Photogenerated charge collection on diamond electrodes with covalently linked chromophore monolayers

Photogenerated charge collection on diamond

electrodes with covalently linked chromophore

monolayers

1

†

2†

1

3

2

J. Raymakers , A. Artemenko , F. Verstraeten , H. Krysova , J. Cermák , S. S.

1

4

2

1

3

1*

Nicley , D. LopezꢀCarballeira , A. Kromka , K. Haenen , L. Kavan , W. Maes , B.

4

*

Rezek

†

These authors contributed equally

Corresponding authors

*

1

2

3

4

UHasselt, Institute for Materials Research (IMO) & IMEC vzw, IMOMEC, 3590 Diepenbeek, Belgium

Institute of Physics, Czech Academy of Sciences, 162 00 Prague, Czech Rep.

J. Heyrovsky Institute of Physical Chemistry, Czech Academy of Sciences, 18 223 Prague 8, Czech Rep.

Faculty of Electrical Engineering, Czech Technical University, 166 25 Prague, Czech Rep.

1

Abstract

Encouraged by its rich surface chemistry and excellent electrochemical properties, boronꢀdoped

nanocrystalline diamond (B:NCD) is a promising pꢀtype photoelectrode in dyeꢀsensitized solar cells. One

method of diamond surface functionalization using stable carbonꢀcarbon bonds involves the

electrochemical grafting of diazonium salts. However, this method typically leads to multilayers that may

complicate the transport of photogenerated charges. Here, we establish functionalization of B:NCD

electrodes

by

a

monolayer

of

ethynylphenyl

molecules

using

sterically

hindered

4ꢀ

(trimethylsilyl)ethynylbenzenediazonium tetrafluoroborate. Both the density and structural orientation of

the grafted layer are investigated by angular resolved Xꢀray photoelectron spectroscopy, confirming the

presence of covalently grafted monolayers. After removal of the trimethylsilyl protective groups, the

resulting ethynyl functionalities are employed to immobilize organic donorꢀacceptor chromophores via

Sonogashira crossꢀcoupling reactions. Homogenous surface coverage is achieved even on the B:NCD

electrode. Atomic scale DFT computing reveals that for the chromophore with the strongest acceptor unit,

efficient charge separation of 20 Å is obtained where photogenerated holes move directly into diamond.

Yet, photocurrent and photovoltage measurements suggest competitive electron recombination to the

diamond electrode via the redox electrolyte. Correlation between the density of the molecular layer and

photocurrents/photovoltage provides better understanding of the charge generation and recombination

pathways in diamondꢀorganic photoelectrochemical cells.

Keywords

Boronꢀdoped diamond

Photoelectrochemistry

– Surface functionalization – Sonogashira crossꢀcoupling – Monolayers –

2

1

Introduction

Already in 1955, scientists successfully synthesized artificial diamond from graphite through the highꢀ

pressure highꢀtemperature method by mimicking the extreme conditions in the Earth’s crust at depths of

1

60 km.[1] Nevertheless, its practical application remained limited since the crystal size of these manꢀ

made diamonds was not bigger than a few millimeters.[2] Progress in the application of synthetic

diamond was achieved only after the development of different chemical vapor deposition (CVD)

techniques, which allow the synthesis of diamond at low pressure from a gas mixture that typically

consists of hydrogen and a carbon source such as methane.[3ꢀ5] Nowadays, high quality films of both

single crystal and polycrystalline diamond can be readily grown on various substrates[6, 7] and different

nanostructures such as nanowires,[8] foams[9] or fibers[10] are accessible. Considering diamond’s

unique properties, i.e. extreme hardness,[11] excellent thermal conductivity,[12] good optical

transparency,[13] chemical stability under harsh conditions,[14] wide electrochemical potential

window,[15] biocompatibility,[16] tunable electron affinity[17] and the possibility to change diamond

from an insulating material to one with semiconducting or even semiꢀmetallic properties upon doping

with boron,[18] these diamond materials have been investigated for numerous applications.[19] Boronꢀ

doped diamond (BDD) electrodes are currently being employed in commercial chlorine sensors,[20]

electrosynthesis,[21] wastewater treatment[22] and the electrocatalytic reduction of CO₂to CO, formic

acid or methanol.[23] Moreover, due to the development of a rich surface chemistry, diamond structures

have been decorated with a variety of biomolecules (DNA, proteins, …) and lightꢀharvesting small

molecules, resulting in the fabrication of biosensors and dyeꢀsensitized solar cells (DSSCs),

respectively.[24, 25]

The application of BDD in pꢀtype DSSCs is appealing since the properties of diamond outperform those of

NiO electrodes, which are currently the gold standard in photocathode materials.[26] For instance,

diamond’s intrinsic chemical inertness is advantageous when corrosive electrolyte solutions such as the

ꢀ

2

ꢀ1

I /I redox couple are being employed. BDD also has a high hole diffusion coefficient (2ꢀ30 cm s ),

3

several times larger than that of NiO electrodes (4 x 10^ꢀ⁸cm²s^ꢀ¹).[27] Even more important is the

possibility to functionalize Hꢀterminated diamond surfaces by means of photochemical grafting[28] or

diazonium chemistry,[29] resulting in stable carbonꢀcarbon bonds. Subsequently, chromophores can be

immobilized on the surface via different types of coupling chemistries.[30ꢀ33] In this way, the lightꢀ

harvesting molecules are anchored onto the electrode material in a fully covalent fashion. These covalent

bonds are more robust than the coordination bonds that are employed to functionalize NiO electrodes and

should lead to an improved stability at the diamondꢀchromophore interface.

Pioneering work on the sensitization of BDD electrodes with an electrolyte solution was performed in

1

999 by Fujishima et al.[34] However, the first covalent functionalization of diamond with bithiopheneꢀ

C₆₀(2TꢀC₆₀) through a combination of diazonium grafting and Suzuki crossꢀcross coupling was only

reported a decade later by Zhong et al.[35] Optimization of this system resulted in photocurrents of ca.

.5 µA cm^ꢀ¹under 1 sun illumination (100 mW cm ) in a 5 mM methyl viologen (MV²⁺) solution.[36] In

ꢀ2

3

2

014, the Suzuki crossꢀcoupling on phenylboronic acid functionalized diamond thin film electrodes was

investigated by Yeap et al., who obtained a surface coverage as high as 0.56 monolayer for the

immobilization of a cyclopenta[2,1ꢀb:3,4ꢀb']dithiophene based chromophore (BrꢀCPDTꢀFur) by employing

highly reactive biarylmonodentate phosphine ligands.[30] This functionalized diamond electrode afforded

photocurrents of 150 nA cm^ꢀ²under 0.15 sun illumination (15 mW cm ) at an external bias of ꢀ0.2 V.

Translation of this functionalization strategy from thin films to mesoporous diamond foam electrodes

ꢀ2

ꢀ

2

resulted in a significant increase of the photocurrent to 15ꢀ22 µA cm .[37] In previous work we also

contributed to the development of diamondꢀbased pꢀtype DSSCs by designing twoꢀstep

functionalization approach based on combination of electrochemical diazonium grafting and

a

Sonogashira crossꢀcoupling.[33, 38] A dithieno[3,2ꢀb:2',3'ꢀd]pyrroleꢀ2,1,3ꢀbenzothiadiazole (DTPꢀBT)

chromophore was immobilized on the diamond surface and photocurrents up to 600 nA cm^ꢀ²were

obtained under 0.9 sun illumination at an applied bias of ꢀ0.3 V.[33] A different route to functionalize

diamond electrodes with lightꢀharvesting molecules was proposed by Krysova et al., using the

photochemical grafting of a short terminal alkene (C3) bearing a trifluoroacetamide group. After

deprotection, the resulting amine was employed to immobilize 4ꢀ(bis(4ꢀ[5ꢀ(2,2ꢀdicyanovinyl)thiopheneꢀ

2

ꢀyl]phenyl)amino) benzoic acid chloride, also known as the “P1 dye”.[39] The photocurrents compared

favorably to a nonꢀcovalent approach for the same dye.[32] More recently, record currents of 6.6 µA cm^ꢀ²

on thin films were reported by decreasing the amount of boron in the gas feed from 4000 to 250 ppm

B/C, which results in a doping concentration (1 x 10²⁰cm ) slightly below the threshold for semiꢀmetallic

conductivity.[40]

ꢀ3

Diazonium grafting is commonly employed to introduce various functional groups on the diamond

surface. This method was developed in 1999 by Swain et al. and holds several advantages over the

photochemical approach.[29] The grafting procedure is mostly performed electrochemically by

submerging a BDD electrode in a solution of an aryldiazonium salt and cycling to negative potentials. In

this manner, the diazonium salt is reduced to its corresponding radical, which is then subsequently

grafted on the surface, resulting in a complete passivation after only one cycle. πꢀπ stacking between the

aromatic rings contributes to the formation of denser layers.[41] Fully conjugated systems are thought to

be beneficial for charge transfer.[35] Nevertheless, electrochemical diazonium grafting also suffers from

a major drawback, which is the lack of control and strong tendency to form thick multilayers (5 to 10 nm)

due to the attack of radicals on the already grafted layer.[42ꢀ44] Since the amount of radicals is directly

related to the charge passing through the electrode, multilayer formation can be suppressed by briefly

applying a constant potential just below the redox potential of the diazonium salt.[45] Hꢀterminated

diamond can also be functionalized with a monolayer through spontaneous diazonium grafting as a result

of its intrinsic pꢀtype surface conductivity due to surface transfer doping.[46] Another way to prevent

multilayer formation was proposed by Leroux et al. and employs diazonium salts that bear bulky

functional groups such as triisopropylsilyl protected ethynyl moieties, as they sterically shield the reactive

ortho positions.[47] However, neither of these precautions against multilayer formation were taken in

earlier work by our group during the sensitization of the BDD electrodes.[33, 38] Therefore, it is safe to

4

assume that in all previous cases a multilayer network of a few nanometers exists in between of the

chromophores and the diamond electrode. Hence, after excitation of the chromophores, the generated

holes have to travel through this layer before they can be injected into the BDD electrode, increasing the

chance of charge recombination.

In this work we aimed to functionalize boronꢀdoped nanocrystalline diamond (B:NCD) with a monolayer

of ethynylphenyl molecules by employing trimethylsilyl (TMS) protected 4ꢀethynylbenzenediazonium

tetrafluoroborate. These TMS protected diazonium salts were chosen over the more bulky triisopropylsilyl

derivatives to obtain a denser layer and a higher surface coverage.[48] The surface composition was

thoroughly analyzed to confirm the presence of this monolayer, which until now has only been assumed

for B:NCD electrodes. After removal of the protecting TMS groups, three different donorꢀacceptor

chromophores, with increasing electronꢀdeficient character of the acceptor unit, were immobilized onto

the surface via Sonogashira crossꢀcoupling. Afterwards, the surface composition was compared with the

inverted Sonogashira crossꢀcoupling approach from our previous study[33] and photoelectrochemical and

photovoltage measurements were performed to study the influence of this monolayer functionalization on

photocurrent generation and charge separation.

2

Materials and methods

2

.1 Materials and instrumentation

All purchased chemicals were of the highest available quality and they were used without further

purification. 2ꢀIodothiophene, 4ꢀiodoaniline, ammonium diethyldithiocarbamate, copper(I) iodide, tertꢀ

butyl nitrite, malononitrile, boron trifluoride etherate, 2ꢀdicyclohexylphosphinoꢀ2′,6′ꢀdimethoxybiphenyl

(

SPhos) and 2ꢀdicyclohexylphosphinoꢀ2′,4′,6′ꢀtriisopropylbiphenyl (XPhos) were purchased from Sigma

Aldrich. Chloroform, dichloromethane, hexanes, dry tetrahydrofuran (THF), dry N,Nꢀdimethylformamide

DMF), dry toluene, dry diisopropylamine (DIPA), nꢀbutyl lithium (2.5 M solution in nꢀhexane; nꢀBuLi),

(

trimethyltin chloride, tetrabutylammonium fluoride (1.0 M solution in THF; TBAF) and glacial acetic acid

were purchased from Acros Organics. Nꢀiodosuccinimide, (trimethylsilyl)acetylene and (+/ꢀ)ꢀ2,2'ꢀ

bis(diphenylphosphino)ꢀ1,1'ꢀdinaphthalene

Tris(dibenzylideneacetone)dipalladium(0),

(BINAP)

were

purchased

from

(0),

Fluorochem.

tetrakis(triphenylphosphine)palladium

tri(orthoꢀ

tolyl)phosphine, transꢀdichlorobis(triphenylphosphine)palladium(II) and palladium(II) acetate were

purchased from Strem Chemicals. All solvents employed for rinsing were of HPLC grade. DTP 1 and the

BT derivatives 2 and 5 were synthesized according to literature procedures.[33, 49, 50]

Preparative (recycling) size exclusion chromatography (prepꢀSEC) was performed on a JAI LCꢀ9110 NEXT

ꢀ

1

system equipped with JAIGEL 1H, 2H and 3H columns (eluent CHCl , flow rate 3.5 mL min ). Nuclear

3

magnetic resonance (NMR) chemical shifts (δ, in ppm) were determined relative to the residual CHCl₃

(

7.26 ppm) absorption or the ¹³C resonance shift of CDCl₃(77.16 ppm). High resolution electrospray

ionization mass spectrometry (ESIꢀMS) was performed using an LTQ Orbitrap Velos Pro mass

spectrometer equipped with an atmospheric pressure ionization source operating in the nebulizer assisted

5

electrospray mode. The instrument was calibrated in the m/z range 220−2000 using a standard solution

containing caffeine, MRFA and Ultramark 1621. UVꢀVis absorption spectroscopy measurements were

ꢀ

1

performed on a VARIAN Cary 5000 UVꢀVis spectrophotometer at a scan rate of 600 nm min . The films

for the UVꢀVis absorption measurements were prepared by drop casting solutions of the chromophores in

chloroform on a quartz substrate. The solidꢀstate UVꢀVis absorption spectra were used to estimate the

optical gap (from the wavelength at the intersection of the tangent line drawn at the low energy side of

the absorption spectrum with the baseline: E_g(eV) = 1240/(wavelength in nm)). Electrochemical

measurements (cyclic voltammetry) were performed with an Eco Chemie Autolab PGSTAT 30

potentiostat/galvanostat using a threeꢀelectrode microcell with a platinum working electrode, a platinum

counter electrode and a Ag/AgNO₃reference electrode (silver wire dipped in a solution of 0.01 M AgNO₃

and 0.1

M NBu PF in anhydrous acetonitrile). The reference electrode was calibrated against

4 6

ferrocene/ferrocenium as an external standard. Samples were prepared by dip coating the platinum

working electrode in the chromophore solution. The CV measurements were done on the resulting films

with 0.1 M NBu PF in anhydrous acetonitrile as electrolyte solution. To prevent air from entering the

4

6

system, the experiments were carried out under a curtain of argon. Cyclic voltammograms were recorded

ꢀ

1

at a scan rate of 100 mV s . For the conversion of V to eV, the onset potentials of the first

oxidation/reduction peaks were used and referenced to ferrocene/ferrocenium, which has an ionization

potential of −4.98 eV vs vacuum.

Xꢀray photoelectron spectroscopy (XPS) was used to determine the chemical composition of the diamond

surfaces using an XPS spectrometer (Kratos, AXIS Supra) equipped with a hemispherical analyzer and a

monochromatic Al Kα Xꢀray source (1486.6 eV). The XPS spectra were acquired from an area of 0.7 ×

2

0

.3 mm with a takeoff angle of 90°. The survey XPS spectra were recorded with a pass energy of 80 eV,

whereas the highꢀresolution spectrum scans were recorded with a pass energy of 20 eV. The obtained

XPS spectra were calibrated on 283.8 eV binding energy (sp³carbon phase).³⁹The relative atomic

concentrations of the chemical elements were calculated from high resolution XPS spectra. The CasaXPS

software with implemented linear baseline and Gaussian line shapes was used for spectral processing.

XPS peak positions were determined with an accuracy of ±0.2 eV.

All geometry optimizations have been performed gas phase within Density Functional Theory (DFT).[51,

5

2] A BLYP functional in combination with the 6ꢀ31G basis set was used to obtain the structures,[53, 54]

while singlet point calculations were carried out with the B3LYP functional,[54, 55] the same basis set

and the RIJCOSX approach. This level of theory was also coupled with timeꢀdependent density functional

theory (TDDFT).[56] All the wavefunctions were obtained with the ORCA suite of programs and further

analyzed with the Multiwfn package.[57]

Macroscopic Kelvin Probe measurements were performed on a Scanning Kelvin Probe system (KP

Technologies) using a gold coated probe (tip of 2 mm in diameter) at a distance defined by a gradient of

3

00. Illumination was provided by a solar simulator, Solar LightLine A1 from Sciencetech Inc. The light

6

was introduced to the sample via a light guide, of which the end was placed at a distance of 190 mm

from the measurement area so that the illumination corresponds to an AM 1.5 solar spectrum.

Photoelectrochemical current measurements were performed in an Arꢀsaturated 0.1 M Na SO₄solution

2

containing 5 mM methyl viologen (MV²⁺) as an electron carrier in a threeꢀelectrode glass cell equipped

with a quartz optical window for illumination of the B:NCD working electrode. The counter electrode was

platinum and the reference electrode was Ag/AgCl (sat. KCl). The glass cell was placed in a dark room

and controlled by a potentiostat (Zahner, Germany). The white light source was an Oriel LSH – 7320 ABA

LED solar simulator. The incident white light intensity was 100 mW cm^ꢀ²(which is roughly 1 sun).

2

.2 Chromophore synthesis

4

ꢀ(Trimethylsilylethynyl)aniline:[58] 4ꢀIodoaniline (2.00 g, 9.13 mmol), Pd(PPh ) Cl (0.32 g, 0.27

3

2

mmol) and CuI (0.087 g, 0.46 mmol) were loaded in a Schlenk vial equipped with a magnetic stirring

bar under inert atmosphere (Ar). Dry THF (30 mL) and dry DIPA (10 mL) were introduced and the

mixture was purged with Ar for 30 min before adding trimethylsilylacetylene (2.58 mL, 18.3 mmol). After

heating overnight at 50 °C, water (75 mL) was added and the product was extracted with chloroform (3 x

7

5 mL). The organic phase was dried over MgSO , filtered and concentrated in vacuo. The crude product

4

was purified via column chromatography (silica, hexanes:dichloromethane, 50:50), yielding the product

as a brown solid (1.53 g, 88%).

4ꢀ(Trimethylsilylethynyl)benzenediazonium

tetrafluoroborate:[58]

A

solution

of

4ꢀ

(trimethylsilylethynyl)aniline (0.772 g, 4.08 mmol) in dry THF (15 mL) was added dropwise to boron

trifluoride etherate (2.01 mL, 16.3 mmol) at ꢀ20 °C under an inert atmosphere. Subsequently a solution

of tertꢀbutyl nitrite (1.94 mL, 16.3 mmol) in dry THF (15 mL) was added over 30 min at ꢀ20 °C. After

stirring for 10 min at ꢀ20 °C, the reaction mixture was brought to 5 °C over 20 min. Cold diethyl ether

(75 mL) was added and after cooling in an iceꢀwater bath for 15 min, the product was collected as a pale

brown solid through filtration (1.02 g, 87%).

4ꢀHexylꢀ2ꢀ(trimethylstannyl)ꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrole: nꢀBuLi (2.5 M solution in n-

hexane; 1.67 mL, 4.18 mmol) was added dropwise to a solution of DTP 1 (1.00 g, 3.80 mmol) in THF (10

mL) at ꢀ78 °C. After stirring for 1 h at ꢀ78 °C, trimethylstannyl chloride (1.0 M solution in n-hexane, 4.75

mL, 4.75 mmol) was rapidly added and the reaction mixture was gradually brought to room temperature.

After stirring overnight, water (50 mL) was added and the aqueous layer was extracted with diethyl ether

(

3 × 50 mL). The combined organic layers were dried over MgSO , filtered and concentrated in vacuo.

4

The crude material was purified via prepꢀSEC (chloroform), yielding the product as a grayꢀgreen oil (1.26

1

mg, 78%). H NMR (400 MHz, Chloroformꢀd): δ = 7.10 (d, J = 5.3 Hz, 1H), 7.02 (s, 1H), 6.99 (d, J = 5.3

Hz, 1H), 4.19 (t, J = 7.1 Hz, 2H), 1.90ꢀ1.81 (m, 2H), 1.37ꢀ1.23 (m, 6H), 0.87 (t, J = 7.1 Hz, 3H), 0.41

(s, 9H).

2ꢀ(Benzo[2,1,3]thiadiazolꢀ4ꢀyl)ꢀ4ꢀhexylꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrole (3): Monostannylated

DTP (1.00 g, 2.35 mmol), monobrominated BT 2 (530 mg, 2.46 mmol), Pd (dba) (64 mg, 0.070 mmol)

2

3

7

and P(oꢀtolyl) (86 mg, 0.28 mmol) were loaded in a Schlenk vial and placed under Ar atmosphere. Next,

3

dry toluene (10 mL) and dry DMF (4 mL) were added, and the mixture was purged with Ar. After

refluxing for 1 h, diethyl ether (100 mL) was added and the mixture was washed with water (2 × 100

mL) and brine (50 mL). The organic layer was dried over MgSO , filtered and concentrated in vacuo. The

4

crude product was purified via column chromatography (silica, hexanes:dichloromethane, 60:40),

1

yielding the product as a red solid (930 mg, 95%). H NMR (400 MHz, Chloroformꢀd): δ = 8.34 (s, 1H),

7

5

0

.86 (dd, J = 8.7, 0.9 Hz, 1H) 7.84 (dd, J = 7.2, 0.9 Hz, 1H), 7.61 (dd, J = 8.7, 7.2 Hz, 1H), 7.20 (d, J =

.3 Hz, 1H), 7.02 (d, J = 5.3 Hz, 1H), 4.29 (t, J = 7.0 Hz, 2H), 1.96ꢀ1.88 (m, 2H), 1.39ꢀ1.23 (m, 6H),

.87 (t, J = 7.0 Hz, 3H).

2ꢀ(Benzo[2,1,3]thiadiazolꢀ4ꢀyl)ꢀ4ꢀhexylꢀ6ꢀiodoꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrole

(4):

Nꢀ

iodosuccinimide (555 mg, 2.47 mmol) was added to a solution of compound 3 (930 mg, 2.35 mmol) in

acetic acid (40 mL) and chloroform (40 mL). The reaction mixture was shielded from light and left to stir

for 3 h, after which water (75 mL) and chloroform (75 mL) were added. The organic layer was washed

with a 20% Na S O solution (3 × 75 mL), water (2 × 75 mL) and brine (75 mL). Next, the organic layer

2

7

was dried over MgSO , filtered and concentrated in vacuo. The crude product was purified via column

4

chromatography (silica, hexanes:dichloromethane, 60:40), yielding the product as a red solid (927 mg,

1

7

2

1

2

5%). H NMR (400 MHz, Chloroformꢀd): δ = 8.29 (s, 1H), 7.87 (dd, J = 8.7, 0.9 Hz, 1H), 7.82 (dd, J =

.1, 0.9 Hz, 1H), 7.60 (dd, J = 8.7, 7.2 Hz, 1H), 7.20 (s, 1H), 4.22 (t, J = 7.1 Hz, 2H), 1.95ꢀ1.85 (m,

H), 1.40ꢀ1.24 (m, 6H), 0.87 (t, J = 6.4 Hz, 3H); ¹³C NMR (100 MHz, Chloroformꢀd): δ = 156.3, 152.6,

47.7, 145.4, 137.6, 130.3, 129.1, 124.8, 120.9, 120.0, 119.9, 115.8, 113.3, 72.1, 48.1, 32.1, 31.0,

+

7.3, 23.2, 14.7; MS (ESI ) Calcd. for C₂₀H₁₈IN S [M] : m/z 522.9708, found: 522.9703.

3

7ꢀ(4ꢀHexylꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrolꢀ2ꢀyl)benzoꢀ2,1,3ꢀthiadiazoleꢀ4ꢀcarbaldehyde (6):

Monostannylated DTP (1.00 g, 2.35 mmol), formylated BT 5 (570 mg, 2.35 mmol), Pd (dba)₃(54 mg,

2

0

.059 mmol) and P(oꢀtolyl)₃(71 mg, 0.23 mmol) were loaded in a Schlenk vial and placed under Ar

atmosphere. Next, dry toluene (10 mL) and dry DMF (4 mL) were added, and the mixture was purged

with Ar. After refluxing for 1 h, diethyl ether (100 mL) was added and the mixture was washed with

water (2 × 100 mL) and brine (50 mL). The organic layer was dried over MgSO , filtered and

4

concentrated in vacuo. The crude product was purified via column chromatography (silica,

hexanes:dichloromethane, 20:80), yielding the product as a purple solid (844 mg, 85%). ¹H NMR (400

MHz, Chloroformꢀd) δ 10.66 (s, 1H), 8.53 (s, 1H), 8.19 (d, J = 7.6 Hz, 1H), 7.95 (d, J = 7.6 Hz, 1H),

7

1

.28 (d, J = 5.3 Hz, 1H), 7.02 (d, J = 5.3 Hz, 1H), 4.29 (t, J = 7.0 Hz, 2H), 1.94 (p, J = 7.0 Hz, 2H),

.42ꢀ1.26 (m, 6H), 0.86 (t, J = 7.1 Hz, 6H).

7

ꢀ(4ꢀHexylꢀ6ꢀiodoꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrolꢀ2ꢀyl)benzoꢀ2,1,3ꢀthiadiazoleꢀ4ꢀ

carbaldehyde (7): Nꢀiodosuccinimide (491 mg, 2.18 mmol) was added to a solution of compound 6

844 mg, 1.98 mmol) in acetic acid (40 mL) and chloroform (40 mL). The reaction mixture was shielded

(

from light and left to stir for 3 h, after which water (75 mL) and chloroform (75 mL) were added. The

organic layer was washed with a 20% Na S O solution (3 × 75 mL), water (2 × 75 mL) and brine (75

2

7

8

mL). Next, the organic layer was dried over MgSO , filtered and concentrated in vacuo. The crude

4

product was purified via column chromatography (silica, hexanes:dichloromethane, 20:80), yielding the

1

product as a purple solid (991 mg, 91%). H NMR (400 MHz, Chloroformꢀd) δ 10.68 (s, 1H), 8.49 (s, 1H),

8

.19 (d, J = 7.6 Hz, 1H), 7.94 (d, J = 7.6 Hz, 1H), 7.22 (s, 1H), 4.24 (t, J = 7.1 Hz, 2H), 1.91 (p, J = 7.0

Hz, 2H), 1.41ꢀ1.27 (m, 6H), 0.91ꢀ0.82 (m, 3H). ¹³C NMR (100 MHz, Chloroformꢀd) δ 188.3, 154.1,

1

3

52.3, 146.3, 145.6, 136.2, 134.4, 132.7, 124.9, 122.8, 120.4, 119.7, 118.1, 114.9, 73.9, 47.7, 31.4,

+

0.3, 26.7, 22.5, 14.0; MS (ESI ) Calcd. for C₂₁H₁₈IN OS [M + Na] : m/z 573.9554, found: 573.9548.

3

2ꢀ((7ꢀ(4ꢀHexylꢀ6ꢀiodoꢀ4Hꢀdithieno[3,2ꢀb:2',3'ꢀd]pyrrolꢀ2ꢀyl)benzoꢀ2,1,3ꢀthiadiazoleꢀ4ꢀ

yl)methylene)malononitrile (8): A solution of compound 7 (400 mg, 0.72 mmol), malononitrile (192

mg, 2.90 mmol) and 3 drops of piperidine in dichloromethane (30 mL) was stirred for 90 min at room

temperature. The crude mixture was purified via column chromatography (silica, dichloromethane),

1

yielding the product as a metallic green solid (320 mg, 74%) that gives a blue solution in chloroform. H

NMR (400 MHz, Chloroformꢀd) δ 8.75 (s, 1H), 8.73 (d, J = 8.1 Hz, 1H), 8.49 (s, 1H), 7.93 (d, J = 8.1 Hz,

1

H), 7.22 (s, 1H), 4.23 (t, J = 7.1 Hz, 2H), 1.91 (p, J = 6.8 Hz, 2H), 1.39ꢀ1.26 (m, 6H), 0.87 (t, J = 6.9

+

Hz, 3H); MS (ESI ) Calcd. for C₂₄H₁₈IN S [M] : m/z 598.9769, found: 598.9742.

5

3

2

.3 Diamond growth

1

50 nm thick B:NCD films were grown on (100)ꢀoriented 10 × 10²mm Si pꢀtype conductive substrates

by MWPECVD, using trimethylborane (TMB) as the boron source, according to previously employed

protocols.[38] Characterization performed on samples grown in the same reactor under similar growth

conditions indicated that the B:NCD films have boron concentrations of approximately 5 × 10²¹ꢀcm^ꢀ³and a

resistivity below 0.005 Ωꢀcm.

2

.4 Electrochemical diazonium grafting

The B:NCD thin films were functionalized

with

4ꢀ(trimethylsilyl)ethynylbenzenediazonium

tetrafluoroborate by means of electrochemical diazonium grafting using cyclic voltammetry. The

electrografting protocol was performed with an Autolab PGSTAT30 potentiostat (Eco Chemie B.V.) and

controlled by the GPES Manager software. A oneꢀcompartment electrochemical cell was employed in a

threeꢀelectrode configuration. Hydrogenꢀterminated B:NCD was used as the working electrode, a

platinum wire as the counter electrode and a Ag/AgCl/KCl(sat) electrode as the reference. All potentials

are reported versus this reference electrode at room temperature. Diamond working electrodes were

functionalized

in

two

steps

by

immersing

each

half

of

the

sample

in

a

4ꢀ

(trimethylsilyl)ethynylbenzenediazonium tetrafluoroborate solution. The electrografting process was

performed by cycling the potential between +600 and ꢀ760 mV vs. Ag/AgCl at a scan rate of 50 mV s^ꢀ¹

for five scans. After grafting, the B:NCD samples were sonicated in MilliQ water and subsequently rinsed

with toluene, THF, ethanol and MilliQ water.

9

2

.5 Sonogashira crossꢀcoupling

All Sonogashira crossꢀcoupling reaction mixtures were prepared in a glovebox under inert atmosphere

N ). The lightꢀharvesting molecule, palladium catalyst, copper(I) iodide, diamond substrate, base and

(

2

solvent were loaded in a tailor made Schlenk vial equipped with a magnetic stirring bar.[33] The Schlenk

vial was closed off with a rubber septum, removed from the glovebox and placed in an oil bath at 60 °C

for 16 h. After surface functionalization, all B:NCD samples were sonicated in toluene (p.a.), a solution of

ammonium diethyldithiocarbamate (25 mg) in THF (10 mL), THF, ethanol and MilliQ water, before drying

under a vigorous stream of nitrogen. Specific reaction conditions are given in Table 3 and S2.

3

Results and discussion

The B:NCD functionalization was performed in three steps, as illustrated in Scheme 1 a. In the first step,

the formation of monolayer is achieved via the electrochemical diazonium grafting of 4ꢀ

trimethylsilyl)ethynylbenzenediazonium tetrafluoroborate (TEBDT). After removal of the TMS protecting

a

(

groups, the resulting ethynylphenyl terminated diamond surface can be employed to covalently link

iodine functionalized chromophores via Sonogashira crossꢀcoupling reactions.

N

X

S

N

a)

NC

O

X :

H

CN

H

2

N

Si

S

C

6

H

13

N

(i)

S

Si

BF

4

N

2

N

(ii)

(iii)

(iv)

Sonogashira

H

_e-

b)

Synthetic procedure toward DTPꢀBTꢀX chromophores:

C

N

6

H

13

C

N

6 13

H

S

N

(i)

N

(iii)

N

Br

I

S

(

ii),

2

O

C

N

6

H

13

3

4, Dye 1

2

5

S

C

N

6

H

13

C

N

6

H

13

C

N

6

H

13

1

S

(i)

N

(iii)

N

(iv)

N

I

S

(

ii),

5

6

O

7, Dye 2

O

8, Dye 3

CN

NC

Scheme 1 | a. General B:NCD functionalization strategy for the development of diamond photocathodes:

i) Diazotization of 4ꢀ(trimethylsilyl)ethynylaniline; (ii) Electrochemical reduction and subsequent grafting

of the corresponding diazonium salt; (iii) Removal of the trimethylsilyl groups with

(

1

0

tetrabutylammoniumfluoride (TBAF); (iv) Dye sensitization of diamond via Sonogashira crossꢀcoupling

with DTPꢀBTꢀX chromophores. b. Synthesis of three iodine functionalized DTPꢀBT derivatives: (i) nꢀBuLi,

Me SnCl, THF; (ii) Pd (dba) , P(oꢀtolyl) , toluene, DMF; (iii) NIS, CHCl , AcOH; (iv) malononitrile,

3

2

3

piperidine, CHCl₃.

3

.1 Chromophore synthesis

Three different pushꢀpull type chromophores were synthesized by changing the substituents on the

electronꢀdeficient benzothiadiazole (BT) acceptor unit from a hydrogen atom to an aldehyde or a

dicyanovinyl group, as shown in Scheme 1 b. These donorꢀacceptor chromophores were selected

because of their envisaged charge separation after excitation, easy tuning of the optical properties

through the introduction of various substituents and compact size, which enables a high diamond surface

coverage. In previous work we functionalized iodophenyl terminated diamond thin films with alkyne

functionalized dithieno[3,2ꢀb:2',3'ꢀd]pyrroleꢀ2,1,3ꢀbenzothiadiazole (DTPꢀBT) molecules via Sonogashira

crossꢀcoupling.[33] By using the same chromophore system, the influence of the initial surface

functionalization strategy (iodophenyl vs. ethynylphenyl or multilayer vs. monolayer) on the photocurrent

generation of the resulting photoelectrodes can be directly compared. The dithienopyrrole (DTP) donor

unit (1) was synthesized in two steps starting from 3ꢀbromothiophene, as described previously.[33] After

monostannylation using nꢀbutyl lithium and trimethylstannylchloride,

a Stille crossꢀcoupling was

performed with 4ꢀbromoꢀ2,1,3ꢀbenzothiadiazole (2). The resulting chromophore, DTPꢀBT (3), was then

prepared for immobilization on the diamond surface via Sonogashira crossꢀcoupling through introduction

of an iodine functionality with Nꢀiodosuccinimide (NIS). The two other chromophores, 7 and 8, were

prepared in a similar fashion. However, instead of acceptor unit 2, 7ꢀbromoꢀ2,1,3ꢀbenzothiadiazoleꢀ4ꢀ

carbaldehyde (5)[49, 50] was employed during the Stille crossꢀcoupling. The dithieno[3,2ꢀb:2',3'ꢀ

d]pyrroleꢀ2,1,3ꢀbenzothiadiazoleꢀ4ꢀcarbaldehyde (DTPꢀBTꢀCHO, 6) compound was also iodinated,

resulting in the second dye (7). Finally, the dicyanovinyl derivative, IꢀDTPꢀBTꢀDCV (8), was prepared

through a Knoevenagel condensation with malononitrile.

The optical and electrochemical properties of the three chromophores were investigated by UVꢀVis

absorption spectroscopy and cyclic voltammetry, respectively (Table 1). The complete UVꢀVis absorption

spectra of the designed chromophores (Figure 1 a) indicate that stronger electronꢀwithdrawing

properties of the acceptor part result in a decrease of the optical gap and hence a red shift of the

absorption maximum (λ_m_a_x). DTPꢀBT (4) shows a λ_m_a_xfor the intramolecular charge transfer band at 474

nm and a moderate molar extinction coefficient (ε) of 18 600 M^ꢀ¹cm . By exchanging the hydrogen end

group by an aldehyde in chromophore 7, the absorption maximum is redꢀshifted to 533 nm and ε

increases to 22 000 M^ꢀ¹cm . For DTPꢀBTꢀDCV (8), which contains the strongest electronꢀwithdrawing

dicyanovinyl group, λ_m_a_xis shifted further to the red part of the spectrum with a λ_m_a_xof 608 nm and a

molar extinction coefficient that is almost twice as large as for DTPꢀBT (ε = 30 800 M^ꢀ¹cm ).

ꢀ1

1

Table 1 | Overview of the experimentally obtained optical and electrochemical data of the DTPꢀBT, DTPꢀ

BTꢀCHO and DTPꢀBTꢀDCV chromophores.

λ

max

ε

E

ox

E

red

E

HOMO

E

LUMO

∆Eopt

(eV)

∆EEC

(eV)

Dye

ꢀ1

(nm)

(M cm

)

(eV)

0.43

0.68

0.69

(eV)

ꢀ1.49

ꢀ1.16

ꢀ0.84

(eV)

ꢀ5.21

ꢀ5.46

ꢀ5.47

(eV)

ꢀ3.29

ꢀ3.62

ꢀ3.94

DTPꢀBT

474

533

608

18 600

22 000

30 800

2.30

2.04

1.80

1.92

1.84

1.53

DTPꢀBTꢀCHO

DTPꢀBTꢀDCV

Eox and Ered represent the oxidation and reduction potential of the chromophores, whereas the optical gap as

determined by UVꢀVis is given by ∆Eopt and the electrochemically calculated gap is given by ∆EEC

a)

b)

-

2.87

-2.96

-

3.29

-

3.31 -3.33

-

3.62

-3.55

3.64

-

3.94

-

4.93

-

5.05

-

5.21

-5.19

5.39

-

5.28

-

5.46

-5.47

-

5.55

DTP-BT

DTP-BT-CHO DTP-BT-DCV

c)

d)

e)

HOMO

LUMO

HOMO

LUMO

HOMO

LUMO

DTPꢀBT

DTPꢀBTꢀCHO

DTPꢀBTꢀDCV

Figure 1 | a. Experimental UVꢀVis absorption spectra of DTPꢀBT, DTPꢀBTꢀCHO, and DTPꢀBTꢀDCV

chromophores. b. Highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital

(LUMO) energy levels and corresponding band gap of the individual chromophores (in eV). For each dye

molecule the experimental data is represented by the leftꢀmost box, the calculated energy levels are

provided by the middle box and the rightꢀmost box includes the effect on the HOMO/LUMO levels when

the chromophores are grafted onto the diamond surface. cꢀe. Arrangement of the chromophores grafted

on Bꢀdoped diamond as obtained by DFT computing. Blue and red surfaces are the frontier orbitals of the

whole system located on B atoms. Cyan and brown colors highlight frontier orbitals of the molecules.

In Figure 1 b the experimentally determined highest occupied molecular orbital (HOMO) and lowest

unoccupied molecule orbital (LUMO) energy levels of the different chromophores are represented by the

leftꢀmost box (darkest color). It is known that structural changes to the acceptor unit have the largest

influence on the location of the LUMO energy level.[59] Hence, the strong decrease of the LUMO level,

from ꢀ3.29 to ꢀ3.62 and ꢀ3.94 eV for the DTPꢀBT, DTPꢀBTꢀCHO and DTPꢀBTꢀDCV series, is in line with our

1

2

expectations. The decrease in the HOMO level was, accordingly, much less pronounced (from ꢀ5.21 to ꢀ

.46 and ꢀ5.47 eV). Note however, that these measurements were performed for standalone

5

chromophores and not on the diamondꢀchromophore complex. In order to obtain a better insight into the

electron configuration and function of this hybrid system, the HOMO and LUMO levels of the

chromophores that are immobilized on the diamond surface were determined theoretically by DFT

calculations. These value are given by the rightꢀmost box in Figure 1 b (lightest color) and for the

theoretical computing, a finite hydrogenated diamond (111) slab was designed including 448 carbon

atoms and 4 boron atoms. The concentration of boron in these slabs corresponds approximately with the

employed B:NCD substrates, and the chosen size was proven to be big enough to model an infinite slab

of diamond.[60] In the case of these diamondꢀchromophore complexes a similar trend was observed for

their HOMO/LUMO levels with regards to the standalone chromophores. However, after grafting of the

chromophores on diamond it is obvious that both the HOMO and LUMO energy is shifted by

approximately 0.3 eV. This may be due to enlargement of the conjugated chain after the linker addition

and grafting to diamond. To ensure fast hole injection from the HOMO of the dye into diamond, the

HOMO energy level should be situated below the valence band maximum (VBM) of diamond, which is

located at ca. ꢀ5.50 eV in contact with an electrolyte solution.[61] In all cases, the HOMO level is located

below the VBM of diamond, however, through the application of an external voltage, this difference can

be overcome. Conversely, all LUMO levels lie well above that of methyl viologen (MV²⁺) (ꢀ4.22 eV), which

is employed as a redox electrolyte. Electrons could be thus easily transferred from the molecule to the

electrolyte after excitation with light. From all chromophores, we expect that DTPꢀBTꢀDCV is the most

suitable for diamond photosensitization, due to a more favorable location of its HOMO energy level with

respect to the diamond VBM and the strong absorption over a broad range.

This assertion is also supported by the molecular orbital localization on an atomic scale. Figure 1 cꢀe

shows the arrangement of the chromophores grafted on boronꢀdoped diamond as obtained by DFT

computing. Blue and red surfaces are the frontier orbitals of the whole system located on B atoms. Cyan

and brown colors highlight frontier orbitals of the molecules. One can see that all three chromophores are

grafted in the same upright arrangement on diamond. No difference in bending, rotation, inclination is

observed. Thus, no structural effect on charge transfer between the molecules and diamond is expected.

The depicted molecular orbitals show that the HOMO of the chromophores is delocalized along the entire

conjugated system including the link to diamond, while the LUMO is centered in the BT moiety. The

spatial separation of HOMO and LUMO seems the most pronounced in the case of the DTPꢀBTꢀDCV

chromophore. This finding supports the idea of exciton dissociation and charge transfer, as an eventual

optical transition centered on the chromophores from the HOMO to the LUMO might lead to charge

separation in which the electron goes to the benzothiadiazole moiety, in agreement with its expected

electron acceptor properties. Therefore, we have performed further computational analysis of the excited

states and related charge transitions. The computed UVꢀVis spectra are shown in the supporting

information (Figure S1) and they reveal a strong interaction between the chromophores and the

diamond. After grafting of the chromophores on diamond the first intense absorption band is redꢀshifted

1

3

and split into several peaks. Key data obtained from the analysis of these split excited states is

summarized in Table 2. For the DTPꢀBT molecule, two major excitations, α and β, were observed and

they transfer 0.15 and 0.16 electrons from diamond to the benzothiadiazole unit, respectively, resulting

in an electronꢀhole couple that is separated by 3ꢀ4 Å. After addition of the formyl group for DTPꢀBTꢀCHO,

two similar excitations were observed. One of these transitions has an increased oscillator strength

(0.73) and is capable of transferring 0.18 electrons across a distance of 4.61 Å. Finally, DTPꢀBTꢀDCV

shows two strong signals with an oscillator strength of 0.57 and 0.60 that transfer 0.18 and 0.19

electrons to the acceptor unit, respectively. This is more than any other excitation on the previous

molecules. Moreover, a third excitation, γ, was observed, which is capable of transferring 0.75 electrons

despite its weaker oscillator strength of 0.12. Thereby it creates an electronꢀhole couple that is separated

by 19.94 Å with an exciton binding energy as low as 0.74 eV. The strong charge separation effect of the

γ excitation on the DTPꢀBTꢀDCV molecule is clearly evidenced in Figure S2, where holes and electrons

are strongly separated. The DFT based calculations thus show that among the investigated molecules,

DTPꢀBTꢀDCV should also provide the most efficient photoꢀgenerated charge separation and injection to

diamond.

Table 2 | Analysis of the excited states for the chromophores grafted on boronꢀdoped diamond.

Dye

Excitation

E (eV)

Osc. Str.

Q

ꢁr (Å)

E_B(eV)

α

β

α

β

α

β

γ

1.82

1.83

1.82

1.83

1.75

1.78

0.30

0.73

0.32

0.57

0.50

0.12

0.15

0.16

0.18

0.11

0.18

0.19

0.75

3.11

3.65

4.61

2.55

3.26

3.47

19.94

1.30

1.37

1.34

1.30

1.29

0.74

DTPꢀBT

DTPꢀBTꢀ

CHO

DTPꢀBTꢀ

DCV

These excited states were obtained due to splitting of the first optically active excited state of the standalone

chromophore after splitting. For each molecule, each transition is denoted with the letters of the Greek alphabet. Only

the transitions with the oscillator strength > 0.10 and high charge transfer were analyzed. Q is an effective electron

charge transferred between the benzothiadiazole to the diamond (positive sign = electrons transferred to BT). Charge

B

transfer distance is denoted as ꢁr. E stands for the exciton binding energy.

3

.2 Primary surface functionalization: Diazonium grafting

The envisaged monolayer functionalization of B:NCD with the various DTPꢀBTꢀX chromophores (where X

is either hydrogen, aldehyde or dicyanovinyl group; Scheme 1) was realized through the

electrochemical diazonium grafting of TMSꢀprotected ethynylbenzenediazonium tetrafluoroborate, as

described by Natsui et al. as shown in Figure a.[62] Diazonium grafting was performed

a

2

electrochemically by employing a freshly Hꢀterminated diamond film as the working electrode in a cyclic

voltammetry setup. The electrode was submerged in a solution of the diazonium salt (5 mM) and

tetrabutylammonium hexafluorophosphate (0.1 M) in acetonitrile. The potential was cycled between

ꢀ

1

+

600 and ꢀ750 mV (vs. a Ag/AgCl reference electrode) at a scan rate of 50 mV s . In this way, electrons

are ejected from diamond into the solution, which results in a reduction of the diazonium salt to the

corresponding radical and its immediate grafting on the surface.[29] During the first scan, a sharp

1

4

reduction peak was observed in the cyclic voltammogram at ꢀ335 mV, as shown in Figure 2 b. In the

subsequent scans, this reduction peak strongly decreased, indicating that after one cycle a near complete

passivation of the diamond surface with TMS protected ethynylphenyl molecules was obtained. The

cathodic current for the reduction peak in the first scan is approximately 4 times lower than the one for

electrochemical grafting of in situ diazotized 4ꢀiodoaniline.[33] Since the amount of generated radicals

and, thus, the thickness of the grafted layer is directly related to the charge that is passed through the

system, this hints already at the formation of thinner layers.[63]

After electrochemical diazonium grafting, the TMS protective groups were removed by stirring the

functionalized diamond samples in a 5 mL THF solution containing 100 mM tetrabutylammonium fluoride

(TBAF). Both the grafting and the deprotection step were analyzed via cyclic voltammetry in an

acetonitrile electrolyte solution containing 1 mM ferrocene and 100 mM TBAF. Cyclic voltammograms for

Hꢀterminated (B:NCDꢀH, black curve), TMS protected (B:NCDꢀPhꢀEthꢀTMS, blue curve) and alkyne

functionalized (B:NCDꢀPhꢀEth, red curve) diamond surfaces were recorded between ꢀ250 and +1000 mV

at a scan rate of 50 mV s^ꢀ¹(Figure 2 c). For the Hꢀterminated B:NCD electrode, the voltammogram

demonstrated two narrow peaks that correspond with the oxidation and reduction of ferrocene. The peak

separation potential (∆E ) of 70 mV is in line with values that are typically obtained for polycrystalline Hꢀ

p

terminated diamond.[64] After electrochemical grafting of the TMS protected diazonium salt, the peak

separation value increased to ∆E_p= 108 mV, accompanied with a decrease of both the oxidation and

reduction peak currents. These changes in electrochemical behavior suggest that the “bulky” TMS groups

impede efficient electron transport between ferrocene and the diamond electrode, thus acting as a barrier

for electric charge transfer.[65] Yet, immediately after treatment with TBAF, the peak separation

potential and currents are restored to the values as obtained for Hꢀterminated diamond, which suggests

successful deprotection leading to an alkyne functionalized surface.

The successful functionalization of the Hꢀterminated diamond electrodes was confirmed by analysis of the

surface composition via Xꢀray photoelectron spectroscopy (XPS). Indeed, after electrochemical diazonium

grafting, two new signals were detected at 151.5 and 100.5 eV, which correspond to the Si 2s and Si 2p

peaks, respectively. Calculation of the relative atomic surface concentration from XPS high resolution

spectra revealed the presence of 2.3 ± 0.5 at.% Si (measured over two samples at two different

2

locations, area = 0.7 × 0.3 mm ). The high resolution Si 2p peak (Figure 2 d) was deconvoluted into

three contributions: SiꢀC at 100.6 eV (79.5%), SiꢀN at 101.7 eV (18.2%) and SiꢀO at 102.4 eV

(2.3%).[66] As expected, the majority of the Si 2p peak is composed of SiꢀC bonds that arise from the

TMS protective groups. Remarkably, there is also a significant contribution from SiꢀN bonds. XPS analysis

showed that there is about 1.8 ± 0.26 at.% of nitrogen on the sample surface. A possible explanation for

this observation is that N₂is adsorbed onto the surface by storing the diamond samples under nitrogen

atmosphere or drying them under a nitrogen stream after functionalization and rinsing. It is worth to

mention that the observed nitrogen should not originate from azoꢀbridges because the bulky TMS groups

shield the ortho positions from multilayer formation.[67] After deprotection of the TMS groups with TBAF,

1

5

the surface Si concentration reduced to 0.15 ± 0.06 at.%, which translates to a reaction efficiency of

4%.

9

a)

b)

c)

d)

0.02

e^-

0

.00

H

ꢀ

0.02

0.04

0.06

0.08

0.10

N N

Si

N₂

Scan 1

Scan 2

Scan 5

Cyclic

voltammetry

ꢀ

800 ꢀ600 ꢀ400 ꢀ200

0

200

400

600

800

High density area

Voltage, mV vs. Ag/AgCl

0

.04

.03

.02

.01

.00

Si

B:NCDꢀH

B:NCDꢀPhꢀEthꢀTMS

B:NCDꢀPhꢀEth

H

ꢀ

0.01

0.02

0.03

ꢀ

400

ꢀ200

0

200

400

600

800

1000 1200

Voltage, mV vs. Ag/AgCl

TBAF

1

8

600

Low density area

B:NCDꢀPhꢀEthꢀTMS

B:NCDꢀPhꢀEth

SiꢀC

200

00

SiꢀN

SiꢀO

H

400

4

2

00

1

04

103

102

101

100

99

98

Binding energy, eV

Figure 2 | a. Preparation of an ethynylphenyl functionalized B:NCD surface. b. Cyclic voltammograms

for the electrochemical grafting of 4ꢀ(trimethylsilyl) ethynylbenzenediazonium tetrafluoroborate. c. Cyclic

voltammograms for Hꢀterminated (black), TMS protected (blue) and alkyne functionalized (red) B:NCD

electrodes in a 1 mM ferrocene solution in acetonitrile. d. Deconvoluted Si 2p high resolution XPS spectra

of TMS protected B:NCD (blue) and the Si 2p peak shape of alkyne functionalized B:NCD (red).

1

6

Additional information on the primary surface functionalization via electrochemical diazonium grafting and

the subsequent deprotection of the TMS groups can be obtained from the C 1s peak at 283.8 eV, which is

further explained in the supplementary information (Figure S3). XPS analysis was also employed to

estimate the density of the functionalized layer and its structural organization, which allows for the

differentiation between monolayer and multilayer functionalization.[45] The density of the obtained layer

after diazonium grafting can be assessed by comparing the relative atomic concentrations of boron after

grafting and after deprotection of the TMS groups. The boron concentration after diazonium grafting was

about 1.67 ± 0.06 at.%, whereas after deprotection of the TMS groups with TBAF the boron content

increased to 2.68 ± 0.22 at.%. Since the TMS group has a thickness, which is les than 3 Å,[47] this

difference in boron concentration cannot be attributed to merely a change in penetration depth of the Xꢀ

rays. We suppose that the observed change in boron content has to be related to a change in efficiency

of the functional surface layer to scatter the electrons from boron atoms from the diamond “bulk”.[47]

Hence, it can be assumed that the density of the grafted layer is determined by the packing density of

the TMS groups rather than that of the benzene molecules since the electrons from boron atoms are

scattered more efficiently before the deprotection step, which results in a lower boron at.%. As a result,

the ethynylphenyl molecules are likely not packed very densely and gaps are present in between the

different molecules, as shown in Figure 2. This observation is in line with the work by Hapiot et al., who

tuned the density of the functionalized surface by changing the protective group from trimethylsilyl (TMS)

to triethylsilyl (TES) and triisopropyl (TIPS) moieties.[48]

By employing angular resolved XPS (ARXPS), the organization of the surface layer could be determined.

In a randomly oriented film, which is associated with multilayer formation, the electrons from the Si and

C atoms have more or less equal trajectories through the functionalized film and undergo equal amounts

of scattering (Figure 3 a).[45] In a proper monolayer, however, all Si atoms are organized at the

topmost part of the functionalized layer. Therefore, when increasing the angle under which the samples

are measured, an increase of the Si/C ratio might be expected (Figure 3 b, c). However, it should be

mentioned that we are not dealing with bare Si atoms, but Si(CH₃)₃groups, where the methyl groups

determine the density of the layer. Hence, rather than an increase of the Si/C ratio, a decrease thereof is

expected due to the presence of these methyl groups at the topmost surface in case of a wellꢀorganized

layer. This hypothesis is rationalized by the fact that for the TMS protected surfaces, the CꢀH component

X

of the C 1s peak has the strongest contribution (Figure S3 and Table S1). ARXPS measurements were

performed at six different substrate angles θ and the results are shown in Figure 3 c.

1

7

a)

c)

2

8

Si

4

.5

.0

O/C

26

B/C

Si/C

N/C

2

4

2

H

3.5

Diamond Electrode

3

2

1

.0

.5

.0

.5

.0

20

18

1

6

4

2

0

b)

Si

0

20

40

60

80

Diamond Electrode

Substrate angle ( ), °

Figure 3 | a. Schematic representation of a wellꢀorganized monolayer of TMS protected ethynylphenyl

molecules. b. Schematic representation of a randomly organized TMS protected ethynylphenyl layer

corresponding with multilayer functionalization. c. B/C (blue), Si/C (red), N/C (green) and O/C (black)

ratios of their respective atomic concentrations for angular resolved XPS measurements at substrate

angles (θ) of 0, 40, 50, 60, 70 and 80° after electrochemical diazonium grafting of TEBDT.

The B/C ratio remained more or less constant until the substrate was tilted at angles higher than 40°,

after which it started to decrease to a value of 0.9 at 80°. This indicates that with increasing substrate

angles, higher volumes of the actual surface are being measured in comparison with the diamond bulk.

As hypothesized, the Si/C ratio also declined with increasing contact angles. This decrease, however, only

starts at angles > 60°, which can be explained by the fact that only at sufficiently high angles, there is

enough difference in the measured volume of the methyl groups and the Si atoms, which are both

located at the interface between the functionalized layer and the atmosphere. The N/C and O/C ratios,

however, demonstrate a rather exponential increase with increasing θ values. This suggests that both

nitrogen and oxygen species are adsorbed onto the surface rather than that they are integrated in the

grafted layer (e.g. through the formation of azoꢀbridges in case of nitrogen).

Based on these results we assume that the electrochemical diazonium grafting of TEBDT on B:NCD also

results in monolayers, which has only been proven for flat surfaces such as pyrolized photoresist films

and glassy carbon until now.[47, 48, 58] AFM scratching experiments were performed as well, but due to

the high roughness of the nanocrystalline diamond films, no clear conclusions could be drawn from those

results.

3

.3 Secondary surface functionalization: Sonogashira crossꢀcoupling

With regards to the development of BDD electrodes for photoelectrochemical applications, and ultimately

the development of pꢀtype DSSCs, two parameters are of utmost importance, i.e. a high surface

coverage with chromophores and minimal surface contamination. The latter can result in charge trapping

of the photogenerated holes and electrons. Charge transport also plays an important role. In this respect,

1

8

fully conjugated systems are desired for the fabrication of photoelectrodes. In contrast to the Stille crossꢀ

coupling, it has been shown that both the Suzuki and Sonogashira crossꢀcoupling, as well as the Cu(I)

catalyzed azideꢀalkyne cycloaddition (CuAAC), yield decent surface coverages without severe surface

contamination.[31, 38, 68] In previous work we demonstrated that BDD films functionalized via

Sonogashira crossꢀcoupling afforded photocurrents that are three times higher than when CuAAC

chemistry was employed.[33] Hence, in this work the Sonogashira crossꢀcoupling was applied to

immobilize the DTPꢀBT, DTPꢀBTꢀCHO and DTPꢀBTꢀDCV chromophores onto diamond electrodes

functionalized with alkyne monolayers, as illustrated in Scheme 2.

Dye 1

N

Dye 2

O

Dye 3

CN

N

S

N

S

Sonogashira

Cross-Coupling

S

N

S

R

N

S

N

I

S

Knoevenagel

Boron-Doped Nanocrystalline Diamond Electrode

Scheme 2 | Sonogashira crossꢀcoupling of 2ꢀiodothiophene, IꢀDTPꢀBT (dye 1), IꢀDTPꢀBTꢀCHO (dye 2)

and IꢀDTPꢀBTꢀDCV (dye 3) on B:NCD electrodes functionalized with a monolayer of ethynylphenyl

molecules.

Before employing the actual chromophores, a simple proof of concept reaction was performed with

commercial 2ꢀiodothiophene, using Pd(PPh ) Cl (5 mol%) and CuI (10 mol%) in a solvent mixture (3:1)

3

2

of THF and diisopropylamine (DIPA). After thorough rinsing and sonication in various solvents, and the

use of a solution of ammonium diethyldithiocarbamate in THF to remove any physisorbed Pd or Cu that

might have remained after the Sonogashira crossꢀcoupling, the surface composition was analyzed by

means of XPS. The relative atomic concentration of S was about 0.35 ± 0.07 at.%. The sulfur content

can be used to estimate the surface coverage since it can only originate from the coupled thiophene

molecules. The observed S concentration is ca. 6 times lower than the values obtained before for the

inverted reaction, wherein 2ꢀethynylthiophene is coupled to iodophenyl functionalized B:NCD

electrodes.[38] Yet, it is not possible to make clearꢀcut conclusions on the degree of functionalization for

the “normal” and “inverted” Sonogashira crossꢀcoupling reaction since in the current experiment the

ethynylphenyl layers are much thinner. Hence, a higher content of carbon atoms from the diamond bulk

is measured by XPS, resulting in a lower relative atomic S concentration.

1

9

The functionalization of the ethynylphenyl decorated diamond electrodes with pushꢀpull type dyes was

then optimized for the iodineꢀfunctionalized DTPꢀBT chromophore 4 (dye 1), since this has the lowest

synthetic complexity. An overview of the employed reaction conditions and resulting XPS analyses of the

surfaces, which gives us an idea on the degree of functionalization and possible contamination, is

presented in Table 3. Various Pdꢀligand catalyst systems were evaluated, while the amount of dye was

kept constant at 50 µmol in a 3:1 mixture of THF and DIPA. The catalytic systems increase in reactivity in

the following order: Pd(PPh ) Cl < Pd dba₃+ P(oꢀtolyl)₃< Pd(OAc)₂+ SPhos/XPhos. The first two

3

2

catalytic systems have a moderate reactivity and are commonly employed for the synthesis of conjugated

small molecules or polymers via Pd catalyzed crossꢀcoupling reactions in solution.[69] The last system,

on the other hand, comprises Pd(OAc) and either one of the two highly reactive ligands SPhos or XPhos,

2

previously employed for diamond surface functionalization via Suzuki crossꢀcoupling.[30] These

monodentate biarylphosphine ligands have been developed by the Buchwald group and facilitate the

oxidative addition step due to their electron rich nature. Since they are rather sterically demanding, only

a 12ꢀelectron Pd complex will be formed, which also increases the reaction rate of the reductive

elimination, allowing for faster crossꢀcoupling reactions.[70, 71]

Table 3 | Analysis of the surface composition by means of XPS for BDD functionalization with IꢀDTPꢀBT

chromophore 4 (dye 1) via Sonogashira crossꢀcoupling.

C

O

N

B

S

Pd

Cu

Entry

Catalyst/Ligand

Dye

at.% at.% at.% at.% at.% at.% at.%

1₁

1₂

2₁

2₂

3₁

3₂

4₁

4₂

89.2

88.1

87.4

87.9

87.4

87.6

7.8

7.2

7.5

7.6

7.9

7.5

7.4

2.4

1.9

2.2

1.7

2.2

1.8

2.1

2.4

ꢀ

0.5

0.4

0.6

0.5

0.4

0.5

ꢀ

1

Pd(PPh ) Cl (5 mol%)

3

2

(

50 µmol)

2.4

2.3

2.4

2.3

2.5

2.3

2.1

Pd dba (3 mol%) /

1

2

3

P(oꢀtolyl) (12 mol%)

(50 µmol)

3

Pd(OAc) (6 mol%) /

1

2

XPhos (6 mol%)

(50 µmol)

Pd(OAc) (6 mol%) /

1

2

XPhos (6 mol%)

(50 µmol)

ꢀ

1

5

88.6

7.2

1.1

3.0

0.1

ꢀ

(

control experiment)

(50 µmol)

General reaction conditions: CuI (10 mol%), dry THF (3 mL), diisopropylamine (1 mL), 60 °C, 18 h. Each sample was

2

measured at two different locations (0.7 × 0.3 mm ), which is indicated by the subscripts in the entry column. Only

traces of Pd and Cu were detected (≤0.05 at.%).

Looking at the relative atomic concentrations for S in Table 3, we can conclude that the degree of

functionalization is independent of the employed catalyst system. All samples were analyzed by XPS at

two different locations and surface contamination with Pd or Cu was always less than 0.05%. The

absence of surface impurities is attributed to the application of

a Pd scavenger, ammonium

diethyldithiocarbamate, which is soluble in organic solvents.[72] Since the chromophores are also easily

solubilized in organic solvents like THF, toluene or chloroform, any trapped transition metals are removed

more efficiently than with the waterꢀsoluble sodium diethyldithiocarbamate employed before. The

2

0

average S concentration (taken over all four samples) after Sonogashira crossꢀcoupling was determined

to be 0.49 ± 0.064 at.%. In comparison with the S concentration for the inverted Sonogashira crossꢀ

coupling of alkyne functionalized DTPꢀBT on iodophenyl terminated diamond (S content of 2.4 at.%), this

value is again ca. 5 times lower, which is similar to the results obtained for the thiophene model

compound.[33] The N concentration after Sonogashira crossꢀcoupling increased from an average value of

1

0

.6 ± 0.14 at.% (for B:NCDꢀPhꢀEth) to 2.08 ± 0.26 at.% (for B:NCDꢀPhꢀEthꢀDTPꢀBT). This increase of

.48 at.% is identical to the increase of the S concentration and corresponds perfectly to the 1:1 ratio of

nitrogen and sulfur atoms in the chromophore. A control experiment without addition of a Pd catalyst

system was also performed to determine whether the obtained results are due to an actual crossꢀ

coupling reaction or mere adsorption on the surface. In this case, only very low S concentrations (0.1

at.% S) were detected, likely attributable to small amounts of physisorbed molecules. Hence, we can

conclude that the dye molecules are covalently immobilized onto the surface of the functionalized

diamond electrodes upon applying Sonogashira crossꢀcoupling reactions. The changes on the diamond

surface during the threeꢀstep functionalization protocol are visualized in an overlay of the XPS survey

spectra at different stages (Figure S4).

The covalent attachment of the two other chromophores, i.e. DTPꢀBTꢀCHO (dye 2) and DTPꢀBTꢀDCV (dye

3

), onto BDD was performed with Pd dba as a catalyst and P(oꢀtolyl)₃as the ligand, since this system

2 3

resulted in the highest surface coverage (even though the difference is not very significant). The exact

reaction conditions are given in Table S2 and every sample was analyzed at three locations to estimate

the homogeneity. The relative atomic concentration of S slightly increased (in comparison with dye 1) for

immobilization of the DTPꢀBTꢀCHO chromophore, up to a value of 0.73 ± 0.058 at.%. This increased

degree of functionalization might be related to an improved oxidative addition rate of the Pd complex

during the crossꢀcoupling reaction due to the additional electronꢀdeficient aldehyde group (although the

aldehyde is small). The N concentration after Sonogashira crossꢀcoupling was about 2.27 ± 0.058 at.%,

which is an increase of 0.67 at.%. This value is again in good agreement with the 1:1 ratio of S:N in the

chromophore system. During functionalization of the diamond electrodes with dye 3 (DTPꢀBTꢀDCV),

however, an unexpected change in color of the reaction solution from blue to purple was observed almost

instantaneously. Matrixꢀassisted laser desorption/ionization timeꢀofꢀflight (MALDIꢀToF) mass analysis of

the reaction mixture revealed that the dicyanovinyl group is susceptible to a nucleophilic attack by DIPA,

which results in degradation of the chromophore, as shown in Figure S5 and S6. By employing a nonꢀ

nucleophilic such as Cs CO , no degradation occurred and DTPꢀBTꢀDCV could be successfully attached

2

3

onto the surface (Table S2, entry 8). The sulfur coverage (1.03 ± 0.23 at.%) was again somewhat

higher, which could be attributed to the even stronger electronꢀwithdrawing properties of the

dicyanovinylꢀsubstituted BT group. Another route toward obtaining DTPꢀBTꢀDCV functionalized surfaces

was explored by transforming the aldehyde group of a DTPꢀBTꢀCHO functionalized diamond electrode into

the corresponding dicyanovinyl group through a Knoevenagel condensation with malononitrile (Table S2,

entry 9). As expected, the S concentration was comparable with that for dye 2 (0.63 ± 0.058 at.%), but

a significant increase in N concentration was observed (3.00 ± 0.17 at.%). This corresponds with a S:N

2

1

ratio of 3:6.6, which is somewhat higher than the expected 3:5 ratio. Nevertheless, based on this

increased N concentration, we can confirm transformation of the surface aldehyde groups to dicyanovinyl

moieties. In Figure 4, the XPS survey spectra upon functionalization with dyes 1, 2 and 3 are shown,

together with the deconvoluted high resolution S 2p peaks. For further photoelectrochemical analysis, the

B:NCDꢀPhꢀEthꢀDTPꢀBTꢀDCV samples obtained via Sonogashira crossꢀcoupling using Cs CO₃as a base

2

have been employed, since this approach resulted in the highest surface coverage.

a) Survey Spectrum Sonogashira

C 1s

CrossꢀCoupling

No pressence

O 1s

of Pd 3d peaks

B:NCDꢀPhꢀEthꢀ

DTPꢀBTꢀDCV

N 1s

S 2s S 2p

B 1s

B:NCDꢀPhꢀEthꢀ

DTPꢀBTꢀCHO

S 2s S 2p

B:NCDꢀPhꢀEthꢀ

DTPꢀBT

S 2s S 2p

7

00

600

500

400

300

200

100

Binding energy, eV

b

)

High resolution S 2p

b

c

)

High resolution S 2p

c

d

)

High resolution S 2p

DTPꢀBT

S 2p₃_/₂

DTPꢀBTꢀCHO

DTPꢀBTꢀDCV

^S²^p3/2

S 2p₁_/₂

^S²^p1/2

S 2p₁_/₂

N-S-N

^S^Ox

S-S

^S^Ox

S-S

SO_x

S-S

1

72

170

168

166

164

162

160

170

168

166

164

162

160

170

168

166

164

162

160

Binding energy, eV

Figure 4 | a. Overlay of the XPS spectra for B:NCD electrodes functionalized with DTPꢀBT (cyan curve),

DTPꢀBTꢀCHO (purple curve) and DTPꢀBTꢀDCV (orange curve) via Sonogashira crossꢀcoupling. b/c.

Deconvoluted high resolution S 2p peak for DTPꢀBTꢀCHO and DTPꢀBTꢀDCV functionalized diamond

electrodes, respectively.

2

3

.4 Photoelectrochemistry and photovoltage measurements

Photoelectrochemical measurements were finally performed to investigate the influence of monolayer

functionalization with the different chromophores. During these measurements, functionalized diamond

electrodes were employed as the working electrode in a typical threeꢀelectrode voltammetry setꢀup. The

samples were irradiated using white light with an intensity of 1 sun in an aqueous electrolyte solution of

methyl viologen (MV²⁺), which serves as the redox electrolyte and transports photoꢀexcited electrons to

the counter electrode. The electrodes were first equilibrated in a dark environment before illuminating

them for intervals of approximately 10 seconds, as shown in Figure 5 a. Application of a negative

voltage has previously been demonstrated to positively influence the cathodic photocurrents due to the

generation of a net vectorial flow of electrons from the photocathode to the counter electrode through the

redox electrolyte.[73, 74] Accordingly,

chronoamperometric measurements.

a

negative bias of ꢀ0.2

V

was applied during the

a)

b)

Figure 5 | a. Photocurrent densities and b. chopped light polarization curves for B:NCD electrodes

functionalized with the three types of DTPꢀBTꢀX chromophores in a 5 mM methyl viologen solution (in 0.1

ꢀ

2

M Na SO ) under white light illumination (100 mW cm ; simulated 1.5G solar spectrum). For the

2

4

photocurrent densities 10 s 1:1 dark light/intervals were applied at bias of ꢀ0.2 V. The black curve for a

multilayer functionalized surface was reproduced from previous work.[33] For the chopped light

polarization curves the applied bias was changed from +0.8 to ꢀ0.4 V. In the shaded areas, the samples

were not illuminated.

The chronoamperometric measurements for the B:NCD electrodes functionalized with DTPꢀBT (cyan

curve), DTPꢀBTꢀCHO (purple curve) and DTPꢀBTꢀDCV (orange curve) all demonstrated a strong decrease

of the initial photocurrent after illuminating the samples with white light. For the B:NCDꢀPhꢀEthꢀDTPꢀBT

electrodes, this decrease occurred relatively slowly, whereas for the other two materials a much faster

decrease of the currents was observed. The observed behavior is probably related to undesired charge

recombination and is likely a direct result of the employed monolayer functionalization strategy.[40]

Multiple control experiments were performed to rule out any influence from the measurement setꢀup or

2

3

potential damaging of the surface caused by Xꢀray irradiation during the XPS measurements. Also,

identical highly doped diamond electrodes on pꢀtype Si, as previously employed for the functionalization

with DTPꢀBT chromophores (black curve) via the inverted strategy, were used in this study.[33] Hence,

the main difference between this work and our previous report is the initial surface functionalization

strategy. Through XPS analysis of the BDD electrodes decorated with TMS protected ethynylphenyl

molecules, it was already demonstrated that the density of the layer is determined rather by the bulky

protecting groups and not the phenyl linkers. We expect that gaps in between the grafted molecules

provide a pathway for recombination of the photogenerated electrons. This hypothesis is supported by

the information obtained from the choppedꢀlight polarization curves for all three types of functionalized

photocathodes, as shown in Figure 5 b. These curves demonstrate relatively large anodic photocurrents

(nꢀtype behavior) upon illumination when an external bias between +0.3 and +0.8 V is applied.

Moreover, cathodic photocurrents (pꢀtype behavior) are only obtained when the bias is equal or lower

than 0.0 V. Although anodic photocurrents have been reported previously for sensitized BDD

electrodes,[33, 39] our values are exceptionally high and the transition to cathodic photocurrents

requires remarkably low potentials. This indicates that the functionalized diamond electrodes rather

behave as photoanodes than photocathodes and that after excitation electrons are injected into diamond.

In case of electrodes functionalized via the multilayer approach, this pathway is less accessible since

there is a more dense and thicker network of phenyl molecules that potentially serves as an electronꢀ

blocking layer. To circumvent this issue for a monolayer functionalization the electron blocking properties

of B:NCD would need to be improved, e.g. by additional surface passivation in between the chromophore

molecules using various surface chemical groups of diamond (oxygenꢀrelated moieties) or some other

small molecules.

The mechanism for recombination can be that either the electrons are directly injected from the

chromophore into diamond or they are first transferred to the methyl viologen electrolyte before

recombination occurs. By employing scanning Kelvin probe (SKP) microscopy, the surface potential can

be determined in dark and under illumination, providing information on the behavior of the generated

charges in the absence of an electrolyte solution. For DTPꢀBT, DTPꢀBTꢀCHO and DTPꢀBTꢀDCV

functionalized diamond electrodes, the surface potential decreased to a value of 35 mV, 64 mV and 68

mV, respectively, as shown in Figure 6 a. This indicates that after photoexcitation, the generated holes

are successfully injected into BDD, while the electrons remain at the interface. Therefore, we can assume

that in the presence of a MV²⁺electrolyte solution, the electrons are at first transferred to the redox

electrolyte, which seems plausible since its redox potential lies well below the LUMO energy level in all

cases. Afterwards the electrolyte is reoxidized at the BDD electrode instead of the counter electrode,

leading to the observed charge recombination (as schematically shown in Figure 6 b).

2

4

3

2

1

00

50

00

50

00

a)

b)

N

B:NCDꢀPhꢀEthꢀDTPꢀBT

B:NCDꢀPhꢀEthꢀDTPꢀBTꢀCHO

B:NCDꢀPhꢀEthꢀDTPꢀBTꢀDCV

N

S

X

N

S

X

S

N

S

N

R

N

5

0

S

SOLAR

ꢀ

50

DARK

200

SIMULATOR

.5AM

DARK

1

ꢀ

100

H

0

400

600

800

1000

Time, s

Diamond Electrode

Figure 6 | a. Macroscopically measured surface potential profiles in the dark (shaded area) and under

illumination (white area) for B:NCD electrodes functionalized with the three different chromophores. b.

Schematic representation of the charge recombination that occurs for functionalized B:NCD electrodes.

After excitation, electrons are first transferred to the methyl viologen redox electrolyte before charge

recombination with the diamond electrode occurs.

4

Conclusions

In conclusion, we have established monolayer functionalized of boronꢀdoped nanocrystalline diamond

electrodes with ethynylphenyl molecules via the electrochemical diazonium grafting of 4ꢀ

(trimethylsilyl)ethynylbenzenediazonium tetrafluoroborate. Monolayer formation on nanocrystalline

diamond electrodes was confirmed by employing angular resolved XPS analysis. From the intensity of the

B 1s peaks it was determined that the density of the grafted layer is controlled by the sterical hindrance

of the trimethylsilyl protective groups, which likely results in small gaps in between the grafted

molecules. Three types of dithienopyrroleꢀbased iodinated organic donorꢀacceptor chromophores were

then synthesized and immobilized onto the surface via Sonogashira crossꢀcoupling reactions. The process

leads again to monolayer, homogeneous chromophore coverage with negligible remaining impurities from

the employed catalyst system. Atomic scale DFT calculations showed that among the investigated

molecules the DTPꢀBTꢀDCV chromophore monolayer should provide superior photoꢀgenerated charge

separation and injection to diamond based on the most suitable distribution of HOMOꢀLUMO orbitals, tight

chromophoreꢀdiamond interaction in excited state transitions, and mutual energy level alignment. This

was indeed confirmed by photovoltage and photoelectrochemical current measurements. The

experimental and theoretically predicted trends were in a good agreement.

Yet, photocurrent and photovoltage measurements also indicated that after photoexcitation of the

chromophore, the electrons were transferred to the redox electrolyte, but some were then regenerated at

the diamond electrode instead of the counter electrode. This pathway seems to be unique for the

monolayer functionalization approach and might be related to the gaps in the initially grafted layer of

ethynylphenyl molecules. Although further research is required to suppress the electrolyte recombination

mechanism (e.g. by additional surface passivation using various surface chemical groups or smaller

2

5

molecules), the presented approach with a monolayer of chromophores appears promising based on the

achieved charge photogeneration and transport characteristics, especially in the case of DTPꢀBTꢀDCV.

Supporting information

The supporting information contains the computational analysis of excited states for the B:NCDꢀ

chromophore systems, additional information obtained by XPS after surface functionalization, a MALDIꢀ

ToF mass spectrum of the degradation product of DTPꢀBTꢀDCV in the presence of DIPA and ¹H/¹³C NMR

spectra of the DTPꢀBT, DTPꢀBTꢀCHO and DTPꢀBTꢀDCV chromophores.

Author Information

Corresponding Authors:

*

Eꢀmail: wouter.maes@uhasselt.be. Phone: +32ꢀ11ꢀ268312

Eꢀmail: rezekboh@fel.cvut.cz. Phone: +420ꢀ22435ꢀ2330

ORCID:

Jorne Raymakers: 0000ꢀ0003ꢀ1468ꢀ3108

Shannon S. Nicley: 0000ꢀ0002ꢀ5960ꢀ7873

Ken Haenen: 0000ꢀ0001ꢀ6711ꢀ7367

Wouter Maes: 0000ꢀ0001ꢀ7883ꢀ3393

Bohuslav Rezek: 0000ꢀ0002ꢀ0378ꢀ4598

Notes:

The authors declare that there is no conflict of interests.

Acknowledgements

The authors thank UHasselt, the Research Foundation – Flanders (FWO Vlaanderen), and the European

Regional Development Fund project CZ.02.1.01/0.0/0.0/15_003/0000464 (CAP) for financial support.

H.K. and L.K. acknowledge the financial support from the Ministry of Education, Youth and Sports of the

Czech Republic – project ProꢀNanoEnviCz (Reg. No. CZ.02.1.01/0.0/0.0/16_013/0001821). Part of this

work has been done using computational resources of the "Research Center for Informatics" supported by

the EU (reg. no. CZ.02.1.01/0.0/0.0/16_019/0000765). This work was also supported by The Ministry of

Education, Youth and Sports through the Large Research Infrastructures IT4Innovations and LNSM. J.

Raymakers and F. Verstraeten thank the FWO for their PhD fellowships.

2

6

References

[1] F.P. Bundy, H.T. Hall, H.M. Strong, R.H. Wentorf, Manꢀmade diamonds, Nature, 176 (1955) 51ꢀ55.

[2] S. Yesudhas, F.P. Bundy, R.C. DeVries, Diamond: Highꢀpressure synthesis, Elsevier, 2016.

[3] S. Matsumoto, M. Hino, T. Kobayashi, Synthesis of diamond films in a rf induction thermal plasma,

Appl Phys Lett, 51 (1987) 737ꢀ739.

4] M. Kamo, Y. Sato, S. Matsumoto, N. Setaka, Diamond synthesis from gasꢀphase in microwave

plasma, J Cryst Growth, 62 (1983) 642ꢀ644.

5] S. Matsumoto, Y. Sato, M. Tsutsumi, N. Setaka, Growth of diamond particles from methaneꢀhydrogen

gas, J Mater Sci, 17 (1982) 3106ꢀ3112.

6] S.S. Nicley, S. Drijkoningen, P. Pobedinskas, J. Raymakers, W. Maes, K. Haenen, Growth of boronꢀ

[

doped diamond films on goldꢀcoated substrates with and without gold nanoparticle formation, Cryst.

Growth Des., 19 (2019) 3567ꢀ3575.

[

7] L. Schafer, M. Hofer, R. Kroger, The versatility of hotꢀfilament activated chemical vapor deposition,

Thin Solid Films, 515 (2006) 1017ꢀ1024.

8] S. Szunerits, Y. Coffinier, R. Boukherroub, Diamond nanowires: A novel platform for electrochemistry

and matrixꢀfree mass spectrometry, Sensors, 15 (2015) 12573ꢀ12593.

9] H. Kato, J. Hees, R. Hoffmann, M. Wolfer, N. Yang, S. Yamasaki, C.E. Nebel, Diamond foam

electrodes for electrochemical applications, Electrochem Commun, 33 (2013) 88ꢀ91.

10] V. Petrák, Z. VlčkováꢀŽivcová, H. Krýsová, O. Frank, A. Zukal, L. Klimša, J. Kopeček, A. Taylor, L.

Kavan, V. Mortet, Fabrication of porous boronꢀdoped diamond on SiO2 fiber templates, Carbon, 114

[

(

2017) 457ꢀ464.

[11] T. Irifune, A. Kurio, S. Sakamoto, T. Inoue, H. Sumiya, Materials: Ultrahard polycrystalline diamond

from graphite, Nature, 421 (2003) 599ꢀ600.

12] K. Yoshida, H. Morigami, Thermal properties of diamond/copper composite material, Microelectron

Reliab, 44 (2004) 303ꢀ308.

13] C.H.Y.X. Lim, Y.L. Zhong, S. Janssens, M. Nesladek, K.P. Loh, Oxygenꢀterminated nanocrystalline

diamond film as an efficient anode in photovoltaics, Adv. Funct. Mater., 20 (2010) 1313ꢀ1318.

[

14] N. Y., Novel aspects of diamond: From growth to applications Springer, 2014.

[15] T.R. Rosa, F.S. Betim, R.D. Ferreira, Development and application of a labmade apparatus using

openꢀsource "arduino" hardware for the electrochemical pretreatment of boronꢀdoped diamond

electrodes, Electrochim. Acta, 231 (2017) 185ꢀ189.

[

16] P.A. Nistor, P.W. May, Diamond thin films: Giving biomedical applications a new shine, J. R. Soc.

Interface, 14 (2017) 20170382.

[

(

[

17] J.B. Cui, J. Ristein, L. Ley, Electron affinity of the bare and hydrogen covered single crystal diamond

111) surface, Phys Rev Lett, 81 (1998) 429ꢀ432.

18] J.V. Macpherson, A practical guide to using boron doped diamond in electrochemical research, Phys.

Chem. Chem. Phys., 17 (2015) 2935ꢀ2949.

19] K. Muzyka, J.R. Sun, T.H. Fereja, Y.X. Lan, W. Zhang, G.B. Xu, Boronꢀdoped diamond: Current

progress and challenges in view of electroanalytical applications, Anal MethodsꢀUk, 11 (2019) 397ꢀ414.

20] T. Watanabe, K. Akai, Y. Einaga, The reduction behavior of free chlorine at boronꢀdoped diamond

electrodes, Electrochem Commun, 70 (2016) 18ꢀ22.

21] T.A. Ivandini, Y. Einaga, Polycrystalline boronꢀdoped diamond electrodes for electrocatalytic and

electrosynthetic applications, Chem. Commun., 53 (2017) 1338ꢀ1347.

22] A. Dirany, I. Sires, N. Oturan, A. Ozcan, M.A. Oturan, Electrochemical treatment of the antibiotic

sulfachloropyridazine: kinetics, reaction pathways, and toxicity evolution, Environ. Sci. Technol., 46

[

(

2012) 4074ꢀ4082.

[23] M. Tomisaki, S. Kasahara, K. Natsui, N. Ikemiya, Y. Einaga, Switchable product selectivity in the

electrochemical reduction of carbon dioxide using boronꢀdoped diamond electrodes, J. Am. Chem. Soc.,

41 (2019) 7414ꢀ7420.

24] S. Szunerits, C.E. Nebel, R.J. Hamers, Surface functionalization and biological applications of CVD

diamond, MRS Bull., 39 (2014) 517ꢀ524.

25] J. Raymakers, K. Haenen, W. Maes, Diamond surface functionalization: From gemstone to

photoelectrochemical applications, J. Mater. Chem. C, 7 (2019) 10134ꢀ10165.

26] H. Ye, L. Shen, S. Zhang, X. Li, F. Yu, R. Diao, J. Hua, Enhanced photocurrent via πꢀbridge

extension of perylenemonoimideꢀbased dyes for pꢀtype dyeꢀsensitized solar cells and

photoelectrochemical cells, ACS Omega, 3 (2018) 14448ꢀ14456.

27] P.ꢀN. Volpe, J. Pernot, P. Muret, F. Omnès, High hole mobility in boron doped diamond for power

device applications, Appl Phys Lett, 94 (2009) 092102.

28] T. Strother, T. Knickerbocker, J.N. Russell, J.E. Butler, L.M. Smith, R.J. Hamers, Photochemical

functionalization of diamond films, Langmuir, 18 (2002) 968ꢀ971.

29] T.C. Kuo, R.L. McCreery, G.M. Swain, Electrochemical modification of boronꢀdoped chemical vapor

deposited diamond surfaces with covalently bonded monolayers, Electrochem Solid St, 2 (1999) 288ꢀ290.

30] W.S. Yeap, D. Bevk, X. Liu, H. Krysova, A. Pasquarelli, D. Vanderzande, L. Lutsen, L. Kavan, M.

Fahlman, W. Maes, K. Haenen, Diamond functionalization with lightꢀharvesting molecular wires:

1

[

2

7

Improved surface coverage by optimized Suzuki crossꢀcoupling conditions, RSC Adv., 4 (2014) 42044ꢀ

2053.

31] W.S. Yeap, X. Liu, D. Bevk, A. Pasquarelli, L. Lutsen, M. Fahlman, W. Maes, K. Haenen,

4

[

Functionalization of boronꢀdoped nanocrystalline diamond with N3 dye molecules, ACS Appl. Mater.

Interfaces, 6 (2014) 10322ꢀ10329.

[

32] H. Krysova, Z. VlckovaꢀZivcova, J. Barton, V. Petrak, M. Nesladek, P. Cigler, L. Kavan, Visibleꢀlight

sensitization of boronꢀdoped nanocrystalline diamond through nonꢀcovalent surface modification, Phys.

Chem. Chem. Phys., 17 (2015) 1165ꢀ1172.

[

33] J. Raymakers, H. Krysova, A. Artemenko, J. Cermak, S.S. Nicley, P. Verstappen, S. Gielen, A.

Kromka, K. Haenen, L. Kavan, W. Maes, B. Rezek, Functionalization of boronꢀdoped diamond with a

pushꢀpull chromophore via Sonogashira and CuAAC chemistry, RSC Adv., 8 (2018) 33276ꢀ33290.

[

34] S. Nakabayashi, N. Ohta, A. Fujishima, Dye sensitization of synthetic pꢀtype diamond electrode,

Phys. Chem. Chem. Phys., 1 (1999) 3993ꢀ3997.

35] Y.L. Zhong, K.P. Loh, A. Midya, Z.ꢀK. Chen, Suzuki coupling of aryl organics on diamond, Chem

Mater, 20 (2008) 3137ꢀ3144.

36] Y.L. Zhong, A. Midya, Z. Ng, Z.K. Chen, M. Daenen, M. Nesladek, K.P. Loh, Diamondꢀbased

molecular platform for photoelectrochemistry, J. Am. Chem. Soc., 130 (2008) 17218ꢀ17219.

37] H. Krysova, L. Kavan, Z. VlckovaꢀZivcova, W.S. Yeap, P. Verstappen, W. Maes, K. Haenen, F. Gao,

[

C.E. Nebel, Dyeꢀsensitization of boronꢀdoped diamond foam: champion photoelectrochemical

performance of diamond electrodes under solar light illumination, RSC Adv., 5 (2015) 81069ꢀ81077.

[38] J. Raymakers, A. Artemenko, S.S. Nicley, P. Štenclová, A. Kromka, K. Haenen, W. Maes, B. Rezek,

Expanding the scope of diamond surface chemistry: Stille and Sonogashira crossꢀcoupling reactions, J

Phys Chem C, 121 (2017) 23446ꢀ23454.

[39] H. Krysova, J. Barton, V. Petrak, R. Jurok, M. Kuchar, P. Cigler, L. Kavan, Efficiency and stability of

spectral sensitization of boronꢀdopedꢀdiamond electrodes through covalent anchoring of a donorꢀacceptor

organic chromophore (P1), Phys. Chem. Chem. Phys., 18 (2016) 16444ꢀ16450.

[40] J. Bartoň, H. Krýsová, P. Janda, H. Tarábková, P. Ashcheulov, V. Mortet, A. Taylor, J. Vávra, P.

Cígler, L. Kavan, Chemical modification of diamond surface by a donor–acceptor organic chromophore

(

[

P1): Optimization of surface chemistry and electronic properties of diamond, Appl. Mater. Today, 12

2018) 153ꢀ162.

41] P. Allongue, C.H. de Villeneuve, G. Cherouvrier, R. Cortes, M.C. Bernard, Phenyl layers on HꢀSi(111)

by electrochemical reduction of diazonium salts: monolayer versus multilayer formation, J Electroanal

Chem, 550 (2003) 161ꢀ174.

[

42] J. Wang, M.A. Firestone, O. Auciello, J.A. Carlisle, Surface functionalization of ultrananocrystalline

diamond films by electrochemical reduction of aryldiazonium salts, Langmuir, 20 (2004) 11450ꢀ11456.

43] J. Pinson, F. Podvorica, Attachment of organic layers to conductive or semiconductive surfaces by

reduction of diazonium salts, Chem. Soc. Rev., 34 (2005) 429ꢀ439.

44] C. Jiang, S.M. Silva, S. Fan, Y. Wu, M.T. Alam, G. Liu, J.J. Gooding, Aryldiazonium salt derived

mixed organic layers: From surface chemistry to their applications, J Electroanal Chem, 785 (2017) 265ꢀ

78.

45] H. Uetsuka, D. Shin, N. Tokuda, K. Saeki, C.E. Nebel, Electrochemical grafting of boronꢀdoped

singleꢀcrystalline chemical vapor deposition diamond with nitrophenyl molecules, Langmuir, 23 (2007)

466ꢀ3472.

46] S.Q. Lud, M. Steenackers, R. Jordan, P. Bruno, D.M. Gruen, P. Feulner, J.A. Garrido, M. Stutzmann,

[

2

[

3

[

Chemical grafting of biphenyl selfꢀassembled monolayers on ultrananocrystalline diamond, J. Am. Chem.

Soc., 128 (2006) 16884ꢀ16891.

[

47] Y.R. Leroux, H. Fei, J.M. Noel, C. Roux, P. Hapiot, Efficient covalent modification of a carbon surface:

use of a silyl protecting group to form an active monolayer, J. Am. Chem. Soc., 132 (2010) 14039ꢀ

4041.

48] Y.R. Leroux, P. Hapiot, Nanostructured monolayers on carbon substrates prepared by electrografting

of protected aryldiazonium salts, Chem Mater, 25 (2013) 489ꢀ495.

49] H.I. Lu, C.W. Lu, Y.C. Lee, H.W. Lin, L.Y. Lin, F. Lin, J.H. Chang, C.I. Wu, K.T. Wong, New molecular

1

[

donors with dithienopyrrole as the electronꢀdonating group for efficient smallꢀmolecule organic solar cells,

Chem Mater, 26 (2014) 4361ꢀ4367.

[

50] Y. Fang, A.K. Pandey, D.M. Lyons, P.E. Shaw, S.E. Watkins, P.L. Burn, S.C. Lo, P. Meredith, Tuning

the optoelectronic properties of nonfullerene electron acceptors, ChemPhysChem, 16 (2015) 1295ꢀ1304.

[

51] P. Hohenberg, W. Kohn, Inhomogeneous Electron Gas, Phys. Rev., 136 (1964) B864ꢀB871.

[52] W. Kohn, L.J. Sham, SelfꢀConsistent Equations Including Exchange and Correlation Effects, Phys.

Rev., 140 (1965) A1133ꢀA1138.

53] A.D. Becke, Densityꢀfunctional exchangeꢀenergy approximation with correct asymptotic behavior,

Phys. Rev. A, 38 (1988) 3098ꢀ3100.

54] C. Lee, W. Yang, R.G. Parr, Development of the ColleꢀSalvetti correlationꢀenergy formula into a

functional of the electron density, Phys. Rev. B 37 (1988) 785ꢀ789.

[

[55] A.D. Becke, Density functional thermochemistry. III. The role of exact exchange, J. Chem. Phys., 98

(1993) 5648ꢀ5652.

2

8

Article Doi

Photogenerated charge collection on diamond electrodes with covalently linked chromophore monolayers

DOI: 10.1016/j.electacta.2020.135762

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.electacta.2020.135762

Full text of DOI:10.1016/j.electacta.2020.135762

Products guided by the article

R&D Labs maybe for 75867-39-9

Relevant to this article

Hot Product