Network-forming liquids from metal-bis(acetamide) frameworks with low melting temperatures

Article abstract of DOI:10.1021/jacs.0c11718

Molten phases of metal-organic networks offer exciting opportunities for using coordination chemistry principles to access liquids and glasses with unique and tunable structures and properties. Here, we discuss general thermodynamic strategies to provide an increased enthalpic and entropic driving force for reversible, low-temperature melting transitions in extended coordination solids and illustrate this approach through a systematic study of a series of bis(acetamide)-based networks with record-low melting temperatures. The low melting temperatures of these compounds are the result of weak coordination bonds, conformationally flexible bridging ligands, and weak electrostatic interactions between spatially separated cations and anions, which collectively reduce the enthalpy and increase the entropy of fusion. Through a combination of crystallography, spectroscopy, and calorimetry, enthalpic trends are found to be dictated by the strength of coordination bonds and hydrogen bonds within each compound, while entropic trends are strongly influenced by the degree to which residual motion and positional disorder are restricted in the crystalline state. Extended X-ray absorption fine structure (EXAFS) and pair distribution function (PDF) analysis of Co(bba)₃[CoCl₄] [bba = N,N′-1,4-butylenebis(acetamide)], which features a record-low melting temperature for a three-dimensional metal-organic network of 124 °C, provide direct evidence of metal-ligand coordination in the liquid phase, as well as intermediate- and extended-range order that support its network-forming nature. In addition, rheological measurements are used to rationalize differences in glass-forming ability and relaxation dynamics. These results provide new insights into the structural and chemical factors that influence the thermodynamics of melting transitions of extended coordination solids, as well as the structure and properties of coordination network-forming liquids.

Full text of DOI:10.1021/jacs.0c11718

pubs.acs.org/JACS
Article
Network-Forming Liquids from Metal−Bis(acetamide) Frameworks
with Low Melting Temperatures
Mengtan Liu, Ryan D. McGillicuddy, Hung Vuong, Songsheng Tao, Adam H. Slavney,
Miguel I. Gonzalez, Simon J. L. Billinge, and Jarad A. Mason*
Cite This: J. Am. Chem. Soc. 2021, 143, 2801−2811
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Molten phases of metal−organic networks offer
exciting opportunities for using coordination chemistry principles to
access liquids and glasses with unique and tunable structures and
properties. Here, we discuss general thermodynamic strategies to
provide an increased enthalpic and entropic driving force for
reversible, low-temperature melting transitions in extended coordi-
nation solids and illustrate this approach through a systematic study
of a series of bis(acetamide)-based networks with record-low melting
temperatures. The low melting temperatures of these compounds are
the result of weak coordination bonds, conformationally flexible
bridging ligands, and weak electrostatic interactions between spatially
separated cations and anions, which collectively reduce the enthalpy
and increase the entropy of fusion. Through a combination of
crystallography, spectroscopy, and calorimetry, enthalpic trends are found to be dictated by the strength of coordination bonds and
hydrogen bonds within each compound, while entropic trends are strongly influenced by the degree to which residual motion and
positional disorder are restricted in the crystalline state. Extended X-ray absorption fine structure (EXAFS) and pair distribution
function (PDF) analysis of Co(bba)3[CoCl4] [bba = N,N′-1,4-butylenebis(acetamide)], which features a record-low melting
temperature for a three-dimensional metal−organic network of 124 °C, provide direct evidence of metal−ligand coordination in the
liquid phase, as well as intermediate- and extended-range order that support its network-forming nature. In addition, rheological
measurements are used to rationalize differences in glass-forming ability and relaxation dynamics. These results provide new insights
into the structural and chemical factors that influence the thermodynamics of melting transitions of extended coordination solids, as
well as the structure and properties of coordination network-forming liquids.
dimensional coordination networks that have been shown to
melt into liquids are predominantly variants of zeolitic
imidazolate frameworks (ZIFs).4a,6a,c,d,e,j Although ZIFs have
provided an initial platform to begin exploring the structures
and properties of liquid phases of metal−organic frameworks,
their melting transitions typically occur at temperatures above
350−400 °Coften preceded by crystalline-to-amorphous
transitions that make it challenging to establish relationships
between structure and melting thermodynamicsand produce
liquid phases that can be prone to decomposition. To realize
the full potential of coordination network-forming liquids,7
new strategies are needed to design metal−organic networks
with lower melting temperatures that yield liquids with
INTRODUCTION
■
Network-forming liquids feature extended structural correla-
tions that influence many important macroscopic properties
including viscosity, conductivity, and densityand can lead to
unique phase behavior.1,2 In metal−organic framework
chemistry, the extended structureand free volumeof solids
is routinely manipulated using coordination chemistry and
reticular design principles to form networks of metals linked by
bridging organic ligands.3 Though in its infancy, the
application of similar ideas to liquids offers exciting possibilities
for expanding the structural and chemical diversity of network-
forming liquids by providing access to new topologies and
longer-range structural correlations mediated by the increased
length of organic bridging ligands relative to the bridging
atomic units of conventional network-forming materials such
as SiO2 and ZnCl2.4 However, even though tens of thousands
of metal−organic networks have been reported in the
literature,5 there are surprisingly few examples of such
compounds that have accessible and stable liquid states
(Table S12).4,6 Indeed, the limited examples of three-
Received: November 8, 2020
Published: February 11, 2021
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811
© 2021 American Chemical Society
2801

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
increased stability and a greater range of structures and
compositions. Beyond harnessing the intrinsic chemical and
physical properties of coordination network-forming liquids to
establish new structure−property relationships and realize new
functionalities, expanded access to molten phases of metal−
organic frameworks would also afford new opportunities for
materials processing,8 growth of large single crystals,9 and glass
formation.10
melting transitions. Melting transitions are governed by the
difference in enthalpy (ΔHfus) and entropy (ΔSfus) between
solid and liquid phases, with the transition temperature given
by ΔHfus = TmΔSfus. The promotion of low melting
temperatures thus requires minimizing ΔHfus, maximizing
ΔSfus, or some combination of both.
Minimizing ΔHfus entails balancing the strength of cohesive
interactions within the solid and liquid phases. In most metal−
organic materials, such interactions involve a convolution of
attractive and repulsive forces, including those mediated by
coordination bonds, hydrogen bonds, van der Waal inter-
actions, Coulombic interactions, and steric repulsions. For a
three-dimensional metal−organic network to transition to a
liquid, the network of coordination bonds holding the
compound together must dissociate to at least some degree
to enable the inherent microscopic fluctuations that give rise to
the macroscopic fluidity of liquids. Indeed, simulations have
indicated that tetrahedral Zn centers in Zn(imidazolate)2 (ZIF-
4) transition to an average coordination number of ∼3.3−3.6
in the liquid phase.4a,c Bridging ligands bearing coordination
groups that form relatively weak bonds with metalssuch as
sulfonates, esters, amides, and nitrilesshould facilitate the
partial breaking of coordination bonds during melting and thus
promote a lower enthalpy changeand lower melting
temperaturethan the carboxylate- and azolate-based ligands
that are typically used to form metal−organic frameworks with
high thermal stability.6f,g,i,k In addition, ΔHfus can be further
reduced by minimizing electrostatic attractions through charge
delocalization and the spatial separation of anions and
cationsa strategy that is routinely applied in the design of
ionic liquids.11
The existence of matter in the liquid state requires that
attractive interactions between constituent atoms, ions,
molecules, or other chemical species are sufficiently counter-
acted by repulsive interactionsimparting fluidity while still
maintaining cohesion. In principle, every metal−organic
network has an equilibrium melting temperature, Tm. In
practice, however, the decomposition temperature, Tdec, of
many compounds falls below Tm. Therefore, the primary
challenge in the synthesis of coordination network-forming
liquids is to design systems for which Tm < Tdec.
6c To date, the
limited efforts to explore melting transitions in two- and three-
dimensional metal−organic networks have focused on
increasing Tdec through the use of ligandssuch as
imidazoleswith high thermal stabilities, while efforts to
design compounds with intrinsically lower Tm have received
comparatively little attention.6c,e−g,j Herein, we establish
generalizable thermodynamic strategies to lower the melting
temperature of three-dimensional metal−organic networks
through the synthesis and characterization of a series of
bis(acetamide)-based compounds that feature stable liquid
phases (Figure 1).
Maximizing ΔSfus requires minimizing solid-state entropy
and maximizing liquid-state entropy. In most cases, the entropy
of metal−organic materials can be partitioned into contribu-
tions from configurational, vibrational, and rotational entropy
for solid phases, with an added contribution from translational
entropy for liquid phases.6e,f,12 Although much remains to be
understood about relationships between the structure,
composition, and entropy of metal−organic networks, their
configurational, vibrational, and rotational entropy should, in
principle, be conducive to predictable manipulation through
ligand design. For instance, low-symmetry, high-flexibility
organic bridging ligands will access more conformations in
the liquid phase relative to the solid,13 and consequently,
compounds constructed from such ligands should experience a
greater increase in entropy upon melting. Similarly, extended
structures that better restrict the residual motion of organic
ligands in the solid state should undergo higher entropy
increases upon melting than structures for which ligands have
more rotational and vibrational degrees of freedom in the solid
state.
Toward the goal of minimizing ΔHfus and maximizing ΔSfus,
we targeted extended networks composed of flexible bridging
ligands with weak, neutral donor groups and noncoordinating
anions to provide charge balance (Figure 1). Although there
are limited examples of three-dimensional coordination
networks formed exclusively from neutral bridging ligands,
select polymethylene bis(amides) are known to form
crystalline two- and three-dimensional extended solids with
several first-row transition metals (Table S13),14 a few of
which even have reported visual melting points.14b We
hypothesized that metal−amide bonds would provide an
interaction weak enough to facilitate melting, but strong
Figure 1. Illustation of design principles to promote the formation of
coordination network-forming liquids from extended metal−organic
materials by minimizing the enthalpy of fusion, ΔHfus, and maximizing
the entropy of fusion, ΔSfus, to achieve low melting temperatures, Tm.
Our approach draws from the well-established “crystal
engineering” design rules of metal−organic framework
chemistry to control structure through directional coordination
bonds and from the “anti-crystal engineering” design rules of
ionic liquid synthesis to promote low melting temper-
atures.3,6f,11 Specifically, we targeted organic ligands predis-
posed to (1) bridge multiple metals and (2) provide an
enthalpic and entropic driving force for low-temperature
2802
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
enough to dictate the orientation and arrangement of bridging
ligands through directional interactions.15 As such, we
anticipated that bis(amide)-based coordination networks
would provide a highly tunable platform to access metal−
organic networks with low melting temperatures and to explore
the thermodynamics, structures, and properties of new classes
of network−forming liquids.
conformational flexibility of hmba and the increased rotation
and mobility of uncoordinated anions in the liquid phase. To
better understand the relative contributions of different
structural and chemical factors to melting thermodynamics,
we endeavored to synthesize and compare the thermal
behavior of a series of metal−organic networks with the
same coordination bonding motifs but different metals,
bridging organic ligands, and uncoordinated anions. Because
of the limited variability and large size of [M(NCS)4]2− anions,
we specifically targeted compounds with similar topologies but
more compact and tunable [MX4]2− (X = Cl, Br) anions.
Although combining metal halides with hmba tends to yield
two-dimensional networks,14a we found that shorter bis-
(acetamide) ligands, which template smaller voids that better
accommodate [MX4]2− anions, promote the formation of
three-dimensional structures. Specifically, combining N,N′-1,4-
butylenebis(acetamide) (bba) with divalent metal chloride
salts led to the new series of three-dimensional networks
M(bba)3[M′Cl4] (M/M′ = Mn, Fe, Co; M = Mn, M′ = Zn;
and M = Mg, M′ = Co, Zn). Single-crystal X-ray diffraction
(SCXRD) revealed that all compounds crystallize in the same
RESULTS AND DISCUSSION
■
Thermal Characterization of Three-Dimensional
Metal−Bis(acetamide) Networks. To begin investigating
melting phenomena in metal−bis(amide) networks, we
synthesized the compound Co(hmba)3[Co(NCS)4] [hmba =
N,N′-1,6-hexamethylenebis(acetamide)] and characterized its
thermal behavior by differential scanning calorimetry (DSC).
In this crystalline extended solid, octahedral Co centers are
bridged through the carbonyl O atoms of hmba ligandswith
two pairs of equatorial ligands connecting to two neighboring
Co centers to form chains that are further connected by axial
hmba ligandsto yield a three-dimensional, cationic net-
work.14b Charge balance is provided by [Co(NCS)4]2− anions
that fill the network voids (Figure S8). Excitingly, DSC
revealed a fully reversible melting transition at 144 °C with a
ΔHfus of 108 kJ/mol (Figure 2). While thermal decomposition,
̅
space group (Pa3) with the same three-dimensional con-
nectivity, wherein octahedral metal centers are each connected
to six other metal centers by bridging bba ligands to form a
three-dimensional cubic network that surrounds tetrahedral
uncoordinated anions (Figure 3a). Moreover, all six com-
pounds undergo reversible melting transitions that span from
124 °C for Co(bba)3[CoCl4] to 262 °C for Mg(bba)3[ZnCl4]
(Figure 3c, Figure S19). To the best of our knowledge, the
melting temperature of Co(bba)3[CoCl4] is the lowest yet
reported for a three-dimensional coordination network (Table
S12).
Prior to melting, TGA with a temperature ramp rate of 2
°C/min indicates minimal mass loss (<1.5 wt %) for all
compounds that have melting temperatures below 200 °C
(Figures S28 and S31). With a melting temperature of 260 °C,
the larger mass loss of Mg(bba)3[CoCl4] prior to melting
(17% at 2 °C/min and 7% at 10 °C/min, Figure S29)as well
as mass loss from the liquid state of other compoundsis
attributed to volatilization of the neutral bridging ligand under
flowing N2 gas. Despite their volatility in an open system, the
liquid phases of each compound exhibit high thermal stability
in a closed system as evidenced by DSC cycling experiments
(Figures S19, S30, and S31).
Trends in Melting Thermodynamics. The M-
(bba)3[M′Cl4] series of compounds illustrate how structure
impacts the interplay between enthalpy and entropy that
ultimately dictates the melting temperature of metal−organic
networks. Within this series, Co(bba)3[CoCl4] has the lowest
melting temperature of 124 °C, while Mn(bba)3[MnCl4] and
Mg(bba)3[CoCl4] have much higher melting temperatures of
172 and 260 °C, respectively. As evidenced by the nearly
identical melting temperatures of Mg(bba)3[CoCl4] and
Mg(bba)3[ZnCl4] (Figure S19), differences in melting
behavior seem to be governed primarily by network metal
atoms, while the identity of the metal in the charge-balancing
anion has minimal influence. This is confirmed by variable
temperature extended X-ray absorption fine structure
(EXAFS) spectra of Mg(bba)3[CoCl4] at the Co K edge,
which show minimal structural changes to [CoCl4]2− anions
upon melting (Figure S33).
Figure 2. Differential scanning calorimetry (DSC) traces for
Co(hmba)3[Co(NCS)4] with heating and cooling rates of 5 °C/
min. The dark purple trace corresponds to the first heating−cooling
cycle, while the two overlapping light purple traces correspond to
subsequent cycles under the same conditions. An image of the liquid
phase inside a glass tube is shown at the bottom right. Inset: An
isothermal hold at 120 °C led to complete recrystallization.
phase separation, large changes in porosity, or high viscosity
preclude recrystallization for all molten three-dimensional
metal−organic networks reported to date,6a,c,d,f Co-
(hmba)3[Co(NCS)4] features high thermal stability in the
liquid state (Figure S23) and can be fully recrystallized upon
holding the supercooled melt at 120 °C for less than 1 h
(Figure 2).
The low melting temperature of Co(hmba)3[Co(NCS)4]
can be attributed, in part, to a large ΔSfus of 260 J/mol·K. In
addition to contributions from an anticipated decrease in the
average coordination number of network Co atoms, this large
entropy change during melting likely arises from the high
Surprisingly, the driving force for the higher melting
temperature of Mn(bba)3[MnCl4] relative to the Co and Fe
2803
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. (a) Crystal structure of the three-dimensional network Co(bba)3[CoCl4]. Note that the bba ligand is disordered over two positions, and
only the higher occupancy atomic positions are shown. Purple, red, gray, blue, and green spheres represent Co, O, C, N, and Cl atoms, respectively;
H atoms have been omitted for clarity. (b) The geometry of hydrogens bonds to [CoCl4]2− anions is compared for Co(bba)3[CoCl4] and
Mg(bba)3[CoCl4] at 100 K to highlight the additional set of moderate hydrogen-bonding interactions present in the former compound. Atomic
displacement parameters are shown at 50% probability for [CoCl4]2− counteranions. Teal spheres represent Mg atoms, and only N−H protons are
shown. For Co(bba)3[CoCl4], note that the lower occupancy position of the bba ligand is shown as this has slightly shorter hydrogen bond
distances. (c) DSC traces for M(bba)3[M′Cl4] compounds (M/M′ = Co, Fe, Mn; and M = Mg, M′ = Co) with a heating rate of 5 °C/min for all
compounds. (d) Top: Trends in ΔHfus and ΔSfus are illustrated for Co(bba)3[CoCl4], Fe(bba)3[FeCl4], Mn(bba)3[MnCl4], and Mg(bba)3[CoCl4].
Error bars reflect a 2% instrumental error. Bottom: The equivalent isotropic displacement parameters, Uequiv, of Cl2 atoms and the N1−Cl1
hydrogen bond donor−acceptor distances at 100 K are generally correlated to one another and inversely correlated to ΔSfus. Note that donor−
acceptor distances were weighted by relative occupancy to account for two-part disorder when present.
analogues is primarily entropic in nature, while that for the Mg
analogue is both enthalpic and entropic. More specifically, the
higher melting temperature of Mn(bba)3[MnCl4], which has a
similar ΔHfus to the Co and Fe analogues, is the result of a
lower ΔSfus, while the even higher Tm of the Mg analogue is
due to both a lower ΔSfus and higher ΔHfus (Figure 3d). These
thermodynamic differences reveal an absence of enthalpy−
entropy compensationwhere changes to ΔH are offset by
proportionate changes to ΔSand indicate that independent
molecular mechanisms contribute to the enthalpy and entropy
of fusion in this system.13b,16
Enthalpic trends can be rationalized, at least in part, on the
basis of M−O coordination bond strength, which is expected
to follow the trend Mn−O < Fe−O < Co−O < Mg−O.17
Stronger Mg−O bonds likely lead to the larger ΔHfus of
Mg(bba)3[CoCl4] through two mechanisms: (1) more energy
will be required to partially break Mg−bba bonds and reduce
the average metal coordination number during melting and (2)
more energy will be required to partially break hydrogen bonds
between amides and uncoordinated anions when the amide
carbonyl is more polarized through a stronger interaction with
Mg centers. The similar ΔHfus of the Mn, Fe, and Co
analogues can be attributed to small differences in metal−
ligand bond strength that may be offset by differences in the
hydrogen bond acceptor strength of [MCl4]2− anions or by
slight differences in the average metal coordination number
after melting.18
Unlike ΔHfus, M−O bond strength does not have a clear
influence on entropic trends within this series of compounds.
Specifically, while Co(bba)3[CoCl4] and Fe(bba)3[FeCl4]
should have M−O bonds of intermediate strength,17 their
entropy changes are considerably higher than both the Mn and
Mg analogues. Explaining this surprising result in a manner
that is consistent with trends in enthalpy is not trivial, as most
molecular mechanisms would predict at least some degree of
enthalpy−entropy compensation. That is, the breaking of
stronger interactions would typically lead to larger increases in
both enthalpy and entropy of fusion within a given system.19 In
this case, the high ΔSfus of the Co and Fe analoguesand,
consequently, their low melting temperaturescan be
rationalized, at least in part, through subtle structural
differences between these compounds and their Mg and Mn
congeners. Although each network crystallizes with the same
three-dimensional connectivity and local coordination environ-
ment around metal centers, a ∼170° difference in the N1−
C3−C4−C5 dihedral angle of bba bridging ligands in
Co(bba)3[CoCl4] and Fe(bba)3[FeCl4] relative to the Mg
and Mn analogues leads to a different arrangement of
hydrogen bonds between the N−H protons of bba and the
Cl atoms of uncoordinated anions (Figure S3). Specifically,
each network contains two crystallographically independent Cl
sites, with each [MCl4]2− anion centered on a 3-fold rotation
axis that lead to three Cl1 atoms and one Cl2 atom (Figure
3b). In Fe(bba)3[FeCl4] and Co(bba)3[CoCl4], each Cl1 is
2804
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
engaged in two hydrogen-bonding interactions of moderate
strength with donor−acceptor distances of less than 3.5 Å and
bond angles greater than 120°.20 In contrast, each Cl1 in the
Mg and Mn analogues forms one moderate hydrogen bond
(3.3 Å donor−acceptor distance) and one very weak hydrogen
bond (≥3.7 Å donor−acceptor distance) (Table S7). This is
corroborated by IR data, where the N−H stretch (∼3290
cm−1) of Co(bba)3[CoCl4] and Fe(bba)3[FeCl4] are both red-
shifted by ∼5 cm−1 compared to Mg(bba)3[CoCl4] and
Mn(bba)3[MnCl4] (Figure S47).
We hypothesized that the additional set of moderate
hydrogen bonds to Cl1 atoms in Co and Fe analogues reduces
the entropy of the solid phaseand thus increases the entropy
change upon meltingthrough two primary effects: (1)
decreased residual motion of the counteranions and ligands,
which reduces vibrational and rotational entropy, and (2) less
disorder in the equilibrium positions of counteranion atoms,
which reduces configurational entropy.12,21 Since the equiv-
Two-Dimensional Metal−Bis(acetamide) Networks.
Beyond metal cation identity, the incorporation of different
anions into metal−bis(acetamide) networks offers additional
insight into the structural and chemical factors that contribute
to their melting thermodynamics. For instance, combining
MgCl2 with bba under suitable conditions affords the
compound Mg(bba)3Cl2, which has uncoordinated Cl− instead
of [MCl4]2− anions and adopts a two-dimensional hexagonal
network with Mg centers in a similar coordination environ-
ment as in the three-dimensional Mg(bba)3[CoCl4] (Figure
4). Notably, this compound has a 70% higher ΔHfus relative to
alent isotropic atomic displacement parameters (Uequiv
)
obtained from crystal structure refinement describe the
deviation of an atom’s electron density from its average
position, accounting for both local motion of atoms (dynamic
disorder) and small deviations in their equilibrium positions
(static disorder),22 we anticipated that Uequiv might serve as an
indicator of the relative entropy of the anions within each
compound. Indeed, we found that Uequiv values for Cl2 atoms,
which do not directly engage in moderate-strength hydrogen
bonding interactions and thus have greater inherent mobility,
were substantially lower in compounds with higher ΔSfus
(Figure 3d). Specifically, Uequiv at both 100 and 298 K is
lower for Cl2 in Co(bba)3[CoCl4] and Fe(bba)3[FeCl4],
which have higher ΔSfus, than in the Mn and Mg analogues,
which have lower ΔSfus (Figure 3d, Figure S6). Assuming that
all [MCl4]2− anions have similar rotational, vibrational, and
translational degrees of freedom in the liquid state, this offers a
compelling potential explanation for the high ΔSfus of the Co
and Fe analogues, whereby the lower solid-state entropy
conferred by more constrained [MCl4]2− anions drives a larger
entropy increase upon conversion to a liquid. Although the
density and strength of hydrogen bonds to [MCl4]2− anions
could impact ΔHfus in addition to ΔSfus, the weak nature of
these interactions likely reduces the magnitude of enthalpic
effects arising from differences in hydrogen bonds relative to
differences in coordination bonds, particularly if hydrogen
bonds to [MCl4]2− anions are largely preserved in the melt.
In addition to the residual motion and positional disorder of
the [MCl4]2− anions, other factors may contribute to the lack
of enthalpy−entropy compensationand, consequently, the
large range of melting temperaturesexhibited by the
M(bba)3[M′Cl4] series of compounds. The high Tm of
Mg(bba)3[CoCl4], for instance, is the result of this compound
having both the highest ΔHfus and lowest ΔSfus of the series.
The high ΔHfus can be attributed to stronger Mg−O bonds,
which need to be at least partially broken in order for melting
to occur. These stronger bonds also appear to restrict the
residual motion of the bba ligand to a greater extent at 298 K
than in other analogues as evidenced by the smaller average
thermal displacement parameters for the bba ligand atoms of
Mg(bba)3[CoCl4] (Figure S7). This does not, however,
provide a large enough entropic effect to compensate for the
higher ΔHfus, likely because differences in the residual motion
of bba ligands are relatively small and decrease with increasing
temperature.
Figure 4. Crystal structure of Mg(bba)3Cl2 highlighting the (a) two-
dimensional coordination network and (b) hydrogen bonds to Cl−
anions (dashed lines). Teal, red, gray, blue, green, and white spheres
represent Mg, O, C, N, Cl, and H atoms, respectively. Only N−H
protons are shown for clarity. (c) DSC traces for Mg(bba)3Cl2 (blue)
and Co(bba)3Br2 (purple) with a heating rate of 5 °C/min.
Mg(bba)3[CoCl4], increasing from 88 to 141 kJ/mol (Figure
5). This increase can be explained by stronger hydrogen bonds
Figure 5. Comparison of ΔHfus, ΔSfus, and Tm for select metal−
bis(acetamide) frameworks. The symbol shape identifies the series of
compounds, while the symbol color designates the metal, and the
border color designates the anion.
2805
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
between bba ligands and the more charge-dense Cl−acceptors.
Despite this much larger enthalpy change, the melting
temperature of Mg(bba)3Cl2 is actually 40 °C below that of
Mg(bba)3[CoCl4] (Figure 4c), indicating that, unlike within
the M(bba)3[M′Cl4] compounds, the increase in ΔHfus is fully
compensated for by an increase in ΔSfus (Figure 5).
ing porosity, might provide an unexplored pathway to lowering
the melting temperature of metal−organic frameworks.
Network-Forming Nature of the Liquid State. To
probe the local structure of the liquid phase of metal−
bis(acetamide) networks and to determine if any intermediate-
or extended-range order is present, we collected variable-
temperature X-ray total scattering data for Co(bba)3[CoCl4]
(Figure S36). The raw data images were integrated, corrected
for experimental effects, and normalized using PDFgetX3 to
obtain total scattering reduced structure functions, F(Q),
(Figure S37), which were Fourier transformed to obtain pair
distribution functions (PDFs), G(r) (Figure 6).26 At short
The large increase in entropy upon melting for Mg(bba)3Cl2
can be at least partially attributed to reduced entropy in the
solid phase, as strong hydrogen bonds restrict the conforma-
tional flexibility of bba bridging ligands and reduce their
residual motion and positional disorder within the crystal to a
greater extent than strong M−O bonds. Indeed, while the bba
ligand is disordered over two positions in M(bba)3[M′Cl4],
there is no disorder in the crystal structure of Mg(bba)3Cl2. In
addition, Uequiv values for the O, C, and N atoms of bba ligands
engaged in hydrogen bonds to Cl anions decrease by more
than 40% compared to those of ligands hydrogen bonded to
[MX4]2− anions, which is consistent with reduced residual
motion and positional disorder in the presence of stronger
hydrogen bonds (Figure S7). The compound Co(bba)3Br2,
which adopts a similar two-dimensional structure,14a also
experiences an increase in ΔHfus and ΔSfus along with lower
Uequiv values. Unlike the reversible melting transition of
Mg(bba)3Cl2, however, Co(bba)3Br2 melts irreversibly. Vari-
able temperature EXAFS spectra (Figure S33) reveal that this
irreversibility is due to Br− anions binding to Co centers in the
liquid phase after melting, and powder X-ray diffraction
suggests the partial formation of a new phase upon cooling,
which appears to adopt the same network structure as
Co(bba)3[CoCl4] (Figure S13).
Figure 6. Pair distribution functions of Co(bba)3[CoCl4] in the
crystalline solid phase just below the melting temperature (black) and
in the liquid phase above the melting temperature (red).
Quasiperiodic oscillations indicative of extended-range order are
shown in the inset.
In addition to reducing solid-state entropy by restricting
residual motion, halide anions should also affect the liquid-
state entropy of Mg(bba)3Cl2 differently from tetrahalome-
tallate anions. Specifically, while the lower rotational symmetry
of tetrahalometallate anions will lead to a larger increase in
rotational entropy upon melting, this is estimated to contribute
just 6 J/mol·K to the overall entropy change.13a,23 On the
other hand, the smaller size of halide anions should result in
greater mobility in the liquid phase and a larger increase in
translational entropy upon melting,24 which will be amplified
by the fact that there are twice as many halide anions present
as tetrahalometallate anions. The combination of restricted
residual motion of bba in the solid state and increased
translational entropy of halide anions in the liquid state
tempered by the slight decrease in rotational entropy in the
liquid from the higher symmetry anioncan explain the
increased ΔSfus of Mg(bba)3Cl2 relative to Mg(bba)3[CoCl4].
While the thermodynamic implications of the different liquid
structure of Co(bba)3Br2 requires further study, the entropic
impact of the restricted residual motion of bba bridging ligands
in the two-dimensional M(bba)3X2 compounds highlights an
opportunity for expanding the diversity of metal−organic
networks with accessible melting transitions. Specifically,
porosity can impart organic ligands with a high degree of
residual motion in the solid state,25 which raises the entropy of
the solid and, consequently, reduces the entropy increase upon
melting. We speculate that this may be, in part, why many
metal−organic compounds have high melting points that
exceed thermal decomposition temperatures, which would be
consistent with a recent study reporting a decrease in the
solid−liquid transition temperature of amorphous ZIF-62 with
increasing pressure.6h As such, strategies to restrict the solid-
state motion of organic bridging ligands, without compromis-
scattering distances (0 < r < 4 Å), the liquid-phase PDF
remains largely unchanged from the crystalline phase, with
sharp features that can be assigned to C−C, C−N, and C−O
bonds within the bba ligands, Co−Cl bonds of the [CoCl4]2−
anions, and Co−O bonds of the coordination network
(Figures S42 and S43). This suggests that the average local
coordination environment around Co centers is not substan-
tially altered after melting.
Coordination of bba ligands in the melt was further
confirmed by variable temperature EXAFS spectra, which can
be well modeled with half of the Co centers in the melt
coordinated to 4 Cl atomsconsistent with the persistence of
[CoCl4]2− anions in the meltand half of the Co centers
coordinated to 4.8(7) O atoms of bba ligands (Figure S35).
The decrease in the coordination number of the network Co
centers by 1.2 (20%) upon melting is consistent with
molecular dynamics studies for ZIF-4 that suggest a decrease
in coordination number upon melting from 4 to 3.3 (18%) or
3.6 (10%)4a,c and is distinct from the ionic liquid-like molten
state of zinc−phosphate−azole coordination polymers, where-
in ligand coordination to Zn is not preserved upon melting.6b
Moreover, the average coordination number of 4.8(7) is well
above the bond percolation threshold of 2.4 for a three-
dimensional aperiodic network,27 which indicates that, in spite
of its dynamic nature, coordination between Co centers and
bridging bba ligands in the melt is persistent enough for there
to be pathways of extended connectivity that span the entire
liquid.
Beyond the local coordination environment, the PDF of
molten Co(bba)3[CoCl4] exhibits peak broadening and
decreased peak intensities that are consistent with the
formation of a liquid phase with no long-range periodicity
2806
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 7. (a) DSC trace for Co(hmba)3[CoBr4] with heating and cooling rates of 5 °C/min. A glass transition is observed at 23 °C during the
second heating run (light purple trace), which leads to the formation of a homogeneous, continuous glass film. (b) Temperature-dependent
viscosity measurements of Co(hmba)3[CoBr4], Mn(bba)3[MnCl4], and Co(hmba)3[Co(NCS)4] are shown as purple triangles, green squares, and
purple circles, respectively. The measurement temperatures are normalized to each respective melting temperature. The dashed black line
represents a VFT fit for Co(hmba)3[CoBr4], while solid black lines represent Arrhenius fits.
and substantially less intermediate-range order than the
crystalline solid phase. However, further PDF analysis reveals
intermediate- and extended-range order that resembles the
underlying network structure of the parent crystalline phase, as
is characteristic of conventional network-forming liquids.
Specifically, a first sharp diffraction peak (FSDP), which is a
common feature of network-forming liquids and glasses, is
observed in the total scattering reduced structure function,
F(Q), at 0.98 Å−1 (Figure S37). Though broadened and
weaker in intensity, the FSDP for the liquid phase is only
shifted slightly with respect to the FSDP of 0.91 Å−1 for the
high-temperature crystalline solid phase, which suggests a small
contraction upon melting and the persistence of some degree
of intermediate-range order. Moreover, broad peaks centered
at r = 8.4 and 14.3 Å in the liquid PDF represent intermediate-
range structural correlations over length scales that are similar
to the nearest-neighbor and second-nearest-neighbor Co−Co
distances in the crystalline solid (Figure S40, Table S25).
Importantly, close examination of the high-r region reveals
quasiperiodic oscillations that persist to at least 80 Å in the
liquid (Figure 6, inset). These oscillations indicate the
presence of extended-range structural correlations that exceed
the common correlation length scale of charge ordering in
ionic liquids and molten salts,28 and resemble the density
fluctuations that arise from topological and chemical ordering
in network-forming liquid and glassy materials such as a-Si and
ZnCl2.2f,29 While quasiperiodic oscillations have been observed
in many metal−organic glasses up to at least 30 Å, and may be
obscured at higher r by low signal to noise,6b,j,30 this
represents, to the best of our knowledge, the first observation
of extended-range order in a metal−organic liquid. We
anticipate that bridging ligand coordination and extended-
range structural correlations will also occur in the liquid phases
of other reversibly melting metal−bis(acetamide) networks,
but additional characterization is necessary to confirm this.
Recrystallization Kinetics and Glass Formation. To
further investigate the properties of liquid phases of metal−
bis(acetamide) networks, we evaluated their behavior upon
cooling through DSC experiments and variable temperature
viscosity measurements. Although melting transitions are fully
reversible for all but two compounds, liquid recrystallization
kinetics vary dramatically during cooling, ranging from rapid
crystallization to easily accessible glass transitions. For
instance, when cooled at 10 °C/min from 20 °C above their
melting temperature, three representative compoundsMn-
(bba)3[MnCl4], Co(hmba)3[Co(NCS)4], and Co-
(hmba)3[CoBr4]undergo rapid recrystallization, slow re-
crystallization, and a glass transition, respectively. More
specifically, molten Mn(bba)3[MnCl4] fully recrystallizes
even when cooled at high rates of 40 °C/min, while
Co(hmba)3[Co(NCS)4] only completely recrystallizes during
cooling when held 20 °C below Tm for 30−60 min and
otherwise undergoes a glass transition. In contrast, Co-
(hmba)3[CoBr4] does not recrystallize under any conditions
evaluated hereincluding after days at ambient temperature
and a homogeneous, transparent glass film is formed upon
cooling, with a glass transition temperature, Tg, of 23 °C
(Figure 7a). The larger Tg/Tm ratio of Co(hmba)3[CoBr4]
(Tg/Tm = 0.78) relative to Co(hmba)3[Co(NCS)4] (Tg/Tm =
0.70) is consistent with the higher glass-forming ability of the
former compound, while all other liquid phases with
measurable Tg in this study have Tg/Tm ratios in between
these values, which are slightly higher than the commonly
observed 2/3 ratio (Table S15).31
To better understand the different recrystallization kinetics
and glass-forming ability of these three metal−bis(acetamide)
liquids, we compared their viscosities just above their melting
temperatures. While all three liquids have comparable or
higher viscosity to many glass-forming ionic liquids (Table
S28), the viscosity of Co(hmba)3[CoBr4] is nearly an order of
magnitude higher than that of the other two liquids (Figure
7b). This higher viscosity leads to slower relaxation dynamics
that, in combination with the smaller amount of thermal
energy available for nucleation at its lower melting temper-
ature, could explain why Co(hmba)3[CoBr4] is the best glass
former.32 Since molten Mn(bba)3[MnCl4] and Co-
(hmba)3[Co(NCS)4] have very similar viscosities near their
respective melting temperatures, the faster recrystallization
kinetics of the former compound is likely the result of its
higher Tm making activation energy barriers for nucleation
easier to overcome because of increased thermal energy.
Further analysis of the temperature dependence of viscosity
provides additional insight into liquid dynamics. Just above
their respective melting temperatures (0.9 < Tm/T < 1), the
viscosity of each melt appears to follow Arrhenius behavior as
described by
2807
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
E
RT
largely dictated by metal−ligand bond strength and ΔSfus
dependent on the degree to which residual motion is restricted
in the solid state. This leads to unexpected situations in which
stronger interactions in the solid state can be associated with
either melting temperature increases or decreases depending
on the nature of entropic effects and highlights the important
role that entropy, in addition to enthalpy, plays in determining
the melting temperatures of metal−organic networks. Addi-
tionally, EXAFS and PDF analyses demonstrate coordination
between metal centers and bridging organic ligands in the
liquid phase of Co(bba)3[CoCl4] and extended-range struc-
tural correlations that persist to at least 80 Å, collectively
demonstrating the network-forming nature of this liquid.
As dynamic systems with short-range order, long-range
disorder, and a combination of solid-like and gas-like
behaviors, liquids have long provided important fundamental
insights relevant to many areas of science.36 Owing to their
tunable and predictable interactions, coordination network-
forming liquids provide a platform to establish new structure−
property relationships and realize functionalities not found in
conventional liquids. The relationships we have identified
between metal−organic network structure, composition, and
melting thermodynamics will support future efforts to
synthesize new classes of metal−organic frameworks that can
be melted into stable and tunable network-forming liquids. For
instance, we are applying similar strategies to achieve low
melting temperatures in frameworks with more rigid bridging
ligands that better fix the relative orientation and arrangement
of metal-binding groups and with noncoordinating anions
covalently bound to organic bridging ligands bearing neutral
donor groups. Both approaches provide opportunities to
achieve predictable control over the local structure of liquids
and, consequently, the spatiotemporal fluctuations that
determine the size, shape, and lifetime of transient voids
within them.
i
y
a
j
j
j
z
z
z
η(T) = A exp
k
{
where A is a temperature-independent constant and Ea is the
activation energy of viscous flow (the energy barrier each
component of the liquid needs to overcome to diffuse pass one
another).33 However, this Arrhenius dependence becomes
nonphysical when extrapolated over a broader temperature
range, significantly underestimating viscosities at both high and
low temperature limits (infinite temperature and near Tg,
respectively; Figure S45). This indicates Ea varies with
temperature, which could be associated with structural
heterogeneity and nonlinearity in relaxation dynamics.34
Though recrystallization of Mn(bba)3[MnCl4] and Co(hmba)-
[Co(NCS)4] occurs too quickly below Tm to allow for viscosity
measurements in the supercooled state, this non-Arrhenius
behavior can be experimentally confirmed in supercooled
Co(hmba)3[CoBr4] (Figure 7b). Over the entire temperature
range measured (0.9 < Tm/T < 1.08), the viscosity of
Co(hmba)3[CoBr4] can be well described (r2 > 0.99) by the
Vogel−Fulcher−Tammann (VFT) equation
B
log η(T) = log η∞ +
T − T
0
where η∞ is the infinite temperature asymptote, B is a
temperature-independent constant, and T0 is the hypothesized
ideal glass transition temperature at which the entropies of the
glass and the crystal become identical. The VFT model for
Co(hmba)3[CoBr4] predicts a Tg of 20 °C when extrapolated
to η = 1012 Pa·s (rheological definition of the glass transition),
which is in excellent agreement with DSC determined Tg of 21
°C upon cooling. However, the VFT model predicts a viscosity
divergence at T0, which has not been shown to occur
experimentally.35 As such, we also fit the viscosity data of
Co(hmba)3[CoBr4] with the Mauro−Yue−Ellison−Gupta−
Allan (MYEGA) equation
ASSOCIATED CONTENT
K
T
C
T
i
y
■
j
z
log η(T) = log η∞ +
expj
z
z
j
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.0c11718.
k
{
where K and C are temperature-independent constants. This
produces a similar high-quality fit (r2 > 0.99) as the VFT
equation but yields a slightly lower predicted Tg of 6 °C
(Figure S45). These differences are likely due to uncertainties
in extrapolating from a relatively narrow range of experimental
data, and viscosity data over a wider range of temperatures
should provide more insight into the dynamics of supercooled
metal−bis(acetamide) liquids.
Additional experimental details, powder X-ray diffraction
data, EXAFS data, total X-ray scattering data, pair
distribution function analysis, calorimetry data, rheo-
logical data analysis, and crystallographic information
(PDF)
Accession Codes
CONCLUSION
CCDC 2055036−2055050 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
■
The foregoing results demonstrate strategies to lower the
melting temperatures of two- and three-dimensional metal−
organic networks through a series of metal−bis(acetamide)
compounds that undergo reversible melting transitions from
106 to 262 °C. The low-melting temperatures of these
compounds are attributed to weak metal−amide coordination
bonds, conformationally flexible bridging ligands, and weak
electrostatic interactions. Moreover, we have elucidated the
specific structural and chemical features that drive differences
in melting thermodynamics within these series of compounds.
For instance, we have shown that entropy and enthalpy of
fusion are decoupled in these systems, with differences in
ΔHfus among isoreticular metal−bis(acetamide) networks
AUTHOR INFORMATION
■
Corresponding Author
Jarad A. Mason − Department of Chemistry & Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States; orcid.org/0000-0003-0328-7775;
Email: mason@chemistry.harvard.edu
2808
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(c) Wilson, M.; Salmon, S. P. Network Topology and the Fragility of
Tetrahedral Glass-Forming Liquids. Phys. Rev. Lett. 2009, 103,
157801.
Authors
Mengtan Liu − Department of Chemistry & Chemical Biology,
Harvard University, Cambridge, Massachusetts 02138,
United States
Ryan D. McGillicuddy − Department of Chemistry &
Chemical Biology, Harvard University, Cambridge,
Massachusetts 02138, United States
Hung Vuong − Department of Chemistry, Columbia
University, New York 10027, United States
Songsheng Tao − Department of Applied Physics and Applied
Mathematics, Columbia University, New York 10027, United
States
Adam H. Slavney − Department of Chemistry & Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States
Miguel I. Gonzalez − Department of Chemistry & Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States
Simon J. L. Billinge − Department of Applied Physics and
Applied Mathematics, Columbia University, New York
10027, United States; Condensed Matter Physics and
Materials Science Department, Brookhaven National
Laboratory, Upton, New York 11973, United States;
orcid.org/0000-0002-9734-4998
(2) (a) Turbull, D.; Bagley, B. G. Transitions in Viscous Liquids and
Glasses. In Changes of State; Hanny, N. B., Ed.; Springer: Boston,
1975; pp 513−554. (b) Roberts, C. J.; Debenedetti, P. G.
Polyamorphism and Density Anomalies in Network-Forming Fluids:
Zeroth- and First-Order Approximations. J. Chem. Phys. 1996, 105,
658−672. (c) Roberts, C. J.; Panagiotopoulos, A. Z.; Debenedetti, P.
G. Liquid−Liquid Immiscibility in Pure Fluids: Polymorphism in
Simulations of a Network-Forming Fluid. Phys. Rev. Lett. 1996, 77,
4386−4389. (d) Poole, P. H.; Grande, T.; Angell, C. A.; McMillan, P.
F. Polyamorphic Phase Transitions in Liquids and Glasses. Science
1997, 275, 322−323. (e) Yarger, J. L.; Angell, C. A.; Borick, S. S.;
Wolf, G. H. Polyamorphic Transitions in Network-Forming Liquids
and Glasses. In Supercooled Liquids; Fourkas, J. T., Kivelson, D.,
Mohanty, U., Nelson, K. A., Eds.; ACS Symposium Series 676;
American Chemical Society: Washington, DC, 1997; pp 214−223.
(f) Wilson, M. Structure and Dynamics in Network-Forming
Materials. J. Phys.: Condens. Matter 2016, 28, 503001−503031.
(3) (a) Yaghi, O. M.; O’Keeffe, M.; Ockwing, N. W.; Chae, H. K.;
Eddaoudi, M. Reticular Synthesis and the Design of New Materials.
Nature 2003, 423, 705−714. (b) Kitagawa, S.; Kitaura, R.; Noro, S.
Functional Porous Coordination Polymers. Angew. Chem., Int. Ed.
2004, 43, 2334−2375. (c) Férey, G. Hybrid Porous Solids: Past,
Present, Future. Chem. Soc. Rev. 2008, 37, 191−214. (d) Yaghi, O. M.
Reticular Chemistry: Molecular Precision in Infinite 2D and 3D.
Molecular Frontiers Journal 2019, 3, 1−18.
(4) (a) Gaillac, R.; Pullumbi, P.; Beyer, K. A.; Chapman, K. W.;
Keen, D. A.; Bennett, T. D.; Coudert, F. X. Liquid Metal−Organic
Frameworks. Nat. Mater. 2017, 16, 1149−1154. (b) Bennett, T. D.;
Horike, S. Liquid, Glass and Amorphous Solid States of Coordination
Polymers and Metal−Organic Frameworks. Nat. Rev. Mater. 2018, 3,
431−440. (c) Gaillac, R.; Pullumbi, P.; Coudert, F.-X. Melting of
Zeolitic Imidazolate Frameworks with Different Topologies: Insight
from First-Principles Molecular Dynamics. J. Phys. Chem. C 2018,
122, 6730−6736. (d) Horike, S.; Nagarkar, S. S.; Ogawa, T.;
Kitagawa, S. A New Dimension for Coordination Polymers and
Metal−Organic Frameworks: Towards Functional Glasses and
Liquids. Angew. Chem., Int. Ed. 2020, 59, 6652−6664.
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.0c11718
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was partially supported by the Arnold and Mabel
Beckman Foundation through a Beckman Young Investigator
grant awarded to J.A.M. and Beckman Postdoctoral Fellow-
ships to M.I.G. and A.H.S. Synthesis, crystallography, thermal
characterization, and EXAFS experiments were supported by
the Office of Naval Research under Award No. N00014-19-1-
2148. Studies of metal−bis(acetamide) glass formation were
supported by the U.S. Department of Energy (DoE) under
Award No. DE-SC0021145. The PDF work was supported by
the NSF MRSEC program through Columbia in the Center for
Precision Assembly of Quantum Materials (PAQM) (DMR-
2011738). We thank Dr. Shao-Liang Zheng for assistance with
crystallography experiments and data analysis. We thank Dr.
Sungsik Lee for assistance with EXAFS experiments and data
analysis. EXAFS measurements were conducted on beamline
12-BM at the Advanced Photon Source, a DoE Office of
Science User Facility operated for the DoE Office of Science
by Argonne National Laboratory under Contract No. DE-
AC02-06CH11357. We thank Dr. Maxwell Terban for
assistance with PDF data analysis. X-ray PDF measurements
were conducted on beamline 28-ID-2 of the National
Synchrotron Light Source II, a U.S. Department of Energy
(DOE) Office of Science User Facility operated for the DOE
Office of Science by Brookhaven National Laboratory under
Contract No. DE-SC0012704.
(5) Moghadam, P. Z.; Li, A.; Wiggin, S. B.; Tao, A.; Maloney, A. G.;
Wood, P. A.; Ward, S. C.; Fairen-Jimenez, D. Development of a
Cambridge Structural Database Subset: A collection of Metal−
Organic Frameworks for Past, Present, and Future. Chem. Mater.
2017, 29, 2618−2625.
(6) (a) Bennett, T. D.; Tan, J.-C.; Yue, Y.; Baxter, E.; Ducati, C.;
Terrill, N. J.; Yeung, H. H. -M.; Zhou, Z.; Chen, W.; Henke, S.;
Cheetham, A. K.; Greaves, G. N. Hybrid Glasses from Strong and
Fragile Metal−Organic Framework liquids. Nat. Commun. 2015, 6,
8079. (b) Umeyama, D.; Horike, S.; Inukai, M.; Itakura, T.; Kitagawa,
S. Reversible Solid-to-Liquid Phase Transition of Coordination
Polymer Crystals. J. Am. Chem. Soc. 2015, 137, 864−870. (c) Bennett,
T. D.; Yue, Y.; Li, P.; Qiao, A.; Tao, H.; Greaves, N. G.; Richards, T.;
Lampronti, G. I.; Redfern, S. A. T.; Blanc, F.; Farha, O. K.; Hupp, J.
T.; Cheetham, A. K.; Keen, D. A. Melt-Quenched Glasses of Metal−
Organic Frameworks. J. Am. Chem. Soc. 2016, 138, 3484−3492.
(d) Zhou, C.; Longley, L.; Krajnc, A.; Smales, G. J.; Qiao, A.; Erucar,
I.; Doherty, C. M.; Thornton, A. W.; Hill, A. J.; Ashling, C. W.;
Qazvini, O. T.; Lee, S. J.; Chater, P. A.; Terrill, N. J.; Smith, A. J.; Yue,
Y.; Mail, G.; Keen, D. A.; Telfer, S. G.; Bennett, T. D. Metal−Organic
Frameworks Glasses with Permanent Accessible Porosity. Nat.
Commun. 2018, 9, 5042. (e) Frentzel-Beyme, L.; Kloß, M.;
Kolodzeiski, P.; Pallach, R.; Henke, S. Meltable Mixed-Linker Zeolitic
Imidazolate Frameworks and Their Microporous Glasses: From
Melting Point Engineering to Selective Hydrocarbon Sorption. J. Am.
Chem. Soc. 2019, 141, 12362−12371. (f) Kimata, H.; Mochida, T.
Chem. Eur. J., Crystal Structures and Melting Behaviors of 2D and 3D
Anionic Coordination Polymers Containing Organometallic Ionic
Liquid Components. Chem. - Eur. J. 2019, 25, 10111−10117.
REFERENCES
■
(1) (a) Agarwal, M.; Chakravarty, C. Relationship between
Structure, Entropy, and Mobility in Network-Forming Ionic Melts.
Phys. Rev. E 2009, 79, No. 030202. (b) Agarwal, M.; Ganguly, A.;
Chakravarty, C. Transport Properties of Tetrahedral, Network-
Forming Ionic Melts. J. Phys. Chem. B 2009, 113, 15284−15292.
2809
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(g) Nagarkar, S. S.; Kurasho, H.; Duong, N. T.; Nishiyama, Y.;
Kitagawa, S.; Horike, S. Crystal Melting and Glass Formation in
Copper Thiocyanate Based Coordination Polymers. Chem. Commun.
2019, 55, 5455−5458. (h) Widmer, R. N.; Lampronti, G. I.; Anzellini,
S.; Gaillac, R.; Farsang, S.; Zhou, C.; Belenguer, A. M.; Palmer, H.;
Kleppe, A. K.; Wharmby, M. T.; Redfern, S. A.; Coudert, F.-X.;
MacLeod, S. G.; Bennett, T. Pressure Promoted Low-Temperature
Melting of Metal−Organic Frameworks. Nat. Mater. 2019, 18, 370−
376. (i) Das, C.; Ogawa, T.; Horike, S. Stable Melt Formation of 2D
Nitrile-Based Coordination Polymer and Hierarchical Crystal−Glass
Structuring. Chem. Commun. 2020, 56, 8980−8983. (j) Hou, J.;
Gómez, M. L. R.; Krajnc, A.; McCaul, A.; Li, S.; Bumstead, A. M.;
Sapnik, A. F.; Deng, Z.; Lin, R.; Chater, P. A.; Keeble, D. S.; Keen, D.
A.; Appadoo, D.; Chan, B.; Vicki, Chen.; Mali, G.; Bennett, T. D.
Halogenated Metal−Organic Framework Glasses and Liquids. J. Am.
Chem. Soc. 2020, 142, 3880−3890. (k) Das, C.; Horike, S. Crystal
Melting and Vitrification Behaviors of the Three-Dimensional Nitrile-
Based Metal−Organic Framework. Faraday Discuss., 2021,
DOI: 10.1039/D0FD00003E.
(7) A brief comment on semantics: many terms, including “metal−
organic frameworks” (MOFs), “coordination polymers”, and “coor-
dination network solids”, are currently used to describe solid-state
structures composed of inorganic units linked by bridging organic
ligands through coordination bonds to form extended networks.
Although “metal−organic framework liquids” has recently been used
to refer to the liquid phases of ZIF compounds, we believe that the
more general descriptors “coordination network-forming liquid” or
“metal−organic network-forming liquid” better capture the inter-
connectedness of inorganic units and bridging organic ligands in the
liquid state and the resulting structural correlations that are imparted
by the directional nature of dynamic coordination bonds. Moreover,
such terminology is consistent with the long-standing use of “network-
forming liquids” in the physical chemistry community to describe
liquids, such as H2O and molten SiO2, with structural correlations that
resemble those of a parent crystalline network and are mediated by
other types of directional interactions, such as hydrogen bonding and
covalent bonds.
(12) Butler, K. T.; Walsh, A.; Cheetham, A. K.; Kieslich, G.
Organised Chaos: Entropy in Hybrid Inorganic−Organic Systems and
Other Materials. Chem. Sci. 2016, 7, 6316−6324.
(13) (a) Dannenfelser, R.-M.; Yalkowsky, S. H. Estimation of
Entropy of Melting from Molecular Structure: A Non-Group
Contribution Method. Ind. Eng. Chem. Res. 1996, 35, 1483−1486.
(b) Gilbert, A. S. Entropy−Enthalpy Compensation in the Fusion or
Organic Molecules: Implications for Walden’s Rule and Molecular
Freedom in Liquid State. Thermochim. Acta 1999, 339, 131−142.
(c) Wei, J. Molecular Symmetry, Rotational Entropy, and Elevated
Melting Points. Ind. Eng. Chem. Res. 1999, 38, 5019−5027.
(14) (a) Goodgame, D. M. L.; Grachvogel, D. A.; Hussain, I.; White,
A. J. P.; Williams, D. J. Formation of Polymeric Network Arrays by
Complexes of Manganese(II) or Cobalt (II) with Alkane Chain
Linked Bis(amide) Ligands of Biological Relevance. Inorg. Chem.
1999, 38, 2057−2063. (b) Chatterton, N. P.; Goodgame, D. M. L.;
Grachvogel, D. A.; Hussain, I.; White, A. J. P.; Williams, D. J.
Influence of the Counteranion on the Formation of Polymeric
Networks by Metal Complexes of Hexamethylenebis(acetamide).
Inorg. Chem. 2001, 40, 312−317. (c) Alexandru, M.-G.; Jitaru, I.;
Madalan, A. M.; Andruh, M. Aggregation of Two Different
Coordination Polymers by Reacting Zinc Nitrate and Cadmium
Chloride with N, N’−Ethylenebisacetamide. J. Coord. Chem. 2011, 64,
3333−3341.
̈
̈
(15) Nimmermark, A.; Ohrstrom, L.; Reedijk, J. Metal-Ligand Bond
Lengths and Strengths: Are They Correlated? A Detailed CSD
Analysis. Z. Kristallogr. - Cryst. Mater. 2013, 228, 311−317.
(16) Ryde, U. A Fundamental View of Enthalpy−Entropy
Compensation. MedChemComm 2014, 5, 1324−1336.
(17) (a) Irving, H.; Williams, R. J. P. The Stability of Transition-
metal Complexes. J. Chem. Soc. 1953, 3192−3210. (b) Helm, L.;
Merbach, A. E. Water Exchange on Metal Ions: Experiments and
Simulations. Coord. Chem. Rev. 1999, 187, 151−181.
(18) McGillicuddy, R. D.; Thapa, S.; Wenny, M. B.; Gonzalez, M. I.;
Mason, J. A. Metal−Organic Phase-Change Materials for Thermal
Energy Storage. J. Am. Chem. Soc. 2020, 142, 19170−19180.
(19) (a) Williams, D. H.; O’Brien, D. P.; Bardsley, B. Enthalpy/
Entropy Compensation as a Competition between Dynamics and
Bonding: The Relevance to Melting of Crystals and Biological
(8) Longley, L.; Collins, S. M.; Li, S.; Smales, G. J.; Erucar, I.; Qiao,
A.; Hou, J.; Doherty, C. M.; Thornton, A. W.; Hill, A. J.; Yu, X.;
Nicholas, T. J.; Smith, A. J.; Cohen, S. M.; Midgley, P. A.; Keen, D. A.;
Telfer, S. G.; Bennett, T. D. Flux Melting of Metal−Organic
Frameworks. Chem. Sci. 2019, 10, 3592−3601.
(9) Mullin, J. B. Single Crystal Growth I: Melt Growth. In Electronic
Materials; Miller, L. S., Mullin, J. B., Eds.; Springer: Boston, 1991; pp
111−126.
̀
Aggregates. J. Am. Chem. Soc. 2001, 123, 737−738. (b) Ventola, L.;
Cuevas-Diarte, M. A.; Ramírez, M.; Oonk, H. A. J.; Mondieig, D.;
Negrier, Ph. Melting Behavior in the n-alkanol Family, Enthalpy−
Entropy Compensation. Phys. Chem. Chem. Phys. 2004, 6, 1786−
1791.
(20) (a) Jeffrey, G.; An, A. Introduction to Hydrogen Bonding; Oxford
University Press: Oxford, 1997. (b) Steiner, T. The Hydrogen Bonds
in the Solid State. Angew. Chem., Int. Ed. 2002, 41, 48−76.
(21) Lambert, F. L. Configurational Entropy Revisited. J. Chem.
Educ. 2007, 84, 1548−1550.
(22) (a) Massa, W. Crystal Structure Determination; Springer:
Wiesbaden, 2015. (b) Neuburger, M. Disorder in Crystal Structures:
New Approaches in Finding the Best Model. Ph.D. Dissertation,
University of Basel, Basel, Switzerland, 2012.
(10) (a) Bennett, T. D.; Cheetham, A. K. Amorphous Metal−
Organic Frameworks. Acc. Chem. Res. 2014, 47, 1555−1562.
(b) Chen, W.; Horike, S.; Umeyama, D.; Ogiwara, N.; Itakura, T.;
Tassel, C.; Goto, Y.; Kageyama, H.; Kitagawa, S. Glass Formation of a
Coordination Polymer Crystal for Enhanced Proton Conductivity and
Material Flexibility. Angew. Chem., Int. Ed. 2016, 55, 5195−5200.
(c) Zhao, Y.; Lee, S.-Y.; Becknell, N.; Yaghi, O. M.; Angell, C. A.
Nanoprous Transparent MOF Glasses with Accessible Internal
Surface. J. Am. Chem. Soc. 2016, 138, 10818−10821. (d) Qiao, A.;
Bennett, T. D.; Tao, h.; Krajnc, A.; Mali, G.; Doherty, C. M.;
Thornton, A. W.; Mauro, J. C.; Greaves, G. N.; Yue, Y. A Metal−
Organic Framework with Ultrahigh Glass-forming Ability. Sci. Adv.
2018, 4, eaao6827. (e) Inukai, M.; Nishiyama, Y.; Honjo, K.; Das, C.;
Kitagawa, S.; Horike, S. Glass-Phase Coordination Polymer
Displaying Proton Conductivity and Guest-Accessible Porosity.
Chem. Commun. 2019, 55, 8528−8531.
(11) (a) Earle, M. J.; Seddon, K. R. Ionic Liquids. Green Solvents for
the Future. Pure Appl. Chem. 2000, 72, 1391−1398. (b) Dean, P. M.;
Turanjanin, J.; Yoshizawa-Fujita, M.; MacFarlane, D. R.; Scott, J. L.
Exploring an Anti-Crystal Engineering Approach to the Preparation of
Pharmaceutically Active Ionic Liquids. Cryst. Growth Des. 2009, 9,
1137−1145. (c) Welton, T.; Hallett, J. P. Room-Temperature Ionic
Liquids: Solvents for Synthesis and Catalysis. 2. Chem. Rev. 2011, 111,
3508−3576.
(23) The rotational entropy difference is estimated by ΔSrfuost = R ln σ,
where R is the ideal gas constant and σ is the external rotational
symmetry number (σ = 12 [MCl4]2−and σ = 100 for Cl− and Br−).
(24) (a) Rosenfeld, Y. Relation Between the Transport Coefficients
and the Internal Entropy of Simple Systems. Phys. Rev. A: At., Mol.,
Opt. Phys. 1977, 15, 2545−2549. (b) Sharma, R.; Chakraborty, S. N.;
Chakravarty, C. Entropy, Diffusivity, and Structural Order in Liquids
with Waterlike Anomalies. J. Chem. Phys. 2006, 125, 204501.
(c) Chakravarty, C.; Nayar, D. Water and Water-like Liquids:
Relationships between Structure, Entropy and Mobility. Phys. Chem.
Chem. Phys. 2013, 15, 14162−14177.
(25) Gould, S. L.; Tranchemontagne, D.; Yaghi, O. M.; Garcia-
Garibay, M. A. Amphidynamic Character of Crystalline MOF-5:
Rotational Dynamics of Terephthalate Phenylenes in a Free-Volume,
Sterically Hindered Environment. J. Am. Chem. Soc. 2008, 130, 3246−
3237.
2810
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(26) (a) Egami, T.; Billinge, S. J. L. Underneath the Bragg Peaks:
Structural Analysis of Complex Materials, 2nd ed.; Elsevier:
́
Amsterdam, 2012. (b) Juhas, P.; Davis, T.; Farrow, C. L.; Billinge,
S. J. L. PDFgetX3: A Rapid and Highly Automatable Program for
Processing Powder Diffraction Data into Total Scattering Pair
Distribution Functions. J. Appl. Crystallogr. 2013, 46, 560−566.
(27) (a) Thorpe, M. F. Rigidity Percolation. In Physics of Disordered
Materials; Alder, D., Fritzsche, H., Ovshinsky, S. R., Eds; Springer:
Boston, 1985; pp 55−61. (b) Phillips, J. C.; Thorpe, M. F. Constraint
Theory, Vector Percolation and Glass Formation. Solid State Commun.
1985, 53, 699−702.
(28) (a) Keblinski, P. Molecular Dynamics Study of Screening in
Ionic Fluids. J. Chem. Phys. 2000, 113, 282−291. (b) Del Pópolo, M.
G.; Voth, G. A. On the Structure and Dynamics of Ionic Liquids. J.
Phys. Chem. B 2004, 108, 1744−1752.
(29) (a) Elliott, S. R. Extended-range Order, Interstitial Voids and
the First Sharp Diffraction Peak of Network Glasses. J. Non-Cryst.
Solids 1995, 182, 40−48. (b) Salmon, P. S.; Martin, R. A.; Mason, P.
E.; Cuello, G. J. Topological versus Chemical Ordering in Network
Glasses at Intermediate and Extended Length Scales. Nature 2005,
435, 75−78.
(30) (a) Tao, H.; Bennett, T. D.; Yue, Y. Melt-Quenched Hybrid
Glasses from Metal−Organic Frameworks. Adv. Mater. 2017, 29,
1601705. (b) Longley, L.; Calahoo, C.; Limbach, R.; Xia, Y.; Tuffnell,
J. M.; Sapnik, A. F.; Thorne, M. F.; Keeble, D. S.; Keen, D. A.;
Wonderaczek, L.; Bennett, T. D. Metal−Organic Framework and
Inorganic Glass Composites. Nat. Commun. 2020, 11, 5800.
(31) Greaves, G. N.; Sen, S. Inorganic Glasses, Glass-Forming
Liquids and Amorphizing Solids. Adv. Phys. 2007, 56, 1−166.
(32) (a) Turnbull, D. Under What Conditions Can a Glass be
Formed? Contemp. Phys. 1969, 10, 473−488. (b) Baird, J. A.;
Santiago-Quinonez, D.; Rinaldi, C.; Taylor, L. S. Role of Viscosity in
Influencing the Glass-Forming Ability of Organic Molecules from the
Undercooled Melt State. Pharm. Res. 2012, 29, 271−284.
(33) Abbasi, A.; Saghir, M. Z.; Kawaji, M. Evaluation of the
Activation Energy of Viscous Flow for a Binary Mixture in Order to
Estimate the Thermodiffusion Coefficient. J. Non-Equilib. Thermodyn.
2011, 36, 23−40.
̈
(34) (a) Bohmer, R.; Angell, C. A. Local and Global Relaxations in
Glass Forming Materials. In Disorder Effects on Relaxational Processes,
Glasses, Polymers, Proteins; Richert, R., Blumen, A., Eds.; Springer-
Verlag: Berlin, 1994; pp 11−54. (b) Mel’cuk, A. I.; Ramos, R. A.;
Gould, H.; Klein, W.; Mountain, R. D. Long-Lived Structures in
Fragile Glass-Forming Liquids. Phys. Rev. Lett. 1995, 75, 2522−2525.
(35) (a) Hecksher, T.; Nielsen, A. I.; Olsen, N. B.; Dyre, J. C. Little
Evidence for Dynamic Divergences in Ultraviscous Molecular Liquids.
Nat. Phys. 2008, 4, 737−741. (b) Zheng, Q.; Mauro, J. C. Viscosity of
Glass-Forming Systems. J. Am. Ceram. Soc. 2017, 100, 6−25.
(36) (a) Angell, A. Thermodynamics: Liquid Landscape. Nature
1998, 393, 521−524. (b) Debenedetti, P. G.; Stillinger, F. H.
Supercooled Liquids and the Glass Transition. Nature 2001, 410,
259−267. (c) Chandler, D. From 50 Years Ago, the Birth of Modern
Liquid-State Science. Annu. Rev. Phys. Chem. 2017, 68, 19−38.
2811
https://dx.doi.org/10.1021/jacs.0c11718
J. Am. Chem. Soc. 2021, 143, 2801−2811