electron affinity states of metal supported phthalocyanines … · 2017. 9. 6. · phthalocyanines...

Journal of Porphyrins and PhthalocyaninesJ. Porphyrins Phthalocyanines 2012; 16: 1–9

DOI: 10.1142/S1088424612004574

Published at http://www.worldscinet.com/jpp/

Copyright © 2012 World Scientific Publishing Company

INTRODUCTION

Phthalocyanines (Pc) are a chemically versatile group of organic semiconductors that have been successfully employed as electroactive elements in a number of electronic devices including advanced photovoltaics [1, 2], sensors [3], field-effect transistors [4], nonlinear optics [5], optical data storage [6], and electrochromic displays [7]. The optoelectronic properties of the Pc systems can be manipulated by incorporating different metal ions in their cavity and covalent attachment of diverse functional side groups [8, 9]. Furthermore, Pc macrocycles are also known to self-aggregate by forming

complex supramolecular assemblies having physical, chemical, electrical, and optical properties different from the discrete molecular units [10–12]. Once these molecular materials are deposited onto an electrode surface (via physisorption or chemisorption) a significant redistribution of electron density in either the molecule/aggregate system or at the adsorbate/electrode interface may result, significantly altering device performance [13,14]. Thus, the knowledge of oxidation and reduction states, the location of the conduction bands, and the bias dependence and mechanism of electrochemical processes of the Pc adsorbates at the substrate interface becomes a crucial prerequisite for their use in electronic devices.

The measurement of oxidation states or ionization potentials (IP) of Pc adsorbates is usually accomplished by employing ultraviolet photoelectron spectroscopy, UPS [15–18]. While inverse photoemission spectroscopy,

Electron affinity states of metal supported phthalocyanines measured by tunneling spectroscopy

K.W. Hipps*◊ and Ursula Mazur*◊

Department of Chemistry and Materials Science and Engineering Program, Washington State University, Pullman, WA 99164-4630, USA

Received 23 November 2011Accepted 28 December 2011

ABSTRACT: Orbital Mediated Tunneling Spectroscopy OMTS (elastic electron tunneling) was employed in measuring electron affinity levels (EA) of unsubstituted, alkylated, sulfonated, and metalated phthalocyanines (Pc) adsorbed as single molecules or aggregates on metal substrates and imbedded in metal-insulator-metal (M-I-M) devices. MPc complexes were vapor deposited, solution phase doped, or transferred as Langmuir–Blodgett films. It was determined that while the nature of the substituents has a large effect on the gas phase electron affinities, they play a minimal role on the electron affinities of metal supported phthalocyanines. Moreover, the orientation of monolayer films and the method of film deposition (vapor, solution, Langmuir–Blodgett) also appear to play only a minor role in determining the electron affinities. Electrochemical reduction potentials obtained for the solution phase molecular systems are compared to the OMTS data and a strong correlation is observed. In contrast, the predicted EA values for the gas phase molecules show little correspondence with their OMTS equivalents for adsorbed phthalocyanines. Inelastic scattering from phthalocyanine p�p* transitions and metal centered d–d transitions are observed for chromophores imbedded in tunnel diodes. Both the observed lowest spin forbidden transitions and the calculated gas phase HOMO–LUMO gaps are only weakly affected by Pc substitution and surface orientation.

KEYWORDS: OMTS, IETS, STM, AFM, tunnel diode, electron affinity, phthalocyanine, adsorbed macrocycles, Langmuir–Blodgett films.

SPP full member in good standing

*Correspondence to: Ursula Mazur, email: [email protected], fax: +1 509-335-8867 and K.W. Hipps, email: [email protected]

S1088424612004574.indd 1 3/19/2012 2:51:04 PM

http://dx.doi.org/10.1142/S1088424612004574

http://dx.doi.org/10.1142/S1088424612004574

Copyright © 2012 World Scientific Publishing Company J. Porphyrins Phthalocyanines 2012; 16: 2–9

2 K.W. HIPPS AND U. MAZUR

IPS, has been successfully applied to the electron affinity (EA) levels (reduction states) of metals and inorganic semiconductors chemisorbed on metal surfaces, applications to large molecular systems such as phthalocyanines have been few [16, 19, 20]. One reason for this scarcity of applications is that most organic compounds cannot survive the experiment; only a few minutes of electron beam exposure (at currents required to generate usable signals) is sufficient to destroy most organic compounds. The most commonly used measure of electron affinities of phthalocyanines is the half-wave reduction potential measured in nonaqueous solvents, E1/2

[21–23]. Given the solvent dependence of some of these potentials, and the importance of coordination state of the metal ion in the Pc complexes, these estimates easily can be off by 1 eV. A more recent and more appropriate spectroscopic method for easily and rapidly determining ionization and affinity states of adsorbed species and thin films is orbital mediated tunneling spectroscopy, OMTS [24–27]. OMTS measures elastic tunneling events that utilize molecular orbitals.

OMTS work has been performed on metal-insulator-adsorbate-metal (M-I-S-M’) tunnel diodes typically used for inelastic electron tunneling spectroscopy (IETS) and in the scanning tunneling microscope, STM, environment [24–27]. In either configuration, the basic physics is essentially the same, although oxide or solvent chemistry may complicate the analysis relative to vacuum STM work.

In this work, we present measurements of the unoccupied molecular orbital energies of different phthalocyanines using OMTS. Unsubstituted, alkylated, and sulfonated phthalocyanines (Fig. 1) adsorbed as single molecules or aggregates on gold substrates and imbedded in aluminum-aluminum oxide-lead (Al-AlOx-Pb) devices are employed. We determine the energy of the unoccupied orbitals of the Pc derivatives as a function of molecular structure, aggregation type, and the influence of interface dipoles. The electron affinities of the phthalocyanines

obtained from the tunneling spectra are compared with their LUMO energies in the gas phase (calculated) and the electrochemical potentials acquired from solution phase measurements.

EXPERIMENTAL

Materials

Copper(II) 2,3,9,10,16,17,23,24-octabuloxyphthalocy- anine, pCuPc(OBu)8, was prepared using the synthetic method reported elsewhere [28, 29]. The compound was recrystallized from THF and ethanol prior to use. The phthalocyanine tetrasulfonic acid, TSPc, was purchased from Frontier Scientific and CoTSPc, CuTSPc and ZnTSPc were obtained from Sigma Aldrich. All TSPc complexes were used without further purification or separation of regioisomers. Gold metal, 99.999% pure, was acquired from Cerac. All metal phthalocyanines (MPc: CoPc, CuPc and ZnPc) were purchased as high-purity compounds from Sigma Aldrich. Because most of the impurities have a higher vapor pressure than the MPc, all the compounds were twice presublimed and then sublimed before use. All metals used were >99.99% purity. The metals used in tunnel diodes were resistively deposited from W wire (Al) or from Mo boats (Pb). All solvents used were either reagent or spectroscopic grade and were acquired either from Sigma-Aldrich or from J.T. Baker. Millipore water (18 MW) was degassed by boiling for 1 h. Corning microscope slides were cleaned in a HNO3/H2O2 bath, and stored in high purity water until use. Mica strips (1 × 3 in and 1 × 4 cm) were purchased from Ted Pella and were freshly cleaved before use.

Preparation of (Al-AlOx-MPc-Pb) tunnel diodes

The fabrication of tunnel diodes was accomplished as follows. Typically, 100 nm of aluminum was deposited on each of three corning glass microscope slides at a pressure of <6 × 10-7 Torr. Oxygen gas (90 milliTorr) was admitted, and a plasma was created with a 400 V ac discharge. Three sequential oxidations of this type gave a total oxidation time of 15 min to 18 min. The vacuum system was then opened, and the substrates were placed over a tray of distilled water (in the vapor) for an additional period of 2 h. This extreme oxidation process was required to obtain junctions with an oxide thickness of about 2 nm that can reliably withstand several volts of bias. Once the oxide growth was completed, the molecules of interest were deposited onto the native aluminum oxide surface. All devices were completed by evaporating 100–150 nm of Pb at ~5 × 10-7 Torr.

Phthalocyanines were incorporated into the diode structures by one of three methods. MPc complexes were either vapor deposited onto the oxide layer, spin-doped from DMSO solution, or transferred as Langmuir–Blodgett films. The unsubstituted MPc complexes were

N

N

N

N

N

N

N

N

R

R' R

R'

R

R'R

R'

M

M: 2H, Co, Cu, Mg, ZnR: H, OCH2CH2CH2CH3, SO3HR': H, OCH2CH2CH2CH3

Fig. 1. Molecular structure of a phthalocyanine bearing different substituents. Phthalocyanine tetrasulfonic acid, TSPc, has several regioisomers, which are not identified in the figure

S1088424612004574.indd 2 3/19/2012 2:51:05 PM


ELECTRON AFFINITY STATES OF METAL SUPPORTED PHTHALOCYANINES 3

deposited from either a baffled W box source (ME-1 from R.D. Mathis, Inc.) or a quartz crucible; no difference in the resulting films was observed. Approximately 0.2–0.4 nm of MPc was deposited at a rate of about 0.02 nm/s. Deposition rate estimates are based on readings from a quartz crystal film thickness monitor. An average value of 1.58 g/cm3 was used for the density of the MPcs. For solution doping, 10-3 M concentrations of MTSPc in DMSO were employed. The LB films of pCuPc(OBu)8 were prepared and transferred onto the AlOx surface by a procedure outlined below.

Langmuir–Blodgett film preparation

The pCuPc(OBu)8 LB layer was prepared using the KSV 5000 LB double trough system by a procedure described earlier [28, 29]. Aliquots (200–900 mL) of 1 × 10-4 M to 2 × 10-4 M solution of the complex dissolved in chloroform were spread on a MilliPore purified water subphase typically kept at approximately 20 °C. The Pc layer was allowed to stabilize for several minutes and was then compressed at a barrier speed of 20 mm/min until the desired surface pressure of 15 mN/m was reached. The molecular monolayer of Pc was transferred via Z-type deposition onto freshly oxidized aluminum substrates. The dipping speed was 10 mm/min and the transfer ratios ranged between 1.0 and 1.10. The films were allowed to dry for 10 min under vacuum and 2 min under a stream of compressed, filtered air. These substrates were then transferred to the vacuum system described above and the Pb counter electrodes were deposited.

AFM imaging

A monolayer of pCuPc(OR)8 LB film was prepared on freshly cleaved mica substrates. Cobalt, copper, and zinc tetrasulfonyl phthalocyanines in DMSO at 10-4 M concentrations were also deposited on mica. The specimen were then mounted on a metal disk with double-stick tape and placed on the top of the scanner tube. The LB samples were always set in such a way that the fast scan direction and the dipping direction were the same. A Nanoscope IIIA multimode microscope from Digital Instruments, DI, was employed. All micrographs were acquired in tapping mode with commercial silicon tips having a diameter of 5–10 nm. The cantilevers had a length of 127 mm and a force constant of the order of 50 N/m. The setpoint varied between 1.85 and 2.75 V and the resonance frequency of the cantilevers varied between 260–340 kHz.

STM based OMTS

Au(111)/mica substrates prepared in-house [30, 31] were used. The TSPc samples were deposited on Au(111) as follows. A 10 mM TSPc aqueous solution at pH = 0 was prepared using HCl for acidification. The pH value was confirmed by direct measurement. A 10 mL aliquot was dropped directly onto the substrate, allowed to

remain for approximately 10 min, and then spun dry for 30 sec at 4400 rpm. Samples were stored in a desiccator before using. Imaging experiments of TSPc on Au(111) were performed in a UHV temperature controlled environment. The samples were transferred via air-lock into the UHV STM chamber (working pressure <1 × 10-10 Torr) where the TSPc aggregate samples were heated to 100–130°C for a period of 3 min or more in order to remove HCl and water.

The STM and controller were purchased from RHK technology and both constant current images and I–V data were acquired with this system. Unless otherwise stated, the images were plane-fit and low-pass filtered. The data were acquired at 19 °C. Both etched W and Pt0.8Ir0.2 tips were used. Generally, the tips required a cleaning step (Argon ion sputtering) in order to produce high quality I(V) curves on a clean gold surface. Spectroscopy was performed by using the RHK software to measure current as a function of sample bias voltage, I(V), at fixed tip-sample separation (feedback off). Multiple curves were acquired at each setting and then averaged. dI/dV curves were obtained as a numerical derivative of the average I(V). Orbital mediated tunneling spectra, OMTS, (dI/dV at fixed height) were measured on well-defined TSPc aggregates on Au(111).

Diode based IETS and OMTS

All diode based tunneling spectra were acquired from Al-AlOx-Pb junctions measured in constant resolution mode. That is, they are normalized tunneling intensities (NTI), (dI2/dV2)/(dI/dV). A modulation voltage of 8–25 mV (rms) was used in collecting spectra. When several spectra are displayed on the same scale, they have either been acquired at the same modulation level or the intensities have been corrected by scaling. Each reported spectrum is the sum of from 16 to 100 repeated scans. In certain cases to be indicated later, the spectrum was simplified by fitting a polynomial (order � 7) to the signal background and then subtracting that smooth function from the data. Throughout, we adopt the convention that normal bias exists when the aluminum electrode is negative. If the top metal is relatively negative, we identify this as reverse bias.

Calculations

All calculations were performed using the commercial program Gaussian 03. All reported results are based on DFT calculations using the B3LYP functional. The basis for most of the systems studied was 6-311++ for C, H, and N, and 6-311+(2d,p) for the metal ions. For the metal free sulfonate, the 6-31+(d,p) basis was used. Ionization potentials and electron affinities were determined by computing the energy differences between initial molecule and the appropriate ionized species, usually in the computed equilibrium geometry of the gas-phase parent molecule (vertical IP and EA). The

S1088424612004574.indd 3 3/19/2012 2:51:05 PM



difference between vertical and relaxed electron affinities is small for phthalocyanines, typically one to two orders of magnitude smaller than the calculated affinity. For example, for CuPc the relaxed ion is calculated to be 0.242 eV more stable than the vertical ionized one. For CoPc the calculated difference is less than 0.1 eV. These small changes are likely due to the rigidity of the Pc ring and the extent of delocalization of the added electron. Solution-phase IP and EA values were determined using the PCM (Polarizable Continuum Model) [32, 33] method with the Gaussian 03 parameters for acetonitrile.

RESULTS AND DISCUSSION

Electron tunneling is a particularly useful probe for measuring electron affinities and ionization potentials of molecular systems because it is easy to control the rate of flow and the energy of electrons used. An electron-tunneling device, M-I-M tunnel diode [34–38] and the corresponding features of a STM [25, 39–45] are depicted in Fig. 2. Both devices rely on the same physics. Within the conductors (metal electrodes in the M-I-M′ case, substrate and atomically sharp tip in the STM case) the electrons are in classically allowed regions I and III. In these regions, their total energy E is greater than their potential energy. However, the potential is greater than E in the gap between conductors (the insulator in the M-I-M case, the vacuum or solvent gap in the STM case), region II in Fig. 2. This region is classically forbidden, but quantum-mechanically allowed.

In our M-I-M′ devices, the adsorbate is deposited onto a native oxide surface of an aluminum base metal (M) and then covered with Pb metal (M′). Lead is usually employed as the top electrode because it can be deposited at low temperatures (<300 °C) and generally does not react with the adsorbate [26, 38]. In a STM, the molecule of interest is placed directly onto a metal substrate such as gold or HOPG. In the case of STM based spectroscopy, it is possible to set up precisely controlled regions of a single molecule through which the electrons must tunnel. In both junction configurations, IETS or STM, adsorbate molecules can be considered to be in electronic equilibrium with one metal electrode (lead in a tunnel diode and the substrate in a STM sample) because a relatively large insulating gap separates it from the other electrode. When the bias across the M-I-M′ junction, or the sample-tip gap in a STM, is such that the Fermi surface of the opposite electrode (tip in STM) matches the energy of a vacant molecular orbital in the adsorbate layer, a large increase in conductance occurs due to orbital-mediated electron transfer. In the opposite bias, any orbital-mediated process involves occupied orbitals [25, 43].

Complicating spectral interpretation is the fact that inelastic excitations show themselves as peaks in (d2I/dV2)/(dI/dV), the so-called normalized tunneling intensity or NTI, while orbital mediated tunneling transitions appear as peaks in dI/dV [46]. Because M-I-M′ tunneling data is historically focused on inelastic processes, diode based tunneling spectra are reported as NTI, irrespective of their nature (elastic vs. inelastic). Elastic tunneling

spectra (including OMTS) are reported as dI/dV. Moreover, the use of NTI in IETS is more than simply convention — the rapidly varying baselines associated with the M-I-M′ structure makes this type of measurement essential for data recovery. To complicate matters further, this voltage dependent baseline in diode IETS conventionally is removed in a rather arbitrary manner. Fortunately, it has been demonstrated that, the error in choosing the maximum in the NTI as the maximum in the density of states associated with an OMT process is approximately the half width at half height of the NTI band independent of the baseline chosen [46]. Thus, OMTS bands are identified as one half width above the apparent maximum in NTI, but at the apparent maximum in STM based dI/dV measurements. Finally, it should be noted that the OMT of adsorbates trapped in tunnel diodes samples 109 molecules, while OMT measurement in a STM environment requires only a single molecule.

In Fig. 3, we compare tunneling spectra obtained from derivatized copper phthalo-cyanines incorporated into Al-AlOx-Pb tunnel diodes with scanning tunneling spectra of a free base phthalocyanine and of CoPc, both

Fig. 2. Schematic drawings of a tunnel diode, a STM, and the electronic energy diagram appropriate for both. U is the height of the potential barrier, E is the energy of the incident electron, d is the thickness of the barrier, A is approximately 1.02 [Angstrom/(eV)1/2]-1 if U and E are in electron volts and d is in Angstroms. y0 is the wavefunction of the incident electron and yd is the wavefunction after transmission through the barrier. I is the measured tunneling current, V is the applied bias, and M and M′ are the electrode metals

S1088424612004574.indd 4 3/19/2012 2:51:14 PM



deposited on Au(111). The IET spectra are shifted by -1.00 eV relative to the STS to account for the difference in the work function between the Au(111) substrate (5.3 eV [47]) and that of lead (4.3 eV). Let us first review the spectroscopic data presented in the figure. In addition to a very intense broad OMTS electron affinity bands (p* LUMO), in the vicinity of 4.0 V relative to the vacuum level in the IETS, there also appear several sharp weaker bands spanning the range of -0.5 to +0.5 V bias (top scale) which can be attributed to a manifold of vibrational states.

A relatively strong feature present in all three IET spectra at approximately -1.2 V bias (top scale) is an electronic singlet to triplet excitation of the chromophore. This transition is spin-forbidden in photon spectroscopy. Note that elastic tunneling bands depend upon the absolute energy of the molecular states (lower axis), while inelastic tunneling bands only depend upon energy differences (upper axis). The STM based elastic tunneling spectrum (dI/dV) of CoPc shows several OMT bands. These include ionization from the Pc ring HOMO at 5.8 eV, ionization from the dz2 orbital near 5.5 eV, and the first affinity level at 3.7 eV relative to the vacuum level [48]. The STM based elastic tunneling spectrum of TSPc (Fig. 3) shows an intense peak at 3.98 eV relative to vacuum level that is attributed to tunneling mediated by the LUMO of the Pc ring. The intense peak (or shoulder) near 6.5 eV relative to vacuum level we attribute to HOMO-1 and HOMO-2 (a nearly degenerate pair) [31]. We could not measure the occupied states of the phthalocyanines imbedded in tunnel diodes because of the limiting value of the work function of the lead electrode and the lack of stability of tunnel junctions below -2.0 V bias. The electron affinity values derived from Fig. 3 are summarized in Table 1 using data from several sources, including the electrochemical literature and our calculations [49–51].

Note that in the Al-AlOx-Pb environment the presence of different substituents on the Pc ring [i.e. sulfonate (electron withdrawing group) vs. butoxy (electron donating group) vs. no substituents] has little effect on the EA that centers around 3.7 eV for all the compounds in Table 1. This shifts in the EA values for all of the forms of Pc thin layers supported on metals is relatively tiny (3.7 ± 0.2 eV) compared with the calculated gas phase electron affinities (Table 1) for these compounds. The

Fig. 3. Comparison of IETS spectra acquired from CuPc and its derivatives incorporated into Al-AlOx-Pb junctions (top three traces) at 4 K. The lower two traces are UHV-OMTS (dI/dV) spectra acquired from TSPc and CoPc deposited onto Au(111) at 295 K. The IET spectra (black traces) are shifted by -1.00 eV relative to the STS (red curves) to account for the difference in the work function of the gold and lead substrates. The work functions of Au(111) and Pb are 5.3 eV (Ref. 47) and 4.3 eV, respectively

Table 1. Electron affinity (eV relative to the vacuum level) and electronic excitation values

Compound On surface EA1IETS EA1

STM1 E1/2

red EAsolcal EAvac

cal b LUMO–HOMOb 1p�3p*d

TSPc cofacial multilayer

3.98 [13] 4.14 [21] 3.19 2.12

CuPc flat monolayer

3.86 4.00 2.74 1.42 2.23 1.15

ZnPc flat monolayer

3.74 3.91 [50] 1.15

MgPc flat monolayer

3.70 3.81 [51] 1.15

CuPc(OBu)8 cofacial monolayer

3.70 3.84a (2.52)c 0.98 (1.19c)

2.20 (2.20c) 1.21

CoTSPc cofacial monolayer

3.73 4.16 [49, 21] 1.21

CuTSPc cofacial monolayer

3.62 3.98 [49] 1.20

CoPc flat monolayer

3.60 3.75 [48] 4.16 [50] 3.14 2.06 2.26 1.28

ZnTSPc cofacial monolayer

3.47 3.88 1.20

a10-4 M in DMSO solutions, 0.1 M tetrabutylammonium perchlorate, Ag/AgCl electrode (UM, unpublished results). bCalculated for this work. cCalculation for the octamethoxy species and larger basis. dIETS data in this paper.

S1088424612004574.indd 5 3/19/2012 2:51:17 PM



vapor phase computed EAs range over more than 2 eV. Note, that this gas phase variation is almost entirely related to the overall molecular potential since the LUMO–HOMO difference is essentially unchanged with changes in substituent.

In addition to having different molecular structures these chromophores also, exhibit different surface structures. Monolayers and submonolayers of vapor deposited CuPc [30] on Au(111), as well as other metallophthalocyanines (CoPc, ZnPc, NiPc, and FePc) [25, 30] all form beautiful close packed 2D structures where the molecules lie flat on the substrate surface. TSPc [31, 53], CuTSPc [52], and pCuPc(OBu)8 [28, 29] on the other hand, all form cofacially packed Pc assemblies facilitated by strong p-p interaction between the adjacent chromophores. In highly acidic solutions,

TSPc exhibits coherent columnar architecture with the phthalocyanine macrocycle planes aligned parallel to the substrate surface forming stacks [31, 53], while the Pc chromophores in pCuPc(OBu)8 [28, 29] and CuTSPc [52] are aligned nearly vertical to the surface normal [28, 29]. The detailed analysis of the Pc structures was reported by us in other publications and will not be presented here [28, 29, 31, 53]. For reference, in Figs 4 and 5, we provide scanning probe microscopy images of the Pc adsorbates on solid supports along with cartoon models of their proposed surface structure. In the case of TSPc it is recognized that the tunneling current observed must pass through strongly p-p coupled phthalocyanine assemblies. Since the LUMO of the phthalocyanines is mainly due to the macrocycles’ affinity states, the cofacially stacked configuration is expected to result in the highest degree

Fig. 4. UHV-STM images of (a) vapor deposited monolayer (0.4 nm apparent height) of CoPc on Au(111) and (b) 10 mM TSPc deposited from pH = 0 solution on Au(111) forming vertical columnar assemblies on the surface (a cross section is included). The setpoint for (a) was -0.8 V and 0.05 nA and for (b) 1.6 V and 5 pA. Below each figure is a cartoon of proposed surface structure for the adsorbed phthalocyanine system

Fig. 5. Ambient tapping mode AFM images of LB monolayer of (a) pCuPc(OBu)8 on mica and (b) spin doped 10-4 M CuTSPc in DMSO also on mica. Each figure contains an inset of a higher resolution image 500 nm � 500 nm in size. The Pc film height in the LB and the CuTSPc films is ~1.2 nm which is consistent with the a Pc rings stacking at an angle relative to the surface normal [28, 29, 52] as depicted in a model drawn below the images

S1088424612004574.indd 6 3/19/2012 2:51:34 PM



of overlap between LUMO’s on adjacent molecules and facilitate efficient electron transport. Thus, the through-space electronic interactions between Pc rings within the linear array of cofacial oligomers may be the origin of the increased energy of the LUMO in the TSPc aggregate relative to the LUMO of the CuPc molecular monolayer that has no such interactions.

We also considered the influence of the metal in the macrocycle core on EA. Figures 6 and 7 depict the IET spectra of cobalt (d7), copper (d9), and zinc (d10) TSPc, respectively. While the LUMO OMT of the Cu and Zn derivatives appears at nearly the same bias voltage, the unoccupied p* level in the d7 system shifts to higher energy relative to EAs of the d9 and d10 systems. This difference may be attributed to more significant mixing of metal and p MOs in the case of CoPc. When the MPc complex is adsorbed on a metal substrate, one may no longer use the

states of the isolated molecule, but must consider those of the new species that includes a significant admixture of the density of states of the substrate metal. To support this, recent XPS, UPS, and X-ray Excited Auger Electron Spectroscopy (XAES) studies provide evidence for a negative charge transfer at the CoPc/Au(111) interface from the substrate into the molecule facilitated by a considerable coupling between the d orbitals of the Co ion and the gold metal surface states [54, 55]. Moreover, the image of monolayer and multilayer CoPc is well known to reflect mixing between the metal substrate and cobalt ion half-filled dz2 orbital [56]. Mixing of metal orbitals and LUMO, when combined with the known Co–Au interaction of the adsorbed CoPc, may results in shifting the LUMO tunneling energy. In the absence of this mixing of the filled dz2 orbitals of copper and zinc ions with Au states, the LUMO tunneling energies may be less unaffected.

Experimental solution phase EAs based on reduction potentials are given for comparison in Table 1 and were obtained by equating the reference SCE potential to 4.71 eV relative to the vacuum level [24, 25]. Table 1 lists the measured values of the first electron affinity (EA1

OMTS) determined for a monolayers and aggregates of the phthalocyanines. Also given in Table 1 are the calculated solution (acetonitrile) and gas phase electron affinities, and the calculated HOMO–LUMO gap.

In the past, we have demonstrated that the ionization potentials and electron affinities for thin films on metal substrates are often closer in energy to values associated with solution-phase redox chemistry than they are to gas-phase values [25]. It has been previously noted that electron affinities (or ionization energies) and reduction (or oxidation) potentials in solution and in the solid state can differ from the gas-phase values by as much as 2 eV [25, 57, 58]. Redox potentials in the solid state are also expected to differ from those in solution. Moreover, there will be shifts in the potentials of a thin film, relative to that of a solid, due to interactions with the metal support and counter electrode, including image-charge effects. There may be an opposite signed shift, due to the absence of a covering layer of adsorbed molecules [59]. Another complication is the fact that electrochemical potentials are equilibrium values, and therefore reflect the energy associated with the formation of an ion in its equilibrium geometry. OMTS transitions, as discussed above, may occur so rapidly, that the ion is formed in an excited state — a vertical transition in the Franck–Condon sense. For a wide range of materials and film thickness (sub-monolayer to about 1 nm) studied to date, a fortuitous cancellation of polarization terms and differences between vertical and equilibrium affinities has resulted in many OMTS bands laying close to the positions predicted from electrochemistry. This correlation is especially good for unoccupied orbitals. OMTS bands associated with occupied orbitals generally lay deeper than predicted by solution phase

Fig. 6. IETS obtained from Al-AlOx-MPc-Pb tunnel junctions. Spectra were collected at 4 K

Fig. 7. IETS obtained from Al-AlOx-MPc-Pb tunnel junctions. Spectra were collected at 4 K

S1088424612004574.indd 7 3/19/2012 2:51:35 PM



electrochemical oxidation potentials, and closer to the gas-phase ionization potentials.

The near transferability of electrochemical values to thin film band positions for affinity levels, but not ionization levels, indicates that the polarization energy terms differ for these processes. This is a failure in the use of a simple model where it was assumed that only the sign of the polarization energy changed [25]. This failure is particularly large for porphyrins [13, 25] where ionization energies measured from a thin film are nearly identical to those reported from the gas phase [60]. These discrepancies also occur with UPS observations (which agree with the OMTS), suggesting that the problem is in the model, not the technique.

CONCLUSION

We presented measurements of the unoccupied molecular orbital energies of different phthalocyanines using OMTS. Unsubstituted, alkylated, and sulfonated metal phthalocyanines adsorbed as single molecules or aggregates on gold substrates and imbedded in aluminum-aluminum oxide-lead (Al-AlOx-Pb) devices were employed. We demonstrated that the behavior of the lowest unoccupied orbital energies of the Pc derivatives depend only weakly on their surface aggregation morphology and the metal ion present in the Pc core with the exception of Co(II). The LUMO in the d7 system shifts to higher energy relative to EAs of the d9 and d10 systems. This difference may be attributed to more significant coupling of the metal ion and p MOs in CoPc and the metal substrate surface states. The cofacial stacking configuration of the MTSPc and CuPc(OBu)8 aggregates resulted in a high degree of overlap between LUMO’s on adjacent molecules and facilitates efficient electron transport — giving rise to an intense tunneling signal.

The electron affinities of the phthalocyanines obtained from the tunneling spectra were compared with their LUMO energies in the gas phase (calculated) and the electrochemical potentials acquired from solution phase measurements. The measured electron affinities for Pc thin films on metal substrates are closer in energy to values associated with solution-phase redox chemistry than they are to gas-phase values. OMTS measured within either an STM or an M-I-M diode can be used to determine electron affinities for single molecules and their aggregates. Gas phase calculated EA energies for molecular systems, on the other hand are not good indicators of those values on a surface.

Acknowledgments

This material is based upon work supported by the National Science Foundation under grants CHE-1112156, CHE-1048600, and CHE-1058435.

REFERENCES

1. de la Torre G, Bottari G and Torres T. Struct. Bond. 2010; 135: 1–44.

2. Martinez-Diaz MV, de la Torre G and Torres T. Chem. Commun. 2010; 46: 7090–7108.

3. Elemans JAAW, van Hameren R, Nolte RJM and Rowan AE. Adv. Mater. 2006; 18: 1251.

4. Karimov KhS, Saleem M, Mahroof-Tahir M, Khan TA and Khan A. Physica E 2010; 43: 547–551.

5. Chen Y, Hanack M, Blau WJ, Dini D, Liu Y, Lin Y and Bai J. J. Mat. Sci. 2006; 41: 2169–2185.

6. Di Bella S, Dragonetti C, Pizzotti M, Roberto D, Tessore F and Ugo R. Top. Organomet. Chem. 2010; 28: 1–55.

7. Armstrong NR, Wang W, Alloway DM, Placencia D, Ratcliff E and Brumbach M. Macromol. Rapid Commun. 2009; 30: 717–731.

8. Brutting W. Physics of Organic Semiconductors, Wiley-VCH: Weinheim, 2005; p 23.

9. Sakamoto K and Ohno-Okumura E. Materials 2009; 2: 127–1179.

10. Liu J-Y, Lo P-C and Ng DKP. Struct. Bond. 2010; 35: 169–210.

11. Zhang Y, Cai X, Bian Y and Jiang J. Struct. Bond. 2010; 135: 275–322.

12. Snow AW. The Porphyrin Handbook, Vol. 17, Kadish KM, Smith KM and Guilard R. (Eds.) Academic Press: San Diego, CA, 2004; pp 130–179.

13. Friesen BA, Wiggins B, McHale JL, Mazur U and Hipps KW. J. Phys Chem. C 2011; 115: 3990–3999.

14. Kumaran N, Veneman PA, Minch BA, Mudalige A, Pemberton JE, O’Brien DF and Armstrong NR. Chem. Mater. 2010; 22: 2491–2501.

15. Schlettwein D, Jaeger NI and Oekermann TP. The Porphyrin Handbook 2003; 16: 247–283.

16. Krause S, Casu MB, Schöll A and Umbach E. New Journal of Physics 2008; 10: 1–16.

17. Petraki F, Peisert H, Biswas I and Chasse T. J. Phys. Chem. C 2010; 114: 17638–17643.

18. Scudiero L, Hipps KW and Barlow DE. J. Phys. Chem. B 2003; 107: 2903–2909.

19. Xiao J and Dowben PA. J. Phys.: Condensed Matter 2009; 21: 052001/1-052001/5.

20. Zahn DRT, Gavrila GN and Gorgoi M. Chem. Phys. 2006; 325: 99–112.

21. Lever ABP, Milaeva ER and Speier G. Phthalo-cyanines 1993; 3: 5–69.

22. Ahmida MM and Eichhorn SH. ECS Transactions 2010; 25: 1–10.

23. Nyokong T. Struct. Bond. 2010; 135: 45–88. 24. Mazur U and Hipps KW. J. Phys. Chem. B 1999;

103: 9721–9727. 25. Hipps KW. Handbook of Applied Solid State

Spectroscopy, Vij DR. (Ed.) Springer Verlag: 2006; pp 305–350.

S1088424612004574.indd 8 3/19/2012 2:51:35 PM



26. Hipps KW and Mazur U. J. Phys. Chem. 1993; 97: 7803–7814.

27. Mazur U and Hipps KW. Handbook of Vibra-tional Spectroscopy, Vol. 1, Chalmers J and Griffiths P. (Eds.) J. Wiley and Sons: 2001; pp 812–829.

28. Oshiro T, Backstrom A, Cumberlidge AM, Hipps KW, Pevovar SP, Bahr DF, Smieja J and Mazur U. J. Mater. Res. 2004; 19: 1461–1470.

29. Mazur U. Dekker Encyclopedia of Nanoscience and Nanotechnology 2007; 1: 1–10.

30. Hipps KW, Lu X, Wang XD and Mazur U. J. Phys. Chem. 1996; 100: 11207–11210.

31. Nishida KRA, Wiggins B, Hipps KW and Mazur U. J. Phys. Chem. C 2011; 115: 16305–16314.

32. Tomasi J, Mannucci B and Cammi R. Chem. Rev. 2005; 105: 2999–3093.

33. Miertus S, Scrocco E and Tomasi J. J. Chem. Phys. 1981; 55: 117–129.

34. Hurst HG and Ruppel W. Zeitschrift Fuer Naturforschung 1964; 19a: 573–579.

35. Lambe J and Jaklevic RC. Phys. Rev. 1968; 165: 821–832. These authors are the fathers of all tunneling spectroscopy.

36. Thomas DE and Klein JM. Rev. Sci. Instrum. 1963; 34: 920–924.

37. Jaklevic RC and Lambe J. Phys. Rev. Lett. 1996; 17: 1139–1140.

38. Tunneling Spectroscopy, Hansma PK. (Ed.) Plenum: New York, 1982.

39. Hamers RJ. J. Phys. Chem. 1996; 100: 13103–13120. 40. Moore AA and Weiss P. Annual. Rev. Anal. Chem.

2008; 1: 857–882. 41. Wahl P, Diekhöner L, Schneider MA and Kern K.

Rev. Sci. Instrum. 2008; 79: 043104/1-043104/4. 42. Giancarlo LC, Fang H, Avila L, Fine LW and Flynn

GW. J. Chem. Educ. 2000; 77: 66–71.

43. Hipps KW and Scudiero L. J. Chem. Ed. 2005; 82: 704–711.

44. Hamers RJ. Ann. Rev. Phys. Chem. 1989; 40: 531–559.

45. Lehmpuhl DW. J. Chem. Educ. 2003; 80: 478–479. 46. Hipps KW and Mazur U. J. Phys. Chem. B. 2000;

104: 4707–4710. 47. Scudiero L, Barlow DE, Mazur U and Hipps KW. J.

Am. Chem. Soc. 2001; 123: 4073–4080. 48. Barlow D, Scudiero L and Hipps KW. Langmuir

2004; 20: 4413–4421. 49. Rollmann LD and Iwamoto RT. J. Am. Chem. Soc.

1968; 90: 1455–1463. 50. Simon J and Andre J-J. Molecular Semiconductors,

Springer-Verlag: New York, 1985. 51. Ougha EA, Crebera KAM, Martin J and Stillman

MJ. Inorg. Chim. Acta 1996; 246: 361–369. 52. Yamashida N, Stevenson K and Mazur U.

Unpublished results. 53. Nishida KRA, Wiggins B, Hipps KW and Mazur U.

J. Porphyrins Phthalocyanines 2011; 15: 1–8. 54. Petraki F, Peisert H, Biswas I and Chasse T. J. Phys.

Chem. C 2010; 114: 17638–17643. 55. Petraki F, Papaefthimiou V and Kennou1 S. J.

Phys.: Conf. Ser. 2005; 10: 135–138. 56. Guo Q, Zang K, Liu C, Qin Z, Yu Y and Cao G.

Langmuir 2010; 26: 11804–11808. 57. Schlettwein D, Hesse K, Gruhn NE, Lee PA,

Nebesny KW and Armstrong NR. J. Phys. Chem. B 2001; 105: 4791–4800.

58. Knupfer M and Schwieger T. Appl. Surf. Sci. 2005; 252: 77–80.

59. Hill IG, Kahn A, Soos ZG and Pascal Jr RA. Chem. Phys. Lett. 2000; 327: 181–188.

60. Westcott BL, Gruhn N, Michelsen L and Lichtenberger D. J. Am. Chem Soc. 2000; 122: 8083–8089.

S1088424612004574.indd 9 3/19/2012 2:51:36 PM

electron affinity states of metal supported phthalocyanines … · 2017. 9. 6. · phthalocyanines...

Documents