Chemistry 5

Chemical Engineering Research and Design 189 (2023) 745–767
Available online at www.sciencedirect.com
Chemical Engineering Research and Design
journal homepage: www.elsevier.com/locate/cherd
Analytical review of the current state of knowledge

of adsorption materials and processes for direct air
]]
]]]]]]
]]
capture
⁎ ⁎
May-Yin (Ashlyn) Low 1, Lucy Victoria Barton 1, Ronny Pini , Camille Petit
Department of Chemical Engineering, South Kensington Campus, London SW7 2AZ, UK
a r t i c l e i n f o a b s t r a c t
Article history: Significant research interest has been directed towards the deployment of direct air
Received 2 August 2022 capture (DAC) as a net-negative CO2 emissions technology to help limit global tempera-
Received in revised form ture rise to below 2 °C. The scope of this review is to outline the advancement of ad-
6 November 2022 sorption-based DAC technologies, as well as to highlight the still-existing data gaps, for
Accepted 25 November 2022 both materials’ development and process design in the period 2016 – 2021. On the material
Available online 28 November 2022 side, we highlight the available and missing data on adsorbent properties in relation to
what is needed for process modelling and design. We cover material densities, textural
Keywords: properties, thermal properties, adsorption isotherms (i.e. CO2, N2, O2, H2O), adsorption
Negative emissions technology kinetics, and adsorbent stability towards humidity, oxidation, and cycling. On the process
Direct air capture side, we provide a detailed look at key process studies conducted in the same time frame
Adsorption by considering the trade-offs to be expected in the design of the adsorption-based DAC
Carbon capture process. We focus on process configuration and contactor design, desorption processes,
CO2 and the need for systematic reporting of key performance indicators to allow for accurate
comparisons and benchmarking. Throughout the review, we identify the lack of synergy
between material and process development which must be addressed to advance the field
of DAC by adsorption.
© 2022 The Authors. Published by Elsevier Ltd on behalf of Institution of Chemical
Engineers. This is an open access article under the CC BY license (http://creative-
commons.org/licenses/by/4.0/).
gas (GHG) emissions and the subsequent impact on the

1. Introduction
global mean surface temperature. Two scenarios – SSP1–2.6
and SSP1–1.9 – predict a surface temperature increase lower
At the 2015 Conference of Parties (COP21) in Paris, 196
than 1.5 – 2 °C before 2100 (relative to 1850–1900 levels) by
countries signed an agreement to limit global warming to
considering the implementation of net-negative CO2 emis-
below 2 °C when compared to pre-industrial levels
sions technologies (NETs). NETs include afforestation, Bioe-
(1850–1900). Since then, anthropogenic carbon dioxide (CO2)
nergy with Carbon Capture and Storage, Soil Carbon
emissions have continued to rise leading to a global mean
Sequestration, and Direct Air Capture with Carbon Seques-
surface temperature increase of ∼1.1 °C (NASA, 2022). Mod-
tration (DACCS). Each technology presents distinct technical,
elling results presented in the recent Intergovernmental
socio-economic, and environmental strengths and chal-
Panel on Climate Change (IPCC) report (Assessment Report 6)
lenges, and a mixed technology portfolio is needed to meet
(Arias et al., 2021) discuss different Shared Socio-economic
the temperature increase target (Minx et al., 2018). Among
Pathways (SSPs) that forecast possible trends of greenhouse
NETs, DACCS offers long-term carbon storage and allows the
amount captured to be accurately measured.
⁎ DAC presents a unique challenge when compared to post-
Corresponding authors.
E-mail addresses: [email protected] (R. Pini), combustion carbon capture due to its ultra-dilute nature
[email protected] (C. Petit). (419 ppm, 0.042 vol% in air) and the inherent variability of
1
These authors contributed equally.
https://doi.org/10.1016/j.cherd.2022.11.040
0263-8762/© 2022 The Authors. Published by Elsevier Ltd on behalf of Institution of Chemical Engineers. This is an open access article
under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
746 Chemical Engineering Research and Design 189 (2023) 745–767
Gas Stream CO2 concentration [%mol]

0 5 10 15 20 25
Steel
Cement
Coal
Natural Gas
DAC
100 101 102 103 104

Minimum Gas Volume @ STP for 1 kg CO2 [m3·kg-1
CO2]
Fig. 2 – Overview of the monthly average temperature and
Fig. 1 – The DAC Challenge. Volume of gas (at STP, bar relative humidity distribution of the different DAC project
chart) containing 1 kg CO2 for CCS applications with locations listed in Table 1, based on data collected between
different inlet process stream compositions (mol%, 1999 and 2019 by the European Centre for Medium-Range
symbols). Flue gas compositions used in calculations from Weather Forecasts (ECMWF) (CLIMATE-DATA.ORG). Box
Bui et al. (2016), Danaci et al. (2020) and Markewitz edges represent the 25th and 75th percentile, whisker
et al. (2019). edges represent the minimum and maximum data points,
the solid line through the box represents the median, and
the filled square symbol represents the mean. Data for
ambient conditions that the process is subject to. To visualise Reykjavik, Aberdeen, Rong and Odessa were collected in
this challenge, we compare in Fig. 1 the volume at standard place of Hellishiedi, North East Scotland, Kollsnes region,
temperature and pressure (STP) of air that contains 1 kg of and Permian Basin, respectively.
CO2 with its flue gas counterpart originating from different
industrial sources. By assuming 100% of CO2 is captured,
1300 m3 of air would need to be processed as opposed to only 0.0007% of the 2065 removal quantity for SSP1–1.9.
13 m3 of gas for a natural gas flue gas stream. Hence, sustained, significant, and consolidated efforts are
Another challenge for DAC systems is that CO2 must be required for the development and scale up of DAC processes.
selectively isolated from the air in the face of wide variations Among the current DAC companies, Carbon Engineering
in key variables such as temperature and relative humidity utilises absorption and uses an aqueous hydroxide (po-
(RH). As shown in Fig. 2, the conditions at which the DAC unit tassium hydroxide, KOH), which is regenerated in a calciner
must operate can vary over a very large range – from 30% to at 674 °C (Keith et al., 2018). Amine-based absorption pro-
90% RH, while monthly average temperatures can go below cesses (e.g., monoethanolamine (MEA) solutions) would offer
0 °C and reach almost 30 °C. Within the same geographical milder regeneration conditions (∼120 °C) (Kiani et al., 2020;
location, the average temperature in a year can vary by al- Barzagli et al., 2020), however, are not yet commercially es-
most 25 °C. The importance of temperature and humidity tablished for DAC. Adsorption-based processes also allow
conditions is investigated in detail in a review by Kong milder regeneration conditions (∼100 °C) and, unlike ab-
et al. (2022). sorption processes, do not suffer from water losses (Kiani
Despite the challenges of DAC, commercial DAC in- et al., 2020). Both Climeworks and Global Thermostat use an
itiatives are being established around the world with the aim amine functionalised polymer as an adsorbent, but employ
of demonstrating the feasibility of processes and business different methods for adsorbent regeneration, i.e. tempera-
models (Erans et al., 2022). These initiatives are largely lim- ture vacuum swing (TVSA) and steam assisted temperature
ited to a small number of start-ups, with Climeworks, Carbon swing (S-TSA), respectively. Overall, in humid environments,
Engineering, and Global Thermostat presenting the most MEA-based absorption processes could perform well, but
mature operations (see Table 1). The DAC technologies em- may require excessive water makeup at low relative hu-
ployed by these emerging companies rely on adsorption or midity conditions. Adsorption processes do not face this
absorption processes. The CO2 recovered by the DAC plants issue, but they must contend with water co-adsorption in
in Table 1 is either stored or utilised in short term CO2 re- humid environments. Although water co-adsorption can
moval applications, such as greenhouse fertilisation. As- improve CO2 uptake for amine-based adsorbents (Young
suming all the operational plants listed in Table 1 are et al., 2021), it can also increase the energy required for re-
running at their reported annual capacity, the current DAC generation (desorption). At this stage, the energy trade-off
CO2 capture capacity amounts to ∼17 ktmetric year-1. To put between aqueous and solid amine-based DAC processes is
this number into perspective, in the IPCC Assessment Report unclear.
6, SSP1–1.9 is based on overall net-negative emissions from Our review, which encompasses over 110 publications,
2058 reaching 2.5 Gt CO2 year-1 by 2065 and 14 Gt CO2 year-1 aims to provide readers with an update on adsorption-based
by 2100. The capacity of operational DAC plants is currently DAC adsorbent materials and processes developed in the
Table 1 – Overview of operating DAC plants and planned projects.
Capture Technology Company (project name) CO2 Capture CO2 Fate Start Status Location Ref
Capacity (ktmetric Date
year-1)
Adsorption Climeworks (Capricon) 9 Utilisation 2017 Operational Hinwil, Switzerland (Beuttler et al., 2019;
Temperature vacuum Climeworks, 2022a)
swing adsorption
(TVSA)
Climeworks (Arctic Fox) 0.05 Storage 2017 Operational Hellisheidi, Iceland (Climeworks, 2022b)
Climeworks, Carbfix, ON 4 Storage 2021 Operational Hellisheidi, Iceland (B. P. C.-D. A. C. A. Council,
power (Orca) + Utilisation 2021; Climeworks)
Climeworks + others 1.05a Utilisation – Operational Italy, Switzerland, (B. P. C.-D. A. C. A. Council,
Iceland 2021; B. Policy, 2021)
Adsorption Global Thermostat 0.5 – 2010 Unknownb Menlo Park, (B. P. C.-D. A. C. A. Council,
Steam Temperature Swing California, U.S. 2021; Chichilnisky, 2016)
Adsorption
(S-TSA)
Global Thermostat 1 – 2013 Unknownb Menlo Park, (B. P. C.-D. A. C. A. Council,
California, U.S. 2021; Chichilnisky, 2016)
Global Thermostat 1c – 2018 Non-Operational Huntsville, (Rathi, 2021; Peters, 2019)
Alabama, U.S.
Absorption Carbon Engineering 0.219 Utilisation 2015 Operational Squamish, B.C., (Viebahn et al., 2019)
Canada
Carbon Engineering Innovation 1.5 Utilisation 2021 Under Squamish, B.C., (B. P. C.-D. A. C. A. Council,
Centre commission Canada 2021; C. Engineering, 2022a)
Carbon Engineering, 1PointFive, 1000 Storage 2024 Planned Permian Basin, (C. Engineering, 2022b; Davis,
Oxy Low Carbon Ventures + Utilisation Texas, U.S. 2021; G. C. Institute)
Carbon Engineering, Storegga 500–1000 Storage 2026 Planned North East (C. Engineering, 2022c)
Chemical Engineering Research and Design 189 (2023) 745–767
Scotland, U.K.
Carbon Engineering, Carbon 500–1000 Storage – Planned Kollsnes region, (C. Engineering, 2022d)
Removal, Oxy Low Carbon Norway
Ventures
-: indicates information was unavailable

a: calculated by subtracting 900 and 50 from the total capacity of 14 Climeworks plants in 2017, reported as 2000 t
b: Ref (Chichilnisky, 2018) reports that the plants are operational, but ref (B. P. C.-D. A. C. A. Council, 2021) suggests that at least one of the two is no longer operational
c: Original goal by Global Thermostat was 4000 t / year
747
Table 2 – Summary of adsorbate and adsorbent properties required in the modelling of an adsorption-based DAC
process. Other properties will influence the design of the process but do not act as direct value inputs into the process
modelling, e.g. textural properties and thermal stability.
Material Property Process Modelling
Equilibrium isotherms considering co-adsorption for ALL adsorbing Determining total amount adsorbed and desorbed
components in ambient air at multiple temperatures
Overall mass transfer coefficient Determining the rate of mass transfer and adsorption
Density (skeletal, particle, bed) Determining mass of solid and volume of process
contactor required
Porosity (particle, bed) derived from pore volume estimation Determining total volume of gas in vessel space and mass
of adsorbent
Heat capacity (adsorbent and adsorbed phase) Determining heat required to perform desorption
Heat of adsorption (of all adsorbing components) Determining required energy input for desorption and
process temperature changes
Overall heat transfer coefficient (related to thermal conductivity of Determining rate of heat transfer in material
adsorbent)
period 2016 – 2021. On the material side, we focus on data the adsorbents being investigated for DAC into six categories:
availability and data gap of adsorbent properties in relation silicas, metal oxides, metal organic frameworks (MOFs),
to what is required for deployment of an adsorption tech- carbons, zeolites, and polymers. These adsorbents can either
nology. Table 2 highlights such properties and how they in- be functionalised or unfunctionalised. Functionalisation ty-
form mass and heat transfer, and ultimately process pically ‘converts’ a physisorbent, which binds adsorbates via
performance. We include to the best of our abilities all can- weak intermolecular forces, into a chemisorbent, which
didate DAC adsorbents developed in the reviewed period, binds adsorbates through covalent or coordination bonds.
and summarise the material properties and equilibrium ad- Amines are the most common chemicals used for functio-
sorption data obtained. On the process side, we focus on nalisation, e.g. polyamines and aminosilanes. Primary and
adsorber/contactor and process designs. We provide a de- secondary amines perform better for DAC adsorption com-
tailed look at all key process studies conducted in the same pared to tertiary amines, which require the presence of
time frame and offer suggestions on decision making and moisture to ensure significant uptake (Sanz-Pérez et al.,
trade-offs in adsorption-based processes. Throughout, we 2016). Amines can either be physically impregnated (class 1),
highlight the connections and current data gaps between covalently grafted to the adsorbent’s surface either by silane
materials development and process studies. Our review linkages (class 2) or via in-situ polymerisation (class 3) (Sanz-
complements past and current reports on DAC. In particular, Pérez et al., 2016). Carbonate functionalisation can also be
it updates the adsorption component of the 2016 review by used for CO2 capture, although in this case, adsorption oc-
Sanz-Pérez et al. (2016) which covers candidate DAC liquid curs only in the presence of moisture (Rodriguez-Mosqueda
and solid adsorbents. The 2022 perspective by Kong et al. et al., 2018). For each material property, we first explain how
(2022) recommends future research on DAC to focus on ma- it can be measured and then discuss the data collected in
terial and process development under sub-ambient and terms of both their availability and the implications of the
humid conditions. We expand on this by also highlighting values reported.
the need for synergy between material and process devel- For perspective on the discussion below related to Fig. 3,
opment. For a broader view on DAC, we direct readers to the we highlight two key points. First, there is currently no
recent article by Erans et al. (2022) which covers the scien- benchmark DAC sorbent, which leads to an absence of re-
tific, engineering, economic, and socio-political aspects of ference DAC sorbent properties. A convenient benchmark
DAC and its role compared to other negative emissions could be an adsorbent used in current DAC commercial
technologies. processes, yet today, their properties remain proprietary.
Second, there also do not exist any calculations of theoretical
optimal properties of an “ideal” DAC adsorbent. While a
2. Material properties priori possible, building such theoretical framework remains
complex, owing to the interdependent nature of the various
Here, we first review the reported material properties of materials properties.
possible DAC adsorbents studied during the period covered.
These properties include: density (skeletal, particle, bed),
textural properties (surface area, total pore volume, micro- 2.1. Density and mechanical properties
pore volume), and thermal properties (specific heat capacity,
thermal conductivity, thermal stability). As explained in 2.1.1. Skeletal density
Table 2, these properties are needed to assess the suitability Skeletal density (ρskeletal) is defined as the mass of a single
of a material for DAC applications via process modelling. particle of adsorbent divided by the volume of the particle
These data are summarised in Fig. 3, where we have classed without including any open pores. The term particle can
Chemical Engineering Research and Design 189 (2023) 745–767 749
Fig. 3 – Comparison of a) skeletal density, b) particle density, c) bed density, d) BET surface area, e) total pore volume, f)
micropore volume, g) specific heat capacity, h) thermal conductivity, and i) thermal stability values between different
adsorbent categories, both with and without additional functionalisation. Box edges represent the 25th and 75th percentile,
whisker edges represent the minimum and maximum data points, the solid line through the box represents the median,
and the filled square symbol represents the mean. Total number of data entries are labelled to the left of each box to better
understand the spread of data. All data entries can be found in Tables S1–S6.
refer either to a single powder particle of the adsorbent, or to can be measured experimentally using mercury intrusion
a single bead or pellet macrostructure. The skeletal density porosimetry (Pini, 2014) or gas pycnometry (Micromeritics,
can be measured experimentally using a variety of methods 2001). For unfunctionalised crystalline materials (e.g. MOFs
such as helium pycnometry (Nguyen et al., 2019), helium and zeolites), the particle density can be calculated based on
gravimetry (Pini, 2014), or a combination of N2 adorption at the crystalline structure of the materials. For process mod-
− 196 °C and mercury intrusion porosimetry (Azzan et al., elling, the pellet or bead density is needed rather than the
2022). Skeletal density is not always a direct input in process density of the powder particle. A higher particle density ty-
models but is needed to calculate other necessary para- pically yields a higher bed density, lower porosity, and
meters such as the total adsorbent porosity. While com- greater mechanical stability. The particle densities presented
monly measured for high pressure adsorption experiments, in Fig. 3b are mostly for powders except for the polymer
only three DAC specific publications reported skeletal den- materials, which are for a single bead, and some zeolites,
sities and most of the values reported in Fig. 3a were col- which have both powder and pellet densities reported (spe-
lected from ‘non-DAC’ publications. Overall, the availability cified in Tables S1 – S6). Fig. 3b highlights the sparsity of
of skeletal density is still sparse for adsorbents considered particle density data for most materials except MOFs and
for DAC. zeolites.
2.1.2. Particle density 2.1.3. Bed density

Particle density (ρparticle) is defined as the mass of a single The bed density (ρbed) of a porous adsorbent is defined as the
particle of adsorbent divided by the envelope volume of the mass of the adsorbent divided by the volume of the vessel it
particle, which includes both the open and closed pores. It occupies. It can be measured by simply filling a container of
known volume with adsorbent and accurately weighing the (Brunauer et al., 1938). While this is the most common ap-
contents (Standard Test Methods for Apparent Density, proach seen in the reviewed DAC literature, we note that this
2017). However, due to the quantity of material needed, ty- adsorbate does not always provide accurate measurements
pically on the gram scale, this technique is not suited to non- due to the quadrupole moment of N2 and possible diffusional
commercial adsorbents. As such, much less data is available limitations in microporous materials at − 196 °C. We direct
compared to other parameters which require less sample readers to the article by Thommes et al. (2015) for a more
mass for measurement, such as the surface area and pore detailed explanation. Some publications have also reported
volume. The lower the bed density, the less adsorbent can be the Langmuir surface area as noted in Tables S1–S6, but these
placed within the vessel volume, and the lower the volu- were not included in Fig. 3d, where we only report the BET
metric CO2 adsorption capacity of the bed. However, a higher surface areas (SBET). Surface area is a parameter which lar-
bed density implies lower adsorbent porosity, which in turn gely influences the adsorption capacity of a material. Al-
could lead to restricted kinetics. Functionalisation fills the though surface area is not a direct input parameter for
adsorbent pores which thus decreases the adsorbent porosity process modelling, it remains one of the most commonly
and increases the bed and particle density. Fig. 3c shows a reported parameters in materials papers. The larger the
rather large span of mean density values, from ∼0.10 – surface area, the higher the total gas uptake. This inevitably
0.70 g cm-3. For context, commercial zeolite 13X beads from leads to a reduction in CO2 selectivity as the uptake of N2, O2,
BASF have a reported bed density of 0.65 – 0.70 g cm-3 (BASF). H2O, and all other gases present in the atmosphere increases.
Approximately half of the bed densities reported in Fig. 3c Supports with larger surface areas can accommodate higher
are for adsorbents in powder form. However, adsorbents in functionalisation loadings via grafting, which can increase
this form are often impractical to use due to their difficulty to the adsorbent’s CO2 capacity. As seen from Fig. 3d, surface
handle at scale and the resulting large pressure drop. Re- areas span a large range depending on the material and
porting the bed density of shaped adsorbents, such as beads, functionalisation typically decreases an adsorbent’s surface
pellets, fibres, or monoliths, is therefore more useful for area, sometimes by over two orders of magnitude.
process modelling. The choice of the best adsorbent struc-
ture depends on the balance of various parameters such as
volumetric working capacity, mass transfer, thermal beha- 2.2.2. Pore volume
viour, and pressure drop. The adsorbent structure selected The total pore volume (Vtot) influences many other material
will also determine the packing porosity in the bed. We direct characteristics of the adsorbent. The pore volume of ad-
readers to other sources such as those by Rezaei and Webley sorbents is often measured using N2 adsorption at - 196 °C,
for further information on the pros and cons of different though as stated in the previous section, this may not always
structured adsorbents (Rezaei and Webley, 2009) and meth- be the most appropriate adsorbate to use. The methods
odologies to compare their performance (Rezaei and employed to extract the pore size information, such as den-
Webley, 2010). sity functional theory (DFT) (Walton and Quirke, 1989;
Lastoskie et al., 2002), Horvath-Kawazoe (HK) (HorvÁTh and
2.1.4. Mechanical stability Kawazoe, 1983), or Barrett-Joyner-Halenda (BJH) (Barrett
Adsorbents with higher mechanical stability can have a et al., 2002), are only reliable in the micropore (pore dia-
longer lifetime, as they are more resistant to vibrational or meter < 2 nm) and/or mesopore (pore diameter 2 – 50 nm)
frictional forces within the bed. The mechanical stability of range. For macroporous materials (pore diameter > 50 nm)
the adsorbent can be assessed through a variety of me- such as polymeric resins, the use of N2 - 196 °C adsorption
chanical properties, such as the tensile strength, compres- alone would lead to an underestimation of the total pore
sion strength, or Youngs modulus, but can only be done for volume. In these cases, other methods capable of accurate
shaped adsorbents. Although these mechanical properties measurements in the macropore region, such as mercury
are not needed for process modelling, they help in appraising intrusion porosimetry, should be used in conjunction with N2
the robustness of the material. Assessment of mechanical - 196 °C adsorption to obtain a complete pore size distribu-
stability remains rare in the DAC literature as most ad- tion (Azzan et al., 2022). The total pore volume of the ad-
sorbents being investigated are in powder form (see Tables sorbent is needed in process modelling to calculate the total
S1-S6). From the remaining studies which involve shaped adsorbent porosity. A higher pore volume implies a higher
adsorbents, only four of them report on their mechanical porosity, which in turn leads to lower bed and particle den-
properties. The materials studied were MOF monoliths sity, and potentially lower thermal conductivity (Taddei and
(Thakkar et al., 2017), zeolite monoliths (Thakkar et al., 2016), Petit, 2021; Babaei et al., 2017) and higher heat capacity
silica-based fibres (Sujan et al., 2019), and polymer fibres (Taddei and Petit, 2021; Wieme et al., 2019), the impacts of
(Armstrong et al., 2018). which are discussed in their respective sections. The higher
the total pore volume, the greater the potential for functio-
2.2. Textural properties nalisation loading. For chemisorbents, this will increase the
CO2 saturation capacity of the adsorbent. However, as seen in
2.2.1. Surface area Fig. 3e, functionalisation can severely reduce the pore vo-
The surface area of an adsorbent is traditionally obtained by lume of an adsorbent. Comparing the data entries in Fig. 3e
collecting N2 isotherms at − 196 °C and extracting the in- vs Fig. 3f, we see that functionalisation often eliminates the
formation using the Brunauer-Emmett-Teller (BET) method presence of micropore volume (Vmicro) due to pore filling.
This could severely limit gas diffusion within the adsorbent process papers or other non-DAC articles. The temperature
and result in poor kinetics for the adsorption process. ranges of the reported thermal conductivities are indicated
in Tables S1–S6 when available. As seen in Fig. 3 h, un-
2.3. Thermal properties modified metal oxides such as alumina and semimetals such
as activated carbon (in monolith form), respectively, conduct
2.3.1. Heat capacity heat well and have high thermal conductivities (35 W m-1 °C-1
The specific heat capacity (Cp) is the amount of heat needed (Vasheghani et al., 2011) and 18 W m-1 °C-1 (Rodriguez-
to produce a unit change of temperature in a unit mass of the Mosqueda et al., 2018), respectively), while materials such as
adsorbent. This can vary with temperature and gas uptake, silica and MOFs have relatively poor thermal conductivities
and can be measured using differential scanning calorimetry (< 0.5 W m-1 °C-1).
(Mu and Walton, 2011; Standard Test Method for
Determining Specific Heat Capacity, 2018). For a temperature 2.3.3. Thermal stability
swing adsorption process, an adsorbent with a lower Cp will The thermal stability of an adsorbent is the maximum tem-
require less energy to be heated to a high temperature for perature the material can withstand before it starts to de-
desorption, leading to lower operational costs. Similarly, for grade, for example due to decomposition of the chemical
the same energy input, an adsorbent with a lower Cp can be structure or leaching of some functionalisation compounds.
heated to a higher desorption temperature, thus increasing This temperature determines the maximum temperature
the working capacity. However, adsorbents with low Cp may which can be used during the desorption process, and is ty-
also undergo a larger temperature swing during adsorption pically measured using a thermogravimetric analyser (TGA)
due to the released exothermic heat, which would reduce the either in a N2 or air environment. To be viable for most
amount adsorbed at and thus the overall working capacity. commercialised DAC processes, adsorbents will need to be
Materials with higher porosity tend to have higher heat ca- thermally stable up to at least 80 °C (Fasihi et al., 2019) to
pacity (Taddei and Petit, 2021; Wieme et al., 2019). Although withstand the regeneration procedure. For unfunctionalised
this parameter is needed for process modelling, it is a mas- materials, metal oxides, silicas, zeolites, and carbons have
sive data gap in the DAC literature, and is rarely reported in high thermal stability above 500 °C, while polymers and
materials papers. Of all the DAC papers reviewed in Tables S1 MOFs exhibit moderate thermal stability between 100 °C and
– S6, only three reported heat capacity values. All other va- 500 °C. Once adsorbents are functionalised however, the
lues shown in Fig. 3g were either found from DAC process thermal stability depends on the functionalisation com-
papers or other non-DAC articles. To provide perspective on pound, and thus functionalisation typically lowers the
the reported values, the heat capacity of commercial zeolite thermal stability of an adsorbent owing to the organic nature
13X is ∼0.8 J g-1 °C-1 (Chan et al., 2015), while that of a 20 wt% of the compounds that are commonly used for functionali-
MEA-based aqueous solvent is ∼3.9 J g-1 °C-1 (Weiland et al., sation. Functionalisation carried out via impregnation leads
1997). We note that all values reported in Fig. 3g and Tables to less stable materials than grafting due to a lack of che-
S1 – S6 are for pristine adsorbents, i.e. not loaded with CO2 mical bonds. As well, functionalised amines of lower mole-
and/or other species. The temperature ranges of the reported cular weights tend to be more prone to evaporation or
heat capacities are also indicated in Tables S1 – S6 when leaching. As seen from Fig. 3i and Tables S1 – S6, most
available. functionalised materials being considered for DAC exhibit
thermal stabilities above 100 °C. One exception is a study
2.3.2. Thermal conductivity done by Pang et al. (2017), who investigated SBA-15 ad-
Thermal conductivity (κ) is a measure of an adsorbent’s sorbents impregnated with small molecule linear or den-
ability to conduct heat. It can vary with temperature and is dritic PEI or PPI which had maximum regeneration
often determined theoretically rather than measured ex- temperatures of 70 °C. Cuesta and Song (2020) also developed
perimentally (Taddei and Petit, 2021). A higher thermal carbon black adsorbents immobilized with bovine carbonic
conductivity is desirable for adsorption if the system is being anhydrase enzymes that could be regenerated at 35 °C, but
operated isothermally, where heat needs to dissipate these needed to be refrigerated or frozen for short or long-
through the adsorbent and out of the bed. However, for term storage, respectively. Overall, thermal stability data is
adiabatic operation, lower thermal conductivity would be available for most of the investigated DAC adsorbents re-
preferable to minimise the increase in the adsorbent tem- viewed here, either from the original study or from other
perature during adsorption, which would decrease the sources.
working capacity. For desorption, a higher thermal con-
ductivity is desirable as heat can diffuse faster through the 3. Equilibrium adsorption data
adsorbent when the bed is heated, thus increasing pro-
ductivity. We note that the contactor/absorber design can An adsorbent for DAC needs to adsorb a high amount of CO2
compensate to some extent the low thermal conductivity of and a low amount of other species. This can be challenging
an adsorbent. Adsorbents with lower porosity tend to have as other species such as N2, O2, and H2O are in much higher
higher thermal conductivity (Babaei et al., 2017). Like heat concentrations in air. The adsorption properties for these
capacity, this parameter is needed for process modelling but species should therefore be investigated and are reviewed
there is limited data available for investigated DAC ad- in the following sub-sections and in Figs. 4 to 6, following
sorbents. The data entries in Fig. 3h were found in DAC the same adsorbent classification adopted in the previous
Fig. 4 – Comparison of a) CO2 uptake values (at concentrations lower than 500 ppm and temperatures between 20 and 30 °C,
except for unfunctionalised SBA-15 whose uptake of 0.01 mmol g-1 was measured at 10,000 ppm (entry 4 in Table S7) and b)
heat of adsorption values for different adsorbent categories, both with and without additional functionalisation. Box edges
represent the 25th and 75th percentile, whisker edges represent the minimum and maximum data points, the solid line
through the box represents the median, and the filled square symbol represents the mean. Total number of data entries are
labelled to the left of each box. Data entries can be found in Tables S7–S12.
Fig. 5 – Overview of CO2, N2, O2, and H2O isotherm availability for different classes of adsorbents. For CO2 and N2 isotherms,
our criterion is met if there are isotherms available at three temperatures and at pressures up to 1 bar. Isotherms that were
collected at cryogenic temperatures for CO2 and N2 were still included in this count, however we note that these are not
suitable for DAC applications and are more useful for porosity characterisation rather than isotherm fittings for process
modelling. Indication was made in Tables S7–S12 if this was the case. For O2 and H2O isotherms, our criterion is met
provided there is an isotherm available, at any temperature and pressure range below 1 bar. An adsorbent was considered to
be studied for DAC if a CO2 uptake between 0.04 vol% and 4 vol% was measured. An adsorbent was not double counted if it
was studied by multiple publications, and adsorbents with different loadings of the same functionalisation were counted as
just one material. Adsorbents of different structural forms (e.g. powder vs. monolith) were counted as two separate
materials. All data entries and their corresponding sources are available in Tables S7–S12. Isotherms were not always found
in DAC related publications. Some adsorbents included in the tables which have isotherm data available (e.g. silica gel, γ-
alumina, activated carbon) were not considered to be studied for DAC and so were not included in these counts.
section. This analysis focuses almost exclusively on single- multiple gas species can be predicted from the single-
component uptake measurements. Multi-component ad- component isotherms using approaches such as ideal ad-
sorption isotherms are experimentally challenging and sorbed solution theory (IAST) (Walton and Sholl, 2015) or
time-consuming to measure, and thus are practically non- empirically extended binary isotherm equations (Schell
existent in the DAC literature. The co-adsorption of et al., 2011).
Fig. 6 – CO2 adsorption isotherms at 25 °C and up to a pressure of 1 bar for materials which have had CO2 isotherms
measured at a minimum of three temperatures as indicated in Tables S7 – S12. Unfunctionalised adsorbents are represented
by filled markers / solid lines, and functionalised adsorbents are represented by hollow markers / dashed lines. Data points
have either been digitised, plotted using a fitted isotherm model, or plotted from tabulated data.
3.1. CO2 adsorption 2017), SIFSIX-3-Cu (Shekhah et al., 2014), and NbOFFIVE-1-Ni
(Bhatt et al., 2016) yield relatively high CO2 adsorption uptake
3.1.1. CO2 uptake at 400 ppm (∼ 1.0 – 1.5 mmol g-1) with the added advantage of lower heats
One of the key requirements for a DAC adsorbent is sufficient of adsorption compared to the functionalised MOFs. Another
CO2 working capacity, i.e. the difference between the CO2 example is the Ca2+ ion-exchanged 5 A type zeolite physi-
uptakes at adsorption and desorption conditions. In this sorbent developed by Oda et al. (2021), which shows a high
section however, we focus our discussion on CO2 adsorption CO2 uptake of 1.8 mmol g-1 at 400 ppm, attributed to each CO2
uptake, rather than working capacity, to identify trends from molecule bridging between two Ca2+ ions. This uptake is
available data in materials research. The CO2 uptake can be much higher than any of the amine-functionalised zeolites
measured using volumetric (e.g. commercial volumetric ad- reviewed in Table S11, which were reported for 5000 ppm CO2
sorption analyser), gravimetric (e.g. TGA), or dynamic set-ups concentrations. Conversely, some functionalised adsorbents
(e.g. packed bed breakthrough columns) as shown in Tables have very low CO2 uptakes (< 0.1 mmol g-1) as seen in Fig. 4a.
S7–S12. The data entries shown in Fig. 4a include uptake This is often due to insufficient loadings or unsuitable
values up to 500 ppm CO2 concentrations (see caption for one functionalisation. For example, tertiary amines have been
exception), and between 20 and 30 °C. shown to have poor CO2 uptakes in dry conditions, and their
As seen in Fig. 4a, the CO2 uptake by different adsorbents uptakes in the presence of humidity may still be lower when
ranges from as low as 0.01 mmol g-1 to over 3.0 mmol g-1. In compared to primary or secondary amines in dry conditions
general, functionalisation of the adsorbent increases its CO2 (Lee et al., 2017). Many of the highest adsorbing materials
uptake compared to its non-functionalised form, and the though are amine-functionalised materials, such as TEPA
higher the functionalisation loading, the higher the CO2 up- impregnated SBA-15 (3.4 mmol g-1) (Miao et al., 2021), mmen-
take. However, past a certain threshold, pore blockage can grafted Mg2(dopbdc) (3.1 mmol g-1) (Darunte et al., 2018), and
occur causing increased diffusion resistance and therefore hydrazine grafted Mg-MOF-74 (3.9 mmol g-1) (Liao et al., 2016).
decreased CO2 uptake. An example is TEPA-functionalised One exception is the unfunctionalised biochar adsorbent
ZSM-5 and SAPO-34, for which amine functionalisation re- synthesized from sawdust by Kua et al. (2019) which is re-
sulted in blockage of the zeolites’ micropores and prevented ported to have a gravimetric CO2 uptake of 3.2 mmol g-1 at
access to both amine and zeolite binding sites for CO2 500 ppm. However, no CO2 uptake curves were provided nor
(Thakkar et al., 2017). any material characterisation data.
It is difficult to distinguish a clear trend in CO2 uptakes As mentioned above, we included only CO2 uptake values
from physisorbents to chemisorbents. For example, some up to a concentration of 500 ppm (or 0.05 vol%) in Fig. 4a (see
unfunctionalised MOFs have comparable CO2 uptake to caption for one exception). However, in Tables S7 – S12, we
functionalised MOFs. Most of the former are classed as ul- cover publications which report the CO2 uptake at con-
tramicroporous MOFs, which have pore sizes smaller than centrations up to 4 vol%, which is the threshold for natural
0.8 nm (Kumar et al., 2015). The small pore size and presence gas applications. Although data at these higher concentra-
of fluorinated ions in MOFs such as TIFSIX-3-Ni (Kumar et al., tions may provide useful insights into the material
performance, effort should be made to conduct experiments (Rakić and Damjanović, 2013). As mentioned earlier, un-
at the relevant conditions for DAC as the trends observed functionalised adsorbents are typically physisorbents, and
may not hold true at 400 ppm CO2 concentrations. functionalised adsorbents are typically chemisorbents. As
seen in Fig. 4b, these classification holds true for most ad-
3.1.2. CO2 isotherms sorbents. Most amine-functionalised adsorbents exhibit
A key process model input is the numerical expression of the heats of adsorption greater than 70 kJ mol-1, while most un-
CO2 adsorption isotherm, which is obtained by fitting ad- functionalised adsorbents exhibit heats of adsorption lower
sorption data collected at constant temperature. The ma- than 50 kJ mol-1.
jority of the uncertainty in a process model stems from the There remain exceptions to the classification mentioned
parameters of the adsorption isotherm model, but this above. In some cases, unfunctionalised materials are che-
variability can be reduced by collecting data at several tem- misorbents and display high heats of adsorption. Examples
peratures and over a wide range of temperatures (Ward and include the bare γ-alumina supports, on which CO2 binds and
Pini, 2022). In Fig. 5, we consider as criterion that at least form bicarbonate and carbonate species on the basic alumina
three temperatures be measured for CO2 isotherms. Data surface (Potter et al., 2017), and the benzotriazolate MOFs
meeting this requirement remains largely unavailable/un- with metal hydroxyl sites which allow for CO2 chemisorption
reported for most proposed DAC adsorbents. This data gap is (Bien et al., 2018; Bien et al., 2019; Cai et al., 2020). Some
one which requires significant attention. Fig. 6 displays CO2 physisorbents, such as the ultramicroporous MOF families of
isotherms at 25 °C for different materials which have iso- TIFSIX (Kumar et al., 2017), SIFSIX (Shekhah et al., 2014, 2015;
therms measured at three or more temperatures and up to Mukherjee et al., 2019), and NbOFFIVE (Bhatt et al., 2016), also
1 bar. We note that materials with the highest CO2 uptake at exhibit heat of adsorption values bordering either side of
1 bar do not necessarily have the highest uptake at 400 ppm 50 kJ mol-1. The alkaline-earth-metal ion-exchanged MFI-
(e.g. zeolites compared to silicas). Rather, materials with type zeolites display even higher heats of adsorption values
steep isotherms in the low-pressure range tend to have high between 80 and 100 kJ mol-1 due to the strong electrostatic
CO2 uptake at the relevant concentrations for DAC. We also interactions between acidic CO2 and the M2+-O-M2+ sites of
observe that some studies do not collect sufficient isotherm the adsorbent, where M2+ is the alkaline-earth-metal ion
points in the low-pressure range. While gathering this data (Itadani et al., 2016). Some chemisorbents have un-
may be time consuming, it is essential to ensure the iso- characteristically low heats of adsorption. One example is a
therm model fitting is accurate in the relevant pressure quaternary ammonium-based polymer adsorbent which
range. Additionally, even if some publications do collect features a heat of adsorption of 30 kJ mol-1 due to the pre-
isotherms at multiple temperatures, they do not always do so sence of moisture to enable CO2 adsorption (Hou et al., 2021).
up to 1 bar. Although adsorption for DAC occurs at 400 ppm, General trends are also observed with the heat of ad-
the adsorption information up to 1 bar becomes relevant for sorption versus CO2 loading. In most cases, the heat of ad-
the desorption process, where a CO2-enriched stream is sorption is highest at zero loading, and decreases as the
produced. loading increases. This pattern implies the presence of het-
erogenous CO2 binding sites, where the stronger sites are
3.1.3. Heat of adsorption filled first. However, some adsorbents maintain a constant
The CO2 heat of adsorption gives an indication of the heat of adsorption over a wide loading range. This is the case
strength of adsorption between the material and CO2, and is with some members of the SIFSIX family, which according to
a necessary input in process models. The higher the value, the authors, suggests the presence of homogenous energetic
the more likely the adsorbent will be able to selectively ad- sites throughout the adsorbent (Shekhah et al., 2015). For
sorb CO2 even at dilute concentrations. Yet, a high heat of other materials, the heat of adsorption increases with
adsorption also implies higher energy to regenerate the ad- loading. An example of this is the diamine-appended
sorbent and thus higher operating costs. The heat of ad- Mg2(dobpdc) MOF developed by McDonald et al., where the
sorption is traditionally obtained by measuring three or more increase in heat of adsorption is attributed to cooperative
CO2 isotherms spaced 10 °C apart, and by either applying the CO2 adsorption (McDonald et al., 2012, 2015).
van’t Hoff equation or virial analysis (Rouquerol et al., 1999; Despite the importance of the heat of adsorption, there is
Nuhnen and Janiak, 2020). For the former, the heat of ad- still a big data gap in this area, as can be seen in Tables
sorption at a specific CO2 loading can be determined from the S7–S12, and when comparing Fig. 4a and b. Except for MOFs
slope of a plot of ln(P) vs 1/T at a given amount adsorbed, and zeolites, most adsorbents in all material classes do not
where P is the pressure and T is the temperature. For the have reported heats of adsorption, which corresponds with
latter, the adsorption isotherm data is fitted directly with a the data gaps in the CO2 isotherms.
virial expansion, generating virial coefficients from which
the heat of adsorption can be determined. The isosteric heat 3.2. N2 adsorption
of adsorption is the heat of adsorption expressed as a func-
tion of the CO2 loading. At zero loading, one refers to it as the N2 isotherm information should be included in process
limiting heat of adsorption (ΔH0). Tables S7 – S12 list the models for DAC as N2 accounts for over 78 vol% of the feed
limiting heats of adsorption of candidate DAC adsorbents, gas composition (i.e. atmospheric air). For physisorbents, N2
along with the heats of adsorption at loadings corresponding is a competitive adsorbate against CO2 and its uptake will
to the adsorbents' CO2 uptake at DAC concentrations if the reduce the CO2 capacity of the adsorbent. For chemisorbents,
values greatly differ. All data entries are included in Fig. 4b to this is often less of an issue, however some N2 adsorption can
show the entire possible range of heat of adsorption values. still occur on unfunctionalised surfaces. Even if only a slight
The general guideline is that materials with heat of adsorp- amount is adsorbed during the process, performance in-
tion values less than 50 kJ mol-1 are classed as physisorbents dicators such as CO2 working capacity and purity may be
and those with higher values are classed as chemisorbents impacted. Therefore, like for CO2, one should collect N2
sorption isotherms at a minimum of three temperatures to low O2 uptake from their measured isotherms. Where pos-
parameterise a suitable adsorption isotherm model and to sible, we include the % RH and temperature conditions used
calculate the N2 heat of adsorption. Isotherms up to 1 bar for these oxidation studies.
should be collected to cover the possible concentration range
of N2 throughout the process. As seen in Fig. 5, it is quite rare 3.4. H2O adsorption
to find N2 isotherms at three temperatures up to 1 bar for
DAC adsorbents, even when considering non-DAC related 3.4.1. H2O isotherms
papers. Most commonly, only one isotherm is measured and Water is always present in the atmosphere to some degree
the impact of N2 on the process is assumed to be negligible. (see Fig. 2) and can compete with CO2, occupying the same
binding sites, and reducing the CO2 saturation capacity. In
3.3. O2 adsorption some cases, H2O increases the CO2 capacity of the adsorbent
(see Hydrolytic stability sub-section below). Regardless of the
3.3.1. O2 isotherms situation – competitive or synergistic adsorption – a material
O2 accounts for over 20 vol% of the gas composition in the adsorbing a high amount of H2O will require higher energy
atmosphere, and like N2, can be a competitive adsorbate for full desorption/regeneration. Therefore, we recommend
particularly for physisorbents. Hence, one should study its for at least one H2O adsorption isotherm to be measured. As
adsorption behaviour. If the adsorption is non-negligible, one for O2, if the H2O uptake is non-negligible, isotherms at three
should collect isotherms at three temperatures up to 1 bar to temperatures up to 1 bar should be measured to obtain the
allow for isotherm fitting and heat of adsorption calculation. necessary information for process modelling. As seen in
In addition, O2 can react with and damage certain adsorbents Fig. 5, collecting at least one H2O isotherm is more frequent
(see Oxidative stability sub-section below). Assessing the ex- than for N2 or O2. Yet, there is still a significant lack of data in
tent of O2 adsorption is therefore useful to understand how this area for candidate DAC adsorbents, even when
much O2 can potentially interact with the material. As shown searching in non-DAC related publications. Due to the com-
in Fig. 5, there is an obvious lack of data in the O2 isotherms plex interactions that can occur between CO2 and H2O as well
available for candidate DAC adsorbents, even when con- as the sorbent, measuring CO2-H2O co-adsorption equili-
sidering non-DAC related publications. brium isotherms is critical. Such measurement represents
more accurately the impact of H2O on DAC processes com-
3.3.2. Oxidative stability pared to predictions from single-component sorption data.
Oxidative stability is a necessary feature of adsorbents for However, this binary sorption isotherms are experimentally
DAC. Amine-functionalised adsorbents can be susceptible to challenging to measure and remain largely unavailable in the
oxidative degradation particularly at high temperatures DAC literature, save for a few studies such as those by Gebald
(i.e. > 75 °C) after just one day, resulting in the loss of CO2 et al. (2014), Elfving and Sainio (2021), and Young et al. (2021).
adsorption sites (Heydari-Gorji and Sayari, 2012). This can Examples of CO2-H2O co-adsorption models that have been
also occur at lower temperatures over longer periods of time, used to model this data include the Guggenheim–Anderson
and under ambient storage conditions (van Paasen et al., de Boer (GAB) model (Stampi-Bombelli et al., 2020), a kinetic
2021). Primary and tertiary amines are known to be more co-adsorption model (Elfving and Sainio, 2021), and a me-
resistant to oxidation compared to secondary amines, which chanistic co-adsorption model and a weighted average dual
can readily degrade to form imines or amides (Bollini et al., site Toth (WADST) co-adsorption model (Young et al., 2021),
2011). For aminopolymers, the presence of additional car- respectively. These models are discussed in further detail in
bons between amine groups (e.g. propylene vs. ethylene Section 7.1.
linkers) also imparts higher oxidation stability (Pang et al.,
2017). Many publications listed in Tables S7–S12 have fo- 3.4.2. Hydrolytic stability
cused specifically on developing DAC adsorbents with im- The presence of humidity can have various impacts on an
proved oxidative stability, such as those by Pang et al. (2018), adsorbent, which we have categorized in Tables S7–S12
Rosu et al. (2020), and Yoo et al. (2019). However, other O2 where data is available. “N” indicates the worst-case scenario
stability studies are limited to either measuring one O2 ad- where the structure of the adsorbent itself has deteriorated
sorption isotherm and claiming minimal adsorption, or due to the collapse of the crystal and/or pore structure and is
running their experiments using an ambient air stream. In unregenerable. “Y” indicates that the adsorbent structure,
these cases, further investigation is still needed to look at the including its functionalisation, is unaffected or is regenerable
impact of O2 at higher temperatures as well as long-term after humidity exposure. Where possible, these are then
exposure, as this will be important for the desorption process further classified into “+ ” or “-” categories, which indicate
and shelf-life, respectively. whether the CO2 uptake increases or decreases in the pre-
We have summarised the impact of O2 on adsorbents in sence of humidity, respectively, along with the % RH and
Tables S7–S12 where data is available, although there is temperatures used for these studies. Where data is available,
clearly a data gap on this topic in the current literature. If the we also identify if the material was stable under steam re-
impact of O2 on the CO2 uptake was provided, we identify if generation (“ss”), which may have benefits from a process
more than (“ > 50% CO2”) or less than 50% (“ < 50% CO2”) of the standpoint (see Section 7.1). For the purposes of simplicity,
CO2 uptake was maintained, and if these studies were con- we did not consider the effect of humidity on the CO2 ad-
ducted at CO2 concentrations higher than 4 vol%. For papers sorption kinetics in this categorisation. This method of ca-
which looked at the structural integrity of the adsorbent, “N” tegorisation follows that proposed by Kong et al. in their
indicates that either the adsorbent support or functionali- recent perspective article on DAC (Kong et al., 2022).
sation material degrades in the presence of oxygen, while “Y” In general, the CO2 uptake for amine-functionalised ad-
indicates that the adsorbent appears to be structurally stable sorbents increases in the presence of humidity due to a
or regenerable. We have also indicated if papers reported a change in the adsorption mechanism. With water, only one
N atom is theoretically needed to adsorb CO2 via bicarbonate carried out at DAC conditions as indicated in the tables. Most
formation, instead of two N atoms in dry conditions via of the studies measure parameters such as rates of adsorp-
carbamate formation (Sanz-Pérez et al., 2016). Tertiary tion, or breakthrough, retention, or saturation times. These
amines cannot react to form carbamates, and only show parameters are dependent on many experimental factors
substantial CO2 adsorption in the presence of water vapour such as gas flow rates, column dimensions, system tem-
via bicarbonate formation (Sanz-Pérez et al., 2016). Moisture perature and pressure, and feed gas concentration. This
also helps to maintain the long-term stability of amine- makes it virtually impossible to compare the kinetic results
functionalised adsorbents by preventing the formation of of different publications. Overall, in-depth kinetic studies are
urea, which would otherwise eliminate CO2 adsorption sites much more common in process-focused papers rather than
(Kim et al., 2016). Some adsorbents even require the presence materials-focused papers.
of moisture for CO2 adsorption to occur. Examples are those
functionalised with sodium or potassium carbonate, which 4.2. Cyclic studies
react with CO2 in the presence of moisture to form sodium or
potassium bicarbonate (Rodríguez-Mosqueda et al., 2019). For To be usable at scale, DAC adsorbents must exhibit sufficient
most physisorbents though, e.g. unfunctionalised MOFs, structural and chemical longevity upon repeated cyclic ex-
zeolites, and carbons, the presence of humidity reduces their posure to adsorption and desorption conditions, while still
CO2 uptake due to the competitive adsorption of water on maintaining their working capacity throughout their life-
CO2 adsorption sites (Kolle et al., 2021). Some adsorbents are time. The less adsorbent that needs to be replaced
also simply not structurally stable in the presence of water throughout the process, the lower the cost. We summarise
vapour, such as SIFSIX-3-Cu (Madden et al., 2017), TIFSIX-3- the presence (“Y”) or absence (“N”) of cyclic studies in the
Co (Kumar, 2018), and Mg-MOF-74 (Kumar et al., 2015). De- reviewed publications for DAC adsorbents in Tables S7–S12.
pending on the extent of the impact of moisture on these Overall, over 85% of the studies report some form of cyclic
adsorbents, the removal of water may need to be in- testing, although not all are carried out at DAC conditions as
corporated into the process which would increase operating indicated in the tables. Among them, the packed bed break-
costs. For further in-depth explanations of the effect of water through method prevails, followed by TGA. These methods
on CO2 adsorption, we direct readers to the recent review by commonly employ a simple temperature swing adsorption
Kolle et al. (2021). (TSA) process, where CO2 adsorption occurs at ambient
For the chemisorbents and physisorbents reviewed here, temperature, followed by desorption at elevated tempera-
the effect of humidity is not always investigated at low CO2 tures (∼70 – 120 °C) using either an inert gas or air stream.
concentrations relevant to DAC, as indicated in Tables These steps are repeated over several cycles, though usually
S7–S12. While an adsorbent may maintain high CO2 capacity no more than ten. Some publications have also investigated
in the presence of moisture at high concentrations, the effect multiple cycles of moisture swing adsorption for adsorbents
of humidity may be more severe when the CO2 concentration (Hou et al., 2021; Song et al., 2019; Wang et al., 2020). A recent
is reduced to 400 ppm. Therefore, one should conduct ex- study has also explored dual functional materials capable of
periments at the relevant conditions whenever possible. CO2 adsorption followed by methanation when exposed to
H2, where the process would cycle between the two steps
4. Beyond CO2 uptake measurements (Veselovskaya et al., 2018; Jeong-Potter and Farrauto, 2021).
We discuss other types of desorption processes further in
Determining if an adsorbent is suitable for DAC requires in- Section 7.1.
formation beyond the measurement of its CO2 uptake. We
have already noted above the importance of considering the 5. Synergy between material and process
impact of other atmospheric gases. Dynamic adsorption development in DAC
studies are needed to determine adsorption kinetics, while
cyclic studies are used to estimate the working capacity of Since ultimately, an adsorbent must be integrated into an
the adsorbent and its potential longevity. adsorption process, material and process development
should occur harmoniously. Applying a promising material
4.1. Kinetic studies in a process model or experiment provides understanding of
how the material properties can influence the process per-
In contrast to other separation processes, early breakthrough formance indicators. Between 2016 and 2021, we counted 84
of CO2 is not a key concern for DAC as any specification on “material studies” focusing on adsorbent development
recovery may be less stringent than with e.g. post-combus- (Fig. 7b) and the number of publications per year roughly
tion capture (Stampi-Bombelli et al., 2020). Therefore, ad- doubled within this timeframe, predominantly influenced by
sorbents with slower kinetics may still be suitable for a DAC an increasing interest in silica, metal oxide, polymer, and
process. Mass transfer coefficients which describe the gas MOF adsorbents. On the other hand, “process studies”
diffusion and the adsorption/desorption kinetics are needed (Fig. 7a), defined as studies reporting process performance
for process modelling. The experiments needed to determine metrics, predominantly use amine functionalised polymers.
these parameters are not systematically conducted. One Many promising MOF- and silica-based adsorbents are not
reason for this might be that they often require a mathe- screened/assessed for a given adsorption process, leaving a
matical model for interpretation. We summarise the pre- gap in understanding their suitability for DAC processes.
sence (“Y”) or absence (“N”) of kinetic studies in the reviewed A common adsorbent applied in process studies is
publications for DAC adsorbents in Tables S7–S12. Over 90% Lewatit® VP 1065, an amine-functionalised polymer. In this
of studies include some form of kinetic study, with packed case, a temperature-dependent isotherm model and heat of
bed breakthrough studies being the most common set up, adsorption (Bos et al., 2019; Veneman et al., 2015), a mass
followed by TGA. However, not all the kinetic studies are transfer coefficient for a kinetic model (Bos et al., 2019), the
Fig. 7 – Comparison of the number of studies using different material classes in (a) process studies and (b) materials studies.
Material studies refer to studies focusing on adsorbent development. Process studies refer to studies reporting process
performance metrics or techno-economic assessments. Some materials studies investigated more than one material
category (e.g. silica and metal oxide), but these were not double-counted in the final total.
heat capacity (Sonnleitner and Hofbauer, 2018), the heat of development. For this reason, when materials exhibit pro-
adsorption (Veneman et al., 2015), and material pellet and mising equilibrium capacity and stability, focus should be on
bed density are available in the literature (Young et al., 2021). obtaining the complete set of information that is required for
Even in the case of Lewatit® VP 1065, the mass transfer ex- process modelling as indicated in our previous sections.
periments completed by Bos et al. (2019) are performed
at > 50 mbar CO2 where DAC conditions are at 0.4 mbar, 6. Adsorber/contactor design
therefore the mass transfer coefficient which is commonly
applied in the Linear Driving force (LDF) model is not at DAC To achieve the large volumes of air flow required for DAC
conditions. An amine-functionalised nanofibrilated cellulose (see Fig. 1), while maintaining a low process specific energy,
material (APDES-NFC) which is utilised by Climeworks has an adsorbent-air contactor with low pressure drop, efficient
been applied in fewer process studies as this material is not heat and mass transfer, low footprint and low capital cost is
as readily accessible. The isotherms and heat of adsorption essential.
of this material are available in the literature (Gebald et al., Here, we consider three categories of contactors for ad-
2014). The full extent of material characterisation that is re- sorption processes (see Fig. 8): fixed bed, moving bed, and
quired for application in a process model is rarely completed, fluidised bed. A fixed bed can be described by a vessel with a
where full characterisation of materials is imperative to im- packing or structured adsorbent material where both the
prove the synergy between material and process vessel and material are stationary. Fixed beds are operated in
Fig. 8 – Schematic representation of the different process configurations and associated contactor designs for continuous
adsorption processes. Contactor configurations are different ways in which these processes can be implemented.
a batch-wise cycle, undergoing adsorption, desorption/ between structured and packed pellet adsorbents. Therefore,
heating, and pressurisation/cooling in at least three separate the benefit of structured adsorbents in DAC likely depends on
stages. Therefore, multiple vessels are required for con- material kinetics and controlling mass transfer resistances.
tinuous operation. In the case of a moving bed, the vessel Additional sensible heat requirements for structured ad-
containing the solid adsorbent and the sorbent move to- sorbents are observed because of the support structure
gether through adsorption and desorption steps. This allows (Sinha et al., 2017), however, heat transfer benefits may be
a continuous process without the requirement of multiple observed from improved uniformity of adsorbent heating
vessels. In the case of a fluidised bed, the pellet, powder, or and thermal conductivity of the structure. The trade-off be-
spherical adsorbent is fluidised and the adsorbent itself tween additional sensible heat requirements and improved
moves to different process stages. This is also a continuous heat transfer will depend on structure properties, for ex-
process where different physical locations of the process are ample thermal conductivity and adsorbent loading (Rezaei
dedicated to adsorption or desorption. and Webley, 2010). Similarly, all improvements in pressure
To understand the requirements of process configuration drop, mass transfer, and heat transfer must be evaluated
and contactor design, it is useful to assess the current com- against the often low volumetric adsorbent loading in
mercial DAC plant designs. The latter have clearly indicated structured adsorbents (Rezaei and Webley, 2010), which once
that contactor optimisation is a key part of the design of the again highly depends on the structure design. Adsorbent
adsorption-based DAC process. The patented Global loading is particularly relevant for DAC where it is already
Thermostat design uses a hybrid moving and fixed bed pro- very difficult to achieve high adsorbate loadings at the ultra-
cess (Eisenberger, 2018). The configuration consists of a dilute feed conditions.
monolithic structured adsorbent in a fixed bed configuration, Many DAC process studies utilise a fixed bed adsorber for
where multiple fixed beds are supported on a track. These modelling and experimentation. Sinha et al. (2017) and Zhu
beds move around the track, albeit not continuously, where et al. (2021) performed studies on two MOF adsorbents coated
one bed is placed in a sealed box for the desorption step once on a monolith structure using amine functionalised MOFs,
saturated. The design of this system, in terms of number of i.e. MIL-101(Cr)-PEI-800 and mmen-Mg2(dobpdc). Both stu-
fixed beds, is dependent on the relative adsorption (time dies utilised a steam assisted temperature vacuum swing (S-
until saturation) and desorption (time until adsorbent re- TVS) process for adsorbent regeneration with the monolith
generation) time of the process. The sequential operation of structure in an axial bed. Sinha et al. (2017) reported the
two of these process configurations allows heat integration monolith wall sensible heat requirement during desorption
of the two regeneration systems where the depressurization as 15% of the total process energy requirement, and dis-
and therefore cooling of one system pre-heats the second cussed that reducing the monolith wall thickness could offer
system. significant energy benefits. Schellevis et al. (2021), Stampi-
Climeworks have patented a fixed bed vacuum chamber Bombelli et al. (2020), and Young et al. (2021) employed a
which is packed with a structured adsorbent (Gebald et al., pancake style packed fixed bed adsorber for process model-
2017). The design of the contactor is intended to minimise ling, where the bed diameter is 8–16 times that of the
pressure drop and thermal mass to improve heat transfer. thickness. In similar fixed bed studies for post-combustion
The design has adequate mechanical strength for low va- carbon capture, this ratio is typically 0.1 or less (Zhao et al.,
cuum pressures (500 mbar) using ribbing to increase the 2019a; Sarkar and Bhattacharyya, 2012; Dantas et al., 2011). A
strength, with an optimised wall thicknesses to reduce pancake style contactor is particularly appropriate for DAC
thermal mass. Inlet and outlet axial walls with sealing ele- as the large thin design allows minimal pressure drop for a
ments are used to allow gas to flow through when open and given amount of adsorbent volume. While a thin process
to provide a seal to allow evacuation to vacuum pressures contactor can result in premature breakthrough, uniquely to
when closed. A gas tight sealing mechanism is also provided DAC, one can envisage a process design without specifying a
for the adsorbent structure to ensure no air bypasses around constraint on recovery. A recent study by Yu and Brilman
the adsorbent. Heat transfer fluids pass through the cir- (2020) proposed to use a radial flow contactor, where the
cumferential walls of the vacuum chamber. Considerations adsorbent is packed in an annulus, and the air flows radially
of this design also include optimisation of void volume based through the bed (from the outside to the inside of the an-
on the trade-off between pressure drop, dilution, and pump nulus). This configuration allows fluid flow to occur across a
work (Gebald et al., 2017). Beyond the selection of a process thin bed, like in the case of a pancake fixed bed, however,
configuration, there remains considerable scope for con- also permits a longer column. This design was reported to
tactor optimisation. achieve a pressure drop of 348–681 Pa for an air flow of
Although structured adsorbents are used in current 180–320 m3h-1 which was compared to 100 Pa for a monolith
commercial DAC process plants, their benefit in the case of bed based on a 2012 study by Kulkarni and Sholl (2012).
DAC remains unclear in the literature. The answer to this Fewer studies have focused on moving bed or fluidised
question may present itself as more complicated than an- bed process configurations. A study by Drechsler and Agar
ticipated. Much of the performance of structured adsorbents, (2019) introducing a moving belt adsorber design using
namely laminates, monoliths, and foams, is dependent on temperature swing adsorption (TSA) presented a highly heat
adsorbent geometry, adsorbent loading, binder, and thermal integrated process with Power-to-Gas technology (methana-
properties (Rezaei and Webley, 2009, 2010; Sinha et al., 2017). tion of CO2 using the Sabatier reaction). The benefit of this
An optimisation study by Rezaei and Webley (2010) high- process was observed in its heat integrated and autothermal
lights that the enhanced productivity that results from re- design. However, the productivity was only 31.2 kg CO2 per
duced pressure drop in structured adsorbents becomes the m3 of sorbent per day (kg CO2 (m)-3.(day)-1) and the maximum
over-arching effect at higher superficial velocities. At lower purity was 20 mol% CO2. Although this design is well adapted
velocities, the smaller mass-transfer zone and low pressure to the specific application presented, the moving belt is not
drop suggests that there is little productivity difference suitable for CO2 storage applications in its current form due
7. Process design
Table 3 – Key desorption methods applied in DAC
process studies.
7.1. Desorption method
TSA Temperature swing adsorption
TCSA Temperature concentration swing adsorption
C-TSA Closed temperature swing adsorption The desorption method significantly impacts process per-
S-TSA Steam-assisted temperature swing adsorption formance and can make up at least 40 – 98% of total process
TVSA Temperature vacuum swing adsorption energy requirements (Young et al., 2021; Sinha et al., 2017;
TVCSA Temperature vacuum concentration swing Zhu et al., 2021; Schellevis et al., 2021; Kulkarni and Sholl,
adsorption 2012; Wilson and Tezel, 2020), not including subsequent CO2
C-TVSA Closed temperature vacuum swing adsorption
compression. Such method should be selected based on the
S-TVSA Steam-assisted temperature vacuum swing
process constraints and location. Specifically, factors that
adsorption
MSA Moisture swing adsorption influence the selection include CO2 product purity and pro-
pHSA pH swing adsorption ductivity requirements, accessibility to waste heat or low
pressure (LP) steam, adsorbent material isotherms and tem-
perature and moisture stability. Several desorption methods
have been studied for DAC and are listed in Table 3.
low purity and productivity. Zhang et al. (2014) performed a We summarise the main desorption options for DAC
study on a fluidised bed using PEI-Silica, where the specific processes and their ranked purity and energy use in Fig. 9,
energy was reported as 6.6 MJ (kg CO2)-1 which is comparable and highlight the trade-off between high purity and low
with other studies (see Table 3), and a cost of $152 (t-CO2)-1 in specific energy. Climeworks, like most reported DAC studies,
2015. The reported productivity was only 7 kg CO2 (m)-3(day)- employs TVSA to achieve a high purity CO2 product
1
, which can be considered to be relatively low. The pressure (Gutknecht et al.). On the other hand, Global Thermostat uses
drop of the fluidised bed reached 1592 Pa compared to only a S-TSA process (Eric Ping et al.). Both desorption methods
348–681 Pa in the case of a radial flow adsorber, indicating a produce high purity CO2, with the intention of long-term
fluidised bed is unlikely to be feasible for a DAC process (Yu subsurface storage. Desorption methods and conditions used
and Brilman, 2020; Zhang et al., 2014). can have a significant impact on process energy require-
Although pancake-like and radial flow bed designs represent ments.
promising options for DAC, there remains significant gaps in Elfving et al. (2021) compared different desorption
assessing process configurations and optimisation of contactor methods for DAC using an aminoresin in a fixed bed ex-
designs. Further research is required to understand the applic- perimental unit. Recognising that not all DAC processes may
ability of different moving bed designs to DAC. Moving bed require high purity CO2, their study highlights that in low
process configurations have the potential to offer improved purity applications, TSA and TCSA offer a viable option with
process cycle time control which may be pertinent in DAC due lower specific energy requirements than desorption methods
to fluctuating ambient conditions. Furthermore, a moving bed that require either vacuum or steam. In this study, in-
can bring in further potential for heat integration as discussed troduction of a mild vacuum (500 mbar) to a TCSA process
in the moving belt design presented by Drechsler and Agar increases the specific energy requirement by 43% but only
(2019) and the patented Global Thermostat technology improves productivity by 14%. When considering desorption,
(Eisenberger, 2018). Moving bed designs may have potential downstream utilisation of CO2 is key. Purities of 5–20 mol%
process intensification benefits, however, the process in- have been reported using TSA, 95–97 mol% using C/S-TVSA
tensification and capital benefit are dependent on the design of and 95–99.6 mol% in desorption processes with a steam
a specific contactor, and data is lacking to allow comparison purge (Young et al., 2021; Stampi-Bombelli et al., 2020; Sinha
between fixed and moving beds for DAC. Although a moving et al., 2017; Zhu et al., 2021; Drechsler and Agar, 2019; Santori
bed can offer the benefit of achieving a continuous process with et al., 2018; Sinha and Realff, 2019; Brilman et al., 2013).
one vessel, continuous operation may not be necessary and is Therefore, the applicability of a TSA process is limited due to
therefore not the sole distinguishing factor between a moving a low purity product. Unsurprisingly, desorption processes
or fixed bed process. Continuous DAC processes may become which use both steam and vacuum conditions tend to result
pertinent if downstream utilisation requires a stable and reli- in higher specific energy requirements than TSA processes.
able feed stream. An interesting possibility of a moving bed TVSA processes can achieve higher productivity when com-
design is changing the contactor conditions for different cycle pared to simply TSA based processes due to rapid desorption
steps. For example, in the Global Thermostat design, the con- rates and higher purity, especially when coupled with a
tactors are only isolated from air during the desorption step and purge flow (Elfving et al., 2021; Wijesiri et al., 2019a). Wijesiri
exposed during the adsorption step (Eisenberger, 2018). et al. (2019a) found through a fixed bed experiment that the
As demonstrated by the Climeworks fixed bed tempera- desorption rate in a S-TVSA process using a PEI impregnated
ture vacuum swing (TVS) contactor design, there is in-depth silica was sixteen times higher than a TVSA process. The
scope for contactor optimisation within different process experimental study performed by Elfving et al. (2021) also
configurations (Gebald et al., 2017). The design of compact found productivity benefits when coupling TVSA with an air
and efficient contactors is essential in DAC. Compact process or inert purge.
designs with efficient heat and mass transfer as well as heat When considering desorption conditions, a wide range of
integration will reduce process specific energy, increase vacuum pressures has been studied from 20 to 900 mbar
productivity, and minimise capital cost. This may be parti- (Stampi-Bombelli et al., 2020; Sinha et al., 2017; Zhu et al.,
cularly important when considering that DAC processes may 2021; Schellevis et al., 2021; Wilson and Tezel, 2020; Wijesiri
become brownfield additions to existing processes where et al., 2019a; Wurzbacher et al., 2016; Zhao et al., 2019b;
excess waste heat is available, and therefore a compact and Wurzbacher et al., 2012). The desorption temperature varies
modular design will be key for easy installation. between 60 and 120 °C (Stampi-Bombelli et al., 2020; Sinha
Fig. 9 – Schematic representation of the different process configurations for DAC and qualitative assessment of their energy
requirement and achieved purity.
et al., 2017; Zhu et al., 2021; Schellevis et al., 2021; Drechsler Zhu et al. (2021) observed the opposite effect when using a
and Agar, 2019; Elfving et al., 2021; Wijesiri et al., 2019a; higher desorption pressure. The higher desorption pressure
Wurzbacher et al., 2016; Zhao et al., 2019b; Wurzbacher et al., resulted in some adsorption (instead of desorption) of water,
2012) for amine functionalised sorbents and 95–260 °C for leading to reduced heating time and improved kinetics. The
zeolites (Wilson and Tezel, 2020; Santori et al., 2018), this impact of these effects in process modelling on process
parameter being often governed by material stability (Young performance indicator results has been shown to vary de-
et al., 2021; Yu et al., 2017). The selected vacuum conditions pending on the type of co-adsorption model used (Young
represent a trade-off between energy use and the pro- et al., 2021). Young et al. (2021) presented two co-adsorption
ductivity (Schellevis et al., 2021; Elfving et al., 2021). models – the mechanistic co-adsorption model and weighted
Schellevis et al. (2021) have shown that a higher adsorption average dual site Toth (WADST) – and compared this with the
temperature (up to the material stability limit) at a constant Guggenheim–Anderson de Boer (GAB) model used by Stampi-
material working capacity, resulted in lower process energy Bombelli et al. (2020). Process key performance indicators
consumption and higher productivity. When a desorption (KPI) were similar for the mechanistic and WADST models,
temperature of 120 °C instead of 60 °C was used the process however, the GAB model varied significantly to these models
specific energy reduced by 40% and productivity increased by (optimum productivity ∼25% lower and optimum specific
a factor of at least two. The observations resulted from the work equivalent was ∼44% more). Young et al. (2021) sug-
energy intensity of vacuum compression and the require- gested that this is due to the lack of temperature dependence
ment of harsher vacuum conditions at lower desorption of the GAB model. Elfving and Sainio (2021) proposed a ki-
temperatures to achieve a desired working capacity. A si- netic based model for water co-adsorption based on the rate
milar result was obtained by Elfving et al. (2021) when com- of reaction of two reactions suspected to occur during che-
paring conditions for a TVSA or TVCSA process, where misorption of CO2 on a primary amine sorbent. The results of
milder vacuum conditions (500 mbar compared to 25 mbar) this model agreed well with the measured isotherms. Use of
resulted in a specific energy (MJ (kgCO2)-1) reduction of more such a model to determine process performance indicators
than 70%. This effect was accompanied by a small difference for comparison with other proposed models would be a
in working capacity (1.6% more for TVSA and 6% less for useful exercise.
TVCSA), and a reduction in productivity by 10–14%. Moisture and pH swing adsorption have also been ex-
A key aspect to consider in optimising desorption condi- plored for DAC (Cuesta and Song, 2020; Hou et al., 2017), yet
tions, particularly when considering amine-functionalised not commonly applied in process studies. Vacuum or tem-
sorbents, is the co-adsorption of water. Wijesiri et al. (2019a) perature swing tends to be applicable to a wide range of
reported that pre-adsorbed water caused increased con- materials, making them the more common desorption
sumption of thermal energy due to desorption of water, and method. Hou et al. (2017) utilised moisture swing adsorption
also resulted in slower desorption kinetics of CO2 at the be- in a greenhouse application, where low purities of approxi-
ginning of the desorption step due to the initial temperature mately 3 mol% were required, making this process at present
drop. Interestingly, mathematical modelling performed by only applicable for low purity CO2 requirements. This process
used an ammonium-based anion exchange resin where CO2 processes. For instance, we observe specific energy values
is displaced upon wetting the adsorbent. Cuesta and Song from 0.7 to 57 MJ (kg CO2)-1. Large differences in performance
(2020) presented an experimental study utilising pHSA on a metrics for seemingly similar processes can be explained
functionalised carbon black. In the first cycle phase, CO2 is either by the process operating parameters, in particular the
captured on the basic sites lowering the pH. A regeneration desorption conditions used, the methods which are used to
solution, i.e. ammonium bicarbonate, is subsequently va- calculate these metrics, or a combination of the two. For
pourised into the system causing desorption of CO2 and in- example, Young et al. (2021) and Schellevis et al. (2021) re-
creasing the pH of the system. The system is then dried and ported a specific energy of 2.5 and 14 MJ (kg CO2)-1 respec-
returned to the original state. The benefits of pHSA include: tively, both using a pancake fixed bed packed with Lewatit®
short cycle times, comparable sorbent capacity to TSA and VP 1065. Young et al. (2021) used a closed instead of open
MSA, and low energy requirements. As there is currently only TVSA process, different desorption conditions and a different
one study discussing a pHSA process, further understanding water co-adsorption model, all of which influence the spe-
is required of the potential scalability of this process. cific energy estimations. The reported productivities of pro-
cesses vary less than the specific energy. However, the
7.2. Performance indicators highest value reported (245.8 kg CO2 (m)-3(day)-1) is seven
times that of the lowest reported value of 7.89 kg CO2 (m)-
3
Assessment, comparison, and optimisation of DAC processes (day)-1. For working capacity, which has fewer reported va-
requires process KPIs (see Table 4). These KPIs typically in- lues, we note a five-fold variation between highest and
clude working capacity of the sorbent (WC), productivity of lowest values. Purity for a given process is either very high
CO2 for a given quantity of sorbent, purity of the product CO2 (> 95–99%) or very low (< 20%), where lower purities are ty-
stream, specific energy consumption per unit of CO2 cap- pically associated with TSA or alternative desorption
tured, and cost per unit of CO2 captured. KPIs provide un- methods such as moisture swing. We note that Table 5 in-
derstanding of how the combination of materials properties cludes an experimental study by Elfving et al. (2021) meant to
(e.g., shape of the equilibrium isotherm, heat of adsorption, compare different desorption methods for DAC based deso-
material kinetics, porosity, density) influence the process rption energy only. It does not present a holistic process
performance. They allow direct comparison between ad- design.
sorbents for a given adsorption process and comparison of Comparison between DAC process studies remains chal-
different process configurations and contactors as well as lenging as the assumptions, constraints, and calculation
desorption methods for a given adsorbent. The intrinsic methods vary from study to study. The units of productivity
properties of materials related to their structure (e.g., por- often differ, and the information needed to convert to con-
osity) and functional properties (e.g., selectivity) can identify sistent units is lacking. Productivity has been reported in (kg
promising materials, however, KPIs are required to provide CO2 (m3.d)-1), (kg CO2 (kgs.d)-1), (kg CO2 (kgs.h)-1), (kg CO2 d-1),
an informative ranking for the more promising sorbents and (kg CO2 h-1), where kgs refers to mass of adsorbent. The
(Farmahini et al., 2021; Rajagopalan et al., 2016). Overall, one mass or volume of adsorbent being referred to in productivity
can use KPIs to understand the best combination of material, values is not consistently specified. As mentioned in Table 4,
process configuration, contactor, and desorption method and productivity informs the size of a unit to capture a given
to optimise the operating parameters of a process. There are amount of CO2 meaning that one should report this KPI ei-
key differences that should be highlighted between assess- ther by dry mass of sorbent or total volume of sorbent (incl.
ment of post-combustion carbon capture and DAC processes. bed and sorbent porosity).
Firstly, there is no recovery specification in the case of DAC The conditions (i.e. RH and temperature) in which pro-
and optimisation is focused on specific energy and processes are explored and possibly compared matter. Focusing
ductivity. Secondly, a DAC process may not be constrained by first on the importance of RH, so far, process studies consider
the upstream or downstream process unlike in pre- or post- either a wet feed stream (25–65% RH), or a dry feed stream.
combustion carbon capture. Given that it is unlikely that the feed to a DAC process will
Table 5 shows process KPIs reported in key process stu- ever have a RH of 0%, one should rather consider a reason-
dies completed since 2016. Many studies are utilising an able baseline RH in the feed stream or include a pre-drying
amine-functionalised polymer adsorbent in a fixed bed con- step in process performance calculations. Wilson and Tezel
figuration, with fewer studies using MOF or silica adsorbents (2020) performed an experimental study on a fixed bed TVSA
or moving bed configurations. The range of reported values is process using Faujasite zeolites and estimated the energy
substantial despite the apparent similarity of some of these impact of a pre-drying step. Additional energy requirements
Table 4 – Description of key process performance indicators and their implication on process design.
Performance Description Impact on Process Design Optimisation
Indicator
Productivity Flow of CO2 over a given time for a fixed unit of Impacts size of unit required to capture a Maximise
adsorbent. given amount of CO2
Purity Mole percentage of CO2 in the product stream Impacts opportunities for downstream Meet target
utilisation of CO2
Working capacity Total CO2 obtained (amount adsorbed – amount Impacts size of unit required to capture a Maximise
desorbed) for a given amount of adsorbent given amount of CO2
Specific energy Energy consumption per unit of CO2 produced Impacts operational costs of DAC process Minimise
and ability to achieve net negative
Cost Cost per unit of CO2 produced Impacts economic feasibility of DAC process Minimise
Table 5 – Reported key performance indicators in process studies completed since 2016.1,2,3
Desorption Material SE Cost Productivity WC Purity Feed Conditions Ref
Method Class [MJ (kg CO2)-1] [$ (t- [kg CO2 [kg CO2 (kgs)-1] [mol%]
CO2)-1] (m)-3day-1]
Fixed Packed Bed

TCSA Polymer 7.1 60 192.9 – – Dry, 25 °C (Yu and Brilman)
C-TVSA MOF 3.9–5.9 86–221 – – 95% Sensitivity on CO2 (Sinha and
to water ratio Realff, 2019)
(1:2, 1:40)
TSA Polymer 4.5–6.4 – 62.1–65.7 0.019–0.024 – Dry, 25 °C (Elfving et al., 2021)
TCSA Polymer 4.4–6.3 – 54.9–58.5 0.019–0.024 – Dry, 25 °C (Elfving et al., 2021)
TVSA Polymer 6.9–38.4 – 62.6–75.6 0.02–0.025 – Dry, 25 °C (Elfving et al., 2021)
TVCSA Polymer 6.2–41.7 – 66.6–75.6 0.02–0.026 – Dry, 25 °C (Elfving et al., 2021)
S-TVSA Polymer 40.3 – 7.89 – 99.6% 50% RH, (Stampi-Bombelli
20 °C et al., 2020)
C-TVSA Polymer 2.5–2.7 – 62.1–78.4 0.040–0.048 > 95% 55% RH, (Young et al., 2021)
15 °C
TVSA Polymer 14–57 – 23.2–142.6 0.014–0.045 – 50% RH, (Schellevis
20 °C et al., 2021)
C-TVSA Polymer 2.7 − 0.2.1 < 200 94.6–143 > 95% 43% RH, (Sabatino et al., 2021)
20 °C
C-TVSA Polymer 2.8–3.7 < 200 92.1–109.7 > 95% 43% RH, (Sabatino et al., 2021)
20 °C
C-TVSA MOF 1.5–2.1 < 200 135.6–245.8 > 95% 43% RH, (Sabatino et al., 2021)
20 °C
C-TVSA Zeolite 17 – – – 95% 0% RH, (Wilson and
23.5 °C Tezel, 2020)
pHSA Carbon – – – 0.116 – 60% RH, (Cuesta and
35 °C Song, 2020)
Fixed Monolith Bed
S-TVSA MOF 6.7 – 97.9 > 99% 0.78% H20, 25 °C (Zhu et al., 2021)
S-TVSA MOF 3.3 75–140 – – 95% 25% RH, (Sinha et al., 2017)
25 °C
S-TVSA MOF 2.6 60–190 – – 95% 25% RH, (Sinha et al., 2017)
25 °C
Fixed Packed Bed in Series
MSA Polymer 0.8 35 – – 3% Dry, 25 °C (Hou et al., 2017)
C-TVSA Zeolite < 50 – – – > 95 Dry, 20 °C (Santori et al., 2018)
C-TVSA Zeolite 50 – – – > 97 Dry, 20 °C (Santori et al., 2018)
Radial Flow Packed Bed
TCSA Polymer 0.7–1.5 – – 0.066 – RH 40–65%, 19–22 °C (Yu and
Brilman, 2020)
Moving Belt
TSA Polymer Autothermal – 31.2 – 20% Dry, 25 °C (Drechsler and
Agar, 2019)
1
Several values obtained from these studies have been converted to allow consistent units.
2
In the case of productivity, the sorbent bulk density or vessel dimensions were used to determine volume conversions (if not already
provided in terms of volume), depending on which information was available.
3
Where producitivy was provided in hours, the per day rate was based on an assumed desorption cycle time from the respective study.
from regeneration of the drying bed increased proportionally (Kong et al., 2022; Zhu et al., 2021; Wijesiri et al., 2019a).
with RH and exponentially with ambient temperature. De- Turning our attention to the effect of temperature, as high-
pending on the RH, the additional drying step was estimated lighted by Table 5, many process studies are performed at
to strongly impact energy requirements (∼100% increase at room temperature (15 – 25 °C). Kong et al. (2022) discuss the
15 °C and 60% RH), making the approach only feasible in cold impact of varying ambient conditions on DAC processes,
dry climates. In addition, RH either causes a reduction in CO2 particularly focusing on sub-ambient conditions. Key find-
uptake or improves CO2 uptake and therefore should be ings from their work highlight that materials that do not
considered in process modelling or experimental work (see have adequate equilibrium uptake at ambient conditions are
discussion in Section 3.4). Water co-adsorption also influ- discounted, but may offer adequate uptake at sub-ambient
ences desorption: depending on the operating conditions, temperatures. The authors also discuss the impact of am-
during the desorption cycle, additional water could be ad- bient temperature on material kinetics, where lower tem-
sorbed or desorbed resulting in either a cooling or heating peratures often result in slower material kinetics.
effect. Additionally, water has faster desorption kinetics than Overall, the ambient conditions of both temperature and
CO2. During desorption the rapidly desorbing water provides RH will have notable impact on material selection and cycle
a purge for impurities, allowing the subsequently desorbed design. There is much scope in the field of DAC processes to
CO2 to have a higher purity. Pre-adsorbed water can also consider the range of ambient conditions that could be ex-
result in overall higher energy consumption for the process perienced in a single location and how one might implement
a suitable control and optimisation system for that process. This should be mandatory for CO2 and N2, and should also be
Although different ambient conditions must be considered, it done for O2 and H2O when uptake of these species is high.
can be beneficial to use a baseline ambient and RH humidity Finally, material selection must go beyond achieving rea-
value for initial comparison of process configurations and sonable equilibrium capacity and consider whether kinetics
cycles, for example 15 °C, 60% RH based on the mid-range are favourable at low CO2 pressures and whether recycling is
values in Fig. 2. achievable. Given the number and sometimes complexity of
In this review, we have focused on technical development the properties to be evaluated, obtaining them for all ad-
of DAC materials and processes. Along this angle, one should sorbents may be impossible. A screening exercise might be
also explore techno-economic assessments (TEA) to assess needed based on limited properties with more in-depth
feasibility and establish benchmarks. TEA studies for ad- characterisation for the most promising adsorbents.
sorption-based DAC remain scarce and we provide here a few The number of process studies on DAC adsorption has
references. Fasihi et al. (2019) compared low temperature increased over the last five years, with models being estab-
adsorption and high temperature absorption processes. lished and initial experimental work performed. Significant
Sinha and Realff (2019) studied the impact of adsorbent focus has been on modelling desorption methods. Going
properties, process conditions and feed streams on the cost forward, these analyses should provide more holistic and
of DAC adsorption processes. Wijesiri et al. (2019b) assessed consistent process assessments, reporting all relevant KPIs to
the cost of DAC for a given adsorbent while varying operation allow comparison. Although process studies should consider
process conditions. They identified material kinetics and use baseline ambient conditions to allow comparison, they
of waste heat as the key factors in cost reduction. Sabatino should also analyse a range of actual ambient conditions,
et al. (2021) performed an optimisation and cost assessment including the effect of moisture to test the robustness of
on the three key DAC technologies: adsorption, low tem- different processes. Process studies have overall paid little
perature MEA based absorption, and high temperature alka- attention to the design of process contactors. The high vo-
line solution absorption. Their works provide comprehensive lumetric flows required by DAC have implications on both
comparison of the three technologies based on cost, energy, the process contactor and material selection. Low pressure
and technical feasibility. Like Wijesiri et al. (2019b), they drop process contactors are needed to allow large volumetric
highlight the importance of sorption kinetics on cost in ad- air flow while minimizing specific energy requirements. The
sorption-based DAC. sustainable operation of the DAC unit from an energy per-
spective is in fact paramount to achieve the climate benefits
8. Summary and perspectives of DAC.
Our review highlights the progress made towards the ad- Funding Sources
vancement of adsorption-based DAC technologies in the
period 2016–2021, for both materials’ development and pro- The authors acknowledge the support from Imperial College
cess design. Yet, it also points towards major gaps in data London via the President’s PhD Scholarship (MYL, CP) and
and process analysis that must be addressed to advance the from the Department of Chemical Engineering at Imperial
field and reach the CO2 removal target fixed by the Paris College London via a PhD Scholarship (LB, RP, CP).
Agreement for NETs. We summarise below what we consider
as key findings and trends from our review.
Declaration of Competing Interest
Among promising candidate adsorbents, we highlight
amine-containing materials (i.e. polymeric resins as well as
The authors declare that they have no known competing fi-
functionalised MOFs and oxides) and ultramicroporous
nancial interests or personal relationships that could have
MOFs. These materials show high CO2 uptake at low CO2
appeared to influence the work reported in this paper.
partial pressure (uptake > 1 mmol g-1 at 500 ppm) and their
reduced porosity and/or their chemical nature limit N2
sorption. All materials exhibit a priori sufficient thermal Acknowledgements
stability for regeneration via a TVSA process, which is the
The authors acknowledge Dr. David Danaci for his insightful
method of choice to reach high CO2 purity.
discussions regarding our written content and figures.
Despite the criticality of adsorbent properties on process
performance, adsorbent development seldom occurs con-
currently with process development. In fact, many process Appendix A. Supporting information
studies do not consider recent and promising DAC ad-
sorbents. The barrier is often the lack of data of material Supplementary data associated with this article can be found
properties that can subsequently be applied to process in the online version at doi:10.1016/j.cherd.2022.11.040.
modelling. To synchronise this work, emphasis should be
placed on collecting material properties of candidate ad- References
sorbents that are necessary inputs for process modelling, e.g.
the heat capacity, thermal conductivity, and bed density. "CLIMATE-DATA.ORG." 〈https://en.climate-data.org/〉 (ac-
This also implies to evaluate adsorbents in a structured form cessed).
such as beads, pellets, or monoliths, rather than powders. Arias, P.A., et al., 2021. Climate Change 2021: The Physical
Science Basis. Contribution of Working Group I to the Sixth
Equilibrium isotherms must be collected with the intention
Assessment Report of the Intergovernmental Panel on
of using them for process modelling. This means measuring
Climate Change. IPCC,.
isotherms at a minimum of three different temperatures to Armstrong, M., Shi, X., Shan, B., Lackner, K., Mu, B., 2018. Rapid
achieve reliable parameterisation of the chosen adsorption CO2 capture from ambient air by sorbent-containing porous
isotherm model and calculation of the heat of adsorption. electrospun fibers made with the solvothermal polymer
additive removal technique. AIChE J., vol. 65 (1), 214–220. updates/new-development-company-1pointfive-formed/〉

https://doi.org/10.1002/aic.16418 (accessed 16/06/2022, 2022b).
Azzan, H., Rajagopalan, A.K., L’Hermitte, A., Pini, R., Petit, C., C. Engineering. "Engineering begins on UK’s first large-scale fa-
2022. Simultaneous estimation of gas adsorption equilibria cility that captures carbon dioxide out of the atmosphere."
and kinetics of individual shaped adsorbents. Chem. Mater. Carbon Engineering. 〈https://carbonengineering.com/news-
https://doi.org/10.1021/acs.chemmater.2c01567 updates/uks-first-large-scale-dac-facility/〉 (accessed 16/06/
B. P. C.-D. A. C. A. Council, The Commercial Case for Direct Air 2022, 2022c).
Capture ", Washington DC, 2021. C. Engineering. "New partnership to deploy large-scale Direct Air
B. Policy. Case Studies: Real-world Companies that are Pioneering Capture in Norway." Carbon Engineering. 〈https://carbo-
Direct Air Capture Technology and Market Applications of nengineering.com/news-updates/partnership-dac-norway/〉
Carbon Dioxide. (2021). Washington DC. [Online]. Available: (accessed 16/06/2022, 2022d).
〈https://bipartisanpolicy.org/download/?file=/wp-content/ Cai, Z., Bien, C.E., Liu, Q., Wade, C.R., 2020. Insights into CO2
uploads/2021/02/Bipartisan_Energy-DAC-Fact-Sheet-Part-3_ Adsorption in M–OH Functionalized MOFs. Chem. Mater., vol.
R01_2.5.2021edits-2.pdf〉. 32 (10), 4257–4264. https://doi.org/10.1021/acs.chemmater.
Babaei, H., McGaughey, A.J.H., Wilmer, C.E., 2017. Effect of pore 0c00746
size and shape on the thermal conductivity of metal-organic Chan, K.C., Chao, C.Y.H., Wu, C.L., 2015. Measurement of prop-
frameworks. Chem. Sci. 8 (1), 583–589. https://doi.org/10.1039/ erties and performance prediction of the new MWCNT-em-
C6SC03704F bedded zeolite 13X/CaCl2 composite adsorbents. Int. J. Heat.
Barrett, E.P., Joyner, L.G., Halenda, P.P., 2002. The Determination Mass Transf., vol. 89, 308–319. https://doi.org/10.1016/j.
of Pore Volume and Area Distributions in Porous Substances. ijheatmasstransfer.2015.05.063
I. Computations from Nitrogen Isotherms. J. Am. Chem. Soc., G. Chichilnisky, Frameworks in Energy Transformation, in Energy
vol. 73 (1), 373–380. https://doi.org/10.1021/ja01145a126 in Western Australia, Australia, 2018: Graciela Chichilnisky.
Barzagli, F., Giorgi, C., Mani, F., Peruzzini, M., 2020. Screening [Online]. Available: 〈https://chichilnisky.com/wp-content/
Study of Different Amine-Based Solutions as Sorbents for uploads/2018/08/Framework-in-Transformation-
Direct CO2Capture from Air. ACS Sustain. Chem. Eng., vol. 8, Presentation.pdf〉. [Online]. Available: 〈https://chichilnisky.
14013–14021. https://doi.org/10.1021/acssuschemeng.0c03800 com/wp-content/uploads/2018/08/Framework-in-
BASF. "BASF 13X Molecular Sieve Product Data Sheet." 〈https:// Transformation-Presentation.pdf〉.
catalysts.basf.com/files/literature-library/BASF_13X- Climeworks. "Capricorn." 〈https://climeworks.com/roadmap/
Molecular-Sieve_Datasheet_Rev.08–2020.pdf〉 (accessed). capricorn〉 (accessed 16/06/2022, 2022a).
Beuttler, C., Charles, L., Wurzbacher, J., 2019. The role of direct air Climeworks. "Arctic Fox." 〈https://climeworks.com/roadmap/
capture in mitigation of anthropogenic greenhouse gas arctic-fox〉 (accessed 16/06/2022, 2022b).
emissions. Front. Clim., vol. 1. https://doi.org/10.3389/fclim. Climeworks. "Orca: the first large-scale plant." 〈https://clime-
2019.00010 works.com/roadmap/orca〉 (accessed).
Bhatt, P.M., et al., 2016. A Fine-Tuned Fluorinated MOF Addresses Cuesta, A.R., Song, C., 2020. pH swing adsorption process for
the Needs for Trace CO2 Removal and Air Capture Using ambient carbon dioxide capture using activated carbon black
Physisorption. J. Am. Chem. Soc., vol. 138 (29), 9301–9307. adsorbents and immobilized carbonic anhydrase biocatalysts.
https://doi.org/10.1021/jacs.6b05345 Appl. Energy, vol. 280. https://doi.org/10.1016/j.apenergy.2020.
Bien, C.E., et al., 2018. Bioinspired Metal-Organic Framework for 116003
Trace CO2 Capture. J. Am. Chem. Soc., vol. 140 (40), Danaci, D., Bui, M., Mac Dowell, N., Petit, C., 2020. Exploring the
12662–12666. https://doi.org/10.1021/jacs.8b06109 limits of adsorption-based CO2 capture using MOFs with PVSA
Bien, C.E., Liu, Q., Wade, C.R., 2019. Assessing the Role of Metal – from molecular design to process economics. Mol. Syst. Des.
Identity on CO2 Adsorption in MOFs Containing M–OH Eng., vol. 5 (1), 212–231. https://doi.org/10.1039/c9me00102f
Functional Groups. Chem. Mater., vol. 32 (1), 489–497. https:// Dantas, T.L.P., et al., 2011. Carbon dioxide-nitrogen separation
doi.org/10.1021/acs.chemmater.9b04228 through adsorption on activated carbon in a fixed bed. Chem.
Bollini, P., Choi, S., Drese, J.H., Jones, C.W., 2011. Oxidative de- Eng. J., vol. 169, 11–19. https://doi.org/10.1016/j.cej.2010.08.026
gradation of aminosilica adsorbents relevant to postcombus- Darunte, L.A., et al., 2018. Moving Beyond Adsorption Capacity in
tion CO2 capture. Energy Fuels, vol. 25 (5), 2416–2425. https:// Design of Adsorbents for CO2 Capture from Ultradilute Feeds:
doi.org/10.1021/ef200140z Kinetics of CO2 Adsorption in Materials with Stepped
Bos, M.J., Kreuger, T., Kersten, S.R.A., Brilman, D.W.F., 2019. Study Isotherms. Ind. Eng. Chem. Res., vol. 58 (1), 366–377. https://
on transport phenomena and intrinsic kinetics for CO2 ad- doi.org/10.1021/acs.iecr.8b05042
sorption in solid amine sorbent. Chem. Eng. J., vol. 377, 120374. C. Davis, Oxy Looks to Ramp Permian Carbon Capture by Early
https://doi.org/10.1016/j.cej.2018.11.072 2024, Consider More JVs in Lower 48, in Natural Gas
Brilman, W., Garcia Alba, L., Veneman, R., 2013. Capturing at- Intelligence, ed, 2021.
mospheric CO2 using supported amine sorbents for micro- Drechsler, C., Agar, D.W., 2019. Simulation and optimization of a
algae cultivation. Biomass-.-. Bioenergy, vol. 53, 39–47. https:// novel moving belt adsorber concept for the direct air capture
doi.org/10.1016/j.biombioe.2013.02.042 of carbon dioxide. Comput. Chem. Eng., vol. 126, 520–534.
Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in https://doi.org/10.1016/j.compchemeng.2019.04.023
multimolecular layers. J. Am. Chem. Soc., vol. 60 (2), 309–319. Eisenberger, P., 2018. Rotating multi-monolith bed movement
https://doi.org/10.1021/ja01269a023 system for removing CO2 from the atmosphere. 14/587. U. S.
Bui, M., Gunawan, I., Verheyen, V., Feron, P., Meuleman, E., 2016. Pat. Appl. 716.
Flexible operation of CSIRO's post-combustion CO2 capture Elfving, J., Kauppinen, J., Jegoroff, M., Ruuskanen, V., Järvinen, L.,
pilot plant at the AGL Loy Yang power station. Int. J. Greenh. Sainio, T., 2021. Experimental comparison of regeneration
Gas. Control, vol. 48, 188–203. https://doi.org/10.1016/j.ijggc. methods for CO2 concentration from air using amine-based
2015.12.016 adsorbent. Chem. Eng. J., vol. 404, 126337. https://doi.org/10.
C. Engineering. "Carbon Engineering Innovation Centre Update." 1016/j.cej.2020.126337
Carbon Engineering. 〈https://carbonengineering.com/news- Elfving, J., Sainio, T., 2021. Kinetic approach to modelling CO2
updates/carbon-engineering-innovation-centre-update-3/〉 adsorption from humid air using amine-functionalized resin:
(accessed 16/06/2022, 2022a). Equilibrium isotherms and column dynamics. Chem. Eng. Sci.,
C. Engineering. "Oxy Low Carbon Ventures, Rusheen Capital vol. 246. https://doi.org/10.1016/j.ces.2021.116885
Management create development company 1PointFive to de- Erans, M., Sanz-Pérez, E.S., Hanak, D.P., Clulow, Z., Reiner, D.M.,
ploy Carbon Engineering’s Direct Air Capture technology." Mutch, G.A., 2022. Direct air capture: process technology,
Carbon Engineering. 〈https://carbonengineering.com/news- techno-economic and socio-political challenges. Energy
Environ. Sci., vol. 15, 1360–1405. https://doi.org/10.1039/ Kua, H.W., Pedapati, C., Lee, R.V., Kawi, S., 2019. Effect of indoor
d1ee03523a contamination on carbon dioxide adsorption of wood-based
M.S.-N. Eric Ping, Peter Eisenberger, Göteborg, Sweden. Global biochar – Lessons for direct air capture. J. Clean. Prod., vol. 210,
Thermostat Low Cost Direct Air Capture Technology. 860–871. https://doi.org/10.1016/j.jclepro.2018.10.206
A.H. Farmahini, S. Krishnamurthy, D. Friedrich, S. Brandani, and Kulkarni, A.R., Sholl, D.S., 2012. Analysis of equilibrium-based
L. Sarkisov, Performance-Based Screening of Porous Materials TSA processes for direct capture of CO 2 from Air. Ind. Eng.
for Carbon Capture, 2021, doi: 10.1021/acs.chemrev.0c01266. Chem. Res., vol. 51, 8631–8645. https://doi.org/10.1021/
Fasihi, M., Efimova, O., Breyer, C., 2019. Techno-economic as- ie300691c
sessment of CO2 direct air capture plants. 2019/07/01/. J. Kumar, A., et al., 2015. Direct air capture of CO2 by physisorbent
Clean. Prod., vol. 224, 957–980. https://doi.org/10.1016/j. materials. Angew. Chem. Int. Ed., vol. 54 (48), 14372–14377.
jclepro.2019.03.086 https://doi.org/10.1002/anie.201506952
G. C. Institute. "Oxy and Carbon Engineering partner to Combine Kumar, A., et al., 2017. Hybrid ultramicroporous materials (HUMs)
Direct Air Capture and Enhanced Oil Recovery storage." with enhanced stability and trace carbon capture perfor-
〈https://www.globalccsinstitute.com/news-media/latest- mance. Chem. Commun. (Camb. ), vol. 53 (44), 5946–5949.
news/oxy-and-carbon-engineering-partner-to-combine-direct- https://doi.org/10.1039/c7cc02289a
air-capture-and-enhanced-oil-recovery-storage/〉 (accessed). Kumar, A., 2018. Crystal engineering of hybrid ultramicroporous
Gebald, C., Meier, W., Repond, N., Ruesch, T., Wurzbacher, J.A., materials for study of direct air capture of carbon dioxide.
2017. Direct air capture device. U. S. Pat. Appl. PhD, Faculty of Science & Engineering. University of Limerick,.
Gebald, C., Wurzbacher, J.A., Borgschulte, A., Zimmermann, T., Lastoskie, C., Gubbins, K.E., Quirke, N., 2002. Pore size distribu-
Steinfeld, A., 2014. Single-Component and Binary CO2 and tion analysis of microporous carbons: a density functional
H2O Adsorption of Amine-Functionalized Cellulose. 2014/02/ theory approach. J. Phys. Chem., vol. 97 (18), 4786–4796.
18. Environ. Sci. Technol., vol. 48 (4), 2497–2504. https://doi. https://doi.org/10.1021/j100120a035
org/10.1021/es404430g Lee, J.J., Chen, C.-H., Shimon, D., Hayes, S.E., Sievers, C., Jones,
V. Gutknecht, S.Ó. Snæbjörnsdóttir, B. Sigfússon, E.S. Aradóttir, C.W., 2017. Effect of Humidity on the CO2 Adsorption of
and L. Charles. Creating a carbon dioxide removal solution by Tertiary Amine Grafted SBA-15. J. Phys. Chem. C., vol. 121 (42),
combining rapid mineralization of CO 2 with direct air cap- 23480–23487. https://doi.org/10.1021/acs.jpcc.7b07930
ture. Liao, P.Q., et al., 2016. Putting an ultrahigh concentration of
Heydari-Gorji, A., Sayari, A., 2012. Thermal, Oxidative, and CO2- amine groups into a metal-organic framework for CO2 capture
Induced Degradation of Supported Polyethylenimine at low pressures. Chem. Sci., vol. 7 (10), 6528–6533. https://doi.
Adsorbents. Ind. Eng. Chem. Res., vol. 51 (19), 6887–6894. org/10.1039/c6sc00836d
https://doi.org/10.1021/ie3003446 Madden, D.G., et al., 2017. Flue-gas and direct-air capture of CO2
HorvÁTh, G., Kawazoe, K., 1983. Method for the calculation of by porous metal-organic materials. Philos. Trans. R. Soc. A:
effective pore size distribution in molecular sieve carbon. J. Math., Phys. Eng. Sci., vol. 375 (2084), 20160025. https://doi.org/
Chem. Eng. Jpn., vol. 16 (6), 470–475. https://doi.org/10.1252/ 10.1098/rsta.2016.0025
jcej.16.470 Markewitz, et al., 2019. Carbon Capture for CO2 Emission
Hou, C., et al., 2021. Porosity and hydrophilicity modulated qua- Reduction in the Cement Industry in Germany. Energies, vol.
ternary ammonium-based sorbents for CO2 capture. Chem. 12 (12). https://doi.org/10.3390/en12122432
Eng. J., vol. 413. https://doi.org/10.1016/j.cej.2020.127532 McDonald, T.M., Lee, W.R., Mason, J.A., Wiers, B.M., Hong, C.S.,
Hou, C. l, et al., 2017. Integrated direct air capture and CO2 utili- Long, J.R., 2012. Capture of Carbon Dioxide from Air and Flue
zation of gas fertilizer based on moisture swing adsorption. J. Gas in the Alkylamine-Appended Metal–Organic Framework
Zhejiang Univ.: Sci. A, vol. 18, 819–830. https://doi.org/10.1631/ mmen-Mg2(dobpdc. J. Am. Chem. Soc., vol. 134 (16),
jzus.A1700351 7056–7065. https://doi.org/10.1021/ja300034j
Itadani, A., et al., 2016. Material Exhibiting Efficient CO2 McDonald, T.M., et al., 2015. Cooperative insertion of CO2 in
Adsorption at Room Temperature for Concentrations Lower diamine-appended metal-organic frameworks. Nature, vol.
Than 1000 ppm: Elucidation of the State of Barium Ion 519 (7543), 303–308. https://doi.org/10.1038/nature14327
Exchanged in an MFI-Type Zeolite. ACS Appl. Mater. Miao, Y., He, Z., Zhu, X., Izikowitz, D., Li, J., 2021. Operating
Interfaces, vol. 8 (13), 8821–8833. https://doi.org/10.1021/ temperatures affect direct air capture of CO2 in polyamine-
acsami.6b00909 loaded mesoporous silica. Chem. Eng. J., vol. 426. https://doi.
Jeong-Potter, C., Farrauto, R., 2021. Feasibility Study of Combining org/10.1016/j.cej.2021.131875
Direct Air Capture of CO2 and Methanation at Isothermal Micromeritics. Density Analysis, 2001. [Online]. Available:
Conditions with Dual Function Materials. Appl. Catal. B: 〈https://www.micromeritics.com/Repository/Files/Density_
Environ., vol. 282. https://doi.org/10.1016/j.apcatb.2020.119416 Brochure.pdf〉.
Keith, D.W., Holmes, G., Angelo St., D., Heidel, K., 2018. A Process Minx, J.C., et al., 2018. Negative emissions - Part 1: Research
for Capturing CO2 from the Atmosphere. Joule, vol. 2, landscape and synthesis. Environ. Res. Lett., vol. 13. https://
1573–1594. https://doi.org/10.1016/j.joule.2018.05.006 doi.org/10.1088/1748-9326/aabf9b
Kiani, A., Jiang, K., Feron, P., 2020. Techno-Economic Assessment Mukherjee, S., et al., 2019. Trace CO2 capture by an ultra-
for CO2 Capture From Air Using a Conventional Liquid-Based microporous physisorbent with low water affinity. Sci. Adv.,
Absorption Process. Front. Energy Res., vol. 8, 1–13. https://doi. vol. 5 (11), eaax9171. https://doi.org/10.1126/sciadv.aax9171
org/10.3389/fenrg.2020.00092 Mu, B., Walton, K.S., 2011. Thermal Analysis and Heat Capacity
Kim, C., Cho, H.S., Chang, S., Cho, S.J., Choi, M., 2016. An ethy- Study of Metal–Organic Frameworks. 2011/11/24. J. Phys. Chem.
lenediamine-grafted Y zeolite: a highly regenerable carbon C., vol. 115 (46), 22748–22754. https://doi.org/10.1021/jp205538a
dioxide adsorbent via temperature swing adsorption without NASA. "World of Change: Global Temperatures." 〈https://
urea formation. Energy Environ. Sci., vol. 9 (5), 1803–1811. earthobservatory.nasa.gov/world-of-change/global-tempera-
https://doi.org/10.1039/c6ee00601a tures〉 (accessed August 1, 2022).
Kolle, J.M., Fayaz, M., Sayari, A., 2021. Understanding the Effect of Nguyen, H.G.T., Horn, J.C., Bleakney, M., Siderius, D.W., Espinal,
Water on CO2 Adsorption. Chem. Rev., vol. 121 (13), 7280–7345. L., 2019. Understanding material characteristics through sig-
https://doi.org/10.1021/acs.chemrev.0c00762 nature traits from helium pycnometry. Langmuir, vol. 35 (6),
Kong, F., et al., 2022. Research needs targeting direct air capture 2115–2122. https://doi.org/10.1021/acs.langmuir.8b03731
of carbon dioxide: Material & process performance char- Nuhnen, A., Janiak, C., 2020. A practical guide to calculate the
acteristics under realistic environmental conditions. Korean J. isosteric heat/enthalpy of adsorption via adsorption iso-
Chem. Eng., vol. 39, 1–19. https://doi.org/10.1007/s11814-021- therms in metal–organic frameworks, MOFs. Dalton Trans. 49
0976-0 (no. 30), 10295–10307. https://doi.org/10.1039/D0DT01784A
Oda, A., et al., 2021. Unprecedented CO2 adsorption behaviour by Sanz-Pérez, E.S., Murdock, C.R., Didas, S.A., Jones, C.W., 2016.
5A-type zeolite discovered in lower pressure region and at 300 Direct Capture of CO2 from Ambient Air. Chem. Rev., vol. 116,
K. J. Mater. Chem. A, vol. 9 (12), 7531–7545. https://doi.org/10. 11840–11876. https://doi.org/10.1021/acs.chemrev.6b00173
1039/d0ta09944a Sarkar, J., Bhattacharyya, S., 2012. Application of graphene and
van Paasen, S., et al., 2021. Development of the solid sorbent graphene-based materials in clean energy-related devices
technology for post combustion CO2 capture towards com- Minghui. Arch. Thermodyn., vol. 33, 23–40. https://doi.org/10.
mercial prototype. Int. J. Greenh. Gas. Control, vol. 109, 103368. 1002/er
https://doi.org/10.1016/j.ijggc.2021.103368 Schellevis, H.M., van Schagen, T.N., Brilman, D.W.F., 2021.
Pang, S.H., Lee, L.C., Sakwa-Novak, M.A., Lively, R.P., Jones, C.W., Process optimization of a fixed bed reactor system for direct
2017. Design of aminopolymer structure to enhance perfor- air capture. Int. J. Greenh. Gas. Control, vol. 110, 103431.
mance and stability of CO2 sorbents: poly(propylenimine) vs https://doi.org/10.1016/j.ijggc.2021.103431
poly(ethylenimine). J. Am. Chem. Soc., vol. 139 (10), 3627–3630. Schell, J., Casas, N., Pini, R., Mazzotti, M., 2011. Pure and binary
https://doi.org/10.1021/jacs.7b00235 adsorption of CO2, H2, and N2 on activated carbon.
Pang, S.H., Lively, R.P., Jones, C.W., 2018. Oxidatively-Stable Adsorption, vol. 18 (1), 49–65. https://doi.org/10.1007/s10450-
Linear Poly(propylenimine)-Containing Adsorbents for CO2 011-9382-y
Capture from Ultradilute Streams. ChemSusChem, vol. 11 (15), Shekhah, O., et al., 2014. Made-to-order metal-organic frame-
2628–2637. https://doi.org/10.1002/cssc.201800438 works for trace carbon dioxide removal and air capture. Nat.
A. Peters, We have the tech to suck CO2 from the air–but can it Commun., vol. 5 (1), 4228. https://doi.org/10.1038/
suck enough to make a difference?, in Fast Company, ed, 2019. ncomms5228
Pini, R., 2014. Interpretation of net and excess adsorption isotherms Shekhah, O., Belmabkhout, Y., Adil, K., Bhatt, P.M., Cairns, A.J.,
in microporous adsorbents. Microporous Mesoporous Mater., vol. Eddaoudi, M., 2015. A facile solvent-free synthesis route for
187, 40–52. https://doi.org/10.1016/j.micromeso.2013.12.005 the assembly of a highly CO2 selective and H2S tolerant
Potter, M.E., Cho, K.M., Lee, J.J., Jones, C.W., 2017. Role of alumina NiSIFSIX metal–organic framework. Chem. Commun., vol. 51
basicity in CO2 uptake in 3-aminopropylsilyl-grafted alumina (71), 13595–13598. https://doi.org/10.1039/C5CC04487A
adsorbents. ChemSusChem, vol. 10 (10), 2192–2201. https:// Sinha, A., Darunte, L.A., Jones, C.W., Realff, M.J., Kawajiri, Y.,
doi.org/10.1002/cssc.201700115 2017. Systems Design and Economic Analysis of Direct Air
Rajagopalan, A.K., Avila, A.M., Rajendran, A., 2016. International Capture of CO2 through Temperature Vacuum Swing
Journal of Greenhouse Gas Control Do adsorbent screening Adsorption Using MIL-101(Cr)-PEI-800 and mmen-
metrics predict process performance ? A process optimisation Mg2(dobpdc) MOF Adsorbents. Ind. Eng. Chem. Res., vol. 56,
based study for post-combustion capture of CO 2. Int. J. 750–764. https://doi.org/10.1021/acs.iecr.6b03887
Greenh. Gas. Control, vol. 46, 76–85. https://doi.org/10.1016/j. Sinha, A., Realff, M.J., 2019. A parametric study of the techno-
ijggc.2015.12.033 economics of direct CO2 air capture systems using solid ad-
Rakić, V., Damjanović, L., 2013. Temperature-Programmed sorbents. AIChE J., vol. 65. https://doi.org/10.1002/aic.16607
Desorption (TPD) Methods. In: Auroux, A. (Ed.), Calorimetry Song, J., Zhu, L., Shi, X., Liu, Y., Xiao, H., Chen, X., 2019. Moisture
and Thermal Methods in Catalysis. Springer Berlin Heidelberg, Swing Ion-Exchange Resin-PO4 Sorbent for Reversible CO2
Berlin, Heidelberg, pp. 131–174. Capture from Ambient Air. Energy Fuels, vol. 33 (7), 6562–6567.
L.K. a A. Rathi, A Carbon-Sucking Startup Has Been Paralyzed by https://doi.org/10.1021/acs.energyfuels.9b00863
Its CEO, in Bloomberg UK, ed, 2021. Sonnleitner, E., Hofbauer, H., 2018. Assessment of zeolite 13X and
Rezaei, F., Webley, P., 2009. Optimum structured adsorbents for Lewatit® VP OC 1065 for application.pdf. Biomass-.-. Convers.
gas separation processes. Chem. Eng. Sci., vol. 64, 5182–5191. Biorefinery, vol. 8, 379–395.
https://doi.org/10.1016/j.ces.2009.08.029 Stampi-Bombelli, V., van der Spek, M., Mazzotti, M., 2020.
Rezaei, F., Webley, P., 2010. Structured adsorbents in gas se- Analysis of direct capture of CO 2 from ambient air via steam-
paration processes. Sep. Purif. Technol., vol. 70, 243–256. assisted temperature–vacuum swing adsorption. Adsorption,
https://doi.org/10.1016/j.seppur.2009.10.004 vol. 26, 1183–1197. https://doi.org/10.1007/s10450-020-00249-w
Rodriguez-Mosqueda, R., Bramer, E.A., Roestenberg, T., Brem, G., Standard Test Method for Determining Specific Heat Capacity by
2018. Parametrical Study on CO2 Capture from Ambient Air Differential Scanning Calorimetry, A. International, 2018.
Using Hydrated K2CO3 Supported on an Activated Carbon Standard Test Methods for Apparent Density, Bulk Factor, and
Honeycomb. Ind. Eng. Chem. Res, vol. 57 (10), 3628–3638. Pourability of Plastic Materials, A. International, 2017.
https://doi.org/10.1021/acs.iecr.8b00566 Sujan, A.R., Pang, S.H., Zhu, G., Jones, C.W., Lively, R.P., 2019. Direct
Rodríguez-Mosqueda, R., Rutgers, J., Bramer, E.A., Brem, G., 2019. CO2 Capture from Air using Poly(ethylenimine)-Loaded Polymer/
Low temperature water vapor pressure swing for the re- Silica Fiber Sorbents. ACS Sustain. Chem. Eng., vol. 7 (5),
generation of adsorbents for CO2 enrichment in greenhouses 5264–5273. https://doi.org/10.1021/acssuschemeng.8b06203
via direct air capture. J. CO2 Util., vol. 29, 65–73. https://doi.org/ Taddei, M., Petit, C., 2021. Engineering metal–organic frameworks
10.1016/j.jcou.2018.11.010 for adsorption-based gas separations: from process to atomic
Rosu, C., Pang, S.H., Sujan, A.R., Sakwa-Novak, M.A., Ping, E.W., scale. Mol. Syst. Des. Eng. no. 11, 841–875. https://doi.org/10.
Jones, C.W., 2020. Effect of extended aging and oxidation on 1039/D1ME00085C
linear poly(propylenimine)-mesoporous silica composites for Thakkar, H., Eastman, S., Al-Naddaf, Q., Rownaghi, A.A., Rezaei,
CO2 capture from simulated air and flue gas streams. ACS F., 2017. 3D-printed metal-organic framework monoliths for
Appl. Mater. Interfaces, vol. 12 (34), 38085–38097. https://doi. gas adsorption processes. ACS Appl. Mater. Interfaces, vol. 9
org/10.1021/acsami.0c09554 (41), 35908–35916. https://doi.org/10.1021/acsami.7b11626
Rouquerol, F., Rouquerol, J., Sing, K., 1999. Thermodynamics of Thakkar, H., Eastman, S., Hajari, A., Rownaghi, A.A., Knox, J.C.,
adsorption at the gas–solid interface. Adsorpt. Powders Porous Rezaei, F., 2016. 3D-Printed Zeolite Monoliths for CO2 Removal
Solids 27–50. from Enclosed Environments. ACS Appl. Mater. Interfaces, vol.
Sabatino, F., Grimm, A., Gallucci, F., van Sint Annaland, M., 8 (41), 27753–27761. https://doi.org/10.1021/acsami.6b09647
Kramer, G.J., Gazzani, M., 2021. A comparative energy and Thakkar, H., Issa, A., Rownaghi, A.A., Rezaei, F., 2017. CO2 capture
costs assessment and optimization for direct air capture from air using amine-functionalized kaolin-based zeolites.
technologies. Joule, vol. 5 (8), 2047–2076. https://doi.org/10. Chem. Eng. Technol., vol. 40 (11), 1999–2007. https://doi.org/10.
1016/j.joule.2021.05.023 1002/ceat.201700188
Santori, G., Charalambous, C., Ferrari, M.C., Brandani, S., 2018. Thommes, M., et al., 2015. Physisorption of gases, with special
Adsorption artificial tree for atmospheric carbon dioxide reference to the evaluation of surface area and pore size dis-
capture, purification and compression. Energy, vol. 162, tribution (IUPAC Technical Report). Pure Appl. Chem., vol. 87
1158–1168. https://doi.org/10.1016/j.energy.2018.08.090 (9–10), 1051–1069. https://doi.org/10.1515/pac-2014-1117
Vasheghani, M., et al., 2011. Effect of Al2O3 phases on the en- capturing CO2 directly from air. Processes, vol. 7 (8). https://
hancement of thermal conductivity and viscosity of nano- doi.org/10.3390/pr7080503
fluids in engine oil. 2011/11/01. Heat. Mass Transf., vol. 47 (11), Wilson, S.M.W., Tezel, F.H., 2020. Direct Dry Air Capture of CO2
1401–1405. https://doi.org/10.1007/s00231-011-0806-8 Using VTSA with Faujasite Zeolites. Ind. Eng. Chem. Res., vol.
Veneman, R., Frigka, N., Zhao, W., Li, Z., Kersten, S., Brilman, W., 59 (18), 8783–8794. https://doi.org/10.1021/acs.iecr.9b04803
2015. Adsorption of H2O and CO2 on supported amine sor- Wurzbacher, J.A., Gebald, C., Brunner, S., Steinfeld, A., 2016. Heat
bents. Int. J. Greenh. Gas. Control, vol. 41, 268–275. https://doi. and mass transfer of temperature-vacuum swing desorption
org/10.1016/j.ijggc.2015.07.014 for CO2 capture from air. Chem. Eng. J., vol. 283, 1329–1338.
Veselovskaya, J.V., Parunin, P.D., Netskina, O.V., Okunev, A.G., 2018. https://doi.org/10.1016/j.cej.2015.08.035
A Novel Process for Renewable Methane Production: Combining J.A. Wurzbacher, C. Gebald, N. Piatkowski, and A. Steinfeld,
Direct Air Capture by K2CO3/Alumina Sorbent with CO2 Concurrent Separation of CO 2 and H 2 O from Air by a
Methanation over Ru/Alumina Catalyst. Top. Catal., vol. 61 Temperature- Vacuum Swing Adsorption/Desorption Cycle,
(15–17), 1528–1536. https://doi.org/10.1007/s11244-018-0997-z 2012.
Viebahn, P., Scholz, A., Zelt, O., 2019. The potential role of direct Yoo, C.-J., Park, S.J., Jones, C.W., 2019. CO2 adsorption and oxi-
air capture in the German energy research program—results dative degradation of silica-supported branched and linear
of a multi-dimensional analysis. Energies, vol. 12. https://doi. aminosilanes. Ind. Eng. Chem. Res., vol. 59 (15), 7061–7071.
org/10.3390/en12183443 https://doi.org/10.1021/acs.iecr.9b04205
Walton, J.P.R.B., Quirke, N., 1989. Capillary condensation: a mo- Young, J., García-Díez, E., Garcia, S., Van Der Spek, M., 2021. The
lecular simulation study. Mol. Simul., vol. 2 (4–6), 361–391. impact of binary water-CO2isotherm models on the optimal
https://doi.org/10.1080/08927028908034611 performance of sorbent-based direct air capture processes.
Walton, K.S., Sholl, D.S., 2015. Predicting multicomponent ad- Energy Environ. Sci., vol. 14, 5377–5394. https://doi.org/10.
sorption: 50 years of the ideal adsorbed solution theory. AIChE 1039/d1ee01272j
J., vol. 61 (9), 2757–2762. https://doi.org/10.1002/aic.14878 Yu, Q., Brilman, W., 2020. A radial flow contactor for ambient air
Wang, T., Wang, X., Hou, C., Liu, J., 2020. Quaternary functiona- CO2 capture. Appl. Sci. (Switz. ), vol. 10. https://doi.org/10.
lized mesoporous adsorbents for ultra-high kinetics of CO2 3390/app10031080
capture from air. Sci. Rep., vol. 10 (1), 21429. https://doi.org/10. Q. Yu and D.W.F. Brilman. Design Strategy for CO2 Adsorption
1038/s41598-020-77477-1 from Ambient Air Using a Supported Amine Based Sorbent in
Ward, A., Pini, R., 2022. Integrated uncertainty quantification and a Fixed Bed Reactor.
sensitivity analysis of single-component dynamic column Q. Yu, J.D.P. Delgado, R. Veneman, and D.W.F. Brilman, Stability
breakthrough experiments. Adsorption, vol. 28 (3), 161–183. of a Benzyl Amine Based CO 2 Capture Adsorbent in View of
https://doi.org/10.1007/s10450-022-00361-z Regeneration Strategies, 2017. doi: 10.1021/acs.iecr.6b04645.
Weiland, R.H., Dingman, J.C., Cronin, D.B., 1997. Heat Capacity of Zhang, W., Liu, H., Sun, C., Drage, T.C., Snape, C.E., 2014.
Aqueous Monoethanolamine, Diethanolamine, N- Capturing CO 2 from ambient air using a polyethyleneimine –
Methyldiethanolamine, and N-Methyldiethanolamine-Based silica adsorbent in fl uidized beds. Chem. Eng. Sci., vol. 116,
Blends with Carbon Dioxide. J. Chem. Eng. Data, vol. 42 (5), 306–316. https://doi.org/10.1016/j.ces.2014.05.018
1004–1006. https://doi.org/10.1021/je960314v Zhao, R., Liu, L., Zhao, L., Deng, S., Li, S., Zhang, Y., 2019a. A
Wieme, J., Vandenbrande, S., Lamaire, A., Kapil, V., Vanduyfhuys, comprehensive performance evaluation of temperature swing
L., Van Speybroeck, V., 2019. Thermal engineering of meta- adsorption for post-combustion carbon dioxide capture.
l–organic frameworks for adsorption applications: a molecular Renew. Sustain. Energy Rev., vol. 114, 109285. https://doi.org/
simulation perspective. ACS Appl. Mater. Interfaces, vol. 11 10.1016/j.rser.2019.109285
(42), 38697–38707. https://doi.org/10.1021/acsami.9b12533 Zhao, R., et al., 2019b. Thermodynamic exploration of tempera-
Wijesiri, R.P., Knowles, G.P., Yeasmin, H., Hoadley, A.F.A., ture vacuum swing adsorption for direct air capture of carbon
Chaffee, A.L., 2019a. Desorption Process for Capturing CO2 dioxide in buildings. Energy Convers. Manag., vol. 183,
from Air with Supported Amine Sorbent. Ind. Eng. Chem. Res., 418–426. https://doi.org/10.1016/j.enconman.2019.01.009
vol. 58, 15606–15618. https://doi.org/10.1021/acs.iecr.9b03140 Zhu, X., Ge, T., Yang, F., Wang, R., 2021. Design of steam-assisted
Wijesiri, R.P., Knowles, G.P., Yeasmin, H., Hoadley, A.F.A., temperature vacuum-swing adsorption processes for efficient
Chaffee, A.L., 2019b. Technoeconomic evaluation of a process CO2 capture from ambient air. Renew. Sustain. Energy Rev.,
vol. 137. https://doi.org/10.1016/j.rser.2020.110651

Chemistry 5

Uploaded by

Copyright:

Available Formats

Chemistry 5

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chemistry 5

Uploaded by

Copyright:

Available Formats

Chemical Engineering Research and Design 189 (2023) 745–767

Available online at www.sciencedirect.com

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

Analytical review of the current state of knowledge

gas (GHG) emissions and the subsequent impact on the

Gas Stream CO2 concentration [%mol]

100 101 102 103 104

-: indicates information was unavailable

2.1.2. Particle density 2.1.3. Bed density

Fixed Packed Bed

additive removal technique. AIChE J., vol. 65 (1), 214–220. updates/new-development-company-1pointfive-formed/〉

You might also like