Porous sorbents for the capture of radioactive iodine compounds: a review

The number of studies on the capture of radioactive iodine compounds by porous sorbents has regained major importance in the last few years. In fact, nuclear energy is facing major issues related to operational safety and the treatment and safe disposal of generated radioactive waste. In particular during nuclear accidents, such as that in 2011 at Fukushima, gaseous radionuclides have been released in the off-gas stream. Among these, radionuclides that are highly volatile and harmful to health such as long-lived 129I, short-lived 131I and organic compounds such as methyl iodide (CH3I) have been released. Immediate and effective means of capturing and storing these radionuclides are needed. In the present review, we focus on porous sorbents for the capture and storage of radioactive iodine compounds. Concerns with, and limitations of, the existing sorbents with respect to operating conditions and their capacities for iodine capture are discussed and compared.


Introduction
Nuclear power plants are one of the main sources of electrical energy. Nuclear energy is safe, clean, without excessive greenhouse gas emissions and economically competitive. Although it is facing major issues related to operational safety and the treatment and safe disposal of generated radioactive waste, 1,2 it remains an important part of today's energy mix. One of the main problems is as follows: during the aqueous reprocessing of used nuclear fuels and/or during a severe nuclear accident (Chernobyl in 1986 and Fukushima in 2011), there is a strong risk of the emission of gaseous radionuclides in the released off-gas stream. 3,4 Gaseous radioactive wastes are a direct threat to the population and the environment owing to their easy dispersion through the atmosphere. The main components of such gaseous waste streams include the ssion products technetium ( 99 Tc), cesium ( 137 Cs) and strontium ( 90 Sr), as well as actinides, lanthanides and various volatile radionuclides ( 129 I, 131 I, 3 H, 14 C, 85 Kr, etc.). [5][6][7][8] Particular attention should be paid to iodine compounds, which are particularly abundant. 129 I is a highly volatile long-lived isotope with a half-life (t 1/2 ) of $1.57 Â 10 7 years, which needs to be captured and reliably stored during its long decay. In contrast, 131 I is a volatile short-lived isotope with a t 1/2 of $8.02 days, which needs to be captured immediately aer being released, as it tends to accumulate and become concentrated in the thyroid gland, which seriously affects human metabolic processes. [9][10][11][12][13][14][15] In addition, radioactive iodine is likely to form organic compounds such as methyl iodide (CH 3 I) with hydrocarbons and other volatile organic compounds present in the gas stream. Therefore, there is a strong interest in the nuclear energy community to develop alternative and effective means of capturing and storing these radionuclides. An efficient solution for iodine capture may consist of combining existing ltration devices such as wet scrubbing methods 8,16 (Mercurex, Iodox, electrolytic and caustic scrubbing), sand bed lters and metallic lters with a supplementary capture step that potentially uses solid inorganic porous materials. Porous sorbents have been in the forefront of the removal of radioactive contaminants thanks to promising results and signicant advantages such as high removal efficiency and adsorption capacities, high thermal stability, low maintenance cost, and the large range of structures and functionalizations that can be obtained. 17 Numerous studies were carried out from the early 1960s to the late 1980s on the use of sorbents for the capture of radioactive iodine compounds with a specic focus on nuclear accidents   (Fig. 1). These studies rst focused on activated carbons and their forms impregnated with KI and/or TEDA 18,22,23,27,36,43,51,57,58 and silver-exchanged zeolites for the capture of iodine. 17,21,[29][30][31][32]34,37,38,40,47,52,56 These sorbents were highly praised for their good performance in iodine capture and were used in industrial applications for the trapping and treatment of iodine Habiba Nouali was born in 1960 in Mulhouse, France. She received her PhD in organic chemistry from Strasbourg University in 1990. Aer her postdoctoral research at Texas A&M University (1990)(1991), she worked at the Maison de la Recherche en Environnement Industriel in Dunkerque, France until 2009. She then moved to the Institut de Science des Matériaux de Mulhouse, University of Haute-Alsace. The research in her laboratory focuses on the synthesis and study of porous and nanostructured materials for applications in the elds of adsorption, heterogeneous catalysis, ion exchange, environmental protection and energy storage.
Grégoire Augé was born in 1964 in Paris, France. He obtained his PhD at INA-PG (France) in 1994. He started his career by teaching applied mathematics, statistics, and data mining. Then he focused his activity on numerical and computational methods for the treatment of nuclear fuel and waste and the development of simulation tools in an IT group. Since 2011 he has worked in the ONET Group and has extended the scope of his activity to focus on methods, chemical processes and tools for nuclear decommissioning and dismantling. He is currently Project Director-Coordinator of R&D Programs and D&D and Waste Expert at ONET Technologies in the Technical Department.
Virginie Lalia was born in Talence, France. She obtained her master's degree in chemistry in 2002 at ECPM (Ecole Européenne de Chimie, Polymères et Matériaux), Strasbourg. Aer her master's specializing in quality and the environment, she joined the group ONET Technologies as a process engineer and worked on the development of various chemical decontamination processes. She is in charge of the laboratory at ONET Technologies where pilots and tests on chemical processes are developed.

Activated carbon
Activated carbons (ACs) are materials that are essentially composed of carbonaceous matter with a porous structure. Thanks to their low production cost, fairly large specic surface area (300-4000 m 2 g À1 ), easily tuned structure and porosity (pore size ranging from 4.5 to 60Å), as well as good adsorption capacities, activated carbons are particularly attractive for the capture of radioactive iodine compounds. 144 Activated carbons are generally produced in two steps: rstly by carbonization (or pyrolysis), and secondly by an activation step. 144,145 The carbonization step generates porosity within the crude carbonaceous matter by removing elements other than carbon. Aerward, an activation procedure is carried out to increase the adsorption capacity of the material. Different kinds of activation procedure can be carried out, namely, physical activation (in the presence of water vapour and air under pressure) or chemical activation (usually by treatment with phosphoric acid). 145 Furthermore, it is possible to increase the adsorption capacities of activated carbons by a wise choice of the crude starting materials (charcoal, bamboo, coconuts and walnuts) and by adjusting the carbonization 146 and activation steps. 145 Iodine species such as iodine (I 2 ) and methyl iodide (CH 3 I) are generally physisorbed on activated carbons, but in real conditions of utilization at a high relative humidity and high temperature (in a nuclear power plant or during a nuclear incident), the efficacity of the adsorption of such materials tends to be reduced owing to competition for adsorption sites between water and iodine species. For this reason, it is necessary to increase the affinity of activated carbons for iodine species, which can be achieved by functionalizing or impregnating the material with organic and/or inorganic compounds. The commonest compounds used for the treatment and capture of iodine compounds are triethylenediamine (TEDA) 6,18,43,69,73,[147][148][149][150][151][152][153][154][155] and potassium iodide (KI). 18,22,23,43,47,58,149 Recent studies from 2010 to 2017 on the subject have essentially focused on the utilization of activated carbon impregnated with TEDA for the capture of radioactive iodine.
2.1.1. Activated carbons impregnated with TEDA. In scientic studies, triethylenediamine (TEDA, C 6 H 12 N 2 ), also known as 1,4-diazabicyclo[2.2.2]octane (DABCO), is the commonest impregnating agent used in activated carbon lters for the capture of radioactive iodine species. TEDA has a melting point and a boiling point of 158 C and 174 C, respectively, and displays strongly hygroscopic properties. To increase the iodine adsorption capacities of activated carbons, these materials are impregnated with TEDA with a concentration of usually between 0 and 10 wt%. 6,17,27,57,148,151,156,157 However, a decrease in the accessibility of the surface and pores was observed by Chinn et al. 158 and González-García et al. 147,157 when the TEDA concentration was increased, which can be explained by the formation of TEDA clusters that blocked the entrances of pores. However, despite this phenomenon, it has been recognized that the presence of TEDA in activated carbons increases their iodine capture capacity (especially for CH 3 I). 147,157 Deitz and Jonas 62 rst thought that the capture of radioactive CH 3 I by TEDA was a rst-order catalytic reaction. However, according to studies by Fessenden et al. 159 and Wilhelm, 57 TEDA impregnated into activated carbons is transformed by reacting with iodine compounds such as CH 3 I via SN2 nucleophilic substitution into a quaternary amine. 57,58,73,159 The general mechanism elaborated by these authors is illustrated in Fig. 2A.
On the other hand, Chun et al. 150 by means of energetics diagrams showed that the impregnation of TEDA into graphene helps to lower the activation energy barriers involved in the dissociation of methyl iodide (CH 3 I) into I À and CH 3 by the formation of a stable quaternary ammonium salt (Fig. 3). Kim 20 highlighted the fact that other neutral species can form during the interaction of methyl iodide with TEDA (Fig. 2B).
Besides, the mechanism of iodine (I 2 ) capture by TEDA has not been the focus of studies thus far. Nonetheless, several authors, such as Fessenden et al. 159 and Takahashi et al., 160 admitted that during a severe nuclear accident radioactive iodine (I 2 ) can decompose in the presence of water and/or oxygen into several iodine species (I À , IO 3 À , HOI) and can readily react with methane (CH 4 ) to form CH 3 I. 161 However, no information on the mechanism of the adsorption of these iodine species (except CH 3 I) by activated carbon or TEDA has been reported. Various studies have been carried out to determine the performance of activated carbon impregnated with TEDA in the capture of iodine species 18,23,27,43,67,73,147 (ESI, Table 1 †). TEDA drastically increases the performance and adsorption capacities of activated carbons for iodine species at low temperatures. However, in these studies the test conditions and corresponding performance evaluations were not always representative of real conditions (temperature > 80 C and relative humidity > 40%) that would be expected during a severe nuclear accident. Hence, major factors such as the temperature (T, C) and relative humidity (R.H.%) should be taken into consideration for data analysis.
Therefore, it is crucial to study the behaviour of activated carbons impregnated with TEDA in the presence of humidity. Upon an examination of all the studies carried out on this subject, humidity was revealed to be a key factor that controls the adsorption performance of activated carbons. 66,69,147 Nakamura et al. 162 demonstrated that water molecules formed clusters that were bound in a network by hydrogen bonds to the surface of activated carbons regardless of their hydrophobicity or hydrophilicity. Furthermore, it was also observed that water lled the porous network of the activated carbons, which hindered the accessibility of active sites for iodine species. Therefore, the adsorption efficiency of these sorbents declined signicantly owing to adsorption competition between water vapour and iodine species. However, the problems caused by humidity can be alleviated by the presence of TEDA in the activated carbons. In fact, methyl iodide (CH 3 I) interacts more strongly with the support. 150 Moreover, computational studies showed that TEDA attracts water molecules (to form hydroxide ions) that are localized on the surface and in the porous network of the adsorbent, which frees the active sites needed for the adsorption of iodine species. 150 Given that during a severe nuclear accident water is present in larger quantities than iodine species, the reaction of TEDA with water will be favoured. 159 During a severe nuclear accident, in a case in which the cooling system would stop running, the process and/or system would eventually be subject to an increase in temperature. Therefore, it is essential to study the adsorption capacities of activated carbon impregnated with TEDA at high temperatures, taking into consideration the fact that TEDA has a melting point of 158 C and an autoignition temperature of around 330 C. However, it was observed several times, notably in the study by Park et al. 73 on activated carbon and in the study by Ampelogova et al. 148 on carbon bres, that even at temperatures that are close to its boiling point TEDA maintains an acceptable capacity for the adsorption of methyl iodide thanks to chemisorption-type interactions (reaction of CH 3 I and TEDA to form a quaternary ammonium salt), in contrast to non-impregnated activated carbon, of which the adsorption capacity is controlled by physisorption-type interactions, although it was shown that the adsorption capacity of activated carbon impregnated with TEDA is reduced by a factor of 4 between 30 and 150 C. The contribution of TEDA to the adsorption (chemisorption) of CH 3 I was determined to be 25.4% at 70 C by the authors, whereas this contribution increases to 73.3% at 150 C. These observations conrm that TEDA plays an important role in the capture of methyl iodide at high temperatures in an irreversible way (chemisorption interactions) and can be used up to a temperature of 150 C (below the melting point of TEDA) in nuclear plants. It still needs to be conrmed that the autoignition temperature of TEDA and the thermal effect of its decomposition in the conditions of a real nuclear accident will not be detrimental to the capture performance of these materials. However, non-impregnated activated carbon loses adsorption efficiency when the temperature rises (weak physisorption interaction). Moreover, it can be noted that at low temperatures (30 C) non-impregnated activated carbon possesses adsorption capacities that are as good as, if not even better than, those when impregnated with TEDA (Park et al. 67,73 and González-García et al. 147 ). TEDA was able to counterbalance the detrimental effect of humidity on the adsorption capacities of activated carbons by freeing the active sites of the support (water has a stronger affinity for TEDA than for the active sites).
2.1.2. Activated carbons impregnated with KI. On an industrial scale, potassium iodide (KI) is an impregnating compound that is widely used with activated carbon lters for the capture of radioactive iodine. It is highly thermostable with a melting point and boiling point of 686 C and 1330 C, respectively. Potassium iodide improves the iodine capture performance of activated carbons via an increase in irreversible interactions. According to Deitz 18 and Zhou et al., 155 the mechanism of the capture of radioactive iodine species on activated carbons impregnated with KI involves isotopic exchange between the radioactive iodine species in the gaseous phase and the stable iodine in KI. The following reversible reaction is believed to occur: CH 3 131 I (g) + K 127 I (ads) 4 CH 3 127 I (g) + K 131 I (ads) .
Because this reaction is reversible, it is necessary to have an excess amount of non-radioactive iodine (K 127 I) to guarantee the efficient capture of radioactive iodine species. Activated carbons are generally impregnated with 0 to 10 wt% KI. 18,20,[24][25][26][27]43,51,148,149,155 Various studies, especially during the 1980s and 1990s, were carried out to determine the performance of the capture of iodine species on activated carbon impregnated with KI (ESI, Table 2 †). Upon an examination of the work by Chien et al., 149 it is clearly observed that potassium iodide signicantly increases the performance of activated carbon in the capture of radioactive iodine species. However, the lack of information on the test conditions used for evaluating the performance of these materials 149 and the fact that there are no comparisons between non-impregnated and KI-impregnated activated carbon 155 in some articles are regrettable. It should be noted that at a high concentration KI drastically limits the accessibility of the pores in the support.
As indicated in Section 2.1.1, the adsorption capacities of activated carbons are particularly affected by humidity. In comparison with TEDA, KI does not possess such a high affinity for water. In 1981, Decourcière 33 studied the evolution of the decontamination factor (DF) ‡ for an activated carbon impregnated with 1 wt% KI in the presence of humidity (in the relative humidity range of 40-96%). A decrease in the decontamination factor from DF ¼ 10 4 to DF < 10 2 for a relative humidity of 40% and 96%, respectively, was observed. The author concluded that the ltration system should not be used if the relative humidity is higher than 40% and should be equipped with a heating apparatus to reduce the relative humidity of air. However, during a severe nuclear accident, the relative humidity will probably exceed 40%. Therefore, it is essential that activated carbons impregnated with KI remain fully operational for the capture of iodine species even in conditions of high relative humidity. Qi-dong et al. 25 reached the same conclusion with an activated carbon impregnated with 2 wt% KI + 2 wt% TEDA. Kitani et al. 43 and Deuber et al. 27 achieved adsorption efficiencies § of 97% and 99%, respectively, at a relative humidity of >90%. In conclusion, thus far it has been difficult to estimate whether KI is directly affected by humidity or the loss of performance is caused only by blocking of the pores and the active sites of the activated carbons. Nevertheless, humidity is detrimental for activated carbons impregnated with KI, which is why a combination of KI and TEDA is generally (or widely) used. Kitani et al. 43 studied the inuence of temperature on activated carbons impregnated with 5 to 10 wt% KI at a relative humidity of 90%. A decrease in the efficiency of iodine adsorption with temperature from 97.8% at 25 C (5 wt% KI) to 93.6% at 70 C (despite an increase in concentration to 10 wt% KI for the latter value) was observed. We could argue that this slight decrease in adsorption efficiency may be due to a decline in the accessibility of the active sites in the activated carbon caused by the increase in the KI loading rather than the increase in temperature. For activated carbons impregnated with KI + TEDA, 25 the decontamination factor (DF) decreases by a factor of 10 2 as the temperature rises from 40 C to 70 C. Therefore, it is clear that the temperature plays a major role in the iodine capture performance of activated carbons impregnated with KI.
2.1.3. Limitations of impregnated activated carbons. There are several problematic issues with the use of activated carbons for the capture of radioactive volatile iodine species other than alterations at high temperatures and in the presence of humidity. Firstly, it is generally accepted in the scientic community that the aging of activated carbons is perfectly capable of altering their adsorption performance. [25][26][27]148 Deuber 26 studied the performance of activated carbons impregnated with TEDA and/or KI over a time period of 0 to 12 months at 30 C and 130 C with a relative humidity of 95% and 2%, respectively, for various bed lengths. It is clear from the penetration proles that were obtained that the iodine adsorption performance was affected over time. In the case of adsorbents impregnated with TEDA, one can observe an increase in the penetration of iodine species through a bed with a length of 25 cm by a factor of 10 2 at 30 C and a factor of 10 2 to 10 1 at 130 C aer 12 months. Similar results were obtained in the case of adsorbents impregnated with KI. Therefore, the aging of activated carbons can become a major issue in the long term, especially if a severe nuclear accident occurs at the end of life of this material. According to Jubin, 16 even though activated carbons are preferentially used in the nuclear eld, especially for the capture of radioactive iodine, these materials cannot be good candidates for several reasons: they are characterized by a low autoignition temperature and in contact with other compounds, such as NO x , their iodine adsorption capacities are reduced and the formation of explosive compounds is likely.
2.2.1. Mechanisms of iodine adsorption. Two thermodynamically viable iodine-containing precipitates are likely to be formed in silver-exchanged zeolites, namely, silver iodide (AgI) and silver iodate (AgIO 3 ). 37 The formation of the silver iodide species (AgI) was reported in various studies, 21,29,34,38,40,47,52,180 whereas Patil et al. 37 showed the formation of a silver iodate precipitate (AgIO 3 ). At quite elevated temperatures (>190 C), the latter precipitate decomposed to form silver iodide (more stable at higher temperatures, eqn (1)). Scheele and Burger 29 identied the various reactions that could potentially be triggered between iodine species (I 2 , CH 3 I) and active silver adsorption sites in silver-exchanged mordenite Ag 0 MOR (in which silver possesses different oxidation states, namely, a charged Ag + form and/or a reduced Ag 0 form). They highlighted the fact that most of these reactions are thermodynamically favourable for the capture of I 2 and CH 3 I by mordenite-type zeolites loaded with silver (AgMOR). The iodine species formed during these reactions are mainly AgI and AgIO 3 (eqn (2) to (8)): More recently, Chapman et al. 171 showed the existence of several AgI species in the same AgMOR/I 2 system. The structure adopted by silver iodide, as well as the distribution of different phases, depends on the temperature and the initial oxidation state of silver. 181 At temperatures below 147 C b-AgI (wurtzite structure) is the more stable phase, whereas at temperatures above 147 C a-AgI (cubic centred structure) becomes the more stable form. However, at high temperatures a third phase can coexist, namely, the metastable g-AgI phase. 181 The authors stated that silver is present in the form of Ag 0 particles with sizes of 3 nm on the zeolite surface. Chapman et al. 171 observed two AgI phases aer the adsorption of I 2 : a-AgI, which was mainly localized in the pores of the silver zeolite, and the metastable g-AgI phase, which was characterized by larger nanoparticles that were mostly localized on the zeolite surface (Fig. 4). According to the authors, these results involve the migration of silver in the zeolite pores during the adsorption of iodine. This migration is made possible by the mobility of both the iodine species (which are initially physisorbed into the pores) and the Ag + ions, which are localized directly inside the zeolite framework. In the absence of pre-treatment with hydrogen, the non-reduced silver localized inside the mordenite gives rise to the formation of AgI conned in the zeolite pores in the form of nanoscale a-AgI clusters.
In 2014, in a study that combined several ex situ characterization techniques, Nenoff et al. 173 proposed a mechanism for the trapping of CH 3 I by silver-exchanged MOR-type zeolite (Ag 0 MOR, in which silver is reduced to its Ag 0 form). This mechanism involves the acidic sites of mordenite and the catalytic decomposition of hydrocarbons derived from CH 3 I. In this study, the emission of various organic compounds was detected, such as dimethyl ether (DME), methanol, and methyl nitrite. These observations are consistent with a previous work by Heafner and Watson, 182 in which the presence of methanol in the off-gas stream was reported. Scheele 29 also suggested that methanol and DME are secondary products formed during the adsorption of CH 3 I onto silver. Therefore, Nenoff et al. 173 explained that CH 3 I decomposed on the Brönsted acid sites (Si-OH-Al) of the zeolite to form adsorbed methoxy species (CH 3 O-), which react with water to form methanol and consequently free the acidic sites of the zeolite. Dimethyl ether was observed aer the reaction of methanol with the adsorbed methoxy species. Finally, the iodine formed by the decomposition of CH 3 I is trapped in the form of an AgI precipitate. A recent work by Chebbi et al. 78 conrmed the observations by Nenoff et al. 173 (Fig. 5).
In addition, it was reported in several works that the formation of AgI in silver-exchanged zeolites was accompanied by the formation of various hydrocarbon compounds. Belapurkar et al. 31 reported the formation of methane (CH 4 ) and ethane (C 2 H 6 ) during the adsorption of CH 3 I on AgX zeolites in dry conditions. The formation of alkanes (CH 4 and C 3 H 8 ) and alkenes (C 2 H 4 and C 3 H 6 ) is linked to the decomposition of chemisorbed CH 3 I species. In the literature, these species were observed during the decomposition of halogenated molecules (CH 3 I, CH 3 Cl) on supports such as AgX zeolites 31 or aluminophosphates (HSAPO-34). 183 The formation of hydrocarbons is a result of the direct coupling of fragmented hydrocarbon species, which is catalyzed by the active sites of the zeolite (silver sites and/or Brönsted acid sites). At higher temperatures, Chebbi et al. 78 also observed the formation of carbon monoxide (T > 420 C) and iodine species such as HI and I 2 (T > 500 C). The formation of carbon monoxide could originate from the partial oxidation of methane (or oxygenated species) by water, whereas the iodine species most probably originated from the reduction-oxidation decomposition of silver iodide (2AgI / 2Ag 0 + I 2 ). Different silver phases (with different oxidation states) can be present in the pores of AgY zeolite, as shown by Chebbi et al. 78 By using diffuse-reectance UV-vis spectroscopy (DR-UVvis), the authors obtained information on the nature of the silver species conned in the zeolite pores. Absorption peaks around 208-328 nm have been attributed to silver-exchanged Ag + species, whereas absorption peaks beyond 305 nm are attributed to Ag 0 nanoparticles. The distinction between Ag n d+ and Ag m 0 (n and m are the numbers of atoms in the clusters) is particularly difficult because the absorption spectra depend greatly on the sizes of the clusters. However, absorption peaks between 240 and 255 nm and at 305 nm are related to Ag n d+ (n < 10) charged clusters and Ag m 0 neutral clusters, respectively. To prove their assumptions and conrm the accurate localization of the silver species inside the zeolite framework, the authors 78 studied the adsorption of CO at 35 C by diffuse-reectance infrared Fourier transform spectroscopy (DRIFT). This technique revealed that the silver species are well dispersed inside the framework of zeolites of the FAU structural type and that they are localized at the level of the exchanged active sites in the form of Ag + cations. Metallic Ag m 0 and charged Ag n d+ aggregates (clusters) were also found near the exchanged active sites, but in lower quantities. During the exposure of the silver-exchanged zeolite to a CH 3 I ow at 100 C, the Ag + cationic species reacted with CH 3 I to form AgI precipitates. The formation of AgI involved the reaction of Ag + with I À aer an initial step of the dissociation of CH 3 I. The dissociation of CH 3 I can occur on the silver active sites or the acidic sites of the zeolite. Then, AgI molecules tend to aggregate to form (AgI) n nanoclusters. The sizes of the (AgI) n aggregates increase continuously until they reach the size of the cavities and/or supercages of the zeolite. Chebbi et al. 78 also studied the effect of water vapour and temperature on (AgI) n clusters. They observed by different methods (XRD, DR-UV-vis) that some of these clusters migrated from the pores of the zeolite to its external surface. It was suggested that wet conditions promoted the formation of a microsolution inside the zeolite pores, which helped the silver to move in the pores and therefore permitted the coalescence or sintering of AgI species. To determine the reaction mechanisms, the desorption proles of the different products and compounds formed during the decomposition of CH 3 I on the active sites of AgY zeolite (Brönsted acid sites, eqn (9) and (10), and/or directly on silver sites, eqn (11)) were also studied.
HI + Ag-zeolite / AgI + H-zeolite (10) In the same context, Bučko et al. 172 used computer simulations at the periodic DFT level to investigate the dissociative adsorption of iodomethane (CH 3 I) onto silver-exchanged mordenite (AgMOR). Using an ab initio molecular dynamics study, the authors investigated the structure, energetics and mobility of Ag(CH 3 I) complexes in the mordenite zeolite structure. In summary, the results indicated that the mechanism of the dissociation of CH 3 I proceeded more in accordance with eqn (11) than (9) + (10). AgI species formed during dissociative adsorption were shown to combine spontaneously into small (AgI) n clusters, of which the dimensions were restricted by the size and geometry of the mordenite cages and pores.
To summarize, the formation of AgI precipitates is initiated by the dissociation of CH 3 I on the silver sites and/or Brönsted acid sites of the zeolite. AgI molecules and then (AgI) n aggregates/clusters are formed inside the cages of zeolites of the FAU structural type. In the presence of humidity and at elevated temperatures, AgI precipitates can readily migrate and form larger entities on the external surface of the zeolite. Furthermore, the thermal and/or catalytic decomposition of CH 3 I (dissociation) leads to the production of various by-products of hydrocarbons such as methanol, DME, alkanes, and alkenes.
2.2.2. Capture performance of Ag-exchanged zeolites. Numerous studies were carried out from the 1960s to the 1980s on the use of zeolites for the capture of radioactive iodine compounds in the event of nuclear accidents. In 1968 and 1970, Maeck et al. worked on more than twenty silver-impregnated zeolites in various operating conditions for the capture of iodine compounds (mainly I 2 and CH 3 I). 38,40 The best adsorption performance was achieved with AgX zeolite (FAU type). In comparison with silver-exchanged zeolites, additional studies conrmed that zeolites in which cations were exchanged for other metals such as Cu, Pd, and Cd have lower adsorption capacities and poorer performance (this phenomenon will be discussed below). Later, numerous studies were carried out on this topic, especially during the 1970s and 1980s 21,29,34,38,40,47,52,180 and then between 2000 and 2017. 5,8,71,[74][75][76][77][78][79][80][169][170][171][173][174][175][176][177][184][185][186][187][188] These studies focused on the determination of the inuence of various parameters such as the temperature ( C), silver content (wt%), Si/Al ratio, supercial velocity, and pretreatment of the material on the adsorption capacities of silver-exchanged zeolites (ESI, Table 3 †). Other works focused on issues such as aging of the material under different atmospheres (NO x , humidity). 178,185 Since 2016, we have observed a resurgence in studies on silver-exchanged zeolites for iodine capture. We can in particular mention the works by Chibani, Chebbi and Azambre, 74-80 who performed systematic studies on the inuence of zeolite parameters on their iodine adsorption capacities. These works, as well as those on the inuence of the other parameters mentioned above, will be discussed in the following section.

Inuence of different parameters on the capture performance of Ag-exchanged zeolites
Inuence of silver content. The silver content in an exchanged zeolite has a major inuence on its adsorption capacity for iodine species. In 2001, Choi et al. 71 studied the inuence of the silver content in a faujasite-type zeolite (AgX) on its iodine adsorption capacity. They incorporated 10, 20 and 30 wt% silver by exchanging Na + in NaX zeolite for Ag + . An increase in the adsorption capacity for iodine was observed when the silver content increased. At 100 C, the adsorption capacity rose from 180 g kg adsorbent À1 to 250 g kg adsorbent À1 for a silver content of 10 wt% and 30 wt%, respectively. However, a large silver content (>20 wt%) entails a decrease in the pore volume, which in consequence limits the accessibility of the pores for iodine species (diffusion limitation). 71 In fact, the micropore volume decreases from 2.37 Â 10 À4 m 3 kg À1 for a non-exchanged zeolite to 1.83 Â 10 À4 m 3 kg À1 in the case of the zeolite with 30 wt% Ag. The use efficiency of Ag was also estimated by calculating the Ag/I ratio. This ratio was optimal (>99%) for the zeolite with 10 wt% Ag, whereas it was only 85% and 68% for the zeolites with 20 wt% and 30 wt% Ag, respectively, at temperatures between 150 and 200 C. Because silver is expensive, there is interest in the incorporation of an optimal silver content with the highest efficiency. More recently, Cheng et al. 188 studied the adsorption of I 2 on a silver-exchanged FAU-type zeolite (AgX) at temperatures between 250 and 650 C. They observed a decontamination factor (DF) in the range of >10 3 for a silver content of greater than 15 wt%, in contrast to 10 1.5 to 10 3 for a silver content of less than 15 wt%. However, only slight differences in the decontamination factor were observed between exchanged zeolites with 15 wt% Ag and 20 wt% Ag, which seems to be in agreement with the study by Choi et al., 71 which implies that there was no real increase in the adsorption capacity at a high silver content. In studies by Chebbi et al. 77 and Azambre et al. 74 on several silver-exchanged zeolites (AgY, AgX, AgMOR, AgZSM-5, etc.) with different silver contents (0 to 35 wt%) at 100 C, an increase in the adsorption capacity of silver-exchanged FAUtype zeolites (AgY and AgX) from 87 mg g adsorbent À1 to 223 mg g adsorbent À1 for a silver content of 9.1 wt% to 22.8 wt% in AgY and from 149 mg g adsorbent À1 to 267 mg g adsorbent À1 for a silver content of 7.3 wt% to 35 wt% in AgX was observed.
Inuence of the zeolite type. Studies for a long time mainly focused on zeolites of the FAU and MOR structural types. 5,8,29,31,34,38,40,47,52,71,[169][170][171]173,175,177,178,182,184,185,187,188 The interest in these two types of zeolites can be explained by their large pore openings, which allow easier diffusion of iodine (I 2 , CH 3 I) and non-iodine (NO x , CO x , H 2 O, etc.) compounds inside the zeolites. Other zeolites, such as the MFI, *BEA and FER structural types, could potentially be suitable for iodine adsorption applications. Aer the study by Chebbi et al. 77 on the inuence of the silver loading, it was concluded that AgX zeolite exhibited the best performance in the capture of iodine compounds, followed closely by AgY, whereas the performance of MFI, *BEA and FER zeolites was not of as much interest. For a similar silver content ($9 wt%), AgZSM-5 and AgY have similar adsorption capacities ($85-87 mg g adsorbent À1 ).
However, these results differ greatly from those observed for AgX zeolite, which exhibited a saturation capacity of 149 mg g adsorbent À1 for a silver content of 7.3 wt%. The discrepancy between these silver-exchanged zeolites indicated that the silver content is not the only parameter that has to be considered for the adsorption of iodine species. Accordingly, AgX and AgY zeolites exhibit the best performance, which is principally due to their higher silver contents, in contrast to the other zeolites with low silver contents studied by Chebbi. 77 Moreover, Chebbi et al. 77 highlighted the fact that AgX zeolite has a cation exchange capacity that is much higher than that of AgY zeolite (iodine adsorption capacities of 234 mg g adsorbent À1 for a silver content of 23.4 wt% and 267 mg g adsorbent À1 for a silver content of 35 wt% in AgY and AgX zeolite, respectively), thanks to its low Si/Al ratio. In conclusion, AgX zeolite was estimated to be the best silver-exchanged zeolite in terms of performance. However, Azambre et al. 74 and Chebbi et al. 77 carried out further studies by examining the quantities of physisorbed and chemisorbed iodine species and those captured in the form of AgI precipitates (AgI) to determine which silver-exchanged zeolites display the best performance in the irreversible capture of iodine. By temperature-programmed desorption experiments in combination with FTIR spectrometry, they studied the adsorption and desorption proles of CH 3 I on different silver-exchanged zeolites (AgX, AgY, AgZSM-5, AgMOR, AgFER and AgBEA) to determine the quantities of physisorbed and chemisorbed iodine (CH 3 I) at saturation (Fig. 6). The reversibility of the capture of CH 3 I differs drastically depending on the structural type of the silver-exchanged zeolite considered. 77 X and Y faujasite zeolites that have large pore openings with a size of 7.4 Â 7.4Å and a silver loading of greater than 20 wt% exhibit the best performance in terms of iodine capture, and more than 90% of CH 3 I is irreversibly adsorbed (chemisorbed + AgI precipitate). In contrast to faujasite-type zeolites, the proportion of physisorbed CH 3 I on impregnated AgFER zeolite (4.2 wt% Ag) with pore sizes of 3.5 Â 4.8Å and 4.2 Â 5.4Å is signicantly higher (76 AE 16%). This behaviour cannot only be attributed to the low loading of silver, given that in the case of Ag*BEA zeolite (3.4 wt% Ag) with pore sizes of 5.6 Â 5.6Å and 6.6 Â 6.7Å the physisorbed component comprises only 33 AE 7%. The kinetic diameter of CH 3 I (between 5 and 6Å) is far greater than the pore sizes of AgFER zeolite (3.5 Â 4.8Å and 4.2 Â 5.4Å), which could partly explain why physisorption is favoured at the expense of the formation of an AgI precipitate. The structure of the zeolite seems to have a non-negligible inuence on the adsorption capacity and retention of iodine species, as zeolites with small pores favour physisorption, whereas zeolites with larger pores favour retention by chemisorption and the formation of an AgI precipitate. 77 Inuence of the Si/Al ratio in the zeolite framework. The amounts of silver that can be exchanged in a zeolite greatly depend on the Si/Al ratio in the zeolite framework. In fact, Chebbi et al. 77 and Azambre et al. 74 obtained silver loadings of greater than 20 wt% aer several cation exchange steps in NaXand NaY-type zeolites. However, in the case of MFI-, MOR-, FERand *BEA-type zeolites the silver loading never exceeded 10 wt%. The discrepancies between these structures can be linked to the fact that FAU-type zeolites have a low Si/Al molar ratio in the framework (1 < Si/Al < 3), which leads to a higher cation exchange capacities. Moreover, a study performed using density functional theory (DFT) 79 has shown that the interaction energies of iodine compounds increase considerably when the Si/Al ratio decreases (Fig. 7). The interaction energy increases from 145 kJ mol À1 to 190 kJ mol À1 and from 135 kJ mol À1 to 300 kJ mol À1 for CH 3 I and I 2 , respectively, as the Si/Al ratio decreases from 47 to 5. Chibani et al. proved via this study that (1) for a zeolite framework with a low Si/Al ratio contaminants such as H 2 O and CO have little or no effect on the adsorption capacities of silver for iodine compounds and (2) at the same Si/ Al ratios iodine compounds interact more strongly with silver, which enables the dissociation of I 2 and formation of AgI for the immobilization of iodine. Therefore, the Si/Al ratio has a strong inuence on the iodine adsorption capacity. Hence, a low Si/Al ratio has two advantages: it provides greater silver exchange capacities and favours interactions between silver and iodine compounds.  Inuence of pretreatment with hydrogen. Several studies have established that the pretreatment of silver-exchanged zeolites with hydrogen signicantly improves their performance in the capture of CH 3 I and I 2 . In 1978, Thomas et al. 60 reported an increase in the adsorption capacity for I 2 by a factor of two aer pretreatment with hydrogen (5 L min À1 H 2 , 500 C, 24 h) for silver-exchanged mordenite (Ag 0 MOR, Q ¼ 138 mg I 2 g Ag À1 , bed length 2.5 cm) in comparison with a non-pretreated silverexchanged mordenite (AgMOR, Q ¼ 71 mg I 2 g Ag À1 , bed length 2.5 cm). Subsequently, these results were conrmed by Jubin 28,64 and Scheele et al. 29 for the same type of zeolite. In 1980, studies led by Jubin 28 showed that pretreatment with H 2 (100% H 2 ) at high temperatures (200, 400 and 500 C) for a long duration (24 h and 48 h) led to a decrease in the adsorption capacity for iodine compounds (CH 3 I). At 200 C, the adsorption capacity decreased from 35.6 mg CH 3 I g À1 to 12.3 mg CH 3 I g À1 for a pretreatment duration of 24 and 48 hours, respectively. The same results were obtained at higher temperatures (400 C and 500 C). This sudden decrease in performance is explained by the formation of silver nanoclusters and large aggregates, which limits the accessibility of silver. Therefore, it is necessary to dene accurately the two parameters of pretreatment (temperature and duration) to maximize the adsorption capacities of silver-exchanged zeolites. With this in mind, Nan et al. 174 recently determined the optimal temperature for the pretreatment with H 2 of silver-supported mordenite (AgMOR, 12 wt%) for the adsorption of iodine (I 2 ). They showed that aer reduction by hydrogen (4% H 2 in argon, 500 mL min À1 ) for 24 hours the iodine adsorption capacity of Ag 0 MOR increased when the pretreatment temperature was increased from 170 C to 400 C. However, aer the temperature reached 400 C the adsorption capacity did not increase further. Studies by Zhao et al. 177 and Aspromonte et al. 169 conrmed these results.
Inuence of the operating temperature. Temperature is a crucial parameter that can signicantly inuence the adsorption capacities for iodine compounds of a silverexchanged zeolite. In 1982, Vance et al. 34 studied the thermal stability in iodine capture of the silver-exchanged zeolites AgY (28 wt%), AgX (37 wt%) and AgMOR (20 wt%). These three zeolites were rst saturated with iodine under dry conditions at 130 C. Then, the temperature was increased from 130 C to 1300 C to observe the effect of temperature on the iodine content in these zeolites. A signicant decrease in the iodine content captured in the zeolites was observed upon an increase in the temperature. The iodine content in AgY zeolite (28 wt%) decreased monotonically from 210 mg I 2 g adsorbent À1 to 50 mg I 2 g adsorbent À1 at temperatures of 130 C and 1300 C, respectively.
However, it can be observed that at temperatures above 700 C the zeolite framework generally tends to collapse and consequently loses crystallinity and becomes amorphous. This could explain the signicant decrease in the iodine content observed by these authors. 34 In 2001, the inuence of temperature on AgX zeolite loaded with 10, 20 and 30 wt% silver was studied by Choi et al. 71 An increase in temperature led to a decrease in the adsorption capacities of the zeolite for iodine compounds. For AgX zeolite (10 wt%), the adsorption capacity decreased from 180 mg g adsorbent À1 to 130 mg g adsorbent À1 at temperatures of 100 C and 400 C, respectively. Similar conclusions were reached by Belapurkar et al. 31 regarding the adsorption capacities for CH 3 I of dehydrated AgX zeolites at temperatures between 25 and 150 C. Furthermore, Cheng et al. observed a relatively signicant decrease in the decontamination factor (DF) for silver-exchanged AgX zeolite between 250 C and 650 C. In the case of a 15 wt% AgX zeolite, the decontamination factor decreased from 10 3.4 to 10 3.0 as the temperature was increased from 250 C to 650 C. From all these studies, it should be expected that the iodine adsorption capacity of a silver-exchanged zeolite decreases with temperature. However, Nan et al. 174 observed an increase in the adsorption capacity of silver-reduced Ag 0 MOR (12 wt%) upon an increase in the temperature from 100 C to 150 C. Nevertheless, the adsorption capacity still decreased at temperatures above 150 C. Nan et al. linked this behaviour to the inuence of water on the adsorption capacity and properties of mordenite. In fact, various studies have reported that water has a negative effect on the adsorption capacity for iodine. It was reported that the inuence of water is less pronounced at higher temperatures than at lower temperatures. The probability that silver will react with water to form silver oxide (Ag 2 O) or its hydroxide decreases. Furthermore, iodine compounds have easier access to the pores and adsorption sites of the zeolite.
Inuence of supercial velocity. Another factor that affects iodine adsorption performance is the ow rate of the carrier gas through the zeolite bed. According to Scheele et al., 29 the ow rate is governed by four parameters, namely, the supercial velocity, bed diameter, bed length and residence time. Few studies have examined the effects of these parameters on the iodine capture performance of silver-exchanged zeolites. The bed diameter and length and the supercial face velocity and contact time are interrelated. In a study by Scheele et al., 29 the authors decided to ignore potential effects of the bed size and contact time by arbitrarily selecting a bed diameter and bed length to exceed the expected zone of mass transfer and thus regrouped these factors into one variable, namely, supercial face velocity. In their work, the authors 29 conrmed that an increase in the supercial velocity from 3.75 m min À1 to 15 m min À1 decreased the adsorption capacities of AgMOR zeolites for CH 3 I from 71 mg g adsorbent À1 to 7 mg g adsorbent

À1
, respectively. This trend was also observed by Pence et al. 40 for AgX-type zeolites. A thorough knowledge of the related operating conditions in reactors and the associated kinetics is crucial for maximising the adsorption capacities of silver-exchanged zeolites.
Inuence of inhibitors. During a severe nuclear accident, several compounds such as H 2 O, CO, CO 2 , nitrogen oxides (NO x ), hydrocarbons and halogenated organic compounds are released. These compounds can act as inhibitors and poisons and thus limit the iodine adsorption capacities of silverexchanged zeolites.
Inuence of water. First of all, the inuence of water and humidity on a silver-exchanged zeolite was studied. In 1976, Thomas et al. 52,60 showed that very little variation in the iodine (I 2 ) adsorption capacity of Ag 0 MOR was observed in the presence of water (dew point 35 C) in the carrier gas ow (4 AE 12 mg I 2 g AgMOR À1 ). They concluded that the presence of water did not affect or only slightly affected the I 2 trapping performance, which was contradicted by more recent studies. The effect of water on the adsorption capacity for CH 3 I was also studied by Scheele and Burger. 29,32 The authors demonstrated that a water concentration of around 5 Â 10 À4 mol L À1 in the carrier gas was benecial for the adsorption capacity for CH 3 I (Q ¼ 139 g kg adsorbent À1 ) rather than a lower concentration of 4.3 Â 10 À6 mol L À1 (Q ¼ 30 g kg adsorbent À1 ). These observations were conrmed by Jubin, 28,64 who asserted that the presence of a moderate amount of water increased the adsorption capacities for CH 3 I of Ag 0 MOR zeolites, whereas a dry atmosphere or high relative humidity led to a decrease in the adsorption capacities. Water in the carrier gas seems to be benecial for increasing the adsorption capacities of silver-exchanged zeolites. However, no advanced studies have determined the exact conditions of humidity for the optimal utilization of silver-exchanged zeolites.
The adsorption capacities for CH 3 I were also studied by Belapurkar et al. 31 for temperatures between 25 and 150 C and degrees of hydration between 0 and 18 wt%. They determined that humidity has a negative effect on performance. This was particularly signicant at low temperatures, as a decrease in capacity of 56% was observed at 25 C in contrast to 17% at 150 C. Further studies by Belapurkar et al. 31 showed that in the absence of silver (NaX) the effect of hydration on adsorption capacities is far more pronounced than in the case of a silverexchanged zeolite (AgX). This discrepancy between the two zeolites can be explained by the fact that in zeolites in which no cations are exchanged for silver (NaX) adsorption phenomena are dominated by physisorption. In fact, notably at low temperatures, competition between water molecules and CH 3 I species occurs for adsorption sites. Consequently, the adsorption capacities for CH 3 I are reduced. In the case of silver-exchanged zeolites, chemisorption phenomena are prevalent. Therefore, the effect of water is reduced, especially at high temperatures.
In a similar way, Choi et al. 185 demonstrated that water (relative humidity of 50%) had a negative effect on the decontamination factor of an AgX zeolite (DF ¼ 3 Â 10 4 to DF ¼ 2 Â 10 3 at 200 C). Finally, Jubin et al. 189 studied the effect of longterm aging (up to 6 months) of a commercial Ag 0 MOR zeolite under dry and humid conditions before the adsorption of I 2 . Aer exposure to a dry atmosphere for 6 months, a decrease of 40% in the iodine adsorption capacity was observed. On the other hand, in the case of exposure to a humid atmosphere a decrease of 45% in the iodine adsorption capacity was observed in only 1 month. The authors explained that aging in dry conditions for 6 months is equivalent to aging for 1 month in humid conditions, which underlined the strong effect of water on the adsorption capacities of silver-exchanged zeolites.
In conclusion, despite a certain amount of contradiction in the literature, the general effect of water and/or humidity on the adsorption capacities of silver-exchanged zeolites seems to be negative. However, a low concentration of water is believed to favour the mechanism of the trapping of CH 3 I on silver.
Inuence of NO x . Nitrogen oxides, which are more commonly known as NO x , are acknowledged by the scientic community to inuence the retention and capture performance of silverexchanged zeolites. 28,63,64 NO x have a more or less adverse effect depending on the iodine species in consideration, their concentrations, the type of zeolite used and the operating conditions (supercial velocity, temperature, presence of other contaminants). Even though several studies reported ambiguous or conicting results, the majority agreed on the fact that NO x impair to a certain extent the adsorption performance of silverexchanged zeolites. In 1979, Holladay 63 studied the effect of NO and NO 2 on the iodine retention capacity of silver-exchanged mordenite (Ag 0 MOR). In the absence of NO x contaminants, the capacity for iodine that was measured was 113 AE 12 mg g Ag 0 MOR

À1
. Surprisingly, when 2% NO was added to the carrier gas the adsorption capacity of silver-exchanged mordenite increased to 129 AE 10 mg g Ag 0 MOR À1 . However, when 2% NO 2 was added to the feed gas, this time a decrease in the adsorption capacity could be observed (68 AE 12 mg g Ag 0 MOR À1 ). The author 63 assumed that NO acts as a reducing agent in the presence of oxygen to maintain silver in its highly reactive reduced metallic form (Ag 0 ). However, the presence of NO 2 allows the slow oxidation of silver to silver oxide (Ag 2 O), which is less reactive and would consequently decrease the adsorption capacity. The presence of both compounds (NO + NO 2 ) in the carrier gas slightly modied the adsorption capacity (119 AE 12 mg I 2 g Ag 0 MOR À1 ). Jubin 28,64 also carried out experiments on the adsorption of CH 3 I onto an Ag 0 MOR zeolite in the presence of NO x (0-3% NO and 0-3% NO 2 ). The results indicated that the presence of NO x slightly modied the adsorption performance. Recent studies led by Bruffey et al. 178 nally helped to conclude that the presence of nitrogen oxides such as NO x ([NO x ] ¼ 10 000 ppm in air) has a negative effect on the adsorption capacities of silver-exchanged zeolites for CH 3 I, with a decrease from 125 mg CH 3 I g Ag 0 MOR À1 to 56 mg CH 3 I g Ag 0 MOR À1 . In the presence of NO and NO 2 in the carrier gas, the adsorption capacity for CH 3 I decreases by 40% in comparison with the adsorption capacity determined in dry air alone. Furthermore, the presence of NO x allows the oxidation of CH 3 I to I 2 (ref. 178) according to the following thermodynamically favoured reactions 29 (eqn (12), DG ¼ À1079 kJ mol À1 , T ¼ 400 K and eqn (13), DG ¼ À853 kJ mol À1 , T ¼ 400 K): However, no experimental studies have proved the occurrence of such reactions during tests on the adsorption of CH 3 I and I 2 in the presence of NO x . Studies of the long-term inuence of NO 2 (200 ppm) on the adsorption capacity of a silverexchanged X zeolite (10 wt% Ag) were carried out by Choi et al. 185 These studies demonstrated that no signicant effect could be observed on the adsorption performance of the zeolite in the presence of NO x in the short term (several weeks). However, in the long term, i.e., aer poisoning by NO x for 16 months, the adsorption efficiency declined from 99.9% (DF ¼ 10 4 ) to 99.0% (DF ¼ 10 2 ) (Fig. 8).
Bruffey et al. 175 studied the iodine trapping stability of an Ag 0 MOR zeolite in the presence of 2% NO 2 for a duration of 1 to 4 months. The authors initially saturated the zeolite with iodine to give an average adsorption capacity of 72 mg I 2 g Ag 0 MOR À1 .
Then, aging and stability tests were carried out at 150 C with 2% NO 2 in the carrier gas for 1, 2, 3 and 4 months. The different tests showed an absence of released iodine aer exposure to NO x for 4 months. In conclusion, the presence of nitrogen oxide (NO x ) in the carrier gas or the surrounding atmosphere is believed to affect negatively the adsorption performance of silver-exchanged zeolites, but the effect is not drastic. According to the literature, this could be attributed to a modication of the oxidation state of silver (Ag 0 to Ag 2 O). However, in the case of a few studies the presence of NO x did not have any important effect on the iodine adsorption capacities of zeolites. Furthermore, aging studies show a more pronounced adverse effect in the long term rather than in the short term.
Inuence of organics and halogen compounds. Organics and halogen compounds are also acknowledged to have an inuence on the adsorption capacities of silver-exchanged zeolites. Jolley et al. 30 studied the effect of various volatile organic compounds (alkanes, alkenes, aromatics, ketones, alcohols, and chlorine-and bromine-containing compounds) on the efficiency of the adsorption of CH 3 I by silver-exchanged X zeolite (AgX). The adsorption capacity for CH 3 I was unaffected or only slightly affected by small molecules (such as hexane or ethanol), whereas in the case of molecules with larger sizes the performance of the silver-containing material was greatly affected in a negative way. Different families of organic compounds had the following inhibitory effects on the adsorption capacity of AgX zeolite: alkynes > alcohols > ketones and aromatics > alkanes; alkynes had the most detrimental inuence on the adsorption capacity. The size of the molecule appears to play an important role, as it sterically hinders access to the zeolite pores. Therefore, the capture of iodine by silver inside the zeolite pores is limited (diffusion limitation). However, this steric hindrance effect is not the only factor in the decrease in the adsorption capacity of the silver-exchanged zeolite. In fact, heteroatoms such as chlorine, bromine and oxygen and/or unsaturated compounds can be adsorbed onto silver sites thanks to the strong interactions that exist between them and silver. Halogenated compounds, especially chlorinecontaining compounds, tend to dissociate and release chlorine, which will poison the silver active sites. Surface analysis techniques (XPS, EDX) have shown that the adsorption of halogenated compounds (such as chloromethane or bromomethane) gives rise to the formation of metallic silver halide precipitates (AgCl and AgBr). From a thermodynamic point of view, chlorine is the main inhibitor of the adsorption of iodine onto silver-exchanged zeolites. Undeniably, because an AgCl precipitate (D f G ¼ À109.9 kJ mol À1 ) is more stable than an AgI precipitate (D f G ¼ À66.3 kJ mol À1 ), the presence of a large excess of Cl 2 , HCl or CH 3 Cl in the environment near a silverexchanged zeolite signicantly affects and impairs its performance. Burchsted et al. 56 reported in 1976 that a silverexchanged zeolite bed became totally ineffective for the capture of iodine nearly instantly aer HCl was introduced into it. Similar studies by Ackley et al. 49 showed that the presence of chlorine (Cl 2 ) signicantly impaired the iodine adsorption performance of an AgX zeolite in terms of the decontamination factor (DF). They reported for a silver-exchanged zeolite that was exposed to Cl 2 concentrations between 0.4 and 0.8 mmol mL À1 that the decontamination factor for a bed length of 10 cm decreased to DF ¼ 14-20 in contrast to DF ¼ 10 2 in the absence of chlorine. In conclusion, organic compounds, especially halogenated compounds, have a strong inuence on the iodine adsorption capacities of silver-exchanged zeolites. It can be observed that chlorine is the most strongly inhibiting compound for this type of material, which is principally due to its high affinity for silver as shown by the formation of a stable AgCl precipitate.
Inuence of other compounds. Other compounds can also be present during a severe nuclear accident, such as CO x (CO 2 and CO), SO x (SO 2 and SO 3 ) and, to a lesser extent, P 4 O 10 . 186 These compounds, which have not been studied in depth, seem to not signicantly affect the iodine adsorption performance of silverexchanged zeolites. However, in their presence, the diffusion of iodine species inside zeolite pores might become limited for steric reasons.
Despite a few contradictions in the literature, most studies acknowledged that the inhibitors mentioned previously (H 2 O, NO x , organics, halogenated compounds, CO x , SO x , etc.) generally have a negative effect on the iodine adsorption capacities of silver-exchanged zeolites. A recent study performed by Chebbi et al. 81 conrmed all the observations cited above.
Inuence of g-radiation. Little information can be found in the literature on the inuence of g-radiation on the iodine adsorption performance of silver-exchanged zeolites. 59 Evans 59 studied an AgX zeolite, which had been saturated with iodine beforehand, in the presence of g-radiation for exposure durations of up to 104 hours. These tests demonstrated that at temperatures of >45 C and in a water-rich atmosphere the destabilisation of trapped I 2 is favoured. The desorption rate was 0.1% per hour at a temperature of 80 C and a relative humidity of 90%. Aer desorption for 105 hours (in various operating conditions: 35 C # T # 80 C and 20% # R.H. # 95%), the total quantity of iodine released remained fairly low (approximately 0.74% of the initial quantity of iodine trapped beforehand). These results show that the stability of iodine trapped in silver-exchanged zeolites (AgX) is not signicantly affected by the presence of g-radiation. Recent studies by Chebbi et al. conrmed the absence of an inuence of g-radiation on the iodine adsorption capacities of AgY and AgX zeolites. 81 All these studies that were performed on silver-exchanged zeolites used for iodine capture and carried out under different operating conditions (inlet iodine concentration, bed length, etc.) are difficult to compare. Furthermore, in some studies the zeolites used were not completely crystalline. In addition, sometimes commercial and home-made zeolites with different loadings of silver were compared, which is not appropriate.
2.2.4. Zeolites loaded with other metals for the capture of radioactive iodine. One of the main drawbacks of silverexchanged zeolites is the inherent cost of the silver metal used. For this reason, other elements have been studied to obtain low-cost metal-exchanged zeolites for the adsorption of iodine. Between 1968 and 1970 Maeck and Pence 38-40 studied the iodine capture performance of approximately twenty zeolites impregnated with metals (Ag, Na, Cu, Pd, Tl, etc.). Their studies on NaX zeolite proved the great superiority of silver in terms of iodine adsorption over the other elements. The silverexchanged zeolite (AgX) exhibited an iodine adsorption efficiency of 99.9%, whereas those of the other exchanged zeolites were less than 30%. No elements other than silver proved to have acceptable performance for the adsorption of iodine. Aer these works, Staples et al. 52 studied the effect of the exchanged cations (Na + , Ag + , Pb 2+ , Cd 2+ ) on the irreversibility of the capture of iodine (I 2 ) by the faujasite-type X zeolite at 150 C. Even though the iodine adsorption capacity of every cation was acceptable, only silver was capable of trapping iodine in an irreversible way. Recent studies, 75,76,79,80 especially those using density functional theory (DFT) simulations, enabled the determination of the inuence of the valence of cations on the selective adsorption of iodine species by zeolites. In an initial study, Chebbi et al. 75 showed that copper (Cu + )-and silver (Ag + )exchanged faujasite zeolites (CuX and AgX, respectively) preferentially adsorbed iodine compounds rather than species such as water, chloromethane and chlorine. Furthermore, the adsorption of iodine species is much more pronounced on the CuX and AgX zeolites rather than on the protonated or sodium forms of faujasite zeolites. However, the presence of inhibitors such as carbon monoxide (CO) or nitrogen oxide (NO) limits the selectivity of Cu + zeolites for the adsorption of iodine. Despite its good performance in the adsorption of iodine, CuX zeolite cannot be used in the presence of elevated quantities of CO or NO, which would be present during a severe nuclear accident. In contrast to CuX zeolite, AgX zeolite displayed good adsorption of iodine species in the presence of NO. It can be observed that these modelling results do not agree totally with the experimental results described in the section on the inuence of inhibitors. In a second study, the inuence of the presence of Cu 2+ , Pb 2+ , and Hg 2+ in mordenite zeolite on its iodine adsorption performance, with a comparison of the adsorption energies of CH 3 I and I 2 , was investigated. In the presence of large quantities of contaminants (H 2 O, CO, CH 3 Cl, Cl 2 ), HgMOR was found to be the best adapted metal-exchanged mordenite for the adsorption of iodine. From a thermodynamic point of view, the tendency of a metal to form the corresponding halide (MI) can be estimated from the Gibbs free energy of formation (D f G , kJ mol À1 ). By comparing the energies of formation of halides and the corresponding oxides, it is possible to identify the most thermodynamically stable compounds. Table 1 lists the Gibbs free energies of formation of several halides and oxides. 190,191 In the majority of cases, the oxide will be preferentially formed rather than the halide form, except for silver (Ag) and mercury (Hg). It is therefore not surprising that in most studies the authors observed that other metals had low adsorption capacities in comparison with that of silver. In conclusion, studies on zeolites in which cations were exchanged for various metals (Na, Cu, Pd, Cd, Tl, etc.) proved that their efficiency in the capture of iodine species is low and showed that silverexchanged zeolites have the highest iodine adsorption capacities. Despite its inherent high cost, silver proved once again its ability to trap iodine compounds efficiently in an irreversible way. Furthermore, silver has the ability to form thermodynamically and chemically stable compounds with iodine.

Titanosilicates
Titanosilicates are a family that is analogous to aluminosilicate zeolites in which aluminium is replaced by titanium. Recently, Wu et al. 83 studied silver-exchanged titanosilicates (named as Table 1 Comparison of Gibbs free energies of formation of iodides and oxides (D f G , kJ mol À1 ) 190,191 Metal Iodide ETS-10 and ETS-2) for trapping iodine (I 2 ). The ETS-10 adsorbent is characterized by a stable chemical structure (isolated octahedral titanium chains in the silica network) with pores exclusively based on silicon. The composition of ETS-10 endows this adsorbent with hydrophobic properties and high chemical stability in acidic conditions. In contrast, ETS-2 is composed of sodium and titanium and possesses a reasonable specic surface area (260 m 2 g À1 ), which corresponds to its external surface, without structural microporosity. Unlike ETS-10, ETS-2 has a high capacity for cation exchange (with sodium), like that of zeolites. It was shown that the iodine adsorption capacities of silver-doped and silver-exchanged titanosilicates were elevated, at around 220 g I 2 kg adsorbent À1 and 243 g I 2 kg adsorbent À1 in dry conditions for ETS-10 (35 wt% Ag) and ETS-2 (40 wt% Ag), respectively. The effect of humidity on the adsorption performance of these materials was also studied. A non-negligible decline of 30% in the adsorption capacity was measured for the Ag-ETS-10 adsorbent. According to the authors, the porous structure of the ETS-10 adsorbent favours the condensation of water inside the pores, which limits the accessibility of the silver adsorption sites for iodine. In the case of the non-porous ETS-2 adsorbent, the performance was much less affected by the presence of water. ETS-10 titanosilicate-type adsorbents supported on a hollow carbon nanostructured polyhedral adsorbent (C@ETS-10) were also studied. 94 An adsorption capacity of 40 g I 2 kg adsorbent À1 at 20 C was recorded.

Porous oxide materials
2.4.1. Silver-doped silica and alumina (Ag/SiO 2 , Ag/Al 2 O 3 ). Silver-exchanged zeolites are not the only inorganic materials that are suitable for the capture of iodine compounds. Oxidetype materials such as alumina (Al 2 O 3 ) and silica (SiO 2 ) that are impregnated with silver, which are designated as Ag/Al 2 O 3 and Ag/SiO 2 , respectively, are another class of adsorbents that can be found in the literature. These materials have been the focus of various studies 19,35,52,61,65,70,[87][88][89][90][91][92][93][94][95][96][97] and have been used for the capture of iodine (I 2 ) on an industrial scale. 65,[88][89][90]96 These adsorbents, which contain silver nitrate (AgNO 3 ), proved their efficiency for the elimination of iodine (I 2 ) and iodomethane (CH 3 I) in a spent fuel reprocessing plant. The mechanism of the capture of iodine by these silver-impregnated materials has been detailed by Wilhelm et al. 48,54 Silver nitrate (AgNO 3 ) reacts with elemental iodine to form stable silver iodide (AgI) or silver iodate (AgIO 3 ) by the following reactions: 2INO 3 + AgNO 3 / AgIO 3 + 3NO 2 + 0.5I 2 (15) AgNO 3 + CH 3 I / CH 3 NO 3 + AgI An amorphous silver-doped silica, which was commercialized under the name of AC-6120, was used in the Karlsruhe spent fuel reprocessing plant (WAK) in Germany for the capture of iodine (I 2 ). 88 This adsorbent has a nominal BET surface area of 65-110 m 2 g À1 , a pore size distribution of 20-40 nm, a pore volume of 0.6 cm 3 g À1 and a silver loading of 8-12 wt%. Decontamination factors (DFs) for iodine of between 100 (>99.0% efficiency) and 50 (98.0% efficiency) were achieved during its utilization on an industrial scale. 48,54,88 On the laboratory scale, this silver-doped silica displayed decontamination factors for I 2 that were higher than 10 4 (>99.99% efficiency) at a temperature of 150 C (bed length of 10 cm, supercial velocity of 25 cm s À1 , in the presence of 1-5% NO 2 ). 48,54 The efficiency declined from 99.9944% to 27% at a relative humidity of 70% and 100%, respectively. In addition, Wilhelm and Schuttelkopf 48 stated that an increase in temperature up to 200 C slightly improved the performance. Nevertheless, an increase in temperature becomes less attractive in terms of balancing cost against efficiency.
On the other hand, Herrmann et al. 89 studied alumina materials that were impregnated with silver (10 wt% Ag and 24 wt% Ag), which were developed in Japan at the Tokai spent fuel reprocessing plant 87,192 for the retention of iodine species (I 2 and CH 3 I). The silver-doped alumina was tested at 150 C with a supercial velocity of 20 cm s À1 and an inlet concentration of CH 3 I of 30 ppm. A decontamination factor (DF) of higher than 500 (which corresponds to an adsorption capacity of 120 mg CH 3 I cm Ag/Al 2 O 3 À3 ) was achieved using alumina with 10 wt% Ag. For the alumina with the higher loading (24 wt% Ag), NO x (1.5%) were introduced into the feed. In this case the decontamination factor exceeded 500 and corresponded to an adsorption capacity of 350 mg CH 3 I cm Ag/Al 2 O 3

À3
. Therefore, this study demonstrated that the silver-doped adsorbents exhibited rather good resistance in the presence of NO x . In addition, it was reported 16 that an elevated concentration of NO 2 (between 1 and 10%) improved the iodine adsorption performance by preventing the reduction of silver (in the form of AgNO 3 ) to metallic silver Ag 0 . The adsorption efficiency was about 99.9961% and 99.9973% in the presence of 1% and 10% NO 2 , respectively, in contrast to 99.9944% in the absence of NO x . It was also shown that a large quantity of organic contaminants, as well as a high relative humidity (>70%), could impair the performance of the material in the absence of a sufficient quantity of NO x .
Finally, the authors highlighted the fact that these types of adsorbents (Ag/Al 2 O 3 and Ag/SiO 2 ) were generally less expensive by a factor of 3 to 10 than silver-exchanged zeolites. Furthermore, an atmosphere of NO 2 is needed to prevent the reduction of silver to its metallic form.
2.4.2. Silver-doped mesoporous silica. Another class of materials, namely, mesoporous structured silica, has aroused the interest of scientists for iodine capture applications. These materials have been the focus of many studies, especially in the elds of catalysis and adsorption. Mesoporous silica materials are characterized by amorphous silica walls that delimit wellordered mesocavities, a regular arrangement of mesopores and a particularly large specic surface area (it can reach values of higher than 1000 m 2 g À1 ). These materials also possess the advantage of having pore sizes (3-10 nm) that are larger than those in zeolites (<1 nm), which ultimately improve the accessibility of the pores and active sites in mesoporous silica. Furthermore, silanol functions, which are localized on the silica surface, can be replaced via functionalization with organic and/ or organometallic functional groups to improve their properties and performance depending on the intended applications. Few studies have focused on these materials for iodine capture applications. 81,84,86 Mainly, the works by Mnasri et al. on silverdoped MCM-41 mesoporous silica, 84 Yang et al. on bismuthdoped SBA-15 86 and Chebbi on silver-doped SBA-15 81 can be found in the literature.
The study by Mnasri et al. 84 focused on the adsorption of iodine compounds by three types of silver-impregnated MCM-41 mesoporous silica with the following pore diameters: 2.4, 3.3 and 3.8 nm. Silver that was introduced into the mesoporous silica was reduced by NaBH 4 to obtain metallic silver (between 1.65 and 2.16 wt% Ag). These materials were tested for the adsorption of iodine (I 2 ) at 35 C in the gas phase. The iodine adsorption capacities of the non-impregnated materials were on average between 90 and 130 g I 2 kg adsorbent À1 . The difference in performance between the non-impregnated materials was associated with the discrepancy in their pore sizes and specic surface areas. 84 In the presence of silver, the performance was signicantly improved as the adsorption capacity reached 760-770 g I 2 kg adsorbent À1 . The irreversibility of the capture of iodine at 120 C and 527 C was also studied by TGA, and no weight losses were indicated. In summary, these forms of mesoporous silica possess interesting and important iodine adsorption capacities, with the addition of a high degree of irreversibility when silver is added. More recently, Chebbi 81 studied, in parallel to his academic works on silver-exchanged zeolites, silver-doped mesoporous silica (SBA-15) for use in the eld of iodine adsorption. Tests on the retention of CH 3 I ([CH 3 I] ¼ 450 ppm, T ¼ 10 C) in the liquid phase were carried out on SBA-15 mesoporous silica doped with 0, 10, 20, and 40 wt% silver. Silver is assumed to be in the form of Ag 0 nanoparticles in the mesopores and on the external surfaces with an average crystallite size of 20 nm. In the case of the non-impregnated parent compound SBA-15, an adsorption capacity of 49 g CH 3 I kg adsorbent À1 was achieved. For the latter material, the capture of CH 3 I is controlled by physisorption-type interactions and is consequently reversible. The presence of Ag 0 nanoparticles at different loadings inside the SBA-15 material signicantly improves its iodine adsorption performance. Chebbi also highlights the quasi-linear relationship between the silver loading and the maximum amount of CH 3 I that can be adsorbed in the SBA-15. However, only a fraction of the silver is used to trap iodine. In fact, if the contribution of the physisorbed fraction of CH 3 I in the parent SBA-15 is considered (49 g CH 3 I kg adsorbent À1 for 0 wt% silver), the I/Ag ratio in the silverimpregnated SBA-15 is particularly low (0.21 and 0.13 for a silver loading of 10 and 40 wt%, respectively). These results were compared with those for a silver-exchanged AgY zeolite in another work (silver loading of 23 wt%, Si/Al ratio of 40), in which the I/Ag ratio was 0.30 for a crystallite size of 9.1 nm. According to the authors, these discrepancies between the performance of silver-doped mesoporous silica and silverexchanged zeolites mainly arise from the high dispersion of silver inside the structure. In addition, the absence of exchange sites in mesoporous silica and the large pore diameters facilitate the migration of silver to the surface, which leads to the aggregation of silver in the form of metallic nanoparticles, which ultimately block access to the pores. In the case of nanoparticles, only a small part of the silver can be used for the capture of iodine, which is probably due to the unfavourable core/shell ratio. In a similar way, the adsorption of I 2 by the same silver-loaded SBA-15 ([I 2 ] ¼ 400 ppm, T ¼ 25 C) was studied. 81 In another study by Chebbi,81 in the case of the parent compound SBA-15 no adsorption of iodine was observed. Once more, the presence of silver signicantly improved the iodine adsorption performance. It should be noted that the efficiency of the capture of I 2 per silver atom is higher than in the case of CH 3 I. A decrease in the efficiency of silver from I/Ag ¼ 0.90 to I/ Ag ¼ 0.59 was also observed when the silver loading was increased from 10 to 40 wt%. In a similar way to zeolites with a high loading of silver, the less efficient utilization of silver can essentially be attributed to accessibility problems that prevented iodine from reaching silver inside pores.

2.4.3.
Other porous oxide materials. SBA-15 that was functionalized with aminopropyltrimethoxysilane (APTMS) was also used for the retention of iodine (I 2 ) in the same standard conditions ([I 2 ] ¼ 400 ppm, T ¼ 25 C) as in previous studies by Chebbi. 81 An iodine adsorption capacity of 179 g I 2 kg adsorbent À1 was achieved with this material. The affinity between I 2 and the APTMS-functionalized silica was explained by the formation of a charge transfer complex. Then, this material was studied in conditions ([I 2 ] > 600 ppm, T ¼ 25 C) such that the saturation capacity for iodine (600 g I 2 kg adsorbent À1 ) of the functionalized SBA-15 was reached. However, it should be noted that the functionalization with APTMS can deteriorate upon an increase in the temperature, and therefore it would not be efficient to use it for the adsorption of iodine in a nuclear power plant in the event of a severe nuclear accident. In a study by Yang et al., 86 the remarkable efficiency for iodine capture of bismuth-doped mesoporous silica (SBA-15) was proved. Bismuth was incorporated into the SBA-15 material by previously modifying the silica surface with thiol groups and subsequent thermal treatment that led to the formation of a BI 2 S 3 phase. The bismuth-doped mesoporous silica displayed a maximum iodine adsorption capacity of around 540 g I 2 kg adsorbent À1 (T ¼ 200 C, 6 h under static air). The good adsorption performance of the material was attributed by the authors to the strong interactions between bismuth sulphide and iodine, the elevated specic surface area and the porosity of SBA-15. Bismuth, when combined with iodine, can form particularly thermodynamically stable compounds such as BiOI and BiI 3 (D f G ¼ À219.5 and À139.7 kJ mol À1 , respectively). The large pores in SBA-15 (ranging from 50 to 300Å, with a uniform distribution), as well as the strong affinity of bismuth for iodine, minimized the physical adsorption of iodine, which allowed chemisorption-type interactions to be the main capture process. Finally, the authors highlighted the low cost of bismuth and the easy preparation of SBA-15 as denite advantages for industrial applications. Another study led by Yang et al. 193 on bismuthbased adsorbents (bismuth oxide) demonstrated an iodine trapping capacity that was approximately 1.9 times higher (0.468 g I 2 kg adsorbent À1 ) than that of commercial AgX zeolite.
Other porous oxide materials, such as metallic oxides (MgO, ZnO, La 2 O 3 , ZrO 2 , etc.) have been studied for trapping iodine in the gas phase. [194][195][196][197] Among these, Glinski et al. 194 reported that La 2 O 3 and MgO had high iodine adsorption capacities of 31 and 48 g I 2 kg adsorbent À1 , respectively, in comparison with other metal oxides at T ¼ 100 C under an anhydrous stream of N 2 saturated with I 2 . More recently, Nandanwar et al. 195 achieved a maximum iodine adsorption capacity of 196 g I 2 kg adsorbent À1 at room temperature with porous microspheres of magnesium oxide (the size of the microspheres was 5-7 mm).
Besides, no studies have focused on the utilization of all these mesoporous materials in conditions closer to those of real utilization (temperature, presence of inhibitors such as NO x and H 2 O).

Silver-functionalized silica aerogels
Several authors have studied aerogels as potential adsorbents for iodine trapping. Aerogels are porous materials that possess an elevated specic surface area (>1000 m 2 g À1 ). Matyáš et al. [100][101][102] developed silver-functionalized aerogels for trapping iodine compounds. These materials exhibited promising results, with an iodine adsorption capacity of 310 g I 2 kg adsorbent À1 and a decontamination factor (DF) of higher than 10 5 . More recently, Riley et al. 105 studied silver-functionalized aluminosilicate aerogels (Na-Al-Si-O) for the capture of iodine. An iodine adsorption capacity of greater than 500 mg I 2 g adsorbent À1 was observed. For comparison, the adsorption capacity of silver-exchanged mordenite (Ag 0 MOR) tested in the same conditions only reached 190 mg I 2 g adsorbent À1 (Fig. 9).
These studies proved that silver-functionalized aerogels can be potential candidates for iodine capture. However, few articles on the subject can be found in the literature and no test was performed in conditions similar to those of real utilization.

Chalcogen-based aerogels (chalcogels)
Aerogels based on chalcogens (S, Se, Te), which are called "chalcogels", have been the focal point of a few studies on iodine trapping. 8 ) interconnected by a secondary metallic species (such as Pt 2+ , Co 2+ , Sn 2+ , Sb 3+ , Bi 3+ , Ni 2+ , and Zn 2+ ). Chalcogens enclosed in chalcogels are classied as weak Lewis bases according to the HSAB concept (hard and so acids and bases). Consequently, they have a high affinity for iodine (I 2 ), which is known to be a weak Lewis acid. This strong affinity for iodine was demonstrated for a large range of chalcogels, including PtGe 2 S 5 , Sn 2 S 3 , CoMoS 4 , NiMoS 4 , CoS 5 , Sb 4 (SnS 4 ) 3 and ZnSn 2 S 6 . 7,104,107,108 Riley et al. 107 developed SnS (Sn 2 S 3 ) structured chalcogels for trapping iodine (I 2 ) with an iodine uptake of up to 32.7 and 68.3 wt% for the best chalcogel formulations. Furthermore, the adsorption mechanism was studied. During adsorption, iodine interacts with SnStype chalcogels by chemical reactions (chemisorption) to form crystalline SnI 4 and SnI 4 (S 8 ) 2 species, which was observed by Xray diffraction analysis. This reaction is particularly favourable, with a Gibbs formation enthalpy of D f G ¼ À215.1 kJ mol À1 at 298.15 K (25 C). Subrahmanyam et al. 7 carried out studies of the adsorption of iodine onto NiMoS 4 , CoMoS 4 , Sb 4 Sn 3 S 12 , Zn 2 Sn 2 S 6 and K 0.16 CoS x (x ¼ 4-5) chalcogel-type adsorbents. A mass uptake of up to 225 wt% was achieved for the Zn 2 Sn 2 S 6 and NiMoS 4 chalcogels. As already mentioned, this high iodine adsorption capacity is due to the chemical reaction between iodine and the adsorbent (by chemisorption) and also the large specic surface areas that range from 200 to 490 m 2 g À1 . Iodine reacts chemically with Zn 2 Sn 2 S 6 , Sb 4 Sn 3 S 12 and K 0.16 CoS x (x ¼ 4À5) chalcogels to form metallic halides such as SnI 4 , SbI 3 and KI, whereas only physisorption interactions are observed for NiMoS 4 and CoMoS 4 chalcogels. TGA analysis 7 showed that Zn 2 Sn 2 S 6 and Sb 4 Sn 3 S 12 chalcogels with stored iodine were stable up to 150 C, but beyond that temperature iodine was released progressively (40 wt% in 30 days). Moreover, the other chalcogels release iodine above 75 C. The weight loss is attributed to the sublimation of iodine molecules and SnI 4 , as well as the release of physisorbed iodine from the surface. In conclusion, this type of material possesses excellent iodine adsorption capacities. However, the thermal stability of iodine capture (or the formed iodides) is quite low.

Macroreticular resins
Synthetic macroreticular resins are macromolecular structures formed by polymerisation reactions (from vinylbenzene monomers, for example). Functionalized chemical groups, such as acrylic esters and polystyrene, can be graed onto the structural framework of the resins. A resin can be dened by the following properties: its crosslinking rate (percentage of crosslinking monomer in the resin), the porosity of the framework (presence of variable pore and/or channel sizes), its granulometric composition (mean size of resin beads), its selectivity and its ion exchange capacity (quantity of ions that can be xed for a given mass or volume of the resin). In general, macroreticular resins possess elevated specic surface areas (up to 900 m 2 g À1 ) and can exhibit cation or anion exchange properties. Resins are known to have high chemical stability and can be restored to their initial state by washing with the appropriate solutions. Furthermore, their crosslinked structure does not limit the Fig. 9 Comparison of the iodine adsorption capacities of different silver-functionalized aerogels (SH-Ag 0 , Ag 0 , SH-Ag + , Ag + , and SH) with that of a silver-loaded mordenite (AgZ). 105  diffusion of molecules or substances and therefore helps to increase their storage and/or trapping capacities. The possibility of employing these kinds of adsorbents for trapping iodine has been the focus of several studies. 16,39,41,42,[44][45][46]50,53,55 However, very few studies have focused on the mechanism of the trapping of iodine (I 2 ) and organic iodine compounds by these adsorbents. In general, macroreticular resins have a strong affinity for any kind of substance thanks to their tunable hydrophobic and hydrophilic properties. Because most of these adsorbents also have non-ionic functions, they can have a strong affinity for uncharged molecules such as iodine (I 2 ) or organic iodine compounds such as CH 3 I. Most studies focused on the trapping of iodine in aqueous solutions. Few studies have focused on the behaviour of these adsorbents in the gaseous phase for iodine adsorption applications because of their limited stability under air. The objective was to develop an iodine ltration medium that was both cheaper than silverdoped adsorbents (e.g., silver-exchanged zeolites) and more stable than activated carbons. 16 In 1968, Hirling 39 studied a cation exchange resin (Varion KS) for iodine trapping applications. Tests showed a strong affinity for iodine with adsorption capacities of between 872 and 1437 g I 2 kg adsorbent À1 .
However, because of the saturation of the adsorbent bed, an efficiency of only 40% was achieved. This resin was also impregnated with silver to capture HI, which endowed the resin with a trapping capacity of 93.1% and an HI adsorption capacity of 1437 g HI kg adsorbent À1 . Resins of another type, namely, Amberlite XAD, were also tested for trapping iodine compounds. These resins are macroporous and can have rigid and three-dimensional structures, depending on the monomer used. These adsorbents can incorporate a large quantity of extractant thanks to their elevated specic surface areas, high mechanical resistance, superior crosslinking and low swelling during impregnation. Amberlite XAD resins have specic surface areas of between 150 and 900 m 2 g À1 with a mean pore diameter of between 4 and 9 nm and a pore volume of 0.6 to 1.1 cm 3 g À1 . The rst studies were carried out by Moore et al. 41,42,44,46 At 21 C in dry conditions, decontamination factors (DFs) of higher than 10 4 were achieved for an iodine (I 2 ) adsorption capacity of 213 mg I 2 g adsorbent À1 . With Amberlite XAD-12 resin, a maximum iodine adsorption capacity of 1.39 g I 2 g adsorbent À1 at 25 C was achieved, but only 84 mg I 2 g adsorbent À1 could be adsorbed at 50 C. In the same way, maximum iodine adsorption capacities of 278 mg I 2 g adsorbent À1 and 15 mg I 2 g adsorbent À1 at 21 C and 25 C, respectively, were achieved with Amberlite XAD-4 resin. However, it is quite surprising to observe such a great discrepancy for a temperature difference of 4 C. At higher temperatures and in the presence of humidity a decrease in the decontamination factor was observed. In conclusion, macroreticular resins possess elevated iodine adsorption capacities (between 200 and 1437 mg g adsorbent

À1
). 41,42,44,46 Furthermore, these materials are fairly stable in the presence of irradiation. 16,50,55 On the other hand, signicant decreases in their adsorption capacities were observed at high temperatures (>50 C) or in the presence of humidity. These drawbacks would drastically limit their utility in the real conditions of a severe nuclear accident.

Metal-organic frameworks (MOFs)
In the past een years, the number of studies on metalorganic frameworks (MOFs) has grown sharply thanks to their very large specic surface areas (up to $10 4 m 2 g À1 ) and well-dened pore sizes. 8 MOFs are organic-inorganic hybrid materials that consist of inorganic parts comprising metal ions, which are referred to as secondary building units, linked by organic entities, which are most commonly referred to as ligands. The secondary building units can be composed of single metallic cations, dimers, trimers, chains, planes and/or a three-dimensional structure. The organic ligands need to possess a charge and/or unbound electron pairs, which ensure strong bonds with the inorganic components of the framework. The most commonly used ligands are organic compounds that contain carboxylic acid or imidazole (C 3 H 4 N 2 ) groups. The best advantages of these types of materials are their large assortment of organic and inorganic entities that can generate a wide range of MOFs with various structures, topologies and pore sizes. In 2003, Abrahams et al. 121 demonstrated the feasibility of introducing iodine into a hydrated zinc saccharate ([Zn(C 6 H 8 O 8 )]$ 2H 2 O). Since then, MOFs have been particularly popular for the retention of iodine. 6,14,109,110,115,116,[118][119][120][121][122][123][124][125][126][127][129][130][131][132] Most of the studies were carried out on materials based on divalent metals such as Zn 2+ and Cu 2+ , as well as aluminium-based structures. In fact, the most commonly studied MOFs for the capture of iodine were the zeolitic imidazolate framework Zn(2methylimidazolate) 2 (ZIF-8) and HKUST-1 (Cu 3 (benzene-1,3,5tricarboxylate) 2 (H 2 O) 3 , Cu-BTC). 14,115,124,125,129,132 These two MOFs have elevated specic surface areas of 1875 and 1798 m 2 g À1 , respectively, with pore diameters of 11.6 and 9Å, respectively. 127 The most prominent studies on this subject were carried out by Sava et al. 14,124,125,129,130 In fact, Sava et al. 14 performed experiments on the iodine capture capacity of ZIF-8 at atmospheric pressure with a vapour pressure of iodine (I 2 ) of 0.014 atm at 350 K. Once equilibrium had been reached (5-12 hours), the iodine adsorption capacity of ZIF-8 was 1.25 g I 2 g adsorbent À1 (1250 g I 2 kg adsorbent À1 ) for an I/Zn ratio of 2.2. According to the results from Sava et al., the capture of iodine is due to the favourable interactions between iodine and the ligand (2-methylimidazolate, MeIM). Only 25% of the iodine is localized at the surface of the MOF, whereas 75% of the iodine is conned in the sodalite cages of ZIF-8. Studies by Hughes et al. 129 give complementary information on the thermal stability of, and chemisorption of iodine by, ZIF-8. They showed that iodine interacts strongly with hydrogen atoms in the methyl group on the one hand and the carbon atom in the ligand methine (] CH-) group on the other hand. Furthermore, they observed that the complex formed inside the sodalite cages is thermally stable at temperatures of up to 300 C, which is the limit of the thermal stability of the ZIF-8 material. 129 However, iodine molecules that are adsorbed on the external surface interact weakly with the material (DH ads ¼ À18.06 AE 2.03 kJ mol À1 ), which ultimately permits the release of iodine at temperatures below 125 C. On the other hand, iodine that is conned in the cages is thermally stable with an adsorption energy of DH ads ¼ À41.47 AE 0.62 kJ mol À1 .
Another study by Sava et al. 130 focused on the iodine adsorption capacity of Cu-BTC (HKUST-1). The study of this material was carried out at atmospheric pressure with a relative humidity of 3.5% and at 77 C. Iodine was introduced in the vapour phase with a ratio to water of 1 : 1 (vapour pressures of I 2 and water of 10.47 and 10.11 mmHg, respectively). An adsorption capacity of 1.75 g I 2 g adsorbent À1 (1750 g I 2 kg adsorbent À1 ) for an I/Cu ratio of 3 was found. Via this study, the authors demonstrated that the adsorption of iodine occurs in two steps. In the rst step, iodine is adsorbed in the small triangular cages of the MOF (5Å), and then interactions take place in the larger cages (11 and 13.55Å) thanks to the strong interactions between iodine and the benzene rings of the tricarboxylate ligand (van der Waals interactions). Finally, in the larger cages, I 2 -I 2 intermolecular interactions could also be observed. Furthermore, the favourable adsorption of iodine in preference to water (0.15 g H 2 O g adsorbent À1 ) was explained by the formation of a hydrophobic barrier due to the presence of iodine (iodine limits the access of water to the adsorption sites), which makes this material very interesting for real applications, including severe nuclear accidents. However, high temperatures would still be harmful to this kind of material. Aluminium-based MOFs have also been studied for iodine trapping applications. 115,116,120 Studies were carried out on the MIL-53 family of MOFs. These materials are particularly attractive owing to their high chemical and thermal stability. Furthermore, these materials are commercially available. They rst studied the MIL-53-X family to examine the effect of functionalization on the iodine adsorption capacity in the liquid phase. Among the above functional groups (H, Cl, Br, CH 3 , NH 2 , NO 2 , (OH) 2 , COOH, and (COOH) 2 ), only MOFs that contained electron-donating groups (NH 2 and OH) exhibited interesting iodine capture properties. In the case of nonfunctionalized MIL-53, only 5% of iodine was adsorbed from a solution in cyclohexane aer 48 hours. In contrast, the best performance was achieved with MIL-53-NH 2 , which exhibited a maximum iodine adsorption of 60% aer 48 hours. For the other MOFs that were studied, only MIL-101-NH 2 displayed excellent iodine adsorption capacities, with an efficiency of 90% aer only 30 hours. Other MOFs such as CAU-1 and MIL-120 also exhibited fairly good performance ($80% aer 30 hours). In the same context, Assaad et al. 120 conrmed that MIL-101 could exhibit an iodine adsorption efficiency of higher than 90% (96.61%).
Some computer simulation studies on the adsorption of iodine on MOFs have been performed. Assfour et al. 119 studied the performance of twelve MOFs for the retention of iodine by molecular modeling. Their simulations obtained denite information and an understanding of the inuence of the pore volumes and specic surface areas on the iodine storage capacity. They showed that MOFs with elevated pore volumes and specic surface areas were most suitable for the capture of iodine in conditions of normal temperature and pressure. However, at lower pressures MOFs with the smallest pore volumes were more suitable for the capture of iodine. Nevertheless, several materials exhibited important iodine adsorption capacities in conditions of normal temperature and pressure (up to $13 g I 2 g adsorbent À1 for NU-110). The modelling studies by Yuan et al. 125 helped to identify ZIF-10 as the MOF with the most important iodine adsorption capacity (2.39 g I 2 g adsorbent À1 at 25 C and in conditions of moderate pressure).
In conclusion, the studies in the literature mainly focused on the adsorption capacities of MOFs for iodine (I 2 ), whereas no studies have focused on organic iodine compounds such as CH 3 I. MOFs with high iodine capacities are a promising alternative for the capture of iodine, principally thanks to their high adsorption capacities. Furthermore, in the case of Cu-BTC, iodine is preferentially adsorbed rather than water, which is a major advantage in the real conditions of a severe nuclear incident. However, these materials have limited thermal stability and are not produced on a large scale, which makes them particularly expensive.

Covalently linked porous organic polymers (POPs)
In parallel to the increasing development of MOFs, porous organic polymers (POPs) have recently attracted attention owing to their highly tunable molecular design, large surface areas, low skeleton densities (lightweight elements), strong covalent linkages, high physicochemical stabilities and tunable porosities. [133][134][135][136][137][138][139][140][141][142][143] Although POPs are usually amorphous solids, unlike crystalline MOFs, their excellent physicochemical stability makes them more suited for real applications. In addition, the synthetic diversity of POPs makes it possible to attain control over their functionality by the rational design and choice of molecular building blocks. On the basis of their physicochemical robustness and pore features, POPs have huge potential to be efficient adsorbents for the capture of volatile radioactive iodine. It is widely accepted that the introduction of active sites such as metallic species, N, S, and C]C into the porous network of POPs can greatly improve their affinity for iodine molecules, which ultimately signicantly increases the uptake of iodine. In general, the uptake of iodine depends on both the pore features (pore size and/or pore volume) and the affinity of the adsorbent for iodine molecules. To date, a certain number of POPs have been employed for this purpose, such as nanoporous organic polymers (NOPs), 134 conjugated microporous polymers (CMPs) 139,141,142 such as metalloporphyrin-based CMPs 138 and hexaphenylbenzene-based CMPs, 135 azo-bridged porous triptycene networks 140,143 and crystalline covalent organic frameworks (COFs). [111][112][113][114]117,128 Chen et al. 134 studied the iodine adsorption performance of a series of hierarchically porous organic polymers based on tetraphenyladamantane (named as NOP-53, NOP-54 and NOP-55). Their pore properties were controlled by adjusting the lengths and rigidities of the linkers; for example, rigid tetrahedral building blocks of 1,3,5,7-tetraphenyladamantane were linked with exible alkyl chains to obtain different hierarchical NOPs. NOP-53 and NOP-54 feature hierarchically porous structures, whereas NOP-55 was found to be only microporous with a broad micropore size distribution. NOP-54 exhibits the largest BET specic surface area and pore volume (1178 m 2 g À1 and 1.32 cm 3 g À1 , respectively), followed by NOP-53 (744 m 2 g À1 and 0.73 cm 3 g À1 , respectively) and NOP-55 (526 m 2 g À1 and 0.42 cm 3 g À1 , respectively). Chen et al. 134 determined their uptakes of iodine (I 2 ) by gravimetric measurements. Samples were loaded into a sealed container in the presence of iodine pellets. The container was degassed and kept at 75 C. According to Chen et al., 134 the iodine uptake increased gradually over a period of 4 hours until the system reached saturation. NOP-54 displayed the highest uptake of iodine of up to 2.02 g I 2 g adsorbent À1 , followed by NOP-53 (1.77 g I 2 g adsorbent À1 ) and NOP-55 (1.39 g I 2 g adsorbent À1 ). According to the authors, NOP-53 and NOP-54 adsorb iodine at a faster rate than NOP-55. They attributed these observations to the existence of the hierarchical porous structures, which facilitated the transport of iodine in the networks.
Li et al. 143 developed in 2016 a novel porous azo-bridged porphyrin-phthalocyanine network, which was synthesized by combining an azo skeleton with p-conjugated building blocks. AzoPPN was synthesized by a catalyst-free coupling reaction between the free base form of 5,10,15,20-tetrakis(4-nitrophenyl) porphyrin (H 2 TPP(NO 2 ) 4 ) and nickel tetraaminophthalocyanine (NiPc(NH 2 ) 4 ) under alkaline conditions. This adsorbent has a BET specic surface area of 400 m 2 g À1 . According to the authors, the porous structure, together with the porphyrin and phthalocyanine units, provides effective sorption sites that can greatly increase the affinity for iodine both physically and chemically. As a result, AzoPPN exhibits an iodine adsorption capacity of up to 2.9 g I 2 g adsorbent À1 . Similarly, Dang et al. 140 studied an azo-linked porous organic network (Azo-Trip), in which triptycene was incorporated as a building block via a facile Zn-induced reductive homocoupling reaction. This adsorbent also has a large BET surface area (510 m 2 g À1 ) and an iodine uptake of up to 2.38 g I 2 g adsorbent À1 .
Conjugated microporous polymers (CMPs) are a class of amorphous materials that permit the linking of building blocks in a p-conjugated fashion and possess three-dimensional (3D) networks. They are usually synthesized using metal-catalyzed cross-coupling chemistry to form 3D networks with extended p-conjugation. Their nely tuned porosity, very large specic surface area, and relatively high thermal and chemical stability, which originates from their rigid p-conjugated structure, and the high affinity of I 2 for p-conjugated CMP networks make them attractive candidates for the adsorption and capture of radioactive iodine molecules.
Sigen et al. 138 developed a new conjugated microporous polymer based on a metalloporphyrin (NiP-CMP), which was synthesized via a homocoupling polymerization reaction. NiP-CMP possesses a large BET surface area of greater than 2600 m 2 g À1 and a large pore volume of 2.3 cm 3 g À1 . To determine its iodine adsorption performance, the NiP-CMP material was placed in a sealed vessel in the presence of solid iodine. The iodine sublimed into the porous adsorbent over time at 77 C and ambient pressure, which are typical conditions in fuel reprocessing. An uptake of iodine (I 2 ) of 2.02 g I 2 g adsorbent À1 from iodine vapour was determined by gravimetric measurements. Similarly, Chen et al. 139 studied conjugated microporous polymer nanotubes (CMPNs) for the capture of iodine. The material exhibited a maximum uptake of 2.08 g I 2 g adsorbent À1 in the adsorption of I 2 . Sigen et al. 138 also studied the reversibility of adsorption and regeneration of the sorbent. In fact, the authors easily removed iodine from the framework by immersing the I 2 @NiP-CMP material in an organic solvent (96% of iodine was desorbed aer 8 hours in ethanol). The authors highlighted the fact that the absorbed iodine can be recovered and that the NiP-CMP sorbent can be easily recycled and reused. The reversibility of the material was also conrmed for the CMPN material studied by Chen et al. 139 Ren et al. 141 studied the iodine adsorption performance of conjugated microporous polymers (SCMPs). Two SCMP networks (SCMP-I and SCMP-II), which are based on the monomer 3,3 0 ,5,5 0 -tetrabromo-2,2 0 -bithiophene and were obtained by a palladium-catalyzed Sonogashira-Hagihara crosscoupling reaction, have interesting honeycomb-like porous 3D network structures. SCMP-I is composed of agglomerated spheres with different sizes, whereas SCMP-II has an intertwining porous structure. Furthermore, the specic surface areas were found to be 2.72 m 2 g À1 for SCMP-I and 119.76 m 2 g À1 for SCMP-II. The uptakes of iodine (I 2 ) were determined by gravimetric measurements, and the capture of iodine vapour was conducted at 80 C and ambient pressure. The authors suggest that the high uptake of iodine by SCMP-II of up to 3.45 g I 2 g adsorbent À1 may be attributed to its unique macroscopically honeycomb-like porous features, as well as the p-conjugated network structure, which has been conrmed to have a relatively strong affinity for iodine molecules. Liao et al. 135 studied a series of conjugated microporous polymers based on hexaphenylbenzene (HCMPs) with secondary amine functional groups. The HCMPs had a moderate microporous BET surface area of up to 430 m 2 g À1 and a narrow pore size distribution with a uniform ultramicropore size of less than 1 nm. These materials exhibit excellent iodine adsorption capacities with an uptake of iodine of up to 3.16 g I 2 g adsorbent À1 , with the possibility of a further increase in the uptake to 3.36 g I 2 g adsorbent À1 when the polymers are reduced with anhydrous hydrazine. The authors studied the desorption rate of the iodine-loaded HCMPs by placing the materials in organic solvents such as ethanol at room temperature. Furthermore, Liao et al. showed that the release of iodine can be induced by heating the iodine-loaded material at 120-200 C. An iodine-loaded HCMP was heated at 125 C in air for 30 minutes. They achieved an iodine release efficiency of 98.8%. In addition, recycling of the HCMP material was studied, and the iodine uptake capacity was found to be 2.95 g I 2 g adsorbent À1 and 2.88 g I 2 g adsorbent À1 upon completion of the rst and second cycles, respectively, which represented a retention of 93.3% and 91.3%, respectively, of the initial capacity. The authors emphasized the fact that these materials are attractive as robust, recyclable and reversible adsorbents for iodine uptake. Zhu et al. 142 developed two novel conjugated porous materials based on BODIPY for studies of the adsorption of iodine (I 2 ). BDP-CPP-1 and BDP-CPP-2 were synthesized via a Sonogashira cross-coupling reaction between 1,3,5-triethynylbenzene (TEB) and dibromo-substituted derivatives. Both materials exhibit high iodine adsorption capacities of 2.83 g I 2 g adsorbent À1 and 2.23 g I 2 g adsorbent À1 , respectively. Owing to the highly pconjugated porous structure (coexistence of triple bonds, phenyl rings and aromatic pyrrole moieties) of BODIPY, BDP-CPP-1 displays a high capacity (2.83 g I 2 g adsorbent À1 ) for volatile iodine at 75 C and a high iodine adsorption rate in an organic solution at 25 C (90% of iodine in a solution in hexane was captured aer 7 hours). According to the authors, these high capacities can be attributed to the large BET surface area (635 m 2 g À1 ) and pore volume (0.78 cm 3 g À1 ) and the chemical substitution reaction at the 2-and 6-positions of the BODIPY core (i.e., adsorption of volatile iodine via a chemical mechanism involving the hydrogen atoms at the 2-and 6-positions). Furthermore, the BDP-CPPs display high thermal stability, with a decomposition temperature of about 300 C. In addition, these CMPs exhibit excellent recyclability aer 4 cycles, which, according to the authors, may result from the p-conjugated porous structure. Finally, covalent-organic frameworks (COFs) are a class of material in which organic ligands are linked together to form a periodic structure via strong covalent bonds. COFs have a well-ordered architecture with particularly interesting structural properties, such as a low structural density, easily tuned porosity, elevated specic surface areas, and quite high thermal (up to 600 C in the best case) and chemical stability, as well as a wide range of functionalization. 117,128,198 Lan et al. 128 used computational studies to determine the theoretical adsorption capacities of COFs for iodine (I 2 ) and iodomethane (CH 3 I) in the conditions of real applications. The results show that 3D-COFs have the best adsorption performance for I 2 and CH 3 I in contrast to 2D-COFs. A pyrene-based 3D-COF (3D-Py-COF) has been found to possess a high adsorption capacity for iodine compounds of 16.7 g I 2 g adsorbent À1 , which is an extremely high value in comparison with those for the other adsorbents described in this review. The morphologies of the pores play a crucial role in the adsorption of CH 3 I. Consequently, 3D-COFs with a ctn topology and a pore size of 9Å display the highest adsorption capacities in comparison with other COFs. COF-103 was identied as the best adsorbent for CH 3 I, with an adsorption capacity that reached 2.8 g I 2 g adsorbent À1 . COFs have not yet been the focus of many studies. [111][112][113][114]117,128 Furthermore, they are still new laboratory materials and are therefore produced in very small quantities at high prices.

Summary and conclusion
Taking into account the severe operating conditions in the case of a nuclear incident, sorbents for the capture of radioactive iodine (CH 3 I and I 2 ) should full the following requirements: a high adsorption capacity over a wide temperature range, irreversibility of capture, selectivity, high thermal stability of the adsorbed species, high stability against irradiation and strong resistance in the presence of molecules of various gases such as water and NO x ( Table 2). Various kinds of porous sorbents have been studied for the capture of iodine. Activated carbons doped with KI and TEDA and porous solids doped with silver (zeolites, porous silica and alumina, aerogels and porous titanosilicates) have been the most widely studied sorbents. Activated carbons proved to exhibit good performance in the adsorption of iodine, especially when loaded with KI and TEDA, with strong capture of radioactive iodine. However, their performance declined signicantly in the presence of humidity (>40%) and at high temperatures (T > 80 C), whereas such operating conditions are probable in the case of a severe nuclear accident. Furthermore, they are particularly affected by NO x and aging. Silver-doped materials, in particular, silver-exchanged zeolites, were meant to overcome the issues with activated carbons by being more resistant to harsh conditions such as high humidity, the presence of NO x and high temperatures. Most studies revealed that silverexchanged zeolites exhibit high performance and capacities for the capture of iodine with strong irreversible adsorption of iodine. The best results were obtained for zeolites with large pores (AgX and AgY) and a high silver content. On the other hand, silver-based sorbents proved to be particularly expensive in comparison with activated carbons and other sorbents. The doping of zeolites with less expensive metals was found to be much less effective in comparison with silver. However, bismuth-doped mesoporous silica displays high capacities for the adsorption of iodine and thermal stability of the captured iodine.
Interesting results have also been obtained for chalcogenbased aerogels, in particular, those based on tin sulphide, but the thermal stability of the captured iodine is relatively low and the inuence of humidity on the capture of iodine has not yet been studied. Other porous sorbents such as metal-organic frameworks (MOFs) and porous organic polymers (POPs) demonstrated great potential for the trapping of iodine. All these sorbents possess good iodine adsorption capacities, which reach 1750 g I 2 kg adsorbent À1 in the case of MOFs. However, although they have promising iodine adsorption properties, most of these sorbents were not fully characterized, and more precise studies on the inuence of various operating parameters (such as temperature, humidity, and the presence of NO x ) are still needed before an accurate judgement regarding the best sorbent for iodine removal applications. Furthermore, some sorbents, such as MOFs and POPs, are laboratory objects and thus not suitable for industrial applications at the moment. It can be concluded that these new sorbents should be studied more thoroughly from the point of view of real applications, which implies cost-effective solutions and tests in a wide range of operating conditions. Thus, the development of new sorbents that combine a high iodine adsorption capacity over a wide temperature range, high stability and selectivity in the capture of iodine at high temperatures and humidities and in the presence of other molecules, and low cost still remains a challenge. The use of several kinds of iodine sorbents to combine their advantages could be a promising way to full the requirements of real applications. Decline in adsorption performance in the presence of humidity (>40%) Low production cost Decline in adsorption performance at high temperatures (T > 80 C) High stability in basic and acidic condition Alteration of the adsorption performance due to aging Low autoignition temperature Strong inuence of NO x (formation of explosive compounds) Silver-exchanged zeolites Good adsorption performance (optimal adsorption temperature: 423 K)

Expensive
Trapping by precipitation (formation of AgI) + high capture stability Alteration of the adsorption performance due to organic compounds (especially chlorides) High irreversibility of the trapping of iodine (especially for AgY zeolite, >80%) Negative effects of NO x and humidity on performance No or little inuence of g-radiation Tunable chemical and structural properties High thermal and chemical stability Silver-doped silica and/or alumina (Ag/SiO 2 and Ag/Al 2 O 3 ) Good adsorption performance (DF > 10 2 ) up to 150 C Loss of efficiency at temperatures of <200 C + higher operating cost In the presence of NO x (1-10%), an increase in performance Loss of performance in the presence of a large excess of organic contaminants Less expensive by a factor of 3 to 10 than silverdoped zeolites (theoretically) Alteration of the adsorption performance with humidity (>70%) Mesoporous silica Adsorption performance similar to that of silverdoped zeolites Few studies on the inuence of humidity and/or inhibitors (NO x , organics, etc.) Elevated adsorption capacities (up to 0.6 g I 2 g adsorbent À1 ) when functionalized (e.g., with APTMS) Limited thermal stability when functionalized Aerogels/chalcogels Excellent adsorption performance (>0.5 g g adsorbent À1 ) Few studies on the inuence of humidity and/or inhibitors (NO x , organics, etc.) Little literature on these materials Expensive Titanosilicates Adsorption performance similar to that of silverdoped zeolites when doped with silver

Low adsorption capacities
Few studies on the inuence of humidity and/or inhibitors (NO x , organics, etc.) Macroreticular resins Excellent adsorption performance (up to 1 g g adsorbent

À1
) Signicant decline in adsorption performance at temperatures of higher than 50 C or in the presence of humidity Stable in presence of g-radiation Resistant in acidic conditions (NO x ) Metal-organic frameworks (MOFs) Excellent adsorption performance (up to 1.75 g I 2 g adsorbent

) theoretically and in ideal conditions
No studies on the inuence of humidity and/or inhibitors (NO x , organics, etc.) High selectivity toward I 2 rather than water No information on the adsorption and capture of CH 3