Covalent organic frameworks based materials: Synthesis

approaches and applications in environmental remediation

Imteaz Ahmed, Sung Hwa Jhung*

Department of Chemistry and Green-Nano Materials Research Center, Kyungpook National

University, Daegu 41566, Republic of Korea

Abstract: Covalent organic frameworks (COFs) are porous organic polymeric

materials. They have been applied for many modern applications in their pristine, modified or

converted forms. COF can be prepared by 1. One of the major application of COFs is in

environmental field. This review mainly focused on the state-ofthe-art development of stable

COFs in environmental applications, such as adsorption, membrane separation and detection of

various environmental pollutants, including heavy metals and radionuclides, toxic organic

pollutants (organic dyes, pharmaceuticals, organophosphorus flame retardants, per- and poly-

fluorinated alkyl substances and endocrine-disrupting chemicals), as well as the catalytic

degradation/conversion and CO2 capture. Their structural characteristics, functional design

approaches and specific environmental applications were introduced, and concluding remarks

and future research directions were proposed.

1
1. Introduction:

In the past few decades indiscrimante use of natural resources, uncontrolled emission

from the manufacturing industry, increasing consumption and energy production increased the

environmental pollution which is resulting a threat to the flora, fauna and human health.

Environment is therefore, one of the biggest concerns in modern world. Some particular

concerns are air, ground water and fuel emission contamination due to the exposure of natural

and synthetic organic dyes, pharmaceutical and personal care products, heavy metals, volatile

organic contaminants (VOCs), and other toxic gases etc. Many of these contaminants are so

harmful that even a trace amount of their presence can cause a huge problem in the

environment and toxicity to the human body, and therefore, required modern techniques to

eliminate. In this situation covalents organic frameworks (COFs)-based materials could be a

potential solution for the environmental remediation in various aspects [Nat. Rev. Mater. 1

(2016) 16068, Coordination Chemistry Reviews 400 (2019) 213046, Adv. Funct. Mater. 2018,

28, 1705553].

COFs have attracted great deal of interests recently because of their potential in the

environmental remediation and many other applications. They are formed by polymerization of

organic monomers with atomically precise porous framework structure [C. S. Diercks, O. M.

Yaghi, Science 2017, 355, 6328, N. Huang, P. Wang, D. Jiang, Nat. Rev. Mater. 2016, 1, 16068,

J. W. Colson, W. R. Dichtel, Nat. Chem. 2013, 5, 453, S.-Y. Ding, W. Wang, Chem. Soc. Rev.

2013, 42, 548]. COF polymers can be predesigned between monomers by geometry matching;

can be synthesized by topology-diagram-directed growth of polymer development. The COF

building blocks, i.e. monomers can be arranged in 2D or 3D space in high precision of

regularity and find tuning to impart specific physical and chemical properties. The design

principle is completely different from the regular polymeric synthesis approaches (e.g. linear,
2
hyperbranched, cross-linked or bio-polymers) which allows unique morphologies by

conformations to create new molecular platforms for functional and structural development.

Due to their specific, highly porous, 2D or 3D geometric structure, COFs can be applied for

various applications including adsorption, catalysis, gas storage, electrochemical and energy

applications [Nat. Rev. Mater. 1 (2016) 16068, Coordination Chemistry Reviews 400 (2019)

213046, Adv. Funct. Mater. 2018, 28, 1705553].

However, COFs can be further modified to increase their applicability in particular

case, or can be converted to other materials for more suitably apply. Post synthetic

modification (PSM) of COFs can be done by Metallation, functionalizaiton, chemical

conversion, composing with other materials and monomer truncation. In this review we will

briefly discuss the synthesis approaches of COFs. The post synthetic modification methods of

the COFs and the synthesis of COF derived materials will also will also be discussed. The

environmental applications of COFs will be discussed elaboratively. Finally, the summary and

prospects of MOF derived materials will be shown.

2. Synthesis of COF based MOFs

2.1. Synthesis of COFs

The synthesis of COFs are much more sophisiticated compared to the regular polymers,

since they have specific geometry and pore size/shape, composition and uniformity. Therefore,

synthesis of particularly designed COF with specific properties is a challenge and a

combination of different research areas including coordination chemistry, topochemial

polymerization, dynamic/reversible bonding, reticular assembly are required for the successful

COF formation.

3
 The first and most prominent COFs are synthesized from boroxines and boronic esters.

Adrien P. Coˆte´ et al. reported two highly crystalline COFs, COF-1 and COF-5 by the self-

condensation reactions of phenyl diboronic acid and hexahydroxythiphenylene [science 310,

2005, 1166] with surface areas of 711 and 1590 m2/g respectively. The crystal structures were

strongly held by strong covalent bonds between B,C and O atoms. Following this report many

other COFs of boron, e.g. COF-6, COF-8, COF-18 Å and COF-11 Å was synthesized [J. Am.

Chem. Soc. 2007, 129, 43, 12914–12915, Chem. Mater. 2006, 18, 22, 5296–5301, Adv Mater.

2008, 20, 2741]. Zhang et al. formed a new class of azobenzene COFs by borate ester

formation reaction [CrystEngComm 2014, 16, 6547]. The COF was formed by trans-to-cis

polymerization of Azo-components under UV irradiation. Yaghi et al. prepared COF-102 and

103 by self-condenstation of borinic acide with a 3D linked network. A very high surface area

was reported for these types of COFs; e.g. 4200 m2/g for COF-103 [science 310, 2005, 1166].

A modern progress of these types of COFs is to use mixed types of linkers. COFs can also be

prepared from more than two linkers. Huang et al. showed a multicomponent (1+2 and 1+3)

COF synthesis using different boronic acid linkers in one COF structure [Nat. Commun. 2016,

7, 12325].

Imine formation is another way to prepare COFs. Imines are formed from the

condensation between aldehydes and amines. In 2009 Yaghi et al. reported the first COF of

imines. The 3D COF (termed as COF-300) crystallized in 5-fold interpenetrated tetrahedral

framework and was prepared from terephthaldehyde and tetra(4-anilyl)methane [J. Am. Chem.

Soc. 2009, 131,4570]. Wang et al. reported the first 2D crystalline imine COF (COF-LZU1)

with a facile synthesis method [J. Am. Chem. Soc. 2011, 133, 49, 19816–19822]. Zhou et al.

reported a star-shaped dual pore 2D imine COF from terephthaldehyde and 4,4’,4”,4”’-(ethene-

1,1,2,2-tetrayl)- tetraaniline) with hexagonal mesopores and trigangular micropores of 26.9 A

4
and 7.1 A respectively. Moreover, using linear linkers of different lengh they could prepare

similar shaped but different sized COFs which also comprised of hexagonal and triangular

pores [J. Am. Chem. Soc. 2016, 138, 4710]. Jiang et al. showed a triangular pored COFs

formed from C6-symmetric linkers. With additional linear terepthaldehyde the length of the

sides of triangles can be extended and accessible pores can be achieved [Nat. Commun. 2015,

6, 7786].

Hydrazones are another class of COFs which are formed by reversible cocondensation

of hydrazines and aldehydes. Azines are similar to hydrazones, however, in this case hydrazine

is used as nitrogen-containing coupling partner for aldehyde. The first report of hydrazone

COFs were published by Yaghi groups in 2011. They reported a COF, denoted as COF-43 from

2,5-diethoxyterephthalohydrazide and 1,3,5-triformylbenzene. Dalapati et al. was first to report

an azine COF from hydrazine and 1,3,5,8-tetrakis(4-formylphenyl)pyrene with rhombic pores

of around 2 nm of diameters [J. Am. Chem. Soc. 2013, 135, 17310]. Mitra et al. developed

COFs with a mixture of ketonamine and hydrazone links. They used 1,3,5-

triformylphloroglucinol and 4-amino-2-hydroxybenzohydrazide as the amine and hydrazine

componensts.

Ketonamine links are particularly important in COFs synthesis. This links make COFs

highly stable in both acidic and basic conditions. Banerjee at el. reported the first Ketonamine

linked COF using 1,3,5-triformylphloroglucinol as the sole linker. This linkers have hydroxyl

groups adjacent to formyl groups. The imine COF is fromed from this linker which is then

converted to irreversible enol-keto tautomarization to form ketonamine links [J. Am. Chem.

Soc. 2012, 134, 19524]. Ketonamine COFs can be synthesized even by mechanical force by

using 1,3,5-triformylphloroglucinol and other linkers [J. Am. Chem. Soc. 2013, 135, 5328]

although these COFs shows low crystallinity and surface area.

5
 Above mentioned linking mechanisms are the most prominent coupling chemistry for

COFs; however, there are other linking mechanisms repored for COFs as well. Borosilicate

linkages are among the tops of them since, Yaghi groups showed a such type of COF synthesis

from tetra(4-dihydroxyboryl-phenyl)methane and tert-butylsilane triol (tBuSi(OH)3). Kuhn et

al. reported triazine frameworks CTF-1 from 1,4-dicyanobenzene [Angew. Chem., Int. Ed.

2008, 47, 3450]. Similar structured COFs can also be synthesized from 2,6-

naphthalenedicarbonitrile, [1,1’-biphenyl]- 4,4’-dicarbonitrile, and 5’-(4-cyanophenyl)-

[1,1’:3’,1”- terphenyl]-4,4”-dicarbonitrile, respectively [Adv. Mater. 2010, 22, 2202, Adv.

Mater. 2012, 24, 2357]. Squaraine-linked COFs were reported from squaric acid and aromatic

amines [Angew. Chem., Int. Ed. 2013, 52, 3770]. Benzimidazole linkage was reported to form

COF from 3-3′ diaminobenzene and 1,3,5-triformylphloroglucinol [Adv. Energy Mater. 2016,

6, 1601189]. Yan et al. introduced pyromellitic dianhydride to triangular amines to prepare

imide linked cryatalline COFs [Nat. Commun. 2014, 5, 4503]. This concept was then extended

to tetrahedral amines for a hexagonal 2D COF [AIChE J. 2017, 63,] and 3D COFs [J. Am.

Chem. Soc. 2015, 137, 8352].

2.2. Post-synthetic modification of COFs

2.2.1. Metallation

Metal sites are vastly used for wide range of catalysis and imbilization of metal site can

be a great advantage to use metals as heterogeneous catalysts for practical applications. COFs

with ligand containing linkers can be successfully used as a porous support for metal ions. Vast

amount of reports has been published for the immobilization of metal ions in COFs including

Sr [19816–19822, ACS Cent. Sci., 2018, 4, 748–754], Ca [J. Am. Chem. Soc., 2011, 133,

19816–19822, ACS Cent. Sci., 2018, 4, 748–754], Ti [Talanta, 2017, 166, 133–140], V [Chem.

6
Commun., 2014, 50, 6161–6163, ACS Appl. Mater. Interfaces, 2019, 11, 3070–3079], Mn [J.

Am. Chem. Soc., 2017, 139, 6042–6045], Fe [J. Mater. Chem. A, 2017, 5, 4343–4351, ACS

Appl. Mater. Interfaces, 2018, 10, 37919–37927], Co [Chem. Mater., 2016, 28, 4375–4379,

Dalton Trans., 2018, 47, 13824–13829], Ni [J. Mater. Chem. A, 2017, 5, 4343–4351, J. Am.

Chem. Soc., 2016, 138, 15134–15137], Cu [J. Am. Chem. Soc., 2016, 138, 15790–15796,

Science, 2016, 351, 365], Zn [J. Am. Chem. Soc., 2017, 139, 6042–6045,], Mo [RSC Adv.,

2014, 4, 51544–51547], Rh [Chem. – Eur. J., 2016, 22, 9087–9091], Pd [Adv. Funct. Mater.,

2018, 28, 1707110, Chem. Sci., 2016, 7, 6015–6020], Re [J. Am. Chem. Soc., 2018, 140,

14614–14618] and Ir [J. Am. Chem. Soc., 2018, 140, 16124–16133], etc. The metallation can

be achieved by post-synthetic modification or pre-metallated linkers can also be used to

prepare metallated COFs with the same results [ACS Appl. Mater. Interfaces, 2018, 10, 8527–

8535]. The post synthetic metallation is a common method and can be easily achieved by

coordination of metals to the ligand sites; however, the incorporation of ligands sites to the

COFs can be carried in different ways. Firstly, the ligands sites can be formed within the COFs

during the polymerization reactions of the linkers. The imine-linked COFs are most prominent

examples of this kind. Here, the ligand sites are formed during the COF formation through the

lone pair electron donation of nitrogen atom. These ligands can bind a wide variety of metal

ions with wide oxidation numbers including Pd2+, Fe2+, Co2+ and Ni2+. Large number of

catalytic reactions are reported using these metal immobilized COF frameworks [Refs]. Often,

one metal atom can be bound with multiple ligand sites through coordination resulting a

stronger and more effective immobilization of metal catalysts for higher stability. In a second

approach, the ligand sites are incorporated by a secondary reaction steps of the linkers after the

formation of primary COF framework. The prominent example of this types includes linkers of

iminophenol, β-ketoneamine and salicylic acid hydrazide [ACS Appl. Mater. Interfaces, 2017,

7
9, 22856–22863, Chem. Soc. Rev., 2004, 33, 410–421]. Imine-linked COFs having –OH

groups formed from salicylaldehyde and diamines to form a chelating Schiff base (named as

Salen ligands) which can strongly bind metals through coordination [Chem. Soc. Rev., 2004,

33, 410–421]. They are also reported as highly thermally stable COFs with minimam stability

temperature of 350 oC. Thirdly, the ligands sites for metal attachment can be incorporated by

the monomers having preincorporated ligand functionality. Tetra-dentate porphyrins,

phthalocynins, tridentate dehydrobenzoannulene, bidentate catechols are the prominent

examples of precursors of this type of COFs [Molecules, 2016, 21, 310, Chem. Rev., 1999, 99,

3153–3180, Adv. Mater., 2013, 25, 653–701]. Among these, porphyrins are most widely used

linkers for the synthesis of COFs with ligand functionality. They are naturally occurred

compounds in living bodies and are vastly investigated for different biomimetic catalyses. A

vast majority of metal ions can be attached to the central pyrrolic nitrogens of porphyrin ring

[Chem. – Eur. J., 2017, 23, 5652–5657, ACS Appl. Bio Mater., 2018, 1, 382–388, ACS Appl.

Mater. Interfaces, 2018, 10, 37919–37927.].

2.2.2. Functionalization

COFs can be post-synthetically modified for the incorporation of additional functional

groups which increases their applicability by imparting particular physical and chemical

properties. This tunability increased the applicability of MOFs vastly. The most commonly

used methods for the functionalization of MOFs are addition and substitution reactions. In

addition reaction, the pre-existing linker functional groups undergo a addition reaction with a

comparatively small molecule and incorporate additional functionality. Some of the most

investigated reactions are click reactions [Nat. Commun., 2011, 2, 536, Chem. Commun.,

2014, 50, 1292–1294, J. Am. Chem. Soc., 2015, 137, 7079–7082, J. Mater. Chem. A, 2017, 5,

17973–17981], ring opening reactions [Angew. Chem., Int. Ed., 2015, 54, 2986–2990, Angew.
8
Chem., Int. Ed., 2018, 57, 6042–6048], amidation [J. Am. Chem. Soc., 2018, 140, 12715–

12719], amidoxidation [Adv. Mater., 2018, 30, 1705479] and inverse vulcanization [J. Mater.

Chem. A, 2018, 6, 17977–17981, Bull. Chem. Soc. Jpn., 2017, 90, 1382–1387]. By click

chemistry reactions, aromatic, aliphatic groups, esters, carboxylic acids, alcohols, thiols,

fullerenes are introduced on the pore surface of COFs [Nat. Commun., 2011, 2, 536, Chem.

Commun., 2014, 50, 1292–1294, J. Am. Chem. Soc., 2015, 137, 7079–7082, J. Mater. Chem.

A, 2017, 5, 17973–17981]. Acid groups can also be incorporated by the reaction of phenol

groups with succinic anhydride in ring opening reactions [Angew. Chem., Int. Ed., 2015, 54,

2986–2990, Angew. Chem., Int. Ed., 2018, 57, 6042–6048]. Amidation and amidoximations

are occurred using nitriles, reacting with NaOH or NH2OH respectively [J. Am. Chem. Soc.,

2018, 140, 12715–12719, Adv. Mater., 2018, 30, 1705479]. Polysulfide groups can be

incorporated to COF’s through strong C-S bond by inverse vulcanization [J. Mater. Chem. A,

2018, 6, 17977–17981, Bull. Chem. Soc. Jpn., 2017, 90, 1382–1387]. In case of substation,

many different approaches have been taken. Etherification [Chem. Commun., 2016, 52, 7082–

7085, ACS Appl. Mater. Interfaces, 2018, 10, 41350–41358], acylation [RSC Adv., 2016, 6,

38774–38781], amidation [Angew. Chem., Int. Ed., 2018, 57, 16754–16759, J. Phys. Chem.

Lett., 2016, 7, 3660–3665], oxidation and reduction are among the most important of them.

Phenolic groups are attached to acyl chlorides or alkyl bromides by acylation and etherification

respectively. On the other hand, amidation usually carried for carbonyl groups by acetic

anhydride. Examples of other substitution reactions includes halogenations of alkyl chain

[CrystEngComm, 2017, 19, 4905–4910], salt formation of alkali metals [J. Am. Chem. Soc.,

2019, 141, 7518–7525, J. Am. Chem. Soc., 2018, 140, 984–992] or aromatic nucleophilic

substitution of aryl fluorides [ACS Appl. Mater. Interfaces, 2018, 10, 42233–42240].

2.2.3. Chemical conversion

9
 To form any kind of porous formation, reversibility is a key factor for ordered and

organized orientation of the components. Without correcting the errors in the frameworks or

healing the defects a successful ordered porous COF material can not be achieved. If the

reversibility can not be achieved properly, the expected high qulity cannot often be prepared.

However, reversibility also can make a COF vulnerable to chemical change and stability may

affect adversely. Therefore, COFs sometimes require a post-synthetic chemical treatment or

converstion better stability, and physical and chemical properties [J. Am. Chem. Soc., 2007,

129, 12914–12915, Angew. Chem., Int. Ed., 2009, 48, 5439–5442]. Yaghi groups first

developed a postsynthetic chemical conversion from imine bonding to amide. By this process

the primary 2D framework structure integrity was remained unchanged while the linkage

conversion increased the stability of the COF [J. Am. Chem. Soc., 2016, 138, 15519–15522]. It

was observed that the primary imine-linked COF is susceptible to disintegration in acidic

solution condition while the chemically converted imide COF remained stable. Cui et al.

applied this technique to a 3D COF successfully and coverted CCOF-5 to CCOF-6 with only

slight change in the unit cell parameters during imine to aminde conversion [J. Am. Chem.

Soc., 2018, 140, 892–895]. Similar to the previous case the amide COF remained stable in

acidic or basic solutions while the primary imine COF dissolves in both of the solutions.

Lotsch et al. converted a imine-linked COF into a thiazole-linked COF with the same porosity

and crystallinity of the parent COF [Nat. Commun., 2018, 9, 2600]. The thiazole-linked COF

showed stability in much higher pH solution compared to the imine-linked one. It was also

showed better resilience in presence of reducing agents, e.g. hydragine and sodium

borohydride. Linker exchange of COFs can be another method to modify a COF for better

physicochemical properties. Since, the reactions involving in the COF formation are dynamic

in nature; therefore, the linking bond allows a postsynthetic linker exchange within COF

10
framework. Zhao et al. studied the linker exchange among to COFs, TP-COF-BZ and TP-COF-

DAB where they showed that the linker benzidine from the two COFs can be exchanged with

1,4-diaminobenzene which makes the COFs more crystalline and hither thermally stable [J.

Am. Chem. Soc., 2017, 139, 6736–6743]. Cycloaddition can be another chemical conversion

method for COFs. Jiang et al. reported a structurally dynamic COF, DaTp prepared by the

condensation of 1,3,5-benzenetriboronic acid and 2,3,6,7-tetrahydroxyanthracene.

Interestingly, the photoinduced anthracene components can produce cycloaddition products

under Ar on a quartz substrate. The converted COF showed a better gas storage capcity than

the pristine COFs. Also, the fluorescence property of the primary COF was also changed.

2.2.4. Composite

In a branched morphology of the TpPa-1 COF with the Fe3O4 particles located at the

tips of the COF agglomerates [ACS Appl. Mater. Interfaces 2017, 9, 2959]. Gao et al. reported

the immobilization of phosphomolybdic acid H3[P(Mo3O10)4] (PMA) in the 3D imine COF-300

(from terephthalaldehyde and tetra-(4-anilyl)methane). For this purpose, the COF was stirred in

a PMA solution for 24 h and subsequently washed. Mo contents of up to 16% were reached.

The material was tested as catalyst for several olefin epoxidation reactions, but revealed a

lower catalytic activity than bare PMA [Micropor. Mesopor. Mater. 2015, 213, 59]. Polymers

and Salts in COFs: In the pores of a COF film (on a gold substrate) made from DAAQ and

TPG (DAAQ–TFPCOF), 3,4-ethylenedioxythiophene was electropolymerized via cyclic

voltammetry. This polymer pore filling resulted in a significantly higher cycling stability and

current response of the COF film in cyclic voltammetry [ACS Cent. Sci. 2016, 138, 5897]. The

ionic liquid 1-ethyl- 3-methylimidazolium bis(trifluoromethylsulfonyl)imide was immobilized

in the 3D COF-320 (from 4,4’-biphenyldicarboxaldehyde and tetra-(4-anilyl)methane) by

mixing the two powders and subsequent heating to 90 oC. Embedded in the COF, the ionic
11
liquid underwent no phase transition in the temperature range from 160 to 190 oC [RSC Adv.

2017, 7, 1697]. The phosphonium salt ethyldiphenyl(4-vinylphenyl)phosphonium bromide was

polymerized in the pores of TpBpy-COF (from TPG and 5,5’-diamino- 2,2’-bipyridine) and the

composite was subsequently treated with Cu(OAc)2 to establish Cu moieties at the bipyridyl

units. The polymer was shown to have some flexibility in the COF pores, which was

advantageous for the application of the composite in catalysis (cycloaddition of epoxides and

CO2) [J. Am. Chem. Soc. 2016, 138, 15790]. COF-Matrix Composites: COF/polymer hybrid

membranes have been prepared by casting a solution of COF and polymer mixture and

subsequent evaporation of the solvent. Biswal et al. prepared COF-polybenzimidazole hybrid

membranes with the COFs TpBD (from TPG and benzidine) and TpPa-1 (from TPG and p-

phenylenediamine), respectively (Figure 31). The membranes were chemically stable and

showed a high gas permeability for H2, CO2, and CH4. With these membranes, the authors

achieved high hydrogen selectivities from gas mixtures, namely 82.7 for H2/N2 (with

TpBD(40%)@PBI-BuI), up to 165.5 for H2/CH4 (with TpPa-1(40%)@PBI-BuI), and 3.9 for

H2/CO2 (with TpBD(40%)@PBI-BuI) [ Chem. - Eur. J. 2016, 22, 4695]. The two COFs NUS-

2 (from TPG and hydrazine) and NUS-3 (from TPG and 2,5-diethoxy-terephthalohydrazide)

were first delaminated in a sonication step and subsequently mixed with the polymer Ultem, a

poly(ether imide) or polybenzimidazole (PBI) to fabricate flexible membranes. Thereby, gas

permeabilities were increased compared to the pristine polymers while the composites also

featured improved gas selectivities. Specifically, for NUS-2@PBI membranes, a remarkably

high H2/CO2 permselectivity of up to 31.4 under higher pressures (for 20 wt% NUS-2@PBI)

could be achieved [Chem. Mater. 2016, 28, 1277]. In a different approach, CTF-1 (from 1,4-

dicyanobenzene) was milled and sonicated to yield delaminated separate COF sheets in a stable

suspension in water. This suspension was then mixed with a suspension of graphene oxide

12
sheets in water and restacked by filtration, yielding COF-graphene oxide mixed membranes

that showed very high H2 permeabilities and H2/CO2 selectivities. For example, a membrane

with an H2 permeability of 1.7 × 10−6 mol m2 Pa−1 s−1 featured an H2/CO2 selectivity of

17.4 [J. MaterChem A 2016, 4, 16288]. TpPa-2 COF (made from 2,5-dimethyl-p-

phenylenediamine and TPG) containing mixed matrix membranes were prepared by mixing

TpPa-2 with polysulfone, where a COF content of only 0.2 wt% significantly improved the

performance of the membranes in the removal of organic foulants from water, presumably due

to the regular porous structure of the COF [ J. Membr. Sci. 2017, 526, 355]. Another mixed

matrix membrane was prepared from the azine-linked COF ACOF-1 (from TFB and hydrazine)

with the polymer Matrimid and utilized for CO2/CH4 separation. The mixed COF and polymer

membrane featured a higher CO2 permeability than Matrimid alone.[ Chem. Eur. J. 2016, 22,

14467] Mixed matrix membranes were also prepared from sulfonic acid functionalized

ketoenamine COFs COF NSU-9 (from TPG and 2,5-diaminobenzenesulfonic acid) and NUS-

10 (from TPG and 2,5-diaminobenzene-1,4-disulfonic acid) blended with polyvinylidene

fluoride for proton conduction (see Section 6.3 for details).[ ACS Appl. Mater. Interfaces 2016,

8, 18505] A TpPa-1 COF blend with magnetic graphene (loaded with Fe) was used for the

immobilization of trypsin. For that, glutaraldehyde was covalently attached to the COF, which

then featured aldehyde anchoring positions to covalently bind trypsin.[ Anal. Bioanal. Chem.

2017, 409, 2179] Methacrylatefunctionalized TpPa-1 was presynthesized (by reacting TPG

with methacrylic anhydride prior to COF synthesis) and then embedded in a

polymethylmethacrylate matrix for chromatography, for details see Section 6.2) [ Chromatogr.

A 2017, 1479, 137]. Several composites intended as membranes have been reported in which a

COF was grown or attached as a coating on another functional material. For example, Lu et al.

reported the modification of a porous α-Al2O3 substrate with the 3D imine COF-320,

13
composed of tetra-(4-anilyl)methane and 4,4′-biphenyldicarboxaldehyde. For this purpose, 3-

aminopropyltriethoxysilane and 4,4′-biphenyldicarboxaldehyde were first used to establish an

aldehyde functionalization at the Al2O3 substrate. The COF synthesis reaction was then

performed under solvothermal conditions, creating an ≈4 µm thick COF film on the Al2O3

substrate. This material was applied in gas permeation experiments and showed moderate H2

selectivities (2.5 for H2/ CH4 and 3.5 for H2/N2) [ Chem. Commun. 2015, 51, 15562]. A

different approach was pursued by Li et al.. They first synthesized COF-1 in a conventional

solvothermal synthesis and subsequently delaminated the COF sheets by sonication in

dichloromethane. A macroporous α-Al2O3 substrate with a thin SiO2–ZrO2 layer was then

coated with the COF sheets. With this method a very smooth and crack-free membrane was

produced that featured a very high H2 permeability of 1 × 10−6 mol m2 Pa−1 s−1 but a low

H2/N2 selectivity of up to 4.2.[ACS Appl. Mater. Interfaces 2017, 9, 8433] COF-MOF

composite membranes were fabricated by Fu et al., by first growing solvothermally a COF-300

film (made from tetra-(4-anilyl)-methane and terephthalaldehyde) on a polyaniline membrane

and subsequently solvothermally growing a film of the MOF ZIF-8 or Zn2(bdc)2(dabco) from

MOF synthesis mother liquors. These hybrid membranes showed excellent performance in gas

separation. They featured H2/CO2 selectivities of 12.6 for a [COF-300]-[Zn2(bdc)2(dabco)]

composite and 13.5 for a [COF-300]-[ZIF-8] composite [ J. Am. Chem. Soc. 2016, 138, 7673].

Using a sonochemical synthesis, Yoo et al. coated carbon nanotubes and graphene flakes with

COF-5. In CO2 adsorption experiments, the COF-5 coated carbon nanotubes showed a higher

CO2 uptake of 1.42 wt% compared to bulk COF-5 (1.25 wt%). However, the CO2 uptake was

lowered in comparison to bare COF-5 when it was coated on graphene [ J. K. Shim, Chem.

Lett. 2015, 44, 560]. Recently the authors reported the use of COF-coated carbon nanotubes for

trapping polysulfide species in lithium–sulfur batteries. In their study, they grew COF-5 and

14
COF-1 solvothermally on carbon nanotubes. In the lithium sulfur batteries, the COF-1 coated

structures were placed between the sulfur cathode and a polyethylene separator membrane and

allowed for the enhanced deposition/dissolution of Li2S species, thereby improving the

performance of Li–S cells [ Nano Lett. 2016, 16, 3292]. For more details see Section 6.4.

Amino-functionalized multiwall carbon nanotubes were coated with COFTTA-DHTA (from

4,4′,4″-(1,3,5-triazine-2,4,6-triyl)trianiline and 2,5-dihydroxyterephthalaldehyde) in a

solvothermal reaction. The resulting composite was investigated regarding its performance as

capacitive electrode material [ Chem. Commun. 2017, 53, 6303]. The ketoenamine COF TpBD

(from TPG and benzidine) was used as a shell for Fe3O4 nanoparticles. This material was

tested for its photothermal conversion efficiency in water, for potential therapeutic

applications. Upon irradiation with a 785 nm laser, the COF coated particles featured a high

photothermal conversion efficiency of 21.5% [Chem. Commun. 2015, 51, 12254]. Fe3O4

particles with a shell of TpBD COF were also applied in water purification (see Section 6.2) [

Chem. Commun. 2017, 53, 2511], and Fe3O4 particles with a shell of an imine-linked COF

made from terephthalaldehyde and 1,3,5-tris(4-aminophenyl)benzene were applied in the

enrichment of peptides (see Section 6.2) [ Chem. Commun. 2017, 53, 3649]. Similarly, the

ketoenamine COF TpPa-1 was solvothermally grown on Fe3O4 nanoparticles, which were then

applied in glycopeptide enrichment [J. Mater. Chem. B 2017, 5, 4052]. Au nanoparticles of 15

nm diameter were coated with polyvinylpyrrolidone (PVP) to facilitate their encapsulation into

COF particles and subsequently introduced in a solvothermal COF synthesis of 1,3,5-tris(4-

aminophenyl)benzene and 2,5-dimethoxyterephthalaldehyde. With this strategy, spherical COF

particles containing one or more Au nanoparticles were obtained. This synthesis pathway

worked also well for PVP-modified 50 nm Au nanoparticles, Au nanorods and 3.8 nm Pt

nanoparticles [ ACS Appl. Mater. Interfaces 2017, 9, 7481]. In another study, a nanofibrous

15
imine-linked COF from 2,6-dihydroxynaphthalene-1,5-dicarbaldehyde and 2,4,6-

tris(aminophenyl)pyridine was coated on graphene oxide.[ J. Mater. Chem. A 2014, 2, 8201]

Mullangi et al. utilized the highly hydrophobic character of two nitrogen-rich COFs (IISERP-

COF 2 from PDA and tris(4- formylphenyl)amine and IISERP-COF 4 (from TA and TAPP) to

prepare functional coatings. By coating the COFs on paper or cloth, highly hydrophobic, thus

self-cleaning materials were obtained [ J. Mater. Chem. A 2017, 5, 8376]. The crystalline

triazine framework TriCF (from melamine and cyanuric chloride) was used as additive in poly-

α-olefin (PAO)-10 oil. The resulting oil blend featured an improved tribological performance in

steel-steel and steelcopper systems [Tribol. Int. 2017, 111, 57]. The numerous examples for the

use of COFs serving as a component in composites discussed in this chapter illustrate the

advantages provided by COFs in combination with other materials due to their regular, porous

structure and their modular synthesis. This becomes particularly apparent in the fields of

catalysis and membrane separations, especially those based on single 2D COF sheets obtained

by sonication.

3. Aplication of COFs in environmental remediation

3.1. Adsorption

COF based catalysts showed a significant amount of progress in recent years. COFs and

their derivatives have been applied to remove a wide range of contaminant from various

mediums. A short note about their applications are given below.

3.1.1. Liquid phase

Liquid phase adsorption can be divided into two parts. Water phase adsorption and gas

phase adsorption.

16
3.1.1.1. Water phase adsorption

Metal ions adosorption: Heavy metals and radionuclides pollution is a great threat to

the health of human beings and environment owing to their chemically toxic effect. The

excessive amounts of these pollutants in aqueous medium should be well treated to the safe

concentration before being discharged into the environment. For the adsorptive removal of

metal ions, stable COFs are usually constructed based on the principle of introducing

adsorptive sites (e.g. affinity groups). Despite of relative high cost of synthesis compared to

commercial adsorbents, various COFs presented fast adsorption equilibrium (< 1 h), high

adsorption capacity, and good adsorption selectivity towards many heavy metal ions and

radionuclides, including Hg2+, UO2 2+, Nd3+, I-, and TcO4-. Coordinative adsorption and

ionexchange are the main adsorption mechanism towards these pollutants [J. Am. Chem. Soc.

139 (2017) 2786–2793, J. Am. Chem. Soc. 139 (2017) 2428–2434, J. Am. Chem. Soc. 138

(2016) 3031–3037, Environ. Sci. Technol. 53 (2019) 5212–5220, J. Mater. Chem. A 5 (2017)

17973–17981, Chem. 9 (2018) 4125–4131, ACS Sustainable Chem. Eng. 7 (2019) 7353–

7361]. Considering the extraordinary binding affinity between sulfur derivatives and Hg2+,

several kinds of sulfur-containing 2 D COFs, including COF-LZU8 [J. Am. Chem. Soc. 138

(2016) 3031–3037], TAPB-BMTTPA-COF [J. Am. Chem. Soc. 139 (2017) 2428–2434], TPB-

DMTP-COF-SH [J. Mater. Chem. A 5 (2017) 17973–17981], COF-S-SH [J. Am. Chem. Soc.

139 (2017) 2786–2793], SCTN-1 [Chem. Eng. 7 (2019) 7353–7361] and NOP-28 [Polym.

Chem. 9 (2018) 4125–4131], were fabricated. These COFs presented good adsorption

capacities and selectivity for mercury. For the selective adsorption of uranyl ions, amidoxime

(AO) functional groups, which are the efficient coordinating groups for uranyl ions [J. Clean.

Prod. 188 (2018) 655–661, Radiochimica Acta, Volume 107, Issue 6, Pages 459–467], were

incorporated into COFs via predesigned bottom-up method [Adv. Mater. 30 (2018) 1705479].
17
With amidoxime functional groups, COF-TpAb-AO presented a good adsorption capacity of

uranyl ions (404 mg/g). Furthermore, its comparison with amorphous porous organic polymers

(POP-TpAb-AO) [Adv. Mater. 30 (2018) 1705479] and amorphous polymer (PAF-1-CH2AO)

[Adv. Mater. 30 (2018) 1705479] with AO groups proved the superiority of COFs with ordered

structure and higher surface area as a decorating platform. COFs were also used for the rere

earth elements capture. Post synthetic method was utilized to prepared 3D COFs with carboxyl

groups, denoted as 3D COF-COOH [Angew. Chem. Int. Ed. 57 (2018) 6042–6048]. The

grafting carboxyl functional groups were obtained by the ring-opening reaction of 3D COFs

with OH and succinic anhydride. The incorporation of succinic anhydride decreased COFs’

BET surface area from 1077 m2/g to 540 m2/g, and pore size from 1.35 nm to 0.68 nm.

Despite of this, the carboxyl functional 3D COFs showed exceptional lanthanide selectivity,

and it could be reused for several cycles with a minimal reduction in adsorption capacity. The

adsorptive removal of radioactive TcO4- from aqueous solutions by cationic covalent organic

nano-sheets (iCON) was reported [Environ. Sci. Technol. 53 (2019) 5212–5220].

Adsorption of organic contaminants: Besides inorganic contaminants, the existence of

organic contaminants (e.g. organic dye and pharmaceutical pollutants) in waters has also

aroused great concern owing to their toxic effect. These toxic organic pollutants may be stable

and soluble, and can readily migrate a long distance, leading to the wide spread of pollution; or

may be easily decomposed and consume dissolved oxygen excessively, resulting in the

collapse of the aquatic ecosystem. Furthermore, even at a low-level exposure, some toxic

organic pollutants may also have adverse effects on ecosystem and human being. Therefore, it

is vital to remove the toxic organic pollutants from solutions The hydrophobic nature of COFs

endows themselves with high affinity towards many lipophilic organic matter [J. Am. Chem.

Soc. 140 (2018) 12677–12681, ACS Appl. Mater. Interfaces 10 (2018) 30265–30272]. Whilst,
18
polar solvents (e.g. methanol, ethanol, isopropanol, and tetrahydrofuran) are effective for the

desorption of these organic pollutants [ACS Appl. Mater. Interfaces 10 (2018) 30265–30272, J.

Phys. D: Appl. Phys. 51 (2018) 345401]. The time for equilibrium varies from 5 min to 12 h

[Chem. Commun. 53 (2017) 2511–2514, ACS Appl. Mater. Interfaces 10 (2018) 30265–30272,

AlChE J. 63 (2017) 3470–3478], depending on the types of COFs and the targeted pollutants.

Various adsorption mechanisms, including pore-size effect [ACS Appl. Mater. Interfaces 10

(2018) 30265–30272, AlChE J. 63 (2017) 3470–3478], H-bonding [Chem. Commun. 53 (2017)

2511–2514], and p– p [Chem. Commun. 53 (2017) 2511–2514, ACS Appl. Mater. Interfaces 10

(2018) 30265–30272], may be involved in adsorption process. Pore size sieving and more

stable adsorption configuration accounted for COFs adsorption affinity [ACS Appl. Mater.

Interfaces 10 (2018) 30265–30272, J. Phys. D: Appl. Phys. 51 (2018) 345401, AlChE J. 63

(2017) 3470–3478]. The study on the adsorption behavior of aryl-organophosphorus flame

retardant onto 3 kinds of COFs with different pore sizes (COF1: 1.81 nm; COF2: 2.57 nm; and

COF3: 3.34 nm) was explored [ACS Appl. Mater. Interfaces 10 (2018) 30265–30272].For the

removal of organic dye pollutants, COFs also showed great potential. It was reported that

triazine-functionalized polyimide COF could effectively adsorb the organic dyes (e.g.

methylene blue (PB), Rhodamine B (RB), and Congo red (CR)) [AlChE J. 63 (2017) 3470–

3478]. A satisfying adsorption capacity of 1691 mg PB/g overtaking the weight of COFs (TS-

COF-1) was recorded. Based on the experimental results and in silico simulations, the

extraordinary adsorption capacity was contributed to intrinsic pore-size effect of adsorbents

and the size of adsorbates [AlChE J. 63 (2017) 3470–3478]. Crystalline 3D COFs based on the

b-cyclodextrin (b-CD) units were successful synthesized and applied for adsorptive removal of

polar micropollutants [Chem. Eur. J. 24 (2018) 10979–10983]. The molecular specific

recognition of b-cyclodextrin was inherited by b-CD COFs, which showed a great adsorption

19
selectivity towards bisphenol A, 4-nonyl phenol, (S)-naproxen, and Rhodamine B. The

adsorptive removal of per- and poly-fluorinated alkyl substances (PFAS) into imine-linked

COFs bearing varying densities of primary amines were reported [J. Am. Chem. Soc. 140

(2018) 12677–12681]. Results showed that partially amine-functionalized ( ⁓20%–28%) imine

COFs presented the highest and fastest removal, indicating that the hydrophobic surfaces and

polar group may work together for the removal of ammonium perfluoro-2-propoxypropionate

(Gen X). COFs bearing 28% amine showed a maximum adsorption capacity of about 200 mg/g

for Gen X. This COFs also outperformed granular/powdered activated carbon [J. Am. Chem.

Soc. 140 (2018) 12677–12681] in adsorption of Gen X and 12 kinds of PFAS.

3.1.2. Gas phase adsorption

COFs with the characteristics of high porosity (e.g. 4210 m2/g BET surface area for

COF-103 [Science 316 (2007) 268–272]) and low density (e.g. 0.17 g/cm3 for COF-108 [78])

has also been evaluated for gas adsorption, and showed great potential in the adsorptive

removal of volatile pollutants, and the storage of combustion gas products.

Adsorption of volatile organic compounds (VOC). A large-scale computational

screening study of 187 kinds of COFs on the adsorption of gaseous I2 and CH3I on the

conditions of 423 K and 1 bar was conducted [CrystEngComm 19 (2017) 4920– 4926]. The

simulated resulted showed that 3D-COFs was super than that of 2D-COFs in the capture of I2

and CH3I. Among these COFs, 3D-Py-COF (Fig. 5d) and COF-103 (Fig. 5c) showed the

highest I2 capture of 16.7 g/g and CH3I capture of 2.8 g/g, respectively. Compared to other

COFs, 3D-COFs with a pore size of ⁓0.9 nm showed a much higher uptake capacity.

Heteropore 2D COFs (SIOC-COF-7) with hierarchical porosities was utilized for iodine

capture [Chem. Commun. 53 (2017) 7266–7269]. It showed an adsorption capacity of 481 wt%

20
towards volatile iodine and 127 wt% towards dissolved iodine in hexane. The occupation of

solvent molecules in the pore structure of COFs led to the decrease in the pore volume for

iodine, accounting for the differences in adsorption capacity.

CO2 capture. The extensive consumption of fossil energy leads to the massive release

of CO2 into the air, making it a highly concerned environmental issue due to its greenhouse

effect. To mitigate the environmental dilemmas, the capture of CO2 has aroused great interests.

It relies on porous materials for gas storage, and COFs have been proved to be good candidates

for the capture of CO2. Theoretical calculation of COFs’ gas capture capability was carried out

before the experimental study [J. Am. Chem. Soc. 130 (2008) 11580–11581]. Inspired by the

theoretical results, the pioneers from the research group of Yaghi, in 2008, firstly reported the

experimental results of prototypical 3D COFs’ sorption performance towards H2, CH4 and

CO2 [J. Am. Chem. Soc. 131 (2009) 8875–8883]. The gas adsorption by COFs’ has been

rapidly progressed. So far, boron-based, triazined-based [Energy Environ. Sci. 6 (2013) 3684–

3692, Environ. Sci. Technol. 50 (2016) 4869–4876], imine-based COFs, as well as

boron/imine-based COFs, have been evaluated as candidate adsorbents for gas capture (e.g.

H2, CO2, and NH3). Several excellent reviews focusing on COFs’ gas capture (e.g. CO2, H2,

CH4, and NH3) have also been published [Chem. Soc. Rev. 42 (2013) 548–568, J. Mater.

Chem. A 1 (2013) 2691–2718, Adv. Mater. 28 (2016) 2855–2873]. COFs showed different

feature in the low-pressure and highpressure adsorption of CO2 [Adv. Mater. 28 (2016) 2855–

2873]. At low-pressure, COFs’ CO2 capture capability was found associated with its pore

structure. COFs with pore size <1 nm usually presented good CO2 adsorption capacity.

Furthermore, grafting polar groups were found effective to boost its CO2 adsorption capacity.

At high-pressure, higher BET surface area and pore volume of COFs was better for higher CO2

adsorption capacity. Compared to boron-based COFs, imine-based and triazine-based COFs

21
showed relative higher CO2 uptake capacity regardless of the CO2 pressure. Nitrogen-doping

and fluorine-doping COFs showed high performance in the capture of H2 and CO2. Among a

series of N-doped covalent triazine frameworks reported by Tuci, et al. [Adv. Funct. Mater. 27

(2017) 1605672], the highest CO2 uptake of 5.97 mmol/g at 273 k and 1 bar was identified.

Whilst, among a series of F-doping covalent triazine frameworks reported [J. Mater. Chem. A 6

(2018) 6370–6375], the highest CO2 uptake of 5.98 mmol/g at 273 k and 1 bar was observed.

The comparison between two 2D COFs with different pore sizes by adjusting the length of

monomers demonstrated that smaller pore size of COFs (8.1 Å and 13.8 Å for SIOC-5) was

superior in the capture of CO2, despite of its lower surface area [Chem. Commun. 52 (2016)

11704–11707]. In fact, these sizes are still bigger than that of the kinetic radius of CO2 ( ⁓0.33

Å). That may be the reason of untramicropore COFs’ (5.2 Å for CTF-FUM and 5.4 Å for CTF-

DCN) excellent CO2 capture ability and selectivity over N2 and CH4 [Environ. Sci. Technol.

50].

3.2. Filtration/Seperation

The study on COFs-based membranes for separation is still in its infancy, far from

industrial application. Several strategies have been adopted to fabricate COFs-based

membranes. As COFs are usually fabricated in powder form, COFbased composite membranes

can be fabricated by incorporating COFs powder within a polymer matrix [J. Mater. Chem. A 6

(2018) 17602–17611, J. Membr. Sci. 526 (2017) 355– 366] or by growing them on supports [J.

Membr. Sci. 523 (2017) 273–281, Environ. Sci. Technol. 51 (2017) 14352–14359]. However,

the cover effect of polymers, poor growth of crystal, and COFs’ poor adhesion to the support

may disfavor COFs-based membranes’ separation performance [Environ. Sci. Technol. 51

(2017) 14352–14359]. The exfoliation of COFs into its nano-sheets by n-BuLi intercalation to

improve its interface compatibility towards polyvinyl alcohol matrix was reported [J. Colloid
22
Interface Sci. 539 (2019) 609–618]. Recently, various methods for the fabrication of thin films

of COFs were reviewed [Chem. Soc. Rev. 48 (2019) 488–516].

Water treatment: membrane-based separation technology (e.g. microfiltration,

ultrafiltration, nanofiltration, and reverse osmosis) has been an important commercial option in

water treatment owing to its low cost, high efficiency, and ease of operation. The recent

developed COFs-based membrane materials have also attracted great attention in this field.

Theoretical studies have revealed COFs’ great potential in desalination [Chem. Commun. 51

(2015) 14921–14924]. Several successful attempts also proved their feasibility as filter

membranes in water treatment [J. Am. Chem. Soc. 139 (2017) 13083–13091, J. Membr. Sci.

523 (2017) 273–281, J. Membr. Sci. 520 (2016) 583–595,97]. 3D COFs (SNW-1) were

incorporated into polyamide (PA) layer of membrane by interfacial polymerization [J. Membr.

Sci. 523 (2017) 273–281]. The addition of COFs with abundant secondary amine groups into

membranes via covalent bonds modified membranes’ hydrophibility and stability. Moreover,

better desalination performance, e.g., higher pure water flux without the sacrifice of the salt-

rejecting ability (>80%), and long-term running stability, was observed after modification. This

was greatly attributed to the presence of SNW-1 with high porosity and suitable pore size,

which provided more passageways for water transport. Similarly, SNW-1 was asymmetrically

immobilized in alginate matrices for ethanol hydration via membranes distillation in other

report [J. Membr. Sci. 520 (2016) 583–595]. The hybrid membranes with incorporating COFs

(pore size 5 Å) were appropriate for the transfer and recognition of water molecules (2.6 Å),

leading to a high permeation flux (2397 g/(m2h)) and separation factor (1293) at the given

operating conditions [J. Membr. Sci. 520 (2016) 583–595]. To improve COFs membrane’s

separation performance, i.e., to enhance their mechanical strength, a mixed-dimensional

assembly strategy to combine 1D cellulose nanofibers (CNFs) with 2D COFs was adopted
23
[Nat. Commun. 10 (2019) 2101]. Higher separation performance in alcohol dehydration (e.g.

separation factor of 3876 for n-butanol dehydration) and salt/dye rejection was achieved, as

well as improved mechanical strength owing to the multiple interaction (Fig. 8a). A hybrid

GO/COF-1 membrane was fabricated and tested for salt/dyes separation, ending with over 310

L m-1 h-1 MPa-1of water permeation and >99% rejection of water soluble dyes, but <12%

rejection for ion salts [J. Membr. Sci. 581 (2019) 321–330]. As shown in Fig. 8b, the physical

size sieving effect arising from aperture size of COF-1 (1.5 nm) and the interlayer spacing of

GO layers (0.33 nm) accounted for its high water permeation and dye rejection performance.

This study unveiled the feasibility of COFs-based hybrid membrane for water treatment. The

study on the influence of TpPa-1 nano-sheets’ stacking fashion and the number of stacked

monolayers on the ion rejection and water transport was conducted [ACS Appl. Mater.

Interfaces 11 (2019) 16847–16854]. Non-equilibrium molecular dynamics model taking into

account the interior and interfacial resistances for water transport was developed to illustrate

TpPa-1 nano-sheets’ desalination behavior. It was found that the offset-eclipsed fashion reduced

its pore size from ⁓1.58 nm (fulleclipsed fashion) to ⁓0.89 nm, but increased its ion

rejections at the expense of water permeance. Furthermore, the increasing stacking number of

both fashions results in lower water permeance but higher ion rejections due to increased

transport resistance. 25 layered offset-eclipsed TpPa -1 nano-sheets demonstrated much better

performance than that of commercial nanofiltration membranes

Gas separation: COFs-based membrane also showed great potential in gas separation.

Various COFs with different pore sizes and linkages have been explored as fillers in the mixed

matrix membranes. The instinct nature of COF fillers is similar to polymer matrices, which is

good for the affinity between the dispersed filler phase and the continuous polymer matrix

phase. COFs as fillers in the membrane materials showed a synergistic effect with the polymer
24
matrix, resulting the composite with a higher selectivity and permeability. The loading

percentages of COFs in the mixed-matrix membranes has great effect on its separation

performance. A small amount of COFs’ loading can improve its performance owing to the

contribution of well-distributed channels. However, excessive COFs in the membranes may

make the gas transport pathways tortuous, which may account for its performance degradation.

For example, the synthesized membranes with 0.4 wt% COF-5 fillers showed the optimal CO2

permeability and CO2/N2 selectivity, compared to other COF-5/Pebax membranes with

different percentages of blending [J. Membr. Sci. 572 (2019) 588–595]. The mixed matrix

membrane composed of azine-linked COF (ACOF-1) as fillers and commercially available

Matrimid 5218 as the supporting polymers was fabricated for CO2/CH2 separation. A more

than doubling increase in CO2 permeability and a slight higher selectivity were observed after

the loading with 16 wt% ACOF-1 [J. Gascon, Chemistry 22 (2016) 14467–14470]. This study

indicates the good potential of COFs in mixed matrix membrane for gas separation. Two water-

stable 2D nano-sheet/monolayer COFs with different pore sizes (NUS-2: 0.8 nm; NUS-3: 1.8

nm) were incorporated into commercial polymers for gas separation, respectively. Owing to the

good compatibility between the fillers and the polymeric matrixes, the as-prepared membranes

showed defect-free structures and presented increase gas permeability. Furthermore, it was

inferred that the smaller pore size of NUS-2 in the membrane increased its CO2 affinity. It

implied the importance of COFs’ pore size in gas separation [Chem. Mater. 28 (2016) 1277–

1285]. In other study, two stable COFs of different pore sizes (TpPa-1: 1.8 nm; TpBD: 2.4 nm)

were homogenously incorporated into PBI-BuI matrix, respectively, and they both showed a

seven-fold increase in gas permeability in relative to pure polymeric membranes [Chemistry 22

(2016) 4695–4699]. It was deduced that larger pore size of TpBD contributed significantly to

permeability with only a small deviation in selectivity. Besides pure COFs as fillers,

25
MOF@COFs hybrids have also been used as fillers to enhance the gas separation performance

of the mixed matrix membrane [J. Membr. Sci. 573 (2019) 97–106]. TpPa-1 COFs layer

increased the compatibility between fillers and polymeric matrixes, resulting in homogenous

texture. An increase of 48% CO2 permeability and 79% CO2/CH4 selectivity was achieved on

only 5 wt% MOF@COF blending mixed matrix membrane in relative to pure PSf membranes.

3.3. Catalysis

COFs-based catalysts are usually utilized in the photo-catalysis or electro-catalytic

reaction for the clean energy conversion (e.g. CO2 reduction and H2 production) or pollutant

degradation For metal-free COFs [J. Mater. Chem. A 4 (2016) 9413–9418], its catalytic activity

can be related to the heteroatoms (e.g., N, P, and S) of the functional moieties in its structure.

Furthermore, most COFs with P-conjugated structures present semiconducting behavior owing

to the in-plane p-electron conjugates and the axial charge transport in the stacking direction,

thus making COFs promising photocatalysts for light collection and charge transmission

[Chem. Soc. Rev. 41 (2012) 6010–6022]; for COFs containing metals, it is usually the metal

ions in the macrocyclic structure that play the catalysis role [Chem. Sci. 9 (2018) 3941– 3947];

for hybrid system, the synergic effect between COFs and the conductive supporting materials

(e.g. graphene and carbon nanotube) increases the catalytic activities. Additionally,

carbonization of COFs has also been reported to improve its electric conductivity [Appl. Catal.,

B 244 (2019) 25–35].

Energy conversion: conversion (e.g. CO2 reduction and H2 production) with the help of

COFs-based catalysts is another promising environmental application of COFs [Nat. Chem. 10

(2018) 1180–1189, Polymers 11 (2019) 31, Chem 5 (2019) 1632– 1647, ACS Mater. Lett.

(2019) 203–208]. Many reviews showed good summaries on it [Polymers 11 (2019) 31, J.

26
Mater. Chem. A 7 (2019) 4689–4698, ACS Appl. Mater. Interfaces 11 (2019) 11029–11060].

Among various COFs, covalent triazine frameworks attract more attention owing to their

nitrogen-rich and robust structures. Various COFs-based catalysts have been utilized in the

photocatalytic or electrocatalytic process for CO2 reduction into CO or for H2 fuel production

via water splitting. Besides CO, it is also possible to reduce CO2 to other high value-added

products. For example, a triazine-based covalent organic framework flexible film as

photocatalyst was utilized for the reduction of CO2 into formic acid (Fig. 8d) [J. Mater. Chem.

A 4 (2016) 9413–9418]. The high catalytic activity of 2D CTF should be due to the suitable

band gap, its good visible light-harvesting capacity, and highly ordered p electron channels.

The electrocatalytic activities of different metals-loading covalent triazine frameworks (CTF)

in CO2 electrochemical reduction were compared, including Cu, Co and Ni coordinated in the

sites offered by pyridine units [Chem. Sci. 9 (2018) 3941– 3947]. Ni-CTF was identified as the

best one compared to other metal-CTF and metal-porphyrin structures. Theoretical calculation

implied the contribution of low coordination of the CTF structure in enhanced performance.

Similar metal-loading strategy was also reported. COFs comprising cobalt porphyrins also

showed a good aqueous electrochemical reduction activity of CO2 to CO, with the Faradaic

efficiency of 90% at given conditions [Science 349 (2015) 1208–1213]. CTFs with rich

atomically distributed and high density active sites (NiN4) for efficient CO2 reduction reaction

was reported [Adv. Funct. Mater. 29 (2019) 1806884]. The as-prepared electrocatalyst (NiPor-

CTF) showed a long-term stability and a high selectivity toward CO2 reduction to CO with the

Faradaic efficiency of >90% at given conditions (⁓0.6 V to ⁓0.9 V). The calculation implied

that the improved performance should be due to the kinetic energy barriers reducing on the

active sites (*CO2 transition to *COOH). COFs-based catalysts have also been constructed for

the water splitting for H2 fuel production. So far, the state-of-the-art catalysts for the reaction

27
of water splitting are metal, especially noble metals (e.g. Pt). A combination of Pt and COFs

for photo-catalysis of H2 evolution under visible light showed a stable H2 production rate of

230 lmmol/(h g) [Chem. Sci. 5 (2014) 2789–2793]. Additionally, a 3678 lmmol/(h g) H2

production was achieved using CdS-COF hybrid as a prototype photo-catalyst [Chemistry 20

(2014) 15961–15965]. In practical application, nonprecious-metal-based catalysts are more

appreciated owing to its low cost. Banerjee, et al. [J. Am. Chem. Soc. 139 (2017) 16228–

16234] fab ricated cobaloximes-loading COFs for hydrogen evolution in the presence of

triethanolamine (TEOA) donor (Fig. 8c). The evolution rate of 782 lmol H2/g was identified.

In this system, outer sphere electron was transferred from the COFs to the co-catalysts,

followed by a monometallic pathway of H2 generation. This study explored the improved H2

evolution efficiency by optimizing electron transfer pathway between COFs and co-catalysts.

Furthermore, the first COFs without metals nor heteronuclear molecular functionalities for

photocatalytic hydrogen generation was reported, and it was found that the diacetylene-based

COFs outperformed acetylene-based COFs in catalytic activity [, J. Am. Chem. Soc. 140

(2018) 1423–1427]. Adjusting the nitrogen content was found to be possible and effective in

improving aryl triphenyl azine COFs’ (Nx-COFs) photocatalytic hydrogen evolution capability

[Nat. Commun. 6 (2015) 8508]. A stepwise increase in N% of the central aromatic ring can

lead to higher hydrogen evolution owing to the electronic and steric variation [Nat. Commun. 6

(2015) 8508]. This strategy was extended to a pyridine-based COFs, and the complex interplay

between extrinsic and intrinsic was explained using the vertical radical anion stabilization

energy and the conformational flexibility of the PTP-COF precursor [Faraday Discuss. 201

(2017) 247–264].

Degradation of pollutants. The study of COFs-based catalysts for pollutants degradation

is still in its infancy. The first COF-based catalyst (Pd/COF-LZU1) for Suzuki-Miyaura
28
coupling reaction was reported in 2011 [J. Am. Chem. Soc. 133 (2011) 19816–19822]. To date,

there are only a few examples of COFs-based catalysts for pollutants degradation [J. Hazard.

Mater. 369 (2019) 494–502, Cryst. Growth Des. 18 (2017) 883–891]. The photocatalytic

degradation of phenol and methyl orange dye by TpMA COFs was studied. This COFs was

prepared by mechanochemical method at ambient temperature, and its optimal synthetic

condition was also studied. Different amounts of solvents/catalyst added during ball milling

process resulted in two morphologies of COFs. The ribbon-like structure was found more

efficient than that of thread-shaped structure in charge transportation. The ribbon-like structure

one showed remarkable photocatalytic activities towards phenol’s degradation by visible light

and presented a good reusability [J. Hazard. Mater. 369 (2019) 494–502]. The photocatalytic

degradation of tetracycline hydrochloride and ciprofloxacin was studied. To increase the

photocatalytic activity, BiOBr nano-flakes was coupled with 3D COF (CTF-3D) by

precipitation. 2 wt% of CTF-3D doping in the composite presented the highest photocatalytic

activity towards the pollutants under visible light. O2- was confirmed as the main radical

species by trapping experiments. The increased photocatalytic activity after CTF- 3D’ doping

may be owing to the synergistic effect of enlarged optical adsorption range, the efficiently

separated photo-generated electron-hole pairs, and the accelerated adsorption and transfer of

antibiotic molecules [Cryst. Growth Des. 18 (2017) 883–891].

4. Summary and future prospects

The study on COFs is still at the very initial stage, but their great potential in

environmental remediation has been gradually noticed. The past years have witnessed the

report of first COFs, the development of stable COFs and luminescent COFs, and the gradually

mature of their synthetic and functional methods. These achievements pave their ways to

versatile applications, especially environmental remediation.

29
 Stable COFs present versatile roles in environmental remediation. They prove to be

good adsorbents, filter membranes, chemosensors, and catalysts for pollutants/gas adsorptive

removal, membrane-separation, selective detection and catalytic degradation from water/air.

Additionally, the utilization of simulation calculation based on DFT methods to explore the

structure of COFs and their interaction with targeted pollutants/gas is becoming more and more

popular. As adsorbents, COFs present good adsorption performance towards many heavy metal

ions, radionuclides, and toxic organic pollutants from aqueous solutions, as well as some gases

(e.g. H2, CO2, and volatile iodine). Their high surface area and abundant porous channels are

good for the fast diffusion of pollutants into their inner structures. Whilst the introduction of

affinity groups via bottom-up or post synthetic method increases their adsorption affinity and

capacity towards the specific metal ions (e.g. mercury and uranyl ions).

The hydrophobic skeleton structures of COFs with tailorable pore size also endow them

with good adsorption performance towards toxic organic pollutants via molecular sieving, H

bonding, pi–pi interaction, etc. As a vital component of filter membrane, COFs with the merits

of nanosize ordered pore structure could provide selective passages for molecule sieving.

Powdery COFs could be utilized as fillers in active layer of membrane materials; whilst COFs

thin membranes fabricated by interfacial polymerization with internal defaults-free could be

sandwiched by porous supporters as assembling membranes. The incorporation of COFs in the

membrane structure also greatly affect the mechanical strength and hydrophilic/hydrophobic

properties. The as-prepared COFs-based membranes could be utilized for water treatment or

gas separation to obtain a higher permeation flux and separation factor than conventional

membranes. The luminescence characteristics of COFs and their luminescence

enhancement/quenching with the combination of targeted ions/molecules make COFs

promising chemical sensors for the detection of many types of cations, anions, and organic
30
molecules. However, owing to the limited reports on luminescent COFs with stable structure

and high emission efficiency, the exploration of luminescent COFs as chemical sensors remains

a long way to go. COFs-based catalysts could also be utilized in the photocatalysis or electro-

catalytic reaction for the clean energy conversion (e.g. CO2 reduction and H2 production) or

pollutant degradation. Various types of COFs-based catalysts, including metal-free, heterometal

and hybrid ones, have been fabricated to increase their catalytic activities.

Compared to clean energy conversion, pollutant degradation with the help of COFs-

based catalysts was less reported. The exploration on COFs’ versatile environmental

applications is still in its infancy and facing great challenging. Some challenges and

suggestions were discussed as follows: Firstly, to date, the construction of functional COFs

with desired characteristics (e.g. adsorption affinity, intact-free, and luminescence) remains a

challenging work. More efforts should be made in the development of functional COFs for

various environmental applications. Currently, COFs particles and their disordered alignment

greatly suppress their performance as filtration membranes for water treatment and gas

separation. Future investigations should also be given into the development of film or

membrane technology for the fabrication of COFs-based materials. Secondly, owing to the

initial stage of research, the structureperformance relationships of COFs remains uncertain.

Based on the understanding of structure-performance relationships, more studies on the

construction COFs-based functional materials and the exploration on their environmental

applications should be performed. Thirdly, to date, the synthetic cost of COFs-based materials

is relative expensive, which hinders their practical utilizations in environmental remediation.

The further development towards synthetic and functional methods to reduce its synthetic cost

and for large scale production is highly desired. Finally, the environmental risk assessment of

COFs has not yet been reported in relevant literature.

31
 With more attentions were given to COFs environmental applications, their

environmental risk assessment should be made in the future to evaluate their feasibility in

practical environmental applications. Metal-organic frameworks (MOFs) are also an emerging

class of porous crystalline materials, demonstrating good prospects in environmental

applications. Both materials show the advantages of porosity and topology, which makes them

super in selective isolation of gas and other pollutants via size effect. Whilst, they can also

adsorb pollutants inside their pore structures via various interaction forces, leading to the

adsorptive removal of pollutants. Furthermore, with luminescence characteristics, both

materials can be utilized for the detection of specific pollutants. Generally speaking, despite the

similar advantages of both materials in environmental applications, the study on COFs’

environmental application is less owing to the initial research stage since 2005. However, they

also show some differences in structure and performance. Compared to MOFs, COFs are

metal-free materials, which makes them lower in density. If pollutants fully occupied the pore

structures of both materials of similar pore volume, COFs of lower density would show a

higher adsorption capacity (mg/g). Furthermore, the pore structure (e.g. pore size and shape) of

COFs is easy to be tailored by various methods, offering abundant options for the selective

isolation of different pollutants. Additionally, pore-wall engineering of COFs is of great

importance to enrich COFs with more active adsorptive sites or catalytic sites. Above all, COFs

are promising materials in various environmental applications.

32