Reviews D. M. D’Alessandro et al.

DOI: 10.1002/anie.201000431
Carbon Dioxide Capture

Carbon Dioxide Capture: Prospects for New Materials

Deanna M. D!Alessandro,* Berend Smit,* and Jeffrey R. Long*

Keywords:
absorbents · adsorption · carbon dioxide ·
membranes ·
metal–organic frameworks

Angewandte
Chemie
6058 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

The escalating level of atmospheric carbon dioxide is one of the most From the Contents
pressing environmental concerns of our age. Carbon capture and
1. Introduction 6059
storage (CCS) from large point sources such as power plants is one
option for reducing anthropogenic CO2 emissions; however, currently 2. Conventional Chemical
the capture alone will increase the energy requirements of a plant by Absorption 6062
25–40 %. This Review highlights the challenges for capture tech-
nologies which have the greatest likelihood of reducing CO2 emissions 3. Emerging Methods for CO2
Capture 6063
to the atmosphere, namely postcombustion (predominantly CO2/N2
separation), precombustion (CO2/H2) capture, and natural gas 4. New Materials for CO2 Capture 6064
sweetening (CO2/CH4). The key factor which underlies significant
advancements lies in improved materials that perform the separations. 5. Future Prospects 6078
In this regard, the most recent developments and emerging concepts in
CO2 separations by solvent absorption, chemical and physical
adsorption, and membranes, amongst others, will be discussed, with
particular attention on progress in the burgeoning field of metal–
organic frameworks.

1. Introduction

The prospect of a worsening climatic situation due to problem of CO2 capture is regarded as one of the grand
global warming is a subject of widespread public concern, challenges for the 21st century.[3]
with annual global emissions of CO2 having escalated by A number of recent high-profile reports and comprehen-
approximately 80 % between 1970 and 2004.[1] This drastic sive articles have considered the engineering feasibility and
rise has been attributed to an increasing dependence on the economics of CO2 capture, and have sought to estimate the
combustion of fossil fuels (coal, petroleum, and natural gas) costs by modeling reference cases of existing postcombustion
which account for 86 % of anthropogenic greenhouse gas capture in coal- and gas-fired power plants.[3–5] Such conven-
emissions, the remainder arising from land use change tional technologies for large-scale capture have been com-
(primarily deforestation) and chemical processing.[2] mercially available for over 50 years and are focused on the
The urgent need for strategies to reduce global atmos- separation of CO2 from flue gases by the use of amine
pheric concentrations of greenhouse gases has prompted absorbers (“scrubbers”) and cryogenic coolers.[6]
action from national and international governments and The IPCC estimates that CO2 emissions to the atmosphere
industries, and a number of high-profile collaborative pro- could be reduced by 80–90 % for a modern conventional
grams have been established including the Intergovernmental power plant equipped with carbon capture and storage
Panel on Climate Change (IPCC), the United Nations technology.[7] A recent analysis has shown that the thermody-
Framework Commission on Climate Change, and the namic minimum energy penalty for capturing 90 % of the CO2
Global Climate Change Initiative. The capture and seques- from the flue gas of a typical coal-fired power plant is
tration of carbon dioxide—the predominant greenhouse approximately 3.5 % (assuming a flue gas containing 12–15 %
gas—is a central strategy in these initiatives, as it offers the CO2 at 40 8C).[8] By comparison, conventional CO2 capture
opportunity to meet increasing demands for fossil fuel energy using amine scrubbers will increase the energy requirements
in the short- to medium-term, whilst reducing the associated
greenhouse gas emissions in line with global targets.[3] Carbon
[*] Dr. D. M. D’Alessandro
capture and storage (CCS) schemes embody a group of School of Chemistry, The University of Sydney
technologies for the capture of CO2 from power plants, Sydney, New South Wales 2006 (Australien)
followed by compression, transport, and permanent storage. E-mail: [email protected]
CCS will complement other crucial strategies, such as Prof. B. Smit
improving energy efficiency, switching to less carbon-inten- Department of Chemical Engineering
sive fuels such as natural gas and phasing in the use of University of California, Berkeley
renewable energy resources (e.g., solar energy, wind, and Berkeley, CA 94720-1460 (USA)
E-mail: [email protected]
biomass).
A critical point is that the deployment of CCS schemes is a Prof. B. Smit, Prof. J. R. Long
Department of Chemistry
multifaceted problem that requires shared vision and world-
University of California, Berkeley
wide collaborative efforts from governments, policy makers Berkeley, CA 94720-1460 (USA)
and economists, as well as scientists, engineers and venture Fax: (+ 1) 510-643-3546
capitalists. From this perspective, it is apparent why the E-mail: [email protected]

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 6059
Reviews D. M. D’Alessandro et al.

of a plant by 25–40 %.[7–9] Other recent cost analysis estimates

based on near-term conventional regenerative amine scrub-
bing systems have predicted an increased cost of electricity of
$0.06 kWh, or an “avoided cost of capture” of $57–60/tonne
CO2 (as an alternative measure).[8] Clearly, the existing
methods of capture are energy intensive and are not cost-
effective for carbon emissions reduction.[7] These economic
and energy comparisons underscore the immense opportuni-
ties and incentives that exist for improved CO2 capture
processes and materials.
Various components of the CCS process chain including
compression, transportation (by pre-existing pipelines for
instance), and storage of CO2 are technologically mature and
available, and a growing number of fully integrated CCS
projects are reaching the pilot and demonstration phases prior
Figure 1. Innovative concepts in CO2 capture and their cost reduction
to commercialization. In addition to three large-scale dem-
benefits versus their remaining development needs/time to commerci-
onstration projects which are currently underway in Sleipner alization.[5, 12]
West (Norway), Weyburn (Canada), and In Salah (Algeria),
several smaller projects have commenced on the Dutch
continental shelf (Netherlands), Snøhvit (Norway), La Barge It is evident that a consideration of the process design
(Wyoming, United States), Fenn Big Valley (Canada), Ketzin economics and costs are required to assess fully the potential
(Germany), and Schwarze Pumpe (Germany).[9–11] All of the of any given new material. The challenge arises due to
current projects demonstrate carbon storage or reuse in necessary assumptions which must be made, and the varia-
enhanced coal-bed methane recovery schemes, although one tions in the technical characteristics (e.g. fuel used, plant
project at Schwarze Pumpe in Germany, captures CO2 at a characteristics), scale and application of a given material,
coal-based plant. A further 40 CCS projects have already which require that capital costs be balanced with the
been proposed worldwide between 2008 and 2020.[11] efficiency of material. Assessing the economics of CCS is a
One explanation for the slow deployment of fully nontrivial task which is outside the scope of the present
integrated commercial CCS schemes is the considerable cost article. Nevertheless, a number of groups have developed
of the capture phase, which represents approximately two advanced cost-analysis models, which permit a number of the
thirds of the total cost for CCS. A recent comprehensive aforementioned parameters to be varied.[5]
report on postcombustion CO2 capture technologies has Here, we focus on the significant challenge of CO2 capture
determined that the regeneration energy, followed by the and highlight recent advances in materials and emerging
capital cost of capture-specific equipment are the two concepts. The emphasis is on designed materials in which a
variables contributing most significantly to the cost of CO2 molecular level of control can be achieved as a means of
capture.[8] One significant contributor to the regeneration tailoring their performance in separating relevant gas mix-
energy is the maximum separation efficiency which can be tures. In this regard, particular attention is directed towards
achieved by a given capture material. Enhancing this the latest developments in CO2 separations using micro-
efficiency will have the greatest potential for lowering the crystalline porous solids or metal–organic frameworks. For
overall cost of capture systems in near-term,[8] with improve- more in-depth discussions on materials for CO2 separations,
ments in the capture phase for new materials representing one the reader is directed to a number of excellent authoritative
of the foremost challenges.[5] As shown in Figure 1, there reviews in the field.[13–21] We also seek to provide some
exists a serious need for research on innovative new materials criteria, measurement parameters and performance standards
in order to reduce the time to commercialization. in which materials developed in the laboratory can be

Deanna M. D’Alessandro received her PhD Berend Smit received his PhD in Chemistry
in Chemistry from James Cook University from Utrecht University. From 1988 to
(Australia) in 2006 for which she received 1997, he worked as a researcher at Shell
the RACI Cornforth Medal and a 2007 Research. In 1997, he returned to academia
IUPAC Prize for Young Chemists. She held a as Professor of Chemistry at the University
postdoctoral position with Prof. Long at of Amsterdam. In 2005, he was elected as
Berkeley (2007–2009) and was a Dow director of CECAM in Lyon, France. Since
Chemical Company Fellow of the Ameri- 2007, he has been Professor of Chemical
can–Australian Association and an 1851 Engineering and Chemistry at the University
Royal Commission Fellow. In 2010 she will of California, Berkeley. Since 2009 he has
commence a University of Sydney Research been director of the US Department of
Fellowship focusing on energy-related appli- Energy EFRC focusing on carbon capture.
cations of microporous materials. His research interests include the develop-
ment and application of advanced molecular
simulation techniques.

6060 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

evaluated and readily compared so as to assess their materials. Table 1 summarizes the typical gas compositions
performance and guide further experimental efforts. Analysis and properties relevant to postcombustion and precombus-
according to benchmark criteria will permit the most promis- tion processes, which should provide some benchmark criteria
ing materials to be determined in a timely fashion, thereby for experimentalists studying gas separation.
accelerating their deployment at the industrial scale. For postcombustion capture from flue gas, a major
obstacle is the low pressure of the flue gas (ca. 1 atm). The
CO2 concentration is low (ca. 15 %), and capture requires
1.1. Key Challenges for CO2 Capture separation from a high volume stream of flue gas containing
other component gases, predominantly N2 (see Table 1).
Two points must first be made with regards to capture Although current postcombustion capture technology is
materials and potential capture technologies, given the sheer suitable for the retrofitting of existing plants, the energy
magnitude of global CO2 emissions.[8] First, any chemical penalty for a coal-fired power station using traditional
employed to capture CO2 will rapidly exhaust its global aqueous monoethanolamine (MEA) for capture has been
supplies if it is used in a once-through manner; and second, estimated as 25 to 40 %.[6, 8, 9]
any chemical produced from CO2 as a reactant will rapidly Natural gas reserves (mainly CH4) are typically contami-
saturate global markets for that chemical. These consider- nated with over 40 % CO2 and N2, and the utilization of such
ations underscore the necessity that capture materials must be gas fields is only acceptable if this additional CO2 is separated
regenerable. In this case, the energy input for regeneration is and sequestered at the source of production. This application
one of the key factors in determining the efficiency and cost. requires an efficient separation of CO2 from the natural gas
Three CO2 separation issues are considered to hold the components at high pressures. CO2 separation from fuel gas
greatest promise for reducing CO2 emissions, namely 1) sep- (e.g., output from gasification, water-gas shift reactors) also
aration from power plant flue streams, 2) separation from occurs under high pressure conditions at high temperatures
sour natural gas wells, and 3) separation from fuel gas (i.e., (250–450 8C), with the relevant precombustion separation
syngas). Each application involves different gas separations being CO2/H2.
which impose distinct requirements and constraints for The key challenge for gas separations materials is that the
differences in properties between the gases that have to be
separated are relatively small, as is evident from the kinetic
Table 1: Benchmark parameters showing typical compositions of gases
(by weight) in postcombustion and precombustion processes, and the diameters provided in Table 1 (note that the kinetic diameter
kinetic diameter of the gas molecules.[22] is a reflection of the smallest effective dimension of a given
molecule). However, differences do exist in the electronic
Postcombustion Precombustion[a] Kinetic
diameter [!]
properties (i.e., quadrupolar moment and polarization) of the
gases: CO2 has a large quadrupole moment (13.4 ! 10!40 C m2
Composition vs. 4.7 ! 10!40 C m2 for N2, CH4 is non-polar), and CH4 adsorbs
CO2 15–16 % 35.5 % 3.30
preferentially over N2 due to its higher polarizability (26.3 !
H2O 5–7 % 0.2 % 2.65
H2 61.5 % 2.89 10!25 cm3 for CO2 vs. 17.6 ! 10!25 cm3 for N2 and 26.0 !
O2 3–4 % 3.45 10!25 cm3 for CH4).[23]
CO 20 ppm 1.1 % 3.75 Novel concepts for capture therefore require a molecular
N2 70–75 % 0.25 % 3.64 level of control that can take advantage of differences in the
SOx < 800 ppm chemical reactivity of the gas molecules. A further challenge
NOx 500 ppm
is that the selectivity of a separation process is determined by
H2S 1.1 %
Conditions
a combination of adsorption and diffusion selectivity, which
Temperature 50–75 8C 40 8C are coupled in most materials. For example, the introduction
Pressure 1 bar 30 bar of a functional group that specifically binds one species, and
improves on the adsorption selectivity, will simultaneously
[a] After water-gas shift reaction.
decrease the diffusion of these molecules. This inverse
relationship between the adsorption and diffusion selectivity
Jeffrey R. Long received a PhD in Chemistry has recently been investigated extensively by Krishna[24] in a
from Harvard University in 1995. Following broad range of meso- and microporous materials including
postdoctoral work at Harvard and the Uni-
zeolites, carbon nanotubes, carbon molecular sieves, and
versity of California, Berkeley, in 1997 he
joined the faculty at Berkeley, where he is metal–organic frameworks. The need therefore exists to
currently Professor of Chemistry. He is also a design materials in which one can independently tune the
Faculty Senior Scientist in the Materials diffusion and adsorption selectivity at the molecular level.
Sciences Division of Lawrence Berkeley The capture of CO2 from ambient air has also been cited as a
National Laboratory. His research involves strategy to reduce atmospheric CO2 levels, particularly with
the synthesis of new inorganic clusters and
respect to small and mobile emission sources (e.g., aircraft
solids, with emphasis on magnetic and
microporous materials.
and home furnaces).[25] Clearly, the low concentration of CO2
in air (0.04 %) presents a significantly higher thermodynamic
barrier to capture compared with postcombustion methods,
while the expense of moving large volumes of air through an

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6061
Reviews D. M. D’Alessandro et al.

absorbing material presents a further challenge. Indeed, the inhibitors to control corrosion and oxidative degradation due
detailed costs of such approaches are the subject of much to residual oxygen in the flue stream. The sensitivity of the
debate, and are outside the scope of this article. The solvents to chemical degradation from other by-products in
uncertainties in future scenarios of desired CO2 levels in the the flue gas streams, such as SOx and NOx, also lead to
atmosphere, however, emphasize the importance of explor- reduced efficiencies and increased costs for electricity pro-
ing, from a research point of view, the synergies between duction. Improved strategies for postcombustion capture
direct air capture and capture from flue gasses. include the use of liquids with lower heats of adsorption,
increasing the concentration of the adsorbent molecules and
improving the mass transfer and reaction kinetics. The use of
2. Conventional Chemical Absorption other amine-based molecules with lower regeneration tem-
peratures have thus been considered for chemical absorption.
Postcombustion “wet-scrubbing” CO2 capture technolo- Secondary amines such as diethanolamine (DEA) possess a
gies have been employed industrially for over 50 years and lower heat of reaction compared with primary amines: the
are based predominantly on the industrially important lower stability of the carbamate formed upon CO2 absorption
primary alkanolamine MEA.[6] The process involves the gives rise to a more economical regeneration step for the
passage of an aqueous amine solution (typically 25–30 secondary amines.
wt. %) down the top of an absorption tower, while a gaseous As shown in Scheme 1 a, the CO2 loading capacity for
stream of flue gas containing CO2 is introduced at bottom. A primary and secondary amines lies in the range 0.5–1 mol of
blower is required to pump the gas through the absorber. At a CO2 per mol of amine, since a fraction of the carbamate
temperature of approximately 40 8C, the reaction of CO2 with species is hydrolyzed to form hydrogen carbonates. The
the amine occurs through a zwitterion mechanism to form reaction of CO2 with tertiary amines such as N-methyldie-
carbamates, a reaction that has been extensively studied thanolamine (MDEA) occurs with a higher loading capacity
(Scheme 1 a).[26] The liquid amine CO2-rich solvent passes of 1 mol of CO2 per mol of amine, albeit with a relatively
lower reactivity towards CO2 compared with the primary
amines. The carbamation reaction (Scheme 1 a) cannot pro-
ceed for tertiary amines, leading instead to a base-catalyzed
hydration of CO2 to form hydrogen carbonate
(Scheme 1 b).[26] MDEA is commonly employed for natural
gas treatment and exhibits lower solvent degradation rates in
addition to a low energy penalty for regeneration of the
solvent in the stripper. In practice, the addition of small
amounts of primary and secondary amines enhances the CO2
absorption rates for tertiary amines.
Specialty amines such as hindered amines have been
formulated to overcome some of the limitations of the
conventional primary, secondary, and tertiary amines. A
number of reports on the thermodynamic capacity and
absorption/desorption rates of CO2–amine reactions have
identified the steric hindrance and basicity of the amine as the
major factors controlling the efficiency of CO2 capture
Scheme 1. General reaction schemes for the chemical absorption of reactions.[26–30] Sterically hindered amines such as 2-amino-2-
CO2 by a) primary or secondary and b) tertiary amine-containing methyl-1-propanol (AMP) which contain bulkier substituents
solvents. have been identified as the most promising absorption
solvents due to the lower stability of their carbamates
(carbamate stability constant: AMP, 0.1 < DEA, 2.0 <
from the absorber column to a stripping tower where the MEA, 12.5 at 303 K).[26] The sterically hindered amines
mixture is heated with steam to liberate the CO2. The allow CO2 loadings well in excess of 0.5 mol equivalents to
regeneration of the chemical solvent is carried out at elevated be attained, with higher regeneration rates (and therefore
temperatures (100–140 8C) and pressures not much higher lower regeneration costs) compared with the conventional
than atmospheric pressure. The high heat of formation alkanolamines (e.g., the CO2 regeneration rate ratio for
associated with carbamate production leads to a considerable AMP/MEA is 1.83[28]). Differences in the absorption rate
energy penalty for regeneration of the solvent. Following within the class of hindered amines have also been
regeneration, the amine solution is cycled back to the observed:[26] for example, 2-piperidineethanol exhibits supe-
absorption tower for additional CO2 absorption. rior performance compared with AMP and the conventional
Whilst postcombustion methods are advantageous in that alkanolamines due to its lower carbamate stability and higher
the technology is commercially mature and can be easily rate constant for reaction with CO2.
retrofitted into existing power plants, they suffer a number of Inorganic solvents such as aqueous potassium and sodium
drawbacks. These include the considerable energy require- carbonate as well as aqueous ammonia solutions have also
ments for solvent regeneration and the necessary use of been considered for chemical absorption. The chilled-ammo-

6062 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

nia process for CO2 capture involves the reversible formation which involve three stages. Initially, a fuel (natural gas in this
of ammonium hydrogen carbonate, with the forward reaction case) is reacted at high pressure and temperature to form a
to capture CO2 as solid NH4HCO3 occurring at temperatures synthesis gas (syngas) containing CO, CO2, and H2 [Eq. (1)].
below 20 8C.[8, 31] In the regeneration stage, the CO2-rich The second stage involves CO conversion using a water-gas
aqueous ammonium carbonate solution is heated to approx- shift reactor in which the CO2 and hydrogen are separated
imately 80 8C to redissolve the solids. The use of potassium prior to combustion [Eq. (2)].
carbonate (20–40 wt. % K2CO3 in water) is also currently
nearing the pilot stage for both natural gas and coal-fired H2 O þ CH4 ! 3 H2 þ CO ð1Þ
power plants, and is already a mature technology in the oil
H2 O þ CO ! H2 þ CO2 ð2Þ
and gas industry for CO2 capture from pressurized well
gases.[8] As a final point, CO2 capture from ambient air using
chemical absorption in aqueous alkali hydroxide solutions has The significant advantage of precombustion capture is
also been proposed.[25] Despite the strongly absorbing nature that the higher component concentrations and elevated
of the solutions, the large energy demands of the regeneration pressures reduce the energy capture penalty of the process
step present a comparable problem to that involved in to 10–16 %, roughly half that for postcombustion CO2
postcombustion CO2 capture from power plants. capture.[33] Although the initial fuel conversion steps are
more elaborate and costly than in postcombustion systems,
the high concentrations of CO2 produced by the shift reactor
3. Emerging Methods for CO2 Capture (typically 15 to 60 % by volume on a dry basis) and the high
pressures often encountered in these applications are more
While the retrofitting of existing power plants using favorable for CO2 separation. A further advantage is that
postcombustion capture methods presents the closest market- precombustion technology generates a hydrogen-rich fuel,
able technology, two major alternatives to postcombustion which can be used as a chemical feedstock, in a fuel cell for
processes exist.[12, 32] The emerging technologies of precom- power generation, or in the development of a hydrogen
economy. While no commercial IGCC
plants with CCS are yet in operation, the
largest proposed projects include the
recently re-instated 275 MW FutureGen
plant in the U.S. and China"s 250 MW
GreenGen project which is scheduled to
commence operation in 2011.[9]
In contrast to post- and precombustion
methods, oxyfuel combustion takes place
in a pressurized CO2-rich recirculating
stream in which nitrogen is completely
excluded from the combustion process
through a preliminary air separation step
(Figure 2).[12] Suitable fuels include coal,
natural gas, light to medium hydrocarbons
or syngas (which could be derived from
coal). The product of combustion is a
concentrated stream of CO2 (in high,
sequestration ready concentrations of 80–
98 %), which is circulated back into the
system. The water product is removed
from the flue gas by condensation.
Oxyfuel technology forms the basis for
Figure 2. Post-, pre- and oxyfuel combustion processes.[32]
“zero emission cycle” plants which are
considered more promising for new instal-
lations compared with postcombustion
bustion capture and oxyfuel processes (Figure 2) are pro- CO2 capture; however, the process requires significant rede-
jected to attain higher efficiencies for CO2 separation and sign of the turbines, such that the retrofitting of existing plants
capture which will compensate for their extensive capital is not considered economical. The world"s first pilot plant for
investments in the longer term.[7] the capture of carbon dioxide using Oxyfuel technology was
Precombustion methods are employed in natural gas inaugurated in September 2008 by the European energy
plants where CO2 capture occurs at high partial pres- company Vattenfall, at Schwarze Pumpe in Germany.[34] One
sures.[9, 12, 33] The precombustion process for converting a shortcoming of the process is the significant energy demands
methane rich fuel into hydrogen, for example, has given rise of the air separation unit, however, this cost should be
to Integrated Gasification Combined Cycle (IGCC) plants mitigated by the elimination of the need for CO2 capture. By

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6063
Reviews D. M. D’Alessandro et al.

reducing the volume of inert gas (i.e., N2) in the furnace, the polyethylene glycol) and Rectisol (methanol chilled to
thermal efficiency of the boiler unit is also increased. Since !40 8C), for example, have been used industrially for
the separation of interest in this case is air separation (O2 40 years for natural gas sweetening and the treatment of
from mainly N2), reducing the cost of oxygen generation is synthesis gas. The advantage in this case is the lower heat
key. As an alternative to cryogenic air separation, oxygen- consumption in the solvent regeneration step, as the stripping
selective membranes offer a promising new technology in this process can be driven by heat or a pressure reduction (i.e.,
regard.[35] “flash distillation”).[36] Physical solvents are suitable for CO2
capture from high pressure streams such as those relevant for
precombustion capture in IGCC plants.
4. New Materials for CO2 Capture Ionic liquids constitute another class of physical solvents
which are also known to be selective for CO2 absorption.[20, 37]
A diverse range of promising new materials exist for CO2 These comprise combinations of large organic cations and
capture from postcombustion, precombustion, and oxyfuel smaller inorganic anions and are typically viscous liquids near
processes (Figure 3). Significant research efforts have been room temperature. In addition to their extremely low vapor
directed, in particular, towards overcoming the energy pressures, they are non-flammable, environmentally benign,
and can exhibit exceptional thermal
stability. The mechanism for capture is
often based on physisorption, and
involves weak association between the
ionic liquid and CO2 molecules (rather
than chemical bonds), with heats of
adsorption of around
!11 kJ mol!1.[38, 39] In view of this low
heat of reaction, the obvious benefit for
CO2 capture is the minimal energy
required for solvent regeneration. The
capacity is directly proportional to the
partial pressure of CO2 and improves at
pressures above 1–2 bar. For this
reason, ionic liquids were initially pro-
posed for precombustion applications.
It should be noted that some ionic
liquids react with CO2 in a chemisorp-
tion mechanism. An additional benefit
is that ionic liquids can simultaneously
serve to remove CO2 and SO2 (“SO2
polishing”), since the SO2 solubility is
8–25 times greater than that of CO2 at
Figure 3. Materials for CO2 capture in the context of postcombustion, precombustion, and the same partial pressure.[40] This may
oxyfuel processes. not necessarily be advantageous, how-
ever, as an additional step to separate
CO2 from SO2 will be required.
intensive solvent regeneration step and chemical degradation While the viscosity of ionic liquids minimizes solvent loss
issues which are characteristic of conventional solvent from the gas stream, this attribute also limits mass transfers,
absorbents in postcombustion flue gas separations. Examples and they often suffer from low rates of absorption. To
of new materials include physical absorbents, adsorption on surmount these shortcomings and increase the capacity of
solids using pressure and/or temperature swing adsorption simple ionic liquids, “task specific ionic liquids”[41] have been
processes, membranes, cryogenic distillation, gas hydrate developed. The introduction of functional groups such as
formation, and chemical-looping combustion using metal amines into TSILs, have allowed higher rates of adsorption to
oxides, amongst others.[36] be achieved at pressures relevant to flue streams (ca. 1 bar). A
number of reports have also demonstrated extremely high
CO2/N2 selectivities in polymerized ionic liquids, which
4.1. Physical Absorbents exhibit enhanced CO2 solubilities relative to the monomeric
ionic liquid itself.[42] These solid materials also show excep-
A promising alternative to chemical absorption using tional promise as facilitated transport ionic liquid mem-
solvents such as MEA is the use of physical solvents in which branes.[43]
the solvent selectively binds CO2 at high partial pressures and
low temperatures (in accordance with Henry"s Law). Physical
absorbents such as Selexol (a mix of dimethylethers of

6064 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

4.2. Adsorption Materials anion (A) is Cl! , NO3! or CO32!. The majority of studies have
focused on the most inexpensive and common naturally
Solid physical adsorbents (in packed or fluidized adsorb- occurring form of hydrotalcites, Mg-Al-CO3, although recent
ent beds) possess significant advantages for energy efficiency work has demonstrated that the substituted hydrotalcites such
compared with chemical and physical absorption approaches. as those containing Ga3+ substituted for some fraction of the
In the case of adsorbents, however, there may be some Al3+ exhibit enhanced CO2 adsorption properties.[49]
practical difficulties in achieving the same heat exchange The “Dry Carbonate Process” is currently in the exper-
efficiency in the cycle between the hot and cold sections of the imental stages of development and is nearing implementation
capture process. Whereas CO2 molecules dissolve into the in postcombustion coal- and natural-gas pilot plants.[50] Here,
bulk of the material in absorption, CO2 adsorption involves flue gas mixes with a solid dry powdered carbonate sorbent
either physisorption (van der Waals) or chemisorption (e.g., Na2CO3 or K2CO3) in a fluidized bed[8] (in the presence
(covalent bonding) interactions between the gas molecules of water) to form the corresponding hydrogen carbonate salt
and the surface of a material. The CO2-laden solid is purified (NaHCO3 or KHCO3). The regenerative decarbonation
in stages using pressure, vacuum, or temperature swing reaction can be achieved at a relatively low temperature of
adsorption cycles to remove and concentrate the CO2. Several ca. 120 8C.
authoritative review articles have discussed characteristics
and examples of physical adsorbents,[12–14, 20, 21, 44, 45] so rather 4.2.1. Microporous and Mesoporous Materials
than being exhaustive, we will focus on the latest and most
promising developments in the field during the past few years. Zeolites are amongst the most widely reported physical
A key concern for physical adsorbents is balancing a adsorbents for CO2 capture in the patent and journal
strong affinity for removing an undesired component from a literature.[13] They constitute the primary adsorption material
gas mixture with the energy consumption required for their for commercial hydrogen production (involving H2/CO2
regeneration. In addition to the adsorption capacity, the separation) using pressure swing adsorption, with the most
selectivity is a principal property relevant to adsorptive gas popular of these based on zeolite 13X.[51] Zeolites are
separation. While both factors are dependent on the opera- typically employed at elevated pressures (above 2 bar), and
tional temperature and pressure, as well as the nature of the their adsorption capacity has been shown to be greatly
adsorbent and the gas adsorbate, the factors which influence reduced by the presence of moisture in the gas, thereby
selectivity are more complicated. Possible mechanisms of necessitating very high regeneration temperatures (often in
adsorptive separation include: 1) the molecular sieving effect, excess of 300 8C).[44, 52] These additional recovery costs for
which is based upon size/shape exclusion of certain compo- their regeneration pose a significant disadvantage.
nents of a gas mixture; 2) the thermodynamic equilibrium Recently, attention has turned to experimental[44] and
effect, due to preferential adsorbate-surface or adsorbate computational[53] screening studies to assess CO2 removal
packing interactions; and 3) the kinetic effect, due to differ- from low pressure flue gas using naturally occurring zeolites,
ences in the diffusion rates of different components of a gas such as X and Y Faujasite systems,[53] as well as synthetic
mixture.[16] zeolites including 5A and 13X.[44] Experimental studies on
A variety of solid physical adsorbents have been consid- several synthetic zeolites have shown that the most promising
ered for CO2 capture including microporous and mesoporous candidates for capture of CO2 from a simulated flue gas
materials (carbon-based sorbents such as activated carbon mixture (consisting of CO2 and N2) are characterized by a
and carbon molecular sieves, zeolites, and chemically modi- near linear CO2 adsorption isotherm. However, a trade-off
fied mesoporous materials), metal oxides, and hydrotalcite- exists here since a linear isotherm is indicative of weak
like compounds, amongst others.[13] adsorbent–adsorbate interactions, which is not compatible with
Metal oxides (such as CaO and MgO) are promising a high CO2/N2 selectivity. The adsorption can be enhanced by a
capture materials given their ability to retain high adsorption low SiO2/Al2O3 ratio and the presence of cations in the zeolite
capacities at temperatures above 300 8C.[13, 46] The operation of structure which exhibit strong electrostatic interactions with
the materials can be defined by a carbonation–calcination CO2.[44] While these potential adsorbents were shown to be
cycle: the carbonation reaction of CO2 with solid CaO at 600– adequate for pressure swing adsorption applications, their
650 8C precipitates calcium carbonate (CaCO3), while the regeneration required significant energy input.
reverse calcination reaction regenerates the oxide at 800– An advantage of porous solid materials is the ability to
850 8C. On the downside, adsorbent degradation has been modify their properties by impregnating or tethering active
observed after several cycles. Improved oxide materials groups such as alkyl-amines onto their internal surfaces. This
containing lithium, such as Li2ZrO3 and Li4SiO4 have recently strategy has often been exploited to improve the gas sorption
attracted attention for their high CO2 adsorption capacities.[47] properties of porous materials for low pressure capture
The class of anionic and basic clays known as hydrotalcites applications, such as those relevant in flue streams and for
and their derivatives have also been shown to be suitable as capture from ambient air. In this regard, numerous amine-
CO2 adsorbents for precombustion processes at temperatures modified silica materials have been prepared.[13, 54] The surface
as high as 400 8C.[13, 48] These compounds belong to a family of modification with primary amines facilitates the adsorption of
layered double hydroxide solids with the general stoichiom- CO2 through the formation of carbamate species, reminiscent
etry M2+1!xM3+x(OH)2Am!x/m·y H2O where the divalent ion is of the amine–CO2 chemistry in conventional liquid phase
typically Mg2+, the trivalent ion is typically Al3+, and the scrubbing (see Scheme 1). The stripping of CO2 can be

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6065
Reviews D. M. D’Alessandro et al.

achieved at lower temperatures than those required for the these processes are already operational worldwide, the use of
regeneration of amine solvents (typically > 100 8C), thus CO2 (rather than the conventional technique requiring vast
decreasing the energy requirements of the process.[55, 56] The quantities of water) has been cited as a more environmentally
impregnation of polyethylenimine into periodic MCM-41 feasible and attractive alternative.[20]
mesoporous molecular sieves has also been shown to lead to a Few studies on carbonaceous adsorbents have investi-
significant enhancement (24-fold) in the CO2 absorption gated their feasibility for low pressure flue gas applications
capacity of the solid support using a pressure swing adsorp- using vacuum or temperature swing regeneration approaches.
tion approach.[13, 57] Increased absorption capacities were also In a recent study, activated carbon and charcoal were shown
observed in a series of amines immobilized in supports such as to exhibit moderate adsorption selectivity for CO2 over N2
poly(methyl methacrylate).[58] However, these materials (defined as the ratio of the pure-component sorption capaci-
impregnated with physisorbed amines often suffer from a ties) at low pressures below 1 atm, while increasing the
lack of stability over repeated cycles. pressure reduced the selectivity.[20] In general, the selectivity
To overcome these limitations, alkylamines have been and capacity of carbonaceous adsorbents is too low for
covalently tethered to the surface of mesoporous supports in postcombustion applications; however, activated carbons are
an attempt to increase their stability. For example, aziridine attractive for precombustion CO2 capture.
polymerization at the surface of mesoporous silica was used to A technique gaining increasing interest for use in CO2
generate a hyperbranched material which was shown to capture with carbon-based sorbents is electrical swing
exhibit reversible CO2 binding (with a capacity of 2 mmol adsorption. Examples include conductive monolithic meso-
CO2/g adsorbent) and multi-cycle stability under simulated porous carbon adsorbents, which rapidly adsorb and desorb
flue gas conditions using a temperature swing adsorption CO2 with the application and removal of electric charge.[61]
(TSA) approach.[54] The influence of the amine type and the
presence of moisture on CO2 adsorption performance have 4.2.3. Organic Solids
also been investigated for SBA-15 grafted with monoamino,
diamino, and triamino ethoxysilanes.[59] Capacities of 0.52, Non-porous self-assembled p-tert-butylcalix[4]arene
0.87, and 1.10 mmol CO2/g adsorbent, respectively, were organic solids have also been considered for CO2 capture.[62, 63]
obtained. In the presence of a moist CO2 stream, the capacity The structure is characterized by cone-shaped calixarene
decreased slightly for the primary amine, but increased by ca. molecules that are stabilized by intramolecular hydrogen
3 and 10 % for the secondary and tertiary amine grafted bonds, as well as the presence of hydrophobic nanodimen-
materials, respectively. sional channels. Upon exposure of the solid to a high pressure
A significant conclusion from these studies was the of CO2 (35 atm) at room temperature, the network undergoes
dependence of the adsorption performance on the surface a gas driven single-crystal-to-single-crystal phase transforma-
density of the amine groups.[44] In considering the adsorption tion from the “guest-free” thermodynamic form to the kinetic
efficiency of the materials, both the rate of adsorption and the form over ca. 2 h.[63] A CO2 capacity of 6.9 wt % and enthalpy
adsorption capacity must be optimal. The amine grafting of adsorption of !16 kJ mol!1 was observed, while exposure
procedure which is employed often involves the addition of an of the solid to 20 atm of H2 gas did not result in any detectable
excess of the grafting reagent to the support. However, the adsorption.[62] The materials may be applicable for high-
amount which is ultimately grafted to the surface may vary pressure CO2/H2 syngas separations.
over repeat syntheses, and may not result in an optimal Recent work has demonstrated the potential of covalent
amount of grafted amine for the particular CO2 capture organic frameworks (COFs) for CO2 capture.[64] These are
process. These considerations highlight the importance of crystalline microporous materials similar to the metal–
studies on the influence of the quantity of grafting reagent organic frameworks discussed below, but with frameworks
added on the actual amount of amine that is covalently built up of only lightweight organic components. COF-102
attached to the surface. Only with this data in hand, can the (C25H24B4O8) is comprised of tetra(4-(dihydroxy)borylphe-
performance of new solid adsorbents for the CO2 capture nyl)methane units and exhibits the highest CO2 uptake, to
applications highlighted here be critically assessed. date, in this class (27 mmol g!1 at 55 bar and 298 K).[64]
Molecular simulations have corroborated the experimental
4.2.2. Carbonaceous Adsorbents findings, predicting exceptionally high uptakes in these
materials.[65, 66]
Adsorption studies on activated carbon, charcoal, and
virgin coal have focused on high pressure CO2 capture 4.2.4. Metal–Organic Frameworks
applications given that the adsorption capacities scale with
pressure.[13, 60] While the surface properties of the adsorbents The past 20 years have seen remarkable progress in the
can vary widely, the materials are advantageous in that they design, synthesis, and characterization of metal–organic
are inexpensive relative to other solid adsorbents (such as frameworks (MOFs) owing to their enormous structural and
zeolite 13X), and are insensitive to moisture. chemical diversity and their potential applications in gas
The majority of studies on carbon-based sorbents are storage, ion exchange, molecular separation, and heteroge-
motivated by the significant industrial potential of enhanced neous catalysis.[67–79] These microporous crystalline solids are
oil recovery schemes, which involve the CO2-driven displace- composed of organic bridging ligands or “struts” coordinated
ment of valuable oil from underground reserves.[7] Although to metal-based nodes to form a three-dimensional extended

6066 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

network with uniform pore diameters typically in the range 3 Other terephthalate-based materials have been exten-
to 20 #.[80–85] The nodes generally consist of one or more sively investigated with regard to their CO2 capture proper-
metal ions (e.g., Al3+, Cr3+, Cu2+, or Zn2+) to which the ties. For example, room temperature CO2 adsorption has been
organic bridging ligands coordinate through a specific func- reported in a series of isoreticular frameworks including
tional group (e.g., carboxylate, pyridyl). The intense current [Zn4O(bdc)3] (MOF-5, bdc2! = 1,4-benzenedicarboxylate)
research efforts towards industrial applications of MOFs in and [Zn4O(NH2bdc)3] (NH2bdc2! = 2-amino-1,4-benzenedi-
gas storage, separation, and catalysis is attributed to their carboxylate).[88] Notably, the isotherms exhibited an interest-
unique structural properties, including: robustness, high ing S-shaped profile in which the presence of inflection points
thermal and chemical stabilities, unprecedented internal indicated electrostatic interactions between the CO2 mole-
surface areas (up to 5000 m2 g!1), high void volumes (55– cules themselves, which dominate as the pressure increases.
90 %), and low densities (from 0.21 to 1.00 g cm!3), which can Measurements on the adsorption kinetics for CO2 in
be maintained upon evacuation of the guest molecules from [Zn4O(bdc)3][88, 90, 91] showed that the activation energy for
the pores.[86] The regular monodisperse nature of the crystal- diffusion is directly proportional to the ratio of the kinetic
line array of micropores is a key feature that distinguishes diameter of the diffusing molecule to the pore opening of the
these systems from other porous materials (e.g., polymers, structure (as is the case for zeolites). Based on this rationale,
mesoporous silicas, carbons, etc.). In addition, the ability to frameworks with smaller pores possess a higher barrier to
modulate systematically the pore dimensions and surface diffusion.[90]
chemistry within metal–organic frameworks[87] is a feature A molecular sieving effect can be exploited for the
that was previously largely absent in zeolite materials. Here, separation of gas mixtures in cases where the pore sizes are
we focus on four sub-sets of metal–organic frameworks, sufficiently small, and are in the range of the kinetic diameters
namely those which contain open metal sites, interpenetrated of the adsorbates. For example, the frameworks [Zn4O(fma)3]
frameworks, flexible frameworks, and surface-functionalized (fma = fumarate)[92] and [Sc2(bdc)3][93] possess smaller pores
frameworks. than those in the aforementioned structures such as [Zn4O-
The high surface area-to-weight ratio of MOFs is such that (bdc)3] which lie in the range required for industrially
they have enhanced capacities for CO2 capture at moderate important gas separations (Table 1).[92] The slow kinetics
pressures compared with zeolites. While zeolites possess associated with diffusion through small pores, however, could
higher storage capacities at pressures of less than 10 bar, it has be problematic for separation materials which are designed
been projected that their maximum capacities are limited to solely on the basis of kinetic size considerations.
one third those of MOFs at pressures greater than 10 bar.[8] In addition to such kinetic size effects, favorable electro-
The capacities of metal–organic frameworks up to high static interactions between CO2 and the framework also
pressures scale with the amount of active area per unit provide enhanced capture, as in the case of [Al(OH)(ndc)]
weight: activated carbon has an active area of 400– (ndc2! = 1,4-napthalenedicarboxylate).[94] Gas uptake meas-
1000 m2 g!1, zeolites of up to 1500 m2 g!1, and frameworks of urements revealed that while CO2 (kinetic diameter 3.3 #) is
1500–4500 m2 g!1.[8] excluded from entering the 3.0 # diameter pores, favorable
In addition to the adsorption capacity, the selectivity is a interactions between the hydroxy groups of the framework
principal property relevant to adsorptive gas separation, and protruding toward the 7.7 # diameter channels give rise to
is determined by an interplay of factors including the preferential CO2 adsorption over N2 (by a factor of ca. 10).
molecular sieving effect, the thermodynamic equilibrium The selective capture of CO2 from CO2/CH4 mixtures can
effect, and the kinetic effect.[16] While the mechanism for be achieved by exploiting the favorable interactions between
CO2 capture and separation can often be determined the framework and the quadrupole moment of CO2 (13.4 !
predominantly by one of these factors, it is more often the 10!40 C m2, compared with CH4 which is non-polar). For
case that a synergistic combination of effects is operative. [Zn2(bpdc)2(dpni)] (dpni = N,N’-di(4-pyridyl)-1,4,5,8-naph-
Table 2 summarizes the examples of metal–organic frame- thalenetetracarboxydiimide, bpdc2! = biphenyl-4,4’-dicarbox-
works reported to date that have been explored for applica- ylate),[95, 96] a CO2/CH4 selectivity of ca. 30 was calculated
tions in CO2/N2 (postcombustion), CO2/H2 (precombustion) using the ideal adsorbed solution theory (IAST). While the
or CO2/CH4 (natural gas sweetening) separations. majority of adsorption studies have reported selectivities
The peak gravimetric adsorption for CO2 has been based on single component adsorption isotherms, IAST
reported in frameworks with high surface areas and pore serves as the benchmark for the simulation and computa-
diameters of greater than 15 #. The framework [Zn4O(btb)2] tional analysis of binary mixture adsorption from the
(MOF-177, btb3! = 1,3,5-benzenetribenzoate) with a surface experimental isotherms.[97]
area (SA) of 4500 m2 g!1 exhibits the highest capacity for CO2, Zeolitic imidazolate frameworks (ZIFs) constitute a sub-
taking up 33.5 mmol g!1 at 32 bar.[88] By comparison, the class of metal–organic frameworks that can adopt zeolite
benchmark material zeolite 13X adsorbs 7.4 mmol g!1 at structure types based on the replacement of: 1) tetrahedral
32 bar (Table 2).[88] High surface area frameworks such as Si4+ and Al3+ ions with tetrahedral transition metal ions such
[Cu2(BPnDC)2(bpy)] (SNU-6, bpy = 4,4’-bipyridine, as Zn2+ or Co2+ and 2) bridging O2! ions with bridging
BpnDC2! = benzophenone-4,4’-dicarboxylate, SABET = imidazolate-based ligands. A recent extensive review arti-
2590 m2 g!1) also possess a high CO2 uptake capacity and cle[98] has highlighted the exceptional selective CO2 capture
preferential CO2 adsorption over CH4.[89] and storage properties of ZIFs.[99–101] High-throughput syn-
thetic methods have been utilized with great utility to produce

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6067
Reviews D. M. D’Alessandro et al.

Table 2: CO2 capture and separation properties of selected metal–organic frameworks categorized as 1) containing open metal sites, 2) inter-
penetrated, 3) flexible, and 4) functionalized.
Material[a] Category CO2 uptake !DHads Separation Ref.
(conditions)[b] [kJ mol!1][c] application,
Selectivity[d]
[Zn4O(btb)2] (MOF-177) 33.5 mmol g!1 (298 K, 32 bar) [88]
[Zn4O(bdc)3] (MOF-5) 2.10 mmol g!1, 9.24 wt % (295 K, 1 atm) [88]
[Zn4O(NH2bdc)3] (4) [88]
[Zn4O(fma)3] 69 wt % CO2 vs. 8.6 wt % CH4 16.1 vs. 12.0 CH4 CO2/CH4 [92]
(300 K, 28 bar)
[Sc2(bdc)3] 0.9 mmol g!1 (235 K, 1 bar), 4.5 mmol g!1 20–25 CO2/CH4 ; CO2/H2 [93]
(50 bar)
[Zn2(bpdc)2(dpni)] CO2/CH4, % 30IAST [95–97]
(296 K, 0–18 bar)
[Cu2(BPnDC)2(bpy)] (SNU-6) 25.8 mmol g!1, 114 wt % (195 K, 1 bar), [89]
2.50 mmol g!1, 11 wt % vs. 0.69 mmol g!1,
1.11 wt % CH4 (273 K, 1 atm)
[Cu3(btc)2] (1) 10.9 mmol g!1 (298 K, 6 bar), 12.7 mmol g!1 30–35 CO2/CH4 ; CO2/N2 [88, 104–
(15 bar) vs. % 4 mmol g!1 CH4 (15 bar) 106]
and % 2 mmol g!1 N2 (15 bar)
[Cr3F(H2O)2O(btc)]3 (MIL-100) (1) 18 mmol g!1 vs. 7.5 mmol g!1 CH4 62 vs. 19 CH4 CO2/CH4 [109]
(303 K, 48.7 bar)
[Cr3F(H2O)2O(bdc)]3 (MIL-101) (1) 40 mmol g!1 vs. 12 mmol g!1 CH4 44 vs. 18 CH4 CO2/CH4 [109]
(303 K, 48.7 bar)
[Ni2(pbmp)] (1) 2.5 mmol g!1 (304 K, 1 bar), 6 mmol g!1 35 vs. 15 CH4 CO2/CH4 ; CO2/N2 [110]
(15 bar)
[Mg2(dobdc)] (1) 23.6 wt % (298 K, 0.1 atm), 35.2 wt % 38–42 vs. 20–22 CH4 CO2/CH4 [111–113]
(1 atm), 63 wt % (50 bar)
[Ni2(dobdc)] (1) 51 wt % (298 K, 50 bar) 41 [111–113]
[Zn3(OH)(p-cdc)2.5] (1) 0.566 mmol g!1 (298 K, 0.5 bar) CO2/CH4, 5–17IAST [97, 114]
vs. 0.0754 mmol g!1 CH4 (0.5–15 bar)
[Co4(m-OH2)4(mtb)2] (1) 1.59 mmol g!1, 7.02 wt % (273 K, 1 atm) CO2/CH4 ; CO2/N2 [115]
[Zn(adc)(Bpe)0.5] (2) CO2/CH4 ; CO2/N2 [116]
[Ni(cyclam)2(mtb)] (2) 2.53 mmol g!1, 11.2 wt % (195 K, 1 atm) CO2/CH4 ; CO2/N2 [89]
[Zn(bdc)(bpy)0.5] (MOF-508b) (2) % 6 mmol g!1, 26.0 wt % vs. 5.5 wt % N2 14.9 vs. 5.6 N2 CO2/CH4 ; CO2/N2 [117]
and 3.2 wt % CH4 (303 K, 4.5 bar) and 5.1 CH4
[Mg(tcpbda)] (1,2) 1.49 mmol g!1, 6.5 wt % (298 K, 1 atm) CO2/CH4 ; CO2/N2 [118]
vs. 0.45 mmol g!1, 0.73 wt % CH4
and negligible N2
[Cr(OH)(bdc)] (MIL-53(Cr)) (3) 8.5 mmol g!1 (10 bar) vs. 2–3 mmol g!1 CO2/CH4 [119]
CH4 (5–10 bar)
Hydrated [Cr(OH)(bdc)] (3) 7.7 mmol g!1 (304 K, 18 bar) CO2/CH4 [120]
vs. 0.2 mmol g!1 CH4 (20 bar)
[Co(F-pymo)2] (3) 7 mmol g!1 (273 K, 20 bar) vs. negligible CH4 CO2/CH4 [121]
[Zn(F-pymo)2] (3) 8 mmol g!1 (273 K, 20 bar) vs. negligible CH4 CO2/CH4 [121]
[(Ni2L1)(bptc)] (ethyl-bridged) (3) 9.3 wt % (298 K, 1 atm), 15 wt % (15 bar) CO2/CH4 ; CO2/H2 ; [122]
CO2/N2
[(Ni2L2)(bptc)] (butyl-bridged) (3) 0 wt % (298 K, 1 atm), 21 wt % (15 bar) [122]
[Zn2(bttb)(py-CF3)2] (4) % 0.2 mmol g!1 (298 K, 0.15 bar), CO2/N2, % 40IAST; [123]
% 3 mmol g!1 (18 bar) CO2/CH4, 6IAST
(298 K)
[Zn2(bttb)] (1) % 0.4 mmol g!1 (298 K, 0.15 bar), CO2/N2, % 23IAST; [123]
% 5 mmol g!1 (18 bar) CO2/CH4, 4.5IAST
(298 K)
[H3O][Zn7(m3-OH)3(bbs)6] (UoC-1) (4) 2 mmol g!1 (273 K, 1 bar) vs. CO2/CH4 ; CO2/N2 ; [124]
0.57 mmol g!1 CH4 and negligible CO2/H2
N2 and H2
[Al(OH)(NH2bdc)] (3,4) 2.3 mmol g!1, 10 wt % (303 K, 5 bar), 38.4 vs. 20 CH4 CO2/CH4 [94, 125]
6.7 mmol g!1, 30 wt % (13 bar)
vs. 2.4 mmol g!1 CH4 (30 bar)
[Al(OH)(bdc)] (MIL-53(Al)) (3) 10 wt % (298 K, 1 atm) 20.1 vs. 20 CH4 CO2/CH4 [25]
[Ni2(NH2bdc)2(dabco)] (4) 14 wt % (298 K, 1 atm), 60 wt % (25 atm) [126]
[Ni2(bdc)2(dabco)] 10 wt % (298 K, 1 atm) [126]
[In(OH)(NH2bdc)] (4) 8 wt % (298 K, 1 atm) [126]
[In(OH)(bdc)] 4 wt % (298 K, 1 atm) [126]

6068 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

Table 2: (Continued)
Material[a] Category CO2 uptake !DHads Separation Ref.
(conditions)[b] [kJ mol!1][c] application,
Selectivity[d]
HCu[(Cu4Cl)3(BTTri)8(en)5] (4) 0.366 mmol g!1, 1.6 wt % (0.06 atm, 298 K), 90 CO2/N2 [127]
1.27 mmol g!1, 5.5 wt % (1 bar)
HCu[(Cu4Cl)3(BTTri)8] (1) 0.277 mmol g!1, 0.92 wt % (0.06 atm, 298 K), 21 CO2/N2 [127]
3.24 mmol g!1, 14.3 wt % (1 atm)
Activated carbon-MAXSORB [e]
0.5 mmol g!1 (1 atm), 14.6 [45, 128]
25 mmol g!1 (298 K, 30 bar)
BPL carbon [e]
% 0.4 mmol g!1 (1 atm), 24.3 CO2/CH4 ; CO2/N2 [45, 98]
4 mmol g!1 (10 atm)
Zeolite 13X [e]
3.3 mmol g!1 (323 K, 1 atm), 6.9 mmol g!1 49 [51, 88, 106]
(295 K, 15 bar), 7.4 mmol g!1 (298 K, 32 bar)
[a] Abbreviations: btb = 1,3,5-benzenetribenzoate, bdc = 1,4-benzenedicarboxylate, NH2bdc = 2-amino-1,4-benzenedicarboxylate, fma = fumarate,
bpdc = biphenyl-4,4’-dicarboxylate, dpni = N,N’-di(4-pyridyl)-1,4,5,8-napthalenetetracarboxydiimide, bpy = 4,4’-bipyridine, BPnDC = benzophenone-
4,4’-dicarboxylate, ndc = 1,4-napthalenedicarboxylate, btc = 1,3,5-benezenetricarboxylate, pbmp = N,N’-piperazinebismethylenephosphonate,
H4dobdc = 2,5-dihydroxyterephthalic acid, p-cdc2! = deprotonated form of 1,12-dihydroxydicarbonyl-1,12-dicarba-closo-dodecaborane, mtb = meth-
anetetrabenzoate, H2tcpbda = N,N,N’,N’-tetrakis(4-carboxyphenyl)-biphenyl-4,4’-diamine, adc = 4,4’-azobenzenedicarboxylate, Bpe = trans-bis(4-pyr-
idyl)ethylene), cyclam = 1,4,8,11-tetraazacyclotetradecane, F-pymo = 5-fluoropyrimidin-2-olate, L1 and L2 = ethyl- and butyl-bridged Ni2 macrocyclic
complexes shown in Figure 5 b, bptc = 1,1’-biphenyl-3,3’,5,5’-tetracarboxylate, bttb = 4,4’,4’’,4’’’-benzene-1,2,4,5-tetrayltetrabenzoate, bbs = dianion of
4,4’-bibenzoic acid-2,2’-sulfone, dabco = 1,4-diazabicyclo[2.2.2]octane, H3BTTri = 1,3,5-tri(1H-1,2,3-triazol-4-yl)benzene, 4’-tert-butyl-biphenyl-3,5-
dicarboxylate. [b] Uptakes of CH4, N2 and H2 shown where appropriate for comparison. [c] DHads = Enthalpy of adsorption for CO2. The initial
isosteric heat of adsorption is reported unless otherwise stated. [d] Reported selectivities from single adsorption isotherms, unless otherwise stated:
IAST = calculated using IAST theory from the experimental single adsorption isotherms assuming equimolar gas mixtures. [e] Benchmark material.

ZIFs with a range of pore aperture metrics. Functionalization The presence of water in such frameworks can lead to
of the imidazolate and benzimidazolate linkers was also striking enhancements in their CO2 capture abilities. Recent
shown to permit fine-tuning of the interactions between the experimental and computational studies[108] showed a significant
pore walls and guest molecules, thereby varying the selectivity increase in the CO2 adsorption capacity as well as enhanced
of adsorption. Importantly, in contrast to many metal–organic CO2/CH4 and CO2/N2 selectivities in the [Cu3(btc)2] framework
frameworks, ZIFs exhibit high thermal stabilities and chem- containing 4 wt % water.[106] In this case, the quadrupole
ical stability in refluxing aqueous and organic media, which moment of CO2 interacts with the electric field created by the
are required for practical separations processes.[99] In partic- water molecules which are coordinated to the open Cu2+
ular, the stability of a framework toward long-term exposure sites.[106] The working capacity of the hydrated [Cu3(btc)2]
to water vapor is a critical issue in determining its suitability material for CO2 capture is almost four times that of the
for CO2 capture from flue gas.[102] benchmark material zeolite 13X.[106] It is not clear, however, if
such an enhancement will extend to other frameworks with
4.2.4.1. Frameworks Containing Open Metal Sites exposed metal cation sites, and in general there is a need for
further study of the effects of water vapor on performance.
The presence of coordinatively unsaturated metal sites in CO2 adsorption has been investigated in the MIL
MOFs also provides a mechanism for the enhanced separa- (Mat$riaux Institut Lavoisier) series of frameworks [Cr3F-
tion of (quadru)polar/non-polar gas pairs such as CO2/CH4. (H2O)2O(btc)]3 (MIL-100, SALangmuir = 3100 m2 g!1) and
Amongst the reported examples, [Cu3(btc)2] (HKUST-1, [Cr3F(H2O)2O(bdc)]3 (MIL-101, SABET = 4100 m2 g!1) which
btc3! = 1,3,5-benzenetricarboxylate), which consists of pad- exhibit high surface areas and large pore sizes in addition to
dlewheel Cu2(COO!)4 units connected through btc3! ligands, coordinatively unsaturated Cr3+ sites.[109] Both frameworks
is amongst the most extensively studied.[103] The results for the possess higher capacities for CO2 relative to CH4, as well as
CO2 uptake capacities reported by various groups differ moderately high enthalpies of adsorption for CO2 at zero
significantly,[88, 104–106] but the highest uptake capacity has been coverage (!44 and !63 kJ mol!1 for MIL-100 and MIL-101,
reported as 12.7 mmol g!1 at 15 bar and 298 K.[106] The respectively). The mechanism of adsorption involves end-on
increase in the CO2 adsorption capacity with increasing coordination (i.e., O=C=O···Cr3+). A recent study has dem-
pressure and decreasing temperature is indicative of phys- onstrated that the Dual Site Langmuir model is useful for
isorption in a microporous solid, and indicates the potential understanding adsorption behaviour in frameworks such as
suitability for application of the material in pressure or MIL-101 which involve heterogeneity due to the presence of
temperature swing adsorption processes. The framework open metal sites.[23] A nearly linear correlation was also
preferentially adsorbs CO2 over both CH4 and N2 with a uncovered between the polarizability of the gas molecule and
moderate enthalpy of adsorption (!30 to !35 kJ mol!1).[106] the resulting enthalpy of adsorption of the adsorbate in the
The mechanism of adsorption can be described by CO2 framework. This finding provides an important design con-
coordination to the metal center in an end-on fashion, i.e., sideration for gas separations in materials containing exposed
O=C=O···Cu2+.[107] metal cation sites.

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6069
Reviews D. M. D’Alessandro et al.

The influence of the identity of the metal center on the IAST calculations have also been employed to predict the
capacity and selectivity of CO2 adsorption has been system- selectivity for CO2/CH4 mixture adsorption in frameworks
atically studied in the series of isostructural frameworks such as [Zn3(OH)(p-cdc)2.5(DMF)4] (p-cdcH2 = 1,12-dihy-
[M2(dobdc)(H2O)2] (M = Ni, Co, Zn, Mg, Mn; H4dobdc = 2,5- droxydicarbonyl-1,12-dicarba-closo-dodecaborane) with and
dihydroxyterephthalic acid).[111] As shown in Figure 4, these without coordinatively unsaturated Zn2+ sites, i.e., with and
without coordinated DMF molecules.[97, 114] As expected, the
fully desolvated framework containing open metal sites
exhibited a higher adsorption capacity due to the favorable
interactions between the partial charges on the framework
and the quadrupole moment of CO2. From the single
component adsorption isotherms, IAST predicts that the
selectivities for the solvated, and desolvated frameworks
approach one another at higher pressures (up to 20 bar), since
the enhanced adsorption of CO2 at open metal sites becomes
less important as the pores are filled with molecules. The
isosteric heat of adsorption for the material containing open
metal sites also decreases significantly after full occupa-
tion.[111]
It must be noted that the morphology of the sample has
been shown to significantly affect gas adsorption. For
example, three samples of the carborane-based framework
Figure 4. The crystal structure of [Mg2(dobdc)] viewed along the one- [Co4(OH)2(p-cdc)3(DMF)2][130] were prepared using slight
dimensional channels.[113] The inset shows the charge-induced inter- variations in the reaction conditions and co-solvents to
action between one of the open Mg2+ coordination sites and a CO2
generate block-like single crystals, polycrystalline tetragonal
molecule. Blue, red, and gray spheres represent Mg, O, and C atoms,
respectively. Hydrogen atoms are omitted for clarity. microrods and a less-crystalline material in the form of
agglomerates. The different porosities, crystallinity, and
internal surface areas of the samples led to variable CO2
frameworks are characterized by a honeycomb structure with uptake data.
one-dimensional channels of 11–12 # diameter, and a high Other frameworks such as [Co4(m-OH2)4(mtb)2], which is
concentration of open metal sites which are generated upon based on the tetrahedral organic linker mtb4! (methane-
removal of the coordinated water molecules. The CO2 uptake tetrabenzoate) and contains one coordinatively unsaturated
of [Mg2(dobdc)] is more than double that for any other site per metal center, also show selective CO2 adsorption over
member of the series,[112] and the high enthalpy of adsorption CH4 and N2.[115] The open Co2+ sites endow the framework
(!47 kJ mol!1) relative to the isostructural frameworks (!41 with exceptional selectivity for O2 over N2, which has
and !37 kJ mol!1 for M = Ni and Co, respectively) suggests relevance to air separation processes for oxyfuel combustion,
preferential adsorption of CO2 on Mg2+. Indeed, infrared amongst others.
studies confirmed an end-on coordination mode for CO2, with
the increased ionic character of the Mg2+–O interaction 4.2.4.2. Interpenetration as a Strategy for Selective Adsorption
accounting for the high adsorption capacity.[112, 113, 129]
Since the dynamic separation capacity of framework Interpenetrated frameworks have potential for applica-
materials provides a more reliable measure of their separa- tion in gas separations processes since their pore sizes often
tion ability, breakthrough experiments were performed on lie in the range for industrially important gas separations, as
[Mg2(dobdc)] and [Ni2(dobdc)] using CO2/CH4 and CO2/N2 shown in Table 1. Their adsorption capacities however, are
mixtures.[113] Such measurements are performed by exposing often inferior to those observed for non-interpenetrated
the material to a mixed gas stream (e.g., 80 % CH4, 20 % structures.
CO2), and detecting the appearance or “breakthrough” of A molecular sieving affect was invoked to explain the
CO2. For both frameworks, quantitative separation of CO2 selective sorption of CO2 over CH4 in [Zn(adc)(Bpe)0.5]
from N2, and substantial retention of CO2 from mixtures with (adc2! = 4,4’-azobenzenedicarboxylate, Bpe = trans-bis(4-pyr-
CH4 was observed. A comparison of the dynamic separation idyl)ethylene), which comprises a triply interpenetrated cubic
capacity with that of a benchmark material such as Zeolite net with pores of 3.4 ! 3.4 #, and a surface area of
13X under the same conditions, revealed that the framework 100 m2 g!1.[116] The four-fold interpenetrating diamondoid
exhibited superior performance. framework [Ni(cyclam)2(mtb)] (cyclam = 1,4,8,11-tetraazacy-
Highly selective CO2 adsorption is also observed in cases clotetradecane), which is characterized by one-dimensional
where a material possesses a number of different potential channels with dimensions of 2.05 ! 2.05 #, and a BET surface
adsorption sites. For example, the nickel phosphonate frame- area of 141 m2 g!1 [89] also shows selective CO2 adsorption over
work [Ni2(pbmp)] (where pbmp4! = N,N’-piperazinebisme- N2 and CH4. In both cases, the origin of the selectivity was
thylenephosphonate) contains three crystallographically dis- ascribed to the smaller kinetic diameter of CO2 (relative to
tinct coordinatively unsaturated metal centers, as well as P=O the other two gases) and its enhanced interactions with the
groups from the organic linkers that project into the pores.[110] host framework due to its quadrupole moment.

6070 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

While adsorption measurements were performed using While high surface area frameworks are often sought for
single component sorption isotherms in the aforementioned gas adsorption to achieve maximum loadings, the penalty in
cases, the doubly interpenetrated framework [Zn(bdc)- many cases may be a lack of selectivity which is required for
(bpy)0.5] (MOF-508b) with one-dimensional micropores of gas separations applications. The flexible interpenetrated
4.0 ! 4.0 # was the first examined for the separation and framework shown in Figure 5 a is based on 4,4’-bipyridine and
capture of CO2 from binary CO2/N2 and CO2/CH4 and ternary the tetrahedral organic linker tetrakis[4-(carboxyphenyl)oxa-
CO2/CH4/N2 mixtures by fixed-bed adsorption.[117] Break- methyl]methane.[132] The material was characterized by a step
through measurements confirmed that CO2 was preferentially in the CO2 isotherm at ca. 10 bar, and a maximum adsorption
adsorbed over N2 and CH4 due to a molecular sieving effect. of 7.1 mmol g!1 at 30 bar. The capacity is well below that
The breakthrough fixed-bed adsorption of the binary and obtained for MOF-177 (35 mmol g!1) and activated carbon
ternary mixtures revealed selectivity factors of 3–4 for CO2/ (MAXSORB, 25 mmol g!1), but comparable to that of zeolite
CH4 and 4–6 for CO2/N2. These were found to improve with 13X (7.4 mmol g!1) at the same pressure.[88] At pressures in
increasing partial pressure and decreasing temperature, the range 0–1 bar relevant to flue streams, the framework
suggesting that further selectivity enhancements could be exhibited a significantly greater uptake of CO2 compared with
achieved at lower temperatures and higher pressures. N2, and at pressures up to 20 bar relevant to syngas
Preferential CO2 adsorption has been observed for separation, virtually no H2 adsorption was detected. The
frameworks which combine the properties of interpenetration mechanism for the selectivity enhancement was ascribed to
with the presence of coordinatively unsaturated sites.[118] The size/shape discrimination, however, it appears that the CO2
framework [Mg(tcpbda)] (H2tcpbda = N,N,N’,N’-tetrakis- molecules must also exhibit a preferential electrostatic
(4-carboxyphenyl)-biphenyl-4,4’-diamine) for example, interaction with the material, possibly with the oxygen
exhibits good CO2 adsorption compared with CH4 and N2.[118] atoms of the linker.
The observation of one or more distinct steps in the
4.2.4.3. Flexible Frameworks adsorption isotherms represents a key feature of flexible

In contrast to the rigid frameworks

discussed above which retain their
porosity upon adsorption and desorp-
tion, flexible and dynamic frameworks
collapse upon removal of guest solvent
molecules, but restore their porous
structures by adsorption of gas mole-
cules at high pressures.[131] The adsorp-
tion isotherms are typically character-
ized by a distinct step, at which point
the material “opens up” as gas mole-
cules enter the pores.[15] In some strik-
ing cases, almost no adsorption occurs
below a threshold or “gate-opening”
pressure for the non-porous structural
phase, until the pores of the material
become accessible and permit signifi-
cant adsorption. The desorption iso-
therms are also typically characterized
by hysteresis. The caveat must be added
that while flexible frameworks operat-
ing by a gate-opening mechanism
appear to be excellent candidates for
gas separations when comparing single
gas isotherms, such materials may not
perform well for gas mixtures. If one
component of a mixture triggers pore
opening through framework breathing,
then all components will be permitted
to enter the structure, thereby negating Figure 5. Examples of flexible metal–organic frameworks. a) The tetrahedral organic linker
the perceived selectivity. This is a tetrakis[4-(carboxyphenyl)oxamethyl]methane and the flexible interpenetrated framework based
upon this ligand with Zn2 paddlewheel cluster units and 4,4’-bipyridine.[132] In one portion of the
serious potential drawback in the utility
structure, blue, red, and gray sticks represent the N, O, and C atoms, respectively, while the
of such breathing frameworks which interpenetrated framework is shown in cyan. Hydrogen atoms are omitted for clarity. b) The bptc
underscores the need for measure- ligand and a schematic diagram of the [(Ni2L)(bptc)] framework, where Ni2L represents the ethyl-
ments using gas mixtures. or butyl-bridged bismacrocyclic complexes shown.[122]

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6071
Reviews D. M. D’Alessandro et al.

frameworks. The MIL-53 series has been extensively inves- [M(bpy)2(BF4)2] (M = Cu, Ni; bpy = 4,4’-bipyridine; BF4! =
tigated for high-pressure CO2 and CH4 adsorption.[119, 133–137] tetrafluoroborate)[139] and [M(bpy)2(CF3SO3)2] (M = Co, Cu;
The framework [Cr(OH)(bdc)] (MIL-53(Cr)), for example, bpy = 4,4’-bipyridine).[139, 140] Here, each octahedral metal
exhibits a small CO2 uptake below 5 bar, however, a distinct center is coordinated by four N atoms from bpy ligands in
step is observed in the adsorption isotherm as the structure equatorial positions and two BF4! or CF3SO3! anions in axial
opens up to provide for a higher gas uptake.[119] This step is positions to form two-dimensional square-grid sheets. In the
absent in the corresponding isotherm for CH4 which exhibits a case of [Cu(bpy)2(BF4)2], negligible CO2 adsorption occurs up
gradual uptake over the same pressure range. On the basis of to a pressure of 0.7 bar at 298 K, at which point the interlayer
diffuse reflectance infrared spectroscopy,[133] microcalorime- spacing increases to accommodate the gas molecules. By
try, and Monte Carlo simulations,[136] the phenomenon was comparison, relatively larger, non-polar CH4 molecules
ascribed to a breathing effect in the flexible porous solid require a higher pressure of 4.5 bar. Since the “gate opening
which occurs in the presence of CO2 (but not CH4).[68] An in- pressure” can be tuned with temperature, a temperature
depth study of MIL-53(Cr) upon adsorption of binary CO2/ adsorption swing approach could be employed to achieve the
CH4 mixtures showed that the breathing effect is mainly separation on an industrial platform.
controlled by the partial pressure of CO2, with CH4 having Mesh-adjustable molecular sieves (MAMS) constitute
only a minor influence on the extent of the transition between another class of materials for gas separations and are based
the open and closed forms.[138] Breakthrough experiments on temperature induced gating phenomena.[141] With the
revealed that the CO2/CH4 selectivity decreases from 15 to 4 caveat added previously about measurements on single gas
as the CO2 partial pressure increases from 0–0.25 (for a 75:25 adsorption isotherms for flexible frameworks, MAMS repre-
mixture). This low selectivity for CO2 at high pressure, in sent a case where the dynamics of substituents at pore
addition to the large hysteresis is such that the desorption openings allow some molecules to pass but not others. On this
occurs at a lower pressure than is desirable for a pressure basis, these materials are likely to be more viable for gas
swing adsorption application, for example. separations applications relative to many of the aforemen-
The presence of water leads to a drastic increase in the CO2 tioned flexible frameworks.
uptake relative to CH4 in the MIL-53(Cr) framework.[120] The MAMS are represented by the isostructural series based
adsorption of CO2 on the hydrated framework occurred to a on 4’-tert-butyl-biphenyl-3,5-dicarboxylic acid and M2-
small extent only at pressures up to 10 bar, while a distinct (COO!)4 paddlewheel units (e.g., M = Cu2+, Co2+, and
uptake occurred over a broad pressure range of 12–18 bar as the Zn2+).[141] The graphitic structures consist of hydrophilic
hydrated structure opens to accommodate gas molecules. It was channels, in which the tert-butyl groups are connected through
inferred that the presence of water within the framework weak van der Waals interactions. It can be inferred from their
impedes the entry of CO2 molecules at low pressures prior to structures that the hydrophobic cages should be the main
the total opening of the structure at higher pressures. The storage space for gas molecules, while the hydrophilic
hydrated framework exhibited negligible adsorption of CH4 due channels should act as passages for gas molecules to enter
to the repulsive effect of the free water in the pores that impede the cages through the tert-butyl groups that serve as gates. The
or block entry to the non-polar molecules. From a practical pore size can be modulated in a linear fashion by varying the
perspective, the results suggest that a pre-adsorber column, temperature, giving rise to an infinite number of mesh sizes
which is often employed for water removal prior to CO2/CH4 between 2.9 and 5.0 # (covering the range of most industrially
separation, is unnecessary given the enhanced separation important gas separations).[141] As the temperature is raised,
performance for the hydrated flexible framework material. the amplitude of thermal vibration of the tert-butyl group also
From these examples, it is evident that the breathing increases, leading to an aperture enlargement which permits
behavior of flexible frameworks is highly dependent on CO2 and N2 to permeate, whilst blocking the relatively larger
cooperative effects which arise from the favorable interaction CH4 molecules. This selective adsorption has been observed
between the quadrupolar CO2 molecule with the polar surface under cryogenic conditions, however the development of
of the framework. For example, the frameworks MAMS for gas separations at ambient temperatures is being
[M(F-pymo)2] (M = Co, Zn, F-pymo = 5-fluoropyrimidin-2- pursued. Strategies that are being considered in this context
olate) were shown to exhibit highly selective adsorption of to tune the mesh sizes and the temperature-dependent
CO2 over CH4 due to preferential CO2 interactions with the properties include changing the functional groups of a
fluorine groups.[121] Highly selective CO2 capture (over N2, H2, linker,[142] adjusting the length and bulkiness of the linkers,[141]
and CH4) has also been observed in the flexible frameworks taking advantage of interpenetration,[143] introducing coordi-
[(Ni2L)(bptc)], where Ni2L represents the ethyl- (L1) or natively unsaturated metal centers,[114] and making use of
butyl-bridged (L2) Ni2+ bismacrocyclic complexes and bptc is postsynthetic functionalization strategies.[144] A recent review
1,1’-biphenyl-3,3’,5,5’-tetracarboxylate (Figure 5 b).[122] The article[16] provides a comprehensive summary of the materials
nature and length of the alkyl linker was also shown to and their ability for application in CO2 separation from N2,
modulate the degree of structural flexibility, leading to CH4, and H2 gases, amongst others.
different gate opening pressures for the ethyl- and butyl-
bridged frameworks. 4.2.4.4. Surface-Functionalized Frameworks
As well as pressure, external stimuli such as temperature
can be used to induce gating effects, particularly in frame- The grafting of functional groups with a high affinity for
works which possess graphitic-type structures such as CO2 onto the surfaces of porous materials through ligand

6072 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

modification, or coordination to unsaturated metal centers

has been employed as a strategy to enhance the capacity and
selectivity for CO2 adsorption. This approach has analogies
with other functionalized solid adsorbents such as amine-
grafted silicas; however, the crystalline nature of metal–
organic frameworks provides for a molecular level of control
in tuning the gas separations properties.
Frameworks containing open metal sites have been
selectively grafted with molecules that have a high affinity
for CO2. For example, [Zn2(bttb)(py-CF3)2] (where bttb4! =
4,4’,4’’,4’’’-benzene-1,2,4,5-tetrayltetrabenzoate) incorporates
the highly polar pyridine derivative py-CF3 (4-(trifluorome-
thyl)pyridine) shown in Figure 6 a.[123] While grafting reduces
the surface area of the framework to 300 m2 g!1 relative to the
non-functionalized form, [Zn2(bttb)] (1370 m2 g!1), function-
alization leads to a remarkable enhancement in the CO2
uptake compared with N2 in the low pressure region
(0–1 bar) relevant to flue gas separations. Indeed, the CO2/
N2 selectivities calculated from binary mixtures using IAST
for the py-CF3 modified framework are higher than the
experimental selectivities for zeolite 13X[51] and activated
carbon.[128]
In contrast to the number of reported frameworks
possessing Lewis acid sites which originate from coordina-
tively unsaturated metal centers, Lewis base functionalized
frameworks are seldom observed owing to the tendency of
Lewis basic sites to coordinate metals during framework
synthesis. The potential for enhanced CO2 adsorption due to
acid–base interactions in the latter case led to the discovery of
the sulfone-functionalized material [H3O][Zn7(m3-OH)3-
(bbs)6] (UoC-1, bbsH2 = 4,4’-bibenzoic acid-2,2’-sulfone, Fig-
ure 6 b).[124] The material exhibited reversible and selective
CO2 adsorption over CH4, N2 and H2 due to a preferential
interaction between CO2 and the polar sulfone groups.
In view of the affinity of amines towards CO2, amine-
functionalized ligands have been incorporated into a number
of frameworks in an attempt to enhance the selectivity and
adsorption. An enhancement in the CO2 uptake relative to
CH4 was observed for the flexible amino-MIL-53(Al) frame-
work [Al(OH)(NH2bdc)] compared with the parent MIL-
53(Al) framework.[125] Pulse chromatographic measurements Figure 6. Examples of functionalized metal–organic frameworks.
revealed a separation factor of 60 for the functionalized a) 4,4’,4’’,4’’’-Benzene-1,2,4,5-tetrayltetrabenzoic acid and a portion of
material at low surface coverage, compared with ca. 5 for the the framework [Zn2(L)(py-CF3)2] as a space-filling model, where L is the
parent framework. The magnitude of the zero-coverage tetra-anionic form of the ligand.[123] The inset shows the interaction
adsorption enthalpy for CO2 also increases from !20.1 to between one of the py-CF3 ligands and a Zn2 cluster unit. Yellow, red,
blue, and gray spheres represent Zn, O, N, and C atoms, respectively.
!38.4 kJ mol!1 upon functionalization. The incorporation of
b) The ligand 4,4’-bibenzoic acid-2,2’-sulfone.[124] c) A portion of the
dangling amino groups reduces the number of apolar sites framework HCu[(Cu4Cl)3(BTTri)8(en)5] showing one of the en-function-
relative to the parent framework, resulting in negligible CH4 alized Cu2+ sites and its interaction with a CO2 molecule.[127] Purple,
adsorption at pressures below 2 bar in the former case. blue, green, and gray spheres represent Cu, Cl, N, and C atoms,
Clearly, the presence of amino and OH groups drastically respectively. Non-framework H and Cu atoms, and hydrogen atoms
enhances the selectivity of the framework towards CO2 are omitted for clarity.
adsorption. Similar enhancements have been reported in the
amine-functionalized frameworks [Ni2(NH2bdc)2(dabco)]
(dabco = 1,4-diazabicyclo[2.2.2]octane) and [In(OH)- (BTTri)8] (H3BTTri = 1,3,5-tri(1H-1,2,3-triazol-4-yl)benzene)
(NH2bdc)], relative to their non-functionalized analogs.[126] contains exposed metal sites which can be reacted postsyn-
Alkylamine-functionalized frameworks have also been thetically with diamines such as ethylenediamine (en).[127] As
shown to enhance the selectivity for CO2 adsorption, espe- shown in Figure 6 c, one amine group in HCu[(Cu4Cl)3-
cially at the low pressures relevant to flue gas separation. For (BTTri)8(en)5] is linked to a coordinatively unsaturated
example, the air- and water-stable framework HCu[(Cu4Cl)3- metal site, while the other is available for selective uptake

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6073
Reviews D. M. D’Alessandro et al.

of CO2. Despite a reduced surface area compared with the in the pores of microporous materials. In these simulations,
parent framework, the en-functionalized material exhibited a one assumes that the interactions between the atoms can be
higher uptake of CO2 at very low pressures (up to 0.01 bar) described accurately using a (classical) force field. The force
compared with the non-grafted material, which gave rise to an fields together with the crystal structure of the frameworks
initial enthalpy of adsorption at zero coverage of are used in molecular dynamics or Monte Carlo simulations to
!90 kJ mol!1 (compared with !21 kJ mol!1 for the non- compute the thermodynamic and transport properties.[150] For
grafted material). This chemisorption interaction was the zeolites, these methods have been very successful in predict-
strongest reported to date in a framework material. There is a ing the adsorption and diffusion properties of many different
need to probe the mechanism of the alkylamine–CO2 types of molecules,[151] and recently these methods have been
chemisorption interactions in the presence and absence of extended to study the adsorption of gases such as CO2 and
water, to search for the formation of carbamate and/or CH4 in metal–organic and covalent–organic frameworks.
hydrogen carbonate species (such as those shown in
Scheme 1). 4.2.5.1. Force Fields
Clearly, extensive possibilities can be envisaged to tune
the selectivity and heat of adsorption in amine-functionalized Compared to zeolites, the structures of metal–organic
frameworks by incorporating secondary, tertiary or sterically frameworks can be controlled to a much higher degree
hindered amines at open metal coordination sites, or on the through variations in the type and nature of the organic
bridging ligands. Such frameworks should overcome some linkers and the metal centers. From a computational perspec-
limitations of other CO2 capture materials, most notably: tive, it is important to describe the entire range of interactions
1) the structural ambiguities present in hyperbranched ami- between all atoms in a given system using force fields. These
nosilica materials produced by surface polymerization[57] and force fields have three components: one that describes the
2) the lack of stability over multiple cycles for amine- interactions between the atoms of the frameworks (or the
impregnated silicas and zeolites.[54] other materials) as well as one that describes the interactions
It is important to note that studies on the viability of between the adsorbed gases, and the interactions between the
metal–organic frameworks as CO2 capture materials under adsorbed molecules and the atoms of the framework.
realistic industrial conditions have received relatively limited Computationally, it is very convenient to assume a rigid
attention to date. While the high robustness and stability of crystal structure for the framework. Indeed, if this assumption
zeolitic imidazolate frameworks has been established,[99] the holds, there is no need for a force field describing the
majority of metal–organic frameworks exhibit comparatively interactions between the atoms of the framework. For many
lower stabilities, and their sensitivity to chemical degradation systems, this assumption provides a reasonable description of
in the presence of air and water, as well as their sensitivity to the adsorption isotherms; however, there are several exam-
other impurities in gas streams (see Table 1) poses a major ples in which the structure of the material changes[152] upon
impediment to their practical application. Ultimately, the adsorption of, for example, water[153] or carbon dioxide,[119] or
integration of metal–organic frameworks into practical CO2 by changing the temperature.[154] Including framework flexi-
capture processes requires a serious assessment of their bility slows the computations by one to two orders of
performance as membranes for gas separation, or in large magnitude. Recently, some progress has been made on the
adsorbent beds where pressure or temperature swing adsorp- development of force fields and methods which describe
tion approaches may be employed.[8] Relatively few exper- structural changes in the materials.[135, 155–158]
imental studies have probed the kinetics of gases diffusing in If the structure of a given framework is known, accurate
MOFs,[145–149] although a growing number of computational force fields which describe the gas–gas and gas–framework
investigations have provided important insights. Low gas interactions must be obtained. The gas–gas interactions are
permeance could be a problem, in particular for MOFs with usually optimized to give an accurate description of the
small, or dynamically restricted (in the case of MAMS) pore vapor–liquid coexistence properties,[159, 160] which provides a
openings. very sensitive test for the accuracy of these force fields.[161]
The gas–framework interactions pose a significantly greater
4.2.5. Simulations of CO2 Capture in Metal–Organic Frameworks challenge to obtain accurately. The typical force fields
and Related Materials describing the interactions include dispersive interaction
and Coulomb interactions. Polarization is also an important
We have highlighted the most recent advances in new contribution;[162] however, its inclusion raises the expense of
materials for CO2 capture. One of the most exciting aspects is the calculations and most studies include the effect indirectly
the nearly infinite array of structures that can be prepared; through the introduction of effective parameters for the
however, it may be impossible, or at least very impractical to Coulomb and dispersive interactions. The parameters of these
synthesize and test all of the conceivable structures. The potentials are either fitted to experimental data or to
development of computational techniques that can deal with quantum calculations.[137, 163] In the case of [Zn4O(bdc)3], for
this eruption of novel materials is critical, and must focus on: example, a computational model which included both Len-
1) describing the thermodynamic and transport properties nard-Jones and Coulombic interactions was found to be
accurately and 2) screening for optimal capture materials. essential to adequately model the adsorption behavior.[91]
Molecular simulation is one of the most promising Both approaches have serious shortcomings. Experimen-
methods to predict the properties of the molecules adsorbed tal data can be fit very accurately for one given system, but

6074 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

this does not guarantee that the potentials can be transferred chloro-, bromo-, and dibromobenzene) leads to weaker
to other systems. Meanwhile, quantum calculations suffer interactions with CO2, while methyl substituents (such as
from the fact that both the dispersive interactions[164] as well mono-, di-, and tetramethylbenzene) enhance these interac-
as charge distribution[135] are very difficult to obtain, and tions.
different quantum methods give remarkably different While most studies have been focused on CO2/N2
results.[160] Despite these limitations, many of the calculations mixtures, the influence of water on the capacity and
performed to date provide a very reasonable description of selectivity has also been considered.[108] One would expect
the experimental adsorption isotherms.[165] that water would compete with CO2 for the “most attractive”
sites, thereby reducing both the selectivity and capacity. The
4.2.5.2. Simulation of the Adsorption and Diffusion simulations, however, showed that water molecules coordi-
nated to open metal sites in [Cu3(btc)2] significantly increase
Here, we focus on those simulations that relate specifically CO2 adsorption. This effect has subsequently been experi-
to CO2 capture and for more general reviews we refer the mentally confirmed.[108]
reader to some excellent literature articles.[166, 167] A significant Clearly, there is tremendous scope for developments in
fraction of the computational work has been directed towards computational methods for screening CO2 capture in frame-
the validation of the force fields.[160, 165] The focus of most work materials.[173] A significant and growing research effort
studies is on pure component data, as for these systems is also underway on the incorporation of frameworks into
experimental data are available to test the force fields.[167] membranes, and the simulation of the permeation selectivities
Often the pure component isotherms exhibit steps, of for such hybrid systems.
which developing a better understanding will lead to The first information on the performance of frameworks
improved force fields. For frameworks, flexibility is an as membranes for gas separations was obtained from a case
essential factor in understanding these steps,[119, 168] however, study on the separation of CO2/N2[176–178] and CO2/CH4[177]
this is not the only contribution. For example, the step in the mixtures by a hypothetical membrane consisting of [Zn4O-
adsorption isotherm of CO2 in MOF-177 has been explained (bdc)3] crystals. In addition to the importance of considering
in terms of a condensation of CO2 in the pores of the both the adsorption and diffusion of gas molecules, binary
materials.[88, 91] The simulations showed excellent agreement (rather than single component) adsorption isotherms were
with the experimental data using a rigid framework. A found to be crucial in order to adequately predict membrane
stepped isotherm for CO2 adsorption was also predicted in the performance in practical applications. IAST theory was
rigid covalent–organic frameworks COF-8 and COF-12.[169] employed to calculate the binary adsorption isotherms for
Different materials including zeolites, metal–organic and an equimolar CO2/CH4 mixture, and on this basis, the
covalent–organic frameworks have also been screened for adsorption selectivity was found to strongly favor CO2 over
their maximum adsorption capacities for CO2.[65, 170] An CH4.
extensive series of Monte Carlo simulations on zeolites, These simulations were subsequently extended to exam-
metal–organic frameworks and covalent–organic frameworks ine the performance of [Cu3(btc)2] membranes for CO2/CH4
has shown that stepped isotherms arise from cluster formation and CO2/H2 separations.[179] A comparison of the ideal
of guest molecules such as CO2, CH4 and Ar below a critical selectivities (calculated from single gas adsorption isotherms)
temperature.[171] In particular, molecular cluster phenomena with the mixture selectivity demonstrated that the values
strongly influence the adsorption and diffusion characteristics differ significantly due to multi-component effects in the
of porous materials with pore sizes larger than 0.75 nm, and diffusing mixtures. The strongly adsorbing component (i.e.,
this strategy may have considerable potential for enhancing CO2 in a CO2/H2 mixture) reduces both the concentration
the CO2/CH4 separation selectivities.[172] gradient across the membrane for the weakly adsorbed
An important application of molecular simulation is to component and the diffusion rate of the more mobile species
screen materials for their optimal separation selectivity.[24, 173] (i.e., H2). The results indicate that the selection of frameworks
The extension of pure component data to gas mixtures is for membrane-based gas separations must focus on both the
relatively straightforward, and has led to many predictions of adsorption and diffusion selectivities of the material.[24, 179]
separation selectivities.[165, 174, 175] Unfortunately, experimental Such simulations are idealized in the sense that they
data are lacking for most of these studies, and it is difficult to describe membranes that are constructed from defect-free
assess directly the reliability of the predictions. From our crystalline materials. Considering the very large number of
experience with the adsorption of hydrocarbons, it appears different frameworks that could be employed to construct a
that once the pure component isotherms are accurately membrane, and the time required for experimental measure-
described, the molecular simulations provide a very reason- ments using gas mixtures under a range of conditions,
able prediction for the mixture isotherms.[151] computational simulations may prove to be critical in
A very high selectivity towards CO2 was predicted in determining the most promising structures.
zeolite-like metal–organic frameworks that possess the RHO
structure and contain Na+ ions.[66] The effect of substitution of
the linker molecules has also been probed using quantum 4.3. Membranes
calculations, and showed that a careful design of the linker
can tune the interactions with CO2.[163] For example, substi- Membrane separation technologies are projected to attain
tution of benzene molecules with halogens (e.g., tetrafluoro-, a high efficiency for CO2 capture owing to their selective

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6075
Reviews D. M. D’Alessandro et al.

extraction of CO2 from mixed gas streams, their low energy stability and high hydrogen flux rates. Indeed, a common
requirements and the flexibility in their possible configura- disadvantage of a number of the aforementioned membranes
tions in industrial plants relative to conventional amine is their lack of stability under the reforming environment
absorber/desorber columns.[18] The constituent materials are present in processing plants (which contains steam and
typically classified as: inorganic (ceramic, metal oxide, hydrogen sulfide).
metallic, molecular sieves, thin layer of Pd on ceramic or Inorganic membranes, including Zeolite T, DDR, silicate-
porous alumina cylinder as support, metal–organic frame- 1 and SAPO-34, have also been investigated for high pressure
works) or organic (cellulose acetate, polymers such as natural gas sweetening, and can exhibit high CO2/CH4
polysulfone, polyamide, polyimide, cross-linked polydime- selectivities due to a combination of differences in diffusivity
thylphenyleneoxide, hollow fibers with high surface area to and competitive adsorption.[184] Further improvements in the
volume ratios of 1500-2000 m2 m!3). selectivity for SAPO-34, for example, were achieved by
A number of mechanisms for membrane separation have synthesizing the material on porous stainless steel tubes, or on
been identified, with the most important being solution a-Al2O3 porous supports. In the latter case, a CO2/CH4
diffusion and molecular sieving.[180] These factors are formally separation selectivity higher than 170 at 295 K and a feed
defined by the gas permeability (transport factor) and pressure of 2.2 atm was achieved, and effective separation
selectivity (separation factor), respectively, and are inversely could still be observed up to 16.7 atm.[184]
related. This trade-off is widely characterized graphically in a To provide an additional level of selectivity beyond that
“Robeson Plot”,[181] from which an upper bound to membrane which may be obtained solely from a molecular sieving effect
performance is observed. The goal of membrane studies has in inorganic membranes, hybrid membrane/chemical absorp-
thus been to reach this maximum to achieve both high tion processes have been considered for low pressure flue gas
permeability and high selectivity, in addition to adequate separations. Since the gas separation mechanism through a
robustness and material lifetime. pure porous support is based mainly on size differences of the
Inorganic membranes are comprised of either porous or permeating molecules, the covalent functionalization of the
non-porous materials, and have the advantage of a high pore walls has been suggested as a strategy to increase the
temperature operation which is relevant for precombustion CO2 selectivity. Examples include inorganic membranes such
separation.[180, 182] Hydrogen transport membranes for pre- as zeolites, or a ceramic support such as Al2O3, onto which a
combustion separation allow one-step reforming, or a single porous separating layer such as silica, alumina or zirconia is
intermediate water-gas shift reaction, and thus permit simul- attached, followed by a functional layer such as MgO which
taneous hydrogen separation (the permeate) and CO2 cap- has a high chemical affinity for CO2. CO2/N2 selectivities as
ture. Since the retentate consists of a concentrated, high- high as 120 at 350 8C have been reported (a value which
pressure CO2 stream, the compression energy for transport approaches Robeson"s upper bound).[180] Hybrid membranes
and storage is reduced. Hydrogen transport membranes are are advantageous in the respect that their degradation is
often based on microporous inorganic materials including minimized through the prevention of oxygen contact with the
zeolites, palladium alloy tubes, or ceramics such as inorganic absorbent.
perovskite oxides.[183] Membranes consisting of metal–organic frameworks
The mechanisms for separation in inorganic membranes immobilized in various supports also show exciting prospects
are typically based upon adsorption selectivity and surface for gas separations. A [Cu3(btc)2] membrane has been
diffusion, which give rise to relatively low separation factors prepared on a copper net,[185] and porous alumina substrates
(e.g., H2/CO2 separation factor of 15). Commercially avail- have been used as supports for [Zn4O(bdc)3][186] and [Zn-
able g-alumina and silica microporous membranes exhibit (bim)2] (ZIF-7, bim! = benzimidazolate).[187] In the latter
significantly higher separation factors (up to 40), however, case, the membrane exhibited an exceptional H2/CO2 sepa-
these are dependent on the stability of the membrane pore ration ability, as well as air, water and thermal stability (up to
size, which is adversely affected by steam in the feed streams. 500 8C). Computational simulation is proving to be an
Dense ceramic membranes based on inorganic perovskite invaluable technique to screen the performance of metal–
oxides (also known as proton conducting membranes) require organic frameworks for membrane separations applications.
high operational temperatures to achieve practical hydrogen These are relatively more difficult than simulations for
flux rates (above 800 8C). Molecular sieves such as zeolites adsorption isotherms, among other reasons because the
(aluminosilicate compositions) or non-zeolites (aluminophos- flexibility of the framework should be taken into account.
phates and silica) have also been considered as inorganic Inorganic ion transport membranes have been proposed
membranes. In these cases, the separations are based on for the separation of oxygen from gas mixtures as a
kinetic size discrimination within the channels of the porous potentially more viable alternative to cryogenic oxygen
structures. removal in oxyfuel combustion processes, for example.[35]
Palladium-based composite membranes have been used The membrane itself uses conductors which are composed
for precombustion applications and are known for their high of mixed-metal oxides at high excitation temperatures of 800–
hydrogen selectivity and permeability over other gases in the 900 8C for both electron and oxygen ion conductivity. Such
temperature range 300–600 8C. Whilst palladium alloy tubes membranes are currently being investigated as part of Zero
have been available for several decades, they are deemed too Emission Power Plant projects, since they enable inherent
expensive for CCS applications due to the substantial thick- CO2 capture directly in the combustion chamber of the gas
ness of the membrane which is required to achieve structural turbine. A major advantage is that the process is driven by the

6076 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

permanent withdrawal of oxygen on the permeate side as a polymeric membranes containing amines in crosslinked
consequence of combustion. polyvinyl alcohol.[196]
Polymeric membranes constitute a second major class Further possibilities for novel polymer membranes
which are widely employed commercially for natural gas involve the creation of hybrid assemblies. For example, the
sweetening (CO2/CH4 separation).[180] They are typically combination of cyclopeptides and synthetic polymers could
based on cellulose acetate and derivatives thereof, whose be used to generate hybrid biomolecular materials which may
structures are tailored by varying the glass transition temper- provide ideal internal structures for gas separation. Mixed-
ature. These polymers are highly susceptible to degradation matrix membranes have also recently been considered, by
through plasticization and compaction under feed stream combining the advantageous features of inorganic and
conditions, and are not deemed to be viable for flue gas polymeric membranes in an effort to overcome the limitations
applications. Polysulfone and polyimide polymer-based mem- of each class.[180] Such membranes are characterized by a
branes are amongst the most widely reported membranes in hetereogeneous gas separation layer comprising discrete
the patent literature. They exhibit excellent advantages over inorganic particles dispersed in a continuous polymeric
cellulose acetate membranes, including improved permeabil- phase. For example, metal–organic frameworks such as
ities and selectivities as well as thermal, chemical and [Cu(bpy)(SiF6)] have been incorporated into Matrimid poly-
placticization resistance.[180] Recently, a new class of micro- mers to form free standing mixed-matrix membranes for CO2/
porous polymeric structures known as polymers of intrinsic CH4 separations.[197] AMH-3/PBI membranes which consist of
microporosity have been reported.[188] These materials are a nanocomposite of the layered silicate AMH-3 and poly-
comprised of randomly contorted rod-like structures which benzimidazole (PBI) have also been prepared, and exhibit H2/
form intrinsic cavities within the membrane, and thus exhibit CO2 selectivities that are doubled compared with the polymer
high free volumes and higher gas solubilities compared with itself.[198]
conventional molecular sieves. Given that a number of Facilitated liquid membranes constitute another impor-
extensive review articles detailing the advantages of poly- tant class of materials which incorporate a number of liquid
meric membranes have appeared in recent years,[14, 18, 180, 189, 190] absorption stages to achieve CO2 separation from flue gas
we focus here instead on emerging concepts in the field of streams.[183] The separations are facilitated by preferential
polymer-based membrane materials. reversible chemical reactions between CO2 and “carriers”
Membrane selectivity currently depends on the relative such as carbonates, amines and molten salt hydrates,[199]
sizes of gas molecules and the relative condensability of the carbonic anhydrase or ionic liquids[200] which are dissolved
gas molecules. From this perspective, two major challenges in the porous membrane. For example, facilitated transport
exist in order for polymeric membranes to become a viable ionic liquid membranes which consist of an amine-function-
route for gas separation. Firstly, the membranes must possess alized ionic liquid encapsulated in a polymeric support (such
relatively small pore sizes to selectively separate the gas as polysulfones or cross-linked Nylon 66) exhibit exception-
molecules of interest. Novel types of assemblies (known as a ally high CO2/H2 selectivities (ca. 10).[201, 202] The cross-mem-
functionalized polymeric composite membranes) have thus brane transport is driven by the difference in the partial
been proposed, and consist of a base porous support and a pressure of CO2 across the membrane,[183] which is higher on
thin layer of permselective materials (typically 0.2–1 mm the side that contacts the flue gas than on the other. This
thick).[191] Although nanoporous polymer thin films can be pressure gradient can be obtained by pressurizing the gas on
routinely prepared, the ability to obtain porous films with one side of the membrane, and/or applying a vacuum on the
sub-nanometer features remains a significant challenge. other. Significantly, the pressure differential supplies the
Secondly, due to the limited heterogeneity of polymers, the energy for separation and is the key to the low energy penalty
spatial arrangement of each individual monomer cannot be of the process.
specifically controlled. Fine-tuning the interactions between Clearly, membranes represent a promising technology for
the gases and polymeric membranes is required to increase gas separation, however, they suffer a number of drawbacks,
the selectivity between similarly sized molecules. particularly with regard to CO2 capture from flue gas. In this
Hypercrosslinked networks hold great promise as new case, the low CO2 partial pressure provides a minimal driving
materials for polymer membranes. Such networks are com- force for gas separation, which creates an energy penalty due
prised of aromatic rings which are connected through linking to the need for compression of the feed gas.[203] Membrane
groups of well-defined lengths to form a porous framework. materials also suffer from a decrease in permeability over
Modification of the network surface could be achieved time due to particulate deposition on the surface.
through functionalization of the precursor polymer and/or
the hypercrosslinked layer.[192] The validity of this concept has
already been demonstrated by the development of hydrogen 4.4. CO2 Fixation: Mineralization, Fuels, and Biofixation
storage materials[192–194] based on hypercrosslinked polyani-
line, which was synthesized by reaction of the linear Chemical fixation through the conversion of CO2 into
polyaniline precursor with a difunctional reagent such as fuels, commodity chemicals, construction materials, or min-
diiodobenzene.[195] A number of other precursor materials eral carbonates represents another promising alternative for
could be explored to tune the pore sizes and polarities. A CO2 capture which has been described in detail in a recent
similar concept has been employed to produce CO2 selective article.[19] The challenge here lies in the fact that CO2 is a
highly stable compound containing a low amount of chemical

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6077
Reviews D. M. D’Alessandro et al.

energy, such that natural conversion processes may be slow have the potential to be less energy intensive and voluminous
and inefficient as a result. than absorption processes. Interestingly, the use of additives
A recently disclosed method for direct mineralization including tetrahydrofuran or propane has also been shown to
involves the production of cement through the reaction of lower the equilibrium pressure required for hydrate forma-
CO2 in flue gas with calcium and magnesium in seawater.[8] tion from CO2/H2 mixtures, without affecting the kinetics and
The process claims to react 0.5 tons of CO2 per ton of cement separation efficiency.[206]
produced, which compares with one ton of CO2 for each ton In chemical looping combustion the combustion step is
of cement produced using current processes. separated into oxidation and reduction reactions, and a
CO2 may also be used to produce fuels such as methanol, transition metal oxide (e.g., of Mn, Fe, Co, or Ni) is employed
formic acid, dimethyl carbonate, methyl formate and higher as an oxygen carrier to circulate between the two reac-
hydrocarbons, as well as polymeric materials and pharma- tors.[12, 46] The oxide particles react with a fuel in a fluidized
ceutical chemicals.[19] Given the sheer magnitude of CO2 bed reactor, producing solid metal particles and a mixture of
emissions and the need for effective catalysts, however, it CO2 and H2O. The reduced metal oxide is subsequently
appears unlikely that this approach will make a major transferred to an air reactor where the metal is oxidized. The
contribution towards reducing CO2 emissions in the near- outlet gas in this case consists of nitrogen and a reduced
term. amount of oxygen. The net chemical reaction for the two
A “sunshine to petrol” approach has recently been reactors is identical to that for normal combustion, with the
proposed, by harnessing photosynthetic energy to recycle advantage that the CO2 is essentially pure after its separation
CO2 into biofuels by mimicking the natural processes of from water. No additional energy demands or costly external
photosynthesis.[204] This biological fixation method uses equipment are therefore required for capture.
microalgae (microscopic aquatic plants that carry out photo-
synthesis) to produce renewable transportation fuels, while
also removing CO2 from large point sources using either open 5. Future Prospects
ponds or enclosed systems such as photobioreactors. Photo-
synthetic CO2 mitigation is advantageous in that it does not Clearly, no unique solution exists currently to solve the
require pure CO2, and does not incur costs for separation, problem of CO2 capture, and this complex challenge will
capture and compression of CO2 gas. almost certainly require the integration of several technology
options. This review article has sought to highlight the
challenges for CO2 separation methods which have the
4.5. Other Promising Capture Technologies greatest likelihood of reducing CO2 emissions to the atmos-
phere, namely postcombustion (low pressure, predominantly
The final category of carbon capture technologies consid- CO2/N2 separation), precombustion (high pressure, predom-
ered here comprises novel concepts based on techniques at inantly CO2/H2 separation) capture and natural gas sweet-
the pilot or laboratory stage of development. Processes that ening (predominantly CO2/CH4 separation). Importantly, the
use chemical looping, or hydration to separate CO2 are requirements for capture materials vary beyond those dis-
examples of such technologies. cussed here depending on the specific technology and stage in
A new approach for CO2 capture involves clathrate or gas a particular process at which CO2 capture occurs. For
hydrate crystallization and is applicable to both post- and example, the selectivity may be critical in some applications,
precombustion capture from flue gas or synthesis gas, but less so in others, tolerance to other components in the gas
respectively.[205, 206] The process relies on the ability of water stream such as water and H2S may or may not be required,
to form non-stoichiometric crystalline compounds in the and the long term chemical and mechanical stability may be
presence of CO2, N2, O2 and H2, as well as natural gas more or less important.
components at high pressures (10–70 atm) and low temper- While improvements to industrial processes and reduc-
atures (near 0 8C). The gas molecules are occluded within a tions in the plant footprints will make some contribution to
network of cavities which are formed by a hydrogen bonded the capture problem, the key factor which underlies signifi-
network of water molecules. A CO2/N2 or CO2/H2 mixture cant advancements lies in improved materials that perform
contacts water at a suitable temperature and pressure to form the separations. The results of research efforts in this area will
hydrate crystals, which are separated and decomposed to have widespread implications not just for CO2 sequestration,
create a CO2-rich stream. The preferential incorporation of but other gas separations, as well as solar-to-fuels conversions,
CO2 over the other gases into the hydrate crystal phase arises H2 production, etc. Although outside the scope of the present
from the difference between the hydrate formation pressure article, the direct capture of CO2 from ambient air represents
for CO2 relative to N2 or H2. another emerging technology option deserving of further
The high pressure and low temperature requirements for research effort.[25]
CO2 capture through hydrate formation impose significant The production of hybrid materials also holds great
challenges for the implementation of the technology.[14] promise. For example, metal–organic frameworks could be
Hybrid separations processes, which combine the advantages closely integrated with hydrophobic polymers to produce
of hydrate crystallization with membrane technologies have block co-polymers which prohibit the permeation of water.
thus been proposed for high pressure CO2/H2 separations Serious advantages over fixed-bed adsorption methods are
relevant to IGCC plants.[207] Such membrane-based processes also expected for the application of metal–organic frame-

6078 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

works to gas separations if reliable methods can be developed of transport and adsorption properties for crystalline materi-
for integrating these free-flowing powder materials into als is a less complex issue than that for amorphous polymer
membranes.[14] membranes.[151, 208] Hence, developing molecular models and
With respect to new materials, the key scientific chal- force fields based on the detailed structures is a prerequisite
lenges are the development of a level of molecular control, for the successful prediction of thermodynamic and transport
and the development of modern characterization and com- properties of new materials.
putational methods that will support, guide and provide Clearly, the combination of material synthesis, character-
further refinement to the most promising structures. Charac- ization, and computation requires a significant critical mass of
terization of these new materials at the molecular level is expertise of a scope only afforded through extensive scientific
essential. To accelerate the process, high-throughput charac- collaborations. Furthermore, scientists and engineers must
terization should be employed in cases where high-through- address the question of how the most promising materials will
put materials synthesis is possible. perform in an actual separation process. Beyond these
For crystalline materials, measurements of the adsorption considerations, the engineering economics of the new materi-
isotherms and breakthrough curves will be essential, while for als must be evaluated upon scaling-up the materials for
polymeric materials, the focus should be on adsorption and industrial applications, and economic models must be estab-
permeation experiments on small polymer films. In combi- lished to cover lifecycle CO2 separation, capture and seques-
nation with gas uptake measurements on powders or films, the tration costs for various technologies. The ease and cost of
structural information should allow issues regarding the large scale synthesis of membrane materials and metal–
loading of potential separation materials with different organic frameworks for example, will be critical in determin-
gases to be addressed. A parameter that must be assessed in ing their applicability on an industrial scale. While a detailed
all cases is the enthalpy of adsorption, since the cost for analysis of these issues was outside the scope of the present
regeneration of any capture material is critically dependent review, this aspect demands serious attention from research-
on the energy required to remove the CO2. ers in the field.
The static properties of the gas-loaded materials could be While no fully integrated, commercial-scale CCS projects
assessed using in-situ techniques such as resonant X-ray are in operation to date, many of the component technologies
absorption spectroscopy, which has the capability to study for the compression and transportation (e.g., through pre-
interactions between gas molecules and the matrix in a existing pipelines for instance) of CO2 are relatively mature.
spatially averaging manner. For crystalline materials, in-situ Progress has clearly been made on the preparation and testing
single crystal X-ray diffraction can be employed to determine (at the pilot plant stage) of new materials for capture,
the material structure under different loading conditions. however, there is an urgent need to develop new concepts and
Chemical information on polymer thin films could be to accelerate the deployment of cost-effective materials to the
obtained at the spatial resolution of a few nanometers using market. The long term storage of CO2 is the subject of a
energy-dispersive spectroscopy and through surface area number of major pilot projects worldwide, and is envisaged in
NMR relaxometry methods. deep geological formations such as saline aquifers or depleted
Characterization of the molecular transport properties of oil/gas fields, by injection into oceans, or by sequestration in
the materials is essential in order to obtain a molecular the form of mineral carbonates.
understanding of transport processes. Important fundamental Despite the numerous challenges surrounding CO2 cap-
questions include: how the structure changes with loading, ture, and the various political, regulatory and economic
how adsorbates bind to the material, and if so where and drivers which will ultimately dictate the time-to-deployment
through which interaction, as well as how different permeates for new CCS schemes, the time is ripe for us as a scientific
influence each other"s solubility. community to play a central role in solving the CO2 capture
Techniques such as resonant soft X-ray spectroscopy to problem.
study the internal chemistry of gas permeates and separation
media, and solid-state NMR may prove of great utility in This work was supported by the Sustainable Products &
relating diffusion to molecular structure. The most significant Solutions Program at the University of California, Berkeley.
conclusion from the measurements will be the ability to We thank the 1851 Royal Commission and the American–
correlate microscopic absorbate dynamics with the structural Australian Association for Research Fellowships for support of
information on loaded materials. A comparison between of D.M.D. B.S. and J.R.L. were supported as part of the Center for
the microscopic mobility and the macroscopic diffusion Gas Separations Relevant to Clean Energy Technologies, an
should provide insights into the mechanism of selective Energy Frontier Research Center funded by the U.S. Depart-
transport through these materials. ment of Energy, Office of Science, Office of Basic Energy
In parallel with experimental studies, computational Sciences under Award Number DE-SC0001015.
modeling methods must be further developed as a tool to
predict the performance of materials or membranes which are Received: January 25, 2010
proposed for a given separation process. Such techniques will Published online: July 22, 2010
enable large scale screening of new materials. Ultimately, a
clear understanding of the structure–function relationships
will direct experimental efforts towards a new generation of [1] N. Stern, Stern Review on the Economics of Climate Change,
materials with improved CO2 capture abilities. The modeling Cambridge University Press, Cambridge, 2006.

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6079
Reviews D. M. D’Alessandro et al.

[2] B. Metz, O. Davidson, H. de Coninck, M. Loos, L. Meyer, [39] C. Cadena, J. L. Anthony, J. K. Shah, T. I. Morrow, J. F.
Intergovernmental Panel on Climate Change. Special Report on Brennecke, E. J. Maginn, J. Am. Chem. Soc. 2004, 126, 5300.
Carbon Dioxide Capture and Storage, Cambridge University [40] J. L. Anderson, J. K. Dixon, E. J. Maginn, J. F. Brenncke,
Press, Cambridge, 2005, http://www.ipcc.ch/. J. Phys. Chem. B 2006, 110, 15 059.
[3] M. Z. Jacobson, Energy Environ. Sci. 2009, 2, 148. [41] E. D. Bates, R. D. Mayton, I. Ntai, J. H. Davis, J. Am. Chem.
[4] United Nations Framework Convention on Climate Change, Soc. 2002, 124, 926.
http://unfccc.int/2860.php. [42] J. Tang, T. H. , W. Sun, M. Plancer, M. Radosz, Y. Shen, Chem.
[5] McKinsey Climate Change Initiative, Carbon Capture & Commun. 2005, 3325.
Storage: Assessing the Economics, McKinsey & Company, 2008. [43] J. Tang, H. Tang, W. Sun, M. Radosz, Y. Shen, J. Polym. Sci. Part
[6] G. T. Rochelle, Science 2009, 325, 1652. A 2005, 43, 5477.
[7] IPCC, IPCC Special Report on Carbon Dioxide Capture and [44] P. J. E. Harlick, A. Sayari, Ind. Eng. Chem. Res. 2006, 45, 3248.
Storage, Cambridge University Press, Cambridge, 2005. [45] K. B. Lee, M. G. Beaver, H. S. Caram, S. Sircar, Ind. Eng.
[8] Electric Power Research Institute, Program on Technology Chem. Res. 2008, 47, 8048.
Innovation: Post-combustion CO2 Capture Technology Devel- [46] B. Feng, H. An, E. Tan, Energy Fuels 2007, 21, 426.
opment, Electric Power Research Institute, Palo Alto, 2008. [47] T. Yamaguchi, T. Niitsuma, B. N. Nair, K. Nakagawa, J. Membr.
[9] R. S. Haszeldine, Science 2009, 325, 1647. Sci. 2007, 294, 16.
[10] Science 2009, 325, 1644. [48] E. Ochoa-Fernandez, H. K. Rusten, H. A. Jakobsen, M. Ron-
[11] The National Mining Association of the American Mining ning, A. Holmen, D. Chen, Catal. Today 2005, 106, 41.
Industry, Status of CCS Development, http://www.nma.org/ccs/ [49] C. T. Yavuz, B. D. Shinall, A. V. Iretskii, M. G. White, T.
ccsprojects.asp. Golden, M. Atilhan, P. C. Ford, G. D. Stucky, Chem. Mater.
[12] J. D. Figueroa, T. Fout, S. Plasynski, H. McIlvried, R. D. 2009, 21, 3473.
Srivastava, Int. J. Greenhouse Gas Control 2008, 2, 9. [50] T. O. Nelson, L. J. Coleman, D. A. Green, R. P. Gupta, Energy
[13] S. Choi, J. H. Drese, C. W. Jones, ChemSusChem 2009, 2, 796. Procedia: Greenhouse Gas Control Technologies 9, Proceedings
[14] D. Aaron, C. Tsouris, Sep. Sci. Technol. 2005, 40, 321. of the 9th International Conference on Greenhouse Gas Control
[15] M. P. Suh, Y. E. Cheon, E. Y. Lee, Coord. Chem. Rev. 2008, 252, Technologies (GHGT-9), 16 – 20 November 2008, Washington,
1007. 2009, 1, 1305.
[16] J.-R. Li, R. J. Kuppler, H.-C. Zhou, Chem. Soc. Rev. 2009, 38, [51] R. V. Siriwardane, M.-S. Shen, E. P. Fisher, J. A. Poston, Energy
1477. Fuels 2001, 15, 279.
[17] R. E. Morris, P. S. Wheatley, Angew. Chem. 2008, 120, 5044; [52] K. T. Chue, J. N. Kim, Y. U. Yoo, S. H. Cho, R. T. Yang, Ind.
Angew. Chem. Int. Ed. 2008, 47, 4966. Eng. Chem. Res. 1995, 34, 591.
[18] A. D. Ebner, J. A. Ritter, Sep. Sci. Technol. 2009, 44, 1273. [53] G. Maurin, P. L. Llewellyn, R. G. Bell, J. Phys. Chem. B 2005,
[19] K. M. K. Yu, I. Curcic, J. Gabriel, S. C. E. Tsang, ChemSu- 109, 16084.
sChem 2008, 1, 893. [54] J. C. Hicks, J. H. Drese, D. J. Fauth, M. L. Gray, G. Qi, C. W.
[20] M. Radosz, X. Hu, K. Krutkramelis, Y. Shen, Ind. Eng. Chem. Jones, J. Am. Chem. Soc. 2008, 130, 2902.
Res. 2008, 47, 3783. [55] O. Leal, C. Bolivar, C. Ovalles, J. Garcia, Y. Espidel, Inorg.
[21] Z. Yong, V. Mata, A. E. Rodrigues, Sep. Purif. Technol. 2002, Chim. Acta 1995, 240, 183.
26, 195. [56] R. V. Sirwardane, U.S. Patent 6,908,497 B1, 1, 2005.
[22] M. C. Trachtenberg, R. M. Cowan, D. A. Smith in Proceedings [57] X. Xu, C. Song, B. G. Miller, A. W. Scaroni, Fuel Process.
of the Sixth Annual Conference on Carbon Capture & Technol. 2005, 86, 1457.
Sequestration, Pittsburgh, 2007. [58] S. Lee, T. P. Filburn, M. Gray, J.-W. Park, H.-J. Song, Ind. Eng.
[23] P. Chowdhury, C. Bikkina, S. Gumma, J. Phys. Chem. C 2009, Chem. Res. 2008, 47, 7419.
113, 6616. [59] N. Hiyoshi, D. K. Yogo, T. Yashima, Chem. Lett. 2004, 33, 510.
[24] R. Krishna, J. Phys. Chem. C 2009, 113, 19756. [60] S. Mazumder, P. van Hemert, A. Busch, K.-H. Wolf, P. Tejera-
[25] D. W. Keith, H.-D. Minh, J. K. Stolaroff, Clim. Change 2006, 74, Cuesta, Int. J. Coal Geol. 2006, 67, 267.
17. [61] C. A. Grande, A. E. Rodrigues, Int. J. Greenhouse Gas Control
[26] P. D. Vaidya, E. Y. Kenig, Chem. Eng. Technol. 2007, 30, 1467. 2008, 2, 194.
[27] G. Sartori, D. W. Savage, Ind. Eng. Chem. Fundam. 1983, 22, [62] P. K. Thallapally, P. B. McGrail, J. L. Atwood, C. Gaeta, C.
239. Tedesco, P. Neri, Chem. Mater. 2007, 19, 3355.
[28] A. Dibenedetto, M. Aresta, C. Fragale, M. Narracci, Green [63] P. K. Thallapally, P. B. McGrail, S. J. Dalgarno, H. T. Schaef, J.
Chem. 2002, 4, 439. Tian, J. L. Atwood, Natutre Mater. 2008, 7, 146.
[29] J. T. Yeh, H. W. Pennline, K. P. Resnik, Energy Fuels 2001, 15, [64] H. Furukawa, O. M. Yaghi, J. Am. Chem. Soc. 2009, 131, 8876.
274. [65] R. Babarao, J. W. Jiang, Langmuir 2008, 24, 6270.
[30] T. Mimura, T. Suda, I. Iwaki, A. Honda, H. Kumazawa, Chem. [66] R. Babarao, J. W. Jiang, J. Am. Chem. Soc. 2009, 131, 11417.
Eng. Commun. 1998, 170, 245. [67] S. R. Batten, R. Robson, Angew. Chem. 1998, 110, 1558;
[31] J. Johnson, Chem. Eng. News 2008, 86(March 3), 7. Angew. Chem. Int. Ed. 1998, 37, 1460.
[32] P. H. M. Feron, C. A. Hendriks, Oil Gas Sci. Technol. 2005, 60, [68] G. F$rey, Dalton Trans. 2009, 4400.
451. [69] G. F$rey, Chem. Soc. Rev. 2008, 37, 191.
[33] L. I. Eide, D. W. Bailey, Oil Gas Sci. Technol. 2005, 60, 475. [70] R. A. Fischer, C. W%ll, Angew. Chem. 2008, 120, 8285; Angew.
[34] Vattenfall!s Project on CCS, http://www.vattenfall.com/www/ Chem. Int. Ed. 2008, 47, 8164.
co2_en/co2_en/index.jsp. [71] S. R. Batten, S. M. Neville, D. R. Turner, Coordination Poly-
[35] P. N. Dyer, R. E. Richards, S. L. Russeka, D. M. Taylor, Solid mers: Design, Analysis and Application, Royal Society of
State Ionics 2000, 134, 21. Chemistry, Cambridge, 2008.
[36] M. M. Abu-Khader, Energy Sources Part A 2006, 28, 1261. [72] D. Tanaka, S. Kitagawa, Chem. Mater. 2008, 20, 922.
[37] M. Smiglak, A. Metlen, R. D. Rogers, Acc. Chem. Res. 2007, 40, [73] S. Kitagawa, R. Matsuda, Coord. Chem. Rev. 2007, 251, 2490.
1182. [74] C. J. Kepert, Chem. Commun. 2006, 695.
[38] J. L. Anthony, S. N. V. K. Aki, E. J. Maginn, J. F. Brennecke, Int. [75] U. Mueller, M. Schubert, F. Teich, H. Puetter, K. Schierle-
J. Environ. Technol. Manage. 2004, 4, 105. Arndt, J. Pastre, J. Mater. Chem. 2006, 16, 626.

6080 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082
Carbon Dioxide Capture
Angewandte
Chemie

[76] D. Bradshaw, J. B. Claridge, E. J. Cussen, T. J. Prior, M. J. [107] S. Bordiga, L. Regli, F. Bonino, E. Groppo, C. Lamberti, B.
Rosseinsky, Acc. Chem. Res. 2005, 38, 273. Xiao, P. S. Wheatley, R. E. Morris, A. Zecchina, Phys. Chem.
[77] S. Kitagawa, R. Kitaura, S. Noro, Angew. Chem. 2004, 116, Chem. Phys. 2007, 9, 2676.
2388; Angew. Chem. Int. Ed. 2004, 43, 2334. [108] A. O. Yazaydın, A. I. Benin, S. A. Faheem, P. Jakubczak, J. J.
[78] M. J. Rosseinsky, Microporous Mesoporous Mater. 2004, 73, 15. Low, R. R. Willis, R. Q. Snurr, Chem. Mater. 2009, 21, 1425.
[79] A. Stein, Adv. Mater. 2003, 15, 763. [109] P. L. Llewellyn, S. Bourrelly, C. Serre, A. Vimont, M. Daturi, L.
[80] S. Natarajan, S. Mandal, Angew. Chem. 2008, 120, 4876; Angew. Hamon, G. Weireld, J.-S. Chang, D.-Y. Hong, Y. K. Hwang,
Chem. Int. Ed. 2008, 47, 4798. S. H. Jhung, G. Ferey, Langmuir 2008, 24, 7245.
[81] O. M. Yaghi, M. O"Keeffe, N. W. Ockwig, H. K. Chae, M. [110] S. R. Miller, G. M. Pearce, P. A. Wright, F. Bonino, S. Chavan,
Eddaoudi, J. Kim, Nature 2003, 423, 705. S. Bordiga, I. Margiolaki, N. Guillou, G. Ferey, S. Bourrelly,
[82] M. Eddaoudi, D. B. Moler, H. Li, B. Chen, T. M. Reineke, M. P. L. Llewellyn, J. Am. Chem. Soc. 2008, 130, 15967.
O"Keeffe, O. M. Yaghi, Acc. Chem. Res. 2001, 34, 319. [111] P. D. C. Dietzel, V. Besikiotis, R. Blom, J. Mater. Chem. 2009,
[83] D. J. Tranchemontagne, Z. Ni, M. O"Keeffe, O. M. Yaghi, 19, 7362.
Angew. Chem. 2008, 120, 5214; Angew. Chem. Int. Ed. 2008, 47, [112] S. R. Caskey, A. G. Wong-Foy, A. Matzger, J. Am. Chem. Soc.
5136. 2008, 130, 10870.
[84] M. O"Keeffe, M. A. Peskov, S. J. Ramsden, O. M. Yaghi, Acc. [113] D. Britt, H. Furukawa, B. Wang, T. G. Glover, O. M. Yaghi,
Chem. Res. 2008, 41, 1782. Proc. Natl. Acad. Sci. USA 2009, 106, 20637.
[85] N. L. Rosi, J. Kim, M. Eddaoudi, B. Chen, M. O"Keeffe, O. M. [114] Y.-S. F. Bae, O. K. Farha, A. M. Spokoyny, C. A. Mirkin, J. T.
Yaghi, J. Am. Chem. Soc. 2005, 127, 1504. Hupp, R. Q. Snurr, Chem. Commun. 2008, 4135.
[86] M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. [115] Y. E. Cheon, M. P. Suh, Chem. Commun. 2009, 2296.
O"Keeffe, O. M. Yaghi, Science 2002, 295, 469. [116] B. Chen, S. Ma, E. J. Hurtado, E. B. Lobkovsky, H.-C. Zhou,
[87] A. G. Wong-Foy, A. J. Matzger, O. M. Yaghi, J. Am. Chem. Soc. Inorg. Chem. 2007, 46, 8490.
2006, 128, 3494. [117] L. Bastin, P. S. Barcia, E. J. Hurtado, J. A. C. Silva, A. E.
[88] A. R. Millward, O. M. Yaghi, J. Am. Chem. Soc. 2005, 127, Rodrigues, B. Chen, J. Phys. Chem. C 2008, 112, 1575.
17998. [118] Y. E. Cheon, J. Park, M. P. Suh, Chem. Commun. 2009, 5436.
[89] Y. E. Cheon, M. P. Suh, Chem. Eur. J. 2008, 14, 3961. [119] S. Bourrelly, P. L. Llewellyn, C. Serre, F. Millange, L. Loiseau,
G. F$rey, J. Am. Chem. Soc. 2005, 127, 13519.
[90] Z. Zhao, Z. Li, Y. S. Lin, Ind. Eng. Chem. Res. 2009, 48, 10015.
[120] P. L. Llewellyn, S. Bourrelly, C. Serre, Y. Filinchuk, G. Ferey,
[91] K. S. Walton, A. R. Millward, D. Dubbeldam, H. Frost, J. J.
Angew. Chem. 2006, 118, 7915; Angew. Chem. Int. Ed. 2006, 45,
Low, O. M. Yaghi, R. Q. Snurr, J. Am. Chem. Soc. 2008, 130,
7751.
406.
[121] S. Galli, N. Masciocchi, G. Tagliabue, A. Sironi, J. A. R.
[92] M. Xue, Y. Liu, R. M. Schaffino, S. Xiang, X. Zhao, G.-S. Zhu,
Navarro, J. M. Salas, L. Mendez-Linan, M. Domingo, M.
S.-L. Qui, B. Chen, Inorg. Chem. 2009, 48, 4649.
Perez, E. Barea, Chem. Eur. J. 2008, 14, 9890.
[93] S. R. Miller, P. A. Wright, T. Devic, C. Serre, G. Ferey, P. L.
[122] H.-S. Choi, M. P. Suh, Angew. Chem. 2009, 121, 6997; Angew.
Llewellyn, R. Denoyel, L. Gaberova, Y. Filinchuk, Langmuir
Chem. Int. Ed. 2009, 48, 6865.
2009, 25, 3618.
[123] Y.-S. Bae, O. K. Farha, J. T. Hupp, R. Q. Snurr, J. Mater. Chem.
[94] A. Comotti, S. Bracco, P. Sozzani, S. Horike, R. Matsuda, J.
2009, 19, 2131.
Chen, M. Takata, Y. Kubota, S. Kitagawa, J. Am. Chem. Soc.
[124] E. Neofotistou, C. D. Malliakas, P. N. Trikalitis, Chem. Eur. J.
2008, 130, 13664. 2009, 15, 4523.
[95] B.-Q. Ma, K. L. Mulfort, J. T. Hupp, Inorg. Chem. 2005, 44, [125] S. Couck, J. F. M. Denayer, G. V. Baron, T. Remy, J. Gascon, F.
4912. Kapteijn, J. Am. Chem. Soc. 2009, 131, 6326.
[96] K. L. Mulfort, J. T. Hupp, J. Am. Chem. Soc. 2007, 129, 9604. [126] B. Arstad, H. Fjellvag, K. O. Kongshaug, O. Swang, R. Blom,
[97] Y. S. Bae, K. L. Mulfort, H. Frost, P. Ryan, S. Punnathanam, Adsorption 2008, 14, 755.
L. J. Broadbelt, J. T. Hupp, R. Q. Snurr, Langmuir 2008, 24, [127] A. Demessence, D. M. D"Alessandro, M. L. Foo, J. R. Long,
8592. J. Am. Chem. Soc. 2009, 131, 8784.
[98] A. Phan, C. J. Doonan, F. J. Uribe-Romo, C. B. Knobler, M. [128] F. Dreisbach, R. Staudt, J. U. Keller, Adsorption 1999, 5, 215.
O"Keeffe, O. M. Yaghi, Acc. Chem. Res. 2010, 43, 58. [129] P. D. C. Dietzel, R. E. Johnsen, H. Fjellvag, S. Bordiga, E.
[99] R. Banerjee, A. Phan, B. Wang, C. Knobler, H. Furukawa, M. Groppo, S. Chavan, R. Blom, Chem. Commun. 2008, 5125.
O"Keeffe, O. M. Yaghi, Science 2008, 319, 939. [130] O. K. Farha, A. M. Spokoyny, K. L. Mulfort, S. Galli, J. T.
[100] K. S. Park, Z. Ni, A. P. Cote, J. Y. Choi, R. Huang, F. J. Uribe- Hupp, C. A. Mirkin, Small 2009, 5, 1727.
Romo, H. K. Chae, M. O"Keeffe, O. M. Yaghi, Proc. Natl. [131] S. Horike, S. Shimomura, S. Kitagawa, Nat. Chem. 2009, 1, 695.
Acad. Sci. USA 2006, 103, 10186. [132] P. K. Thallapally, J. Tian, M. R. Kishan, C. A. Fernandez, S. J.
[101] H. Hayashi, A. P. Cote, H. Furukawa, M. O"Keeffe, O. M. Dalgarno, P. B. McGrail, J. E. Warren, J. L. Atwood, J. Am.
Yaghi, Nat. Mater. 2007, 6, 501. Chem. Soc. 2008, 130, 16842.
[102] P. K&sgens, M. Rose, I. Senkovska, H. Fr%de, A. Henschel, S. [133] A. Vimont, A. Travert, P. Bazin, J.-C. Lavalley, M. Daturi, C.
Siegle, K. Kaskel, Microporous Mesoporous Mater. 2009, 120, Serre, G. Ferey, S. Bourrelly, P. L. Llewellyn, Chem. Commun.
325. 2007, 3291.
[103] S. S.-Y. Chui, S. M.-F. Lo, J. P. H. Charmant, A. G. Orpen, I. D. [134] C. Serre, S. Bourrelly, A. Vimont, N. A. Ramsahye, G. Maurin,
Williams, Science 1999, 283, 1148. P. L. Llewellyn, M. Daturi, Y. Filinchuk, O. Leynaud, P. Barnes,
[104] Q. M. Wang, D. Shen, M. B&low, M. L. Lau, S. Deng, F. R. G. Ferey, Adv. Mater. 2007, 19, 2246.
Fitch, N. O. Lemcoff, J. Semanscin, Microporous Mesoporous [135] N. A. Ramsahye, G. Maurin, S. Bourrelly, P. Llewellyn, T.
Mater. 2002, 55, 217. Loiseau, G. Ferey, Phys. Chem. Chem. Phys. 2007, 9, 1059.
[105] P. Chowdhury, C. Bikkina, D. Meister, F. Dreisbach, S. Gumma, [136] N. A. Ramsahye, G. Maurin, S. Bourrelly, P. L. Llewellyn, T.
Microporous Mesoporous Mater. 2009, 117, 406. Loiseau, C. Serre, G. Ferey, Chem. Commun. 2007, 3261.
[106] Z. Liang, M. Marshall, A. L. Chaffee, Energy Fuels 2009, 23, [137] N. A. Ramsahye, G. Maurin, S. Bourrelly, P. L. Llewellyn, C.
2785. Serre, T. Loiseau, T. Devic, G. Ferey, J. Phys. Chem. C 2008,
112, 514.

Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082 ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6081
Reviews D. M. D’Alessandro et al.

[138] L. Hamon, P. L. Llewellyn, T. Devic, A. Ghoufi, G. Clet, V. [172] R. Krishna, J. M. van Baten, Langmuir 2010, 26, 8450.
Guillerm, G. D. Pirngruber, G. Maurin, C. Serre, G. Driver, W. [173] R. Krishna, J. M. van Baten, Chem. Eng. J. 2007, 133, 121.
van Beek, E. Jolimaitre, A. Vimont, M. Daturi, G. Ferey, J. Am. [174] Q. Yang, C. Xue, C. Zhong, J.-F. Chen, AIChE J. 2007, 53, 2832.
Chem. Soc. 2009, 131, 17490. [175] A. Martin-Calvo, E. Garcia-Perez, J. M. Castillo, S. Calero,
[139] H. Kanoh, A. Kondo, H. Noguchi, H. Kajiro, A. Tohdoh, Y. Phys. Chem. Chem. Phys. 2008, 10, 7085.
Hattori, W.-C. Xu, M. Inoue, T. Siugiura, K. Morita, T. Ohba, K. [176] S. Keskin, D. S. Sholl, Ind. Eng. Chem. Res. 2009, 48, 914.
Kaneko, J. Colloid Interface Sci. 2009, 334, 1. [177] S. Keskin, D. S. Sholl, J. Phys. Chem. Lett. C 2007, 111, 14055.
[140] A. Kondo, A. Chinen, H. Kajiro, T. Nakagawa, K. Kato, M. [178] D. Sholl, S. Keskin, S. E. Jee, H. Chen, Abstracts of Papers,
Takata, Y. Hattori, F. Okino, T. Ohba, K. Kaneko, H. Kanoh, 234th ACS National Meeting, Boston, August 19 – 23 2007,
Chem. Eur. J. 2009, 15, 7549. FUEL216.
[141] S. Ma, D. Sun, X.-S. Wang, H.-C. Zhou, Angew. Chem. 2007, [179] S. Keskin, J. Liu, J. K. Johnson, D. S. Sholl, Microporous
119, 2510; Angew. Chem. Int. Ed. 2007, 46, 2458. Mesoporous Mater. 2009, 125, 101.
[142] B. Wang, A. P. Cote, H. Furukawa, M. O"Keeffe, O. M. Yaghi, [180] C. A. Scholes, S. E. Kentish, G. W. Stevens, Recent Pat. Chem.
Nature 2008, 453, 207. Eng. 2008, 1, 52.
[143] S. Ma, X.-S. Wang, E. S. Manis, C. D. Collier, H.-C. Zhou, Inorg. [181] L. Robeson, J. Membr. Sci. 1991, 62, 165.
Chem. 2007, 46, 3432. [182] L. Zhang, I.-S. Park, K. Shqau, W. S. W. Ho, H. Verweij, JOM
[144] K. K. Tanabe, Z. Wang, S. M. Cohen, J. Am. Chem. Soc. 2008, 2009, 61(April), 61.
130, 8508. [183] Carbon Capture and its Storage: an Integrated Assessment
[145] F. Salles, H. Jobic, A. Ghoufi, P. L. Llewellyn, C. Serre, S. (Eds.: S. Shackley, C. Gough), Aldershot, Ashgate, 2006, p. 313.
Bourrelly, G. Ferey, G. Maurin, Angew. Chem. 2009, 121, 8485; [184] M. A. Carreon, S. Li, J. L. Falconer, R. D. Noble, J. Am. Chem.
Angew. Chem. Int. Ed. 2009, 48, 8335. Soc. 2008, 130, 5412.
[146] D. Saha, S. Deng, J. Chem. Eng. Data 2009, 54, 2245. [185] H. Guo, G. Zhu, I. J. Hewitt, S. Qui, J. Am. Chem. Soc. 2009,
[147] D. Saha, Z. Wei, S. Deng, Int. J. Hydrogen Energy 2008, 33, 131, 1646.
7479. [186] Y. Liu, Z. Ng, E. A. Khan, H.-K. Jeong, C. Ching, Z. Lai,
[148] S. Keskin, J. Liu, J. K. Johnson, D. S. Sholl, Langmuir 2008, 24, Microporous Mesoporous Mater. 2009, 118, 296.
8254.
[187] Y.-S. Li, F.-Y. Liang, H. Bux, A. Feldhoff, W.-S. Yang, J. Caro,
[149] V. Krungleviciute, K. Lask, A. D. Migone, J.-Y. Lee, J. Li,
Angew. Chem. 2010, 122, 558; Angew. Chem. Int. Ed. 2010, 49,
AIChE J. 2008, 54, 918.
548.
[150] D. Frenkel, B. Smit, Understanding Molecular Simulations:
[188] N. B. McKeown, P. M. Budd, Chem. Soc. Rev. 2006, 35, 675.
From Algorithms to Applications, 2nd ed., Academic Press, San
[189] V. Abetz, T. Brinkmann, M. Dijkstra, K. Ebert, D. Fritsch, K.
Diego, 2002.
Ohlrogge, D. Paul, K.-V. Peinemann, S. Pereira-Nunes, N.
[151] B. Smit, T. L. M. Maesen, Chem. Rev. 2008, 108, 4125.
Scharnagi, M. Schossig, Adv. Eng. Mater. 2006, 8, 328.
[152] G. F$rey, C. Serre, Chem. Soc. Rev. 2009, 38, 1380.
[190] C. E. Powell, G. G. Qiao, J. Membr. Biol. 2006, 279, 1.
[153] C. Serre, F. Millange, C. Thouvenot, M. Nogues, G. Marsolier,
[191] R. W. Baker, K. Lokhandwala, Ind. Eng. Chem. Res. 2008, 47,
D. Louer, G. Ferey, J. Am. Chem. Soc. 2002, 124, 13519.
2109.
[154] Y. Liu, J. H. Her, A. Dailly, A. J. Ramirez-Cuesta, D. A.
[192] J. Germain, J. M. J. Frechet, F. Svec, J. Mater. Chem. 2007, 17,
Neumann, C. M. Brown, J. Am. Chem. Soc. 2008, 130, 11813.
4989.
[155] D. Dubbeldam, R. Krishna, R. Q. Snurr, J. Phys. Chem. C 2009,
[193] J. Germain, J. M. J. Fr$chet, F. Svec, PMSE Preprints 2007, 97,
113, 19317.
272.
[156] F. Salles, A. Ghoufi, G. Maurin, R. G. Bell, C. Mellot-
Draznieks, G. Ferey, Angew. Chem. 2008, 120, 8615; Angew. [194] J. Germain, J. Hradil, J. M. J. Frechet, F. Svec, Chem. Mater.
Chem. Int. Ed. 2008, 47, 8487. 2006, 18, 4430.
[157] D. S. Coombes, F. Cora, C. Mellot-Draznieks, R. G. Bell, [195] J. Germain, J. M. J. Frechet, F. Svec, Chem. Mater. 2008, 20,
J. Phys. Chem. C 2009, 113, 544. 7069.
[158] F. X. Coudert, C. Mellot-Draznieks, A. H. Fuchs, A. Boutin, [196] J. Zou, W. S. W. Ho, J. Membr. Sci. 2006, 286, 310.
J. Am. Chem. Soc. 2009, 131, 11 329. [197] Y. Zhang, I. H. Musselman, J. P. Ferraris, K. J. Balkus,
[159] A. Garcia-Sanchez, C. O. Ania, J. B. Parra, D. Dubbeldam, J. Membr. Sci. 2008, 313, 170.
T. J. H. Vlugt, R. Krishna, S. Calero, J. Phys. Chem. C 2009, 113, [198] S. Choi, J. Coronas, E. Jordan, W. Oh, S. Nair, F. Onorato, D. F.
8814. Shantz, M. Tsapatsis, Angew. Chem. 2008, 120, 562; Angew.
[160] R. B. Rankin, J. C. Liu, A. D. Kulkarni, J. K. Johnson, J. Phys. Chem. Int. Ed. 2008, 47, 552.
Chem. C 2009, 113, 16906. [199] R. Quinn, J. B. Appleby, G. P. Pez, J. Membr. Sci. 1995, 104, 139.
[161] J. I. Siepmann, S. Karaborni, B. Smit, J. Am. Chem. Soc. 1993, [200] J. E. Bara, C. J. Gabriel, E. S. Hatakeyamaa, T. K. Carlisle, S.
115, 6454. Lessmann, R. D. Noble, D. L. Gin, J. Membr. Sci. 2008, 321, 3.
[162] J. L. Belof, A. C. Stern, B. Space, J. Chem. Theory Comput. [201] C. Myers, H. Pennline, D. Luebke, J. Ilconich, J. K. Dixon, E. J.
2008, 4, 1332. Maginn, J. F. Brennecke, J. Membr. Sci. 2008, 322, 28.
[163] A. Torrisi, C. Mellot-Draznieks, R. G. Bell, J. Chem. Phys. 2009, [202] J. Ilconich, C. Myers, H. Pennline, D. Luebke, J. Membr. Sci.
130, 194703. 2007, 298, 41.
[164] K. Sillar, A. Hofmann, J. Sauer, J. Am. Chem. Soc. 2009, 131, [203] E. Favre, J. Membr. Sci. 2007, 294, 50.
4143. [204] C. Stewart, M.-A. Hessami, Energy Convers. Manage. 2005, 46,
[165] B. Liu, B. Smit, Langmuir 2009, 25, 5918. 403.
[166] C. Mellot-Draznieks, J. Mater. Chem. 2007, 17, 4348. [205] P. Linga, R. Kumar, P. Englezos, J. Hazard. Mater. 2007, 149,
[167] T. D&ren, Y. S. Bae, R. Q. Snurr, Chem. Soc. Rev. 2009, 38, 1237. 625.
[168] A. J. Fletcher, K. M. Thomas, M. J. Rosseinsky, J. Solid State [206] R. Kumar, P. Englezos, I. Moudrakovski, J. A. Ripmeester,
Chem. 2005, 178, 2491. AIChE J. 2009, 55, 1584.
[169] Q. Y. Yang, C. L. Zhong, Langmuir 2009, 25, 2302. [207] R. Kumar, P. Linga, J. A. Ripmeester, P. Englezos, J. Environ.
[170] R. Babarao, J. W. Jiang, Energy Environ. Sci. 2008, 1, 139. Eng. 2009, 411.
[171] R. Krishna, J. M. van Baten, Langmuir 2010, 26, 3981. [208] B. Smit, R. Krishna, Chem. Eng. Sci. 2003, 58, 557.

6082 www.angewandte.org ! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 6058 – 6082