Origin of Photocarrier Losses in Iron Pyrite(FeS2) NanocubesSudhanshu Shukla,†,‡ Guichuan Xing,§ Hu Ge,‡ Rajiv Ramanujam Prabhakar,∥ Sinu Mathew,⊥,¶

Zhenghua Su,∥ Venkatram Nalla,# Thirumalai Venkatesan,⊥,¶ Nripan Mathews,‡ Thirumany Sritharan,‡

Tze Chien Sum,§ and Qihua Xiong*,§,+

†Energy Research Institute, Interdisciplinary Graduate School, and ‡School of Materials Science and Engineering, NanyangTechnological University, Singapore 637371, Singapore§Division of Physics and Applied Physics, School of Physical and Mathematical Sciences, Nanyang Technological University,Singapore 637371, Singapore∥Energy Research Institute, Nanyang Technological University, Singapore 637371, Singapore⊥Department of Electrical and Computer Engineering, National University of Singapore, Singapore 117576, Singapore¶NUSNNI-NanoCore, National University of Singapore, Singapore 117576, Singapore#Centre for Disruptive Photonic Technologies (CDPT), Nanyang Technological University, Singapore 639798, Singapore+NOVITAS, Nanoelectronics Centre of Excellence, School of Electrical and Electronic Engineering, Nanyang TechnologicalUniversity, Singapore 637371, Singapore

*S Supporting Information

ABSTRACT: Iron pyrite has received significant attention due to its highoptical absorption. However, the loss of open circuit voltage (Voc) preventsits further application in photovoltaics. Herein, we have studied thephotophysics of pyrite by ultrafast laser spectroscopy to understandfundamental limitation of low Voc by quantifying photocarrier losses in highquality, stoichiometric, and phase pure {100} faceted pyrite nanocubes. Wefound that fast carrier localization of photoexcited carriers to indirect bandedge and shallow trap states is responsible for major carrier loss. Slowrelaxation component reflects high density of defects within the band gapwhich is consistent with the observed Mott-variable range hopping (VRH)conduction from transport measurements. Magnetic measurementsstrikingly show the magnetic ordering associated with phase inhomoge-neity, such as FeS2−δ (0 ≤ δ ≤ 1). This implies that improvement of ironpyrite solar cell performance lies in mitigating the intrinsic defects (such as sulfur vacancies) by blocking the fast carrierlocalization process. Photocarrier generation and relaxation model is presented by comprehensive analysis. Our resultsprovide insight into possible defects that induce midgap states and facilitate rapid carrier relaxation before collection.

KEYWORDS: iron pyrite, nanocubes, transient absorption, variable range hopping, photovoltaics, magnetization, carrier dynamics

Low cost and clean solar photovoltaic electricitygeneration is one of the key challenges of modern erasince the demand for power is continuously increas-

ing.1,2 Solar energy conversion via photovoltaic (PV) systems iscertainly a viable option as it does not generate any pollution inthe conversion process. However, the major impediment towidespread PV use is the high cost of the modules as well as therequirement for large land area. Consequently, the majormotivating factor for PV research has been to achieve highefficiency and low cost with the use of environmentally benignmaterials.3,4 Earth-abundant semiconducting materials offergreat opportunity for large-scale PV deployment by leveragingthe cost concerns and environmental issues.5−8 Extensiveresearch effort has been devoted to the development of such

materials including a-Si, CZTS, CuO, Cu2O, Cu2S, CoO3,Zn3P2, and FeS2.

9 Of those, FeS2 has received much attentionlately due to its abundance and attractive semiconductingproperties.Iron pyrite (FeS2) has a cubic crystal structure with good

semiconducting and optical absorption properties that make itattractive as a material for solar energy conversion. It has aremarkably high optical absorption coefficient (α > 105 cm−1

for hν > 1.3 eV), a nearly ideal energy band gap Eg of ∼0.95 eV

Received: January 5, 2016Accepted: March 10, 2016Published: March 10, 2016

Artic

lewww.acsnano.org

© 2016 American Chemical Society 4431 DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

and sufficiently long minority carrier diffusion length (100−1000 nm). Furthermore, a high carrier mobility of μ ∼ 360 cm2

V−1 s−1 has been reported for n-type bulk single crystals.10

Nonetheless, the efficiency for a pyrite single crystal photo-electrochemical cell has not exceed 2.8% despite anextraordinarily high short circuit current (Jsc = 42 mA/cm2),fill factor of 50% and external quantum efficiency reaching ashigh as 90%. This low efficiency is attributed to its low opencircuit voltage of Voc ∼ 187 mV which has still not beenexceeded, even after nearly 30 years since this original report.11

Early extensive research on iron pyrite, primarily led byTributsch and co-workers, attributed the problem of poorphotovoltage to possible defect states within the forbiddenenergy region, and the challenge of control on phase purity andstoichiometry.10 Recently, independent studies confirmed theexistence of a surface inversion layer on pyrite surface andconclude that the surface states and ionization of high densitydeep donor defect states are responsible for low voltages,consistent with the prediction of Jaegermann et al.12−14

Interestingly, recent reports of synthesis of pure pyritenanostructures by solution processing caused a resurgence ofinterest in pyrite as a PV material as this might make it possibleto avoid impurities and the surface inversion problem.15−17

Solution processed nanoparticles offer a unique platform tostudy fundamental as well as device physics due to their bettercontrol of quality, composition, size and shape, and the relativeease of making films by simple dip or spin coating.18 However,post treatment is generally required to remove any residualorganics, such as attached insulating ligands, before deviceapplications to get appreciable currents in films throughenhanced electronic interparticle coupling.19 Nanostructurediron pyrite has been deployed, (i) in a heterojunction and p−i−n configuration photodiode, (ii) as a catalyst in dye sensitizedsolar cell (DSSC), (iii) as a photoconductor, and (iv) in bulkheterojunction inorganic−organic hybrid solar cells.20−25

However, regardless of their application as a photoresponsivelayer in such devices, pyrite nanocrystals have not proven theireffectiveness in solar cell configuration which is a setback forthe adoption of pyrite nanoparticles in solar cells technology.26

Despite this setback, we believe that the high responsivity tooptical excitation in pyrite nanocrystals, coupled with theirgood crystallinity and stability, gives rise to the potential fortheir use as the absorber layer in solar cells if adequatephotovoltages could be derived through appropriate posttreatments and better heterojunctions fabrication. To achievethis, the physics of photoexcitation and charge carrier dynamicsin pyrite nanocrystals must be well understood, because loss ofphotovoltage clearly implies major loss of photocarriersthrough trapping pathways within the crystal. Understandingthe fundamental photophysics of charge carriers in pyritenanocrystals is key to improve their poor photovoltaicperformance.In this work, we extensively investigate and correlate the

physical property characterizations with the charge carriergeneration and relaxation dynamics in as-prepared (withoutsulfurized as-coated films) and sulfur treated {100} faceted ironpyrite nanocubes by transient absorption (TA) measurements.Since {100} surface inhibits surface reconstruction, contrary to{111} or {210} face, {100} faceted nanocubes were chosen forthe present study.27−29 In addition, spin-coated as-preparediron pyrite nanocubes were sulfur treated to make hetero-junction solar cells. Iron pyrite films were characterized by X-ray diffraction (XRD), high resolution transmission electron

microscopy (HRTEM), and Raman spectroscopy for crystal-linity and phase evaluation. Charge carrier relaxation pathwayswere identified from exponential decays of photoexcited chargecarriers in TA dynamics. On the basis of the slow and fastrelaxation components of photoexcited carriers, a recombina-tion model is formulated here that could account for therecombination mechanisms of charge carriers in iron pyritethrough shallow and deep defect trap states. Temperaturedependence of electrical charge transport measurementrevealed a disordered semiconductor behavior with Mott-VRH type transport, implying a high density of defects. Also,temperature dependence of magnetization revealed a weakmagnetic order at low temperatures which could be attributedto the defects and their influence. These findings support acarrier relaxation model where the relaxation pathway isthrough multiple electronic states present within the band gap.

RESULTS AND DISCUSSIONPanels a and b of Figure 1 show the X-ray diffraction (XRD),Raman spectra of as-prepared and sulfurized films obtainedfrom nanocubes, respectively. All XRD peaks could be indexedto the cubic pyrite phase (JCPDS 42-1340). No peaks frompossible impurity phases such as pyrrhotite (Fe1−xS), ironmonosulfide (FeS), marcasite (orthorhombic polymorph ofFeS2), or iron oxide were detected. It is evident from Figure 1that sulfurization process retains the shape and phase purity ofthe as-prepared sample. This allows us to improve theconductivity of the films without affecting the phase purityand the cube morphology by removing the insulating organicligands and enhancing inter particle coupling (Figure S1,Supporting Information). The size of the nanocubes producedin our case is between 80 and 150 nm, which is desirable forsolar cell applications.15,30,31 This size range takes advantage ofthe long diffusion length of carriers in pyrite and minimizes thescattering of charge carriers. A film made from larger nanocubesalso minimizes excess contact surface area where undesirablecarriers scattering and recombination impede the chargetransport. Lattice constants computed for the as-coated andsulfurized films from the XRD spectra are 5.392 and 5.412 Å,respectively. This is very close to the value of 5.418 Å reportedfor a high quality, pyrite sample by Birkholz et al.32 Since thesensitivity of the XRD technique in detecting small fractions ofimpurity phases is limited, we resorted to confocal micro-Raman spectroscopy with laser excitation at 532 nm with 0.14mW power. Lower laser illumination intensity used as oxidationoccurred at higher intensities (Figure S2, SupportingInformation). Three characteristic phonon modes wereobserved in the spectra at wavenumber 341 (Ag), 377 (Eg),and 425 cm−1 (Tg(3)), respectively. The Ag and Eg modes arefrom sulfur−sulfur bond vibrations and Tg (3) is vibrationalmode.33,34 Out of five active Raman active modes, the Tg (1)and Tg (2) phonon modes are generally not observed due tosignificant light extinction caused by other modes andinefficient light scattering at room temperature.35 Both as-prepared and sulfurized films showed only pyrite peaksconfirming the high phase purity of the films. No signalswere obtained from possible impurity phases such as FeS,Fe3S4, Fe1−xS including the polymorph marcasite phase.Nanocube morphology of as-prepared and sulfurized films isevident from SEM images of Figure 1c,d with clear facets. Sincethe cube shape remained intact even after sulfurization, this is agood high temperature heat treatment platform to performpassivation and ligand removal without affecting the morphol-

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4432

ogy for basic studies retaining the cube shape. To furtherconfirm the crystalline quality and phase purity of thenanocubes, high-resolution transmission electron microscopy(HRTEM) was performed as shown in Figure 2. Singlecrystallinity of the as-coated (Supporting Information) andsulfurized nanocubes is evident in the HRTEM images. Thelattice-resolved, high-resolution image yields a lattice constantof 5.48 Å, in congruence with the (100) plane lattice spacingshown in Figure 2b. Selected area electron diffraction (SAED)pattern shows two pairs of planes which could be indexed to(200) and (2 00), and (020) and (02 0) as shown in Figure 2c.The zone axis of the cube is [001] which proves that they areterminated with {100} facets. Planar defects such as stackingfaults could be discerned in the lattice as marked with yellowlines and arrows in Figure 2b. Although such defects have beenpreviously reported in nanoparticles, their presence in thesecubes may have significance in this study, as they may have animportant role in governing the electronic properties of pyritenanoparticles.17,36,37

Optical absorption of sulfurized pyrite nanocubes thin filmswas measured by UV−Vis absorption spectrophotometer and isshown in Figure 3a. An optical absorption coefficient α > 1 ×105 cm−1 for hν > 1.3 eV was obtained. Such a high opticalabsorption is a notable characteristics of iron pyrite.38,39 InFigure 3a, absorption below band gap is observed whichcorresponds to defect level excitations and transitions withinthe exponential band tails (region left to the dashed redline).40−42 The shape of the curve below the band gapabsorption (referred as Urbach tail) also indicates the degree ofdisorder in the semiconducting material.43,44 From Taucanalysis, shown in Figure 3b, intercepts of the linearextrapolations of the different slopes on the energy axis couldbe interpreted as optical transitions corresponding to thosespecific energy levels.45 The electronic behavior of pyrite isgoverned by these energy level transitions.46 The band gapestimated is 1.02 eV, which is well within the range (0.80−1.10eV) commonly reported for pyrite.10,29,47 However, this bandgap value appears to be direct as it fits linearly with n = 2 valuein the parameter (αhν) n in the Tauc equation (α is calculated

Figure 1. Structural characterization. (a) XRD pattern, (b) confocal micro-Raman spectra, and SEM morphology of (c) as-prepared and (d)sulfurized iron pyrite nanocubes.

Figure 2. HRTEM image of sulfurized iron pyrite nanocube. (a) Edge of the nanocube, inset shows full nanocube image; (b) lattice fringesand stacking faults marked with yellow dashed lines and arrows; (c) selected area electron diffraction (SAED) pattern indicating the pointscorresponding to marked planes (red spots).

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4433

by absorption spectra). Such indirect to direct transition is notunusual in semiconductors with high density of defect statesand disorder, and has been reported even for iron pyrite singlecrystals.12,48 Notably, the discrepancies in commonly reportedoptical band gap of pyrite could be an indication of thepresence of fuzzy band edges in pyrite, a quintessential featurein disordered semiconductors.49,50 Potential fluctuations arisingfrom defects perturb the electronic band edges and could resultin broadening of electronic states leading to band tailsextending below the band edges, which could manifest assub-band gap optical absorption.51−53 Assessment of the carrier

dynamics of these states are critical in order to establish carrierloss mechanism in iron pyrite.Given the presence of high density defect states, the next

important step is to understand the carrier dynamics in ironpyrite as defect states within the band gap will haverepercussions on the optoelectronic performance. Pumpprobe or transient absorption (TA) spectroscopy wasconducted on as-prepared and sulfurized samples to understandcharge carrier relaxation in pyrite. TA spectroscopy allows us tomonitor the photogenerated charge carrier dynamics in as-prepared and sulfurized pyrite samples. The broad banddynamical TA spectra in the femtosecond and microsecond

Figure 3. (a) UV−Vis absorption spectra and (b) Tauc plot analysis of sulfurized iron pyrite film. Indirect and direct band edge intercept at0.72 and 1.02 eV, respectively.

Figure 4. Differential transient absorption spectra after photoexcitation of sulfurized iron pyrite nanocubes as a function of time delay in (a)femto−picosecond range, (c) nano−microsecond range. Carrier decay dynamics probed at 950 nm (photobleaching) and 1400 nm(photoinduced absorption) by fitting the transients in (b) picosecond range, fitted with single exponential decay function (solid black line)with decay time constant (charge transfer time, τct) 1.8 ps and (d) microsecond range, fitted with biexponential decay function with timeconstants 50 ns (τd1) and 990 ns (τd2) associated with the long-lived trap states and the eventual recombination process.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4434

probe time delay ranges were measured with a HELIOS andEOS from ultrafast systems. Following photoexcitation withabove bandgap 650 nm pump pulses, the hot charge carriersrapidly relax to the band edge and trap states. Figure 4a,c(Figure S4a,c) shows the pseudo color plots of the change inabsorption (ΔA represented by the color scale) againstwavelength (nm) and probe time delay (ps and μs) for thesulfurized sample (as-prepared sample). Similar TA spectra anddynamics for the as-prepared and sulfurized samples suggestthat the intrinsic electronic property (crystallization) of pyrite isnot changed much by the sulfurization. This result furtherconfirms that the dominant effect of this post treatment is toremove the attached insulating ligands for better interparticlecoupling. Figure 4a also shows a prominent negative ΔA bandor photobleaching (PB) signature near the bandgap at around950 nm. This PB band shows a fast decay with a fitted lifetimeof 1.8 ps. This lifetime closely matches with the building uptime of a broad positive ΔA band or photoinduced absorption(PIA) signature at wavelengths >1000 nm. The band edge PBsignal should originate from near bandgap stimulated emissionof the photogenerated carriers and state-filling of the band edgestates; the below band gap broad PIA band should originatefrom the localized charge carrier absorption. Therefore, theband edge PB fast decay represents photogenerated chargecarrier localization from the direct gap edge to indirect bandedge and midgap trap states. Since, the sample is of p-typenature (Hall measurement, Supporting Information TableST1), fast charge carrier localization most likely originatesfrom the photogenerated electron localization and trapping.The recombination time of the trapped electrons with holes isof the order of few hundreds of nanoseconds (Figure 4d). Onthe basis of our analysis, we have drawn a representativeschematic of the complete photophysical processes involved iniron pyrite nanocubes as shown in Scheme 1. These processesclearly indicate the high carrier losses in iron pyrite due to fastlocalization and recombination assisted by intermediate gapstates. Our model also supports the mechanism of rapid

electron relaxation through intraband states proposed by Cederet al. through theoretical and STS analysis.49

Next, we examine the impact of such high defect states oncharge carrier transport in iron pyrite. To study the physics ofcarrier transport and the effects of intrinsic defects, weconducted temperature-dependent electrical transport andmagnetization measurements. Figure 5a shows the dependenceof resistivity (Ln ρ) on 1/T (inset shows the resistance versustemperature data) which resembles a typical semiconductingbehavior. When the results are plotted against T−1/4, a goodstraight line could be fitted from room temperature (300 K) toabout 50 K as shown in Figure 5b. This indicates a goodagreement with Mott variable range hopping (VRH)mechanism for charge transport given by the equation ρ = ρoexp(To/T)

1/d+1 where d in the exponent is the dimensionality ofthe system, To is the characteristic temperature, and ρo is aweak temperature dependent pre-exponential factor (value of ρat T → ∞) which depends on the electron−phononinteractions.54 It is to be highlighted that the density of states(DOS) is assumed to be non-zero and constant at the Fermilevel within the thermal energy interval between localized statesin the equation. The resistance below 50 K was high andtherefore difficult to measure; hence, we focused in the range300−50 K. A high characteristic temperature of To = 4.5 × 106

K was obtained by linear fitting of Ln(ρ) vs T(−1/4). Suchconducting behavior is marked by charge transport through thelocalized states via phonons (phonon assisted transport). Atfinite temperature, VRH conduction is caused by the localizedstates that depend on dimensionality (d) of the system.55 Sincethe charge transport in our films is through nanocubes, thevalue of d = 3 could be intuitively anticipated due to three-dimensional isotropy of nanocubes. Moreover, deviation from d= 3 is not anticipated as the characteristic length scales are toosmall for iron pyrite to observe such transitions or crossover inpyrite nanostructures. This argument agrees well with theexperimental data fitted with three- dimensional VRH equation.Previous transport studies on iron pyrite single crystals and

Scheme 1. Representative Schematic of the Photophysical Processes Involved in Iron Pyrite Based on Optical Pump ProbeSpectroscopya

a(1) Optical excitation of electron from valence to conduction band, (2) rapid carrier localization of the excited carrier to indirect band edge and lowlying shallow defect states, (3) slower electron relaxation to midgap deep defect states/band (long lived trap states),and (4) electron recombinationprocess with the valence band holes.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4435

nanostructures also showed Mott-VRH type conductionphenomena.12,56 The fact that Mott-VRH type conductionbehavior is often observed in different pyrite nanostructures,regardless of the synthesis procedures used, reinforces theexistence of smaller localization lengths and intrinsic defectmechanisms in iron pyrite. Disorder and intrinsic defects causelocalized states (Anderson localization) within the band gapand electron hops from one site to another spatially separatedsite. The selection of the site is based on the criteria ofminimum activation energy (minimum energy difference)configuration that maximizes the hopping probability.57 Thepresence of high density of defect states reflected by our opticalpump probe studies is in agreement with the charge transportbehavior observed in the temperature dependence of resistancein the iron pyrite films.It is thus clear that the disorder and intrinsic defects in the

pyrite crystals cause the defect states that trap the chargecarriers and consequently assist recombination. To eliminatesuch defects, the nature and characteristics of such defects needto be understood. A very rational and logical argument wouldbe to look for the possible existence of any phaseinhomogeneities in iron pyrite. Intrinsic defects are arguablythe most plausible explanation for the universal conductionbehavior of iron pyrite and its sluggish photoresponse (FigureS5, Supporting Information). We conducted temperature-dependent magnetic measurements on our sulfurized ironpyrite nanocube film samples. In general, FeS2 is a diamagneticsemiconductor having all the electrons filled in t2g state in lowspin configuration.58,59 Fe and other Fe−S phases exhibitmagnetic order.60 Temperature-dependent magnetization datais shown in Figure 5c. It can be clearly seen that below about 50K, magnetic ordering is evident. This might occur if theordering energy dominates the thermal energy. Isothermal M−

H measurements conducted in the two different regimes areshown in Figure 5d. Typical diamagnetic behavior is observedat 300 K, though at 10 K, a very weak magnetic ordering isexhibited. This raises the possibility of the presence ofunexpected phases such as FeS2−δ (0 ≤ δ ≤ 1) which couldbe the consequence of nonstoichiometry at localized regionssuch as sulfur point defects within the bulk and the surface.Such concerns have been stressed in number of studies.42,61

Since such phases were not detected at room temperature, thiscould be attributed to the point defects clustering at lowtemperatures which might create nanoscale regions of non-stoichiometry and consequently local magnetic ordering.62 Atlow temperature, magnetism may be due to freezing of defectspins which enable magnetic exchange to occur versus thermalfluctuations which tend to kill the alignment and hencemagnetism.Moreover, the origin of magnetic signal can be explained in

terms of geometrical consideration of the Fe−S coordination inthe crystal. LFT and MO theory predicts the formation ofmidgap defect states/band due to sulfur vacancies. The Fe atomin iron pyrite is in octahedral coordination (Fe−S6) with S. Dueto strong ligand field (S atoms), Fe d-states split and formvalence and conduction bands which are derived from t2g and egstates, respectively. Electron distribution in these states is in lowspin configuration, and hence, pyrite is nonmagnetic. When Svacancies are formed, the local coordination of Fe (such as Fe−S5, FeS4,..., etc.) is reduced which leads to symmetry reduction.This causes further splitting of Fe d-states and introducesadditional states in the forbidden energy gap. Electrondistribution is no longer in low spin configuration (fully filledorbitals) giving rise to some high spin configurations due tosplitting of t2g state, resulting in localized magnetic spins.

32,63 Atlow temperatures when the thermal vibrations are low, the

Figure 5. Electrical transport of iron pyrite. (a) Resistivity vs 1/T plot. Inset shows resistance vs temperature plot; (b) Ln ρ vs T−1/4 plot fittedlinearly with Mott-VRH type transport. Magnetic measurements on iron pyrite, (c) magnetization vs temperature plot from 300 to 10 K and(d) M−H curve at temperature 10 and 300 K showing superparamagnetic and diamagnetic response, respectively.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4436

crystals tend to reduce their total energy by S vacancy diffusioninto some low energy configurations where exchange couplingof the unpaired electrons could occur through the Fe ions.64

This could lead to ferromagnetism in localized regions of thecrystal where it is detectable by a sensitive magnetometer. Sincepoint defects are a source of localized gap states, the magneticinformation is also consistent with the photophysical processesproposed above in iron pyrite. The characteristics of defectsdepend mostly on the growth conditions. Our results suggestimmediate link between localized gap states and existence ofsuch phases intrinsic to pyrite chemistry.Iron pyrite heterojunction solar cells were evaluated using

CdS as the heterojunction partner for charge extraction due toits favorable energy level configuration for electron transfer.Figure 6a shows the current density versus voltage (J−V) curveof the solar cell under 1 Sun (100 mW/cm2) illumination. Aphotovoltage of around 94 mV was obtained with FF of 0.28and current density of 0.4 mA/cm2 showing that the sulfurizediron pyrite nanoparticles could be used as solar absorbermaterial. Figure 6b shows the external quantum efficiency(EQE) of the FeS2−CdS heterojunction solar cell. Photo-current spectra consistently followed the optical absorptionspectra of pyrite, which reflects the ability of photogeneratedcarriers to be converted to electrical current. Only electroniccontribution is considered for generation of photocurrents asthermal contributions would be negligible at femtosecondexcitation followed by relaxation in approximately microsecondtime scale (Figure S5, Supporting Information). This furthercorroborates that current flows in pyrite through trappedcharges that are distributed within the band gap which is clearly

the reason for the lower photovoltage. Figure 6c shows theschematic of charge transfer process in a pyrite nanocubes/CdSheterojunction device. Energy band position of CdS and FeS2NCs are favorable for charge carriers to be extracted by the holeand electron selective contacts. Cross-sectional SEM image ofthe device is shown in Figure 6d. Active layer, top and bottomelectron and hole collecting layers can be distinctly seen. Thelow photovoltage value obtained is consistent with the chargecarrier dynamics of pyrite elucidated previously. Our modelpredicts the recombination pathway through trapping in bandtail localized states followed by deep state relaxation andsubsequent recombination via hole capture or directly throughnonradiative processes.

CONCLUSION

We have studied high purity as-prepared and sulfurized ironpyrite nanocubes prepared by hot-injection method. We havedemonstrated that sulfurization treatment of nanocubes do notaffect the shape and phase purity of nanocubes while increasingthe conductivity, making them suitable for fundamental opticaland transport studies and device applications. Nature of opticalabsorption of iron pyrite corresponds to that of a disorderedsemiconductor. Optical pump−probe studies conducted onpyrite sample to study photocarrier generation and relaxationprocess have found strong photocarrier generation occurring atthe direct band edge (∼950 nm). Decay of the photoexcitedcarriers consisted of a fast and a slow component. The fastcomponent was attributed to rapid carrier localization ofphotoexcited carriers to indirect band edge and shallow defect

Figure 6. (a) Current density vs voltage (J−V) curve of iron pyrite heterojunction solar cell, (b) external quantum efficiency of theheterojunction solar cell, (c) schematics of energy band alignment of different layers of the solar cell, and (d) cross-sectional SEM of themeasured solar cell.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4437

states with τct of 1.8 ps. The slow component was attributed tocarrier relaxation through deep midgap states and recombina-tion process via direct and hole capture process. Temperature-dependent transport measurements showed the operation ofVRH type conduction mechanism from 50 to 300 K, which isconsistent with high density of defect states detected by opticalTA measurements. Temperature-dependent magnetic measure-ments showed magnetic ordering arising in iron pyrite at lowtemperatures, contrary to nonmagnetic behavior at roomtemperature. These magnetic impurities directly correlatewith phase inhomogeneity in iron pyrite, such as localizedformation of FeS2−δ (0 ≤ δ ≤ 1) that induces defect states/band within the energy gap that traps the charge carriers by fastlocalization of carriers and promote recombination losses. Ourfindings suggest that photovoltages can still be obtained frompyrite nanocubes, though heterojunction solar cells preparedusing sulfurized iron pyrite nanocubes showed photovoltage ofonly 94 mV with current density of 0.4 mA/cm2. However,significant increases to efficiency may be obtained by twofactors, (i) boosting the current through enhanced interparticlecoupling or realizing a single domain films free of boundaries,(ii) increment in photovoltage by mitigating the defect statespresent within the band gap by not only limiting to posttreatment process but also finding adequate growth strategythat results in low bulk defects. We believe that our findings willhelp in the understanding of the rarely explored photophysicsof iron pyrite nanocubes in detail and provide deep insight onintricate relationship between defects (intrinsic and disorderinduced) and gap states responsible for photocarrier losses. Ourfuture work will be focused on quantifying these defects,eliminating them varying growth conditions and post treatmentprocesses, and subsequently, studying their relationship topyrite solar cell performance.

METHODSSynthesis. The {100} faceted iron pyrite nanocubes were

synthesized by hot-injection method as reported previously.65 Thestarting materials for the synthesis were 98% anhydrous iron(II)chloride (FeCl2), 70% oleylamine (OLA), and sulfur powder fromSigma-Aldrich. A total of 0.5 mmol (63.5 mg) of FeCl2 with 5 mL ofOLA were mixed in a trineck flask and degassed for 30 min, andsubsequently held at 110 °C for 1 h to form Fe−OLA complex.Thereafter, the temperature was raised to 180 °C and 3 mmol (96 mg)of sulfur mixed in OLA was injected into the flask while maintainingthe temperature at 180 °C. The duration of reaction was 24 h. Afterreaction, the solution was cooled to room temperature naturally; alarge amount of methanol was added to precipitate the FeS2nanocubes followed by centrifugation and dispersion in hexane.Then, 50 μL of Pyrite NCs suspension with a concentration of (0.1M)was spin-coated on the substrate. Sulfurization process was carried outin a two zone furnace. Sulfur powder was kept at 200 °C and nanocubefilm was kept at 500 °C in argon atmosphere for 30 min. Tube wasevacuated and purged with argon three times before ramping up thetemperature for sulfurization process.Device Fabrication. At first pyrite nanocubes were spin coated on

molybdenum bottom contact and sulfurized by above-mentionedprocess. Then, CdS buffer layer of ∼60 nm thickness was deposited onsulfurized pyrite nanocube thin film on Mo by chemical bathdeposition (CBD). Subsequently, 50 nm i-ZnO followed by 600 nmZnO/Al layer were deposited by RF and DC magnetron sputtering.Finally, silver glue was printed on AZO layer to form top contactfingers.Characterization. Thin films X-ray diffraction (XRD) were taken

from Shimadzu 6000 using Cu Kα (λ = 1.54178 Å) radiation. T64000micro-Raman spectrometer with an incident power of 0.14 mW andlaser illumination of 532 nm wavelength was used for Raman analysis.

SEM images were acquired using JEOL JSM-7600F field-emissionscanning electron microscope. HRTEM analysis was performed usingJEOL TEM 2010 and 2100F. Optical absorption measurement wasdone using PerkinElmer Lambda 950 spectrophotometer. Four probeHall measurements were performed using MMR variable temperatureHall measurement system. Transient photocurrent measurementswere conducted using mode-locked Ti:sapphire laser seededTi:sapphire regenerative amplifier pumped OPA (spectra physics,Mai Tai-Spitfire Ace-TOPAS Prime). Current density versus voltage(J−V) curves of the solar cell were measured using solar simulator(San-EI electric, XEC-301S) equipped with a 450 W xenon lamp,which was coupled to Agilent semiconductor parameter analyzer(4155C). Solar cell performance was measured under illumination ofAM 1.5 (100 mW/cm2) calibrated using silicon reference cell(Fraunhofer).

ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsnano.6b00065.

Additional experimental data (PDF)

AUTHOR INFORMATIONCorresponding Author*E-mail: qihua@ntu.edu.sg.NotesThe authors declare no competing financial interest.

ACKNOWLEDGMENTSAuthors would also like to thank Kalon Gopinadhan for help intransport measurements. V.N. would like to thank the Ministryof Education, Singapore, for financial support (GrantMOE2011-T3-1-005). T.C.S. gratefully acknowledge thefinancial support from Nanyang Technological Universitystart-up grant M4080514, the Ministry of Education AcRFTier 1 grant RG101/15 and AcRF Tier 2 grants MOE2013-T2-1-081 and MOE2014-T2-1-044; and the NRF through theSingapore-Berkeley Research Initiative for Sustainable Energy(SinBeRISE) CREATE Programme. S.M would like toacknowledge SinBeRISE and Mid-IR photonics project bridgefunding (MIPS R-398-000-082-646) from ODPRT, NUS. Q.X.gratefully acknowledges Singapore National Research Founda-tion via an Investigatorship Award (NRF-NRFI2015-03), andSingapore Ministry of Education through two AcRF Tier2grants (MOE2011-T2-2-051 and MOE2015-T2-1-047).

REFERENCES(1) Chu, S.; Majumdar, A. Opportunities and Challenges for aSustainable Energy Future. Nature 2012, 488, 294−303.(2) Lewis, N. S.; Nocera, D. G. Powering the Planet: ChemicalChallenges in Solar Energy Utilization. Proc. Natl. Acad. Sci. U. S. A.2006, 103, 15729−15735.(3) Fthenakis, V. Sustainability of Photovoltaics: The Case for Thin-Film Solar Cells. Renewable Sustainable Energy Rev. 2009, 13, 2746−2750.(4) Razykov, T. M.; Ferekides, C. S.; Morel, D.; Stefanakos, E.; Ullal,H. S.; Upadhyaya, H. M. Solar Photovoltaic Electricity: Current Statusand Future Prospects. Sol. Energy 2011, 85, 1580−1608.(5) Lee, Y. S.; Bertoni, M.; Chan, M. K.; Ceder, G.; Buonassisi, T. InEarth Abundant Materials for High Efficiency Heterojunction Thin FilmSolar Cells, Photovoltaic Specialists Conference (PVSC), 2009 34thIEEE, 7−12 June 2009; 2009; IEEE; pp 002375−002377.(6) Todorov, T.; Gunawan, O.; Chey, S. J.; de Monsabert, T. G.;Prabhakar, A.; Mitzi, D. B. Progress Towards Marketable Earth-

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4438

Abundant Chalcogenide Solar Cells. Thin Solid Films 2011, 519,7378−7381.(7) Wadia, C.; Alivisatos, A. P.; Kammen, D. M. Materials AvailabilityExpands the Opportunity for Large-Scale Photovoltaics Deployment.Environ. Sci. Technol. 2009, 43, 2072−2077.(8) Woo, K.; Kim, Y.; Moon, J. A Non-Toxic, Solution-Processed,Earth Abundant Absorbing Layer for Thin-Film Solar Cells. EnergyEnviron. Sci. 2012, 5, 5340−5345.(9) Unold, T.; Schock, H. W. Nonconventional (Non-Silicon-Based)Photovoltaic Materials. Annu. Rev. Mater. Res. 2011, 41, 297−321.(10) Ennaoui, A.; Fiechter, S.; Pettenkofer, C.; Alonso-Vante, N.;Buker, K.; Bronold, M.; Hopfner, C.; Tributsch, H. Iron Disulfide forSolar Energy Conversion. Sol. Energy Mater. Sol. Cells 1993, 29, 289−370.(11) Ennaoui, A.; Tributsch, H. Iron Sulphide Solar Cells. Sol. Cells1984, 13, 197−200.(12) Caban-Acevedo, M.; Kaiser, N. S.; English, C. R.; Liang, D.;Thompson, B. J.; Chen, H.-E.; Czech, K. J.; Wright, J. C.; Hamers, R.J.; Jin, S. Ionization of High-Density Deep Donor Defect StatesExplains the Low Photovoltage of Iron Pyrite Single Crystals. J. Am.Chem. Soc. 2014, 136, 17163−17179.(13) Liang, D.; Caban-Acevedo, M.; Kaiser, N. S.; Jin, S. Gated HallEffect of Nanoplate Devices Reveals Surface-State-Induced SurfaceInversion in Iron Pyrite Semiconductor. Nano Lett. 2014, 14, 6754−6760.(14) Limpinsel, M.; Farhi, N.; Berry, N.; Lindemuth, J.; Perkins, C.L.; Lin, Q.; Law, M. An Inversion Layer at the Surface of n-type IronPyrite. Energy Environ. Sci. 2014, 7, 1974−1989.(15) Puthussery, J.; Seefeld, S.; Berry, N.; Gibbs, M.; Law, M.Colloidal Iron Pyrite (FeS2) Nanocrystal Inks for Thin-FilmPhotovoltaics. J. Am. Chem. Soc. 2011, 133, 716−719.(16) Wadia, C.; Wu, Y.; Gul, S.; Volkman, S. K.; Guo, J.; Alivisatos, A.P. Surfactant-Assisted Hydrothermal Synthesis of Single Phase PyriteFeS2 Nanocrystals. Chem. Mater. 2009, 21, 2568−2570.(17) Gong, M.; Kirkeminde, A.; Ren, S. Symmetry-Defying IronPyrite (FeS2) Nanocrystals through Oriented Attachment. Sci. Rep.2013, 3, 2092.(18) Yin, Y.; Alivisatos, A. P. Colloidal Nanocrystal Synthesis and theOrganic-Inorganic Interface. Nature 2005, 437, 664−670.(19) Kovalenko, M. V.; Scheele, M.; Talapin, D. V. ColloidalNanocrystals with Molecular Metal Chalcogenide Surface Ligands.Science 2009, 324, 1417−1420.(20) Gong, M.; Kirkeminde, A.; Xie, Y.; Lu, R.; Liu, J.; Wu, J. Z.; Ren,S. Iron Pyrite (FeS2) Broad Spectral and Magnetically ResponsivePhotodetectors. Adv. Opt. Mater. 2013, 1, 78−83.(21) Kirkeminde, A.; Scott, R.; Ren, S. All Inorganic Iron PyriteNano-Heterojunction Solar Cells. Nanoscale 2012, 4, 7649−7654.(22) Wang, D.-Y.; Jiang, Y.-T.; Lin, C.-C.; Li, S.-S.; Wang, Y.-T.;Chen, C.-C.; Chen, C.-W. Solution-Processable Pyrite FeS2 Nano-crystals for the Fabrication of Heterojunction Photodiodes with Visibleto NIR Photodetection. Adv. Mater. 2012, 24, 3415−3420.(23) Wang, Y.-C.; Wang, D.-Y.; Jiang, Y.-T.; Chen, H.-A.; Chen, C.-C.; Ho, K.-C.; Chou, H.-L.; Chen, C.-W. FeS2 Nanocrystal Ink as aCatalytic Electrode for Dye-Sensitized Solar Cells. Angew. Chem., Int.Ed. 2013, 52, 6694−6698.(24) Wu, J.; Liu, L.; Liu, S.; Yu, P.; Zheng, Z.; Shafa, M.; Zhou, Z.; Li,H.; Ji, H.; Wang, Z. M. High Responsivity Photoconductors Based onIron Pyrite Nanowires Using Sulfurization of Anodized Iron OxideNanotubes. Nano Lett. 2014, 14, 6002−6009.(25) Yang, Z.; Wang, M.; Shukla, S.; Zhu, Y.; Deng, J.; Ge, H.; Wang,X.; Xiong, Q. Developing Seedless Growth of ZnO Micro/NanowireArrays towards ZnO/FeS2/CuI P-I-N Photodiode Application. Sci.Rep. 2015, 5, 11377.(26) Steinhagen, C.; Harvey, T. B.; Stolle, C. J.; Harris, J.; Korgel, B.A. Pyrite Nanocrystal Solar Cells: Promising, or Fool’s Gold? J. Phys.Chem. Lett. 2012, 3, 2352−2356.(27) Alfonso, D. R. Computational Investigation of FeS2 Surfaces andPrediction of Effects of Sulfur Environment on Stabilities. J. Phys.Chem. C 2010, 114, 8971−8980.

(28) Murphy, R.; Strongin, D. R. Surface Reactivity of Pyrite andRelated Sulfides. Surf. Sci. Rep. 2009, 64, 1−45.(29) Macpherson, H. A.; Stoldt, C. R. Iron Pyrite Nanocubes: Sizeand Shape Considerations for Photovoltaic Application. ACS Nano2012, 6, 8940−8949.(30) Bi, Y.; Yuan, Y.; Exstrom, C. L.; Darveau, S. A.; Huang, J. AirStable, Photosensitive, Phase Pure Iron Pyrite Nanocrystal Thin Filmsfor Photovoltaic Application. Nano Lett. 2011, 11, 4953−4957.(31) Li, W.; Doblinger, M.; Vaneski, A.; Rogach, A. L.; Jackel, F.;Feldmann, J. Pyrite Nanocrystals: Shape-Controlled Synthesis andTunable Optical Properties via Reversible Self-Assembly. J. Mater.Chem. 2011, 21, 17946−17952.(32) Birkholz, M.; Fiechter, S.; Hartmann, A.; Tributsch, H. SulfurDeficiency in Iron Pyrite (FeS2‑x) and its Consequences for Band-Structure Models. Phys. Rev. B: Condens. Matter Mater. Phys. 1991, 43,11926−11936.(33) Verble, J. L.; Wallis, R. F. Infrared Studies of Lattice Vibrationsin Iron Pyrite. Phys. Rev. 1969, 182, 783−789.(34) Shukla, S.; Loc, N. H.; Boix, P. P.; Koh, T. M.; Prabhakar, R. R.;Mulmudi, H. K.; Zhang, J.; Chen, S.; Ng, C. F.; Huan, C. H. A.;Mathews, N.; Sritharan, T.; Xiong, Q. Iron Pyrite Thin Film CounterElectrodes for Dye-Sensitized Solar Cells: High Efficiency for Iodineand Cobalt Redox Electrolyte Cells. ACS Nano 2014, 8, 10597−10605.(35) Ushioda, S. Raman Scattering from Phonons in Iron Pyrite(FeS2). Solid State Commun. 1972, 10, 307−310.(36) Caban-Acevedo, M.; Faber, M. S.; Tan, Y.; Hamers, R. J.; Jin, S.Synthesis and Properties of Semiconducting Iron Pyrite (FeS2)Nanowires. Nano Lett. 2012, 12, 1977−1982.(37) Birkholz, M.; Hartmann, A.; Fiechter, S.; Tributsch, H. DefectChemistry and Homogeneity Range of Pyrite and their Influence onthe Optoelectronic Behaviour. In Tenth E.C. Photovoltaic Solar EnergyConference; Luque, A., Sala, G., Palz, W., Dos Santos, G., Helm, P.,Eds.; Springer: The Netherlands, 1991; pp 96−99.(38) Kou, W. W.; Seehra, M. S. Optical Absorption in Iron Pyrite(FeS2). Phys. Rev. B: Condens. Matter Mater. Phys. 1978, 18, 7062−7068.(39) Seefeld, S.; Limpinsel, M.; Liu, Y.; Farhi, N.; Weber, A.; Zhang,Y.; Berry, N.; Kwon, Y. J.; Perkins, C. L.; Hemminger, J. C.; Wu, R.;Law, M. Iron Pyrite Thin Films Synthesized from an Fe(acac)3 Ink. J.Am. Chem. Soc. 2013, 135, 4412−4424.(40) Krishnamoorthy, A.; Herbert, F. W.; Sidney, Y.; Krystyn, J. V. V;Bilge, Y. Electronic States of Intrinsic Surface and Bulk Vacancies inFeS2. J. Phys.: Condens. Matter 2013, 25, 045004.(41) Hu, J.; Zhang, Y.; Law, M.; Wu, R. First-Principles Studies of theElectronic Properties of Native and Substitutional Anionic Defects inBulk Iron Pyrite. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 85,085203.(42) Yu, L.; Lany, S.; Kykyneshi, R.; Jieratum, V.; Ravichandran, R.;Pelatt, B.; Altschul, E.; Platt, H. A. S.; Wager, J. F.; Keszler, D. A.;Zunger, A. Iron Chalcogenide Photovoltaic Absorbers. Adv. EnergyMater. 2011, 1, 748−753.(43) John, S.; Soukoulis, C.; Cohen, M. H.; Economou, E. N. Theoryof Electron Band Tails and the Urbach Optical-Absorption Edge. Phys.Rev. Lett. 1986, 57, 1777−1780.(44) Urbach, F. The Long-Wavelength Edge of PhotographicSensitivity and of the Electronic Absorption of Solids. Phys. Rev.1953, 92, 1324−1324.(45) Tauc, J.; Grigorovici, R.; Vancu, A. Optical Properties andElectronic Structure of Amorphous Germanium. Phys. Status Solidi B1966, 15, 627−637.(46) Davis, E. A.; Mott, N. F. Conduction in Non-CrystallineSystems V. Conductivity, Optical Absorption and Photoconductivity inAmorphous Semiconductors. Philos. Mag. 1970, 22, 0903−0922.(47) Meester, B.; Reijnen, L.; Goossens, A.; Schoonman, J. Synthesisof Pyrite (FeS2) Thin Films by Low-Pressure MOCVD. Chem. Vap.Deposition 2000, 6, 121−128.(48) Simmons, J. H.; Potter, K. S. Chapter 5 - Optical Properties ofSemiconductors. In Optical Materials; Potter, J. H. S. S., Ed. AcademicPress: San Diego, CA, 2000; pp 191−263.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4439

(49) Lazic, P.; Armiento, R.; Herbert, F. W.; Chakraborty, R.; Sun,R.; Chan, M. K. Y.; Hartman, K.; Buonassisi, T.; Yildiz, B.; Ceder, G.Low Intensity Conduction States in FeS2: Implications for Absorption,Open-Circuit Voltage and Surface Recombination. J. Phys.: Condens.Matter 2013, 25, 465801.(50) Lucas, J. M.; Tuan, C.-C.; Lounis, S. D.; Britt, D. K.; Qiao, R.;Yang, W.; Lanzara, A.; Alivisatos, A. P. Ligand-Controlled ColloidalSynthesis and Electronic Structure Characterization of Cubic IronPyrite (FeS2) Nanocrystals. Chem. Mater. 2013, 25, 1615−1620.(51) PopovA.Disordered Semiconductors; Pan Stanford: Singapore,2011.(52) Unger, K.; Brodsky, M. H. Amorphous Semiconductors, Topicsin Applied Physics 36 Krist. Tech. 1981, , 16, 12−12.10.1002/crat.19810160104(53) Thouless, D. J. Electrons in Disordered Systems and the Theoryof Localization. Phys. Rep. 1974, 13, 93−142.(54) Shore, K. A.; Mott, N.F.; Davis, E.A. Electronic Processes inNon-crystalline Materials. Contemp. Phys. 2014, 55, 337−337.(55) Nandi, U.; Jana, D.; Talukdar, D. Scaling Description of Non-Ohmic Direct Current Conduction in Disordered Systems. Prog.Mater. Sci. 2015, 71, 1−92.(56) Caban-Acevedo, M.; Liang, D.; Chew, K. S.; DeGrave, J. P.;Kaiser, N. S.; Jin, S. Synthesis, Characterization, and Variable RangeHopping Transport of Pyrite (FeS2) Nanorods, Nanobelts, andNanoplates. ACS Nano 2013, 7, 1731−1739.(57) Lee, P. A.; Ramakrishnan, T. V. Disordered Electronic Systems.Rev. Mod. Phys. 1985, 57, 287−337.(58) Montano, P. A.; Seehra, M. S. Magnetism of Iron Pyrite (FeS2) A Mossbauer Study in an External Magnetic Field. Solid StateCommun. 1976, 20, 897−898.(59) Burgardt, P.; Seehra, M. S. Magnetic Susceptibility of Iron pyrite(FeS2) Between 4.2 and 620 K. Solid State Commun. 1977, 22, 153−156.(60) Zhang, X.; Scott, T.; Socha, T.; Nielsen, D.; Manno, M.;Johnson, M.; Yan, Y.; Losovyj, Y.; Dowben, P.; Aydil, E. S.; Leighton,C. Phase Stability and Stoichiometry in Thin Film Iron Pyrite: Impacton Electronic Transport Properties. ACS Appl. Mater. Interfaces 2015,7, 14130−14139.(61) Zhang, X.; Manno, M.; Baruth, A.; Johnson, M.; Aydil, E. S.;Leighton, C. Crossover From Nanoscopic Intergranular Hopping toConventional Charge Transport in Pyrite Thin Films. ACS Nano2013, 7, 2781−2789.(62) Herbert, F. W.; Krishnamoorthy, A.; Ma, W.; Van Vliet, K. J.;Yildiz, B. Dynamics of Point Defect Formation, Clustering and PitInitiation on the Pyrite Surface. Electrochim. Acta 2014, 127, 416−426.(63) Bronold, M.; Tomm, Y.; Jaegermann, W. Surface States onCubic d-Band Semiconductor Pyrite (FeS2). Surf. Sci. 1994, 314,L931−L936.(64) Zhang, Y. N.; Law, M.; Wu, R. Q. Atomistic Modeling of SulfurVacancy Diffusion Near Iron Pyrite Surfaces. J. Phys. Chem. C 2015,119, 24859−24864.(65) Ge, H.; Hai, L.; Prabhakar, R. R.; Ming, L. Y.; Sritharan, T.Evolution of Nanoplate Morphology, Structure and Chemistry DuringSynthesis of Pyrite by a Hot Injection Method. RSC Adv. 2014, 4,16489−16496.

ACS Nano Article

DOI: 10.1021/acsnano.6b00065ACS Nano 2016, 10, 4431−4440

4440