Document [original]

www.advmat.de

CommuniCation

Phosphonate Metal–Organic Frameworks: A Novel Family

of Semiconductors

Konrad Siemensmeyer, Craig A. Peeples, Patrik Tholen, Franz-Josef Schmitt,

Bünyemin Çoşut, Gabriel Hanna,* and Gündoğ Yücesan*

Dr. K. Siemensmeyer

Helmholtz-Zentrum

Berlin 14109, Germany

C. A. Peeples, Prof. G. Hanna

University of Alberta

Edmonton T6G 2R3, Canada

E-mail: [email protected]

P. Tholen, Prof. G. Yücesan

Technische Universität

Berlin 13355, Germany

E-mail: [email protected]

F.-J. Schmitt

Marthin-Luther-Universität Halle-Wittenberg

Halle (Saale) 06120, Germany

B. Çoşut

Gebze Technical University

Gebze 41400, Turkey

The ORCID identification number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/adma.202000474.

DOI: 10.1002/adma.202000474

geometries and functional groups, and

the option of postsynthetic modification

of pore surfaces,[8–10] many diverse appli-

cations of MOF chemistry have emerged,

including small molecule storage,[11]

greenhouse gas sequestration,[12] drug

delivery,[13,14] and detoxifying agents;[15] and

the presence of IBUs has led to applica-

tions in heterogeneous catalysis,[16,17] mag-

netism,[18–21] and conductivity.[21,22] Among

these applications, magnetic MOFs have

recently attracted a great deal of atten-

tion due to the possibility of creating tun-

able magnetic materials by varying the

host–guest interactions at pore sites or

exploiting structural changes induced by

MOF breathing; and conductive MOFs

are expected to serve as next-generation

porous electrode materials with higher and customizable sur-

face areas compared to active carbon electrodes.[21–23]

Traditional MOFs have primarily relied on molecular IBUs

known widely as paddle wheel patterns.[1–3] To synthesize mag-

netic MOFs based on molecular IBUs, the IBUs must be close

enough to each other to generate the desired magnetic interac-

tions. The design patterns for such magnetic MOFs have been

summarized in two recent review articles.[18,19] Typically, shorter

linkers such as CN or azolate linkers make magnetic interac-

tions possible between the inorganic components.[24,25] Also,

linkers that can generate free radicals may be used to create

magnetically significant MOFs.[26,27] However, because MOF

chemistry has evolved toward the use of longer organic linkers

for larger surface areas, the distances between the molecular

IBUs has increased and thereby diminished the possibility of

constructing magnetic MOFs with molecular IBUs.[28] Thus,

new architectural strategies for synthesizing magnetic MOFs

are in need. Along these lines, 1D and 2D IBUs can provide a

more suitable platform for magnetic interactions as the metal

centers may come into close proximity to each other in such

geometries.[29–31] MOFs, which are synthesized at high tempera-

tures and under hydrothermal reaction conditions, usually form

1D chain IBUs.[32–34] In contrast, conventional MOFs, which are

synthesized at low temperatures in the presence of organic sol-

vents,[1–6] usually form molecular IBUs. 1D magnetic chains are

well known,[29] but relatively few porous magnetic MOFs with

1D IBUs have been reported in the literature.[18,19,25,31,35]

The construction of conductive MOFs requires more subtle

design elements. For example, highly conjugated linkers

such as phthalocyanine or porphyrins with orthodiimine,

Herein, the first semiconducting and magnetic phosphonate metal–organic

framework (MOF), TUB75, is reported, which contains a 1D inorganic

building unit composed of a zigzag chain of corner-sharing copper dimers.

The solid-state UV–vis spectrum of TUB75 reveals the existence of a narrow

bandgap of 1.4eV, which agrees well with the density functional theory (DFT)-

calculated bandgap of 1.77eV. Single-crystal conductivity measurements for

different orientations of the individual crystals yield a range of conductances

from 10−3 to 103 S m−1 at room temperature, pointing to the directional nature

of the electrical conductivity in TUB75. Magnetization measurements show

that TUB75 is composed of antiferromagnetically coupled copper dimer

chains. Due to their rich structural chemistry and exceptionally high thermal/

chemical stabilities, phosphonate MOFs like TUB75 may open new vistas in

engineerable electrodes for supercapacitors.

Metal–organic frameworks (MOFs) emerged as a new family

of microporous materials at the beginning of the 21st cen-

tury.[1–5] They are composed of inorganic building units (IBUs)

and organic linkers, which combine to create microporous

frameworks.[1,3,6,7] Owing to the vast range of organic linker

KGaA, Weinheim. This is an open access article under the terms of the

Creative Commons Attribution-NonCommercial-NoDerivs License,

which permits use and distribution in any medium, provided the original

work is properly cited, the use is non-commercial and no modifications

or adaptations are made.

Adv. Mater. 2020, 32, 2000474

www.advmat.dewww.advancedsciencenews.com

orthodihydroxy, and azolate metal-binding units connected via

molecular IBUs composed of a single metal ion, are known

to give rise to conductivity.[21–23] However, due to the limited

number of metal-binding modes for the single nitrogen and

oxygen donating linkers and the highly conservative nature of

metal binding in these systems, progress in the design of con-

ductive MOFs has been limited. Alternative metal binding units

capable of yielding both rich structural diversity and conduc-

tivity are needed for constructing next-generation conductive

MOFs.[34,36]

Phosphonate MOFs are known for their high structural diversity

due to the multiple metal-binding modes and protonation states

of the phosphonic acid group.[37] They are known to contain com-

plex molecular clusters and 1D/2D IBUs.[30,33–36] Recently, Yücesan

and co-workers synthesized the phosphonate MOF TUB75 (where

TUB stands for Technische Universität Berlin) at temperatures

above 180°C and under hydrothermal reaction conditions.[38] The

crystal structure revealed that this MOF contains 1D copper dimer

chains linked by polyaromatic 1,4-naphthalenediphosphonic acid

linkers (see Figure1). This chain structure is unique compared to

that of previously reported 1D IBUs in the literature with respect to

the presence of three different (and relatively short) characteristic

Cu–Cu distances along the 1D IBU (see Figure1B).[30,33–36] As seen

in Figure1B, TUB75 is composed of zigzag copper dimer chains

(one of which is portrayed in Figure 1B; the experimental and

calculated Cu–Cu distances are given in Figure S2 and Table S1,

Supporting Information, respectively). Its surface area was previ-

ously measured to be 132.1 m2 g−1.[38] The thermal decomposition

patterns of phosphonate MOFs constructed using 4,4′-bipyri-

dine as the auxiliary linker with 1,4-naphthalenediphosphonic

acid, 2,6-naphthalendiphosphonic acid, and different aromatic

phosphonic acids were also previously reported.[38,39] MOFs in

this family have a general tendency to be thermally stable up to

≈375°C, after which thermal decomposition begins. In our pre-

vious work with 2,6-naphthalendiphosphonic acid, we found that

removing the 4,4′-bipyridines to obtain pure metal phosphonates

increases the thermal stability to ≈400°C.[37] As seen in Figure1B,

all of the 1,4-naphthalenediphosphonic acids in TUB75 are fully

deprotonated, which leads to substantial electron delocalization

within the 1D IBU (see electronic population analyses in Table S2

of the Supporting Information).

In light of the above, in this study, we revisit our phospho-

nate MOF TUB75 to explore its conductive and magnetic prop-

erties. As we will discuss below, we find that TUB75 has an

indirect bandgap of 1.4 eV (based on a Tauc plot of the UV–

Vis spectrum), making it a semiconductor. In addition, we find

that TUB75 possesses an antiferromagnetic chain-type IBU.

Our density functional theory (DFT) calculations of TUB75’s

bandgap, band structure, partial density of states, and rela-

tive energies of the ferromagnetic (FM) and antiferromagnetic

Adv. Mater. 2020, 32, 2000474

Figure 1. A) One layer of the [{Cu2(4,4’-bpy)0.5}(1,4-NDPA)] (TUB-75) MOF, showing nine 1D copper dimer IBUs and four void channels (which extend

into and out of plane). B) Side view of the 1D IBU consisting of a zigzag chain of corner-sharing copper dimers, with Cu–Cu distances of less than 3 Å.

Dimers are colored based on their CuCu bond distances. Color definitions: O—red; N—orange; Cu—cyan; C—black; P—blue.

www.advmat.dewww.advancedsciencenews.com

(AFM) configurations, provide detailed insight into its elec-

tronic structure. The details of our calculations, which employed

Slater-type orbital (STO) basis sets, can be found in the Compu-

tational Details section of the Supporting Information.

In Figure3B, we present an indirect Tauc plot derived from

the UV–vis spectrum (shown in Figure S2, Supporting Infor-

mation) of a sample of pure handpicked (under a microscope)

TUB75 crystals—thin green needles with an average length

of 0.5 mm. Linear extrapolation of this Tauc plot[40,41] yields

an estimate of the optical bandgap of 1.4 eV, indicative of a

semiconductor. To shed light on the origin of the semiconduc-

tivity, we performed DFT calculations, the details of which

are provided in the Supporting Information. The density of

states calculation yielded a HOMO–LUMO gap of Eg= 1.77eV,

which is in good agreement with the experimental bandgap

of 1.4eV. Based on the projected density of states (pDOS) in

Figure2, we see that the HOMO–LUMO gap is predominantly

due to atomic orbitals associated with the carbon atoms in the

π-conjugated 1,4-naphthalenediphosphonic acid (1,4-NDPA)

and 4,4’-bipyridine (4,4′-bpy) auxiliary linker groups. It also

appears that there is some contribution from the nitrogen

orbitals to the LUMO and a very small contribution from the

oxygen orbitals to the HOMO. As for copper, there is effec-

tively no contribution from the copper orbitals to the HOMO

and LUMO. There is indication of spin dependence in the

higher energy virtual orbitals, primarily associated with the

copper atoms and smaller contributions coming from carbon,

nitrogen, phosphorous, and oxygen. We further projected

the carbon pDOS into the individual contributions from the

1,4-NDPA and the 4,4′-bpy carbons (Figure3). This projection

reveals that the HOMO and LUMO are spatially separated,

with the HOMO localized on the 1,4-NDPA carbons and the

LUMO localized on the 4,4′-bpy carbons. In such a case, a

photoexcited electron (in the LUMO) would be spatially sepa-

rated from its hole (in the HOMO).

Electrical conductivity measurements on MOFs have been

mainly based on polycrystalline pellets. However, such meas-

urements may greatly underestimate the conductance of the

MOF due to contact/grain boundary resistances and aniso-

tropic electrical conduction. On the other hand, single-crystal

measurements can provide much more accurate conductance

values, provided that the crystals are large enough. In light of

this, we carried out a number of single-crystal measurements

on TUB75 by clamping the individual crystals between two gold

surfaces of a relay. From room-temperature measurements, we

obtained a range of resistances from 10 Ω to 10 MΩ, depending

on the orientation of the crystal with respect to the gold sur-

faces. Assuming that the TUB75 crystal makes perfect contact

with the gold surfaces, these resistances yield a maximum

conductance of ≈103 S m−1 and a minimum conductance of

≈10−3 S m−1 (see Supporting Information for the details of the

calculations). However, since the TUB75 crystals do not make

perfect contacts with the gold surfaces, the actual conductances

could even be higher than our reported values. Nevertheless,

our results show that TUB75 is a semiconductor and provide

strong evidence of the directional nature of the electrical con-

ductivity of TUB75. We are currently working on growing larger

crystals to maximize the contact surface area in order to better

understand TUB75’s directional conductivity

Next, we report the results of our magnetiza-

tion measurements on TUB75 in Figure4A (see Sup-

porting Information for the details of the measurements).

The magnetization data exhibits a Néel temperature of

T≅ 30 K (i.e., the temperature corresponding to the maximum

Adv. Mater. 2020, 32, 2000474

Figure 2. Spin-up and spin-down projected density of states for TUB75 in the AFM configuration. A) Copper. B) Carbon. C) Nitrogen. D) Phosphorous.

E) Oxygen.

www.advmat.dewww.advancedsciencenews.com

magnetization) that increases with increasing field strength,

a behavior that is characteristic of a material with AFM

correlations (see Figure5 for a depiction of the AFM correla-

tions of spins in the geometry-optimized structure). This is

corroborated by our DFT-calculated exchange energy (i.e., the

energy difference between the AFM and FM configurations) of

Eex= EAFM –EFM= –37.3 meV, indicating that the AFM configu-

ration is more stable than the FM one. Moreover, the position

of the maximum does not vary with temperature, which is char-

acteristic of the presence of short-range order in the 1D spin

chains and is consistent with the zigzag chains observed in

the crystal structure (Figure1B). At 2 K, we observe a non-zero

magnetization that increases with increasing field strength and

appears to plateau at higher fields; for the lower field strengths,

the magnetization initially decreases and then increases with

increasing temperature (up to 10 K), while for the higher field

strengths the magnetization simply increases with increasing

temperature (up to 10 K). The upturn of the magnetization

below 10 K (known as a Curie tail) observed for the lower field

strengths is suggestive of the presence of a small amount of

paramagnetic impurity, e.g., Cu ions that are not embedded in

the TUB75 crystal structure.[31] At high temperatures, the mag-

netization decreases to zero with increasing temperature, as

expected (see Figure4A).

Given the underlying 1D chain geometry, we fit our high-

temperature (>30 K) magnetic susceptibility data (Figure 4B)

to Heisenberg chain and dimer chain models (depicted in

Figure 4D). As shown in Figure 4B, close fits to the data

were obtained with coupling constants of Jchain= 16.8 K and

J′= –22 K for the Heisenberg chain model,[42,43] and coupling

constants of Jdimer= 54 K and Jchain= –2.6 K for the dimer

chain model.[44–46] The signs and magnitudes of the Heisenberg

chain coupling constants are consistent with those observed

in antiferromagnetically coupled 1D chains, while those of the

dimer chain model are suggestive of another type of coupling.

The diamagnetic contribution is small in both cases. As the

temperature is lowered to 1 from 30 K, the magnetic suscep-

tibility predicted by the dimer chain model drops rapidly (ulti-

mately to negative values), in contrast with the experimental

susceptibility which drops less rapidly and remains positive.

On the other hand, the magnetic susceptibility predicted by

the Heisenberg chain model drops less rapidly than the experi-

mental susceptibility and remains positive. As mentioned ear-

lier, these deviations may be due to the presence of impurity

spins; thus, we fit the low-temperature (<10 K) magnetic sus-

ceptibility data to a field-dependent Brillouin function plus a

baseline signal (dotted black line) that is due to the 1D chains.

From Figure 4B, we see that this combination of functions

(with an impurity content of ≈8%, S= 1⁄2 and g= 2) closely fits

the low-temperature data. Given that the magnetic susceptibili-

ties of TUB75, the Heisenberg chain, and the dimer chain are

positive, positive, and negative, respectively, at very low temper-

atures (see Figure4B), it appears that its IBU is best described

in terms of Heisenberg chains. A Curie–Weiss fit to the high

temperature portion of the inverse magnetic susceptibility is

shown in Figure4C. All details of the data fitting are given in

the Supporting Information.

Finally, in Figure6, we present the DFT-calculated spin

density isosurface, focusing on a portion of the IBU in the

AFM configuration. The antiparallel spin density along

the copper dimer chain is clearly seen and it is delocalized

onto the coppers and their nearest-neighbor oxygens and

Adv. Mater. 2020, 32, 2000474

Figure 3. A) Spin-up and spin-down projected density of states of the 1,4-NDPA and 4,4′-bpy carbons for TUB75 in the AFM configuration. B) Indirect

Tauc plot derived from the UV–Vis absorption spectrum of TUB75 (shown in Figure S2, Supporting Information), revealing a typical semiconductor

pattern and a bandgap of Eg= 1.4eV.

www.advmat.dewww.advancedsciencenews.com

nitrogens. This result suggests that the magnetic behavior

of TUB75 is likely dependent on the shared spin density of

these three atoms.

Herein, we have reported on the conductive and mag-

netic properties of the phosphonate MOF, TUB75. With

an experimental bandgap of 1.4 eV and room-temperature

(orientation-dependent) conductances ranging from ≈10−3 to

≈103 S m−1, TUB75 is the first semiconducting phosphonate

MOF in the literature, paving the way for a new family of

semiconductors with an extremely rich structural chemistry.

The metal-binding modes of the phosphonic acid group

in TUB75 support a 1D IBU composed of a zigzag copper

dimer chain, which was found to be antiferromagnetically

coupled. The temperature-dependent magnetic suscepti-

bility data was well fit using a combination of a Heisenberg

chain model at higher temperatures and Brillouin functions

at very low temperatures. Our experimental measurements

were accompanied by DFT calculations, which yielded a

bandgap of 1.77eV in good agreement with the experimental

one and support the AFM nature of the IBU. Given the high

thermal/chemical stabilities of phosphonate MOFs and the

numerous metal-binding modes of phosphonates, our find-

ings suggest that they could be used in next-generation elec-

trodes and supercapacitors capable of withstanding harsh

operating conditions. The vast structural diversity of phos-

phonate MOFs could lead to a new generation of porous

materials with engineerable surface areas and magnetic/

conductive properties. Currently, we are working on the

Adv. Mater. 2020, 32, 2000474

Figure 4. Magnetic response data for TUB75. A) Magnetization versus temperature data for TUB75 in different applied magnetic fields. B) Magnetic

susceptibility, χ, (colored circles) obtained from the magnetization data along with fits (solid blue lines) to the Heisenberg chain and dimer chain models.

The upturn in the low-temperature signal (<10 K), which is suggestive of the presence of paramagnetic impurities, is fit by Brillouin functions (solid colored

lines) with a baseline signal (dotted black line). C) Magnetic susceptibility (green circles) obtained in a 5 T magnetic field, corrected for the diamagnetic

background, and the inverse susceptibility (purple circles) on which a Curie–Weiss high-temperature linear fit is shown (solid purple line). D) Schematic

pictures of Heisenberg chains and dimer chains, the models used to fit the magnetic susceptibility data. See Supporting Information for details.

Figure 5. Minimum energy structure of the 1×3 × 1 supercell depicting

the antiferromagnetic configuration of the electrons on each copper (α-

spin: red, β-spin: blue). Color definitions: O—red; N—orange; Cu— cyan;

C—black; P—blue; H—white.

www.advmat.dewww.advancedsciencenews.com

Adv. Mater. 2020, 32, 2000474

reticular chemistry of phosphonate MOFs to explore these

possibilities.

Supporting Information

Supporting Information is available from the Wiley Online Library or

from the author.

Acknowledgements

The authors thank Dr. Pradip Pachfule from TU-Berlin for his help with the

UV–vis measurements. G.Y. thanks the DFG for funding his work with grant

number DFG YU 267/2-1 and DAAD for supporting Prof. Dr. Bünyemin

Çoşut’s visit to his lab at TU-Berlin. G.H. acknowledges support from the

Natural Sciences and Engineering Research Council of Canada (NSERC).

The DFT calculations were enabled by support provided by WestGrid (www.

westgrid.ca) and Compute Canada (www.computecanada.ca).

Conflict of Interest

G.Y. has a patent pending protecting some of the presented results.

Author Contributions

K.S. and C.P. contributed equally to this work. G.Y. created the

hypothesis, supervised the project, and wrote, synthesized, and revised

the non-computational parts of the manuscript. P.T. generated the crystal

structure figure (Figure1) and resynthesized the crystals for the electrical

conductivity measurements. C.A.P. performed the DFT calculations,

generated the computational figures/tables, and wrote the initial drafts

of the computational parts of the manuscript. G.H. supervised the

calculations, wrote/revised the computational parts of the manuscript,

and performed extensive critical revisions of the entire manuscript. K.S.

performed the magnetic measurements and wrote the initial drafts of the

corresponding methods and results sections. B.Ç. prepared the Tauc plot

(Figure3B). F.-J.S. performed the electrical conductivity measurements.

Keywords

DFT calculations, electrodes, magnetic MOFs, metal–organic

frameworks, semiconductive MOFs, supercapacitors

Received: January 20, 2020

Revised: March 21, 2020

Published online: May 6, 2020

[1] M.Eddaoudi, J.Kim, N.Rosi, D. Vodak, J.Wachter, M.O’Keeffe,

O. M.Yaghi, Science 2002, 295, 469.

[2] H. C.Zhou, J. R.Long, O. M.Yaghi, Chem. Rev. 2012, 112, 673.

[3] O. M.Yaghi, M.O’Keeffe, N. W.Ockwig, H. K.Chae, M.Eddaoudi,

J.Kim, Nature 2003, 423, 705.

[4] H.Furukawa, K. E.Cordova, M.O’Keeffe, O. M.Yaghi, Science 2013,

341, 1230444.

[5] H.-C.“Joe” Zhou, S.Kitagawa, Chem. Soc. Rev. 2014, 43, 5415.

[6] M. J.Kalmutzki, N.Hanikel, O. M.Yaghi, Sci. Adv. 2018, 4, eaat9180.

[7] A. Schoedel, M. Li, D. Li, M. O’Keeffe, O. M. Yaghi, Chem. Rev.

2016, 116, 12466.

[8] S. M.Cohen, J. Am. Chem. Soc. 2017, 139, 2855.

[9] Z.Wang, S. M.Cohen, J. Am. Chem. Soc. 2007, 129, 12368.

[10] K. K.Tanabe, S. M.Cohen, Chem. Soc. Rev. 2011, 40, 498.

[11] M. T. Kapelewski, T. Runčevski, J. D. Tarver, H. Z. H. Jiang,

K. E. Hurst, P. A. Parilla, A. Ayala, T. Gennett, S. A. Fitzgerald,

C. M.Brown, J. R.Long, Chem. Mater. 2018, 30, 8179.

[12] M.Ding, R. W.Flaig, H. L.Jiang, O. M.Yaghi, Chem. Soc. Rev. 2019,

48, 2783.

[13] P.Horcajada, R.Gref, T.Baati, P. K.Allan, G.Maurin, P.Couvreur,

G.Férey, R. E.Morris, C.Serre, Chem. Rev. 2012, 112, 1232.

[14] K. J.Hartlieb, D. P. Ferris, J. M.Holcroft, I.Kandela, C. L.Stern,

M. S.Nassar, Y. Y.Botros, J. F.Stoddart, Mol. Pharmaceutics 2017,

14, 1831.

[15] S.Rojas, T.Baati, L.Njim, L.Manchego, F.Neffati, N. Abdeljelil,

S.Saguem, C.Serre, M. F.Najjar, A.Zakhama, P.Horcajada, J. Am.

Chem. Soc. 2018, 140, 9581.

[16] M.Ranocchiari, J. A.Van Bokhoven, Phys. Chem. Chem. Phys. 2011,

13, 6388.

[17] A.Dhakshinamoorthy, Z.Li, H.Garcia, Chem. Soc. Rev. 2018, 47, 8134.

[18] E. Coronado, G. Mínguez Espallargas, Chem. Soc. Rev. 2013, 42,

1525.

[19] G. Mínguez Espallargas, E. Coronado, Chem. Soc. Rev. 2018,

47, 533.

[20] E.Coronado, M.Giménez-Marqués, G. M.Espallargas, L.Brammer,

Nat. Commun. 2012, 3, 828.

[21] C. Yang, R. Dong, M. Wang, P. S. Petkov, Z. Zhang, M. Wang,

P. Han, M. Ballabio, S. A. Bräuninger, Z. Liao, J. Zhang,

F. Schwotzer, E. Zschech, H. H. Klauss, E. Cánovas, S. Kaskel,

M. Bonn, S. Zhou, T. Heine, X. Feng, Nat. Commun. 2019, 10,

3260.

[22] L.Sun, M. G.Campbell, M.Dincaˇ, Angew. Chem., Int. Ed. 2016, 55,

3566.

[23] D. Sheberla, J. C. Bachman, J. S. Elias, C. J. Sun, Y. Shao-Horn,

M.Dincaˇ, Nat. Mater. 2017, 16, 220.

[24] S. I.Ohkoshi, H.Tokoro, Acc. Chem. Res. 2012, 45, 1749.

[25] W.Ouellette, A. V.Prosvirin, K.Whitenack, K. R.Dunbar, J.Zubieta,

Angew. Chem., Int. Ed. 2009, 48, 2140.

[26] T. B.Faust, D. M.D’Alessandro, RSC Adv. 2017, 4, 17498.

Figure 6. Spin density isosurface of a portion of the IBU. β/α spin density is shown in blue/red and corresponds to a difference between the spin-up

and spin-down density of 0.005 electrons per Å3. Color definitions: O—red; N—orange; Cu—cyan; C—black; P—blue; H—white.

www.advmat.dewww.advancedsciencenews.com

Adv. Mater. 2020, 32, 2000474

[27] D.Maspoch, D.Ruiz-Molina, K.Wurst, N.Domingo, M.Cavallini,

F. Biscarini, J. Tejada, C. Rovira, J. Veciana, Nat. Mater. 2003,

2, 190.

[28] O. K. Farha, I. Eryazici, N. C. Jeong, B. G. Hauser, C. E. Wilmer,

A. A.Sarjeant, R. Q.Snurr, S. T.Nguyen, A. Ö.Yazaydin, J. T.Hupp,

J. Am. Chem. Soc. 2012, 134, 15016.

[29] G. Yücesan, V. Golub, C. J. O’Connor, J. Zubieta, Solid State Sci.

2005, 7, 133.

[30] M. A. AlDamen, S. Cardona-Serra, J. M. Clemente-Juan,

E.Coronado, A.Gaita-Ariño, C.Martí-Gastaldo, F.Luis, O.Montero,

Inorg. Chem. 2009, 48, 3467.

[31] D. T.Tran, X.Fan, D. P.Brennan, P. Y.Zavalij, S. R. J.Oliver, Inorg.

Chem. 2005, 44, 6192.

[32] K. J.Gagnon, H. P.Perry, A.Clearfield, Chem. Rev. 2012, 112, 1034.

[33] M.Taddei, F.Costantino, R.Vivani, Eur. J. Inorg. Chem. 2016, 27, 4300.

[34] G.Yücesan, Y.Zorlu, M.Stricker, J.Beckmann, Coord. Chem. Rev.

2018, 369, 105.

[35] Y.Zorlu, D.Erbahar, A.Çetinkaya, A.Bulut, T. S.Erkal, A. O.Yazaydin,

J.Beckmann, G.Yücesan, Chem. Commun. 2019, 55, 3053.

[36] S. J. I. Shearan, N. Stock, F. Emmerling, J. Demel, P. A. Wright,

K. D. Demadis, M. Vassaki, F. Costantino, R. Vivani, S. Sallard,

I. R.Salcedo, A.Cabeza, M.Taddei, Crystals 2019, 9, 270.

[37] R. A.Coxall, S. G.Harris, D. K.Henderson, S.Parsons, P. A.Tasker,

R. E. P.Winpenny, J. Chem. Soc., Dalton Trans. 2000, 14, 2349.

[38] A. Bulut, Y. Zorlu, M. Wörle, A. Çetinkaya, H. Kurt, B. Tam,

A. Ö.Yazaydın, J.Beckmann, G.Yücesan, ChemistrySelect 2017, 2, 7050.

[39] A. Bulut, M. Maares, K. Atak, Y. Zorlu, B. Çoşut, J. Zubieta,

J. Beckmann, H. Haase, G. Yücesan, CrystEngComm 2018, 20,

2152.

[40] A. R.Zanatta, I.Chambouleyron, Phys. Rev. B 1996, 53, 3833.

[41] J.Tauc, R.Grigorovici, A.Vancu, Phys. Status Solidi B 1966, 15, 627.

[42] W. E.Hatfield, R. R.Weller, J. W.Hall, Inorg. Chem. 1980, 19, 3825.

[43] J. C.Bonner, M. E.Fisher, Phys. Rev. 1964, 135, A640.

[44] D. C. Johnston, in Handbook of Magnetic Materials, Vol. 10

(Ed: K. H. J. Buschow), Elsevier, Amsterdam, The Netherlands 1997.

[45] E. S.Klyushina, A. T. M. N.Islam, J. T. Park, E. A.Goremychkin,

E.Wheeler, B.Klemke, B.Lake, Phys. Rev. B 2018, 98, 104413.

[46] S.Eggert, I.Affleck, M.Takahashi, Phys. Rev. Lett. 1994, 73, 332.