Probing Isoreticular Expansions in Phosphonate MOFs and their Applications

This review is divided into three different parts. In the introduction, the current applications of metal-organic frameworks (MOFs) are shown. Furthermore, the overall stability, a comparison of phosphonate MOFs with carboxylate MOFs, the structural richness of phosphonate MOFs with the challenge in synthesis and different linker geometries is discussed.


Introduction
The systematic synthesis of microporous metal organic frameworks (MOFs) has spawned wide interest in early 2000s for their gas storage properties and sequestration of greenhouse gases. [1,2] As MOF research kept growing, distinct applications in diverse fields, such as catalysis, gas adsorption, drug delivery, food transportation, solar cells, conductivity, proton conductivity, biomimetic chemistry, magnetism etc., have emerged.  The limitless linker geometries, [29] possibility of attaching reactive organic functional groups around linker scaffolds have created new chemistry namely post synthetic modifications to further optimize the MOF properties. [30][31][32][33] To date, the greatest majority of the reported MOFs are synthesized using the phenylcarboxylate derivatives that has trigonal planar metal binding unit fabricating precise metal binding modes. [34,35] Therefore, carboxylate MOFs produced persistent molecular inorganic building units (IBUs) in the shape of paddle wheels or persistent rod shaped IBUs to form the basics or reticular chemistry in this system. [35] Despite the high number of carboxylate MOFs reported in the structure, the industrial applications of MOFs somewhat remained limited due to the presence of carbonyl bonds in the carboxylate metal binding unit, which makes the In the second part, several phosphonate MOFs with Y-shaped planar, X-shaped planar, and X-shaped tetrahedral linker systems are described. The final part discusses different structures of inorganic building units (IBUs) with different metal atoms and a comparison of the formed MOF structures.
carboxylate MOFs more prone to hydrolysis. [36] Therefore, one of the growth points of MOF chemistry should be the use of new metal binding units, which could generate chemically and thermally stable MOFs to perform the functions under harsh conditions. The ideal MOF family should also provide extreme structural richness compared to traditional MOFs.
In this sense, phosphonate metal organic frameworks (phosphonate MOFs) form a distinct family of MOFs that exhibit exceptional thermal and chemical stabilities compared to the conventional MOFs. [37][38][39] Amongst all known MOFs, the phosphonate-based frameworks play only a niche role so far. [38] Despite the limited number of phosphonate MOFs reported in the literature they have already created diverse applications in proton conductivity, catalysis, magnetism etc., [40][41][42][43] and they exhibit potential in forming future non-toxic MOFs, which are suitable for biological and food chemistry applications. [44] The recent interest in phosphonate MOFs have been summarized in several review articles. [37][38][39][40][45][46][47] One of the distinctive natures of phosphonic acid metal binding group is the rich structural diversity observed in this system. [38,48,49] The structural richness of phosphonate metal binding modes could reflect in more diverse applications compared to the traditional MOF families. [38] On the other hand, uncontrolled structural diversity could lead to unexpected growth of inorganic building units of phosphonate MOFs creating random compounds. [49,50] To date, design patterns or strategies in forming predictable porous structures in phosphonate linkers have not been identified yet. In order to create isoreticular expansion or create well-defined pore sizes in this system, such design patterns are needed within the MOF community to create new phosphonate MOFs that could combine near zeolite stabilities with the organic functionalities.
The unique and numeric nomenclature of metal binding modes of phosphonate group have been summarized by Harris and Chandrasekhar. [48,49] The rich metal binding modes of phosphonate MOFs originate from the multiple protonation modes of R-PO 3 H 2 , the possibilities of coordinate covalent and ionic bonding combinations and rich charge balance efforts to compensate the multiple protonation modes of the R-PO 3 H 2 metal binding unit. In addition, the strong non-covalent interactions, such as the presence of strong hydrogen bonding between the phosphonate metal binding units, is another structural factor contributing to the structural diversity. [42,51] Traditional lamellar or pillared layered dense phosphonate metal organic solids are constructed with alkyldiphosphonic acid, or linear phenyldiphosphonic acids. Once such phosphonate linkers form metal organic solids, hydrophilic R-PO 3 H 2 metal binding units and the hydrophobic linker cores create layered and pillared motifs similar to phospholipid bilayers observed in the cell membranes. [37,39,[51][52][53][54][55] Therefore, in order to introduce porosity into phosphonate MOFs our group and others have employed two different types of linker geometries to disrupt the dense lamellar and pillared-layered patterns observed in phosphonate MOFs. [56] The first strategy is to use tetrahedral linker cores expanding into three dimensions based one methane and silane cores. [56][57][58][59] Such linker geometries prohibited the formation of hydrophobic layers and created the large pore sites. The second approach relied on planar tritopic and tetratopic linkers where the phosphonate arms are separated by 90°and 120°a ngles. [44,45,[60][61][62][63][64][65][66] Although, such planar linkers follow the same mechanism of crystallization observed in pillared layered and lamellar phosphonates. The unique geometry of the Y and X shaped linkers generate hexagonal or rectangular void channels with one-dimensional inorganic building units. [38] These recent efforts to synthesize phosphonate MOFs with linkers having Y, X-tetrahedral and X-planar geometries created the first persistent building units in phosphonate-MOF chemistry to achieve isoreticular expansions in this system.
In this minireview, we are summarizing the molecular and one-dimensional IBUs observed in phosphonate MOFs with Y, X-tetrahedral and X-planar geometries to derive persistent inorganic building units to establish the reticular chemistry in this system. We further comment on the conformations of observed IBUs and how the structural evolution of phosphonate MOFs could be achieved. While explaining the structural chemistry of the phosphonate MOFs, we also comment about the applications of the phosphonate MOFs and structural relations in detail in this work.

Y-Shaped Phosphonate Linkers
Linkers with Y-shaped geometry have 120°of separation between the phosphonic acid metal binding groups. Due to the nature of the planar linker, it is not expected to observe lamellar networks with one-dimensional IBUs and three-dimensional pillared-layer networks with two-dimensional sheet-like IBUs. It is more expected, once they form MOFs with one-dimensional IBUs to produce hexagonal void channels. A first example can be seen in the in the following pictures ( Figure 1). Figure 1. Linker molecule, inorganic building unit and MOF structure of: CALF-33-Et 2 H, [61] CAU-14, [66] UPG-1, [70] [Pb 3 (H 3 -PMP) 2 (H 2 O)] [71] and PCMOF-3. [72] CALF-33-Et 2 H [61] is based on the linker 1,3,5-tri-(phenylene-4-triethoxy phosphane) benzene which was synthesized in three steps according to the literature. [67][68][69] Suitable crystals for X-ray diffraction were grown from a mixture of Cu(NO 3 ) 2 , ligand Et 3 H 3 -BBP, ethanol, water and 1,3-diisopropylbenzene at 120°C in a hydrothermal vessel. The blue needle-like crystals were composed of copper-phosphonate pillars which contain two types of copper-phosphonate chains. Square planar copper phosphonate chains and square pyramidal copper-phosphonate mono ethyl ester double chains were connected by stacked Et 2 H-BBP 3units to form the three-dimensional framework. One-dimensional pores measuring 7.2 Å × 16.1 Å suitable for N 2 gas sorption were generated. Experimental N 2 gas sorption isotherm gave a calculated Langmuir surface area of 950 m 2 g -1 and a BET surface area of 810 m 2 g -1 .
The CALF-33-Et 2 H is composed of two distinct one-dimensional inorganic building units ( Figure 1). The first one-dimensional IBU is composed of edge touching 8 membered rings of Cu-O-P-O-Cu-O-P-O (further methodically called s.pl.-Cu-8mpr, square planar-Cu-8 membered phosphonate ring) and the second IBU structure is composed of two of IBUs composed of two s.py.-Cu-8mpr, which are connected via copper dimers to the final shape of the second IBU. The second IBU is more complex and the copper centers are square pyramidal due to the presence copper dimers. Therefore, the CALF-33-Et 2 H structure has two distinct type of hexagonal and parallelogram void channels.
CAU-14 [66] is based on the linker 2,4,6-tri-(phenylene-4-phosphonic acid)-s-triazine linker (H 6 PPT) and was obtained by the reaction of the linker (H 6 -PPT) in a mixture of Cu(NO 3 ) 2 , NaOH solution and water at 190°C. The crystal structure shows corner-sharing trimeric Cu 3 O 9 (H 2 O) 3 units composed of three distorted square-pyramidal Cu 2+ ions. The Cu 2+ ions are coordinated by four different phosphonate groups and an additional water molecule. Each Cu 3 O 9 (H 2 O) 3 cluster combines into a one-dimensional chain which is connected to adjacent chains via the linker H 3 -TPP 3to form a three-dimensional honey-comb like framework. The generated pores are measuring 9.4 Å in diameter ( Figure 1). Interestingly, the crystals are thermal stable up to 380°C and experimental N 2 gas sorption isotherm gave a calculated BET surface area of 647 m 2 g -1 . Also CAU-14 is suitable to reversible take up water without loss of crystallinity.
UPG-1 [70] is based on the linker 2,4,6-tri-(4-(phosphonomethyl)phenyl)-1,3,5-triazine linker (H 6 -TTBMP) and a similar IBU observed in CALF-33-Et 2 H. The only difference is that the 1D IBU in CALF-33-Et 2 H is composed of tetrahedral copper while UPG-1 is composed of octahedral zirconium within the chain structure ( Figure 1). UPG-1 was obtained by the reaction of the linker (H 6 -TTBMP) in a mixture of water, methanol, ZrOCl 2 and HF as mineralizer at 80°C. The three-dimensional framework is based on one-dimensional zirconia-phosphonate chains with o-Zr-8mpr connected by H 6 -TTBMP ligands. Each ZrO 6 octahedral consists of two monodentate and four bidentate phosphonate tetrahedral from six different phosphonate groups. In addition to the monodentate and bidentate phosphonate groups there are non-coordinating phosphonate groups in the crystal lattice. There are two types of void channels observed in UPG-1, the first one is formed via the hydrogen bonding interactions of the fully protonated phosphonic acid ends lined by free P-O groups and a diameter of about 5 Å. The second void channel is observed as distorted hexagonal void with a diameter of about 10 Å and is composed of four linkers molecules coordinating to the 1D IBU nodes (Figure 1). In comparison with the H 6 -PPT ligand (benzene 1,3,5-tri-4-phenylphosphonic acid) the H 6 -TTBMP is more flexible and the functional groups are out-ofplane with respect to the aromatic ring which allows different connections to the Zr 4+ ion. Experimental data of CO 2 adsorption at 195 K shows a calculated BET and Langmuir surface area of 410 m 2 g -1 and 514 m 2 g -1 . In addition to that the framework shows efficient adsorption of hydrocarbons due to CH -π inter-actions and a thermal stability up to 430°C. Overall, the structure of UPG-1 is marvelous with starshaped hydrogen bonded pattern is surrounded by hexagonal void channels observed with Y-shaped linkers. UPG-1 is the clear indication of additional functionalities like hydrogen-bonded networks and hexagonal MOF channels could be together in the same MOF due to the structural diversity of phosphonate MOFs.
Taddei recently continued the work to synthesize UPG-2 with the same Y shaped linker where the methylphosphonate groups are located at meta position. Due to the slightly hindered meta position, the molecular IBU of two-dimensional UPG-2 is composed of the isolated 8 membered rings observed in 1D IBU of UPG-1. [60] [Pb 3 [71] is based on the linker benzene 1,3,5-tri-(4-(phosphonomethyl)phenyl) and was obtained by the reaction with Pb(CH 3 COO) 2 . The framework consists of two types of Pb 2+ ions, two triply deprotonated trisphosphonate ligands (H 3 -PMP) 3 [72] is based on the linker 1,3,5-benzenetriphonsphonic acid and was obtained by diffusion of acetone into a mixture of Zn(ClO 4 )·6H 2 O, linker and water. The framework consists of two different types Zn 2+ ions. One is coordinated by four oxygen atoms of four different triphosphonate ligands and forms t-Zn-8mpr which are connecting to one-dimensional chains ( Figure 1). The other Zn 2+ ion is coordinated by two oxygen atoms of two different triphosphonate ligands and two water molecules. The one-dimensional chains are connected by the linker molecules and the second Zn 2+ ion forming a threedimensional framework. In addition, the linkers are forming πstacked dimers with a distance of ⊥d = 3.718(5) Å. Proton conductivity in H 2 at 25°C and 98 % relative humidity was measured with 3.5 × 10-5 S cm -1 and 4.5 × 10-8 S cm -1 at 25°C and 44 % relative humidity. The well-ordered chains of aqua ligands serve as main source of protons. With an activations energy for proton transfer with E = 0.17 eV, the Grotthuss proton transfer mechanism was proven.

PCMOF-3
The two-dimensional IBU of PCMOF-3 is composed of the classical one-dimensional IBU with edge touching 8 membered M-O-P-O-M-O-P-O rings, which are connected by the second Zn 2+ ion to give the final structure of the two-dimensional building units. The short length of the Y-shaped linker in this study was influential for the oligomerization of inorganic components into two dimensions, while longer tether lengths prove longer distances between the one-dimensional IBUs and there is not enough space for the second zinc ion to connect the one-dimensional chains.

Planar X-shaped organophosphonic acid linker 5,10,15,20-Tetrakis(p-phenylphosphonic acid) porphyrin (H 8 TPPA)
H 8 -TPPA is an interesting linker where the four phosphonic acid metal binding units are separated by 90°of angle. Therefore, it is expected that once H 8 -TPPA is connected to four linear IBUs, it would produce a MOF structure with rectangular void channels. It is related to the tetrakis(p-carboxyphenyl) porphyrin which porphyrin MOFs are also used for catalytic applications. [73][74][75][76] Even a heme-like biomimetic system with ultra-large pore size, high water stability and catalytically active centers [77] and thermally high stable Zr-MOFs [78] are possible. Also the most recent review article on stable MOFs should be mentioned at this point [79] This could lead to an interesting field for the more stable phosphonate porphyrin MOFs. Due to the unique nature of the planar linker, it is not expected to observe lamellar networks unless H 8 -TPPA has two fully protonated phosphonic acids in the final framework structure. In addition, H 8 -TPPA can provide up to 10 negative charges with the deprotonation of hydrogen in the porphyrin ring. Therefore, H 8 -TPPA is a very suitable linker to form MOFs with trivalent and tetravalent metals in addition to the divalent metals commonly used in MOF synthesis.
Co-CAU-36 [80] is based on the linker nickel 5,10,15,20-tetrakis(p-phenylphosphonic acid) porphyrin and was obtained by reaction with a mixture of linker, CoCl 2 ·6H 2 O, DABCO, water and HCl solution at 160°C in a sealed Teflon insert ( Figure 2). The Crystal structure was determined of small sized crystals of less than 1 μm using cRED data at low temperature (96 K). The Co atoms are coordinated by four oxygen atoms of four different bidentate phosphonate groups. Within the IBU there are t-Co-8mpr forming a one-dimensional chain. The chains are connected by the porphyrin linkers Ni-H 4 -TPPA 4forming a threedimensional framework. Within this framework pores with a diameter of about 9 Å when occupied by DABCO and about 12.5 Å when empty were generated. Thermal stability up to 350°C was observed without loss of crystallinity. The BET surface area after gas adsorption with N 2 is 700 m 2 g -1 .
Zr-CAU-30 [65] is based on the linker nickel 5,10,15,20-tetrakis(p-phenylphosphonic acid) porphyrin and was obtained by reaction of a mixture of linker, Zr 4+ salt, NaF and NaOH solution in a sealed Teflon inlet at 180°C. The crystal structure consists of corner sharing ZrO 6 octahedral forming a one-dimensional chain with o-Zr-8mpr. Adjacent chains are connected by partly deprotonated Ni-H 2 -TPPA 6linker forming a three-dimensional framework. Within this framework pores with the size of 13 Å (edge to edge) or 20 Å (diagonal) were generated. Sorption measurements using N 2 at 77 K were performed and a BET surface area of 970 m 2 g -1 was calculated. The compound is chemical stable in a pH range of pH = 0 -12 in HCl/NaOH and organic solvents like pure acetic acid, MeOH, EtOH, acetone, DMF and DCM. Surprisingly the crystals are also stable in a 0.1 M phosphate buffer at pH = 7.2. Additionally, thermal stabil- Figure 2. Linker molecule, inorganic building unit and MOF structure of: Co-CAU-36, [80] Zr-CAU-30, [65] Ni-CAU-29, [63] GTUB-1 [43] and IPCE-1. [81] ity was achieved up to 400°C in air. Cyclic voltammograms of Zr-CAU-30 showed a reversible redox process at a half-wave potential of E 1/2 = -0.649 V. Cyclic stability was shown by measuring 50 cycles with a scan rate of 250 mV s -1 .
Ni-CAU-29 [63] is based on the linker nickel 5,10,15,20-tetrakis(p-phenylphosphonic acid) porphyrin and it was obtained by reaction of the linker and NiCl 2 ·6H 2 O in water at 180°C (Figure 2). The crystal structure consists of edge-sharing NiO 6 polyhedral forming a Ni 2 O 10 cluster. Eight Ni 2 O 10 units are connected by four different Ni-H 6 -TPPA 2linkers forming the threedimensional framework with one-dimensional pores with sizes of 2.8 Å × 5.0 Å and 3.8 Å × 3.8 Å. During the reaction at 180°C and 48 h no cleavage of the central Ni 2+ was observed. The sorption properties were obtained by employing N 2 and H 2 O at 77 K and 298 K, respectively. At 77 K no uptake of N 2 was observed probably due to low kinetics. The H 2 O uptake was the highest in a series of Ni-, Mn-, Co-and Cd-CAU-29 with 181 mg g -1 , 140 mg g -1 , 166 mg g -1 and 116 mg g -1 respectively. Thermal stability was measured by TG up to the collapse of the framework at temperatures higher than 350°C. Chemical stability was proven by stirring Co-CAU-29 in different solvents for 24 h. Tolerated were aqueous solutions of HCl/NaOH in a pH range of 1 -11, the organic solvents DCM, EtOH, DMF, acetone, pure acetic acid. At pH of 0 a loss of crystallinity and above pH 11 as well as in phosphate buffer complete decomposition is observed. The proton conductivity at different temperatures and relative humidities and the Arrhenius activation energies E A were measured by electrochemical impedance spectroscopy and are given in Table 1. Due to the data the Grotthuss mechanism is the prominent pathway for proton conductivity.
GTUB-1 [43] is based on the linker copper 5,10,15,20-tetrakis(p-phenylphosphonic acid) porphyrin and was obtained by reaction of the linker with Cu(NO 3 ) 2 ·3H 2 O, DMF and MeOH at 120°C (Figure 2). The crystal structure contains edge-sharing  4-linkers and generates rectangular void channels with dimensions of 9.78 Å × 7.73 Å. In addition, the linkers tend to πstacking within the three-dimensional framework with a high separation of 3.786 Å and 4.038 Å between the layers. The simulated N 2 adsorption shows a BET surface area of 697 m 2 g -1 .
Thermal stability was observed up to 400°C without decomposition. Furthermore, the toxicity on an intestinal cell line, namely Caco-2, was investigated. Cells were treated with different concentrations of GTUB-1 and were analyzed by measuring mitochondrial enzyme activity. [51,82] No cytotoxicity was observed even at higher concentrations. IPCE-1Ni [81] is based on the linker 5,10,15,20-tetrakis(p-phenylphosphonic acid) porphyrin and was obtained by reaction of the linker with a mixture of Zn(NO 3 ) 2 ·6H 2 O, phenylphosphonic acid, DMF and H 2 O. The crystal framework consists of discrete dinuclear Zn 2 P 6 O 18 units forming t-Zn-8mpr ( Figure 2). The Zn 2+ ions are tetragonal coordinated by oxygen atoms of four different Ni-H 3 -TPPA 5linker molecules and two bidentate phosphonate groups. Therefore, the linker acts tritopic and is bonding adjacent inorganic building units to form the tree-dimensional framework. The remaining two phosphonate groups are involved in H-bondings with adjacent linkers. The threedimensional framework possesses pores of 6.52 Å in diameter which are filled with CH 3 NH 2 + ions to compensate the anionic charge. Thermal stability was proven up to 150°C without change in the framework geometry due to the loss of (CH 3 )NH 2 + ions. Between 150°C and 210°C a change in the framework geometry and above 300°C the collapse of the framework is observed. Adsorption of CO 2 at 195 K with activation at 140°C gave an uptake of 60.9 cm 3 g -1 and calculated BET and Langmuir surface areas of 208.7 m 2 g -1 and 280.3 m 2 g -1 . Proton conductivity was proven by impedance spectroscopy at different temperatures and relative humidities with 1.3 × 10 -5 S cm -1 at 25°C and 80 % RH as well as 1.55 × 10 -3 S cm -1 at 75°C and 80 % RH and an activation energy E A of 0.86 eV.
CALF-25 [83] is based on the linker octaethyl-1,3,6,8-pyrenetetraphosphonate and was obtained by the reaction of the linker with a mixture of BaBr 2 ·2H 2 O, EtOH and water through partially hydrolysis in situ at 120°C. This is the only MOF synthesized using this type of planar tetratopic phosphonate linker in the literature (Figure 3). The crystal structure consists of edgesharing barium phosphonate polyhedral forming one-dimensional chains which are cross-linked to four additional chains by the pyrene linkers giving a three-dimensional porous framework. Every barium atom is coordinated and saturated by phosphonate monoethyl ester oxygen atoms. The one-dimensional pores are shaped rectangular with site lengths of 4.59 Å × 3.89 Å and are lined with phosphonate ethyl ester groups in the corners as well as the aromatic system of the pyrene core on their walls. Due to a barium atom to linker ratio of 1:1 the three-dimensional framework is acidic. Adsorption of N 2 at 77 K after activation at 100°C and a pressure of 10 -5 Torr gave a calculated BET surface area of 385 m 2 g -1 . For characterization of the pore surface area properties CO 2 and CH 4 adsorption were performed. Despite the acidic nature of CALF-25 the zeroloading heats of adsorption were found to be 23.9 kJ mol -1 for CO 2 and 14.6 kJ mol -1 for CH 4 . To prove the hydrophobic and weakly polarizing character water vapor sorption was performed. With a calculated heat of adsorption of around 45 kJ mol -1 it is comparable to graphite and other hydrophobic MOFs. [84][85][86][87] Therefore it suggests that the acidic barium phosphonate chains are shielded by the ethyl ester groups.

Tetrahedral X-shaped organophosphonic acid linkers with methane, silane and adamantane cores
Tetrahedral geometry opens up into three dimensions with ca. 108°of angle where the core is composed of methane, silane and adamantane cores. The first MOF with adamantane core was reported in 2000s by Shimizu and our group reported the first MOF with the methane core in 2015. [38,40] The unique tetrahedral three-dimensional structure prohibits the formation of highly dense hydrophobic hydrocarbons to form lamellar or pillared layered structures observed in traditional phosphonate metal organic solids.  [88] is based on the linker methane tetra-p-phenylphosphonic acid and was obtained by reaction of the linker with a mixture of CuSO 4 ·5H 2 O, Na 3 VO 4 , 2,2′:6′:2′′-terpyridine, HF and water at 180°C (Figure 4). The obtained three-dimensional framework consists of a one-dimensional chain containing octahedral coordinated Cu(II) atoms, square-pyramidal coordinated Cu(II) atoms and tetrahedral coordinated V(V) atoms. The octahedral coordinated Cu(II) atom is surrounded by four phosphonate and two VO 4 tetrahedrons. The square-pyramidal coordinated Cu(II) atom is surrounded by two phosphonate tetrahedrons and three nitrogen atoms of the tridentate terpy ligand. The V(V) atoms are surrounded by two phosphonate tetrahedrons, one Cu(II) octahedron and an aqua ligand. While the square pyramidal [{Cu(terpy)}] 2+ units are exclusively coordinated by MTPPA oxygen atoms and act like dead-end clamps, the connection in the three-dimensional framework is due to the tetrahedral V(V) and the octahedral Cu(II) sites. Adjacent one-dimensional chains are connected by MTPPA ligands coordinating to the V(V) sites. For the purpose of this study, the [{Cu(terpy)}] 2+ units should be used as moldcast templates which can be removed due to decomposition at around 250°C. In a TGA measurement the sample was heated to 375°C with 5°C min -1 and kept at this temperature for 75 minutes. A weight loss of 8.6 % due to the loss of water and half of the terpyridine molecules was observed. For final results further investigations are currently done.
Zn 2 H 4 -MTPPA [56] is based on the linker methane tetra-pphenylphosphonic acid and was obtained by the reaction of the linker with ZnSO 4 ·7H 2 O in DMF at 180°C (Figure 4). A onedimensional chain containing edge-sharing t-Zn-8mpr is formed. The Zn(II) atoms are tetrahedrally coordinated to four R-PO 3 H 1units derived from H 4 -MTPPA 4 -linkers. The three-dimensional framework is build up by connecting the chains via the H 4 -MTPPA 4 -molecules. The calculated BET surface area by simulated N 2 adsorption at 77 K shows 927 m 2 g -1 . The void volume was computed with 51.3 % with a 0 Å sized probe and the accessible pore volume was computed with 46.0 % with a helium probe. SZ-2 [59] is based on the linker methane tetra-p-phenylphosphonic acid and was obtained by slow diffusion and ionothermal reaction of separated ZrCl 4

and H 8 -MTPPA into an ionic liquid containing N-butyl-N-methylpyrrolidinium bromide ([C 4 mpyr][Br]) and N-ethylpyridinium bromide ([C 2 py][Br]
) in the presence of HF. ZrO 3 F 3 and ZrO 4 F 2 moieties are found which are coordinated by three and four MTPPA ligands. Two ZrO 3 F 3 and one ZrO 4 F 2 units are connected by four bridging CPO 3 groups forming a Zr 3 P 4 O 10 F 8 cluster (Figure 4). One MTPPA ligand connects four Zr 3 P 4 O 10 F 8 cluster to form a three-dimensional framework. Adsorption of N 2 resulted in a calculated BET surface area of 225 m 2 g -1 and Langmuir surface area of 375 m 2 g -1 . The thermal stability up to 300°C was found by thermogravimetric analyses. The compound is chemical stable against boiling water and aqueous solutions in a pH range of 1-11. SZ-3 [59] is based on the linker 1,3,5,7(p-phenylphosphonic acid) adamantine and was obtained by ionothermal reaction of ZrCl 4

and H 8 -PPA in an ionic liquid containing N-butyl-Nmethylpyrrolidinium bromide ([C 4 mpyr][Br]) and N-ethylpyridinium bromide ([C 2 py][Br]
) in the presence of HF. The ZrO 6 and ZrO 3 F 3 moieties are found which are coordinated by six and three bridging PPA ligands. Six ZrO 3 F 3 units are connected to one ZrO 6 unit by six corner-sharing phosphonate groups generating a Zr 7 P 8 O 22 F 16 cluster (Figure 4). These Zr 7 P 8 O 22 F 16 clusters are connected by PPA linkers to form the three-dimensional framework. Adsorption of N 2 resulted in a calculated BET surface area of 572 m 2 g -1 and Langmuir surface area of 695 m 2 g -1 . The thermal stability up to 210°C was found by thermogravimetric analyses. The compound is chemical stable against boiling water, aqueous solutions in a pH range of 1-11 and fumic acids including aqua regia, conc. HCl and conc. HNO 3 . When SZ-3 is soaked in aqua regia the adsorption of N2 resulted in slightly higher BET and Langmuir surface areas of 594 m 2 g -1 and 732 m 2 g -1 .

Tetrahedral X-shaped organophosphonic acid linkers with silane core
Zr-H 4 -STPPA [41] is based on the linker tetraphenylsilane tetrakis-4-phosphonic acid, H 8 -STPPA, and was obtained by the reaction of the linker with a mixture of ZrOCl 2 ·8H 2 O, water, methanol and HF at 180°C. The framework consists of adjacent ZrO 6 octahedrons each connected by four R-PO 3 H 1tetrahedrons from the H 4 -STPPA linkers ( Figure 5). In addition, two monodentate R-PO 3 H 1tetrahedrons are connecting adjacent chains to form the three-dimensional framework. Within the chains o-Zr-8mpr were found. The crystals are chemical stable in organic solvents like DMA, n-hexane, acetonitrile, 2-propanol, DCM, DMF, methanol, chloroform, ethanol, tetrachloromethane, acetone and propanol as well as in aqueous solutions in a pH range of 1 -12, boiling water, conc. HCl, aqua regia and conc. HNO 3 . Due to Lewis acid and Brønsted acid sites in the framework heterogeneous catalyzation was tested using styrene oxide and CO 2 in a cycloaddition reaction. Under optimal reaction conditions yields of 95 %, a TON of 2850 with a TOF of 238 h -1 were obtained. Figure 5. Linker molecule, inorganic building unit and MOF structure of: Cu 2 H 4 STPPA 2H 2 O, [56] Zn 2 H 4 STPPA [56] and Zr(H 4 STPPA). [41] Zn 2 -H 4 -STPPA [56] is based on the linker tetraphenylsilane tetrakis-4-phosphonic acid and was obtained by the reaction of the linker with a mixture of ZnNO 3 , water and methanol at 150°C ( Figure 5). The framework consists of adjacent ZnO 4 tetrahedrons each connected by four R-PO 3 H 1tetrahedrons from the H 4 -STPPA linkers forming one-dimensional chains with t-Zn-8mpr. The three-dimensional framework is build up by connecting the chains via the STPPA molecules. The calculated BET surface area by simulated N 2 adsorption at 77 K shows 565 m 2 g -1 . The void volume was computed with 48.1 % with a 0 Å sized probe and the accessible pore volume was computed with 33.8 % with a helium probe. [56] is based on the linker tetraphenylsilane tetrakis-4-phosphonic acid and was obtained by the reaction of the linker with a mixture of Cu(NO 3 ) 2 ·3H 2 O, NaOH, water and methanol at 150°C. The framework shows three distinct phosphonate protonation modes ( Figure 5). Reduced to one inorganic building unit it shows one full deprotonated RPO Eur. J. Inorg. Chem. 2020, 1542-1554 www.eurjic.org

Structural Comparison of Inorganic Building Units (IBUs) Tetrahedral Metal Coordination
The with a less straightened chain (156°> 142°> 125°) due to less distance between the linker molecules forming the connection (Figure 7). Therefore, a promising way to maximize the void volume is to generate straight one-dimensional chains to maximize the distance between the organic linkers. The same structural motif is given in the compound Co-CAU-36 with a t-Co-8mpr forming a one-dimensional chain. Compared to the same linker with a discrete Zn 2 P 6 O 18 unit in the IPCE-1Ni, forming a t-Zn-8mpr as well, the surface area drops from 700 m 2 g -1 to 209 m 2 g -1 . Therefore, it is believed that the formation of a one-dimensional chain gives access to a more ordered and bigger void volume (12.5 Å in diameter when unoccupied) [80] compared to a denser structure with a smaller void volume (6.5 Å). [81] To exclude the effect of the metal in both complexes the tetrahedral covalent radii of Co and Zn are given with 1.256 Å and 1.304 Å [89] and are very comparable in size.

Square Pyramidal or Trigonal Bipyramidal Metal Coordination
The common structure in the components Pb 3 (H 3 L) 2

Octahedral Metal Coordination
The structural motif of the following compounds UPG-1, Zr-CAU-30 and Zr-H 4 -STPPA is an octahedral coordination around the Zr atom and the formation of o-Zr-8mpr in one-dimensional chains ( Figure 11). This structure is also regarded with the tetragonal coordinated metal atoms. The difference is that a one-dimensional chain composed of tetrahedrons can exist in a bend form as described previously. Thanks to the four equatorial and two axial ligands that is not the case for octahedral coordinated metal atoms in one-dimensional chains. In most cases the equatorial ligands make up the connection in the chain and the axial coordination sites are filled with water molecules or connect to other linker molecules. The compared Zrbtbp-a [90] consists of a layered structure of short chains containing three Zr octahedrons with o-Zr-8mpr ( Figure 12). The extended inorganic connectivity is missing. Within the short chain the central Zr atom is coordinated by six bidentate phosphonate groups and the two terminal Zr atoms are connected by three monodentate phosphonate groups each. Due to the missing connectivity the layers are shifted  and the voids are capped by adjacent layers. Therefore, the BET surface area was calculated to be as low as 10 m 2 g -1 . Compared to UPG-1 with a one-dimensional chain containing octahedral Zr atoms and void channels through the whole system, the BET surface area is considerably larger with 410 m 2 g -1 .
The difference between Zr-CAU-30 and Co-CAU-36 is the coordination of the metal atom. The M-8mpr, M = o-Zr or t-Co, differ in their constitution due to the different metal coordination. Therefore, the Zr-CAU-30 is more uniform and allows more organized and bigger voids then the compared Co-CAU-36. The crystallographic data shows rectangular voids for both components with different sizes. While Co-CAU-36 (Figure 13 left) got a void diameter of 12.5 Å the Zr-CAU-30 (Figure 13 right) got a slightly bigger voids with a diameter of 13 Å. This reflects the BET surfaces with 700 m 2 g -1 compared to 970 m 2 g -1 . Therefore, the previously mentioned straightening of the chain produces more uniform and bigger voids allowing higher surface areas.
On the other hand tetrahedral linkers provide different mechanism on porosity compared to the planar porphyrin cored linkers. Octahedral coordination environment of Zr-H 4 -STPPA allowed the coordination of more H 4 -STPPA units around octahedral Zr leading more dense network while tetrahedral, Zn 2 -H 4 -STPPA opened up the distance between the linkers resulting in a higher BET surface area while Zr-H 4 -STPPA is non-porous. Previous results [56] show that the known H 8 -MTPPA linkers generate higher surface areas compared to the MOFs synthesized with H 8 -STPPA linker.

Metal Clusters
There are still very limited number of phosphonate-MOFs that have metal clusters as IBUs. Two of such phosphonate-MOFs also have the monomeric 8-membered rings observed in the 1D chains as IBUs (See Figure 7). IBUs of Ni-CAU-29 ( Figure 14a) and IPCE-1Ni (Figure 14b) are connected with the same linker Ni-H 8 -TPPA but do not form any chain structures like Zr-CAU-30 and Co-CAU-36. In Ni-CAU-29 there are Ni-Ni distances of 3.19 Å and 5.12 Å. The smaller distances belong to the Ni 2 O 2 cluster and therefore a formation of one-dimensional chains is not given to the bigger distances between the linkers. The same goes for IPCE-1Ni where the Ni-Ni distance between the linkers changes frequently between 6.22 Å and 4.98 Å. The distances between Zn-Zn in the distinct t-Zn-8mpr with 4.64 Å compared is smaller. A possible explanation appears by looking at the linker. The tetratopic porphyrin core with its π-conjugated system is planar with except for the phenyl rings ( Figure 15). These can rotate along the central axis and extend to the required  . Schematic visualization of planar porphyrin core and saddle conformation. Double bonds omitted for clarity. [44] space. Adding a suitable metal atom, the porphyrin core will bend into a saddle conformation and takes up more space, which is not provided by the metal-metal interactions.
In the other two structures, molecular IBUs have been synthesized via coordinated fluorine atoms in at least two positions around the octahedral coordination sphere of Zr to block the oligomeriation of the IBU into 1D chains. In comparison with other Zr-MOFs generated with a solvothermal or hydrothermal reaction the compounds SZ-2 and SZ-3 were generated by an ionothermal reaction providing different reaction dynamics. [59]

Conclusion
In conclusion, we have presented the known and persistent design patterns observed in the inorganic building units of phosphonate-MOFs. The most persistent structural motif observed in phosphonate-MOFs is the eight-membered M-O-P-O-M-O-P-O rings and the one dimensional IBUs comprising the corner shared such rings as monomers. We have further compared the different zero-dimensional inorganic building units and the resulting MOFs different linkers geometries. The rich conformational changes between tetrahedral and octahedral metal centers in M-O-P-O-M-O-P-O rings have generated flexible BET surface areas in phosphonate-MOFs. The observed conformational flexibility of one-dimensional IBUs of phosphonate MOFs, presence of tetrahedral metal centers facing the pore sites and square planar metals in porphyrin rings are clear indications for the small molecule storage or separation such as CO 2 and H 2 S. Furthermore, we showed possible pathways to build up desired framework patterns using multi branched phosphonate linkers in Y-, X-, and X-tetrahedral geometries to reach well-defined pores. The current phosphonate MOFs has already generated variety of applications ranging from proton conductivity, catalysis, magnetism as well as biological and food chemistry applications. The identification of such persistent IBUs would pave new ways to isoreticular expansions of phosphonate MOFs and therefore creating a broader catalogue of synthetic approaches and applications.
In addition, we provided a new nomenclature for one-dimensional IBUs for phosphonate MOFs. The abbreviations used for the coordination of the metal atoms are: t = tetrahedral; s.pl. = square planar; s.py = square pyramidal; o = octahedral; 8mpr = 8-membered phosphonate ring(s).