Increasing the Complexity in the MIL‐53 Structure: The Combination of the Mixed‐Metal and the Mixed‐Linker Concepts

Abstract The isoreticular mixed‐component concept is a promising approach to tailor the material properties of metal–organic frameworks. While isoreticular mixed‐metal or mixed‐linker materials are commonly synthesized, the combination of both concepts for the development of isoreticular materials featuring both two metals and two linkers is still rarely investigated. Herein, we present the development of mixed‐metal/mixed‐linker MIL‐53 materials that contain different metal combinations (Al/Sc, Al/V, Al/Cr, Al/Fe) and different linker ratios (terephthalate/2‐aminoterephthalate). The possibility of changing the metal combination and the linker ratio independently from each other enables a large variety of modifications. A thorough characterization (PXRD, ATR‐IR, TGA, 1H NMR, ICP‐OES) confirmed that all components were incorporated into the framework structure with a statistical distribution. Nitrogen physisorption measurements showed that the breathing behavior can be tailored by adjusting the linker ratio for all metal combinations. All materials were successfully used for post‐synthetic modification reactions with maleic anhydride.


Introduction
Multicomponent metal-organic framework (MOF) materials are gaining growing attention in this decade. [1] The toolbox-like design enables ahigh variability of components and, thus, various functionalities, often in combination with high porosity. Within this field, the synthesis of multicomponent metal-organic framework materials, in which several comparatively small and structurally simple building blocks are combined into one framework structure, is one of the most promising concepts. [2] The advantage of isoreticular mixed-component metal-organic framework materials, in whicht he metals or linkersa re distributed over crystallographically equivalent lattice positions (Figure 1a,b ), is that the given framework structure can be retained while different properties thereof can be altered. [2d, g, 3] This way,l inker molecules can be replaced by another type with similarg eometry,b ut different functionality;t his enables the development of al arge library of materials from easily availablea nd cheap starting materials. To date,t he majority of publications on isoreticularm ixed-component metal-organic frameworks used either the mixed-metalo rt he mixed-linker concept. [2d, g, 3c-f] Only af ew publications combined both concepts for the synthesis of isoreticular mixed-metal/mixed-linker metal-organic framework materials ( Figure 1c). [4] However, only if such ac ombination of these two concepts is possible, all benefits of the toolbox-like design can truly be exploited to tailor the properties of metal-organic frameworks. For as elected framework structure, the number of possible modifications exponentiates if the mixed-metal and the mixed-linker concepts are combined, even for as mall number of availablec omponents. Different one-step or multi-step synthesisp rocedures have been reported for the synthesis of isoreticular mixed-component materials. [3c, 5] One-step or direct methods are usuallyt he most convenient ones. All startingm aterials are mixed during the synthesis and the resulting product is purified afterwards. Multistep methods usually use either ap artial post-synthetic modification or post-synthetic exchange reactions for the synthesis of mixed-component materials. [3c] While the controlo ver the whole process is much better in am ulti-step method, the effort is often significantly higher due to severals ynthesis and purifications steps. Furthermore, the formation of core-shell ar- chitectures has often been observed forthese multi-step methods, while as tatistical distribution of linkersi sf requently found in one-step procedures. [3c] The family of MIL-53 materials is well-investigated and has been synthesized with av ariety of differentm etals and linker molecules. [6] In the MIL-53 structure ( Figure 1d), the trivalent metals have an octahedral coordination sphere with four oxygen atoms from linker molecules and two oxygen atoms from bridging m 2 -OH ions that are connecting neighboring metal centers. The resulting infinite M-OH chains are connected by terephthalate linker molecules resulting in one-dimensional diamond-shaped channels. The vanadium analogue was named MIL-47 and contains vanadium centers in the 4 + oxidation state and O 2À as the bridging moieties. Various mixedmetal [6,7] or mixed-linker [8] materials with MIL-53 structure have been reported, which were shown to have tunable properties depending on the ratio of the incorporated metals or linkers.
In the present contribution,t he MIL-53 topology has been used for the formation of previously unknown mixed-metal/ mixed-linker MIL-53 materials,i nw hich the respective components have been incorporated in as tatistically distributed fashion into the framework structure. Various metal combinations (Al/Sc, Al/V,A l/Cr,A l/Fe) have been successfully used in combination with am ixture of terephthalate and 2-aminoterephthalate as linker molecules. Furthermore, all materials were used for post-synthetic modificationso ft he amine groups with maleic anhydride.

Results and Discussion
The development of ad irect synthesis method for the herein presented mixed-metal/mixed-linker MIL-53 materials seemed to be desirable to achieve the targeted statistical distribution of linkersa nd metals with al ow effort. The previously developed direct synthesis procedures for MIL-53(Al 0.8 M 0.2 )-NH 2 materials (M = Sc, V, Cr or Fe) were used for the synthesis of mixed-metal/mixed-linker materials (Scheme1). [7a] For each of the metal combinations, differentratios of terephthalate/2-aminoterephthalate were used during the synthesis without any other changes to the procedures. The obtained MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials were labeled accordingt ot he percentage of 2-aminoterephthalate with respect to the total amount of linkers. The ratios of incorporated terephthalate and 2-aminoterephthalate linkersw ere determined by using liquid-phase 1 HNMR spectroscopy of digested samples. The found ratios were close to the expected values for all MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials (Table S1 in the Supporting information).
Recorded ATR-IR spectrao fall MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials ( Figure 2) showedt hat no or only smalla mounts of residual linker molecules (1690-1675 cm À1 )o rN,N-dimethylformamide (DMF,1 675-1665 cm À1 )w ere presentw ithin the pores of the MIL-53s tructure. However,f or MIL-53(Al 0.8 V 0.2 )-NH 2 (X) materials, this state was only achieved after additional suspension in DMF at 90 8Ca nd as ubsequent filtration and washing process (see the Experimental Section), since bands of residual linker molecules were observed in the IR spectra directly after the synthesis ( Figure S2). The presence of various terephthalate/2aminoterephthalate ratios was visible from the changing intensities of severalb ands. With increasing amount of the incorporated terephthalate linker,t he bands centered at 1387, 1492 and 1620 cm À1 decreased and the bands at 1410 and 1507 cm À1 increased in intensity.F urthermore, the intensity of the bands of amine vibrations (3497 and 3387 cm À1 )d ecreased with decreasing 2-aminoterephthalate content. Although quantitative statements based on the IR spectra are difficult, these observedt rends were in accordance with the determined linker ratios based on 1 HNMR measurements.
The band centered at 1125 cm À1 ,w hich could be assigned to the d(OH)v ibration of the m 2 -OH group interacting with water molecules within the pores, [9] was similarf or all materials within the series of Cr andF e. However,t he band intensity constantly decreased for V-and Sc-containing materials with an increasing amount of terephthalate in the framework. Furthermore, ab and close to 990 cm À1 ,w hich has previously been ascribed to the d(OH) vibration of the m 2 -OH group without any interaction of solvent molecules, [9] increased in intensity with increasinga mount of terephthalate for all materials. Based on the currently available set of data, ac onclusive explanation for this observation could not be provided.
The metal ratios of all MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials were determined using inductively coupled plasma optical emission spectroscopy (ICP-OES; Table S2). For iron and chromium, the ratios werec lose to the expected values, which were present during the synthesis. In the case of vanadium, slightly higher aluminumcontentsa nd, for scandium, slightly lower aluminum contentst han expected were found. Overall,t he different linker ratios did not have any influence on the incorporated metal ratios compared to the materials containing only 2-aminoterephthalate as linker.
Powder X-ray diffraction patterns ( Figure 3) showed that crystalline materials were obtained for all metal and linker combinations. The observed diffraction patterns for scandiumand iron-containing materials weres imilar to reported patterns of MIL-53(Al)-NH 2 in the narrow pore (np) form [10] and showed only few changes upon the variation of the linker ratios, which was similar to previously reported mixed-linkerM IL-53(Al)-NH 2 (X). [8a, 11] Thus, the MIL-53 structure was successfully obtained for all of these materials. For MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) materials, the reflectionso ft he MIL-53(Al)-NH 2 np-form dominated the diffractionp atterns for all linker ratios.H owever,r eflections of presumably al arge pore (lp) form (2q = 8.78,1 5.08) increased in intensity with increasing contento ft erephthalate in the framework. This observation indicated that an increasing proportion of the particles featured al p-state. In the case of  the vanadium-containing materials, reflectionsc orresponding to np-and lp-formsw ere also visible forthe mixed-linker materials, but in contrast to the chromium-containing materials, the changes were more pronounced. The np-form was dominating for MIL-53(Al 0.8 V 0.2 )-NH 2 (80), whereas the lp-form dominated for MIL-53(Al 0.8 V 0.2 )-NH 2 (60). The reflectionso ft he np-state were only hardly visible for MIL-53(Al 0.8 V 0.2 )-NH 2 (40) and the obtained reflection pattern was similart ot he one observed for MIL-47(V). [12] Therefore,t he lp-form seemed to be favoreda lready at comparatively low terephthalate contents. Thus,t he incorporation of terephthalica cid into the MIL-53-NH 2 structure had the strongesti nfluence for vanadium-containing materials, whereas only small changes were observed for chromium, scandiuma nd iron.
Pawley refinements of the measured powder X-ray diffraction patterns were performed to determine the unit cell parameters of the observed np-and lp-phases startingf rom reported values (Table S3). The incorporationo ft erephthalate resulted in only small changes of the unit cell parameters for the correspondingphases.
The thermal stability of monometallic mixed-linker MIL-53(Al)-NH 2 (X) materials was previously shown to increase with an increasing amount of terephthalate and ad ecreasing amount of 2-aminoterephthalate. [8a] Thist rend was also found for the MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials ( Figure 4). However, the differences of the stabilities depended on the metal combinations. Comparatively large shifts of the maximum decomposition rate to highert emperatures were observedf or the MIL-53(Al 0.8 Sc 0.2 )-NH 2 (X) (450-520 8C) and the MIL-53(Al 0.8 V 0.2 )-NH 2 (X) (400-470 8C) series, while mediums hifts were visible for the MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) (395-420 8C) an MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) (360-390 8C) series. The onset of the major decomposition step shiftedi ns imilar extents. Shoulders or additional small local minimaw ere observed for several of these materials in the derivatives of the TG curves. However,t hese features also shifted depending on the linker ratios. In total, the continuous shift of the thermals tabilityc orroborated the above claimed statistical incorporation of all components into the MIL-53 structure. All materials showedagood stabilityu pt o2 50 8Ca nd can, therefore, also be potentially used for gas phase applications at elevated temperatures.
X-ray absorption spectra ( Figure 5) were recorded for the MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials to investigate whether the incorporation of terephthalate had an influence on the local chemicale nvironment of the iron centers. The resultsc onfirmedt hat the oxidation state of the iron centers remained unchanged (+ III) for all linker ratios. Apart from that, only small changes in the k-and R-space were visible.
As reported earlier, [7a] we proposed that ab reathing transition occurred only at high relative pressures, which would explain the missing S-shaped adsorption isotherm that is usually relatedt oabreathing behavior andt he differenceo ft he adsorbedv olume betweent he adsorption and desorption branches.
The incorporation of different amounts of terephthalate in addition to 2-aminoterephthalate resulted in significantly larger adsorbed volumes in the adsorption and the desorption branches of the MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) and MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials.T he adsorbed volumes increased with an increasingc ontento ft erephthalate for both the chromium-and iron-containing materials. Furthermore, an S-shaped adsorption isotherm was visible for all of these mixed-linker materials.T he step steepeneda nd shifted towards lower relative pressures with increasing terephthalate contents. As imilarb ehavior has previously been found for MIL-53(Al)-NH 2 (X) materials. [13] However,t he position and the steepness of the step in the adsorp-
In situ high-resolutionp owderX -ray diffraction patterns have been recorded forM IL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials in nitrogen flow to investigate their temperature-dependent breathing behavior.T hese materials have been inserted as powders into ac apillary set-up and heatedt o3 10 8Cu sing a hot air blower.M IL-53(Al 0.8 Fe 0.2 )-NH 2 (100) was heated only to 250 8C, as it started to decompose at highert emperatures. MIL-53(Al 0.8 Fe 0.2 )-NH 2 (100) (Figure 7a)s howedo nly small changes upon heatingt o2 50 8C. With decreasinga mount of 2aminoterephthalate and increasing amount of terephthalate as linker molecule, the observed changes in the recorded powder diffraction pattern upon heating to 310 8Cw ere more pronounced (Figure 7b-d). While the position of the (200) reflection at 0.66 À1 did not change at elevated temperatures, the (110) reflection close to 0.90 À1 shifted to higher values until 130 8Ca nd subsequently to lower values at higher temperatures. These shifts were more pronouncedw ith increasing terephthalate content. Furthermore, the intensity of the small reflectiona t0 .61 À1 ,w hich was visible only as as houlder at room temperature for all MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials, increased with increasing temperature and increasing terephthalate content.
Additionally,areflection at 1.07 À1 was only observed at elevated temperatures for MIL-53(Al 0.8 Fe 0.2 )-NH 2 (60) and MIL-53(Al 0.8 Fe 0.2 )-NH 2 (40). The reflections at 0.61 and 1.07 À1 could be assigned to the (001) and (101) reflections of al p-form. The increasing intensity of the (001) reflection indicated that an increasingn umber of crystallites underwent an p !lp transition at elevated temperatures. Hence, the fraction of the lp-form increasedw ith increasing terephthalate content.H owever,n o complete phase transition was observed in the availablet emperaturer ange fora ny of these materials. Similart endencies have been observed for MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) materials (Fig-Figure 7. High-resolution powder X-ray diffraction patterns of MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials recorded in situ at different temperatures undern itrogen at ambient pressure (X = a) 100, b) 80, c) 60 or d) 40). Chem. Eur.J.2021, 27,1 724 -1735 www.chemeurj.org 2020 The Authors. Published by Wiley-VCH GmbH ure S3, lab powder X-ray diffractometer)a nd were also reported for monometallic MIL-53(Al)-NH 2 (X) materials. [11] In the case of the vanadium-containing MIL-53(Al 0.8 V 0.2 )-NH 2 (X) materials, the temperature-dependent flexibility has been investigated using al ab powder X-ray diffractometer. The results showed that the diffraction patternso fa ll materials changed only slightly at elevated temperature in comparison to the patterns at 30 8C ( Figure 8). The position of severalr eflections of the np-form shifteds imilarly to the already discussed MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials at elevated temperatures (e.g., Q = 0.88, 1.26, 1.28, 1.63, 1.76 À1 ). With increasing amount of terephthalate in the framework, theses hifts were slightly more pronounced, whichi ndicated as lightly higher flexibility of the framework structure.T he intensity of these reflections didn ot change significantly,w hich indicated that the fraction of the np-phase remained constanto ver the investigated temperature range. Thei ntensity of the characteristic reflections of the lp-form (Q = 0.61, 1.07 À1 )r emained also almostconstant and their positionsd id not showany shifts.
Only for MIL-53(Al 0.8 V 0.2 )-NH 2 (40), shifts of the small reflections at Q = 0.76, 1.18, and 1.54 À1 towards lower values were visible. Furthermore, these reflections were visible up to 250 8C, but not at 300 8C. Ap ossible explanation would be that these reflections originated from as mall proportion of an pphase, whicht ransformed into al p-form above 250 8C, but no unambiguous prove can be provided so far.I ns ummary,p resumably no significant np!lp transitions occurred for all MIL-53(Al 0.8 V 0.2 )-NH 2 (X) materials at temperatures up to 300 8C, which was particularly surprising for MIL-53(Al 0.8 V 0.2 )-NH 2 (60), since already al arge fraction of the lp-phase was presenta t room temperature and it was expected that the np-form would easily undergo at ransformation into the lp-from.
Uncoordinated amine-groups of metal-organic frameworks are commonly used for post-synthetic modification reactions. [14] For this purpose, maleic anhydride is ac ommons ubstrate as the modification proceeds by an addition reaction and no water is eliminated in the courseo ft he modification process. [13a, 14a, b, 15] This way,c helating side groups or Brønsted acid functionalities can be introduced within the pores. The chelating group is particularly interesting for the immobilization of metal ions to obtain single-site catalyst materials. [3a, 13a, 15b, 16] Therefore, all MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials were modified with maleic anhydride (Scheme 2) to show that the mixed-metal/mixed-linker concept is compatible with postsynthetic modifications and to investigate whether different combinationso rr atios of components have an influence on the modificationp rocess. In analogy to our previouss tudy, [7a] all materials were activated at 130 8Ci nv acuum before adding as olutiono fm aleic anhydride in acetonitrile to the dehydrated materials.
The metal ratios of all materials did not change during the post-synthetic modification process (Table S2). The number of functionalized amine linkers( modification degree) was determined based on the recorded 1 HNMR spectra of digested materials.T he recorded spectra (example shown in Figure S4) showeds ignals originating from protons of both terephthalate (d = 7.85 ppm (s, 4H,H 4 )) and unmodified 2-aminoterephthalate (d = 7.16 (d, 1H,H 1b ), 7.24 (s, 1H,H 1c ), 7.67 ppm (d, 1H, H 1a )) linkers. Furthermore, three signals of protons from the benzene ring of modified 2-aminoterephthalate linkers were visible (d = 7.59 (d, 1H,H 2b ), 7.85 ppm (d, 1H,H 2a ), 8.58 (s, 1H, H 2c )), of which the signalw ith the highest downfield shift overlapped with the strong singlet of terephthalate. The protons of the maleate side group of the modified linker were visible as two individual signals (d = 6.08 (d, 1H,H 2e ), 6.49 ppm (d, 1H, H 2d )). In addition, as inglet at d = 5.98 ppm (s, 2H,H 3 )w as ascribed to free maleate molecules. Because as trong acid (DCl) and as trong base (NaOD) were involved in the digestion process, these free maleate molecules were most likely formed during the digestion process by cleavage of the amide band of the post-synthetically modified linker.Ameasurement of the prepared solution after four weeks also supported this assumption, since the number of free maleate molecules relative to the number of modified linkersi ncreased ( Figure S6), thus, indicatinga no ngoing cleavage of the amide bond. As the recorded ATR-IR spectra after post-synthetic modification did not show any bands of free maleic acid or maleic anhydridew ithin the pores (characteristic bands expected close to 2900 and 1780 cm À1 ,r espectively), it is assumed that all free maleate molecules found in the NMR spectra originated from the cleavage of post-synthetically modified linker molecules.
The determined percentages of modifieda mine groups (Table S5)w ere in the range 16-86 %. Only for the scandiumcontaining materials, ac lear correlation of the modification degree and the linker ratio was found. The number of modified amine groups increased with an increasing fraction of terephthalate in the framework. The overall modification degree of mixed-metal/mixed-linker materials, which is the number of modified linker molecules with respect to the total number of linkers( terephthalate + 2-aminoterephthalate), was in the range 16-36 %.
The recorded powder X-ray diffraction patterns (Figure 9) confirmed that the MIL-53 structure remained intact during the post-synthetic modification process for all materials. The characteristic reflectionso fe ither np-or lp-forms were clearly visible.Incontrasttothe diffraction patterns beforethe modification (Figure 3), for which only few materials showed reflections of lp-phases, most materials were found to be in al pform after post-syntheticm odification (one characteristici ntense reflection close to 2q = 8.78). However,t he iron-containing materials were found predominantly in the np-form (two characteristic reflectionsc lose to 2q = 9.4 and1 2.28). The transition of the np-form to the lp-form during the post-synthetic modification process seemed to be ar esult of the space requirementoft he incorporated maleate side group.
The recorded diffractionp atterns were used to determine the unit cell parameters after the modification process by using Pawley refinements. The obtained pore dimensions confirmed the qualitativelyd erived assumption on the presence of np-and lp-forms ( Table S4). However,c onclusive refinements were not possible for all materials, since asymmetric and/or broadened reflectionsw erep resent for severalm aterials, which presumably resulted from non-uniform modification within the crystallites.
Nitrogenp hysisorptioni sotherms after post-synthetic modification showedd ramatically decreased nitrogen uptakes in comparison to the pristine MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials (cf. Figures 6a nd 10). Furthermore, no S-shaped adsorptioni sotherms were found, which indicated that none of these materials showed any flexibility under the applied measurement conditions anymore. The reduced nitrogen uptake can be explained by the incorporation of the maleate side groups, which require space in the pores and,t hus, reduce the accessible pore volume. As mall gap between the adsorption and desorptionb ranches reaching to the lowest measured relative pressure was visible for severalm aterials, but an unambiguous explanation cannot be provided so far.W hereas the specific surfacea reas (BET method) were in the range of 30-260 m 2 g À1 ,t he micropore volumes (t-plot method) were relatively small ( 0.05 cm 3 g À1 ,T able S5);t his suggested that the majority of the accessible surfacea rea resulted from interparticle void space.

Conclusions
This study has shown that mixed-metal/mixed-linker MIL-53(Al 0.8 M 0.2 )-NH 2 (X) materials were successfully synthesized by using ad irect synthesis procedure at ambient pressure. The thoroughc haracterization revealed that all components were statistically distributed within the MIL-53 framework structure with ratios close to the expectedv alues, which were applied during the synthesis. Based on the powder X-ray diffraction patterns,t he framework structuresc ontinuously changed with varying linker ratios.W hereas the np-form was dominant for the majority of the materials, reflections of al p-form appeared and increased in intensity with highert erephthalate content for the MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) materials. Furthermore, the lpform wast he dominant phase for MIL-53(Al 0.8 V 0.2 )-NH 2 (60) and exclusively presentf or MIL-53(Al 0.8 V 0.2 )-NH 2 (40). The thermal stabilitys trongly depended on the linker ratio for all metal combinationsa nd increased with increasing terephthalate content. Nitrogen physisorption measurements showedt hat the breathing behavior strongly depended on the linker ratio of terephthalate and 2-aminoterephthalate. An increased adsorption volume with increasing terephthalate content was ob- Remarkably,m ost of the adsorptioni sotherms of mixed-metal/mixed-linker materials showedastep that was characteristic fort he breathing behavior and was not observed for most single-linker MIL-53(Al 0.8 M 0.2 )-NH 2 (100) materials containing only 2-aminoterephthalate. The height, the positiona nd the steepness of this step depended on both the metal combination and the linker ratio. Te mperature-dependent powder X-ray diffractionm easurements showedt hat only small changes occurred in the framework structure between room temperature and 300 8Ca nd that no complete np!lp transition could be achieved. All developedmaterials were used for post-synthetic modification reactions with maleic anhydride. The framework structure remained intact during the modification process, but mostm aterials adopted al p-form after the modification. Due to the presence of the maleate side group, the nitrogen uptake decreased significantly after the modification. Thus, the properties of the MIL-53 structure can be carefullyt uned by selecting appropriate metal combinationsand linker ratios.
Synthesis of MIL-53(Al 0.8 V 0.2 )-NH 2 (X) materials:2 -Aminoterephthalic acid and terephthalic acid were dissolved in 10 mL of water and 10 mL of N,N-dimethylformamide (DMF) at 90 8Cw ith stirring. After complete dissolution, as olution containing aluminum(III) nitrate nonahydrate (0.7998 g, 2.13 mmol, 0.80 equiv.) and vanadium(III) chloride (0.0838 g, 0.53 mmol, 0.20 equiv.) in 5mLo fw ater was added under continuous stirring. The resulting reaction solution was stirred at 90 8Cu nder reflux cooling for 72 hours. Afterwards, the hot suspension was filtrated by using ag lass filter, washed (3 20 mL DMF,1 30 mL H 2 O) and dried in air atmosphere (overnight at room temperature and for another 3days at 130 8C). Ta ble 1c ontains the used amounts of 2-aminoterephthalic acid and terephthalic acid. To remove residual linker molecules from within the pores, the obtained materials were suspended in 20 mL of DMF at 90 8Cf or 4h,s ubsequently filtrated by using a glass filter,w ashed (2 20 mL DMF,1 20 mL H 2 O) and dried in air (at room temperature overnight and at 130 8Cf or another 3days).
Synthesis of MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials:2 -Aminoterephthalic acid and terephthalic acid were dissolved in 20 mL of water and 20 mL of N,N-dimethylformamide (DMF) at 90 8Cw ith stirring. After complete dissolution, as olution containing aluminum(III) nitrate nonahydrate (0.7998 g, 2.13 mmol, 0.80 equiv.) and iron(III) chloride hexahydrate (0.1443 g, 0.53 mmol, 0.20 equiv.) in 10 mL of water was added under continuous stirring. The resulting reaction solution was stirred at 90 8Cu nder reflux cooling for 120 h. Afterwards, the hot suspension was filtrated by using ag lass filter, washed (3 20 mL DMF,1 30 mL H 2 O) and dried in air (at room temperature overnight and at 130 8Cf or another 3days). Table 1 contains the used amounts of 2-aminoterephthalic acid and terephthalic acid.
Powder X-ray diffraction:P owder X-ray diffraction patterns were collected using aB ruker D8 Discover powder diffractometer with Cu Ka radiation in the 2q range from 4t o5 0 8 with as tep width of 2q = 0.01028 and an accumulation time of 1.60 ss tep À1 .
The assignment of lattice planes to reflections in the measured powder X-ray diffraction patterns was performed based on avail- . [17] In situ powder X-ray diffraction:T emperature-dependent in situ powder X-ray diffraction measurements of MIL-53(Al 0.8 Cr 0.2 )-NH 2 (X) and MIL-53(Al 0.8 V 0.2 )-NH 2 (X) materials were performed on the Bruker D8 Advance powder diffractometer recorded with Cu Ka radiation in Bragg-Brentano geometry.A ni ns itu measurement chamber XRK 900 from Anton Paar was used with ac ontinuous gas flow of nitrogen. The samples were heated from 30 to 300 8Ci nintervals of 50 K( 30, 50, 100, …3 00 8C). Fast diffraction patterns (range:2 q = 4t o5 08,s tep width:2 q = 0.02048,t ime/step = 0.100 s) were recorded after an equilibration time (25 min) at each temperature step.
Te mperature-dependent high-resolution in situ powder X-ray diffraction measurements of MIL-53(Al 0.8 Fe 0.2 )-NH 2 (X) materials were performed at PETRA III Extension beamline P24 (operated at 20 keV = 0.62 )a tD ESY (Deutsches Elektronensynchrotron in Hamburg, Germany). The samples were inserted in an in situ capillary setup, which was attached to aKappa-diffractometer.Acontinuous stream of nitrogen (100 mL min À1 )w as passed through the capillary.F or temperature-dependent measurements, the samples in the capillaries were heated in steps of 15 Kf rom room temperature to am aximum of 310 8Cw ith ah ot air blower.D iffraction patterns were collected for each temperature step after an equilibration time of 5min.
For ab etter comparability of the measurements at the lab diffractometer and at the synchrotron, all in situ recorded diffraction patterns are presented in reciprocal space (Q).
Structure refinements:T he structure refinement was performed using TOPAS4 .2. The complete measured range 2q = 4.0 to 50.08 of the PXRD measurements was used for the refinement. The Pawley method was used to determine the unit cell parameters starting from reported values of MIL-53(Al)-NH 2 (lt)( space group Cc), [10] MIL-47(V) (space group Pnma) [12] or MIL-53(Cr)ht (space group Imma). [18] Infrared spectroscopy:A TR-IR spectra were recorded on aN icolet 6700 FTIR spectrometer from ThermoFisher Scientific equipped with aSmart iTX ATR-IR accessory with Ge crystal and aMCT detector operating at liquid nitrogen temperature. As can range of 400-4000 cm À1 with ar esolution of 4cm À1 was chosen and 400 spectra were accumulated for each measurement.
NMR spectroscopy:L iquid-phase 1 HNMR spectra were recorded on aB ruker Ascend 400 MHz spectrometer.C hemical shifts were referenced to internal solvent resonances and reported relative to tetramethylsilane (TMS). However,d ue to slightly different pH values of the prepared solutions, the residual solvent signal shifted slightly,w hich led to minor shifts of peaks. Scandium-, vanadiumand chromium-containing materials were digested prior to the measurement using the following procedure:5mg of the MOF material were digested by adding 40 mLofDCl (20 wt %inD 2 O) followed by 5min of ultrasonication and, subsequently,t he addition of 70 mLo fNaOD (30 wt %inD 2 O) and 590 mLo fD 2 O.
For the iron-containing materials, as lightly modified procedure was necessary:5 0mLo fD Cl (20 wt %i nD 2 O) were added to 5mg of MOF material and the mixture was treated for several minutes in an ultrasound bath. Afterwards, 140 mLo fN aOD (30 wt %i n D 2 O) and 140 mLo fD 2 Owere added and the mixture was placed in the ultrasound bath for several minutes. The suspension was centrifuged at 10,000 min À1 for 2min and the clear liquid was filled in an NMR tube. 50 mLN aOD (30 wt %i nD 2 O) and 150 mLo fD 2 O were added to the residual solid, the mixture was placed in the ul-trasound bath, centrifuged and the clear liquid was combined with the previously separated liquid in the NMR tube.
The degree of post-synthetically modified amine groups was calculated based on the following Equation (1) For the calculation of the overall modification degree, the calculated degree of modification obtained from Equation 1) was multiplied by the percentage of 2-aminoterephthalate with respect to the total amount of linkers.
Thermogravimetric analysis:T he thermal stability and the decomposition of the materials was investigated in oxidative atmosphere (20 %O 2 /He) on at hermo balance Cahn TG-2131. 15 mg of sample were heated from 40 to 1000 8Cw ith ah eating ramp of 5Kmin À1 .
Nitrogen physisorption:T he samples were activated for 20 ha t 130 8Cp rior to the nitrogen physisorption measurements. The measurements were performed on an Autosorb 6s etup from Quantachrome. Micropore volumes were estimated by using the tplot method [19] and specific surface areas by using the BET method. [20] Elemental analysis:M etal ratios were determined using ICP-OES (inductively coupled plasma optical emission spectroscopy). An iCAP 6500 Duo from Thermo Scientific with standard equipment was used for the measurements. As ix-point standard was used for the calibration curve. Prior to the measurement, the samples were dissolved in diluted nitric acid. The data analysis was performed with the device's own software "iTEVA9.8".
X-ray absorption spectroscopy:X AS experiments were performed at PETRA III Extension beamline P65 (energy range:4 -44 keV) at DESY. [21] For the measurements at the Fe K-edge, aS i(111)C -type double crystal monochromator was used. The beam current was 100 mA with ar ing energy of 6.08 GeV.A ll samples were prepared as pellets using cellulose as ab inder.A ll spectra were recorded in continuous scan mode both in transmission and in fluorescence mode at ambient temperature and pressure in the range of À150 to 1000 eV around the edge within 180 s. For the data analysis, transmission data were used. For calibration, an Fe foil was measured as areference simultaneously with the samples.
The data treatment was performed using the Demeter software package. [22] In order to compensate for the oversampling of the continuous scan mode, the data points of the obtained spectra were reduced with the help of the "rebin" function of the Athena software (edge region: À50 to + 50 eV;p re-edge grid:5 eV; XANES grid:0 .5 eV;E XAFS grid:0 .05 À1 ). For data evaluation, a Victoreen-type polynomial was subtracted from the spectrum to remove the background using the Athena software. The first inflection point of the edge was taken as edge energy E 0 .