Electronic Structure of a Spin Crossover Molecular Adsorbate

Oct 10, 2012 - Patrick Rosa,*. ,‡,§. Jean-François ... Street, University of Nebraska Lincoln, Lincoln, Nebraska 68588-0299, United States. ‡. C...
0 downloads 0 Views 537KB Size
Article pubs.acs.org/JPCC

Electronic Structure of a Spin Crossover Molecular Adsorbate Xin Zhang,† Tatiana Palamarciuc,‡,§ Patrick Rosa,*,‡,§ Jean-François Létard,*,‡,§ Bernard Doudin,∥,⊥ ZhengZheng Zhang,† Jian Wang,# and Peter A. Dowben*,† †

Nebraska Center for Materials and Nanoscience, Department of Physics and Astronomy, Theodore Jorgensen Hall, 855 North 16th Street, University of NebraskaLincoln, Lincoln, Nebraska 68588-0299, United States ‡ CNRS, ICMCB, UPR 9048, F-33600 Pessac, France § Univ. Bordeaux, ICMCB, UPR 9048, F-33600 Pessac, France ∥ CNRS, IPCMS, UMR 7504, 23 Rue du Loess, F-67034 Strasbourg, France ⊥ Univ. Strasbourg, 23 Rue du Loess, F-67034 Strasbourg, France # Canadian Light Source Inc., University of Saskatchewan, 101 Perimeter Road, Saskatoon, Saskatchewan, Canada S7N 0X4 S Supporting Information *

ABSTRACT: We have investigated the occupied and unoccupied electronic structure of ultrahigh vacuum (UHV) evaporated molecular thin films of the spin crossover [Fe(H2B(pz)2)2(bipy)] complex (with H2B(pz)2 = bis(hydrido)bis(1H-pyrazol-1-yl)borate and bipy = 2,2′-bipyridine) by ultraviolet photoelectron spectroscopy (UPS), inverse photoemission (IPES), and X-ray absorption spectroscopy (XAS). A bandgap of 2−3 eV is deduced from combined UPS and IPES measurements of the molecular films on Au substrates. The matching Fe XAS and IPES spectra indicate that the electronic unoccupied states have a significant Fe weight. The shift of the unoccupied density of states seen in inverse photoemission is consistent with the thermally induced spin crossover transition for [Fe(H2B(pz)2)2(bipy)] deposited on the organic ferroelectric copolymer poly(vinylidene fluoride) with trifluoroethylene (PVDF− TrFE).

1. INTRODUCTION In nature there is a vast class of molecules for which the magnetic structure can be altered at the atomic level by an external stimulus. These are spin crossover (SCO) molecules, for which low-spin (LS) to high-spin (HS) transitions can be induced by pressure, temperature, illumination with light, or magnetic pulses.1−4 Recently, it has also been proposed that SCO transition can be induced by an electric field applied to some of these molecular system.5−7 The spin crossover phenomenon is a consequence of the splitting in energy of the transition metal d orbitals into the t2g and eg sets in an octahedral ligand field. As a result, octahedral complexes of some of the transition metal ions from the first transition series with configurations d4 to d7, especially Fe(II),2,3,7−9 may exist in either of the HS or LS states, depending on the nature of the ligand field around the metal ion. The electrical conductivity of a spin crossover (SCO) molecular thin film may also be modified when the transition occurs, as in the case of the SCO [Fe(HB(pz)3)2] (pz = 1Hpyrazol-1-yl) complex sublimated on a surface.10 There is a growing interest in the electrical transport properties of SCO materials for molecular electronics studies,11 with photoconductivity effects also occurring in nanocrystals of a polymeric SCO material drop-casted over gold nanoelectrodes.12,13 One can expect that the electronic density of states differs in the HS and LS states and has important consequences for the electronic transport properties.14 Experimental insight © 2012 American Chemical Society

into the density of states (DOS) of these materials is thus of paramount importance. Thin films of SCO materials, made by sublimation under high and ultrahigh vacuum, make possible the use of surface spectroscopy techniques under high vacuum conditions to gain access to electronic spectroscopic properties, free of ambient contamination. The SCO complex [Fe(H2B(pz)2)2(bipy)], with bipy = 2,2′-bipyridine, is known to display a rather abrupt SCO centered at about 160 K (see Figure 1)15 and meet the rather stringent requisites for sublimation under ultrahigh vacuum conditions,16 as do the related [Fe(phen)2(NCS)2] and [FeH2B(pz)2phen] SCO molecular complexes with phen = 1,10-phenanthroline and H2B(pz)2 = bis(hydrido)bis(1Hpyrazol-1-yl)borate.7,8,17,18 The [Fe(H2B(pz)2)2(bipy)] spin crossover complex shows photoinduced conversion at low temperatures,19 but the experimental electronic structure of this organometallic compound has not been well characterized. Ultimately, the goal is to look at the interaction of this species in a local electric field, using the ferroelectric domain switching to change the molecular band offsets and possibly the molecular configuration.20,21 This substrate mediated voltage controlled spin crossover transition requires a dielectric substrate and a Received: April 11, 2012 Revised: August 26, 2012 Published: October 10, 2012 23291

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296

The Journal of Physical Chemistry C

Article

2.2. Film Characterization. The occupied and unoccupied molecular orbitals of the iron complex [Fe(H2B(pz)2)2(bipy)] thin films were characterized by combined photoemission and inverse photoemission spectroscopies, as described elsewhere.20,21,25,29−39 The IPES spectra were obtained by using variable kinetic energy incident electrons while detecting the emitted photons at a fixed energy (9.7 eV) using a Geiger− Müller detector.20,21,25,29−39 The IPES was limited by an instrumental line width of ∼400 meV, as described elsewhere.20,21,25,29−39 The angle integrated photoemission (UPS) studies were carried out using a helium lamp at hv = 21.2 eV (He I) and a Phi hemispherical electron analyzer with an angular acceptance of ±10° or more, as described in detail elsewhere.20,21,25,29−39 The photoemission experiments were made with the photoelectrons collected along the surface normal, while the inverse photoemission spectra were taken with the incident electrons normal to the surface. This restriction of the electron emission (photoemission) or electron incidence (inverse photoemission) to the surface normal was done to preserve the highest point group symmetry and eliminate any wave vector component parallel with the surface. In both photoemission and inverse photoemission measurements, the binding energies are referenced with respect to the Fermi edge of gold in intimate contact with the sample film and the photoemission (UPS, XPS) data are expressed, in terms of E − EF (thus making occupied state energies negative). The core level X-ray photoemission spectra were taken with a SPECS X-ray source with a Mg anode (hv = 1253.6 eV), but data collection times and the X-ray flux were specifically limited to avoid X-ray induced molecular fragmentation. It was valuable to investigate the unoccupied molecular orbital states by both inverse photoemission and X-ray absorption. The key difference is that inverse photoemission probes the delocalized unoccupied states according to matrix element effects while X-ray absorption tends to emphasize those unoccupied states with weight associated with the element whose core excitation is exploited, in this case the Fe (iron) L3 shell. The Fe L2,3 X-ray absorption spectra (XAS) of [Fe(H2B(pz)2)2bipy] were measured using the scanning transmission X-ray microscope (STXM) at the soft X-ray spectromicroscopy (SM) beamline of the Canadian Light Source (CLS, Saskatoon, Canada). The X-ray source is an APPLE II type elliptically polarizing undulator (EPU) which can provide circularly polarized light (130−1000 eV) and linearly polarized light (130−2500 eV) with polarization adjustable from −90° to +90°.40 A 500 l/mm plane grating monochromator (PGM) was used for the Fe L2,3-edge. In STXM, the monochromatic X-ray beam is focused by a Fresnel zone plate to a 30 nm spot on the sample, and the sample was scanned, with a raster in both the in plane x and y directions, with synchronized detection of transmitted X-rays to generate image sequences, i.e., image stacks,41 over a range of photon energies.

Figure 1. [Fe(H2B(pz)2)2(bipy)] complex schematic and the product of molecular magnetic susceptibility with temperature, versus temperature, from the powder bulk material.

switching process not influenced by the interface dipole,8 conditions best ensured by an organic ferroelectric.

2. EXPERIMENTAL DETAILS 2.1. Synthesis and Film Preparation. The [Fe(H2B(pz)2)2(bipy)] complex (scheme 1 in Figure 1), with H2B(pz)2 = bis(hydrido)bis(1H-pyrazol-1-yl)borate and bipy = 2,2′bipyridine, was synthesized as described earlier.16 The magnetic properties of the powder (Figure 1) indicate a thermal SCO transition in agreement with earlier reports.12,13,19 The [Fe(H2B(pz)2)2(bipy)] complex was deposited on gold and the organic ferroelectric copolymer poly(vinylidene fluoride) with trifluoroethylene (PVDF−TrFE) substrates, in a UHV system equipped with ultraviolet photoemission (UPS) and inverse photoemission (IPES) characterization tools. The SCO films grown were ∼5 nm thick, with the film growth calibrated by the attenuation of the substrate photoemission signal (see Supporting Information). Infrared (IR) microspectroscopy measurements for the molecular layers deposited on Au and LiTaO3 were carried out at the Center for Advanced Microstructures and Devices (CAMD)22−24 by means of the IR reflection mode of a Thermo Nicolet continuum microscope.25−28 Spectra were recorded on a Nicolet 6700 FT-IR spectrometer (in the ATR mode, diamond crystal). Raman spectroscopy was taken with a Renishaw InVia Raman microscope with an excitation wavelength of 514 nm and lateral resolution of ∼1 μm. The evaporated films of the [Fe(H2B(pz)2)2bipy] deposited on the ferroelectric copolymer poly(vinylidene fluoride) with trifluoroethylene (PVDF−TrFE: 70−30) have Raman spectral features that match those of the spectra taken of the powder as well as the prior Raman data of the high-spin (HS) state16 (see the Supporting Information). The agreement of the IR data with calculated dipole active modes (Supporting Information) and of the Raman spectra with prior results16 provides compelling evidence that the molecular structure of the SCO complex is preserved after evaporation on both gold and ferroelectric substrates like the inorganic ferroelectric LiTaO3 (Supporting Information) or the organic ferroelectric copolymer poly(vinylidene fluoride) with trifluoroethylene (PVDF−TrFE: 70−30).15,16,26 The changes in the Raman and infrared vibrational mode intensities suggest that the evaporated films of the Fe(H2B(pz)2)2bipy] favor a preferential orientation when deposited on a ferroelectric (see Supporting Information).

3. MOLECULAR ORBITAL AND MOLECULAR BAND OFFSETS Insight into the molecular orbital energies of both the occupied and unoccupied states was obtained from the UPS and IPES spectra for samples deposited on conductive gold substrates (Figure 2). The final state binding energies can be deduced from the spectra and compared to the orbital energies obtained from theory, corrected for the work function. The orbital 23292

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296

The Journal of Physical Chemistry C

Article

2−3 eV. This is in close proximity to the value deduced from the single molecule DFT calculation (2.7 eV) but much smaller than the value obtained from the ground state semiempirical calculation (6.5 eV). The difference between our theoretical expectations and experiment could be indicative of intermolecular hybridization. In any case, the placement of the chemical potential (Fermi level) close to the LUMO indicates that as a molecular semiconductor, [Fe(H2B(pz)2)2(bipy)] thin film conduction properties are likely dominated by electron, not hole, conduction. Such a finding may be of importance in the light of recent results12,13 where it was found that nanocrystals of a SCO polymer showed conductivity values larger than 10−2 Ωm, corresponding to either a bulk semiconductor or a conjugated polymer of nonoptimized conductivity. The combined photoemission and inverse photoemission HOMO−LUMO gap of 2−3 eV is larger than the metal-toligand charge transfer (MLCT) gap, at about 1.9 eV (650 nm), responsible for the deep violet color of the compound. This disagreement is still reasonable, when one considers that intramolecular Coulomb interactions should place the MLCT gap well below the ground state HOMO−LUMO gap, much like an exciton. Nonetheless, the 5T2 → 5E d−d transition at 1.49 eV (830 nm), which is implicated in photoinduced processes for HS Fe(II) spin crossover compounds, lies below the metal-to-ligand charge transfer (MLCT) gap and the calculated 2.7 eV ground state HOMO−LUMO gap, as well as the 2−3 eV range HOMO−LUMO gap determined from experiment. In the LS compounds, the 1A1 → 1T1 d−d transition, however, is seen to occur at 2.34 eV. The X-ray absorption spectroscopy (XAS) data at the iron L3 (2p3/2) edge is in qualitative agreement with the unoccupied density of states measured in IPES (Figure 2). The subpeaks in the 2p3/2 peak in the X-ray absorption spectra are due to the unoccupied orbitals with strong iron weight. This shift of weighting in XAS, when compared to inverse photoemission, is the result of the core to bound state excitation process which gives greater weight to the unoccupied states that have a strong iron component in XAS. This is a key point since low-energy excitations into these low-lying unoccupied states provide paths in the energy landscape to trigger a spin transition. Our spectroscopic findings are therefore consistent with the picture of photoexcited states related to a modification of the Fe ligand field. Overall, the unoccupied states with strong Fe weight lie at energies well above the lowest unoccupied molecular orbital (LUMO) state, as indicated in the XAS spectra (Figure 2). The iron L3 (2p3/2) edge at roughly 708 eV photon energy, in XAS, is consistent with the Fe 2p3/2 core level binding energy. Furthermore, the density of states indicated by the XAS spectral intensity, above the absorption edge, has an energy placement that is consistent with the satellite features found in the Fe 2p core level X-ray photoemission for [Fe(H2B(pz)2)2(bipy)], as detailed below.

Figure 2. Combined photoemission (UPS) and inverse photoemission (IPES) of [Fe(H2B(pz)2)2(bipy)] deposited on Au (with the background subtracted), compared to a semiempirical single molecule calculation (PM3) and density functional theory (DFT). The vertical bars indicate the orbital energies for each calculation. The molecular orbital diagrams of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) are shown as insets. Binding energies are denoted in terms of E − EF, so occupied energies are negative. The L3 (Fe 2p3/2) X-ray absorption spectrum (XAS) is included for comparison, shifted so the core absorption edge lines up with the chemical potential, thus illustrating the Fe weighted unoccupied states (see text). The components contributing to the inverse photoemission spectra, shown underneath the IPES spectra, line up with the major components of the XAS spectra.

energies of the single molecule unpaired spins were calculated using both the semiempirical (PM3) and the hybrid density functional theory (DFT B3LYP) methods, as has been undertaken successfully elsewhere.21,26,28,35 Geometric optimization of the system was performed by obtaining the lowest unrestricted Hartree−Fock (UHF) energy states, and the DFT calculations were done with the Spartan package 06, with the standard 6-31G* basis set. A model density of states was obtained by applying equal Gaussian envelopes of 1 eV width to each molecular orbital at the ground state binding energies to account for the solid state broadening in photoemission. This extrinsic width applied to the molecular orbital eigenstates, together with a rigid energy shift of a typical value of 5.4 and 3.5 eV was applied to the calculated electronic structure by PM3 and DFT, respectively.25,30,32,39 The applied shift in the orbital energies has a magnitude similar to the Au work function value in the case of the semiempirical calculations. We expect a number of differences between experiment and theory, since both types of calculations estimate the ground state electronic structure, while photoemission and inverse photoemission are final state spectroscopies and subject to matrix element effects which have not been included in the ground state calculations. Furthermore, these are single molecule calculations, and the molecules studied here are well screened in the final state by the gold substrate. The highest occupied molecular orbital (HOMO) to lowest unoccupied molecular orbital (LUMO) gap for the [Fe(H2B(pz)2)2(bipy)] complex on Au, as determined from the combined photoemission and inverse photoemission, is about

4. LOCAL CHARGE OCCUPATION Figure 3 details the iron Fe 2p3/2 core level binding energy of 709.5 ± 0.3 eV, obtained for [Fe(H2B(pz)2)2(bipy)] deposited on gold and PVDF−TrFE. This value is somewhat higher than the one found for the iron oxide Fe3O4 (708.3 eV,42 708.2 eV43) but is close to the values found for FeO (709.4 eV,43 709.542 to 709.8 eV44) and smaller than observed for Fe2O3 (710.6 eV,44,45 710.8 eV,42 710.9 eV43). The 6 eV satellite feature of the 2p iron core level, found about 715 eV binding energy, is characteristic of a two hole bound state as could 23293

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296

The Journal of Physical Chemistry C

Article

Because the SCO molecular thin films are of significant thickness (5 nm), it is reasonable to expect that the electronic structure is independent of the substrate, consistent with the core level spectra. Indeed the magnetic properties are similar to those seen with the purified power, illustrated in Figure 1. As seen in Figure 5, the magnetization of thin films of

Figure 3. X-ray photoemission Fe 2p core level spectra of [Fe(H2B(pz)2)2(bipy)] deposited on (a) gold and (b) the organic ferroelectric PVDF−TrFE. The fittings (c) show the main Fe 2p3/2 and Fe 2p1/2 components as well as the satellite features due to the two hole bound states (see text). All binding energies here are below EF. Figure 5. Product of the magnetic susceptibility with temperature, shown as a function of temperature, for [Fe(H2B(pz)2)2(bipy)] thin film (∼5 nm thickness) deposited on PVDF−TrFE, in arbitrary units, with the paramagnetic contributions of the substrate subtracted.

occur with photoemission accompanied by an excitation from the highest occupied molecular orbital (HOMO) to the unoccupied states.46,47 A corresponding photoemission satellite feature is not evident in the B 1s core level (Figure 4), suggesting that this shakeup satellite is localized on the ironweighted unoccupied molecular orbital states within [Fe(H2B(pz)2)2(bipy)].

[Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE show a variation with temperature that is similar to the behavior of the powder. It is not clear whether [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE remains a large local moment paramagnet in the high-spin state, owing to the difficulty in calibrating the magnetic moment for thin films measurement. Nevertheless, the similarities between the shape of the magnetization curves of Figure 1 (powder) and Figure 5 (thin film), in conjunction with the similarities in the Raman spectra between powder and thin films samples, provide us strong confidence that the molecule can be deposited without degradation from the vapor.

5. CHANGES IN THE UNOCCUPIED ELECTRONIC STRUCTURE IPES data indicate that there are shifts in the unoccupied states for [Fe(H2B(pz)2)2(bipy)] with temperature as well as differences between [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE and on Au (Figure 6). A shift of about 1.7 eV

Figure 4. B 1s X-ray core level photoemission spectrum for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE. All binding energies here are below EF in terms of E − EF.

The B 1s binding energy of 189.2 ± 0.5 eV, of Figure 4, is less than the one found for BN (190.3 eV48,49 to 190.5 eV50,51). However, the local boron environment is hydrogenated in [Fe(H2B(pz)2)2(bipy)], and the resulting charge donation should result in a smaller B 1s core level binding energy than observed in BN. Thus, apart from the satellite feature, the local charge occupation on the boron and iron sites are about what we would expect. Note, however, that the core level binding energies could also be in error as the final state screening for photoemission and inverse photoemission from [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE should be weak. The Fe 2p core level binding energies are sensitive to final state effects and may contribute to the increased Fe 2p3/2 core level binding energies obtained for [Fe(H2B(pz)2)2(bipy)], although such an assessment is not consistent with the observed B 1s core level binding energy. In fact, the final state effects are also difficult to separate out from the observed changes in the unoccupied density of states distribution that accompanies the temperature-driven spin crossover (SCO) transition, as discussed below.

Figure 6. Changes in the unoccupied states from inverse photoemission for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE. The orbital energies (a) and calculated unoccupied density of states (b) from density functional theory (DFT-B3LYP) are compared to inverse photoemission of [Fe(H2B(pz)2)2(bipy)] deposited on gold (c) and on the organic ferroelectric PVDF−TrFE at 300 K (d) and 170 K (e). 23294

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296

The Journal of Physical Chemistry C

Article

but the shift in the distribution of the unoccupied density of states, much in the same way as has been seen for other SCO systems8,51,52 that makes this temperature-dependent change with decreasing temperature so compelling (Figure 6). Other prior results also suggest that there is a redistribution of the density of states that accompanies the SCO transition,14 but again, a change in molecular packing or molecular configuration cannot, a priori, be excluded.

in the unoccupied state peak intensity away from the Fermi level to higher binding energies is observed for [Fe(H2B(pz) 2 ) 2 (bipy)] films deposited on PVDF−TrFE when compared to films deposited on Au (Figure 6). This shift in the unoccupied state density is in fact similar to the shift of about 1.9 eV of the LUMO seen for the largely nonpolar Co phthalocyanine when adsorbed on conducting gold and the insulating organic ferroelectric poly(vinylidene fluoride).31,52 A larger HOMO−LUMO gap, as determined by combined photoemission and inverse photoemission, is expected for a molecular adsorbate on a dielectric as a consequence of final state effects and the influence of the substrate polarization on the electronic structure of the molecular adsorbate.53 A shift in the unoccupied density of states due to charging is not observed here. We exclude charging effects because an appreciable density of unoccupied states just above the Fermi level exists, as seen in Figure 6 for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE. Furthermore, the iron core level spectra are very similar for the two substrates (Figure 3). Neglecting background, unoccupied density of states, as inferred from inverse photoemission, for Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE appears similar, when compared to the films deposited on Au. The unoccupied density of states for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE appears to be in much closer agreement with our expectations from density functional theory (using the hybrid B3LYP functional). Excluding charging effects, a redistribution of the unoccupied density of states for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE and Au (Figure 6) may occur for one of several reasons. The molecular films on PVDF−TrFE substrate are expected to differ significantly from films on Au substrates. A large dipole from the ferroelectric substrate, with the dipole largely pointing “up”, is expected at the interface between the [Fe(H2B(pz)2)2(bipy)] and the poly(vinylidene fluoride).29 Because the PVDF−TrFE ferroelectric substrate is insulating, it should be possible to take advantage of the intrinsic large stray electric field from the substrate to induce a high-spin state of the SCO complex, following the implications of recent reports of electric field switching with a STM tip.8 The [Fe(H2B(pz)2)2(bipy)] molecule could easily adopt a different molecular packing arrangement and/or a different molecular configuration resulting from the differences in electrical properties between conducting Au and insulating PVDF−TrFE. While changes in molecular configuration are usually associated with the spin crossover transition, this need not be universally true, and the former may happen independently of the latter. Upon cooling, the unoccupied density of states for [Fe(H2B(pz)2)2(bipy)] deposited on PVDF−TrFE is significantly modified, with the density of states shifting to energies well above the Fermi level. This shift in the unoccupied density of states away from the Fermi level (chemical potential) is in fact very similar the changes in the unoccupied density of states observed for other iron-based SCO molecular systems using XAS.54,55 Because the conductivity is expected to be lower in the SCO systems at lower temperatures, especially below the HS−LS transition,10 final state screening effects cannot be a priori excluded by our data alone: the large shift in the unoccupied molecular orbital density of states (Figure 6) seen in inverse photoemission may occur because the [Fe(H2B(pz)2)2(bipy)] molecular film rests on the insulating organic ferroelectric PVDF−TrFE substrate which is not metallic.34 In fact, it is not just a shift to energies away from the Fermi level

6. SUMMARY The SCO molecular [Fe(H2B(pz)2)2(bipy)] complex can be deposited “intact” on both gold and on PVDF−TrFE substrates from the vapor. The [Fe(H2B(pz)2)2(bipy)] HOMO−LUMO gap determined from combined photoemission and inverse photoemission is consistent with expectations and slightly larger than the observed d−d transition and optical gap energy values. This suggests that optical excitations are strongly influenced by local Coulomb interactions with the photohole. The presence of a shake-up satellite in the Fe core level spectrum, localized at the iron atom, is also consistent with optical d−d transitions localized at the iron atom. The [Fe(H2B(pz)2)2(bipy)] molecular films exhibit significant energy shift of the unoccupied density of states with temperature, indicative that some thermal spin crossover is observed on a poled organic ferroelectric substrate.



ASSOCIATED CONTENT

S Supporting Information *

Infrared (IR) microspectroscopy measurements and Raman spectroscopy showing the vibrational spectroscopic signatures of the molecule as synthesized and after vapor deposition; attenuation of the substrate signal indicating a molecular film thickness of 5 nm. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Tel +1-402-472-9838, Fax +1-402-472-6148, e-mail [email protected] (P.A.D.); Tel +33(0)540002678, Fax +33(0)540002649, e-mail [email protected] (P.R.), [email protected] (J.L.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported by the National Science Foundation through grants numbered CHE-0909580 as well as the Nebraska MRSEC (DMR-0820521). The XAS-STXM work described in this paper was performed at the Canadian Light Source, which is supported by the NSERC, NRC Canada. Support from the French funding agency (ANR projects MOSE and MultiSelf) and Région Aquitaine is also gratefully acknowledged. The authors would like to thank Ning Wu for significant technical assistance.



REFERENCES

(1) Spin Crossover in Transition Metal Compounds; Gütlich, P., Goodwin, H. A., Eds.; Top. Curr. Chem. 2004, 233−235. (2) Bousseksou, A.; Molnar, G.; Salmon, L.; Nicolazzi, W. Chem. Soc. Rev. 2011, 40, 3313−3335. (3) Létard, J.-F.; Guionneau, P.; Laurence, G.-C. Top. Curr. Chem. 2004, 235, 221−249. (4) Halcrow, M. A. Chem. Soc. Rev. 2011, 40, 4119−4142. 23295

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296

The Journal of Physical Chemistry C

Article

(33) Xu, B.; Choi, J.; Caruso, A. N.; Dowben, P. A. Appl. Phys. Lett. 2002, 80, 4342−4344. (34) Xiao, J.; Dowben, P. A. J. Phys.: Condens. Matter 2009, 21, 052001. (35) Choi, J.; Borca, C. N.; Dowben, P. A.; Bune, A.; Poulsen, M.; Pebley, S.; Adenwalla, S.; Ducharme, S.; Robertson, L.; Fridkin, V. M.; Palto, S. P.; Petukhova, N.; Yudin, S. G. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61, 5760. (36) Sokolov, A.; Yang, C.-S.; Yuan, L.; Liou, S.-H.; Cheng, R.; Jeong, H.-K.; Komesu, T.; Xu, B.; Borca, C. N.; Dowben, P. A.; Doudin, B. Europhys. Lett. 2002, 58, 448−454. (37) Zhang, J.; McIlroy, D. N.; Dowben, P. A.; Zeng, H.; Vidali, G.; Heskett, D.; Onellion, M. J. Phys.: Condens. Matter 1995, 7, 7185− 7194. (38) McIlroy, D. N.; Zhang, J.; Dowben, P. A.; Heskett, D. Mater. Sci. Eng., A 1996, 217/218, 64. (39) Balaz, S.; Caruso, A. N.; Platt, N. P.; Dimov, D. I.; Boag, N. M.; Brand, J. I.; Losovyj, Ya. B.; Dowben, P. A. J. Phys. Chem. B 2007, 111, 7009−7016. (40) Kaznatcheev, K. V.; Karunakaran, C.; Lanke, U. D.; Urquhart, S. G.; Obst, M.; Hitchcock, A. P. Nucl. Instrum. Methods Phys. Res., Sect. A 2007, 582, 96−99. (41) Jacobsen, C.; Wirick, S.; Flynn, G.; Zimba, C. J. Microsc. 2000, 197, 173−184. (42) Brion, D. Appl. Surf. Sci. 1980, 5, 133−152. (43) McIntyre, N. S.; Zetaruk, D. G. Anal. Chem. 1977, 49, 1521. (44) Barr, T. L. J. Phys. Chem. 1978, 82, 1801−1810. (45) Dzhurinskii, B. F.; Gati, D.; Sergushin, N. P.; Nefedov, V. I.; Salyn, Ya. V. Russ. J. Inorg. Chem. 1975, 20, 2307−2314. (46) Chandesris, D.; Lecante, J.; Petroff, Y. Phys. Rev. B 1983, 27, 2630. (47) Chandesris, D.; Lecante, J.; Petroff, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 34, 8971−8972. (48) Riviere, J. P.; Cahoreau, M.; Pacaud, Y. Thin Solid Films 1993, 227, 44−53. (49) Benyagoub, A.; Faussemagne, A.; Marest, G.; Moncoffre, N.; Delichere, P. Surf. Coat. Technol. 1996, 83, 70−73. (50) Hamrin, K.; Johansson, G.; Gelius, U.; Nordling, C.; Seigbahn, K. Phys. Scr. 1970, 1, 277. (51) Kohiki, S.; Ohmura, T.; Kasuo, K. J. Electron Spectrosc. Relat. Phenom. 1983, 31, 85. (52) Kong, L.; Medina, G. J. P.; Santana, J. A. C.; Wong, F.; Bonilla, M.; Amill, D. A. C.; Rosa, L. G.; Routaboul, L.; Braunstein, P.; Doudin, B.; Lee, C.-M.; Choi, J.; Xiao, J.; Dowben, P. A. Carbon 2012, 50, 1981−1986. (53) Garcia-Lastra, J. M.; Rostgaard, C.; Rubio, A.; Thygesen, K. S. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 245427. (54) Collison, D.; Garner, C. D.; McGrath, C. M.; Mosselmans, J. F. W.; Roper, M. D.; Seddon, J. M. W.; Sinn, E.; Young, N. A. Dalton Trans. 1997, 4371−4376. (55) Lee, J.-J.; Sheu, H.-S.; Lee, C.-R.; Chen, J.-M.; Lee, J.-F.; Wang, C.-C.; Huang, C.-H.; Wang, Y. J. Am. Chem. Soc. 2000, 122, 5742− 5747.

(5) Baadji, N.; Piacenza, M.; Tugsuz, T.; Sala, F. D.; Maruccio, G.; Sanvito, S. Nat. Mater. 2009, 8, 813. (6) Mahfoud, T.; Molnár, G.; Bonhommeau, S.; Cobo, S.; Salmon, L.; Demont, P.; Tokoro, H.; Ohkoshi, S. I.; Boukheddaden, K.; Bousseksou, A. J. Am. Chem. Soc. 2009, 131, 15049. (7) Gopakumar, T. G.; Matino, F.; Naggert, H.; Bannwarth, A.; Tuczek, F.; Berndt, R. Angew. Chem. 2012, 51, 6262. (8) Miyamachi, T.; Gruber, M.; Davesne, V.; Bowen, M.; Boukari, S.; Joly, L.; Scheurer, F.; Rogez, G.; Yamada, T. K.; Ohresser, P.; Beaurepaire, E.; Wulfhekel, W. Nat. Commun. 2012, 3, 938. (9) Gütlich, P.; Garcia, Y.; Goodwin, H. A. Chem. Soc. Rev. 2000, 29, 419−427. (10) Mahfoud, T.; Molnár, G.; Cobo, S.; Salmon, L.; Thibault, C.; Vieu, C.; Demont, P.; Bousseksou, A. Appl. Phys. Lett. 2011, 99, 053307. (11) Prins, F.; Monrabal-Capilla, M.; Osorio, E. A.; Coronado, E.; van der Zant, H. S. J. Adv. Mater. 2011, 23, 1545. (12) Etrillard, C.; Faramarzi, V.; Dayen, J.-F.; Létard, J.-F.; Doudin, B. Chem. Commun. 2011, 47, 9663. (13) Létard, J.-F.; Etrillard, C.; Doudin, B.; Faramarzi, V.; Dayen, J.-F. Patent FR 2011/11 50949, 2011-02-07. (14) Aravena, D.; E. Ruiz, E. J. Am. Chem. Soc. 2012, 134, 777−779. (15) Real, J. A.; Muñoz, M. C.; Faus, J.; Solans, X. Inorg. Chem. 1997, 36, 3008. (16) Palamarciuc, T.; Oberg, J. C.; Hallak, F. E.; Hirjibehedin, C. F.; Serri, M.; Heutz, S.; Létard, J.-F.; Rosa, P. J. Mater. Chem. 2012, 22, 9690. (17) Shi, S.; Schmerber, G.; Arabski, J.; Beaufrand, J.-B.; Kim, D. J.; Boukari, S.; Bowen, M.; Kemp, N. T.; Viart, N.; Rogez, G.; Beaurepaire, E.; Aubriet, H.; Petersen, J. Appl. Phys. Lett. 2009, 95, 043303. (18) Naggert, H.; Bannwarth, A.; Chemnitz, S.; v. Hofe, T.; Quandt, E.; Tuczek, F. Dalton Trans. 2011, 40, 6364. (19) Moliner, N.; Salmon, L.; Capes, L.; Munoz, M. C.; Létard, J.-F.; Bousseksou, A.; Tuchagues, J.-P.; McGarvey, J. J.; Dennis, A. C.; Castro, M.; Burriel, R.; Real, J. A. J. Phys. Chem. B 2002, 106, 4276− 4283. (20) Xiao, J.; Sokolov, A.; Dowben, P. A. Appl. Phys. Lett. 2007, 90, 242907. (21) Dowben, P. A.; Rosa, L. G.; Ilie, C. C.; Xiao, J. J. Electron Spectrosc. Relat. Phenom. 2009, 174, 10−21. (22) Hormes, J.; Scott, J. D.; Suller, V. P. Synchrotron Radiat. News 2006, 19, 27. (23) Roy, A.; Morikawa, E.; Bellamy, H.; Kumar, C.; Goettert, J.; Suller, V.; Morris, K.; Ederer, D.; Scott, J. Nucl. Instrum. Methods Phys. Res., Sect. A 2007, 582, 22−25. (24) Morikawa, E.; Scott, J. D.; Goettert, J.; Aigeldinger, G.; Kumar, Ch. S. S. R.; Craft, B. C.; Sprunger, P. T.; Tittsworth, R. C.; Hormes, F. J. Rev. Sci. Instrum. 2002, 73, 1680−1683. (25) Xiao, J.; Zhang, Z.; Wu, D.; Routaboul, L.; Braunstein, P.; Doudin, B.; Losovyj, Ya. B.; Kizilkaya, O.; Rosa, L. G.; Borca, C. N.; Gruverman, A.; Dowben, P. A. Phys. Chem. Chem. Phys. 2010, 12, 10329−10340. (26) Kizilkaya, O.; Scott, J. D.; Morikawa, E.; Garber, J. D.; Perkins, R. S. Rev. Sci. Instrum. 2005, 76, 13703. (27) Kizilkaya, O.; Prange, A.; Steiner, U.; Oerke, E.-C.; Scott, J. D.; Morikawa, E.; Hormes, J. Nucl. Instrum. Methods Phys. Res., Sect. A 2007, 582, 274. (28) Zhang, Z.; Sharma, P.; Borca, C. N.; Dowben, P. A.; Gruverman, A. Appl. Phys. Lett. 2010, 97, 243702. (29) Zhang, Z.; González, R.; Díaz, G.; Rosa, L. G.; Ketsman, I.; Zhang, X.; Sharma, P.; Gruverman, A.; Dowben, P. A. J. Phys. Chem. C 2011, 115, 13041−13046. (30) Feng, D.; Losovyj, Ya.; Tai, Y.; Zharnikov, M.; Dowben, P. A. J. Mater. Chem. 2006, 16, 4343−4347. (31) Xiao, J.; Dowben, P. A. J. Mater. Chem. 2009, 19, 2172−2178. (32) Zhang, Z.; Alvira, J.; Barbosa, X.; Rosa, L. G.; Routaboul, L.; Braunstein, P.; Doudin, B.; Dowben, P. A. J. Phys. Chem. C 2011, 115, 2812−2818. 23296

dx.doi.org/10.1021/jp3034962 | J. Phys. Chem. C 2012, 116, 23291−23296