Article pubs.acs.org/JPCA
Photodissociation of Cobalt and Nickel Oxide Cluster Cations C. J. Dibble, S. T. Akin, S. Ard, C. P. Fowler, and M. A. Duncan* Department of Chemistry, University of Georgia, Athens, Georgia 30602-2556, United States S Supporting Information *
ABSTRACT: Cobalt and nickel oxide cluster cations, CoxOy+ and NixOy+, are produced by laser vaporization of metal rods in a pulsed nozzle cluster source and detected using time-of-flight mass spectrometry. The mass spectra show prominent stoichiometries of x = y for CoxOy+ along with x = y and x = y − 1 for NixOy+. The cluster cations are mass selected and multiphoton photodissociated using the third harmonic (355 nm) of a Nd:YAG laser. Although various channels are observed, photofragmentation exhibits two main forms of dissociation processes in each system. CoxOy+ dissociates preferentially through the loss of O2 and the formation of cobalt oxide clusters with a 1:1 stoichiometry. The Co4O4+ cluster seems to be particularly stable. NixOy+ fragments reveal a similar loss of O2, although they are found to favor metal-rich fragments with stoichiometries of NixOx−1. The Ni2O+ fragment is produced from many parent ions. The patterns in fragmentation here are not nearly as strong as those seen for early or mid-period transition-metal oxides studied previously.
■
INTRODUCTION The diverse properties of transition-metal oxides make them useful for applications in electronics,1−4 catalysis,3−7 and magnetic materials.1−4 These systems have been studied extensively, and it has been documented that the size of the material can strongly influence its properties.8−14 Cobalt and nickel form particularly interesting oxides, which are used in many industrial and experimental applications.10−24 The catalytic properties of cobalt oxide have attracted much interest as this material appears to be an active ingredient in water splitting.10,15−18 Similarities between this system and the Mn4O4 cluster at the heart of Photosystem II suggest that Co4O4 might be the active site for this oxidation.15 Cobalt oxide has also been shown to form magnetic nanoparticles that are useful in storage media and biomedical sensing.22,23 Nickel oxide has applications as an inexpensive material for capacitors and anodes in electrochemical cells11,24 and in chemical sensors.12 Like cobalt oxide, nickel oxide is used as a catalyst in water splitting,19,20 as well as other reactions.21 Small gasphase clusters of these metal oxides allow fundamental concepts of structure and bonding to be investigated for finite-sized systems. A number of previous experiments and theoretical studies have focused on the structure and bonding of these systems and their reactivity.8,9,25−45 Here, we perform photodissociation studies of mass-selected cations of cobalt and nickel oxide in order to investigate their growth and stability patterns. Many experiments have attempted to determine the stability of gas-phase clusters such as metal oxides.8,9 In the past, relatively intense mass peaks were often thought to signify clusters that were more stable than others present in a distribution. However, as we have discussed before,32 mass spectrometry alone cannot determine cluster structures nor their relative stabilities. This is because the intensities of mass © 2012 American Chemical Society
peaks are influenced not only by relative stability but also by growth conditions (e.g., plasma temperature) and reagent concentrations, as well as size-dependent ionization potentials, ionization cross sections, and fragmentation dynamics.32 A single cluster mass or stoichiometry may have more than one isomeric structure. We have shown that mass spectrometry followed by mass-selected photodissociation measurements is a more reliable way to identify cluster sizes that are more stable than others.32,46,47 Although many factors also affect cluster dissociation, more stable species tend to appear repeatedly as fragments from larger clusters and are themselves difficult to dissociate. We have used this method to probe several metal oxide,32 metal carbide,46 and metal silicon clusters.47 For several oxide systems, unexpected stable clusters were found with stoichiometries quite different from those of the most common bulk phases.32 Here, we apply this methodology to cobalt and nickel oxides and examine their growth and stability patterns.
■
EXPERIMENTAL SECTION Metal oxide clusters are produced by laser vaporization in a pulsed nozzle source and mass analyzed in a reflectron time-offlight spectrometer (RTOF-MS) described previously.48 The second (532 nm) or third harmonic (355 nm) of a Nd:YAG laser (Spectra Physics GCR-11) is used for ablation of a rotating and translating metal rod.49 Helium gas seeded with 0.5−5% oxygen is pulsed using a General Valve with a 180 μs pulse duration and 150 psi backing pressure. The gas is pulsed through a “standard” rod holder,49 which has a 0.5−1 in. long growth channel with a 5 mm bore to promote cluster growth. The oxide cluster cations grow in the laser plasma and then the Received: March 16, 2012 Revised: April 24, 2012 Published: May 18, 2012 5398
dx.doi.org/10.1021/jp302560p | J. Phys. Chem. A 2012, 116, 5398−5404
The Journal of Physical Chemistry A
Article
peaks are assigned to x = y or x = y − 1, where x and y are the number of nickel and oxygen atoms, respectively. Masses with slightly more or slightly less oxygen are also observed. The greater width of these mass peaks is attributed to the multiple isotopes of nickel. Like the cobalt oxide results, these mass spectra of nickel oxide are consistent with those presented in the literature.25,29 Variations in the mass spectra obtained by different laboratories are believed to arise from differences in the cluster growth conditions, such as rod holder design, vaporization laser power, or oxygen concentration. Remarkably, even though many variables may affect these spectra, our mass spectra in many ways resemble those measured previously. Unfortunately, as we have discussed previously,32 mass spectra such as these do not provide any definitive information about the cluster growth mechanism nor the stable clusters. It is impossible to distinguish between oxygen that is integrated into a metal cluster framework and that which is weakly adsorbed as surface O2 molecules. Therefore, additional measurements are necessary to probe this information. Previous experiments have also shown that it is extremely difficult to make measurements on neutral clusters26 because their detection in a mass spectrometer requires some form of energetic ionization that might confuse the result. To obtain size-specific information, ionized clusters must be mass selected before their study. Photoelectron spectroscopy has been employed for massselected anion species,34−36 and collisional dissociation30 or photodissociation measurements32 have proven useful for massselected cations. In some of the most insightful measurements, mass-selected cations have been studied with infrared spectroscopy,40 complemented by computational studies. These measurements can provide specific structural information but unfortunately have been applied to only a limited number of systems. To investigate the photodissociation behavior of cobalt and nickel oxide cations here, we mass select various clusters and excite them at high power with a pulsed laser at 355 nm. 532 nm laser excitation was also explored but yielded much lower fragment ion intensities. High fluences of up to 40 mJ/pulse/ cm2 are needed to observe significant dissociation, which is consistent with conditions used in our previous studies on other transition-metal oxides.32 This high laser power indicates that a multiphoton process is required to break the bonds in these clusters. The bond energies of small cobalt oxide and nickel oxide clusters were found previously to be 4−5 and 7− 10 eV/bond, respectively, by Armentrout and co-workers.30 These high values are consistent with the need for multiple photons of 355 nm light to cause dissociation. Because absorption spectra of these clusters have not been measured, the large fluences required could also be needed to overcome weak absorption at the wavelengths used. Selected examples of photofragmentation spectra for cobalt and nickel oxides are shown in Figures 3−5 and 6−8, respectively. Additional fragmentation spectra for other clusters are presented in the Supporting Information for this paper. A comprehensive list of the photofragments produced from all selected cluster ions is presented in Tables 1 and 2, with the prominent photofragments indicated in bold. Our photodissociation data are collected in a difference mode of operation, where spectra with the dissociation laser off are subtracted from those with it on. This results in peaks with negative intensity for parents and positive intensity for fragments. In some cases, the parent ion peaks are presented off-scale to expand the plotting range for the fragment peaks.
molecular beam is skimmed into a differentially pumped mass spectrometer chamber. Ions are sampled perpendicular to the molecular beam flight direction with the RTOF-MS using pulsed acceleration fields. Pulsed deflection plates in the flight tube allow mass selection of specific clusters. Selected ions are excited in the turning region of the reflectron using a second Nd:YAG laser (Spectra Physics DCR-3 at 355 nm). The parent ion and any fragment ions are mass analyzed in the second flight tube and detected with an electron multiplier tube and a digital oscilloscope (LeCroy 6051A). Here, 10−40 mJ/pulse of unfocused 355 nm radiation (Spectra Physics DCR-3) with a laser spot size of roughly 1 cm2 was used for photodissociation.
■
RESULTS AND DISCUSSION The mass spectrum of CoxOy+ clusters is shown in Figure 1. It contains clusters up to 12 metal atoms with varying numbers of
Figure 1. Time-of-flight mass spectrum of CoxOy+ clusters using a vaporization wavelength of 532 nm.
oxygen atoms. Prominent mass peaks correspond to stoichiometries of x = y, where x is the number of cobalt atoms and y is the number of oxygen atoms. Other mass peaks are assigned to CoxOx+1+, CoxOx+2+, CoxOx−1+, and CoxOx−2+. These mass distributions are comparable to those of cation clusters produced previously by Castleman and co-workers25 and Ge and co-workers.28 Neutral and anionic clusters of cobalt oxide have also been reported to form mass distributions similar to this.25−28 Mass spectra of NixOy+ clusters are presented in Figure 2. The length of the growth channel was decreased to promote smaller clusters (top, 0.5 in.) or increased to grow larger clusters (bottom, 1 in.). The most intense mass
Figure 2. Time-of-flight mass spectrum of NixOy+ clusters in the lower (top) and higher (bottom) mass range using a vaporization wavelength of 532 nm. 5399
dx.doi.org/10.1021/jp302560p | J. Phys. Chem. A 2012, 116, 5398−5404
The Journal of Physical Chemistry A
Article
Table 2. Stoichiometries of Nickel Oxide Photofragments (NixOy+) Detected at 355 nma parent cluster 2,6 2,7 3,3 3,4 3,6 4,4 4,5 4,6 5,5 5,6 5,7 6,6 6,7 7,7 7,8
Figure 3. Photodissociation mass spectra of CoxOx+ (x = 5,6,8) clusters at 355 nm. Parent ion peaks are presented off-scale to highlight fragmentation peaks.
Table 1. Stoichiometries of Cobalt Oxide Photofragments (CoxOy+) Detected at 355 nma
a
parent cluster
fragment ions
4,4 4,5 4,6 4,7 5,4 5,5 5,6 5,7 6,6 6,7 6,8 7,7 7,8 7,9 8,8 8,9 8,10 9,9 12,12
3,4; 3,3; 3,2 4,4; 4,3; 3,3 4,4; 3,4 4,5; 4,4; 4,3; 3,3 4,5; 4,4; 3,4; 3,3; 3,2 4,5; 4,4; 4,3; 3,3 5,4; 4,4 5,5 5,6; 5,5; 5,4; 4,4 6,6; 6,5; 5,7; 5,6; 5,5; 5,4; 4,4 6,7; 6,6; 6,5; 5,6; 5,5; 5,4 6,7; 6,6; 6,5 7,6; 6,8; 6,6; 6,5 7,8; 7,7; 7,6; 6,7; 6,6; 6,5 7,8; 7,6; 6,8; 6,5; 5,5; 4,4; 3,3 8,7; 7,9; 7,7 8,8; 7,9; 7,8 8,9 11,11; 11,10
8,9 9,9 9,10 10,10 10,11 a
fragment ions 2,2; 2,1 2,5; 2,3; 2,1; 2,0; 1,0 3,3; 3,2; 2,1; 1,0 3,3; 3,2; 3,1; 2,2; 2,1; 2,0; 1,0 3,3; 3,2; 2,1; 1,0 4,2; 3,3; 3,2; 3,1; 2,1; 1,0 4,3; 3,3; 3,2; 3,1; 2,1; 1,0 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1; 1,0 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,2; 2,1; 2,0; 1,0 5,5; 5,4; 4,4; 4,3; 4,2; 3,2; 3,1; 2,1; 1,0 5,5; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1 6,5; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,2; 2,1; 2,0; 1,0 6,5; 5,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1; 2,0 7,5; 6,5; 6,4; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1; 2,0 7,6; 7,5; 6,5; 6,4; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1; 2,0 8,8 7,7; 6,5; 5,4; 4,3 8,7; 7,6; 6,5; 5,4; 4,4; 4,3; 4,2; 3,2; 2,1 9,7; 8,6; 6,5; 6,4; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1; 2,1 9,8; 8,7; 8,6; 7,6; 7,5; 6,5; 6,4; 5,5; 5,4; 5,3; 4,4; 4,3; 4,2; 3,3; 3,2; 3,1 10,9; 8,6; 7,6; 6,5; 5,4; 5,3; 4,4; 4,3; 4,2; 3,2; 3,1; 2,1 10,9; 8,7; 7,6; 6,5; 5,4; 4,3; 2,1
The most prominent stoichiometries are indicated in bold.
Photodissociation mass spectra of Co5O5+, Co6O6+, and Co8O8+, hereafter denoted as 5,5, 6,6, and 8,8, are shown in Figure 3. The 5,5, 6,6, and 8,8 species each lose a cobalt atom to form the 4,5, 5,6, and 7,8 fragments, respectively. The 4,4 fragment ion is produced by all three of these parent cluster ions. Additionally, the 8,8 parent forms 5,5 and 3,3 fragments, while the 5,5 species dissociates to form the 3,3 species. A 1:1 ratio of cobalt to oxygen is therefore favored for many fragment ions produced from these parents. We have observed this same stoichiometry for iron oxide clusters in the past.32 The photodissociation spectra of clusters containing one more oxygen than cobalt, that is, CoxOx+1+ (x = 4−7), are shown in Figure 4. The 7,8, 6,7, 5,6, and 4,5 parents lose either two oxygen atoms or molecular O2 to form the 7,6, 6,5, 5,4, and
The most prominent stoichiometries are indicated in bold.
Because mass discrimination makes it impossible for us to focus on the parent and fragment ions simultaneously, we cannot report quantitative branching ratios. We therefore distinguish only between strong and weak intensities in a qualitative way. It is important to note that these photodissociation experiments, like all others, measure only the identities and stoichiometries of the fragment ions produced. They provide no direct information about the structure of the precursor ion because extensive reorganization of the structures may occur before fragmentation. They also provide no information about the structures of the abundant fragment ions produced. Instead, we hope to identify specific stoichiometries of clusters produced more often than others, if any, and then, these can be compared to the stoichiometries of the common bulk phases. If unusual stoichiometries are observed, then these can be investigated further with computational chemistry to suggest likely structures.
Figure 4. Photodissociation mass spectra of CoxOx+1+ (x = 4−7) clusters at 355 nm. Parent ion peak for Co4O5+ are presented off-scale to highlight fragmentation peaks. 5400
dx.doi.org/10.1021/jp302560p | J. Phys. Chem. A 2012, 116, 5398−5404
The Journal of Physical Chemistry A
Article
4,3 species, respectively. Of these two channels, molecular O2 loss, is more likely as it is the lower-energy process, and we do not see many examples of the elimination of a single oxygen atom. However, we do not detect the neutral fragments and cannot experimentally distinguish between the loss of two O atoms as opposed to molecular O2. Therefore, we designate inferred neutral fragments with brackets to indicate this uncertainty. It is conceivable that the loss of [O2] here comes from the elimination of surface-bound O2 molecules that are not integrated into the cluster framework. However, oxygen atoms could also be eliminated from the framework and recombine into O2 molecules before separation from the cluster. The present photodissociation experiment does not allow us to distinguish between these structural alternatives. Although the most prominent fragments from these parents are those resulting from [O2] loss, we also detect many fragments with a 1:1 stoichiometry. Again, the 4,4 species seems to be relatively stable as it is formed from each of the 6,7, 5,6, and 4,5 parents. Photodissociation mass spectra of CoxOx+2+ (x = 5−8) clusters are shown in Figure 5. The most intense fragment ions
Figure 6. Photodissociation mass spectra of NixOx+ (x = 4−7) clusters at 355 nm. Parent ion peaks are presented off-scale to highlight fragmentation peaks.
ions that we assign to NixOx+ and NixOx+1+, but these always appear with lower intensity than NinOn−1 fragments. Photodissociation mass spectra of the oxygen-rich nickel oxide clusters 5,7, 5,6, 4,6, and 4,5 are shown in Figure 7. The
Figure 7. Photodissociation mass spectra of Ni4Oy+ (y = 5−6) and Ni5Oy+ (y = 6−7) clusters at 355 nm. Figure 5. Photodissociation mass spectra of CoxOx+2+ (x = 5−8) clusters at 355 nm. Parent ion peaks are presented off-scale to highlight fragmentation peaks.
fragments here are similar to those in Figure 6, with metal-rich clusters preferred and the 2,1 fragment particularly intense. For all of these oxygen-rich clusters, the loss of [O2] seems favorable. The 5,7 and 5,6 parents produce similar fragments and so do 4,6 and 4,5, suggesting that the NixOx+1+ and NixOx+2+ clusters have similar structures and that the extra oxygen does not greatly affect fragmentation. Figure 8 shows photodissociation mass spectra of several larger NixOx+1+ clusters (x = 6−10). The observation of intense metal-rich fragment ions is similar to the data shown in Figures 6 and 7. As before, [O2] loss remains the primary dissociation channel. Again 2,1 is a relatively intense photofragment, particularly for x = 6−7. Although we cannot directly observe the mechanism by which these clusters fragment, we can gain some insight by monitoring the intensities of fragment ions as a function of the dissociation laser power. For the cobalt oxide 5,5 cluster, which produces 4,4 and 3,3 charged fragments, the dissociation can be viewed as either a concerted process, producing the fragments
from these clusters again come from [O2] loss, but in these cases, this process also results in the formation of clusters with the preferred 1:1 cobalt to oxygen ratio. Fragment ions are also observed here with stoichiometries of CoxOx−1, CoxOx, CoxOx+1, and CoxOx+2, although these clusters have much lower intensities. Figure 6 shows photodissociation mass spectra of nickel oxide clusters with equal numbers of nickel and oxygen atoms (Ni x O x + , x = 4−7). The prominent fragments have stoichiometries of NixOx−1, which are metal-rich. We observe [O2] loss from the 4,4 and the 7,7 clusters, although this channel is not observed for 5,5 and 6,6. The 2,1 fragment is particularly intense in these spectra, and its repeated occurrence suggests that it is a stable cation unit. We also observe fragment 5401
dx.doi.org/10.1021/jp302560p | J. Phys. Chem. A 2012, 116, 5398−5404
The Journal of Physical Chemistry A
Article
probe the distribution of neutral clusters and found that CoxOy species prefer an x = y stoichiometry.26 They also showed that clusters with equal numbers of cobalt and oxygen have smaller rates for reactions with CO, NO, C2H2, and C2H4 than those clusters where x < y. The 1:1 cobalt oxide clusters therefore appear to be both chemically inert and thermodynamically stable, independent of charge state. The NixOy cluster cations favor a metal-rich, NixOx−1, stoichiometry, and the 2,1 fragment ion seems particularly stable. Nix O y cluster cations, like cobalt oxides, also preferentially lose [O2]. Nickel oxide cluster formation has been studied by several groups using laser vaporization coupled to flow tube reactors,25,31 Fourier Transform ion cyclotron resonance mass spectrometers (FT-ICR),29 or guided ion beams.30 Castleman and co-workers studied the reactions of nickel oxide cluster cations and anions with nitrogen oxides.25 Their mass spectra of cations and anions have distributions similar to ours shown in Figure 2. Reactions of nickel oxide cation clusters with NO produced metal-rich clusters, and Ni2O1(NO)3+ was a particularly intense product, which can be compared to the 2,1 fragment ion that we observe. Riley and co-workers photoionized neutral nickel oxide clusters and found that they formed with oxygen to metal ratios of around 0.75, indicating a preference for metal-rich clusters.31 Similar results were obtained by Koga and co-workers, who found that oxidation reactions produced oxygen to metal ratios of approximately 0.7.41 Our photodissociation experiments and these previous studies are in agreement that metal-rich clusters of nickel oxide form preferentially. It is interesting to compare the stoichiometries seen here for small clusters to those known for the solids of these oxides.1−3 The common bulk phases of cobalt are cobalt(II) oxide (CoO), cobalt(II,III) oxide (Co3O4), and cobalt(III) oxide (Co2O3). Although Co3O4 is the most common stoichiometry found in bulk materials, we find that the 1:1 ratio of CoO predominates for small gas-phase cations. The only well-characterized bulk phase of nickel oxide is NiO, although Ni2O3 and NiO2 have been observed. Our results indicate that small clusters prefer to form metal-rich NixOx−1+, which is not similar to any bulk nickel oxide phase. These new results are similar to our previous ones on other metal oxides, where we observed several examples of cluster stoichiometries different from those of the most common bulk phases.32 In those studies, iron oxide cation clusters were found to form (FeO)x, even though Fe2O3 is favored in the bulk. Furthermore, vanadium, tantalum, and niobium oxides did not form clusters corresponding to any bulk phase but instead formed a variety of different stoichiometries in the gas phase. Theoretical studies on cobalt and nickel oxides represent a serious challenge. Compounds featuring these elements are among the least studied of the first row (3d) transition metals using computational methods.41,42 The difficulty here arises from the occurrence of many low-lying excited states, often differing only in the electron spin configurations.25,26,28,41−45 Despite these difficulties, several groups have investigated cobalt and nickel oxides using density functional theory (DFT), although most have examined species with two or less metal atoms. CoOy species, where y = 1−4, have been studied by Uzunova et al.42 and Castleman et al.25 The ground state of CoOy (y = 1−4) was found to be a high-spin sextet, while the ground state of its anion analogue was found to be a quintet. Larger complexes have been investigated by Ge et al.28 who found two bonding schemes for 1:1 CoO clusters. Towers
Figure 8. Photodissociation mass spectra of NixOx+1+ (x = 6−10) clusters at 355 nm. Parent ion peaks for Ni7O8+, Ni8O9+, Ni9O10+, and Ni10O11+ are presented off-scale to highlight fragmentation peaks.
in parallel, as in (a) and (b), or as a sequential process, as in (c): (a)
5,5+ → 4,4+ + 1,1
(b)
5,5+ → 3,3+ + 2,2
(c)
5,5+ → 4,4+ + 1,1; 4,4+ → 3,3+ + 1,1
If (a) and (b) occur simultaneously to produce the photodissociation mass spectra that we observe, then the relative intensities of the two fragment ions should remain constant under varying laser power. However, if a sequential process is occurring, then the smaller fragment ions, which require multiple steps for their formation, could be influenced by the dissociation laser power. If the relative intensities of fragment ions change with respect to laser power, a sequential process is indicated. However, power dependence studies may be misleading because of saturation and bottlenecks in the kinetics. These types of dissociation processes have been described in detail in our previous work on transition-metal oxides.32 For all clusters, the dissociation laser power was varied between 10 and 40 mJ/pulse at 355 nm. These power dependence studies resulted in no change in the relative intensity ratios of the fragmentation peaks. Therefore, our results are consistent with a concerted process, but we cannot rule out a sequential fragmentation scheme with absolute certainty. Our data suggest that small cations of CoxOy favor a 1:1 stoichiometry and that the 4,4 cluster is produced repeatedly from many precursors, suggesting that is particularly stable. We also find that CoxOy cluster cations where x < y preferentially lose [O2]. These results can be compared to reactivity studies, dissociation pathways, and neutral cluster distributions explored in previous experiments.25−28,30,36 Castleman and co-workers performed collision-induced dissociation studies of both anionic and cationic clusters of cobalt oxide.25 They determined that clusters with either charge fragment by losing [O2]. Clusters larger than three cobalt atoms were not studied in this previous work, but smaller ions produced fragments with 1:1 ratios of cobalt to oxygen. Both of these results are completely consistent with our photodissociation measurements. Bernstein and co-workers used near-threshold VUV photoionization to 5402
dx.doi.org/10.1021/jp302560p | J. Phys. Chem. A 2012, 116, 5398−5404
The Journal of Physical Chemistry A
■
composed of small linked rings and cubic structures were determined to be close in relative energy (
