Diamond & Related Materials 55 (2015) 131–138

Contents lists available at ScienceDirect

Diamond & Related Materials

j ourna l homepage: www.e lsev ie r .com/ locate /d iamond

Surface immobilization of Mo6I8 octahedral cluster cores onfunctionalized amorphous carbon using a pyridinecomplexation strategy

C. Godet a,⁎, S. Ababou-Girard a, B. Fabre b, Y. Molard b, A.B. Fadjie-Djomkam a, S. Députier b,M. Guilloux-Viry b, S. Cordier b,⁎a Institut de Physique de Rennes, CNRS UMR 6251, Université de Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, Franceb Institut des Sciences Chimiques de Rennes, CNRS UMR 6226 CNRS, Université de Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, France

⁎ Corresponding authors.E-mail addresses: christian.godet@univ-rennes1.fr (C.

Stephane.cordier@univ-rennes1.fr (S. Cordier).

http://dx.doi.org/10.1016/j.diamond.2015.03.0110925-9635/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 22 January 2015Received in revised form 28 February 2015Accepted 9 March 2015Available online 23 March 2015

Keywords:Tetrahedral amorphous carbonPulsed laser deposition (PLD)Bonding configurationsSurface characterizationElectrical properties

Tetrahedral amorphous carbon (a-C) films have been grown by pulsed laser deposition to investigate a liquidphase process for surface immobilization of electroactive [Mo6Ii8]4+ transition metal cluster cores using a com-plexation reaction with a pyridine-terminated alkyl monolayer covalently bonded to the a-C surface (PyS–alkyl/a-C). These films are stable against thermally-assisted grafting of alkene molecules and the covalent C\C inter-face provides a robust monolayer/a-C assembly. Octahedral [Mo6I

i8]

4+ cluster cores with iodine inner ligandsand labile triflate apical ligands [Mo6Ii8(CF3SO3)a6]2− have been immobilized through partial complexa-tion in apical positions by surface pyridine groups (PyS). The remaining CF3SO3

− apical ligands of [Mo6Ii8(PyS)ay(CF3SO3)

a6 − y] cluster units were further substituted with bromopyridine (Py-Br) to obtain air sta-

ble surface with expected final composition [Mo6Ii8 (PyS)ay(Py-Br)a6 − y]. The yield of the different reactionsteps is followed by X-ray photoelectron spectroscopy, providing cluster coverage ΣMo6I8 = 9 × 1012 cm−2.Each [Mo6I8]

4+ cluster is bound to the carbon surface through multiple anchoring metal sites (NPYR = 3 or 4), in-dicating that pyridine-terminated alkyl chains areflexible enough to accommodate four bonds. Electrical transportthrough Hg//Mo6I8–PyS–alkyl/a-C/p-Si(111) junctions shows rectifying current–voltage characteristics but doesnot reveal any signature of cluster immobilization.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

In addition to its early applications as hard protective coatings andsolid lubricant [1,2], amorphous carbon (a-C) is receiving a growing in-terest as a chemically inert and biocompatible electronic material.Amorphous carbon films have been studied as injecting layers in organ-ic light emitting diodes [3], as electrodes for analytical electrochemistrywith a large overpotential in aqueous solvent [4–6] and as electronicbio-sensing substrates [7,8].

For analytical and sensing applications, modification of carbon sur-faces by molecular grafting has been strongly developed using various(photo-, electro-) chemical methods [5,7–12]. In these fields, stabilityagainst temperature and electrical bias stress are important issues.Molecular systems with interesting redox properties have beenimmobilized on carbon to evaluate their potential interest formolecularmemories and catalytic systems based on carbon-supported metals [13,14]. Several families ofmolecular systemswithmultiple oxidation levels

Godet),

have been considered, including porphyrins, ruthenium complexes,octahedral metal clusters [15–17].

In octahedral cluster chemistry, the basic building blocks are[(M6Li8)La6] nanosized units (M = Mo; L = halogen, chalcogen; La =apical ligand and Li = inner ligand) [16–18]. The M6 cluster is face-capped by eight inner ligands to form rigid cluster cores characterizedby covalentM\Li bonds (Scheme 1) [19]. The cluster cores are ionicallybound to apical ligands located in terminal positions. The intrinsicphysicochemical properties (optical, electronic and redox) of theseoctahedral transition metal clusters are strongly related to the numberof electrons available for metal–metal bonds within the cluster [20].These electrons are localized on a set of twelve metal-bonding molecu-lar orbitals for the [(M6Li8)La6] unit. Consequently, full occupationof these orbitals leads to a closed-shell configuration with twentyfour valence electrons. These octahedral clusters can be furtherfunctionalized by substituting one or several labile apical ligand(s),e.g. trifluoromethanesulfonate (or triflate) groups (CF3SO3

−), withcoordinating organic or inorganic moieties [21–25]. Indeed, metalatom cluster units are versatile building blocks that exhibit a well-defined organization and monodispersity particularly useful fornanoarchitectonics [26]. Beyond carrying specific luminescence,

Scheme 1. Surface immobilization of a [Mo6I8]4+ metallic cluster core on amorphous carbon through complexation with a pyridine-terminated alkyl monolayer. For clarity, a single an-choring site is shown; however complexation with more than one anchoring site is also possible. In the final reaction step (Scheme 2), the remaining apical sites are substituted withbromopyridine.

132 C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

sensitization, catalytic and photo-catalytic properties [27–32], molecu-lar assemblies and nanomaterials built up fromM6 clusters could be in-volved in the material structuration by self-assembling of organiccationic frameworks in charge transfer salts thanks to their ability toform hydrogen bonding. The versatility of metal atom clusters canalso be used for the functionalization of surfaces [33] including semi-conducting electrodes.

The immobilization of octahedral molybdenum-based cluster cores[Mo6I8]4+ on crystalline silicon Si(111) substrates (n-type and p-type)has been previously reported [34,35]. It was shown that the reactionof [(n-C4H9)4N]2[Mo6I8(CF3SO3)6] with pyridine derivatives (Py-R)leads easily to the formation of [Mo6I8(Py-R)a6]4+ species both in solu-tion and on surfaces; triflates are substituted by pyridine groups whichare known to form strong covalent and stable Mo\N bonds with Mo6clusters. The electronic properties of Mo6I8–pyridine–alkyl–a-C systemare likely to depend on the surface coverage of [Mo6I8]4+ cluster cores(ΣMo6I8) while the robustness of the immobilization is related withthe anchoring configuration basically defined by the number NPYR ofsurface pyridine (PyS) moieties attached to each cluster.

This work aims at understanding the chemical reaction path of the[Mo6I8]4+ cluster core complexation by a pyridine-functionalized sur-face in order to quantify the number of surface anchoring sites per[Mo6I8]4+ cluster core. X-ray photoelectron spectroscopy (XPS) analysisof the immobilized cluster monolayer is performed to estimate the sur-face coverage and to determine the number, NPYR, of surface pyridinemoieties attached to each cluster and the chemical nature of the (6-NPYR) remaining apical groups. Bromine from the Py-Br ligand servesas a local XPS probe in order to estimate the number of molybdenumatoms per cluster not involved in the immobilization and thus to indi-rectly determine the number of PyS involved in the attachment of Mo6on the surface. However, in the case of Si substrates, it has beenshown that quantitative XPS data analysis is hampered by the intenseSi 2s plasmon losses (168 eV, 185 eV) overlapping the S 2p (sulfurfrom triflate moieties), I 4s (inner and apical ligands) and Br 3p (bro-mine from the labeling bromopyridine groups) core level peaks [35].Hence, theMo6I8–pyridine–alkyl–a-C systemprovides a better opportu-nity for a quantitative XPS analysis of the above chemical elements. Inaddition, the very low surface roughness (0.3–0.6 nm) of amorphouscarbon [36] is also interesting for quantitative XPS analysis of the clustercoverage.

Pulsed laser deposition (PLD) has been chosen to produce dense andsmooth tetrahedral a-C films with a high sp3 hybridization; since thissp3 fraction does not exceed 60% in the few nm from the surface, themain reactive chemical groups are probably C\H terminations andC_C double bonds [37]. a-C films can be reproducibly grafted with al-kene molecules with no previous surface preparation, using the samethermally assisted process as used with hydrogenated crystallineSi(111):H surfaces [36], leading to the covalent immobilization of adense alkyl monolayer (with ester or pyridine end-capping).

The coverage and the composition of the grafted molecular layersare estimated from quantitative XPS analysis at different steps of thegrafting process. The oxidation state, derived from the binding energyof Mo core levels, is also compared to that of the precursor (polycrystal-line powder). The number of surface anchoring sites is also deducedfrom XPS analysis to estimate the efficiency of the immobilization pro-cess. Electrical transport properties of metal–insulator–semiconductor(MIS) diodes obtained from functionalized a-C films were also investi-gated using a mercury drop as the soft electrical top contact to checkpossible changes in the electronic properties of a-C after the alkylmono-layer grafting and whether some signature of the monolayer ofelectroactive [Mo6I8]4+ clusters can be detected after immobilization.

2. Experimental

The cluster immobilization strategy (Schemes 1 and 2) isbased on a complexation reaction of the electrically neutral pre-cursor [(n-C4H9)4N]2[Mo6Ii8Ia6 − x(CF3SO3)ax] in which a numberNPYR of the labile triflate groups in apical positions are displaced bythe surface pyridine groups to form Mo\PyS bonding. To obtainderivatized surfaces sufficiently stable for subsequent characterizations,in a final step III, the (6-NPYR) remaining triflate groups (CF3SO3) aredisplaced by reaction with a more stable complexation moiety.Bromopyridine has been chosen because Br atoms are useful as labelingmoieties to improve the XPS analysis accuracy. The electrical neutralityof the cluster monolayer requires the presence of counter ions, e.g.tetrabutylammonium TBA+ ((n-C4H9)4N+) (step I) or triflate CF3SO3

−

(step III). Owing to the octahedral geometry of the cluster and the sterichindrance of the pyridine–alkyl chains bonded to the semiconductorsurface, all six Mo atoms cannot be simultaneously attached to surfacepyridine groups. Scheme 1 shows the minimal anchoring configuration

I) ((n-C4H9)4N)2[Mo6Ii8 I

a6] + x AgCF3SO3 →→ ((n-C4H9)4N)2[Mo6I

i8 I

a6-x(CF3SO3)

ax] + x AgI

II) ((n-C4H9)4N)2[Mo6Ii8 I

a6-x(CF3SO3)

ax] + y (PyS) →

(CF3SO3)y-2[Mo6Ii8I

a6-x (PyS)a

y(CF3SO3)ax-y] + 2 (n-C4H9)4N)CF3SO3

III) (CF3SO3)y-2[Mo6Ii8 I

a6-x(PyS)a

y(CF3SO3)a

x-y] + (x-y) Py-Br →

(CF3SO3)x-2[Mo6Ii8 I

a6-x (PyS)a

y(Py-Br)ax-y]

Scheme 2. Three-step reaction corresponding to the immobilization of [Mo6I8]4+metallic cluster cores on a-C surfaces through complexationwith pyridine-terminated alkylmonolayers.Partial substitution (x b 6) is observed in step I. The surface pyridine anchoring sites (y = NPYR) are denoted as (PyS) in steps II and III.

133C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

where each cluster is linked to the surface through a single pyridine(PyS) moiety.

2.1. Pulsed laser deposition of amorphous carbon

Carbon films (30 nm thick) were grown at room-temperature bypulsed laser deposition using a KrF laser (Tuilaser Excistar E20, λ =248 nm, 2 Hz, 40 ns pulse width) in a turbo-pumped high vacuumchamber (base pressure ~ 10−4 Pa). Crystalline silicon (c-Si (111), p-type) substrates were cleaned in ultrasonic bath (acetone, isopropanol,water) and dropped in NH4F (40%) to remove their native oxide beforea-C film growth. They were placed at 60 mm from the rotating glassycarbon target (Sigradur G, HTW). The deposited carbon thickness is0.03 nm/pulse for a pulse energy of 150 mJ (fluence = 7.3 J.cm−2). Be-fore deposition, the target was pre-cleaned for 5 min at 2 Hz with ashutter in front of the substrate. After deposition, nitrogen gas was in-troduced in the chamber and the carbon films were quickly transferredto the XPS setup or to the grafting setup under a dry nitrogenatmosphere.

2.2. Preparation of the molybdenum cluster-terminated carbon surfaces

Similar to our previous work with Si(111):H surfaces [35], ourstrategy for anchoring [Mo6I8]4+ cluster cores is based on the covalentderivatization of amorphous carbon surfaces by complexing pyridine-terminated alkyl monolayer. Linear ω-functionalized alkenes are usedto obtain densely packed monolayers with a robust covalent bondingto the a-C substrate [38]. Thefirst part of this thermally-induced graftingprocess includes successive surface reactions [39]: (i) covalent graftingof C10 linear alkenes, ω-functionalized with ester moieties, (ii) conver-sion of terminal ester to reactive acid functionalities, and (iii) amidationreaction with the amino-substituted pyridine. Briefly, ethylundecylenate CH2_CH(CH2)8\COOC2H5 (Aldrich, 97%) was passedthrough a neutral, activated alumina column to remove residual waterand peroxides, before liquid-phase reaction at 160 °C for at least 2 h.Using a protected ester (rather than an acid) functionality offers the ad-vantage of limiting the possible reactivity of acid functionswith the car-bon surface. The terminal ester groups were further converted to acidicgroups using dimethylformamide (DMF) + 0.25 M potassium tert-butoxide (t-BuOK) for 10 min, followed by HCl 1 M [40]. The acid-terminated surface was rinsed with ultrapure water and activatedwith N-hydroxysuccinimide (NHS) (Acros, 98+%) by immersion in afreshly prepared mixture of 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide (EDC) (Acros, 98+%) and NHS in DMF. The covalentattachment of the functional units was performed by immersing theNHS-activatedmolecular layer/a-C film in a deaerated acetonitrile solu-tion containing 0.1M of 4-aminomethylpyridine (Aldrich, 98%) at roomtemperature for 3.5 h.

The cluster chemistry includes three major preparation steps, assummarized in Scheme 2.

Step I The reaction of [(n-C4H9)4N]2[Mo6Ii8Ia6] and AgCF3SO3 inrefluxing THF is performed to remove and replace apical io-dides by triflate (CF3SO3

−) anions in apical positions [23,41];the [(n-C4H9)4N]2[Mo6I8(CF3SO3)6] formula is expected for acomplete reaction, where electrical neutrality is provided bytwo TBA+ counter cations. However previous XPS character-izations of the precursor (crystallized by diffusion of ether indry dichloromethane) [35] indicate incomplete replacementof apical iodine atoms (Ia) by triflate groups, leading to the pre-cursor formula [(n-C4H9)4N]2[Mo6Ii8Ia6 − x(CF3SO3)ax] (x b 6)instead of [(n-C4H9)4N]2[Mo6Ii8(CF3SO3)a6].

Step II The [Mo6I8]4+ cluster core is anchored on the pyridine-terminated a-C surface by immersing the modified a-C filmovernight in a distilled dichloromethane solution (15mL) con-taining 8 × 10−3 M [(n-C4H9)4N]2[Mo6Ii8Ia6 − x(CF3SO3)ax] at40 °C; the replacement of some labile triflate groups by thecomplexation reaction with surface pyridine groups (PyS),leads to a (CF3SO3)y − 2[Mo6Ii8Ia6 − x(PyS)ay(CF3SO3)ax − y]composition for the electrically neutral immobilized cluster,with y= NPYR b x. At this step II, the number of triflate counteranions, y − 2, depends on the number of anchoring sites, y.

Step III To obtain derivatized surfaces sufficiently stable for subsequentcharacterization, the triflate anions remaining in apical posi-tions of the cluster after the immobilization step weresubstituted by 3-bromopyridine (Py-Br) by dipping themod-ified surface in neat 3-bromopyridine (99%, Aldrich) for30–40 min. Assuming that all apical triflates have beensubstituted and that electrically neutrality is provided bytriflate counter anions, the expected final composition is(CF3SO3)x − 2[Mo6Ii8Ia6 − x(PyS)ay(Py-Br)ax − y]. After this stepIII, in order to remove physisorbed metallic aggregates, a thor-ough rinsing of the Mo6 cluster-modified surface with HPLCgrade CH2Cl2 (previously distilled over P2O5)was performed [35].

2.3. XPS analysis

After exposure to ambient atmosphere for a fewminutes, the graftedcarbon surfaces were introduced in the UHV chamber and kept at1× 10−9mbar for several hours before XPS analysis. XPSmeasurementswere performedwith aMg Kα (hν=1254 eV) X-ray source, at an emis-sion angle α, using a VSW HA100 photoelectron spectrometer with ahemispherical photoelectron analyzer, at 22 eV pass energy, providingan experimental resolution of 1.0 eV. Spectral analysis included a Shirleybackground subtraction and peak separation using mixed Gaussian–Lorentzian functions. Peak area ratios quoted in this work have beencorrected for the photoemission cross section of each chemical element.Additionalmeasurementswere performedusing amonochromatizedAlKα (hν=1486.6 eV) X-ray source to obtain a better peak separation forspecific core levels.

In addition to the XPS analysis of metal (Mo) and inner (I) ligands toobtain their oxidation state, determination of the nature and number of

800 700 600 500 400 300 200 100 00

5

10

15

20

25

b)

Br 3p Br 3d

I 3d

Mo 3d

I 4dMo 3p

F 1s

Auger I

Auger O N 1sO 1s

XP

S s

igna

l (10

4 CP

S)

Binding energy (eV)

a-C + Pyridinea-C + Pyridine + Mo cluster

C 1s

S 2p

a)

Fig. 1. XPS survey spectra (α = 0°) before (a) and after (b) immobilization of [Mo6I8]4+

cluster cores on a pyridine-functionalized a-C surface (step II) followed by the complexa-tion with bromopyridine (step III).

134 C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

ligands of the immobilized cluster [Mo6Ii8Ia6 − x(PyS)ay(Py-Br)ax − y] re-quires an accurate value of the bromine relative intensity. Once x hasbeen determined using some iodine core level, the Br 3p or Br 3d corelevels provide x− y, and hence the number of anchoring sites y=NPYR.

Molybdenum Mo 3d (rather than Mo 3p) and sulfur S 2p (ratherthan S 2s) photoelectron emission lines were chosen for their largerphotoemission cross section; in contrast, the iodine I 4d (rather than I3d) linewas used to avoid corrections for the dependence of the analyz-er sensitivity at small photoelectron kinetic energies. The bromine Br 3dintensity can be compared with the stronger Br 3p intensity, althoughthe latter peaks are partially overlapped by the stronger I 4s band.

The [Mo6I8]4+ cluster core coverage, ΣMo6I8 (cm−2), was derivedfrom the Mo 3d intensity through a normalization to the C1s intensityof a clean HOPG (highly oriented pyrolytic graphite, Neyco) substrate.Assuming that self-attenuation of the Mo signal by the cluster itselfcan be neglected, the density ΣMo6I8 of [Mo6I8]4+ cluster cores (m = 6Mo atoms) per unit surface:

Σ ¼ 1m

AMo

AHOPG

λHOPGC1s

5:62NHOPG ¼ 6:01� 1014

AMo

AHOPGð1Þ

is derived from the measured peak areas AMo and AHOPG

(eV × counts·s−1), taking into account their relative photoemissioncross sections σ(Mo 3d5/2)/σ(C 1s) = 5.62, the C atom density(NHOPG=11.4×1022 at·cm−3) and the inelasticmean freepath,λC1sHOPG=1.78 nm, of C 1s photoelectrons (with 970 eV kinetic energy) in graphite[42]. A similar relationship applies for the coverage (ΣPYR) of the pyridine-functionalized a-C surface, using the N 1s signal with σ(N 1s)/σ(C 1s) =1.80, and m = 2 (pyridine and amide functions) if self-attenuation isneglected.

2.4. Electrical transport measurements

Electrical transport measurements were performed on Hg//Mo6I8/pyridine–alkyl/a-C/p-Si junctions, using a mercury drop as a top elec-trode because its large surface tension (480 mJ m−2) limits the occur-rence of mercury diffusion through pinholes in the molecular layer.Current (I)–voltage (V) characteristics were obtained at room tempera-ture using a Keithley 6487 picoammeter and a home-made setup withhanging Hg drop and a micromanipulator to adjust the contact area at~0.3mm2. A freshHg drop (99.999%, Fluka)was used for eachmeasure-ment, at several locations on the sample surface. An ohmic contact wasmade on the scratched rear side of the p-type Si substrate (resistivity =1–5 Ω·cm) by applying a silver paste electrode. In this work, the directcurrent transport characteristics were measured in a glove box with ahumidity level of 22–25% of the water saturation pressure, after XPSmeasurements (under UHV) and a brief exposure to the ambient.

3. Results

3.1. Amorphous carbon

The a-C film thickness, da-C = 30 nm, and optical properties weredetermined from spectroscopic ellipsometry. The refractive index n(2 eV) = 2.63 is characteristic of a high density material [43], asconfirmed by previous X-ray reflectivity measurements (ρ =2.84 g.cm−3) [37] and bulk plasmon energy (EP = 29.5 ± 0.2 eV)given by photoelectron energy loss spectroscopy; the surface hybridiza-tion of the as-grown a-C film (sp3/sp2 + sp3) was found to be 0.5–0.6from XPS measurements using a monochromatic X-ray source [37,44].

Under optimized PLD conditions including the pre-cleaning of thetarget, amorphous carbon films suitable for monolayer grafting experi-ments are obtained, with a very smooth surface (root-mean-squareroughness ~ 0.3 nm) as shown by Atomic Force Microscopy (AFM).

3.2. Pyridine-functionalized a-C

The XPS survey spectrum of the pyridine-modified carbon surfacereveals characteristic peaks from the carbon substrate itself and fromthe C 1s, O 1s and N 1s core levels of the attached organic molecule(Fig. 1a). The shape and the intensity of these spectra aswell as their rel-ative positions in energy are consistent with the monolayer composi-tion, with some additional O 1s contribution due to some adventitiousadsorbed oxygen or water molecules.

The N 1s core level is characterized by a main peak at 400.4 eV cor-responding to the pyridine (C5H4N) component and a shoulder at399.2 eV attributed to the amide (N\C_O) environment (notshown). The pyridine surface coverage is estimated to be at ΣPYR =2.9 × 1013 cm−2.

3.3. Mo6I8–pyridine–alkyl–a-C assembly

After attachment of molybdenum clusters and treatment with 3-bromopyridine, as expected the survey spectrum (Fig. 1b) shows thepresence of iodine and molybdenum atoms characterized by theirrespective d levels (I 3d3/2, I 3d5/2, Mo 3d3/2, Mo 3d5/2 and I 4d5/2 at632.5, 621.1, 232.8, 229.6 and 51.0 eV respectively) and 3p components(Mo 3p3/2 at 395.4 eV).

Fig. 2 shows the (I 3d3/2, I 3d5/2) core level with the expected spinorbit splitting of 11.5 eV. The main component corresponds to theeight inner iodine ligands, while a weaker component attributed tosome apical ligands (Ia), unreacted at step I, is observed at smaller bind-ing energy (ΔE ≈ 1.75 eV). Monochromatized X-ray source data wereused to avoid baseline correction problems in the fitting of Ii and Ia com-ponents; the area ratio (Ia/Ii) = 0.12 corresponds to an average of0.96 ± 0.1 apical iodine per cluster. Qualitatively similar results are ob-served with the (I 4d3/2, I 4d5/2) core level (not shown). Hence, onlyclusters with one apical iodine ligand were immobilized on a-C surface,which confirms previous results [35].

In Fig. 3, the strong (Mo 3d3/2, Mo 3d5/2) doublet is observed withthe expected 3.2 eV spin orbit splitting. The positions of the main Mo3d5/2 and 3d3/2 peaks are identical to that observed for the startingproduct [(n-C4H9)4N]2Mo6I14 and are perfectly consistent with a MoII

oxidation state. In addition, a smaller component (indicated as Mo⁎,red arrows) is observed at larger binding energy (ΔE ≈ 2.75 eV) withan intensity ratio (Mo⁎/MoII) = 0.23 ± 0.05. Similarly, the broad Mo3p level shows a peak tailing towards higher binding energies; as a con-sequence of the overlap between theN 1s andMo 3p core levels (Fig. 4),the fitted ratio (N 1s/Mo 3p) = 1.7 is likely overestimated.

640 635 630 625 6200

1

2

3

4RX mono (0°)

IaIa

Ia / Ii = 0.12

I 3d (5/2)I 3d (3/2)

XP

S s

igna

l (10

3 CP

S)

Binding energy (eV)

Exp.Fit

Fit Ii

Fit Ia

Fit Ii

Fit Ia

Fig. 2. XPS data (α = 0°, monochromatized X-ray source) showing the (I 3d3/2, I 3d5/2)core level after [Mo6I8]4+ cluster immobilization. Some unreacted apical ligands are ob-served at smaller binding energy than the inner iodine ligands with an intensity ratio(Ia/Ii) = 0.12.

420 410 400 3900

1

2

3

4

5

6

Mo3p (3/2)

Mo3p (1/2)N1s

XP

S s

igna

l (10

3 CP

S)

Binding energy (eV)

Exp. Mo 3p + N 1sFit

Fit Mo 3p (1/2) Fit Mo 3p (3/2) Fit Fit N 1s

Fig. 4. XPS data (α = 45°) after [Mo6I8]4+ cluster core immobilization showing the (Mo3p1/2, Mo 3p3/2) doublet and the N 1s core level. A broad tail is visible on the Mo 3p1/2peak at larger binding energies. The red arrow denotes a source satellite (absent withmonochromatized X-ray source).

135C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

The ratio between the total Mo 3d5/2 area and the main Ii 4d peakarea is estimated to be at 0.77 ± 0.05 which is close to the expectedvalue for a [Mo6I8]4+ cluster core (0.75). However, the ratio betweenthe MoII (main peak) area and the Ii 4d peak area is estimated to be at0.66 ± 0.05, indicating that the Mo⁎ component must be included inthe cluster composition. The cluster coverage, Σ=9.2 × 1012 cm−2, de-duced from theMo 3d intensity using Eq. (1), is consistentwith amono-layer of bound [Mo6I8]4+ cluster cores.

Besides the Mo 3d core level, the Br 3d core level clearly shows twocomponents with different binding energies, in contrast with the iodinecore atom and the counter anion elements (S and F).

Let us consider first the bromine labeling element core levels. Boththeweaker (Br 3d3/2, Br 3d5/2) doublet near 70 eV (Fig. 5) and the stron-ger (Br 3p1/2, Br 3p3/2) doublet near 183 eV (Fig. 6) were used to esti-mate the average number of triflate anions which were not exchangedby surface pyridine (PyS). (i) With the Br 3d signal at 70 eV beingbroader than expected from the doublet separation (≈1 eV), datadecomposition reveals the presence of two doublets separated byΔE ≈ 2.0 eV with an intensity ratio of ≈0.6 ± 0.1. The ratio (Br 3d/I4d) = 0.17 provides an average number of 1.5 bromopyridine unitsper bound metallic cluster. (ii) Decomposition of the Br 3p signal near

240 238 236 234 232 230 228 2260.0

0.2

0.4

0.6

0.8

1.0

1.2

Mo*

XP

S s

igna

l (10

4 CP

S)

Binding energy (eV)

Exp. Mo 3dFit

Fit Mo 3d (3/2) Fit Mo 3d (3/2) Fit Mo 3d (5/2) Fit Mo 3d (5/2)

Fig. 3. XPS data (α=45°) after [Mo6I8]4+ cluster immobilization showing the Mo 3d corelevel. Besides the strong (Mo 3d3/2, Mo 3d5/2) doublet, a smaller componentMo⁎ (doubletwith red arrows) is observed at larger binding energy (ΔE = 2.74 eV).

183 eV also requires two (Br 3p1/2, Br 3p3/2) doublets (Fig. 6) with atotal intensity consistent with the Br 3d fitting.

As far as the counter anions are concerned, the C 1s spectrum shows ahigh binding energy (292.7 eV) component (not shown) assigned to CF3moieties pertaining to triflate anions (CF3SO3

−) which remain on the sur-face after treatment with 3-bromopyridine. This result consistent withstep III (Scheme 2) is confirmed by the presence of F 1s and S 2p signalsat 684.2 eV and 169.0 eV, respectively. Fig. 6 shows the (S 2p1/2, S 2p3/2)doublet with the expected spin orbit separation of 1.37 eV and an inten-sity ratio of 0.58. Its position is consistent with triflate moieties [45]. Thepeak area ratio (S 2p/I 4d) is 0.21, which means that, on average, eachbound metallic cluster is surrounded by 1.8 triflate units. This ratio is ob-tained with a good precision; however, this value might beunderestimated because some triflate moieties may be lost during XPSmeasurements, as reported previously [35].

In summary, the Br 3d intensity provides (x − y) = 1.5 ± 0.3bromopyridine unit per bound metallic cluster and I 4d decompositiongives (6 − x) = 1 ± 0.1 apical iodine per cluster, on average. It can beconcluded that each metallic cluster is bound to the organic monolayersurface through either three or four pyridine anchoring groups, corre-sponding respectively to (CF3SO3)3[Mo6Ii8Ia1 (PyS)a3(Py-Br)a2] and

74 72 70 68 660

1

2

3

4

5

6

XP

S s

igna

l (10

2 CP

S)

Binding energy (eV)

Exp. Br 3dFit

Fit Br 3d (b) Fit Br 3d (a)

Fig. 5.XPS data (α=45°) after [Mo6I8]4+ cluster immobilization showing the (Br 3d3/2, Br3d5/2) doublet. Besides the main peak (black line), a second component (red line) is ob-served at larger binding energy (ΔE = 2.0 eV).

195 190 185 180 175 170 1650

5

10

15

Br 3p

XP

S s

igna

l (10

2 CP

S)

Binding energy (eV)

Exp. Fit

Fit I 4s Fit I 4s Fit Br 3p Fit Br 3p Fit S 2p Fit S 2p

S 2p

I 4s

Fig. 6. XPS data (α = 45°) after [Mo6I8]4+ cluster core immobilization with fitted peakswhich include X-ray source satellites. A single component is observed with the sulfur(S 2p1/2, S 3p3/2) doublet (green), while the broad I 4s (black) and the Br 3p (3p1/2, 3p3/2)(red) doublet were fitted with two components.

Fig. 7. Current density (a) and conductance (b) characteristics of Hg//molecular layer/a-C/p-type Si junctions (T = 296 K): before (blue line) and after (green circles) grafting of apyridine-terminated alkyl monolayer to a-C and after complexation of a monolayer ofcluster cores (red triangles).

136 C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

(CF3SO3)3[Mo6Ii8Ia1 (PyS)a4(Py-Br)a1]. Considering that one apical Ia isattached to the cluster core, a (−3) charge must be carried by threetriflate moieties in order to maintain the total charge balance.

3.4. Electrical transport

The I–V characteristics were measured at room temperature on thebare Hg//a-C–p Si(111) junction, after grafting of the pyridine-terminated monolayer and further immobilization of the [Mo6I8]4+

cluster core monolayer. Current density J(V) and numerically derivedconductanceG(V)= (dJ/dV) characteristicswere reproducible at sever-al points of the sample, measured over several days.

Fig. 7 shows strong rectification for all junctions with a rectificationfactor, J(+2 V)/J(−2 V), exceeding four decades for the Hg//a-C–pSi(111) heterojunction. The saturation conductance value, observed atpositive bias as expected for p-type Si substrates, corresponds to thebulk resistance of the Si substrate. After thermal grafting of the alkylmonolayer and pyridine functionalization, an increase in the reversebias current by about one decade is observed.

In the forward bias regime, I–V characteristics remain nearly un-changed after Mo cluster monolayer immobilization. The sharp expo-nential rise in the current density observed for increasing positivebias, corresponds to a decrease of the hole injection barrier from thep-type semiconductor into a-C. For the Hg//pyridine–alkyl–a-C–p Sijunction, the low forward bias (0–0.15 V) range can be accurately fittedwith a transport mechanism limited by a Schottky barrier, using an ef-fective barrier ΦEFF = 0.65 eV and an ideality factor n = 1.23.

In the reverse bias regime, the current density increases after ther-mal grafting of the pyridine-terminated alkyl monolayer, but no furthereffect can be detected after the cluster immobilization step. This regimecorresponds to injection of minority carriers (electrons) from p-type Siand transport through the a-C film.

4. Discussion

A monolayer of octahedral [Mo6I8]4+ cluster cores has beenimmobilized through complexation with a pyridine functionalizedamorphous carbon surface. Eachmetallic cluster is bound to the organicmonolayer surface through either three or four pyridine anchoringgroups, corresponding respectively to (CF3SO3)3[Mo6Ii8Ia1 (PyS)a3(Py-Br)a2] and (CF3SO3)3[Mo6Ii8Ia1 (PyS)a4(Py-Br)a1], where three triflatemoieties are required to maintain the total charge balance.

This work shows that [Mo6I8]4+ cluster cores are strongly bound tothe pyridine functionalized carbon surface through multiple pyridineanchoring sites. Conversely, it might also indicate that clusters withfewer binding sites (being less stable) have been untied from the sur-face during the grafting process and that the complexation reaction isthus partly reversible. On the other hand, it cannot be excluded thatthe less stable cluster core configurations, e.g. with a low number of an-choring sites, have been removed during step III of the process.

A large fraction (more than 60%) of available pyridine sites at thefunctionalized carbon surface (ΣPYR = 2.9 × 1013 cm−2) are participat-ing in cluster complexation, as derived from the average number of an-choring sites per cluster (NPYR= 3.5± 0.4). Considering the topologicalconstraints related to the octahedral cluster geometry, cluster attach-ment to the alkyl monolayer surface through 3 or 4 pyridine anchoringsites indicates that the pyridine-terminated alkyl chains are flexibleenough to accommodate up to four bonds. This work confirms previousresults for Mo6 cluster immobilization on Si(111) where the number ofanchoring groups was closer to 4 [35].

From crystallographic data, the footprint of the [Mo6I8]4+ cluster corewith bromopyridine complexation can be estimated at 2.2 nm2, i.e. thedensely packedmonolayer ideally represents 4.5 × 1013 clusters/cm2. Ex-perimentally, the Mo6I8–pyridine–alkyl–a-C system shows a moderatelylarge density of clusters (ΣMo6I8 = 9.2 × 1012 cm−2) providing a specificarea of Amol = 10.9 nm2 per bound cluster.

137C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

The binding energy of themain (Mo 3d3/2, Mo 3d5/2) doublet is iden-tical to that observed for the starting product [(n-C4H9)4N]2Mo6I14, thusbeing consistent with the expected oxidation number of +II. However,a fraction of the photoelectrons emitted byMo atoms have a smaller ki-netic energy (ΔE ≈ 2.75 eV) with an intensity ratio (Mo⁎/MoII) =0.23 ± 0.05. A similar effect could be detected after grafting of[Mo6I8]4+ cluster cores on pyridine-functionalized Si(111), howeverwith an extremely weak intensity of the Mo⁎ peak [35]. Note that twodifferent environments are found in the Mo6I8–pyridine–alkyl–a-C as-sembly with either one or two (Py-Br) unit(s) at apical positions,while a single (Py-Br) is essentially obtained in the Mo6I8–pyridine–alkyl–Si assembly. Two tentative explanations can be proposed:

(i) complexation by oneor two bromopyridine group(s) possibly in-duces different local charge values on each cluster element, witha larger binding energy (apparent cluster core oxidation) arisingfrom a larger number of electron withdrawing (Py-Br) units;however, with a single peak being found in the 3d and 4d spectraof inner iodine atoms, the hypothesis of a global charging of thecluster core seems unlikely.

(ii) in XPS studies of transition metal complexes, additional struc-tures at higher apparent binding energy have been reportedand attributed to charge transfer between the ligand and thecentral atom upon creation of a core hole [46]. In the case of oc-tahedral [Mo6X8]4+ cluster cores, although such effect has notbeen reported, different numbers of (Py-Br) and (PyS) complex-ation groups could induce different charge transfer mechanismsin the final electronic state after the photo-hole generation.

Electrical transport measurements performed with Hg//Mo6I8/pyridine–alkyl/a-C/p-Si(111) junctions show rectifying current–voltage characteristics, nearly independent of the pyridine func-tionalization or [Mo6I8]4+ cluster core immobilization. However,they do not reveal any signature of the [Mo6I8]4+ cluster core immo-bilization, in contrast with the Hg//Mo6I8–pyridine–alkyl–p-Si junc-tions investigated previously. This indicates that band bendinginduced by the [Mo6I8]4+ cluster core monolayer is allowed by thelow density of states at the OML–Si interface, while it is forbiddenby the high density of localized (π and π*) states in amorphous car-bon [1,47,48]. The increase in reverse bias current after thermalgrafting of the ω-functionalized monolayer at 160 °C, observed inFig. 7, may be attributed to an increase in the hopping conductivityof a-C as a result of partial sp3-to-sp2 C conversion (larger densityof localized π states) [48].

5. Conclusions

Molecular junctions incorporating electroactive (Mo6Ii8Xa6) transi-

tion metal cluster units have been obtained by covalent grafting of apyridine-terminated alkyl monolayer and subsequent pyridine com-plexation at apical cluster positions of the clusters. The very smoothsurface of tetrahedral a-C films grown by pulsed laser deposition allowsquantitative XPS determination of the cluster packing density (ΣOML =9.2 × 1012 cm−2) and the corresponding specific area Amol = 10.9 nm2

per bound cluster.Considering topological constraints due to the octahedral geometry,

cluster attachment to the surface through 3 or 4 pyridine anchoringsites indicates that the pyridine-terminated alkyl chains are flexibleenough to accommodate up to four bonds, providing robust surface im-mobilization. The pyridine complexation strategywithmultiple anchor-ing sites on the surface of this functional architecture may thus beinteresting for potential applications of octahedral clusters in photo-chemistry and catalysis, including different carbon materials (amor-phous carbon, glassy carbon, pyrolyzed photoresist films, graphene…).

Prime novelty statement

This work elucidates the binding mechanisms and determines thenumber of anchoring sites of octahedral metal clusters (potentially use-ful in photochemistry and catalysis) immobilized on tetrahedral a-Csurfaces, using XPS characterizations. The pyridine complexation strat-egy (PyS–alkyl/a-C) combines the robustness of the interface C\Cbonding and the multiplicity of anchoring sites on the functionalizedsurface.

Acknowledgment

This work was supported by Agence Nationale de la Recherche(ANR-07-BLAN-0170-02, project “CLUSTSURF”).

References

[1] J. Robertson, Diamond-like amorphous carbon,Mater. Sci. Eng. R37 (2002) 129–281.[2] C. Donnet, A. Erdemir, Tribology of Diamond Like Carbon Films: Fundamentals and

Applications, Springer, New York, 2008.[3] K. Lmimouni, C. Legrand, C. Dufour, A. Chapoton, C. Belouet, Diamond-like carbon

films as electron-injection layer in organic light emitting diodes, Appl. Phys. Lett.78 (2001) 2437–2439.

[4] H. Cachet, C. Deslouis, M. Chouiki, B. Saidani, N.M.J. Conway, C. Godet, Electrochem-istry of nitrogen-incorporated amorphous carbon films, J. Electrochem. Soc. 149(2002) E233–E241.

[5] R.L. McCreery, Advanced Carbon ElectrodeMaterials forMolecular Electrochemistry,Chem. Rev. 108 (2008) 2646–2687.

[6] B. Khadro, A. Sikora, A.S. Loir, A. Errachid, F. Garrélie, C. Donnet, et al., Electrochem-ical performances of B doped and undoped diamond-like carbon (DLC) films depos-ited by femtosecond pulsed laser ablation for heavy metal detection using squarewave anodic stripping voltammetric (SWASV) technique, Sensors Actuators BChem. 155 (2011) 120–125.

[7] B. Sun, P.E. Colavita, H. Kim, M. Lockett, M.S. Marcus, L.M. Smith, et al., Covalentphotochemical functionalization of amorphous carbon thin films for integratedreal-time biosensing, Langmuir 22 (2006) 9598–9605.

[8] M.D. Angione, R. Pilolli, S. Cotrone, M. Magliulo, A. Mallardi, G. Palazzo, et al., Carbonbased materials for electronic bio-sensing, Mater. Today 14 (2011) 424–433.

[9] J. Pinson, F. Podvorica, Attachment of organic layers to conductive orsemiconductive surfaces by reduction of diazonium salts, Chem. Soc. Rev. 34(2005) 429–439.

[10] S. Szunerits, R. Boukherroub, Different strategies for functionalization of diamondsurfaces, J. Solid State Electrochem. 12 (2008) 1205–1218.

[11] S. Ababou-Girard, F. Solal, B. Fabre, F. Alibart, C. Godet, Covalent grafting of organicmolecular chains on amorphous carbon surfaces, J. Non-Cryst. Solids (2006)2011–2014.

[12] R.L. McCreery, A.J. Bergren, Progress with molecular electronic junctions: meetingexperimental challenges in design and fabrication, Adv. Mater. 21 (2009)4303–4322.

[13] A. Yesildag, D. Ekinci, Covalent attachment of pyridine-type molecules to glassycarbon surfaces by electrochemical reduction of in situ generated diazonium salts.Formation of ruthenium complexes on ligand-modified surfaces, Electrochim. Acta55 (2010) 7000–7009.

[14] C. Moreno-Castilla, F.J. Maldonadao-Hódar, Carbon aerogels for catalysis applica-tions: an overview, Carbon 43 (2005) 455–465.

[15] S. Zaman, K.J. Smith, A review of molybdenum catalysts for synthesis gas conversionto alcohols: catalysts, mechanisms and kinetics, Catal. Rev. Sci. Eng. 54 (2012)41–132.

[16] P. Braunstein, L.A. Oro, P.R. Raithby, Metal Clusters in Chemistry, vol. I and II, Wiley,Weinheim, 1999.

[17] S. Cordier, F. Dorson, F. Grasset, Y. Molard, B. Fabre, H. Haneda, T. Sasaki, M. Mortier,S. Ababou-Girard, C. Perrin, Novel nanomaterials based on inorganic molybdenumoctahedral clusters, J. Clust. Sci. 20 (2009) 9–21.

[18] S. Cordier, Y. Molard, K. Brylev, Y. Mironov, F. Grasset, B. Fabre, N. Naumov, Advancesin the Engineering of Near Infrared Emitting Liquid Crystals and Copolymers, Ex-tended Porous Frameworks, Theranostic Tools and Molecular Junctions Using Tai-lored Re6 Cluster Building Blocks, J. Clust. Sci. 26 (2015) 53–81, http://dx.doi.org/10.1007/s10876-014-0734-0.

[19] T. Hughbanks, R. Hoffman, Molybdenum chalcogenides: clusters, chains, and ex-tended solids. The approach to bonding in three dimensions, J. Am. Chem. Soc.105 (1983) 1150–1162.

[20] A.W. Maverick, J.S. Nadjdzionek, D. MacKenzie, D.G. Nocera, H.B. Gray, Spectroscop-ic, electrochemical, and photochemical properties of molybdenum (II) and tungsten(II) halide clusters, J. Am. Chem. Soc. 105 (1983) 1878–1882.

[21] D.H. Johnston, C.L. Stern, D.F. Shriver, Synthesis of 12-metal clusters based on the([Mo6Cl8]4+) core. X-ray structure of (PPN)2 [Mo6Cli8{(μ-NC)Mn(CO)2Cp}a6],Inorg. Chem. 32 (1993) 5170–5175.

[22] J.H. Golden, H. Deng, F.J. DiSalvo, J.M. Fréchet, P.M. Thompson, Monodisperse metalclusters 10 Angstroms in diameter in a polymeric host: the ‘monomer as solvent’ ap-proach, Science 268 (1995) 1463–1466.

138 C. Godet et al. / Diamond & Related Materials 55 (2015) 131–138

[23] D. Mery, L. Plault, C. Ornelas, J. Ruiz, S. Nlate, D. Astruc, et al., From simplemonopyridine clusters [Mo6Br13(Py-R)][n-Bu4N] and hexapyridine clusters[Mo6X8(Py-R)6][OSO2CF3]4 (X = Br or I) to cluster-cored organometallic stars,dendrons, and dendrimers, Inorg. Chem. 45 (2006) 1156–1167.

[24] K. Kirakci, H. Hosoda, S. Cordier, C. Perrin, G. Saito, A hybrid material based on[Mo6Br14]2− inorganic cluster units and [BEDO-TTF]+ organic monocationic radi-cals: synthesis, structure and properties of (BEDO-TTF)2Mo6Br14(PhCN)4, J. SolidState Chem. 179 (2006) 3628–3635.

[25] M. Amela-Cortes, S. Cordier, N. Naumov, C. Meriadec, F. Artzner, Y. Molard,Hexacyano octahedral metallic clusters as versatile building blocks in the designof extended polymeric framework and clustomesogens, J. Mater. Chem. C 2(2014) 9813–9823.

[26] S. Cordier, F. Grasset, Y. Molard, M. Amela-Cortes, R. Boukherroub, S. Ravaine, et al.,Inorganic molybdenum octahedral nanosized cluster units, versatile functional build-ing block for nanoarchitectonics, J. Inorg. Organomet. Polym. 25 (2015) 189–204,http://dx.doi.org/10.1007/s10904-014-0112-2.

[27] A. Garreau, F. Massuyeau, S. Cordier, Y. Molard, E. Gautron, P. Bertoncini, et al., Colorcontrol in coaxial two-luminophore nanowires, ACS Nano 7 (2013) 2977–2987.

[28] Y. Molard, C. Labbe, J. Cardin, S. Cordier, Sensitization of Er3+ infraredphotoluminescence embedded in a hybrid organic–inorganic copolymer containingoctahedral molybdenum clusters, Adv. Funct. Mater. 23 (2013) 4821–4825.

[29] S. Kamiguchi, S. Nagashima, T. Chihara, Characterization of catalytically active octa-hedral metal halide cluster complexes, Metals 4 (2014) 84–107.

[30] A. Barras, S. Cordier, R. Boukherroub, Fast photocatalytic degradation of rhodamineB over [Mo6Br8(N3)6]2− cluster units under sun light irradiation, Appl. Catal. B 123(2012) 1–8.

[31] P. Kumar, S. Kumar, S. Cordier, S. Paofai, R. Boukherroub, S.L. Jain, Photoreduction ofCO2 to methanol with hexanuclear molybdenum [Mo6Br14]2− cluster units undervisible light irradiation, RSC Adv. 4 (2014) 10420–10423.

[32] S. Kumar, O.P. Khatri, S. Cordier, R. Boukherroub, S.L. Jain, Graphene oxide supportedmolybdenum cluster: first heterogenized homogeneous catalyst for the synthesis ofdimethylcarbonate from CO2 and methanol, Chem. Eur. J. 21 (2015) 3488–3494.

[33] A. Barras, M.R. Das, R.R. Devarapalli, M.V. Shelke, S. Cordier, S. Szunerits, et al., One-pot synthesis of gold nanoparticle/molybdenum cluster/graphene oxide nanocom-posite and its photocatalytic activity, Appl. Catal. B 130 (2013) 270–276.

[34] S. Ababou-Girard, S. Cordier, B. Fabre, Y. Molard, C. Perrin, Assembly ofhexamolybdenum metallic clusters on silicon surfaces, ChemPhysChem 8 (2007)2086–2090.

[35] B. Fabre, S. Cordier, Y. Molard, C. Perrin, S. Ababou-Girard, C. Godet, Electrochemicaland charge transport behavior of molybdenum-based metallic cluster layers

immobilized on modified n- and p-type Si(111) surfaces, J. Phys. Chem. C 113(2009) 17437–17446.

[36] H. Sabbah, S. Ababou-Girard, A. Zebda, B. Fabre, S. Deputier, A. Perrin, et al., Thermalgrafting of organic monolayers on amorphous carbon and silicon (111) surfaces: acomparative study, Diamond Relat. Mater. 18 (2009) 1074–1080.

[37] A. Zebda, H. Sabbah, S. Ababou-Girard, F. Solal, C. Godet, Surface energy and photo-electron spectroscopy studies of amorphous carbon surfaces, Appl. Surf. Sci. 254(2008) 4980–4991.

[38] C. Godet, H. Sabbah, M. Herve, S. Ababou-Girard, S. Deputier, A. Perrin, et al., Thermalstability of perfluorinated molecular monolayers immobilized on PLD amorphouscarbon surfaces, Mater. Sci. Eng. 16 (2010) 012003.

[39] S. Ababou-Girard, H. Sabbah, B. Fabre, K. Zellama, F. Solal, C. Godet, Organic mono-layer grafting on amorphous carbon: surface preparation and packing density,J. Phys. Chem. C 111 (2007) 3099–3108.

[40] T. Strother, W. Cai, X. Zhao, R.J. Hamers, L.M. Smith, Synthesis and characterizationof DNA-modified silicon (111) surfaces, J. Am. Chem. Soc. 122 (2000) 1205–1209.

[41] D.H. Johnston, D.C. Gaswick, M.C. Lonergan, C.L. Stern, D.F. Shriver, Preparation ofbis(tetrabutylammonium) octa(μ3-chloro)hexakis(trifluoromethanesulfonato)-octahedro-hexamolybdate(2-), (Bu4N)2[Mo6Cli8(CF3SO3)a6]: a versatile startingmaterial for substituted molybdenum(II) clusters containing the [Mo6Cli8]4+ core,Inorg. Chem. 31 (1992) 1869–1873.

[42] S. Tanuma, C.J. Powell, D.R. Penn, Calculations of electron inelastic mean free paths.VIII. Data for 15 elemental solids over the 50–2000 eV range, Surf. Interface Anal. 36(2004) 1–14.

[43] T. Katsuno, C. Godet, J.C. Orlianges, A.S. Loir, F. Garrelie, A. Catherinot, Optical prop-erties of high-density amorphous carbon films grown by nanosecond and femtosec-ond pulsed laser ablation, Appl. Phys. A 81 (2005) 471–476.

[44] C. Godet, D. David, H. Sabbah, S. Ababou-Girard, F. Solal, Bulk and surface plasmonexcitations in amorphous carbon films measured by core-level loss spectroscopy,Appl. Surf. Sci. 265 (2009) 6598–6606.

[45] B.J. Lindberg, K. Hamrin, G. Johansson, U. Gelius, A. Fahlman, C. Nordling, et al., Mo-lecular spectroscopy by means of ESCA II. Sulfur compounds. Correlation of electronbinding energy with structure, Phys. Scr. 1 (1970) 286–298.

[46] P. Bagus, E. Ilton, C. Nelin, Review: the interpretation of XPS spectra: insights intomaterials properties, Surf. Sci. Rep. 68 (2013) 68273–68304.

[47] C. Arena, B. Kleinsorge, J. Robertson, W.I. Milne, M.E. Welland, Electronic propertiesof tetrahedral amorphous carbon investigated by scanning tunneling microscopy,J. Appl. Phys. 85 (1999) 1609–1615.

[48] C. Godet, Physics of bandtail transport in disordered carbons, Diamond Relat. Mater.12 (2003) 159–165.