We present an IR-PD study of tantalum cluster adsorbate complexes [Tan(N2)m]+, abbreviated (n,m), n = 5–8. We utilize infrared spectroscopy of isolated and size selected clusters as prepared and characterized by a cryogenic tandem ion trap setup, and we augment our experiments with quantum chemical simulations at the level of density functional theory. The cluster adsorbate complexes (n,m) reveal vibrational bands above 2000 cm−1, which indicate end-on coordinated μ1-N2 oscillators, and bands below 2000 cm−1, which indicate side-on μ2-κN:κN,N coordinated ones. We observe a general increase in spectral complexity and an inhomogeneous broadening, mainly towards the red, at certain points of N2 loading m, which originates from an increasingly higher amount of double and triple N2 coordination at Ta sites, eventually at all of them. Other than the small tantalum clusters Tan+, n = 2–4, the IR-PD spectra of the initial N2 adsorbate species (n,1), n = 5–8, provide strong evidence for a lack of spontaneous N2 cleavage. Spontaneous N2 cleavage by Tan+, n = 5–8, seems suppressed. Therefore, the ability of a small Ta cluster to cleave dinitrogen disappears with one more tantalum core atom. The study of stepwise N2 adsorption on size selected Tan+, n = 5–8 clusters revealed adsorption limits m(max) of [Tan(N2)m]+ that are independent of cluster size within this size range. Cryo-adsorption kinetics at 26 K allowed for kinetic fits to consecutive N2 adsorption steps, and the fits revealed significant N2 desorption rates upon higher N2 loads, and the cluster adsorbate complexes eventually reached equilibrium. Some enhanced N2 desorption rates point towards likely adsorbate shell reorganization, and there is also some evidence for the coexistence of isomeric cluster adsorbate complexes.

The application of transition metals (TM) in heterogeneous catalysis is the basis for numerous industrial processes.2–7 Mechanistic understanding of the catalytic reactions was early on postulated to take place at so-called active sites8 that are part of the solid surface of a catalyst, which is most often less characterized when in operando. Transition metal clusters serve as model systems for these active sites in order to provide insights into elemental catalytic processes at the molecular level.9–15 All of this research elucidates cluster size effects on their structure and on their reactivity.

In particular, there are numerous studies that demonstrate the interaction and reactivity of tantalum clusters and their alloys, carbides and chalcogenides, with small molecules such as N2,1,16–23 CH4,17,24–26 CO2,27 and O2.28 Interaction, leading towards adsorption, is understood to yield adsorbates on the cluster surfaces. The cluster adsorbate complexes may act as precursor states for the activation and bond cleavage of these adsorbates. The issue of N≡N bond cleavage, for example, plays a role in the ongoing research efforts to develop an energetically optimized protocol, e.g., for Haber–Bosch NH3 production, a catalytic process of paramount significance to mankind on this planet and beyond.29,30

N2 activation has also been observed to proceed by small clusters of Co and Rh when assisted by surface plasma radiation,31,32 and bare TM clusters of V, Gd, and Sc,33–35 and homo- and heterometallic cluster carbides, with some subsequent C–N bond formation,16,17,36,37 and ultimately through lithiated TM complexes.38 

Size effects, an intrinsic propriety of clusters, prevail in tantalum clusters as well. The governing thermochemistry drives tantalum cluster cations of six atoms and less to reduce CO2 to CO, while larger ones form adducts.27 There are further size effects in the oxidation of such tantalum cluster cations, with the Ta8+ cluster cation sticking out by its oxidative cleavage towards Ta4O4+.28 Furthermore, size effects arise from small tantalum clusters (n = 1–4) dehydrogenating CH4 and concurrently eliminating H2, yielding the product of the stoichiometry [Tan,C,2H]+, while larger clusters (n = 5–10) enable mere molecular adsorption.26 While the bare Ta5+ cluster is inert towards CH4, its monoxide, Ta5O+, activates CH4.24 

The elucidation of such size effects asks for a structural characterization of participating species that may arise through spectroscopic experiments, which elucidate characteristic vibrations of, e.g., cluster adsorbate complexes and eventual reaction products.

Historically, inelastic electron scattering off N2 layered surfaces has proved valuable, but it has suffered from limited spectral resolution.39,40 The technical advancement of infrared photon dissociation (IR-PD) spectroscopy by either free electron lasers (FEL) or Optical Parametric Oscillator (OPO) technology has helped to identify characteristic vibrational bands, in particular when applied to isolated clusters and their complexes, and thereby achieve sufficient spectral resolution (as low as ±5 cm−1),41–44 and this method is widely applicable and well established by now.45–54 

In any case, it is mandatory to augment recorded vibrational spectra with quantum chemical modeling by density functional theory (DFT) methods in order to enable sound structural and dynamic interpretation. A successful example of such an approach is, e.g., the subsequent elucidation of the afore-mentioned tantalum product complex [4Ta,C,2H]+24 by IR-MPD spectroscopy and first principles quantum calculations, which identified a dihydride carbide formation, H2Ta4C+, and ruled out any other CH bond containing species.25 Another study complements experimental results from vibrational spectra and theoretical modeling of bare Tan+ clusters, n = 6–20, and it managed to identify cluster geometries, their relative stabilities, and electronic properties.55,56 An early study on the chemisorption properties of small tantalum clusters has revealed an oxygen adatom effect on the N2 uptake rate.18 Several studies including DFT modeling of small tantalum clusters provide strong evidence for dissociative N2 adsorption.19–23,57 A 2022 perspective summarizes recent progress in the field of nitrogen activation by TM species in the gas phase, and it shows that a systematic understanding of cluster reactivity towards N2 is lacking to date.58 

With this work, we build on our earlier study of Ta4+1 and complement it with a study of cationic clusters of n = 5–8 tantalum atoms. We reacted the bare clusters with N2 at 26 K and recorded IR-PD spectra of the adsorption species [Tan(N2)m]+, similar to our earlier studies on Nin+, Rhn+, RhiFej+, and Fen+ clusters.59–64 In the adjoined IR-PD study [IRS1], we discuss the N2 adsorption onto the three small clusters Ta2-4+. In our current study [IRS2], we actually present findings for the N2 adsorption onto the cluster sizes Ta5-8+. The findings of both studies, which we refer to in the following as [IRS1] and [IRS2], respectively, find strong support through our complementary cryogenic kinetic study (cf., supplementary material). The combination of IR-PD measurements and DFT modeling with kinetic studies provides insights into the structures of the N2 adsorbate species (5,m), (6,m), (7,m), and (8,m). We gain insight into the coordination motifs of the N2 ligands and find strong size effects, especially in terms of the first few adsorbates. We clarify and extend the overall picture and provide a systematic study of N2 adsorption on small tantalum clusters Tan+, n = 2–8. While we find complete cleavage of the N2 triple bond of the very first adsorbed N2 molecule on Ta2+, Ta3+, and Ta4+, the present study provides strong evidence for a lack of cleavage for clusters from Ta5+ on and beyond. Instead, adsorption of the first N2 molecule occurs in end-on orientation to the cluster core exclusively.

It is a tandem cryo-ion trap instrument that serves as the platform for all of the experiments conducted. It consists of a customized Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer (Apex Ultra, Bruker Daltonics) that encompasses cluster ion generation, size selection, N2 coordination, cluster adsorbate complex selection, and eventually infrared photon dissociation (IR-PD) spectroscopy of the respective cluster adsorption species. Note that this protocol comprises an MS3 procedure. Detailed descriptions of the ion generation and adsorption processes of the metal clusters can be found elsewhere1,59,60,64–67 and also in the supplementary material, where we present and discuss adjoined kinetic studies.

Briefly, we trap the cluster ions in an RF hexapole ion trap using the routine described in Text S3 in the supplementary material and guide the thus generated cluster N2 adsorbate complexes into the cryogenic (∼10 K) FT-ICR cell. For recording IR-PD spectra, we trap and select each cluster adsorbate species of interest in the IRC cell and irradiate each isolated ion packet by a tunable IR laser system with 7–10 laser pulses with an energy of 0.5–3.0 mJ each. These IR photon pulses stem from a pulsed injection seeded Nd:YAG laser (10 Hz, PL8000, Continuum), which pumps a KTP/KTA optical parametric oscillator/amplifier (OPO/OPA, LaserVision). An AgGaSe2 crystal generates the difference frequency (DF) of the OPA’s signal and the idler waves, whereby we generate the applied IR irradiation in the range of 1100 and 2400 cm−1. Absolute values of the DFM frequencies were derived from the online monitored OPO signal frequencies (continuous online wavemeter monitoring, Bristol Instruments, 8721B-NIR) and the pre-recorded value of the Nd:YAG frequency. Fragmentation mass spectra are recorded during continuous scans of the IR wavelength, and the recorded IR-PD signal transforms into a fragmentation efficiency via evaluation of iFiiFi+iPi, where Fi and Pi are fragment and parent ion intensities. The plotted fragmentation efficiency as a function of laser frequency finally yields the IR-PD spectrum. In all cases, single or multiple N2 losses were the only observed fragmentation channels.

Quantum chemical modeling by Density Functional Theory (DFT) served to obtain optimized cluster (adsorbate) geometries and vibrational eigenmodes, as achieved by the program packages Gaussian 09 and 16.68,69 We employed the hybrid functional PBE070,71 along with the Def2-TZVP basis set72,73 for all atoms in (5,0) cases, the Def2-TZVP basis set72 for N atoms, and the cc-pVTZ-pp basis set74 for Ta atoms in all (6,m) and (7,m) cases. We did so in continuity with prior studies that have succeeded to model N2 adsorption before60,61,64,66,67,75—particularly for the adsorption, activation, and cleavage of N2 on Ta4+ clusters.1 We tolerated relaxed SCF convergence criteria of 10−6 (as compared to 10−8 in “standard” DFT calculations) in order to achieve swift SCF convergence. The vibrational eigenmode spectra of all optimized geometries were inspected for the absence of any imaginary frequency. We applied optimized scaling factors of 0.9586 and 0.9512 for (6,m) and (7,m), respectively, for the cluster sizes to account for the prevailing anharmonicities, documenting unscaled IR frequencies in the supplementary material. We arrive at theoretical IR absorption spectra by convolution of the predicted and scaled IR frequencies with a Gaussian profile of FWHM = 5 cm−1. Natural Population Analysis (NPA) was performed via the NBO 5.9 option as implemented in Gaussian 09.75 

The current IR-PD spectroscopic study of the N2 adsorbates of Tan+ clusters, n = 5–8, was conducted under adiabatic conditions, and it is complemented by some N2 adsorption kinetics studies (cf. supplementary material). Here, we extend and complete our IR-PD study on the small cluster adsorbates (n,m), n = 2–4, as discussed in the adjoined infrared study [IRS1]. Once more, we recorded IR-PD patterns that vary remarkably by the size of the Tan+ clusters, n, which in turn vary by the amount of N2 adsorbate, m. In keeping with the adjoint study [IRS1], we arrange the studied cluster adsorbates in a (n,m) matrix and discuss the respective IR-PD spectra of each cluster size n in separate chapters.

The adsorbed N2 chromophores in each of the cluster adsorbate complexes possess vibrational bands, which we record as IR-PD features. Therefore, for all cluster adsorbate complexes (n,m) studied, we observe a specific and individual variation of IR-PD features as a function of the tantalum cluster size n and the nitrogen adsorbate loading m. In all cases, the recorded N2 fingerprint features range from 2000 to 2350 cm−1, and they most likely correspond to end-on N2 adsorbates that coordinate in a μ1 fashion to individual tantalum centers, with little to no interaction with other neighboring tantalum centers. This interpretation is in line with previously reported cases for Fen+,64 Con+,65 Nin+,59,60 Run+,66 Rhn+,67 and even Ta4+.1 In addition to these end-on bands, there are some cases where we observe IR-PD bands in the range of 1300–1700 cm−1, which indicate the most likely presence of a side-on μ2-κN:κN,N N2 ligand within the respective cluster adsorbate complex. This conclusion follows from the results of our previous study of N2 on a Ta4+ cluster.1 

We recorded IR-PD spectra of all cluster adsorbate complexes from (5,1) up to (5,13) (Fig. 1). We observe bands in the end-on N2 stretching range at 2000–2300 cm−1 for all investigated cluster adsorbate complexes, while there are little and weak IR-PD bands in the side-on N2 range below 2000 cm−1. There are IR-PD signals in the end-on N2 stretching range at all levels of N2 adsorption, m = 1–13. There is no indication for N2 activation and nitride formation whatsoever. While Nitride formation would be highly exothermic (cf. Table S3 in the supplementary material), the recorded IR-PD signatures signify intact N2 adsorbates, and it seems to prevail an extensive kinetic hindrance by unsurmountable barriers. If activated nitrides form at all, they seem to be minor and leave no fingerprint. Instead of N2 cleavage, we find a manifold of bands that persist at any level of N2 adsorption and shift slightly to the red with increasing N2 adsorption. In parallel, at (5,8) and beyond, we observe some broad and some sharp bands appearing to the red and spreading out to almost 2000 cm−1. It seems appropriate to account for the observed redshifts with respect to the free N2 stretching vibration at 2330 cm−1 in terms of a σ-donor and π-acceptor synergism as before. Up to species (5,8), we found no evidence of a N2 stretching vibrational band in the side-on range. For species (5,9) to (5,12), the IR-PD spectra indicate the presence of some weak adsorptions in the range between 1400 and 1700 cm−1. In the following, we will briefly discuss the recorded IR-PD spectra (5,m) and some of their IR-PD bands in some more detail.

FIG. 1.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta5(N2)m]+, m = 1–13, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

FIG. 1.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta5(N2)m]+, m = 1–13, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

Close modal

For single N2 adsorption m = 1, (5,1), we found a sharp single band at 2268 cm−1, which is red-shifted by 62 cm−1 with respect to the free N2 stretching vibration at 2330 cm−1.76 The (5,2) spectrum shows a more complex IR-PD band pattern, and we identify bands at 2236, 2257, 2262, and 2273 cm−1. Instead of merely two bands, e.g., from a single isomer with two end-on coordinated N2 molecules, we observe a manifold of at least four weak bands between 2200 and 2300 cm−1 and thus conclude on some co-existing isomers with comparable stabilization energies. The recorded IR-PD spectra of the species (5,3), (5,4), and (5,5) reveal no significant changes, and we observe three bands: An intense single band shifts slightly to the red from 2234, (5,3), to 2229 cm−1, (5,5); another intense band shifts to the red from 2281, (5,3) to 2277 cm−1, (5,5). Finally, there is another weak band at 2289 cm−1 (5,3) that shifts less with increasing m and thus merges into a mere shoulder of the second band in the IR-PD spectrum (5,5) (2282 cm−1). The slight redshift of the measured IR-PD bands with m cross-references to our prior studies on the N2 adsorption to other transition metal clusters such as Fen+64 and Nin+.59 In both cases, we found similar effects, and we took them as an indication of rough cluster surfaces. The adsorption of one and two more N2 molecules onto the formal monolayer at (5,5) yields a more complicated IR-PD band pattern for species (5,6)—and even more so for (5,7). The prior (5,5) IR-PD spectrum is augmented by one more band at 2259 cm−1 in (5,6) and another band on top at 2240 cm−1 in (5,7). The spectral broadening continues through the adsorption of the eighth N2 molecule, and the recorded IR-PD bands for (5,8) stretch out across a significantly wider range, from 2150 to 2300 cm−1. Nevertheless, we identify some partially resolved maxima at 2159, 2177, 2214, 2225, 2236, 2241, 2246, 2252, 2269, and 2277 cm−1. The new, most red-shifted bands might arise either from a variation of Ta center next neighbor coordination or from a triple end-on N2 occupation at a single Ta center. The IR-PD spectra of (5,9), (5,10), and (5,11) largely resemble that of (5,8), and we observe broad IR-PD patterns spread out across 2130–2310 cm−1, with some IR-PD bands partially resolved. The observed excess of IR-PD bands over present N2 oscillators clearly indicates co-existing isomers, which likely differ in the very structure of the N2 adsorbate shell or less likely differ in some parameters of the relaxed cluster core. In contrast to the prior IR-PD spectra, the spectra of the three species (5,9), (5,10), and (5,11) reveal a weak IR-PD band at 1504, 1516, and 1558 cm−1, respectively, which indicates the most likely presence of a side-on μ2-κN:κN,N N2 ligand. In addition, (5,11) reveals two more bands at 1422 and 1694 cm−1 that indicate further side-on N2 coordination at other sites and with varying coordination parameters such as, e.g., the tilting angle. The IR-PD spectra of species (5,12) and (5,13) reveal another increase in the maximum of redshifts and an eventual gain in spectral structure and reduction of broadening. In (5,13), the end-on coordinated N2 ligands seems to separate into two classes, as indicated by high and low red-shifted bands at 2010–2080 cm−1 and 2208–2306 cm−1, with less to no options in between other than before. A side-on μ2-κN:κN,N N2 indicating band of (5,12) at 1563 cm−1 continues the trend of such bands in (5,9) – (5,12), and it terminates this series, as (5,13) does not reveal any such or similar IR-PD signal. It seems as if at this point the shell of N2 adsorbates has become so sterically crowded that a two Ta center coordinating N2 tilt is not possible anymore. Instead, the lack of wagging space leads to an increase in stiffness in this N2 adsorbate shell, which might manifest in a highly structured and well-resolved IR-PD spectrum of (5,13).

In summary, we found numerous IR-PD bands in the IR-PD spectra of species (5,m), m = 1–13, which are located mainly in the end-on coordinated N2 stretching range of 2000–2310 cm−1. Some of these species (m = 9–12) indicate weak IR-PD signals below 2000 cm−1, which we assign to side-on μ2-κN:κN,N N2 ligands. Starting with a few single and distinguishable end-on N2 bands of the less occupied species (5,m), m small, we observe with increasing m an overall increase of spectral complexity and an asymmetric broadening, largely towards the red. Finally, the spectrum of (5,13) reveals some sharp bands. While intermediate N2 loads (moderate m values) allow for much spectral complexity and likely isomerism, the highly loaded species with much of a filled N2 shell exhibit reduced spectral complexity and less isomerism. This indicates the onset of orientational stiffness and a less wagging disorder. Side-on μ2-κN:κN,N N2 coordination seems to die out.

Other than the smaller tantalum clusters Tan+, n = 2–4 (cf. Fries et al.1 and the adjoined study [IRS1]), the (5,m) spectra provide strong evidence for a lack of any spontaneous N2 cleavage. Note that we found IR-PD fingerprints of end-on bound N2 molecules on otherwise bare Ta5+ clusters and a lack of side-on μ2-κN:κN,N N2 fingerprints. It appears as if an enhanced entrance channel barrier prevents spontaneous tilting and further activation. It takes further N2 adsorption to find some weak side-on μ2-κN:κN,N N2 fingerprints. Spontaneous N2 cleavage by Ta5+ seems suppressed.

In order to investigate the effects of stepwise N2 adsorption on the cluster structure and possible spin relaxations in the course of adsorption, we optimized the Ta5+ cluster core together with the adsorbate shell of (5,m) in the cases of m = 0,1,2,5,10 comprising end-on N2 ligands. For each m, we started with several conceivable adsorption shell isomers, which adsorb N2 onto a reasonable trigonal bipyramid (tbp) cluster framework. We obtained fully converged structures for singlet to nonet multiplicities in most cases (Fig. 2). Vertical offsets of the (5,m) curves depict N2 adsorption and desorption energies—each of these curves resembling the spin valley61 of a single (5,m) complex. We find triplet and singlet isomers most stable in all cases and sometimes almost degenerate.

FIG. 2.

Total energies of computed [Ta5(N2)m]+ cluster adsorbate complexes (5,m) as a function of the spin multiplicity 2S + 1, and normalized to the most stable (5,0) spin isomer (trigonal bipyramid, 3tbp, triplet). The calculated minimum structures for (5,0), (5,5), and (5,10) are shown as insets; the geometries of other multiplicities may differ slightly.

FIG. 2.

Total energies of computed [Ta5(N2)m]+ cluster adsorbate complexes (5,m) as a function of the spin multiplicity 2S + 1, and normalized to the most stable (5,0) spin isomer (trigonal bipyramid, 3tbp, triplet). The calculated minimum structures for (5,0), (5,5), and (5,10) are shown as insets; the geometries of other multiplicities may differ slightly.

Close modal

The bare Ta5+ cluster finds itself most stable in a triplet trigonal bipyramidal (tbp) minimum structure 3tbp, independent of starting geometries. Two apical Ta atoms coordinate three next neighbors each, and the three equatorial Ta atoms coordinate four next neighbors each. This most stable isomer is close to D3h symmetry. The singlet electromer is slightly less stable, by a mere 14 kJ/mol. Higher multiplicities are more expensive by at least 90 kJ/mol and much beyond. Through single N2 adsorption, singlet and triplet multiplicities become degenerate. This degeneracy persists throughout further N2 adsorption, up to m = 5. Note that high multiplicities seem to stabilize as well, and the spin valleys become more shallow. The energy difference between each curve’s minimum and the quintet state decreases from 63 kJ/mol (m = 1) over 48 kJ/mol (m = 2) to 29 kJ/mol (m = 5).

The most stable (5,5) cluster adsorbate complex has its N2 adsorbate shell concentrating on one side of the tbp cluster core: two of the equatorial Ta atoms are singly occupied, and one of the two apical Ta atoms has adsorbed three N2 molecules. Upon adsorption of ten N2 molecules, yielding the most stable (5,10) cluster adsorbate complex, the tbp cluster core has received some significant distortion towards an oblate geometry. Two opposite (apical) Ta atoms adsorb only a single N2 molecule. Two other Ta atoms adsorb a triplet of N2 molecules each. In (5,10), the singlet state is favored over the triplet state by 16 kJ/mol. We thus assume that adsorption-induced spin quenching could occur at high N2 coverage, m ≥ 5.

We extract adsorption energies from the spin valley curves by evaluating their offsets, which yield both vertical (spin conserving) and adiabatic (including relaxation of the spin state) adsorption energies (cf. Fig. S2 and Table S2). In most cases, the vertical and adiabatic energies are very similar, as singlet and triplet states are close to degenerate. The adsorption energies increase from about −40 kJ/mol (m = 0) to about −50 kJ/mol (m = 1), irrespective of multiplicity. From m = 2 to m = 4, we interpolate values of about −32.4 kJ/mol per N2 adsorbate. From m = 5 to m = 9, we interpolate adiabatic adsorption energies of −41.5 kJ/mol per N2, which are slightly larger than the vertical ones (−38.4 kJ/mol per N2).

Through natural population analysis (NPA), we obtained more information on the charge and spin distributions of some selected clusters and on the influence of the first N2 adsorption.

The NPA of the most stable (5,0) 3tbp isomer reveals a concentration of the positive charge on the equatorial Ta atoms (Fig. S4). The two apical Ta atoms possess less positive charge density and more spin density than the three equatorial ones. A spin density of ∼1 at each of the two apical Ta atoms indicates the localization of a single unpaired electron at each of them.

All of the three considered (5,1) isomers a, b, and c (Fig. S5) are tbp type structures, and they are equally stable (within the uncertainty of the DFT modeling, about 4 kJ/mol), with all other structures being less stable by +18 kJ/mol and beyond. The isomers a, b, and c reveal minor changes in singlet or triplet spin preference and in the predicted orientation of the end-on coordinated N2 adsorbate. In effect, isomers a and b find themselves in close proximity to CS symmetry, whereas c possesses C2v symmetry. Their computed linear absorption spectra are depicted in the supplementary material, Fig. S3. Application of the successfully applied scaling factor of our prior Ta4+ study (see the adjoined study [IRS1]) 0.9736 finds isomer b interpreting the recorded IR-PD spectrum well, isomer a less likely, while isomer c may be ruled out. These assumptions find nice correspondence in the spin valley prediction—1(5,1) is slightly less stable than 3(5,1)—and in structural motifs of N2 coordination to TM centers such as Rh and Ta, where we found a preferred “90° to surface normal” N2 coordination that persists for many species along the adsorption chain.1,77

As in the naked (5,0) Ta cluster, the (5,1) N2 adsorbate complex locates a large fraction of positive charge on the three equatorial Ta atoms and the majority of electron density in triplet states at the two apical Ta atoms. Note that the N2 molecule prefers to adsorb onto either of the three equatorial Ta atoms, which serve as better electron acceptors than the two apical ones.

In the case of the triplet isomers b and c, both apical Ta atoms comprise an unpaired electron each, and their charge density is reduced with respect to the singlet isomer a by 0.04e per Ta atom. Upon adsorption, the N2 molecule is polarized (∆q < 0.29e) but obtains little net charge and little net spin density. The electron population of the adsorbing Ta atom strongly depends on the isomer (spin and mainly angle), in particular if compared with the other equatorial Ta atoms: it varies between 0.255, b, and 0.174e, c. In the singlet isomer a, the adsorbing Ta atom is the most positive equatorial atom with a charge density of 0.255e. The charge density of the adsorbing Ta atom in triplet isomer b is in between the other equatorial atoms 0.219e < 0.259e < 0.279e. In the highly symmetric triplet isomer c, the adsorbing isomer comprises the least charge density of the equatorial Ta atoms (0.174e < 0.26e).

In summary, the observed weak IR-PD bands below 2000 cm−1 of several species (5,m), m ≥ 9, indicate isomers that might contain a side-on coordinated μ2-κN:κN,N N2 ligand. However, the measured IR-PD spectra for the adsorbate species (5,m) provide no evidence for N2 cleavage. We found IR-PD bands of end-on-bound N2 molecules to single Ta centers for all species along the adsorption chain. It is remarkable, that the addition of a single metal atom to the Ta4+ cluster framework eliminates the ability of the cluster to cleave N2 molecules.

We speculate that the adsorption of a single N2 molecule induces little change in the electron distribution within the Ta5+ cluster core, likely the computed isomer b, and it has almost no effect on the spin density distribution. These effects are specific for each coordination isomer of the (5,1) complex, and they depend strongly on the orientation of the adsorbed N2 molecule. It remains to elucidate how this changes along the further N2 adsorption chain.

Next, we inspect in detail the IR-PD spectra of the tantalum cluster adsorbate species Ta6(N2)m+: (6,m), m = 1–15. We took IR-PD spectra for (6,m) in the 1200–2400 cm−1 range (cf. Fig. 3), as we did before for smaller tantalum clusters. That is, we analyzed each adsorbate species for IR-PD vibration bands in the end-on region above 2000 cm−1 and in the side-on region below 2000 cm−1. The recorded IR-PD spectra reveal an end-on N2 fingerprinting fragmentation signal in all investigated cases (6,m) up to m = 15–even for the very first N2 adsorbate species (6,1). All of these bands are shifted to the red with respect to the stretching frequency of free N2 at 2330 cm−1,76 but by varying amounts. All of these shifts originate from an interplay of σ-donor and π-acceptor interactions at work, and in a cluster size modulated manner.

FIG. 3.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta6(N2)m]+, m = 1–15, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

FIG. 3.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta6(N2)m]+, m = 1–15, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

Close modal

In the following, we briefly discuss the recorded IR-PD spectra (6,m) (cf. Fig. 3) and the obtained characteristic bands separately. The IR-PD spectrum of (6,1) reveals merely a single weak IR-PD induced fragmentation at 2293 cm−1. The presence of this IR-PD band reveals strong evidence for a lack of N2 cleavage on the Ta6+ cluster. This is well in line with the observed lack of N2 cleavage by Ta5+, which makes for a trend: the N2 cleavage ability of small tantalum clusters vanishes upon an increase in cluster size. Finally, there is no IR-PD evidence for tilted N2 side coordination in (6,1).

The IR-PD spectrum of (6,2) reveals a single band at 1425 cm−1 and another one at 2292 cm−1. The former one is in accordance with similar bands of smaller tantalum clusters, which we assigned to N2 molecules acting as μ2-κN:κN,N ligands that bind side-on across a Ta–Ta edge of the cluster cores. Our DFT calculations will confirm this interpretation, the presence of a single side-on N2 ligand in the (6,2) complex (cf. below). The other IR-PD band of (6,2), above 2000 cm−1, corresponds to a like band in (6,1), and we assign it to the N2 stretching vibration of the other N2 adsorbate, which binds end-on to single tantalum centers.

The IR-PD spectrum of (6,3) reveals evidence for the onset of co-existing adsorption isomers: There are three IR-PD bands in the side-on range, at 1416, 1468, and 1625 cm−1, and a strong band in the end-on range, at 2289 cm−1, with an additional weak and broad IR-PD induced adsorption at 2255 cm−1. Therefore, the number of IR-PD bands (6,3) exceeds the amount of adsorbed N2 molecules. We interpret this necessarily as strong evidence for the co-existence of adsorbate isomers in a cluster adsorbate complex with a persistent Ta6+ cluster core framework. This would be similar to our findings for N2 adsorbate complexes of the Ta4+ (cf. [IRS1]) and Ta5+ clusters. The IR-PD spectrum reveals a wide spread of IR-PD bands in the side-on range, from 1400 to 1650 cm−1. This reminds us of the Ta4+ case in our adjoined study [IRS1], where we assigned some observed bands with a diminished redshift to N2 complexes with a smaller tilt of the side-on μ2-κN:κN,N N2 ligand. Therefore, we speculate on multiple isomers, each of which contains a single side-on μ2-κN:κN,N N2 ligand across an edge of the cluster core and which likely differ slightly in the degree of tilting. However, co-existing isomers that contain two of these side-on μ2-κN:κN,N N2 ligands are conceivable as well. The subsequent DFT calculations will help to clarify.

We observe five distinguishable IR-PD bands for species (6,4): four of them are located in the side-on range at 1406, 1423, 1462, and 1620 cm−1 (weak), and a single one is located at 2289 cm−1 in the end-on range. Note that the integrated IR-PD signals of the side-on region (below 2000 cm−1) exceed those of the end-on region for all species (6,3) to (6,7), whereas it is the other way around for all cluster adsorbate species with N2 loads as in (6,8) and beyond.

The IR-PD spectrum of (6,5) comes somewhat as a surprise: it reveals merely two bands, one of which is significantly broadened, around 1448 cm−1, and the other is a sharp IR band at 2287 cm−1, the latter one known as an end-on coordinated N2 ligand, as observed in all spectra of less loaded clusters. There are little to no shifts with respect to the equivalent bands of (6,4). The broadening of the side-on peak suggests the presence of multiple coordination isomers of two or up to three side-on coordinated N2 ligands.

The IR-PD spectrum of (6,6) resembles very much that of (6,5), with merely an additional weak band at 2300 cm−1. This new band is located 14 cm−1 to the blue of the persistent N2 end-on band, which might signify a splitting that originates from two N2 oscillators and their symmetrically/asymmetrically coupled stretching vibrations. Note that our prior investigations of related species have ongoingly found such effects in N2 adsorbates to other types of transition metal clusters.59,64 The broad adsorption below 2000 cm−1 persists, as observed in (6,5).

The IR-PD spectrum of species (6,7) shows a similar band pattern as (6,6). However, the side-on region of (6,7) yields two well separate bands at 1445 and 1464 cm−1. The ratio of integrated IR-PD signals in the side-on region to those in the end-on region indicates once again a strong contribution of side-on coordinated N2 chromophores, two or three of them in total.

The side-on IR-PD signal diminishes significantly in (6,8), with a single weak adsorption at 1438 cm−1. In contrast, the IR-PD signal in the end-on range builds up and reveals a strong quartet at 2276, 2284, 2289, and 2294 cm−1, which might either signify various types of symmetric/asymmetric couplings or distinguishable coordination sites and motifs. In any case, all of these IR-PD bands are located in the same frequency range (2250–2300 cm−1) as the singlet and doublet bands of the less loaded adsorbate species.

Starting with the adsorption of the ninth N2 molecule, the range of observed end-on IR-PD bands widens to the red; in (6,9), it extends down to 2200 cm−1. Features of (6,8) largely persist in (6,9) through three partially overlapping bands at 2269, 2283, and 2294 cm−1. Two new features arise: a sharp band at 2245 cm−1 and a further red-shifted broad adsorption around 2202 cm−1. There is a significant change in the IR-PD signal in the side-on range: any discernible IR-PD signal of (6,9) diminishes to some very weak and widely broadened signal around 1525 cm−1, which is less to the red than the intensive peaks of the less loaded species (6,2) to (6,7).

The IR-PD spectrum of species (6,10) merely hints at some broad and very weak adsorption in the side-on range that is hardly discernible and might peak around 1515 cm−1. On its own, this feature might be questionable. Similar broad IR-PD bands in (6,9) and (6,11) lend some credibility to our speculation. In the end-on region, the weak band of (6,9) at 2202 cm−1 rises to a prominent doublet of (6,10) at 2180 and 2199 cm−1. Quite the opposite, the four strong bands of (6,9) above 2200 cm−1 reduce to a single band of (6,10) at 2283 cm−1.

The observable IR-PD signal of (6,11) in the end-on range broadens to the red and extends from 2140 to 2300 cm−1. There are five distinguishable IR-PD bands at 2143, 2171, 2273, 2280, and 2288 cm−1. The side-on spectral range of the recorded IR-PD spectrum of (6,11) reveals a weak but sharp adsorption at 1470 cm−1, which hints towards a single (6,11) isomer with a side-on μ2-κN:κN,N coordinated N2 ligand.

The IR-PD spectra of species (6,12) and (6,13) reveal band patterns that are most similar to those of (6,11). Some IR-PD bands of end-on coordinated N2 stretching modes reach out slightly further to the red, down to 2100 cm−1. The weak side-on adsorption at 1474 cm−1 persists in (6,12), as observed in (6,11), and it reveals a small but discernible splitting in (6,13).

In the IR-PD spectrum of species (6,14), there are no IR-PD bands in the side-on range below 2000 cm−1, and despite a sufficient signal-to-noise ratio. This indicates that the addition of the fourteenth N2 molecule suppresses any side-on μ2-κN:κN,N N2 ligand, and it likely converts any prior one. We speculate that spatial crowding of the many ligands may well favor the slimmer end-on N2 coordination over the space consuming side-on N2 coordination. In the IR-PD region of end-on N2 coordination, above 2000 cm−1, we observe numerous resolved IR-PD bands: 2100, 2106, 2133, 2151, and 2161 cm−1, as well as at 2266, 2272, 2284, 2300, and 2308 cm−1. All in all, these bands are significantly better resolved than in the spectra of (6,12) and (6,13). We take this as a hint towards a reduced degree of disorder in N2 wagging coordinates.

Reaching out for feasibility limits, we managed to record an IR-PD spectrum of species (6,15). This spectrum revealed numerous IR-PD bands at the end-on coordinated N2 stretching range above 2000 cm−1. We observe intensive adsorptions at 2098, 2108, 2145, 2276, 2287, 2295, and 2310 cm−1. Moreover, there are some broad IR-PD absorptions without resolved structures between 2168 and 2245 cm−1. It appears as if the fifteenth N2 ligand induces some enhanced couplings towards and among the other N2 ligands. This might be well understood in terms of threefold N2 coordinated Ta centers. At medium N2 load, m = 9–14, there are one or two such centers, and they may assume distal positions in the octahedral Ta scaffold. Any third center with a threefold N2 load must, of course, come close to either of the previous ones. Which induces many opportunities for inter-center N2 stretching mode couplings. In effect, the recorded IR-PD spectrum yields broad and largely unstructured bands—as observed and on display in the (6,15) spectrum of Fig. 3.

In summary, we found numerous IR-PD bands in various wavenumber ranges for the N2 adsorption species (6,m), m = 1–15, mainly in the range of 1400–1650 cm−1 and 2000 to 2310 cm−1. We assign the discernible IR-PD bands below 2000 cm−1 to side-on bound μ2-κN:κN,N N2 molecules and those above 2000 cm−1 to μ1 end-on coordinated N2 molecules. The end-on bound N2 coordination signature in (6,1) corresponds to a similar finding in (5,1). Therefore, Ta6+ provides strong evidence for a lack of any spontaneous N2 cleavage as well. For some species, explicitly (6,3) and (6,4), the total number of IR-PD bands exceeds the total number of N2 adsorbates, which provides strong evidence of coordination isomers. The μ1 coordination dominates the end-on range of the IR-PD spectra at low N2 loads up to m = 5. With increasing N2 coverage, we observe an overall asymmetrical spread out to the red of the end-on bands. At the highest recorded N2 load, in the IR-PD spectrum of (6,15), the end-on bands assume a range of more than 200 cm−1, between 2100 and 2310 cm−1. We attribute the observation of such a wide spreading out to an increasingly higher amount of double and triple N2 coordination at Ta sites of the cluster scaffold, eventually at all of them. The higher the N2 coordination of Ta, the more the vibrational frequencies of the corresponding N2 ligands shift to the red.

The side-on μ2-κN:κN,N N2 coordination motifs emerge in most of the IR-PD spectra, except (6,1), (6,14), and (6,15). At low N2 loads, up to (6,7), the side-on fingerprints are roughly as strong as the end-on ones. At high N2 loads (6,8) and beyond, there is an undisputable propensity for favorable end-on signatures, and side-on ones diminish. Therefore, it appears as if highly loaded clusters prefer, upon saturation, an exclusive end-on N2 coordination, and partially loaded ones enable a mixture of some side-on N2 coordination with some end-on N2 coordination. Steric overcrowding seems to enforce the spatially less demanding end-on N2 coordination in (6,14) and (6,15).

We undertook DFT modeling for conceivable clusters (6,m) in order to investigate the structure and geometry of the bare Ta6+ cluster (m = 0) and its first N2 adsorbate species up to m = 3.

We found the bare Ta6+ cluster most stable in a slightly compressed octahedral structure, with four slightly elongated equatorial bonds (cf. Table S9) and a doublet state (M = 2). We also found some less stable structure, a nido-pentagonal pyramid with a quartet state (M = 4, +66 kJ/mol). Other test structures were optimized into either of these two. Energetics of other spin states/multiplicities and further details are to be found in the supplement (cf. Table S4). Our present conclusions on the naked Ta6+ cluster are well in line with a previous study.56 Beyond that study, we went on and modeled the stepwise adsorption of N2 molecules to the favored octahedral Ta6+ structure.

The slightly compressed octahedron provides for two distinguishable adsorption sites of likely end-on coordinated N2 molecules, with either of the two polar Ta sites of the octahedron proving most favorable (cf. Table S4, isomer a). Note that the compression-distortion of the octahedral cluster did receive slight changes from the polar N2 adsorption. In such (6,1), comprising a polar N2 adsorbate, the prior equal equatorial bonds elongate (+0.10 Å) and compress (−0.12 Å) in concert with a minor tilting of 2° of the N2 ligand with respect to the polar symmetry axis. This, in turn, makes the four equatorial Ta sites pairwise distinguishable. Alternatively, an N2 adsorption onto either of the other four equatorial Ta centers (cf. Table S4, isomer b) would stabilize less, by +7 kJ/mol, and the according quartet states even less, +22 and +17 kJ/mol, respectively. Spin states seem to persist upon N2 adsorption in the present case, and any spin raise seems unlikely. This is even more so in view of our spin-valley curves of the Ta5+ cluster species (5,m) (see above) and of several prior studies of adsorbate induced spin quenching in other TM clusters.61,64 From here on, we constrain ourselves to consider doublet states of (6,m).

The recorded IR-PD bands of the (6,2) cluster complex at 1425 cm−1 motivate a computational quest for either of two N2 molecules coordinating in a side-on (μ2-κN:κN,N) motif. Indeed, such a motif proved more favorable by at least +85 kJ/mol [cf. Fig. 4 and Table S6(b)], than that of mere end-on coordination, whose predicted IR spectrum deviates from the measured IR-PD spectrum anyway (cf. Fig. 4). The identified most stable isomers (6,2)c and (6,2)h both comprise a side-on μ2-κN:κN,N N2 ligand that is coordinated to two equatorial Ta centers, next neighbors to the polar N2 end-on coordinating Ta center (cf. Fig. 4). They differ by chirality, and they are degenerate, as expected [cf. Table S6(b) in the supplementary material]. The bond of the side-on N2 coordinating Ta pair shortens by −0,07 Å (cf. Table S9).

FIG. 4.

Experimental IR-PD spectrum of (6,2) and DFT modeled spectra of selected structural isomers in the doublet state, their relative energies indicated with respect to the most stable isomer (6,2)c and the optimized cluster adsorbate structures. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (6,2), see Tables S6(b) and S7(b).

FIG. 4.

Experimental IR-PD spectrum of (6,2) and DFT modeled spectra of selected structural isomers in the doublet state, their relative energies indicated with respect to the most stable isomer (6,2)c and the optimized cluster adsorbate structures. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (6,2), see Tables S6(b) and S7(b).

Close modal

We identified further N2(side-on)/N2(end-on) ligand combinations, and we found them less stable by +11 to +20 kJ/mol. [cf. Tables S6(b) and S7(b)]. Obviously, both N2 ligands seek proximity, and they benefit from a cooperative gain on the order of up to 20 kJ/mol in achieving such proximity. Indeed, our modeling confirms that the most stable degenerate isomer pairs [isomers (6,2)b/g and (6,2)c/h] reveal identical IR spectra, respectively. Moreover, all four IR spectra match the recorded IR-PD spectra (Fig. 4). In contrast, the computed IR spectra of the isomer pair (6,2)e,j, and further ones do not coincide with the recorded IR-PD spectrum [cf. Table S6(b)]. Most notably, it shows that the side-on N2 orientation proves negligible, and hence, only the location of the coordinating Ta pair over which the side-on ligand is located is more relevant for searching for a matching minimum structure for the measured (6,2) IR-PD spectrum. The orientation of the side-on ligand over the edge seems to be less relevant. Furthermore, we have identified a cooperative effect between the end-on N2 and the side-on N2.

In order to model the observed stepwise uptake of a third N2 ligand by Ta6+, we added a second end-on bound N2 ligand to the prior end-on and side-on μ2-κN:κN,N N2 ligands [cf. Tables S6(c1) and S7(c1)]. It emerges from a favorable arrangement of the two end-on N2 ligands at Ta centers on the opposite Ta corners of the octahedral core [cf. Fig. 5, structures (6,3)j and (6,3)w]. Both end-on N2 ligands are located at distal coordination sites. The side-on μ2-κN:κN,N N2 ligand is located in between and proximal to either end-on N2 ligand. The orientation of the side-on N2 ligand across its coordinating Ta–Ta edge is degenerate. By virtue of the high symmetry, the modeled IR spectrum provides for merely two bands that match well the observed IR-PD bands at 2289 and 1468 cm−1. Furthermore, observed IR-PD bands at 1416, 1625, and 2255 cm−1 point towards other isomers.

FIG. 5.

Experimental IR-PD spectrum of (6,3) and DFT modeled spectra of selected structural isomers in the doublet state, their relative energies indicated with respect to isomers (6,3)j/w, and optimized cluster adsorbate structures included as insets. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (6,3), see Tables S6(c) and S7(c).

FIG. 5.

Experimental IR-PD spectrum of (6,3) and DFT modeled spectra of selected structural isomers in the doublet state, their relative energies indicated with respect to isomers (6,3)j/w, and optimized cluster adsorbate structures included as insets. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (6,3), see Tables S6(c) and S7(c).

Close modal

Up to now, we have not yet succeeded in identifying such candidates that would provide for two end-on and a single side-on ligand [cf. Fig. 5 and Fig. S7 (a)], and we wonder about some isomer with another degree of tilting of the N2 side-on ligand. Note that our adjoined study of Ta4+ [IRS1] confirms such effects. We find support for our hypothesis by a closer look at the bond distances Taκ(N,N)-Nside-on(μ1) and the angles Taκ(N,N)-Taκ(N)-Nside-on(μ1) of the twenty-two isomers at hand (6,3)a-w. Indeed, there is a correlation between an increase in the tilt of the side-on μ2-κN:κN,N N2 ligand and the observed redshift of its IR-PD vibrational frequency (cf. Table S8). This correlation seems independent of cluster size and cluster core structure, and it seems that the local Ta-(N–N)–Ta geometry dominates the observed effects.

In order to assign the IR-PD bands at 1416, 1625, and 2255 cm−1, we consider isomers with two side-on μ2-κN:κN,N N2 adsorbates. Our modeling of several conceivable arrangements [cf. Tables S6(c) and 6(c)] yielded three isomers (6,3)XIV, (6,3)XVI, and (6,3)XX with a bunch of IR frequencies that agree quite well with the experimental IR-PD bands [cf. Fig. 5 and Fig. S7 (b)]. Within these three isomers, the arrangement of the end-on ligand and one of the side-on ligands corresponds to the structure of the most favorable isomer (6,2), as described earlier. Differences arise from varying positions of the second side-on ligand.

In the two isomers (6,3)XIV and (6,3)XVI, all three N2 ligands are located in proximity to each other, and they share common Ta sites for coordination. These two isomers suffice to interpret all five experimental IR-PD bands at 1416, 1468, 1625, 2255, and 2289 cm−1 (cf. Fig. 5 and Fig. S7). The third isomer (6,3)XX, which is actually the most stable one at −69 kJ/mol, manages to locate its three N2 ligands while avoiding shared Ta sites.

This yields a highly symmetric arrangement of the N2 ligands, and it causes merely a single IR band in the side-on range and merely another single IR band in the end-on range. Both match the experimental IR-PD bands at 1468 and 2289 cm−1. Note that this isomer (6,3)XX on its own does not explain all observed IR-PD bands. However, we interpret the recorded IR-PD spectrum of (6,3) in terms of likely isomers (6,3) XIV and XVI, and we acknowledge possible contributions from other isomers.

In summary, DFT modeling provided insights into the structure of the first three adsorbate species (6,1), (6,2), and (6,3) on the Ta6+ cluster octahedron. The first N2 ligand yielding (6,1) adsorbs most likely end-on to the cluster core followed by two N2 adsorption steps giving two μ2-κN:κN,N N2 ligands in (6,3). This sequence of end-on adsorption before side-on adsorption of the first three N2 ligands differs from the results of N2 adsorption on the Ta4+ cluster (cf. Fries et al.1 and the adjoined study [IRS1]) and highlights once more the differential behavior between cluster sizes Tan+, n ≤ 4, and Tan+, n ≥ 5.

At this point, we want to address the question of whether the DFT predictions and the recorded IR-PD spectra are in line with the recorded adsorption kinetics (cf. supplementary material, Text S9, Fig. S22). The kinetics show an N2 adsorption cascade to the bare Ta6+ cluster. The possible isomers for (6,1), (6,2), and (6,3) identified by DFT modeling correspond in structure/geometry to this chain of successive N2 adsorption steps. The isomer identified as most likely to be populated in each case serves as the initial geometry for the adsorption of the next N2 molecule.

From our kinetic studies presented and discussed in the supplementary material (cf. Text S10, Fig. S23), we know that the Ta7+ cluster adsorbs up to 16 N2 molecules onto its cluster surface. We found an equilibrium between adsorbed species (7,10) to (7,16) at long reaction times. However, (7,16) is a very small, low-intensity product in equilibrium. Therefore, (7,15) became the largest N2 adsorbate complex in our current IR-PD study of N2 adsorption on Ta7+ (cf. Fig. 6). As before, we searched for IR-PD induced bands in the 1200 to 2400 cm−1 range and found numerous vibrational bands in two frequency ranges: below 2000 cm−1 for N2 oscillators binding in a side-on μ2-κN:κN,N motif to the Ta7+ cluster core, and beyond 2000 cm−1 for those binding end-on to single Ta centers. In the following, we discuss the experimental IR-PD spectra of the species in detail.

FIG. 6.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta7(N2)m]+, m = 1–15, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

FIG. 6.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta7(N2)m]+, m = 1–15, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line).

Close modal

The N2 adsorption chain on Ta7+ starts with the initial adsorbate species (7,1). For this species, we observe a single IR-PD band at 2294 cm−1 and thus a redshift with respect to the free N2 stretching oscillation of –36 cm−1. Once more, it seems reasonable to explain this redshift by a σ-donor and π-acceptor synergism. The presence of this IR-PD band provides strong evidence that spontaneous N2 cleavage by Ta7+ does not proceed. Instead, Ta7+ appears to continue the trend that at a cluster size of n = 5 and beyond, adsorption of the first N2 molecule occurs in end-on orientation to the cluster core and without further activation.

There are strong similarities in the IR-PD spectra of the subsequent two adsorbate species (7,2) and (7,3). Both spectra reveal a doublet with a splitting of 9 cm−1: at 2282 and 2291 cm−1 for (7,2) and at 2280 and 2289 cm−1 for (7,3), respectively. The second N2 molecule adsorbs to a site of the cluster such that another IR-PD band arises. A detailed examination of the (7,3) spectrum shows a very weak IR-PD sideband at 2277 cm−1 (Fig. 7). Our DFT modeling will reveal conceivable geometries and vibrational modes that will help to clarify.

For all of the species from (7,4) to (7,10), we observe a similar IR-PD band pattern in the end-on range above 2000 cm−1. These spectra reveal induced fragmentation from 2234 cm−1 [the lowest end-on IR-PD fingerprint in the case of species (7,6)] to 2292 cm−1 [the highest end-on IR-PD fingerprint in the case of species (7,8)]. The number count of clearly identifiable IR-PD bands in the respective range varies between four and six.

The experimental IR-PD band patterns in the side-on region below 2000 cm−1 vary somewhat by the N2 load, but they also show some recurring band patterns. For the adsorbate species (7,4), we observe a broad and weak IR-PD band around 1457 cm−1 with some partially resolved band structure. The IR-PD spectrum of the next larger adsorbate complexes (7,5) and beyond reveals a much more complex band pattern of side-on N2 oscillators. Instead of a single broad IR-PD band around 1457 cm−1 (7,4), we now observe up to three bands for (7,5), (7,6), (7,7), and (7,8), which partially overlap and reveal some variation in their width but little shifts. This trend does not continue for the adsorbate species (7,9) and (7,10); instead, the spectra of both species show four and merely two sharp bands, respectively, all of which reveal clearly less bandwidth than the smaller adsorbate species.

In addition, for (7,5) and beyond, we observe some IR-PD adsorptions at up to 1600 cm−1: There is some side-on N2 stretching doublet structure that shifts to the red with N2 load from 1548 to 1563 cm−1 in (7,5) to 1498 and 1513 cm−1 in (7,10). The IR-PD spectra of (7,5), (7,9), and (7,10) reveal an additional single band just below 1600 cm−1, which also shifts to the red with N2 load, from 1594 in (7,5) to 1576 cm−1 in (7,10).

With the adsorption of the eleventh N2 molecule onto the cluster adsorbate complex, the range of the end-on IR-PD band pattern successively expands to the red, down to 2100 cm−1 in (7,13). Within this range, the IR-PD spectra of the three species (7,11), (7,12), and (7,13) show an average of ten bands, which is twice as much as for the previous adsorbate species (7,4) to (7,10).

In the side-on region below 2000 cm−1, the adsorption of the eleventh N2 preserves the band pattern of the (7,10) species, and the same largely holds for (7,13). In contrast, the IR-PD spectra of the species (7,12) reveal particularly different IR-PD fingerprints: There is a single intensive band at 1431 cm−1, which is accompanied by some weak bands at 1530 and 1554 cm−1. The occurrence of such a sharp and intense band indicates multiple side-on N2 adsorbates in equivalent arrangements, likely μ2-κN:κN,N type, that experience like next neighbor interactions and concomitant couplings of their stretching bands. Astonishing as it is, larger and smaller adsorbate complexes, namely (7,11) and (7,13), behave differently and provide for disliked next neighbor interactions among their side-on coordinated N2 ligands.

The experimental IR-PD spectrum of the species (7,14) is conspicuous for its low number count of IR-PD bands. The (7,14) species provides for a formal N2 double layer of two ligands per Ta site. If truly occupied in such a stoichiometric manner, it would provide for all equivalent ligands and high symmetry. Indeed, the IR-PD spectrum provides for a single intensive band at 1467 cm−1, indicating side-on μ2-κN:κN,N N2 oscillators, and a single intensive broad band around 2301 cm−1, originating from some end-on N2 oscillators. It is thus clearly evident that the formal “double layer” species does provide for two types of clearly distinguishable N2 adsorbates, with the exact count of each of them remaining undisclosed.

From (7,4) to (7,13), the number of measured IR-PD bands corresponded to or even exceeded the number of adsorbed N2 molecules, which indicates the likely presence of multiple isomers of these species. In contrast, the IR-PD band number count of species (7,14) strongly suggests the occurrence of a single isomer.

It is surprising to recognize the positions and number count of the IR-PD bands of (7,15). There are at least twelve distinguishable adsorption maxima in the end-on range between 2088 and 2304 cm−1. Even more surprisingly, the prior intense IR-PD band of (7,14) at 1467 cm−1 shifts by adsorption of the 15th N2 ligand down to 1307 cm−1, by as much as −160 cm−1, which is quite significant. Some more weak bands arise between 1475 and 1568 cm−1.

The recorded IR-PD spectra of (7,m) reveal numerous bands that provide information about the structure of the coordinated N2 ligands within the respective cluster adsorbate complexes. The spectra of the first three adsorbate species, (7,1) to (7,3), indicate N2 oscillators that bind end-on to single Ta centers. From (7,4) and beyond, we find evidence for N2 oscillators binding in a side-on μ2-κN:κN,N motif to the Ta7+ cluster core. We recognize the large number of IR-PD bands below 2000 cm−1. Together with the IR-PD bands above 2000 cm−1, the total number of IR-PD bands for all species from (7,4) and beyond, but (7,12) and (7,14), exceeds the number of N2 oscillators of the respective cluster adsorbates. We take this as strong evidence for several co-existing adsorbate isomers. However, there is the case of (7,14), where the number of IR-PD bands clearly falls below the number of adsorbed N2 ligands. This in turn indicates a highly symmetric arrangement of the N2 ligands on the Ta7+ cluster framework within the formal double layer of N2 oscillators.

In order to gain deeper insights into the N2 adsorption on the Ta7+ cluster and the involved structures of the cluster adsorbate complexes, we performed DFT modeling of conceivable geometries.

First of all, we wondered about the bare framework of the Ta7+ cluster. We chose four geometries as starting points: a capped octahedron (coh), a pentagonal bipyramid (pbp), and two types of capped trigonal prism (ctp1: capped over square plane, ctp2: capped over triangle plane). The pentagonal bipyramid (pbp) in a triplet state turned out to be the most stable (cf. Table S10). Our conclusions on this naked Ta7+ cluster are well in line with a previous study.56 Singlet and quintet states of the same geometry are less stable by +35 and +37 kJ/mol, respectively. As listed in the supplement (cf. Table S10), all other investigated geometries are less stable, e.g. +162 kJ/mol for the capped octahedral (3coh) geometry. We, therefore, take the pentagonal bipyramidal (pbp) geometry for the Ta7+ framework as a starting point to investigate N2 adsorption.

The highly symmetric pentagonal bipyramidal core structure of Ta7+ provides two distinguishable binding sites for the adsorption of a first N2 molecule in end-on orientation. DFT modeling predicts that the two preferred triplet structures differ by 11 kJ/mol [cf. Tables S11(a) and S12(a)]. N2 adsorption on one of the two tips of the pentagonal bipyramid is preferred as compared to adsorption on one of the Ta atoms of the pentagonal plane of the framework. The predicted IR band of the end-on N2 oscillator comes close to the single experimental IR-PD band at 2294 cm−1 (cf. Fig. 7 and Fig. S8).

FIG. 7.

Experimental IR-PD spectra of (7,1), (7,2), and (7,3) and DFT modeled spectra of selected structural isomers in the triplet state and the corresponding optimized cluster adsorbate structures as insets. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (7,m), see Tables S11(a)–S11(c) and S12(a)–S12(c).

FIG. 7.

Experimental IR-PD spectra of (7,1), (7,2), and (7,3) and DFT modeled spectra of selected structural isomers in the triplet state and the corresponding optimized cluster adsorbate structures as insets. The predicted IR frequencies are scaled by a factor of 0.9586. For calculated energies and IR frequencies of all modeled isomers of (7,m), see Tables S11(a)–S11(c) and S12(a)–S12(c).

Close modal

DFT modeling of conceivable geometries of species (7,2) [cf. Tables S11(b) and S12(b)] yielded merely a single isomer whose predicted IR band pattern matches the experimentally observed twin IR-PD bands at 2282 and 2291 cm−1 (cf. Fig. 7 and Fig. S8) but for some shift. The characteristic splitting of the IR-PD band results from an arrangement of both N2 oscillators onto a single Ta atom. In consistency with the most stable isomer (7,1)b, which might serve as a precursor to this isomer (7,2)f, both N2 molecules adsorb to either of the two tips of the pentagonal bipyramid. Unfortunately, this (7,2) isomer with this arrangement of N2 ligands is 29 kJ/mol less stable than the most stable one in our DFT study. Nevertheless, we take this isomer to be the most likely, and we presume unknown issues on the DFT modeling side.

For species (7,3), our DFT modeling yields several isomers that differ in stability by a mere 19 kJ/mol [cf. Tables S11(c) and S12(c)]. None of the predicted IR band patterns reveal satisfactory predictions. The third most stable isomer (+6 kJ/mol) shows some resemblance with the observed IR-PD spectrum of (7,3) (cf. Fig. 7 and Fig. S9). In this isomer, the three N2 oscillators would be distributed over the two tips of the pentagonal bipyramid. At this point, it becomes obvious that DFT modeling with our methods at hand reaches its limits, and we look very much forward to seeing other researcher pick up this topic, hopefully in the near future.

In summary, we conclude from our DFT modeling for conceivable geometries of the bare Ta7+ cluster (7,0) and of the first three adsorbate species (7,1), (7,2), (7,3) that the N2 molecules adsorb in an end-on motif to the pentagonal bipyramid prior to an adsorption to one of the five Ta atoms forming the pentagonal plane of the framework.

At this point, our IR-PD spectra and our concomitant DFT modeling strongly suggest a consecutive uptake of N2 adsorbates into end-on coordination geometries onto a presumed most stable Ta7+ isomer without any indication of its scaffold to rearrange or the spin states to relay. These findings are much in line with our kinetic studies of N2 adsorption to Ta7+, as documented in the supplementary material (cf. Text S10, Fig. S23). There, we do observe stepwise N2 uptake by the Ta7+ cluster, and we managed to perform a pseudo-first-order kinetic fit that does not indicate neither any isomerization nor any activation throughout the first four N2 adsorption steps. The second and third N2 adsorption processes proceed somewhat slower than the first step. This might correspond to the reasonable conclusion of sterically limited N2 adsorption to the tips of the pentagonal bipyramidal framework of the Ta7+ cluster core.

The almost abrupt increase in the number of IR-PD bands from (7,4) and beyond, which we interpret as a clear indication of the presence of multiple isomers for each of the species, coincides with the increasing desorption rates from k-(7,4) on and beyond. Therefore, the isomers for all species from (7,4) onward appear to adsorb additional N2 molecules with likely simultaneous isomerization of the adsorbate shell. There is further coincidence of a particular feature in the (7,11)th IR-PD spectra with the kinetics (cf. supplementary material, Fig. S23) of the eleventh N2 adsorption onto Ta7+: The IR-PD spectrum of (7,11) reveals a significant extension of the end-on coordinated IR-PD bands to the red, and the recorded N2 adsorption kinetics demonstrate an equilibrium of all cluster adsorbates from (7,10) and beyond. Therefore, our IR-PD studies and our kinetic studies come to the same conclusions. The increase in multiple N2 coordination at the same tantalum centers causes likely isomerization, e.g., N2 adsorbate site hopping.

The Ta8+ cluster is the largest one that we characterized by recording the IR-PD spectra of (8,m), m = 1–8 (Fig. 8 and Fig. S10). We found no IR induced fragmentation in the low wavenumber range below 2000 cm−1 and, therefore, we have no indications for the formation of side-on μ2-κN:κN,N coordinated N2 oscillators in the investigated cluster adsorbate complexes. We observe several IR-PD bands above 2000 cm−1, which likely stem from μ1 end-on coordinated N2 ligands of the Ta8+ cluster. The low yields of higher loaded adsorbate species (m > 8, cf. Fig. S25, supplementary material) hampered the recording of their IR-PD spectra, and the obtained signal-to-noise ratio is meager. Once more, it seems reasonable to assume weak σ-donation and considerable π-acceptance at work.

FIG. 8.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta8(N2)m]+, m = 1–8, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line). For IR-PD spectra in the range of 1200–2400 cm−1, see Fig. S10 in the supplementary material.

FIG. 8.

IR-PD spectra of sequentially increasing cluster N2 adsorbate complexes [Ta8(N2)m]+, m = 1–8, as recorded after prior 26 K He buffer gas cooling. Note the significant red-shift of the IR-PD bands with respect to the free N2 stretching vibrational frequency (2330 cm−1,76 red dotted line). For IR-PD spectra in the range of 1200–2400 cm−1, see Fig. S10 in the supplementary material.

Close modal

We observe a strong IR-PD band at 2261 cm−1 and a very weak band at 15 cm−1 to the blue, which indicates a second less likely isomer. With the current DFT modeling at hand, we cannot assign particular geometries to these isomers. We are left with the speculation that some coordination isomers form, which finds some support in our kinetic studies (cf. supplementary material, Text S11, Figs. S24 and S25). It is noteworthy to emphasize at this point that the presence of end-on coordinated N2 fingerprinting IR-PD bands is a compelling argument for the lack of N2 cleavage ability of the Ta8+ cluster.

The IR-PD band pattern of (8,1) persists in the IR-PD spectrum of (8,2), slightly shifted to 2259 and 2274 cm−1. The latter one intensifies significantly. With the adsorption of the third N2 molecule onto the Ta8+ cluster, this very IR-PD band pattern is preserved in (8,3), but for some shoulder of the blue band.

The species (8,4) reveals two intensive IR-PD bands at 2253 and 2269 cm−1. The vaguely discernible blue shoulder of (8,3) has evolved into a clear feature of (8,4) at 2274 cm−1. In addition, one notes a weak IR-PD band at 2292 cm−1. It is this band pattern that persists in the IR-PD spectrum of (8,5) but with some intensity variations and minor shifts.

Besides an inferior signal-to-noise ratio, the (8,5) IR-PD bands persist in (8,6), and an additional band arises at 2241 cm−1, which is to the red of all previous (8,m) bands. We found such behavior in (5,m), (6,m), and (7,m) at some point as well. The IR-PD spectra of species (8,6), (8,7), and (8,8) reveal six to eight bands at similar locations.

The cross-reference to our cryo-kinetics study of N2 adsorption on the Tan+ cluster, n = 2–8, supplementary material, yields further insights. The kinetic curve of the first adsorbate species (8,1), Fig. S25, provides strong indications for two isomers, which signify their different reactivity. Our interpretation of the IR-PD spectrum (8,1) concludes as well that there are two isomers, one of which is less populated.

We present an IR-PD study of tantalum cluster adsorbate complexes [Tan(N2)m]+, abbreviated (n,m), n = 5–8, [IRS2], and in the adjoined IR-PD study [IRS1], we present findings for the N2 adsorption to the smaller cluster adsorbate complexes [Ta2-4(N2)m]+.

We found numerous bands in the IR-PD spectra of species (5,m), m = 1–13, which are located mainly in the end-on coordinated N2 stretching range. Some of these species (m = 9–12) indicate weak IR-PD signals, which we assign to side-on μ2-κN:κN,N N2 ligands. DFT calculations allowed us to tentatively assign cluster structures on the basis of trigonal bipyramidal (tbp) cluster cores. The evaluation of spin valley curves allows us to evaluate the spin-state-dependent stability of cluster adsorbate complexes in the course of N2 adsorption. Triplet and singlet isomers reveal themselves to be most stable in all cases. We assume some spin quenching from triplet to singlet states upon adsorption of the fifth to tenth N2 molecule. It is the subject of ongoing work to conduct extensive quantum chemical modeling of the seemingly hindered reaction path of (5,1) and of all of the stationary points.

We found numerous IR-PD bands of the N2 adsorption species (6,m), m = 1–15, in two spectral ranges, 1400 to 1650 cm−1 and 2000 to 2310 cm−1. DFT modeling revealed a sequence of initial N2 end-on coordination and a subsequent mix of N2 side-on adsorption. The recorded IR-PD spectra of (7,m) reveal numerous bands indicating N2 oscillators that bind end-on to single Ta centers in complexes (7,1) to (7,3). From (7,4) on and beyond, we find evidence for N2 oscillators binding in a side-on μ2-κN:κN,N motif to the Ta7+ cluster core. We observe several IR-PD bands of (8,m), which likely stem from μ1-end-on coordinated N2 ligands. Starting from a few single and distinguishable end-on N2 bands of the less occupied adsorbate complexes (n,m), n = 5–7, we observe a general increase in spectral complexity and an inhomogeneous broadening of bands, mainly towards the red, at certain points of N2 loading m. We attribute this observation to an increasingly higher amount of double and triple N2 coordination of Ta sites, eventually at all of them. The more N2 ligands coordinate with single Ta sites, the more the vibrational frequencies shift to the red.

Other than the IR-PD spectra of the small tantalum clusters Tan+, n = 2–4 (cf. Fries et al.1 and the adjoined study [IRS1]), the present IR-PD spectra of the initial adsorbate species (n,1), n = 5–8, provide strong evidence for a lack of spontaneous N2 cleavage. Instead, there are significant IR-PD fingerprints of end-on bound N2 molecules on otherwise bare Tan+, n = 5–8, clusters throughout. It appears as if an enhanced entrance channel barrier prevents spontaneous tilting and further activation. In the cases of n = 5–8, it takes further N2 adsorption to find some weak side-on μ2-κN:κN,N N2 fingerprints, and the reasoning for the onset of these features at (5,9), (6,2), and (7,4) is an open question as of now. One may speculate on a conceivable N2–N2 cooperativity that is highly sensitive to cluster surface morphologies and thereby also to the cluster sizes. In any case, it is a sound finding that the spontaneous N2 cleavage by Tan+, n = 5–8, is suppressed. Therefore, the ability of a small Ta cluster to cleave dinitrogen (cf. Fries et al.1 and [IRS1]) disappears by adding a single, fifth metal atom to the cluster core.

The above-mentioned findings find strong support in our cryogenic kinetic studies, as documented in the supplementary material. There, we present a study of stepwise N2 adsorption on size selected Tan+, n = 5–8 clusters, yielding [Tan(N2)m]+. The N2 adsorption reaches a limit m(max) = 16, which does not depend on cluster size n, and we observe a final equilibrium among those adsorbate complexes that possess up to three to five N2 adsorbates less than m(max), and the kinetic fits revealed significant N2 desorption rates upon higher N2 loads. The variations of absolute rates find reasonable interpretation in terms of simple thermodynamic arguments. We find some evidence for spin quenching, and it mixes in a conceivable coexistence of isomeric cluster adsorbate complexes.

Online supplementary material comprises a nomenclature for the clusters and their adsorbates, comprehensive listings of energetics, structures, and IR frequencies of computed structures, further experimental and computed IR spectra, and multiple text blocks and viewgraphs on the conducted kinetic experiments, kinetic fits, and their interpretation in detail.

This work was supported by the German research foundation DFG within the trans-regional collaborative research center Grant No. SFB/TRR 88 (Cooperative effects in homo and heterometallic complexes, 3MET.de) and by the research center OPTIMAS. We acknowledge Thomas Kolling for his technical assistance and valuable discussion. Quantum chemical modeling took place at the “Regionales Hochschulrechenzentrum Kaiserslautern” (RHRK).

The authors have no conflicts to disclose.

The experiments were performed by D.V.F.—partly together with A.S., M.P.K., and M.E.H. D.V.F. and M.P.K. performed the quantum chemical calculations. The evaluation of the measured data was carried out by D.V.F., and it was accompanied by discussions with A.S., M.P.K., and G.N.-S. D.V.F. and G.N.-S. wrote the manuscript, which all the authors revised and agreed to.

Daniela V. Fries: Data curation (lead); Formal analysis (lead); Investigation (lead); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Matthias P. Klein: Conceptualization (supporting); Data curation (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting); Writing – review & editing (supporting). Annika Straßner: Data curation (supporting); Investigation (supporting); Methodology (supporting); Resources (supporting); Validation (supporting); Writing – review & editing (supporting). Maximilian E. Huber: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Project administration (supporting); Validation (supporting); Writing – review & editing (supporting). Gereon Niedner-Schatteburg: Conceptualization (lead); Formal analysis (equal); Funding acquisition (lead); Investigation (supporting); Methodology (lead); Project administration (lead); Resources (lead); Supervision (lead); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
D. V.
Fries
,
M. P.
Klein
,
A.
Steiner
,
M. H.
Prosenc
, and
G.
Niedner-Schatteburg
,
Phys. Chem. Chem. Phys.
23
(
19
),
11345
11354
(
2021
).
2.
A.
Mittasch
,
Geschichte der Ammoniaksynthese
(
Verlag Chemie
,
1951
).
3.
K.
Ziegler
,
H.
Breil
,
E.
Hilzkamp
, and
H.
Martin
, Germany Patent No. DE973626C (14 April
1960
).
4.
W.
Ostwald
, United Kingdom Patent No. GB190200698A (20 March
1902
).
5.
BASF
, Germany Patent No. DE235421C (13 October
1908
).
6.
F.
Fischer
and
H.
Tropsch
, Germany Patent No. DE524468C (7 May
1931
).
7.
F.
Fischer
and
H.
Tropsch
, Germany Patent No. DE484337C (16 October
1929
).
8.
H. S.
Taylor
,
Proc. R. Soc. London, Ser. A
108
(
745
),
105
111
(
1925
).
9.
S. M.
Lang
and
T. M.
Bernhardt
,
Phys. Chem. Chem. Phys.
14
(
26
),
9255
9269
(
2012
).
10.
D. K.
Böhme
and
H.
Schwarz
,
Angew. Chem., Int. Ed.
44
(
16
),
2336
2354
(
2005
).
11.
R. A. J.
O’Hair
and
G. N.
Khairallah
,
J. Cluster Sci.
15
(
3
),
331
363
(
2004
).
12.
P.
Armentrout
,
Annu. Rev. Phys. Chem.
52
(
1
),
423
461
(
2001
).
13.
M. M.
Kappes
,
Chem. Rev.
88
(
2
),
369
389
(
1988
).
14.
P. B.
Armentrout
,
Catal. Sci. Technol.
4
(
9
),
2741
2755
(
2014
).
15.
M. D.
Morse
,
M. E.
Geusic
,
J. R.
Heath
, and
R. E.
Smalley
,
J. Chem. Phys.
83
(
5
),
2293
2304
(
1985
).
16.
L.-H.
Mou
,
Y.
Li
,
Z.-Y.
Li
,
Q.-Y.
Liu
,
H.
Chen
, and
S.-G.
He
,
J. Am. Chem. Soc.
143
(
45
),
19224
19231
(
2021
).
17.
L.-H.
Mou
,
Y.
Li
,
G.-P.
Wei
,
Z.-Y.
Li
,
Q.-Y.
Liu
,
H.
Chen
, and
S.-G.
He
,
Chem. Sci.
13
(
32
),
9366
9372
(
2022
).
18.
M. R.
Zakin
,
D. M.
Cox
, and
A.
Kaldor
,
J. Chem. Phys.
87
(
8
),
5046
5048
(
1987
).
19.
G.-D.
Jiang
,
L.-H.
Mou
,
J.-J.
Chen
,
Z.-Y.
Li
, and
S.-G.
He
,
J. Phys. Chem. A
124
(
38
),
7749
7755
(
2020
).
20.
C.
Geng
,
J.
Li
,
T.
Weiske
, and
H.
Schwarz
,
Proc. Natl. Acad. Sci. U. S. A.
115
(
46
),
11680
11687
(
2018
).
21.
C.
Geng
,
J.
Li
,
T.
Weiske
, and
H.
Schwarz
,
Proc. Natl. Acad. Sci. U. S. A.
116
(
43
),
21416
21420
(
2019
).
22.
F.
Mafuné
,
Y.
Tawaraya
, and
S.
Kudoh
,
J. Phys. Chem. A
120
(
24
),
4089
4095
(
2016
).
23.
M.
Kumar Yadav
and
A.
Mookerjee
,
Physica B
405
(
18
),
3940
3942
(
2010
).
24.
J. F.
Eckhard
,
T.
Masubuchi
,
M.
Tschurl
,
R. N.
Barnett
,
U.
Landman
, and
U.
Heiz
,
J. Phys. Chem. C
122
(
44
),
25628
25637
(
2018
).
25.
J.
Lengyel
,
N.
Levin
,
F. J.
Wensink
,
O. V.
Lushchikova
,
R. N.
Barnett
,
U.
Landman
,
U.
Heiz
,
J. M.
Bakker
, and
M.
Tschurl
,
Angew. Chem., Int. Ed.
59
(
52
),
23631
23635
(
2020
).
26.
J. F.
Eckhard
,
T.
Masubuchi
,
M.
Tschurl
,
R. N.
Barnett
,
U.
Landman
, and
U.
Heiz
,
J. Phys. Chem. A
125
(
24
),
5289
5302
(
2021
).
27.
N.
Levin
,
J. T.
Margraf
,
J.
Lengyel
,
K.
Reuter
,
M.
Tschurl
, and
U.
Heiz
,
Phys. Chem. Chem. Phys.
24
(
4
),
2623
2629
(
2022
).
28.
J. F.
Eckhard
,
D.
Neuwirth
,
C.
Panosetti
,
H.
Oberhofer
,
K.
Reuter
,
M.
Tschurl
, and
U.
Heiz
,
Phys. Chem. Chem. Phys.
19
(
8
),
5985
5993
(
2017
).
29.
R.
Schlögl
,
Angew. Chem., Int. Ed.
42
(
18
),
2004
2008
(
2003
).
30.
H.
Liu
,
Chin. J. Catal.
35
(
10
),
1619
1640
(
2014
).
31.
L.
Geng
,
Y.
Jia
,
H.
Zhang
,
C.
Cui
, and
Z.
Luo
,
ChemPhysChem
23
(
17
),
e202200288
(
2022
).
32.
C.
Cui
,
Y.
Jia
,
H.
Zhang
,
L.
Geng
, and
Z.
Luo
,
J. Phys. Chem. Lett.
11
(
19
),
8222
8230
(
2020
).
33.
M.
Zhou
,
X.
Jin
,
Y.
Gong
, and
J.
Li
,
Angew. Chem., Int. Ed.
46
(
16
),
2911
2914
(
2007
).
34.
X.
Cheng
,
Z.-Y.
Li
,
L.-H.
Mou
,
Y.
Ren
,
Q.-Y.
Liu
,
X.-L.
Ding
, and
S.-G.
He
,
Chem. – A Eur. J.
25
(
72
),
16523
16527
(
2019
).
35.
Y.
Gong
,
Zhao
, and
M.
Zhou
,
J. Phys. Chem. A
111
(
28
),
6204
6207
(
2007
).
36.
Z.-Y.
Li
,
Y.
Li
,
L.-H.
Mou
,
J.-J.
Chen
,
Q.-Y.
Liu
,
S.-G.
He
, and
H.
Chen
,
J. Am. Chem. Soc.
142
(
24
),
10747
10754
(
2020
).
37.
Z.-Y.
Li
,
F.
Horn
,
Y.
Li
,
L.-H.
Mou
,
W.
Schöllkopf
,
H.
Chen
,
S.-G.
He
, and
K. R.
Asmis
,
Chem. Eur. J.
29
(
14
),
e202203384
(
2022
).
38.
Y.-Q.
Ding
,
Z.-Y.
Chen
,
Z.-Y.
Li
,
X.
Cheng
,
M.
Wang
, and
J.-B.
Ma
,
J. Phys. Chem. A
126
(
9
),
1511
1517
(
2022
).
39.
M. C.
Tsai
,
U.
Ship
,
I. C.
Bassignana
,
J.
Küppers
, and
G.
Ertl
,
Surf. Sci.
155
(
2
),
387
399
(
1985
).
40.
H. J.
Freund
,
B.
Bartos
,
R. P.
Messmer
,
H.
Grunze
,
H.
Kuhlenbeck
, and
M.
Neumann
,
Surf. Sci.
185
(
1
),
187
202
(
1987
).
41.
A.
Fielicke
,
A.
Kirilyuk
,
C.
Ratsch
,
J.
Behler
,
M.
Scheffler
,
G.
Von Helden
, and
G.
Meijer
,
Phys. Rev. Lett.
93
(
2
),
023401
(
2004
).
42.
J.
Oomens
,
B. G.
Sartakov
,
G.
Meijer
, and
G.
von Helden
,
Int. J. Mass Spectrom.
254
(
1
),
1
19
(
2006
).
43.
J. R.
Eyler
,
Mass Spectrom. Rev.
28
(
3
),
448
467
(
2009
).
44.
P.
Maitre
,
D.
Scuderi
,
D.
Corinti
,
B.
Chiavarino
,
M. E.
Crestoni
, and
S.
Fornarini
,
Chem. Rev.
120
(
7
),
3261
3295
(
2020
).
45.
J.
Roithová
,
Chem. Soc. Rev.
41
(
2
),
547
559
(
2012
).
46.
G.
Altinay
and
R. B.
Metz
,
Int. J. Mass Spectrom.
297
(
1-3
),
41
45
(
2010
).
47.
B.
Chiavarino
,
M. E.
Crestoni
,
M.
Schütz
,
A.
Bouchet
,
S.
Piccirillo
,
V.
Steinmetz
,
O.
Dopfer
, and
S.
Fornarini
,
J. Phys. Chem. A
118
(
34
),
7130
7138
(
2014
).
48.
O.
Dopfer
,
J. Phys. Org. Chem.
19
(
8-9
),
540
551
(
2006
).
49.
N. I.
Hammer
,
J. W.
Shin
,
J. M.
Headrick
,
E. G.
Diken
,
J. R.
Roscioli
,
G. H.
Weddle
, and
M. A.
Johnson
,
Science
306
(
5696
),
675
679
(
2004
).
50.
J. M.
Headrick
,
E. G.
Diken
,
R. S.
Walters
,
N. I.
Hammer
,
R. A.
Christie
,
J.
Cui
,
E. M.
Myshakin
,
M. A.
Duncan
,
M. A.
Johnson
, and
K. D.
Jordan
,
Science
308
(
5729
),
1765
1769
(
2005
).
51.
A. M.
Rijs
and
J.
Oomens
,
Top. Curr. Chem.
364
,
1
42
(
2015
).
52.
T. R.
Rizzo
and
O. V.
Boyarkin
,
Top. Curr. Chem.
364
,
43
97
(
2015
).
53.
T. R.
Rizzo
,
J. A.
Stearns
, and
O. V.
Boyarkin
,
Int. Rev. Phys. Chem.
28
(
3
),
481
515
(
2009
).
54.
W. H.
Robertson
and
M. A.
Johnson
,
Annu. Rev. Phys. Chem.
54
,
173
213
(
2003
).
55.
P.
Gruene
,
A.
Fielicke
, and
G.
Meijer
,
J. Chem. Phys.
127
(
23
),
234307
(
2007
).
56.
J.
Du
,
X.
Sun
, and
G.
Jiang
,
J. Chem. Phys.
136
(
9
),
094311
(
2012
).
57.
X.
Sun
and
X.
Huang
,
ACS Omega
7
(
26
),
22682
22688
(
2022
).
58.
L.-H.
Mou
,
Z.-Y.
Li
, and
S.-G.
He
,
J. Phys. Chem. Lett.
13
(
18
),
4159
4169
(
2022
).
59.
J.
Mohrbach
,
S.
Dillinger
, and
G.
Niedner-Schatteburg
,
J. Phys. Chem. C
121
(
20
),
10907
10918
(
2017
).
60.
S.
Dillinger
,
J.
Mohrbach
, and
G.
Niedner-Schatteburg
,
J. Chem. Phys.
147
(
18
),
184305
(
2017
).
61.
A. A.
Ehrhard
,
M. P.
Klein
,
J.
Mohrbach
,
S.
Dillinger
, and
G.
Niedner-Schatteburg
,
Mol. Phys.
119
(
17-18
),
e1953172
(
2021
).
62.
A. A.
Ehrhard
,
M. P.
Klein
,
J.
Mohrbach
,
S.
Dillinger
, and
G.
Niedner-Schatteburg
,
J. Chem. Phys.
156
(
5
),
054308
(
2022
).
63.
A.
Straßner
,
M. P.
Klein
,
D. V.
Fries
,
C.
Wiehn
,
M. E.
Huber
,
J.
Mohrbach
,
S.
Dillinger
,
D.
Spelsberg
,
P. B.
Armentrout
, and
G.
Niedner-Schatteburg
,
J. Chem. Phys.
155
(
24
),
244306
(
2021
).
64.
A.
Straßner
,
C.
Wiehn
,
M. P.
Klein
,
D. V.
Fries
,
S.
Dillinger
,
J.
Mohrbach
,
M. H.
Prosenc
,
P. B.
Armentrout
, and
G.
Niedner-Schatteburg
,
J. Chem. Phys.
155
(
24
),
244305
(
2021
).
65.
S.
Dillinger
,
J.
Mohrbach
,
J.
Hewer
,
M.
Gaffga
, and
G.
Niedner-Schatteburg
,
Phys. Chem. Chem. Phys.
17
(
16
),
10358
10362
(
2015
).
66.
S.
Dillinger
,
M. P.
Klein
,
A.
Steiner
,
D. C.
McDonald
,
M. A.
Duncan
,
M. M.
Kappes
, and
G.
Niedner-Schatteburg
,
J. Phys. Chem. Lett.
9
(
4
),
914
918
(
2018
).
67.
M. P.
Klein
,
A. A.
Ehrhard
,
J.
Mohrbach
,
S.
Dillinger
, and
G.
Niedner-Schatteburg
,
Top. Catal.
61
(
1
),
106
118
(
2018
).
68.
M. J.
Frisch
,
G. W.
Trucks
,
H. B.
Schlegel
,
G. E.
Scuseria
,
M. A.
Robb
,
J. R.
Cheeseman
,
G.
Scalmani
,
V.
Barone
,
B.
Mennucci
,
G. A.
Petersson
,
H.
Nakatsuji
,
M.
Caricato
,
X.
Li
,
H. P.
Hratchian
,
A. F.
Izmaylov
,
J.
Bloino
,
G.
Zheng
,
J. L.
Sonnenberg
,
M.
Hada
,
M.
Ehara
,
K.
Toyota
,
R.
Fukuda
,
J.
Hasegawa
,
M.
Ishida
,
T.
Nakajima
,
Y.
Honda
,
O.
Kitao
,
H.
Nakai
,
T.
Vreven
,
J. A.
Montgomery
, Jr.
,
J. E.
Peralta
,
F.
Ogliaro
,
M.
Bearpark
,
J. J.
Heyd
,
E.
Brothers
,
K. N.
Kudin
,
V. N.
Staroverov
,
R.
Kobayashi
,
J.
Normand
,
K.
Raghavachari
,
A.
Rendell
,
J. C.
Burant
,
S. S.
Iyengar
,
J.
Tomasi
,
M.
Cossi
,
N.
Rega
,
J. M.
Millam
,
M.
Klene
,
J. E.
Knox
,
J. B.
Cross
,
V.
Bakken
,
C.
Adamo
,
J.
Jaramillo
,
R.
Gomperts
,
R. E.
Stratmann
,
O.
Yazyev
,
A. J.
Austin
,
R.
Cammi
,
C.
Pomelli
,
J. W.
Ochterski
,
R. L.
Martin
,
K.
Morokuma
,
V. G.
Zakrzewski
,
G. A.
Voth
,
P.
Salvador
,
J. J.
Dannenberg
,
S.
Dapprich
,
A. D.
Daniels
,
Ö.
Farkas
,
J. B.
Foresman
,
J. V.
Ortiz
,
J.
Cioslowski
,
and D. J.
Fox
, GAUSSIAN 09, Revision E.01,
Gaussian, Inc.
,
Wallingford, CT
,
2009
.
69.
M. J.
Frisch
,
G. W.
Trucks
,
H. B.
Schlegel
,
G. E.
Scuseria
,
M. A.
Robb
,
J. R.
Cheeseman
,
G.
Scalmani
,
V.
Barone
,
G. A.
Petersson
,
H.
Nakatsuji
,
X.
Li
,
M.
Caricato
,
A. V.
Marenich
,
J.
Bloino
,
B. G.
Janesko
,
R.
Gomperts
,
B.
Mennucci
,
H. P.
Hratchian
,
J. V.
Ortiz
,
A. F.
Izmaylov
,
J. L.
Sonnenberg
,
D.
Williams-Young
,
F.
Ding
,
F.
Lipparini
,
F.
Egidi
,
J.
Goings
,
B.
Peng
,
A.
Petrone
,
T.
Henderson
,
D.
Ranasinghe
,
V. G.
Zakrzewski
,
J.
Gao
,
N.
Rega
,
G.
Zheng
,
W.
Liang
,
M.
Hada
,
M.
Ehara
,
K.
Toyota
,
R.
Fukuda
,
J.
Hasegawa
,
M.
Ishida
,
T.
Nakajima
,
Y.
Honda
,
O.
Kitao
,
H.
Nakai
,
T.
Vreven
,
K.
Throssell
,
J. A.
Montgomery
, Jr.
,
J. E.
Peralta
,
F.
Ogliaro
,
M. J.
Bearpark
,
J. J.
Heyd
,
E. N.
Brothers
,
K. N.
Kudin
,
V. N.
Staroverov
,
T. A.
Keith
,
R.
Kobayashi
,
J.
Normand
,
K.
Raghavachari
,
A. P.
Rendell
,
J. C.
Burant
,
S. S.
Iyengar
,
J.
Tomasi
,
M.
Cossi
,
J. M.
Millam
,
M.
Klene
,
C.
Adamo
,
R.
Cammi
,
J. W.
Ochterski
,
R. L.
Martin
,
K.
Morokuma
,
O.
Farkas
,
J. B.
Foresman
, and D. J.
Fox
, GAUSSIAN 16, Revision C.01,
Gaussian, Inc.
,
Wallingford, CT
,
2016
.
70.
J. P.
Perdew
,
K.
Burke
, and
M.
Ernzerhof
,
Phys. Rev. Lett.
78
(
7
),
1396
(
1997
).
71.
C.
Adamo
and
V.
Barone
,
J. Chem. Phys.
110
(
13
),
6158
6170
(
1999
).
72.
F.
Weigend
and
R.
Ahlrichs
,
Phys. Chem. Chem. Phys.
7
(
18
),
3297
3305
(
2005
).
73.
D.
Andrae
,
U.
Häußermann
,
M.
Dolg
,
H.
Stoll
, and
H.
Preuß
,
Theor. Chim. Acta
77
(
2
),
123
141
(
1990
).
74.
D.
Figgen
,
K. A.
Peterson
,
M.
Dolg
, and
H.
Stoll
,
J. Chem. Phys.
130
(
16
),
164108
(
2009
).
75.
E. D.
Glendening
,
J. K.
Badenhoop
,
A. E.
Reed
,
J. E.
Carpenter
,
J. A.
Bohmann
,
C. M.
Morales
, and
F.
Weinhold
, NBO 5.9,
2009
.
76.
W. M.
Haynes
,
CRC Handbook of Chemistry and Physics
(
CRC Press Taylor & Francis Group
,
2014
).
77.
M. P.
Klein
, Ph.D. thesis,
Technische Universität Kaiserslautern
,
2021
.

Supplementary Material