Review

pubs.acs.org/CR

Hydrogen Sulfide Capture: From Absorption in Polar Liquids to

Oxide, Zeolite, and Metal−Organic Framework Adsorbents and
Membranes
Mansi S. Shah,† Michael Tsapatsis,† and J. Ilja Siepmann*,†,‡
†
Department of Chemical Engineering and Materials Science, University of Minnesota, 421 Washington Avenue SE, Minneapolis,
Minnesota 55455-0132, United States
‡
Department of Chemistry and Chemical Theory Center, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota
55455-0431, United States

ABSTRACT: Hydrogen sulfide removal is a long-standing economic and environ-

mental challenge faced by the oil and gas industries. H2S separation processes using
reactive and non-reactive absorption and adsorption, membranes, and cryogenic
distillation are reviewed. A detailed discussion is presented on new developments in
adsorbents, such as ionic liquids, metal oxides, metals, metal−organic frameworks,
zeolites, carbon-based materials, and composite materials; and membrane technologies
for H2S removal. This Review attempts to exhaustively compile the existing literature on
sour gas sweetening and to identify promising areas for future developments in the field.

CONTENTS 4.1. Polymeric Membranes 9785

4.2. Membranes for Gas−Liquid Contact 9786
1. Introduction 9755 4.3. Ceramic Membranes 9789
2. Absorption 9758 4.4. Carbon-Based Membranes 9789
2.1. Alkanolamines 9758 4.5. Composite Membranes 9790
2.2. Methanol 9758 5. Cryogenic Distillation 9790
2.3. N-Methyl-2-pyrrolidone 9758 6. Outlook and Perspectives 9791
2.4. Poly(ethylene glycol) Dimethyl Ether 9759 Author Information 9792
2.5. Sulfolane and Diisopropanolamine 9759 Corresponding Author 9792
2.6. Ionic Liquids 9759 ORCID 9792
3. Adsorption 9762 Notes 9792
3.1. Metal Oxides 9762 Biographies 9792
3.2. Metals 9767 Acknowledgments 9792
3.3. Metal−Organic Frameworks 9768 References 9792
3.4. Zeolites 9772
3.4.1. All-Silica Zeolites 9772
3.4.2. Aluminosilicate Zeolites 9773
3.4.3. Titanosilicates 9775 1. INTRODUCTION
3.5. Carbon-Based Adsorbents 9775 Hydrogen sulfide (H2S) is a colorless gas with a characteristic
3.5.1. Activated Carbons 9775 odor of rotten eggs. It is highly toxic, corrosive, and flammable
3.5.2. Structured Carbons 9775 in nature. Since H2S is heavier than air, it has a tendency to
3.6. Composite Materials 9778 settle in low-lying and poorly ventilated areas (see Table 1 for
3.6.1. Metal Oxides/Mesoporous Silica 9778 the properties of H2S). It causes irritation to the eyes, nose, and
3.6.2. Supported Amines 9778 throat at concentrations as low as 5 ppm and is fatal over 1000
3.6.3. Zeolite/Polymer 9780 ppm.1 Inhalation is the primary route for H2S exposure. At ∼30
3.6.4. Metal−Organic Framework/Graphite ppm, H2S paralyzes the sense of smell and is rapidly absorbed
Oxide 9780
3.6.5. Metal (Hydr)oxide/Graphite Oxide 9782 Special Issue: Carbon Capture and Separation
3.6.6. Metal Oxide/Activated Carbon 9782
4. Membrane Separation 9782 Received: February 14, 2017
Published: July 5, 2017

© 2017 American Chemical Society 9755 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 1. Some Physical Properties of H2S that can, in turn, give rise to acid rain. Hydrodesulfurization is
also important since sulfur compounds, often at very low
property value
concentrations, can deactivate noble metal catalysts for
molar mass 34.081 g/mol downstream catalytic reforming units. Gasification of coal and
specific gravity 1.2 biomass for syngas (CO + H2) production, under high
lower explosive Limit (LEL) 4.5%
temperature and low oxygen concentrations, results in a
upper explosive limit (UEL) 46%
conversion of the sulfur content of these resources into H2S.
normal boiling point 212.9 K
For any of the hydrocarbon streams, such as natural gas,
melting point 190.9 K
biogas, or syngas, to be useful for energy generation or
critical temperature 373.4 K
chemicals production, they need to be purified to remove
critical pressure 89.7 bar
unacceptable amounts of acid gases such as H2S and CO2. The
critical density 0.349 g/cm3
kinetic diameter 3.6 Å
acceptable H2S levels can be quite different depending on the
dipole moment 0.97 D
end use; for example, 5−15 ppm H2S is acceptable in natural
odor rotten eggs
gas depending on the country of use, but <1 ppm H2S is
appearance colorless gas required for fuel cell applications, where the electrodes are very
threshold limit value (TLV) 1 ppm susceptible to sulfur poisoning. The magnitudes of different
immediate danger limit 100 ppm streams that are processed globally are also quite different: for
instance, over 120 trillion standard cubic feet (scf) of natural
gas is produced each year, while only about 1 trillion scf of
by the blood, resulting in poor oxygen uptake at the cellular biogas is produced annually. The optimal materials and
level. Therefore, one should not fully trust their sense of smell processes for H2S removal from different sources and for
for the presence of H2S because alarming concentrations may different end applications can be quite different. In this Review,
be present with no perceivable odor. Handling of H2S requires we will focus more on the technical aspects of the different
stringent safety regulations such as trained personnel, complete
alternative methods and processes. Detailed economic analysis
set of personal protective equipment including a gas mask,
of competing H2S-removal processes for specific feeds and
strategically located sensors, and emergency evacuation plans.
applications is beyond the scope of this Review. While there is a
Experimental studies with H2S at concentrations of a few
hundred parts per million are challenging enough; the challenge significant overlap in materials for different H2S-capture
is even greater at the percentage-level concentrations in actual problems, we will base our discussions mainly in the context
gas-processing facilities. of natural gas purification.
H2S occurs naturally, at varied concentrations, in various Raw natural gas is a complex mixture of methane (CH4),
hydrocarbon sources such as crude petroleum, natural gas, and ethane, and other light alkanes, H2S, carbon dioxide (CO2),
biogas (see Table 2). It is also formed by the hydrolysis nitrogen, and water (H2O) vapor. The U.S. Environmental
reaction of metal sulfides present in volcanic rocks. Anaerobic Protection Agency classifies natural gas as sour at H2S
respiration of sulfate-reducing bacteria, whereby these organ- concentrations above 4 ppm;5 sweetening refers to removal of
isms breathe sulfate to derive energy and release sulfide as acidic sulfur compounds, primarily H2S. Natural gas emerging
waste, constitutes another major natural source of H2S. at the reservoir well head is subjected to low temperatures to
Landfills, rotten vegetation, and composting sites can also condense out heavier hydrocarbons (see Figure 1). Overhead
contribute to H2S generation; caution should be exercised in vapors are sent to the acid gas removal unit to selectively strip
dealing with these not-so-obvious sources of H2S. off H2S and CO2; amine-based absorption is most commonly
Anthropogenically, the main source of H2S is in oil refineries, used for this step. The H2S-rich stream is sent to the sulfur
wherein desulfurization of various gas streams using hydrogen recovery unit (SRU), while the CH4-rich stream, after some
gas generates H2S (also known as hydrodesulfurization). This post-processing steps such as dehydration, is sent to the
process is essential to reduce the sulfur content of automobile pipeline as sales gas. In the SRU, sulfur is recovered by the well-
and heating fuels, since combustion of sulfur compounds will known Claus process, where H2S undergoes high-temperature
result in emission of harmful sulfur oxides to the atmosphere (∼1000 °C) thermal oxidation,

Table 2. Properties and Compositions of Raw H2S-Containing Gases from Different Sources (Data from Refs 2−4)

natural gas biogas landfill gas syngas

Ekofisk, Panhandle, Wustrow, Süd-Oldenburg, Tenguiz, anaerobic coal
parameter unit Norway US Germany Germany USSR digestion gasification
calorific MJ/N·m3 40 36 16 30 70 21−25 12−23 8−10
value
CH4 mol% 85 73.2 43 77 42 60−70 35−65 0−5
C2+ mol% 12.5 11.5 0.7 0.1 39
CO2 mol% 2 0.3 0.3 8 2.6 30−40 15−40 5−15
H2S mol% 0.001 8 16 0−0.4 0−0.01 0−0.01
N2 mol% 0.4 14.3 56 7 0.8 0−0.5 15
H2 mol% 25−30
CO mol% 30−60
NH3 ppm 100 5
He mol% 0.7 0.04

9756 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

In addition to other papers in this special issue, Chemical

Reviews has published three papers on the subject of CO2
capture in the past 5 years: Sumida et al. reviewed CO2 capture
in metal−organic frameworks (MOFs),7 Pera-Titus reviewed
the capture process using porous inorganic membranes,8 and
Sanz-Pérez et al. focused on chemical sorbents for the direct
capture of CO2 from air.9 However, similar reviews on H2S
capture are few and limited in scope.10−16 Abatzoglou and
Boivin reviewed biogas purification processes for removal of
contaminants such as H2S, NH3, and siloxanes.11 They divided
the purification methods into two generic categories: those
involving physicochemical phenomena and those involving
biological processes. Most biotechnologies still require further
optimization and scale-up in order to compete with their
Figure 1. Block diagram for natural gas processing.
physical/chemical counterparts.17 We will focus here on the
physicochemical techniques for selective H2S removal; the
H 2S + 1.5O2 → SO2 + H 2O reader interested in microbiological alternatives is referred to
−ΔH = 518−576 kJ/mol of H 2S (1) earlier reviews on the subject.18,19 The abundant literature
available on H2S removal may be broadly classified into four
prior to low-temperature (200−300 °C) catalytic oxidation, categories: absorption, adsorption, membranes, and cryogenic
distillation (see Figure 2). In this Review, we focus on materials
2H 2S + SO2 ⇄ 3S + 2H 2O and chemistries involved in each of these processes. Extensive
−ΔH = 88−146 kJ/mol of SO2 review articles focus on solutions to tackle concentrated H2S by
(2)
methods like the Claus process20−22 or the production of
(−ΔH data from ref 6), in a series of reactors at progressively hydrogen from H2S,23 and these are not considered here.
lower temperatures, accompanied by sulfur removal in Since H2S is a polar molecule, while CH4 is a nonpolar
intermediate condensers. Since the melting point of elemental molecule, it is obvious to imagine a polar solvent or a polar
sulfur is 115 °C and sulfur deposition leads to catalyst adsorbent surface to carry out this separation. However, the
deactivation, the temperature of the final Claus reactor is presence of other polar natural gas impurities such as H2O,
generally maintained above 200 °C. This results in an which has a dipole moment about twice that of H2S, presents
incomplete conversion of H2S, and the resulting gas stream is challenges to a simple polarity-based CH4/H2S separation.
sent to the tail gas treatment unit for additional sulfur recovery Exploiting the higher acidity of H2S compared to that of H2O
before the waste gases are released to the atmosphere. Thus, using basic chemicals such as amines is also an alternative for
highly sour natural gas is usually processed for H2S removal in sour gas sweetening. However, this principle of separation also
two different units, at two very different concentrations. suffers from some limitations due to the acidity of CO2 being

Figure 2. Different alternatives for sour gas sweetening.

9757 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

higher than that of H2S and the concentration of CO2 in natural H 2S + R3N ⇌ R3NH+ + HS− (5)
gas being generally much higher than that of H2S. The kinetic
diameter of CH4 (3.8 Å) is slightly larger than that of H2S (3.6 MDEA not only offers kinetic selectivity toward H2S but also
Å); this might be another property that can be exploited for this allows an equilibrium that favors the absorption of H2S over
particular gas separation using nanoporous adsorbents or CO2.25,27
membranes. Thus, while several possible solutions exist, Besides MDEA, diisopropanolamine (DIPA), which is used
challenges are associated with each of them, and it is important in the Adip process developed by Shell, has also been reported
to evaluate these on a case-by-case basis. to have a greater selectivity for H2S over CO2 than either
monoethanolamine (MEA) or DEA.28 In 1983, Sartori et al.
2. ABSORPTION investigated 2-amino-2-methyl-1-propanol (AMP), a sterically
hindered amine, for CO2 removal at Exxon Research and
The natural gas industry has relied on gas absorption for
Engineering Company.29 Introducing steric hindrance by a
purification of raw natural gas for nearly 100 years, and this
bulky substituent adjacent to the amino group lowers the
separation technique dominates most sour gas separations
stability of the carbamate formed by the CO2−amine reaction.
around the globe even today. A variety of different liquid
Using kinetic modeling, the authors described as to how
solvents have been employed in a tray or a packed-bed
decreased carbamate stability leads to higher thermodynamic
configuration to selectively absorb H2S and CO2 from CH4 and
CO2 loadings approaching 1 mol of CO2 per mole of amine.
light alkanes. Depending on the strength of interaction of gas
Additionally, this leads to a higher free-amine concentration;
molecules with the solvent and the nature of the absorption
therefore fast CO2−amine reaction rates are obtained despite
isotherm, these solvents maybe broadly classified into chemical
some reduction of the rate constant owing to steric
and physical categories. For chemisorption, solubility of the gas
interference. These hindered amines were found to be selective
in the liquid increases sharply at very low pressures and starts to
toward H2S over CO2 and have been commercialized as the
plateau at higher pressures; the plateau is indicative of the
Flexsorb process by Exxon Research and Engineering
stoichiometric capacity for the chemical reaction. In physi-
Company. It is claimed that the new hindered amine-based
sorption, the plateau is shifted to much higher pressures, and in
processes are potentially attractive replacements for the existing
many cases there is no plateau for the pressure range of interest,
selective H2S removal processes, including the MDEA-based
so solubility continues to increase, more or less linearly, with
process.28
pressure. A sharp distinction between these categories does not
Thorough knowledge of the vapor−liquid equilibria for a
exist (e.g., hydrogen bonding and weak charge transfer), and in
gas/liquid combination is a prerequisite for accurate design of
reality, all solute−solvent combinations have some contribution
absorption columns. Classical thermodynamic modeling of acid
from physisorption. For sour gas sweetening, alkanolamines are
gas absorption in aqueous alkanolamines, using either semi-
the most widely used class of chemical solvents. All other
empirical equilibrium constant approaches or activity coefficient
solvents usually absorb the gas physically, and their very high
methods, has been a subject of interest since the 1930s.
capacities make them of interest if a feed with high acid gas
Recently, Suleman et al. reviewed these models, discussing their
partial pressure is available.
classification, chronological development, salient features, and
2.1. Alkanolamines recommendations of appropriate models depending on the
Bottoms for the first time, in 1930, developed an absorption pressure and loading of the acid gases.30
process for acid gases using alkanolamine solutions.24 For 2.2. Methanol
several years, diethanolamine (DEA) was commonly used in
refineries to remove acid gases. However, not only do H2S and The Rectisol process, developed independently by Linde and
CO2 concentrations in raw natural gas vary over a wide range of Lurgi in the late 1950s, involves physical absorption using
concentrations, but also since the toxicity of H2S is much higher methanol at temperatures between −40 and −60 °C. Since
than that of CO2, the upper-bound specifications for these methanol has a low viscosity, it can be used for very low-
impurities in the pipeline gas are also quite different (4 ppm for temperature absorption processes, in particular for feeds with
H2S and 2% for CO2). Clearly, the extent of H2S and CO2 low partial pressures of acid gases. Since the high vapor
removal required can be very different too. For example, nearly pressure of methanol demands washing of all streams with
complete removal of H2S without any major removal of CO2 is liquid water in order to recover the methanol vapors, all
desired for a feed gas containing high H2S concentration but streams are humidified during the process; the purified gas
CO2 concentration being already at 2%. Traditional alkanol- stream will also be humidified, thus requiring an additional step
amines absorb CO2 in preference to H2S; Frazier and Kohl first to meet the upper-bound H2O specifications for natural gas.
proposed the use of tertiary amines such as methyldiethanol- The main application for the Rectisol process is purification of
amine (MDEA) for selective removal of H2S.25 The CO2− syngas derived from the gasification of heavy oil and coal rather
tertiary amine reaction, than natural gas treating applications.31 The multi-stage Ifpexol
process, developed by Institut Francais du Petrole, can be used
CO2 + R3N + H 2O ⇌ R3NH+ + HCO3− (3) for natural gas applications.32 Physical properties of various
commercially used physical solvents for gas sweetening are
proceeding via the bicarbonate formation, is relatively slow listed in Table 3.33,34
compared with the CO2−primary and CO2−secondary amine 2.3. N-Methyl-2-pyrrolidone
reactions, resulting in a carbamate ion,
Like the Rectisol process, the Purisol process is also a physical
CO2 + 2R 2NH ⇌ R 2NH 2+ + R 2NCOO− (4) absorption process based on N-methyl-2-pyrrolidone (NMP).
However, due to its much lower vapor pressure, it can be
while the H2S−tertiary amine reaction, like all other amines, operated at relatively higher temperatures without excessive
involves a proton transfer and is faster,26 solvent loss. Its operation at sub-ambient temperatures allows
9758 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 3. Physical Properties of Common Sweetening tetrafluoroborate, hexafluorophosphate, acetate, methyl sulfate,
Solvents ethyl sulfate, triflate, etc.43 Due to their polar character, ILs
show a high affinity for polar molecules such as H2O, CO2, and
MeOH NMP PEGDME
H2S, while nonpolar molecules like CH4 and other alkanes
process Rectisol Purisol Selexol exhibit a much smaller solubility. In this section, we will discuss
viscosity at 25 °C (cP) 0.6 1.65 5.8 various advantages and shortcomings of ILs for H2S capture.
density at 25 °C (kg/m3) 785 1027 1030 The reader is also referred to earlier reviews on the
molecular weight (g/mol) 32 99 280 subject.44−46
vapor pressure at 25 °C (Pa) 16700 53 0.097 Jou and Mather first investigated H2S absorption in an IL, 1-
freezing point (°C) −92 −24 −28 N-butyl-3-methylimidazolium hexafluorophosphate ([bmim+]-
normal boiling point (°C) 65 202 275 [PF6−]), and observed that H2S only physisorbed to the IL.47
maximum operating temperature (°C) N/A N/A 175
This further suggests that ILs are unlikely to replace
specific heat at 25 °C (cal/(g·°C)) 0.566 0.40 0.49
alkanolamines completely but may be useful either for the
H2S solubility at 25 °C (mol/L) 1.00 1.62 1.43
bulk removal of high-partial-pressure acid gases or for mixing
CO2 solubility at 25 °C (mol/L) 0.142 0.159 0.162
with a chemical solvent to increase capacity. Pomelli et al.
CH4 solubility at 25 °C (mol/L) 0.0072 0.011 0.011
extended this investigation to a wide range of cations and
anions and demonstrated ILs with extremely high H2S
for simultaneous water removal since water is not required to solubilities.62 They also carried out gas-phase electronic
wash the treated gas stream for solvent (NMP) recovery. structure calculations elucidating the strong binding of H2S to
Various advantages and limitations of the Rectisol and Purisol the anion of the IL, with interaction energies of these
processes are covered in great detail elsewhere.34,35 complexes in the range of 30−60 kJ/mol, indicating traditional
2.4. Poly(ethylene glycol) Dimethyl Ether hydrogen bonds. Experimental data for the Henry’s law
Poly(ethylene glycol) dimethyl ether (PEGDME) is a mixture constants for H2S, CO2, and CH4 in various ILs are
of CH 3 O(CH 2 CH 2 O) n CH 3 oligomers with n = 2−9. summarized in Table 4.
Compared to the other solvents, PEGDME has a higher
viscosity, which reduces mass-transfer rates and tray efficiencies Table 4. Henry’s Law Constants for H2S, CO2, and CH4 in
and increases packing or tray requirements, especially at Ionic Liquids
reduced temperatures.34 However, it benefits from lower
Henry’s law constant (bar)
solvent vapor pressure, which prevents solvent losses in various
streams, thus eliminating the need for water wash to recover the ionic liquid T (°C) H2S CO2 CH4 reference
solvent and allowing for simultaneous dehydration of the feed [bmim ][PF6−]
+
25 14 53.4 1700 47, 48
to the desired dew point of sales gas. The Selexol process, [bmim+][Tf2N−] 30 13 33 44, 49
developed by Allied Chemical Corporation, was first utilized in 40 500 50
1970. [bmim+][BF4−] 30 16 61.6 840 49, 51
2.5. Sulfolane and Diisopropanolamine hmim+][PF6−] 30 18 52
[hmim+][Tf2N−] 30 17 31.6 350 52, 53
Unlike other physical solvents such as methanol, sulfolane are [hmim+][BF4−] 30 12 52
rarely used alone but rather in admixture with a chemical [hemim+][PF6−] 30 28 54
solvent such as an alkanolamine solution.36 Shell Oil’s Sulfinol [hemim+][Tf2N−] 30 19
process, first developed by Dunn et al. in 1964, employs a [hemim+][BF4−] 30 31 108 54, 55
mixed solvent consisting of DIPA, sulfolane, and water.37 The [emim+][PF6−] 80 47 56
presence of the physical solvent appreciably enhances the [emim+][Tf2N−] 30 15 37 570 53, 56
solution capacity, especially when the gas stream to be treated is [emim+][EtSO4−] 30 60 57
available at high pressure and the acidic components are [hmim+][Tf2N−] 30 12 58
present in high concentrations.38 There have been several [omim+][Tf2N−] 30 10 24
subsequent studies to improve the chemical solvent; one such [emim+][Ace−] 30 1.2 59
candidate is MDEA, a tertiary amine that has thermodynamic [emim+][Pro−] 30 0.85
and kinetic selectivity toward H2S in the presence of CO2.39−41 [emim+][Lac−] 30 1.4
Currently, three different solvents are available for the Sulfinol [bmim+][Ace−] 30 0.90
process: Sulfinol-X consists of MDEA and piperazine, Sulfinol- [hmim+][Ace−] 30 0.85
M consists of MDEA, and Sulfinol-D consists of DIPA.42 [MDEAH][Ac] 30 5.5 49 60
2.6. Ionic Liquids [MDEAH][For] 30 11.5 117
Ionic liquids (ILs) are salts having moderately low melting [DMEAH][Ac] 30 3.5 53
points (<100 °C), with their ionic character imparting them a [DMEAH][For] 30 5.9 115
large cohesive energy in the liquid phase, which in turn results [BDMAEE][Tf2N] 25 2.47 29 61
in extremely low vapor pressures over a wide temperature
range. This is a benefit for absorptive gas separation since the Since sorption experiments are much more difficult to
loss of solvent to the gas stream can be minimized. conduct for gas mixtures than for pure gases, an assumption
Additionally, many ILs are thermally very stable and are less needs to be invoked for defining absorption selectivity. There
combustible than other absorbents. Typically, an IL comprises a are at least two possible ways to define ideal absorption
large, low-symmetry organic cation derived by alkyl substitution selectivity that have been used in the literature: (i) ratio of the
on parent cations such as imidazolium, pyridinium, ammonium, solubilities at infinite dilution (Henry’s law coefficients) and (ii)
and phosphonium and an inorganic or organic anion such as computing a mixture selectivity,
9759 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

xi /xj that the solubility of H2S in ILs is governed predominantly by

Sij = entropic effects and that the most promising approach to
yi /yj (6) enhance the sour gases/CH4 selectivities at high pressures is by
assuming that the amount absorbed for a mixture at a given lowering the methane solubility in ILs and not by increasing the
partial pressure is equal to that for a pure case at the same total sour gas solubility, which is essentially dependent on the
pressure. The answers obtained from the two definitions can be solvent molar volume. 69 These authors measured CH4
quite different. In order to assess the validity of ideal absorption solubility in four different ILs and demonstrated that it was
selectivity approximation, Singh et al.63 carried out molecular highly dependent on the polarity of the IL and that, contrary to
simulations for mixture absorption of CO2/H2 in [C4mim+]- the previous selectivity models in the literature,70,71 the molar
[Tf2N−] and [C6mim+][Tf2N−] and found that definition (ii) volume of the IL alone was not a sufficient criterion to tune
for selectivity is in much better agreement with mixture selectivity.72
selectivity. For CO2/CH4 absorption in [C4mim+][Tf2N−], While high solubility and selectivity for the desired molecule
Budhathoki et al. observed definition (i) to be in better form the prerequisites for an IL to be useful for separation
agreement with the mixture selectivity.64 These differences purposes, additional criteria such as thermal and mass transport
demonstrate a lack of quantitative predictive capabilities to through the IL can also contribute to the overall operating cost
determine mixture absorption from pure-component data and of the absorber. Due to their high cohesive energies, ILs
suggest that either mixture measurements or mixture demonstrate high viscosities, 1−3 orders of magnitude higher
simulations with reliable force fields should indeed be carried than for the conventional solvents.73 Higher viscosities imply
out for at least the more promising ILs. higher power consumption for gas−liquid contact, but also an
Trends in affinity of H2S toward the anion of the IL were unusual possibility for diffusion-based selectivity in a gas−liquid
found to depend on the cation under consideration: [emim+]- system. Under such circumstances, it will be helpful if the
[Tf2N−] > [emim+][PF6−],56 [bmim+][Tf2N−] > [bmim+]- diffusivity of H2S is higher than that of CO2 in order to benefit
[BF4−] > [bmim+][PF6−],49 [hmim+][BF4−] > [hmim+][PF6−] from a transport selectivity. Reports investigating the diffusion
≈ [hmim+][Tf2N−],52 and [hemim+][Tf2N−] ≥ [hemim+]- of acid gases in ILs are limited.55,57,74 At room temperature,
[OTf−] > [hemim+][PF6−] > [hemim+][BF4−].54 These results [emim+][EtSO4−] shows a high diffusion selectivity, + H2S/CO2
indicate that H2S solubility is affected by both the cation and ≈ 40, while [hemim+][BF4−] shows + H2S/CO2 ≈ 1.
the anion of the IL, and it may not be possible to find the best In 2002, Bates et al. first reported an IL designed for CO2
anion and the best cation without trying all possible pairs. With capture (task-specific ionic liquids (TSILs)) where a primary
the aid of molecular simulations, deeper insights can be gained amine moiety is covalently tethered to an imidazolium cation
about the entropic and enthalpic contributions of H2S for enhanced interaction with CO2.75 Huang et al. employed
solubilities in different ILs. Recently, Sánchez-Badillo and co- carboxylate-based ILs for H2S capture, the Lewis base character
workers computed solvation thermodynamics of H2S in ILs and of which resulted in higher absorption capacities.59 These ILs
demonstrated qualitative agreement with experimental data.65 absorb about 0.3 mol of H2S per mole of IL under ambient
These calculations showed that, at room temperature, the pressure at 60 °C, about 15 times higher than conventional ILs.
absolute solvation enthalpy is 2 times higher than the absolute The enthalpy of this reactive absorption varies between 30 and
entropic contribution for [bmim+][PF6−] and [bmim+][BF4−] 65 kJ/mol, depending on the cation and the anion, but is still
and 3 times higher for [bmim+][Cl−]. The solubility of H2S in lower than typical enthalpies of absorption in alkanolamines
[bmim+][Tf2N−] from molecular simulations was shown to be (≥80 kJ/mol). However, in spite of these favorable character-
sensitive to the choice of the H2S force field; the TraPPE model istics, the SH2S/CO2 in the carboxylate ILs is even worse than in
resulted in the most accurate predictions.66,67 Jalili et al. also
studied [C8mim+][Tf2N−] and found that the solubility of H2S the normal ILs due to the strong affinity of carboxylate ions
in [Cnmim+][Tf2N−] ILs increases as the number of carbon toward CO2.76 This challenge was overcome with the design of
atoms in the alkyl substituent of the methylimidazolium cation, dual Lewis base functionalized ILs (DLB-ILs),77 which have a
n, increases.58 The authors explained that since the molar tertiary amine group tethered to the carboxylate anion. It was
volume of the IL increases with n, possibly due to poorer found that the DLB-ILs could reversibly absorb 0.39−0.84 mol
interactions between bulkier cations and the anion, higher free of H2S per mole of IL at 1 bar and 60 °C, which is significantly
volumes may result, in turn increasing the solubility. One must higher than for other ILs. These DLB-ILs show SH2S/CO2 ≈ 20 at
remember that such increased acid gas solubility due to 1 bar and SH2S/CO2 > 100 at pressures less than 0.1 bar. Density
increased free volumes will also lead to increased CH4 functional theory (DFT) calculations demonstrated the
solubility, in turn reducing SH2S/CH4. As described previously favorable geometry and energetics of the carboxylate−H2S−
with regard to how the necessity for selective H2S over CO2 amine complex and the reduced CO2−IL binding.
absorption led to the development of absorbents such as Although TSILs can be modified for high H2S capacity and
MDEA, new solvents with SH2S/CO2 > 1 that also exert a low SH2S/CO2, their high costs and viscosity disfavor their application
vapor pressure to reduce solvent losses will be interesting in industry.60 This limitation encouraged the same authors to
targets, and ILs, with their tunable polarity, can be ground- explore protic ionic liquid (PILs), which are a class of low-cost
breaking candidates to achieve this selectivity.68 Investigations ILs that can be easily prepared from simple materials through
of ternary systems consisting of H2S, CO2, and [C8mim+]- direct neutralization between the corresponding acid and
[Tf2N−] showed SH2S/CO2 ≈ 3 under ambient pressure and base.60,78 For instance, mixing dimethylethanolamine and acetic
temperature.58 acid yields dimethylethanolammonium acetate ([DMEAH]-
All the studies conducted so far focused only on H2S and not [Ac]). It was demonstrated that [DMEAH][Ac] and dimethyl-
on CH4 absorption. After all, for gas sweetening, selectivity is ethanolammonium formate ([DMEAH][For]) have SH2S/CO2
more important than solubility alone. Carvalho et al. suggested greater than those of normal ILs and any other organic solvent.
9760 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Another issue with TSILs is that they are highly hydrophilic the past few years,87,88 and CO2 removal from natural gas was
and miscible with water, thus increasing the expense of reviewed recently.89 To the best of our knowledge, there is no
regeneration by requiring energy for removing bound water. report in the literature on H2S capture, and this might be an
Both the cation and anion can affect the mutual solubilities unexplored opportunity to discover novel solvents for this
between water and ILs, but the anion plays the major role in application.
determining their phase behavior.79 Huang et al.61 synthesized Since 2010, several research groups have started to focus on
half-protonated diamines with a free tertiary amine group equations of state (EoS) modeling of H2S + ILs phase
(disfavoring CO2) tethered on the cations and bis(trifluoro- behavior.90−98 A major drawback in the use of the most
methylsulfonyl)imide ([Tf2N−]) as the anion, which is a well- prevalent cubic EoS is that they require critical parameters of
known building block for hydrophobic ILs with low viscosity.79 the ILs that can only be obtained indirectly and with large
Although these ILs have H2S absorption capacity similar to that
uncertainties, limiting their utility to correlation purposes
of other TSILs reported in the literature on a mole ratio basis,
only.99 In contrast, contributions from various energetic
they have a molecular weight nearly twice that of other TSILs,
making capacity on molality basis (per mass of IL) nearly half. interactions, such as dipole−dipole, quadrupole−quadrupole,
As described earlier, compared to common organic solvents and dipole−quadrupole, can be explicitly included in a class of
with high vapor pressures, ILs contribute negligible volatile equations based on the statistical associating fluid theory
organic compounds to the atmosphere, thus reducing the issue (SAFT) formalism.90,92 Several other correlation techniques
of atmospheric emissions. However, it has been noted in the that use molecular weight, critical temperature and pressure,
literature that most ILs are not intrinsically “green”,80 and and acentric factor as input variables have been applied to H2S
similar to other commercial organic solvents, irresponsible absorption in ILs. 100−103 Group contribution methods,
release of them will pose a threat to the environment. One of requiring only the number and type of each fragment type as
the key issues here is that, unlike the common solvents, the the input, were used to establish a comprehensive database on
toxicity of ILs is not fully characterized. However, at least in the solubility of H2S in ILs using 1334 pieces of data covering
principle, the large design space of ILs should allow for the period from 2007 to 2016.104
designing solvents that not only are effective for the application Lei et al. employed the conductor-like screening model for
of interest but also pose no serious ecological threat. In an real solvents (COSMO-RS) to predict the solubilities of
attempt to explore environmentally benign solvents, Guo et al. different gases, including H2S, in ILs and observed deviations of
synthesized and measured H2S absorption in a series of less than 40%.43 This approach requires only the molecular
caprolactam tetrabutylammonium bromide (CPL−TBAB) ILs structure and is independent of any experimental correlation. In
with different mole ratios of CPL and TBAB.81 For the 1:1 another study, Mortazavi-Manesh et al. used COSMO-RS to
CPL−TBAB mixture, the solubilities of pure H 2 S at
screen 425 ILs for sour gas separations and, based on binary
atmospheric pressure are 0.69 and 0.44 mol of H2S per mole
selectivities, suggested 58 promising ILs with H2S/CH4, H2S/
of IL at 30 and 90 °C, respectively.
Heldebrant et al. developed a new class of CO2-binding C2H6, and H2S/CO2 selectivities in the ranges 51−75, 11−23,
organic liquids (CO2BOLs) that chemically and reversibly and 2.6−3.4, respectively.105
absorb CO2.82 These organic liquids are mixtures of organic In 1977, Tennyson and Schaaf provided some guidelines for
alcohols and amidine/guanidine bases. Since these organic the choice of appropriate physical/chemical solvent depending
liquids have lower specific heats compared to H2O and also on the partial pressures of acid gases in the feed and product
since they require much lower regeneration temperatures, a gas streams.106 Process optimization for absorption using ILs in
higher energy efficiency can be achieved during absorption/ comparison to the commercial solvents can help identify the
regeneration cycles. Following on the capture of acid gases such partial pressure regimes under which certain ILs may be
as CO2, SO2, CS2, and COS using such organic liquids,82,83 it preferred over conventional solvents. The past decade has seen
was hypothesized that weak bases such as tertiary alkanol- several reports demonstrating high H2S−IL affinities, but ILs
amines (anhydrous) could react with H2S to form liquid for selective acid gas removal are still in the design stage.
hydrosulfide salts that could be used as reversible H2S-binding Several design variables, such as length of the alkyl chains on
organic liquids (H2SBOLs).84 These authors demonstrated the cation, branching of the alkyl chain, task-specific functional
desorption of H2S by either inert gas sparging or mild heating, groups, pairing with appropriate anions, etc., make this a multi-
or even at room temperature by the addition of low-specific- dimensional combinatorial problem. As is shown in several
heat antisolvents such as hexanes that can be boiled at 50−70 fundamental studies for discovery of novel solid adsorbents,
°C. BOLs are non-ionic before reacting with the gas molecule; insights from molecular simulations can potentially speed up
however, after chemically reacting with the gas, they become discovery of optimal ILs and DESs for H2S capture.
ILs.
Additionally, simulations can also help understand the
Very similar in properties to ILs, deep eutectic solvents
differences between selectivities calculated from pure-compo-
(DESs) are a relatively new class of solvents that are formed by
complexing a large, asymmetric quaternary ammonium salt with nent absorption isotherms and those observed in an actual
a metal salt or a hydrogen bond donor.85 Hydrogen bonding in mixture measurement. In conclusion, while ILs enjoy clear
the system results in charge delocalization, thus lowering the advantages such as design flexibility and negligible vapor
lattice energy and decreasing the melting point of the mixture pressures, properties such as viscosity, diffusivity, thermal
relative to the melting points of the individual components.86 conductivity, and reaction kinetics in ILs under varying
In addition to most of the better solvent properties of ILs, temperatures and partial pressures of acid gases require
DESs are generally more biodegradable, less toxic, and thorough investigations in order to compare these new solvents
inexpensive, thus making these solvents more sustainable. with the conventional solvents that have been successfully
Articles on CO2 capture using DESs have started to appear in commercialized for several decades.
9761 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

3. ADSORPTION
Adsorptive separation involves the use of a selective and high-
capacity solid substrate that shows a preferential affinity for one
or more molecule types from their mixtures. Similar to
absorption, adsorption may also be classified as physisorption
and chemisorption, depending on the strength of sorbate−
sorbent interactions. In addition to interaction-based separa-
tion, different sizes and shapes of molecules can also be
exploited for selective removal of one compound over another.
In this section, we will discuss a variety of adsorbents, such as
metal oxides, MOFs, zeolites, carbons, and composite materials,
for H2S capture from gas mixtures containing CH4, CO2, N2,
H2, and/or H2O. For any adsorption application, it is desirable
that the adsorbent has a high separation selectivity for the
molecule of interest (primarily H2S in this case) and also a high
capacity for the adsorbate. In addition to capacity and
selectivity, stability under extreme acid gas environments as
well as performance and structural regenerability are important
for longer lifetime of the adsorbent bed.
3.1. Metal Oxides
Since the 1960s, metal oxides have been used for desulfuriza-
tion of coal gas, syngas, hydrogen for fuel cells, and Claus tail
gas.107,108 Metal oxides for hot-gas desulfurization have been
investigated extensively;109,110 this large body of literature is
beyond the scope of this Review, and we focus here mostly on
low to moderate temperatures (25−300 °C) that are of Figure 3. Total coverage of sulfur (S + SH) as a function of H2S
relevance to natural gas sweetening, including the Claus tail gas. exposure at 300 K on Al2O3, MgO, α-Cr2O3(0001), ZnO, Cu2O, and
In 1976, Westmoreland and Harrison screened for the high- Cr3O4(111). One monolayer (ML) corresponds to the deposition of
temperature (400−1200 °C) desulfurization potential of oxides 1.43 × 1015 atoms/cm2. The numbers in parentheses denote the band
gap of each oxide. Reprinted with permission from ref 114. Copyright
of 28 elements with computational free energy minimizations 1999 AIP Publishing.
of a C-H-N-O-S elemental system in the presence of the metal
oxide using temperature-dependent free energies from the
literature.111 Eleven candidate solids, based upon the metals Fe, better desulfurization using CuO compared to Cu2O. Room-
Zn, Mo, Mn, V, Ca, Sr, Ba, Co, Cu, and W, showed temperature experimental screening of oxides of Ag, Cu, Zn,
thermodynamic feasibility (at least 95% desulfurization and Co, Ni, Ca, Mn, and Sn showed a conversion of 66% for CuO
solid stability) for high-temperature desulfurization. Another but <10% for the other oxides.117
simulation study at 400 °C with feed containing 12.5% CO2, Several DFT studies on H2S reaction with oxides of
42.5% CO, 32.175% H2, 12.5% H2O, and 3250 ppm H2S Zn,118,119 Fe,120 Ce,121 Ti,122 Mn,123 V,124,125 Cu,126,127
showed that oxides of Zn, Mo, Mn, W, Sn, Co, Cu, Ni, and Fe Co,128 and Mg129 have been carried out in the past decade.
should be successful in reducing the residual H2S content to While these studies provide some details of possible reaction
0.03, 0.6, 2.5, 3.4, 19, 24, 39, 41, and 91 ppm, respectively.112 mechanisms for each metal oxide, a comparative study of
Rodriguez et al. studied adsorption of H2S on a series of different pure and mixed oxides for adsorption and
metal oxide surfaces using synchrotron-based high-resolution regeneration from first-principles calculations, at the same
photoemission.113,114 It was observed that the smaller the band level of theory, can provide insights into optimal metal oxide
gap in the oxide, the bigger its reactivity toward S-containing selection. Jia et al. carried out a comparative study of the
molecules (see Figure 3). Ab initio self-consistent field (SCF) reactions of MmOn− (M = Fe, Co, Ni, Cu, and Zn) with H2S
theory calculations and experimental data indicated that the using mass spectrometry and DFT calculations.130 The authors
reactivity of an oxide mainly depends on how well its bands mix tested several density functionals by calculating the bond
with the orbitals of H2S or HS.113 Oxides possessing a large enthalpies of M−O, M−S, HS−H, and S−H species, and
ionicity, like Al2O3 and MgO, displayed the lowest reactivity except for Zn−S, the B3LYP functional was found to be
toward H2S, indicating that the electrostatic interactions generally in better agreement with the experimental enthalpies.
between the dipole of H2S and the ionic field generated by Adsorption of H2S by ZnO is stoichiometric at 350 °C but
the charges in an oxide may play only a secondary role in the falls off rapidly as the temperature is reduced.131 Since this is an
adsorption process.114 As can be seen from Figure 3, Cr3O4 exothermic reaction, this decline in conversion is due to slower
interacts strongly with H2S. Even before this study, the most reaction kinetics. Baird et al. reported 91% conversion of Co3O4
significant evolution of the Claus process, the Superclaus at room temperature,132 while Xue et al. reported only a 1%
process, employed an α-alumina support coated with a mixture conversion.117 These studies suggest the possible implications
of iron and chromium oxides as the catalyst.115 However, due of synthesis recipes on the size, shape, and surface area of the
to the severe toxicity of chromia, continuing research to metal oxide particles that can, in turn affect the H2S
develop new catalysts for this important reaction is under- breakthrough capacity. In fact, there are several reports
way.116 CuO and Cu2O have band gaps of 1.2 and 2.1 eV, exploring desulfurization using different ZnO morpholo-
respectively; the contribution by Rodriguez et al. should imply gies.133−138 While it is hard to draw a head-to-head comparison
9762 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 5. BET Surface Areas and Percentage Utilization in ZnO for H2S Adsorption
morphlogy SBET (m2/g) % utilization T (°C) composition reference
imp-ZnO, 25−600 nm 3.1 0.29 <200 H2S in N2 Baird et al.131
cp-ZnO, 10−100 nm 37 3.7

cp-ZnO, 20−50 nm 39 14 30 2% H2S in N2 Baird et al.132

44 nm, commercial 3.9 1.4 25 N/A Carnes et al.133

33 nm, commercial 20 1.9
4 nm, nanocrystalline 130 24
29 100
42 250

64 × 283 nm, rod-like 12 57 250 0.5−1% H2S, 59% CH4 in He Habibi et al.134
49 150
56 nm, spherical 50 22 250
8.3 150

4000 nm, commercial 45 38 200 1% H2S in He Garces et al.135

34 69 300

Zn(acac)2, microwave 52 38 200 1% H2S in He Garces et al.136

Zn(acac)2, heat 14 11
Zn(NO3)2, heat 14 5.1

between independent studies due to different gas velocities, oxidation state of the sulfur remains as S2− throughout the
reactor volume, feed compositions, and definitions of break- process.143 The X-ray photoelectron spectroscopy (XPS) and
through concentration, data for these are compiled in Table 5. X-ray diffraction (XRD) data for the final states of the
The % utilization for a metal oxide-based adsorbent is defined adsorbent confirmed the theory of a two-phase material
as the ratio of the actual amount of H2S adsorbed to the consisting of crystalline ZnO (wurtzite) and a less ordered
theoretical amount of H2S adsorbed by complete conversion of phase of ZnS (wurtzite), in spite of the fact that ZnS blend
metal oxide to metal sulfide, multiplied by 100. Note that the structure is thermodynamically more stable.144 The reader is
value can change depending on the choice of the oxidation state also referred to a review paper by Samokhvalov et al. on the
of the metal in its oxide and sulfide forms. This number can be low-temperature desulfurization using ZnO.145
greater than 100 due to H2S adsorbing without chemical Until recently, ZnO sulfidation models assumed inward
reaction with the metal oxide (see, for instance, the H2S diffusion of H2S, but Neveux et al. showed evidence for an
adsorption on metal oxides impregnated on high-surface-area outward growth mechanism (see Figure 4).146,147 Using
silica supports).
The reaction between H2S and a metal oxide in the absence
of O2 belongs to the class of non-catalytic gas−solid reactions.
These reactions have been most commonly analyzed using
either the unreacted shrinking core model (USCM)139 or the
grain model.140 USCM is built on the assumption that the
reaction is much faster than the diffusion and, hence, the
reaction occurs at the solid reactant surface. This model clearly
assumes inward propagation of the reaction front. In the grain
model, each of the large number of spherical grains in a solid
pellet is assumed to react independently according to the Figure 4. Two mechanisms of ZnS growth starting from pure ZnO
USCM. Early reports on sulfidation kinetics of ZnO, employing (center): (a) inward ZnS growth and (b) outward ZnS growth.
the grain model with constant or varying properties (size Reprinted with permission from ref 150. Copyright 2012 American
changes due to sintering and reaction) of the adsorbent, agreed Chemical Society.
quantitatively with experiments.141,142
From the time when empirical commercialization of ZnO as transmission electron microscopy (TEM) and electron
a desulfurization sorbent took precedence until the early 1990s, tomography, the authors observed voids inside the ZnS
very little was known about the structural changes and the particle, suggesting that the Zn2+ and O2− species diffuse
mechanism for this reaction. Using in situ sulfur K-edge X-ray through the product ZnS layer. Since the external particle
absorption spectroscopy (XAS) in the soft X-ray region (<5000 surface increases during the course of the reaction, the decrease
eV), Evans and co-workers first showed that the reaction of in reaction rate with time cannot be explained if the rate-
ZnO and H2S occurs via island formation of ZnS on the oxide determining step (RDS) is located at the external surface
surface and, as the reaction proceeds, eventually bulk regions of (adsorption, desorption, external interface reaction). Thus, the
ZnS are formed.143,144 Since there was no apparent shift in the authors proposed that diffusion through the ZnS layer is most
edge energies of the S K-edge spectra, it was concluded that the likely to be the RDS. Eight elementary steps of the sulfidation
9763 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 6. BET Surface Areas and Percentage Utilization in Mixed Zinc-Based Metal Oxides for H2S Adsorption
adsorbent SBET (m2/g) % utilization T (°C) composition reference
ZnO 29 10 200 0.2% H2S, 49.9% H2, 49.9% N2 Skrzypski et al.148
Cu0.03Zn0.97O 44 47
Cu0.06Zn0.94O − 61
ZnO 29 28 250
Cu0.03Zn0.97O 44 70
Cu0.06Zn0.94O − 76
ZnO 29 71 350
Cu0.03Zn0.97O 44 79

cp-ZnO 36 12 40 0.3% H2S in N2 Jiang et al.149

cp-Cu33.3Zn66.7 67 22
cp-Cu67.7Zn33.3 54 32
cp-Cu88.9Zn11.1 48 38
cp-CuO 27 41
cp-Zn90Al10 82 17
cp-Cu30Zn60Al10 93 25
cp-Cu60Zn30Al10 76 39
cp-Cu80Zn10Al10 61 52
cp-Cu90Al10 46 74
mp-Zn90Al10 61 13
mp-Cu30Zn60Al10 93 30
mp-Cu60Zn30Al10 98 48
mp-Cu80Zn10Al10 81 67
mp-Cu90Al10 46 48
mp-Cu80Zn10Al10 81 99 80

imp-ZnO 3.1 0.29 <200 H2S in N2 Baird et al.131

imp-5.7%Fe/ZnO 5.0 0.31
imp-4.7%Co/ZnO 10 0.58
imp-6.1%Cu/ZnO 4.8 0.42
cp-ZnO 37 3.7
cp-4.4%Fe/ZnO 17 1.7
cp-6.2%Co/ZnO 46 4.7
cp-4.7%Cu/ZnO 57 5.4

Zn6Al2O9 69 8.9 30 2% H2S in N2 Baird et al.151

Co3Zn3Al2O10 110 12
Co6Al2O11 93 15

ZnO (0/100) 39 14 30 2% H2S in N2 Baird et al.132

18.66%Co-59.26%Zn (20/80) 65 24
29.13%Co-40.36%Zn (40/60) 66 33
31.69%Co-40.67%Zn (50/50) 60 20
43.80%Co-16.42%Zn (70/30) 57 46
64.74%Co-7.16%Zn (90/10) 83 63
Co3O4 (100/0) 87 92

reaction were proposed, with the outward growth of ZnS and indicating the appearance of Cu+ in a sulfide environment. The
diffusion of zinc and oxygen through ZnS.147 The authors sulfidation mechanism of these mixed oxides was further
derived a rate expression that also considered the evolution of investigated using TEM and in situ synchrotron XRD.150 While
the voids by a nucleation−growth mechanism, and the model non-doped ZnO at 250 °C resulted in void formation at the
predictions agreed quantitatively with experiments. ZnO/ZnS interface, Cu-doped ZnO showed no cavity
To enhance the H2S capacity at lower temperatures, another formation; this suggests an inward growth mechanism where
approach of varying the ZnO morphology involves doping with S2− anions diffuse from the external surface to the internal
other metallic species (see Table 6).117,131,132,148−153 Skrzypski ZnO/ZnS interface and exchange with O2− anions. This radical
et al. studied mixed Cu−Zn oxides and found a 6-fold increase change in mechanism was attributed to the presence of sulfur
in capacity at 200 °C by doping 6 mol% Cu.148 XPS of pristine vacancies in the ZnS layer due to partial replacement of Zn2+ by
Cu0.03Zn0.97O showed a Cu 2p3/2 peak (933.5 eV), character- Cu+; these anionic vacancies accelerate the diffusion of sulfur to
istic of Cu2+ surrounded by oxygen; after only partial yield higher conversions of Cu−Zn oxides. XRD measurements
sulfidation, this peak shifted to lower energy (932.6 eV), showed that the size of the ZnS crystallites formed from non-
9764 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

doped ZnO remains constant during the reaction. In contrast, for 10Mn−45Zn−45Ti−O using a feed of 600 ppm H2S, 25%
the pronounced growth of Cu-doped samples enhances the H2, and 7.5% CO2 in helium. The presence of 1% H2O in the
mobility of sulfur anions in Cu-containing ZnS particles. feed leads to a significantly higher capacity (160 mg/g).
Jiang et al. also studied Cu−Zn and Cu−Zn−Al mixed metal Davidson et al. have demonstrated the autocatalytic role of
oxides prepared via co-precipitation (cp) or multi-precipitation water during desulfurization on ZnO, suggesting a mechanism
(mp) methods.149 Better dispersion by aluminum results in involving proton transfer from H2S to chemisorbed OH groups
smaller mean crystalline sizes and higher specific surface area of from hydration of the surface.137,155 Addition of Fe showed a
Al-containing sorbents; breakthrough sulfur capacities also pronounced effect of Fe/Mn molar ratio on the morphology of
improve as a result of reducing the thickness of the potential the Fe−Mn−Zn−Ti mixed oxide crystallites.116 H2S uptakes of
sulfide shell on the outer layer of the oxide, in turn reducing the 82 mg/g and 17 mg/g were observed for 5Fe−15Mn−40Zn−
diffusion path length. The Cu-rich adsorbents were more 40Ti−O and 20Fe−40Zn−40Ti−O, respectively; in this case,
suitable for H2S adsorption than the Zn-rich adsorbents, with 1% H2O led to only a small increase in capacity to 100 mg/g.
mp-Cu80Zn10Al10 reaching nearly complete conversion at 80 Mössbauer spectroscopy is sensitive to changes in the energy
°C. XRD of sulfidated samples revealed that the CuS peaks levels of an atomic nucleus in response to its environment;
were much more dominant compared to the CuO peaks, but quadrupole splitting (Δ) is induced by an electric field gradient
high-intensity ZnO peaks persisted for sulfidated Zn-rich of nearby electrons. Changes in Δ during sulfidation and
samples. This suggests that CuS penetrated deep into the regeneration in the case of 5Fe−15Mn−40Zn−40Ti−O solid,
crystal, while ZnS likely formed only surface layers. Since the compared to no changes for 20Fe−40Zn−40Ti−O solid,
anionic sites in ZnO need to rearrange from hexagonal to cubic suggest that Mn makes the mixed oxide structures more
lattice on sulfidation, while those in CuO need only to susceptible to structural distortions, in turn enhancing H2S
transform from monoclinic to hexagonal, a faster sulfidation adsorption. Regeneration in 20% O2 in He at 500 °C for 4 h
rate was proposed for CuO than that for ZnO. Similar resulted in over 80% of capacity recovered for the Fe-
structural arguments were invoked for Zn−Co oxides containing solids, while regeneration at 750 °C halved the
(discussed below) by Baird et al.132 It should be noted here H2S capacity. The structure was preserved for the regenerated
that the sulfidation reaction in the study by Jiang et al. was 5Fe−15Mn−40Zn−40Ti−O solid with clear peaks in the XRD
carried out at 40 °C,149 as opposed to a much higher pattern.
temperature (370 °C) in the study by Hayter et al. (discussed Liu et al. investigated copper-based cp-mixed oxides (see
above)144 that reported the wurtzite (hexagonal) phase of ZnS. Table 7) and observed high capacities for Fe-doped oxides.157
This suggests that temperature can play a role in stabilizing
different crystal structures of the product metal sulfide, which in Table 7. BET Surface Areas and Breakthrough Capacity in
turn can influence the overall sulfidation kinetics. Mixed Copper-Based Metal Oxides for H2S Adsorption
Co-precipitated mixed oxides showed smaller particle sizes,
SBET mg S/
higher specific surface areas, and a higher H2S capacity adsorbent (m2/g) g sorbent T (°C) composition reference
compared to their impregnated counterparts that are cp-CuO 16.4 17.7 40 0.1% H2S in Liu et
synthesized by immersing the parent oxide in a salt solution cp-Cu3Al 72.3 86.0 CO2 al.156
of the dopant metal and calcining this precursor at ∼350 °C.131 cp-Co0.44Cu3Al 87.4 73.4
These authors also studied adsorption on single-phase Co− cp-Ni0.44Cu3Al 86.0 55.7
Zn−Al−O prepared from co-precipitated hydroxycarbonate cp-Ce0.44Cu3Al 82.1 91.1
precursors (hydrotalcite-type structure).151 The presence of cp-Fe0.44Cu3Al 96.7 114
aluminum increased the specific surface areas, but the modified 207 70
compounds did not adsorb significantly more H2S per metallic 212 100
(Zn, Co) species. H2S adsorption in Co−Zn oxides with 220 130
varying Co/Zn atomic ratios showed that the increase in % cp-CuO 16.4 25.3 40 0.1% H2S in
utilization correlates well with the increase in surface area; the cp-Cu3Al 72.3 106 N2
main function of Co doping seems to be increasing the surface cp-Fe0.44Cu3Al 96.7 125
area by reducing the particle size.132 Thermodynamically,
Co3O4 has a more negative free energy for sulfidation than
ZnO. Additionally, sulfiding cubic Co3O4 normal spinel to Capacities were 10−40% higher for H2S in bulk N2 compared
spinel Co3S4 or Co8S9 will require lesser lattice rearrangement to the case of bulk CO2, suggesting some degree of competitive
than that for hexagonal ZnO to cubic ZnS. adsorption with CO2; the inhibition effect was lower for doped
Westmoreland et al. showed that the intrinsic reaction rate of CuO compared to pure CuO. The authors report that
H2S with MnO over a temperature range of 300−800 °C is regenerated CuO exhibited a breakthrough capacity up to
approximately 1 order of magnitude greater than that with 75% of its initial value when regenerated between 100 and 200
ZnO, even though its surface area was an order of magnitude °C but decreased rapidly with further increase of regeneration
smaller.154 Higher conversions were observed for mixed oxides temperature. The authors explain that CuS tends to decompose
of Zn containing Fe, Ni, Co, Mn, Cu, Al, Ti, and Zr at room into the more stable Cu2S when the temperature is above 220
temperature, especially for Zn2.9MnO4.2, but the reason for poor °C. The authors carried out five adsorption−regeneration
conversion for pure Mn 3O4 solid is not yet clear.117 cycles for pure CuO, Cu−Al−O, and Fe−Cu−Al−O with
Regeneration of Zn−Ti-based mixed metal oxides using 20% regeneration using air at 100 °C. For CuO, the regenerated
O2 in He at 750 °C for 4 h showed broad XRD peaks of very capacities were maintained at 75% of the fresh adsorbent
low intensity, indicating transformation to an amorphous capacity; however, regenerated capacities decreased for the
material.152 Doping this oxide with Mn, Cu, or Mo led to an Cu−Al−O and Fe−Cu−Al−O adsorbents and became steady
improvement in H2S uptake for the regenerated solid: 41 mg/g during the last two cycles at 50% and 43% of the fresh capacity,
9765 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 8. BET Surface Areas and Percentage Metal Utilization in Mesoporous Metal Oxides for H2S Adsorption
adsorbent SBET (m2/g) % utilization T (°C) composition reference
3DOM-F65 44 68 350 300 ppm H2S, 5% H2 in N2 Fang et al.158
3DOM-F90 43 70
3DOM-F158 34 93
3DOM-F205 31 90
3DOM-F550 16 79
3DOM-FS156 110 97

FS-1 125 30 80 360 ppm H2S in N2 Huang et al.159

FS-2 120 42
FS-4 120 75
FS-8 113 76
FS-16 72 21
γ-Fe2O3 82 18
α-Fe2O3 36 0
Fe2O3 (commercial) 87 −
γ-Fe2O3 82 5.6 60
2.6 40
1.0 20

ZnO (commercial) 60 16 30 360 ppm H2S, 3% H2O in N2 Wang et al.160

SZ50 220 11
3DOM-SZ50 340 69
3DOM-SZ20 360 60
3DOM-SZ73 140 59

Co3O4 (commercial) 3 4 200 1% H2S in He Pahalagedara et al.161

m-Co-150 23 9
m-Co-250 140 130
m-Co-350 99 130
m-Co-450 38 40
m-Co-250 140 30 25
50 100
100 150
120 175
130 250
m-Cr-250 N/A 50 200
m-Cu-250 N/A 60
m-Mn-250 N/A 110
m-Fe-250 N/A 60

respectively. These studies bring to light the difficulty in 44 m2/g for 3DOM-F and from 80 to 220 m2/g for 3DOM-FS.
regenerating the high capacities of mixed metal oxides; future Mesopore-sized void spaces between the crystallites are
work in this area should focus not only on high-capacity observed in the TEM images and also from the pore size
sorbents but also on better regenerability. distribution (20−33 and 6−17 nm for 3DOM-F and 3DOM-
Researchers have investigated H2S adsorption in three- FS, respectively) obtained from the N2 isotherms. It can be
dimensionally ordered macroporous (3DOM) structures of seen that the % utilization of Fe improves with increase in
iron oxide (3DOM-F) and of silica-supported iron oxide macropore diameter up to about 150 nm due to possible
(3DOM-FS) prepared by a colloidal crystal templating method reduction in diffusion limitations, but further increasing the
(see Table 8).158,159 Highly uniform polystyrene (PS) beads macropore diameter leads to a lowering of Fe utilization. Nearly
ordered into close-packed domains were used as a template for complete regeneration of 3DOM-FS samples using 2% O2 in
the growth of the inorganic 3DOM structure; the resulting pore N2 at 300 °C was shown for three cycles. Scanning electron
diameters were considerably smaller than the original PS microscopy (SEM) illustrated that, despite some cracking, the
sphere.158 The shrinkage in size resulting from sintering during macroporosity and ordering of the structure was still intact and
calcination ranged from 45 to 59% for the 3DOM-F sorbents the overall 3DOM structure was retained after second
and from 37 to 45% for the 3DOM-FS sorbents. 3DOM-FS regeneration.
showed somewhat smaller nanoparticles (5 nm) than 3DOM-F Huang et al. studied the effect of Fe:Si ratio from 1 to 16
(25 nm), which the authors explain as silicon keeping the iron (and also the case of no Si), temperature, and humidity on the
grains separate and preventing the latter from growing larger low-temperature desulfurization performance of 3DOM
during calcination. Depending on the macropore diameter, structures.159 Breakthrough capacities showed an increase
these structures have variable surface areas, ranging from 16 to with increase in Fe:Si ratio up to 8; low capacity at even
9766 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 5. Schematic illustration of the impregnation and thermal treatment process for stabilizing metal nanoparticles in the cage structure of SBA-
16. Reprinted with permission from ref 171. Copyright 2009 American Chemical Society.

higher Fe:Si ratios was attributed to the presence of α-Fe2O3. physically adsorbed H2S beyond the chemically reacted
SiO2 enhanced the dispersion of Fe2O3, which stabilized the γ- capacity. The authors also investigated other mesoporous
Fe2O3 and hindered its transformation to α-Fe2O3 during materials (Cr2O3, CuO, Mn2O3, and Fe2O3) that showed
calcination. The authors further explain that the vacancy site in superior activities compared to their nonporous analogues (as
the lattice of γ-Fe2O3, accompanied by its lower density, larger high as 200 times). Using field-emission scanning electron
surface area, and smaller grain size, facilitates ion diffusion. microscopy (FESEM), the mesostructure was shown to be
Relative humidity of ∼10% favored H2S removal. It was also preserved after sulfidation, but the mesostructure collapsed
found that conventional regeneration at 300 °C using 5% O2 in during calcination at 450 °C, and regeneration at 400 °C in air
N2 resulted in a drastic loss of capacity for the regenerated resulted in a 45% drop in capacity. An alternative regeneration
material. As an alternative regeneration technique, the authors technique for ZnO catalysts using a 3:1 H2O/N2 mixture at 800
proposed an in situ regeneration process by adding 5% O2 to °C resulted in nearly 100% capacity regeneration without loss
the feed stream (desulfurization and regeneration occur at the of porosity, as opposed to only about 10% regeneration and
same time); this resulted in breakthrough capacities twice those significant increase in bulk density of the adsorbent when air
observed when O2 was absent from the inlet gas. In the case of was used at the same temperature.163 A significant scope still
conventional regeneration, a significant fraction of γ-Fe2O3 remains for developing better regeneration procedures for
transformed to α-Fe2O3, while no such change was observed metal oxides that retain not only the adsorption capacity
for the case where O2 was added to the feed. It is unclear if this reversibility but also the structural reversibility. The latter is
simultaneous desulfurization−regeneration process is resulting more important from the perspective of adsorbent bed stability
in a transformation of H2S to SO2; monitoring only H2S to prevent attrition of the solids that will increase the pressure
concentration in the outlet stream may hint at high H2S drop across the bed.
breakthrough capacity, but monitoring SO2 concentration may 3.2. Metals
reveal that little or no sulfur is adsorbed in such a process. Early literature focused on H2S poisoning of metal catalysts and
In another study, Wang et al. synthesized 3DOM structures the threshold concentration below which H2S will not unduly
of ZnO and silica, with different weight percent of ZnO ranging poison a given metal catalyst. McCarty and Wise investigated
from 23 to 70%.160 Samples exhibited a typical amorphous the thermodynamics of H2S chemisorption on different metallic
structure, except 3DOM-SZ73, which showed peaks corre- catalysts over a wide range of temperatures and reported the
sponding to ZnO in the XRD pattern. Highest utilization of order of increasing activity toward H2S at 300 °C and 50%
69% was observed for the 3DOM structure with 50 wt% ZnO; surface coverage to be Ni > Co > Ru > Fe > Pt.164−167 Using
nearly 70% of this capacity was retained for four regeneration magnetization measurements, Den Besten and Selwood showed
cycles (air at 500 °C for 4 h). These adsorbents hold significant that H2S is dissociatively chemisorbed on the Ni surface.168 The
promise for desulfurization at ambient temperature, and enthalpy of adsorption of H2S on Ni was measured to be about
investigation of structural stability over several regeneration 130−160 kJ/mol.164,169 Due to such high enthalpies of
cycles will help reveal the true potential of these sorbents for adsorption, adsorbent regeneration will result in a very
commercial application. expensive natural gas desulfurization process. Metal adsorbents
Even though cobalt oxide is one of the thermodynamically may be useful as a final polishing step to limit the energy duty
more favorable metals for desulfurization of H2S, attempts at for H2S adsorption/regeneration. For example, they can be
making it a major component for removing H2S are useful in certain applications such as direct-methane fuel
limited.132,162 Due to larger particle sizes, bulk cobalt oxide cells170 to achieve ultra-low sulfur content methane from pre-
showed slower sulfidation and incomplete conversion even at treated natural gas.
500 °C, but when it was supported on either Al2O3 or TiO2, Usually, when metals are employed for H2S removal, they are
higher rates and complete conversion could be observed even used as sacrificial sorbents since, during a regeneration process,
at 300 °C, primarily due to smaller grain size.162 TiO2- metals will have a strong tendency to sinter and aggregate,
supported Co3O4 showed lower activation barriers, and it may resulting in a complete loss of surface area and active sites. By
serve as an effective H2S removal sorbent, even at lower isolating and stabilizing small metal particles in cubic structured
temperatures. Pahalagedara et al. investigated mesoporous mesoporous silica, Li et al. were probably the first to
cobalt oxides (m-Co-X, where X = 150, 250, 350, or 450 °C, demonstrate the regeneration of the sulfur-loaded adsorbents
the calcination temperature) as desulfurization sorbents and with minimal sintering (see Figure 5).171 Adsorption was
could achieve 100% capacity at 150 °C.161 The reason for carried out at 300 °C using a coal syngas composition of 23%
breakthrough capacities higher than 100% metal utilization for H2, 29% CO, 8% CO2, 30% H2O, 10% He, and 10 ppm H2S,
mesoporous Co and Mn is not clear but may be attributed to and regeneration was carried out at 550 °C using a five-cycle
9767 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 6. (a)View of the MIL-53(Cr)LP/MIL-47(V) structures along the chain (z axis), highlighting the 1D pores system with M = Cr3+ or V4+.
View perpendicular to the pores of the MIL-53(Cr) (b) and MIL-47(V) (c) with μ2-OH and μ2-O atoms linked to the metal atom (M), respectively.
Reprinted with permission from ref 179. Copyright 2011 American Chemical Society.

alternating oxidation−reduction treatment of pure H2 and 10% respectively).178 SP MILs are built from corner-sharing chains
air in Ar with a short Ar purge in between. A fully regenerable of metal oxide (MIVO6) or hydroxide (MIIIO4(OH)2) octahedra
capacity of 7 mg of S/g adsorbent over five adsorption− connected through terephthalate moieties (SBET ≈ 1000 m2/g)
regeneration cycles was achieved for 16.6 wt% Ni loaded on (see Figure 6). LP MILs comprise of trimers of chromium
SBA-16, with the ratio of chemisorbed H2S to Ni being 0.07. It octahedra linked with trimesate (MIL-100) or terephthalate
is unclear, however, whether H2S adsorption was carried out on (MIL-101) (SBET > 2000 m2/g). MIL-100, MIL-101, and MIL-
the hydrogen-regenerated form of the adsorbent or on the air- 47 showed the type-I-shaped (no adsorption step) isotherm,
regenerated form that would result in a metal oxide. indicating that these are rigid structures; MIL-53(Al,Cr,Fe)
3.3. Metal−Organic Frameworks exhibited two-step adsorption isotherms, which the authors
proposed to be caused by the polar H2S molecules interacting
MOFs are porous structures resulting from coordination of strongly with the OH groups of the inorganic chains at the pore
metal ions or clusters to organic ligands.172 Bare metals exist at opening, leading to closure of the pores at low loading. This
most as nanoparticles formed out of clusters of several
hypothesis was later claimed to be verified using infrared (IR)
thousands of atoms and suffer from activity loss by way of
spectroscopy and molecular simulations.179 It should be noted
sintering during high-temperature regeneration. MOFs provide
here that, for the breathing MOF (MIL-53(Cr)), isotherms
a scheme for isolating the metal atoms or the metal oxide
were simulated for two different rigid MOF structures,
clusters by being chemically coordinated within a porous
representing the narrow and the large-pore regimes in the
framework. This allows design of adsorbents where not only
experimental adsorption isotherm. By imposing the exper-
each metal atom exclusively interacts with a single adsorbate
molecule but also each of these interaction sites may be imental fractions of each pore type as a function of pressure, a
regenerated without significant loss of activity. Additionally, composite isotherm was calculated for the simulated isotherms.
appropriately functionalized organic ligands can also contribute While this strategy yields a good match with the experimental
in promoting the selectivity toward the desired gas molecule. data, the study did not investigate the ability to predict the
MOFs offer a large parameter space to rationally design the structural transformation. Measurement of excess CH 4
size, shape, and functionality of pores with high adsorption adsorption before and after high-pressure H2S adsorption
and/or diffusion selectivity toward the molecule of interest. In followed by vacuum desorption showed full recovery of CH4
addition to selectivity, performance and structural reversibility capacity for MIL-47(V) and MIL-53(Al,Cr), partial recovery for
on regeneration are important factors in judging the perform- MIL-100 and MIL-101, and complete loss of capacity for MIL-
ance. The presence of polar impurities such as H2O on the 53(Fe) due to its decomposition into iron sulfide. At 30 °C and
adsorption performance as well as stability of the adsorbent 20 bar, MIL-101 showed an unprecedented H2S loading of 38.4
structure can also play determining roles in the successful mmol/g; MIL-53(Al,Cr), MIL-47, and MIL-100 showed
application of MOFs for desulfurization applications. Typical saturation loadings of 11.8, 13.1, 14.6, and 16.7 mmol/g,
pore diameters in MOFs are in the microporous regime ranging respectively, which are also quite high compared to those of
from 3 to 20 Å,173 although more recently, MOFs with pore other H2S adsorbents. However, in cases such as off-shore
sizes as high as 85 Å have been synthesized for inclusion of natural gas sweetening, where having a smaller footprint is
large biological molecules.174 Due to the highly porous nature important, volumetrically bulky porous MOFs suffer a draw-
of most MOFs, their lower conductivity (0.3 W/m·K for MOF- back. Adsorption being an exothermic process, sharp increases
5 at 300 K175) compared to that of crystalline materials in in temperatures for MOFs with ultra-high capacity raise
general can be a potential concern for strong host−guest concerns about bed stability in the absence of adequate heat
interactions. Desulfurization using MOFs has been reviewed in transfer. Attempts to compute thermal conductivities of MOFs
the past;14,15,176,177 here we try to summarize some of these in the presence of adsorbed gases are starting to elucidate that
topics, provide our interpretation, and identify areas for further the thermal conductivity of MOFs decreases with increasing
investigation. pore sizes.180,181
Hamon et al. carried out the first experimental study of H2S Vaesen et al. explored functionalization of MOFs for
adsorption in MOFs; these included the small-pore (SP) MIL- desulfurization application by comparing amino-functionalized
53(AlIII, CrIII, FeIII) and MIL-47(VIV) (∼11 Å) and the large- titanium terephthalate MIL-125(Ti)-NH2 versus its parent
pore (LP) MIL-100(Cr) and MIL-101(Cr) (25 and 30 Å cage MIL-125(Ti) analogue.182 The frameworks consists of
diameter and 4.8×5.8 and 12.5×12.5 Å pore apertures, tetrahedral (4.7 Å) and octahedral (10.7 Å) cages, accessible
9768 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

through triangular windows of 5−7 Å. The cage diameters for additional saturation capacity, that is predicted from
the functionalized material are slightly smaller those of the simulations, may or may not be accessible; the difference in
parent material (6 and 12 Å for the tetrahedral and octahedral the adsorbed amounts of CH4 and CO2 for the material
cages, respectively). Grand Canonical Monte Carlo (GCMC) activated at 260 or 330 °C was negligible. Although MIL-
simulations were carried out using force fields that were shown 68(Al) is a polymorph of MIL-53(Al) (described above), unlike
to yield a decent agreement with the single-component the latter, the former does not seem to exhibit framework
experimental adsorption isotherms and enthalpies. The flexibility, as evidenced by its type-I adsorption isotherm. This
GCMC-predicted binary H2S/CH4 selectivity values (at 10 brings to light the importance of MOF framework topology in
bar, 30 °C, and yH2S = 0.001) for MIL-125(Ti) and MIL- addition to the choice of the metallic cluster and the organic
125(Ti)-NH2 were reported to be 40 and 67, respectively. The ligands, further iterating the importance of powerful screening
H2S adsorption enthalpy for these materials (∼30 kJ/mol) is of adsorbents using existing and new computational techniques
comparable to that for the aluminosilicate zeolite 4A (∼35 kJ/ that capture these subtle differences.
mol).183 These MOF materials are thermally stable up to 300 Cu-BTC (or HKUST-1 or MOF-199) comprises of an
°C, but only up to 100 °C in water vapor. Their performance intersecting three-dimensional system of large square-shaped
and structural stability in the presence of both H2S and H2O pores (9 × 9 Å) and small tetrahedral pockets (5 Å) formed
require further investigation. Furthermore, the moderately high with copper as the inorganic metal and benzene-1,3,5-
CO2/CH4 selectivity (∼7) allows for envisaging a one-step tricarboxylic acid (BTC), also known as trimesic acid, as the
process for the concomitant elimination of CO2 and H2S. organic linker (see Figure 8) with a reasonable degree of
Obtaining separate CO2-rich and H2S-rich streams unnecessa- thermal stability (up to 240 °C).185 In the past few years,
rily increases the separation costs if the end goal for H2S is several research groups have investigated various aspects of H2S
high-pressure reinjection into empty wells, instead of the adsorption in this material.186−192 Petit et al. first studied
conventional Claus process for sulfur recovery. HKUST-1 and its composites with graphite oxide under
Yang et al. studied H2S adsorption in MIL-68(Al) at room ambient conditions using 1000 ppm H2S in moist air.186 The
temperature using molecular simulations and experiments.184 authors commented that bed pre-humidification does not
This framework consists of two kinds of channels, triangular prevent the adsorption of H2S (2.7 mmol/g); on the contrary,
(6.0−6.4 Å) and hexagonal (16−17 Å), running along the c axis it favors its retention via dissolution in the water film.
(see Figure 7). It was found that only some of the hydroxyls Additional H2S measurements in absence of moisture will
help better understand the role of H2O. The XRD pattern
revealed that exposure to H2S results in a collapse of the MOF
structure. From the decrease in pH on H2S adsorption and
from Fourier-transform infrared (FTIR) spectroscopy, it was
concluded that the BTC ligands were released (no more
coordinated to copper) to their acidic form.
Watanabe and Sholl investigated chemisorption of several
molecules, including H2S and H2O, in Cu-BTC using plane
wave periodic DFT.187 H2O and H2S showed similar binding
energies of 51 and 47 kJ/mol, respectively. Single-component
GCMC adsorption simulations at room temperature, computed
by restricting the chemisorbed molecules in the geometries
obtained from prior DFT calculations, showed that the onset
pressure for H2O adsorption is nearly 3 orders of magnitude
lower than that for H2S. The influence of physisorbed
molecules on the adsorption isotherm was not considered in
this study. H2S models in the literature, trained to reproduce
the bulk liquid densities and vapor pressures, significantly over-
estimate the gas-phase dipole moment of H2S. Gutiérrez-
Sevillano et al. developed new H2S models to test whether a
model with an accurate gas-phase dipole moment could predict
Figure 7. View of the crystalline structure of the MIL-68(M) (M = V,
a strong interaction of H2S with Cu-BTC, in line with the
Ga, Fe, or Al) along the c axis: green and red circles denote the experimental observation.188 The simulations used generic
triangular and hexagonal channels, respectively (metal polyhedra, pink; force fields for the MOF structure that have been shown to
C, gray; O, red; H, white). Reproduced with permission from ref 184. perform reasonably well for some of the other adsorption
Copyright 2012 The Royal Society of Chemistry. studies in MOFs.184 Additionally, their hybrid DFT calculations
showed that indeed H2O has a higher affinity for the copper
were available for interaction, as evidenced by the computed sites (extra 3.3 kJ/mol) compared to H2S. Ab initio molecular
isotherm, obtained by blocking the triangular channels, dynamics (MD) studies of the molecules adsorbed on the
matching exactly with the experimental isotherm. It can also model cluster showed H2O to be closer to the copper center
be observed that the saturation loadings are less sensitive to the compared to H2S by about 2.6 Å. Another hybrid DFT study,
choice of force field parameters than the Henry’s law region with a damped 1/r6 term to account for London dispersion,
slope. This partially activated material was fully regenerable for showed that the interaction energies for H2O and H2S
five consecutive cycles with a pure-component H2S capacity of adsorption on Cu-BTC are −60.9 and −52.2 kJ/mol,
12 mmol/g at 30 °C and 12 bar and an adsorption enthalpy of respectively, reiterating the dominance of H2O.189 Different
∼21 kJ/mol. Even if the material can be fully activated, the 25% theoretical treatments are in agreement that the interaction of
9769 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 8. Crystal structure of dehydrated Cu3(BTC)2 view along the [100] direction (left). Carbon, hydrogen, copper, and oxygen atoms are
represented by the blue, white, brown, and red spheres, respectively. The paddlewheel structure of a Cu dimer in Cu3(BTC)2 with water molecules
coordinated to the metal centers is shown on the right. Reprinted with permission from ref 187. Copyright 2010 AIP Publishing.

Figure 9. H2O (left) and H2S (right) molecules as arranged in the CPO-27-Ni channels. Metal, sulfur, oxygen, carbon, and hydrogen atoms are blue,
yellow, red, gray, and white, respectively. Reprinted with permission from ref 195. Copyright 2013 American Chemical Society.

Cu-BTC with H2S is weaker than that with H2O, thus failing to (0% relative humidity) and wet (80% relative humidity)
explain the experimental observation that H2S converts Cu in conditions.191 In spite of a lot of research effort to understand
Cu-BTC to black CuS in the presence of moisture.186 An the role of moisture in desulfurization using Cu-BTC, this
important point to note here is that the theoretical studies only problem remains unresolved.
considered the host−guest interaction; however, H2S reacts Eddaoudi et al. synthesized a series of 16 isoretucular (having
chemically with copper to form CuS, and the thermodynamics the same network topology) MOFs (IRMOF-n, n = 1−16)
of product formation have not been attempted in any of the based on a skeleton of Zn-based MOF-5, wherein the pore
theoretical studies reported so far. functionality and size were varied without changing the original
Li et al. experimentally investigated the same system by cubic topology.173 Theoretical investigation of H2S was
activating the material before carrying out H2S adsorption.190 performed in IRMOF-1 (or MOF-5) having 1,4-benzene-
They found that the breakthrough capacity increased by 38% dicarboxylate organic linkers; the heat of adsorption was ∼ −15
on activation at 180 °C compared to the unactivated sample kJ/mol, much lower than that for Cu-BTC (∼ −30 kJ/mol)
and decreased by 10% if the sample was activated further at 200 due to wider (11.2 Å) pores of IRMOF-1 compared to confined
°C. The breakthrough capacity also increased by 40% when the (5 Å) tetrahedral-shaped pockets in Cu-BTC. IRMOF-1 also
adsorption temperature was raised from 30 to 80 °C. Both of results in ∼6 kJ/mol weaker binding than H2O, suggesting the
these observations suggest that there is a competitive difficulty in using this material for H2S adsorption from moist
adsorption between H2O and H2S, and lowering the adsorbed natural gas.188 Wang et al. studied IRMOF-3 (free pore
quantity of moisture, either with an increased activation diameter of 9.6 Å) having 2-amino-1,4-benzenedicarboxylate
temperature or with an increased adsorption temperature, ligands and showed that the breakthrough capacity increased by
increases the H2S capacity. Contrary to this inference, Peterson 46% on activation at 150 °C compared to that of the
et al. observed no change in H2S adsorption capacity under dry unactivated sample and decreased by 93% if the sample was
9770 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

activated further at 200 °C.193 The breakthrough capacity for such highly charged systems, and experiments will help to
decreased to 19% when the adsorption temperature was confirm the validity of the predictions as regards such high
increased from 30 to 60 °C. It would be interesting to study selectivity and capacity as well as stability of MOF in the
comparative H2S adsorption in the two MOFs (IRMOF-1 and presence of impurities such as moisture.
IRMOF-3) to understand the role of the −NH2 group in Very recently, Belmabkhout et al. investigated In3+-, Fe3+-,
IRMOF-3 for selective binding with acid gases. Ga3+-, and Al3+-based socMOFs as adsorbents and thin-film
Allan et al. studied desulfurization using Ni-CPO-27 and Zn- membranes for the production of high-quality hydrocarbons
CPO-27, where Ni/Zn cations are linked by 2,5-dihydroxy- (CH4, C3H8 and n-C4H10, and olefins) from H2S-containing gas
terephthalic acid organic linkers, forming a honeycomb lattice streams. Only Ga3+- and Al3+-based socMOFs could preserve
with channels of ∼11 Å pore diameter.194 In the hydrated state, the crystal structure on exposure to H2S.197 The authors
each metal is coordinated by one H2O molecule, but upon measured pure-component adsorption isotherms for H2S, CO2,
activation at 150 °C under vacuum, H2O is removed, leaving and CH4 in Ga3+-socMOF at 25 °C and found that, unlike CH4
the metal in square pyramidal geometry with an open metal and CO2, H2S showed a steep adsorption. The authors also
site, resulting in SBET = 1193 and 379 m2/g, respectively. H2S carried out breakthrough measurements for a ternary mixture
exposure did not degrade the crystalline structure of Ni-CPO- containing 5% H2S, 5% CO2, and 90% CH4 for six adsorption−
27 but amorphized Zn-CPO-27.194,195 Ni-CPO-27 showed a desorption cycles with a combination of temperature and
steep rise in loading up to 6.4 mmol/g (corresponding to each vacuum swing regeneration. Optimal regeneration was achieved
metal center interacting with one gas molecule, see Figure 9), at 160 °C under a helium flow for 8 h with complete
indicating a sharp chemisorption regime followed by a slower recyclability of the adsorbent. Since the mass of adsorbent used
uptake likely due to physisorption up to 12 mmol/g at for the experiment is unclear (between 0.1 and 0.4 g), the
atmospheric pressure. Adsorbents with such high loadings at reported H2S breakthrough time of 40 min translates to a
low pressures are rare, and it may be useful in effectively breakthrough capacity between 2 and 8 mmol/g. CH4 and CO2
removing large amounts of H2S from gas streams with a low breakthrough occurred at nearly 0 and 5 min, respectively,
H2S partial pressure. However, the material cannot be fully suggesting very high H2S/CH4 and H2S/CO2 selectivities. The
regenerated at outgassing temperature of 150 °C and a vacuum narrow window aperture (4.5 Å) to reach the cages of the of
of 0.1 Pa,194 but it could be nearly completely regenerated at Ga3+-socMOF may have an influence on the breakthrough
200 °C.195 The authors hint at some damage to the framework capacity and selectivity. Al3+-socMOF was shown to be stable
by this high-temperature regeneration; additional microscopy up to temperatures as high as 300 °C and relative humidity as
and diffraction data for the regenerated samples should help high as 95%, and even soaking in liquid water for 2 weeks did
understand the degree of damage and whether the structural not affect the XRD pattern. However, the H2O stability of the
damage increases with the number of regeneration cycles. The Ga3+-socMOF was not reported. These materials show promise
heat of adsorption is nearly constant (56−58 kJ/mol) up to a for regenerative sour gas sweetening using MOFs.
loading of 0.7 H2S molecule per atom of Ni, assuming all Ni Stability of adsorbent in the corrosive environment of acid
atoms are activated. The authors comment that, beyond this gases in the presence/absence of moisture is arguably the most
loading, the drop in interaction strength can be justified with important test that the adsorbent must meet in order to qualify
typical weak physisorption with no specific interaction between as a potential candidate for sweetening applications. The not-
the host/guest species. However, it is unclear how the so-strong coordination bond between the metal and the organic
physisorption regime can start without reaching a loading of ligand in a MOF structure make it very susceptible to
1 H2S molecule per Ni atom, and this could possibly indicate hydrolysis and poor stability. Burtch et al. have recently
incomplete activation of the MOF (∼70% activated). published a very detailed paper in Chemical Reviews on the
Compared to other MOFs investigated so far for H2S hydrothermal stability of MOFs.198 Apart from a few studies on
adsorption, Ni-CPO-27 shows the highest uptake and high acid gas stability of certain MOFs,199 this area is not very
enthalpy of adsorption (indicative of possibly higher selectivity mature for a comprehensive review. Future efforts in this
over methane) with structural stability. However, the direction will not only help advance the development of
experimental heat of adsorption of H2O for this material is adsorbents for commercial purposes but also enhance the
< −100 kJ/mol, which may or may not make this material an fundamental understanding of MOF stability.
ideal adsorbent for natural gas in the presence of moisture. Zeolitic imidazolate frameworks (ZIFs), a sub-class of MOFs,
Nonetheless, the material appears to be very promising and composed of transition metal ions that replace the silicon/
should be investigated for H2S adsorption in the presence of aluminum atoms in a zeolite topology and organic linkers in
H2O. place of the oxygen atoms in the zeolite lattice, are a relatively
Using GCMC simulations with generic force fields that have new type of materials that benefit from the structural variety of
not been thoroughly validated for any particular property, Peng MOFs and stability of zeolites.200,201 ZIF-8, with a BET surface
and Cao carried out a screening of some select porous carbons, area of 1630 m2/g, a high thermal stability (up to 550 °C), and
zeolites, and MOFs for a feed containing 0.2% H2S, 29.8% CO2, remarkable chemical resistance to boiling alkaline water and
and 70% CH4 at 50 bar and 30 °C.196 Among the MOFs, organic solvents,201 is one of the very extensively studied ZIF
indium-based rho-zMOF (a zeolite-like MOF with the topology materials in the literature. Recently, Ethiraj et al. investigated
of rho-zeolite) and socMOF showed high selectivity (175 and the stability of HKUST-1 and ZIF-8 MOFs in an H2S
100, respectively) and capacity (2.6 and 1.8 mmol/g, environment using various techniques, including XRD, and
respectively). The capacity for rho-zMOF at temperatures of observed a significantly higher stability of ZIF-8 over HKUST-
60 °C and higher is reported to be 0.5 mmol/g, suggesting a 1.192 Other microporous materials, such as covalent organic
working capacity of about 2 mmol/g with very low temper- frameworks202 and porous aromatic frameworks,203 that are
atures of regeneration. While this is an important discovery, low-density, high-surface-area materials but do not contain any
there is a considerable uncertainty in the force-field parameters selective metallic sites, can be explored for bulk H2S removal.
9771 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 10. Selectivity (left axis) and ΔHads (right axis) in top-performing zeolite structures at yF = 0.50, T = 343 K, and p = 50 bar. SH2S/CH4 cyan
triangles, SH2S/C2H6 magenta squares, and ΔHads (for the H2S/CH4 mixture) green bars. Reprinted with permission from ref 205. Copyright 2016 John
Wiley & Sons, Inc.

As noted earlier, feed streams with varying overall pressure as of the chabazite (CHA) framework.207 This material demon-
well as concentration of H2S are likely to require very different strated a high H2S capacity of 4 mmol/g at 25 °C and 2 bar,
sweetening procedures. While low-concentration (10−1000 suggesting that all-silica zeolites maybe promising materials for
ppm) breakthrough measurements provide important data for bulk H2S removal. Shah et al. carried out Monte Carlo
applications such as Claus tail gas cleanup, data at higher H2S simulations of H2S and CH4, both pure-component and
partial pressures are also important for applications involving mixture, over a wide range of temperatures, pressures, and
ultra-sour gas streams. A major difficulty in drawing a fair compositions in seven different all-silica zeolites.208 The
comparison between different MOFs in the literature for H2S authors showed quantitative agreement between predictive
adsorption arises due to the unavailability of complete molecular simulations without any fitting parameters and the
adsorption data over a wide range of partial pressures, from a experimental data reported by Maghsoudi et al., thus validating
few ppm at atmospheric pressure to 100% H2S at several the force fields for H2S,209 CH4,210 and the zeolite.211 Shat et al.
atmospheres in the presence/absence of moisture, different also assessed the application of ideal adsorbed solution theory
definitions of breakthrough capacity, etc. Future studies should (IAST) to pure-component data in order to predict mixture
account for factors such as moisture, regeneration for several data and found differences as high as 60% in case of some
cycles, and XRD measurements before and after each (or at zeolites. In order to probe whether the presence of H2S induces
least, after every few) regeneration. H2O coadsorption and reduces the hydrophobic character of
3.4. Zeolites
all-silica zeolites, binary H2S/H2O mixture adsorption in MFI
was also studied. The simulations showed preferential
For several decades, zeolites have been used as catalysts in the adsorption of H2S over H2O with a selectivity of about 18,
petrochemical industry for cracking, isomerization, and suggesting that these zeolites can be promising materials for
alkylation applications. These materials are also widely used sour gas sweetening. In a subsequent study, Shat et al. carried
for water purification and softening in the detergent industry. out a large-scale computational screening of the adsorption of
Structurally, zeolites are crystalline porous aluminosilicates binary H2S/CH4 and H2S/C2H6 mixtures in the all-silica forms
consisting of a three-dimensional framework of SiO4 and AlO4 of 386 zeolitic frameworks (see Figure 10).205 The authors also
tetrahedra. Their pores are defined by their crystal structure investigated the adsorption of four- and five-component
and have precise sizes and shapes, allowing them excellent mixtures at the conditions representing sour gas fields for 16
sieving properties at a molecular level. As of today, the database promising all-silica zeolites selected from binary adsorption
of the International Zeolite Association−Structure Commission selectivities and capacities. It was found that, depending on the
identifies 232 unique zeolite frameworks, and each framework fractions of CH4, C2H6, and CO2, different sorbents allow for
is represented by a three-letter code.204 The International optimal H2S removal and hydrocarbon recovery.
Union of Pure and Applied Chemistry classifies nanoporous Liu et al. carried out a multi-scale modeling study for
materials into three classes: microporous (<2 nm), mesoporous combined material screening and process optimization to
(2−50 nm), and macroporous (>50 nm). Zeolites are reduce the overall process cost with atomistic and mesoscopic
microporous in nature, and depending on the particular scales for the selection of zeolites and the macroscopic scale for
framework structure, the maximum diameter of a sphere that the selection of optimal conditions for pressure swing
can diffuse through the zeolite framework can range from 1 to 8 adsorption.212 Several high-performing zeolites were identified,
Å. The maximum diameter of a sphere that can be included in and zeolite ABW ranks as one of the top materials for the
the zeolite structures can range from 3 to 16 Å. removal of both H2S and CO2 (acid gases) from natural gas.
3.4.1. All-Silica Zeolites. All-silica zeolites have a very low There is a considerable degree of overlap between the zeolites
polarity due to negligible amounts of polar cations and silanol identified from the two screening studies in the literature for
defects, imparting a highly hydrophobic character to this class high H2S/CH4 selectivities.205,212 The best-performing zeolites
of zeolites.206 Maghsoudi et al. first investigated pure- will likely always be from this pool of structures identified with
component adsorption of H2S and CH4 in an all-silica zeolite high selectivity over methane; however, the optimal adsorbent
9772 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 11. Schematic of dissociative H2S adsorption on an aluminosilicate zeolite.

will depend on the feed conditions (wet/dry, lean/rich in light higher H2S adsorption capacities for X-type zeolites than for Y-
hydrocarbons, CO2/H2S proportions). type zeolites were observed; moreover, the uptake changed in
Of all the top-performing structures presented in Figure 10, the following sequence: K-FAU > Na-FAU > Li-FAU,
CAS, DOH, ITE, ITH, ITW, MEL, MEP, MFI, MTN, RRO, suggesting that H2S adsorption improves with FAU basicity.223
RWR, and SFF frameworks have been synthesized in their all- Garcia and Lercher studied adsorption of H2S in ZSM-5
silica form, IWV with a Si/Al ratio of 29, SEW with a Si/B ratio zeolites (Si/Al = 35.5) and observed molecular adsorption on
of 13, and ABW, BIK, EDI, EPI, GIS, JBW, LTJ, LTL, MER, both Na- and H-ZSM-5 zeolite surfaces.221 In conclusion, lower
MON, MOZ, PAU, PHI, and UFI with a low to moderate Si/Al Si/Al ratios promote dissociative adsorption of H2S on cation-
ratio.205 Although DOH, MEP, and MTN show a high exchanged zeolites. Recently, Sigot et al. attempted to
selectivity, it is unlikely that the pores in these structures investigate the retention mechanism of H2S on NaX zeolites
with only six-membered ring openings will allow accessibility to with Si/Al ratio of 1.4.224 The authors observed poor thermal
the selective sites. Development of experimental syntheses for desorption at 350 °C, indicating no physisorption and the
the all-silica defect-free forms of these top-performing zeolite formation of stable sulfur polymers which further compromised
topologies can further the development of these hydrophobic thermal regeneration.
adsorbents for the bulk acid gas removal from ultra-sour gas The SH vibration band of H2S adsorbed on NaA showed a
fields. large frequency shift; however, in the case of CaNaA, over a
3.4.2. Aluminosilicate Zeolites. Aluminosilicate zeolites wide range of coverage, the characteristic frequency of the SH
are crystalline microporous frameworks, based on a three- band remained constant at 2540 cm−1.217 Thus, the authors
dimensional network of silica and alumina, that can concluded that no significant sorbate−sorbate interactions
accommodate a variety of cations such as Na+, K+, Ca2+, etc. occur and that all H2S molecules occupy equivalent sites in the
In addition to the precise sizes and shapes of pores in all-silica case of CaNaA. No bands corresponding to adsorbed species
zeolites that allow them excellent sieving properties at a were observed for NiNaA and CuNaA, and CoNaA adsorbed
molecular level, the cations in aluminosilicate zeolites may non-dissociatively, while ZnNaA and MnNaA demonstrated
allow the possibility of strong selective interactions with the both dissociative and non-dissociative adsorption.220 In agree-
desired polar gas molecules. There is a significant amount of
ment with previous studies of H2S adsorption on various
literature exploring the Claus reaction between H2S and SO2 to
zeolites,216−218 water formation was observed on adsorption of
produce elemental sulfur using cation-exchanged zeolites.213,214
H2S onto MnNaA, ZnNaA, and CoNaA. It is not very clear,
While this is a very important aspect of sour gas sweetening, it
however, whether H2O is formed using either framework
has been reviewed in the past,20 and we will focus here on the
oxygen or oxygen impurities in the gas feed. It is unclear if
interaction of these zeolites with H2S in the absence of oxygen.
Ozekmekci et al. have presented a short review on H2S metal oxide clusters of the exchanged cations are present in the
removal using zeolites.215 We try to expand on this subject system and whether these form a source of oxygen for water
here. Identifying the adsorption mechanism of H2S in cation- formation. Very recently, Wynnyk and co-workers carried out
exchanged zeolites has been a challenge, whether it is a precise and detailed measurements up to high pressures for
physisorption, chemisorption by coordination to the cations, or adsorption of sour gas species (CO2, COS, CH4, and H2S) on
a chemical reaction forming H2O using either framework zeolite 4A (LTA) with Si/Al ratios between 1.14 and 1.15 and
oxygen or gas-phase oxygen impurities or possibly even metal activated under ultra-high vacuum (10−10 bar) up to 400 °C.183
oxides of the extra-framework cations. In an attempt to The adsorption isotherms of these compounds were found to
understand the adsorption mechanism, several zeolitic systems be quite different, with saturation loadings extrapolated from
based on faujasite (FAU; NaX, NaY, HY), LTA (NaA, CaNaA, the fitted adsorption isotherm for CH4, CO2, H2S, and COS
ZnNaA, MnNaA, CoNaA, NiNaA, CuNaA), and MFI (H- being 9.16, 7.25, 5.87, and 3.11 mmol/g, respectively. The low
ZSM-5, Na-ZSM-5) were probed using IR,216−221 UV/vis,219 saturation capacity of COS was not expected; computational
and nuclear magnetic resonance (NMR) spectroscopies.222 calculations for HOMO bond orientation indicated that COS
On aluminum-rich FAU up to Si/Al ratios of about 2.5 (NaX should be excluded from zeolite 4A, despite the experimental
zeolites), the first H2S molecules are adsorbed dissociatively, results. The reason for decreasing saturation loadings for CO2
producing OH groups (3650 cm−1) similar to those of with decreasing temperatures (0, 25, and 50 °C) was not
hydrogen FAU; however, NaY-type zeolites with Si/Al > 2.5 explained by the authors. The authors also calculated the
adsorb the H2S molecules practically without dissociation.218 enthalpy of adsorption using isotherms at different temper-
Similar conclusions were also reported by Lechert and Henning atures; at low loadings, CH4 and COS yield similar values of 20
using the NMR technique.222 The first molecules adsorbed on kJ/mol, while CO2 and H2S interact more strongly, with values
NaX dissociated completely with formation of S2− and 2OH, of ∼35 kJ/mol (near saturation loading). These enthalpies of
but at higher coverages, dissociation into HS− and H+ (see adsorption seem too small to suggest a chemisorption behavior;
Figure 11) as well as physical adsorption occurs.219 Four times however, it should be noted that these interactions can be
9773 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

sensitive to even trace amounts of H2O that may accompany

the feed gas.
H2S adsorption on cluster models of metal-exchanged ZSM-
12 (M = Fe, Co, Ni, Cu, and Zn atoms) was investigated using
DFT calculations, and the reported enthalpies of adsorption
were −77, −77, −90, −58, and −76 kJ/mol, respectively.225
This suggests that Ni binds most strongly to H2S, while the
binding with Cu is the weakest. One can see that the trends for
different metals are not the same as reported for cation-
exchanged LTA zeolites described above.220 A key issue here is
that we are not certain of the fraction of the cations in the ion-
exchanged sites versus that in the metal oxide form. Several
other factors, such as difference in the confinement of LTA and
MTW zeolites for the same metallic species, presence of
impurities in experimental measurements, or inadequate level of
theory to describe the system using DFT, can also contribute to
the different trends in the two zeolite types. In the absence of
more data, it is hard to draw concrete conclusions about the
affinity of various transition metal cations toward H2S.
SAPO-43, a small-pore silicoaluminophosphate, is a micro-
porous sorbent with appropriate dimensions for sour gas
sieving; however, its low thermal stability up to about 300 °C
prevents its utilization for adsorptive separations. Elemental
Figure 12. H2S adsorption capacities of NaX, AgY, AgX, CuX, and
analysis and CO2 heats of adsorption point to the presence of CuY zeolites for 10 ppmw H2S in He and N2 in the absence and
amine-like compounds entrapped in the surface of SAPO-43, presence of CO2 (20%), CO (2%), and H2O (2%) in the feed at (a)
which are believed to keep the framework from collapsing.226 room temperature and (b) 150 °C. Reprinted with permission from ref
Hernández-Maldonado et al. demonstrated the use of partially 229. Copyright 2011 Elsevier.
calcined SAPO-43 to separate CO2, H2S, and H2O from natural
gas.226 Regeneration with pure helium at 180 °C was sufficient CO2, CO, or H2O, while CuX and CuY showed similar losses
to remove almost all of the adsorbed water, but only 40% of the in selectivity in the presence of CO. AgX and AgY zeolites
adsorption capacity was restored for H2S. The authors suggest maintained H2S capacities of about 30−40 mg/g at 10 ppm
that regeneration using chemicals such as hydrogen peroxide H2S, even in the presence of all three impurities. The authors
(H2O2) may help in entirely removing H2S without the need to could also explain the observed experimental trends using DFT
thermally regenerate the adsorbent. calculations; comparisons of the structures and energetics
Crespo et al. compared the H2S breakthrough capacities between the cluster and periodic calculations led to the
among Cu(I)Y, Cu(II)Y, CuCl/MCM-41, CuCl/SBA-15, and conclusion that the confinement effect from the complete
AgY.227 It was found from both ab initio molecular orbital zeolitic framework is insignificant for the adsorbates studied.
calculations and the experimental temperature dependence of Further DFT calculations showed that, while the presence of
adsorption isotherms that the energy of adsorption decreased in the zeolite does not change radically the trend in the adsorption
the following order: Cu(I)Y > AgY > Cu(II)Y. The chemical energies for alkali metal cations, in the case of transition metal
composition of AgY zeolites, Ag63Al57Si135O384, clearly shows cations it enhances the CO adsorption markedly relative to all
that the Ag/Al ratio is greater than unity; this suggests thats other adsorbates.230 The authors also investigated other cation-
some of the cations are located at sites other than the cation- exchanged zeolites such as Zn(II)Y, Ni(II)Y, and Ga(III)Y and
exchange sites. Structural characterization of cation-exchanged found that they are all subject to strong adsorption of H2O.231
zeolites228 may help in better understanding some of the It should be noted here that strict ion exchange (without
anomalies in trends in affinity of H2S toward various cations. formation of metal oxides) has not been demonstrated for these
Nearly 70% regeneration of Cu(I)Y zeolite was achieved by cations.
heating at 450 °C in a helium environment, and 100% In addition to the competitive adsorption in the presence of
regeneration could be accomplished by reacting with air at 350 impurities, Bülow and co-workers reported on reactive
°C; this is another indication of cations in their oxide form. adsorption of H2S and CO2 on cation-exchanged LTA and
With a high H2S capacity at concentrations as low as 100 ppm, FAU zeolites.232−234 Due to strongly preferred sorption of H2O
high selectivity over CH4, and being fully regenerable at a on these zeolites (H2O ≫ H2S ≥ CO2 ≈ COS), the
moderate temperature (350 °C), Cu(I)Y, CuCl/SBA-15, and equilibrium of the reaction H2S + CO2 ⇄ COS + H2O is
CuCl/MCM-41 materials hold some promise for natural gas shifted to the right. As a result, one sulfur-containing species
desulfurization. (H2S) is transformed to another (COS), with nearly zero
In addition to the H2S impurity, most natural gas fields also enrichment of sulfur on the molecular sieve. The authors have
contain other impurities such as CO2, CO, and H2O, and these extensively studied different variations of the adsorbent such as
polar molecules may have a tendency to compete with, and type of zeolite, Si/Al ratio, choice of cation for exchange, degree
sometime even supersede, H2S adsorption. Kumar et al. carried of ion exchange (percentage of Na + exchanged), and
out a detailed investigation of such mixture systems for Na-, presorption of NH3. At room temperature, H2S conversion
Ag-, and Cu-exchanged FAU at high and low Si/Al ratios (see on LTA can be as high as 50% for a Si/Al ratio of 1, and drops
Figure 12).229 The authors report a nearly complete loss in H2S to nearly zero for a Si/Al ratio of 3.232 At higher degrees of ion
adsorption capacity for NaX zeolite in the presence of either exchange, there is a considerable kinetic barrier, and one can
9774 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

observe H2S conversions anywhere between 20 and 75%, Turk et al. have demonstrated the injection of basic
depending on the reaction time.232,233 Thus, in the absence of chemicals such as ammonia and caustic solutions to enhance
any known method to suppress this reaction, the authors H2S adsorption.239,240 There are numerous investigations from
suggest short-time techniques such as PSA/VSA.234 This study the Bandosz group on removal of H2S using ACs that differ in
suggests that future investigations of H2S breakthrough their surface areas, pore volumes, and surface acidities.241−246
behavior in cationic zeolites should also monitor COS The heat of H2S adsorption on a variety of ACs, differing in
breakthrough. surface chemistry, pH, or the presence of specific oxygenated
3.4.3. Titanosilicates. In 1989, several synthetic titano- groups, was shown to be between 39 and 47 kJ/mol, about 2
silicates were introduced as oxides or molecular sieves that have times that obtained for a nonporous surface.242 Adib et al. tried
not only tetrahedrally coordinated sites that are characteristic of to increase the basicity of AC by impregnating it with urea;
zeolites such as aluminosilicates or aluminophosphates but also however, the H2S conversion decreased by about 10% from 298
octahedrally coordinated framework sites.235 These materials to 272 mg/g for the impregnated samples compared to the
possess moderately high surface areas (SBET ≈ 200 m2/g), and initial sample, possibly due to an 18% decrease in surface area
if active metals are well-dispersed on these supports, high metal on urea impregnation.244 Tian et al. chemically modified the
utilization for reactive adsorption can be attained. Rezaei et al. surface of AC using 3-aminopropyltriethoxysilane (APTES),
have recently investigated low-temperature reactive H2S thus imparting a basic character to the sorbent with the goal of
removal using copper-exchanged Engelhard titanosilicates enhancing interaction with acidic H2S.247 The authors observed
(ETS-2, ETS-4, ETS-10).236 While ETS-4 and ETS-10 are poor breakthrough capacities of <1 mg/g at 1600 ppm, maybe
microporous, ETS-2 is most probably just a layered titanate and due to blockage of pores because of the bulky APTES
is too dense to be considered a molecular sieve.235 Cu-ETS-2, impregnation. At 8200 ppm the equilibrium adsorption capacity
Cu-ETS-4, and Cu-ETS-10 show H2S breakthrough capacities of H2S is 18.13, 8.88, and 6.03 mg/g for APTES-AC, Na2CO3-
of 1.4, 0.32, and 1.3 mmol/g and copper utilization of 71, 44, AC, and AC, respectively. Moisture and temperature were
found to have adverse effects on H2S removal, while oxygen
and 98%, respectively. The high and low copper utilization in
favored H2S removal. The negative role of humidity in this
ETS-10 and ETS-4 has been attributed to their large and small
study is contradictory to previous studies of H2S adsorption on
pores, respectively, leading to pore blockage by adsorbed H2S
AC,241 while the positive effect of oxygen has been explained by
in case of ETS-4. Since ETS-10 allows near complete utilization
the authors to be due to the catalytic role of oxygen during
of copper sites at 8.4 wt% copper loading, higher loadings oxidation.
should be investigated to possibly improve breakthrough Regeneration of ACs using cold and hot water washing and
capacities. In a follow-up work, the authors investigated the using heating at 300 °C in air atmosphere showed that, after the
effect of exchange metals such as Ag, Ca, Cu and Zn on ETS-2 first adsorption run, the capacity of carbon for hydrogen sulfide
and compared to a fully developed commercial H2S adsorbent adsorption significantly decreased (around 60%) due to
(R3-11G, 36 wt% CuO, BASF).237 The results indicated H2S irreversible adsorption in the most active adsorption cites
uptake capacities in the following order: Cu-ETS-2 > Ag-ETS-2 located in small pores; subsequent runs revealed a more or less
> Zn-ETS-2 ≈ R3-11G > Ca-ETS-2 ≈ Na-ETS-2. In a recent constant capacity.248,249 Boudou et al. could regenerate the
study, Roller et al. studied Cu-ETS-2, Ce-ETS-2, and Cu-Ce- H2S-adsorbed AC sorbents by heating for 10 min in flowing N2
ETS-2 over a wide range of temperatures and under hydrogen- at 500 °C.250 Although inert atmosphere can be maintained at
rich and water-rich atmospheres and found that, contrary to laboratory scales, the possibility of self-ignition at such high
Ce-modified sorbents in the literature,153 Ce-based ETS-2 temperatures of regeneration cannot be eliminated. Aging of
showed a very low H2S breakthrough capacity.238 ACs has shown that the sorbents having significant capacity
3.5. Carbon-Based Adsorbents after the first adsorption run are practically unwashable with
water due to the deposition of bulky sulfur polymers that are
3.5.1. Activated Carbons. Activated carbon (AC) with resistant to oxidation.251
surface areas easily in excess of 1000 m2/g due to a high degree 3.5.2. Structured Carbons. Structured nanoporous carbon
of microporosity is a well-known material for adsorption and materials may be broadly classified into two main kinds:
catalysis. It is most commonly manufactured by carbonizing graphene-based materials and carbon nanotubes (CNTs).
easily available carbon sources such as coconut shell, wood, Graphene is an allotrope of carbon that consists of a two-
coal, etc. at high temperatures in an inert atmosphere. The dimensional hexagonal lattice of single atomic layer of sp2-
resultant coke has pores too small for use as an adsorbent, and hybridized carbon atoms. Graphene-based materials may be
usually steam activation is carried out to enlarge the pore further classified as pristine graphene, graphene oxide (GO),
structure, making it accessible for adsorption. Several studies and reduced graphene oxide (rGO).252 CNTs are cylindrical
have been reported in the literature on the use of ACs for H2S nanostructures that may be imagined as seamlessly rolled-up
adsorption as well as Claus reaction between H2S and O2 to graphene sheets with channel diameters as small as 0.4 nm,253
form elemental sulfur. Unlike crystalline materials such as ∼1.4 nm in several studies,254 and can be as high as 50 nm.252
MOFs and zeolites, ACs are amorphous and hence The CNTs can be open or closed with half-fullerene caps at
characterized using N2 adsorption, pH measurements, thermal either ends of the cylindrical structure. These single-walled
analysis, ion chromatography, elemental analysis, and FTIR CNTs (SWCNTs) are usually observed as closely packed
techniques. It is nearly impossible to rationalize the large bundles on a honeycomb lattice with typically 100 nanotubes
amounts of data in this field with various carbon sources, (of 1.4 nm diameter) forming bundles of <20 nm diameter.255
activation conditions, pH of adsorption, and presence of The adsorption sites in SWCNT bundles maybe classified into
unidentifiable impurities in the ACs. An attempt is made to three kinds: inside each tube, interstitial channels between the
highlight some of the studies that show high adsorption of H2S, tubes, and outer surface peripheral to the bundle. The surface
but the reasons for most of these are not well-understood. area for these materials may range from over 2500 m2/g for a
9775 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 9. H2S Breakthrough Capacities and Percentage of Theoretical Capacity Utilized in Metal Oxides Supported on Silica
adsorbent mg S/g sorbent % utilization T (°C) composition reference
0.50%Zn-SBA-15 39 1600 25 0.1% H2S in air Wang et al.275
1.24%Zn-SBA-15 65 1100
3.04%Zn-SBA-15 220 1500
4.45%Zn-SBA-15 42 190
9.00%Zn-SBA-15 34 80

0.58%Fe-SBA-15 97 2900 25 0.1% H2S in air Wang et al.276

1.36%Fe-SBA-15 120 1600
3.59%Fe-SBA-15 170 810
7.14%Fe-SBA-15 190 470
11.4%Fe-SBA-15 190 300
19.6%Fe-SBA-15 240 210
31.3%Fe-SBA-15 320 180

BASF ZnO 11 3.3 25 1−2% H2S, 33% CO or CO2 Dhage et al.277,278

Sud-Chemie ZnO 24 7.1 in H2
16.9%Zn-SiO2 51 62
16.9%Zn-SiO2 53s 64
12.1%Zn-SiO2 32s 54
0.82%Cu-16.0%Zn-SiO2 77s 93
0.82%Cu-16.0%Zn-SiO2 51 −17 −375−179
0 1
620−211−455−219
0.59%Cu-11.4%Zn-SiO2 43s 73
3.28%Cu-13.5%Zn-SiO2 78s 94
2.34%Cu-9.64%Zn-SiO2 45s 76
0.71%Mn-16.0%Zn-SiO2 50s 61
0.72%Fe-16.0%Zn-SiO2 69s 84
0.76%Co-16.0%Zn-SiO2 64s 78
0.76%Ni-16.0%Zn-SiO2 53s 64

ZnO (commercial) 290−3.21 6.60−0.731 20 2% H2S in H2 Yang and

13.7%Zn-SiO2 140−131 220−201 Tatarchuk279
0.66%Cu-13.0%Zn-SiO2 280−221 420−331
0.62%Co-13.0%Zn-SiO2 14 21
0.61%Ni-13.0%Zn-SiO2 17 25
0.57%Mn-13.0%Zn-SiO2 17 25
1.13%Ag-13.0%Zn-SiO2 28 42

BASF ZnO 11 3.3 25 1% H2S in H2 Dhage et al.280

Sud-Chemie ZnO 24 7.1
12.1%Zn-SiO2 26 44
0.26%Fe-11.75%Zn-SiO2 35 59
0.25%Mn-11.75%Zn-SiO2 37 63
0.25%Fe-0.25%Mn-11.45% 370−205−379−2310 630−345−639−3910
Zn-SiO2

12%Zn-MCM-41 9.6 16 25 200 ppm H2S in H2 Hussain et al.281

16%Zn-KIT-6 9.1 12
12%Zn-SBA-15-sphere 21 36
12%Zn-SBA-15-fiber 22 37
2.4%Zn-TNP (TiO2 NP) 5 43
8.0%Zn-ROZ3 (AC Norit) 8 20

11.2%Cu-6%Zn-SBA-15 75 87 150 102 ppm H2S in He Elyassi et al.282

10.6%Cu-5%Zn-SiO2 34 44

10%Cu-MSU-1 11 22 25 5% H2S in CH4 Montes et al.283

20%Cu-MSU-1 19 19
30%Cu-MSU-1 19 13
10%Zn-MSU-1 42 86
20%Zn-MSU-1 15 15
30%Zn-MSU-1 12 7.8

9776 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 9. continued
adsorbent mg S/g sorbent % utilization T (°C) composition reference

24%Zn-SiO2 3.8 3.2 200 100 ppm H2S in N2 Liu et al.284

46 39 300
960−611−652−533−524−535 810−521−552−453−444−455 400

Katalco 32-5 130−31−32 0.68 300 1.5% H2S in He Mureddu et al.285

IWI-24.1%Zn-SBA-15 90−151−182 15
TS-24.1%Zn-SBA-15 90−121−122 10

Katalco 32-5 60−21−22 0.45 300 1.5% H2S in He Mureddu et al.286

TS-14.5%Zn-SBA-15 100−131−142 20
TS-11.2%Fe-SBA-15 800−401−422 65

16.1%Zn-SBA-16 8.0 10 25 800 ppm H2S in N2 Li et al.287

24.1%Zn-SBA-16 5.7 4.8
32.1%Zn-SBA-16 3.4 2.2
8.03%Zn-MCM-48 4.6 12
16.1%Zn-MCM-48 6.8 8.6
24.1%Zn-MCM-48 15 13
32.1%Zn-MCM-48 8.0 5.1
8.03%Zn-KIT-6 11 29
16.1%Zn-KIT-6 14 17
24.1%Zn-KIT-6 38 32
32.1%Zn-KIT-6 29 18
a
Key to superscripts: 0, fresh sample; n (= 1−5), after n cycles; s, saturation capacity.

SWCNT to as low as 800 m2/g for bundles of infinite theories employed for the independent studies. A consistent
diameter.255 SWCNTs of varying channel diameters can also treatment of all defect types will be important in order to
arrange in the form of several concentric cylinders to form understand, quantitatively, the role of graphene defects in H2S
multi-walled nanotubes (MWNTs). adsorption.
Several groups have theoretically investigated the role of There are several studies employing DFT for adsorption of
defects in the adsorption of H2S on the graphene surface, H2S on metal-doped graphene.256,260−265 Strong adsorption
mainly for sensor applications.256−259 Depending on the was found due to a large binding energy (615 kJ/mol for 1
position of binding, adsorption energies of H2S on pristine H2S/Pt system and 75 kJ/mol of H2S for the 7 H2S/Pt system)
graphene were found to be between 10 and 16 kJ/mol, but and short Pt−H2S bond length compared to those of the H2S/
having a defect site with the sulfur atom of H2S pointing toward graphene system.260 Zhang et al. reported binding energies of
the defect resulted in a much stronger binding (88 kJ/mol).256 16, 64, 174, and 185 kJ/mol for pristine, Ca-, Co-, and Fe-
In another study, three different graphene structures (pristine, doped graphene, respectively.256 Sharma and Verma calculated
with divacancy, and with the inclusion of nitrogen in graphene) binding energies of 33, 42, 123, and 98 kJ/mol for pristine, B-,
were investigated, and the binding energies were found to be Al-, and Ga-doped graphene, respectively.261 Pang et al.
1.6, 1.3, and 2.5 kJ/mol, respectively.257 Borisova et al. reported binding energies of 15, 81, 102, and 112 kJ/mol for
investigated the effect of vacancy in graphene and found that
pristine, Cr-, Cu-, and W-doped graphene, respectively.265
H2S adsorbed by forming weak S−C covalent bonds with the
These highly energetic interactions of H2S with metal-doped
atoms surrounding the single vacancy.258 In addition to the
graphene are mainly of interest for sensor applications;
single vacancy and divacancies in graphene described above,
there can be several other defect types; one such defect is the however, they may have potential usefulness in adsorbents for
Stone−Wales (SW) defect that has four hexagons fused into ultra-deep sulfur removal of natural gas as the final polishing
two alternating pentagons and heptagons. Hegde et al. studied step for direct-methane fuel cells.170
H2S binding to the carbon site of pristine graphene, SW defect The effect of the presence of S atom on graphene on the
site, and BN sites of BN-doped graphene through first- adsorption of H2S was investigated using DFT.266 Two types of
principles DFT with van der Waals (vdW) interactions.259 systems were investigated: S-functionalized graphene, where an
While the adsorption of H2S is weak at both C and BN sites, S atom was added on pristine graphene, and S-doped graphene,
with a binding energy of 15 kJ/mol, it is significantly stronger at where an S atom was added to the single-vacancy defect of
the SW defect site with a binding energy of 26 kJ/mol. In the graphene. Pristine graphene showed adsorption energies of 34
relaxed geometries, the sulfur atom of H2S points toward kJ/mol, while S-functionalized graphene showed energies as
graphene (at a distance 3.14 Å) at C and BN sites, while the high as 170 kJ/mol due to covalent bond formation between S
molecule’s H atoms point toward graphene (at a distance 2.84 and H2S for configurations with H atoms pointing away from
Å) at the SW defect site. There seem to be significant the graphene plane. Both physisorption and chemisorption
differences in the various predictions for H2S adsorption on were observed, depending upon C−S bond distances in the
graphene; these differences may be due to different levels of H2S/S-doped graphene complex before optimization.
9777 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Molecular simulations for adsorption on structured carbons cally increasing or decreasing. Using dopants such as Mn, Fe,
in the literature give very inconsistent answers and use idealized Co, Ni, and Cu improved the utilization of ZnO, with the effect
structures that may not be experimentally realizable.267−270 from copper being most prominent, changing the ZnO
Neglect of electrostatic interactions for the adsorbent in studies utilization at saturation from an already high 64% to about
involving polar sorbate molecules may result in incorrect 93%. This increase does not seem to improve with the amount
predictions of sorbate−sorbent interactions. The potential of of copper, since the saturation capacities remained nearly
GO has not yet been extensively explored by molecular unchanged upon increasing the Cu:Zn ratio from 0.05 to 0.25
modeling, mainly due to the lack of an atomistic GO model for at the same overall metal oxide loading. DFT studies showed
theoretical calculations.271 Huang et al. have reported the that there is a significant energetic discrimination between H2S
controllable generation of atomistic GO models by a binding to a Cu2+ versus a Zn2+ site; this difference is smaller
temperature-programmed reactive MD simulation. They for H2O.278 On a ZnO surface with some Zn2+ sites substituted
further calculated the reactive adsorption of H2S and H2O/ with Cu2+, there can be a displacement of H2O with H2S at the
H2S mixtures on the as-generated atomistic GO structures.271 Cu2+ surface sites. This is yet another explanation for the
H2S was observed to dissociate on the carbonyl functional promoter effect by copper, in addition to several other
groups, in agreement with the dynamic adsorption experiments; arguments such as surface area and diffusion invoked earlier.
for the H2O/H2S mixtures, H2O molecules were preferentially Yang and Tatarchuk279 also carried out similar studies and
adsorbed to the carbonyl sites, thus blocking the availability of found significantly enhanced utilization by doping ZnO with
active sites for H2S decomposition, again in agreement with the copper or silver. In another study, they studied Fe- and Mn-
experimental findings. promoted H2S sorbents supported on silica and found higher
Mohamadalizadeh et al. investigated the sorption of pure H2S metal utilization compared to undoped ZnO; however, no
on AC, MWNTs, CNTs decorated with tungsten nanoparticles significant preference was observed for Fe or Mn.280 For the
(W-CNTs), and amino-functionalized CNTs (F-CNTs).272 first time, real-time in situ electron spin resonance (ESR)
Although these CNT-based materials have a surface area (∼100 spectroscopy was performed during reactive adsorption of H2S.
m2/g) lower than that of AC (∼1000 m2/g), the saturation It was inferred that the surface of the sorbents is enriched with
capacities for these systems (at 10 bar and 20 °C) did not differ Fe promoter cations, while the Mn3+ promoter cations are
drastically, and CNTs showed better regenerability compared located within the supported ZnO nanocrystallites. In situ
to AC. experiments over multiple cycles may help to further elucidate
3.6. Composite Materials
the mechanism behind the oscillatory behavior.
Hussain et al. investigated the effect of different supports for
3.6.1. Metal Oxides/Mesoporous Silica. While metal ZnO; SBA-15 allows much higher breakthrough capacities as
oxide sorbents show high affinity for H2S, they suffer from well as ZnO utilization.281 This further highlights the
limitations such as loss in surface area and porosity, and this importance of designing increasingly improved supports for
subsequently results in degraded performance during cyclic metal oxides and suggests that Cu-Zn mixed oxides doped on
sulfidation and regeneration. Ordered mesoporous molecular SBA-15 may demonstrate a high capacity. In addition to
sieves that allow narrow pore sizes (16−100 Å) tailored capacity, stable performance of the adsorbent over several
through the choice of surfactant, auxiliary chemicals, and regeneration cycles is an important factor that may impact the
reaction conditions were first discovered in the early economic viability of the process; in general, it is hard to
1990s.273,274 This seminal discovery enabling uniform pore reproducibly achieve the capacity of fresh adsorbents.284−286
sizes of mesoporous materials with high surface areas opened Generally, nanoparticles exhibit a tendency to grow into larger
up a new area for novel heterogeneous catalysts. crystallites during the process of high-temperature regeneration.
Wang et al. for the first time investigated adsorption of H2S Inspired by the work of Prieto et al., who developed a highly
on metal-oxide-impregnated mesoporous silica, ZnO/SBA- stable methanol synthesis catalyst of Cu-ZnO nanoparticles on
15275 and Fe2O3/SBA-15.276 The breakthrough capacities SBA-15,288 Elyassi et al. investigated this adsorbent for H2S
mentioned by those authors are not in accord with the removal, resulting in extremely high and fully regenerable
breakthrough times they report; using the details of the breakthrough capacities and ZnO utilization over five
experimental setup, we calculated new breakthrough capacities regeneration cycles.282
which are reported in Table 9. In an inert medium, H2S reacts The major drawback of such composite sorbents is that the
chemically with metal oxides, and hence, at most the adsorbed active component forms only a small fraction of the overall
amount can correspond to 100% utilization of the metal oxide mass of the material, so even though high metal oxide
on the high-surface-area substrates. However, these two studies utilizations can be achieved, the absolute amount of H2S that
show H2S adsorption far exceeding the theoretical maximum, can be removed per unit mass of the adsorbent remains low.
suggesting that Claus reaction occurs between H2S and O2 Attempts to increase the loading of metal oxide have not
from the air in the feed and the metal oxide functions as catalyst succeeded in increasing the H2S capacity without loss in %
for this reaction rather that itself reacting with H2S to form utilization.283,284,287 Future research to increase the fraction of
metal sulfide. the active material in the composite with high utilization of the
Dhage et al. studied room-temperature H2S adsorption on active component and high regenerability of the sorbents will
transition-metal-promoted ZnO, impregnated on high-surface- help develop even better materials.
area (∼550 m2/g) SiO2.277 The ZnO/SiO2 system showed 3.6.2. Supported Amines. Supported amine adsorbents
much higher absolute breakthrough capacities as well as can be broadly classified into three classes: Class 1 constitutes
percentage Zn utilization compared to their unsupported adsorbents with physically loaded amine species, Class 2
commercial counterparts from BASF and Sud-Chemie (see adsorbents contain a covalent link between the amine and the
Table 9). It was observed that the sulfur capacity oscillated, solid support, mostly via silane chemistry or via polymeric
reproducibly, with the number of cycles rather than monotoni- supports with amine-containing side chains, and Class 3
9778 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

adsorbents comprise porous supports upon which amino- desorption rates. The same authors also studied the effects of
polymers are polymerized in situ, starting from an amine- the operating temperature (22−75 °C), gas hourly space
containing monomer.289 Class 3 materials combine the higher velocity (GHSV; 337−1011 h−1), amount of PEI loading, and
amine loadings of class 1 materials and the enhanced stability of different molecular sieve supports (MCM-41, MCM-48, and
class 2 materials due to amine−support covalent linkages.290,291 SBA-15).299 The structure of the mesoporous molecular sieves
Huang et al. first reported on H2S removal using supported influences the kinetic capacity of the supported PEI sorbents;
amines by employing high-surface-area silica supports (silica large pore size and three-dimensional channel structure favor
xerogel (816 m2/g) and MCM-48 (1389 m2/g)) functionalized higher breakthrough capacity. A decrease in temperature favors
with with 3-aminopropyl (−CH2CH2CH2NH2) groups.292 both breakthrough and saturation capacities, while a decrease in
Adsorption/desorption was measured for pure-component GHSV enhances the breakthrough capacity. The amount of PEI
H2 S, CO2 , and CH4 , with and without moisture. At loading has a strong synergetic influence on the sorption
atmospheric pressure and temperature, the CO2/amino group performance: 50 wt% PEI on SBA-15 gives the best
ratio was observed to be 0.7 and 0.9 for xerogel and MCM-48, breakthrough capacity, while 65 wt% PEI on SBA-15 shows
respectively. These ratios are 0.08 and 0.05 for CH4 and 0.29 the highest saturation capacity. The developed sorbents exhibit
and 0.25 for H2S at the highest measurement pressure of 0.001 a breakthrough capacity as high as 0.79 mmol/g and a
atm (1000 ppmv) for xerogel and MCM-48, respectively. The saturation capacity of 3.02 mmol/g at 22 °C using 4000 ppm
sorbents could be almost completely regenerated by pressure H2S, 20% H2 in bulk N2. In addition, the developed sorbents
swing or temperature swing at 75 °C. The results indicated that can be regenerated easily at mild conditions (75−100 °C), with
the presence of water vapor doubled the amount of CO2 sorption capacity maintained over eight cycles.
adsorbed and barely affected the H2S adsorption. The studies discussed so far did not involve H2S adsorption
Zhou et al. explored Class 1 adsorbents by physically loading in the presence of CO2, which is known to be much more acidic
triethanolamine (C6H15O3N; TEA) on silica gel (335 m2/g).293 compared to H2S. Ma et al. have made significant progress by
Due to the weaker basicity of TEA, its sorption capacity is finding an exceptional dependence of these sorbents on
nearly 100 times less than that of amine-modified silicas.292 The temperature for CO2 and H2S competitive sorption and
breakthrough capacity doubled on loading the silica gel with 20 developing an innovative sorption process.300 These PEI/
wt% of TEA and increased 5-fold when the loading was SBA-15 sorbents show an exceptionally high CO2 loading at 75
increased to a mere 27.5 wt%. There is no clear explanation for °C (3.2 mmol/g at 0.15 atm), but the lower kinetic barrier for
this sudden rise in adsorption, except that small selective pores H2S compared to CO2, which may be attributed to the faster
that are otherwise unexposed to TEA may get filled with TEA proton-transfer reaction mechanism in the case of H2S, results
at a higher loading of 27.5 wt%. In a subsequent study, the in higher H2S breakthrough capacities at a lower temperature of
authors used this system for studying a multi-cycle pressure 22 °C. The authors proposed a two-stage process, with the first
swing adsorption setup.294 In a follow-up work, the authors column being operated at 75 °C for CO2 removal and the
synthesized SBA-15 as a high specific surface area TEA carrier second column at 22 °C for H2S removal (see Figure 13).
and reported that the amount of H2S adsorbed on the TEA-
modified SBA-15 is 1.85 times more than that on the TEA-
modified silica gel.295 Xue and Liu tried another tertiary amine,
methyl-diethyl-amine, loaded on SBA-15, and found an optimal
amine loading of 60 wt% for an H2S breakthrough capacity of
0.1 mmol/g.296 Although the adsorbed amine is stated to be
methyldiethylamine, such low loadings, comparable to those for
TEA, are more representative of methyldiethanolamine.
In another study, physically loaded polyethylenimine (PEI)
particles on mesoporous molecular sieves MCM-41 at 50 wt%
loading were explored for low-temperature H2S adsorption at
90 °C and 10 ppmv inlet concentration.297 The adsorbent
could maintain an outlet concentration of 0.04 ppmv. The
utility of these sorbents for natural gas sweetening at higher
H2S concentrations needs to be investigated. In yet another
study, linear PEI was dispersed inside the meso-channels of a Figure 13. Scheme of the experimental two-stage process for removal
SBA-15 framework, allowing interaction of amine groups in PEI of CO2 and H2S from a model fuel gas. Reprinted with permission
from ref 300. Copyright 2009 American Chemical Society.
with silanol groups on SBA-15.298 The breakthrough and
saturation capacities increased with inlet H2S concentrations of
2000, 4000, and 6000 ppmv in the ranges 0.68, 1.18, 1.48 and Yoosuk and co-workers carried out similar two-stage studies for
1.50, 2.06, 2.38 mmol/g, respectively. The saturation value for PEI loaded on low-cost fumed silica (fSi) and also investigated
amine-grafted sorbents at 1000 ppmv is 0.56 mmol/g.292 So regeneration studies over multiple cycles.301 Although the
unlike TEA, PEI-modified supports show higher affinity for authors observed high breakthrough capacities for both CO2
H2S, probably due to higher basicity of PEI, but possibly due to and H2S for pure-component adsorption at 80 and 30 °C,
higher swelling of PEI compared to TEA. This is an open respectively, in the mixture adsorption with two-column
question that can be addressed perhaps by molecular configuration, the working capacity of CO2 was enhanced
simulations. The presence of moisture proved to enhance the from 2.2 to about 3.5 mmol/g, while that of H2S decreased
adsorption capacity since only one amine group is needed per from the expected 1.2 to about 0.05 mmol/g. This reduced
H2S molecule in the presence of moisture, versus two in its performance for H2S compared to the system introduced by
absence. However, moisture also resulted in a decrease in H2S Ma et al. can be attributed to two factors: (i) the difference in
9779 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

CO2 and H2S pure-component adsorption at 22 °C for PEI/ sour gas stream into empty gas fields is preferred over
SBI-15 is much higher compared to the difference at 30 °C for conversion of H2S to elemental sulfur by the Claus process.
PEI/fSi, and (ii) the mixture feed contained a very high H2S adsorption isotherms in the temperature range of 25−80
concentration of CO2 (20%) compared to H2S (0.36%) in the °C are available from two different studies, and both predict an
work of Yoosuk et al.; these values were 2.4% and 0.40%, isosteric heat of adsorption between −40 and −20 kJ/mol with
respectively, for the work of Ma et al. These differences increasing loading; the low-loading heats of adsorption can be
demonstrate the importance of choice of process conditions in as high as −100 kJ/mol.305,306
selecting optimal sorbents. Similar to the discovery of novel Another Class 1 adsorbent study employed SBA-15 loaded
zeolites for natural gas purification through a combined with hexamethylenetetraamine for H2S removal from a stream
material screening and process optimization approach,302 containing 9000 ppm H2S in CH4; the highest loading was
process optimization in materials selection for supported observed for 50 wt% amine loading with a capacity of 0.75
amines can help discover optimal sorbents. mmol/g, stable over 10 cycles when overnight regeneration was
In spite of these exciting properties, the powder nature of conducted using N2 at 100 °C.307 In general, all sorbents in this
ordered mesoporous silica inhibits its direct usage in industrial category can be fully regenerated under mild temperature
packed columns due to very high pressure drops. Chen et al. conditions and are shown to maintain a constant capacity for
investigated hierarchical porous silica supports for PEI. They several cycles in most of the studies. Efforts toward reducing
can be easily formed as monoliths, which can be directly used in the regeneration time will increase the productivity of these
the fixed bed without any shaping process.303 The largest sorbents.
breakthrough capacity of these sorbents was 1.27 mmol/g, 3.6.3. Zeolite/Polymer. Bhandari et al. developed hollow
about 60% higher than that of PEI/SBA-15 (0.79 mmol/g) or fibers with polymer “binder”, impregnated with high loadings of
MCM-41 (0.81 mmol/g), although the test for the former was sulfur-selective zeolite sorbent “fillers” (see Figure 15).308
carried out at a lower inlet H2S concentration and a higher
GHSV.299 A new parameter, polymerization degrees of PEI
(423, 600, 15 000, 750 000), was investigated, and it was found
that the sorbent with a PEI molecular weight of 600 had the
best sorption property. The authors propose that PEI with
higher polymerization degree should have very large steric
hindrance and low fluidity, restraining the internal diffusion of
H2S and lowering the amine utilization, while PEI (423), a
mixture of linear monomers and branch polymers, may be
partially consumed or volatilized during the impregnation and
drying, resulting in the decrease of efficient amine groups.
Belmabkhout et al. studied triamine-grafted (−CH2CH2-
CH2NHCH2CH2NHCH2CH2NH2) pore-expanded mesopo-
rous silica (TRI-PE-MCM-41) for adsorption of dry CO2,
H2S, and CH4 and found a much higher affinity for CO2 and
H2S compared to CH4 (see Figure 14).304 Binary SH2S/CO2
varied from 70 to 10 upon varying the H2S concentrations
from 5 to 100 ppm.305 These materials thus provide a
promising solution to single-step H2S and CO2 removal. Such
simultaneous removal is desirable if re-injection of concentrated
Figure 15. Typical dimensions and mode of creation in a pellet packed
bed vs fiber sorbent module. Reprinted with permission from ref 308.
Copyright 2010 American Chemical Society.

Cellulose acetate formed the binder and zeolite NaY formed

the filler for these hollow fibers with high sorbent loadings and
good strength. Although the hollow fiber morphology is
representative of contact in a membrane process, it should be
noted that the fibers here are used as adsorbents. The fiber−
sorbent morphology was characterized using equilibrium
sorption, gas permeation, SEM, and EDX. These sorbents
were shown to be very effective for low-concentration (30
ppm) H2S removal, especially when compared to a packed-bed
process, and were found to be fully regenerable by application
of heat (120 °C under N2).
3.6.4. Metal−Organic Framework/Graphite Oxide. It
was envisioned that the GO and MOF composites would
Figure 14. Schematic representation of the synthesis of amine- increase the dispersive forces in a highly porous MOF structure
functionalized mesoporous silica. Reprinted with permission from ref via dense nanoporous GO layers (5−46 wt%).186 Petit and co-
306. Copyright 2010 Elsevier. workers studied GO composites with HKUST-1 and found a
9780 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 16. Schematic view of the glucose-promoted MOF-5/GO structure unit: (A) GO layer, (B) MOF-5, and (C) glucose polymer. Reprinted
with permission from ref 310. Copyright 2012 American Chemical Society.

Figure 17. TEM images of (a) pristine ZnO nanoparticles, (b) TRGZ-1 (10 wt% ZnO), (c) TRGZ-2 (30 wt% ZnO), and (d) TRGZ-3 (50 wt%
ZnO) graphene/ZnO nanohybrids synthesized by an in situ thermo-annealing method. Reproduced with permission from ref 311. Copyright 2016
The Royal Society of Chemistry.

synergistic effect on sorption capacity (9 mg/g for GO, 92 mg/ overall sorption. Indeed the adsorbent bed changed color from
g for HKUST-1, and 199 mg/g for composite with 5 wt% GO). light blue (hydrated copper) to black (CuS). Ebrahim et al.
The breakthrough capacities measured at room temperature found that adsorption capacities increased by 85% for
using 1000 ppm H2S in moist air were not found to correlate composites with sulfur-doped graphite due to the presence of
with the measured porosity of the composites, suggesting a moisture compared to dry conditions.309 Those authors
mechanism in addition to physisorption contributing to the suggested that, under moist conditions, water acts as a
9781 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

“screening” factor, blocking the direct attack of H2S/SO2 on the Table 11. H2S Breakthrough Capacities in Metal Oxide/
copper centers and preserving the structure to some degree. Activated Carbon
They found that the catalytic effect of the modified graphene
SBET
phase results in the formation of SO2 with a much higher H2S adsorbent (m2/g) mg S/g T (°C) composition reference
breakthrough capacity of 241 ± 6 mg/g.
18Fe-CNF 139 8 100 500 ppm Fauteux-
In another study, Huang et al. prepared composites using H2S in He Lefebvre
22 300
MOF-5 and GO in the presence of glucose.310 GO enhanced 28Fe-CNF 145 8 100
et al.325
the dispersive forces in MOF-5, as demonstrated by higher H2S
uptake, but beyond the optimal loading of 5.25%, GO led to AC 580 14 40 20 ppm H2S Zhang et
collapse of the MOF crystal structure. At 5.25% GO loading, 50MnO2-AC 391 110 in N2 al.326
composites with and without glucose have surface areas of 1062
and 295 m2/g, respectively; the respective breakthrough AC 641 4.6 30 3000 ppm Balsamo et
capacities of these materials using 100 ppm H2S in N2 are Cu0Zn1.0/AC 558 34 H2S in N2 al.327
130.1 and 31.9 mg/g. The authors suggest that glucose can Cu0.05Zn0.95/AC 520 37
restrain the distortion of the composite material due to Cu0.1Zn0.9/AC 549 40
polymerization of glucose under hydrothermal conditions; Cu0.25Zn0.75/AC 555 40
these short polymers fill in the structure to support the Cu0.5Zn0.5/AC 570 47
framework, and some may even react with the functional
groups on the GO (see Figure 16). While this approach
surface area was hardly affected due to the high dispersion of
provides a route to stabilize such composite materials,
the metal oxides; moderately high H2S capacities could be
performance and structural reversibility of these sorbents on achieved. Only 18−27% of the capacity could be regenerated at
regeneration need to be investigated. 400 °C, and even at a temperature of 620 °C, only 40−46% of
3.6.5. Metal (Hydr)oxide/Graphite Oxide. Several the capacity could be regenerated.
studies report on the Claus reaction between H2S and air on Complete regeneration and reusability of an adsorbent may
composites of metal (hydr)oxide and GO,312−323 with or may not be feasible; therefore, it is important to consider the
breakthrough H2S capacities as high as 150 mg/g of adsorbent eventual disposal of spent adsorbents. This becomes partic-
at 1000 ppm H2S in air. Lonkar et al. demonstrated an easy in ularly important for H2S-removal applications, since contact of
situ fabrication of nanostructured graphene/CuO324 and an improperly disposed adsorbent with air or water may result
graphene/ZnO311 composites without the use of any solvents, in a concentrated release into the environment and can have
toxic-reducing agents, or organic modifiers. The metal oxide very serious health implications. Some of the spent metal
particles in the presence of graphene had a size of sub-20 nm, oxides, such as iron sulfide, can be pyrophoric and can release
compared to metal oxide particles of ∼50 nm synthesized by acid gases into the atmosphere. Dissolution of these gases in
the same method (see Figure 17). At ambient temperature and water bodies can have unacceptable effects on the ecosystem.
with a feed composition of 0.15% H2S, 0.41% CO2, and 99.4% Some of the methods for treating spent adsorbents include
CH4, a 50 wt% composite material led to over twice the regeneration followed by inert purge, flooding with liquid
capacity of pristine ZnO (see Table 10). Regeneration of these water, and steam purging.328 The cost of an adequate spent-
high-capacity materials with the advantage of easy synthesis is adsorbent treatment method may have large bearings on this
an important aspect that should be considered in the future. additional cost.
Table 10. H2S Breakthrough Capacities of ZnO/GO 4. MEMBRANE SEPARATION
Composites311 As natural gas consumption is expected to nearly double in the
wt% GO SBET (m2/g) mg S/g coming 25 years and the raw natural gas has a varying
0 12.4 52.2 composition depending on its origin, there is an opportunity
50 119 111 for membrane technology to remove impurities including
70 208 85.8 water, carbon dioxide, nitrogen, hydrogen sulfides, and other
90 257 33.9 hydrocarbons.329 Bhide and Stern investigated the relative
100 223 2.9 economics of cellulose acetate (CA) membrane technology and
gas absorption using DEA and found that the membrane
process for CO2 removal is more economical to meet the 2%
3.6.6. Metal Oxide/Activated Carbon. Fauteux-Lefebvre specification over the entire range of CO2 feed concentrations
et al. embedded iron nanoparticles in carbon nanofilaments and (>5%).330 Presence of H2S in the feed stream resulted in an
tested these for H2S removal under a wide range of operating increased separation cost for meeting the specifications of both
conditions.325 Fe K-edge spectra showed that iron was in its H2S and CO2 (4 ppm and 2%, respectively). Unlike the
oxidized states (Fe2O3 and Fe3O4). Even at higher temperature superior performance of membranes over the entire range of
of 300 °C, low breakthrough capacities were observed. Higher acid gas concentration for CO2 alone, the presence of H2S
metal loadings of 28 wt% could not improve the H2S uptake allowed membranes to be economical only at higher total acid
(see Table 11). Zhang and co-workers studied various loadings gas (H2S + CO2) concentrations. While this analysis was
of MnO2 on AC under a wide range of synthesis and operating
carried out for CA membranes with SCO2/CH4 = 21 and SH2S/CH4
conditions and observed capacities as high as 110 mg/g for a 50
wt% MnO2 loading at only 20 ppm H2S.326 In another study, = 50,330 it should be noted that the SH2S/CH4 values for such
sorbents were prepared by finely dispersing mixed ZnO and membranes range between only 15 and 30. Membranes with
CuO onto a commercial AC at fixed total metal loading (10 wt higher selectivities and permeabilities than the CA membranes
%) and Cu:Zn atomic ratios ranging from 0:1 to 1:1.327 The employed in this economic analysis may significantly expand
9782 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 12. H2S Permeability and Permselectivity in Polymeric Membranes

polymera PH2S (mol/(m2·s·Pa)) SH2S/CH4 SH2S/CO2 T (°C) p (bar) CH4:CO2:H2S reference
Nylon-6 (polyamide) 3.0 × 10−9 30 0.94 pure Heilman et al.339
rubber hydrochloride 5.3 × 10−9 30 0.99
PVB 3.6 × 10−7 30 0.85
poly(vinyl trifluoroacetate) 1.4 × 10−8 30 1.0
polyethylene 5.8 × 10−7 30 0.97
PET 6.1 × 10−10 0 0.95
7.3 × 10−9 60 1.0
PVDC 1.4 × 10−9 30 0.93

PDMS + PEG + PSf 2.6 × 10−8 140 5.7 25 4.4 pure Kulprathipanja et al.340
PDMS + PEG + PSf (glycerol) 3.6 × 10−8 162 8.1

PDMS (Cl3C−CF3) + PEG + PSf (1.2−3.1) × 10−8 115−185 5.4−5.9 pure Kulprathipanja et al.341
PDMS + PSf 4.7 × 10−8 9.9 2.3

CA 15−20 40 14.6 pure Schell et al.332

PDMS 2.9 × 10−10 6.5 1.8 35 1.1 pure Stern et al.342

Pebax 4011 + PSf 139−190 7.7−9.5 20 4.5 pure Blume et al.343

Pebax 4011 + PVDF 51 23 28 99.9:0:0.097 Lokhandwala et al.344

41 68
52 3.7 68 95.79:4.12:0.087
48 3.4 68 94.9:4.12:0.986
50 3.6 68 87.34:10.8:1.83
48 3.4 68 91.77:8.14:0.095

poly(ether urethane) [PPG] 1.1 × 10−9 21 3.0 35 10.1 70.8:27.9:1.3 Chatterjee et al.333
poly(ether urethane) [PEG] 1.2 × 10−9 58 4.5 70.8:27.9:1.3
poly(ether urethane urea) [PPG] 2.7 × 10−9 19 3.1 70.8:27.9:1.3
poly(ether urethane urea) [PEG] 8.8 × 10−10 74 4.4 70.8:27.9:1.3
Pebax MX 1074 2 × 10−9 54 4.5 70.8:27.9:1.3
CA 9.4 × 10−12 19 0.9 65:29:6
poly(ether urethane urea) [PEG] 4.5 × 10−10 102 4.6 20 70.8:27.9:1.3

PVBTAF 9.3 × 10−9 7000 30 1.2 89.5:0:10.5 Quinn et al.345

4.3 × 10−9 950 5.0
5.4 × 10−9 3300 11 1.1 79.7:10:10.3
2.8 × 10−9 2100 8.7 8.5

PVBTAF-4CsF 3.06 × 10−8 n.d. 10 1.1 80.1:10:9.9 Quinn et al.346

1.41 × 10−8 n.d. 4.7 4.4

PC/siloxane 4.85 × 10−11 10 2.8 20 ± 2 pure Amirkhanov et al.347

poly(4-methyl pentene) 3.35 × 10−13 7.6 1.2

PDMS 1.7 35 10.1 pure Wilks et al.348

Pebax 2533 2.1

substituted-PPZ (Tg = 55 °C) 6 × 10−11 4 0.74 30 2.1 pure Orme et al.349

substituted-PPZ (Tg = 4 °C) 1 × 10−10 27 4.1 3:3:3 Orme et al.350

substituted-PPZ (Tg = −63 °C) 3 × 10−9 60 5.0
substituted-PPZ (Tg = 4 °C) 4 × 10−11 28 1.9
substituted-PPZ (Tg = −63 °C) 3 × 10−9 79 4.4

PTMSP 7.1 × 10−8 1.2 23 0:10.5:1.5 Merkel et al.351

PDMS 1.1 × 10−8 1.5 21 0:11.7:0.7
Pebax 1657 3.7

9783 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 12. continued

polymera PH2S (mol/(m2·s·Pa)) SH2S/CH4 SH2S/CO2 T (°C) p (bar) CH4:CO2:H2S reference
PC 6.3 × 10−12 0.23
PSf 1.1 × 10−11 0.26
Cytop 3.8 × 10−12 0.037 0:0:15
TFE/PMVE/8CNVE 0.13 37 0:10.5:1.5
Teflon AF 1600 0.15 23

polyamide + PSf (before annealing) 1.7 × 10−8 40.5 3.16 30 10 pure Sridhar et al.352
polyamide + PSf (after annealing) 1.7 × 10−8 49.1 3.41

poly(ester urethane urea) + Teflon 4.4 × 10−8 43 3.79 35 10 91.6:5.4:3 Mohammadi et al.353
3.0 × 10−8 27 2.72 30 91.6:5.4:3
8.2 × 10−8 43 3.91 55 10 91.6:5.4:3
2.9 × 10−8 12 1.90 30 91.6:5.4:3

butadiene−sulfone 5.7 × 10−9 3.8 35 5 96.1:0:3.9 Saeidi et al.354

8.4 × 10−9 8.1 15

Cardo-type PIM 7.3 × 10−9 3.8 23 4.5 99.96:0:0.04 Chenar et al.355

9.4 × 10−9 6.1 7.9
PPO 2.6 × 10−8 3.1 23 4.5
2.5 × 10−9 4.1 7.9

6F-PAI-1 12 ± 1 0.35 35 43 70:20:10 Vaughn et al.356

6F-PAI-2 6.9 ± 0.5 0.49 60:20:20
6F-PAI-1 8.1 0.19 4.5 pure
6F-PAI-2 10.3 0.21

CA 3.8 × 10−11 30 1.0 35 35 60:20:20 Achoundong et al.334

GCV-modified CA 6.8 × 10−10 34 1.6

(6FDA-DAM)m(6FDA-DABA)n 1.3 × 10−10 12 0.35 70.15:19.9:9.95 Kraftschik et al.357

PEGMC [DEG] 2.0 × 10−10 19 0.59 35 35 60:20:20 Kraftschik et al.358

PEGMC [TEG] 2.0 × 10−10 21 0.64
PEGMC [TetraEG] 1.6 × 10−10 20 0.59
(6FDA-DAM)m(6FDA-DABA)n 3.3 × 10−10 25 0.83
9.0 0.30 4.5 pure

6F-PAI-2 9 0.32 35 15 60:20:20 Vaughn et al.359

6F-PAI-1 12 0.27
22 0.40 35 15 50:45:5
Pebax SA01 MV 3000 (commercial PAI) 42 4.2 35 15 60:20:20
42 3.8 35 15 50:45:5
43 35 2.4 pure

PIM-6FDA-OH 7 35 4.5 pure Yi et al.360

−10
1.7 × 10 13 0.28 7 70:15:15
4.2 × 10−10 30 1.2 48.3

(6FDA-mPDA)-(6FDA-durene) 15 0.41 22 34.5 60:10:1 Yahaya et al.361

23 0.84 60:10:20

PDMS + PESf 4.3 × 10−9 4.0 35 30 98.32:0:1.68 Saedi et al.362

3.1 × 10−9 3.2 0.66 97.5:2.1:0.4
a
Abbreviations: DABA, 3,5-diaminobenzoic acid; DAM, 2,4-diaminomesitylene; DEG, diethylene glycol; GCV, grafting and cross-linking using
vinyltrimethoxysilane; PAI, polyamide−imide; PC, polycarbonate; PEGMC, poly(ethylene glycol) monoesterified cross-linkable; PESf, poly(ether
sulfone); PET, poly(terephthalic ester); PIM, polymer of intrinsic microporosity; PMVE, perfluoromethyl vinyl ether; PPO, poly(2,6-dimethyl-1,4-
phenylene oxide); PTMSP, poly(1-trimethylsilyl-1-propyne); PVB, poly(vinyl butyral); PVDC, poly(vinylidine chloride); PVDF, poly(vinylidine
fluoride); Pebax, poly(ether amide) block copolymers; TEG, triethylene glycol; TFE, tetrafluoroethylene; TetraEG, tetraethylene glycol; mPDA, 1,3-
phenylenediamine; n.d., CH4 not detectable; 4CsF, four cesium fluoride per repeat unit; 6FDA, 2,2′-bis(3,4-dicarboxyphenyl)hexafluoropropane

9784 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Table 12. continued

dianhydride (same as 4,4′-hexafluoroisopropylidene diphthalic anhydride); 6F-PAI, 6FDA-3-aminobenzoic acid-4,4′-hexafluoroisopropylidene
dianiline; 8CNVE, perfluoro-8-cyano-5-methyl-3,6-dioxa-1-octene.

the regime for superior membrane performance over gas sophisticated approach to include the effects of multi-
absorption. Additionally, hybrid processes involving mem- component sorption and diffusion.335,336 It is well known in
branes for bulk acid gas removal, followed by chemical the literature that the assumption of ideality grossly over-
absorption or adsorption in the polishing step, can drive the predicts selectivity for polymeric membranes if one of the
overall process toward better economics. species strongly absorbs/adsorbs onto the polymer. Absorp-
4.1. Polymeric Membranes tion-induced swelling of the polymer increases chain mobility
Approximating solubility with the Henry’s law and diffusivity and gas flux, a phenomenon commonly known as membrane
with the Fick’s law, and approximating the concentrations on plasticization. Increased chain mobility lowers Tg, potentially
the permeate side to be negligible compared to that on the feed transforming a membrane from glassy to rubbery and adversely
side, the permselectivity of a polymer membrane, Sij, in affecting the size-selective term, Di/Dj. CO2-induced swelling of
differentiating two species, i and j, can be expressed as CA membranes results in a much lower CO2/CH4 mixture
selectivity compared to the ideal selectivty (35 vs 50 at 4 atm);
Pi Fi /pi DK the effect is more pronounced at higher CO2 partial pressures
Sij = = = i i (32 vs 80 at 12 atm).337 Other impurities, such as light
Pj Fj/pj DjKj (7) hydrocarbons, water, very low concentration of aromatics, etc.,
where Pi, Fi, pi, Di, and Ki are the permeability, flux, partial can also have a plasticization effect on membrane permeances.
pressure, Fick’s diffusion constant, and Henry’s constant for In addition to plasticization, glassy materials suffer from
species i. The ratio Di/Dj can be viewed as the transport physical aging where the polymer chains slowly relax into a
selectivity indicating the relative speed of the two components preferred high-density form by a reduction in free volume.338
and is mainly governed by the molecular shapes and sizes. The This type of aging effect can reduce membrane permeabilities
ratio Ki/Kj can be viewed as the sorption selectivity and is significantly over the average 2−5-year life of a membrane used
governed by the relative gas solubilities in the membrane in natural gas service.331
material. Kulprathipanja et al. synthesized two-component membranes
The glass transition temperature, Tg, characterizes the range comprising of a mixture of polydimethylsiloxane (PDMS) and
of temperatures over which a hard and relatively brittle “glassy” poly(ethylene glycol) (PEG) plasticizers, having molecular
amorphous state of the polymer transitions into a soft and weights between 200 and 600, cast upon a porous polysulfone
viscous “rubbery” state. Polymer chains are trapped in a frozen (PSf) support that was pre-treated with glycerol.340,341 Glycerol
state at T < Tg, but T > Tg allows for significant motion of the alters the pores of the support, thus increasing the permeability
chains. Glassy polymers offer size-selective separation by virtue of hydrogen-bonding H2S (see Table 12). These membranes
of differential diffusivity; in contrast, rubbery polymers suffer are of the asymmetric type; a very thin, dense surface layer or
reduced size selectivity from increased elasticity of the polymer skin is supported upon an integrally attached, much thicker
chains. The kinetic diameters of H2S (3.6 Å) and CO2 (3.3 Å) supporting layer. In this case, the selective skin is rubbery, while
are smaller than that of CH4 (3.8 Å); this suggests the the support is glassy. The authors reported very high H2S/CH4
possibility of a diffusion-based selectivity. The critical temper- ideal selectivities between 115 and 185; however, the mixed gas
atures for these gases are 373, 304, and 191 K, respectively, selectivities measured later are smaller by a factor of 4−8.344,363
suggesting that, compared to methane, the acid gases will Both PDMS and PEG belong to the class of rubbery
interact more strongly with the polymer and also allow for membranes with Tg = −127 and −67 °C, respectively;
sorption-based selectivity. Although both rubbery and glassy therefore, it is unlikely that plasticization results in lower
polymers can separate CO2 and H2S from natural gas, in mixture selectivities. It is possible, however, that these strongly
practice, CO2 is best separated by glassy polymer membranes absorbing membranes are being operated at both feed and
(maximizing size selectivity), whereas H2S, which is larger and permeate pressures in the saturated loading regime. Several
more condensable than CO2, is best separated by rubbery studies report ideal and mixture H2S/CH4 selectivities in
polymer membranes (maximizing sorption selectivity).331 PDMS membranes of ≤11,341,342,362 indicating the crucial role
Although plants based on glassy CA membranes (Tg ≈ 110 of PEG plasticizers in enhancing the selectivity toward H2S.
°C) are widely used for CO2 removal from natural gas, these Unlike CA and PDMS membranes, these two-component
membranes also possess SH2S/CH4 ≈ 15−30.332−334 The membranes present high SH2S/CO2 as well, possibly by
contribution of diffusion and sorption to this overall selectivity benefitting from the hydrogen-bonding character of H2S.
is unclear; H2S/CH4 sorption isotherms will help understand However, this raises concerns about the influence of H2O
the relative proportions of the two factors. Using CA to reduce impurities on such membranes, and this can be additionally
H2S to pipeline specifications (4 ppm) results in vast over- important considering that PEG is soluble in H2O. Molecular
processing as far as the CO2 concentrations are concerned. simulations can possibly explain the role of PEG and also the
Thus, just like liquid absorbents with a higher H2S/CO2 influence of impurities on membrane performance.
selectivity, there is a need for membranes with higher SH2S/CO2. Another class of polymer membranes, commercially known
Permselectivity can be obtained either from a real mixture as the Pebax membranes, are composed of ether−amide block
experiment or by modeling mixture transport using pure-gas copolymers, where the polyether block imparts the flexible
permeabilities and sorption isotherms. While it is most character to the rigid polyamide block. Various grades of Pebax
common to report the ideal selectivity (ratio of pure-gas membranes have been tested for acid gas removal, and these
permeabilities), it may be possible to adopt a more membranes yield ideal and mixture SH2S/CH4 values in the ranges
9785 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

140−190 and 35−70, respectively.333,343,344,359 Similar to operated in the saturated loading regime; permselectivity data
PDMS and PEG, the Pebax membranes also belong to the imply that both CO2 and H2S compete for the same reaction
class of rubbery membranes with Tg < −50 °C; as mentioned sites, hydrated fluoride ions.345 The authors also report that the
earlier, the drop in mixture selectivity compared to ideal membranes show no deterioration after 30 days of continuous
selectivity may be due to operation in the saturated loading operation but react irreversibly with trace levels of sulfur-
regime. These membranes demonstrate SH2S/CO2 between 3.4 containing contaminants, such as CS2 and COS, that are
and 4.5. The presence of H2O impurities was examined by common to cylinder H2S. The permselectivity of these
using a feed gas stream saturated with H2O vapor; no membranes was increased by 3−4 times without reduction in
detrimental effect on the selectivities was observed, but the selectivity by using fluoride-containing salts such as cesium
fluxes decreased by 40−45%.364 Chatterjee et al. synthesized fluoride.346 The performance of these membranes is very
two poly(ether urethanes) and two poly(ether urethane ureas) sensitive to the relative humidity of the feed stream.
using PEG or poly(propylene glycol) (PPG) for the ether Unlike conventional microporous materials such as zeolites
block; the polyurethanes are structurally related to the Pebax and activated carbon, polymeric materials possess advantageous
polymers.333 Membranes with PEG blocks yield much higher properties, such as solution-processability and precise control
H2S/CH4 and H2S/CO2 selectivities compared to the ones over chemical functionality. However, polymeric membranes
with PPG blocks, but with no significant difference in the suffer from large inherent densities and, as a result, low
permeabilities; this maybe attributed to the higher mass fraction permeabilities in general. Budd and co-workers developed a
of polar hydroxyl groups in PEG compared to PPG. Very high new class of polymeric membranes that not only are solution-
selectivities reaching up to 100 for H2S/CH4 and 4.8 for H2S/ processable but also possess intrinsic microporosity.366 This
CO2 were observed by reducing the temperature from 35 to 20 new class of materials, known as polymers of intrinsic
°C. microporosity (PIMs), are a result of contorted shape and
Yet another class of rubbery membranes that offer high extreme rigidity of planar units fused together by rigid
selectivity for H2S/CH4 and H2S/CO2 separation (60−80 and spirocyclic linkers. These kinks at the spiro-centers prevent
4−9, respectively) are substituted polyphosphazenes (PPZ), efficient packing to form a dense polymer matrix, which in turn
results in a material with intrinsic microporosity.367 These
where the substituents are predominantly 2-(2-methoxy-
materials are microporous (pores smaller than 2 nm) and
ethoxy)ethanol (MEE) and 4-methoxyphenol.349,350 Orme et
possess a considerable fraction of pores that are below 7 Å.368
al. explained that the flexible backbone of PPZ has a low
Hydroxyl-functionalized PIMs (PIM-6FDA-OH) demonstrated
rotational energy barrier due to pπ−dπ bonding between
higher permeabilities than most traditional polyimides for sour
nitrogen and phosphorus. Polyether pendant groups, such as
gas sweetening application.360,369 Annealed PIM-6FDA-OH
MEE, are linear chains of sp3-hybridized atoms which have
films were shown to resist plasticization up to much higher
many degrees of rotational freedom; thus, they do not
pressures compared to unannealed films and other 6FDA-based
significantly hinder the inherent flexibility of the backbone.350
polyimide materials. These films were dried under vacuum at
On the other hand, aromatic substituents hinder the backbone
100 °C to remove any moisture prior to testing. However, since
flexibility, in turn increasing Tg. This work showed a surprising
processed natural gas, although with low dew point temper-
result that lowering Tg resulted in increased permeabilities and
atures, may contain finite moisture levels, future research for
separation factors at the same time. Similar ideal and mixture
these very promising, but hydrophilic, membranes should
selectivities suggest that these rubbery membranes are operated
consider the effects of moisture on gas sweetening.
in the regime where the polymer is not saturated with the more Several other research groups have been investigating
strongly absorbing gas molecules: the operating pressure is only polymeric membranes for H2S removal, from the 1950s right
2.1 bar as opposed to several tens of atmospheres for previously up until today,334,339,347,348,351−359,361,370−373 and the data from
mentioned mixture experiments. MEE is hydrophilic and makes some of these papers is also presented in Table 12. To
the PPZ hydrophilic; at 100% MEE substitution, the resulting maximize the membrane area per unit volume, two main
PPZ is soluble in water. In contrast, aromatic substituents configurations are widely used: hollow fibers and flat sheets in
impart hydrophobicity. The effect of moisture in the feed the form of spiral-wound modules. Both these module types are
stream on the performance of PPZ membranes remains to be commercially applied for CO2 removal from natural gas.331
investigated. These 100−200 μm thick membranes were
supported on ceramic supports with 0.2 μm pores. Permeation 4.2. Membranes for Gas−Liquid Contact
properties of thin membranes are very different from those of In general, synthesizing highly selective gas membranes that
thick films, and since the selective layer of membranes in also offer high flux is an active area of industrial and academic
industrial use is usually only 0.1−1.0 μm thick,365 performance research. An alternative technology that combines the
for thinner substituted-PPZ membranes remains to be seen. advantages of high surface area per unit volume of membrane
Quinn and co-workers have explored facilitated transport contactors with the selective absorption in amine-based liquid
membranes for selective removal of H2S from natural gas solutions was first developed for CO2 capture in 1985 by Qi
mixtures.345,346 These membranes are based on the poly- and Cussler.375,376 Such gas−liquid interfaces may be achieved
electrolyte, poly(vinylbenzyltrimethyl ammonium fluoride) by flowing a gas on the outside of a membrane configuration
(PVBTAF), and show exceptionally high mixture selectivity such as a hollow fiber and with liquid flowing within its lumen
for selective removal of H2S and CO2 from CH4 (>800 and (see Figure 18). These gas−liquid membrane contactors offer
>400, respectively). Using NMR data, the authors show that several advantages over conventional contacting devices, such
CO2 reacts with hydrated fluoride ions to form mobile as high surface area per unit contactor volume; independent
bicarbonate ions, HCO3−. The permeance of both CO2 and control of gas and liquid flow rates without any flooding,
H2S increased with decreasing feed partial pressure of the loading, weeping, foaming, or entrainment problems; small size;
respective gases, a phenomenon observed for membranes known gas−liquid interfacial area; and being modular and easy
9786 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

absorbent, mass transfer, and membrane modules, are critically

discussed elsewhere by Mansourizadeh and Ismail.377
Asymmetric hollow-fiber membrane modules made of PSf
and PESf, coupled with a concentrated alkaline NaOH solution,
were investigated both experimentally and theoretically for the
removal of H2S (see Table 13).378 In general, the gas-phase
diffusivity is about 4 orders of magnitude higher than the liquid-
phase diffusivity. As a result, in the case of non-reactive
absorption, the additional mass-transfer resistance contributed
by a non-wetted membrane is significantly smaller than the
resistance in the liquid phase. Thus, addition of a non-wetting
membrane to a non-reactive gas−liquid system results in a
negligible decrease in the overall mass-transfer coefficient. A
non-wetting solvent−membrane combination can therefore
greatly enhance the overall mass transfer (Koverall × A̲ ) of the
Figure 18. Cross section of a hollow fiber (gas on the shell side and soluble species from the gas phase to the liquid phase by virtue
liquid on the tube side). P and C are the gas partial pressure and liquid of higher surface area per unit volume ( A̲ ). Since reactive
concentration, respectively. The subscripts gb, gm, lm, and lb represent absorption in many cases shifts the rate-limiting step to the gas
gas bulk, gas−membrane interface, liquid−membrane interface, and phase, having a wetted membrane can be beneficial over a non-
liquid bulk, respectively. Reprinted with permission from ref 374. wetted membrane in such cases. These are general guidelines
Copyright 2016 Elsevier. and may vary from case to case, depending on the extent of
equilibrium between a gas−liquid pair. In a subsequent study,
Li et al. developed a PVDF membrane, and due to its excellent
to scale up or down.377 Important aspects of membrane chemical resistance, much higher H2S feed concentration could
contactors for acid gas removal, including liquid absorbents, be investigated.379 Unlike other membrane-based processes
membrane characteristics, combination of membrane and such as ultra-filtration or molecular sieving, very small pressure

Table 13. H2S Capture by Gas−Liquid Contact Using Hollow-Fiber Membrane Modules

membrane liquid T (°C) p (bar) yH2S (ppm) reference

PSf 2.5 M NaOH 25 1−3.5 16−24 Li et al.378
PESf

PVDF 2.5 M NaOH 25 1−3.5 Li et al.379

2 M Na2CO3 0.9−1.1 17.9−1159 Wang et al.380
1.2−1.7 1000 Wang et al.381

ePTFE H2O 24 1−50 20000 Marzouk et al.382

0.1−2.0 M NaOH
0.1−2.0 M MEA
0.5 M DEA
0.5 M TETA

PVDF 0.84−2.0 M MDEA 25 1.3−1.6 1500−3100 Hedayat et al.383

PSf 0.84−1.84 M MDEA 25−45 1 1100−2200
0.84−1.0 M MDEA
0.84 M MDEA, 0.16 M DEA
0.84 M MDEA, 0.16 M MEA

ePTFE H2O 22 1−50 20000 Marzouk et al.384

0.1−2.0 M NaOH
0.5 M MEA
0.5 M DEA
0.5 M DETA
PFA H2O
0.1−2.0 M NaOH
0.5 M MEA
0.5 M DEA
0.5 M DETA

PVDF H2O 25 N/A 250−1000 Rongwong et al.385

0.5 M MEA

9787 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

gradients need to be maintained across the gas−liquid interface, wetting of the membrane. The effect of temperature on the
just enough to prevent bubbling of the gas in the liquid phase, H2S removal efficiency was small, but it substantially decreased
since the driving force for mass transfer is the chemical the CO2 removal efficiency and consequently enhanced the
potential (concentration) gradient. The effect of an alternate H2S selectivity.383 Absorbent mixtures as well as acid gas
alkaline medium such as sodium carbonate (Na2CO3) was also mixtures affect the H2S and CO2 removal efficiencies in
investigated, and complete removal of H2S could be achieved complex ways. While the logic presented in the paper by
with short residence times (<0.1 s).380 H2S removal in the Hedayat et al. makes sense, there are multiple phenomena
presence of CO2 showed SH2S/CO2 = 180−200, about 1 order of going on in this system at the same time, and it is hard to
magnitude higher than that reported for packed towers (10− deconvolute the problem; insights from molecular simulations,
30).381 Note that here the selectivity is defined as the ratio of both equilibrium and non-equilibrium, can play a role in
overall mass-transfer coefficients for H2S and CO2 and may improving the understanding of the underlying phenomena.
vary with the operating conditions of the packed tower and the Recently, Esquiroz-Molina et al. studied the effect of pH for
membrane module. With the hollow fibers, the selectivity alkaline NaOH solution on the mass transfer of H2S in
increased by 33% for a 6-fold increase in gas/liquid ratio and by commercially available hollow-fiber membrane contactors made
of hydrophobic polypropylene.386 Theoretical gas side
50% for a 3-fold increase in gas pressure, and was insensitive to
residence times were varied between 0.37 and 17 s, while the
the inlet concentration of CO2 over a wide range (5−25%).381
liquid side residence time was about 12 s. The authors
Marzouk et al. explored expanded polytetrafluoroethylene
determined a threshold pH of 10 (10−4 M), above which the
(ePTFE) hollow fibers for removal of high-concentration H2S
reaction could be described as pseudo-first order with respect
from a binary gas mixture (2% H2S in CH4) at high pressures
to H2S due to an excess of reactive hydroxide ions. At pH > 10,
(up to 50 bar).382,384 PTFE is a highly non-reactive polymer,
the reactive absorption in the liquid phase increases the liquid-
mainly due to the strong strength of carbon−fluorine bonds;
side mass-transfer coefficient enough to make the reaction gas-
this non-reactivity allows the polymer to withstand highly
phase controlled. The authors propose a NaOH pH of 11
corrosive H2S environments. Additionally, fluorocarbons are
(0.001 M) to be economically attractive; this value is several
known to be very hydrophobic, thus repelling water or water- orders of magnitude smaller than the concentrations previously
containing solutions and in turn preventing wetting of the investigated in the literature. It should be noted, however, that
hollow fibers. It was observed that the ratio of flux in the case of dilute solvents can reduce the cost of solvents but will greatly
NaOH solution to that in the case of distilled water (a physical increase the regeneration cost of the spent absorbent. Such a
absorbent) decreases from ∼4.5 at 1 bar to ∼1.35 at 50 bar. reduction in concentration may be justifiable only for a very
This seemingly small effect of alkaline solution at higher high value solvent.
pressure amounts to ∼75% H2S removal for water and ∼100% Several groups have modeled H2S removal using gas−liquid
H2S removal for 2 M NaOH solution. Performance for MEA contact over hollow-fiber membrane modules.387−394 Some of
was observed to be vey similar to that for NaOH. H2S flux rates the models considered non-wetting and partial wetting
were obtained with three different amines, i.e., MEA, DEA, and conditions, where the gas mixture and the solvent, respectively,
triethylenetetraamine (TETA), and it was shown that the flux fill the membrane pores.388,394 Axial and radial diffusion were
followed the order TETA > MEA > DEA. TETA has four considered inside the fiber, through the membrane, and within
amino (Lewis base) sites, explaining its having the largest H2S the shell using a two-dimensional (2D) mathematical model.388
flux. The steric hindrance in MEA is lower compared to that in This 2D model was modified to account for the reversible
DEA, allowing better interaction of the Lewis base site with reactions of all species involved in the chemical absorption of
H2S. The effect of simultaneous absorption of CO2 and H2S H2S in aqueous carbonate solution389 and then further
from a pressurized CO2−H2S−CH4 gas mixture was inves- extended to systems involving the presence of CO2.390 Amrei
tigated using ePTFE and a commercially available diblock et al. presented a comprehensive model considering the
polymer, poly(tetrafluoroethylene-co-perfluorinated alkyl vinyl entrance regions of momentum, energy, and mass transfers;
ether) (PFA).384 The results proved that the PFA fibers are compressibility of the gas phase as a function of gas-phase
superior to the more common ePTFE mainly because of their temperature, pressure, and composition; and variation of
smaller dimensions, higher porosity, and higher hydrophobicity. physical solubility of CO2 and H2S as a function of gas−liquid
Utilizing PFA fibers in the future would lead to substantial interface temperature, pressure, and compositions.393 Dynamic
reduction in the membrane area (10 times) and/or significant response of the membrane reactor to various types of pulses on
enhancement in the acid gas removal efficiency. Since the the feed and solvent showed that the system was dynamically
contact angles for ePTFE and PFA are 111.3 and 118.6°, stable and reverted back to its previous operating point. It is
respectively, it is quite possible that a polymer with even higher worth noting that the modeling of flow in hollow-fiber
hydrophobicity may further increase the membrane flux. contactors is not trivial. Inside the fiber, the assumption of a
While liquid-phase resistance dominates mass transfer in the fully developed laminar parabolic velocity profile in a tube
case of physisorption,384 the gas-phase resistance dominates for geometry works very well. However, due to poor knowledge of
reactive absorption in the liquid phase.383 An increase in amine the shell-side geometry, the models for what happens on the
concentration in the absence of CO2 resulted in a decrease in outside of the hollow fibers may correlate one set of data but
the mass-transfer coefficient, and this was attributed to the are rarely effective in general.395,396 Another issue about
enhanced wetting of the membrane resulting from reduced hollow-fiber membranes that is not addressed in many studies
surface tension due to organic amines.383 In the presence of is the unstable gas−liquid interface. To overcome this, Prasad
CO2, however, MDEA concentration enhanced the removal and Sirkar put a higher static pressure on the phase which did
efficiency of H2S as well as CO2; since MDEA is a tertiary not wet the membrane, pushing the interface back to a more
amine, the CO2 molecules are not absorbed and stay in the stable location.397 However, this may not work as well for
gas−liquid interface within the membrane, hindering the polydisperse pores, and it is an issue that, if not addressed at the
9788 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

design stage, may lead to considerable performance problems in H2S/CH4 mixture using a porous graphene membrane.411 The
actual practice. pore was formed by deleting 12 C atoms from the center of the
In conclusion, hollow-fiber membrane contactors have sheet in the simulation box, and partial charges of −0.241 and
shown some promise for selective H2S removal; especially the +0.241 e were assigned to the carbon atoms in the first and
uncommon H2S/CO2 selectivities are promising. A better second 12-carbon shells. For the charged-pore membrane, the
understanding of the flow patterns may be obtained from flux nearly doubled for H2S and nearly halved for CH4,
reactive computational fluid dynamics. This will enable further resulting in a 4-fold increase in selectivity compared to the
insights into whether operating and design parameters such as uncharged-pore membrane. In another simulation study, Sun et
wettability of the membrane have positive or detrimental effects al. investigated different pore sizes and shapes for graphene
in optimizing not only H2S capacity but also H2S/CH4 and membranes and found H2S/CH4 selectivities in excess of 100 at
H2S/CO2 selectivities. reasonably high permeances.412
4.3. Ceramic Membranes Another class of carbon-based membranes, carbon molecular
sieve (CMS) membranes, which are prepared by the carbon-
Ceramic membranes have the potential to precisely control ization of polymeric precursors, have shown promising CO2/
transport properties of molecules with different sizes, shapes, CH4 selectivity (35−190) at high CO2 flux of 10−6−10−8 mol/
and interaction strength, in turn allowing sieving at molecular (m2·s·Pa).413,414 The utility of such membranes for selective
length scales. Maghsoudi and Soltanieh investigated the H2S removal from natural gas streams remains to be explored.
simultaneous separation of H2S and CO2 from CH4 using a Very recently, Wenz and Koros introduced a postsynthetic
ceramic membrane of CHA-type zeolite.398 Since polar modification method using aromatic amines to bridge the sp2
impurities such as H2O, due to their large permanent dipole domains in CMS membranes with the goal to prevent
moments, may affect membrane performance, there is a relaxation of the CMS to suppress physical aging (see Figure
considerable value attached to exploring high-silica zeolite 19).415 However, the modifications were found insufficient to
membranes. An approximately 20 μm thick silica membrane
was synthesized on a porous α-alumina disk, with repeating the
crystallization process thrice to seal defects as judged from N2
permeability tests. The elemental profiles for Al and Si along
the cross section by EDX line scanning view showed that the
zeolite layer has Si/Al ≈ 5. Before the actual permeance
measurement, the feed and permeate sides were purged for 1 h
with the desired feed and Ar, respectively, and Ar was also used
as a carrier gas during the permeation measurements. Ternary
mixture (2.13% CO2, 0.30% H2S in CH4) permeance was
measured at room temperature, and the ratios of permeances
for H2S/CH4 and CO2/CH4 were observed to be 3.24 and 5.3,
respectively. When normalized by the ratio of feed concen-
trations, these values are 1050 and 240, respectively, suggesting
that the potential for acid gas separation using zeolite
membranes should be further explored. Figure 19. Schematic of amine doping concept, using the carbonyl
4.4. Carbon-Based Membranes functional groups as example reaction site. Reprinted with permission
from ref 415. Copyright 2017 American Institute of Chemical
Pristine graphene is a two-dimensional, single-atomic-layer, Engineers.
hexagonal lattice of sp2-hybridized carbon atoms that is
completely impervious to liquids and gases, including atoms
as small as helium (2.6 Å).399 Perforated graphene membranes yield aging-resistant membranes. This is an emerging field, and
with sub-nanometer micropores, also known as “porous hopefully new synthetic techniques to prevent collapsing of
graphene membranes”, have shown promise for numerous CMS will emerge in the future.
separation applications.400−408 Nanoporous defect formation in Gilani and co-workers used vertically aligned CNT
single-atomic layers of graphene using high-resolution TEM (VACNT) membranes,416 functionalized with different amine
was reported by Hashimoto et al.409 In another approach, groups, for separation of binary H2S/CH4 mixtures (see Figure
“bottom-up” synthesis has also been demonstrated using 20).417 Permselectivity values between 2.5 and 3 and H2S
phenylene rings as building units to form porous graphene.410 permeabilities between 3 × 10−7 and 8 × 10−7 mol/(m2·s·Pa)
Although porous graphene membranes have been explored for were observed, depending on the choice of functional groups.
the selective molecular sieving of He, H2, CO2, Ar, O2, N2, CO, Linear amines with a short carbon chain length lead to a higher
and CH4, there is no experimental measurement for H2S, which selectivity and H2S permeability; the order of selectivity and
should be explored in the future. H2S permeability of membranes was obtained as butylamine >
It is important to note here that a successful gas separation dodecylamine > sec-butylamine > octadecylamine.
using a single-layer graphene membrane has not been reported. Controlling aggregation of individual carbon membrane
Most measurements are for a few-layers-thick graphene oxide units, ensuring uniformity of thickness, and tuning precise
membranes. This is likely to have strong implications on distribution of defects are important challenges in the area of
separations involving polar species such as H2S due to the carbon-based membranes that are still not fully mature. Goh et
different binding affinities of polar species toward pristine al. have recently reviewed carbon-based membranes with a
graphene surface compared to the functionalized surface of focus on water purification and identified five strategic areas for
graphene oxide. Lei et al. carried out MD simulations to probe future research, namely, concentration polarization, fouling,
the effect of the charges around the pore on the separation of a stability, scalability, and cost.252 Stacked graphene and surface-
9789 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Figure 20. FESEM micrograph of (a) surface of anodic aluminum oxide (AAO) template, (b) cross section of AAO, and (c) VACNTs after
removing AAO. Reprinted with permission from ref 417. Copyright 2016 Elsevier.

modified graphene are much more stable and scalable Note that, for modeling the separation performance of MMMs,
compared to single-layer graphene. Similarly, randomly aligned the permeability values of H2S/CH4 mixtures through
CNTs are more scalable compared to VACNT. Like in most polymeric membranes were taken from the experimental data
research fields, even in carbon-based membranes, the pursuit of available in the literature, while those through the MOF phase
advanced performance faces competing characteristics such as were taken from the simulation results. It was found that, for
scalability, stability, and cost-effectiveness. polymers with much lower H2S permeabilities compared to the
4.5. Composite Membranes MOF, MOF addition into its matrix increases the permeability
by about an order of magnitude for 50% volume fraction of the
Supported ionic liquid membranes (SILMs) consist of an IL MOF and the selectivity remains almost unchanged at the
embedded in a solid polymer matrix (support). SILMs with 1- polymer permselectivity. However, for polymers with H2S
butyl-3-methylimidazolium tetrafluoroborate as the IL for its permeabilities not lower than 2 orders of magnitude compared
superior affinity to acidic gases, and PVDF as the polymer to the MOF permeability, both selectivity and permeability of
matrix for its hydrophobicity as well as stable physical and the polymer increase upon MOF addition. In a subsequent
chemical properties, were tried for selective H2S removal.418,419 study, Li et al. carried out molecular simulations using generic
The permeability coefficients increased dramatically, especially MOF force fields (shown to have worked reasonably accurately
for H2S, with an increase in the IL content of the membrane. in many previous studies) to study the adsorption and
This increase may be due to enhancement in diffusion, since membrane performance of UiO-66(Zr) and its functionalized
solubility does not vary much with IL loading. derivatives for H2S removal from binary mixtures with CH4 and
Wang et al. computationally studied the separation of an CO2.422 UiO-66-(COOH)2 and UiO-66-COOH performed
H2S/CH 4 mixture in the UiO-66(Zr)-(COOH)2 MOF better than other functionalized derivatives; however, the
membrane and the mixed matrix membranes (MMMs) performance of these adsorbents/membranes containing
composed of polymer as the continuous phase and MOF as hydrophilic functional groups, in the presence of moisture,
the dispersed phase.420 GCMC simulations using generic force remains to be seen.
fields that are known to show reasonable adsorption estimates
for MOFs were used to obtain equilibrium loading at the
conditions of interest. MD simulations were then employed to 5. CRYOGENIC DISTILLATION
calculate diffusivities considering framework flexibility, and Although the phase behavior of the CH4/H2S system was
equivalent results were obtained from simulations in the studied extensively in the 1950s,423−425 there has been renewed
canonical and the microcanonical ensembles. With a combina- interest in the study of these systems at additional state points,
tion of atomistic simulations and the Maxwell−Stefan especially at higher pressures, in the past couple of years.426−428
formulation for permeation,421 H2S/CH4 permselectivities of Coquelet et al. investigated phase equilibria from −87 to 40 °C
∼200 for UiO-66(Zr)-(COOH)2, 2−50 for the different up to 140 bar,426 and Perez et al. measured the mixture
polymers, and between 2 and 60 for MMMs were estimated. densities between −20 and 20 °C and up to pressures of 300
9790 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

bar.428 Lange et al. also explored the complete pT diagram of was overcome for liquid absorbents. Similar breakthroughs for
the methane−hydrogen sulfide system down 70 K and for supported amines will be essential for efficient H2S capture.
pressures up to 2500 bar by tuning the parameters within a One such promising attempt employs variations in the
solid−liquid−vapor equation of state and validating against the operating conditions for a two-column setup: the first column
available experimental data.427 at high temperature for absorption of CO2, and the second
Total, Institut Francais du Petrole, and its fully owned column at low temperature for removal of H2S without over-
subsidiary, Prosernat, have developed a cryogenic process for processing for CO2.300 Adsorbents where aminopolymers are
bulk H2S removal from very sour gases, the SPREX (Special polymerized in situ on a porous support have not yet been tried
PRe-Extraction) process.429 The pre-treatment step reduces the for H2S capture, and this could constitute an important future
H2S content in the gas by producing a partially sweetened gas investigation.
which can then be processed by a smaller sized amine Ionic liquids are an emerging class of absorbents for selective
sweetening unit; it also produces high-pressure liquid H2S acid gas removal; emphasis on understanding their structure
(50−65 bar) that can be easily re-injected into a geological and performance should remain the focus of fundamental
reservoir. The minimum overhead temperature is limited to investigations, and cost comparisons with commercial liquid
about −30 °C, so as to remain above the hydrate formation solvents will help understand the true potential of these new
temperature, ruling out the necessity of installing a dehydration liquids for sweetening applications. Deep eutectic solvents are
unit upstream of the Sprex column. This temperature of −30 another new class of absorbents; although these are explored
°C limits the H2S concentration in the distillate to about 10%. only for CO2 capture as of today, future studies may investigate
However, the authors also describe an enhanced version, where H2S removal using these solvents. In addition to the
the temperature can be reduced to −65 °C without crystallizing performance metrics of capacity and selectivity, operational
CO2 or H2S, thereby reducing the H2S content in the overhead aspects, such as lowering the viscosity and aquatic toxicity of
stream to 2−3%. The triple point of H2S is about −85 °C, these new classes of solvents, are also likely to play an
suggesting that the process is very close to where it can important role in the development of these new technologies.
practically be, even if pre-drying is carried out to completely Mixed metal oxides for reactive H2S removal have evolved
remove the moisture. empirically in most cases, and very little is known in the
For many chemical separations, distillation has been the most literature about the fundamental mechanisms governing
straightforward and often long-lasting choice, unless an obvious improved performance. Some progress has been shown in the
better choice of separation was available. It is intriguing that, in ZnO and Cu−Zn−O mixed oxide sulfidation mechanisms using
the development of separation processes for gas sweetening, TEM and in situ synchrotron XRD. An even better under-
distillation has come into the picture only in the past decade or standing of these reaction mechanisms may assist in designing
so. Distillation, in general, is a very energy-intensive process mixed oxides with not only a high degree of regenerability but
that requires boiling and condensing of the entire feed stream also structural reversibility on regeneration. The latter is more
(a number of times for systems with low relative volatilities). important for a stable bed performance over the life of the
However, for a raw gas stream (4.5 million S·m3/day) adsorbent. Some of the MOFs investigated in the literature
containing 35% H2S and 7% CO2, the capital expenditure, have shown very high H2S capacities. Knowledge of thermal
power consumption, and steam consumption of the amine conductivities and specific heats of these materials becomes
process and the SPREX + amine process were reported to be essential in the design of adsorption beds involving high
153 and 128 MM US$, 52 and 30 MW, and 46 and 34 MW, exothermicity. In case of cation-exchanged zeolites, structural
respectively.429 This shows that distillation can indeed be more characterization of metallic species to elucidate whether these
cost-effective and energy-efficient as a pre-treatment step, are located at catalytic ion-exchange sites or just as extra-
compared to amine absorption, for bulk acid gas removal.429 framework metal oxides should help in understanding the
performance on more fundamental grounds. Most literature
6. OUTLOOK AND PERSPECTIVES studies on utilizing solids for mass separation focus on
Considering the wide range of hydrogen sulfide concentrations designing sorbents with higher selectivities and capacities;
in different natural gas reservoirs, a need for a portfolio of emphasis should also be levied on factors such as structural and
efficient (multi-step) separation processes is evident. In performance regenerability, stability in the presence of strongly
addition to H2S removal, other competing goals such as CO2 adsorbing/reacting impurities, heats of adsorption, specific heat
removal, removal of other organic sulfur compounds, and capacities, and thermal conductivities. Future adsorbent
moisture removal need to be considered where necessary. screening should try to incorporate these diverse properties
Before one can quantitatively compare many of the emerging in the selection criteria.
materials described in this Review, further detailed inves- Significant development has occurred in the scientific and
tigations of these materials under conditions of interest (as technological improvement of polymeric membranes for sour
discussed at appropriate places throughout the Review) are gas sweetening. Synthesis and performance studies of thin
essential. membranes made from some of the promising polymeric
As far as liquid absorbents are concerned, processes based on materials should be undertaken in order to asses the potential
alkanolamines have matured and been perfected over the past of these membranes for commercial application. There is only
several decades. New research on amines for acid gas capture is one report in the literature using zeolitic membranes for H2S
evolving around developing solid supports, such as mesoporous separation;398 the results from this study seem to be promising
silica, to uniformly disperse the amine functionalities, either and should be explored further. Use of porous graphene and
physically or chemically bonded by silane chemistry, and to carbon molecular sieve membranes for H2S capture also
benefit from the advantages of adsorptive separations over constitute an area of future investigation in the field. Mixed-
absorptive separations. Since CO2 is more acidic than H2S, matrix membranes composed of MOFs dispersed in polymeric
developing H2S-selective alkanolamines was a big challenge that materials have been modeled using theory and simulations.
9791 DOI: 10.1021/acs.chemrev.7b00095
Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Experimental and theoretical development of these next- Royal/Shell Laboratory in Amsterdam, and the University of
generation membranes for gas sweetening will remain the Pennsylvania’s Laboratory for the Research on the Structure of
research focus for the next decade or so. Matter. His scientific interests are focused on particle-based
Sulfur is mainly used in the production of sulfuric acid. Other simulations of complex chemical systems, including the predictions
uses of sulfur include manufacture of vulcanized rubber, of phase and sorption equilibria and of thermophysical properties, the
building materials, inorganic chemicals, dyes, skin treatment understanding of retention in chromatography, and the investigation
ointments, matches, explosives, fertilizers, agrochemicals, of microheterogeneous fluids and nucleation phenomena.
fungicides, and fumigants. It is also useful in the bleaching of
paper, extraction of minerals, and processing of wastewater. In
recent years, increased processing of highly sour gas fields in
ACKNOWLEDGMENTS
the Middle East regions has resulted in a large surplus of sulfur, The authors gratefully acknowledge discussions with Edward
and research is being focused on developing new commercial Cussler on membranes for gas−liquid contact. This research is
applications of sulfur, such as battery materials and increased supported by the U.S. Department of Energy, Office of Basic
sulfur content of common plastics and building materials. Thus, Energy Sciences, Division of Chemical Sciences, Geosciences
the overall economics of highly sour gas wells may depend not and Biosciences under Award DE-FG02-12ER16362. M.T.
only on the economics of acid gas removal but also on what acknowledges support through the Center for Gas Separations
one decides to do with the sulfur. Whether one can find high- Relevant to Clean Energy Technologies, an Energy Frontier
value applications for sulfur, or store mountains of sulfur as Research Center funded by the U.S. Department of Energy,
inventory for future increase in demand, or maybe even re- Office of Science, Basic Energy Sciences under Award DE-
inject the H2S gas into depleted wells are the different aspects SC000105. M.S.S. gratefully acknowledges financial support
that may influence the choice of an H2S-removal process. from University of Minnesota through a Graduate School
Detailed economic analysis of competing bulk H2S removal Doctoral Dissertation Fellowship.
processes, such as polymeric membrane separations, cryogenic
distillation, all-silica zeolitic adsorbents, and combinations REFERENCES
thereof, may help researchers arrive at the commercially
optimal solution. (1) United States Department of Labor, Occupational Safety and
Health Administration. Hydrogen Sulfide, Hazards, https://www.osha.
gov/SLTC/hydrogensulfide/hazards.html [accessed May 5, 2017].
AUTHOR INFORMATION
(2) Hammer, G.; Lübcke, T.; Kettner, R.; Pillarella, M. R.; Recknagel,
Corresponding Author H.; Commichau, A.; Neumann, H.; Paczynska-Lahme, B. Natural Gas.
*E-mail: siepmann@umn.edu. Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag
GmbH & Co. KGaA: Weinheim, 2006.
ORCID (3) Awe, O. W.; Zhao, Y.; Nzihou, A.; Minh, D. P.; Lyczko, N. A
J. Ilja Siepmann: 0000-0003-2534-4507 Review of Biogas Utilisation, Purification and Upgrading Technolo-
Notes gies. Waste Biomass Valorization 2017, 8, 267−283.
(4) U.S. Department of Energy, The National Energy Technology
The authors declare no competing financial interest. Laboratory. Gasification Introduction, Syngas Composition, https://
Biographies www.netl.doe.gov/research/coal/energy-systems/gasification/
gasifipedia/syngas-composition [accessed May 5, 2017].
Mansi Shah received her bachelors degree (2012) in Chemical (5) Emission Factor and Inventory Group, Office of Air Quality
Engineering from the Institute of Chemical Technology (formerly Planning and Standards. Petroleum Industry. Compilation of Air
known as UDCT), Mumbai. She is currently a Chemical Engineering Pollutant Emission Factors. Vol. 1: Stationary Point and Area Sources, AP-
Ph.D. candidate at the University of Minnesota and is advised by M. 42, 5th ed.; U.S. Environmental Protection Agency: Research Triangle
Tsapatsis and J. I. Siepmann. Her doctoral thesis focuses on Park, NC, January 1995; Chapter 5.
(6) Gamson, B. W.; Elkins, R. H. Sulfur from Hydrogen Sulfide.
integration of molecular simulations and experiments for the
Chem. Eng. Prog. 1953, 49, 203−215.
development of new materials and processes for gas separation. (7) Sumida, K.; Rogow, D. L.; Mason, J. A.; McDonald, T. M.; Bloch,
Michael Tsapatsis joined the Department of Chemical Engineering E. D.; Herm, Z. R.; Bae, T.-H.; Long, J. R. Carbon Dioxide Capture in
and Materials Science at the University of Minnesota in September Metal-Organic Frameworks. Chem. Rev. 2012, 112, 724−781.
2003 as a professor and currently holds the Amundson Chair. He (8) Pera-Titus, M. Porous Inorganic Membranes for CO2 Capture.
received an Engineering Diploma (1988) from the University of Chem. Rev. 2014, 114, 1413−1492.
Patras, Greece, and M.S. (1991) and Ph.D. (1994) degrees from the (9) Sanz-Pérez, E. S.; Murdock, C. R.; Didas, S. A.; Jones, C. W.
Direct Capture of CO2 from Ambient Air. Chem. Rev. 2016, 116,
California Institute of Technology (Caltech) working with G. R.
11840−11876.
Gavalas. He was a postdoctoral fellow with M. E. Davis at Caltech (10) Tabe-Mohammadi, A. A Review of the Applications of
(1993−1994). Before joining the University of Minnesota, he was a Membrane Separation Technology in Natural Gas Treatment. Sep.
faculty member in the Chemical Engineering Department at the Sci. Technol. 1999, 34, 2095−2111.
University of Massachusetts Amherst (1994−2003). He is a fellow of (11) Abatzoglou, N.; Boivin, S. A Review of Biogas Purification
the American Association for the Advancement of Science and a Processes. Biofuels, Bioprod. Biorefin. 2009, 3, 42−71.
member of the National Academy of Engineering. (12) Wiheeb, A. D.; Shamsudin, I. K.; Ahmad, M. A.; Murat, M. N.;
Kim, J.; Othman, M. R. Present Technologies for Hydrogen Sulfide
J. Ilja Siepmann is a Distinguished McKnight University Professor and
Removal from Gaseous Mixtures. Rev. Chem. Eng. 2013, 29, 449−470.
a Distinguished Teaching Professor and holds the Merck Professorship (13) Wang, L.; Yang, R. T. New Nanostructured Sorbents for
in Chemistry at the University of Minnesota. He received his Ph.D. in Desulfurization of Natural Gas. Front. Chem. Sci. Eng. 2014, 8, 8−19.
Chemistry from the University of Cambridge. Before joining the (14) Ahmed, I.; Jhung, S. H. Adsorptive Desulfurization and
University of Minnesota in 1994, Dr. Siepmann carried out Denitrogenation Using Metal-Organic Frameworks. J. Hazard. Mater.
postdoctoral research at the IBM Zurich Research Laboratory, the 2016, 301, 259−276.

9792 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(15) Vellingiri, K.; Deep, A.; Kim, K.-H. Metal-Organic Frameworks and Hydrogen Sulfide in Mixtures of Physical and Chemical Solvents.
as a Potential Platform for Selective Treatment of Gaseous Sulfur AIChE J. 1979, 25, 975−984.
Compounds. ACS Appl. Mater. Interfaces 2016, 8, 29835−29857. (39) Roberts, B. E.; Mather, A. E. Solubility of CO2 and H2S in a
(16) George, G.; Bhoria, N.; AlHallaq, S.; Abdala, A.; Mittal, V. Mixed Solvent. Chem. Eng. Commun. 1988, 72, 201−211.
Polymer Membranes for Acid Gas Removal from Natural Gas. Sep. (40) Macgregor, R. J.; Mather, A. E. Equilibrium Solubility of H2S
Purif. Technol. 2016, 158, 333−356. and CO2 and Their Mixtures in a Mixed Solvent. Can. J. Chem. Eng.
(17) Muñoz, R.; Meier, L.; Diaz, I.; Jeison, D. A Review on the State- 1991, 69, 1357−1366.
of-the-Art of Physical/Chemical and Biological Technologies for (41) Jenab, M. H.; Abdi, M. A.; Najibi, S. H.; Vahidi, M.; Matin, N. S.
Biogas Upgrading. Rev. Environ. Sci. Bio/Technol. 2015, 14, 727−759. Solubility of Carbon Dioxide in Aqueous Mixtures of N-
(18) Jensen, A. B.; Webb, C. Treatment of H2S-Containing Gases: A Methyldiethanolamine + Piperazine + Sulfolane. J. Chem. Eng. Data
Review of Microbiological Alternatives. Enzyme Microb. Technol. 1995, 2005, 50, 583−586.
17, 2−10. (42) Mokhatab, S.; Poe, W. A.; Mark, J. Y. Handbook of Natural Gas
(19) Syed, M.; Soreanu, G.; Falletta, P.; Béland, M. Removal of Transmission and Processing; Gulf Professional Publishing, 2012.
Hydrogen Sulfide from Gas Streams Using Biological Processes - A (43) Lei, Z.; Dai, C.; Chen, B. Gas Solubility in Ionic Liquids. Chem.
Review. Can. Biosyst. Eng. 2006, 48, 2. Rev. 2014, 114, 1289−1326.
(20) Piéplu, A.; Saur, O.; Lavalley, J.-C.; Legendre, O.; Nédez, C. (44) Karadas, F.; Atilhan, M.; Aparicio, S. Review on the Use of Ionic
Claus Catalysis and H2S Selective Oxidation. Catal. Rev.: Sci. Eng. Liquids (ILs) as Alternative Fluids for CO2 Capture and Natural Gas
1998, 40, 409−450. Sweetening. Energy Fuels 2010, 24, 5817−5828.
(21) Pandey, R. A.; Malhotra, S. Desulfurization of Gaseous Fuels (45) Bara, J. E. Green Solvents II; Springer: Berlin, 2012; pp 155−167.
with Recovery of Elemental Sulfur: An Overview. Crit. Rev. Environ. (46) Kumar, S.; Cho, J. H.; Moon, I. Ionic Liquid-Amine Blends and
Sci. Technol. 1999, 29, 229−268. CO2 BOLs: Prospective Solvents for Natural Gas Sweetening and CO2
(22) Eow, J. S. Recovery of Sulfur from Sour Acid Gas: A Review of Capture Technology-A Review. Int. J. Greenhouse Gas Control 2014,
the Technology. Environ. Prog. 2002, 21, 143−162. 20, 87−116.
(23) Reverberi, A. P.; Klemeš, J. J.; Varbanov, P. S.; Fabiano, B. A (47) Jou, F.-Y.; Mather, A. E. Solubility of Hydrogen Sulfide in
Review on Hydrogen Production from Hydrogen Sulphide by [bmim][PF6]. Int. J. Thermophys. 2007, 28, 490−495.
Chemical and Photochemical Methods. J. Cleaner Prod. 2016, 136, (48) Anthony, J. L.; Maginn, E. J.; Brennecke, J. F. Solubilities and
72−80. Thermodynamic Properties of Gases in the Ionic Liquid 1-n-Butyl-3-
(24) Bottoms, R. R. Process for Separating Acidic Gases. U.S. Patent Methylimidazolium Hexafluorophosphate. J. Phys. Chem. B 2002, 106,
US1,783,901, 1930. 7315−7320.
(25) Frazier, H. D.; Kohl, A. L. Selective Absorption of Hydrogen (49) Jalili, A. H.; Rahmati-Rostami, M.; Ghotbi, C.; Hosseini-Jenab,
Sulfide from Gas Streams. Ind. Eng. Chem. 1950, 42, 2288−2292. M.; Ahmadi, A. N. Solubility of H2S in Ionic Liquids [bmim][PF6],
(26) Carey, T. R.; Hermes, J. E.; Rochelle, G. T. A Model of Acid Gas [bmim][BF4], and [bmim][Tf2N]. J. Chem. Eng. Data 2009, 54,
Absorption/Stripping Using Methyldiethanolamine with Added Acid. 1844−1849.
Gas Sep. Purif. 1991, 5, 95−109. (50) Raeissi, S.; Peters, C. J. High Pressure Phase Behaviour of
(27) Mandal, B. P.; Biswas, A. K.; Bandyopadhyay, S. S. Selective Methane in 1-Butyl-3-Methylimidazolium Bis-
Absorption of H2S from Gas Streams Containing H2S and CO2 into (trifluoromethylsulfonyl)imide. Fluid Phase Equilib. 2010, 294, 67−71.
Aqueous Solutions of N-Methyldiethanolamine and 2-Amino-2- (51) Jacquemin, J.; Gomes, M. F. C.; Husson, P.; Majer, V. Solubility
Methyl-1-Propanol. Sep. Purif. Technol. 2004, 35, 191−202. of Carbon Dioxide, Ethane, Methane, Oxygen, Nitrogen, Hydrogen,
(28) Saha, A. K.; Bandyopadhyay, S. S.; Saju, P.; Biswas, A. K. Argon, and Carbon Monoxide in 1-Butyl-3-Methylimidazolium
Selective Removal of Hydrogen Sulfide from Gases Containing Tetrafluoroborate Between Temperatures 283 and 343 K and at
Hydrogen Sulfide and Carbon Dioxide by Absorption into Aqueous Pressures Close to Atmospheric. J. Chem. Thermodyn. 2006, 38, 490−
Solutions of 2-Amino-2-methyl-1-propanol. Ind. Eng. Chem. Res. 1993, 502.
32, 3051−3055. (52) Rahmati-Rostami, M.; Ghotbi, C.; Hosseini-Jenab, M.; Ahmadi,
(29) Sartori, G.; Savage, D. W. Sterically Hindered Amines for A. N.; Jalili, A. H. Solubility of H2S in Ionic Liquids [hmim][PF6],
Carbon Dioxide Removal from Gases. Ind. Eng. Chem. Fundam. 1983, [hmim][BF4], and [hmim][Tf2N]. J. Chem. Thermodyn. 2009, 41,
22, 239−249. 1052−1055.
(30) Suleman, H.; Maulud, A. S.; Man, Z. Review and Selection (53) Finotello, A.; Bara, J. E.; Camper, D.; Noble, R. D. Room-
Criteria of Classical Thermodynamic Models for Acid Gas Absorption Temperature Ionic Liquids: Temperature Dependence of Gas
in Aqueous Alkanolamines. Rev. Chem. Eng. 2015, 31, 599−639. Solubility Selectivity. Ind. Eng. Chem. Res. 2008, 47, 3453−3459.
(31) Kohl, A. L.; Nielsen, R. B. Gas Purification, 5th ed.; Gulf (54) Sakhaeinia, H.; Taghikhani, V.; Jalili, A. H.; Mehdizadeh, A.;
Publishing Co.: Houston, TX, 1997. Safekordi, A. A. Solubility of H2S in 1-(2-Hydroxyethyl)-3-
(32) Minkkinen, A.; Larue, J. Y. M.; Patel, S.; Levier, J.-F. Methanol Methylimidazolium Ionic Liquids with Different Anions. Fluid Phase
Gas-Treating Scheme Offers Economics, Versatility Oil Gas J. 1992.90 Equilib. 2010, 298, 303−309.
(33) Ranke, G.; Mohr, V. H. The Rectisol Wash: New Developments (55) Shokouhi, M.; Adibi, M.; Jalili, A. H.; Hosseini-Jenab, M.;
in Acid Gas Removal from Synthesis Gas. In Acid and Sour Gas Mehdizadeh, A. Solubility and Diffusion of H2S and CO2 in the Ionic
Treating Processes; Newman, S. A., Ed.; Gulf Publishing Co.: Houston, Liquid 1-(2-Hydroxyethyl)-3-Methylimidazolium Tetrafluoroborate. J.
TX, 1985; pp 80−111. Chem. Eng. Data 2010, 55, 1663−1668.
(34) Burr, B.; Lyddon, L. A Comparison of Physical Solvents for Acid (56) Sakhaeinia, H.; Jalili, A. H.; Taghikhani, V.; Safekordi, A. A.
Gas Removal, 87th Annual Gas Processors Association Convention, Solubility of H2S in Ionic Liquids 1-Ethyl-3-methylimidazolium
Grapevine, TX, March 2008. Hexafluorophosphate ([emim][PF6]) and 1-Ethyl-3-methylimidazo-
(35) Hochgesand, G. Rectisol and Purisol. Ind. Eng. Chem. 1970, 62, lium Bis (trifluoromethyl) sulfonylimide ([emim][Tf2N]). J. Chem.
37−43. Eng. Data 2010, 55, 5839−5845.
(36) Roberts, B. E.; Mather, A. E. Solubility of H2S and CO2 in (57) Jalili, A. H.; Mehdizadeh, A.; Shokouhi, M.; Ahmadi, A. N.;
Sulfolane. Can. J. Chem. Eng. 1988, 66, 519−520. Hosseini-Jenab, M.; Fateminassab, F. Solubility and Diffusion of CO2
(37) Dunn, C. L.; Freitas, E. R.; Goodenbour, J. W.; Henderson, H. and H2S in the Ionic Liquid 1-Ethyl-3-Methylimidazolium Ethylsulfate.
T.; Papadopoulous, M. N. Sulfinol Process. Oil Gas J. 1964, 62, 95−98. J. Chem. Thermodyn. 2010, 42, 1298−1303.
(38) Rivas, O. R.; Prausnitz, J. M. Sweetening of Sour Natural Gases (58) Jalili, A. H.; Safavi, M.; Ghotbi, C.; Mehdizadeh, A.; Hosseini-
by Mixed-Solvent Absorption: Solubilities of Ethane, Carbon Dioxide, Jenab, M.; Taghikhani, V. Solubility of CO2, H2S, and Their Mixture in

9793 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

the Ionic Liquid 1-Octyl-3-Methylimidazolium Bis (Trifluoromethyl) (75) Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H. CO2 Capture
Sulfonylimide. J. Phys. Chem. B 2012, 116, 2758−2774. by a Task-Specific Ionic Liquid. J. Am. Chem. Soc. 2002, 124, 926−927.
(59) Huang, K.; Cai, D.-N.; Chen, Y.-L.; Wu, Y.-T.; Hu, X.-B.; Zhang, (76) Shiflett, M. B.; Yokozeki, A. Phase Behavior of Carbon Dioxide
Z.-B. Thermodynamic Validation of 1-Alkyl-3-Methylimidazolium in Ionic Liquids:[emim][Acetate],[emim][Trifluoroacetate], and
Carboxylates as Task-Specific Ionic Liquids for H2S Absorption. [emim][Acetate]+[emim][Trifluoroacetate] Mixtures. J. Chem. Eng.
AIChE J. 2013, 59, 2227−2235. Data 2009, 54, 108−114.
(60) Huang, K.; Zhang, X.-M.; Xu, Y.; Wu, Y.-T.; Hu, X.-B.; Xu, Y. (77) Huang, K.; Cai, D.-N.; Chen, Y.-L.; Wu, Y.-T.; Hu, X.-B.; Zhang,
Protic Ionic Liquids for the Selective Absorption of H2S from CO2: Z.-B. Dual Lewis Base Functionalization of Ionic Liquids for Highly
Thermodynamic Analysis. AIChE J. 2014, 60, 4232−4240. Efficient and Selective Capture of H2S. ChemPlusChem 2014, 79, 241−
(61) Huang, K.; Zhang, X.-M.; Hu, X.-B.; Wu, Y.-T. Hydrophobic 249.
Protic Ionic Liquids Tethered with Tertiary Amine Group for Highly (78) Greaves, T. L.; Drummond, C. J. Protic Ionic Liquids:
Efficient and Selective Absorption of H2S from CO 2. AIChE J. 2016, Properties and Applications. Chem. Rev. 2008, 108, 206−237.
62, 4480−4490. (79) Freire, M. G.; Santos, L. M. N. B. F.; Fernandes, A. M.;
(62) Pomelli, C. S.; Chiappe, C.; Vidis, A.; Laurenczy, G.; Dyson, P. J. Coutinho, J. A. P.; Marrucho, I. M. An Overview of the Mutual
Influence of the Interaction Between Hydrogen Sulfide and Ionic Solubilities of Water-Imidazolium-Based Ionic Liquids Systems. Fluid
Liquids on Solubility: Experimental and Theoretical Investigation. J. Phase Equilib. 2007, 261, 449−454.
Phys. Chem. B 2007, 111, 13014−13019. (80) Rogers, R. D.; Seddon, K. R. Ionic Liquids-Solvents of the
(63) Singh, R.; Marin-Rimoldi, E.; Maginn, E. J. A Monte Carlo Future? Science 2003, 302, 792−793.
Simulation Study To Predict the Solubility of Carbon Dioxide, (81) Guo, B.; Duan, E.; Zhong, Y.; Gao, L.; Zhang, X.; Zhao, D.
Hydrogen, and Their Mixture in the Ionic Liquids 1-Alkyl-3- Absorption and Oxidation of H2S in Caprolactam Tetrabutyl
Methylimidazolium Bis(trifluoromethanesulfonyl) Amide ([Cnmim+]- Ammonium Bromide Ionic Liquid. Energy Fuels 2011, 25, 159−161.
[Tf2N−], n = 4, 6. Ind. Eng. Chem. Res. 2015, 54, 4385−4395. (82) Heldebrant, D. J.; Yonker, C. R.; Jessop, P. G.; Phan, L. Organic
(64) Budhathoki, S.; Shah, J. K.; Maginn, E. J. Molecular Simulation Liquid CO2 Capture Agents with High Gravimetric CO2 Capacity.
Study of the Solubility, Diffusivity and Permselectivity of Pure and Energy Environ. Sci. 2008, 1, 487−493.
Binary Mixtures of CO2 and CH4 in the Ionic Liquid 1-n-Butyl-3- (83) Heldebrant, D. J.; Yonker, C. R.; Jessop, P. G.; Phan, L.
Methylimidazolium Bis(trifluoromethylsulfonyl) Imide. Ind. Eng. Reversible Uptake of COS, CS2, and SO2: Ionic Liquids with O-
Chem. Res. 2015, 54, 8821−8828. Alkylxanthate, O-Alkylthiocarbonyl, and O-Alkylsulfite Anions. Chem. -
(65) Sánchez-Badillo, J.; Gallo, M.; Alvarado, S.; Glossman-Mitnik, Eur. J. 2009, 15, 7619−7627.
D. Solvation Thermodynamic Properties of Hydrogen Sulfide in (84) Koech, P. K.; Rainbolt, J. E.; Bearden, M. D.; Zheng, F.;
[C4mim][PF6],[C4mim][BF4], and [C4mim][Cl] Ionic Liquids, Heldebrant, D. J. Chemically Selective Gas Sweetening Without
Determined by Molecular Simulations. J. Phys. Chem. B 2015, 119, Thermal-Swing Regeneration. Energy Environ. Sci. 2011, 4, 1385−
10727−10737.
1390.
(66) Ramdin, M.; Balaji, S. P.; Vicent-Luna, J. M.; GutiŽ rrez-
(85) Abbott, A. P.; Capper, G.; Davies, D. L.; Munro, H. L.; Rasheed,
Sevillano, J. J.; Calero, S.; de Loos, T. W.; Vlugt, T. J. H. Solubility of
R. K.; Tambyrajah, V. Preparation of Novel, Moisture-Stable, Lewis-
the Precombustion Gases CO2, CH4, CO, H2, N2, and H2S in the Ionic
Acidic Ionic Liquids Containing Quaternary Ammonium Salts With
Liquid [bmim][Tf2N] from Monte Carlo Simulations. J. Phys. Chem. C
Functional Side Chains. Chem. Commun. 2001, 2010−2011.
2014, 118, 23599−23604.
(86) Smith, E. L.; Abbott, A. P.; Ryder, K. S. Deep Eutectic Solvents
(67) Jamali, S. H.; Ramdin, M.; Becker, T. M.; Torres-Knoop, A.;
(DESs) and Their Applications. Chem. Rev. 2014, 114, 11060−11082.
Dubbeldam, D.; Buijs, W.; Vlugt, T. J. H. Solubility of Sulfur
(87) Li, X.; Hou, M.; Han, B.; Wang, X.; Zou, L. Solubility of CO2 in
Compounds in Commercial Physical Solvents and an Ionic Liquid
a Choline Chloride + Urea Eutectic Mixture. J. Chem. Eng. Data 2008,
from Monte Carlo Simulations. Fluid Phase Equilib. 2017, 433, 50−55.
(68) Heintz, Y. J.; Sehabiague, L.; Morsi, B. I.; Jones, K. L.; Luebke, 53, 548−550.
(88) Sarmad, S.; Xie, Y.; Mikkola, J.-P.; Ji, X. Screening of Deep
D. R.; Pennline, H. W. Hydrogen Sulfide and Carbon Dioxide
Removal from Dry Fuel Gas Streams Using an Ionic Liquid as a Eutectic Solvents (DESs) as Green CO2 Sorbents: From Solubility to
Physical Solvent. Energy Fuels 2009, 23, 4822−4830. Viscosity. New J. Chem. 2017, 41, 290−301.
(69) Carvalho, P. J.; Coutinho, J. A. P. Non-Ideality of Solutions of (89) Aissaoui, T.; Al Nashef, I. M.; Qureshi, U. A.; Benguerba, Y.
NH3, SO2, and H2S in Ionic Liquids and the Prediction of Their Potential Applications of Deep Eutectic Solvents in Natural Gas
Solubilities Using the Flory-Huggins Model. Energy Fuels 2010, 24, Sweetening for CO2 Capture. Rev. Chem. Eng. 2017, DOI: 10.1515/
6662−6666. revce-2016-0013.
(70) Camper, D.; Bara, J.; Koval, C.; Noble, R. Bulk-Fluid Solubility (90) Rahmati-Rostami, M.; Behzadi, B.; Ghotbi, C. Thermodynamic
and Membrane Feasibility of Rmim-Based Room-Temperature Ionic Modeling of Hydrogen Sulfide Solubility in Ionic Liquids Using
Liquids. Ind. Eng. Chem. Res. 2006, 45, 6279−6283. Modified SAFT-VR and PC-SAFT Equations of State. Fluid Phase
(71) Scovazzo, P. Determination of the Upper Limits, Benchmarks, Equilib. 2011, 309, 179−189.
and Critical Properties for Gas Separations Using Stabilized Room (91) Llovell, F.; Marcos, R. M.; MacDowell, N.; Vega, L. F. Modeling
Temperature Ionic Liquid Membranes (SILMs) for the Purpose of the Absorption of Weak Electrolytes and Acid Gases with Ionic
Guiding Future Research. J. Membr. Sci. 2009, 343, 199−211. Liquids Using the Soft-SAFT Approach. J. Phys. Chem. B 2012, 116,
(72) Carvalho, P. J.; Coutinho, J. A. P. The Polarity Effect Upon the 7709−7718.
Methane Solubility in Ionic Liquids: A Contribution for the Design of (92) Shahriari, R.; Dehghani, M. R.; Behzadi, B. A Modified Polar
Ionic Liquids for Enhanced CO2/CH4 and H2S/CH4 Selectivities. PHSC Model for Thermodynamic Modeling of Gas Solubility in Ionic
Energy Environ. Sci. 2011, 4, 4614−4619. Liquids. Fluid Phase Equilib. 2012, 313, 60−72.
(73) Alcalde, R.; Garcia, G.; Atilhan, M.; Aparicio, S. Systematic (93) Jalili, A. H.; Shokouhi, M.; Maurer, G.; Hosseini-Jenab, M.
Study on the Viscosity of Ionic Liquids: Measurement and Prediction. Solubility of CO2 and H2 S in the Ionic Liquid 1-Ethyl-3-
Ind. Eng. Chem. Res. 2015, 54, 10918−10924. Methylimidazolium Tris (Pentafluoroethyl) Trifluorophosphate. J.
(74) Koller, T. M.; Heller, A.; Rausch, M. H.; Wasserscheid, P.; Chem. Thermodyn. 2013, 67, 55−62.
Economou, I. G.; Fröba, A. P. Mutual and Self-Diffusivities in Binary (94) Safavi, M.; Ghotbi, C.; Taghikhani, V.; Jalili, A. H.; Mehdizadeh,
Mixtures of [EMIM][B(CN)4] with Dissolved Gases by Using A. Study of the Solubility of CO2, H2S and Their Mixture in the Ionic
Dynamic Light Scattering and Molecular Dynamics Simulations. J. Liquid 1-Octyl-3-Methylimidazolium Hexafluorophosphate: Experi-
Phys. Chem. B 2015, 119, 8583−8592. mental and Modelling. J. Chem. Thermodyn. 2013, 65, 220−232.

9794 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(95) Faúndez, C. A.; Díaz-Valdés, J. F.; Valderrama, J. O. Testing Streams in the Presence of H2, CO2, and H2O. J. Catal. 2005, 236,
Solubility Data of H2S and SO2 in Ionic Liquids for Sulfur-Removal 205−220.
Processes. Fluid Phase Equilib. 2014, 375, 152−160. (117) Xue, M.; Chitrakar, R.; Sakane, K.; Ooi, K. Screening of
(96) Ji, X.; Held, C.; Sadowski, G. Modeling Imidazolium-Based Adsorbents for Removal of H2S at Room Temperature. Green Chem.
Ionic Liquids With ePC-SAFT. Part II. Application to H2S and 2003, 5, 529−534.
Synthesis-Gas Components. Fluid Phase Equilib. 2014, 363, 59−65. (118) Casarin, M.; Maccato, C.; Vittadini, A. An LCAO-LDF Study
(97) Haghtalab, A.; Kheiri, A. High Pressure Measurement Equation of the Chemisorption of H2O and H2S on ZnO (0001) and ZnO (10
of State for Solubility of Carbon Dioxide and Hydrogen Sulfide in 1- 1̅0. Surf. Sci. 1997, 377-379, 587−591.
Butyl-3-Methylimidazolium Acetate. J. Chem. Thermodyn. 2015, 89, (119) Goclon, J.; Meyer, B. The Interaction of H2S with the ZnO (10
41−50. 1̅0) Surface. Phys. Chem. Chem. Phys. 2013, 15, 8373−8382.
(98) Llovell, F.; Oliveira, M. B.; Coutinho, J. A. P.; Vega, L. F. (120) Song, J.; Niu, X.; Ling, L.; Wang, B. A Density Functional
Solubility of Greenhouse and Acid Gases on the [C4mim][MeSO4] Theory Study on the Interaction Mechanism Between H2S and the α-
Ionic Liquid for Gas Separation and CO2 Conversion. Catal. Today Fe2O3 (0001) Surface. Fuel Process. Technol. 2013, 115, 26−33.
2015, 255, 87−96. (121) Chen, H.-T.; Choi, Y.; Liu, M.; Lin, M. C. A First-Principles
(99) Vega, L. F.; Vilaseca, O.; Llovell, F.; Andreu, J. S. Modeling Analysis for Sulfur Tolerance of CeO2 in Solid Oxide Fuel Cells. J.
Ionic Liquids and the Solubility of Gases in Them: Recent Advances Phys. Chem. C 2007, 111, 11117−11122.
and Perspectives. Fluid Phase Equilib. 2010, 294, 15−30. (122) Huang, W.-F.; Chen, H.-T.; Lin, M. C. Density Functional
(100) Ahmadi, M. A.; Haghbakhsh, R.; Soleimani, R.; Bajestani, M. B. Theory Study of the Adsorption and Reaction of H2S on TiO2 Rutile
Estimation of H2S Solubility in Ionic Liquids Using a Rigorous (110) and Anatase (101) Surfaces. J. Phys. Chem. C 2009, 113, 20411−
Method. J. Supercrit. Fluids 2014, 92, 60−69. 20420.
(101) Shafiei, A.; Ahmadi, M. A.; Zaheri, S. H.; Baghban, A.; (123) Jia, M.-Y.; Xu, B.; Ding, X.-L.; He, S.-G.; Ge, M.-F.
Amirfakhrian, A.; Soleimani, R. Estimating Hydrogen Sulfide Solubility Experimental and Theoretical Study of the Reactions Between
in Ionic Liquids Using a Machine Learning Approach. J. Supercrit. Manganese Oxide Cluster Anions and Hydrogen Sulfide. J. Phys.
Fluids 2014, 95, 525−534. Chem. C 2012, 116, 24184−24192.
(102) Ahmadi, M.-A.; Pouladi, B.; Javvi, Y.; Alfkhani, S.; Soleimani, R. (124) Jia, M.-Y.; Xu, B.; Ding, X.-L.; Zhao, Y.-X.; He, S.-G.; Ge, M.-F.
Connectionist Technique Estimates H2S Solubility in Ionic Liquids Experimental and Theoretical Study of the Reactions Between
Through a Low Parameter Approach. J. Supercrit. Fluids 2015, 97, 81− Vanadium Oxide Cluster Cations and Hydrogen Sulfide. J. Phys.
87. Chem. C 2012, 116, 9043−9048.
(103) Hamzehie, M.; Fattahi, M.; Najibi, H.; Van der Bruggen, B.; (125) Jia, M.-Y.; Xu, B.; Deng, K.; He, S.-G.; Ge, M.-F. Consecutive
Mazinani, S. Application of artificial neural networks for estimation of Oxygen-for-Sulfur Exchange Reactions between Vanadium Oxide
solubility of acid gases (H 2 S and CO 2) in 32 commonly ionic liquid Cluster Anions and Hydrogen Sulfide. J. Phys. Chem. A 2014, 118,
and amine solutions. J. Nat. Gas Sci. Eng. 2015, 24, 106−114. 8106−8114.
(104) Zhao, Y.; Gao, H.; Zhang, X.; Huang, Y.; Bao, D.; Zhang, S. (126) Zhang, R.; Liu, H.; Li, J.; Ling, L.; Wang, B. A Mechanistic
Hydrogen Sulfide Solubility in Ionic Liquids (ILs): An Extensive Study of H2S Adsorption and Dissociation on Cu2O (111) Surfaces:
Database and a New ELM Model Mainly Established by Imidazolium- Thermochemistry, Reaction Barrier. Appl. Surf. Sci. 2012, 258, 9932−
Based ILs. J. Chem. Eng. Data 2016, 61, 3970−3978. 9943.
(105) Mortazavi-Manesh, S.; Satyro, M. A.; Marriott, R. A. Screening (127) Sun, S.; Zhang, D.; Li, C.; Wang, Y. DFT Study on the
Ionic Liquids as Candidates for Separation of Acid Gases: Solubility of Adsorption and Dissociation of H2S on CuO (111) Surface. RSC Adv.
Hydrogen Sulfide, Methane, and Ethane. AIChE J. 2013, 59, 2993− 2015, 5, 21806−21811.
3005. (128) Jia, M.-Y.; Luo, Z.; He, S.-G.; Ge, M.-F. Oxygen-Sulfur
(106) Tennyson, R. N.; Schaaf, R. P. Guidelines can Help Choose Exchange and the Gas-Phase Reactivity of Cobalt Sulfide Cluster
Proper Process for Gas-Treating Plants. Oil Gas J. 1977, 75, 78−86. Anions with Molecular Oxygen. J. Phys. Chem. A 2014, 118, 8163−
(107) Bureau, A. C.; Olden, M. J. F. Operation of the Frodingham 8169.
Desulfurizing Plant at Exeter Chem. Eng. (London). 1967.206 (129) Bagheri, Z.; Moradi, M. DFT Study on the Adsorption and
(108) Phillipson, J. J. Desulfurization. Catalyst Handbook; Wolf Dissociation of Hydrogen Sulfide on MgO Nanotube. Struct. Chem.
Scientific Books: London, 1970. 2014, 25, 495−501.
(109) Vamvuka, D.; Arvanitidis, C.; Zachariadis, D. Flue Gas (130) Jia, M.-Y.; Ding, X.-L.; He, S.-G.; Ge, M.-F. Experimental and
Desulfurization at High Temperatures: A Review. Environ. Eng. Sci. Theoretical Study of the Reactions between MO−2 (M = Fe, Co, Ni,
2004, 21, 525−548. Cu, and Zn) Cluster Anions and Hydrogen Sulfide. J. Phys. Chem. A
(110) Cheah, S.; Carpenter, D. L.; Magrini-Bair, K. A. Review of Mid- 2013, 117, 8377−8387.
to High-Temperature Sulfur Sorbents for Desulfurization of Biomass- (131) Baird, T.; Denny, P. J.; Hoyle, R.; McMonagle, F.; Stirling, D.;
and Coal-Derived Syngas. Energy Fuels 2009, 23, 5291−5307. Tweedy, J. Modified Zinc Oxide Absorbents for Low-Temperature
(111) Westmoreland, P. R.; Harrison, D. P. Evaluation of Candidate Gas Desulfurisation. J. Chem. Soc., Faraday Trans. 1992, 88, 3375−
Solids for High-Temperature Desulfurization of Low-Btu Gases. 3382.
Environ. Sci. Technol. 1976, 10, 659−661. (132) Baird, T.; Campbell, K. C.; Holliman, P. J.; Hoyle, R. W.;
(112) Elseviers, W. F.; Verelst, H. Transition Metal Oxides for Hot Huxam, M.; Stirling, D.; Williams, B. P.; Morris, M. Cobalt-zinc oxide
Gas Desulphurisation. Fuel 1999, 78, 601−612. absorbents for low temperature gas desulfurisation. J. Mater. Chem.
(113) Rodriguez, J. A.; Chaturvedi, S.; Kuhn, M.; Hrbek, J. Reaction 1999, 9, 599−605.
of H2S and S2 with Metal/Oxide Surfaces: Band-Gap Size and (133) Carnes, C. L.; Klabunde, K. J. Unique Chemical Reactivities of
Chemical Reactivity. J. Phys. Chem. B 1998, 102, 5511−5519. Nanocrystalline Metal Oxides Toward Hydrogen Sulfide. Chem. Mater.
(114) Rodriguez, J. A.; Jirsak, T.; Chaturvedi, S. Reaction of H2S with 2002, 14, 1806−1811.
MgO (100) and Cu/MgO (100) Surfaces: Band-Gap Size and (134) Habibi, R.; Rashidi, A. M.; Daryan, J. T.; Zadeh, A. M. A. Study
Chemical Reactivity. J. Chem. Phys. 1999, 111, 8077−8087. of the Rod-Like and Spherical Nano-ZnO Morphology on H2S
(115) Van Nisselrooy, P. F. M. T.; Lagas, J. A. Superclaus Reduces Removal from Natural Gas. Appl. Surf. Sci. 2010, 257, 434−439.
SO2 Emission by the Use of a New Selective Oxidation Catalyst. Catal. (135) Garces, H. F.; Galindo, H. M.; Garces, L. J.; Hunt, J.; Morey,
Today 1993, 16, 263−271. A.; Suib, S. L. Low Temperature H2S Dry-Desulfurization with Zinc
(116) Polychronopoulou, K.; Galisteo, F. C.; Granados, M. L.; Fierro, Oxide. Microporous Mesoporous Mater. 2010, 127, 190−197.
J. L. G.; Bakas, T.; Efstathiou, A. M. Novel Fe-Mn-Zn-Ti-O Mixed- (136) Garces, H. F.; Espinal, A. E.; Suib, S. L. Tunable Shape
Metal Oxides for the Low-Temperature Removal of H2S from Gas Microwave Synthesis of Zinc Oxide Nanospheres and Their

9795 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

Desulfurization Performance Compared with Nanorods and Platelet- (155) Davidson, J. M.; Denny, P. J.; Lawrie, C. H. Autocatalysis by
Like Morphologies for the Removal of Hydrogen Sulfide. J. Phys. Water in the Reaction of Hydrogen Sulphide with Zinc Oxide. J. Chem.
Chem. C 2012, 116, 8465−8474. Soc., Chem. Commun. 1989, 1695−1696.
(137) Davidson, J. M.; Lawrie, C. H.; Sohail, K. Kinetics of the (156) Liu, C.; Zhang, R.; Wei, S.; Wang, J.; Liu, Y.; Li, M.; Liu, R.
Absorption of Hydrogen Sulfide by High Purity and Doped High Selective Removal of H2S from Biogas Using a Regenerable Hybrid
Surface Area Zinc Oxide. Ind. Eng. Chem. Res. 1995, 34, 2981−2989. TiO2/Zeolite Composite. Fuel 2015, 157, 183−190.
(138) Novochinskii, I. I.; Song, C.; Ma, X.; Liu, X.; Shore, L.; (157) Liu, D.; Chen, S.; Fei, X.; Huang, C.; Zhang, Y. Regenerable
Lampert, J.; Farrauto, R. J. Low-Temperature H2S Removal from CuO-Based Adsorbents for Low Temperature Desulfurization
Steam-Containing Gas Mixtures with ZnO for Fuel Cell Application. 1. Application. Ind. Eng. Chem. Res. 2015, 54, 3556−3562.
ZnO Particles and Extrudates. Energy Fuels 2004, 18, 576−583. (158) Fan, H.-L.; Sun, T.; Zhao, Y.-P.; Shangguan, J.; Lin, J.-Y.
(139) Yagi, S.; Kunii, D. Studies on Combustion of Carbon Particles Three-Dimensionally Ordered Macroporous Iron Oxide for Removal
in Flames and Fluidized Beds. Symp. (Int.) Combust., [Proc.] 1955, 5, of H2S at Medium Temperatures. Environ. Sci. Technol. 2013, 47,
231−244. 4859−4865.
(140) Szekely, J.; Evans, J. W. A Structural Model for Gas-Solid (159) Huang, G.; He, E.; Wang, Z.; Fan, H.; Shangguan, J.; Croiset,
Reactions with a Moving Boundary. Chem. Eng. Sci. 1970, 25, 1091− E.; Chen, Z. Synthesis and Characterization of γ-Fe2O3 for H2S
1107. Removal at Low Temperature. Ind. Eng. Chem. Res. 2015, 54, 8469−
(141) Gibson, J. B., III; Harrison, D. P. The Reaction Between 8478.
Hydrogen Sulfide and Spherical Pellets of Zinc Oxide. Ind. Eng. Chem. (160) Wang, L.-J.; Fan, H.-L.; Shangguan, J.; Croiset, E.; Chen, Z.;
Process Des. Dev. 1980, 19, 231−237. Wang, H.; Mi, J. Design of a Sorbent to Enhance Reactive Adsorption
(142) Ranade, P. V.; Harrison, D. P. The Variable Property Grain of Hydrogen Sulfide. ACS Appl. Mater. Interfaces 2014, 6, 21167−
Model Applied to the Zinc Oxide-Hydrogen Sulfide Reaction. Chem. 21177.
Eng. Sci. 1981, 36, 1079−1089. (161) Pahalagedara, L. R.; Poyraz, A. S.; Song, W.; Kuo, C.-H.;
(143) Evans, J.; Corker, J. M.; Hayter, C. E.; Oldman, R. J.; Williams, Pahalagedara, M. N.; Meng, Y.-T.; Suib, S. L. Low Temperature
B. P. In Situ Sulfur K-Edge X-Ray Absorption Spectroscopy of the Desulfurization of H2S: High Sorption Capacities by Mesoporous
Reaction of Zinc Oxide with Hydrogen Sulfide. Chem. Commun. 1996, Cobalt Oxide via Increased H2S Diffusion. Chem. Mater. 2014, 26,
1431−1432. 6613−6621.
(144) Hayter, C. E.; Evans, J.; Corker, J. M.; Oldman, R. J.; Williams, (162) Chung, J. B.; Chung, J. S. Desulfurization of H2S Using Cobalt-
B. P. Sulfur K-Edge X-Ray Absorption Spectroscopy Study of the Containing Sorbents at Low Temperatures. Chem. Eng. Sci. 2005, 60,
Reaction of Zinc Oxide with Hydrogen Sulfide. J. Mater. Chem. 2002,
1515−1523.
12, 3172−3177. (163) Pouralhosseini, S. Making ZnO Adsorption More Affordable to
(145) Samokhvalov, A.; Tatarchuk, B. J. Characterization of Active
Remove H2S by Changing the Catalysts Regeneration Method. Pet. Sci.
Sites, Determination of Mechanisms of H2S, COS and CS2 Sorption
Technol. 2013, 31, 62−67.
and Regeneration of ZnO Low-Temperature Sorbents: Past, Current
(164) McCarty, J. G.; Wise, H. Thermodynamics of Sulfur
and Perspectives. Phys. Chem. Chem. Phys. 2011, 13, 3197−3209.
Chemisorption on Metals. I. Alumina-Supported Nickel. J. Chem.
(146) Neveux, L.; Chiche, D.; Bazer-Bachi, D.; Favergeon, L.; Pijolat,
Phys. 1980, 72, 6332−6337.
M. New Insight on the ZnO Sulfidation Reaction: Evidences for an
(165) McCarty, J. G.; Wise, H. Thermodynamics of Sulfur
Outward Growth Process of the ZnS Phase. Chem. Eng. J. 2012, 181-
Chemisorption on Metals. II. Alumina-Supported Ruthenium. J.
182, 508−515.
(147) Neveux, L.; Chiche, D.; Perez-Pellitero, J.; Favergeon, L.; Gay, Chem. Phys. 1981, 74, 5877−5880.
A.-S.; Pijolat, M. New Insight into the ZnO Sulfidation Reaction: (166) McCarty, J. G.; Wise, H. Thermodynamics of Sulfur
Mechanism and Kinetics Modeling of the ZnS Outward Growth. Phys. Chemisorption on Metals. III. Iron and Cobalt. J. Chem. Phys. 1982,
Chem. Chem. Phys. 2013, 15, 1532−1545. 76, 1162−1167.
(148) Skrzypski, J.; Bezverkhyy, I.; Heintz, O.; Bellat, J.-P. Low (167) McCarty, J. G.; Sancier, K. M.; Wise, H. Thermodynamics of
Temperature H2S Removal with Metal-Doped Nanostructure ZnO Sulfur Chemisorption on Metals: IV. Alumina-Supported Platinum. J.
Sorbents: Study of the Origin of Enhanced Reactivity in Cu- Catal. 1983, 82, 92−97.
Containing Materials. Ind. Eng. Chem. Res. 2011, 50, 5714−5722. (168) Den Besten, I. E.; Selwood, P. W. The Chemisorption of
(149) Jiang, D.; Su, L.; Ma, L.; Yao, N.; Xu, X.; Tang, H.; Li, X. Cu- Hydrogen Sulfide, Methyl Sulfide, and Cyclohexene on Supported
Zn-Al Mixed Metal Oxides Derived from Hydroxycarbonate Nickel Catalysts. J. Catal. 1962, 1, 93−102.
Precursors for H2S Removal at Low Temperature. Appl. Surf. Sci. (169) Oliphant, J. L.; Fowler, R. W.; Pannell, R. B.; Bartholomew, C.
2010, 256, 3216−3223. H. Chemisorption of Hydrogen Sulfide on Nickel and Ruthenium
(150) Bezverkhyy, I.; Skrzypski, J.; Safonova, O.; Bellat, J.-P. Catalysts: I. Desorption Isotherms. J. Catal. 1978, 51, 229−242.
Sulfidation Mechanism of Pure and Cu-Doped ZnO Nanoparticles (170) Murray, E. P.; Tsai, T.; Barnett, S. A. A Direct-Methane Fuel
at Moderate Temperature: TEM and In Situ XRD Studies. J. Phys. Cell with a Ceria-Based Anode. Nature 1999, 400, 649−651.
Chem. C 2012, 116, 14423−14430. (171) Li, L.; King, D. L.; Liu, J.; Huo, Q.; Zhu, K.; Wang, C.; Gerber,
(151) Baird, T.; Campbell, K. C.; Holliman, P. J.; Hoyle, R.; Stirling, M.; Stevens, D.; Wang, Y. Stabilization of Metal Nanoparticles in
D.; Williams, P. B. Mixed Co-Zn-Al Oxides as Absorbents for Low- Cubic Mesostructured Silica and its Application in Regenerable Deep
Temperature Gas Desulfurisation. J. Chem. Soc., Faraday Trans. 1995, Desulfurization of Warm Syngas. Chem. Mater. 2009, 21, 5358−5364.
91, 3219−3230. (172) Yaghi, O. M.; Li, G.; Li, H. Selective Binding and Removal of
(152) Polychronopoulou, K.; Fierro, J. L. G.; Efstathiou, A. M. Novel Guests in a Microporous Metal-Organic Framework. Nature 1995,
Zn-Ti-based Mixed Metal Oxides for Low-Temperature Adsorption of 378, 703−706.
H2S from Industrial Gas Streams. Appl. Catal., B 2005, 57, 125−137. (173) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.;
(153) Fen, L.; Bo, Y.; Jie, Z.; Anxi, J.; Chunhong, S.; Xiangji, K.; Xin, O’Keeffe, M.; Yaghi, O. M. Systematic Design of Pore Size and
W. Study on Desulfurization Efficiency and Products of Ce-Doped Functionality in Isoreticular MOFs and Their Application in Methane
Nanosized ZnO Desulfurizer at Ambient Temperature. J. Rare Earths Storage. Science 2002, 295, 469−472.
2007, 25, 306−310. (174) Deng, H.; Grunder, S.; Cordova, K. E.; Valente, C.; Furukawa,
(154) Westmoreland, P. R.; Gibson, J. B.; Harrison, D. P. H.; Hmadeh, M.; Gándara, F.; Whalley, A. C.; Liu, Z.; Asahina, S.;
Comparative Kinetics of High-Temperature Reaction Between Kazumori, H.; O’Keeffe, M.; Terasaki, O.; Stoddart, J. F.; Yaghi, O. M.
Hydrogen Sulfide and Selected Metal Oxides. Environ. Sci. Technol. Large-Pore Apertures in a Series of Metal-Organic Frameworks. Science
1977, 11, 488−491. 2012, 336, 1018−1023.

9796 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(175) Huang, B. L.; Ni, Z.; Millward, A.; McGaughey, A. J. H.; Uher, (194) Allan, P. K.; Wheatley, P. S.; Aldous, D.; Mohideen, M. I.;
C.; Kaviany, M.; Yaghi, O. Thermal Conductivity of a Metal-Organic Tang, C.; Hriljac, J. A.; Megson, I. L.; Chapman, K. W.; De Weireld,
Framework (MOF-5): Part II. Measurement. Int. J. Heat Mass Transfer G.; Vaesen, S.; Morris, R. E. Metal-Organic Frameworks for the
2007, 50, 405−411. Storage and Delivery of Biologically Active Hydrogen Sulfide. Dalton
(176) DeCoste, J. B.; Peterson, G. W. Metal-Organic Frameworks for Trans. 2012, 41, 4060−4066.
Air Purification of Toxic Chemicals. Chem. Rev. 2014, 114, 5695− (195) Chavan, S.; Bonino, F.; Valenzano, L.; Civalleri, B.; Lamberti,
5727. C.; Acerbi, N.; Cavka, J. H.; Leistner, M.; Bordiga, S. Fundamental
(177) Barea, E.; Montoro, C.; Navarro, J. A. R. Toxic Gas Removal - Aspects of H2S Adsorption on CPO-27-Ni. J. Phys. Chem. C 2013, 117,
Metal-Organic Frameworks for the Capture and Degradation of Toxic 15615−15622.
Gases and Vapours. Chem. Soc. Rev. 2014, 43, 5419−5430. (196) Peng, X.; Cao, D. Computational Screening of Porous
(178) Hamon, L.; Serre, C.; Devic, T.; Loiseau, T.; Millange, F.; Carbons, Zeolites, and Metal Organic Frameworks for Desulfurization
Férey, G.; de Weireld, G. Comparative Study of Hydrogen Sulfide and Decarburization of Biogas, Natural Gas, and Flue Gas. AIChE J.
Adsorption in the MIL-53 (Al, Cr, Fe), MIL-47 (V), MIL-100 (Cr), 2013, 59, 2928−2942.
and MIL-101 (Cr) Metal-Organic Frameworks at Room Temperature. (197) Belmabkhout, Y.; et al. Metal-Organic Frameworks to Satisfy
J. Am. Chem. Soc. 2009, 131, 8775−8777. Gas Upgrading Demands: Fine-tuning the soc-MOF platform for the
(179) Hamon, L.; Leclerc, H.; Ghoufi, A.; Oliviero, L.; Travert, A.; Operative Removal of H2S. J. Mater. Chem. A 2017, 5, 3293−3303.
Lavalley, J.-C.; Devic, T.; Serre, C.; Férey, G.; De Weireld, G.; Vimont, (198) Burtch, N. C.; Jasuja, H.; Walton, K. S. Water Stability and
A.; Maurin, G. Molecular Insight into the Adsorption of H2S in the Adsorption in Metal-Organic Frameworks. Chem. Rev. 2014, 114,
Flexible MIL-53 (Cr) and Rigid MIL-47 (V) MOFs: Infrared 10575−10612.
Spectroscopy Combined to Molecular Simulations. J. Phys. Chem. C (199) Pang, S. H.; Han, C.; Sholl, D. S.; Jones, C. W.; Lively, R. P.
2011, 115, 2047−2056. Facet-Specific Stability of ZIF-8 in the Presence of Acid Gases
(180) Han, L.; Budge, M.; Greaney, P. A. Relationship Between Dissolved in Aqueous Solutions. Chem. Mater. 2016, 28, 6960−6967.
Thermal Conductivity and Framework Architecture in MOF-5. (200) Huang, X.-C.; Lin, Y.-Y.; Zhang, J.-P.; Chen, X.-M. Ligand-
Comput. Mater. Sci. 2014, 94, 292−297. Directed Strategy for Zeolite-Type Metal-Organic Frameworks: Zinc
(181) Babaei, H.; McGaughey, A. J. H.; Wilmer, C. E. Effect of Pore (II) Imidazolates with Unusual Zeolitic Topologies. Angew. Chem.
Size and Shape on the Thermal Conductivity of Metal-Organic 2006, 118, 1587−1589.
Frameworks. Chem. Sci. 2017, 8, 583−589. (201) Park, K. S.; Ni, Z.; Côté, A. P.; Choi, J. Y.; Huang, R.; Uribe-
(182) Vaesen, S.; Guillerm, V.; Yang, Q.; Wiersum, A. D.; Marszalek, Romo, F. J.; Chae, H. K.; O’Keeffe, M.; Yaghi, O. M. Exceptional
B.; Gil, B.; Vimont, A.; Daturi, M.; Devic, T.; Llewellyn, P. L.; Serre, Chemical and Thermal Stability of Zeolitic Imidazolate Frameworks.
C.; Maurin, G.; De Weireld, G. A Robust Amino-Functionalized Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 10186−10191.
Titanium (IV) Based MOF for Improved Separation of Acid Gases. (202) Liu, Y.; Ma, Y.; Zhao, Y.; Sun, X.; Gándara, F.; Furukawa, H.;
Chem. Commun. 2013, 49, 10082−10084. Liu, Z.; Zhu, H.; Zhu, C.; Suenaga, K.; Oleynikov, P.; Alshammari, A.
(183) Wynnyk, K. G.; Hojjati, B.; Pirzadeh, P.; Marriott, R. A. High- S.; Zhang, X.; Terasaki, O.; Yaghi, O. M. Weaving of Organic Threads
Pressure Sour Gas Adsorption on Zeolite 4A. Adsorption 2017, 23, into a Crystalline Covalent Organic Framework. Science 2016, 351,
149−162. 365−369.
(184) Yang, Q.; Vaesen, S.; Vishnuvarthan, M.; Ragon, F.; Serre, C.; (203) Ben, T.; Ren, H.; Ma, S.; Cao, D.; Lan, J.; Jing, X.; Wang, W.;
Vimont, A.; Daturi, M.; De Weireld, G.; Maurin, G. Probing the Xu, J.; Deng, F.; Simmons, J. M.; Qiu, S.; Zhu, G. Targeted Synthesis
Adsorption Performance of the Hybrid Porous MIL-68 (Al): A of a Porous Aromatic Framework with High Stability and Exception-
Synergic Combination of Experimental and Modelling Tools. J. Mater. ally High Surface Area. Angew. Chem. 2009, 121, 9621−9624.
Chem. 2012, 22, 10210−10220. (204) Baerlocher, C.; McCusker, L. B. Database of Zeolite Structures,
(185) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P. H.; Orpen, A. G.; http://www.iza-structure.org/databases/; 2017.
Williams, I. D. A Chemically Functionalizable Nanoporous Material (205) Shah, M. S.; Tsapatsis, M.; Siepmann, J. I. Identifying Optimal
[Cu3(TMA)2(H2O)3]n. Science 1999, 283, 1148−1150. Zeolitic Sorbents for Sweetening of Highly Sour Natural Gas. Angew.
(186) Petit, C.; Mendoza, B.; Bandosz, T. J. Hydrogen Sulfide Chem., Int. Ed. 2016, 55, 5938−5942.
Adsorption on MOFs and MOF/Graphite Oxide Composites. (206) Flanigen, E. M.; Bennett, J. M.; Grose, R. W.; Cohen, J. P.;
ChemPhysChem 2010, 11, 3678−3684. Patton, R. L.; Kirchner, R. M.; Smith, J. V. Silicalite, a New
(187) Watanabe, T.; Sholl, D. S. Molecular Chemisorption on Open Hydrophobic Crystalline Silica Molecular Sieve. Nature 1978, 271,
Metal Sites in Cu3(Benzenetricarboxylate)2: A Spatially Periodic 512−516.
Density Functional Theory Study. J. Chem. Phys. 2010, 133, 094509. (207) Maghsoudi, H.; Soltanieh, M.; Bozorgzadeh, H.;
(188) Gutiérrez-Sevillano, J. J.; Martín-Calvo, A.; Dubbeldam, D.; Mohamadalizadeh, A. Adsorption Isotherms and Ideal Selectivities of
Calero, S.; Hamad, S. Adsorption of Hydrogen Sulphide on Metal- Hydrogen Sulfide and Carbon Dioxide over Methane for the Si-CHA
Organic Frameworks. RSC Adv. 2013, 3, 14737−14749. Zeolite: Comparison of Carbon Dioxide and Methane Adsorption with
(189) Supronowicz, B.; Mavrandonakis, A.; Heine, T. Interaction of the All-silica DD3R Zeolite. Adsorption 2013, 19, 1045−1053.
Small Gases with the Unsaturated Metal Centers of the HKUST-1 (208) Shah, M. S.; Tsapatsis, M.; Siepmann, J. I. Monte Carlo
Metal Organic Framework. J. Phys. Chem. C 2013, 117, 14570−14578. Simulations Probing the Adsorptive Separation of Hydrogen Sulfide/
(190) Li, Y.; Wang, L.-J.; Fan, H.-L.; Shangguan, J.; Wang, H.; Mi, J. Methane Mixtures using All-silica Zeolites. Langmuir 2015, 31,
Removal of Sulfur Compounds by a Copper-Based Metal Organic 12268−12278.
Framework Under Ambient Conditions. Energy Fuels 2015, 29, 298− (209) Shah, M. S.; Tsapatsis, M.; Siepmann, J. I. Development of
304. Transferable Potentials for Phase Equilibria Model for Hydrogen
(191) Peterson, G. W.; Britt, D. K.; Sun, D. T.; Mahle, J. J.; Browe, Sulfide. J. Phys. Chem. B 2015, 119, 7041−7052.
M.; Demasky, T.; Smith, S.; Jenkins, A.; Rossin, J. A. Multifunctional (210) Martin, M. G.; Siepmann, J. I. Transferable Potentials for Phase
Purification and Sensing of Toxic Hydride Gases by CuBTC Metal- Equilibria. 1. United-Atom Description of n-Alkanes. J. Phys. Chem. B
Organic Framework. Ind. Eng. Chem. Res. 2015, 54, 3626−3633. 1998, 102, 2569−2577.
(192) Ethiraj, J.; Bonino, F.; Lamberti, C.; Bordiga, S. H2S Interaction (211) Bai, P.; Tsapatsis, M.; Siepmann, J. I. TraPPE-zeo: Transferable
with HKUST-1 and ZIF-8 MOFs: A Multitechnique Study. Micro- Potentials for Phase Equilibria Force Field for All-silica Zeolites. J.
porous Mesoporous Mater. 2015, 207, 90−94. Phys. Chem. C 2013, 117, 24375−24387.
(193) Wang, X.-L.; Fan, H.-L.; Tian, Z.; He, E.-Y.; Li, Y.; Shangguan, (212) Liu, T.; First, E. L.; Hasan, M. M. F.; Floudas, C. A. A Multi-
J. Adsorptive Removal of Sulfur Compounds Using IRMOF-3 at Scale Approach for the Discovery of Zeolites for Hydrogen Sulfide
Ambient Temperature. Appl. Surf. Sci. 2014, 289, 107−113. Removal. Comput. Chem. Eng. 2016, 93, 63.

9797 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(213) Kerr, G. T.; Johnson, G. C. Catalytic Oxidation of Hydrogen (234) Bülow, M.; Lutz, W.; Suckow, M. The Mutual Transformation
Sulfide to Sulfur over a Crystalline Aluminosilicate. J. Phys. Chem. of Hydrogen Sulphide and Carbonyl Sulphide and its Role for Gas
1960, 64, 381−382. Desulphurization Processes with Zeolitic Molecular Sieve Sorbents.
(214) Dudzik, Z.; Ziólek, M. The Specific Catalytic Activity of Stud. Surf. Sci. Catal. 1999, 120, 301−345.
Sodium Faujasites in H2S Oxidation. J. Catal. 1978, 51, 345−354. (235) Kuznicki, S. M. Large-pored Crystalline Titanium Molecular
(215) Ozekmekci, M.; Salkic, G.; Fellah, M. F. Use of Zeolites for the Sieve Zeolites. U.S. Patent US4,853,202, 1989.
Removal of H2S: A Mini-Review. Fuel Process. Technol. 2015, 139, 49− (236) Rezaei, S.; Tavana, A.; Sawada, J. A.; Wu, L.; Junaid, A. S.;
60. Kuznicki, S. M. Novel Copper-Exchanged Titanosilicate Adsorbent for
(216) Deo, A. V.; Dalla Lana, I. G.; Habgood, H. W. Infrared Studies Low Temperature H2S Removal. Ind. Eng. Chem. Res. 2012, 51,
of the Adsorption and Surface Reactions of Hydrogen Sulfide and 12430−12434.
Sulfur Dioxide on Some Aluminas and Zeolites. J. Catal. 1971, 21, (237) Rezaei, S.; Jarligo, M. O. D.; Wu, L.; Kuznicki, S. M.
270−281. Breakthrough Performances of Metal-Exchanged Nanotitanate ETS-2
(217) Förster, H.; Schuldt, M. Infrared Spectroscopic Study of the Adsorbents for Room Temperature Desulfurization. Chem. Eng. Sci.
Adsorption of Hydrogen Sulfide on Zeolites NaA and NaCaA. J. 2015, 123, 444−449.
Colloid Interface Sci. 1975, 52, 380−385. (238) Roller, D.; Bläsing, M.; Dreger, I.; Yazdanbakhsh, F.; Sawada, J.
(218) Karge, H. G.; Raskó, J. Hydrogen Sulfide Adsorption on A.; Kuznicki, S. M.; Müller, M. Removal of Hydrogen Sulfide by
Faujasite-Type Zeolites with Systematically Caried Si-Al Ratios. J. Metal-Doped Nanotitanate under Gasification-Like Conditions. Ind.
Colloid Interface Sci. 1978, 64, 522−532. Eng. Chem. Res. 2016, 55, 3871−3878.
(219) Karge, H. G.; Ziołek, M.; Łaniecki, M. U.V./Vis and I.R. (239) Turk, A.; Sakalis, E.; Lessuck, J.; Karamitsos, H.; Rago, O.
Spectroscopic Study of Hydrogen Sulphide Adsorption on Faujasite- Ammonia Injection Enhances Capacity of Activated Carbon for
Type Zeolites. Zeolites 1987, 7, 197−202. Hydrogen Sulfide and Methyl Mercaptan. Environ. Sci. Technol. 1989,
(220) Howard, J.; Kadir, Z. A. The Adsorption of H2S on Some 23, 1242−1245.
Transition Metal Exchanged Zeolites: An Infrared Study. Spectrochim. (240) Turk, A.; Sakalis, E.; Rago, O.; Karamitsos, H. Activated
Acta, Part A 1985, 41, 825−831. Carbon Systems for Removal of Light Gases. Ann. N. Y. Acad. Sci.
(221) Garcia, C. L.; Lercher, J. A. Adsorption of Hydrogen Sulfide on 1992, 661, 221−228.
ZSM5 Zeolites. J. Phys. Chem. 1992, 96, 2230−2235. (241) Adib, F.; Bagreev, A.; Bandosz, T. J. Effect of Surface
(222) Lechert, H.; Hennig, H. J. The Behavior of the Proton and Characteristics of Wood-Based Activated Carbons on Adsorption of
23
Na-Resonances in Zeolites of the Faujasite Type Depending on the Hydrogen Sulfide. J. Colloid Interface Sci. 1999, 214, 407−415.
Occupancy of Hydrogen Sulfide. Z. Phys. Chem. 1971, 76, 319−327. (242) Bagreev, A.; Adib, F.; Bandosz, T. J. Initial Heats of H2S
(223) Ziolek, M. Adsorption of H2S and SO2 over Faujasite Type Adsorption on Activated Carbons: Effect of Surface Features. J. Colloid
Zeolites. Bull. Polym. Acad. Sci., Chem. 1984, 32, 285−291. Interface Sci. 1999, 219, 327−332.
(243) Bandosz, T. J. Effect of Pore Structure and Surface Chemistry
(224) Sigot, L.; Ducom, G.; Germain, P. Adsorption of Hydrogen
of Virgin Activated Carbons on Removal of Hydrogen Sulfide. Carbon
Sulfide (H2S) on Zeolite (Z): Retention Mechanism. Chem. Eng. J.
1999, 37, 483−491.
2016, 287, 47−53.
(244) Adib, F.; Bagreev, A.; Bandosz, T. J. Adsorption/Oxidation of
(225) Fellah, M. F. Adsorption of Hydrogen Sulfide as Initial Step of
Hydrogen Sulfide on Nitrogen-Containing Activated Carbons.
H2S Removal: A DFT Study on Metal Exchanged ZSM-12 Clusters.
Langmuir 2000, 16, 1980−1986.
Fuel Process. Technol. 2016, 144, 191−196.
(245) Adib, F.; Bagreev, A.; Bandosz, T. J. Analysis of the
(226) Hernández-Maldonado, A. J.; Yang, R. T.; Chinn, D.; Munson,
Relationship Between H2S Removal Capacity and Surface Properties
C. L. Partially Calcined Gismondine Type Silicoaluminophosphate
of Unimpregnated Activated Carbons. Environ. Sci. Technol. 2000, 34,
SAPO-43: Isopropylamine Elimination and Separation of Carbon 686−692.
Dioxide, Hydrogen Sulfide, and Water. Langmuir 2003, 19, 2193− (246) Bandosz, T. J. On the Adsorption/Oxidation of Hydrogen
2200. Sulfide on Activated Carbons at Ambient Temperatures. J. Colloid
(227) Crespo, D.; Qi, G.; Wang, Y.; Yang, F. H.; Yang, R. T. Superior Interface Sci. 2002, 246, 1−20.
Sorbent for Natural Gas Desulfurization. Ind. Eng. Chem. Res. 2008, 47, (247) Tian, S.; Mo, H.; Zhang, R.; Ning, P.; Zhou, T. Enhanced
1238−1244. Removal of Hydrogen Sulfide from a Gas Stream by 3-Amino-
(228) Hutson, N. D.; Reisner, B. A.; Yang, R. T.; Toby, B. H. Silver propyltriethoxysilane-Surface-Functionalized Activated Carbon. Ad-
Ion-Exchanged Zeolites Y, X, and Low-Silica X: Observations of sorption 2009, 15, 477−488.
Thermally Induced Cation/Cluster Migration and the Resulting (248) Bagreev, A.; Rahman, H.; Bandosz, T. J. Study of H2S
Effects on the Equilibrium Adsorption of Nitrogen. Chem. Mater. Adsorption and Water Regeneration of Spent Coconut-Based
2000, 12, 3020−3031. Activated Carbon. Environ. Sci. Technol. 2000, 34, 4587−4592.
(229) Kumar, P.; Sung, C.; Muraza, O.; Cococcioni, M.; Al Hashimi, (249) Bagreev, A.; Rahman, H.; Bandosz, T. J. Thermal Regeneration
S.; McCormick, A.; Tsapatsis, M. H2S Adsorption by Ag and Cu Ion of a Spent Activated Carbon Previously Used as Hydrogen Sulfide
Exchanged Faujasites. Microporous Mesoporous Mater. 2011, 146, 127− Adsorbent. Carbon 2001, 39, 1319−1326.
133. (250) Boudou, J.-P.; Chehimi, M.; Broniek, E.; Siemieniewska, T.;
(230) Sung, C.; Al Hashimi, S.; McCormick, A.; Tsapatsis, M.; Bimer, J. Adsorption of H2S or SO2 on an Activated Carbon Cloth
Cococcioni, M. Density Functional Theory Study on the Adsorption Modified by Ammonia Treatment. Carbon 2003, 41, 1999−2007.
of H2S and Other Claus Process Tail Gas Components on Copper-and (251) Bagreev, A.; Rahman, H.; Bandosz, T. J. Wood-Based Activated
Silver-Exchanged Y Zeolites. J. Phys. Chem. C 2012, 116, 3561−3575. Carbons as Adsorbents of Hydrogen Sulfide: A Study of Adsorption
(231) Sung, C.-Y.; Al Hashimi, S.; McCormick, A.; Cococcioni, M.; and Water Regeneration Processes. Ind. Eng. Chem. Res. 2000, 39,
Tsapatsis, M. A DFT Study on Multivalent Cation-Exchanged Y 3849−3855.
Zeolites as Potential Selective Adsorbent for H2S. Microporous (252) Goh, K.; Karahan, H. E.; Wei, L.; Bae, T.-H.; Fane, A. G.;
Mesoporous Mater. 2013, 172, 7−12. Wang, R.; Chen, Y. Carbon Nanomaterials for Advancing Separation
(232) Fellmuth, P.; Lutz, W.; Bülow, M. Influence of Weakly Membranes: A Strategic Perspective. Carbon 2016, 109, 694−710.
Coordinated Cations and Basic Sites Upon the Reaction of H2S and (253) Guan, L.; Suenaga, K.; Iijima, S. Smallest Carbon Nanotube
CO2 on Zeolites. Zeolites 1987, 7, 367−371. Assigned with Atomic Resolution Accuracy. Nano Lett. 2008, 8, 459−
(233) Lutz, W.; Suckow, M.; Bülow, M. Adsorption of Hydrogen 462.
Sulphide on Molecular Sieves: No Enrichment in the Presence of (254) Rinzler, A. G.; Liu, J.; Dai, H.; Nikolaev, P.; Huffman, C. B.;
Carbon Dioxide. Gas Sep. Purif. 1990, 4, 190−196. Rodriguez-Macias, F. J.; Boul, P. J.; Lu, A. H.; Heymann, D.; Colbert,

9798 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

D. T.; Lee, R. S.; Fischer, J. E.; Rao, A. M.; Eklund, P. C.; Smalley, E. Their Conversion to Microporous Materials. Bull. Chem. Soc. Jpn.
R. Large-Scale Purification of Single-Wall Carbon Nanotubes: Process, 1990, 63, 988−992.
Product, and Characterization. Appl. Phys. A: Mater. Sci. Process. 1998, (274) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.;
67, 29−37. Beck, J. S. Ordered Mesoporous Molecular Sieves Synthesized by a
(255) Williams, K. A.; Eklund, P. C. Monte Carlo Simulations of H2 Liquid-Crystal Template Mechanism. Nature 1992, 359, 710−712.
Physisorption in Finite-Diameter Carbon Nanotube Ropes. Chem. (275) Wang, X.; Sun, T.; Yang, J.; Zhao, L.; Jia, J. Low-Temperature
Phys. Lett. 2000, 320, 352−358. H2S Removal from Gas Streams with SBA-15 Supported ZnO
(256) Zhang, Y.-H.; Han, L.-F.; Xiao, Y.-H.; Jia, D.-Z.; Guo, Z.-H.; Li, Nanoparticles. Chem. Eng. J. 2008, 142, 48−55.
F. Understanding Dopant and Defect Effect on H2S Sensing (276) Wang, X.; Jia, J.; Zhao, L.; Sun, T. Mesoporous SBA-15
Performances of Graphene: A First-Principles Study. Comput. Mater. Supported Iron Oxide: A Potent Catalyst for Hydrogen Sulfide
Sci. 2013, 69, 222−228. Removal. Water, Air, Soil Pollut. 2008, 193, 247−257.
(257) Castellanos Á guila, J. E.; Hernández Cocoletzi, H.; Hernández (277) Dhage, P.; Samokhvalov, A.; Repala, D.; Duin, E. C.; Bowman,
Cocoletzi, G. A Theoretical Analysis of the Role of Defects in the M.; Tatarchuk, B. J. Copper-Promoted ZnO/SiO2 Regenerable
Adsorption of Hydrogen Sulfide on Graphene. AIP Adv. 2013, 3, Sorbents for the Room Temperature Removal of H2S from Reformate
032118. Gas Streams. Ind. Eng. Chem. Res. 2010, 49, 8388−8396.
(258) Borisova, D.; Antonov, V.; Proykova, A. Hydrogen Sulfide (278) Dhage, P.; Samokhvalov, A.; McKee, M. L.; Duin, E. C.;
Adsorption on a Defective Graphene. Int. J. Quantum Chem. 2013, Tatarchuk, B. J. Reactive Adsorption of Hydrogen Sulfide by
113, 786−791. Promoted Sorbents Cu-ZnO/SiO2: Active Sites by Experiment and
(259) Hegde, V. I.; Shirodkar, S. N.; Tit, N.; Waghmare, U. V.; Simulation. Surf. Interface Anal. 2013, 45, 865−872.
Yamani, Z. H. First Principles Analysis of Graphene and Its Ability to (279) Yang, H.; Tatarchuk, B. Novel-Doped Zinc Oxide Sorbents for
Maintain Long-Ranged Interaction with H2S. Surf. Sci. 2014, 621, Low Temperature Regenerable Desulfurization Applications. AIChE J.
168−174. 2010, 56, 2898−2904.
(260) Ganji, M. D.; Sharifi, N.; Ardjmand, M.; Ahangari, M. G. Pt- (280) Dhage, P.; Samokhvalov, A.; Repala, D.; Duin, E. C.;
Decorated Graphene as Superior Media for H2S Adsorption: A First- Tatarchuk, B. J. Regenerable Fe-Mn-ZnO/SiO2 Sorbents for Room
Principles Study. Appl. Surf. Sci. 2012, 261, 697−704. Temperature Removal of H2S from Fuel Reformates: Performance,
(261) Sharma, S.; Verma, A. S. A Theoretical Study of H2S Active Sites, Operando Studies. Phys. Chem. Chem. Phys. 2011, 13,
Adsorption on Graphene Doped with B, Al and Ga. Phys. B 2013, 427, 2179−2187.
12−16. (281) Hussain, M.; Abbas, N.; Fino, D.; Russo, N. Novel
(262) Ganji, M. D.; Sharifi, N.; Ahangari, M. G. Adsorption of H2S Mesoporous Silica Supported ZnO Adsorbents for the Desulphuriza-
Molecules on Non-Carbonic and Decorated Carbonic Graphenes: A tion of Biogas at Low Temperatures. Chem. Eng. J. 2012, 188, 222−
van der Waals Density Functional Study. Comput. Mater. Sci. 2014, 92, 232.
127−134. (282) Elyassi, B.; Al Wahedi, Y.; Rajabbeigi, N.; Kumar, P.; Jeong, J.
(263) Zhang, H.-p.; Luo, X.-g.; Song, H.-t.; Lin, X.-y.; Lu, X.; Tang, Y. S.; Zhang, X.; Kumar, P.; Balasubramanian, V. V.; Katsiotis, M. S.;
DFT Study of Adsorption and Dissociation Behavior of H2S on Fe- Mkhoyan, K. A.; Boukos, N.; Hashimi, S. A.; Tsapatsis, M. A High-
Doped Graphene. Appl. Surf. Sci. 2014, 317, 511−516. Performance Adsorbent for Hydrogen Sulfide Removal. Microporous
(264) Mohammadi-Manesh, E.; Vaezzadeh, M.; Saeidi, M. Cu-and Mesoporous Mater. 2014, 190, 152−155.
CuO-Decorated Graphene As a Nanosensor for H2S Detection at (283) Montes, D.; Tocuyo, E.; González, E.; Rodríguez, D.; Solano,
Room Temperature. Surf. Sci. 2015, 636, 36−41. R.; Atencio, R.; Ramos, M. A.; Moronta, A. Reactive H2S
(265) Pang, J.; Ma, X.; Yang, Q.; Meng, R.; Tan, C.; Sun, X.; Chen, Chemisorption on Mesoporous Silica Molecular Sieve-Supported
X. Enhancement of H2S Detection in Impurity-Doped Graphene. Int. CuO or ZnO. Microporous Mesoporous Mater. 2013, 168, 111−120.
Conf. Electron. Packag. Technol. 2016, 803−806. (284) Liu, G.; Huang, Z.-H.; Kang, F. Preparation of ZnO/SiO2 Gel
(266) Faye, O.; Raj, A.; Mittal, V.; Beye, A. C. H2S Adsorption on Composites and Their Performance of H2S Removal at Room
Graphene in the Presence of Sulfur: A Density Functional Theory Temperature. J. Hazard. Mater. 2012, 215-216, 166−172.
Study. Comput. Mater. Sci. 2016, 117, 110−119. (285) Mureddu, M.; Ferino, I.; Rombi, E.; Cutrufello, M. G.; Deiana,
(267) Wang, W.; Peng, X.; Cao, D. Capture of Trace Sulfur Gases P.; Ardu, A.; Musinu, A.; Piccaluga, G.; Cannas, C. ZnO/SBA-15
from Binary Mixtures by Single-Walled Carbon Nanotube Arrays: A Composites for Mid-Temperature Removal of H2S: Synthesis,
Molecular Simulation Study. Environ. Sci. Technol. 2011, 45, 4832− Performance and Regeneration Studies. Fuel 2012, 102, 691−700.
4838. (286) Mureddu, M.; Ferino, I.; Musinu, A.; Ardu, A.; Rombi, E.;
(268) Chen, J.-J.; Li, W.-W.; Li, X.-L.; Yu, H.-Q. Improving Biogas Cutrufello, M. G.; Deiana, P.; Fantauzzi, M.; Cannas, C. MeOx/SBA-
Separation and Methane Storage with Multilayer Graphene Nano- 15 (Me = Zn, Fe): Highly Efficient Nanosorbents for Mid-
structure via Layer Spacing Optimization and Lithium Doping: A Temperature H2S Removal. J. Mater. Chem. A 2014, 2, 19396−19406.
Molecular Simulation Investigation. Environ. Sci. Technol. 2012, 46, (287) Li, L.; Sun, T. H.; Shu, C. H.; Zhang, H. B. Low Temperature
10341−10348. H2S Removal with 3-D Structural Mesoporous Molecular Sieves
(269) Gholampour, F.; Yeganegi, S. Molecular Simulation Study on Supported ZnO from Gas Stream. J. Hazard. Mater. 2016, 311, 142−
the Adsorption and Separation of Acidic Gases in a Model 150.
Nanoporous Carbon. Chem. Eng. Sci. 2014, 117, 426−435. (288) Prieto, G.; Zečević, J.; Friedrich, H.; de Jong, K. P.; de Jongh,
(270) Lei, G.; Liu, C.; Li, Q.; Xu, X. Graphyne Nanostructure as a P. E. Towards Stable Catalysts By Controlling Collective Properties of
Potential Adsorbent for Separation of H2S/CH4 Mixture: Combining Supported Metal Nanoparticles. Nat. Mater. 2013, 12, 34−39.
Grand Canonical Monte Carlo Simulations with Ideal Adsorbed (289) Li, W.; Choi, S.; Drese, J. H.; Hornbostel, M.; Krishnan, G.;
Solution Theory. Fuel 2016, 182, 210−219. Eisenberger, P. M.; Jones, C. W. Steam-Stripping for Regeneration of
(271) Huang, L.; Seredych, M.; Bandosz, T. J.; van Duin, A. C. T.; Supported Amine-Based CO2 Adsorbents. ChemSusChem 2010, 3,
Lu, X.; Gubbins, K. E. Controllable Atomistic Graphene Oxide Model 899−903.
and Its Application in Hydrogen Sulfide Removal. J. Chem. Phys. 2013, (290) Hicks, J. C.; Drese, J. H.; Fauth, D. J.; Gray, M. L.; Qi, g.;
139, 194707. Jones, C. W. Designing Adsorbents for CO2 Capture from Flue Gas-
(272) Mohamadalizadeh, A.; Towfighi, J.; Rashidi, A.; Mohajeri, A.; Hyperbranched Aminosilicas Capable of Capturing CO2 Reversibly. J.
Golkar, M. Modification of Carbon Nanotubes for H2S Sorption. Ind. Am. Chem. Soc. 2008, 130, 2902−2903.
Eng. Chem. Res. 2011, 50, 8050−8057. (291) Jones, C. W. C. O. CO2 Capture from Dilute Gases as a
(273) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. The Component of Modern Global Carbon Management. Annu. Rev.
Preparation of Alkyltriinethylaininonium-Kaneinite Complexes and Chem. Biomol. Eng. 2011, 2, 31−52.

9799 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(292) Huang, H. Y.; Yang, R. T.; Chinn, D.; Munson, C. L. Amine- (313) Seredych, M.; Mabayoje, O.; Bandosz, T. J. Visible-Light-
grafted MCM-48 and Silica Xerogel as Superior Sorbents for Acidic Enhanced Interactions of Hydrogen Sulfide with Composites of Zinc
Gas Removal from Natural Gas. Ind. Eng. Chem. Res. 2003, 42, 2427− (Oxy)hydroxide with Graphite Oxide and Graphene. Langmuir 2012,
2433. 28, 1337−1346.
(293) Zhou, L.; Zhong, L.; Yu, M.; Zhou, Y. Sorption and Desorption (314) Mabayoje, O.; Seredych, M.; Bandosz, T. J. Cobalt (Hydr)-
of a Minor Amount of H2S on Silica Gel Covered with a Film of oxide/Graphite Oxide Composites: Importance of Surface Chemical
Triethanolamine. Ind. Eng. Chem. Res. 2004, 43, 1765−1767. Heterogeneity for Reactive Adsorption of Hydrogen Sulfide. J. Colloid
(294) Zhou, L.; Zhong, L.; Su, W.; Sun, Y.; Zhou, Y. Experimental Interface Sci. 2012, 378, 1−9.
Study of Removing Trace H2S Using Solvent Coated Adsorbent for (315) Mabayoje, O.; Seredych, M.; Bandosz, T. J. Enhanced Reactive
PSA. AIChE J. 2006, 52, 2066−2071. Adsorption of Hydrogen Sulfide on the Composites of Graphene/
(295) Chu, X.; Cheng, Z.; Zhao, Y.; Xu, J.; Zhong, H.; Zhang, W.; L, Graphite Oxide with Copper (Hydr)oxychlorides. ACS Appl. Mater.
J.; Zhou, S.; Zhu, F.; Zhou, v; Zhou, L. Study on Sorption Behaviors of Interfaces 2012, 4, 3316−3324.
H2S by Triethanolamine-Modified Mesoporous Molecular Sieve SBA- (316) Mabayoje, O.; Seredych, M.; Bandosz, T. J. Enhanced
15. Ind. Eng. Chem. Res. 2012, 51, 4407−4413. Adsorption of Hydrogen Sulfide on Mixed Zinc/Cobalt Hydroxides:
(296) Xue, Q.; Liu, Y. Removal of Minor Concentration of H2S on Effect of Morphology and an Increased Number of Surface Hydroxyl
MDEA-Modified SBA-15 for Gas Purification. J. Ind. Eng. Chem. 2012, Groups. J. Colloid Interface Sci. 2013, 405, 218−225.
18, 169−173. (317) Mabayoje, O.; Seredych, M.; Bandosz, T. J. Reactive
(297) Xu, X.; Novochinskii, I.; Song, C. Low-Temperature Removal Adsorption of Hydrogen Sulfide on Visible Light Photoactive Zinc
of H2S by Nanoporous Composite of Polymer-Mesoporous Molecular (Hydr)oxide/Graphite Oxide and Zinc (Hydr)oxychloride/Graphite
Sieve MCM-41 as Adsorbent for Fuel Cell Applications. Energy Fuels Oxide Composites. Appl. Catal., B 2013, 132-133, 321−331.
2005, 19, 2214−2215. (318) Arcibar-Orozco, J. A.; Wallace, R.; Mitchell, J. K.; Bandosz, T.
(298) Wang, X.; Ma, X.; Sun, L.; Song, C. A Nanoporous Polymeric J. Role of Surface Chemistry and Morphology in the Reactive
Sorbent for Deep Removal of H2S from Gas Mixtures for Hydrogen Adsorption of H2S on Iron (Hydr)oxide/Graphite Oxide Composites.
Purification. Green Chem. 2007, 9, 695−702. Langmuir 2015, 31, 2730−2742.
(299) Wang, X.; Ma, X.; Xu, X.; Sun, L.; Song, C. Mesoporous- (319) Giannakoudakis, D. A.; Bandosz, T. J. Zinc (Hydr)oxide/
Molecular-Sieve-Supported Polymer Sorbents for Removing H2S from Graphite Oxide/AuNPs Composites: Role of Surface Features in H2S
Hydrogen Gas Streams. Top. Catal. 2008, 49, 108−117. Reactive Adsorption. J. Colloid Interface Sci. 2014, 436, 296−305.
(300) Ma, X.; Wang, X.; Song, C. Molecular Basket Sorbents for (320) Florent, M.; Bandosz, T. J. Effects of Surface Heterogeneity of
Separation of CO2 and H2S from Various Gas Streams. J. Am. Chem. Cobalt Oxyhydroxide/Graphite Oxide Composites on Reactive
Soc. 2009, 131, 5777−5783. Adsorption of Hydrogen Sulfide. Microporous Mesoporous Mater.
(301) Yoosuk, B.; Wongsanga, T.; Prasassarakich, P. CO2 and H2S
2015, 204, 8−14.
Binary Sorption on Polyamine Modified Fumed Silica. Fuel 2016, 168,
(321) Florent, M.; Wallace, R.; Bandosz, T. J. Removal of Hydrogen
47−53.
Sulfide at Ambient Conditions on Cadmium/GO-Based Composite
(302) First, E. L.; Hasan, M. M. F.; Floudas, C. A. Discovery of Novel
Adsorbents. J. Colloid Interface Sci. 2015, 448, 573−581.
Zeolites for Natural Gas Purification through Combined Material
(322) Giannakoudakis, D. A.; Jiang, M.; Bandosz, T. J. Highly
Screening and Process Optimization. AIChE J. 2014, 60, 1767−1785.
Efficient Air Desulfurization on Self-Assembled Bundles of Copper
(303) Chen, Q.; Fan, F.; Long, D.; Liu, X.; Liang, X.; Qiao, W.; Ling,
Hydroxide Nanorods. ACS Appl. Mater. Interfaces 2016, 8, 31986−
L. Poly(ethyleneimine)-Loaded Silica Monolith with a Hierarchical
Pore Structure for H2S Adsorptive Removal. Ind. Eng. Chem. Res. 2010, 31994.
(323) Song, H. S.; Park, M. G.; Ahn, W.; Lim, S. N.; Yi, K. B.;
49, 11408−11414.
(304) Belmabkhout, Y.; de Weireld, G.; Sayari, A. Amine-Bearing Croiset, E.; Chen, Z.; Nam, S. C. Enhanced Adsorption of Hydrogen
Mesoporous Silica for CO2 and H2S Removal from Natural Gas and Sulfide and Regeneration Ability on the Composites of Zinc Oxide
Biogas. Langmuir 2009, 25, 13275−13278. with Reduced Graphite Oxide. Chem. Eng. J. 2014, 253, 264−273.
(305) Belmabkhout, Y.; Heymans, N.; de Weireld, G.; Sayari, A. (324) Lonkar, S. P.; Pillai, V. V.; Stephen, S.; Abdala, A.; Mittal, V.
Simultaneous Adsorption of H2S and CO2 on Triamine-Grafted Pore- Facile In Situ Fabrication of Nanostructured Graphene-CuO Hybrid
Expanded Mesoporous MCM-41 Silica. Energy Fuels 2011, 25, 1310− with Hydrogen Sulfide Removal Capacity. Nano-Micro Lett. 2016, 8,
1315. 312−319.
(306) Serna-Guerrero, R.; Belmabkhout, Y.; Sayari, A. Modeling CO2 (325) Fauteux-Lefebvre, C.; Abatzoglou, N.; Braidy, N.; Hu, Y.
Adsorption on Amine-Functionalized Mesoporous Silica: 1. A Semi- Carbon Nanofilaments Functionalized with Iron Oxide Nanoparticles
Empirical Equilibrium Model. Chem. Eng. J. 2010, 161, 173−181. for In-Depth Hydrogen Sulfide Adsorption. Ind. Eng. Chem. Res. 2015,
(307) Anbia, M.; Babaei, M. Novel Amine Modified Nanoporous 54, 9230−9237.
SBA-15 Sorbent for the Removal of H2S from Gas Streams in the (326) Zhang, J.; Wang, L.; Song, H.; Song, H. Adsorption of Low-
Presence of CH4. Int. J. Eng., Trans. B 2014, 27, 1697−1704. Concentration H2S on Manganese Dioxide-Loaded Activated Carbon.
(308) Bhandari, D. A.; Bessho, N.; Koros, W. J. Hollow Fiber Res. Chem. Intermed. 2015, 41, 6087−6104.
Sorbents for Desulfurization of Natural Gas. Ind. Eng. Chem. Res. 2010, (327) Balsamo, M.; Cimino, S.; de Falco, G.; Erto, A.; Lisi, L. ZnO-
49, 12038−12050. CuO Supported on Activated Carbon for H2S Removal at Room
(309) Ebrahim, A. M.; Jagiello, J.; Bandosz, T. J. Enhanced Reactive Temperature. Chem. Eng. J. 2016, 304, 399−407.
Adsorption of H2S on Cu-BTC/S-and N-Doped GO Composites. J. (328) UOP LLC. Handling of Zeolite Molecular Sieve Adsorbents in
Mater. Chem. A 2015, 3, 8194−8204. Process Units, UOP5085, April 2009; https://www.uop.com/
(310) Huang, Z.-H.; Liu, G.; Kang, F. Glucose-Promoted Zn-Based ?document=uop-molsiv-handling-procedures-brochure&download=1.
Metal-Organic Framework/Graphene Oxide Composites for Hydro- (329) Pullumbi, P. Gas Separation by Polymer Membranes: Research
gen Sulfide Removal. ACS Appl. Mater. Interfaces 2012, 4, 4942−4947. Activity and Industrial Applications. In Membrane Reactor Engineering:
(311) Lonkar, S. P.; Pillai, V.; Abdala, A.; Mittal, V. In Situ Formed Applications for a Greener Process Industry; Basile, A., De Falco, M.,
Graphene/ZnO Nanostructured Composites for Low Temperature Centi, G., Iaquaniello, G., Eds.; John Wiley & Sons: New York, 2016;
Hydrogen Sulfide Removal from Natural Gas. RSC Adv. 2016, 6, p 259.
81142−81150. (330) Bhide, B. D.; Stern, S. A. Membrane Processes for the Removal
(312) Seredych, M.; Bandosz, T. J. Reactive Adsorption of Hydrogen of Acid Gases from Natural Gas. II. Effects of Operating Conditions,
Sulfide on Graphite Oxide/Zr(OH)4 Composites. Chem. Eng. J. 2011, Economic Parameters, and Membrane Properties. J. Membr. Sci. 1993,
166, 1032−1038. 81, 239−252.

9800 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(331) Baker, R. W.; Lokhandwala, K. Natural Gas Processing with (355) Chenar, M. P.; Savoji, H.; Soltanieh, M.; Matsuura, T.; Tabe, S.
Membranes: An Overview. Ind. Eng. Chem. Res. 2008, 47, 2109−2121. Removal of Hydrogen Sulfide from Methane Using Commercial
(332) Schell, W. J.; Wensley, C. G.; Chen, M. S. K.; Venugopal, K. Polyphenylene Oxide and Cardo-Type Polyimide Hollow Fiber
G.; Miller, B. D.; Stuart, J. A. Recent Advances in Cellulosic Membranes. Korean J. Chem. Eng. 2011, 28, 902−913.
Membranes for Gas Separation and Pervaporation. Gas Sep. Purif. (356) Vaughn, J.; Koros, W. J. Effect of the Amide Bond Diamine
1989, 3, 162−169. Structure on the CO2, H2S, and CH4 Transport Properties of a Series
(333) Chatterjee, G.; Houde, A. A.; Stern, S. A. Poly(Ether of Novel 6FDA-Based Polyamide-Imides for Natural Gas Purification.
Urethane) and Poly(Ether Urethane Urea) Membranes with High Macromolecules 2012, 45, 7036−7049.
H2S/CH4 Selectivity. J. Membr. Sci. 1997, 135, 99−106. (357) Kraftschik, B.; Koros, W. J.; Johnson, J. R.; Karvan, O. Dense
(334) Achoundong, C. S. K.; Bhuwania, N.; Burgess, S. K.; Karvan, Film Polyimide Membranes for Aggressive Sour Gas Feed Separations.
O.; Johnson, J. R.; Koros, W. J. Silane Modification of Cellulose J. Membr. Sci. 2013, 428, 608−619.
Acetate Dense Films as Materials for Acid Gas Removal. Macro- (358) Kraftschik, B.; Koros, W. J. Cross-Linkable Polyimide
molecules 2013, 46, 5584−5594. Membranes for Improved Plasticization Resistance and Permselectivity
(335) Brun, J. P.; Larchet, C.; Melet, R.; Bulvestre, G. Modelling of in Sour Gas Separations. Macromolecules 2013, 46, 6908−6921.
the Pervaporation of Binary Mixtures Through Moderately Swelling, (359) Vaughn, J. T.; Koros, W. J. Analysis of Feed Stream Acid Gas
Non-Reacting Membranes. J. Membr. Sci. 1985, 23, 257−283. Concentration Effects on the Transport Properties and Separation
(336) Story, B. J.; Koros, W. J. Comparison of Three Models for
Performance of Polymeric Membranes for Natural Gas Sweetening: A
Permeation of CO2/CH4 Mixtures in Poly(phenylene oxide). J. Polym.
Comparison Between a Glassy and Rubbery Polymer. J. Membr. Sci.
Sci., Part B: Polym. Phys. 1989, 27, 1927−1948.
(337) Donohue, M. D.; Minhas, B. S.; Lee, S. Y. Permeation Behavior 2014, 465, 107−116.
of Carbon Dioxide-Methane Mixtures in Cellulose Acetate Mem- (360) Yi, S.; Ma, X.; Pinnau, I.; Koros, W. J. A High-Performance
branes. J. Membr. Sci. 1989, 42, 197−214. Hydroxyl-Functionalized Polymer of Intrinsic Microporosity for an
(338) Huang, Y.; Paul, D. R. Physical Aging of Thin Glassy Films Environmentally Attractive Membrane-Based Approach to Decon-
Monitored by Gas Permeability. Polymer 2004, 45, 8377−8393. tamination of Sour Natural Gas. J. Mater. Chem. A 2015, 3, 22794−
(339) Heilman, W.; Tammela, V.; Meyer, J. A.; Stannett, V.; Szwarc, 22806.
M. Permeability of Polymer Films to Hydrogen Sulfide Gas. Ind. Eng. (361) Yahaya, G. O.; Qahtani, M. S.; Ammar, A. Y.; Bahamdan, A. A.;
Chem. 1956, 48, 821−824. Ameen, A. W.; Alhajry, R. H.; Sultan, M. M. B.; Hamad, F. Aromatic
(340) Kulprathipanja, S. Separation of Polar Gases from Nonpolar Block Co-Polyimide Membranes for Sour Gas Feed Separations.
Gases. U.S. Patent US4,606,740, 1986. Chem. Eng. J. 2016, 304, 1020−1030.
(341) Kulprathipanja, S.; Kulkarni, S. S. Separation of polar gases (362) Saedi, S.; Madaeni, S. S.; Shamsabadi, A. A. PDMS Coated
from nonpolar gases. U.S. Patent US4,608,060, 1986. Asymmetric PES Membrane for Natural Gas Sweetening: Effect of
(342) Stern, S. A.; Bhide, B. D. Permeability of Silicone Polymers to Preparation and Operating Parameters on Performance. Can. J. Chem.
Ammonia and Hydrogen Sulfide. J. Appl. Polym. Sci. 1989, 38, 2131− Eng. 2014, 92, 892−904.
2147. (363) Li, N. N.; Funk, E. W.; Chang, Y. A.; Kulkarni, S. S.;
(343) Blume, I.; Pinnau, I. Composite Membrane, Method of Swamikannu, A. X.; White, L. S. Membrane Separation Processes in the
Preparation and Use. U.S. Patent US4,963,165, 1990. Petrochemical Industry: Phase II, April 1, 1984−October 30, 1985, DOE/
(344) Lokhandwala, K. A.; Baker, R. W.; Amo, K. D. Sour Gas ID/12422-2; U.S. Department of Energy, Sept 30, 1987.
Treatment Process. U.S. Patent US5,407,467, 1995. (364) Lokhandwala, K. A.; Baker, R. W. Sour Gas Treatment Process
(345) Quinn, R.; Laciak, D. V. Polyelectrolyte Membranes for Acid Including Membrane and Non-Membrane Treatment Steps. U.S.
Gas Separations. J. Membr. Sci. 1997, 131, 49−60. Patent US5,407,466, 1995.
(346) Quinn, R.; Laciak, D. V.; Pez, G. P. Polyelectrolyte-Salt Blend (365) Baker, R. W.; Low, B. T. Gas Separation Membrane Materials:
Membranes for Acid Gas Separations. J. Membr. Sci. 1997, 131, 61−69. A Perspective. Macromolecules 2014, 47, 6999−7013.
(347) Amirkhanov, D. M.; Kotenko, A. A.; Tul’skii, M. N.; Chelyak, (366) Budd, P. M.; Elabas, E. S.; Ghanem, B. S.; Makhseed, S.;
M. M. Hollow Fibres for Removal and Concentration of Hydrogen McKeown, N. B.; Msayib, K. J.; Tattershall, C. E.; Wang, D. Solution-
Sulfide from Gas Mixtures by the Membrane Method. Fibre Chem. Processed, Organophilic Membrane Derived from a Polymer of
2001, 33, 67−72. Intrinsic Microporosity. Adv. Mater. 2004, 16, 456−459.
(348) Wilks, B.; Rezac, M. E. Properties of Rubbery Polymers for the (367) McKeown, N. B.; Budd, P. M. Polymers of Intrinsic
Recovery of Hydrogen Sulfide from Gasification Gases. J. Appl. Polym. Microporosity (PIMs): Organic Materials for Membrane Separations,
Sci. 2002, 85, 2436−2444. Heterogeneous Catalysis and Hydrogen Storage. Chem. Soc. Rev. 2006,
(349) Orme, C. J.; Klaehn, J. R.; Stewart, F. F. Gas Permeability and
35, 675−683.
Ideal Selectivity of Poly[Bis-(Phenoxy)Phosphazene], Poly[Bis-(4-
(368) Swaidan, R.; Ghanem, B. S.; Litwiller, E.; Pinnau, I. Pure- and
Tert-Butylphenoxy)Phosphazene], and Poly[Bis-(3, 5-di-Tert-Butyl-
Mixed-Gas CO2/CH4 Separation Properties of PIM-1 and an
phenoxy)1.2(Chloro)0.8Pphosphazene]. J. Membr. Sci. 2004, 238, 47−
55. Amidoxime-Functionalized PIM-1. J. Membr. Sci. 2014, 457, 95−102.
(369) Ma, X.; Swaidan, R.; Belmabkhout, Y.; Zhu, Y.; Litwiller, E.;
(350) Orme, C. J.; Stewart, F. F. Mixed Gas Hydrogen Sulfide
Permeability and Separation Using Supported Polyphosphazene Jouiad, M.; Pinnau, I.; Han, Y. Synthesis and Gas Transport Properties
Membranes. J. Membr. Sci. 2005, 253, 243−249. of Hydroxyl-Functionalized Polyimides with Intrinsic Microporosity.
(351) Merkel, T. C.; Toy, L. G. Comparison of Hydrogen Sulfide Macromolecules 2012, 45, 3841−3849.
Transport Properties in Fluorinated and Nonfluorinated Polymers. (370) Heyd, R. L.; McCandless, F. P. Separation of H2S from N2 by
Macromolecules 2006, 39, 7591−7600. Selective Permeation through Polymeric Membranes. J. Membr. Sci.
(352) Sridhar, S.; Smitha, B.; Mayor, S.; Prathab, B.; Aminabhavi, T. 1977, 2, 375−389.
M. Gas Permeation Properties of Polyamide Membrane Prepared by (371) Hao, J.; Rice, P. A.; Stern, S. A. Upgrading Low-Quality
Interfacial Polymerization. J. Mater. Sci. 2007, 42, 9392−9401. Natural Gas with H2S- and CO2-Selective Polymer Membranes: Part I.
(353) Mohammadi, T.; Moghadam, M. T.; Saeidi, M.; Mahdyarfar, Process Design and Economics of Membrane Stages without Recycle
M. Acid Gas Permeation Behavior through Poly(Ester Urethane Urea) Streams. J. Membr. Sci. 2002, 209, 177−206.
Membrane. Ind. Eng. Chem. Res. 2008, 47, 7361−7367. (372) Hao, J.; Rice, P. A.; Stern, S. A. Upgrading Low-Quality
(354) Saeidi, M.; Moghadam, M. T.; Mahdyarfar, M.; Mohammadi, Natural Gas with H2S- and CO2-Selective Polymer Membranes: Part
T. Gas Permeation Properties of Seragel Membrane. Asia-Pac. J. Chem. II. Process Design, Economics, and Sensitivity Study of Membrane
Eng. 2010, 5, 324−329. Stages with Recycle Streams. J. Membr. Sci. 2008, 320, 108−122.

9801 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(373) Ren, X.; Ren, J.; Deng, M. Poly(Amide-6-b-Ethylene Oxide) and H2S by MEA Solution in Hollow Fiber Membrane Reactors.
Membranes for Sour Gas Separation. Sep. Purif. Technol. 2012, 89, 1− AIChE J. 2014, 60, 657−672.
8. (394) Faiz, R.; Li, K.; Al-Marzouqi, M. H2S Absorption at High
(374) Zhao, S.; Feron, P. H. M.; Deng, L.; Favre, E.; Chabanon, E.; Pressure Using Hollow Fibre Membrane Contactors. Chem. Eng.
Yan, S.; Hou, J.; Chen, V.; Qi, H. Status and Progress of Membrane Process. 2014, 83, 33−42.
Contactors in Post-Combustion Carbon Capture: A State-of-the-Art (395) Schlünder, E. U. On the Mechanism of Mass Transfer in
Review of New Developments. J. Membr. Sci. 2016, 511, 180−206. Heterogeneous SystemsIn particular in Fixed Beds, Fluidized Beds and
(375) Qi, Z.; Cussler, E. L. Microporous Hollow Fibers for Gas on Bubble Trays. Chem. Eng. Sci. 1977, 32, 845−851.
Absorption: I. Mass Transfer in the Liquid. J. Membr. Sci. 1985, 23, (396) Cussler, R. L. Diffusion: Mass Transfer in Fluid Systems, 3rd ed.;
321−332. Cambridge University Press, New York, 2009.
(376) Qi, Z.; Cussler, E. L. Microporous Hollow Fibers for Gas (397) Prasad, R.; Sirkar, K. K. Dispersion-Free Solvent Extraction
Absorption: II. Mass Transfer Across the Membrane. J. Membr. Sci. with Microporous Hollow-Fiber Modules. AIChE J. 1988, 34, 177−
1985, 23, 333−345. 188.
(377) Mansourizadeh, A.; Ismail, A. F. Hollow Fiber Gas-Liquid (398) Maghsoudi, H.; Soltanieh, M. Simultaneous Separation of H2S
Membrane Contactors for Acid Gas Capture: A Review. J. Hazard. and CO2 from CH4 by a High Silica CHA-Type Zeolite Membrane. J.
Mater. 2009, 171, 38−53. Membr. Sci. 2014, 470, 159−165.
(378) Li, K.; Wang, D.; Koe, C. C.; Teo, W. K. Use of Asymmetric (399) Bunch, J. S.; Verbridge, S. S.; Alden, J. S.; van der Zande, A. M.;
Hollow Fibre Modules for Elimination of H2S from Gas Streams via a Parpia, J. M.; Craighead, H. G.; McEuen, P. L. Impermeable Atomic
Membrane Absorption Method. Chem. Eng. Sci. 1998, 53, 1111−1119. Membranes from Graphene Sheets. Nano Lett. 2008, 8, 2458−2462.
(379) Li, K.; Kong, J. F.; Wang, D.; Teo, W. K. Tailor-Made (400) Jiang, D.; Cooper, V. R.; Dai, S. Porous Graphene as the
Asymmetric PVDF Hollow Fibers for Soluble Gas Removal. AIChE J. Ultimate Membrane for Gas Separation. Nano Lett. 2009, 9, 4019−
1999, 45, 1211−1219. 4024.
(380) Wang, D.; Teo, W. K.; Li, K. Removal of H2S to Ultra-Low (401) Koenig, S. P.; Wang, L.; Pellegrino, J.; Bunch, J. S. Selective
Concentrations Using an Asymmetric Hollow Fibre Membrane Molecular Sieving Through Porous Graphene. Nat. Nanotechnol. 2012,
Module. Sep. Purif. Technol. 2002, 27, 33−40. 7, 728−732.
(381) Wang, D.; Teo, W.; Li, K. Selective Removal of Trace H2S (402) O’Hern, S. C.; Stewart, C. A.; Boutilier, M. S. H.; Idrobo, J.-C.;
from Gas Streams Containing CO2 Using Hollow Fibre Membrane Bhaviripudi, S.; Das, S. K.; Kong, J.; Laoui, T.; Atieh, M.; Karnik, R.
Modules/Contractors. Sep. Purif. Technol. 2004, 35, 125−131. Selective Molecular Transport Through Intrinsic Defects in a Single
(382) Marzouk, S. A. M.; Al-Marzouqi, M. H.; Abdullatif, N.; Ismail, Layer of CVD Graphene. ACS Nano 2012, 6, 10130−10138.
Z. M. Removal of Percentile Level of H2S from Pressurized H2S-CH4 (403) Kim, H. W.; Yoon, H. W.; Yoon, S.-M.; Yoo, B. M.; Ahn, B. K.;
Gas Mixture Using Hollow Fiber Membrane Contactors and Cho, Y. H.; Shin, H. J.; Yang, H.; Paik, U.; Kwon, S.; Choi, J.-Y.; Park,
Absorption Solvents. J. Membr. Sci. 2010, 360, 436−441. H. B. Selective Gas Transport Through Few-Layered Graphene and
(383) Hedayat, M.; Soltanieh, M.; Mousavi, S. A. Simultaneous
Graphene Oxide Membranes. Science 2013, 342, 91−95.
Separation of H2S and CO2 from Natural Gas by Hollow Fiber (404) Li, H.; Song, Z.; Zhang, X.; Huang, Y.; Li, S.; Mao, Y.; Ploehn,
Membrane Contactor Using Mixture of Alkanolamines. J. Membr. Sci.
H. J.; Bao, Y.; Yu, M. Ultrathin, Molecular-Sieving Graphene Oxide
2011, 377, 191−197.
Membranes for Selective Hydrogen Separation. Science 2013, 342, 95−
(384) Marzouk, S. A. M.; Al-Marzouqi, M. H.; Teramoto, M.;
98.
Abdullatif, N.; Ismail, Z. M. Simultaneous Removal of CO2 and H2S
(405) Boutilier, M. S. H.; Sun, C.; O’Hern, S. C.; Au, H.;
from Pressurized CO2-H2S-CH4 Gas Mixture Using Hollow Fiber
Hadjiconstantinou, N. G.; Karnik, R. Implications of Permeation
Membrane Contactors. Sep. Purif. Technol. 2012, 86, 88−97.
Through Intrinsic Defects in Graphene on the Design of Defect-
(385) Rongwong, W.; Boributh, S.; Assabumrungrat, S.;
Laosiripojana, N.; Jiraratananon, R. Simultaneous Absorption of CO2 Tolerant Membranes for Gas Separation. ACS Nano 2014, 8, 841−
and H2S from Biogas by Capillary Membrane Contactor. J. Membr. Sci. 849.
2012, 392-393, 38−47. (406) Celebi, K.; Buchheim, J.; Wyss, R. M.; Droudian, A.; Gasser, P.;
(386) Esquiroz-Molina, A.; Georgaki, S.; Stuetz, R.; Jefferson, B.; Shorubalko, I.; Kye, J.-I.; Lee, C.; Park, H. G. Ultimate Permeation
McAdam, E. J. Influence of pH on Gas Phase Controlled Mass Across Atomically Thin Porous Graphene. Science 2014, 344, 289−
Transfer in a Membrane Contactor for Hydrogen Sulphide 292.
Absorption. J. Membr. Sci. 2013, 427, 276−282. (407) Joshi, R. K.; Carbone, P.; Wang, F. C.; Kravets, V. G.; Su, Y.;
(387) Boucif, N.; Favre, E.; Roizard, D.; Belloul, M. Hollow Fiber Grigorieva, I. V.; Wu, H. A.; Geim, A. K.; Nair, R. R. Precise and
Membrane Contactor for Hydrogen Sulfide Odor Control. AIChE J. ultrafast molecular sieving through graphene oxide membranes. Science
2008, 54, 122−131. 2014, 343, 752−754.
(388) Faiz, R.; Al-Marzouqi, M. Mathematical Modeling for the (408) Sun, C.; Boutilier, M. S. H.; Au, H.; Poesio, P.; Bai, B.; Karnik,
Simultaneous Absorption of CO2 and H2S Using MEA in Hollow R.; Hadjiconstantinou, N. G. Mechanisms of Molecular Permeation
Fiber Membrane Contactors. J. Membr. Sci. 2009, 342, 269−278. Through Nanoporous Graphene Membranes. Langmuir 2014, 30,
(389) Faiz, R.; Al-Marzouqi, M. H2S Absorption via Carbonate 675−682.
Solution in Membrane Contactors: Effect of Species Concentrations. J. (409) Hashimoto, A.; Suenaga, K.; Gloter, A.; Urita, K.; Iijima, S.
Membr. Sci. 2010, 350, 200−210. Direct Evidence for Atomic Defects in Graphene Layers. Nature 2004,
(390) Faiz, R.; Al-Marzouqi, M. Insights on Natural Gas Purification: 430, 870−873.
Simultaneous Absorption of CO2 and H2S Using Membrane (410) Bieri, M.; Treier, M.; Cai, J.; Aït-Mansour, K.; Ruffieux, P.;
Contactors. Sep. Purif. Technol. 2011, 76, 351−361. Gröning, O.; Gröning, P.; Kastler, M.; Rieger, R.; Feng, X.; Müllen, K.;
(391) Sohrabi, M. R.; Marjani, A.; Moradi, S.; Davallo, M.; Shirazian, Fasel, R. Porous Graphenes: Two-Dimensional Polymer Synthesis
S. Simulation Studies on H2S Absorption in Potassium Carbonate with Atomic Precision. Chem. Commun. 2009, 6919−6921.
Aqueous Solution Using a Membrane Module. Asian J. Chem. 2011, (411) Lei, G.; Liu, C.; Xie, H.; Song, F. Separation of the Hydrogen
23, 4227. Sulfide and Methane Mixture by the Porous Graphene Membrane:
(392) Mahdavian, M.; Atashi, H.; Zivdar, M.; Mousavi, M. Simulation Effect of the Charges. Chem. Phys. Lett. 2014, 599, 127−132.
of CO2 and H2S Removal Using Methanol in Hollow Fiber Membrane (412) Sun, C.; Wen, B.; Bai, B. Application of Nanoporous Graphene
Gas Absorber (HFMGA). Adv. Chem. Eng. Sci. 2012, 2, 50−61. Membranes in Natural Gas Processing: Molecular Simulations of
(393) Amrei, S. M. H. H.; Memardoost, S.; Dehkordi, A. M. CH4/CO2, CH4/H2S and CH4/N2 Separation. Chem. Eng. Sci. 2015,
Comprehensive Modeling and CFD Simulation of Absorption of CO2 138, 616−621.

9802 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803
Chemical Reviews Review

(413) Jones, C. W.; Koros, W. J. arbon Molecular Sieve Gas

Separation Membranes-I. Preparation and Characterization Based on
Polyimide Precursors. Carbon 1994, 32, 1419−1425.
(414) Sedigh, M. G.; Onstot, W. J.; Xu, L.; Peng, W. L.; Tsotsis, T.
T.; Sahimi, M. Experiments and Simulation of Transport and
Separation of Gas Mixtures in Carbon Molecular Sieve Membranes.
J. Phys. Chem. A 1998, 102, 8580−8589.
(415) Wenz, G. B.; Koros, W. J. Tuning Carbon Molecular Sieves for
Natural Gas Separations: A Diamine Molecular Approach. AIChE J.
2017, 63, 751−760.
(416) Hinds, B. J.; Chopra, N.; Rantell, T.; Andrews, R.; Gavalas, V.;
Bachas, L. G. Aligned Multiwalled Carbon Nanotube Membranes.
Science 2004, 303, 62−65.
(417) Gilani, N.; Towfighi, J.; Rashidi, A. Performance Enhancement
of Vertically Aligned Carbon Nanotube Membranes for Separation of
Binary Mixtures of H2S/CH4 Using Different Amine Groups. Mater.
Res. Bull. 2016, 77, 155−165.
(418) Lee, S.-H.; Kim, B.-S.; Lee, E.-W.; Park, Y.-I.; Lee, J.-M. The
Removal of Acid Gases from Crude Natural Gas by Using Novel
Supported Liquid Membranes. Desalination 2006, 200, 21−22.
(419) Park, Y.-I.; Kim, B.-S.; Byun, Y.-H.; Lee, S.-H.; Lee, E.-W.; Lee,
J.-M. Preparation of Supported Ionic Liquid Membranes (SILMs) for
the Removal of Acidic Gases from Crude Natural Gas. Desalination
2009, 236, 342−348.
(420) Wang, S.; Wu, D.; Huang, H.; Yang, Q.; Tong, M.; Liu, D.;
Zhong, C. Computational Exploration of H 2 S/CH4 Mixture
Separation Using Acid-Functionalized UiO-66 (Zr) Membrane and
Composites. Chin. J. Chem. Eng. 2015, 23, 1291−1299.
(421) Skoulidas, A. I.; Sholl, D. S.; Krishna, R. Correlation Effects in
Diffusion of CH4/CF4 Mixtures in MFI Zeolite. A Study Linking MD
Simulations with the Maxwell-Stefan Formulation. Langmuir 2003, 19,
7977−7988.
(422) Li, Z.; Liao, F.; Jiang, F.; Liu, B.; Ban, S.; Chen, G.; Sun, C.;
Xiao, P.; Sun, Y. Capture of H2S and SO2 from Trace Sulfur
Containing Gas Mixture by Functionalized UiO-66 (Zr) Materials: A
Molecular Simulation Study. Fluid Phase Equilib. 2016, 427, 259−267.
(423) Reamer, H. H.; Sage, B. H.; Lacey, W. N. Phase Equilibria in
Hydrocarbon Systems - Volumetric and Phase Behavior of the
Methane-Hydrogen Sulfide System. Ind. Eng. Chem. 1951, 43, 976−
981.
(424) Kohn, J. P.; Kurata, F. Heterogeneous Phase Equilibria of the
Methane-Hydrogen Sulfide System. AIChE J. 1958, 4, 211−217.
(425) Kohn, J.; Kurata, F. Volumetric Behavior of the Methane-
Hydrogen Sulfide System at Low Temperatures and High Pressures. J.
Chem. Eng. Data 1959, 4, 33−36.
(426) Coquelet, C.; Valtz, A.; Stringari, P.; Popovic, M.; Richon, D.;
Mougin, P. Phase Equilibrium Data for the Hydrogen Sulphide +
Methane System at Temperatures from 186 to 313 K and Pressures
Up to About 14 MPa. Fluid Phase Equilib. 2014, 383, 94−99.
(427) Lange, S.; Campestrini, M.; Stringari, P. Phase Behavior of
System Methane + Hydrogen Sulfide. AIChE J. 2016, 62, 4090−4108.
(428) Perez, A. G.; Valtz, A.; Coquelet, C.; Paricaud, P.; Chapoy, A.
Experimental and Modelling Study of the Densities of the Hydrogen
Sulphide + Methane Mixtures at 253, 273 and 293 K and Pressures up
to 30 MPa. Fluid Phase Equilib. 2016, 427, 371−383.
(429) Lallemand, F.; Lecomte, F.; Streicher, C. Highly Sour Gas
Processing: H2S Bulk Removal With The Sprex Process. International
Petroleum Technology Conference, Doha, Qatar, Nov 21−23, 2005;
DOI: 10.2523/IPTC-10581-MS.

9803 DOI: 10.1021/acs.chemrev.7b00095

Chem. Rev. 2017, 117, 9755−9803