4646
J. Phys. Chem. C 2010, 114, 4646–4651
Oxidation of Thiol Anchor Groups in Molecular Junction Devices: A Density Functional Theory Study Yun Hee Jang*,† and William A. Goddard III.‡ Department of Materials Science and Engineering, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea, and Materials and Process Simulation Center, California Institute of Technology, Pasadena, California 91125 ReceiVed: NoVember 13, 2009; ReVised Manuscript ReceiVed: January 14, 2010
Thiols (RSH) or thiolates (RS) are one of the most popular anchor groups used for attaching molecules to gold electrodes in molecular junction devices. Operation of these devices under oxidizing conditions may lead to oxidation of the thiolate anchor groups. Herein we investigate the plausibility of this process as a potential source of current fluctuation. Density functional theory calculations on various oxide derivatives of ethanethiolate on gold clusters (EtSOn/Au13; n ) 0-3) suggest that oxidation of thiolate anchor groups (Au-S) into sulfoxides (Au-SO) should be unlikely under ambient conditions. Under oxidizing conditions, sulfoxides can form and oxidize further into sulfinates (Au-SO2) and sulfonates (Au-SO3) favorably via oxygen transfer from surface oxides or other active oxygen species. Nonequilibrium Green’s function calculations on model devices with thiolate, sulfoxide, and sulfonate anchor groups suggest that thiolates show essentially the same insulating current-voltage characteristics before and after oxidation. However, oxidation of thiolates into sulfonates can increase the length of the electron transfer pathway (that is, the molecule), and this type of oxidation-induced change (molecular lengthening, SAM thickening, and possibly SAM disordering) can affect the robustness of the molecular junction devices. 1. Introduction In molecular electronics involving bottom-up manufacturing of single-molecule or self-assembled monolayer (SAM) junction devices,1–7 one of the essential components is an anchor group that attaches each molecule to electrodes at the contact. Understanding the behavior of the anchor group is crucial for the development of robust molecular junctions. Anchor groups such as thiols, sulfides, and disulfides are popular alligator clips to make robust SAMs on gold surfaces.1–9 However, molecular junction devices often operate under a wide range of bias voltages (up to (2 V) or under redox conditions, and these conditions can lead to redox processes involving Au-S bonds at the contact.9–11 Formation or breakage of the disulfide linkage (-SS-) can be induced,12–14 and the thiolate group (-S) can be oxidized into oxide derivatives such as sulfoxide (-SO), sulfinate (or sulfone; -SO2), and sulfonate (-SO3) groups.15–22 In our previous study23 we suggested that the formation of the disulfide bonds would not affect the switching behavior of the junction at moderate bias voltages, unless it induces a thermal desorption. Then, how about the oxide formation? Would the conversion between a thiolate group and a sulfonate group at the contact alter the electron transport through the junctions and affect the robustness of the junction? To test this idea, we carried out quantum mechanics (density functional theory; DFT) calculations on model junctions which have a thiolate anchor group and its oxide derivative (sulfoxide, sulfinate, and sulfonate) to establish (1) whether these oxide derivatives can form favorably on gold surfaces and (2) how this oxide formation affects the current-voltage (I-V) char* To whom correspondence should be addressed. Phone: +82-62-9702323. Fax +82-62-970-2304. E-mail
[email protected]. † Gwangju Institute of Science and Technology. ‡ California Institute of Technology.
acteristics of the junction. We first calculated their adsorption structures and binding energies on gold clusters which represent Au(111) surfaces (Section 2). The most stable adsorption structures were employed to build model junctions comprised of a SAM of oxidized thiolate group sandwiched between two Au(111) slabs, which were used for the I-V curve calculations (Section 3). 2. EtSOn/Au13 Clusters (n ) 0-3): Oxidation Energies 2.1. Calculation Details. The geometries of ethanethiolate (CH3CH2S; EtS) and its oxide derivatives (EtSO, EtSO2, and EtSO3) were optimized both in the gas phase and on finite gold clusters. A 3-layer 13-atom gold cluster (Au13) was employed to represent the Au(111) surface (or a protruding small island on corrugated surfaces) (Figure 1). All the Au atoms were kept fixed at their bulk positions (with the nearest-neighbor Au-Au distance of 2.88 Å)24 throughout the geometry optimization of bare and adsorbed clusters. At each optimum geometry, the binding energy (BE) was evaluated as [E(adsorbate) + E(Au13) - E(adsorbate/Au13)] (Table 1). While the singlet ground states of EtSOn/Au13 were calculated with the spin-restricted DFT, the doublet ground states of the gas-phase EtSOn and the bare Au13 cluster were calculated using the spin-unrestricted DFT. The triplet ground state of O2 in the gas phase and the doublet ground state of an oxygen-covered Au13 cluster, which are involved in the oxidation process (Tables 2-4 and Figure 2), were also calculated using the spinunrestricted DFT. The B3LYP (hybrid) functional of DFT25–29 was used for all the geometry optimizations. A PBE (GGA) functional of DFT30,31 was also used for single-point energy calculations on the optimized structures in order to estimate the DFT functional dependence of the calculated results. The Jaguar v6.5 quantum chemistry software was used for all the calculations.32 The core electrons of Au atoms were replaced with the
10.1021/jp910801v 2010 American Chemical Society Published on Web 02/24/2010
Oxidation of Thiol Anchor Groups
J. Phys. Chem. C, Vol. 114, No. 10, 2010 4647
Figure 1. Top views (first row) and side views (second and third rows) of the optimized structures of ethanethiolate and its oxide derivatives on Au13 clusters. The positions of the S and O atoms above the Au surface [h(S) and h(O)] as well as their distances from the nearestneighbor Au atoms [r(Au-S) and r(Au-O)] are shown together. Other plausible adsorption structures identified for the thiolate (but 7.5 kcal/ mol less stable than 1) and for the sulfinate (as stable as 3) are not shown but their features are described in the text (Section 2.2). Color code: H (black), C (gray), O (red), S (yellow), and Au (gold).
TABLE 1: Binding Energy of EtSOn on Au13a B3LYP BE (kcal/mol) 1 2 3 4
(EtS) (EtSO) (EtSO2) (EtSO3) a
PBE
LACVP**
LACV3P**
LACVP**
LACV3P**
50.9 25.2 37.0 70.2
48.6 23.5 34.5 66.7
62.4 35.8 41.8 70.3
59.7 32.0 37.4 65.2
E(EtSOn) + E(Au13) - E(EtSOn/Au13).
Hay-Wadt small-core LACVP effective core potential and only the outermost 19 electrons were treated explicitly using the LACVP basis set.33,34 For the other atoms (H, C, O, and S), all electrons were treated explicitly with the 6-31G** basis set. An extended triple-ζ basis set (LACV3P** and 6-311G**)32 was also used for single-point energy calculations on the optimized structures in order to estimate the basis set dependence of the calculation results. 2.2. Adsorption Structure. Figure 1 shows the most stable adsorption structures of EtS and its oxide derivative (EtSOn; n ) 0-3) on the Au13 clusters, which were optimized at the B3LYP/LACVP** level. The most stable adsorption state of EtS is 1, where the sulfur atom sits on a 2-fold bridge (edge) site (2.0 Å above the Au surface) and the C-S bond is tilted significantly from the surface normal. The adsorption state where the sulfur atom sits on a 3-fold hollow (hcp) site and the C-S bond stands along the surface normal (not shown here) is less stable than 1 [∆(BE) ) 7.5 kcal/mol]. It has been known35 that the sulfur atom of an alkanethiolate on Au(111) is positioned on a 3-fold hollow site36 (fcc37,38) or a 2-fold bridge site39 or somewhere between them.40–42 The preference is shifted toward the 2-fold bridge (edge) site on our cluster model (Au13) due to the low coordination number of this site, but this is not entirely a size
effect because the adsorption on the 1-fold on-top (corner) site with the lowest coordination number (not shown here) is less stable than that on the edge site (by 15.6 kcal/mol at the B3LYP/ LACVP** level). The tilt of the C-S bond from the surface normal of this small cluster is more significant than the value (∼30°) reported at the 1 ML coverage overage (θ ) 1/3).35 The most stable adsorption state of EtSO is 2, where the sulfur and oxygen atoms sit on the on-top sites along the edge of the Au cluster and the C-S bond is still tilted significantly from the surface normal. This (O,S)-bidendate adsorption geometry is consistent with experimental findings on similar systems, dimethyl sulfoxide on Au(100) and Au(111).43,44 On the other hand, the most stable adsorption states of EtSO2 and EtSO3, 3 and 4, have the C-S bond pointing away from the surface along the surface normal. The oxygen atoms sit on the on-top sites (2.2-2.3 Å above the Au surface) making direct covalent bonds with Au atoms, while the sulfur atom is positioned on the 3-fold hollow sites further away from the surface (2.6-2.7 Å above the Au surface). This bidentate (or tridentate) O-coordination geometry has been suggested from previous studies on related systems, benzenesulfinate (PhSO2-) and benzenesulfonate (PhSO3-) on gold,17 sulfate (SO42-) on Au(111),45,46 and benzenesulfinate on silver.47 [The intermediate oxidation product, EtSO2/Au13, has another adsorption structure 3′ (not shown here) that is as stable as 3. This structure 3′ also has the bidentate O-coordination geometry (as in 3), but the whole adsorbate (EtSO2) has rotated around the line connecting the two oxygen atoms until the OSO plane is perpendicular to the Au(111) surface, the sulfur atom is positioned as high as 3.0 Å away from the surface, and the C-S axis is significantly tilted away from the surface normal (as in 1 and 2). We expect that these two states (3 and 3′) may coexist on Au(111) and fluctuate between each other by the pivoting motion.] On the basis of the vertical position of the sulfur atom from the surface and the tilt angle of the C-S bond, we expect that the SAM of alkanesulfinate and alkanesulfonate, if they stay close-packed and ordered, should be slightly thicker than the SAM of alkanethiolate with the same number of carbon atoms. 2.3. Binding Energy. The BE of EtSOn (n ) 0-3) on the Au13 cluster was calculated for the optimized structure at four different calculation levels [differing in the basis set (LACVP** and LACV3P**) and the DFT functional (B3LYP and PBE)]. They are listed in Table 1. Even though the BE’s calculated with the smaller basis set (LACVP**) are slightly higher than those calculated with the larger basis set (LACV3P**), the difference is essentially constant over all the compounds and thus the relative stabilities of the species exhibit essentially no basis set dependence. On the other hand, the DFT functional dependence of BE is rather significant. The BE’s calculated with the PBE functional are consistently higher than those calculated with the B3LYP functional, as was the case in our previous study.23 Discrepancy between the two BE’s decreases dramatically with increasing number of oxygen atoms or with decreasing interaction between Au and S [EtS (11.5) > EtSO (10.6) > EtSO2 (4.8) > EtSO3 (0.1) with LACVP**; EtS (11.1) > EtSO (8.5) > EtSO2 (2.9) > EtSO3 (-1.5) with LACV3P**]. A common feature found at all the calculation levels is that all the compounds are stable on the surface with the BE increasing in the following order: EtSO (2) < EtSO2 (3) < EtS (1) < EtSO3 (4). Sulfoxide (EtSO) was found to be the least stable among the four species, and it is consistent with the reports that only sulfinates and sulfonates, not sulfoxides, have been detected as the oxidation products of thiolates on
4648
J. Phys. Chem. C, Vol. 114, No. 10, 2010
Jang and Goddard, III
TABLE 2: Oxidation (or Oxide Formation) Energy on Au13a B3LYP ∆E (kcal/mol) 1(EtS) f 2(EtSO) 2(EtSO) f 3(EtSO2) 3(EtSO2) f 4(EtSO3) a
PBE
LACVP**
LACV3P**
LACVP**
LACV3P**
-8.5 -29.9 -45.7
-7.7 -27.7 -43.0
-11.0 -29.7 -45.2
-8.7 -27.7 -43.1
LACVP**
LACV3P**
LACVP**
LACV3P**
-54.7 -76.2 -92.0
-55.7 -75.6 -90.9
-43.0 -61.6 -77.1
-42.3 -61.2 -76.6
LACVP**
LACV3P**
LACVP**
LACV3P**
-4.0 -25.4 -41.2
-5.4 -25.3 -40.7
2.1 -16.6 -32.1
5.1 -13.9 -29.3
E(EtSOn+1/Au13) - E(EtSOn/Au13) - 1/2E[O2(g)].
TABLE 3: Oxygen Transfer Energy (from Ozone) on Au13.a B3LYP ∆E (kcal/mol) 1(EtS) f 2(EtSO) 2(EtSO) f 3(EtSO2) 3(EtSO2) f 4(EtSO3) a
PBE
E(EtSOn+1/Au13) + E[O2(g)] - E(EtSOn/Au13) - E[O3(g)].
TABLE 4: Oxide Transfer Energy (from Gold Oxide) on Au13a B3LYP ∆E (kcal/mol) 1(EtS) f 2(EtSO) 2(EtSO) f 3(EtSO2) 3(EtSO2) f 4(EtSO3) a
PBE
E(EtSOn+1/Au13) + E(Au13) - E(EtSOn/Au13) - E(O/Au13).
Figure 2. Top views (first row) and side views (second and third rows) of a model process where a thiolate anchor group (EtS/Au13) is oxidized into its first oxide derivative (EtSO/Au13) through a transfer of surface oxide [EtSOn/Au13 + O/Au13 f EtSOn+1/Au13 + Au13 (n ) 0)].
gold.10,15–19,22 The highest BE of sulfonate (EtSO3) among the four species is consistent with the experimental findings that a prolonged exposure to air converts sulfinates to sulfonates, the major oxidation product,17–19 but also surprising because many previous studies have suggested that sulfonates should be less stable than thiolates on gold surface according to their exchange behavior in aqueous or enthanolic thiol solutions (sulfonates are replaced by thiolates).16–19,21 This exchange behavior, however, does not contradict our result, because it most likely comes from the balance between their relative stability on gold (BE) and their relative stability in solution (solubility) which was not considered in the current calculation.17,22 2.4. Oxide Formation Energy under Ambient Conditions. The oxidation energy of EtSOn/Au13 (n ) 0-2) under ambient conditions can be defined as the energy change during the following oxide formation process: EtSOn/Au13 + 1/2O2 (g) f EtSOn+1/Au13 (n ) 0-2). These energies were estimated at the four different calculation levels and are listed in Table 2.
The calculations at all the levels consistently suggest that the oxidation along this pathway (the formation of EtSOn+1 from EtSOn by capturing gaseous O2) is always exothermic. The exothermicity increases with increasing degree of oxidation (-8.5, -29.9, and -45.7 kcal/mol to form EtSO, EtSO2, and EtSO3, respectively, at the B3LYP/LACVP** level). However, the exothermicity, especially for the formation of EtSO, is not high enough to compensate the activation barrier of O2 dissociation (∼118 kcal/mol ) ∼59 kcal/mol per O).48 We therefore expect that the formation of EtSO (and in turn EtSO2 and EtSO3) should be unlikely or extremely slow under ambient conditions. 2.5. Oxygen Transfer Energy under Oxidizing Conditions. In fact, thiolates on gold are known to oxidize when exposed to air (even in the absence of light).10,15,16,18,19,22 Ozone, either atmospheric or UV-photosynthesized, is believed to be the primary cause of the oxidation.19–22 Therefore, oxidation of EtSOn/Au13 (n ) 0-2) in the presence of ozone was modeled by the process of oxygen transfer from ozone [EtSOn/Au13 + O3 (g) f EtSOn+1/Au13 + O2 (g)]. The energy changes during these processes were estimated at the four different calculation levels and are listed in Table 3. All the processes are indeed very exothermic [well enough to compensate the tiny activation barrier of O3 (∼0.1 kcal/mol)49]. This condition should lead to a rapid and complete oxidation of thiolates into sulfonates.15–22 Although gold is one of the most resistant elements against oxidation, gold oxides can form by anodic oxidation in an electrochemical cell or by exposure to highly reactive chemical environment such as an oxygen plasma employed to remove polymer patterns in a nanolithography process.50,51 Such residual oxygens (or oxides) on gold surface can be a source of active oxygens.51 Oxidation of the thiolate anchor groups via transfer of these surface oxygens was modeled by the following processes: EtSOn/Au13 + O/Au13 f EtSOn+1/Au13 + Au13 (n ) 0-2). See Figure 2 for the first step (n ) 0). The energy changes during these processes were estimated at the four different calculation levels and are listed in Table 4.
Oxidation of Thiol Anchor Groups While the calculated results show little basis set dependence, the DFT functional dependence is again quite significant. The calculations with the B3LYP functional predict exothermic formation of each oxide derivative (2-4) via this process, but those with the PBE functional predict endothermic formation of EtSO (2) followed by exothermic formation of EtSO2 (3) and EtSO3 (4). However, the energy needed to oxidize EtS (1) into EtSO (2) is moderate (∆E ) 2-5 kcal/mol) and should be easily compensated by the next favorable oxidation steps (∆E ) -14 and -29 kcal/mol) as long as there are plenty of surface oxygens (high O2 pressure) and the oxygen diffusion on the surface is sufficiently fast (high temperature). We therefore expect that the thiolate anchor group could be slowly oxidized into EtSO and then rapidly into EtSO2 and EtSO3, the major oxidation products, at elevated temperatures. 2.6. Electronic Structure: EtS/Au13 versus EtSO3/Au13. Figure 3 shows the electronic structure of EtS/Au13 and its final oxidation product, EtSO3/Au13, which was calculated at the B3LYP/LACVP** level. The density of states (DOS; Figure 3a,b) shows the energy level distribution of each system, and the projected DOS (PDOS; Figure 3a′,b′) shows the projection of the DOS onto (or the contribution from) each component of the system.52,53 The DOS curves (Figure 3a,b) and the frontier molecular orbitals (MO) of free EtS and free EtSO3 are shown together. Even though alkanethiolate and alkanesulfonate are different in the electronic structure and the interaction with gold surface, they share a few common features on the Au cluster, which can be related to electron transport through them in a junction. First, the alkyl chains (the molecular part in the junction) have no contribution to the DOS near the Fermi level of the system (EF ≈ -5 eV). Their contribution appears only in the energy levels below -9 eV. See the PDOS(Et) curves (the third panel from the top) for the adsorbed states and the MO’s at -9.1 and -9.2 eV (the bottom panel) for the free states. Second, the singly occupied MO’s of the free EtS and EtSO3, which correspond to the energy levels of -4.1 and -6.2 eV, respectively (see the bottom panel), become completely delocalized upon adsorption on Au13, indicating strong interaction with the surface [see the PDOS(S,O) curves on the fourth panel from the top]. As a result, the contribution of the anchor part to the DOS near EF of the system is negligible in both cases. 3. Model Junction Devices: I-V Curves 3.1. Models. The optimized structures of EtS/Au13 and EtSO3/Au13 were employed to build periodic models of molecular junction devices containing thiolate anchor groups before and after oxidation. To build the alligator clip to the bottom electrode, the structure of EtS (or EtSO3) optimized on Au13 in Section 2 was positioned on an equivalent adsorption site (a 2-fold bridge site for the sulfur atom of EtS and three 1-fold on-top sites for the oxygen atoms of EtSO3) of a three-layer Au(111) slab in the hexagonal (3×3)R30° unit cell, which leads to one anchor group per three surface Au atoms (coverage θ ) 1/3).35,39,54–56 Even though sulfonates may form a SAM with a coverage lower than the coverage for thiolates (θ ) 1/3), we keep the coverage constant for both systems for a fair comparison of intrinsic anchor properties. The alligator clip to the bottom electrode was then replicated in a symmetrical fashion to build the alligator clip to the top electrode, which was again modeled by a three-layer Au(111) slab. The top and bottom layers were connected to each other via alkyl chains whose structure was optimized after the connection (Figure 4). The model devices constructed in this manner, one with a butyl dithiolate SAM (5) and the other with a butyl disulfonate
J. Phys. Chem. C, Vol. 114, No. 10, 2010 4649
Figure 3. Electronic structure of EtSOn/Au13 (n ) 0 and 3) near the Fermi level (EF indicated by blue arrows): (top) DOS of EtSOn/Au13, (middle) PDOS of each part (Au, Et, S, O, and EtSOn), and (bottom) DOS of free EtSOn along with the frontier MO’s (whose different phases are represented in red and blue) and their energy levels (in eV).
SAM (6), have different sizes of unit cells: 5.0 Å × 5.0 Å × 22.3 Å for butyl dithiolate (5) and 5.0 Å × 5.0 Å × 23.6 Å for butyl disulfonate (6). Since the butyl dithiolate in 5 has been fully extended to minimize the difference between the two devices in the tilt of the molecular axis from the surface normal (based on the experimental observations that the dominant electron transport mechanism in molecular junctions is throughbonding tunneling, in which the length of the electron transfer pathway is determined by the length of the molecule rather than the separation between the electrodes2,6,57,58), the difference in the cell height (22.3 Å versus 23.6 Å) comes mostly from the difference in the molecular length (which is the distance between the two contact points of a molecule to the top and the bottom electrodes, that is, the S-to-S distance of 7.0 Å for 5 versus the O-to-O distance of 9.1 Å for 6). Since the I-V characteristic is sensitive to the length of the electron transfer pathway determined by the molecular length2,6,57,58 (and possibly to the separation between the two electrodes), it might be important to keep the molecular length and the junction gap constant for the comparison of the intrinsic I-V charac-
4650
J. Phys. Chem. C, Vol. 114, No. 10, 2010
Jang and Goddard, III
Figure 5. I-V curves of the model junction devices (5-8) depicted in Figure 4.
Figure 4. A set of model junction devices (5-6) with thiolate anchor groups before and after oxidation, which are modeled by a SAM of butyl dithiolate (5) and butyl disulfonate (6), respectively, sandwiched by three-layer Au(111) slabs in periodic (3×3)R30° unit cells (coverage θ ) 1/3). In another set of model junction devices (6-8), two extra CH2 groups were inserted between the EtS and EtSO anchor groups to make a SAM of hexyl dithiolate (7) and hexyl disulfoxide (8), respectively, in order to make all the models have the same unitcell size, the same junction gap, and similar molecular lengths.
teristics of different anchor groups. Thus we built another set of model junction devices 6-8 with similar junction gap sizes and molecular lengths. Two extra units of CH2 groups were introduced when connecting the top and bottom EtS anchor groups to build the device with a hexyl dithiolate SAM (7). The device with the intermediately oxidized sulfoxide anchor groups (8) was also built in the same manner. These three model junction devices with butyl disulfonate, hexyl dithiolate, and hexyl disulfoxide (6-8) have the same unit-cell size of 5.0 Å × 5.0 Å × 23.6 Å, the same junction gap size of 16.5 Å, and essentially the same molecular lengths [9.1 Å for the O-to-O distance in 6, 9.2 Å for the S-to-S distance in 7, and 9.6 Å for the O-to-O distance in 8]. 3.2. Calculation Details. The I-V curves were calculated with use of the Landauer-Bu¨ttiker formalism and the nonequilibrium Green’s function (NEGF) approach59–61 coupled with the LCAO (linear combination of atomic orbitals) type of DFT as implemented in SeqQuest.60,62,63 We employed a non-selfconsistent NEGF approach that is valid only in the low-bias regime.61 The PBE functional of DFT was used.30,31 The core electrons of each atom were replaced by norm-conserving pseudopotentials64–67 and the outer electrons were described by a double-ζ-plus-polarization (DZP) basis set.68 A smaller singleζ-plus-polarization (SZP) basis set was used for Au. The pseudopotentials and basis sets for Au were taken as optimized and tested in closely related previous studies.23,61,69,70 The 33 × 33 × 150 (and 142) points were used for the real space grid for the devices 6-8 (and 5) (∼0.157 Å/grid). The 3 × 3 × 1 Monkhorst-Pack grid was used for the Brillouin zone sampling. 3.3. I-V Curves. The I-V curves calculated for the four model devices are shown in Figure 5. First comparing the I-V curves of the devices 6-8 with the same junction gap and similar molecular lengths [9.1 Å (6), 9.2 Å (7), and 9.6 Å (8)], we notice that the zero-bias conductance G ()dI/dV at 0 V) was calculated as 214, 122, and 39 nS for 6, 7, and 8, respectively. The current I was calculated as 105, 60, and 19
nA at 0.5 V and as 533, 287, and 86 nA at 2 V for 6, 7, and 8, respectively. The conductance G of 122 nS (or 0.0016 G0, where G0 is 77.4 µS) calculated for the device 7 with the hexyl dithiolate is in good agreement with experimental observations of 0.0012-0.002 G0.2,58 Compared to this resting thiolate state (7), lower conductance was calculated for the intermediately oxidized SO state (8) while higher conductance was calculated for the completely oxidized SO3 state (6). Moreover, the variations between them are quite small (with the ratios of 2-3). It appears that this small variation of conductance (G ) 214, 122, and 39 nS) comes from the small variation of the molecular length (L ) 9.1, 9.2, and 9.6 Å) rather than from the change in the intrinsic nature of the anchor group due to oxidation [They follow the relationship G ∝ exp(-βL) with high correlation (R2 ) 0.98)]. That is, the three junction devices with different anchor groups show essentially the same insulating characteristics at least over a moderate range of bias voltages up to (2 V. This should be because there is essentially no contribution from the molecule and the anchor parts in the energy levels near the Fermi leVel, more specifically within the window of [EF - 1 eV, EF + 1 eV], of the junction systems in both cases (see Figure 3). Since the bias voltage typically stays between (2 V, this result implies that the thiolate anchor group would stay intrinsically insulating during the redox processes. However, the oxidation of the thiolate anchor group can still affect the robustness of the device by inducing a change in the SAM structure, as can be seen by comparing 5 and 6, the devices with butyl dithiolate and disulfonate, respectively. In the resting state 5, the dithiolates are bonded to the two electrodes through the sulfur atoms and the molecular length (S-to-S) is estimated as 7.0 Å. In the oxidized state 6, the sulfur atoms are pushed away from the surfaces and the disulfonates are bonded to the electrodes through the oxygen atoms with the molecular length (O-to-O) estimated as 9.1 Å. This increase of the molecular length (and also of the SAM thickness) induces an 11-fold decrease of the zero-bias conductance G (from 2389 ns of 5 to 214 nS of 6). Moreover, the oxidation of thiolate may decrease the SAM density (from the surface coverage of 1/3 of thiolates) due to the tridentate O-coordination of the sulfonate to the surface, and a disordering has been observed in the SAM of sulfonates.16–19,21 This type of change in the SAM structure can also affect the robustness of the molecular junction with thiolate anchor groups during redox processes. 4. Conclusion According to our DFT calculations, oxidation of thiolate (S) groups leading to sulfoxide (SO) derivatives is not favorable on Au(111) surfaces in ambient conditions. However, under oxidizing conditions, the sulfoxide derivatives can be formed owing to active oxygens (surface oxide or ozone) and then oxidized further into sulfinate (SO2) and sulfonate (SO3) derivatives as long as active oxygen species are supplied. Even
Oxidation of Thiol Anchor Groups though the thiolate anchor groups show essentially the same insulating I-V characteristics before and after such oxidation, an oxidation-induced molecular lengthening and SAM thickening (and possibly SAM disordering) can result in hindered electron transport through the molecular junction. Acknowledgment. This work was supported by PIMS (GIST, Korea), NRF (Korea), and MARCO-FENA. In addition, the facilities used were supported by KISTI (Korea), ONR-DURIP, and ARO-DURIP. References and Notes (1) McCreery, R. L.; Bergren, A. J. AdV. Mater. 2009, 21, 1. (2) Chen, F.; Hihath, J.; Huang, Z.; Li, X.; Tao, N. J. Annu. ReV. Phys. Chem. 2007, 58, 535. (3) Lindsay, S. M.; Ratner, M. A. AdV. Mater. 2007, 19, 23. (4) Tao, N. J. Nat. Nanotechnol. 2006, 1, 173. (5) McCreery, R. L. Chem. Mater. 2004, 16, 4477. (6) Salomon, A.; Cahen, D.; Lindsay, S.; Tomfohr, J.; Engelkes, V. B.; Frisbie, C. D. AdV. Mater. 2003, 15, 1881. (7) Nitzan, A.; Ratner, M. A. Science 2003, 300, 1384. (8) Ulman, A. Chem. ReV. 1996, 96, 1533. (9) Gooding, J. J.; Mearns, F.; Yang, W.; Liu, J. Electroanalysis 2003, 15, 81. (10) Dong, Y.; Abaci, S.; Shannon, C.; Bozack, M. J. Langmuir 2003, 19, 8922. (11) Antonello, S.; Holm, A. H.; Instuli, E.; Maran, F. J. Am. Chem. Soc. 2007, 129, 9836. (12) Howie, J. K.; Houts, J. J.; Sawyer, D. T. J. Am. Chem. Soc. 1977, 99, 6223. (13) Fujihara, H.; Nakai, H.; Yoshihara, M.; Maeshima, T. Chem. Commun. 1999, 737. (14) Maksymovych, P.; Yates, J. T., Jr. J. Am. Chem. Soc. 2006, 128, 10642. (15) Li, Y.; Huang, J.; McIver, R. T., Jr.; Hemminger, J. C. J. Am. Chem. Soc. 1992, 114, 2428. (16) Tarlov, M. J.; Newman, J. G. Langmuir 1992, 8, 1398. (17) Garrell, R. L.; Chadwick, J. E.; Severance, D. L.; McDonald, N. A.; Myles, D. C. J. Am. Chem. Soc. 1995, 117, 11563. (18) Scott, J. R.; Baker, L. S.; Everett, W. R.; Wilkins, C. L.; Fritsch, I. Anal. Chem. 1997, 69, 2636. (19) Schoenfisch, M. H.; Pemberton, J. E. J. Am. Chem. Soc. 1998, 120, 4502. (20) Zhang, Y.; Terrill, R. H.; Tanzer, T. A.; Bohn, P. W. J. Am. Chem. Soc. 1998, 120. (21) Poirier, G. E.; Herne, T. M.; Miller, C. C.; Tarlov, M. J. J. Am. Chem. Soc. 1999, 121, 9703. (22) Yang, G.; Amro, N. A.; Starkewolfe, Z. B.; Liu, G. Langmuir 2004, 20, 3995. (23) Jang, Y. H.; Goddard, W. A., III. J. Phys. Chem. C 2008, 112, 8715. (24) Maeland, A.; Flanagan, T. B. Can. J. Phys. 1964, 42, 2364. (25) Slater, J. C. Quantum Theory of Molecules and Solids, The SelfConsistent Field for Molecules and Solids, Vol. 4; McGraw-Hill: New York, 1974. (26) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (27) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200. (28) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (29) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200. (30) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865.
J. Phys. Chem. C, Vol. 114, No. 10, 2010 4651 (31) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1997, 78, 1396. (32) Jaugar 6.5; Schrodinger Inc., Portland, OR, 2005. (33) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299. (34) Melius, C. F.; Olafson, B. D.; Goddard, W. A., III. Chem. Phys. Lett. 1974, 28, 457. (35) Vericat, C.; Vela, M. E.; Salvarezza, R. C. Phys. Chem. Chem. Phys. 2005, 7, 3258. (36) Sellers, H.; Ulman, A.; Shnidman, Y.; Eilers, J. E. J. Am. Chem. Soc. 1993, 115, 9389. (37) Groenbeck, H.; Curioni, A.; Andreoni, W. J. Am. Chem. Soc. 2000, 122, 3839. (38) Yourdshahyan, Y.; Zhang, H. K.; Rappe, A. M. Phys. ReV. B 2001, 63, 081405. (39) Vargas, M. C.; Giannozzi, P.; Selloni, A.; Scoles, G. J. Phys. Chem. B 2001, 105, 9509. (40) Hayashi, T.; Morikawa, Y.; Nozoye, H. J. Chem. Phys. 2001, 114, 7615. (41) Akinaga, Y.; Nakajima, T.; Hirao, K. J. Chem. Phys. 2001, 114, 8555. (42) Gottschalck, J.; Hammer, B. J. Chem. Phys. 2002, 116, 784. (43) Schroter, C.; Roelfs, B.; Solomun, T. Surf. Sci. 1997, 380, L441. (44) Ikemiya, N.; Gewirth, A. A. J. Phys. Chem. B 2000, 104, 873. (45) Patrito, E. M.; Olivera, P. P.; Sellers, H. Surf. Sci. 1997, 380, 264. (46) Simeone, F. C.; Kolb, D. M.; Venkatachalam, S.; Jacob, T. Angew. Chem., Int. Ed. 2007, 46, 8903. (47) Cao, Y.; Li, Y.-S.; Tseng, T.-L.; Desiderio, D. M. Spectrochim. Acta, Part A 2001, 57, 27. (48) Lide, D. R. CRC Handbook of Chemistry and Physics, 87th ed.; CRC Press: Boca Raton, FL, 2006-2007. (49) Siebert, R.; Schinke, R.; Bittererova, Phys. Chem. Chem. Phys. 2001, 3, 1795. (50) Koslowski, B.; Boyen, H.-G.; Wilderotter, C.; Kastle, G.; Ziemann, P.; Wahrenberg, R.; Oelhafen, P. Surf. Sci. 2001, 475, 1. (51) Yan, C.; Golzhauser, A.; Grunze, M. Langmuir 1999, 15, 2414. (52) Hoffmann, R. Solids and Surfaces: a Chemist’s View of Bonding in Extended Structures; VCH: New York, 1988. (53) Jang, Y. H.; Hwang, S.; Kim, Y.-H.; Jang, S. S.; Goddard, W. A., III. J. Am. Chem. Soc. 2004, 126, 12636. (54) Widrig, C. A.; Alves, C. A.; Porter, M. D. J. Am. Chem. Soc. 1991, 113, 2805. (55) Chidsey, C. E. D.; Loiacono, D. N. Langmuir 1990, 6, 682. (56) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546. (57) Haiss, W.; van Zalinge, H.; Higgins, S. J.; Bethell, D.; Hobenreich, H.; Schiffrin, D. J.; Nichols, R. J. J. Am. Chem. Soc. 2003, 125, 15294. (58) Xu, B.; Tao, N. J. Science 2003, 301, 1221. (59) Datta, S. Quantum Transport: Atom to Transistor; Cambridge University Press: Cambridge, UK, 2005. (60) Williams, A. R.; Feibelman, P. J.; Lang, N. D. Phys. ReV. B 1982, 26, 5433. (61) Kim, Y.-H.; Tahir-Kheli, J.; Schultz, P. A.; Goddard, W. A., III. Phys. ReV. B 2006, 73, 235419. (62) Schultz, P. A. SeqQuest Electronic Structure Code (Sandia National Laboratories; http://dft.sandia.gov/Quest). (63) Feibelman, P. J. Phys. ReV. B 1987, 35, 2626. (64) Hamann, D. R. Phys. ReV. B 1989, 40, 2980. (65) Redondo, A.; Goddard, W. A., III.; McGill, T. C. Phys. ReV. B 1977, 15, 5038. (66) Troullier, N.; Martins, J. L. Phys. ReV. B 1991, 43, 1993. (67) Fuchs, M.; Scheffler, M. Comput. Phys. Commun. 1999, 119, 67. (68) Feibelman, P. J. Phys. ReV. B 1988, 38, 1849. (69) Kim, Y.-H.; Jang, S. S.; Jang, Y. H.; Goddard, W. A., III. Phys. ReV. Lett. 2005, 94, 156801. (70) Kim, Y.-H.; Jang, S. S.; Goddard, W. A., III. J. Chem. Phys. 2005, 122, 244703.
JP910801V
