Diamond functionalization with light-harvesting molecular wires : improved surface coverage by optimized Suzuki cross-coupling conditions

Donor-acceptor type light-harvesting molecular wires are covalently attached to a borondoped diamond surface via a combination of diazonium electrografting and Suzuki crosscoupling. For the Suzuki reaction, various catalytic systems are compared with respect to their imposed surface coverage. Combining 2-dicyclohexylphosphino-2’,6’-dimethoxybiphenyl (SPhos) and Pd(0), the diamond coverage improves considerably (by 98%) as compared to the standard tetrakis(triphenylphosphine)palladium(0) (Pd(PPh3)4) catalyst. As the energy levels between the molecular chromophores and the diamond film align well, the sophisticated functionalized diamond surfaces present a first step towards the development of fully carbonbased devices for light to electricity conversion.


Introduction
In recent years, boron-doped diamond (BDD) has become increasingly important within science and technology. 1 By adding boron, diamond conductivity can be tuned, with values ranging from 1 × 10 -9 to 100 Ω -1 cm -1 , the latter resulting in metallic behavior. 2Boron-doped nanocrystalline diamond (B:NCD) prepared in a thickness of ~150 nm is highly transparent (with a transmission of 70-80%). 3Despite its extraordinary physical properties and numerous applications (chemical and biosensors, electrosynthesis etc.), 4 earlier works focused mostly on tailoring the diamond's surface properties.Photochemical modification, 5 plasma treatment, 6 and diazonium coupling 7 afford a diversity of diamond surface modification schemes, allowing the attachment of various molecular structures.The next challenge, however, is to develop a versatile chemistry that can impart useful sensory, optoelectronic and/or photophysical properties to the B:NCD material.

b']dithiophen-2-yl)vinyl]-3-cyano-5,5-dimethylfuran-2(5H)-ylidene}malononitrile
(Br-CPDT-Fur) (Scheme 1), via optimized Suzuki cross-coupling reactions.Both molecular wires possess a bromoaryl moiety to enable reaction with an arylboronic ester functionalized B:NCD surface.The most effective catalytic system was pursued to increase the coupling efficiency and hence the surface coverage.Surface functionalization as well as coverage were determined by XPS.The CPDT-Fur molecular wire has suitable properties to be used effectively in future molecular photovoltaic systems, i.e. high absorbance, wide absorption and good charge transfer abilities through the π-system. 19Furthermore, the highest occupied molecular orbital (HOMO) of the dye seems to match well with the valence band of the B:NCD film, as evaluated by ultraviolet photoelectron spectroscopy (UPS), allowing hole injection from the dye to diamond.

Results and discussion
As shown in Scheme 1, the anchoring of the molecular wires onto the diamond surface begins with the functionalization of hydrogen-terminated B:NCD (see Experimental) with arylboronic ester moieties through diazonium electrografting. 16Two light-harvesting molecular wires with a donor-acceptor structure (to increase light-harvesting and enhance charge separation), Br-BT-Rho and Br-CPDT-Fur (Scheme 1), were pre-synthesized in solution before undergoing Suzuki coupling to the B:NCD surfaces.Both molecular wires contain an electron-rich (bridged) bithiophene 'donor' part and an electron-deficient subunit and they carry a reactive arylbromide group on the donor part.As the B:NCD film is finally intended to be used as the hole conductor in photovoltaic devices, the donor part of the chromophores has to be attached to the diamond surface.The Br-BT-Rho chromophore was synthesized by reacting commercially available 5-bromo-5'-formyl-2,2'-bithiophene with rhodanine-3-acetic acid in a standard condensation reaction.On the other hand, the Br-CPDT-Fur molecule was synthesized according to the procedure outlined in Scheme 2.

Scheme 2
Synthetic protocol towards the Br-CPDT-Fur molecular wire.Successful electrografting of the arylboronic ester groups on the B:NCD surface by the diazonium approach was confirmed by CV (Fig. 1).The first CV cycle shows two reduction peaks.The broader peak at -0.6 V (vs Ag/AgCl) can be assigned to the formation of arylboronic ester radicals.During the second cycle, this reduction wave disappears and the cyclic voltammogram only exhibits a very small reduction current, which suggests passivation of the diamond surface by the grafted layer, hence blocking the access of the diazonium cations to the surface.A smaller reduction peak appearing at -0.1 V (vs Ag/AgCl) was previously noticed as well for electrografting of 4-nitrophenyl-and 4-carboxyphenyldiazonium cations, but its origin is not yet clearly understood. 7 Fig. 1 In situ electrochemical generation of boronic ester derivatized phenyldiazonium cations, showing five CV reduction cycles (first cycle, black solid line; second cycle, blue dashed line; fifth cycle, red dotted line) with applied voltage between +0.5 and -0.8 V (scan rate 100 mVs -1 ).
XPS spectra of the hydrogenated and arylboronic ester functionalized B:NCD films are shown in Fig. 2. As can be seen clearly from the high-resolution XPS spectra, the B 1s peak of the dopant boron atoms in B:NCD is found at ~186 eV.After electrografting of the arylboronic ester groups on the B:NCD surface, an additional intense signal appears at ~191 eV.This peak is assigned to the B 1s core level in boron atoms from the grafted arylboronic ester layer. 16High-resolution XPS clearly shows the presence of two chemically distinct boron species.The arylboronic ester surface provides an excellent solid phase platform for further C-C extension via Suzuki cross-coupling.As such, an 'all-carbon' molecular photovoltaic system could be realized.However, a diamond surface with high chromophore loading is required for this purpose and this high coverage is not trivially obtained.To tackle this, a careful search for the optimum Suzuki conditions on arylboronic ester functionalized B:NCD was conducted.
Table 1 and 2 gather some of the different catalytic systems (changing the Pd catalyst, ligand, base and/or solvent) investigated for the two different chromophores (or their precursors).It should be mentioned at this point that the 'opposite' strategy, coupling 5-(5-formyl-2thienyl)-2-thiopheneboronic acid ((HO) 2 B-BT-CHO) to a bromophenyl decorated B:NCD surface (obtained via spontaneous grafting of 4-bromobenzenediazonium tetrafluoroborate), was employed as well, but the obtained results were inferior to the procedure outlined here.Successful coupling was confirmed by monitoring the S and N markers present in the dye molecules by XPS (Fig. 3).To determine the surface coverage, the S 2p peak intensities were calibrated against the XPS 2p intensity of a dodecanethiol self-assembled monolayer (SAM) on Au, 17 where the coverage of the latter was assumed to be 1 ML (see Experimental).The coverages for both molecular wires as determined by this method are listed in Table 1 and 2.
When the standard catalytic system outlined in Entry 1 (Pd(PPh 3 ) 4 , NaOAc, THF) was used, the obtained surface loadings (0.01 and 0.02 ML, respectively) were relatively low as compared to the results obtained by Zhong et al. 17 There are three possible reasons for the low surface coverage.First, the relative reactivity for the Suzuki cross-coupling is known to be R-I (Zhong and co-workers used an iodinated bithiophene) > R-Br >> R-Cl. 21Secondly, Pd(PPh 3 ) 4 is known to be easily degrading, even in the presence of the smallest amount of oxygen, thus reducing the coupling efficiency. 22Thirdly, for sterically demanding molecules as the two molecular wires used here, catalytic systems with electron-rich ligands are often more appropriate. 23A slightly better coverage was observed when an in situ generated Pd catalyst, via combination of tri(o-tolyl)phosphine (P(o-tol) 3 ) and palladium(II) acetate (Pd(OAc) 2 ), was used (Entry 6, Table 1; Entry 2, Table 2).The more electron-donating tri(otolyl)phosphine ligand enhances the oxidative insertion of Pd into the Ar-Br bond. 23Moreover, the steric congestion of the ligand can also facilitate dissociation of the ligand from the Pd complex.Over the past 20 years, the Buchwald group has made important contributions to the design of sterically hindered electron-rich biarylmonodentate phosphine ligands such as SPhos and XPhos (2-dicyclohexylphosphino-2',4',6'-triisopropylbiphenyl) which, when combined with catalytic amounts of Pd(0), enable to prepare extremely hindered biaryls via Suzuki-Miyaura cross-coupling reactions in an efficient manner. 24Recently, Zhou et al. successfully applied these conditions towards extension of the CPDT chromophore. 25Triggered by the effectiveness of the method and the similar CPDT building blocks used, we applied this catalytic system to the Suzuki cross-coupling on the B:NCD surface.Zhou and co-workers applied a dioxane:water mixture as the reaction solvent and K 3 PO 4 as a base.This catalytic system might, however, cleave our acceptor groups from the molecular wires due to hydrolysis.We thus first tested the catalytic system for two 'precursor' systems (Br-BT-CHO and Br-CPDT-CHO) devoid of the respective acceptor groups.As shown in ), the overall coupling efficiency was noticeably higher when SPhos was used as the supporting ligand.A surface coverage of 0.58 and 0.56 ML was achieved for the Br-2T-Rho and Br-CPDT-Fur chromophores, respectively, the highest reported surface coverages so far for Suzuki cross-coupling on B:NCD films (Fig. 3).An increase in coverage of 17% is achieved despite the use of a less reactive Br-Ar function and a more complex molecular structure (0.56 ML for Br-CPDT-Fur) as compared to the results of Zhong and co-workers (0.41 ML for 2-(dicyanovinyl)-5iodobithiophene). 17Moreover, the Pd catalyst usage is also quite low (3 mol%) as compared to standard Pd(PPh 3 ) 4 , which requires a loading of 10 mol%. 17 studies have shown that E f -VBM is 0.25 eV in bulk B-doped diamond.[28] Accordingly, the HOMO levels of the Br-BT-Rho and Br-CPDT-Fur functionalized diamond samples lie 0.3 and 0.5 eV below the VBM of B:NCD, respectively.These energy differences have implications on the possibility for hole transport between the B:NCD thin film and the functionalized molecular wires.Based on the principle of charge transfer in dyesensitized solar cells with p-semiconducting photocathode (p-DSSCs), upon photogeneration of an exciton, the electron is transferred to an acceptor site and subsequently to the oxidized form of the electrolyte redox mediator.In turn, the hole is transferred from the donor site to the semiconductor electrode.Hence, a good match between the HOMO of the donor molecule and the valence band of the electrode is important for efficient hole injection.For B:NCD, the VBM is situated rather close to the HOMO of both molecular wires, which should favor hole injection from the light-harvesting chromophores to the B:NCD thin film. 17,26 a proof of concept, photoelectrochemical measurements were performed using either a hydrogen-terminated, Br-CPDT-Fur or Br-BT-Rho functionalized B:NCD film as the working electrode in a standard three-electrode electrochemical system (Fig. 5).After equilibration in darkness, the light source was switched on and off approximately every 10 s for several cycles.At Time = 0, the electrode was in dark, and each light-on triggered the cathodic photocurrent.The process was fully reversible, and was stopped at Time ≈ 300 s, when the electrode was relaxed again in dark.We observed an increased photocurrent intensity for both molecular wire functionalized B:NCD electrodes as compared to the control system.In accord with previous work, 17 negative bias enhances the photocurrent considerably.
This agrees well with the mechanism that, upon light generation of electron-hole pairs in the molecular wires, the separated electrons flow toward the MV 2+ (dimethyl viologen) electron carrier and the holes toward the B:NCD electrode (at negative bias). 16,17r photocurrents are comparable to those reported earlier for similar systems: Zhong et al. 16 observed a photocurrent of ca. 150 nA/cm 2 under white light (150 W halogen lamp) illumination of B:NCD sensitized by dicyanovinyl-bithiophene and C 60 -bithophene.Later on, the same group reported ca.4-6 µA/cm 2 under 1 sun illumination (AM1.5G; 100 mW/cm 2 ). 17 general, photocurrents on sensitized diamond electrodes are very small compared to those observed for the state-of-art n-DSSCs with titania photoanode (≈ 20 mA/cm 2 at 1 sun).
Among other reasons, the difference can be rationalized in terms of surface morphology of the aforementioned semiconductor photoelectrodes.Early works on flat titania surface, sensitized with Ru-bipyridine complexes, reported photocurrents of several nA under white light (450<λ/nm<650) illumination from a 150 W Xe-lamp. 27 In principle, the flat electrode cannot deliver external quantum efficiencies larger than ≈ 0.3%, whereas mesoporous electrodes can give values exceeding 90% for the same dye/electrolyte systems. 28The B:NCD films made by CVD are usually quite compact, with double-layer capacitances approaching those of flat surfaces (ca. 3 μF/cm 2 ). 29Also our SEM image shown below (Fig. 6) confirms quite dense  Curves are offset for clarity, but the intensity scale is identical in all cases.

Conclusions
In summary, we have demonstrated that a proper choice of the catalytic system used for Suzuki cross-coupling can effectively increase the coupling efficiency, also on a diamond surface.Best results, leading to unprecedented surface loadings, were obtained for the Pd(OAc) 2 /SPhos combination.The resulting B:NCD films decorated with light-harvesting molecular dyes can potentially be applied for 'all-carbon' photovoltaic applications.
Preliminary photoelectrochemical measurements using the novel molecular wire systems show photocurrent generation.More detailed studies concerning full devices and quantum efficiencies will be carried out and reported in near future.Hydrogenation.Hydrogenation was performed using the same PECVD reactor (ASTeX 6500).To create a hydrogen terminated surface, all diamond samples were treated with hydrogen plasma under the following conditions: i) 500 sccm hydrogen flux, 30 Torr reactor pressure with a 3500 W microwave power for 2 min, ii) 500 sccm hydrogen flux, 15 Torr reactor pressure with a 2500 W microwave power for 5 min, iii) at the end of the plasma treatment, the microwave power was switched off and the samples were allowed to cool down under hydrogen flux for 40 min.

Experimental
Diazonium coupling.The functionalization of the diamond surfaces started with diazonization of the targeted aniline (4-aminophenylboronic acid pinacol ester) followed by electrochemical reduction of the in situ generated diazonium salt by cyclic voltammetry (Scheme 1). 16The electrografting reaction was performed with an Autolab PGSTAT30 potentiostat (Eco Chemie B.V.) and controlled by the GPES Program.A one-compartment electrochemical cell with three-electrode configuration was used.The working electrode was a hydrogen-terminated B:NCD film.Platinum wire and Ag/AgCl/KCl (sat) electrodes were used as counter and reference electrodes, respectively.Half of the diamond sample was immersed in the reaction solution for larger area functionalization.As stated above, 5 mM of 4aminophenylboronic acid pinacol ester was diazotized with an equimolar amount of NaNO 2 in a (N 2 gas purged) 0.5 M HCl solution, which was directly used for CV scanning with a potential from +0.5 to -0.8 V (vs Ag/AgCl) at a scan rate of 100 mV/s for 5 scans. 16After the modifications, the substrates were sequentially rinsed and sonicated in ultrapure water, THF and n-hexane to remove any physisorbed molecules.
Suzuki cross-coupling.All preparations for the Suzuki cross-coupling reactions were performed in a glove box.To achieve the highest possible surface coverage, the most efficient conditions were pursued.In general, a 15 mL ACE pressure tube (Sigma Aldrich) containing a magnetic stirring bar and the functionalized diamond film was charged with the dye (Br-BT-CHO/Br-BT-Rho/Br-CPDT/Br-CPDT-Fur), Pd catalyst, base and solvent.The resulting mixture was then heated at 80 °C for 18 h.Detailed information on the catalyst systems is provided in Table 1 and 2.
X-ray/ultraviolet photoelectron spectroscopy.Photoemission experiments were carried out using a Scienta ESCA 200 spectrometer in ultrahigh vacuum with a base pressure of 1x10 -10 mbar.The measurement chamber is equipped with a monochromatic Al Kα X-ray source and He discharge lamp providing photons with 1486.6 eV for XPS and 21.22 eV for UPS, respectively.The XPS experimental conditions were set so that the full width at half maximum of the clean Au 4f 7/2 line was 0.65 eV.The total energy resolution of the measurements in UPS, determined by the Fermi edge of clean gold, is about 0.1 eV.All spectra were measured at a photoelectron take-off angle of 0° normal emission.The UPS spectra have been corrected for the contributions from He I satellites radiation.
The surface coverage with the thiophene moieties was estimated using the same method as used in Ref. 17.This is done by calibrating the XPS S 2p signal of the thiophene moieties against the measured XPS S 2p intensity of a dodecanethiol SAM on Au, where the coverage of the latter was assumed to be 100% (1 ML) (Fig. 7).Several samples of dodecanethiol SAMs were prepared on Au, i.e. at different functionalization times, to make sure the same S 4H-Cyclopenta[1,2-b:5,4-b']dithiophenes (CPDTs) have been proven to be attractive electron donating building blocks for applications in organic electronics due to the planarity of the bridged bithiophene system and the excellent charge transfer capabilities through the molecule.19For the purpose of B:NCD functionalization, a novel CPDT chromophore has been developed, employing a furan group as the acceptor part attached to the CPDT core via a vinyl bridge.The synthesis of the CPDT core was done according to two different previously published procedures.19a,20CPDT formylation was performed using a standard Vilsmeier-Haack procedure in 1,2-dichloroethane at 0 °C.This reaction yielded exclusively monoformylated CPDT derivative 2, which was used in the next step without prior purification.Treatment of CPDT-CHO 2 with N-bromosuccinimide (NBS) at room temperature furnished Br-CPDT-CHO derivative 3 in a smooth manner.To attach the furan acceptor, a Knoevenagel condensation reaction was performed in the presence of a catalytic amount of piperidine.The Br-CPDT-Fur dye precipitated from the hot reaction mixture and was recrystallized from the corresponding solvent and used as such.UV-Vis analysis of the dye showed a high extinction coefficient ( = 73100 L mol -1 cm -1 at λ = 582 nm), thus proving suitability of the molecule for the envisaged application.The frontier orbital energy levels, as determined from the oxidation and reduction onsets derived from cyclic voltammetry (CV), are E HOMO = -5.70eV and E LUMO = -4.03eV.
UPS was applied to probe the electronic properties of the valence band of the functionalized B:NCD samples.Figs 4(a) and (c) show the UPS spectra of the Br-BT-Rho and Br-CPDT-Fur functionalized B:NCDs, respectively, compared to the respective pure molecular wire systems (drop-casted on an Au substrate).When the UPS intensity is plotted on a logarithmic scale, the weak valence band feature can be distinguished.It can clearly be seen that the functionalized B:NCD films reproduce the spectral features of the pure molecular wires.

Fig. 4
Fig. 4 (a,c) Comparison of the UPS spectra of the molecular wire-functionalized B:NCDs (blue line) and the pure molecular wires (drop-casted on an Au substrate) (black line) for (a) Br-BT-Rho and (c) Br-CPDT-Fur on an logarithmic scale.(b,d) Comparison of the UPS spectra of the molecular wire-functionalized B:NCDs (blue line) and a bare hydrogenterminated B:NCD film (red line) for (b) Br-BT-Rho and (d) Br-CPDT-Fur.Linear extrapolation of the valence band edge yields an estimate of the VBM the of B:NCD thin film.
packing of the diamond crystals without any noticeable mesoporosity.Hence, surface engineering of the B:NCD photoanode is a straightforward task in the future development of diamond-based p-DSSCs.

Fig. 6
Fig. 6 Typical SEM image of a B:NCD film.
2p signal was obtained by XPS.The S intensity in dodecanethiol has been corrected for the attenuation factor due to the carbon chain.Attenuation factor = exp( −     = 0.71 (n = 12 is the number of carbon atoms in the alkyl chain, d C = 1.1 Å is the length of one methylene group (1.27 Å) at a tilt angle of 30° to the surface normal, θ is the angle made by the electron energy analyzer to the surface normal, and λ S = 9.0 + 0.022 KE). 33All samples were measured with short collection time and on several different spots to avoid any beam damage.However, single-collected S2p data are shown here.

Fig. 7
Fig. 7 XPS spectrum showing the S 2p signal of dodecanethiol on a Au surface.

Table 1
Optimization of the Suzuki cross-coupling conditions for Br-

BT-CHO and Br-BT- Rho.
a General conditions: 80 °C, 18 h, 6 mL solvent, 100 μmol Ar-Br.b The XPS S 2p intensity was used to estimate the surface coverage.c Precursor used as a model compound.

Table 2
Optimization of the Suzuki cross-coupling conditions for Br

-CPDT-CHO and Br- CPDT-Fur.
a General conditions: 80 °C, 18 h, 6 mL solvent, 100 μmol Ar-Br.b The XPS S 2p intensity was used to estimate the surface coverage.c