Advertisement

Aluminum cluster for CO and O2 adsorption

  • Bipasa SamantaEmail author
  • Turbasu Sengupta
  • Sourav PalEmail author
Original Paper
  • 171 Downloads
Part of the following topical collections:
  1. International Conference on Systems and Processes in Physics, Chemistry and Biology (ICSPPCB-2018) in honor of Professor Pratim K. Chattaraj on his sixtieth birthday

Abstract

Low temperature oxidation of CO to CO2 is an important process for the environment. Similarly adsorption of CO from the releasing sources is also of major concern today. Whereas the potential of gold and silver clusters is well proven for the catalysis of the above mentioned reaction, the potential of aluminum (Al) clusters remains unexplored. The present study proves that, similar to the transition metals, Al clusters can also be used for adsorption of gases. We first tested the potential of Al cluster as adsorbents for CO. The high binding energy (BE) values prove that Al clusters can be used for adsorbing both CO and O2. Since oxygen binding is more facile, we adsorbed oxygen on Al and then checked the effect of this O2 on the BE of CO. The results were obtained by DFT calculations at M062X/TZVP level of theory.

Graphical abstract

Activation of carbon monoxide (CO) on oxygen-adsorbed aluminum (Al) cluster

Keywords

DFT Aluminum clusters CO adsorption O2 adsorption Charge decomposition analysis Wiberg bond indices 

Introduction

The modification of nanoclusters as adsorbents and catalysts to adjust the selectivity and efficiency of reactions is of major research interest nowadays [1, 2, 3, 4, 5]. Considering all the materials currently used in catalysis and adsorption, e.g., transition metal catalyst, metal organic framework (MOF), covalent organic framework (COF), etc., the most promising are the nanoscale catalysts. Nanocluster chemistry represents an area of science that bridges the gap between the very reactive microscopic level and the bulk matter. Nanoclusters contain only a few atoms, expanding over the nanometer range or more, yet they have attracted much attention due to their unique and unanticipated properties, which do not have linear relationship with either bulk or micro level properties [6, 7]. Their fascinating behavior is due to their very high surface to volume ratio, electronic and geometric shell closing, superatomic character, and the distribution of electrons among the atoms, all of which are quite different from the bulk material due to quantum confinement [8]. These properties, and the characteristics and reactivity of clusters can be handled very selectively by addition or removal of atoms. In a cluster, the high surface to volume ratio plays a crucial role in adsorption as well as catalysis, as the number of under-coordinated atoms [9] and unsaturation increases, which in turn increases the adsorption and catalytic activity of the clusters. Furthermore, the activation barrier is lowered due to large surface area [9]. A detailed study of adsorbates and metal cluster surface interactions is important in the study of adsorption and heterogeneous catalysis.

Carbon monoxide (CO) is a polar molecule with almost zero dipole moment that is used widely in organometallic synthesis, but excess CO in the atmosphere is harmful. When inhaled, it causes hemoglobin poisoning and may eventually lead to death. It has been observed that the presence of CO in fuel cell poisons platinum (Pt) anodes [10, 11] and decreases power density [12]. Thus abatement of such pollutants is necessary. One of the most effective means to remove CO is via oxidation to carbon dioxide (CO2) or by removal of such gases by adsorption onto some other metal surface. Furthermore, this oxidation is important in combustion exhaust, development of CO detection devices and improving the efficiency of CO2 lasers [13]. Also, the adsorbed CO can be used in methanation or steam reforming of hydrocarbons and thus is of great economic importance. CO oxidation to CO2 is exothermic by 283 kJ mol−1, and has a high activation barrier due to the need for O–O bond breakage. Thus, it is not possible to oxidize CO to CO2 in gas phase under ambient conditions. To reduce this activation barrier, transition metal catalyst is employed [14, 15, 16, 17, 18].

Transition metals are have been applied widely as adsorbent materials and catalysts in the study of this reaction [14, 15, 16, 17, 18] but very few studies have been done on clusters of the main group. For our purposes, we have chosen the aluminum (Al) cluster. Al is abundantly available on the Earth’s surface and has a simple electronic configuration (consisting only of s and p electrons), which makes it a simple model to study using first principle calculations that can easily provide information on the origin of catalytic activity and detailed insights into reaction mechanisms. Al surfaces and clusters are easy to oxidize; the oxidation reaction is highly exothermic, which drives the reaction in the forward direction and mostly causes dissociation of oxygen molecules to oxygen atoms [19, 20, 21, 22, 23, 24]. The oxygen atoms formed can be used for CO to CO2 oxidation. Further, using transition metals is an expensive affair, as compared to Aluminium and other transition metals have certain problems in CO oxidation selectivity [25]. Al clusters (Al13) have unique properties like they behave as halogens [26], forms interhalogenic bonds and polyhalides [27]. \( {\mathrm{Al}}_7^{-} \)cluster shows multiple valency similar to carbon groups [28]. \( {\mathrm{Al}}_{13}^{-} \)and \( {\mathrm{Al}}_7^{-} \) acts like inert noble gases [29, 30, 31, 32]. Moreover, Al clusters help in the activation of various bonds, like nitrogen (N2) [33], oxygen (O2) [34, 35], etc. They have also shown improved adsorption of thiols [36], ammonia [37], water (H2O) [38], etc. Though studies are present for O2 and CO adsorption on the Al surfaces, very few reports in the literature is present on CO adsorption on pre-oxidized Al clusters and study of mechanism of CO to CO2 catalysis on such clusters. To mention a few, Bagus et al. [39] calculated the Al–CO distance as 3.75 Bohrs, and the interaction and binding energies to be 0.23 eV and 3.85 eV, respectively. They also reported the interaction energy of Al4CO to be −3.84 eV. Using pulsed fast flow reactor techniques, Cox et al. [40] studied CO adsorption on various transition metals, like vanadium (V), iron (Fe), cobalt (Co), nickel (Ni), copper (Cu), niobium (Nb), molybdenum (Mo), ruthenium (Ru), palladium (Pd), tungsten (W), iridium (Ir), and Pt as well as Al, for clusters containing up to 14 atoms. They successfully assigned the CO bond strength as Co, Pt, Pd > V, Ni, Nb, Ru, Ir, W > Mo, Fe > Cu, Al. Interaction of adsorbates on the Al surface has been studied by Pireaux et al. [41] by plasmon spectroscopy. Shiraka et al. [42] studied the two stages of adsorption using auger electron spectroscopy on the Al surface. The first step is Al–C bond formation, and then dissociation of CO occurs to form an Al–O-type bond. Thus, a CO molecule is adsorbed onto the Al surface with the carbon end first. This same group further studied O2 and CO adsorption on the Al surface, observing that, with increasing O2 and CO content, the intensity of metal auger transition decreases, ensuring that the metal surface is covered by the passed O2 and CO gas with oxygen being adsorbed much more quickly on the Al surface than CO [43]. Post et al. [44] showed weak adsorption at hollow sites and strong adsorption at the top site for Al atoms. Ruatta et al. [21] studied \( {\mathrm{Al}}_n^{+} \) with n = 1–9 with O2 and N2O. They observed exoergic oxidation of the cluster ion above the collision threshold with the activation barrier being highly dependent on cluster size. Collison-induced dissociation of AlnO+ and AlnO2+, which involves fragmentation of Al clusters with loss of Al2O molecules, was reported by Jarrold and Bower [19]. Cox et al. [45] reported rate constants of different size clusters at 350–500 K and high pressure. It was seen from their data that atoms and dimers have high reactivity, and Al3−Al7 has low reactivity; reactivity increases again from Al8−Al13. Crowell et al. [46] and Hoffman et al. [47] both observed that O2 is dissociatively chemisorbed on the Al surface, and the corresponding activation barrier is very small. Wang et al. [24] reported the binding energy (BE) of Al13 with O2 as 3.36 eV, 3.46 and 1.06 eV on hollow surface, bridged and atop positions, respectively. Cox et al. [48] studied the reactivity of various molecules towards Al clusters. They found that reactivity is highly dependent on the number of atoms in the Al cluster. They ordered the reactivity as O2 > CH3OH > CO > D2O > D2 > CH4, with CH4 showing no reaction under their experimental conditions. It was shown experimentally by Castleman and co-workers [49] that, in even electron anion cluster (both Al and Al-hydrogen based), there is spin transfer when they react with oxygen. This transfer is to combined with electrons of the antibonding orbital on triplet oxygen to form a singlet state. Their theoretical investigations showed that although \( {\mathrm{Al}}_{13}^{-} \) have high electron affinities, when they react with O2, they cleave the O−O bond. They finally concluded that magic clusters like \( {\mathrm{Al}}_{13}^{-} \) , which have high spin excitation energy, are resistant to any kind of reaction, whereas odd electron clusters have low excitation energy and are highly reactive. Through their experimental results, Jarrold et al. [19] have bracketed the interaction energy of Aln/O. They found a value of 8.0 ± 1.0 eV for \( {\mathrm{Al}}_6^{+} \) and 7.5 ± 1.0 eV for \( {\mathrm{Al}}_{19}^{+} \). Jarrold and Bower [20] measured the reaction cross section of Al3 to Al26, and found that reaction cross section increases with cluster size. They estimated the heat of chemisorption of oxygen atom on an Al cluster to be order of 10 eV. Castleman and co-workers [22] studied\( {\mathrm{Al}}_{\mathrm{n}}^{+} \), n = 1–33 and\( {\mathrm{Al}}_{\mathrm{n}}^{-} \), n = 5–37 reaction with oxygen and noticed that \( {\mathrm{Al}}_7^{+},{\mathrm{Al}}_{13}^{-},\kern0.5em \mathrm{and}\ {\mathrm{Al}}_{23}^{-} \) are unreactive. The same groups [50] have studied the reactivity of Al oxide clusters (Al2O3) towards CO and N2O. They concluded that anionic clusters are less reactive than cationic clusters but have higher selectivity towards transfer of oxygen atom during oxidation. They observed that oxygen-deficient clusters having one oxygen atom less than bulk alumina, i.e., Al2O2, when reacted with N2O, readily form both stoichiometric and oxygen-rich species.

From the above discussion, it is clear that, although a large number of studies have been done for oxygen atom and CO separately, there is a definite deficiency of studies that include both molecules.

In this paper, we present a density functional theory (DFT) study of CO and O2 adsorption on Al clusters and test the potential of oxidized cluster as an adsorbent for CO. The paper is divided into four sections. Computational details outlines the procedure adopted. Results and discussion describes the thermodynamic and kinetic parameters and other data obtained, and comparisons with other metal clusters, and, finally, Conclusions sums up our observations.

Computational details

The initial starting geometry of the pristine Al clusters was obtained from the literature [51, 52, 53]. The initial coordinates of Al clusters as given in the literature were optimized at M062X level of theory with Grimme’s dispersion added (DFT-D), as implemented in Gaussian 09 [54]. The basis set Triple Zeta Valence plus Polarization (TZVP) was applied for all atoms and used for the entire calculation. For this study, we used Al6 and Al13 clusters. Both neutral and anionic clusters of Al6 and Al13 were studied here. Fukui functions were calculated on initial optimized structures to get the different reactive sites. On these different reactive sites, CO and O2 were attached separately in different positions to get the lowest possible minima. While we were trying to find out the lowest possible minima, we optimized structures with different spin multiplicity (S = 1,2,3,4,5,6,7,8,9) and checked which spin multiplicity gave the lowest minima. With the lowest possible minima of oxygen attached Al cluster, CO was adsorbed at different positions to obtain the pre-reaction complex of the oxidation process of CO to CO2. The isosurface representation of the Fukui function (f+) of Al clusters was calculated by subtracting the corresponding density volume data of (N + 1) and (N) electronic systems at equilibrium geometry using the M062X/TZVP level of theory. The condensed Fukui function (f+)was calculated via the following equation [55]
$$ {f}_A^{+}={q}_{A,N+1}-{q}_{A,N} $$
(1)
where qA, N + 1 and qA, N are the Hirshfeld population (HP) of atom A for the (N + 1) and (N) electronic systems, respectively. When we wanted to attach CO to the O2 attached cluster, we chose sites near to the oxygen atom in order to facilitate the oxidation reaction, which might occur on the surface. At every step of optimization, we ensured that there was no imaginary frequency and all structures taken were local minima and not saddle points. We then calculated the BE of CO and O2 adsorption, respectively,
$$ \varDelta \mathrm{E}={\mathrm{E}}_{\mathrm{CO}-\mathrm{CLUSTER}}-{\mathrm{E}}_{\mathrm{CLUSTER}}-{\mathrm{E}}_{\mathrm{CO}\kern2.5em } $$
(2)
or
$$ \varDelta \mathrm{E}={\mathrm{E}}_{\mathrm{CO}-\mathrm{CLUSTER}-{\mathrm{O}}_2}-{\mathrm{E}}_{\mathrm{CLUSTER}-{\mathrm{O}}_2}-{\mathrm{E}}_{\mathrm{CO}} $$
(3)
where the subscripted terms denote the energy of the respective species. ECO − CLUSTER is the energy of the CO adsorbed Al cluster, \( {\mathrm{E}}_{\mathrm{CO}-\mathrm{CLUSTER}-{\mathrm{O}}_2} \) is the energy of the CO and O2 adsorbed clusters, \( {\mathrm{E}}_{\mathrm{CLUSTER}-{\mathrm{O}}_2} \) is the energy of the oxygen molecule adsorbed clusters, ECLUSTER is the energy of the pristine cluster and ECOis the energy of the CO molecule. Zero point energy correction is included in all the reported values. Further, we have used the Multiwfn [56] package to calculate the charge transfer, charge decomposition analysis [57] and charge density deformation plots. The second-order perturbative estimation of the donor−acceptor stabilization energy (Es) within the Natural Bond Orbital (NBO) basis and the Wiberg Bond Indices (WBI) were computed by the NBO 3.1 package implemented in Gaussian 09 [58]. The expression of Es is given as,
$$ {E}_s=\Delta {E}_{ij}={q}_I\frac{F_{ij}^2}{\Delta {\in}_{ji}} $$
(4)
where qi is donor orbital occupancy number. Fij is off-diagonal NBO Fock matrix elements. ji = ∈j − ∈i is the orbital energy difference between acceptor (j) and donor (i) NBO. We have also calculated the Basis Set Superposition Error (BSSE) and found negligible changes from reported values. The results are presented in supporting information Tables S1S3 and are not discussed further.

Results and discussion

We will divide the results and discussion into three parts. First, we discuss the adsorption of CO and O2 onto the cluster separately, and finally the effect of oxygen on CO adsorption. We have taken Al6 and Al13 anionic and neutral clusters for our study. The reason for doing so is that Al6 is the first three dimensional (3D) cluster, Al13 is close to a “magic” cluster and \( {\mathrm{Al}}_{13}^{-} \) is a double magic number cluster.

CO adsorption

After determining the reactive sites in the cluster using the Fukui function, CO was attached at all sites and the structures were optimized at M062X/TZVP level of theory. From the literature [42, 43], we know that when CO is adsorbed onto a metal surface, the carbon end is adsorbed first, and when O2 and CO are both present in the gas mixture, O2 is adsorbed first. Thus, when we adsorb CO, we did not take into account the oxygen end adsorption site. Moreover, Landman et al. [61] have also shown that only mono-carbonyl peaks are formed and no poly carbonyl peaks are detected. Thus, we attached only one CO in our study. Among all possible orientations taken for CO binding, e.g., head-on, parallel, side, atop, bridged, etc. (see supporting information Fig. S1), the lowest energy orientation was taken as the ground state of the CO adsorbed complex. From the optimized structures obtained, we see that, although we tried different types of bonding, the minima turn out to be very specific. In the case of the Al6 neutral cluster, only one specific minima was obtained, where CO is attached to three Al atoms in bridged form with BE −1.331 eV. \( {\mathrm{Al}}_6^{-} \) has two minima, where the lowest one is in perfect bridged form with BE −0.799 eV and the second one is a slightly distorted bridged structure with BE −0.762 eV. For Al13, two structures were obtained, the first, which is also the lowest one (BE = − 0.377 eV), is in head on position and the second, which is the second lowest one (BE = −0.344 eV), a slightly distorted bridged structure. In case of \( {\mathrm{Al}}_{13}^{-} \) the structures differ in BE very slightly from each other. The highest and lowest are listed in Table 1. The structures are also almost the same with CO bound in side-on form but are small distance away from the cluster. This is because \( {\mathrm{Al}}_{13}^{-} \)is a magic cluster and shows the lowest BE. Hence, the distance between the cluster and CO is large. The respective optimized structures are shown in Fig. 1, with the lowest conformer named as (a), the second as (b) and so on. From Fig. 1, we can say that CO mostly prefers the bridged form to bind to an Al cluster. The lowest energy complexes for carbon monoxide adsorption on Al clusters in terms of BE, bond length (BL), charge decomposition analysis (CDA), WBI and Es are shown in Table 1.
Table 1

Binding energies (B.E) of different conformers of aluminium cluster - CO complex are presented. Few other important parameters like Wiberg Bond Indices (WBI), results of Natural Bond Orbital (NBO) analysisa and bond length (BL) of C-O and Al-C are also included

System

Configuration

BE (eV)

BL (Å)

Charge transfer (cluster→CO) |e|

WBI

NBO

C–O

Al–C

C–O

Al–C

Donor

Acceptor

Es (kcal mol−1)

Al 6

(a)

−1.331

1.166

2.248

0.2607

1.75

0.47

BD–C–O

LP*–Al

64.99

\( {Al}_6^{-} \)

(a)

−0.791

1.177

2.014

0.3143

1.71

0.79

CR–C

BD*–Al–C

55.09

(b)

−0.762

1.175

2.098

0.3780

1.69

0.66

LP–C

LP*–Al

67.31

Al 13

(a)

−0.377

1.172

2.108

0.2769

1.90

0.58

LP–O

BD*–Al–C

12.81

(b)

−0.344

1.132

2.015

0.0053

2.14

0.67

LP–C

LP*–Al

93.62

\( {Al}_{13}^{-} \)

(a)

−0.095

1.123

3.649

0.0138

2.25

0.05

LP–C

LP*–Al

2.35

(b)

−0.089

1.123

3.550

0.0059

2.25

0.06

LP–C

LP*–Al

3.40

*Atoms connected by dotted bonds in all figures are participants in donor–acceptor interactions

Fig. 1

Different binding modes of carbon monoxide (CO) are presented with the lowest energy conformer; (a) represents the lowest energy conformer, (b) the second lowest and so on. The bond length (BL) of Al–C and C–O are also represented here in Å

According to the jellium model [62, 63, 64], structures with closed shell configuration are the most stable. They are generally unreactive and show low BE and charge transfer. \( {\mathrm{Al}}_{13}^{-} \) is one such magic cluster. It is expected that it will show low BE as compared to the other clusters. If we look at Table 1, we see that indeed the \( {\mathrm{Al}}_{13}^{-} \) cluster exhibits the lowest BE among all the clusters taken for the study; this may be accounted for by the fact that it is the first icosahedral structure of the Al family cluster to have a closed shell electronic configuration. That is to say that it is a magic cluster not only electronically, but also geometrically. Therefore, the observed low binding energies of \( {\mathrm{Al}}_{13}^{-} \) cluster can easily be explained. Furthermore, the CDA analysis also shows a low amount of charge transferred from the cluster to CO, i.e., 0.0059 |e|. The BL of CO is almost like that of free CO, hence, we can say that \( {\mathrm{Al}}_{13}^{-} \) is not activating CO like the other clusters are doing. Al6 is one of the clusters that is expected to have a filled shell after adsorption of CO. After adsorbing CO, the Al6 cluster will accept two electrons from a CO molecule (CO is a two electron donor molecule) and it will then have 20 electrons, making this complex a magic cluster. Thus, the observed high BE of this cluster can be explained as an effect of attainment of closed shell configuration. The charge transferred in these cases is also high, which is in agreement with the high BE. The other two clusters i.e., \( {\mathrm{Al}}_6^{-} \) and Al13 exhibit intermediate BE. This is because they do not exactly form a magic number cluster after CO adsorption. Upon accepting two electrons from CO\( {\mathrm{Al}}_6^{-} \) and Al13 attain a jellium configuration, with one electron more than the closet electronic magic number. The exact order of BE is Al6>\( {\mathrm{Al}}_6^{-} \)>Al13>\( {\mathrm{Al}}_{13}^{-} \). The CO bond is activated to 1.77 \( \overset{{}^{\circ}}{A} \) with bond order (WBI) 1.69 by \( {\mathrm{Al}}_6^{-} \) cluster. WBI and BL values from Table 1 also suggest that CO is being adsorbed and is activated to a good extent. The stabilization energy value also suggests that a certain amount of interaction is taking place with a sufficient Es value. NBO analysis revealed that this occurs mainly by a donation from LP of the C or O atom of CO to LP* of Al of the cluster.

Kaldor co-workers [40] and Cox et al. [45] have also shown that BE decreases with increasing size. From Table 1, we see that Al6 BE is greater than that of Al13; thus, we also get a similar trend. The reason may be that smaller clusters have more dangling bonds than larger clusters, and smaller clusters have a greater surface-to-volume ratio compared to larger clusters. Also in the literature, smaller cluster shows better adsorption and binding energies [19].

Although we obtained low BE for Al cluster as compared to gold (Au18 = −0.83 eV) [66] and Pd [67], we have much higher BE for the Al cluster as compared to Fe (−0.73), Ni (−0.23) and Cu (−0.44) [68] . Thus, with the exception of Au and Pd, Al clusters are better at binding CO than any other noble or precious metal.

Moreover, the dipole moment calculated for the complexes was higher than that of their pristine parts. Before adsorption the dipole moment was almost zero, whereas after adsorption the dipole moment increases. This increased dipole moment is important in adsorption studies as it helps enhance long-range interaction between the cluster and the CO molecule. This long-range interaction helps trap the CO molecule on the cluster surface [65]. The greater the trapping of CO, the greater the BE seen for the clusters and CO. Cluster complexes that show high to good binding energies are also characterized by high vibrational frequencies along with good dipole moment values. Thus, from all the parameters calculated, we can say that Al clusters can be used for CO bond activation and adsorption.

Oxygen adsorption

We know that oxygen exists in the ground state as a triplet. Unlike the case of CO adsorption, where adsorption depends on a jellium model, oxygen adsorption actually depends on the way the number of electrons is arranged in the cluster (i.e., the number of unpaired electrons). Castleman and co-workers [49] have shown that, for both pristine and H-adsorbed Al clusters, oxygen reacts faster with systems that have an odd number of electrons than those with an even number of electrons. This is because of the spin excitation that is required for even electron cluster. With this concept, we optimized the O2-adsorbed complex with different spin multiplicities (M = 2S + 1) to see which spin state gives the lowest energy complex. For Al6, we considered all possible spin multiplicities within the range 1–9. Similarly, for Al13, we considered spin multiplicities within the range 1–8. The spins are noted in superscript in second bracket (e.g., \( {\mathrm{Al}}_6^{-} \) with M = 2 represented as \( {\mathrm{Al}}_6^{-(2)} \)). All possible modes of binding of oxygen, e.g., atop, parallel mode, side on overlap, bridge form, etc., were optimized at M062X/TZVP level of theory and from all the optimized structures, only the lowest one was taken, with no negative frequency, ensuring that it is a minima. The lowest minima structures for \( {\mathrm{Al}}_6^{(1)} \),\( {\mathrm{Al}}_6^{(3)},{\mathrm{Al}}_6^{(5)},{\mathrm{Al}}_6^{(7)} \) \( \mathrm{and}\ {\mathrm{Al}}_6^{(9)} \) are presented in Fig. 2. From all the structures obtained, we infer that, although different conformers were attempted, oxygen prefers to become localized in three forms: mainly in side-on position with one oxygen atom below the Al–Al bond and the other above the Al–Al bond, secondly in a parallel bridged form where the O2 molecule is placed parallel to the surface held by three Al atoms, and finally atop mode where both O atoms are attached to only one Al atom. From the optimized structures, it can also be seen that the lowest energy structures are different in different spin. The lowest energy structure in spin multiplicity 1 may not be the lowest one in spin multiplicity 3. From the different optimized spin complexes, we note the trend in BE as spin 1 < 3 < 5 < 7 < 9 (Table 2). For BE comparison we took similar structures in each spin. The similar structures in different spin for \( {\mathrm{Al}}_6^{(1)} \),\( {\mathrm{Al}}_6^{(3)},{\mathrm{Al}}_6^{(5)},{\mathrm{Al}}_6^{(7)} \) \( \mathrm{and}\ {\mathrm{Al}}_6^{(9)} \) are presented in Fig. 3. Furthermore, the more we move to higher spin, more of the binding site shows positive BE for the oxygen molecule and the structures obtained move towards greater distortion. For Al6, the lowest energy complexes are obtained in spin 1 with the highest BE for the oxygen molecule. The highest BE (−3.717 eV) is shown by conformer (a) . Upon adsorbing oxygen, it activates the O–O bond to 1.51 \( \overset{{}^{\circ}}{A} \). WBI calculation also supports activation to single bond order with the value 1.01. The CDA value also shows transfer of charge from the cluster to the oxygen molecule, which helps activation of the bond. For \( {\mathrm{Al}}_6^{-} \) we optimized structure with spin multiplicities 2, 4, 6 and 8. From all the optimized structures, we noticed that in \( {\mathrm{Al}}_6^{-(6)} \) and \( {\mathrm{Al}}_6^{-(8)} \) highly distorted structures are obtained with positive BE in most of the minima obtained. In fact, in the case of \( {\mathrm{Al}}_6^{-(6)} \) no proper minima were obtained to compare with the \( {\mathrm{Al}}_6^{-(2)} \) minima structures. For \( {\mathrm{Al}}_6^{-(8)} \) all properly obtained structures exhibit positive BE. The ones with negative BE are too highly distorted to be of interest. The lowest energy structures for \( {\mathrm{Al}}_6^{-(2)} \), \( {\mathrm{Al}}_6^{-(4)} \), \( {\mathrm{Al}}_6^{-(6)} \) and \( {\mathrm{Al}}_6^{-(8)} \) are presented in Fig. 4 and similar structures in different spin are present in Fig. 5. From all the structures obtained, we note that the lowest ones are similar to Al6. The lowest structure is the side-on adsorbed form with one O atom below the Al–Al bond and the other above the Al–Al bond. The second lowest minima structure is the bridged form, where the O2 molecule becomes adsorbed on the surface, held by three Al atoms. The O2 molecule becomes adsorbed on this surface in parallel mode. The BE range obtained is −3.3 to −3.2 eV for spin 2 only, i.e., the variation in BE is less for spin 2. For higher spin, it is going up to −0.208 eV. The order for BE (with sign) for O2 is \( {\mathrm{Al}}_6^{-(2)} \)<\( {\mathrm{Al}}_6^{-(4)} \)<\( {\mathrm{Al}}_6^{-(6)} \)<\( {\mathrm{Al}}_6^{-(8)} \). The lowest energy bonded complex is \( {\mathrm{Al}}_6^{-(2)} \) conformer (a) where the BE is −3.303 eV with O–O BL 1.50 \( \overset{{}^{\circ}}{A} \) with WBI value 0.97. Thus, we see more activation of the O–O bond in anionic than in neutral state. As mentioned earlier, the number of unpaired electrons plays an important role in oxygen adsorption. In the case of Al6 and \( {\mathrm{Al}}_6^{-} \), \( {\mathrm{Al}}_6^{-} \) has unpaired electrons and exhibits more BE than Al6 (lowest spin, i.e., the ground state is considered in both cases). This is because, in the case of \( {\mathrm{Al}}_6^{-} \) no spin excitation is required, whereas for Al6 spin excitation is required. Even the stabilization energy Es is higher in the case of the anion than in the neutral state. From CDA, we see that the net charge transferred is greater in \( {\mathrm{Al}}_6^{-} \) than in Al6. The O2 bond is activated more by \( {\mathrm{Al}}_6^{-(2)} \) than \( {\mathrm{Al}}_6^{(1)} \). The WBI values from Table 3 suggest the same pattern.
Fig. 2

Different binding modes of O2 with Al6  are presented with the lowest energy conformers; (a) represents the lowest energy conformer, (b) the second lowest and so on. The BLs of Al–O and O–O are given in Ångstroms. Superscripted numbers Spin multiplicity

Table 2

BE of different spin complexes (in eV). Same conformers with different spin multiplicities were taken for comparison.

System

Conformer

M = 1

M = 3

M = 5

M = 7

Al6

(a)

−3.657

−3.079

a

−1.621

(b)

−3.591

−2.798

1.500

−1.620

(c)

a

−1.342

1.332

−0.338

\( {\mathrm{Al}}_{13}^{-} \)

(a)

−2.613

−1.790

a

a

(b)

−2.613

−1.720

−0.555

0.472

(c)

−2.447

−1.480

−0.317

0.878

(d)

−2.413

−1.470

−0.608

a

(e)

a

−0.840

0.374

1.552

(f)

a

−0.830

0.593

a

System

Conformer

M = 2

M = 4

M = 6

M = 8

\( {\mathrm{Al}}_6^{-} \)

(a)

−3.303

−2.534

a

a

(b)

−3.219

a

a

−0.165

(c)

a

−1.355

a

0.060

Al13

(a)

−2.893

−1.920

−0.661

0.453

(b)

−1.362

−1.360

0.340

1.746

(c)

−1.059

−0.190

1.073

2.280

aNo comparable structures were obtained

Fig. 3

The same geometric conformer of Al6 in different spin multiplicities is presented. Blank spaces mean the corresponding conformer in that spin was not obtained. (a) represents the lowest energy conformer, (b) the second lowest and so on. The respective binding energy (BE) is presented in eV. Superscripted numbers Spin multiplicity

Fig. 4

Different binding modes of O2 with \( {\mathrm{Al}}_6^{-} \) with the lowest energy conformer are presented; (a) represents the lowest energy conformer, (b) the second lowest and so on. The BLs of Al–O and O–O are given in Ångstroms. Superscripted numbers Spin multiplicity

Fig. 5

The same geometric conformer of \( {\mathrm{Al}}_6^{-} \) in different spin multiplicity is presented. Blank spaces mean the corresponding conformer in that spin was not obtained. (a) represents the lowest energy conformer, (b) the second lowest and so on. The respective BE is presented in eV. Superscripted numbers Spin multiplicity

Table 3

Binding energies (BE) of the different conformers of aluminium cluster -O2 complex are presented. Few other important parameters like Wiberg Bond Indices (WBI), results of Natural Bond Orbital (NBO) analysisa and bond length (BL) of O-O and Al-O are also included

System

Configuration

BE (eV)

BL (Å)

Charge transfer (cluster→O2) |e|

WBI

NBO

O–O

Al–O

O–O

Al–O

Donor

Acceptor

Es (kcal mol−1)

\( {Al}_6^{(1)} \)

(a)

−3.717

1.507

1.911

0.591

1.01

0.19

LP–O

LP*–Al

17.71

(b)

−3.657

1.483

1.860

0.609

0.99

0.20

LP–O

LP*–Al

26.30

(c)

−3.591

1.494

1.790

0.632

1.00

0.31

LP–O

LP*–Al

34.18

\( {Al}_6^{(3)} \)

(a)

−3.079

1.478

1.814

0.643

0.96

0.33

LP–O

LP*–Al

38.17

(b)

−2.798

1.499

1.767

0.603

0.97

0.39

LP–O

LP*–Al

46.98

(c)

−1.342

1.331

1.878

0.281

1.48

0.29

LP–O

LP*–Al

37.73

\( {Al}_6^{(5)} \)

(a)

−2.451

1.500

1.770

0.650

0.97

0.38

LP–O

LP*–Al

44.64

(b)

−1.159

1.550

1.728

0.673

1.01

0.46

LP–O

LP*–Al

55.87

(c)

−1.013

1.332

1.888

0.273

1.42

0.30

LP–O

LP*–Al

37.85

\( {Al}_6^{(7)} \)

(a)

−1.621

1.474

1.823

0.664

0.95

0.34

LP–O

LP*–Al

39.10

(b)

−0.338

1.333

1.891

0.270

1.18

0.42

LP–O

LP*–Al

42.80

\( {Al}_6^{(9)} \)

(a)

−0.003

1.471

1.809

0.648

0.94

0.36

LP–O

LP*–Al

39.15

\( {Al}_6^{-(2)} \)

(a)

−3.303

1.500

1.780

0.685

0.97

0.36

LP–O

LP*–Al

42.25

(b)

−3.219

1.493

1.800

0.705

0.93

0.33

LP–O

LP*–Al

41.65

\( {Al}_6^{-(4)} \)

(a)

−3.303

1.502

1.780

0.685

0.97

0.36

LP–O

LP*–All

42.24

(b)

−2.534

1.505

1.781

0.681

0.97

0.36

LP–O

LP*–Al

51.02

(c)

−1.355

1.331

1.920

0.704

1.50

0.26

LP–O

LP*–Al

36.64

\( {Al}_6^{-(6)} \)

(a)

−2.113

1.458

1.764

0.708

0.99

0.41

LP–O

LP*–Al

61.80

(b)

−2.025

1.462

1.757

0.724

0.98

0.41

LP–O

LP*–Al

56.87

(c)

−1.580

1.548

1.743

0.792

1.02

0.44

LP–O

LP*–Al

58.65

\( {Al}_6^{-(8)} \)

(a)

−0.208

1.468

1.780

0.680

0.99

0.39

LP–O

LP*–Al

40.22

(b)

−0.165

1.471

1.759

0.722

0.99

0.41

LP–O

LP*–Al

54.50

\( {Al}_{13}^{(2)} \)

(a)

−7.676

2.921

1.729

1.126

0.01

0.40

LP–O

LP*–Al

49.27

(b)

−2.893

1.467

1.827

0.664

0.97

0.34

LP–O

LP*–Al

50.74

(c)

−2.674

1.462

1.769

0.664

1.00

0.38

LP–O

LP*–Al

37.02

(d)

−1.362

1.338

1.877

0.298

1.46

0.30

LP–O

LP*–Al

45.34

(e)

−1.059

1.330

1.917

0.495

1.25

0.21

LP–O

LP*–Al

22.52

\( {Al}_{13}^{(4)} \)

(a)

−1.92

1.467

1.807

0.664

0.97

0.35

LP–O

LP*–Al

52.42

(b)

−1.90

1.468

1.814

0.652

0.97

0.34

LP–O

LP*–Al

32.19

(c)

−1.85

1.470

1.816

0.665

0.97

0.33

LP–O

LP*–Al

30.29

(d)

−1.51

1.452

1.772

0.634

1.00

0.41

LP–O

LP*–Al

37.79

(e)

−1.36

1.338

1.879

0.299

1.46

0.29

LP–O

LP*–Al

45.39

(f)

−0.71

1.234

1.885

0.124

1.48

0.44

LP–O

LP*–Al

47.38

(g)

−0.19

1.321

1.928

0.433

1.32

0.20

LP–O

LP*–Al

17.85

\( {Al}_{13}^{(6)} \)

(a)

−0.846

1.469

1.802

0.710

0.96

0.34

LP–O

LP*–Al

40.47

(b)

−0.686

1.474

1.818

0.703

0.93

0.34

LP–O

LP*–Al

38.43

(c)

−0.677

1.467

1.820

0.684

0.96

0.33

LP–O

LP*–Al

29.00

(d)

−0.675

1.467

1.819

0.651

0.97

0.34

LP–O

LP*–Al

29.98

(e)

−0.661

1.475

1.804

0.662

0.95

0.35

LP–O

LP*–Al

28.02

(f)

−0.611

1.460

1.820

0.650

0.97

0.35

LP–O

LP*–Al

34.54

\( {Al}_{13}^{-(1)} \)

(a)

−2.613

1.478

1.819

0.712

0.96

0.33

LP–O

LP*–Al

35.09

(b)

−2.447

1.457

1.766

0.684

1.00

0.40

LP–O

LP*–Al

45.85

(c)

−2.413

1.500

1.777

0.665

0.97

0.36

LP–O

LP*–Al

53.81

\( {Al}_{13}^{-(3)} \)

(a)

−1.79

1.503

1.910

0.668

0.94

0.23

LP–O

LP*–Al

21.68

(b)

−1.72

1.472

1.818

0.699

0.97

0.33

LP–O

LP*–Al

33.57

(c)

−1.48

1.462

1.787

0.689

1.00

0.38

LP–O

LP*–Al

35.64

(d)

−1.47

1.461

1.783

0.688

1.00

0.39

LP–O

LP*–Al

37.57

(e)

−0.84

1.334

1.906

0.340

1.50

0.27

LP–O

LP*–Al

38.85

(f)

−0.83

1.322

1.821

0.448

1.43

0.37

LP–O

LP*–Al

55.91

\( {Al}_{13}^{-(5)} \)

(a)

−0.598

1.468

1.821

0.700

0.97

0.33

LP–O

LP*–Al

34.67

(b)

−0.579

1.470

1.828

0.694

0.97

0.34

LP–O

LP*–Al

35.33

(c)

−0.569

1.478

1.821

0.696

0.96

0.34

LP–O

LP*–Al

28.08

(d)

−0.555

1.474

1.822

0.693

0.97

0.34

LP–O

LP*–Al

52.15

(e)

−0.317

1.458

1.782

0.676

1.00

0.39

LP–O

LP*–Al

36.91

*Atoms connected with dotted bonds in all figures are participants in donor–acceptor interactions

For Al13, a larger number of conformers were obtained after adsorbing oxygen in different forms. The corresponding structures are presented in Figs. 6 and 7. The lowest conformer is marked (a), the second lowest is marked (b) and so on. Here also we considered different spin multiplicities, i.e., for Al13 and \( {\mathrm{Al}}_{13}^{-} \) spin multiplicities considered are 1–8.
Fig. 6

Different binding modes of O2 with Al13 are presented with the lowest energy conformers; (a) represents the lowest energy conformer, (b) the second lowest and so on. BLs of Al–O and O–O are given in Ångstroms. Superscripted numbers Spin multiplicity

Fig. 7

Different binding modes of O2 with \( {\mathrm{Al}}_{13}^{-} \) are presented with the lowest energy conformers; (a) represents the lowest energy conformer, (b) the second lowest and so on. BLs of Al–O and O–O are given in Ångstroms. Superscripted numbers Spin multiplicity

In the case of the Al13 cluster, all forms of binding initially considered were obtained. For \( {\mathrm{Al}}_{13}^{(2)} \)neutral structure the lowest conformer actually shows O2 in dissociated form. The next two structures have O2 in side-on adsorbed mode. The next structure has O2 in atop mode and the highest energy conformer has O2 in head-on adsorbed mode. For \( {\mathrm{Al}}_{13}^{(4)} \)case, the first four structures have O2 adsorbed in parallel mode onto the surface held by three Al atoms, and the last three structures have O2 in atop mode with bent oxygen. For \( {\mathrm{Al}}_{13}^{-(1)} \) the lowest conformer has oxygen adsorbed in parallel mode on the surface held by three Al atoms. The second structure is side on overlap mode and the third is bridged, with one O atom above the Al–Al bond and the other below the Al–Al bond. In case of \( {\mathrm{Al}}_{13}^{-(3)} \) the two lowest structures have O2 in atop mode. The first four lowest energy structures have O2 adsorbed in parallel mode onto the surface held by three Al atoms. The corresponding BE, BL, WBI and Es values are given in Table 3. From Table 3 we see that \( {\mathrm{Al}}_{13}^{-(1)} \) exhibits more BE than \( {\mathrm{Al}}_{13}^{-(3)} \). As we move towards higher spin we see that the BE is gradually decreasing. The trend for Al13is \( {\mathrm{Al}}_{13}^{(2)} \)< \( {\mathrm{Al}}_{13}^{(4)} \), \( {\mathrm{Al}}_{13}^{(6)} \)< \( {\mathrm{Al}}_{13}^{(8)} \) and, in the case of \( {\mathrm{Al}}_{13}^{-} \) the trend is \( {\mathrm{Al}}_{13}^{-(1)} \)<\( {\mathrm{Al}}_{13}^{-(3)} \)<\( {\mathrm{Al}}_{13}^{(5)} \)<\( {\mathrm{Al}}_{13}^{(7)} \) (see Table 2, Figs. 8 and 9). For \( {\mathrm{Al}}_{13}^{(8)} \) and \( {\mathrm{Al}}_{13}^{-(7)} \)no negative BE complex was obtained. For spin 8 and 7, oxygen adsorption is thus not possible. For higher spin complexes, the distortion was less as compared to Al6. Coming to BE analysis, Al13 exhibits more BE than \( {\mathrm{Al}}_{13}^{-}. \) \( {\mathrm{Al}}_{13}^{(2)} \) conformer (a) has BE − 2.674 eV and \( {\mathrm{Al}}_{13}^{-(1)} \) conformer (a) has BE − 2.613 eV. This is because \( {\mathrm{Al}}_{13}^{-} \) has an even number of electrons and Al13has an odd number of electrons. Thus, spin excitation is not required in case of Al13 which is required for \( {\mathrm{Al}}_{13}^{-}. \)
Fig. 8

The same geometric conformer for Al13 in different spin multiplicity is presented; (a) represents the lowest energy conformer, (b) the second lowest and so on. The respective BE is presented in eV. Superscripted numbers Spin multiplicity

Fig. 9

Same geometric conformer of \( {\mathrm{Al}}_{13}^{-} \) in different spin multiplicity is presented; (a) represents the lowest energy conformer, (b) the second lowest and so on. Blank spaces means corresponding conformer in that spin is not obtained. The respective BE is presented in eV. Superscripted numbers Spin multiplicity

In general, the O=O BL is 1.2 Å and O−O bond is 1.4 Å. Table 3 clearly indicates that the O=O bond is highly activated, almost to the extent of the O−O bond and even sometimes activated to a single O atom like \( {\mathrm{Al}}_{13}^{(2)} \) conformer (a) where the O−O BL is 2.9 Å and the WBI value is 0.01, which indicates no bond. Furthermore, a value of 0.93 to 0.96 is seen most often in the WBI column of Table 3, which indicates high activation of the O–O bond, i.e., from second bond order to less than bond order 1. Mijoule and co-workers [60] used the B3PW91 functional and calculated the BE for Al13 to be −4.09 eV; we calculated it to be −7.676 eV for the lowest structure. Mijoule et al. [60] reported values for Al−O BL of 0.85–1.90 Å, whereas we calculated it to be 1.7 to 1.9 Å. Wang et al. [24] calculated the binding energies at various positions of Al13 and, using generalized gradient approximation (GGA), estimated the range to be −1.06 to −3.36 eV. Our range is −0.71 eV to −7.676 eV. Castleman and co-workers [49] have shown that \( {\mathrm{Al}}_{13}^{-} \) and \( {\mathrm{Al}}_5^{-} \) cleaves the O−O bond. Figure 6 also shows that both \( {\mathrm{Al}}_{13}\ \mathrm{and}\ {\mathrm{Al}}_{13}^{-} \) indeed cleave the O−O bond. They noted that O−O is activated to 1.5 \( \dot{\mathrm{A}} \). Our computed data suggests the same. Salahub and co-workers [59] bracketed the O2 BE on the Al surface to be −5.5 to −8.1 eV; our range is −0.71 to −7.676 eV. Crowell et al. [46] and Hoffman et al. [47] mentioned that O2 undergoes dissociative chemisorption on the Al surface. From Figure 6, we note that indeed O2 undergoes dissociative chemisorption. The WBI value is reduced to 0.94, which is not even a single bond. Thus, we can say that Al clusters highly activate the O–O bond and that chemisorption is occurring.

When compared to CO BE, the Al cluster shows greater BE with O2 than with CO. Even the CDA values and Es values are higher than with CO adsorption. Thus, it can be concluded that O2 adsorbs more than CO on the Al cluster surface. Also, when O2 and CO are both present in the mixture, at first O2 will become attached to the Al surface then become activated, then CO will come and take the O atom to form CO2, which then moves away.

The WBI value for Al–C and Al–O also suggest that there is no proper bond between the adsorbent and adsorbate. The values lie between 0.3 and 0.5, which suggests that a weak physical attraction between the two molecules is holding it together. Thus, when an extra CO enters, these oxygen atoms can be easily donated to form CO2.

From NBO analysis we see that the donor is always lone pair (LP) of O and the acceptor is LP* of Al. The Es values show that sufficient interaction takes place in between oxygen and Al.

CO adsorption on O2-adsorbed Al clusters

Thus far, we have seen that Al clusters are effective for CO and O2 adsorption, and that the CO and O2 bonds are pretty well activated. After this individual addition of O2 and CO, we proceeded to study the adsorption of CO on an O2-pre-adsorbed Al cluster. As mentioned earlier, CO to CO2 oxidation is an important process, hence studying the effect of oxygen on CO adsorption becomes a topic of interest. It has already been seen that pre-adsorbed oxygen on metal surface facilitates the oxidation process.

We started off with oxygen-adsorbed clusters as it has been seen that, when a mixture of oxygen and CO are both present, at first oxygen is absorbed on the surface and then CO. Following this observation, after the adsorption of oxygen at different position in Al clusters, the lowest energy geometry was taken for further adsorption of CO. Other possibilities may also exist for adsorbing CO onto higher energy conformers but, due to computational constraints, we followed the lowest energy isomer of the O2-adsorbed cluster for attaching CO. For finding the lowest energy structure, we adsorbed CO onto the oxygen adsorbed clusters at different positions. From Fig. 10, it can be seen that CO is adsorbed mainly close to the oxygen atom, and in atop form. This affinity for sites near oxygen is required for easy CO to CO2 oxidation. The presence of O and CO near to each other facilitates CO oxidation. Furthermore, it is seen that CO always adsorbs in atop mode, attaching to single Al atom.
Fig. 10

Different binding modes of CO with oxygen adsorbed clusters are presented here with the lowest energy conformer. BLs of C–O, O–O, Al–C and Al–O are given in Ångstroms. Superscripted numbers Spin multiplicity

The data shown in Table 4 reveal that CO adsorption on oxygen adsorbed clusters is quite feasible. CO binds to these clusters with BE in the range ~ −0.2 to 0.5 eV. Except for \( {\mathrm{Al}}_6^{-} \), the BE for CO increases for oxygen-adsorbed clusters. The BL of CO also increases compared with the BL of free CO. The amount of charge transferred from the cluster to CO also remains almost the same when adsorbed on a pristine cluster. Of note is the sign of CDA value. For Al6, the values are positive except for \( {\mathrm{Al}}_6^1 \), the charge of which is transferred from cluster to CO. For \( {\mathrm{Al}}_6^1 \), it may be that, upon adsorbing O2, the complex is nearing an electron count of 20. Thus, addition of extra CO to the complex is not feasible. This is further reflected by the low BE value of −0.14 eV and the WBI of the C–O bond, which is 2.17. Thus, the extent of activation seen in other Al6 cluster is absent from \( {\mathrm{Al}}_6^1. \)Al13 is exactly the opposite, with charge transferred from CO to cluster. This is reflected in the BL of CO, which is reduced to 1.119 \( \overset{{}^{\circ}}{A} \) i.e., slightly lower than the length of the free CO bond. From the WBI calculation, it is seen that the bond order for CO is <1. It is >1 in Table 1 but <1 in Table 4. NBO analysis revealed that there is sufficient interaction between LP of C of CO and the LP* of the Al atom of the cluster, as indicated by the Es value. The NBO picture of CO, O2 and both adsorbed clusters are represented in Fig. 11. The figure clearly represents the good amount of interaction between the cluster and the molecules and the corresponding Es values are present in Table 4.
Table 4

BE of CO on O2-adsorbed clusters

Cluster

Configuration

BE (eV)

BL (Å)

Charge transfer (complexCO) |e|

WBI

NBO

C–O

O–O

Al–C

Al–O

C–O

O–O

Donor

Acceptor

Es (kcal mol)−1

\( {\boldsymbol{Al}}_{\mathbf{6}}^{\left(\mathbf{1}\right)} \)

(a)

−0.148

1.130

1.519

2.155

1.903

−0.024

2.17

0.98

LP–O

LP*–Al

19.80

\( {\boldsymbol{Al}}_{\mathbf{6}}^{\left(\mathbf{3}\right)} \)

(a)

−0.40

1.152

1.479

2.074

1.818

0.209

0.95

2.05

LP–C

LP*–Al

22.28

\( {\boldsymbol{Al}}_{\mathbf{6}}^{-\left(\mathbf{2}\right)} \)

(a)

−0.506

1.171

1.501

2.069

1.787

0.291

0.97

1.95

LP–C

LP*–Al

38.66

\( {\boldsymbol{Al}}_{\mathbf{6}}^{-\left(\mathbf{4}\right)} \)

(a)

−0.030

1.164

1.513

1.992

1.807

0.264

0.98

1.94

LP–C

LP*–Al

82.23

\( {\boldsymbol{Al}}_{\mathbf{13}}^{\left(\mathbf{2}\right)} \)

(b)

−0.415

1.119

1.465

2.158

1.821

−0.176

0.97

2.31

LP–C

LP*–Al

63.87

\( {\boldsymbol{Al}}_{\mathbf{13}}^{\left(\mathbf{4}\right)} \)

(a)

−0.420

1.119

1.465

2.148

1.809

−0.174

0.97

2.31

LP–C

LP*–Al

72.31

\( {\boldsymbol{Al}}_{\mathbf{13}}^{-\left(\mathbf{1}\right)} \)

(a)

−0.090

1.123

1.478

3.704

1.821

0.013

2.24

0.96

LP–O

LP*–Al

37.77

\( {\boldsymbol{Al}}_{\mathbf{13}}^{-\left(\mathbf{3}\right)} \)

(a)

−0.270

1.169

1.520

2.043

1.901

0.208

0.93

2.09

LP–C

LP*–Al

29.89

*Atoms connected with dotted bonds in all figures are participants in donor–acceptor interactions

Fig. 11

Natural bond orbital (NBO) interaction between the donor and acceptor orbitals is presented. Green Positive surface, pink negative surface

We also studied electron deformation for the Al6 cluster after adsorption of CO, O2 or both (Fig. 12). From the counter plot and the isosurface diagram it is clear that there is significant change in electron density when CO is adsorbed onto an oxygen-adsorbed cluster. For CO- and O2-adsorbed clusters, the electron density is more in the vicinity of the CO or O2 molecules and decreases in vicinity of the Al atom. In Fig. 12, the red lines are electron rich regions and blue lines indicate electron depleted regions. The figure also indicates that electrons are being transferred from the cluster to these molecules, which results in activation of the molecule. In the case of CO adsorption onto an oxygen-adsorbed cluster, the scenario is quite different. The electron density is fully reduced in the vicinity of the oxygen molecule and increases near the CO molecule. Around the Al atom, the electron density is slightly reduced. This also indicates that electrons are transferred from the whole complex to the CO molecule. The low WBI value of CO also indicates this electron transfer and activation of the bond. The greater the transfer of electrons, the greater the BL and lower the bond order.
Fig. 12

Density deformation plot for CO, O2 and both adsorbed complexes. Top Contour plot of electron density deformation; red solid lines electron-rich zones, blue dotted lines electron depletion zone. Bottom Corresponding isosurface diagram; green positive surface lobes, red negative surface lobes

Thus, we see that O2 adsorbed cluster can still facilitate CO adsorption.

Conclusion

The metal clusters chosen for our study, Al6 and Al13 (and their neutral and anion counterparts) are potential candidates for adsorption of molecules like CO and O2. The BE of these molecules is quite appreciable. The BL and bond order from Tables 1 and 3 suggests that an activation of these bonds takes place upon adsorption to the cluster. The stability of the adsorbed complex is governed by different factors. For CO adsorption it is mainly the jellium model but for oxygen adsorption it is governed by unpaired electrons of the system. Al6 shows the highest BE for CO, with the lowest value shown by magic cluster \( {\mathrm{Al}}_{13}^{-} \). For O2, the highest BE is shown by Al13, where O2 is dissociated into O atoms on the cluster surface. The lower spin indicates greater BE for the oxygen molecule. Furthermore, it is seen that the BE for CO adsorption increases slightly on oxygen-adsorbed clusters. The optimized structures suggest that CO always prefers the atop position close to the O atom. This preference for site might help in CO to CO2 oxidation, the reaction mechanism of which can be another topic of concern in the near future.

Supporting information

The Cartesian coordinates of structures of CO, oxygen and both adsorbed clusters, different mode of adsorption, BSSE calculation, Fukui and Natural Population Analysis (NPA) calculations.

Notes

Acknowledgments

B.S. would like to thank IITB for providing a Teaching Assistant fellowship and high performance computation facilities. T.S. would like to thank the University Grant Commission (UGC) for a Senior Research Fellowship (SRF). The authors acknowledge the Center of Excellence in Scientific Computing at CSIR-NCL. S.P. acknowledges the J.C. Bose Fellowship grant of SERB, India, towards partial fulfillment of this work.

Compliance with ethical standards

Conflict of interest

The authors declare no competing financial interest.

Supplementary material

894_2018_3869_MOESM1_ESM.docx (449 kb)
ESM 1 (DOCX 449 kb)

References

  1. 1.
    Bernhardt TM (2005) Int J Mass Spectrom 243:1–29CrossRefGoogle Scholar
  2. 2.
    Yin S, Bernstein ER (2012) Int J Mass Spectrom 321–322:49–65CrossRefGoogle Scholar
  3. 3.
    Bohme DK, Schwarz H (2005) Angew Chem Int Ed 44:2336–2354CrossRefGoogle Scholar
  4. 4.
    Schlangen M, Schwarz H (2012) Catal Lett 142:1265–1278CrossRefGoogle Scholar
  5. 5.
    Richard A, Hair JO, Khairallah GN (2004) J Clust Sci 15:3CrossRefGoogle Scholar
  6. 6.
    Castleman Jr A, Khanna S (2009) J Phys Chem C 113:2664–2675CrossRefGoogle Scholar
  7. 7.
    Luo Z, Castleman AW (2014) Acc Chem Res 47:2931–2940CrossRefGoogle Scholar
  8. 8.
    Alonso J (2005) Structure and properties of atomic nanoclusters. World Scientific, LondonGoogle Scholar
  9. 9.
    Eric CT, Stefan V (2015) NAT Nanotechnology 10:577–588CrossRefGoogle Scholar
  10. 10.
    Vogel W, Lundquist JG, Ross P, Stonehart P (1975) Electrochim Acta 20:79CrossRefGoogle Scholar
  11. 11.
    Dhar HP, Christner LG, Kush AK (1987) J Electrochem Soc 134:302CrossRefGoogle Scholar
  12. 12.
    Lemons RA (1990) J Power Sources 29:251CrossRefGoogle Scholar
  13. 13.
    Haruta M (1997) Catal Today 36:153CrossRefGoogle Scholar
  14. 14.
    Lang SM, Bernhardt TM (2012) Phys Chem Chem Phys 14:9255–9269CrossRefGoogle Scholar
  15. 15.
    Pascucci B, Otero GS, Belelli PG, Illas F, Branda MM (2014) J Mol Model 20:2448CrossRefGoogle Scholar
  16. 16.
    Falsig H, Hvolbæk B, Kristensen IS, Jiang T, Bligaard T, Christensen CH, Nørskov JK (2008) Angew Chem 120:4913–4917CrossRefGoogle Scholar
  17. 17.
    Manzoor D, Krishnamurty S, Pal S (2014) J Phys Chem C 118:7501–7507CrossRefGoogle Scholar
  18. 18.
    Van Rijn R, Balmes O, Felici R, Gustafson J, Wermeille D, Westerström R, Lundgren E, Frenken JWM (2010) J Phys Chem C 114:6875–6876CrossRefGoogle Scholar
  19. 19.
    Jarrold MF, Bower JE (1987) J Chem Phys 87:1610CrossRefGoogle Scholar
  20. 20.
    Jarrold MF, Bower JE (1987) J Chem Phys 87:5728CrossRefGoogle Scholar
  21. 21.
    Ruatta SA, Anderson SL (1988) J Chem Phys 89:273CrossRefGoogle Scholar
  22. 22.
    Leuchtner RE, Harms AC, Castleman Jr AW (1991) J Chem Phys 94:1093CrossRefGoogle Scholar
  23. 23.
    Batra IP, Kleinman L (1984) J Electron Spectrosc Relat Phenom 33:175–241CrossRefGoogle Scholar
  24. 24.
    Wang L, Kuklja MM (2010) J Phys Chem Solids 71:140–144CrossRefGoogle Scholar
  25. 25.
    Huang Y, Wang A, Wang X, Zhang T (2007) Int J Hydrog Energy 32:3880–3886CrossRefGoogle Scholar
  26. 26.
    Bergeron DE, Castleman AW, Morisato T, Khanna SN (2004) Science 304:84–87CrossRefGoogle Scholar
  27. 27.
    Bergeron D, Roach P, Castleman A, Jones N, Khanna S (2005) Science 307:231–235CrossRefGoogle Scholar
  28. 28.
    Reber AC, Khanna SN, Castleman AW (2007) J Am Chem Soc 129:10189–10194CrossRefGoogle Scholar
  29. 29.
    Bergeron D, Castleman A (2003) Chem Phys Lett 371:189–193CrossRefGoogle Scholar
  30. 30.
    Leuchtner R, Harms A, Castleman Jr A (1989) J Chem Phys 91:2753–2754CrossRefGoogle Scholar
  31. 31.
    Burgert R, Schnöckel H (2008) Chem Commun 18:2075–2089CrossRefGoogle Scholar
  32. 32.
    Burgert R, Stokes ST, Bowen KH, Schnöckel H (2006) J Am Chem Soc 128:7904–7908CrossRefGoogle Scholar
  33. 33.
    Leskiw B, Castleman Jr A, Ashman C, Khanna S (2001) J Chem Phys 114:1165–1169CrossRefGoogle Scholar
  34. 34.
    Burgert R, Schnöckel H, Grubisic A, Li X, Stokes ST, Bowen KH, Ganteför G, Kiran B, Jena P (2008) Science 319:438–442CrossRefGoogle Scholar
  35. 35.
    Neumaier M, Olzmann M, Kiran B, Bowen KH, Eichhorn B, Stokes ST, Buonaugurio A, Burgert R, Schnöckel H (2014) J Am Chem Soc 136:3607–3616CrossRefGoogle Scholar
  36. 36.
    Luo Z, Smith JC, Berkdemir C, Castleman A (2013) Chem Phys Lett 590:63–68CrossRefGoogle Scholar
  37. 37.
    Grubisic A, Li X, Gantefoer G, Bowen KH, Schnöckel H, Tenorio FJ, Martinez A (2009) J Chem Phys 131:184305CrossRefGoogle Scholar
  38. 38.
    Reber AC, Khanna SN, Roach PJ, Woodward WH, Castleman JA (2010) J Phys Chem A 114:6071–6081CrossRefGoogle Scholar
  39. 39.
    Bagus PS, Hermann CK, Bauschlicher Jr W (1984) J Chem Phys 80:4378CrossRefGoogle Scholar
  40. 40.
    Cox DM, Reichmann KC, Trevor DJ, Kaldor A (1988) J Chem Phys 88:111CrossRefGoogle Scholar
  41. 41.
    Pireaux JJ, Ghijsen J, JWm M, Verbist J, Caudano R (1979) Surf Sci 80:488–502CrossRefGoogle Scholar
  42. 42.
    Shiraki Y, Kobayashi KLI, Katayama Y (1978) Surf Sci 77:458–464CrossRefGoogle Scholar
  43. 43.
    Katayama Y, Kobayashi KLI, Shiraki Y (1979) Surf Sci 86:549–554CrossRefGoogle Scholar
  44. 44.
    Post D, Baerends EJ (1982) Surf Sci 116:177–187CrossRefGoogle Scholar
  45. 45.
    Cox M, Trevor DJ, Whetten RL, Kaldor A (1988) J Phys Chem 92:421CrossRefGoogle Scholar
  46. 46.
    Crowell JE, Yates Jr T (1986) Surf Sci 37:165Google Scholar
  47. 47.
    Hoffman A, Maniv T, Folman M (1987) Surf Sci 182:57CrossRefGoogle Scholar
  48. 48.
    Cox DM, Trevor DJ, Whetten RL, Kaldor A (1988) J Phys Chem 92:421–429CrossRefGoogle Scholar
  49. 49.
    Reber AC, Khanna SN, Roach PJ, Woodward WH, Castleman Jr AW (2007) J Am Chem Soc 129:16098–16101CrossRefGoogle Scholar
  50. 50.
    Johnson GE, Tyo EC, Castleman Jr AW (2008) J Phys Chem A 112:4732–4735CrossRefGoogle Scholar
  51. 51.
    Rondina GG, Da Silva JL (2013) J Chem Inf Model 53:2282–2298CrossRefGoogle Scholar
  52. 52.
    Candido L, Rabelo JT, Da Silva JL, Hai GQ (2012) Phys Rev B: Condens Matter Phys 85:245404Google Scholar
  53. 53.
    Drebov N, Ahlrichs R (2010) J Chem Phys 132:164703CrossRefGoogle Scholar
  54. 54.
    Frisch M, Trucks G, Schlegel H, Scuseria G, Robb M, Cheeseman J, Scalmani G, Barone V, Mennucci B, Petersson GJ, Fox DJ (2009) Gaussian 09, revision A.01. Gaussian, Inc, Wallingford, CTGoogle Scholar
  55. 55.
    Yang W, Mortier WJ (1986) J Am Chem Soc 108:5708–5711CrossRefGoogle Scholar
  56. 56.
    Xiao M, Lu T (2015) J Adv Phys Chem 4:111–124CrossRefGoogle Scholar
  57. 57.
    Lu T, Chen F (2012) J Comput Chem 33:580–592CrossRefGoogle Scholar
  58. 58.
    Glendening E, Badenhoop J, Reed A, Carpenter J, Weinhold F (1996) NBO, version 3.1. Theoretical Chemistry Institute University of Wisconsin, Madison, WIGoogle Scholar
  59. 59.
    Duarte HA, Salahub DR (1998) J Chem Phys 108:743CrossRefGoogle Scholar
  60. 60.
    Lacaze-Dufaure C, Blanc C, Mankowski G, Mijoule C (2007) Surf Sci 601:1544–1553CrossRefGoogle Scholar
  61. 61.
    Socaciu LD, Hagen J, Bernhardt TM, Woste L, Heiz U, Hakkinen H, Landman U (2003) J Am Chem Soc 125:10437CrossRefGoogle Scholar
  62. 62.
    Wedler G, Papp H, Schroll G (1974) Surf Sci 44:463CrossRefGoogle Scholar
  63. 63.
    Helms CR, Madix RJ (1975) Surf Sci 52:677CrossRefGoogle Scholar
  64. 64.
    Knight WD, Clemenger K, de Heer WA, Saunders WA, Chou M, Cohen ML (1984) Phys Rev Lett 52:2141–2143Google Scholar
  65. 65.
    Veldeman N, Lievens P, Andersson M (2005) J Phys Chem A 109:11793–11801CrossRefGoogle Scholar
  66. 66.
    Dar M, Krishnamurty S, Pal S (2016) Phys Chem Chem Phys 18:7068CrossRefGoogle Scholar
  67. 67.
    Kalita B, Deka RC (2009) Eur Phy J D 53:51–58CrossRefGoogle Scholar
  68. 68.
    Bagus PS, Nelin CJ, Bauschlicher Jr CW (1984) J Vac Sci Technol A 2:905CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Department of ChemistryIndian Institute of TechnologyMumbaiIndia
  2. 2.Physical Chemistry DivisionCSIR National Chemical LaboratoryPuneIndia
  3. 3.Department of Chemical ScienceIndian Institute of Science Education and ResearchNadiaIndia

Personalised recommendations