Metal-hydride anions of main group elements, such as BaH3 - and InH4 -, were generated by dissociating formate adducts of the respective metal formates. Upon activation, these adducts fragment by formate-ion ejection or by decarboxylation. For adducts of alkali-metal formates, the formate-ion ejection is the preferred pathway, whereas for those of alkaline-earth and group 13-15 metals, the expulsion of CO2 is the more favorable pathway. Decarboxylation is deemed to yield a metal-hydrogen bond presumably by a hydride transfer to the metal atom. For example, the decarboxylation of Al(η-OCOH)4 - and Ga(η-OCOH)4 - generated AlH4 - and GaH4 -, respectively. The initial fragment-ion with a H-M bond formed in this way from adducts of the heavier metals of group 13 (Ga, In, and Tl) undergo a unimolecular reductive elimination, ascribable to the "inert-pair" effect, to lower the metal-ion oxidation state from +3 to +1. As group 13 is descended, the tendency for this reductive elimination process increases. PbH3 -, generated from the formate adduct of lead formate, reductively eliminated H2 to form PbH-, in which Pb is in oxidation state zero. In the energy-minimized structure [H-Pb(η2-H2)]-, proposed as an intermediate for the process, a H2 molecule is coordinated with PbH- as a dihapto ligand. The formate adducts of strontium and barium produce monoleptic ions such as [M(0)(η2-O2CH)1]-, in which the formate ion is chelated to a neutral metal atom. The bismuth formate adduct undergoes a double reductive elimination process whereby the oxidation state of Bi is reduced from +3 to +1 and then to -1. Upon activation, the initially formed [H-Bi-H]- ion transforms to an anionic η2-H2 complex, which eliminates dihydrogen to form the bismuthide anion (Bi-).
Metal-hydride anions of main group elements, such as BaH3 - and InH4 -, were generated by dissociating formate adducts of the respective metal formates. Upon activation, these adducts fragment by formate-ion ejection or by decarboxylation. For adducts of alkali-metal formates, the formate-ion ejection is the preferred pathway, whereas for those of alkaline-earth and group 13-15 metals, the expulsion of CO2 is the more favorable pathway. Decarboxylation is deemed to yield a metal-hydrogen bond presumably by a hydride transfer to the metal atom. For example, the decarboxylation of Al(η-OCOH)4 - and Ga(η-OCOH)4 - generated AlH4 - and GaH4 -, respectively. The initial fragment-ion with a H-M bond formed in this way from adducts of the heavier metals of group 13 (Ga, In, and Tl) undergo a unimolecular reductive elimination, ascribable to the "inert-pair" effect, to lower the metal-ion oxidation state from +3 to +1. As group 13 is descended, the tendency for this reductive elimination process increases. PbH3 -, generated from the formate adduct of lead formate, reductively eliminated H2 to form PbH-, in which Pb is in oxidation state zero. In the energy-minimized structure [H-Pb(η2-H2)]-, proposed as an intermediate for the process, a H2 molecule is coordinated with PbH- as a dihapto ligand. The formate adducts of strontium and barium produce monoleptic ions such as [M(0)(η2-O2CH)1]-, in which the formate ion is chelated to a neutral metal atom. The bismuthformate adduct undergoes a double reductive elimination process whereby the oxidation state of Bi is reduced from +3 to +1 and then to -1. Upon activation, the initially formed [H-Bi-H]- ion transforms to an anionic η2-H2 complex, which eliminates dihydrogen to form the bismuthide anion (Bi-).
Binary hydrides are
well-known in inorganic chemistry.[1] Alkali
and alkaline-earth metals, except beryllium
and magnesium, form ionic hydrides with discrete H– ions. By contrast, the hydrides of the heavier main group metals
and of the nonmetals are covalent in character. Boron, in particular,
forms a broad array of covalent hydrides (boranes). In addition, there
are many complex hydrides such as NaBH4 and LiAlH4, which are important reducing agents in the chemical synthesis.
The hydrides of calcium, strontium, and barium, doped with divalent
europium, exhibit interesting luminescence properties.[2,3] Moreover, metal hydrides have attracted serious research attention
because they have been detected in stellar atmospheres, and they have
been suggested as potential hydrogen-storage materials.[4−6] For example, solid-state magnesium hydride is able to store hydrogen
reversibly.[7] Thus, novel, multinuclear
neutral and ionic magnesium hydride complexes are being investigated
rigorously as model systems for hydrogen-storage materials.[8,9]Because metal-hydride bonds are highly susceptible to air
and moisture,
the handling of hydrides, especially for analytical purposes, is a
challenging task. However, analytical techniques such as photoelectron
spectroscopy,[10−16] nuclear magnetic resonance,[2,17] a variety of vibrational
spectroscopies,[3,18−20] and X-ray diffraction[21] methods have been used to study metal hydrides
under inert and anhydrous conditions.In general, the gas-phase
generation and analysis of complex hydride
anions have been accomplished only under very rigorous experimental
conditions. For example, AlH–-type clusters have been generated
in a discharge between an anode and a grounded aluminum sample cathode
engulfed with hydrogen gas at 200 psi.[10,22−24] Mass spectra of a number of diatomic transition-metal hydrides have
been recorded under field evaporation[25] by Cs-beam sputtering on metal surfaces while spraying hydrogen
or ammonia[26] or by spraying an aqueous
solution of the metallic chloride into a 2200 K methane–oxygen
flame.[27] Extensive investigations have
been carried out with gas-phase hydride species[28] generated from silicon,[29] boron,[30,31] aluminum,[31,32] silver,[33−38] copper,[35,36,39,40] magnesium,[41] vanadium,[42] chromium,[42] iron,[42] cobalt,[42] and molybdenum;[42] however, there is still a paucity of information
about hydrides of many other metals.The formation of anion
adducts of inorganic salts under electrospray
ionization (ESI) conditions is a well-known phenomenon. Such adducts,
which bear a metal–organic framework, are routinely used for
the mass-scale calibration of mass spectrometers. The formate ion
is the smallest carboxylate ion that can coordinate with metal cations
in a monodentate or bidentate fashion.[43] Generally, collision-induced dissociation (CID) investigations of
carboxylate salt clusters are not very rewarding because most spectra
show only a single product-ion peak for the ejected carboxylate anion.
Similar results have been observed under positive ionization conditions.
For example, the CID spectrum of the sodium adduct of sodium benzoate
showed only one product-ion peak at m/z 23 for the ejected sodium ion.[44] On the
other hand, certain inorganic carboxylate complexes have been shown
to undergo decarboxylation and generate organometallic ions.[45] An extensive study of CID spectra of formate
adducts of main group metal formates revealed that the decarboxylation
reaction shows distinct periodic trends. Herein, we report not only
the formation of metal–hydrogen bonds during the CO2 ejection process of activated formate adducts of metal formates
but also that the core metal atom sometimes undergoes a reduction
of its oxidation state. For example, Bi– can be
generated in this way by two consecutive reductive elimination steps.
Results
and Discussion
Alkali-Metal Formates
Under negative-ion-generating
ESI conditions, a range of gaseous adducts form between the formate
anion and the ion pairs of low-molecular-weight carboxylate anions
with monovalent metal cations. For example, an ESI mass spectrum recorded
from an aqueous solution of lithiumformate shows a series of peaks
representing the [Li(OCOH)]− and [Li(OCOH)]2– series
of adducts (Figure A). Such complexes, upon mass selection and activation under tandem
mass spectrometry (MS) conditions, readily eject ion pairs of the
neutral salt to generate the formate anion, which registers as a peak
at m/z 45 (1). For
instance, the CID spectrum recorded from the m/z 97 ion [Li(η2-O2CH)2]− (2a) showed an intense peak at m/z 45 (1) (Figure B). According to density functional
theory (DFT) calculations, the two formate anions in the m/z 97 ion coordinate with the central lithium cation
in a bidentate manner (Supporting Information Table S1 2a). In addition to the m/z 45 peak, the spectrum showed a peak at m/z 53 (3a) to indicate the
existence of a second fragmentation channel (Figure B). Accurate mass measurements (observed
53.0224 u; calculated for LiC1O2H2 53.0220 u) confirmed that this ion corresponds to the formula LiC1O2H2. Presumably, this m/z 53 ion (3a) originates by the decarboxylation
of the precursor [Li(η2-O2CH)2]− complex. The m/z 53 ion (3a) generated in this way is expected to bear
a newly formed H–Li bond, and the structure of the ion can
be represented as [H–Li(η2-O2CH)]− because this ion was demonstrated to react with water
to give an ion of m/z 69 in a manner
expected for a metal hydride, details of which will be discussed in
another paper. For the formation of the m/z 53 ion (3a), one of the formate moieties
in the precursor complex, upon activation, should first become monodentate
and then donate a hydride ion to the lithium atom by pivoting the
H–C bond in the formate moiety. The subsequent expulsion of
a neutral CO2 molecule then generates a H–Li bond.
Analogous decarboxylations have been described previously for the
acetate adduct of magnesium acetate[46] and
recently for the formate adduct of copper formate.[34] Although less conspicuous than the peak for the [H–Li(η2-O2CH)]− ion (3a), the product-ion spectra recorded from the m/z 113 (2b), 129 (2c), and 175
(2d) ions for the respective formate adducts of Na, K,
and Rbformates also showed peaks at m/z 69 (3b), 85 (3c), and 131 (3d) for the respective purported [H–Na(η2-O2CH)]−, [H–K(η2-O2CH)]−, and [H–Rb(η2-O2CH)]− ions (Supporting Information Figures S1, S2, and S3D). However, the product-ion
spectrum recorded from m/z 223 (2e) for the formate adduct of Csformate did not show a peak
at m/z 179 for [H–Cs(η2-O2CH)]− even under the most
sensitive instrumental settings (Supporting Information Figure S3E). On the basis of these results, we propose that [M(η2-O2CH)2]− ions generated
from alkali-metal formates, upon activation, undergo fragmentation
by two competitive mechanisms (Scheme ), the more favorable of which is the ejection of a
neutral ion pair to generate the formate ion (pathway “b,” Scheme ). In fact, the role
of the minor fragmentation pathway becomes less significant as the
size of the alkali metal increases down the group (pathway “a,” Scheme ). Apparently, this
observation is the first indication of a periodic trend in the fragmentation
patterns of gas-phase metal formate adducts.
Figure 1
MS[1] spectrum recorded from an aqueous
solution of lithium formate (A) (the peaks in blue represent the [Li(OCOH)]− series of ions, and those in red depict the [Li(OCOH)]2– series) and the product-ion spectrum of the m/z 97 ion for [Li(η2-O2CH)2]− recorded at a transfer
collision energy setting of 5.5 eV (B) on a SYNAPT G2 instrument under
negative-ion-generating ESI mode.a
Scheme 1
Fragmentation Pathways of Formate Adducts of Alkali-Metal
Formates
by Two Competitive Mechanisms
MS[1] spectrum recorded from an aqueous
solution of lithiumformate (A) (the peaks in blue represent the [Li(OCOH)]− series of ions, and those in red depict the [Li(OCOH)]2– series) and the product-ion spectrum of the m/z 97 ion for [Li(η2-O2CH)2]− recorded at a transfer
collision energy setting of 5.5 eV (B) on a SYNAPT G2 instrument under
negative-ion-generating ESI mode.aIn a somewhat analogous series of experiments, conducted
with acetate
adducts of alkali-metalacetates, Jacob et al.[47] have previously shown that such gas-phase acetate adducts
fragment primarily to yield the acetate ion. An additional fragmentation
channel, which eliminates ketene, has also been noted for acetate
adducts such as [Mo2O6(OCOCH3)]− and [Ca(O2CCH3)3]−.[47−49] However, the ketene elimination is not a significant
dissociation pathway for the acetate adducts of magnesium acetate.
Although insignificant CO2 losses have been noted for the
acetate adducts of sodium and potassium salts, the sole fragmentation
reaction of [Li(O2CCH3)2]− was the acetate-ion loss.[47] By contrast,
our study revealed that the formate adduct [Li(η2-O2CH)2]− (2a) underwent a more significant CO2 loss (decarboxylation)
(Figure B).
Formates
of the Alkaline-Earth Metals
Similar to the
behavior of alkali-metal cations, the divalent alkaline-earth-metal
cations are also known to generate a range of gaseous adducts with
low-molecular-weight carboxylate anions under negative-ion-generating
ESI conditions.[50] For example, ESI mass
spectra recorded from aqueous solutions of alkaline-earth-metal formates
showed a series of peaks representing [M(OCOH)2]− adducts
(Figure ).
Figure 2
MS[1] spectrum recorded from an aqueous
solution of magnesium formate which shows [Mg(OCOH)2]− series
of ions (A) and the product-ion spectrum of the m/z 159 ion for [24Mg(η2-O2CH)3]− (4a) recorded at a transfer collision energy setting of 12 eV (B) on
a SYNAPT G2 instrument under negative-ion-generating ESI mode.
MS[1] spectrum recorded from an aqueous
solution of magnesiumformate which shows [Mg(OCOH)2]− series
of ions (A) and the product-ion spectrum of the m/z 159 ion for [24Mg(η2-O2CH)3]− (4a) recorded at a transfer collision energy setting of 12 eV (B) on
a SYNAPT G2 instrument under negative-ion-generating ESI mode.
Magnesium and Calcium Formates
Tandem MS experiments
conducted with formate adducts of alkaline-earth-metal formates demonstrated
that the periodic trend of the relative contributions of the two major
fragmentation pathways to the overall fragmentation is completely
reversed compared to that observed for the alkali-metal formates.
In general, decarboxylation is the preferred dissociation channel
for the group 2 metal formates. For example, in the product-ion spectra
of the mass-selected m/z 159 ion
for [Mg(η2-O2CH)3]− (4a) and the m/z 175
ion for [Ca(η2-O2CH)3]− (4b), the base peaks were observed at m/z 115 (5a) and 131 (5b), respectively (Figures and S4).In fact,
under conditions of higher collision energy, a second decarboxylation
could be instigated on the adduct of Mgformate (Figure ) and Ca formate (Supporting Information Figure S4). In an analogous
experiment conducted with the formate adduct of magnesiumchloride,
Khairallah and O’Hair have demonstrated that its decarboxylation
leads also to a magnesium hydride anion.[41] Although it was mentioned that the competing loss of HCOO– does not appear to operate, it is most likely that a peak for HCOO– was not observed because the experiment was conducted
on an ion-trap mass spectrometer, which is not efficient in collecting
low-mass product ions. Although complex Mghydrides such as LiMgH3, NaMgH3, and KMgH3 have been synthesized
under certain experimental conditions and evaluated by spectroscopic
methods, no MS data from such compounds have been reported.[3,51,52]The acetate adduct of magnesiumacetate has been reported to undergo
an analogous CO2 loss and generate organomagnesates.[46,47] In contrast to the fragmentation of magnesium acetate, which underwent
only one CO2 loss under ion-trap MS conditions, we were
able to instigate two consecutive decarboxylations upon collisional
activation of [24Mg(η2-O2CH)3]− (4a) (Figure ). We must mention that the acetate adduct
of silver acetate has been demonstrated to undergo two consecutive
CO2 losses in an analogous manner to generate the dimethyl
argentate anion on a quadrupole ion-trap instrument.[53]
Strontium Formate
A spectrum recorded
from an aqueous
solution of strontium formate showed a series of peaks representing
the [Sr(OCOH)2]− and [Sr(OCOH)2]2– adducts
(Figure A). When the m/z 223 ion (4c) for [Sr(η2-O2CH)3]− was mass-selected
and subjected to low-energy CID (2 eV), the product-ion spectrum recorded
showed an intense peak at m/z 179
(5c) for the elimination of CO2 from the precursor
ion, whereas the peak at m/z 45
(1) was negligibly small (Figure B). Similar to the mechanism proposed for
the decarboxylation of [Li(η2-O2CH)2]− (2a) (Scheme ) for the formation of the m/z 179 ion (5c), one of the formate
moieties attached to the Sr cation is deemed to rearrange and donate
its H atom to Sr, forming a H–Sr bond.
Figure 3
MS[1] spectrum recorded from an aqueous
solution of strontium formate (A) (the peaks in blue represent the
[Sr(OCOH)2]− series of ions, and those in red depict the
[Sr(OCOH)2]2– series) and the product-ion spectra of the m/z 223 ion (4c) for [Sr(η2-O2CH)3]− recorded
at a transfer collision energy setting of 2 eV (B) and 25 eV (C) on
a SYNAPT G2 instrument under negative-ion-generating ESI mode.
MS[1] spectrum recorded from an aqueous
solution of strontium formate (A) (the peaks in blue represent the
[Sr(OCOH)2]− series of ions, and those in red depict the
[Sr(OCOH)2]2– series) and the product-ion spectra of the m/z 223 ion (4c) for [Sr(η2-O2CH)3]− recorded
at a transfer collision energy setting of 2 eV (B) and 25 eV (C) on
a SYNAPT G2 instrument under negative-ion-generating ESI mode.To shed more light on the formation
of a H–Sr bond by this
rearrangement, a sample of Sr(O2CD)2 was made,
and its adduct spectrum was recorded. The CID spectrum of the m/z 226 ion for [Sr(η2-O2CD)3]− showed a peak at m/z 182 for the loss of CO2 from
the precursor, which confirmed the transfer of a deuteride to the
Sr atom prior to decarboxylation (Supporting Information Figure S5B). Moreover, under more energetic CID conditions (25 eV), three consecutive losses of CO2 could be instigated
from the [Sr(η2-O2CH)3]− (4c) precursor ion (Scheme ; Figure C). Of the peaks observed at m/z 179 (5c), 135 (6c),
and 91 (7a) for the three consecutive decarboxylations,
the peak at m/z 91 (7a) is believed to represent the trihydridostrontate anion [SrH3]− ion (note: in Figure C, the ions for the peaks at m/z 107, 123, 139, 151, 167, and 195 originate from
the interaction of residual water with ions bearing H–Sr bonds).
A spectrum recorded from the m/z 235 ion for [Sr(η2-18O2CH)3]− generated from a sample of Sr(18OC18OH)2, which showed peaks at m/z 187, 139, and 91 for three consecutive C18O2 losses, supported the proposed mechanism (Supporting Information Figure S6). To the best
of our knowledge, an MS peak has not been reported for the [SrH3]− ion (7a), although the existence
of this species has been recognized.[54] Previously,
a procedure to generate [SrH3]− (7a) has been described as a method for the detection of 80Sr in environmental samples.[55] To form [SrH3]− by that procedure,
the samples were initially converted to SrCO3, which was
then reduced by Ti powder in vacuum to metallic Sr. After distillation,
Sr was allowed to react with H2 at 450 °C, and the
SrH2 pellets produced in this way were used as the sputter
cathode for the generation of the [SrH3]− ion.[55]
Scheme 2
Fragmentation Pathways
of Formate Adducts of Alkaline-Earth-Metal
Formates
Although we were able
to effect three consecutive CO2 losses from [Sr(η2-O2CH)3]− (4c), we could instigate only two
consecutive CO2 losses from [Ca(η2-O2CH)3]− (4b) and
[Mg(η2-O2CH)3]− (4a). Evidently, the periodic trend in group 2 is that
a complete decarboxylation to the ultimate metal-hydride anion can
be achieved only when the size of the core metal atom becomes sufficiently
large.
Barium Formate
A spectrum recorded from an aqueous
solution of bariumformate showed a series of peaks at m/z 273, 501, 729, 957... for [Ba(OCOH)2]− complexes
(Supporting Information Figure S7A). On
the basis of DFT calculations carried out at B3LYP level with 6-311G
and LanL2DZ basis sets, each formate ion can chelate to the core Ba
cation in a bidentate fashion (Supporting Information Table S1 4d). Similar to the fragmentation pathway
followed by the formate adduct of Srformate, the m/z 273 ion (4d) for [Ba(η2-O2CH)3]− upon activation
underwent three consecutive decarboxylations. The peak observed at m/z 141 (7b) represents the
[BaH3]− ion (Supporting Information Figure S7C).Although complex barium hydrides
such as LiBaH3 have been reported,[54,56] no recorded information was found on mass spectrometric identification
of [BaH3]−. In summary, all tested alkaline-earth-metalformate adducts underwent at least two decarboxylations, but those
with central cations of larger radius, such as Sr2+ and
Ba2+, decarboxylated completely to generate previously
unreported MH3– ions (7, Scheme ).
Group 13 Formates
A spectrum recorded from an aqueous
solution of aluminum formate showed a peak at m/z 207 (17) for the [Al(η-OCOH)4]− complex. Upon mass selection and activation,
the m/z 207 (17) ion
initially lost either a molecule of Al(η2-O2CH)3 to release the HCOO– ion (m/z 45) (1) or CO2 to produce an ion of m/z 163 (9a) (Figure A). Under more energetic conditions, the m/z 207 ion (17) underwent four consecutive decarboxylations, and the product-ion spectrum recorded
showed a peak at m/z 31 (12a). Albeit small, this peak is believed to represent the tetrahydridoaluminate
ion ([AlH4]−; alanate anion). The alanate
anion is one of the most familiar reducing agents in synthetic chemistry.
Its existence under certain experimental conditions has been noted
previously.[31] For example, its vibrational
spectrum has been recorded only under cryogenic conditions by embedding
it in a solid hydrogen matrix.[18,57−59] In general, many Al–H cluster anions have been generated
under pulsed-arc discharge conditions[23,60] and laser
ionization methods.[58]
Figure 4
Product-ion spectrum
of the m/z 207 [Al(η-OCOH)4]− ion (8a) recorded at a transfer
collision energy setting of 15
eV on a SYNAPT G2 instrument (A) and a plot of relative ΔG (at 298.15 K and 1 atom) values calculated by B3LYP/6-311++G(2d,2p)
for intermediates and transition states participating in CO2 and Al(η2-O2CH)3 losses from
[Al(η-OCOH)4]− (B).
Product-ion spectrum
of the m/z 207 [Al(η-OCOH)4]− ion (8a) recorded at a transfer
collision energy setting of 15
eV on a SYNAPT G2 instrument (A) and a plot of relative ΔG (at 298.15 K and 1 atom) values calculated by B3LYP/6-311++G(2d,2p)
for intermediates and transition states participating in CO2 and Al(η2-O2CH)3 losses from
[Al(η-OCOH)4]− (B).According to DFT calculations, the addition of
HCOO– to Al(η2-O2CH)3 (16) is an exergonic reaction (Figure B). Unlike the coordination
in adducts of alkali and
alkaline-earth formates, the formate ions in [Al(η-OCOH)4]− (8a) appear to bind to the
core Al atom in a monodentate manner to form a tetrahedral complex
(Supporting Information Table S1 8a). During the fragmentation process, one of the H atoms
in a formate moiety in the complex pivots toward the aluminum cation
(17, Supporting Information Table S1 17) and then via a four-member transition
state (TS, Supporting Information Table S1 TS) transfers a hydrogen atom to the Al core (Figure B). Then the ion–neutral
complex (18), formed in this way, dissociates by releasing
CO2 from the m/z 163
product-ion (9a), which subsequently undergoes three
more decarboxylation steps until the alanate anion is formed as the
final product.Analogous to the fragmentation behavior of [Al(η-OCOH)4]− (8a), the m/z 249 ion (8b) for [69Ga(η-OCOH)4]− also underwent four consecutive decarboxylations
upon activation (Figure ). The peak at m/z 73 (12b) is believed to represent the tetrahydridogallate ion ([69GaH4]−; Figure ). Gallium has two isotopes (69Ga and 71
Ga) at an abundance ratio of about 60:40. The relative intensities
of peaks recorded for isotopologues of [GaH4]− aided the characterization of this ion (Figure ).
Figure 5
Product-ion spectrum recorded from mass-selected
formate adducts
of Ga(η2-O2CH)3 isotopologues
(8b) at a transfer collision energy setting of 12 eV
on a SYNAPT G2 instrument (the mass-isolation window of the quadrupole
analyzer was set wide enough to transmit both m/z 249 and 251 ions).
Product-ion spectrum recorded from mass-selected
formate adducts
of Ga(η2-O2CH)3 isotopologues
(8b) at a transfer collision energy setting of 12 eV
on a SYNAPT G2 instrument (the mass-isolation window of the quadrupole
analyzer was set wide enough to transmit both m/z 249 and 251 ions).Analogously, the formate adduct of indium(III) formate [115In(η-OCOH)4–, m/z 295, 8c] also underwent
four successive
decarboxylations [m/z 295 (8c) → 251 (9c) → 207 (10c) → 163 (11c) → 119 (12c)]
resulting in the formation of the [115InH4]− ion (m/z 119, 12c) (Scheme , Figure B). The
peak at m/z 207 in the spectrum
of the adduct of indium(III) formate was accompanied by a significant
satellite peak at m/z 205 (13b), signifying the existence of a competing fragmentation
mechanism. Apparently, the m/z 251
ion (9c), formed by the initial loss of CO2, which bears a H–In bond, can fragment further by two distinct
mechanisms (Scheme ). In addition to the anticipated decarboxylation to form an ion
of m/z 207 (10c; Scheme ), the m/z 251 ion (9c; Scheme ) can also lose elements of HCOOH to generate
an ion of m/z 205 (13b). The latter ion originates from an unanticipated fragmentation
mechanism. To form the m/z 205 ion
(13b), the [H–In(III)(η-OCOH)3]− ion (9c) should undergo a reductive
elimination reaction. Even though there are many examples of reductive
eliminations from organometallic cations in the gas phase,[61,62] the reductive elimination of elements of HCOOH described here is
rather unusual. Apparently, the m/z 251 ion (9c) undergoes simultaneous oxidation to H2 (or HCOOH), whereas the indium core undergoes a reduction
from oxidation state +3 to +1 (Scheme ). Although attempts were not made to find the rate-limiting
steps, thermochemical data from the literature show that the loss
of H2 and CO2 is less endothermic than the elimination
of an intact molecule of HCOOH.[63,64] The m/z 205 ion (13b) formed in this way
can then fragment further by eliminating In(I)(η-OCOH) to form
the formate ion (m/z 45) (1) or eject one more CO2 to yield m/z 161 ion for [H–In(I) (η-OCOH)]− (14a) (Supporting Information Figure S8C). On the other hand, the m/z 207 ion for [H2–In(III)(η-OCOH)2]− (10c) can undergo two more
successive CO2 losses to yield the m/z 119 ion for In(III)H4– (12c) (Supporting Information Figure
S8B). The tetrahydridoindium anion (12c) was generated
previously only under very rigorous conditions.[65] Even the neutral species InH3 was observed only
by matrix-isolation experiments conducted below −90 °C.[66−68] As a prerequisite for designing catalytic systems, the H2 production from formic acid on a Pd(111) surface has been investigated
by Wang et al. by DFT computations.[69] By
a process similar to that described in the present study, the use
of hydrido cuprates, generated from copper formate, as catalysts to
transform formic acid to hydrogen and carbon dioxide has been reported
recently by Zavras at al.[39] Thus, the reductive
elimination process described here could provide additional insights
into the development of metal-catalyzed decomposition of formic acid
via formates.
Scheme 3
Fragmentation Pathways of Formate Adducts of Group 13 Metal
Formates
Figure 6
Product-ion spectra recorded from the m/z 295 ion (8c) for [115In(η-OCOH)4]− at a transfer collision
energy setting
of 2 (A) and 28 eV (B); product-ion spectra recorded from the m/z 385 ion (8d) for [205Tl(III)(η2-O2CH)4]− at a transfer collision energy setting of 0
(C), and 10 eV (D); product-ion spectra recorded from the m/z 343 ion for [208Pb(η-OCOH)3]− at a transfer collision energy setting
of 0 (E), and 20 eV (F); product-ion spectra recorded from the m/z 389 ion (28) for [209Bi(III)(η-OCOH)4]− at
a transfer collision energy setting of 0 (G), and 15 eV (H). All spectra
were recorded on a SYNAPT G2 instrument.
Product-ion spectra recorded from the m/z 295 ion (8c) for [115In(η-OCOH)4]− at a transfer collision
energy setting
of 2 (A) and 28 eV (B); product-ion spectra recorded from the m/z 385 ion (8d) for [205Tl(III)(η2-O2CH)4]− at a transfer collision energy setting of 0
(C), and 10 eV (D); product-ion spectra recorded from the m/z 343 ion for [208Pb(η-OCOH)3]− at a transfer collision energy setting
of 0 (E), and 20 eV (F); product-ion spectra recorded from the m/z 389 ion (28) for [209Bi(III)(η-OCOH)4]− at
a transfer collision energy setting of 0 (G), and 15 eV (H). All spectra
were recorded on a SYNAPT G2 instrument.Further scrutiny
of recorded data from fragmentation of metal formate
adducts revealed that the reductive elimination process described
above for the indium adduct is in fact another general periodic phenomenon.
The facile formation of a lower oxidation state of the central metal
atom by reductive elimination of formate adducts can be ascribed to
the so-called inert-pair effect.[68,70] It is known
that group 13 elements exhibit +3 and +1 oxidation states. Our results
support the generalization that the stability of the +1 oxidation
state of group 13 elements increases in the following sequence: Al(I)
< Ga(I) < In(I) < Tl(I).[68] The
inert-pair effect is a general phenomenon observed in the metals of
the p block,[68] and our results exemplify
that descending along a main group in the periodic table, the influence
of this effect increases. For example, the spectrum of [Al(η-OCOH)4]− (8a) showed no peaks for
the products of a reductive elimination reaction, and that of [Ga(η-OCOH)4]− (8b) showed only a low-intensity
peak at m/z 159 (13a) for the simultaneous loss of H2 and CO2 (Figure ). By contrast, the
spectrum of the formate adduct of indium formate showed a significant
peak at m/z 205 (13b) for the reductive elimination product (Figure A). The reductive elimination trend was especially
pronounced with [Tl(III)(η2-O2CH)4]− (m/z 385, 8d): the base peak in the spectrum appeared at m/z 295 (13c) for the reductive
elimination product (Figure C). Apparently, the first-stage decarboxylation [H–Tl(III)(η-OCOH)3]− ion (9d) is such a short-lived
species that a peak for it is not observed in the mass spectrum: it
immediately undergoes reductive elimination to restore the 6S2 inert pair and yield the [Tl(I)(η2-O2CH)2]− ion (13c)
represented by the base peak at m/z 295 in Figure C.
Computations predict that the transient [H–Tl(III)(η-OCOH)3]− ion (9d) bears a tetrahedral
coordination in which the three formate groups are attached to the
thallium core in a monodentate manner (Supporting Information Table S1 9d). By contrast, in the
precursor [Tl(III)(η2-O2CH)4]− ion (8d), the four formate moieties
are chelated in a bidentate fashion (Supporting Information Table S1 8d). In the m/z 295 ion (13c) formed by the reductive
elimination reaction, the two formate groups are also attached to
the central atom in a bidentate mode. Upon further activation of the m/z 295 ion (13c), we were
able to instigate only one more CO2 loss and generate an
ion of m/z 251 (14c) for [H–Tl(I)(η2-O2CH)]− (Supporting Information Figure S9 and
Scheme S1). In fact, the elimination of Tl(I)(O2CH) to
generate a formate ion is the preferred fragmentation pathway of the m/z 295 ion (Supporting Information Figure S9). In this respect, the performance of
the m/z 295 [Tl(I)(η2-O2CH)2]− ion (13c) is very similar to the fragmentation properties exhibited by the
alkali-metal formate adducts. In fact, this is not surprising because
it is known in inorganic chemistry that the properties of Tl(I) salts
resemble somewhat those of their sodium and potassium counterparts.[68]Relative reaction energies computed for
the reductive elimination
reaction [H–M(III)(OCOH)3]− →
[M(I)(OCOH)2]− + CO2 + H2 indicated that for thallium, the reduction is exergonic,
whereas for aluminum, gallium, and indium, the conversions are endergonic
(Table ). In fact,
the endergonicity of the reaction progressively decreases when moving
downward along the main group 3 elements (Table ). This periodic trend is clearly manifested
in the CID spectra recorded from the adducts of group 13 elements.
Table 1
Reaction Energies (kJ/mol) of [H–M(III)(OCOH)3]− → [M(I)(OCOH)2]− + CO2 + H2 Conversion (M = Al,
Ga, In, or Tl) at 1 atm and 298.15 K
The formate
adducts of Pb, Sb, and Bi formates also undergo decarboxylation upon
activation. For example, the mass-selected m/z 343 ion (19) for [208Pb(II)(η-OCOH)3]− underwent three successive decarboxylations
[m/z 343 (19) →
299 → 255 → 211 (20)] resulting in the
formation of the [208PbH3]− ion (m/z 211, 20)
(Figures E, Supporting Information Figure S10B). According
to computations, the singlet state is the energetically favored form
of [PbH3]− (Supporting Information Table S2). Interestingly, the PbH3– ion, upon mass-selection and activation, undergoes
an additional reductive elimination of dihydrogen to yield the [208Pb(0)H]− ion (m/z 209, 22, Scheme , Supporting Information Figure S11). An analogous dihydrogen loss could not be instigated
from the activated AlH4–, GaH4–, InH4–, SrH3–, or BaH3– ions. In the process of reductive elimination, the oxidation state
of Pb changes from +2 to zero. The observation that the intensity
of the m/z 209 (22)
peak for [208PbH]− progressively increases
when the collision energy is raised, supported the proposed H2 elimination mechanism (Scheme ). Apparently, the transformation of [PbH3]− to [PbH]− takes place via
the transition state TS to form
an intermediate complex (HPb)−–(H2) (21), which then loses H2. The [PbH3]− ion could be isolated by ion-mobility
separation and subjected to fragmentation in the Transfer collision
cell of the SYNAPT G2 instrument (Supporting Information Figure S11). According to calculations, upon activation, two of
the H–Pb bonds in [PbH3]− undergo
an elongation with a concomitant decrease of the H–Pb–H
bond angle from 91.4° to 33.6° (Supporting Information Table S3). In the optimized structure, the internuclear
distance between two participanthydrogen atoms in the transition
species TS is 1.1 Å. Once
it turns to complex 21, the internuclear distance between
the two hydrogen atoms is short enough (0.8 Å) for us to consider
that the two hydrogen atoms are now linked by a chemical bond. Many
examples are known of a H2 molecule acting as a dihapto
ligand and occupying the coordination sphere of a metal atom in a
low oxidation state.[71,72] Such complexes in which dihydrogen
acts as a ligand are generally known for transition metals.[80] However, even some main group elements have
been reported to form transient dihydrogen complexes.[74,73]
Scheme 4
Formation of the [208PbH]− Ion Starting
from [208Pb(η-OCOH)3]− Ion (Spin States, Relative Gibbs Free Energies, and Structures Were
Evaluated at DFT/B3LYP Level of Theory)
Presumably, the (HPb)−–(H2)
interaction in the HPb (H2)− complex
is a weak “side-on” donation of the bonding σ
electrons of the dihydrogen to the metal atom.[74] The zero oxidation state of the Pb atom facilitates the
interaction. The computed vibrational frequency of 536 cm–1 for the HPb–(H2) bond compared to 1423 cm–1 for the H–PbH2 bond indicated that
the former is a weak interaction [the computed binding energy on the
singlet surface was 49.8 kJ/mol; an optimized structure for a triplet
complex was not found (Supporting Information Table S2)]. The dH–H distance
of 0.8 Å suggests that the intermediate [PbH3]− ion could be considered as a “Kubas-type”
complex in which a dihydrogen molecule is coordinated as a dihapto
ligand to a PbH– ion.[71,72] Upon activation,
the H2 ligand in [H–Pb(η2-H2)]− separates to generate [PbH]− as the product. Previously, anionic Pb species was recognized only
under laser-ionization conditions.[75]Of the numerous conformations evaluated by DFT calculations, the
trigonal pyramidal (C3) geometry was found to be the most stable minimum-energy structure
for PbH3–. This ion, which in its valence
electrons is isoelectronic to nitrogen in NH3, assumes
a singlet state in its lowest-energy configuration. By contrast, its
reductive elimination product, the [PbH]− ion, bears
a triplet ground state (Supporting Information Table S2). DFT calculations, which returned a Mulliken spin density
of 2.054e for the Pb atom, supported the notion that
[PbH]− is a triplet that bears a formal Pb0 atom with a 6s26p2 electronic configuration.
The charge on the Pb atom in [PbH]− is −0.565e, and the charge on the H atom is −0.435e. Conceptually, the bonding in [PbH]− can be envisioned to originate from an electron transfer from the
1s2 orbital of a hydride ion to the empty 6p orbital of the Pb0 atom (Figure ). The energy of the σ-type
molecular orbital (MO) formed in this way lies between that of the
orbital formed by a lone pair of s-electrons and those of the two
singly occupied p-orbitals of Pb. The conclusion is that the bond
between H– and Pb in PbH– is essentially
a single σ-bond.
Figure 7
Occupied MOs of triplet PbH– (isosurface
value
= ±0.005 au). Note that the p-featured
MO has been rotated vertically by 90° to visually illustrate
the orbital.
Occupied MOs of triplet PbH– (isosurface
value
= ±0.005 au). Note that the p-featured
MO has been rotated vertically by 90° to visually illustrate
the orbital.Upon the revelation that
the hydride ion can coordinate with Pb0, we reexamined
the spectra recorded from [Sr(η2-O2CH)3]− and [Ba(η2-O2CH)3]− complexes
(Figure C, and Supporting Information Figure S7C). The product-ion
spectrum recorded from the m/z 223
ion for [Sr(η2-O2CH)3]− showed a peak at m/z 133 (Figures C and S5C), and that from [Sr(η2-O2CD)3]− displayed a peak at m/z 134 (Supporting Information Figure S5D). Accurate-mass measurements (observed
132.9041 u; calculated for SrCO2H 132.9038 u) supported
the notion that the m/z 133 peak
represents the [Sr(η2-O2CH)1]− ion, in which the oxidation state of strontium
is zero. In other words, during the fragmentation process, the strontium
cation undergoes a reduction. Analogously, the spectrum of [Ba(η2-O2CH)3]− showed a
peak at m/z 183 for the [Ba(0)(η2-O2CH)1]− ion (observed
182.9032 u; calculated for BaCO2H 182.9034 u) (Supporting Information Figure S7C). In both [Sr(η2-O2CH)3]− and [Ba(η2-O2CH)3]− ions, the
formate ion is chelated to a metal atom of oxidation number zero.
DFT calculations indicated the change of Mulliken charge of the barium
atom from positive (1.5e) to negative (−0.3e) density during the transformation of H2Ba(η2-O2CH)− to Ba(η2-O2CH)− (Supporting Information Table S4).The fragmentation of the [121Sb(III)(η-OCOH)4]− ion
(m/z 301, 23) followed
a trend similar to that exhibited
by the [115In(η-OCOH)4]− ion (Supporting Information Figure S12).
However, in contrast to the fragmentation propensities of the [H–In(η-OCOH)3]− ion (9c), which fragmented
further by decarboxylation or reductive elimination, the [H–121Sb(III)(η-OCOH)3]− ion
(m/z 257, 24) underwent
further fragmentation exclusively by the reductive elimination mechanism
to generate an ion of m/z 211 (25) for the [121Sb(I)(η-OCOH)2]−. The latter then loses two more CO2 molecules to generate the [121Sb(I)H2]− ion (m/z 123, 27) (Supporting Information Figure
S12). In stark contrast to the reductive elimination reaction of [H–In(η-OCOH)3]− (9c), the Sb analogue undergoes
reductive elimination with such ease that the product-ion formed by
H2 and CO2 elimination corresponds to the base
peak in the spectrum (Scheme ).
Scheme 5
Proposed Fragmentation Pattern of the Activated [Sb(III)(η-OCOH)4]− (m/z 301; 23) Ion
Of all fragmentation reactions studied herein, the dissociation
of the [Bi(III)(η-OCOH)4]− ion
(28), which was the heaviest adduct-ion investigated,
proved to be the most dramatic because it demonstrated the participation
of two reductive-elimination steps (Scheme ). In this case also, the dissociation
of the initial ion formed by the first CO2 loss occurs
so rapidly that a peak is not observed at m/z 345 for the [H–Bi(III)(η-OCOH)3]− ion as with thallium formate (Figure G). Presumably, the m/z 345 ion (29) reductively
eliminates immediately by losing H2 and CO2 to
yield the ion of m/z 299 (30) for [H–Bi(I)(η-OCOH)]− (Figure G). Upon further
activation, the m/z 299 ion (30) undergoes two consecutive eliminations of CO2 to form an ion of m/z 211 for
[Bi(I)H2]− (32). The m/z 211 ion then further reductively eliminates
by the loss of a dihydrogen molecule to yield an ion of m/z 209 (Supporting Information Figure S13). Evidently, the m/z 209 peak represents the bare bismuthide anion, Bi– (34), and not an isotope, because Bi is a monoisotopic
element. During the overall fragmentation process, the Bi core undergoes
two reductions: first from oxidation state +3 to +1 and then +1 to
−1. The relative intensity of the m/z 209 peak for the bismuthide anion (Bi–) increased when the mass-isolated m/z 211 (32) was subjected to increasing collision energies
(Supporting Information Figure S13). Hitherto,
the generation of the bismuthide anion has been achieved only by a
discharge or sputter-ion source,[76] photoelectronic, or laser ionization sources.[75] The generation of Bi– in this way adds to the list of methods available
to form negatively charged metallic ions.[47,48,78,79]
Scheme 6
Proposed
Fragmentation Pattern of the Activated [Bi(III)(η-OCOH)4]− (m/z 389; 28) Ion
Computations predicted that the singlet electronic state
is the
preferred configuration for [Bi(I)H2]− (Supporting Information Table S6). Analogous
to the transformation demonstrated for [PbH3]−, the [BiH2]− ion upon activation (32) forms a [Bi(H2)]− complex
(33) via the transition state (TS). The dH–H internuclear
distance 0.75 Å computed for the [Bi(H2)]− complex is closer than the value (0.82 Å) predicted for the
[PbH3]− complex (Supporting Information Tables S7 and S3). However, this [Bi(H2)]− complex (33) is still a
weak complex [the computed binding energy of the complex on the triplet
surface was 18 kJ/mol and that on the singlet surface was 76.6 kJ/mol
(Supporting Information Table S6)].[72] Upon activation, the [BiH2]− ion reductively eliminates a dihydrogen molecule to generate the
bismuthide anion with 6s26p4 electronic configuration,
for which the triplet state is more favorable than the singlet state
(Supporting Information Table S6).
Conclusions
In summary, we have described an approach to form M–H bonds
(M = Li, Na, K, Rb, Mg, Ca, Ba, Sr, Al, Ga, In, Tl, Pb, Sb, and Bi)
in the gas phase. Upon activation, the formate adducts of main group
metal formates fragment either by ejecting the neutral salt to release
the formate ion or by decarboxylation to form a product ion bearing
a metal–hydrogen bond. Furthermore, we have demonstrated a
trend in preferred fragmentation pathways: for alkali-metal adducts,
the formate-ion ejection is the preferred pathway, which becomes progressively
more favorable down group 1 of the periodic table; however, the expulsion
of CO2 becomes the more favorable pathway for adducts of
group 2, 13–15 metals. The formate anions bind to the central
metal cation in a bidentate manner in both neutral ion pairs and their
formate complexes of group 1 and 2 metals. For aluminum, the bidentate
arrangement of formate moieties in Al(η2-O2CH)3 becomes monodentate upon complexation. Upon activation,
the bidentate adducts undergo a conversion to a monodentate arrangement
before a molecule of CO2 is eliminated. We have shown that,
during the fragmentation, the adducts of group 1 formates do not undergo
a change of the oxidation state of the metal; by contrast, a trend
is observed for the adducts of higher members of group 2, 13, 14,
and 15 metals to undergo a progressively facile reduction of the oxidation
state. Some of the observed periodic trends evinced through formate
adduct fragmentations can be attributed to the increased inert-pair
effect as we move vertically down in a main group of the periodic
table. To recapitulate, during collision-induced fragmentation, the
oxidation states of In, Tl, and Sb were reduced from +3 to +1, Pb
from +2 to zero, and Bi most dramatically from +3 to +1 to −1.
Most hitherto reported η2-H2 complexes
are made from cationic transition metals. The existence of anionic
H2 or D2 complexes such as X––H2 and F––D2 (X = F, Cl, and Br) has been investigated by infrared, photoelectron,
and theoretical studies.[81−91] However, to our knowledge, the [H–Pb(η2-H2)]− complex (21) and [Bi(η2-H2)]− complex (33) species recognized in the current study are among the first anionic
η2-H2 type metal–hydrogen complexes.
Experimental
Section
MS
High-resolution and accurate mass CID mass spectra
were recorded on a SYNAPT G2 HDMS (Waters, UK) mass spectrometer equipped
with an ESI source. Typically, the capillary voltage was set to 2.0–3.0
kV. Sample solutions were infused to the source at a rate of 10–20
μL/min. Nitrogen was used as the nebulizer and the desolvation
gas. Argon was used as the collision gas. The source and desolvation-gas
temperatures were held at 100 and 150 °C, respectively. Mass
calibration (m/z 20 to 1200) was
performed using a 100 ppm sodium formate solution in 50% isopropyl
alcohol: 50% water (v/v). Typically, for CID experiments, the trap
collision energy was kept low (4 eV) and the transfer collision energy
was varied between 2 and 30 eV. The pressure in the transfer collision
cell was about 10–5 bar (under these conditions,
about 10 collisions are expected to take place before an ion exits
the 130 mm collision cell). The instrument was operated at a resolving
power of 10 000 measured at m/z 205. Ion-mobility separation of m/z 211 was conducted on the SYNAPT G2 HDMS instrument using N2 as the drift gas at a gas pressure of 3.43 mbar, an ion-mobility
spectrometry (IMS) wave velocity of 2796 m·s–1, and an IMS wave height of 32.3 V.
Chemicals
Acetonitrile
was purchased from PHARMCO-AAPER
(Brookfield, CT, USA). Formic acid (88%), barium hydroxide, and calciumoxide were purchased from Fisher Scientific (Fair Lawn, NJ, USA).
Metalmagnesium turnings, lithium hydroxide, potassium hydroxide,
sodium formate, rubidium chloride, cesium carbonate, gallium metal,
indium metal, thallium(III) acetate, lead(II) carbonate, antimony(III)
oxide, and formic acid-d2 (95 wt % in
D2O, 98 atom % D) were purchased from Sigma-Aldrich (Saint
Louis, MO, USA). Strontium carbonate was obtained from J.T. Baker
(Avantor, Center Valley, PA). H218O was purchased
from Cambridge Isotope Laboratories (Cambridge, MA). All chemicals
were used without further purification.
Metal Formates
LiOH, KOH, Mgmetal, CaO, SrCO3, Ba(OH)2, Al,
Cs2CO3, PbCO3, and Sb2O3 were dissolved in excess
formic acid (88%). RbCl was mixed with a stoichiometric amount of
lead formate. Gametal was first dissolved in concentrated HNO3 to obtain Ga(NO3)3, from which Ga(OH)3 was precipitated by ammonium hydroxide; Ga(OH)3 was then dissolved in HCOOH. Indium was dissolved in concentrated
HCl to obtain InCl3, from which In(OH)3 was
precipitated by ammonium hydroxide; In(OH)3 was then dissolved
in HCOOH. Bi was first dissolved in concentrated HNO3 to
obtain Bi(NO3)3, from which BiOCO3 was precipitated by action of Na2CO3; BiOCO3 was then dissolved in HCOOH. (CH3COO)3Tl(III) was precipitated by adding ammonium hydroxide to obtain Tl(III)(OH)3, and then Tl(III)(OH)3 was dissolved in HCOOH.
Each mixture was diluted with Millipore water to obtain aqueous solutions
of the corresponding metal formate for direct-infusion mass spectrometric
analysis. To prepare 18O-substitutedstrontium formate,
formic acid (96%, 10.00 μL) was dissolved in H218O (25 μL). After 12 h, the HC18O2H prepared in this way was added to SrCO3 (0.0105 g, 0.07
mmol), and the mixture was diluted with Millipore water. Analogously,
deuterium-substituted strontium formate was synthesized by adding
formic acid-d2 (95 wt % in D2O, 98 atom % D) to SrCO3.
DFT Calculations
DFT calculations were carried out
using the Gaussian 09W program at the hybrid density functional of
B3LYP and the 6-311++g(d,p), 6-311++g(2d,2p), or LanL2DZ basis sets.
The 6-311++g(2d,2p) basis set was used for lithium, sodium, potassium,
magnesium, and aluminum formates; the 6-311++g(d,p) basis set was
used for calciumformates. For strontium, barium, gallium, indium
and thallium, antimony, lead, and bismuthformates, the LANL2DZ basis
set was used for the metal atom and the 6-311++G(2d,2p) basis set
was used for all other atoms. Complete geometry optimizations were
conducted for all species with subsequent frequency calculations at
ambient pressure (1 atm) and room temperature (298.15 K) to verify
the nature of the corresponding stationary states on their potential
energy surfaces and to derive zero-point energy corrected electronic
energies (EZPE), enthalpies (H), and Gibbs free energies (G), in addition to the
electronic energies (ESCF). The atomic
charges were calculated using the natural population analysis scheme
in Gaussian 09.
Authors: J D Graham; A M Buytendyk; X Zhang; E L Collins; B Kiran; G Gantefoer; B W Eichhorn; G L Gutsev; S Behera; P Jena; K H Bowen Journal: J Phys Chem A Date: 2014-02-25 Impact factor: 2.781
Authors: A Grubisic; X Li; S T Stokes; K Vetter; G F Ganteför; K H Bowen; P Jena; B Kiran; R Burgert; H Schnöckel Journal: J Chem Phys Date: 2009-09-28 Impact factor: 3.488
Authors: Athanasios Zavras; George N Khairallah; Marjan Krstić; Marion Girod; Steven Daly; Rodolphe Antoine; Philippe Maitre; Roger J Mulder; Stefanie-Ann Alexander; Vlasta Bonačić-Koutecký; Philippe Dugourd; Richard A J O'Hair Journal: Nat Commun Date: 2016-06-06 Impact factor: 14.919