Wen-Ping
Peng
a,
Grant E.
Johnson
b,
Ivy C.
Fortmeyer‡
b,
Peng
Wang
b,
Omar
Hadjar
b,
R. Graham
Cooks
*c and
Julia
Laskin
*b
aDepartment of Physics, National Dong Hwa University, Hualien, 974, Taiwan, Republic of China
bChemical and Materials Sciences Division, Pacific Northwest National Laboratory, Richland, Washington 99352, USA. E-mail: Julia.Laskin@pnl.gov
cDepartment of Chemistry and Center for Analytical Instrumentation Development, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907, USA. E-mail: cooks@purdue.edu
First published on 12th November 2010
Soft landing (SL) of mass-selected ions is used to transfer catalytically-active metal complexes complete with organic ligands from the gas phase onto an inert surface. This is part of an effort to prepare materials with defined active sites and thus achieve molecular design of surfaces in a highly controlled way. Solution-phase electrochemical studies have shown that VIVO(salen) reacts in the presence of acid to form VVO(salen)+ and the deoxygenated VIII(salen)+ complex—a key intermediate in the four electron reduction of O2 by vanadium–salen. In this work, the VVO(salen)+ and [NiII(salen) + H]+ complexes were generated by electrospray ionization and mass-selected before being deposited onto an inert fluorinated self-assembled monolayer (FSAM) surface on gold. A time dependence study after ion deposition showed loss of O from VVO(salen)+ forming VIII(salen)+ over a four-day period, indicating a slow interfacial reduction process. Similar results were obtained when other protonated molecules were co-deposited with VVO(salen)+ on the FSAM surface. In all these experiments oxidation of the VIII(salen)+ product occurred upon exposure to oxygen or to air. The cyclic regeneration of VVO(salen)+ upon exposure to molecular oxygen and its subsequent reduction to VIII(salen)+ in vacuum completes the catalytic cycle of O2reduction by the immobilized vanadium–salen species. Moreover, our results represent the first evidence of formation of reactive organometallic complexes on substrates in the absence of solvent. Remarkably, deoxygenation of the oxo-vanadium complex, previously observed only in highly acidic non-aqueous solvents, occurs on the surface in the UHV environment using an acid which is deposited into the inert monolayer. This acid can be a protonated metal complex, e.g. [NiII(salen) + H]+, or an organic acid such as protonated diaminododecane.
Preparation of organometallic films typically involves chemical vapor deposition or self-assembly from solution.7 Here we investigate the use of soft landing (SL) of mass-selected ions, i.e. ions of well-controlled mass, and hence chemical composition to create organometallic films. Ion SL instrumentation and various applications have been reviewed.8,9 The SL methodology is used to prepare high purity materials with well-defined structures and active sites and thus to achieve molecular design of surfaces in a highly controlled way. The ion deposition approach has advantages in relation to solution-phase techniques, avoiding clumping of materials seen as solvent evaporates and allowing deposition of uniform films while preserving the structure of deposited molecules.10–15
Our experiments present a first step towards the development of catalytically active organometallic materials using SL and also extend the work of others16–21 on ion deposition as an approach to catalyst preparation. Metal–salen complexes [salen = N,N′-ethylenebis(salicylideneaminato) ligand] were chosen as model systems in this study because they are effective homogeneous catalysts for the epoxidation of a variety of non-functionalized olefins with high levels of enantioselectivity,22 for ring-opening polymerization of six-membered cyclic carbonates,23 for the oxygenation of heteroatom containing organic compounds24 and for the hydroxylation of organic substrates.25 In a previous study, we showed that metal–salen complexes display redox activity in the gas phase.26 The protonated mixed complex [NiIIVIVO(salen) + H]+, generated by atmospheric pressure thermal desorption ionization, was found to undergo formal oxidation (hydrogen atom loss) to produce Ni(salen) monomer as the molecular radical cation upon collision-induced dissociation (CID). In this paper, we study redox chemistry in thin layers of vanadium–salen complexes initiated by the presence of a proton donor deposited from the gas-phase.
First studies of the reactivity of vanadium–salen complexes in strongly acidic solution were focused on understanding of vanadium bioaccumulation in blood cells of ascidians.27 Bonadies et al. demonstrated that in an acidic solution VIVO(salen) undergoes a disproportionation reaction shown in eqn (1):28
2VIVO(salen) + 2H+ → VIII(salen)+ + VVO(salen)+ + H2O | (1) |
2VIII(salen)+ + O2 → 2VVO(salen)+ + 4e− | (2) |
In this study we examine the immobilization and reactivity of VVO(salen)+ deposited on an inert fluorinated self-assembled monolayer surface (FSAM) on gold using SL of mass-selected ions. [We use the term immobilization to indicate that the landed species are bound to the surface strongly enough that desorption from the substrate is very slow.] Ions were generated by electrospray ionization (ESI) and guided to a mass analyzer where species with m/z values corresponding to the desired chemical entities were selected and gently deposited onto an FSAM surface. SAM surfaces are convenient substrates, the chemical properties of which can be easily tailored.33 It has been demonstrated that FSAM surfaces prevent neutralization of immobilized ions34 including metal–salen complexes that undergo very slow charge reduction through electron transfer to the surface.35 SL also allows convenient co-deposition of protonated ions which serve as proton donors on the surface enabling investigation of acid-mediated reactivity. Secondary ion mass spectrometry (SIMS) was chosen for the analysis of the SL prepared surfaces both before and after exposure to gaseous reagents. We demonstrate facile proton-initiated reduction of VVO(salen)+ to VIII(salen)+ in the absence of solvent and regeneration of VVO(salen)+ by exposure to O2 (Scheme 1). The role of a proton donor is examined by depositing protonated [NiII(salen) + H]+ complexes or protonated diaminododecane molecules into the SAM containing the VVO(salen)+ species.
Scheme 1 |
In situ analysis of surfaces following SL was performed by combining 8 keV Cs+ secondary ion mass spectrometry with FT-ICR detection of the sputtered ions (FT-ICR-SIMS) as described elsewhere.37–39 Static SIMS conditions with a total ion flux of about 2 × 1010 ions per cm2 (current 4 nA, duration 80 μs, spot diameter 4.6 mm, 15 shots per spectrum, 100–200 data points) were used in these experiments. Data acquisition was accomplished using a MIDAS data station.40
For ex situTOF-SIMS experiments mass-selected ion currents of 300 pA and 200 pA were obtained for VVO(salen)+ and [NiII(salen) + H]+, respectively. During the deposition (deposition times of 20 min and 30 min for VVO(salen)+ and [NiII(salen) + H]+, respectively) 2 × 1012 ions of each species were delivered onto a 5 mm diameter spot on the target. In situTOF-SIMS experiments utilized a different apparatus, in which 1012 ions of VVO(salen)+ and [NiII(salen) + H]+ were deposited onto a 3 mm spot. Ion currents of 50 pA and 20 pA were obtained and deposition times of 60 min and 120 min were used for VVO(salen)+ and [NiII(salen) + H]+, respectively.
In these experiments sequential deposition of metal–salen complexes on the FSAM surface was followed by characterization using 15 keV Ga+ SIMS and subsequent reactivity studies. Ex situTOF-SIMS analysis was performed after a brief exposure of surfaces to laboratory air because it took approximately 10 min to remove the sample from the soft-landing apparatus and introduce it into the TOF-SIMS.41In situTOF-SIMS experiments were performed ca. 5 min after soft-landing—the time required to transfer the surface from the ion deposition apparatus into the TOF-SIMS and position it on the stage. ESI conditions for ion soft-landing experiments were as follows: spray voltage, 3 kV; flow rate, 30 μL h−1; capillary temperature, 160°.
Fig. 1 Comparison of ex situTOF-SIMS spectra obtained after the deposition of 2.2 × 1012 ions of VO(salen)+ (blue) and 1.6 × 1012 ions of [Ni(salen) + H]+ (red) on the FSAM surface in two separate experiments; black trace corresponds to the FSAM background. |
Sequential deposition of VVO(salen)+ and [NiII(salen) + H]+ onto the same FSAM surface yields a TOF-SIMS spectrum shown as the red trace in Fig. 2 that closely resembles the superposition of individual VVO(salen)+ and [NiII(salen) + H]+TOF-SIMS spectra shown in Fig. 1. However, the TOF-SIMS spectrum of the same surface re-examined after 4 days in UHV (blue trace in Fig. 2) shows reactivity of the deposited molecules, while the spectra of the individually deposited VVO(salen)+ and [NiII(salen) + H]+ do not change a great deal with time (data not shown). This reactivity is manifested by substantial increases in the abundances of VIII(salen)+ (m/z 317), [VIVOVIII(salen)2]+ (m/z 650), [NiIIVIII(salen)2]+ (m/z 641), [Ni2V(salen)3 + 3H]+ (m/z 968), and [Ni2VO(salen)3 + 2H]+ (m/z 983) ions in the TOF-SIMS spectrum recorded four days after ion deposition. In this paper we will limit the discussion to the monomeric species observed in SIMS spectra. Deposition and analysis of dimer ions will be discussed in a separate publication.
Fig. 2 Ex situ TOF-SIMS characterization of the FSAM surface before the deposition (black), 10 min after the sequential deposition of 2.5 × 1012 ions of VO(salen)+ and 2.5 × 1012 ions of [Ni(salen) + H]+ on the FSAM surface (red), and 4 days later (blue). |
The observed increase in the abundance of VIII(salen)+ with time and the concomitant decrease in the abundance of VVO(salen)+ could be attributed to a slow interfacial reduction of the VVO(salen)+ on the FSAM surface in UHV. It is remarkable that the large increase in the VIII(salen)+ peak is seen only when [NiII(salen) + H]+ is present on the surface, but such an effect is not observed in its absence.
Fig. 3 (a) FT-ICR spectrum of the primary beam of a mixture of VVO(salen)+ and [NiII(salen) + H]+; FT-ICR-SIMS spectra of the FSAM surface after (b) 170 min and (c) 320 min of soft-landing (at the end of the deposition), (d) 87 h after the end of soft-landing, and (e) after exposure of the sample to laboratory air for 1 h. |
The intensity ratio of VIII(salen)+ and VVO(salen)+, I(317)/I(333), increases from the value of 0.85 ± 0.15 at the beginning of the deposition to 2.0 ± 0.2 at the end of the deposition and continues to increase after the deposition is finished reaching the value of 3.8 ± 0.2 after 100 h (data not shown). A dramatic decrease in the I(317)/I(333) ratio is observed after exposure of the surface to laboratory air (Fig. 3e); the I(317)/I(333) ratio drops to a value of ca. 0.2.
VIII(salen)+ ions observed in the SIMS spectrum may be produced through several processes. First, a small fraction of VIII(salen)+ ions is produced as a result of dissociation of VVO(salen)+ in the SIMS analysis. The percentage of VIII(salen)+ fragments is expected to remain constant at low surface coverage used in our experiments. The upper limit for the amount of fragmentation in the FT-ICR-SIMS plume of 20% was estimated from the I(317)/I(333) ratio obtained after exposure of the FSAM surface to air. This value is higher than 3.5% fragmentation efficiency observed in the TOF-SIMS experiment (Fig. 1). Lower fragmentation efficiency in TOF-SIMS spectra was previously attributed to the shorter residence time (few microseconds) for ions in the TOF-SIMS as compared to hundreds of milliseconds residence time in FT-ICR SIMS experiments.43
The second process involves dissociation (crash landing)39,44,45 of VVO(salen)+ during deposition followed by trapping of the VIII(salen)+ fragment in the monolayer. Similar to dissociation in SIMS, the fraction of VIII(salen)+ ions resulting from crash landing should remain constant during ion deposition and cannot contribute to the observed increase in the yield of VIII(salen)+ after the deposition is complete. Finally, VIII(salen)+ can be produced through reactions between ions and neutral molecules trapped on the surface. Surface reactivity is the only process that can rationalize the increase in the I(317)/I(333) ratio after the end of ion deposition. From the above discussion it follows that a significant fraction of the VIII(salen)+ (m/z 317) is produced as a result of the loss of O from the VVO(salen)+ (m/z 333) deposited on the surface and the I(317)/I(333) ratio is a measure of the extent of the reduction of VVO(salen)+ to VIII(salen)+. A significant drop in the I(317)/I(333) ratio upon exposure of the surface to laboratory air (Fig. 3e) indicates a facile oxidation of VIII(salen)+ to VVO(salen)+ upon exposure to O2.
The abundance ratio of NiIII(salen)+ (m/z 324) and [NiII(salen) + H]+ (m/z 325) also increases with time. It is reasonable to assume that the NiIII(salen)+ ion is produced in SIMS through ionization of the neutral NiII(salen). Our previous studies demonstrated that FSAM surfaces are characterized by the highest efficiency for charge retention when protonated molecules are deposited on the substrate.43 We found that both instantaneous proton loss during ion deposition and a slow charge reduction process that continues for many hours after SL are responsible for neutralization of the deposited protonated species.38 It is reasonable to assume that [NiII(salen) + H]+ undergoes a similar slow charge reduction on the FSAM surface. The resulting neutral NiII(salen) complex is subsequently re-ionized in SIMS and observed as NiIII(salen)+ species.
Fig. 4a shows the time dependence of the I(317)/I(333) ratio. In the first 165 min of the kinetics experiment the ratio remains almost constant while an almost linear increase is observed after the deposition of ca. 7% of a monolayer of metal–salen species on the surface (4.5 × 1011 ions). This trend continues after the deposition is finished. Our results can be rationalized assuming that VVO(salen)+ deposited onto the FSAM surface is slowly converted into VIII(salen)+. The observed threshold behavior shown in Fig. 4a indicates that slow diffusion-limited interfacial reduction occurs on the inert FSAM surface and confirms that the formation of the reaction product, VIII(salen)+, is not associated with the SIMS analysis step.
Fig. 4 Time dependent intensities of (a) VIII(salen)+ (m/z 317), (b) [NiII(salen) + H]+ (m/z 325), and (c) NiIII(salen)+ (m/z 324) in FT-ICR SIMS spectra normalized to the intensity of the VVO(salen)+ (m/z 333). The dashed line shows the end of ion deposition. |
Table 1 lists the I(317)/I(333) ratios observed in TOF-SIMS spectra obtained in situ immediately after sequential deposition of 1012 ions of VVO(salen)+ and 1012 ions of a proton donor. The lowest initial yield of VIII (salen)+ was obtained when [NiII(salen) + H]+ was used as a proton donor. The I(317)/I(333) ratio of 0.5–0.7 was obtained in a number of repeated experiments. This ratio increases to 1.0 when VVO(salen)+ is co-deposited with [VIVO(salen) + H]+ which may also serve as a proton donor. Interestingly, the I(317)/I(333) ratio of 2.5–3.2 was obtained when soft-landing of VVO(salen)+/[VIVO(salen) + H]+ was followed by a fairly long deposition (2–3 h) of [NiII(salen) + H]+. However, a much lower ratio of 1.2 was obtained when the deposition time of [NiII(salen) + H]+ was reduced to 1.5 hours consistent with the presence of an initial induction period for the reduction reaction.
Deposition 1 (ion) | Time/h | Deposition 2 (ion) | Time/h | I(317)/I(333) |
---|---|---|---|---|
a In these experiments both VVO(salen)+ and [VIVO(salen)+H] were produced in the ESI source and were not separated prior to deposition. | ||||
VVO(salen)+ | 0.75 | [NiII (salen) + H]+ | 2 | 0.5 |
VVO(salen)+ | 0.75 | [NiII (salen) + H]+ | 3 | 0.7 |
VVO(salen)+ | 0.75 | [Diaminododecane + H]+ | 7 | 3.1 |
[Diaminododecane + H]+ | 12 | VVO(salen)+ | 0.75 | 1.6 |
VVO(salen)+ | 1 | [RGDGG + H]+ | 2 | 3.7 |
[RGDGG + H]+ | 3 | VVO(salen)+ | 0.75 | 1.3 |
VVO(salen)+/[VIVO(salen) + H]+a | 0.75 | — | — | 1.0 |
VVO(salen)+/[VIVO(salen) + H]+a | 1 | [NiII (salen) + H]+ | 2 | 3.2 |
VVO(salen)+/[VIVO(salen) + H]+a | 2 | [NiII (salen) + H]+ | 3 | 2.5 |
VVO(salen)+/[VIVO(salen) + H]+a | 3 | [NiII (salen) + H]+ | 3 | 2.9 |
VVO(salen)+/[VIVO(salen) + H]+a | 3 | [NiII (salen) + H]+ | 1.5 | 1.2 |
Deposition of VVO(salen)+ with protonated diaminododecane resulted in a fairly high initial ratio of 3.1. However, a much lower ratio of 1.6 was obtained when the order of the deposition was reversed and diaminododecane was deposited prior to VVO(salen)+. Similar results were obtained when a small protonated pentapeptide, RGDGG, was used as a proton donor. We attribute the observed dependence on the order of deposition to the shorter residence time of VVO(salen)+ on the surface when it is deposited after the proton donor.
Fig. 5 In situ TOF-SIMS characterization of the FSAM surface (a) after the deposition of 1 × 1012 ions of VO(salen)+, (b) after 4 h in vacuum, (c) after exposure to 10−4 Torr of O2 for 20 min, (d) after 65 h in vacuum. |
Fig. 6a shows the I(317)/I(333) ratio derived from TOF-SIMS spectra obtained following deposition of VVO(salen)+ and [VIVO(salen) + H]+. As discussed earlier, the ratio changes only upon initial exposure to oxygen and remains constant when the system is kept in vacuum. In contrast, both the reduction of VVO(salen)+ to VIII(salen)+ and the oxidation of VIII(salen)+ to VVO(salen)+ by exposure to O2 were observed when VVO(salen)+ was co-deposited with [NiII(salen) + H]+. The cyclic regeneration of VVO(salen)+ upon repeated exposures to O2 is evident in the oscillating abundance ratios shown in Fig. 6b. Similar oxidation/reduction cycles were observed when protonated diaminododecane was used as a proton donor. The I(317)/I(333) ratio obtained in this experiment is shown in Fig. 6c. We note that the initial yield of VIII(salen)+ obtained after SL was quite high. Exposure to oxygen resulted in a significant decrease in the I(317)/I(333) ratio. Subsequent increase in the I(317)/I(333) ratio indicates that the reduction of VVO(salen)+ to VIII(salen)+ on the surface continues even after exposure of the surface to oxygen. This behavior is reproducible and continues for at least three cycles when [NiII(salen) + H]+ is used as a proton donor. However, the reactivity is clearly suppressed upon multiple exposures of the surface to oxygen. This is likely due to depletion of the population of [NiII(salen) + H]+ on the surface which results in a smaller number of proton donors, and consequently, a lower overall acidity in the monolayer.
Fig. 6 The intensity ratio of peaks at m/z 317 and m/z 333 observed in situTOF-SIMS spectra showing oxidation and reduction cycles on the FSAM surface for (a) VO(salen)+, (b) sequential deposition of VO(salen)+ and [Ni(salen) + H]+, (c) sequential deposition of the singly protonated diaminododecane and VO(salen)+. Individual points were obtained immediately after SL (labeled as End of SL), after one or multiple O2 exposures (labeled as O2), and at various times after each processing step, SL or O2 deposition (labels indicate the time between the corresponding processing step and the analysis). Lines are shown to guide the eye. |
Liu and Anson re-examined the mechanism of the acid-mediated reactivity of VO(salen) in solution.47 They provided further support for the disproportionation reaction and showed that VIII(salen)+ in acetonitrile reacts with O2via the reaction shown in eqn (2). The reaction is fast and the conversion of VIII(salen)+ into VVO(salen)+ is quantitative. They also showed that the disproportionation reaction is a multistep process as shown in eqn (3) and (4).
VIVO(salen) + 2H+ → VIV(salen)2+ + H2O | (3) |
VIVO(salen) + VIV(salen)2+ → VIII(salen)+ + VVO(salen)+ | (4) |
From the above discussion it follows that the slow interfacial reactivity observed in this study could be attributed to reactions between the ionic and neutral species trapped on or in the FSAM. The reaction rate is determined either by the rate of migration of vanadium–salen complexes on the surface or by the rate of the proton loss by the protonated nickel–salen complex. Previous studies demonstrated that both small ions and organometallic clusters can be trapped between adjacent SAM chains.21,34 Penetration of the soft-landed species into the layer may significantly reduce the rate of migration and slow down the reaction. Alternatively, the reaction rate may be determined by the availability of protons in the layer. Previous work showed that the proton loss from protonated molecules soft-landed onto FSAM surfaces is rather slow.38 Similarly, the decay of the [Ni(salen) + H]+ signal observed in FT-ICR SIMS experiments occurs on a timescale of several hours.
The reverse reaction in which VIII(salen)+ is oxidized upon exposure to molecular oxygen regenerating the original VVO(salen)+ complex is fairly efficient and continues for several cycles. This cyclic regeneration of VVO(salen)+ was evident in the oscillating abundance ratios of VIII(salen)+ and VVO(salen)+. The yield of the reduction reaction is limited by the amount of protonated molecules co-deposited with VVO(salen)+ and the rate of the proton loss by the proton donor. Our results indicate that in the presence of proton donors soft-landed VVO(salen)+ initiates reduction of molecular oxygen that is similar to the catalytic electrochemical reduction of O2 observed in strongly acidic solutions. Although this study focused on a specific reduction reaction we believe that the SL methodology can be utilized for preparation of model catalysts of controlled chemical composition and hence, in conjunction with SIMS analysis and other surface characterization techniques, it should enable detailed fundamental studies of the surface reactivity of metal complexes on solid supports and perhaps allow screening and increased understanding of a wider variety of catalysts.48
Footnotes |
† Electronic supplementary information (ESI) available: Electrospray ionization mass spectra of solutions used in soft-landing experiments. See DOI: 10.1039/c0cp01457e |
‡ Undergraduate student from Columbia University |
This journal is © the Owner Societies 2011 |