism is revealed by combining quantum chemical calculations.

2. Experimental and Computational Details

2.1. Materials Preparation

Three types of activated carbon (Lu Zhiyuan, Ka Ergang, and Zhu Xi) were ground and sieved to 100 mesh, serving as typical HI catalyst samples, denoted as CAC-1, CAC-2, and CAC-3, respectively. Furthermore, CAC-2 was selected as the original sample for H₂ reduction modification at 500 °C and 800 °C. The specific process was as follows: 5 g of CAC-2 sample was placed in a tubular furnace, heated to the specified temperature at a heating rate of 10 °C min⁻¹ under a N₂ atmosphere (250 mL min⁻¹), then switched to a 5% H₂/Ar (150 mL min⁻¹) mixed atmosphere and maintained for 1 h, and finally naturally cooled to room temperature under a N₂ atmosphere (250 mL min⁻¹). The activated carbon samples modified at different temperatures were denoted as CAC-2-500 and CAC-2-800, respectively.

2.2. Material Characterizations

The relative contents of carbon, oxygen, hydrogen, nitrogen, and sulfur in the activated carbon samples were determined using an elemental analyzer (vario MACRO cube, Elementar Analysensysteme GmbH, Hanau, Germany). N₂ adsorption experiments were conducted at 77 K using an automatic surface analyzer (Quantachrome Autosorb 1C, Quantachrome Instruments, Boynton Beach, FL, USA) to determine the pore structure characteristics of the activated carbon samples. The samples were degassed under vacuum at 523 K for 6 h before the N₂ adsorption–desorption tests were performed at 77 K. The specific surface area was calculated using the Brunauer–Emmett–Teller (BET) method; the total pore volume was calculated from the adsorption amount at a relative pressure (P/P₀) of 0.99; the micropore surface area and volume were calculated using the t-plot method; and the pore size distribution curve was determined using the non-local density functional theory (NLDFT) model. The chemical structures of nitrogen and oxygen atoms in the samples were analyzed using X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250XI, Thermo Fisher Scientific, Waltham, MA, USA). The O1s spectra were deconvoluted using XPS Peak 4.1 software. The peaks were fitted according to the binding energies corresponding to different chemical structures of oxygen, and the proportion of each chemical structure was determined based on the integrated peak areas. During the deconvolution process, it is necessary to ensure that the Σx² (representing the residual sum of squares, and serving as a key parameter for evaluating the fitting quality) values are maintained below 10, with the fitted curves exhibiting excellent agreement with the original experimental curves, thereby guaranteeing the reliability and accuracy of the deconvolution results. N₂ adsorption experiments and XPS analyses were performed in triplicate, with the results reported as average values.

2.3. HI Decomposition Tests

The experimental system for hydrogen iodide decomposition catalyzed by activated carbon is shown in our previous study [17]. The catalytic experiment was carried out in a fixed-bed reactor at 500 °C under atmospheric pressure. The reaction was conducted in a quartz tube with an inner diameter of 16 mm. An appropriate amount of activated carbon was fixed in the middle of the tube with quartz wool, and a suitable volume of coarse quartz particles was packed above it. A 55 wt% hydroiodic acid solution was injected into the quartz reactor at a rate of 0.7 mL min⁻¹ by a syringe driven, which instantly evaporated to form gaseous hydrogen iodide. Nitrogen at 60 mL min⁻¹ was used as the carrier gas throughout the experiment. Most of the HI, I₂, and H₂O were condensed into a conical flask by a condenser, while any remaining HI, I₂, and H₂O were absorbed by NaOH solution, Na₂S₂O₃ solution, and silica gel, respectively. The mixture of generated hydrogen and carrier gas nitrogen was collected in a gas sampling bag every 10 min, and the hydrogen volume was measured by a refinery gas analyzer. The HI decomposition test for each sample was performed in triplicate, and the error ranges for the HI decomposition rate and H₂ production rate were indicated with error bars. The decomposition rate of hydrogen iodide was calculated as follows:

$$\eta ={\displaystyle \frac{Q_{{\mathrm{H}}_2}}{Q_{\mathrm{HI}}}}={\displaystyle \frac{273M_{\mathrm{HI}}{Q}_{{\mathrm{N}}_2}{C}_{{\mathrm{H}}_2}}{4.48\times {10}^4\rho T{Q}_{\mathrm{HI}}{C}_{\mathrm{HI}}{C}_{{\mathrm{N}}_2}}}$$

(1)

where Q_H2 and Q_HI represent the molar amounts of product hydrogen and injected hydroiodic acid, respectively (mol min⁻¹). M_HI is the molar mass of hydroiodic acid (128 g mol⁻¹). Q_N2 is the nitrogen flow rate (mL min⁻¹). ρ is the density of hydroiodic acid (1.7 g mL⁻¹). T is the room temperature (K), Q_HI is the injection flow rate of hydroiodic acid (mL min⁻¹). C_HI is the concentration of hydroiodic acid. C_H2 and C_N2 are the concentrations of hydrogen and nitrogen in the gas sampling bag, respectively.

2.4. Computational Details

For amorphous carbon materials, especially the activated carbons in this study, their basic units are finite-size aromatic carbon clusters [22,23]. Therefore, to ensure the consistency between the calculation model and our catalyst system, the carbon cluster models, as shown in Figure 1, were constructed to represent the basic structure of activated carbon. The two adjacent zigzag unsaturated carbon atoms in the uppermost layer were taken as active sites. Similar carbon-based cluster models have been used to calculate heterogeneous catalytic reactions involving carbon and have proven to be effective and convenient [24,25,26]. Additionally, different types of oxygen-containing functional groups were modified on the surface of the original carbon cluster model according to the XPS peak deconvolution results. All density functional theory (DFT) calculations were performed using Gaussian 09 D.01 [27]. Geometric optimizations and vibrational frequency analyses employed the M06-2X method paired with the 6-31G(d,p) basis set [28]. Transition state structures were identified and rigorously validated through intrinsic reaction coordinate (IRC) calculations at the same theoretical level.

The Gibbs free energy G was calculated using the following equation:

$$G=E+ZPE-TS$$

(2)

where E corresponds to the single-point energy of the optimized configuration, while ZPE and TS represent the zero-point energy and entropy correction, respectively, for the optimized configuration.

The Gibbs free energy variations for the physical and chemical adsorption of the HI molecule, ΔG_P and ΔG_C, were calculated as follows:

$$\Delta G_P=G_{P,HI-carbon}-G_{HI}-G_{carbon}$$

(3)

$$\Delta G_C=G_{C,HI-carbon}-G_{HI}-G_{carbon}$$

(4)

where G_{P, HI-carbon} and G_{C, HI-carbon} represent the Gibbs free energies of the physical and chemical adsorption configurations, respectively. G_HI and G_carbon represent the Gibbs free energies of the HI and carbon surface, respectively.

The Gibbs free energy barrier and variation for the H₂ desorption, ΔG and ΔG’, were calculated as follows:

$$\Delta G=G_{transition state}-G_{reactant}$$

(5)

$$\Delta G’=G_{product}-G_{reactant}$$

(6)

where G_reactant, G_{transition state}, and G_product represent the Gibbs free energy of the reactant, transition state, and product, respectively.

3. Results and Discussion

3.1. Physicochemical Structure Characterization and HI Decomposition Activity of Typical Activated Carbons

The elemental analysis results of three commercial activated carbons are shown in

Table 1. Elemental analyses of commercial activated carbon samples.

Sample	C (wt-%)	O (wt-%)	N (wt-%)	H (wt-%)
CAC-1	39.48	14.91	0.43	0.91
CAC-2	82.04	10.68	0.33	0.54
CAC-3	88.73	6.12	0.47	0.35

. The carbon content of sample CAC-1 is only 39.48%, which is significantly lower than that of samples CAC-2 and CAC-3. This discrepancy may be attributed to the selection of raw materials. Generally, coconut shell-based activated carbons exhibit higher carbon contents than other types, leading to better catalytic performance. Samples CAC-1 and CAC-2 have relatively high oxygen contents of 14.91% and 10.68%, respectively, making them both potential candidates for H₂ reduction modification. However, the final selection of the original sample should be determined by its catalytic activity analysis. The sum of C and O contents in sample CAC-1 is less than 60%, indicating a high impurity content in this sample.

Figure 2a displays the adsorption–desorption isotherms of different commercial activated carbon samples. All three samples exhibit a composite characteristic of type I/IV adsorption–desorption. The adsorption capacity rises rapidly at low pressures, and a slight hysteresis loop exists in the middle-pressure range. However, the hysteresis loops of CAC-1 and CAC-2 samples belong to type H2, while that of the CAC-3 sample belongs to type H4. This indicates that CAC-1 and CAC-2 samples show complex pore structures and uniform pore size distributions, while the CAC-3 sample shows an irregular slit-like pore structure. Overall, the N₂ adsorption capacities are in the order of CAC-3 > CAC-2 > CAC-1, reflecting the order of specific surface areas.

Table 2. Pore parameters of CAC-1, CAC-2, CAC-3 from N₂ adsorption.

Samples	SBET a (m2 g−1)	Smic b (m2 g−1)	Smic/SBET	Vt c (cm3 g−1)	Da d (nm)
CAC-1	394	280	0.71	0.24	3.94
CAC-2	1075	956	0.89	0.47	3.67
CAC-3	1264	1137	0.90	0.58	1.30

^a Calculated by the BET model from the adsorption branches of the isotherms. ^b Calculated from the t-plot method. ^c Calculated from adsorption date at P/P₀ = 0.982. ^d Calculated from 4 V_t/S_BET.

lists the detailed parameters of the pore structures of different commercial activated carbons. The specific surface area of the CAC-1 sample is 394 m² g⁻¹, the pore volume is 0.24 cm³ g⁻¹, and the micropore proportion is 0.71. The specific surface area of the CAC-2 sample is 1075 m² g⁻¹, the pore volume is 0.47 cm³ g⁻¹, and the micropore proportion is 0.89. The specific surface area of the CAC-3 sample is 1264 m² g⁻¹, the pore volume is 0.58 cm³ g⁻¹, and the micropore proportion is 0.90. It can be seen that all three samples have the pore structure dominated by micropores, and the distribution of micropores determines the size of the specific surface area. As can be seen from the pore size distribution diagram of different commercial activated carbon samples in Figure 1b, the CAC-3 sample has more micropore distributions, the CAC-1 sample has fewer micropore distributions, and the CAC-2 sample has micropore distributions between the two, which is consistent with the specific surface area data.

The XPS survey spectra of three commercial activated carbon samples are shown in Figure 3a, and the relative contents of C and O elements are consistent with the results of elemental analysis. To determine the types and quantities of oxygen-containing functional groups in each sample, as shown in Figure 3b–d, O1s can be deconvoluted into carbonyl (C=O, 531.4 eV), carboxyl and ester (O-C=O, 532.5 eV), and hydroxyl (C-OH, 533.7 eV). It can be seen from Figure 3e that although the oxygen-containing functional groups in samples CAC-1 and CAC-2 both include carbonyl, carboxyl, ester, and hydroxyl, their distributions are different. The carboxyl and ester account for the majority of all oxygen-containing functional groups in CAC-1, while the proportions of carbonyl and hydroxyl are equal; in sample CAC-2, the types of oxygen-containing functional groups are more evenly distributed. Sample CAC-3 contains a large amount of carbonyl, while the contents of carboxyl, ester, and hydroxyl are extremely low.

As shown in Figure 4a, the catalytic performance of different commercial activated carbon samples for HI decomposition follows the order of CAC-1 < CAC-2 < CAC-3. The three commercial activated carbons exhibit significant differences in HI decomposition activity. The decomposition efficiency of samples CAC-1, CAC-2, and CAC-3 are 8.0%, 17.0%, and 22.8%, respectively. Both samples CAC-2 and CAC-3 show good catalytic capabilities, especially sample CAC-3, whose HI decomposition efficiency approaches the equilibrium decomposition efficiency at 500 °C. The difference in their HI decomposition activity is not determined by a single factor but results from the comprehensive effect of parameters such as elemental composition, functional group content, and pore structure. The correlation analysis reveals a significant positive correlation between HI decomposition rate and pore structure parameters. As demonstrated in Figure 4b and Figure 4c, the HI decomposition rates of representative activated carbon samples (CAC-1, CAC-2, and CAC-3) exhibit linear positive correlations with both BET surface area and pore volume, with correlation coefficients of 0.96 and 0.99, respectively. The increase in specific surface area leads to enhanced HI decomposition rates, which can be attributed to the substantially increased number of active sites (e.g., unsaturated edge carbon atoms) on high-surface-area activated carbon, providing more reactive interfaces for HI adsorption and decomposition. Similarly, the expansion of pore volume improves HI decomposition rate, indicating that the enlarged pore channels significantly optimize the diffusion kinetics of HI reactant molecules within the porous structure, thereby facilitating subsequent catalytic conversion at the active sites.

When performing H₂ reduction modification on commercial activated carbons, the selection of the original sample for modification should balance both oxygen content and catalytic activity. Samples with higher oxygen content can exhibit more obvious gradient changes during H₂ reduction treatment at different temperatures, which is conducive to studying the effect of oxygen-containing functional groups on HI decomposition. Samples with moderate catalytic activity can provide elastic space for activity increase or decrease after H₂ reduction treatment. In summary, sample CAC-2 has a high oxygen content and suitable catalytic activity, so it is further subjected to H₂ reduction modification in this work to investigate the effect of oxygen-containing functional groups on HI decomposition.

3.2. Effects of H₂ Reduction Modification on Physicochemical Structure and HI Activity of Activated Carbon

After H₂ reduction modification, the adsorption–desorption curves of activated carbon samples are shown in Figure 5a. All three samples exhibit the same type I/IV composite adsorption isotherms with type H2 hysteresis loops, indicating the simultaneous presence of micropores and mesopores. As shown in

Table 3. Pore parameters of CAC-2, CAC-2-500, and CAC-2-800 from N₂ adsorption.

Samples	SBET a (m2 g−1)	Smic b (m2 g−1)	Smic/SBET	Vt c (cm3 g−1)	Da d (nm)
CAC-2	1075	956	0.89	0.47	3.67
CAC-2-500	975	902	0.93	0.44	3.78
CAC-2-800	982	903	0.92	0.44	3.30

^a Calculated by the BET model from the adsorption branches of the isotherms. ^b Calculated from the t-plot method. ^c Calculated from adsorption date at P/P₀ = 0.982. ^d Calculated from 4 V_t/S_BET.

, the specific surface areas of modified activated carbon samples have small differences. The original specific surface area of sample CAC-2 is 1075 m² g⁻¹, which changes to 975 m² g⁻¹ and 982 m² g⁻¹ after H₂ reduction treatment at 500 °C and 800 °C, respectively. In addition, the pore volume of CAC-2 changes from 0.47 cm³ g⁻¹ to 0.44 cm³ g⁻¹ and 0.44 cm³ g⁻¹, and the pore diameter changes from 3.67 nm to 3.78 nm and 3.30 nm, respectively. Meanwhile, there is no obvious difference in the pore size distribution of these three samples (Figure 5b). This indicates that H₂ reduction treatment has little effect on the pore structure of activated carbon.

The XPS spectra of activated carbon after H₂ reduction modification (Figure 6a) show the relative contents of carbon atoms and oxygen atoms. After treatment with H₂ reduction at 800 °C, the oxygen contents decreased from 10.68% to 4.35%. To further determine the effect of H₂ reduction treatment on the types and quantities of oxygen-containing functional groups, deconvolution processing of the O1s peak was performed. It can be decomposed into carbonyl (C=O, 531.4 eV), carboxyl or lactone (O-C=O, 532.5 eV), and hydroxyl (C-OH, 533.7 eV). Figure 3c shows that the oxygen-containing functional groups in the CAC-2 sample are C=O, O-C=O, and C-OH structures. It should be noted that due to the spectral overlap of the carboxyl and lactone groups containing O-C=O structures in XPS, their specific type cannot be directly distinguished solely through peak deconvolution. Figure 6b indicates that the oxygen-containing functional groups in the CAC-2-500 sample are mainly carbonyl and hydroxyl groups. This indicates that the O1s peak at 532.5 eV completely disappeared in the CAC-2-500 sample after H₂ thermal treatment at 500 °C. According to literature-reported desorption temperature ranges, carboxyl groups (125–525 °C) decompose at significantly lower temperatures than lactones (425–825 °C), which confirms that the O-C=O structures in the pristine CAC-2 sample predominantly originate from carboxyl groups [21]. In addition, it can be seen from Figure 6c that the oxygen-containing functional group in the CAC-2-800 sample is only carbonyl, which is due to its highest desorption temperature (825–1025 °C). Therefore, as shown in Figure 6d, the carboxyl groups are completely removed at 500 °C, and the hydroxyl groups are removed at 800 °C. Due to the high desorption temperature requirement of carbonyl, a small amount of that is removed at 500 °C and 800 °C, respectively.

Based on the above characterization results, we found that H₂ reduction treatment can gradually remove oxygen-containing functional groups in activated carbon, but has little effect on the pore structure. Therefore, CAC-2, CAC-2-500, and CAC-2-800 obtained by H₂ reduction treatment have gradient changes in the types of oxygen-containing functional groups and consistent pore structures, which can be used as model catalysts to study the effect of oxygen-containing functional groups on HI decomposition.

Figure 7 shows that the HI decomposition performance of activated carbon samples after H₂ reduction treatment follows the order of CAC-2 < CAC-2-500 < CAC-2-800. The HI decomposition efficiency of samples CAC-2-500 and CAC-2-800 reaches 18.4% and 21.0%, respectively, which are higher than that of the original sample CAC-2 (17.0%). The H₂ production rates of CAC-2, CAC-2-500, and CAC-2-800 are 3.48, 3.76, and 4.32 mmol min⁻¹ g⁻¹, respectively. This indicates that H₂ treatment of activated carbon can improve its catalytic activity for HI decomposition. Due to the different oxygen contents and similar pore structures, the relationship between oxygen content and HI decomposition activity was analyzed. It was found that as the oxygen content decreases, the decomposition efficiency of HI catalyzed by activated carbon samples increases. This suggests that the presence of oxygen-containing functional groups is detrimental to the catalytic decomposition of HI by activated carbon, so directional removal of oxygen-containing functional groups can enhance the HI decomposition efficiency. It can be seen that changes in oxygen-containing functional groups significantly influence the catalytic activity of active sites, which is closely correlated with the HI decomposition rate. Thus, the oxygen atom was used as an indicator to quantify the intrinsic catalytic activity, thereby normalizing the catalytic performance. The calculated number of HI molecules converted per unit time was 32 h⁻¹ for the CAC-2 sample. After H₂ reduction treatment at 800 °C, the CAC-2-800 sample achieved a conversion number of 96 h⁻¹, tripling the performance of CAC-2. The results further confirm that reducing oxygen content enhances the intrinsic activity of HI decomposition. It should be emphasized that the evaluation of HI decomposition performance for all catalysts was conducted under strictly identical reaction conditions (500 °C, atmospheric pressure, fixed feed rate, etc.), ensuring the comparability of experimental data. Although the single-temperature experimental design cannot directly determine the reaction activation energy, the systematic comparison of activity differences among different catalysts under identical conditions, combined with the further density functional theory (DFT) calculations of energy barriers for key elementary steps, can effectively reveal the regulatory mechanism of oxygen-containing functional groups on catalytic activity, providing theoretical guidance for designing high-efficiency catalysts.

We further compare CAC-2-800 with other known catalysts.

Table 4. The comparison of CAC-2-800 with other known catalysts.

Sample	Catalyst Type	Treatment Method	HI Decomposition Rate (%)	H2 Production Rate (mmol min−1 g−1)	Reference
CAC-2-800	Activated carbon	H2 reduction (O contents: 4.35%)	21.0 (500 °C)	4.3 (500 °C)	Our work
COAL-90HN	Activated carbon	HNO3 treatment (O contents: 6.68%)	18.5 (500 °C)	1.0 (500 °C)	[15]
NAC-9	Activated carbon	Nitrogen doping (N contents: 6.01%)	20.7 (500 °C)	4.2 (500 °C)	[17]
50N800-4h	Activated carbon	Nitrogen doping (N contents: 7.79%)	23.0 (500 °C)	4.7 (500 °C)	[29]
Pt/Al2O3	Pt/Al2O3	Pt loading of 5 wt-%	1.0 (500 °C)	0.3 (500 °C)	[12]

compiles the catalyst types, treatment methods, and HI decomposition performance reported in the literature and includes the modified activated carbon catalyst (CAC-2-800) in our work for comparison. It can be observed that compared to other activated carbons and supported catalysts reported in the literature, the HI catalytic activity of CAC-2-800 is in the medium-to-high range, especially considering its origin as a commercial activated carbon modified only by a relatively simple one-step H₂ reduction treatment. Data from the COAL-90HN sample in the literature further demonstrates that an increase in oxygen content is detrimental to HI decomposition. Comparison with samples like NAC-9 and 50N800-4h shows that nitrogen-containing functional groups can also significantly improve HI catalytic activity. However, achieving this improvement requires relatively complex nitrogen doping treatments and a relatively high nitrogen content. In addition, comparison with catalysts like Pt/C-5 and Pt/Al₂O₃ reveals that activated carbon is indeed a low-cost, high-performance catalyst and also serves as an effective support for active metals.

3.3. Effect Mechanism of Oxygen-Containing Functional Groups on HI Decomposition

Zhang et al. calculated the catalytic decomposition pathway of HI molecules on zigzag carbon structures using the Gaussian software [30]. The first HI molecule preferentially adsorbs on an unsaturated carbon atom, forming the adsorption structure P1. Then, the second HI molecule continues to adsorb, accompanied by the formation and subsequent desorption of I₂. Following this, the third HI molecule adsorbs to form the product H₂, which then desorbs. This study has demonstrated that the unsaturated edge carbon atoms serve as the active sites for HI decomposition, with the chemical adsorption of HI molecules as the initial step driving the subsequent formation of products. Additionally, the desorption of H₂ molecules from the carbon surface is the rate-determining step of the entire HI decomposition reaction, exhibiting the highest energy barrier. Therefore, to elucidate the effect mechanism of oxygen-containing functional groups on HI decomposition, this work employs DFT calculations to analyze the influence of different oxygen-containing functional groups on the key elementary steps (HI molecule adsorption and H₂ desorption) in the HI decomposition process.

First, the presence of oxygen-containing functional groups directly occupies the active sites (unsaturated carbon atoms), and then the effect of oxygen-containing functional groups on the adsorption of HI molecules was studied to determine whether they could serve as a new type of active site in addition to unsaturated carbon atoms. As shown in Figure 8, we attempted to place HI molecules near various oxygen-containing functional groups adjacent to saturated carbon atoms and found that HI molecules could not form adsorption complexes with oxygen atoms. Instead, they were physically adsorbed near oxygen atoms, indicating that oxygen-containing functional groups themselves cannot serve as active sites for HI chemisorption.

Furthermore, the non-bonding interaction analysis was applied to determine the positions and types of weak interactions. The weak interaction positions were obtained by mapping the isosurface of the reduced density gradient at 0.5 based on the wavefunction after geometric optimization. The types of weak interactions were colored according to the sign(λ₂)ρ values on the isosurface, where sign(λ₂)ρ ranges from −0.05 to 0.05 and the coloring range is −0.04 to 0.02. A bluer color indicates the dominantly strong attractive interactions, such as hydrogen bonds and strong halogen bonds, and a redder color signifies steric hindrance in rings or cages with dominant strong repulsive interactions. A greener color represents van der Waals interactions between molecules. As shown in Figure 8, steric hindrance exists at the center of all aromatic rings, corresponding to the tip at sign(λ₂)ρ = 0.02 a.u. in the scatter plot. Blue regions between HI molecules and O atoms in carbonyl, hydroxyl, and carboxyl groups in Figure 8b–d correspond to the leftmost tips in the scatter plots, indicating hydrogen bond interactions stronger than van der Waals forces. The leftmost tips in the scatter plots correspond to different sign(λ₂)ρ values (0.015 a.u., 0.03 a.u., 0.025 a.u.), suggesting differences in hydrogen bond strength, which is reflected in the physical adsorption energy of HI molecules. After calculation, carboxyl and carbonyl increase the physical adsorption energy of HI molecules on the carbon surface from −23 kJ mol⁻¹ to −31 kJ mol⁻¹ and −41 kJ mol⁻¹, respectively. Additionally, the adsorption energy of HI molecules on the hydroxyl-containing carbon surface (−22 kJ mol⁻¹) is close to that on the original carbonaceous surface, as HI molecules interact with hydroxyl groups via van der Waals forces. Therefore, although oxygen-containing functional groups are not active sites for activated carbon-catalyzed HI decomposition, HI molecules can still undergo physical adsorption near various oxygen-containing functional groups. Among them, the incorporation of carbonyl and carboxyl groups enhances the physical adsorption energy of HI molecules on the carbonaceous surface due to the presence of strong or weak hydrogen bond interactions.

Furthermore, the chemisorption process of HI molecules on unsaturated carbon atoms of the original carbon surface and various oxygen-containing functional group-modified carbon surfaces was analyzed. As shown in Figure 9, the chemisorption configurations and adsorption energies indicate that the adsorption energy of HI molecules on the original unsaturated carbon atoms is −599 kJ mol⁻¹. When oxygen-containing functional groups such as carbonyl, hydroxyl, and carboxyl are close to unsaturated carbon atoms, the adsorption energies decrease to −340 kJ mol⁻¹, −509 kJ mol⁻¹, and −489 kJ mol⁻¹, respectively, with only slight changes in the bond lengths of C-I and C-H bonds. Therefore, this chemical adsorption demonstrates that compared to oxygen functional groups, unsaturated carbon atoms remain the active sites for HI decomposition, thereby validating their thermodynamically preferred role. The intensity of various oxygen-containing functional groups in inhibiting the chemisorption of HI molecules shows the order of carbonyl > carboxyl > hydroxyl. The inhibiting effect is fundamentally attributed to the strong electron-withdrawing effect of oxygen atoms, which reduces the electron density of neighboring unsaturated carbon atoms, thereby weakening their bonding ability with HI (manifested as a decrease in chemisorption energy). The Mulliken charge distribution depicted in Figure 10 further confirms this mechanism—the introduction of carbonyl, hydroxyl, and carboxyl groups reduces the atomic charge of adjacent unsaturated carbon atoms from −0.129 a.u. to −0.058, −0.061, and −0.044 a.u., respectively.

Furthermore, the influence of oxygen-containing functional groups on the H₂ desorption was investigated, and their reaction pathways on the original carbon surface are shown in Figure 11. M is an intermediate, which is the product of the adsorption of the third HI molecule and the reactant for H₂ molecule desorption. In M, two hydrogen atoms are symmetrically connected to the C atom relative to the carbonaceous surface, and then absorb excitation energy to break the C-H bond and combine to form H₂ molecules, forming a transition state (TS). Finally, H₂ molecules leave the carbon surface to form stable products P. The Gibbs free energy calculations were performed on the intermediate M, transition state TS, and product P, respectively, and the energy barrier for H₂ molecule desorption on the pristine carbon surface was 360 kJ mol⁻¹, with an endothermic heat of 334 kJ mol⁻¹.

Figure 11 shows the desorption reaction pathways of H₂ molecules on carbon surfaces modified by carbonyl, hydroxyl, and carboxyl groups, respectively. After the addition of various oxygen-containing functional groups, the desorption pathways of H₂ molecules on each carbon surface remain unchanged, while the Gibbs free energy barrier and endothermic heat increase or decrease relatively. The addition of carbonyl groups reduces the energy barrier for H₂ molecule desorption on the carbon surface from 360 kJ mol⁻¹ to 341 kJ mol⁻¹, and the endothermic heat from 334 kJ mol⁻¹ to 278 kJ mol⁻¹. The addition of hydroxyl groups reduces the Gibbs free energy barrier for H₂ molecule desorption on the carbon surface from 360 kJ mol⁻¹ to 353 kJ mol⁻¹, while the endothermic heat increases from 334 kJ mol⁻¹ to 336 kJ mol⁻¹. The addition of carboxyl groups increases the Gibbs free energy barrier of H₂ molecule desorption on the carbon surface from 360 kJ mol⁻¹ to 368 kJ mol⁻¹, and the endothermic heat increases from 334 kJ mol⁻¹ to 340 kJ mol⁻¹. Therefore, different types of oxygen-containing functional groups have different effects on H₂ desorption: carbonyl and hydroxyl groups accelerate the desorption of H₂ molecules on the carbon surface, while carboxyl groups hinder the desorption of H₂ molecules on the carbon surface.

The above calculation results indicate that the incorporation of oxygen-containing functional groups occupies the active sites (unsaturated carbon atoms), and significantly alters the chemisorption of reactant HI molecules and the chemical desorption of product H₂ molecules on the neighboring unsaturated carbon atoms. However, different types of oxygen-containing functional groups vary in their degrees of promoting or inhibiting HI molecule adsorption and H₂ molecule desorption.

Table 5. Adsorption energies of HI and H₂ desorption energy parameters on carbon surfaces modified with different oxygen-containing functional groups.

Functional Groups	ΔGP (kJ mol−1)	ΔGC (kJ mol−1)	ΔG (kJ mol−1)	ΔG’ (kJ mol−1)
Pristine carbon	−23	−599	360	334
Carboxyl (C-OOH)	−31	−489	368	340
Hydroxyl (C-OH)	−22	−509	353	336
Carbonyl (C=O)	−41	−340	341	278

summarizes the adsorption energies of HI and H₂ desorption energy parameters on carbon surfaces modified with different oxygen-containing functional groups. The results demonstrate that carboxyl, hydroxyl, and carbonyl groups all significantly inhibit HI chemisorption, reducing the adsorption energies by 110, 90, and 259 kJ·mol⁻¹, respectively. For H₂ desorption, the carboxyl group raises the energy barrier by 8 kJ mol⁻¹, indicating the inhibitory effect. Although hydroxyl and carbonyl groups lower the H₂ desorption barriers by 7 and 19 kJ·mol⁻¹, their inhibitory effect on HI adsorption is more pronounced. This confirms that oxygen-containing functional groups primarily suppress the overall reaction by hindering HI chemisorption at active sites. Consistent with experimental results, the ability of activated carbon to catalytically decompose HI is enhanced after removing oxygen-containing functional groups due to the restoration of both the quantity and electron density of unsaturated carbon atoms. Therefore, we conclude that common oxygen-containing functional groups, including carbonyl, hydroxyl, and carboxyl groups, are unfavorable for activated carbon-catalyzed HI decomposition. For activated carbon applied in the direction of HI decomposition, its preparation should reduce the relative content of oxygen-containing functional groups.

4. Conclusions

In this work, the catalytic activity of activated carbon for HI decomposition was improved by gradient removal of oxygen-containing functional groups, and the inhibition mechanism of oxygen-containing functional groups on the HI decomposition process was revealed. Experiments showed that H₂ reduction treatment can directionally remove oxygen-containing functional groups on the surface of activated carbon, maintaining the pore structure while enhancing the HI catalytic activity, confirming that the reduction of oxygen-containing functional groups is the key factor for the increase in HI decomposition efficiency. DFT calculation results showed that various oxygen-containing functional groups can weaken the chemisorption of HI on unsaturated carbon atoms, and carboxyl groups will increase the desorption activation energy of H₂, while carbonyl groups and hydroxyl groups, although reducing the desorption energy barrier of H₂, have a more significant inhibitory effect on HI adsorption. This study provides a theoretical basis and modification strategy for the design of high-activity carbon-based HI decomposition catalysts, that is, optimizing their catalytic performance by regulating the process to reduce the content of oxygen-containing functional groups.