Metal-specific interactions of H2 adsorbed within isostructural metal-organic frameworks

Article abstract of DOI:10.1021/ja2071384

Diffuse reflectance infrared (IR) spectroscopy performed over a wide temperature range (35-298 K) is used to study the dynamics of H₂ adsorbed within the isostructural metal-organic frameworks M₂L (M = Mg, Mn, Co, Ni and Zn; L = 2,5-dioxidobenzene-1,4-dicarboxylate) referred to as MOF-74 and CPO-27. Spectra collected at H₂ concentrations ranging from 0.1 to 3.0 H₂ per metal cation reveal that strongly red-shifted vibrational modes arise from isolated H₂ bound to the available metal coordination site. The red shift of the bands associated with this site correlate with reported isosteric enthalpies of adsorption (at small surface coverage), which in turn depend on the identity of M. In contrast, the bands assigned to H₂ adsorbed at positions >3 A from the metal site exhibit only minor differences among the five materials. Our results are consistent with previous models based on neutron diffraction data and independent IR studies, but they do not support a recently proposed adsorption mechanism that invokes strong H₂???H₂ interactions (Nijem et al. J. Am. Chem. Soc.2010, 132, 14834-14848). Room temperature IR spectra comparable to those on which the recently proposed adsorption mechanism was based were only reproduced after contaminating the adsorbent with ambient air. Our interpretation that the uncontaminated spectral features result from stepwise adsorption at discrete framework sites is reinforced by systematic red shifts of adsorbed H₂ isotopologues and consistencies among overtone bands that are well-described by the Buckingham model of molecular interactions in vibrational spectroscopy.

Full text of DOI:10.1021/ja2071384

ARTICLE
pubs.acs.org/JACS
Metal-Specific Interactions of H2 Adsorbed within Isostructural
MetalÀOrganic Frameworks
Stephen A. FitzGerald,*,† Brian Burkholder,† Michael Friedman,† Jesse B. Hopkins,† Christopher J. Pierce,†
Jennifer M. Schloss,† Benjamin Thompson,† and Jesse L. C. Rowsell*,‡
†Department of Physics and Astronomy, Oberlin College, Oberlin, Ohio 44074, United States
‡Department of Chemistry and Biochemistry, Oberlin College, Oberlin, Ohio 44074, United States
S Supporting Information
b
ABSTRACT: Diffuse reflectance infrared (IR) spectroscopy
performed over a wide temperature range (35À298 K) is used
to study the dynamics of H2 adsorbed within the isostructural
metalÀorganic frameworks M2L (M = Mg, Mn, Co, Ni and Zn;
L = 2,5-dioxidobenzene-1,4-dicarboxylate) referred to as MOF-
74 and CPO-27. Spectra collected at H2 concentrations ranging
from 0.1 to 3.0 H2 per metal cation reveal that strongly red-
shifted vibrational modes arise from isolated H2 bound to the
available metal coordination site. The red shift of the bands
associated with this site correlate with reported isosteric en-
thalpies of adsorption (at small surface coverage), which in turn depend on the identity of M. In contrast, the bands assigned to H2
adsorbed at positions >3 Å from the metal site exhibit only minor differences among the five materials. Our results are consistent
with previous models based on neutron diffraction data and independent IR studies, but they do not support a recently proposed
adsorption mechanism that invokes strong H2 H2 interactions (Nijem et al. J. Am. Chem. Soc. 2010, 132, 14834À14848). Room
3 3 3
temperature IR spectra comparable to those on which the recently proposed adsorption mechanism was based were only
reproduced after contaminating the adsorbent with ambient air. Our interpretation that the uncontaminated spectral features result
from stepwise adsorption at discrete framework sites is reinforced by systematic red shifts of adsorbed H2 isotopologues and
consistencies among overtone bands that are well-described by the Buckingham model of molecular interactions in vibrational
spectroscopy.
and Zn-CPO-27,17,16 has aroused considerable interest because of its
1. INTRODUCTION
relatively large density of available coordination sites and a respect-
able micropore volume. This M2L framework (L = 2,5-dioxidoben-
zene-1,4-dicarboxylate) is especially attractive to those investigating
the interactions of small molecules in cavities because isostructural
It is envisioned by many that supplanting the inveterate
combustion engine with modern fuel cell technology employing
hydrogen as an energy carrier will greatly decrease our impact on
the global environment. Efficient and safe storage of hydrogen is
just one of the unsolved puzzles that comprise this enticing, yet
formidable goal. Although the past decade of research has yielded
an inventory of microporous materials with very large specific
materials may be prepared from several metal dications (M =
Mg,18,19 Mn,20 Fe,21 Co,22 Ni,23 Zn16), offering the opportunity to
examine site-specific alterations to the adsorption surface. Isostruc-
tural series are rare among MOFs, and it has already been shown that
surface areas and pore volumes for reversibly containing H2
the isosteric enthalpy of H2 adsorption (at small surface coverage)
molecules,1À9 the average adsorption enthalpies of these physi-
ranges from À8.5 to À13 kJ/mol among the members of the
sorbents fall short of the targeted range of À15 to À25 kJ/mol
that is theoretically necessary for operation at room temperature.10À12
Zeolites and some metalÀorganic frameworks (MOFs) contain
available cation coordination sites to which stronger binding of
H2 has been observed.14À15 These highly attractive sites typically
constitute a small fraction of the surface sites within a given
material, obviating their practical use for hydrogen storage. In
spite of this, these materials are of great interest as model systems
because the quantum dynamics of H2 in their pores provides
critical information for predicting the viability of other known or
proposed structures.
M2L series.20 These values suggest that microporous adsorbents
with average adsorption enthalpies surpassing À15 kJ/mol may be
within reach.
Despite several experimental studies and theoretical treatments
of the adsorption properties of M2L, uncertainties remain regard-
ing the mechanism by which H2 binds at different sites and the
degree to which H2 molecules on neighboring sites interact.14,24
One technique that has proven effective for revealing details
of the dynamics of H2 within porous structures is infrared (IR)
The structure type adopted by the material catena-(μ8-2,5-
dioxidobenzene-1,4-dicarboxylato)dizinc(II), referred to as MOF-74
Received: July 29, 2011
Published: November 10, 2011
r
2011 American Chemical Society
20310
dx.doi.org/10.1021/ja2071384 J. Am. Chem. Soc. 2011, 133, 20310–20318
|

Journal of the American Chemical Society
ARTICLE
spectroscopy.11,24À33 Unperturbed, gaseous H2 is IR inactive,
but upon adsorption weak IR signals arise due to interaction-
induced dipole moments. The frequency and relative intensity of
these spectral bands provide a measure of the different interac-
tion mechanisms (e.g., inductive, dispersive, and electrostatic) at
a particular adsorption site, as well as a means for distinguishing
between competing theoretical descriptions of these interactions.
In general, MOFs containing available coordination sites exhibit
IR spectra with H2 vibrational bands that are significantly red-
shifted compared to frameworks lacking these highly attractive
sites.24 In the case of M2L it has been independently shown that
spectra measured at low temperature (and low H2 concentra-
tion33) are dominated by these strongly red-shifted bands.11,24,33
Furthermore, low temperature neutron diffraction studies re-
vealed that adsorbed H2 first occupies the available metal co-
ordination site, and following saturation of this site other secondary
sites become occupied.34 The conclusion has thus emerged that the
IR bands that are most red-shifted are assignable to isolated H2 at
the metal site.
2. MATERIALS AND EXPERIMENTAL METHODS
Crystalline samples of M2L (M = Mg, Mn, Co, Ni, and Zn) were
prepared using reported procedures with minor modifications.18,20,36
Reaction solutions were prepared in 100 mL volumes, filtered, and
transferred to 250 mL glass bottles sealed with heat-resistant screw caps
fitted with Teflon liners. Solvents were thoroughly sparged with nitrogen
gas prior to use and during dissolution of the precursors. Reaction
vessels were placed in a preheated forced-convection oven (100 °C for
M = Co, Ni, Zn; 125 °C for M = Mg, Mn) and removed after 24 h. The
solid products were immediately separated by filtration and rinsed with
N,N-dimethylformamide followed by copious amounts of deionized
water. Exchange of water for adsorbed species within the pores of the
materials was effected by immersing the solids in large volumes
(∼100 mL) of deionized water that were decanted and replenished
every 12 h over a three day period. The solids were then separated by
filtration and transferred to an airtight glass vessel connected to a
vacuum manifold. After evacuating for several hours until the samples
were free-flowing powders, the vessel was immersed in a silicone oil bath
and heated at 2 °C/min to 200 °C and maintained at this temperature for
at least 4 h under dynamic vacuum. Samples were agitated periodically to
improve temperature homogeneity. After cooling to room temperature,
the sealed vessel was transferred to an argon glovebox to recover the
dehydrated material.
The composition and crystalline phase purity of each sample was
confirmed by a combination of Fourier transform infrared (FTIR)
spectroscopy, thermogravimetric analysis (TGA), powder X-ray diffrac-
tion (PXRD), and hydrogen gas adsorption measurements. FTIR
spectra were measured with 2 to 4 cmÀ1 resolution using a Nicolet
6700 spectrometer equipped with an attenuated total reflectance
accessory (Smart iTR, Thermo Scientific). TGA was performed by
heating ∼10 mg of sample in an alumina crucible at a constant rate of
2 °C/min under flowing prepurified nitrogen or dry compressed air and
monitoring the mass changes (TA Instruments SDT 2960). Powder
X-ray diffraction was performed on samples packed into flat plates using
Cu Kα radiation with BraggÀBrentano or parallel beam geometries
(Rigaku Ultima IV diffractometer). Samples were dehydrated in situ
under flowing nitrogen by raising the sample temperature by 2 °C/min
to specified set points for data collection using a programmable low/
medium temperature attachment (Rigaku). Hydrogen adsorption iso-
therms were measured at 77 K up to a pressure of 1 atm using a custom-
fabricated Sieverts apparatus. Dehydrated samples (∼0.1 g) were loaded
in a stainless steel adsorption cell within an argon glovebox and
evacuated on the manifold of the apparatus for several hours before
analysis. The manifold was flushed with hydrogen multiple times before
immersing the adsorption cell in liquid nitrogen and dosing the sample
with hydrogen in a stepwise manner. The pressure changes were
monitored using a transducer (Omega PX-303), with 15 min observed
to be generally sufficient for equilibration of each dosing step.
Recently, Chabal and co-workers challenged this conclusion
by invoking an unusual H2 pairing interaction to explain their IR
spectra of H2 in M2L.35 In this scenario, H2 at the metal site
exhibits a moderate shift in its vibrational frequency, and it is only
through interaction with another H2 at an adjacent secondary site
that the IR bands become strongly red-shifted to the value reported
by ourselves and others.11,24,33 In this scenario, H2 H2 interac-
3 3 3
tions are hypothesized to dominate over differences in the site-
specific binding potentials. If this were true the consequences
would be profound, both for theoretical modeling, in which
H2
H2 interactions are often neglected, but also (as the
3 3 3
authors point out) for the utility of variable-temperature IR
spectroscopy in estimating site-specific adsorption enthalpies.
Calculations using van der Waals density functional theory
(vdW-DFT) were presented to support the authors’ claim for
such a large frequency shift resulting from H2 H2 interactions;
3 3 3
however, the magnitude of this shift contradicts our previous
interpretation of the spectra.33 Consequently, a definitive experi-
mental analysis is required to test the capability of vdW-DFT to
accurately model H2 adsorption.
Given that the M2L structure type is animportant model system
for understanding the interactions of H2 within microporous
physisorbents, we have attempted to reproduce the observations
of the Chabal group (being particularly mindful of the long
equilibration times emphasized in their paper) to scrutinize their
proposed H2 pairing interaction. Through careful sample prepara-
tion and handling we have acquired spectra from H2 in five
different M2L materials (M = Mg, Mn, Co, Ni, Zn), all of which
are consistent with the original adsorption model developed by
ourselves and others. We find that, regardless of temperature or
equilibration time, the spectra acquired with the smallest concen-
trations of adsorbed H2 display only the strongly red-shifted
vibrational bands originally reported.11,24,33 Moreover, we are only
able to reproduce the spectra presented by the Chabal group after
deliberately exposing our samples to air, which is known to
introduce adsorbed species (particularly water) that tenaciously
bond to available metal coordination sites. To complete our study
of this system, we have measured the spectra of the adsorbed
isotopologues HD and D2 in these materials and confirmed
consistencies among their fundamental and overtone bands that
adhere to the well-established Buckingham model of quantized
rotational dynamics. We present the results of this study to
accelerate the development of accurate models of H2 adsorption.
Infrared spectra were acquired using a Bomem DA3 Michelson inter-
ferometer equipped with quartz-halogen and glowbar sources, CaF2
beamsplitter, and a liquid nitrogen cooled mercuryÀcadmiumÀtelluride
detector. A custom-built diffuse reflectance system with a sample chamber
that allows both the temperature and atmosphere of the material to be
controlled was utilized for all experiments.37 Samples of M2L (∼10 mg)
were transferred under inert atmosphere to a cup affixed to a copper slab
providing thermal contact to a coldfinger cryostat (Janis ST-300T). The
sample temperature was monitored by a Si-diode thermometer placed
directly within the sample cup. Prior to introduction of the hydrogen gas,
the samples were evacuated for several hours at room temperature.
Volumes of research grade H2, HD, and D2 (99.9999% purity) were
dispensed from a calibrated gas manifold. In a typical experiment, a
reference spectrum was first obtained of the evacuated material at 35 K.
The sample was then heated to 150 K, exposed to a known quantity of
20311
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
H2, and cooled slowly to 35 K before acquiring a spectrum. The quantity
of H2 adsorbed was determined by changes in the manifold pressure.
Control experiments were also performed with a nonadsorbing sample
such as CaF2 and by exposing M2L samples to He gas.
pores, samples were immersed in water, which extracts and
replaces initially adsorbed species.
Results of PXRD and TGA of the hydrated materials were in
good agreement with those reported by the Dietzel group,17,19,23
who employed water and tetrahydrofuran as the reaction
solvents. Variable temperature PXRD revealed the expected
changes in the relative intensities of reflections during the
desolvation process, below 150 °C. Between 150 and 300 °C
there were negligible changes in the diffraction patterns of each
material, indicating the samples were solvent-free and their
crystallinity was maintained. Comparison of the patterns of the
five samples with each other (see Figure S1) and with calculated
patterns of the desolvated structures confirmed the phase purity
of the crystalline fractions. The thermal behavior observed by
PXRD was consistent with observations by TGA, which indi-
cated that desolvation of the samples was complete by 180 °C, as
shown in Figures S2ÀS6. Together with results from previous
reports, these observations lead us to choose 200 °C as the
desolvation temperature for the bulk samples. In some previous
reports methanol has been used as the exchange solvent for the
M2L structure type, yet we found that exchange with water was
just as effective for producing high-quality samples, in agreement
with the observations of the Dietzel group.
The compositional purity of the samples was checked by TGA
under flowing air. Oxidation of the organic ligand leaves the
corresponding metal oxide (MgO, Mn3O4, Co3O4, NiO, or
ZnO) as residue, and the observed mass changes were in good
agreement with calculated values (see Figures S2ÀS6). For
Mn2L, Co2L, and Ni2L the oxidation step was observed to begin
around 200 °C; while these materials have confirmed stabilities
above this temperature under inert atmospheres, it is noted that
small partial pressures of oxygen present while heating above
200 °C will likely cause sample degradation.
Desolvated samples were further characterized by FTIR and
H2 adsorption measurements. Small aliquots of the materials
were examined using attenuated total reflectance by compres-
sing the powder on the surface of a ZnSe crystal. IR spectra
collected within 60 s after exposing a sample to ambient air
display bands characteristic of the structure type (see Figure S7
and Table S1) and no bands corresponding to water or residual
solvent in the pores. Within minutes, however, the absorbance
features of water became readily apparent (see Figure S8).
Monitoring the rate of rehydration was improved by releasing
the anvil compressing the sample, stirring the powder to
homogenize exposed grains, and repressing the sample before
collecting the next spectrum. Our observations reinforce re-
commendations that desolvated MOF samples should be stored and
handled under an inert atmosphere to ensure optimal performance.
We have confirmed the requisite porosity of our samples by
measuring their adsorption of H2 at 77 K up to a pressure of 1 atm
(see Figure S9).
Adsorbed H2 IR Band Assignments. Figure 1 compares
representative spectra for H2 loaded in M2L at 35 K and acquired
at 35 K, the lowest temperature examined in this study. For each
material, spectra are shown for both a small H2 concentration
(∼1 H2/M) and a large concentration (>2 H2/M). Each
spectrum can be divided into three regions: (1) the Q-band
region arising from purely vibrational motion of the adsorbed H2
and occurring between 4000 and 4150 cmÀ1, (2) the region from
4150 to 4300 cmÀ1 associated with changes in the center-of-mass
translational quantum number, and (3) the S-band region from
4250 to 4700 cmÀ1, where one observes rovibrational bands.
3. THEORETICAL BACKGROUND
Free H2 vibrates at an IR inactive frequency of 4161 cmÀ1 and
acts as a near-perfect quantum rotor with rotational energy levels
EJ = BvJ(J + 1), where J is the rotational quantum number and
Bv is the rotational constant for a given vibrational state v.38 As a
result of quantum statistics, the H2 molecule exists as either
ortho-H2 with total nuclear spin I = 1 and odd J quantum
numbers or as para-H2 with I = 0 and even J. These restrictions
fundamentally limit all photon-induced transitions to ΔJ = 0
(Q transitions) or ΔJ = ( 2 (S and O transitions).39 In the
absence of ortho to para conversion there are four possible
transitions at low temperature: purely vibrational Q(0) and
Q(1), which differ by only 6 cmÀ1 in the gas phase, and the
rovibrational transitions S(0) and S(1) (numbers in parentheses
indicate the initial rotational state). The interaction of H2 with
surface atoms modifies both the vibrational and rotational
constants and lifts the 2J + 1 degeneracy of the rotational levels.
This leads to frequency shifts, splittings, intensity changes, and
activation of IR absorption bands, all of which are correlated with
the symmetry and the interaction energy of the binding site.
Adsorbed H2 can also undergo translational transitions in which
it changes its center-of-mass quantum state. Bands associated
with these translational transitions tend to be broader than those
arising from pure rovibrational transitions.29,40,41
There are two basic mechanisms by which interaction-induced
dipole moments can lead to IR activity. In the first mechanism,
framework atoms interact with the H2 polarizability to induce a
dipole moment on the molecule. The H2 polarizability is
essentially isotropic, and so to a good approximation this
mechanism only activates Q transitions.42 Herein, we refer to
this as the isotropic mechanism. The second mechanism involves
the quadrupole moment of H2 inducing a dipole moment on the
framework atoms, and is denoted the quadrupole-induction
mechanism. The quadrupole moment of H2 is highly anisotropic,
and this term leads to activation of rovibrational S(0), S(1), and
purely vibrational Q(1) transitions. In contrast, Q(0) transitions
are not activated because the J = 0 state is spherically symmetric
and possesses no quadrupole moment in the initial and final
states.43
4. RESULTS AND DISCUSSION
Sample Preparation and Characterization. Samples of the
isostructural materials were prepared by solvothermal reaction of
a soluble metal(II) precursor with the protonated form of the
structural ligand. Differences in the solvent composition (which
includes N,N-dimethylformamide, ethanol, and water) and reac-
tion temperature were used to improve product yield and crystal
quality, as described in the literature.18,20,36 It was observed that
thoroughly sparging the solutions with nitrogen gas prior to
heating minimized solvent oxidation, which is evident by brown
discoloration of the product mixtures, particularly in reactions
containing nickel(II) and cobalt(II). It is hypothesized that
oxidative solvent degradation yields byproducts that are strongly
adsorbed within the framework pores and the formation of these
byproducts should be prevented. To facilitate desolvation of the
20312
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
Figure 1. Diffuse reflectance IR spectra of adsorbed H2 in isostructural M2L frameworks at 35 K. Light-colored traces correspond to adsorbed H2
concentrations less than 1 H2/M, and dark gray traces correspond to >2 H2/M. Spectra are offset for clarity and have a resolution of 1 cmÀ1. The spectra
of H2 in Ni2L have been magnified by a factor of 2 for comparison.
In all cases the Q-bands are more intense than the S-bands,
indicating that the isotropic mechanism is dominant over the
quadrupole-induction mechanism. Similar intensity relationships
have been observed for H2 in other MOFs and zeolites.25,30,31
In contrast, H2 in C60 gives S-bands of comparable intensity to
the Q-bands, which was attributed to the large polarizability of
the C60 molecules.29,41 As expected for low temperature spectra,
the translational sidebands associated with Q-transitions are
observed on the high frequency side of the Q-bands, and these
are broader than the parent Q-bands. The translational sidebands
are most likely composed of multiple overlapping bands, reflecting
the asymmetry of the adsorption sites. We did not detect transla-
tional sidebands associated with the S-transitions, probably due
to their smaller intensities.
The spectra acquired with smaller H2 concentrations (light-
colored spectra in Figure 1) exhibit notable differences among
the five materials, particularly in the Q-region between 4000 and
4100 cmÀ1. In contrast, the secondary bands that emerge after
more H2 is adsorbed (dark gray spectra in Figure.1) appear
at similar frequencies and with similar relative intensities for
each material. These include new Q-bands between 4110 and
4150 cmÀ1 and S-bands between 4440 and 4500 cmÀ1. To
explore the concentration dependence of the spectral features in
more detail, we performed experiments in which the system was
loaded with a known quantity of H2 at 150 K, and then cooled
slowly to 35 K. It has been suggested that a few hours may be
required for such a system to attain thermal equilibrium;35
however, we observed no significant differences between spectra
acquired after temperature ramps lasting 30 min or 3 h. Repre-
sentative spectra for Mg2L are presented in Figure 2. At the
smallest concentration, the only bands are an ortho/para pair in
the region 4080À4100 cmÀ1. The assignment of these to an
ortho/para pair was confirmed previously using para-enriched H2
in Zn2L.33 The integrated intensity of these bands increases with
H2 concentration until saturating at a concentration of 1.0 H2/M.
At this concentration a second ortho/para pair appears at
4127 and 4136 cmÀ1, and the intensity of these bands similarly
increases with concentration until saturation at 2.0 H2/M. The
pattern is repeated by a third pair of bands at 4131 and
4141 cmÀ1 for concentrations between 2.0 and 3.0 H2/M.
While other researchers have presented IR spectra of H2 in
MOFs as a function of loading pressure,11,24,35 none have
analyzed these data stoichiometrically in terms of the number
of adsorbed H2 per metal ion. Doing so allows us to directly
compare our spectra with reported neutron diffraction models.34
These indicated that at 4 K an adsorbed D2 molecule (an
isotopologue of H2) first occupies the available coordination site
of the metal ion, with its center-of-mass 2.6 Å from the metal.
This is the only site reportedly occupied up to a loading of
1.0 D2/M. Above this concentration, the occupancy of this
primary site remains essentially complete, while a second site
located above a triangle of oxygen atoms begins to fill. Similarly, a
third site close to the benzene ring becomes occupied for
loadings greater than 2 D2/M. On the basis of these results,
it can be stated with confidence that the IR ortho/para pair
centered at 4090 cmÀ1 is associated with the metal site, while the
20313
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
Figure 3. Concentration dependencies of Q-bands in the spectra of H2
in desolvated Zn2L (upper traces, blue/gray) and the same sample after
exposure to ambient air for 5 min (lower traces, orange/brown). Spectra
were acquired at 298 K for loading pressures of 25, 40, 50, and 100 bar
(bottom to top of each grouping).
predicts the 4090 cmÀ1 band would reach a maximum intensity
at concentrations above 2 H2/M as molecular pairs are formed.
The main data cited by Chabal and co-workers to justify their
band assignments are spectra acquired at room temperature with
high pressures of H2.
To resolve this disagreement, we attempted to reproduce the
experiments performed by the Chabal group using Zn2L that had
been carefully desolvated at 200 °C and handled under an inert
atmosphere. Spectra collected at room temperature (Figure 3,
upper traces) show that at the smallest H2 loading pressure (25
bar) a single broad band is observed, centered at 4095 cmÀ1
.
Increasing the H2 pressure to 40 bar causes the intensity of the
band to increase, but produces no other significant changes.
Further H2 loading (50 bar) leads to the appearance of a new
band at 4135 cmÀ1. The intensity of this band increases
significantly at the largest loading pressure of 100 bar. These
results are consistent extrapolations of our low-temperature data.
The open metal coordination site is still the first populated; however,
its enthalpic advantage is less dominant at room temperature.
Entropic effects are more apparent, in that the secondary sites
begin to be occupied before the primary site is completely filled.
The larger widths of the bands at room temperature make it
impossible to distinguish the bands of the secondary sites from
each other, and most likely the band at 4135 cmÀ1 arises from H2
in multiple environments. Notably, these spectra are different
from those reported by Chabal and co-workers for (ostensibly)
the same material under the same conditions.
It was not until we deliberately exposed our Zn2L sample to air
for 5 min, evacuated it at room temperature, and then repeated
the H2 loading experiment that we acquired spectra comparable
to those in ref 35. At each pressure, the spectra are dominated by
a single broad band centered at 4126 cmÀ1 (see Figure 3, lower
traces). During preparation of this manuscript, a second paper
was published by Chabal and co-workers with analogous findings
for Co2L.44 We thus repeated our analyses with Co2L both before
Figure 2. Upper plot: IR spectra demonstrating the concentration
dependencies of Q-bands in the spectra of H2 in Mg2L at 35 K. The
appearance of bands assigned to H2 adsorbed at the primary (1°),
secondary (2°), and tertiary (3°) sites are noted. The translational
sideband of the primary transition is labeled Qtrans. Spectra are offset for
clarity and have a resolution of 1 cmÀ1. Lower plot: the deconvoluted,
integrated intensities of the primary (b), secondary (9), and tertiary
(2) ortho/para paired bands with increasing adsorbate concentration,
determined by parallel manometry measurements.
bands in the region 4125À4145 cmÀ1 arise from H2 adsorbed at
the oxygen and benzene sites.
In the alternative scheme for the adsorption process proposed
by Chabal and co-workers,35 the isolated occupancy of the metal
site produces initial bands around 4125 cmÀ1, and only after
occupancy of secondary sites do these bands shift to the
4090 cmÀ1 region, due to appreciable H2 H2 interactions.
3 3 3
This interpretation is irreconcilable with the low-temperature
spectra presented in Figure 2, since it precludes the presence of a
4090 cmÀ1 band without an accompanying band at ∼4125 cmÀ1
arising from H2 in the secondary sites. Moreover, their scheme
20314
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
Table 1. Frequencies of the Fundamental Vibrational Mode
Q(0) of Molecular Hydrogen Isotopologues at the Primary
Adsorption Site of M2L, Along with the Frequency Shift
Relative to Gas Phase, Percentage Frequency Shift, and
Frequency Shift of the Overtone Mode
ν1r0
(cmÀ1
Δν1r0
Δν1r0 /ν1r0
Δν2r0 /2
(cmÀ1
gas
)
(cmÀ1
)
(%)
)
Mg2L
Mn2L
Co2L
Ni2L
H2
HD
D2
4092
3570
2940
À69
À61
À52
À1.66
À1.68
À1.74
À68.5
À63.0
À53.5
H2
HD
D2
4088
3566
2936
À73
À65
À56
À1.75
À1.79
À1.87
À74.5
À67.5
À58.0
H2
HD
D2
4047
3529
2904
À114
À102
À88
À2.74
À2.81
À2.94
a
À108
À92.0
H2
HD
D2
4036
3518
2889
À125
À113
À103
À3.00
À3.11
À3.44
À134
À118
a
Zn2L
H2
HD
D2
4096
3518
2889
À65
À58
À50
À1.56
À1.60
À1.67
À66.5
À61.0
À52.5
a Unobserved due to low signal-to-noise.
Correlations Between the Vibrational Frequency Shift
and Adsorption Enthalpy. Figure 4 shows the purely vibra-
tional Q-region of the spectrum for adsorbed H2 in M2L
materials, with frequencies shown as red shifts from the vibra-
tional mode Q(0) of gaseous H2, 4161 cmÀ1. Our observed
frequencies are summarized in Table 1 and are consistent with
the findings of other groups.11,24,35,44 At low concentration,
where only the primary site is filled, the red shift of the Q(0)
band almost doubles upon substitution of Ni for Zn in the
structure. The bands from Ni2L and Co2L are significantly more
red-shifted than the other members of the series. In contrast, the
bands associated with the secondary sites exhibit a much smaller
variation. This reduced dependence on the metal cation is to be
expected, since the secondary sites are located farther from the
metal ion (>4.8 Å compared to 2.6 Å).34 The slightly lower
frequency bands associated with each type of site arise from
ortho-H2. The reason for two or more such bands is the
anisotropic nature of the interaction potential, which leads to a
lifting of the degeneracy of the J = 1 rotational levels.34
Figure 4. Upper plot: the purely vibrational Q-region of the spectra of
H2 in M2L with the frequency scale referenced to that of the gas phase
Q(0) value, 4161 cmÀ1. Spectra correspond to those in Figure 1. Lower
plot: positive correlation between the frequency shifts of the lowest
concentration Q(0) band and reported isosteric enthalpies of adsorption
at zero surface coverage (data from refs 20, 30, 31, 36, 45).
and after air-exposure and observed an even more pronounced
effect, with the dominant room-temperature band changing from
4050 cmÀ1 in a pristine sample to 4125 cmÀ1 after exposing the
sample to air (see Figure S10). Again, the spectra from the
contaminated sample are similar to those published in ref 44.
We have observed this sensitivity to atmospheric moisture in
many different MOF samples. Whenever the sample is exposed
to air the intensities of the bands associated with the primary sites
are diminished to the point where secondary site bands dominate
the spectra. It is sensible that the primary site, being the most
reactive, is the one most affected by contaminants. On the basis
of the relative area of the two set of spectra in Figure 3 it appears
that contamination of the primary site also increases the intensity
of the bands associated with the secondary sites. This is most
likely due to an increase in the induced dipole moment via
Several groups have shown that the first H2 bands to appear are
those with the largest vibrational red shifts and that there is
a correlation between large red shifts and large adsorption
enthalpies.11,24,28 This was disputed, however, in a recent paper
water H2 interactions.
3 3 3
20315
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
wherein room temperature spectra from a diverse set of MOFs
displayed only minor variations.46 The shortcoming of room
temperature spectra is their tendency to exhibit broad features
resulting from overlapping bands of H2 in many different sites
(vide supra). In contrast, low temperature spectra acquired with
small concentrations of H2 contain bands corresponding to the
most favorable site and therefore can be directly compared with
isosteric enthalpies of adsorption (in the limit of zero coverage)
measured independently. Data compiled from this study and
reports from the literature are shown in Figure 4 (bottom). While
a positive correlation between binding energy and frequency shift
is not always obeyed, the overall the trend is clear: IR band red
shifts increase with adsorption enthalpy. The largest uncertain-
ties in these data correspond to the isosteric enthalpies of
adsorption, which show a variation on the order of 1 kJ/mol
among reports from different groups.45,47
Evidence for Intermolecular H2 H2 Interactions. From
3 3 3
inspection of the spectra in Figure 1 it is clear that bands assigned
to H2 at the primary site display small changes with increasing H2
concentration. This effect is most pronounced for the S(0) band
in the region 4300À4380 cmÀ1. For each of the materials we
observe a red shift of approximately 6 cmÀ1 as the concentration
is increased from less than 0.1 H2/M to greater than 3 H2/M.
Rather than a continuous change in frequency with concentra-
tion, each band is supplanted with a new band at the lower
frequency and the transfer of intensity is continuous with
increasing concentration. In the Q-region the frequency change
is smaller, between 2 and 5 cmÀ1, with the largest changes being
observed for the Co and Ni materials.
Intermolecular interactions between the adsorbed H2 are the
probable origin of these band shifts. The fact that the spectral
changes in the primary site bands are observed only for con-
centrations greater than 1 H2/M points to interactions between
molecules at the primary site with those at secondary sites. This
interpretation is consistent with the neutron diffraction
models,34 which show the distance between primary sites is
large (5.3 Å), while the distance between adjacent primary and
secondary sites is less than 2.9 Å, comparable to the kinetic
diameter of H2. These weakly attractive interactions increase the
total binding energy and alter the anisotropy of the orientational
potential energy surface of each site. The additional anisotropy
manifests itself through changes in the rotational J sublevels and
the observed large shifts of the S bands.
An additional result of increasing the H2 concentration is the
substantial broadening of the center-of-mass translational modes
in the region 4200À4300 cmÀ1. This is most recognizable in
Figure 2, where the full-width-at-half-maximum of the transla-
tional band in the largest concentration spectrum (top) is more
than double that in the smaller concentration spectra. This
behavior reflects the changes in the center-of-mass potential
energy surface at the primary site. The additional anisotropy
resulting from interactions with H2 on the secondary sites
increases the spread of the different translational modes, broad-
ening the observed band.
Figure 5. Fundamental (bold traces) and overtone (fine traces) por-
tions of the same spectrum for H2 in each of Mn2L, Mg2L, and Zn2L
loaded with 2 H2/M at 35 K. The frequency shifts of the overtone
spectra were divided by 2, and their intensities are magnified by a factor
of 4 for comparison with the fundamental bands.
Isotope Effects. In his seminal paper, Buckingham showed
that for a diatomic molecule with rotational constant Be and
2
harmonic frequency ωe the ratio of Be/ωe is unaffected by
isotopic substitution.48 Thus, to second-order perturbation the
fractional vibrational shift of different hydrogen isotopologues
should be equal, i.e., Δν /νfree = constant.
To examine the validity of the Buckingham model for describ-
ing the dynamics of H2 isotopologues in M2L, we acquired low
temperature spectra for each of the materials exposed to HD and
D2. The results, summarized in Table 1 and Figure S11, reveal
only small deviations from the Buckingham prediction with a
systematic increase in the fractional frequency shift with increas-
ing molecular mass. The excellent consistency among the
different isotopologues removes any possibility that the highly
red-shifted bands arise from the pairing mechanism proposed by
ref 35, since they accept that pairs should not be formed with D2.
The minor systematic increase in the fractional frequency shift
may be explained by vibrationalÀtranslational coupling, in which
the greater D2 mass leads to a smaller translational zero-point
energy and a deeper residence within the confining potential
energy well. The heavier isotopologues experience slightly larger
adsorption enthalpies and consequently larger fractional red
shifts. This enhanced binding has been observed in the adsorp-
tion isotherms for other MOFs,31,49 and it has been speculated
that it could be employed in separation technology.50À51
Overtone Bands. In the absence of increased anharmonicity
of an adsorbed H2, the Buckingham model also predicts the
vibrational shift of the overtone mode should be double that of
the fundamental, Δν(2r0) = 2 Δν(1r0).48 Discrepancies from
this behavior can be used to assess the degree to which surface
interactions mix the vibrational energy states of an adsorbed
molecule. Despite the value of this information, overtone bands
are rarely reported for adsorbed H2,54À56 and to our knowledge
they have not been discussed in the literature surrounding
MOFs. Detecting these weak transitions may be quite challeng-
ing; since ab initio calculations predict overtone intensities to be
only a few percent of those of the corresponding fundamentals.57,58
An advantage of the diffuse reflectance optical geometry used here is
We are aware of only one theoretical approach aimed at
modeling the spectral perturbation due to H2 H2 interactions
3 3 3
in a porous material. The paper by the Chabal group, discussed
above, describes vdW-DFT calculations that predict concentra-
tion-dependent Q-band frequency shifts of 21 and 16 cmÀ1 for
Zn2L and Mg2L, respectively.35 These are significantly larger
than our experimentally observed shifts of 2.0 and 1.6 cmÀ1 for
the primary site bands of the respective systems.
20316
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
the amplification of weak bands. This has afforded us the opportunity
to observe and compare H2 overtones in the ∼8000 cmÀ1 region.
To directly compare the red shifts of the overtone bands with
those of the corresponding fundamentals of H2 in the different
materials, we have prepared the reduced plots shown in Figure 5.
Only data for the Mg2L, Mn2L, and Zn2L are presented, as the
overtone bands from H2 in Ni2L were too weak to be observed
and Co2L has a strong framework absorption in this region. The
frequency shifts, summarized in Table 1, are in good agreement
with predictions from the Buckingham model, indicating that any
induced anharmonicity is minor. The overtone modes of H2 at the
primary site are remarkably intense, with integrated areas ranging
from 10% to 20% of that of the fundamental. This is significantly
higher than the 3% ratio observed for H2 in zeolites using the
transmission geometry.54 Interestingly, we did not detect overtone
bands arising from H2 at any secondary sites. While overtone
enhancement appears to be limited to the primary sites, it occurs
for all of the M2L materials, and it is even stronger for the HD and
D2 isotopologues. We are presently working on a theoretical
model to explain this large increase in the overtone intensity.
’ ACKNOWLEDGMENT
This work was partially funded by the American Chemical
Society Petroleum Research Fund and the Oberlin Department
of Chemistry and Biochemistry.
’ REFERENCES
(1) Ferey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F.; Dutour,
J.; Surble, S.; Margiolaki, I. Science 2005, 309, 2040.
(2) Dinca, M.; Dailly, A.; Liu, Y.; Brown, C. M.; Neumann, D. A.;
Long, J. R. J. Am. Chem. Soc. 2006, 128, 16876.
(3) Furukawa, H.; Miller, M. A.; Yaghi, O. M. J. Mater. Chem. 2007,
17, 3197.
(4) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.; Lamberti, C.;
Bordiga, S.; Lillerud, K. P. J. Am. Chem. Soc. 2008, 130, 13850.
(5) Ma, S. Q.; Eckert, J.; Forster, P. M.; Yoon, J. W.; Hwang, Y. K.;
Chang, J. S.; Collier, C. D.; Parise, J. B.; Zhou, H. C. J. Am. Chem. Soc.
2008, 130, 15896.
(6) Lin, X.; Telepeni, I.; Blake, A. J.; Dailly, A.; Brown, C. M.;
Simmons, J. M.; Zoppi, M.; Walker, G. S.; Thomas, K. M.; Mays, T. J.;
Hubberstey, P.; Champness, N. R.; Schroder, M. J. Am. Chem. Soc. 2009,
131, 2159.
(7) Koh, K.; Wong-Foy, A. G.; Matzger, A. J. J. Am. Chem. Soc. 2009,
131, 4184.
(8) Sumida, K.; Hill, M. R.; Horike, S.; Dailly, A.; Long, J. R. J. Am.
Chem. Soc. 2009, 131, 15120.
(9) Furukawa, H.; Ko, N.; Go, Y. B.; Aratani, N.; Choi, S. B.; Choi, E.;
Yazaydin, A. O.; Snurr, R. Q.; O’Keeffe, M.; Kim, J.; Yaghi, O. M. Science
2010, 329, 424.
(10) Bhatia, S. K.; Myers, A. L. Langmuir 2006, 22, 1688.
(11) Areꢀan, C. O.; Chavan, S.; Cabello, C. P.; Garrone, E.; Palomino,
G. T. ChemPhysChem 2010, 11, 3237.
5. SUMMARY
We have determined that for adsorbed hydrogen within the
M2L series the most highly red-shifted vibrational features arise
from isolated H2 molecules at the primary site associated with the
coordinatively unsaturated metals and not from a proposed
H2 H2 pairing mechanism.35 We have presented further
3 3 3
evidence that the magnitudes of these red shifts correlate with
the adsorption enthalpies of the sites. As expected, the secondary
sites, which become occupied only after the primary site is filled,
give rise to IR features that are largely insensitive to metal
substitution. Frequency shifts revealed by isotopologue substitu-
tion spectra (HD and D2) and overtone spectra show good
agreement with the Buckingham model predictions.48 From this
we can conclude that interactions with the primary site cause only
minor changes to the H2 anharmonicity. This is an important
consideration for accurately modeling the interactions at each
adsorption site. Finally, we observed small, but distinct, changes
in the primary site IR bands after population of the secondary
(12) Bae, Y. S.; Snurr, R. Q. Microporous Mesoporous Mater. 2010,
132, 300.
(13) Stephanie-Victoire, F.; Goulay, A. M.; de Lara, E. C. Langmuir
1998, 14, 7255.
(14) Dinca, M.; Long, J. R. Angew. Chem., Int. Ed. 2008, 47, 6766.
(15) Zavorotynska, O.; Vitillo, J. G.; Spoto, G.; Zecchina, A. Int. J.
Hydrogen Energy 2011, 36, 7944.
(16) Rosi, N. L.; Kim, J.; Eddaoudi, M.; Chen, B. L.; O’Keeffe, M.;
Yaghi, O. M. J. Am. Chem. Soc. 2005, 127, 1504.
(17) Dietzel, P. D. C.; Johnsen, R. E.; Blom, R.; Fjellvag, H. Chem.—
Eur. J. 2008, 14, 2389.
(18) Caskey, S. R.; Wong-Foy, A. G.; Matzger, A. J. J. Am. Chem. Soc.
2008, 130, 10870.
sites. We attribute this behavior to weak H2 H2 interactions;
3 3 3
however, the magnitude of the effect is much smaller than re-
cently published theoretical predictions.35 We hope this work
resolves some of the uncertainties surrounding this system and
inspires the development of an improved theoretical model to
describe H2 adsorption in future materials, be they established
materials with known structures or hypothetical candidates.
(19) Dietzel, P. D. C.; Blom, R.; Fjellvag, H. Eur. J. Inorg. Chem.
2008, 3624.
(20) Zhou, W.; Wu, H.; Yildirim, T. J. Am. Chem. Soc. 2008, 130, 15268.
(21) Bhattacharjee, S.; Choi, J. S.; Yang, S. T.; Choi, S. B.; Kim, J.;
Ahn, W. S. J. Nanosci. Nanotechnol. 2010, 10, 135.
(22) Dietzel, P. D. C.; Morita, Y.; Blom, R.; Fjellvag, H. Angew.
Chem., Int. Ed. 2005, 44, 6354.
(23) Dietzel, P. D. C.; Panella, B.; Hirscher, M.; Blom, R.; Fjellvag,
H. Chem. Commun. 2006, 959.
(24) Vitillo, J. G.; Regli, L.; Chavan, S.; Ricchiardi, G.; Spoto, G.;
Dietzel, P. D. C.; Bordiga, S.; Zecchina, A. J. Am. Chem. Soc. 2008, 130, 8386.
(25) Kazansky, V. B.; Borovkov, V. Y.; Karge, H. G. J. Chem. Soc.,
Faraday Trans. 1997, 93, 1843.
’ ASSOCIATED CONTENT
S
SupportingInformation. Powder X-ray diffraction patterns
b
of the samples; results of thermogravimetric analyses; mid-IR
spectra of the desolvated samples; table of characteristic IR absorp-
tion frequencies for each sample; mid-IR spectra of rehydrated
Zn2L; hydrogen adsorption isotherms; diffuse reflectance IR spectra
of H2 in Co2L at room temperature; diffuse reflectance IR spectra
of H2, HD, and D2 in the samples. This material is available free
of charge via the Internet at http://pubs.acs.org.
(26) Kazansky, V. B.; Jentoft, F. C.; Karge, H. G. J. Chem. Soc.,
Faraday Trans. 1998, 94, 1347.
(27) Borovkov, V. Y.; Serykh, A. I.; Kazansky, V. B. Kinet. Catal.
2000, 41, 787.
(28) Bordiga, S.; Vitillo, J. G.; Ricchiardi, G.; Regli, L.; Cocina, D.;
Zecchina, A.; Arstad, B.; Bjorgen, M.; Hafizovic, J.; Lillerud, K. P. J. Phys.
Chem. B 2005, 109, 18237.
(29) FitzGerald, S. A.; Churchill, H. O. H.; Korngut, P. M.; Simmons,
C. B.; Strangas, Y. E. Phys. Rev. B 2006, 73, 155409.
’ AUTHOR INFORMATION
Corresponding Author
stephen.fitzgerald@oberlin.edu; jesse.rowsell@oberlin.edu
20317
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318

Journal of the American Chemical Society
ARTICLE
(30) Bordiga, S.; Regli, L.; Bonino, F.; Groppo, E.; Lamberti, C.;
Xiao, B.; Wheatley, P. S.; Morris, R. E.; Zecchina, A. Phys. Chem. Chem.
Phys. 2007, 9, 2676.
(31) FitzGerald, S. A.; Allen, K.; Landerman, P.; Hopkins, J.;
Matters, J.; Myers, R.; Rowsell, J. L. C. Phys. Rev. B 2008, 77, 224301.
(32) Kong, L. Z.; Cooper, V. R.; Nijem, N.; Li, K. H.; Li, J.; Chabal,
Y. J.; Langreth, D. C. Phys. Rev. B 2009, 79, 081407.
(33) FitzGerald, S. A.; Hopkins, J.; Burkholder, B.; Friedman, M.;
Rowsell, J. L. C. Phys. Rev. B 2010, 81, 104305.
(34) Liu, Y.; Kabbour, H.; Brown, C. M.; Neumann, D. A.; Ahn,
C. C. Langmuir 2008, 24, 4772.
(35) Nijem, N.; Veyan, J. F.; Kong, L. Z.; Wu, H. H.; Zhao, Y. G.; Li,
J.; Langreth, D. C.; Chabal, Y. J. J. Am. Chem. Soc. 2010, 132, 14834.
(36) Rowsell, J. L. C.; Yaghi, O. M. J. Am. Chem. Soc. 2006, 128, 1304.
(37) FitzGerald, S. A.; Churchill, H. O. H.; Korngut, P. M.; Simmons,
C. B.; Strangas, Y. E. Rev. Sci. Instrum. 2006, 77, 093110.
(38) Bragg, S. L.; Brault, J. W.; Smith, W. H. Astrophys. J. 1982,
263, 999.
(39) Herzberg, G. Infrared and Raman Spectra; D. Van Nostrand
Company Inc.: New York, 1996.
(40) FitzGerald, S. A.; Forth, S.; Rinkoski, M. Phys. Rev. B 2002, 65.
(41) Herman, R. M.; Lewis, J. C. Phys. Rev. B 2006, 73, 155408.
(42) Gush, H. P.; Hare, W. F.; Allen, E. J.; Welsh, H. L. Can. J. Phys.
1960, 38, 176.
(43) McKellar, A. R.; Clouter, M. J. Can. J. Phys. 1994, 72, 51.
(44) Nijem, N.; Kong, L. Z.; Zhao, Y. G.; Wu, H. H.; Li, J.; Langreth,
D. C.; Chabal, Y. J. J. Am. Chem. Soc. 2011, 133, 4782.
(45) Murray, L. J.; Dinca, M.; Long, J. R. Chem. Soc. Rev. 2009, 38,
1294.
(46) Nijem, N.; Veyan, J. F.; Kong, L. Z.; Li, K. H.; Pramanik, S.;
Zhao, Y. G.; Li, J.; Langreth, D.; Chabal, Y. J. J. Am. Chem. Soc. 2010, 132,
1654.
(47) Thomas, K. M. Dalton Trans. 2009, 1487.
(48) Buckingham, D. A. Trans. Faraday Soc. 1960, 56, 753.
(49) Xiao, B.; Wheatley, P. S.; Zhao, X. B.; Fletcher, A. J.; Fox, S.;
Rossi, A. G.; Megson, I. L.; Bordiga, S.; Regli, L.; Thomas, K. M.; Morris,
R. E. J. Am. Chem. Soc. 2007, 129, 1203.
(50) Garberoglio, G.; DeKlavon, M. M.; Johnson, J. K. J. Phys. Chem.
B 2006, 110, 1733.
(51) Chen, B.; Zhao, X.; Putkham, A.; Hong, K.; Lobkovsky, E. B.;
Hurtado, E. J.; Fletcher, A. J.; Thomas, K. M. J. Am. Chem. Soc. 2008,
130, 6411.
(52) Wang, Y.; Bhatia, S. K. Mol. Simul. 2009, 35, 162.
(53) Nguyen, T. X.; Jobic, H.; Bhatia, S. K. Phys. Rev. Lett. 2010,
105, 085901.
(54) Forster, H.; Frede, W. Infrared Phys. 1984, 24, 151.
(55) Knippers, W.; Vanhelvoort, K.; Stolte, S. Chem. Phys. Lett. 1985,
121, 279.
(56) Bose, H.; Forster, H.; Frede, W. Chem. Phys. Lett. 1987, 138,
401.
(57) Karl, G.; Poll, J. D. J. Chem. Phys. 1967, 46, 2944.
(58) Schwartz, C.; Leroy, R. J. J. Mol. Spectrosc. 1987, 121, 420.
20318
dx.doi.org/10.1021/ja2071384 |J. Am. Chem. Soc. 2011, 133, 20310–20318