Along with interaction–based separation, the shape and size of molecules may also affect the adsorption selectivity of one compound over another. Generally, it is desirable for the adsorbent to exhibit increased separation selectivity and adsorption capacity towards target molecules (e.g., H2S) as well as stability under acid conditions, and structural regenerability in order to increase its cost-effectiveness.
2.1. Zeolites
Zeolites are widely used as catalytic materials in the petrochemical industry and as water softeners in detergents [
20]. They are also considered molecular sieves with the capacity to selectively remove a target molecule from a fluid mixture, e.g., they have been extensively used in order to remove H
2S from biogas. Structurally, zeolites are microporous crystalline materials principally comprised of a three-dimensional framework of SiO
4 and AlO
4 tetrahedra. Their frameworks are typically anionic and have charge–compensating cations such as Na
+ and Ca
2+ to retain electrical neutrality. These cations can be a part of ion-exchange processes, making them ideal selective H
2S adsorbents, as is the case in the well-known Claus process [
21].
Several papers published so far have attempted to shed light on the retention mechanism of H
2S in different types of zeolites, such as zeolite–A (LTA), zeolite–Y (FAU), zeolite–X (FAU), and ZSM–5 (MFI), as well as natural zeolites (i.e., mordenite, clinoptilolite, erionite, phillipsite, and ferrierite), without however reaching a consensus, as some studies support that the formation of elemental sulfur predominates [
22], while others propose an acid–base reaction [
23]. Moreover, whether the water formation during the adsorption of H
2S is attributed to the presence of oxygen contained either in the inlet gas mixture or in the zeolite framework or to the metal oxides of the extra-framework cations remains unclarified [
24]. Sigot et al. [
25] attempted to provide an exegesis by studying the retention mechanism of H
2S on NaX zeolites with a Si/Al ratio of 1.4, suggesting a mechanism based on adsorption-oxidation and the formation of elemental sulfur. The authors argued that the formation of stable sulfur polymers affects the thermal regeneration of the zeolite, and they proposed regeneration by high-temperature desorption at 500–600 °C in an inert atmosphere, which is nevertheless an energy-consuming process damaging the zeolite’s crystalline pore structure. Desorption-oxidation at a temperature <400 °C was also considered as an alternative. The authors suggestion of chemical dissolution with organic solvents maybe feasible but is not an environmetally sustainable solution.
As mentioned above, most industrial gases contain both acid (H
2S, CO
2) and polar species (CO and H
2O), which compete and at times supplant H
2S adsorption. Kumar et al. [
26] throrougly investigated the effect of the presence of molar compounds on the adsorption of H
2S for Na, Ag, and Cu–exchanged FAU at high and low Si/Al ratios. The authors concluded that NaX exhibited almost zero capability for capturing H
2S in the presence of CO, CO
2, or H
2O, while CuX and CuY demonstrated reduced H
2S selectivity in the presence of CO. AgX and AgY zeolites showed H
2S adsorption capacities of approximately 30–40 mg/g
ads when H
2S concentration in the mixture was 10 ppm, even in the presence of all three components.
An interesting study in the context of polarity-based adsorption on zeolites was reported by Mohammed and Nassrullah [
27], who prepared zeolite 5A using kaolin clay and tested H
2S removal from liquefied petroleum gases (LPG) in the presence of H
2O. The authors used breakthrough curves to draw conclusions with regard to the characteristics of the zeolite, and as expected, found that H
2O was a strongly adsorbed component, in comparison to H
2S which was a weakly adsorbed component. As has been argued, even traces of water may essentially decrease the adsorption capacity for other gasses on zeolites and lower their intracrystalline diffusivity [
28]. Mohammed and Nassrullah [
27] also demonstrated better adsorption capacities for H
2S when increasing flow rates from 3 to 5 liters/minute due to the fact the contact time between the water vapor and adsorbent was decreased. The adsorption capacity for H
2O when the flow rate was 5 liters/minute reached 7.547 g/g while for H
2S it reached 1.734 g/g.
Bulow et al. [
29] examined H
2S/CO
2 reactive adsorption on LTA and FAU zeolites. Owing to the highly preferred adsorption of H
2O on these materials, which follows the order H
2O ≫ H
2S ≥ CO
2 ≈ COS, the equilibrium of the reaction H
2S + CO
2 ⇄ COS + H
2O moves to the right; therefore, according to the reaction stoichiometry, one H
2S molecule is converted to another COS molecule, with infinitesimal enrichment of sulfur on the adsorbent. The impact of various parameters such as, Si/Al ratio, type of zeolite, cation exchange degree, pre-adsorption of appropriate gases (NH
3) and temperature was also investigated during this work. At ambient temperature, H
2S conversion on LTA reached 50% for a Si/Al ratio of 1, but it was decreased to almost zero for a Si/Al ratio of 3. The authors reported that when ion exchange increased, wherein higher activation energy was required, the conversion of H
2S was between 20% and 75%, depending on the time of reaction. Thereby, due to the difficulty in repressing this reaction, the authors suggested short–time adsorption techniques such as pressure swing adsorption (PSA)/vacuum swing adsorption (VSA).
Tomadakis et al. [
30] used three different types of molecular sieve adsorbents (4A, 5A, 13X) to separate mixtures containing bulk quantities of H
2S and CO
2 by means of PSA. Following the adsorption step, CO
2 was removed from the column by blowdown to increase the purity of the H
2S stream during desorption. According to the study results, 5A and 13X exhibited higher selectivity (84–85%) in comparison to 4A (61%) for adsorbing H
2S over CO
2. The authors also observed that H
2S was the strongest adsorbed component in all cases. Desorption recovery for 4A zeolites was the greatest at 76–91%, followed by 5A and 13X (40–44% and 12–14%, respectively).
Another study focused on short-time adsorption techniques was reported by Wynnyk et al. [
31], who performed extensive high-pressure experiments for selective adsorption of CO
2, COS, CH
4, and H
2S using 4A zeolite (LTA). Two different Si/Al ratios were used, i.e., 1.14 and 1.15. Prior to H
2S adsorption tests, the 4A zeolite was activated under ultra-high vacuum (10
−10 bar) up to 400 °C. The comparison of the saturation loadings derived from the fitted adsorption isotherms for CH
4 (9.16 mmol/g), CO
2 (7.25 mmol/g), H
2S (5.87 mmol/g), and COS (3.11 mmol/g), revealed considerable differences between these species. Computational calculations for highest occupied molecular orbital (HOMO) bond orientation suggested that zeolite 4A did not perform well in the presence of COS in the gas stream. The positive correlation between CO
2 saturation loading and temperature was not clarified by the authors. In particular, lower temperatures (0, 25, and 50 °C) resulted in reduced saturation loadings for CO
2. Moreover, the enthalpy of adsorption was calculated using isotherms at different temperatures. It was found that at low loadings, COS as well as CH
4 provided identical values of 20 kg/mol, whilst the vigorously interacting H
2S and CO
2 yielded values of ~35 kJ/mol (close to saturation loading). The low enthalpy values indicated that chemical adsorption was favored. Nevertheless, it is worth mentioning that these interactions may be susceptible to even very small concentrations of H
2O contained in the feed gas.
However, despite the fact that the latter two separation methods (PSA/VSA) are of relevance with respect to acidity/based separations (i.e. CO
2/H
2S), they are economically unattractive due to their high operating costs. For this reason, numerous works attempted to achieve selective H
2S adsorption by applying more cost-efficient methods. Sigot et al. [
32], carried out both laboratory-scale (with synthetic gas) and on-site (with real biogas) adsorption tests. The laboratory-scale adsorption tests were performed using 80 ppmv of H
2S, simulating a partly desulfurized gas, and compared the efficiencies of 13X zeolite (Z) and a coal-based impregnated activated carbon (IAC) in different humidity conditions (up to 80%). With respect to IAC, suppliers generally recommend 40–80% relative humidity (RH) for an optimal efficiency, and this is confirmed by data in the literature [
14,
33]. The absolute gas humidity at the column inlet was ~14.5 g·m
−3. At the laboratory scale, dry N
2 (containing ~1% O
2 due to incomplete air separation in the N
2 generator) was used as diluent stream. A 25 mm diameter and 25 cm high stainless steel adsorption column was used. For each experiment, 10–30 g of adsorbent was introduced in the adsorption column. The 13X zeolite (Z) demonstrated a sufficient H
2S uptake on dry gas but its capacity decreased almost to zero in the presence of humidity due to its high affinity for water. On the other hand, IAC, which was treated with alkaline solution, showed enhanced performance in the presence of humidity owing to its lower affinity for water. In summary, the presence of water decreased the adsorption efficiency of Z towards H
2S, but enhanced the removal of H
2S by IAC. Another conclusion of the study was that CH
4 was not adsorbed regardless of the adsorbent. Moreover, Z had the ability to retain CO
2 (corresponding to approximately 90 mg CO
2/g Z adsorbed), which is expected to reduce its H
2S adsorption capacity. This was anticipated, considering that zeolites are commonly used for biogas upgrading (biogas to biomethane) [
34]. Concerning the on-site adsorption tests (H
2S concentration of 4000 ppmv), the authors, taking into to account the laboratory-scale experiments, concluded that the H
2S adsorption capacity of Z was 142 mg H
2S/g Z. With respect to IAC, an adsorption capacity of 785 mg H
2S/g IAC was demonstrated, without reaching saturation of the adsorbent.
Barelli et al. [
9] also tested a 13X zeolite treated with Cu ions (13X Ex–Cu) by impregnation or ion exchange as sorbent for H
2S removal from biogas. Material behavior was characterized over several operating conditions typical of biogas mixtures (high H
2S contents, 200 and 1000 ppmv CH
4 (60%)/CO
2 (40%) matrices) and compared to specific conventional adsorbents. The impact of changing parameters such as reactor temperature, space velocity, gas matrix compositions, and particles size was investigated. The material demonstrated its best performance at reduced space velocity (850 h
−1), at increased reactor temperature (120 °C), and in high CH
4 concentrations in the gas mixture (1000 ppmv), with H
2S uptake of 40 mg/g. Brunauer-Emmet-Teller (BET) and X–ray diffraction (XRD) analysis showed that the improved performance was related to the high concentrations of Cu
2+ ions, confirming the well–known affinity of Cu
2+ for S
2− ions [
23].
Another work that highlighted the increased capacity of 13X in H
2S adsorption was reported by Melo et al. [
35], who used the material as adsorbent for the removal of H
2S from natural gas. The composition of the zeolite was Na
2O:Al
2O
3:2.8 ± 0.2SiO
2:XH
2O, and X depended on the dryness or activation of the material according to X–ray diffraction peaks of zeolite 13X similar to the Faujasite structure. The authors concluded that its maximum adsorption capacity, at a surface area of 16.15 m
2/g, was 53 mg H
2S/g
ads at 25 °C.
As was previously mentioned, there are numerous types of zeolites used for H
2S removal. Lee et al. [
36] conducted an intriguing study focusing on the adsorption capacity of LTA. They tested iron-incorporated Na–A zeolites prepared from FeCl
3 solution and melting slag, which contained sodium silicate solution and sodium aluminate (main constituent), as adsorbents for hydrogen H
2S removal. The gas mixture comprised of N
2 and H
2S (1%) and the flow rate was 100 ml/min. The samples were packed into a bed (length 30 cm and diameter 15 mm) which was placed inside an oven maintained at 30 °C. The tests were stopped at a breakthrough concentration of 2000 ppm for H
2S. With respect to H
2S adsorption capacity, it was directly proportional to Fe
3+ concentration and inversely proportional to calcination temperature. In particular, there was an increase in the surface area from 20 m
2/g to 85 m
2/g when the Fe
3+ concentration increased from 0 nM to 78 nM. The highest H
2S adsorption capacity (1.5%) was found at 78 nM Fe
3+ solution calcinated at 200 °C. When the adsorption capacity of Fe–Na–A (0.2–1.5%) was compared to other adsorbents, it was found to be lower than that of the commercial activated carbons (1.2–2.4%) and higher than that of commercial zeolites 13X (0.91%) and 4A (0.15%), owing to acidic surface properties presented by the Fe–Na–A, which is attributed to the incorporation of the Fe
3+ ion.
Abdullah et al. [
37] studied the adsorption of H
2S from a gas stream (200 ppmv H
2S in 30 mL min
−1 N
2 flow) in a fixed–bed reactor at room temperature (28 °C, 1 atm) using zinc oxide (ZnO) (10–30 wt% loading) impregnated on Na–A zeolite prepared from kaolin. Increasing zinc oxide loading concentration over a critical dispersion capacity led to the formation of bulk zinc oxide at the external surface of the material. The material provided its optimum adsorption capacity, 15.75 mg S/g sorbent, when 20 wt% ZnO was loaded on Na–A zeolite. Crystallinity was negatively affected when wt% ZnO increased above 20 wt%, resulting in lower H
2S adsorption capacity.
Liu et al. [
38] also investigated LTA zeolites. In particular, they prepared and tested a 4A zeolite using attapulgite in different conditions in order to remove H
2S from various industrial gases such as natural gas, coal gasification gas, and digester gas. The gas mixture used for the experiment contained nitrogen as carrier gas and 1000 mg/m
3 of H
2S, while the flow rate was 1000 mL/min. The temperature range was between 25 and 75 °C. The authors concluded that the highest amount of adsorbed H
2S onto zeolite was 8.36 mg/g
ads, accomplished at 50 °C. The regeneration performance of the sorbent was also studied and significant damage of the particle structure during the regeneration process was observed, leading to reduced adsorption capacity. Finally, it was revealed that the adsorption of H
2S molecules was hindered at high temperatures due to the exothermic reaction resulting in decreased adsorption capacity at 75 °C (6.5 mg/g
ads).
In a number of papers in the literature, comparative assessments of the performance of different types of zeolites were performed. For example Yokogawa et al. [
39] prepared LTA (zeolite–A), MFI (ZSM–5), Ag–grafted LTA, and Ag–grafted MFI zeolites in order to remove volatile sulfur compounds (VSCs), and compared their capacity in capturing H
2S. Experiments were conducted for an initial H
2S concentration of 30 ppm and the authors showed that H
2S adsorption capacity of LTA was higher than that of MFI. In addition, Ag–grafted LTA and Ag–grafted MFI exhibited higher adsorption capacity over LTA and MFI. This was due to the exchange of alkali metal or alkaline earth metal atoms of zeolite material to univalent metal ions, which improved the H
2S reduction effect. The authors considered that the etiology of the lower H
2S adsorption of the MFI was due to its lower surface potential in comparison to that of the LTA zeolite framework. Results showed that after 24 h, the H
2S concentration fell to 60% for the LTA and to 87% for the MFI. For the Ag–grafted zeolites, the H
2S concentration dropped to almost 0%; an important difference was that this was achieved after 4 h for the Ag–grafted LTA and after 8 h for the Ag–grafted MFI.
It is also worth briefly mentioning the work carried out by Karge et al. [
40], which pioneered the investigation of the properties of FAU–type zeolites. They found that similar to hydrogen FAU (HY) zeolites, on FAU (NaX) zeolites (Si/Al ~ 2.5), OH groups and S
2− are produced due to the fact that the first H
2S molecules are adsorbed dissociatively. However, the authors also showed that increased coverage not only led to the dissociation of H
2S into HS
− and H
+ (
Figure 2), but also to physical adsorption. Nevertheless, on NaY–type zeolites (Si/Al > 2.5) H
2S molecules are adsorbed without dissociation. The authors concluded that H
2S adsorption capacity of X–type zeolites was four times higher relative to Y–type zeolites.
Recently, De Oliveira et al. [
41] published an interesting study in which they used a Na–Y zeolite (hydrophilic structure due to Si–OH groups) as an adsorbent for H
2S removal from simulated natural and biogas mixtures in equilibrium and fixed bed conditions. Adsorption and desorption isotherms were acquired for various temperatures and H
2S partial pressures and modeled using Tόth and Dubinin-Astakhov equations with a view to determining the interactions that occurred. In particular, experiments were conducted at 20, 30, and 40 °C and partial pressures from 0.01 to 1.2 bar. The ranges of operating conditions were typical of natural gas and biogas processing in Brazil. A linear driving force (LDF) model was used to evaluate the adsorbent particle’s effective mass transfer coefficient. Equilibrium results specified maximum adsorption capacities above 6.0 mol kg
−1 for the highest partial pressures. Isotherm modeling revealed the prevalence of physisorption of H
2S in the zeolite. Nevertheless, a hysteresis effect was detected for desorption curves and ascribed to H
2S chemisorption on the zeolite surface. The breakthroughs recorded were non-symmetrical and illustrated the presence of internal and external mass transfer resistances. Mass transfer zones showed no major changes with inlet flow rate, while higher availability of H
2S at the sorbent’s surface, promoted the mass transfer in the intraparticle region. The authors concluded that the results obtained demonstrated that the LDF model could accurately depict the behavior of fixed bed adsorption data.
Ratnasamy et al. [
42] also studied the adsorption capacity of FAU–type zeolites. They fabricated Ca–X and Na–X synthetic zeolites with a view to purifying natural gas at ambient conditions of pressure and temperature for use in polymer electrode membrane (PEM) fuel cells. Additionally, mixed metal oxides (Cu–Mn, Fe–Mn) supported on Al
2O
3 were also studied. As expected, none of the materials tested managed to remove all the sulfur species, which led to the use of a sequential bed system. The results showed that the removal of H
2S was achieved in the third bed. It is worth mentioning that the sequential location and quantity of the sorbents also affected the overall adsorption performance. More specifically, a sequential bed system composed of zeolite Ca–X followed by Fe–Mn–Al
2O
3 oxides had notable desulfurization capacity at 38 °C, 15 psig, and a natural gas space velocity of 6000 h
–1. This work divulged that the adsorption capacity of alkali exchange zeolites was Ca–X > Na–X, but also showed that Al–supported Cu–Mn and Fe–Mn adsorbents demonstrated the best performance toward the H
2S capture.
Another interesting work was reported by Rong et al. [
43], who carried out regeneration and sulfidation experiments with cerium ions and active copper supported onto zeolite–Y (Ce–Y, Cu–Y) with a view to examining the removal of hydrogen sulfidefrom fuels (such as JP–8). Ion exchange was carried out by exchanging Na
+ with Cu
2+. The Cu–exchanged zeolite was further exchanged by a triple positively charged cerium (Ce
3+) ion, thus preparing a mixed metal supported sorbent. Sulfidation experiments were performed for the conditions of 50% H
2–10% H
2O–balanced He with a space velocity of 80,000 h
−1. Results showed that the overall sulfidation capacity of the mixed metal oxides was higher compared to the corresponding zeolite–supported counterparts (i.e., at 800 °C the maximum capacity for Cu–Ce–Y was 1.78 mg S/g
ads, while the maximum capacity for 5% CuO/CeO
2 was 2.45 mg S/g
ads). However, neither copper nor cerium oxide were able to resist sintering above 650 °C, thus resulting in a reduced capacity. On the other hand, the sulfidation capacity of the zeolite–supported mixed metal oxide (Cu–Ce–Y), seemed unaffected by the temperature at both 650 and 800 °C, suggesting that the Cu–Ce–Y zeolite is resistant to the sintering at high temperatures. In particular, during 12 sulfidation/regeneration cycles at 650 and 800 °C the capacity ranged between 1.38 and 1.58 mg S/g
ads. Ce–Y and Cu–Ce–Y exhibiting increased stability in surface area than Cu–Y at 800 °C may represent good candidates for H
2S removal.
Crespo et al. [
44] provided a comparison of H
2S breakthrough capacities of four different molecular sieves, namely Cu(I)Y, Cu(II)Y, CuCl/MCM–41, CuCl/SBA–15, and AgY, and showed that the adsorption energy degraded as follows: Cu(I)Y > AgY > Cu(II)Y. Considering the chemical composition of AgY zeolites, Ag
63Al
57Si
135O
384, it was detected that the Ag/Al ratio was above unity, suggesting that there are cations found at sites other than cation exchange sites. Regeneration of the Cu(I)Y zeolite reached about 70% when heating at 450 °C using He; complete regeneration was achieved when Cu(I)Y reacted with air at 350 °C, indicating the presence of cations in an oxide form. Due to their high CH
4 selectivity, high H
2S adsorption capacity at low H
2S concentrations (100 ppm), and capability to fully regenerate at 350 °C, CuCl/SBA–15 and CuCl/MCM–41 can also be regarded as promising adsorbents.
A noteworthy work in the context of silicoaluminophosphate zeolites was reported by Hernandez–Maldonado et al. [
45] who employed SAPO–43 (partially calcined), to separate CO
2, H
2O, and H
2S, gas molecules typically found in natural gas streams. A major drawback of this material was its lack of thermal stability over 300 °C, which limits its potential use in separation achieved by adsorption. The authors performed elemental analysis as well as CO
2 heats of adsorption, which helped them identify amine-containing compounds; the presence of the latter helped the framework from being dilapidated. For SAPO–43 and at ambient temperature and pressure, the authors recorded capacities of 1.1, 4.93, and 2.52 mmol/g towards the adsorption of CO
2, H
2O, and H
2S, respectively. Regeneration using pure helium at low temperatures (180 °C) removed nearly all the captured water, while 40% of the captured H
2S was restored. Using chemicals such as hydrogen peroxide (H
2O
2) instead of thermal regeneration may also be an alternative for the regeneration of the sorbent.
Several works in the literature have shown that compared to synthetic zeolites, natural zeolites can exhibit sufficient H
2S adsorption capacities, easier regeneration, and lower operating costs. For example, Yasyerli et al. [
46] probed natural clinopltilolite for the capture of 1.06% H
2S from carrier gas at different temperatures between 100 and 600 °C, and found that its adsorption capacity was 0.087 g S/g
ads at 100 °C, which was higher compared to that at 600 °C (0.03 g S/g
ads). Thus, the authors concluded that increasing temperature reduced the capacity of the zeolite in capturing H
2S. Pourzolfaghar et al. [
47] studied H
2S removal efficiency of a virgin zeolite for palm oil mill effluent (POME) biogas desulfurization at ambient temperature and pressure. One of the key points of the investigation was that at higher Si/Al ratios, the adsorption efficiency was increased. Moreover, it is worth mentioning that after the zeolite had been exposed to fresh air for four days or more, regeneration was not only achieved without the need for an extra regeneration process, but the adsorption capacity of the sorbent also increased. Experimetal results showed that the natural zeolite reached 94% capacity of H
2S removal.
2.2. Activated Carbons
Activated carbons (ACs) represent another group of materials extensively used as adsorbents or catalysts for H
2S removal due to their surface chemistry, high degree of microporosity, and developed surface area (that can exceed 1000 m
2/g) [
48]. They are amorphous materials usually produced from easily available carbonaceous materials such as wood, coal, coconut shells, and rice husks at high temperatures in an inert atmosphere. Steam activation of the highly microporous resultant carbon is usually performed to expand the pore structure in order to facilitate the adsorption. Although the retention mechanism of H
2S on activated carbons has been thoroughly investigated, it is almost impossible to rationalize all the emerging phenomena. This is due to differences in the pH of adsorption, the presence of multiple species in the activated carbon, and the wide variety of carbonaceous materials with different functional groups. Several studies reported high H
2S adsorption using different kinds of activated carbons, but the precise reasons that led to it are not perfectly comprehended.
Pure activated carbon is considered as an attractive candidate for adsorbing sulfur compounds due to its cost–effectiveness [
49,
50], but its performance can be further increased by surface modification with chemicals, such as transition metals and/or their salts, leading to the production of metal sulfide [
51,
52] or basic compounds [
53] which promote H
2S removal through acid-base reactions or strong oxidizing agents, which result in H
2S oxidation. For example, Turk et al. [
54] suggested that the impregnation with compounds that are basic in nature (e.g., ammonia) and/or caustic solutions can improve the adsorption of H
2S. However, impregnated activated carbons may demonstrate a decreased capacity for physical adsorption as a result of the reduction of the available micropore volume.
Adib et al. [
55] prepared an activated carbon impregnated with urea in order to increase basicity, but the H
2S conversion dropped from 298 to 272 mg/g for the impregnated sorbents relative to the bare AC, which was probably caused by a substantial drop in the surface area of the sorbent (18%). Similarly, Tian et al. [
56] tried to enhance the removal of H
2S by employing 3-aminopropyltriethoxysilane-surface-functionalized activated carbon (APTES), which is basic in nature, hoping that it would improve interaction with acidic H
2S. However, they noticed decreased breakthrough capacities (lower than 1 mg/g at 1600 ppm), possibly owing to the occlusion of the pores by APTES impregnation. The authors also reported that at a concentration of 8200 ppm of H
2S, the adsorption capacity values for APTES–AC, Na
2CO
3–AC, and AC were 18.13, 8.88, and 6.03 mg/g, respectively.
The average pH of the carbon surface coupled with the pH of the areal environment can also affect the adsorption capacity of the sorbent. An acidic environment hinders the dissociation of H
2S, leading to low concentrations of H
2S ions and thereby to the formation of highly dispersed sulfur species. Notwithstanding, for a less acidic surface, the degree of H
2S dissociation is higher, and the creation of polymeric elemental sulfur species resistant to further oxidation is favored. In summary, an alkaline environment promotes the formation of elemental sulfur, whilst an acidic environment favors the formation of sulfur oxides and sulfuric acid. Bandosz et al. [
57,
58,
59] conducted various studies on the removal of H
2S using ACs that differ in their acidity, pore volumes, and surface areas.
Regeneration of activated carbons via cold and hot water washing at 300 °C in air atmosphere following the first adsorption run resulted in a considerable decrease (approximetely 60%) in H
2S adsorption capacity of the carbon owing to the irrevesible chemical adsorption in the active sites of the material [
60,
61]. Boudou et al. [
62] tried to regenerate an activated carbon used as adsorbent for H
2S removal, in nitrogen by heating for 10 min at 500 °C. However, while it is possible to maintain inert atmosphere at laboratory scale (thus eliminating the chance of AC self-ignition), this is not the case at industrial scale. Moreover, the authors observed that used activated carbons were virtually unwashable with water, owing to the agglomeration of sulfur polymers that are resistant to oxidation, making them very difficult to be reused. Therefore, replacement with fresh activated carbon is a frequently used industry practice for lowering the operational cost and overcoming conventional regeneration issues [
63,
64].
The presence of moisture and oxygen (
Figure 3) may also affect the adsorption capacity of activated carbons, and numerous studies have investigated their impact on H
2S uptake. It is generally accepted that the presence of water may have a positive or negative effect on the adsorption mechanism depending on the type of adsorbent used [
32]. When activated carbons are impregnated with metal oxides, water competes with H
2S on carbon active sites, resulting in the gradual catalyst deactivation [
65,
66]. In contrast, for carbons treated with basic compounds, humidity considerably enhances H
2S adsorption capacity [
67]. Moreover, the presence of oxygen also increases the breakthrough time of H
2S adsorption for the latter adsorbents [
68,
69,
70].
Several studies published so far have dealt with the effect of oxygen and humidity on the adsorption capacity of activated carbons. As an example, Papurello et al. [
72] tested three commercial activated carbons, namely AirDep CKC (impregnated with 5% potassium bicarbonate (KHCO
3)), Sulfa Trap R&G (Fe: 8.3 wt%, Cu: 19.8 wt%), and AirDep CKI (impregnated with 2% potassium iodide (KI)), for the removal of H
2S from low-H
2S level biogases (50–200 ppm) at 30 °C. They found that SulfaTrap R&G exhibited the best adsorption capacity (10.87 mg/g) under anoxic conditions due to a catalytic oxidation mechanism promoted by copper and iron particles identified on the surface of activated carbon. Results also revealed that by adding 0.1% of oxygen in the gas mixture, the alkaline AirDep CKC increased its adsorption capacity from 3.91 to 84.87 mg/g, surpassing both Sulfa Trap R&G (47.91 mg/g) and AirDep CKI (27.04 mg/g) (
Figure 4). Finally, the authors concluded that the presence of oxygen led to the formation of elemental solid sulfur that is adsorbed on carbon pores yielding water release, as depicted by the following equations (Equation (1)–(2)):
In a very interesting study, Ciahotny et al. [
73] tested several commercially available carbons impregnated with oxidants to emphasize the functionality of impregnated adsorbent materials for H
2S removal from biogas in the presence of O
2 at ambient temperature. In terms of the experimental test conditions, the adsorption capacity of the selected impregnated sorbents was evaluated under four different atmospheres i.e., (1) containing approximately 20% O
2, 15 g/m
3 H
2O, and 7.5 g/m
3 H
2S, balanced with nitrogen, (2) without oxygen, (3) without moisture, and (4) without oxygen and moisture. Experimental results showed that the presence of oxygen in the gas mixture was beneficial for H
2S removal. Moreover, the authors concluded that copper, iodine, and potassium salts outperformed the other impregnating agents. In contrast, sorbents impregnated with Mn or Mg were not regarded as suitable candidates due to their decreased sulfur adsorption capacity (below 5%) (
Table 3).
In particular, in the presence of humidity and under oxic conditions, iodine–impregnated sorbents reached a sulfur adsorption capacity of close to 25% (
Table 3), followed by sorbents impregnated with copper or potassium salts (close to 10%). Under anoxic conditions, a considerable decrease in adsorption capacity of the adsorbent to values as low as one–tenth of the initial ones was noticed. This behavior is probably related to changes on the H
2S bond mechanism of the sorbent. Under anoxic conditions, H
2S bonds as a result of physisorption instead of chemisorption. The authors also concluded that the CuO-impregnated sorbent was an exception, demonstrating a lower adsorption capacity decrease to one–third of the initial value, probably due to the strong oxidation properties of CuO. Compared to the absence of oxygen (
Table 3), the absence of humidity did not have such a strong influence on the adsorption capacity of the impregnated sorbents. A dry environment (
Table 3) led to a decrease of adsorption capacities for H
2S to almost half the value reached in a humid environment. CuO once again demonstrated exceptional behavior, with minor changes in adsorption capacities or no changes at all.
This subject has been studied recently by Barelli et al. [
74], who compared two commercial activated carbons, namely AC Ultra DS (treated with KOH–KI) and Norit RGM1 (which contains Cu (II)–Cr (IV) salts), and performed several dynamic adsorption tests in order to study their performance under the influence of humidity and oxygen content in concentration and conditions typical of biogas mixtures (50% CH
4/50% CO
2, RH 90%, and O
2 2%
v). Unlike Norit RGM1, humidity, in the presence of oxygen, enhanced the adsorption capacity (at RH 90% and O
2 2%v, with the highest capacity found for the N
2 carrier at 65.60 mg/g) of alkaline–treated carbon (AC Ultra DS). The authors reported that a humid atmosphere was required in order to ensure the dissociation of H
2S and to facilitate the completeness of chemical reactions needed for H
2S oxidation on the carbon surface. In contrast, the negative effect of humidity on AC RGM1 was attributed to sorbent deactivation, caused by the reduction of amorphous copper oxide (Cu
2O) species [
75]. Moreover, considering that H
2S adsorption by Cu–coated AC is an exothermic reaction, when humidity is present in the inlet gas mixture, water vapor condensation and water film formation may not occur. Therefore, moisture (water vapor) and H
2S species compete with each other for carbon active sites, leading to decreased H
2S adsorption. The authors also investigated and confirmed the competitive adsorption between H
2S and CO
2, and highlighted the negative effect of CO
2 towards H
2S capture. Concerning the impact of gas matrix composition, enhanced performance was noted when using high percentages of CH
4 under dry and anoxic conditions. Thus, the latter three works clearly explain the determining effect of water and oxygen on H
2S adsorption capacity of activated carbons.
The increasing interest in developing new cost–effective materials as sorbents for H
2S removal from industrial gases (i.e. biogas treatment) has led to the investigation of industrial and agricultural wastes and sewage sludge as alternative sorbents. In terms of this, the group of Awe [
76] used solid wastes as adsorbents for H
2S removal from air and biogas. More specifically, they used CaCO
3–based solid wastes (CCW–D and CCW–S (a calcium carbonate–based form of solid waste, D and S indicate the number of the sample), commercial activated carbon, and pure commercial calcite as adsorbents for H
2S removal from air and biogas. Experimental tests were carried out in two phases, firstly with synthetic waste gas (dry air containing 200 ppmv H
2S) and secondly with synthetic biogas (CH
4: 64%, CO
2: 31%, H
2S: 200 ppmv) balanced with N
2 (approximately 5%) under room temperature and atmospheric pressure. It was observed that, in the first case, AC showed the highest adsorption efficiency toward H
2S, followed by CCW–D, CCW–S, and pure commercial calcite. Physisorption was promoted when using AC, whilst the acid–base reaction and catalytic oxidation was favored with calcium carbonate-based adsorbents. Unlike calcium carbonate–based wastes, AC gradually lost its reactivity. With respect to biogas mixture, only CCW–D and AC were studied. CCW–D rapidly lost its reactivity because no oxidation reaction occurred, due to the anoxic conditions in the biogas matrix used. Besides, AC again demonstrated increased performance for H
2S removal through physisorption, though its reactivity was decreased in the biogas matrix compared to the air matrix, which may be explained by the competitive adsorption of CO
2 and CH
4 present in the biogas matrix. The authors concluded that that the adsorption of H
2S was dependent on the porous structure and presence of alkali elements, which is because of the acidic nature of H
2S. Both solid wastes presented higher performance as opposed to the pure commercial calcite for H
2S removal. The pairing of AC and CaCO
3 led to better results in terms of sorption performance compared to pure commercial CaCO
3 and AC used alone.
Another study devoted to low–cost adsorbents for H
2S removal was reported by Skerman et al. [
77]. The authors, aiming at reducing H
2S concentrations in piggery biogas, tested and compared H
2S removal by a commercial iron–oxide H
2S scavenger (cg5) with alternative solid media i.e., granulated steel furnace slag (GSFS), red soil, compost, composted beef feed lot manure, granular activated carbon (GAC), and biochar. Experiments measured single-pass H
2S removal from a pre–humidified standard gas (2000 ppm H
2S in nitrogen) onto solid media contained in a cylindrical plastic column (DN 25 mm, depth 110–147 mm). Results showed that all carbon–based materials exhibited poor physisorption, leading to decreased H
2S adsorption capacity. On the other hand, the commercial medium (cg5) performed significantly better than the other media, achieving sulfur removal of 143 g S/kg medium at breakthrough (>10 ppm outlet H
2S). The red soil was regarded as the most promising alternative medium (2–12 g S/kg medium at breakthrough).
From an economical and technical perspective, petcoke is also considered as a competitive raw material for activated carbon production owing to its high carbon content, low volatile matter, and ash content. As has been extensively reported in the literature, KOH chemical activation is regarded as a prominent activation method for preparing highly porous activated carbon from different raw materials, including petcoke [
78]. In this regard, Mochizuki et al. [
79] used activated carbon fabricated from petcoke via KOH chemical activation (the KOH/petcoke weight ratio and activation temperature were varied in the ranges of 2–6 and from 873 K to 1073 K, respectively). Surface chemistry, pore structure, and adsorption capacity towards H
2S of the synthesized material were studied. It was found that rising activation temperature led to an increased specific area of activated carbon, and at the same time to a reduction in the acidic surface functional groups. Enlargement in the KOH/petcoke weight ratio favored the development of the pore structure. Regarding the H
2S adsorption, experiments were carried out at room temperature, showing that the adsorption of H
2S proceeded slowly. It was also reported that the adsorption capacity of H
2S was correlated with distinguishing features of pores of activated carbons.
Considering the recent momentum towards promoting cost–effective promising materials for H
2S removal, Asaoka et al. [
80] explored and tested carbonated steel slag in order to reveal the removal mechanism of H
2S by these materials and to quantify its removal performance. The authors reported that the reduction of MnO
2 in carbonated steel slag emerged at the same time as the decomposition of H
2S, whereby H
2S adsorption proceeded on the carbonated steel slag with an adsorption capacity of 7.5 mg/g. Results also showed that H
2S was captured via two reactions such as the formation of pyrite, and the oxidation to elemental sulfur combined with the reduction of manganese oxide on the carbonated steel slag. The promising functions of steel slag as an adsorbent for the removal of H
2S have also been reported by other studies [
81].
In a subsequent work by the same group [
82], granulated coal ash (GCA), a mixture of coal fly ash (CFA) and blast furnace cement (BFC), was employed in order to reduce H
2S concentration in water. This group demonstrated that GCA adsorbed and oxidized H
2S efficiently [
83]. The authors focused on optimizing the mixing ratio of GCA with a view to improving the rate of H
2S removal. Results showed that the highest removal rate of H
2S was detected for 87 wt% of coal fly ash owing to reduced adsorption competition between hydroxyl ions and sulfide.
As mentioned above, activated carbons can be derived by a variety of materials such as wood, peat, coal, coconut shells, fruit stones, and peels. Bazan-Wozniak et al. [
84] used pistachio nutshells in order to create carbonaceous sorbents for H
2S capture. The activated carbons were fabricated using physical, chemical, and direct activation, and their surface area ranged from 277 to 1204 m
2/g. According to the results reported, physical activation of pistachio nutshells could result in carbonaceous sorbents with high sorption capacity for acid gases.
Similarly, Zulkefli et al. [
85] used commercial coconut shell activated carbon (CAC) treated with zinc acetate (ZnAc2), copper sulfate (CuSO
4), potassium hydroxide (KOH), sodium carbonate (Na
2CO
3), and potassium iodide (KI) in order to enhance its selectivity towards H
2S removal. The authors used two types of gas mixtures (containing H
2S/N
2/CO
2 and H
2S/N
2), and the operating parameters for the H
2S adsorption varied in L/D ratio (0.5–2.5) and feed flow rate (1.5–5.5 L/min), corresponding to gas hourly space velocity (GHSV) (212.4–780.0 h
−1). Characterization results showed that the impregnated solution homogeneously covered the adsorbent surface and morphology. BET analysis further revealed that the treated sorbents surface area lessened by up to 96%. Among the tested sorbents, ZnAc2–CAC not only demonstrated the best performance ranging within 1.3–1.7 mg H
2S/g, but also exhibited “reuse potential” for the adsorption–desorption process in numerous cycles.
Very recently, Nam et al. [
86] employed an activated carbon derived from rice husks as an adsorbent for H
2S removal. The experimental results showed that the surface area of activated rice husk carbon as well as their adsorption capacity were largely dependent on carbonization temperature. Moreover, as is usually the case, when activated carbon was modified via impregnation with copper, the adsorption capacity of the Cu–treated material was increased, reaching its maximum efficiency at 450 °C, as it eliminated H
2S in 15 minutes. An interesting observation was that the regenerated Cu–treated activated carbon successfully re-adsorbed the H
2S gas, making this carbonaceous material a promising sorbent for the removal of H
2S.
Generally, the structure of activated carbons depends on the activation processes, conditions of carbonization, and type of precursor [
87]. Conventional heating is considered as the most extensively used method for obtaining activated carbons. Nevertheless, this technique is associated with certain limitations such as non–uniform heating, and thus increased pyrolysis and activation temperatures are required. In this context, the group of Kazmierczak–Razna [
88] applied radiation as a possible alternative for heating trying to obtain carbonaceous adsorbents from sawdust in order to remove H
2S from biogas. This method of conversion of electromagnetic energy into thermal energy is fast, occurs in the whole volume, and can be tuned. Hence, the authors used sawdust pellets and coniferous wood to fabricate (through pyrolysis at 400 °C using N
2) activated carbon and assessed its adsorption capacity for H
2S in dry and wet conditions (70% humidity). They verified that the adsorption capacity of activated carbon is reliant on the pH and activation procedure and can reach up to 4.1 mg H
2S/g
ads and 6.2 mg H
2S/g
ads in dry conditions, respectively.
Obviously, the adsorbent used in adsorption applications must exhibit high separation selectivity for the compound targeted (i.e., H
2S). In this context several studies have investigated the competitive adsorption of CO
2 and H
2S on the porous structure of the sorbent since they are both acid gases and concluded that this adverse effect can be partly eliminated by adding O
2 and humidity in the gas stream. In a recent and very promising study, De Oliveira et al. [
89] examined the selectivity of an activated carbon, fabricated by Brazilian babassu coconut hush, towards a H
2S/CO
2/CH
4 mixture as well as H
2S and CO
2 partition coefficients (K) at different pressures and temperatures. The data obtained were also correlated by Langmuir and Toth models, and multicomponent adsorption was predicted using the extended Langmuir, extended Toth, and ideal adsorption solution theory (IAST) methods. Results demonstrated that the activated carbon adsorbed roughly 26 wt% of H
2S or CO
2. Partition coefficients (K) values for H
2S and CO
2 reached more than 26 and 3, respectively, according to the predictive model applied, and were superior for diluted mixtures (high CH
4 content in the gas phase). Selectivity values for H
2S and CO
2 were greater than 25 for Toth and IAST. Moreover, selectivity toward CO
2 was approximately 5.6 times less than H
2S, which can be ascribed to a chemisorption phenomenon. The multicomponent results signaled S and K value degradation as the temperature rose. Results showed that the separation of H
2S and CO
2 from mixture containing CH
4 using this activated carbon as adsorbent is possible, especially for low H
2S and CO
2 concentrations and lower temperatures.
2.3. Metal Oxides
Metal oxides have exhibited high efficiency with respect to the desulfurization of coal gas, syngas, H
2 for fuel cells, and Claus tail gas etc., depending on their acid–base properties. However, pure metal oxides without a framework are characterized by low porosity, low surface area, and metal evaporation, and they are not able to resist sintering at higher temperatures, resulting in a reduced capacity for capturing H
2S [
90,
91]. For example, Ganwal et al. [
92] mentioned that the exposure of the metal oxides to temperatures over 500 °C led to the formation of iron carbide due to excessive reduction and limited operating conditions caused by decreasing H
2S adsorption reactivity. The general reaction between H
2S and a metal oxide is as follows (where M is the representative metal):
These limitations have led to mixed metal oxides in order to improve sulfur efficiency, prevent or delay the reduction to elemental metal, prevent sulfate formation (not reactive but occupies active material), improve dispersion and modify porosity and pore size, and impart structural strength and stability. Therefore, mixed metal oxides have gained in popularity in desulfurization processes owing to their wide range of operating temperatures (300−800 °C) and feasibility of compositional changes [
93]. Zinc–, manganese–, copper–, iron–, cobalt–, nickel–, chromium–, and calcium–based oxides are the most extensively studied oxides for desulfurization in solid−gas reactions due to their high affinity to S. Of note is the work of Westmoreland et al. [
94], who pioneered studies of metals and metal oxides for high temperature desulfurization. In particular, they screened the high–temperature desulfurization potential of 28 solids, mainly metal oxides, and came up with 11 candidate solids based upon the metals Fe, Zn, Mo, Mn, V, Ca, Sr, Ba, Co, Cu, and W, as these exhibited thermodynamic feasibility for high-temperature desulfurization.
Portela et al. [
95] showed that among different tested metal oxides, ZnO exhibited the highest equilibrium constant for sulfidation, reducing H
2S levels to fractions of 1 ppm. However, ZnO has certain limitations associated with the desulfurization process. In the highly reducing atmosphere of syngas it is partially reduced to elemental zinc, which is volatile above 600 °C, and thus a part of the sorbent may be lost. Hence, adding TiO
2 as a stabilizer is recommended when it comes to high-temperature applications. In contrast, for low operating temperatures such as in coal gas desulfurization, pure ZnO may be a candidate sorbent. A number of studies have investigated desulfurization using ZnO with different morphologies, including an interesting review by Samokhvalov [
96], but it is extremely difficult to provide a direct comparison between them owing to the different operating parameters used in each study, for example with respect to GHSV, gas-feed compositions, and definitions of breakthrough concentration.
In an early work, Jalan et al. [
97] studied the initial rate of sulfidation of ZnO at 650 °C and showed that the adsorption capacity was low as the high temperatures prompted an occlusion of the material’s pores. The authors tried to regenerate the metal oxide by means of a mixture comprised of steam and air at 650 °C, but the pore structure was damaged by this process. Davidson et al. [
98] published results on the adsorption of H
2S onto 300–710 μm particles of ZnO, studied in flow reactors at low H
2S concentrations and in the presence of H
2O. These reactions were conducted at ambient conditions (25–45 °C), and approximately 40% conversion of H
2S was noted. Nevertheless, the reactions were not performed to saturation, so the total percentage of ZnO converted to ZnS was not reported. Another study focusing on ZnO was reported by Sasaoka et al. [
99], who discussed the effects of the presence of CO, CO
2, and H
2O on the reaction. The authors concluded that CO inhibited the reaction by adsorbing onto the active sites, while H
2O inhibited the reaction by causing a back reaction with the ZnO to form H
2S and ZnO. Rodriguez et al. [
100] published a work in which a direct comparison between the adsorption of H
2S on ZnO and MgO was attempted. The results showed that the reaction of ZnO toward H
2S was more effective. The same group compared their results to previously studied Al
2O
3, Cr
2O
3, Cr
3O
4, and Cu
2O and found a trend between the band gap energy and chemical reactivity (availability to adsorb H
2S). In particular, they investigated H
2S adsorption on different metal oxide frameworks by means of synchrotron-based high-resolution photoemission and argued that H
2S principally interacts with the metal centers of the oxides. Moreover, they also stated that at nearly ambient temperature (26.85 °C), it undergoes complete decomposition [
101,
102]. The authors concluded that for these systems, the smaller the band gap in the oxide, the greater its reactivity over S–containing species. The theoretical results of ab initio self consistent field (SCF) calculations and experimental results suggested that the oxide’s reactivity mostly depended on the degree of the interaction between its band and the orbitals of H
2S or HS. Oxides with high ionic character, such as Al
2O
3 (band gap ~9 eV) and MgO (band gap ~6.5 eV), exhibited low reactivity in relation to H
2S, illustrating that the electrostatic interactions taking place between the dipole of H
2S and the ionic field being generated by the charges in an oxide were of subordinate importance in the adsorption process. On the other hand, Cr
3O
4 (band gap ~0 eV) interacted strongly with H
2S, CuO, and Cu
2O. Therefore, it was found that the lower the band gap energy, the more H
2S was adsorbed. Another effort in drawing comparisons of the adsorption of H
2S on ZnO and other metal oxides was reported by Carnes et al. [
103], who investigated the adsorption of H
2S on microcrystalline metal oxides of Zn, Ca, Mg, and Al at high (250 ~ 500 °C) and low temperatures (25 ~ 100 °C) and they concluded that at elevated temperatures, CaO was the best choice for H
2S adsorption and ZnO appeared to be superior at temperatures lower than 100 °C.
In a more recent study, Rosso et al. [
104] reported an interesting contribution devoted to the impact of calcination temperature on H
2S adsorption capacity of the ZnO sorbent. The authors tested several pure ZnO materials prepared with different methods and calcined at different temperatures (ZnO
600, ZnO
500, ZnO
400, ZnO
300) as low–temperature desulfurizing sorbents for syngas. The sulfidation performance was evaluated in a fixed–bed reactor by breakthrough curves at 250 °C. ZnO
300 and ZnO
400 sorbents, calcined at the lowest temperatures and hence with the highest specific surface area (SSA) and pore volume, exhibited the highest capacity in capturing sulfur, with breakthrough times varying from 430 to 530 min measured at a space velocity of 105 h
−1. The ZnO
400 sorbent, due to its higher structural stability, was identified as the most effective candidate for application in syngas purification. The authors evaluated the regenerability of the sorbent by subsequent sulfidation–thermal regeneration cycles (for six in particular). They found that the ZnO
400 sorbent recovered about 50% of its initial sulfur–capturing capacity following the first regeneration and about 25% after the second regeneration, while it stabilized to a value of about 15% from the third to the sixth regeneration. A numerical model was developed and validated using the experimental data of the ZnO
400. The ZnO
400 sulfidation rate was limited by the internal mass transfer resistance (which results in the rate determining step (RDS) of the process), due to the difficulty of for H
2S molecules to reach the adsorbent core, while the impact of the external mass transfer became rapidly irrelevant. Therefore, only the external layers of the sorbent participated in the sulfidation process.
However, despite the fact that the ZnO desulfurization reaction has been extensively investigated, little is known in terms of the transformation at the crystal scale, its detailed mechanism, and kinetics. Most works found in the literature presume that the diffusion of H
2S on ZnO is inward; however, Neveux et al. [
105] argued that the growth mechanism was an outward one. The authors noticed the presence of voids in the interior of ZnS particles, indicating Zn
2+ and O
2− diffusion from the ZnO–ZnS internal interface to the surface of the ZnS particle. Nevertheless, the formation of voids at the internal interface led to a decrease in the ZnO–ZnS internal interface and to a strong slowing down of the reaction rate while the external interface increased. Thereby, the authors suggested that diffusion through the ZnS layer was probably the rate limiting step. They also suggested a sequence of steps for the sulfidation reaction that included the outward growth of ZnS and diffusion of Zn and O
2 through ZnS. A rate expression was obtained that took into account the void evolution by means of random nucleation. The authors were able to show that the model predictions were in good agreement with the experimental rates obtained. A schematic representation of an inward and an outward growth mechanism is shown in
Figure 5 [
106].
Apart from ZnO–based adsorbent, CuO–based adsorbents have also received considerable attention in recent decades due to their high equilibrium sulfidation constants, which can allow extremely low-equilibrium H
2S concentration [
107]. However, CuO is readily reduced to metallic copper in the presence of CO and H
2 in the fuel, downgrading its desulfurization efficiency [
108,
109].
In a recent study, Pola-Albores et al. [
110],compared ZnO and Cu–based adsorbents for biogas desulfurizaion at room temperature. Powder–X ray diffraction and X–ray fluorescent spectroscopy analysis showed a unique zincite phase for ZnO and three phases in copper samples: Cu
0, Cu
2O, and CuO (76.3, 18.3, and 5.4 wt%). Experiments were conducted in a fixed bed glass reactor using a simulated gas mixture containing N
2 and H
2S (100 ppm) measured at a space velocity of 9120 h
−1. Both sorbents became saturated after 5 min, while the presence of metallic sulfides at amounts of less than 0.05 wt% was observed. An interesting point was that following gas–phase desulfurization, desorption tests did not reveal the presence of H
2S, indicating that the adsorbed compound was chemisorbed, forming solid metallic sulfide phases.
However, the growing demand for low temperature desulfurization for environmental and economic reasons is driving the search for more efficient solutions. In this regard, Xue et al. [
111] prepared a series of mixed oxide adsorbents of Zn containing Fe, Ni, Co, Mn, Cu, Al, Ti, and Zr; the H
2S adsorptive capacity of ZnO could be enhanced considerably by the doping of these metal ions. An important contribution was reported by Skrzypski et al. [
112], who found that the breakthrough capacity of Cu–doped ZnO was higher than that of pure ZnO. More specifically, they probed mixed Cu–Zn oxides (Cu
0.03Zn
0.97O) and found that at 200 °C the breakthrough capacity of the sample containing 6 mol% Cu (Cu
0.06Zn
0.94O) was six times higher than that of pristine ZnO. Moreover, there was an absence of voids on the Cu–doped ZnO, which is a good indication that the growth mechanism was an inward one, in which the anions of sulfur diffused from the external surface to the internal ZnO/ZnS interface and exchanged with active oxygen species (O
2–). This was in contrast to pure ZnO, which at 250 °C showed the formation of voids at the interface of ZnO/ZnS. These vacancies resulted in the diffusion enhancement of sulfur and thus, in increased conversions of Cu–Zn oxides. XRD analysis revealed that the ZnS crystallite size remained unchanged during the reaction. On the other hand, the presence of copper increased the mobility of sulfur anions in Cu–containing ZnS particles. Several other studies [
113,
114] also attributed the positive role of copper in improving the conversion degree of ZnO to reduced diffusional resistance in the ZnS product layer with a transition from an outward (for pure ZnO) to an inward (for Cu–doped ZnO) growth mechanism.
Very recently, Cimino et al. [
115] studied the reactive adsorption of H
2S from gaseous streams at room temperature (30 °C) and 500–3000 ppm H
2S concentration (N
2 carrier gas) using a mixed Zn–Cu active phase supported onto γ–alumina (Zn–Cu/Al) in order to highlight the role of the support and the possible synergic effect between the two metals on the adsorption mechanism. XRD, diffuse reflectance infrared fourier transform spectroscopy DRIFTS, and temperature programmed desorption (TPD) NOx studies on fresh sorbents calcined at 250 °C showed that the active phase contained both metal hydroxides and hydroxynitrates, owing to the alumina support acting as a stabilizer preventing the complete decomposition of metal nitrates precursors during the calcination stage. The deposition of Cu–Zn on the support determined a negligible decrease of the pore size distribution and BET surface area in proportion to the weight increase, testifying to the high dispersion of the active face, that, in contrast, considerably increased the H
2S adsorption capacity of the bare support. The interaction between Zn and Cu entailed a gradual enhancement in the H
2S adsorption capacity, which, on the contrary, was barely affected by the inlet H
2S concentration. The sorbent exhibited its best capacity (0.82 mmol g
−1) in trapping H
2S at the maximum copper content (Cu
0.5Zn
0.5/Al) under both dry and wet environments. Moreover, the authors attempted to draw comparisons between the sorbent (Cu
0.5Zn
0.5/Al) and a respective material with similar nominal active phase composition, supported on activated carbon (Cu
0.5Zn
0.5/AC). The Cu
0.5Zn
0.5/AC exhibited an immensely higher breakpoint time and adsorption capacity at saturation compared to Cu
0.5Zn
0.5/Al. This behavior is most likely attributable to the significantly larger surface area of AC promoting dispersion of the Cu–Zn nanoparticles, and to the fixation of metal hydroxynitrates on the surface of the alumina support, which were less reactive compared to the corresponding oxides formed on the activated carbon.
However, inorganic supports such as alumina and silica usually offer the advantage of high thermal stability and hence potential regenerability after sulfidation/regeneration cycles, frequently carried out under oxygen/air streams [
116]. Elyassi et al. [
113] reported that a sequence of oxidation (in a 5%/95% O
2/N
2 mixture at 400 °C) followed by reduction (in a 10%/90% H
2/He mixture at 400 °C to reduce CuSO
4 species) of a spent silica-supported (Cu–ZnO–SBA–15) sorbent was efficient in fully restoring the capacity of the sorbent in trapping H
2S. On the other hand, Wang et al. [
117] noticed that the CuO–SiO
2 3DOM sorbents, whose behavior on adsorption applications will be further analyzed later, exhibited poor H
2S capture capacity at ambient temperature when performing the regeneration stage at 650 °C in air, owing to the agglomeration of CuO upon exposure to high temperature. Therefore, the issues associated with the sorbent’s regenerability has to be evaluated depending on different factors including the nature of the substrate/sulfation products, their chemical interaction, and possible sintering phenomena.
The preparation of metal–supported catalysts by co-precipitation (cp) or multi–precipitation (mp) was also investigated. Jiang et al. [
118] examined the performance of Cu–Zn and Cu–Zn–Al synthesized by means of cp and mp and found that adsorbents with high Cu concentrations were more efficient in capturing H
2S compared to adsorbents with high Zn concentrations. As an example, at 80 °C, mp–Cu
80Zn
10Al
10 showed almost complete conversion. The etiology for the increased efficiency of Cu–rich adsorbents was the faster sulfidation rate of CuO as compared to ZnO.
In general, mixed metal oxides that were prepared using the co–precipitation method exhibited higher SSAs, smaller particle sizes, and improved H
2S capacity in comparison to their counterparts that were prepared by simple impregnation [
119]. Baird et al. [
120] also probed H
2S adsorption on a Co–Zn–Al–O that was synthesized using co-precipitation. As expected, the presence of alumina led to a higher SSA, but the doped materials were not able to increase their capacity in trapping H
2S. Notwithstanding, mixed metal oxides (Co–Zn) with different Co/Zn ratios exhibited positive correlations between the increase in percent utilization and the increase in the surface area. Generally, Co doping leads to increased SSA by decreasing the size of the particle. From a thermodynamic perspective, ZnO has more positive free energy for sulfidation in comparison with Co
3O
4 [
121].
Westmoreland et al. [
94] also defined the initial rates for the reaction between H
2S and MnO, ZnO, CaO, and V
2O
3 over a temperature range of 300 to 800 °C, using a thermos-balance reactor. It was shown that at temperatures between 300 and 800 °C the intrinsic reaction rate of H
2S with ZnO was almost 10 times less than that of MnO, despite the fact that the SSA of ZnO was almost 10 times higher than that of the MnO.
Zn metal oxides doped with Fe, Ni, Co, Mn, Cu, Al, Ti, and Zr exhibited higher conversions at ambient temperature, notably for Zn
2.9MnO
4.2. It is worth mentioning that pure manganese tetroxide (Mn
3O
4) is associated with poor conversion, but the reasons for this behavior remain unclear [
111].
Polychronopoulou et al. [
122] investigated the regeneration capacity of Zn–Ti–based mixed metal oxides and concluded that its doping with Mn, Cu, or Mo resulted in an enhancement in H
2S capture for the regenerated material, notably 41 mg/g for 10Mn–45Zn–45Ti–O (600 ppm H
2S in the feed). Moreover, the addition of 1% H
2O in the gas mixture resulted at a considerable increase in capacity (160 mg/g).
Liu et al. [
123] studied CuO–based mixed oxides that were prepared by co–precipitation (
Table 4) as adsorbents for H
2S removal and noted increased capacities for oxides that were doped by Fe. It was shown that CuO–based adsorbents were able to remove H
2S from a CO
2 stream to less than 0.1 ppm at 40 °C. Capacities were higher for H
2S in N
2/H
2S gas mixture (1000 ppm H
2S) in comparison to a CO
2/H
2S gas mixture (1000 ppm H
2S), indicating the competitive adsorption between these acid gases. Doped CuO (Fe–Cu–Al–O) exhibited the highest breakthrough capacity of 113.9 mg g
−1 at 40 °C. In the temperature range of 40–130 °C, the breakthrough capacity of the modified CuO increased with adsorption temperature, which nonetheless would have been below 100 °C in the presence of CO
2 owing to the possibility of generating COS impurities. As expected, at lower space velocities, increased breakthrough capacities were observed. It was also reported that the inhibitory effect (CO
2 inhibits the adsorption of H
2S) was higher for pristine CuO in comparison to doped CuO. Five cycles of adsorption and regeneration were performed using CuO, Cu–Al–O, and Fe–Cu–Al–O at 100 °C in air. Results showed that only CuO was able to maintain its regeneration capacity (75% of the capacity of the fresh adsorbent). The latter doped materials showed reduced regenerated capacities and became constant during the last two cycles at 50% and 43% of the initial value, respectively.
Jung et al. [
124], with a view to enhancing the poor regeneration properties of the ZnO–Al
2O
3 (ZA) sorbent featuring high sulfur removal capacity and fast H
2S adsorption rate, added 5–10 wt% of metal oxides such as Fe
2O
3, Co
3O
4, NiO, and CeO
2. The method used for the preparation of the sorbents was co–precipitation, whilst their trapping capacities and regeneration properties were evaluated in a fixed–bed reactor during multiple regeneration–sulfidation cycles at temperatures between 480 and 580 °C. All the sorbents exhibited sufficient sulfur removal (0.17–0.20 g S/g) and regeneration properties, which were maintained even after 10 cycles of sulfidation and regeneration, as compared with the conventional ZnO–Al
2O
3 and ZnO–TiO
2 prepared by the physical mixing method. Fe–, Co–, and Ni–containing adsorbents formed their aluminates, which did not transform to sulfide during sulfidation. The additives/aluminates strongly promoted the transformation of S contained in ZnS into SO
2. Cerium dioxide exhibited similar behavior to aluminates in terms of the sulfur oxidation.
The studies mentioned above clearly demonstrate that it is extremely difficult to achieve high regeneration capacities for mixed metal oxides. Thus, further research in this area should concentrate not only on increasing the capacity of these materials in capturing H2S but also on enhancing regenerability issues.
Another very interesting study in the context of Fe
2O
3 desulfurization properties was reported by the group of Thanakunpaisit [
125] who studied the removal of H
2S contained in biogas using Fe
2O
3–rich laterite materials as adsorbents, which promote potential reactivity with H
2S. Initially, to predict the interaction among five principle compositions of biogas (CH
4, H
2S, CO
2, and N) and one composition of laterite (Fe
2O
3), the density functional theory with the B3LYP/6–31G(d,p) level was applied. Study results depicted that energy adsorption between Fe
2O
3 and H
2S was –26.06 kcal/mol, which was higher in comparison to that of other compositions of biogas and Fe
2O
3 due to the strong affinity between the sulfur atom of H
2S and the iron atom of Fe
2O
3. Afterward, the adsorption of H
2S and CH
4 on laterite was investigated. Adsorption efficiency values for H
2S and CH
4 were 91.67% and 1.84%, respectively.
Thermodynamic investigations also concluded that sorbents doped with transition metal oxides such as oxides of Co, Cu, Mn, Cr, Fe, Ni, and W are potentially attractive replacements for low temperature desulfurization [
126,
127,
128]. Among these, Co–doped sorbents are thought to have the highest degree of H
2S removal rates and subsequent metal sulfide conversions.
The group of Pahalagedara [
129] studied mesoporous cobalt oxides calcinated in different temperatures (150, 250, 350, or 450 °C) as desulfurization sorbents and found that they may attain 100% capacity at 150 °C. A significantly high sulfur sorption capacity was also noted at room temperature (13.4 g S/100 g sorbent) as well as in the temperature range of 175–250 °C (65.0–68.9 g S/100 g sorbent). According to the literature, even though the sulfidation at low temperatures is kinetically slow, the mesoporous cobalt oxide materials showed high sulfur capacities and reached high oxide to sulfide conversions. The high breakthrough capacities recorded are probably due to H
2S physical adsorption. Other mesoporous materials (Cr
2O
3, CuO, Mn
2O
3, and Fe
2O
3) also showed superior H
2S removal capacity compared to their nonporous analogues (as high as 200 times).
Van Nisselrooya et al. [
130] used a catalyst comprised of chromium and iron oxides supported on α–alumina in order to enhance the thermodynamically limited Claus process, introducing the term “Superclaus catalyst”. Nevertheless, due to the toxic effects of chromium, further research into the development of new catalysts for this reaction is needed [
131].
Several works have investigated vanadium–based catalysts supported on different metal oxides for H
2S abatement in the temperature range of 50–250 °C. It was reported that the catalyst supported on CeO
2 exhibited the higher catalytic performance at 150–250 °C with a moderate selectivity to SO
2 (13%) [
132]. Having said that, Palma et al. [
133] investigated vanadium–based catalysts supported on ceria for the direct and selective oxidation of H
2S to sulfur and water at low temperature. Two different vanadium-grafted catalysts (20–50 wt% of V
2O
5) were fabricated, and tested at temperatures varying from 150–200 °C with a view to identifying the optimum catalytic conditions. The most effective catalyst was that with the 20 wt% of V
2O
5, which exhibited 99% sulfur selectivity and equilibrium H
2S conversion at different operating conditions. Palma et al. [
133] also studied the catalytic performance in the presence of three principle compositions of biogas stream (CH
4, CO
2, and H
2O) at 150 °C, in an attempt to investigate the possible formation of by–products such as COS and CS2. Results revealed that the presence of CO
2 and CH
4 had no impact regarding the formation of both SO
2 and undesired secondary products. Moreover, the authors studied the effect of the H
2S inlet concentration, contact time, molar feed ratio of O
2/H
2S, and temperature at 80 °C. The results obtained by changing the H
2S concentration in the feed stream verified that low concentrations (H
2S = 80–100 ppm) led to a complete H
2S conversion at 100 °C, yet with increased SO
2 selectivity. In the presence of higher H
2S concentrations (200 ppm) no catalyst deactivation occurred, while the deposition of the sulfur on the catalyst surface resulted in clogged pores of the carrier. The temperature at low contact time affected to an important extent the catalytic activity and catalyst stability, while strong deactivation and low H
2S conversion were detected at 80 °C as a result of the low reaction temperature that did not favor the reaction from a kinetic perspective. At the end, a stability experiment was performed at the minimum temperature (80 °C) and at the maximum contact time (240 ms) by using a specific molar feed ratio (O
2/H
2S = 1). Results showed a slight deactivation starting from 20 h of time on stream, probably owing to sulfur blocking the pores of the CeO
2. It is worth mentioning that when performing the same experiment at a higher temperature of 120 °C no deactivation occurred.
Researchers also investigated H
2S adsorption in three-dimensionally ordered macroporous (3DOM) structures. Recently, Fan et al. [
134] synthesized an iron oxide sorbent with a structure composed of three-dimensionally ordered macropores (3DOM). The 3DOM media feature high porosity with interconnected macropores, large surface area, and nanosized active particles, and represent an alternative at medium–temperature desulfurization applications. Furthermore, stable γ–Fe
2O
3 proved to be easily prepared via a colloidal crystal templating method when silicon was added to the precursors.
Huang et al. [
135] studied H
2S adsorption using iron oxide γ–Fe
2O
3 and silica-supported iron oxide γ–Fe
2O
3/SiO
2 (3DOM–FS) with different Fe:Si molar ratios from 1 to 16, at low temperatures (20–80 °C) in the presence and/or absence of humidity. Then, the performance of these materials was compared with that of a–Fe
2O
3 and the commercial sorbent HXT–1 (amorphous hydrated iron oxide). Adding the amorphous and highly dispersed silicon in γ–Fe
2O
3 improved the dispersion of Fe
2O
3 and hampered its transformation to α–Fe
2O
3 during calcination. The activity of γ–Fe
2O
3 was higher than that of HXT–1 at temperatures over 60 °C, whereas a–Fe
2O
3 had little activity. The vacancies in the lattice of γ–Fe
2O
3 promoted the ion diffusion. Moreover, the reference γ–Fe
2O
3 featuring lower density, larger surface area, and smaller particle size than that of α–Fe
2O
3 exhibited increased performance in trapping H
2S as compared with the reference α–Fe
2O
3, but lower performance when evaluated against 3DOM–FS. Decreased capacity at high Fe:Si ratios was ascribed to the fact that α–Fe
2O
3 was present. The authors also reported that relative humidity (approximately 10%) had a positive role in H
2S capture. Moreover, it was reported that at 300 °C, conventional regeneration by means of 5% O
2 in N
2 led to a significant drop in the capacity of the regenerated sorbent. As an alternative regeneration technique, the authors suggested the use of 5% O
2 in the feed stream and showed that desulfurization and regeneration could occur synchronously; this resulted in substantially increased breakthrough capacities (79.1% from 40.61%) in comparison to those achieved in the absence of oxygen from the gas mixture. It is not quite clear whether this alternative regeneration method promotes H
2S oxidation to SO
2. Monitoring SO
2 concentration may provide evidence of the presence or absence of sulfur during this process.
As already mentioned, zinc oxide–based materials have exhibited promising desulfurization properties. ZnO has the ability to react with H
2S, leading to the formation of zinc sulfide by means of an exothermic reaction. Accordingly, decreased temperatures give rise to higher desulfurization efficiency. Wang et al. [
136] synthesized, via a colloidal crystal template method, a number of ZnO silica-based 3DOM structures with different weight ratios of ZnO (3D–SZx–500, where x (ZnO weight ratio) = 20%, 50%, and 73%) with a view to capturing H
2S at ambient temperature. The synthesized materials exhibited a well–ordered 3D–interconnected network consisting of macropores, mesopores, and micropores. It was reported that both the Zn/Si ratio and calcination temperature strongly affected structural parameters. Moreover, the surface areas and pore volumes reduced with increasing the Zn/Si ratio, presumably owing to the agglomeration of ZnO species in the pores. The same decreasing effect was observed for the surface areas and pore volumes at high calcination temperatures (>500 °C) due to the shrinkage of the channels. Even lower calcination temperatures (400 °C) led to the production of carbon impurities, which affected the ZnO surface and occluded the pore channels, resulting in reduced sulfur adsorption capacity. Furthermore, the breakthrough time increased with increased ZnO content. In particular, 3DOM sorbents exhibited the highest utilization (69%) and sulfur capacity (135 mg g
−1) with 50 wt% of ZnO. It was also reported that moisture could have a positive effect on the removal of H
2S.
Another work in the context of 3DOM structure H
2S adsorbents was reported by the group of Wang [
137] who prepared two types of cobalt oxide-silica composites via the sol-gel method using different alcohols (n–butyl alcohol and ethylene glycol) as sol precursors. The respective adsorbents with the 3DOM structure were also synthesized by means of the colloidal crystal template method. In order to understand the physicochemical properties of the materials, several characterization techniques were used, while the efficiency of the sorbent in capturing H
2S was evaluated at room temperature. It was shown that use of n–butyl alcohol in sol precursors could result in a sorbent (SCN57–500) featuring a very large surface area (314.5 m
2/g) and ordered mesoporocity, with a diameter of 2–5 nm in grains. However, it exhibited extremely decreased capacity in capturing H
2S (12 mg/g). On the other hand, (SCE57–500) had a smaller surface area (137.2 m
2/g) but well–distributed particles, resulting in increased H
2S breakthrough capacity of 28 mg/g. It was reported that use of ethylene glycol could benefit the formation of 3DOM materials. Unlike 3D–SCN57–500 (159.5 m
2/g), the surface area of 3D–SCE57–500 was considerably enhanced (257.4 m
2/g). The decreased surface area of 3D–SCN57–500 was due to the sintering of the sorbent’s grains, which affected the pore volume. It can be deduced that the increased surface area of 3DOM composites was attributed to the high dispersion, i.e., the newly formed puny nanoscale grains. The desulfurization performance of 3DOM composites was strongly affected by the calcination temperature as well as the Co
3O
4 content of the material. The 3DOM sorbents achieved higher capacities for trapping H
2S compared to their respective bulk counterparts. Study results suggested that the crystalline structure or dispersion contributed to the adsorbent’s reactivity, along with the interconnected and well–ordered macropores. Despite the fact that sulfidation at low temperature was kinetically slow, the breakthrough sulfur capacity of 3D–SCE57–500 could reach as high as 189 mg/g with Co
3O
4 utilization of 63% at room temperature. The 3D–SCE57–500 showed increased performance compared with the reference mesoporous cobalt oxide. Using XRD and selected area (electron) diffraction (SAED) analysis, the presence of sulfide cobalt in amorphous form was observed, and regarded as CoS by XPS and thermogravimetric–mass spectrometric TG–MS analysis. Moreover, XPS and TG–MS analysis was carried out for 3D–SCE57–500S and the involved reactions for H
2S removal by Co
3O
4 in the discussed experimental conditions. The results exhibited the presence of at least five distinct sulfidation products (CoS, CoSO
4, CoSOH, CoOOH, and elemental sulfur) that were produced during the sulfidation and storage process.