Thin Solid Films 518 (2009) 138–146

Contents lists available at ScienceDirect

Thin Solid Films

j ourna l homepage: www.e lsev ie r.com/ locate / ts f

Photoelectrochemical reduction of nitrate on p-Si coated with metallic Re thin films

Eduardo Muñoz a,⁎, Ricardo Schrebler a, Rodrigo Henríquez a, Cristopher Heyser a,Patricia A. Verdugo b, Ricardo Marotti c

a Instituto de Química, Facultad de Ciencias, Pontificia Universidad Católica de Valparaíso, Casilla 4059, Valparaíso, Chileb Departamento de Química y Bioquímica, Facultad de Ciencias, Universidad de Valparaíso, Av. Gran Bretaña 1111, Valparaíso, Chilec Instituto de Física, Facultad de Ingeniería, Universidad de La República, Herrera y Reissig 565, C.C. 30, 11000 Montevideo, Uruguay

⁎ Corresponding author. Tel./fax: +56 32 3170861.E-mail address: eduardo.munoz.c@ucv.cl (E. Muñoz)

0040-6090/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.tsf.2009.06.010

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 January 2009Received in revised form 1 June 2009Accepted 4 June 2009Available online 12 June 2009

Keywords:Photoelectrochemical reductionNitrateRheniumElectrodeposition, siliconIntensity modulated photocurrent spectroscopyElectrical properties and measurements

In this study we examined the rhenium electrodeposition process onto p-Si(100) from acidic media. Thestudy was carried out by means of cyclic voltammetry and the potential-steps method from which thecorresponding nucleation and growth mechanism were determined. Both methods were performed underillumination using a solar simulator for electron photogeneration. A 3D progressive nucleation, diffusion-controlled growth of rhenium films was found. Likewise, a morphologic analysis was completed for thedeposits obtained at different potential values by means of atomic force microscopy. An energeticcharacterization through capacitance measurements (Mott–Schottky plots and parallel capacitance) of the p-Si/NO3

− and p-Si/Re/NO3− interfaces was done.

The photoelectrochemical reduction of nitrate ions, PERN, on the different p-Si/Re electrode systemssynthesized was studied. An overpotential decrease of 0.3 V and a photocurrent increase for the PERN on p-Si(100)/Re electrode systems compared with p-Si(100) and metallic Re was found. Finally, the kineticparameters of the cathodic reactions in the p-Si and p-Si/Re acidic media were estimated using intensitymodulated photocurrent spectroscopy. A brief analysis from this technique was done. According to theseresults, the p-Si/Re electrode system could be a potential photoelectrocatalyst for the PERN.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Nitrate ions are found naturally in vegetables. Their presence is aresult of the nitrogen cycle in which plants assimilate inorganicnitrogen, in the form of nitrates, for use in vegetable-protein synthesis.Nitrates are also found in drinkable water, principally due to a naturalpresence or as a consequence of contamination from agriculturalpractices (fertilizers and manure). Nitrates are also employed asadditives in cured flesh, due to their bacteriostatic properties(prevents botulism) and their capacity to fix red color. Therefore,there are three principal sources of nitrate ingestion in our diet:vegetables, water and cured meat. Vegetables constitute the mostimportant source of nitrates in the diet, contributing 80% of dailyingestion. The public and the government concern regarding nitratepresence in food has two aspects: first, nitrite toxicity, which derivatesfrom nitrate reduction. This species is capable of oxidizing thehemoglobin in globules red, transforming it into methemoglobin [1].Methemoglobin has difficulty transporting oxygen to the cells. Thesecond concern is that nitrates and nitrites can form N-nitrosoaminesand N-nitrosoamides (precursors to cancer causing agents) [2].

.

ll rights reserved.

Water can be considered the most valued resource on earth,therefore, it is necessary to propose methodologies that allow for thereduction of nitrate levels from water sources: drinkable water,residual water, ionic exchange resins, reverse osmosis, etc. In thiscontext, the direct reduction of nitrate onmetallic electrodes has beeninvestigated. This research focused on the thermodynamic, kinetic,and mechanistic aspects of nitrate reduction on this type of electrode.The most commonly used electrode has been Pt [3,4], due to itsevident electrocatalytic effect. This is attributed to the vacancy thatoffers a d band, which converts in a metal with a strong adsorbcapacity in front of chemical species such as reactants andintermediates.

From the thermodynamic point of view, if we analyze the standardequilibrium potentials we have [5]

NO−3 →

0:76V NO2→1:12V HNO2→

1:00V NO→1:59V N2O→1:77 N2→

0:27 NHþ4

And it is possible to go fromNO−3 →

0:96V NO, and fromNO−3 →

1:25V N2.From this, it can be concluded that independent of the reductionproduct, this must be attained in the interval of water dischargepotentials. However, in practice, the nitrate reduction requires largeoverpotentials. This means that the problem in this reaction isprincipally due to kinetic aspects. Due to this, the most commonlyusedmaterials in this reaction have been Pt, Pd, Rh, and Ru [3,4,6] due to

139E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

their well-known electrocatalyst character. Also, nitrate reduction hasbeen studied on naked semiconductor materials as well as thosemodifiedwithmetals deposited on its surface. In fact, studies have beendone on nitrate and nitrite reduction to ammonia on ZnS [7], Ru/TiO2

[6,8], Ag/TiO2 [9,10], and Cu/TiO2 [11]. In this context, the p-Si surfacemodification bymeans of Rewas studied and tested using the hydrogenevolution reaction (HER) [12]. The data indicated that the modifiedsubstrates present promissory results due to an overpotential decreaseof 0.25V and a current increase of one order ofmagnitude. This behaviorcould be associated with a photoelectrocatalytic effect of this electrodesystem(p-Si/Re) in theHER. Therefore, thegoal of this current studywasto investigate the photoelectrochemical reduction of nitrate ions (PERN)on theelectrode systemformedbyp-Si, as a semiconductor, andmetallicrheniumfilms deposited bymeans of an electrodeposition technique onthis substrate.

2. Experimental details

Electrodeposition was performed on monocrystalline p-Si (100)with a resistivity between 0.01–0.3 Ω cm (NA≅5×1017 cm−3), B-doped and polished/etched surfaces (Int. Wafer Service, CA, USA). Thesilicon wafer was cut into rectangles (1.0×2.5 cm2) that weredegreased in boiling acetone for 10 min. Next they were cleanedultrasonically for 10 min in acetone, then ethanol and finally in water.The electrodes were treated for 10minwith a 1:1 H2SO4:H2O2mixtureheated to 80 °C, in order to remove any trace of heavy metals andorganic species. Afterwards, the oxide film was removed by etchingwith a 4% HF solution for 2 min and thoroughly rinsed with ultra purewater. The ohmic contact was made with InGa eutectic on the etchedface of the samples and the electrodes were mounted onto a Teflonholder. The silicon area exposed to the solutionwas 1.0 cm2. Before theelectrochemical experiments, the electrode surface was again etchedfor 2 min in 4% HF solution. For each measurement, a new p-Si(100)electrode was used, due to the well known fact that some metals candiffuse toward the inside of silicon [13,14].

For the voltammetric studies on rhenium electrodeposition andthe photoelectrochemical reduction of nitrates ions, a platinum wirewas used as a counterelectrode and a mercury/mercury sulfateelectrode (Hg/Hg2SO4, K2SO4 (saturated), 0.640 V vs. normal hydro-gen electrode) was used as a reference electrode. All the potentialsreported in this study refer to this reference electrode.

The electrolytic solutions were prepared using distilled anddeionized water with a resistivity of 18 MΩ cm. Analytical gradereagents from Fluka (NH4ReO4) and Merck (NaNO3, H2SO4) wereused.

Two electrochemical cells were used in all experiments. A pulsestep potential technique was employed on cell (1) for the rheniumelectrodeposition. The choroamperograms were recorded inNH4ReO4/H2SO4 solutions. In cell (2) the photoelectrocatalytic activityof the p-Si(100)/Re electrodic system in the PERN was tested. Theelectrodeposition baths were aqueous solutions constituted of 50 mMNH4ReO4 in 0.1 M H2SO4 in Cell (1). For the PERN process an aqueoussolution of 0.1 M NaNO3 at pH=2 in Cell (2) was used.

The following procedures have been employed. The p-Si(100)treated electrode was mounted in a teflon holder and transferred toCell (1). In a first stagewas realized a voltammetric analysis and later anew p-Si substrate for the rhenium electrodeposition process wasused (choroamperometric study). These last experiments werestopped after a constant total cathodic charge passed between0.02 C cm−2–4.5 C cm−2. Considering that in these conditions thefaradaic efficiency for rhenium electrodeposition corresponds to approxi-mately 16% [15,16], these charges are equivalent to 1.2–281monolayers. Ifwe consider that eachmonolayer is constituted only of metallic rhenium,the thickness of the films can be estimated between 0.3 nm–77 nmapproximately. However, it must be considered that the rhenium filmsobtained are not continuous (atomic force microscopy images, AFM);

therefore this is only estimation. After this procedure, the electrode waswashed and transferred to Cell (2) in order to record the cyclicvoltammogram corresponding to a p-Si(100)/Re electrodic system forthe PERN (photoelectrocatalytic activity test).

For the electrochemical impedance spectroscopy (EIS) measure-ments (Mott–Schottky plots) a new p-Si sample, and a p-Si/Reelectrode system in Cell (1) with 0.1 M NaNO3 at pH=2 solution wasemployed. The experiments were recorded at potentials valuesfrom−0.8 V to−0.3 V, at an AC frequency of 25 kHz and AC amplitudeof 10 mV.

Cyclic voltammetry experiments (NH4ReO4/H2SO4 solutions andPERN) were carried out at room temperature and at a scan rate of0.010 V s−1. All the measurements were carried out under illumina-tion conditions, unless the opposite was indicated, using a SolarSimulator system. A Xenon-UV Lamp of 150 W (Newport OrielInstruments 6254) mounted in a lamp holder (Oriel 67005) wasconnected to the cell through a water filter (Oriel 61945) and a 1 mlength optical fiber (Oriel 77578) was used. A power supply of 40–300 W (Newport Oriel Instruments 69907) was used to generate thearc in the lamp. The illumination power was quantified inside of thecell by means of an energy radiant meter (Oriel 70260). The samplesilluminated with the Xe-UV lamp reached a light intensity of 0.1 suns(1 sun=1 kW m−2). A pure argon stream was passed through thesolution for 30 min before measurements, and over the solutionwhenthe experiments were under way.

The electrochemical measurements (cyclic voltammetry andchronopotentiometry) were done using Princeton Applied Research(PAR)model 273A equipment. The EIS measurements (Mott–Schottkyplots) were made using a Zahner model IM6e equipment. All samplesused for ex situ AFM were prepared in the electrochemical Cell (1).The AFM images were obtained with a Digital Instrument NanoscopeIIIa series employed in tapping mode at a scan rate of 0.02 µm/s. Thetips used were 1–10 Ω cm Phosphorus (n) doped Si, with resonancefrequency between 200–400 kHz, model Olympus AC240TS fromAsylum Research.

The Intensity Modulated Photocurrent Spectroscopy (IMPS)experimental setup consisted of the Cell (2) illuminated by a 5 mWred (635 nm) Lasiris laser, provided by Stocker Yale as the light source.The laser was modulated by a Model 2030 Global SpecialitiesInstruments Function Generator. A 10X beam expander provided byEdmund Optics was used to uniformly illuminate the cell window. Theilluminated area over the sample at the cell was 50 mm2. A circularneutral density filter with variable optical density was used to adjustthe light intensity reaching the sample. This intensity mean value wasmeasured with a 70260 Oriel radiant power meter with a 70288silicon probe, while a 53375 silicon photodiode provided by EdmundOptics was used to detect the instantaneous value of this signal. Theintensity mean value was kept at ~300 µW/cm−2, which was theminimum value [17] that allowed a good signal to noise ratio.

A two channel lock-in amplifier Model 5210 from Signal Recoverywasused tomeasure the complex (modulus and phase)AC current fromthe photodiode. In order to calibrate the lock-in before each measure-ment, a PDA 750 Photodiode Preamplifier from Terahertz TechnologyInc.was used. An analog to digital converter and a PC acquired the signalfrom the preamplifier to assure a good sinusoidal shape. This shapewasadjusted with the DC offset of the function generator, whose amplitudewas kept at itsminimumvalue. This adjustmentwas carried out at 1 kHzwhere the preamplifier had a constant response. This preamplifier wasnot used when measuring signal at varying frequency. The photodiodewas checked to have an almost flat response fromDC to 30 kHz, with nodelay time, by comparing its response against a very fast 1 GHzavalanche photodiode.

A second two channel lock-in amplifier Model 5210 from EG&GPAR measured the photocurrent from the electrochemical cell. Thissignal was obtained from the current intensity monitor output of thepotentiostat (PAR model 273A). This instrument was configured to

Fig. 1. Potentiodynamic j/E profile of a p-Si(100) electrode in 50 mM NH4ReO4 at pH 2under solar simulator illumination of 0.1 suns. The inset shows the applied E/tperturbation program. Scan rate, 0.010 V s−1.

140 E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

assure its maximum frequency response (N1 MHz for −3 dB). Bothlock-in amplifiers can work from 0.5 to 120 kHz.

3. Results and discussion

3.1. Cyclic voltammetry analysis of the p-Si(100)

Fig. 1 shows the potentiodynamic jph/E profile for a p-Si(100)electrode in 50 mM ReO4

− solution (NH4ReO4/H2SO4 at pH 2) underillumination conditions using the solar simulator. This profile wasmeasured considering an E/t program that begins from open circuitpotential (Ei=−0.4 V) and moves towards the negative direction andthen in the reverse direction (the arrows inserted in the figure indicatethe scan direction). In the initial negative scan, a cathodic contributionappears at−0.5 V, attaining amaximumat−0.84V,which is attributedto the perrhenate electroreduction. According to X-ray photoelectronspectroscopy measurements it is indicated that at pH 2 and over, in thissubstrate, the species formed aremetallic Re, ReO2, and ReO3 [16,18], thesame results as observed by us in a previous study.

After this process, a cathodic contribution appears that corre-sponds to the HER. This reaction does not occur on the silicon surface,

Fig. 2. (a) jPh/t transients obtained on a p-Si(100) electrode in 50 mM NH4ReO4 at pH 2, f(b) Nondimensional plots for the jPh/t transients from (a). Dashed and dotted lines correspoand progressive nucleation mechanisms, respectively. Light intensity of 0.1 suns.

but rather, on the nascent rhenium species. In the reverse scan, thepeak localized at−0.8 V is attributed to the formation and growth of arhenium nucleus onto the film previously deposited.

3.2. Study of nucleation and growth mechanisms of rhenium ontop-Si(100)

The jph/t transients obtained by the potential step method wereperformed in the potential region where the electrodeposition ofrhenium species occurs. The potential steps were carried out from aninitial potential Ei=−0.40 V to a final deposition potential (Ed)ranging between −0.650 V≥Ed≥−0.800 V. Fig. 2(a) shows the jph/ttransients corrected by subtracting the double layer contribution,considering the induction time as point (t0, j0). In the jph/t transients,it can be seen that an initial increase in the photocurrent produces anincrease in the electrode area. In transients 2–4, a maximum (jphmax)is reached, and then the photocurrent density begins to decay. Thislast effect is not observed at the initial potential (−0.650 V) due to thefact that a diffusion control is not reached at this value. In order todetermine the nucleation and growth mechanisms of rhenium on p-Si(100), we sought to distinguish between instantaneous and progres-sive nucleation; the experimental data are represented in a non-dimensional form. Fig. 2(b) represents the data from Fig. 2(a) oncoordinates (I/Im)2 versus t/tm. The above mentioned is in accordancewith the theoretical models of nucleation and diffusion controlledgrowth of hemispherical clusters. The models for instantaneous andprogressive nucleation are given by Eqs. (1a) and (1b), respectively[19,20],

iim

� �2=

1:9542t = tm

� 1−exp −1:2564 � ttm

� �� �� �2ð1aÞ

iim

� �2=

1:2254t = tm

� 1−exp −2:3367 � ttm

� �2� �� �2

ð1bÞ

where im is the current and tm is the time coordinate of a chroampero-metric peak.

As can be seen from Fig. 2(b) in the initial portion, theexperimental data follow a three-dimensional progressive nucleationmechanism. After tmax, deviations from this model are not observedconsidering the time scale of the measurement. However, in previousstudies deviations from the model after tmax have been observed,

or potential steps from −0.40 V to a deposition potential (Ed) indicated in the figure.nd to the calculated curve for the growth laws for 3D diffusion controlled instantaneous

141E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

attributed to the presence of a parallel reaction to the film formation,such as proton reduction [12,21,22].

3.3. Morphological study by atomic force microscopy, AFM

As discussed above, from the j/t transient analysis, a three-dimensional progressive nucleation mechanism has been determinedand this should be reflected in the morphology of the film. Tocorroborate this observation, a morphological study of the films wascarried out by AFM in tapping mode technique. Fig. 3 shows10 µm×10 µm AFM images of rhenium films onto p-Si(100) depositedat different potential values under solar simulator illumination. Theelectrodeposition process was stopped after a constant charge value of20 mC cm−2 had been reached, which is approximately 75 mono-layers. Considering the 3D view (Fig. 3a–d), isolated 3D nuclei ofdifferent sizes with a hemispherical geometry can be seen. If weconsider the 2D view (Fig. 3a–d), the nucleus density increased whenthe potential was more negative. Given that the charge value passagewas always constant, when nucleus density increased, their sizedecreased. These features are consistent with a three-dimensionalprogressive nucleation mechanism.

3.4. Analysis of the p-Si and p-Si/Re in nitrate media: high and lowfrequency capacitance measurements (Mott–Schottky plots and surfacestates density)

Fig. 4a shows theMott–Schottky (MS) plots andFig. 4b showsparallelcapacitance measurements for the p-Si and p-Si/Re electrodes in 0.1 MNaNO3 at pH 2 solutions. In the case of MS plots, the capacitances havebeen measured at an AC frequency of 25 kHz in darkness and with ACamplitude of 10 mV. The parallel capacitance was measured at differentfrequencies, from 1 kHz to 10 Hz, with the same AC amplitude. The p-Si/Re sample corresponding to an electrode system with rheniumelectrodeposited at −0.800 V, with a total cathodic charge densitypassed of 20mC cm−2 (see Fig. 3d). FromMS plots for the two electrodicsubstrates, negative slopes were obtained, indicating p-type character.The systemspresent a linear behavior in the range from−0.8V to−0.3Vfor the p-Si sample and between −0.65 V to −0.4 V for the p-Si/Re

Fig. 3. 2DAFM images 10×10 µm of rhenium thin films onto p-Si(100), deposited at differentsimulator illumination of 0.1 suns: (a) −0.650 V; (b) −0.700 V; (c) −0.750 V; (d) −0.800

sample. The flat band potential,VFB, for a givenp-type semiconductor canbe obtained from the intercept of the MS plot formed at C−2=0 and theNA can be calculated from the following equation [17,23]

1C2 =

2eee0NA

−V + VFB − kbTe

� �ð2Þ

where C is the space charge capacitance, ε the dielectric constant of thesemiconductor (11.9), εo the permittivity of the vacuum (8.85×10−14 Fcm−1), NA the majority carrier density, V the applied potential, kbBoltzmann's constant, and T the temperature. From the slope of the MSplots, by using Eq. (2), themajority carrier concentrationwas calculated.In our case, for the p-Si and p-Si/Re samples the majority carriers, NA,were 3.8×1017 cm−2 and 2.2×1017 cm−2, respectively, which are inagreement with that given by the supplier. The parallel capacitancemeasurements were used to determine surface state density and theinfluence of the Re deposited (Fig. 4(c–d)). These results were obtainedby subtracting the space charge capacitance (measured at 10 kHz) fromthose capacitance measurements at low frequencies [24]. Fig. 4(c–d)indicates that when the frequency decreases, a broad capacitancemaximumdevelops that is associatedwith thepresence of electronicallyactive surface states [24]. Also important to notice is that when Re ispresent, an increase in the capacitance maxima occurs. From themaximumvalues of the ΔCp/E curves (CSS,max), the surface state densitycan be calculated with the following equation,

CSS;max =e2

4kTNSS ð3Þ

where NSS is the surface state density, and e, k and T represent theelectron charge, the Boltzmann constant, and the temperature,respectively. Furthermore, if we consider the influence of Helmholtzlayer capacity to benegligible, theenergetic positionof the surface states(ESS) can be obtained from the potential value in the ΔCp/E curves. Thispotential value is located at−0.740 V for both systems, and the densityof the surface states were 1.22×1014 cm−2 and 3.7×1014 cm−2 for p-Siand p-Si/Re substrates, respectively. For a better view of the differencesbetween p-Si and p-Si/Re substrates the inset of Fig. 4(c) shows curvesof the Lorentzian fit at low frequencies.

potential values until reaching a total constant charge value of 20mC cm−2. Under solarV.

Fig. 4. Capacitance measurements at high and low frequencies for p-Si and p-Si/Re in 0.1 M NaNO3 solution at pH 2. Top: Mott–Schottky plots of p-Si(100) (a) and p-Si/Re (b); themeasurements were done at a AC frequency of 25 kHz. Bottom: parallel capacitance measurements of p-Si(100) (c) and p-Si/Re (d); (-◯-) 1 kHz; (-△-) 333 Hz; (-▽-) 100 Hz; (-v-)10 Hz. All samples measured at an AC amplitude of 10 mV. In the inset (c) Lorentzian fit to the 10 Hz measurements.

142 E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

Considering the MS results, the VFB value presents a shift in thepresence of rhenium, Fig. 4(b). The valueswere−0.154V and−0.365Vfor p-Si and p-Si/Re samples respectively. This shift can be attributed tothe potential distribution in the semiconductor in the presence of ametal deposit. This effect has been reported by other authors [25–31],and is referred to as charge and potential distributions and effectivebarrier heights in semiconductor electrodes coated with discontinuousmetal layers. All of these explanations are related with n-typesemiconductors/metal interfaces. Therefore, we make associationswith the p-type semiconductor considering that the potential barrierconcept reflects the mismatch in the energy position of the majoritycarrier band edge of the semiconductor and themetal Fermi level acrossthe metal/semiconductor interface. In this case, it is possible to identifythe following: in the naked surface area of the semiconductor, apotential barrier height at the semiconductor/electrolyte interface isgenerated by thedifference between thevalence bandedgeVVBS (E) at theinterface and Vfermi(redox) as follows,

/ Eð ÞB = VS

VB Eð Þ− VfermiðredoxÞ ð4Þ

where (E) is the quantity related with the semiconductor surface incontact with the electrolyte. Accordingly, ϕB

(E) increases linearly withthe shifting of Vfermi(redox). On the other hand, in the areas where thesemiconductor is coated with a metal deposit, a potential barrierheight ϕB

(M) is given for the difference between the valence band edgeVVBS (M) at the metal-covered surface and the Fermi level Vfermi(M) of

the metal layer as follows,

/ Mð ÞB = VS

VB Mð Þ− Vfermi Mð Þ ð5Þ

where (M) denotes that the quantity is for the metal-covered part ofthe semiconductor surface. This behavior will produce a surface-bandmodulation because the surface valence-band edge in the metal-covered parts VVB

S (M) as defined by Eq. (5) and that in the naked partsVVBS (E) as defined by Eq. (4) are, in general, different from each other.

This last relationship is due to that in the vicinity of the metal/electrolyte interface is produces a double layer, yielding a potential

difference that makes Vfermi(M) equal to Vfermi(redox). Accordingly,force lines are created in the metal coated sample and they runthrough the space-charge layer of the semiconductor creating thesame potential difference as that which exists at the metal/electrolytebut this time in the space-charge layer near the metal island [27].

According to the above mentioned, and considering that the shift ofthe flat band potential is attributed to the rhenium islands and theenergetic position of the surface states (SS), thebanddiagrams for nakedp-Si and Re-coated p-Si are shown in Fig. 5. This Figure was drawnconsidering the following parameters: flat band potentials of naked andmodified p-Si, silicon bandgap (1.1 eV), work function for metallicrhenium (4.9 eV), redox Fermi level (open circuit potential,−0.7 V) andsurface states location (SS) −0.740 V. It is possible to observe that thepresence of the metallic deposit on p-Si produces a shift of the flat bandpotential to a more negative potential value (−0.154 V to−0.365 V). Ifwe consider that this shift is the consequence of the band modulationgenerated in the surface-band edges, the presence of rhenium on p-Siproduces a potential barrier for the electrons in the conduction band,and a potential trap for holes in the valence band, according to Scheme1.This is a case corresponding to the formation of a contact between ametal and a semiconductor with a high density of surface states. In thiscase, the Fermi level occurs at the energy of the surface states beforemaking the contactwould indicate that the surface levels are about half-filled. If the concentrationof the surface states is sufficiently large, to takefurther charges without much change of the occupation level (semi-conductor Fermi level remains constant at the surface), the space-chargebelow the surface of the semiconductor remains unchanged uponcontact between semiconductor and metal. Accordingly, if some holesare transferred from the surface to themetal, the barrier height remainsunaffected. This situation it is describe as “Fermi level pinning” [32].

3.5. Photoelectrochemical reduction of nitrate

3.5.1. Voltammetric analysis of nitrate reduction onto p-Si andp-Si/Re surfaces

Fig. 6 shows the response of modified p-Si with Re electrodepos-ited from NH4ReO4 (pH 2) solutions, (samples synthesized with same

Fig. 5. Band diagrams for p-Si/NO3− and p-Si/Re/NO3

− interfaces. (a) p-Si flat bands; (b) p-Si/Re flat bands; (c) p-Si/NO3− cathodic polarization (depletion layer); (d) p-Si/Re/NO3

−

cathodic polarization.

143E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

conditions as those employed in morphological analysis of Fig. 3) in a0.1 M NaNO3 at pH 2 solution.

From this figure it is possible to observe that, independent of thepotential value employed in the rhenium electrodeposition process,the overpotential and the photocurrent was decreased and increased,respectively. In fact, similar results were obtained for the HER when

Scheme 1. Two-dimensional energy-band diagram for p-Si electrode coated withmetallic Re thin film.

rhenium was electrodeposited onto p-Si [12], and when it wasdeposited by an electroless process [18]. In terms of the overpotentialapplied and the photocurrent obtained, all modified samples presentapproximately the same behavior: there exists a shift of the PERNonset toward a more positive potential in 0.4 V and an increase of thephotocurrent approximately one order of magnitude compared to thesame process onto p-Si. Likewise, a shift of 0.3 V and an increase of thecurrent are observed when the p-Si/Re samples are compared withmetallic Re. The best result in the PERN was obtained for the p-Si/Resample synthesized at −0.80 V. This sample showed the largestnucleus density and the smaller size. According to these features, the

Fig. 6. Potentiodynamic j/Eprofiles of p-Si, Re andp-Si/Re systems in0.1MNaNO3 solution atpH 2. Electrodes formed from the same conditions of those in Fig. 3. (—) p-Si; (–––) metallicRe; (····) p-Si/Reelectrodeposited at−0.65 V; (–·–) p-Si/Reelectrodeposited at−0.70 V; (-–··– ) p-Si/Reelectrodeposited at−0.75 V; (—) p-Si/Reelectrodeposited at−0.80 V; Scan rate, 0.010 V/s.

144 E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

displacement toward more positive potentials and an increase of thephotocurrent of the PERN is produced in those samples that present ahigh nucleus density with small size and with spacing between thembeing the smallest when it is compared to the space charge layerthickness (≔50 nm). This indicates a photoelectrocatalytic effect of p-Si/Re in front of PERN [25–29]. Furthermore, these facts can beexplained by taking into account the Fermi level pinning observed inthe contact between the semiconductor and a metal. This means thatunder these conditions, a depletion layer remains unchanged in thesemiconductor. When illuminated, electrons appear in the conductionband, which are accumulated in the surface and can be transferred tothe metal and then to the nitrate ions. Additionally, electrons can betransferred to the surface states and then to the solution. Bothprocesses could be favoring the charge transfer and with this anincrease in the current density can be observed.

Fig. 7. IMPS plots of different electrodes immersed in a 0.1 M NaNO3 solution at pH 2.(◯) bare p-Si; (△) p-Si/Reelectrodeposited at−0.65 V (5) p-Si/Reelectrodeposited at −0.80 V.The data were obtained at −0.8 V and for frequencies in the range 1 Hz–32 kHz.

3.5.2. Preliminary analysis of reaction products in the PERNFor this experiment, a photoelectrolysis at controlled potential was

done. The p-Si/Re modified electrode was immersed in a ≔100 ppmNaNO3 at pH2 solution and a constant potential of−0.8 Vwas imposed.During the experiment, the sample was illuminated using the solarsimulator. Samples were extracted from the cell at 5–30–60 and120 min. Each experiment began with a fresh NaNO3 solution. Thereaction products analyzed were: NO3

−, HNO2, and NH4+. They were

quantified by the following analytic techniques: (i) NO3− ions were

complexedwith resorcine and indophenol red-violet colorant, and thendetermined photometrically by NOVA60 (275 nm); (ii) NO2 specieswere complexed by the addition of sulfonamide and naphtylethyldia-mine hydrochloride, considering standard methods 4500-NO2 B, anddeterminated photometrically by NOVA60 (543 nm); (iii) for determi-nation of NH4

+ ions, hypochlorite was added at pHN8, for obtainingmonochloramine which forms a derivate of blue-indophenol for theaddition of a substituted phenol. This was determined photometricallyby NOVA60 (630 nm). Analogue procedure at EPA350. Standardmethods 4500-NH3D and ISO 7150/1 were employed.

The results obtained are summarized in Table 1.Observable in Table 1, all the species were detected by the

analytical methods. In this context and considering the short timeemployed in the photoelectrolysis process, the most important resultwas the detection of amoniacal–nitrogen, indicating that the electrodesystem, p-Si/Re, can adsorb nitrogen's intermediates and transformthem down to N(3-). However, we continue investigations in this areabecause of the large amount of other reaction-products, which couldbe indicative of the presence of nitrogen gas. Additionally, thisexperiment must carried out for longer periods of time. For thisreason, we are working on a special photoelectrochemical cell designfor analysis of gas products.

3.6. Analysis of PERN by means of Intensity Modulated PhotocurrentSpectroscopy (IMPS)

Fig. 7 shows the IMPS results for the Nitrate reduction process ontop-Si and p-Si/Re electrodes. The measurement done on the baresubstrate (pSi) can be taken as a comparison for thosewith Re depositedonto it. The two samples chosen for the IMPS measurements were two

Table 1Nitrogen reaction products from controlled potential photoelectrolysis technique.

Sample NO3−/ppm NO2

−/ppm NH4+/ppm Other/ppm

Blank 112 – – –

5 min 33.6 1.232 4.48 72.730 min 30.2 1.3 5.49 75.060 min 28.0 7.84 6.27 69.9120 min 25.8 9.74 7.2 69.2

limiting cases: thefirst onewith Redeposited at−0.65Vand the secondwithRedeposited at−0.80V, bothwith a chargepassedof 20mCcm−2.

The IMPS results represent a plot in the complex plane of the ratiojph ωð Þ= g ωð Þ where jph(ω) is the modulated photocurrent component,and g(ω) is the modulated component of the charge flux photogener-ated by the incident photonflux [33]; beingw themodulation frequency(in s−1) of themonochromatic laser illumination. The ratio jph ωð Þ= g ωð Þis a dimensional and represents a transfer function that, as the frequencyis varied, describes a locus in the complex plane (imaginary vs. real partsplot). A detailed generalized theoretical model was developed byPonomarev and Peter [34] taking into consideration the influence ofseveral kinetic processes in the semiconductor/electrolyte interface.Ponomarev and Peter deduced typical features in the shape of IMPSplots. According to their models, the processes occurring in theSemiconductor/Electrolyte interface, as shown in Scheme 2, arecompatible with the single loops observed in Fig. 7.

Scheme 2. Processes occurring in the p-Si/0.1 M NaNO3 interface according to the IMPSmodel considered.

145E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

In this simplemodel, the kinetic processes aremodeled by two rateconstants, k1 and k2, which correspond to a direct charge transferfrom the semiconductor to the interface and an electron-hole pairrecombination at the interface, respectively. Although the models arederived for n-type semiconductors, they can also be used in p-typesemiconductors [17]. Within this model, the transfer function formodulated photocurrent response would correspond to:

jph ωð Þg ωð Þ =

k1 + iω � C = CSC

k1 + k2 + iω� 11 + iωτ

ð6Þ

where CSC corresponds to the semiconductor space-charge capaci-tance, C is the total interface capacitance (1C = 1

CSC+ 1

CH, being CH the

Helmholtz capacitance), and τ=RC is a cell time constant, R being thecell resistance (solution and ohmic contact). Within the context of Eq.(6), the upper loop of Fig. 7 (positive real part) is related to the firstfactor, while the time constant τ generates the lower loop (negativereal part) at high frequencies. Assuming that τ can be neglected, thereal and imaginary parts in Eq. (6) become:

Rejph ωð Þg ωð Þ

� �=

CCSC

ω2 + CSCC k1 k1 + k2ð Þ

k1 + k2ð Þ2 + ω2 ð7Þ

Imjph ωð Þg ωð Þ

� �=

CCSC

ω 1− CSCC

� k1 + k2

h ik1 + k2ð Þ2 + ω2 ð8Þ

Therefore the plots obtained have a maximum

ddω

Imjph ωð Þg ωð Þ

� �� �= 0Z ωmax = k1 + k2 ð9Þ

On the other hand, the crossing with the real axis Im jph ωð Þg ωð Þ

h i= 0 the

values of Re jph ωð Þg ωð Þ

h i= k1

k1 + k2ð Þ for ω=0, and Re jph ωð Þg ωð Þ

h i= C

CSCfor ωNN

k1+k2, although in this limit the influence of the cell time constantcannot be denied.

According to these IMPSmodels, the frequency corresponding to themaximum in the imaginary vs. real plots is related with the constantrates of the recombination and charge transfer processes k1+k2,while the crossing of the real axis at very low frequency is related tothe ratio k1

k1 + k2ð Þ. Therefore from Fig. 7, due to that all loops of thesesamples beginning at the origin and having a direct charge transfer fromthe semiconductor to the electrolyte, k1 is much slower than therecombination rate k2. This second constant can be directly obtainedfrom the position of the maxima, which was close to 10.1 kHz, 7.53 kHzand 12.2 kHz for the p-Si, p-Si/Reelectrodeposited at −0.65 V, and p-Si/Reelectrodeposited at −0.80 V, respectively. Therefore the correspondingvalues for the time constant k2≈ωmax=2πfmax, with fmax the frequencymeasured in Hz, are 63 ms−1, 77 ms−1, and 47 ms−1, respectively.Although previous reports of similar time constants in p-Si silicon areslower than these values [17], it was already shown that therecombination rate constant increases with carrier concentration [35].Therefore, for the present p-Si substrate with very low resistivity anincrease in this constant rate is expected in comparisonwith low dopedp-Si. On the other hand, the results show that there is an increase in thetime constant for the sample with Re electrodeposited at a morenegative potential. These results can be tentatively interpreted in thefollowing way. For the sample electroprepared at more positivepotentials than −0.65 V, larger Re islands are expected (see AFMimages). Metallic Re islands act as electron sources, which decrease thiscarrier concentration in the semiconductor, therefore reducing therecombination rate. Meanwhile for the sample prepared at a morenegative potentials (−0.80 V), a smaller Re nucleus is obtained. Thisnucleus may act as a recombination center, increasing the recombina-

tion rate. As no direct charge transfer from the semiconductor to thesample was observed in the present samples, the recombinationthrough the surface states may be responsible for the photoelectro-chemical response of the samples, also considering that the presence ofthe rhenium deposit increases the density of the surface states.Furthermore, as was observed in the energy-band diagrams in presenceof metallic Re, if a barrier height for electrons appears in the conductionband, then the recombination process probabilitywould be increased inpresence of these films.

4. Conclusions

The study of rhenium electrodeposition deposition processes on p-Si(100) from acidic media solutions indicates that the nucleation andgrowth mechanism corresponds to a three-dimensional progressivenucleation and diffusion-controlled growth, which was confirmed byAFM. According to the MS plots and to the theory of the semiconduc-tors coated with metal thin films, it was found that the presence of Reproduces a band-modulation due to the irregular potential distribu-tion in the coated areas. Considering the Re work function, theposition of the conduction and valence band-edge, and the redoxFermi level, the p-Si/Re system would have a modulation in theconduction band like an energy barrier to electrons (like Schottkycontact between n-Si/Metal) and traps for holes in the valence band.The p-Si/Re electrode system demonstrated to be a potentialphotoelectrocatalyst for the PERN due to the displacement of 0.4 Vtowards more positive potentials and an increase of the photocurrentdensity by one order of magnitude versus naked p-Si(100). It also hada displacement of 0.3 V and a photocurrent increase of two timescompared with the metallic Re substrate. Finally, the IMPS techniquedemonstrated that rhenium deposited at more negative potentialtimes increased the recombination rate, considering that under theexperimental conditions the transfer charge is neglected.

Acknowledgments

We thank FONDECYT, Chile, for financial support for this study(grant No. 11060514 and 7080186) and the Dirección de Investigacióne Innovación of the Pontificia Universidad Católica de Valparaíso(grant No. 037.108/2008 DII-PUCV). E. Muñoz and R. Henríquez thankPrograma Bicentenario de Ciencia y Tecnología, PSD82. R. Marotti alsoacknowledges support received from PEDECIBA — Física, and the CSIC(Comisión Sectorial de Investigación Científica) of the Universidad dela República, in Montevideo, Uruguay. Special thanks to Jennifer A.Kersten of the University of Minnesota for her collaboration on thediscussion of this work.

References

[1] C.J. Johnson, B.C. Kross, Am. J. Ind. Med. 18 (1990) 449.[2] K.P. Cantor, Cancer, Causes Control Pap. 8 (1997) 292.[3] G. Ritzoulis, J. Electroanal. Chem. 327 (1992) 209.[4] A. Ahmadi, E. Bracey, R. Wyn Evans, G. Attard, J. Electroanal. Chem. 350 (1993) 297.[5] M. Pourbaix, N. de Zoubov, Nitrogen, Atlas of Electrochemical Equilibria in Aqueous

Solutions, National Association of Corrosion Engineer, Houston, 1974, p. 493.[6] K.T. Ranjit, B. Viswanathan, J. Photochem. Photobiol. A 108 (1997) 73.[7] K.T. Rangit, R. Krishnamoorthy, B. Viswanathan, J. Photochem. Photobiol. A 81

(1994) 55.[8] K.T. Ranjit, T.K. Varadarajan, B. Viswanathan, J. Photochem. Photobio. A 89 (1995)

67.[9] F. Zhang, R. Jin, J. Chen, C. Shao, W. Gao, L. Li, N. Guan, J. Catal. 232 (2005) 424.[10] L.-F. Liu, X.-Y. Dong, F.-L. Yang, C.Y. Jimmy, Chin. J. Inorg. Chem. 24 (2008) 211.[11] J. Sá, C. Alcaraz, S. Gross, J.A. Anderson, Appl. Catal. B 85 (2009) 192.[12] E. Muñoz, R. Schrebler, M. Orellana, R. Córdova, J. Electroanal. Chem. 611 (2007) 35.[13] G. Oskam, P.M. Vereecken, P.C. Searson, J. Electrochem. Soc. 146 (1999) 1436.[14] A.A. Istratov, E.R. Weber, J. Electrochem. Soc. 149 (2002) G21.[15] R. Schrebler, P. Cury, M. Orellana, H. Gómez, R. Córdova, E.A. Dalchiele, Electrochim.

Acta 46 (2001) 4309.[16] R. Schrebler, P. ury, E. Muñoz, F. Vera, R. Córdova, H. Gómez, J.R. Ramos-Barrado, D.

Leinen, E.A. Dalchiele, Thin Solid Films 483 (2005) 50.[17] G. Schlichthörl, E.A. Ponomarev, L.M. Peter, J. Electrochem. Soc. 142 (1995) 3062.

146 E. Muñoz et al. / Thin Solid Films 518 (2009) 138–146

[18] E. Muñoz, R. Schrebler, P. Grez, R. Henríquez, C. Heyser, P. Verdugo, R.E. Marotti, J.Electroanal. Chem. 633 (2009) 113.

[19] B. Scharifker, G. Hills, Electrochim. Acta 28 (1983) 879.[20] G. Gunawardena, G. Hills, I. Montenegro, B. Scharifker, J. Electroanal. Chem. 138

(1982) 225.[21] M. Palomar-Pardavé, B.R. Scharifker, E.M. Arce, M. Romero-Romo, Electrochim.

Acta 50 (2005) 4736.[22] D. Grujicic, B. Pesic, Electrochim. Acta 49 (2004) 4719.[23] S. Ottow, G.S. Popkirov, H. Föll, J. Electroanal. Chem. 455 (1998) 29.[24] G. Oskam, P.M. Hoffmann, J. Schmidt, P. Searson, J. Phys.Chem. 100 (1996) 1801.[25] Y. Nakato, K. Ueda, H. Tsobomura, J. Phys. Chem. 90 (1986) 5495.[26] R. Dominey, N. Lewis, J. Bruce, D. Bookinder, M. Wrighton, J. Am. Chem. Soc. 104

(1982) 467.

[27] Y. Nakato, H. Tsobomura, Electrochim. Acta, 37 (1992) 897.[28] Y. Nakato, K. Ueda, H. Yano, H. Tsobomura, J. Phys. Chem. 92 (1988) 2316.[29] M. Szklarczk, J. O'M Bockris, J. Phys. Chem. 88 (1984) 1808.[30] Y. Nakato, H. Tsobomura, J. Photochem., 29 (1985) 257.[31] P. llongue, E. Souteyrand, J. Electranal. Chem. 269 (1989) 361.[32] R. Memming, Semiconductor Electrochemistry, Wiley-VCH,Weinheim, 2001, p. 27.[33] L.M. Peter, Chem. Rev. 90 (1990) 753.[34] E.A. Ponomarev, L.M. Peter, J. Electroanal. Chem. 396 (1995) 219.[35] G. Oskam, J.C. Schmidt, P.C. Searson, J. Electrochem. Soc. 143 (1996) 2538.