Solar Energy Materials & Solar Cells 95 (2011) 2907–2913

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells

0927-02

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/solmat

Structure, composition and optical properties of Cu2ZnSnS4 thin filmsdeposited by Pulsed Laser Deposition method

Lin Sun n, Jun He, Hui Kong, Fangyu Yue, Pingxiong Yang, Junhao Chu

Key Laboratory of Polar Materials and Devices, Ministry of Education, Department of Electronic Engineering, East China Normal University, Shanghai 200241

a r t i c l e i n f o

Article history:

Received 16 February 2011

Received in revised form

12 June 2011

Accepted 19 June 2011

Keywords:

Cu2ZnSnS4

Pulsed Laser Deposition

Structure

Composition

Optical properties

Thin-film solar cells

48/$ - see front matter & 2011 Elsevier B.V. A

016/j.solmat.2011.06.026

esponding author. Tel.: þ86 21 54345123; fa

ail addresses: lsun@ee.ecnu.edu.cn, bolissun@

a b s t r a c t

Polycrystalline Cu2ZnSnS4 (CZTS) thin films have been directly deposited on heating Mo-coated glass

substrates by Pulsed Laser Deposition (PLD) method. The results of energy dispersive X-ray spectro-

scopy (EDX) indicate that these CZTS thin films are Cu-rich and S-poor. The combination of X-ray

diffraction (XRD) results and Raman spectroscopy reveals that these thin films exhibit strong

preferential orientation of grains along [1 1 2] direction and small Cu2�xS phase easily exists in CZTS

thin films. The lattice parameters and grain sizes have been examined based on XRD patterns and Atom

Force Microscopy (AFM). The band gap (Eg) of CZTS thin films, which are determined by reflection

spectroscopy varies from 1.53 to 1.98 eV, depending on substrate temperature (Tsub). The optical

absorption coefficient of CZTS thin film (Tsub¼450 1C) measured by spectroscopic ellipsometry (SE) is

above 104 cm�1.

& 2011 Elsevier B.V. All rights reserved.

1. Introduction

CZTS thin film is one of the most promising photovoltaicmaterials as the absorber of thin-film solar cells because it hasoptimal band gap (EgE1.4–1.5 eV) and high absorption coeffi-cient of 104 cm�1. More importantly, CZTS consists of abundantand non-toxic elements, so researches on CZTS thin-film solarcells have been increasing significantly in recent years. There aresome reports in which CZTS thin-film solar cells were prepared byseveral typical techniques, such as thermal evaporation [1,2], thesulfurization of electron-beam-evaporated precursors [3], sput-tering [4–6], electrodeposition [7,8], spray pyrolysis [9], solution-based synthesis [10] PLD [11–13], etc. Up to now, the highestconversion efficiency of CZTS-based thin-film solar cell has reachedto 9.6%, which was reported by Todorov et al. and it was fabricatedby chemical solution-based method [14]. However compared withthe conversion efficiency of Cu(In,Ga)Se2 thin-film solar cell, whichhas approached to 20% for singe junction solar cell [15], CZTS thin-film solar cell needs more systematic researches to further improveconversion efficiency.

PLD technique can be used to deposit high-quality films withcomplex compositions because it has the advantages of offeringstoichiometric preservation during materials transfer from targetto substrate and good crystallinity due to the highly energetic

ll rights reserved.

x: þ86 21 54345119.

hotmail.com (L. Sun).

species [16]. Thus CZTS thin film as a quaternary compoundshould be very suitable to produce from a quaternary CZTS targetby PLD. In previous studies, Moriya and Pawar et al. preparedpolycrystalline CZTS thin films using a two-step process by PLDand Moriya group even fabricated a thinfilm solar cell based onCZTS thin-film as absorber layer. However, the two-step processinvolves amorphous CZTS thin films deposited at room tempera-ture by PLD and subsequently annealing treatment at N2 or H2Satmosphere to obtain polycrystalline CZTS thin films. It is obviousthat the post-annealing adds to the fabrication procedure ofpolycrystalline CZTS thin films, solar cells and time-cost.

In this article, we report the successful growth of polycrystal-line quaternary CZTS thin films deposited on Mo-coated glasssubstrates using a one-step process by PLD without post-anneal-ing treatment, and have investigated the effect of substratetemperature on structure, composition and optical properties ofthese CZTS thin films.

2. Experimental

CZTS thin films were deposited on Mo-coated glass substrates byPLD from a sintered pellet target (30 mm diameter and 4 mm thick).The CZTS pellet was synthesized by the solid-state reaction of Cu2S,ZnS and SnS2 powders mixed at 1:1:1 mol ratio. These compoundpowders were pressed into a pellet and sintered at 700 1C for 4 hoursin a sealed alumina tube-furnace under an atmosphere of Ar. TheMo-coated glass substrates were ultrasonically cleaned in acetone,

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–29132908

ethanol, and distilled water, and dried in a nitrogen gas streambefore being put into vacuum chamber for depositing CZTSthin films. The deposition chamber was initially evacuated to2.0�10–4 Pa using turbo molecular pump. The substrates wereplaced on a rotating heater and the distance between target andsubstrate was fixed as 5 cm. KrF excimer laser pulses with awavelength of 248 nm, pulse duration of 30 ns, pulse energy of200 mJ, and a repetition rate of 5 Hz were converged onto thesintered CZTS pellet target, and the plume–like ablated fragmentswere deposited on a Mo-coated glass substrate. The target wasrotated slowly during depositing CZTS thin films in order to avoidpit formation and to ensure uniform ablation. The substrate tem-perature varied from 300 to 450 1C in the increment of 50 1C. Thethickness of the deposited films is approximately 1.2 mm.

The chemical composition, surface morphology and structuralproperties of these CZTS thin films prepared on Mo-coated glasssubstrates were examined by EDX (Inca, Oxford, UK, AFM (Veeco,Dimension 3100, USA), Raman spectroscopy (Jobin Yvon LabRAMHR 800UV MicroPL, 514.5 nm laser) and XRD ( Rigaku DMAX2500,Japan). The optical properties of these CZTS thin films weremeasured by reflectance spectroscopy (PerkinElmer Lambda950) and SE (SC630UVN, Shanghai Sanco Instrument, Co., Ltd.,China). All measurements were performed at room temperature.

3. Results and discussion

3.1. Composition analysis

EDX measurement for CZTS samples was performed with anacceleration voltage of 20 kV and the size of the investigated areais above 100 mm�100 mm. Fig. 1 gives the atomic percent of CZTSthin films as the function of substrate temperature. It can beobviously seen that all CZTS thin films deposited at differentsubstrate temperatures are non-stoichiometric and all thin filmsare of Cu-rich, Zn-poor and S-poor states. The enrichment of Cucan be attributed to its relatively lighter mass. During PulsedLaser Deposition, the velocities of elements are dependent ontheir mass [17]. Cu as a lighter-mass element has a higher flowspeed and more forward-peaked distribution than other higher-mass elements, which leads to the enrichment of Cu in CZTS thinfilms deposited by PLD [18]. Although Zn and S elements are alsolighter mass elements, they are volatile elements and thus theloss of Zn and S in CZTS thin films is possibly due to the fact that

Fig. 1. Atomic percent of CZTS thin films as the function of substrate temperature.

The line serves only as a guide to the eye.

the re-evaporation of Zn and S elements surpasses the role of lightmass during film growth at heating substrates. It is notable thatCu-rich and Zn-poor conditions are required to form single-phaseCZTS samples, according to CZTS-related first-principle calcula-tions [19,20]. Therefore, Cu-rich and Zn-poor conditions in ourCZTS samples are beneficial to the single-phase growth of highquality CZTS thin films. However, it should be noted that the CZTSthin film solar cells with the highest conversion efficiency areusually grown with the opposite compositional requirements(i.e. Cu-poor and Zn-rich) [14].

3.2. Structure of CZTS thin films

3.2.1. XRD characterization and Raman spectroscopy

Fig. 2 shows XRD patterns of CZTS thin films with differentTsub. It can be seen that [1 1 2]-oriented texture is dominant for allCZTS thin films with different Tsub. For the CZTS thin filmsdeposited at Tsub¼350 and 400 1C, the peak near 2y¼321 (markedwith ‘‘n’’ symbol in Fig. 2 ) cannot be indexed by CZTS standardXRD patterns (JCPDS 26-0575) and possibly originates fromCu2�xS phase, which can be confirmed by the following Ramanspectroscopy. Note that EDX analysis for these samples shown inFig. 1 demonstrates a significantly increased Cu-share. The pre-sence of the Cu2�xS impurity phase in CZTS thin films withTsub¼350 and 400 1C is likely to be related with the Cu-rich inthese CZTS thin films. However, no Cu2�xS impurity phase can bedetected in the XRD pattern of the CZTS thin films with Tsub¼ 300and 450 1C, though they are also Cu-rich. This observation can beexplained from the local inhomogeneous composition-distribu-tion in CZTS target used by PLD. It has been reported that theconventional target processing technology to fabricate quaternarycompound such as Cu(In,Ga)Se2 usually has some problems aboutlocal deviation of the chemical composition, inhomogeneous micro-structures and porous micro-regions [21]. The quaternary CZTStarget used by PLD was prepared with Cu2S, SnS2 and ZnS powdersusing conventional solid-state reaction sintering method, so thisCZTS target maybe has inhomogeneous composition-distribution insome local micro-regions, which ultimately leads to the existence ofsmall Cu2�xS as impurity phase in CZTS thin films deposited atTsub¼350 and 400 1C. This result implies that it is important toensure the homogeneous chemical composition–distribution in thewhole quaternary CZTS target for depositing pure CZTS thin films.There is another possible explanation for the presence of Cu2�xSphase. Although theoretical calculations have demonstrated thatthe Cu-rich and Zn-poor requirement benefits the growth of CZTSsingle phase, more excessive Cu is likely to induce the presence ofthe secondary phase like Cu2�xS because the stable chemical

Fig. 2. XRD patterns of CZTS thin films deposited at different substrate tempera-

tures (‘‘n’’ symbols represent Cu2�xS secondary phase).

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–2913 2909

potential region for CZTS is very small [19,20]. Thus, compared tothe samples with Tsub¼300 and 450 1C, too excessive Cu content insamples with Tsub¼350 and 400 1C easily leads to phase segrega-tion or the formation of Cu2�xS secondary phase. In fact, similarresult has also been experimentally observed by other researchersin the fabrication of CZTS thin films. For instance, it has beenreported that Cu2�xS impurity phase also easily occurs in Cu-richCZTS thin films [22].

As for the stability of quaternary CZTS compound, someexperimental researches have pointed that at higher tempera-tures (above 550 1C) quaternary CZTS phases decompose andCu2�xS and ZnS remain as solid phases in the films, while SnSand S as gas phases evaporate significantly [23,24]. But in thefabrication of our CZTS samples, the deposition temperature (Tsub)does not surpass 450 1C and no post-annealing treatment isperformed after growing thin films in order to avoid the possibledecomposition of CZTS. More importantly, no significant loss of Snelement is observed in EDX results shown in Fig. 1, implying ourCZTS samples are stable and no decomposition reaction happens.

Fig. 3 shows room temperature Raman scattering results ofCZTS films deposited at different substrate temperatures. Thedominant Raman peak at about 329 cm�1 are evidently observedfor CZTS thin films with Tsub¼300, 350 and 450 1C, whereas forthe CZTS thin-film with Tsub¼400 1C there is a small peak at465 cm�1 besides the main Raman peak at 329 cm�1. Theshoulder peaks at about 350 cm�1 may indicate a small amountof ZnS occurs in these samples. The main Raman peak of CZTS thinfilm or monograin powder reported in previous literature is near338 cm�1 [25]. However, the internal compressive stress canexist in CZTS films without annealing under sulfur atmosphereand this internal strain will induce the Raman peak near

Fig. 3. Raman spectra of CZTS films deposited at different substrate temperatures.

Table 1Lattice constants and crystallite sizes of CZTS thin films obtained from XRD patterns.

Substrate

temperature Tsub (1C)

FWHM of (1 1 2)

peak (deg)

Crystallite

size (d) (nm)

300 0.158 57

350 0.182 50

400 0.181 50

450 0.201 45

Note: For standard bulk CZTS powder data from JCPDS 26-0575, a¼5.427 A, c¼10.84 A

338 cm�1 to shift slightly towards the low wave number direc-tion [9]. Considering that our CZTS thin films deposited by PLD donot experience post-annealing, the internal strain is likely tooccur in these CZTS samples. Moreover, the shrinking of substratewhen cooling down may also contribute to the internal strain inthese samples. In fact, the presence of the internal strain will beconfirmed by the shrunk lattice constants calculated from XRDpatterns (Table 1). Thus the peak at 329 cm�1 can be attributed tothe Raman characteristic peak of CZTS and the internal strainproduced in film-deposition process makes the position of Ramanpeak at 338 cm�1 move to 329 cm�1. Although the Raman peakof Cu2�xS is near 475 cm�1 [26], for CZTS thin film withTsub¼400 1C the small Raman peak at 465 cm�1 corresponds tothe Cu2�xS secondary phase based on the same reason. The XRDpatterns have demonstrated that a little diffraction peak ofCu2�xS near 2y¼321 is also detected for the CZTS thin film withTsub¼350 1C, the Raman peak at 465 cm�1 cannot observed in itscorresponding Raman spectrum, which is possibly due to only asmall amount of Cu2�xS impurity phase. Chen et al. [19,20] havepredicted that it is important to control the elemental chemicalpotentials during crystal growth of quaternary CZTS compoundsto prevent the formation of secondary phases like CuS, ZnS andCu2SnS3, and further pointed out that Cu-rich/Zn-poor conditionswere necessary for the growth of single-phase CZTS compoundusing first-principle calculation . Chen’s calculation just explainswhy our samples with Tsub¼300 and 450 1C have the dominantCZTS kesterite phase because these two samples are Cu-rich andZn-poor. For these samples with Tsub¼350 and 400 1C, they alsoshow the dominant CZTS kesterite phase except the occurrence ofsmall Cu2�xS secondary phase, which originates from moreexcessive Cu element in these two samples compared to thoseof Tsub¼300 and 450 1C.

All of XRD patterns in Fig. 2 show predominant [1 1 2] orienta-tion regardless of deposition temperature (Tsub). In order to evaluatethe preferred orientation degree dependent on Tsub in these films,we define the variable R1 as the ratio of intensity of (1 1 2) peak tothe sum of intensities of all peaks in the XRD pattern [27]

R1 ¼I112P

all peaksIhklð1Þ

For the randomly oriented CZTS powder samples given byJCPDS file no. 26-0575, R1¼0.37. For a complete [1 1 2]-texture ofCZTS thin film, R1 should be equal to 1. Similarly we use thevariable R2 to describe the [2 2 0] orientation degree of CZTS thinfilms with variable Tsub.

R2 ¼I220P

all peaksIhklð2Þ

Fig. 4 demonstrates the variation of R1 and R2 as the depen-dence of Tsub. It can be observed that R1 value gradually decreasesfrom 0.968 for Tsub¼300 1C to 0.741 for Tsub¼450 1C whereasR2 value increases slowly from 0.017 for Tsub¼300 1C to 0.079for Tsub¼450 1C. It can be concluded that although [1 1 2]

Lattice

constant a (A)

Lattice

constant c (A)

Volume of

unit cell v (A3)

5.365 10.705 308.124

5.291 11.046 309.229

5.362 10.508 302.116

5.355 10.638 305.056

˚ and v¼319.26 A3.

Fig. 4. Degree of [1 1 2] orientation ( R1) and [2 2 0] orientation (R2) of CZTS thin

films deposited at different substrate temperatures by PLD. The line serves only as

a guide to the eye.

Fig. 5. AFM topography images of the CZTS thin films grown at different substrate

temperatures: (a) Tsub¼300 1C, 1�1 mm2 scan area and (b) Tsub¼450 1C,

2�2 mm2 scan area.

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–29132910

preferential orientation is absolutely dominant in all CZTS thinfilms deposited at various Tsub, [2 2 0] orientation degreeincreases gradually with the enhancement in deposition tem-perature (Tsub), accompanied with the decrease in [1 1 2] orienta-tion degree. For CuInSe2 or Cu(In,Ga)Se2 thin films grown onMo-coated soda-lime glass, [2 2 0]/[2 0 4] orientation can bemanipulated by controlling growth temperature (i.e. Tsub) [28],and (2 2 0)-textured Cu(In,Ga)Se2 thin films have some specialfeatures such as lower density of non-radiative recombinationcenters [29], more open structure [30] and stronger inactive grainboundary [31], in contrast with [1 1 2] texture thin films. CZTShas the crystal-structural descendant from CuInSe2 by the sub-stitution of Zn, Sn for In and S for Se, so we expect that CZTS thinfilms inherit the feature of Cu(In,Ga)Se2 thin films and CZTS thinfilms with more [2 2 0] orientation degree have better optoelec-tronic properties in thin-film solar cell applications.

Calculations of the mean sizes of crystallites in CZTS thin filmsdeposited at various substrate temperatures are performedfrom the strongest (112) diffraction peak broadening using theScherrer’s formula [32]

d¼0:9l

Br cos y: ð3Þ

Where l is the wavelength of CuKa radiation, and Br is the full-width half-maximum (FWHM) of (1 1 2) peak. In addition, latticeconstants a, c and volume of unit cell v are calculated according toBragg’s law for CZTS thin films with various Tsub. The FWHM of(1 1 2) peak, calculated values of crystallite size and latticeconstants for all CZTS thin film samples are listed in Table 1. Ascan be seen in Table 1, all lattice constants and volumes of unitcell v of CZTS thin films deposited on various Tsub are smaller thanthat of bulk CZTS powder data from JCPDS 26-0575. This behaviorimplies that the compressive internal stress exists in all CZTS thinfilms. It should be emphasized that the existence of the smallamount of Cu2�xS secondary phase in CZTS thin films withTsub¼350 and 400 1C seems not to affect the character of internalstress. The evidence for it is that the main Raman peaks of allsamples appear at 329 cm�1 whether there is Cu2�xS phase inthese samples or not.

3.2.2. Surface morphology of CZTS thin films

These crystallite sizes of CZTS thin films, calculated fromScherer’s formula, are between 40 and 60 nm, which are thevolume averaged values. The surface morphology of CZTS thinfilms deposited on various Tsub is carried out using AFM and it can

give the surface grain sizes directly. Fig. 5 shows AFM morphologyimages of CZTS thin films grown at 300 and 450 1C, respectively.The surface of CZTS thin films grown at 300 1C shows a texturedsurface with uniform island-like topography. It suggests that suchtopography originates from the island growth of the Volmer–Weber mode and the kinetic energy at low substrate temperatureis not sufficient for the coalescence of island-like crystallites [33].When the substrate temperature increases up to 450 1C, somesuper-structure of clusters with sizes of roughly 500 nm diametercan be clearly observed in Fig. 5(b). These crystallite clusters arethe result of crystallites coalescence because the higher substratetemperature increases the surface mobility. The grain sizes of subgrains in clusters for CZTS thin films with Tsub¼450 1C are in therange of several tens of nanometers as well as grain sizes of CZTS

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–2913 2911

thin films with Tsub¼300 1C, which correspond in order ofmagnitude with crystallite sizes calculated from XRD patterns.

3.3. Optical properties

3.3.1. Reflection spectroscopy

Fig. 6 gives optical Eg of CZTS thin films deposited on variousTsub and their respective reflection spectra curves. Eg values ofCZTS thin films are estimated from reflection spectra by means ofthe envelope method, which have been successfully used toprecisely determine Eg values of Cu(In,Ga)Se2 thin films [34]. Asshown in Fig. 6 (inset), it is obvious that CZTS thin films aretransparent in the pronounced oscillation regions, which areoriginated from the interference of light reflected at the air/CZTSfilm and CZTS film/Mo-coated substrate interfaces. No oscillationsappear for the high energy regions (4Eg) because the photons arefully absorbed by the CZTS thin films. Thus, Eg approximatelycorresponds to the photon energy where the oscillations in thereflection spectra begin to disappear. According to the aboveapproximated approach, Eg values are actually larger than truevalues due to the insufficient thickness of CZTS thin film (filmthickness E1.2 mm). In order to determine accurately Eg values,the method to determine Eg (evaluated error: 70.05 eV) fromthe ‘‘oscillation-free’’ reflectance (R) is conducted. In oscillationregions, the interference maxima and minima are fitted by the

Fig. 6. Optical band gap (Eg) determination using an approximated absorption coeffi

(d) 450 1C. Note that absorption coefficient (a) is not measured but derived from the re

Tsub, along with the respective envelopes of the interference extrema and the construc

figure legend, the reader is referred to the web version of this article).

second order polynomials and R is obtained by the mean of thesetwo fitting curves. Additionally, the absorption coefficient (a) isproportional to ln[(R-Rmin)�1] [34,35], where Rmin is the minimalreflectance in the considered spectra region. As CZTS is a directsemiconductor, Eg can be eventually determined by extrapolatingthe linear (ahn)2 vs. hn plots to (ahn)2

¼0, as depicted in Fig. 6.Therefore, Eg of CZTS thin films is 1.98, 1.79, 1.67 and 1.53 eV forTsub¼300, 350, 400 and 450 1C, respectively.

The band gap (EgE1.5 eV) of CZTS thin films with Tsub¼450 1Cis in good agreement with the experimental and theoreticalvalues reported by other researchers [4,6,11,36,37]. However,Eg of CZTS thin films grown at Tsub¼300, 350 and 400 1C is largerthan that of Tsub¼450 1C and Eg increases gradually with thedecrease in Tsub. It is reported that secondary phases such as ZnS,SnS2 and Cu2SnS3 easily occur under the lower growth tempera-tures during the CZTS synthesis process [26]. In addition, there isa recent experimental study that has strongly demonstrated theexistence of ZnSe in Cu2ZnSnSe4 (CZTSe) thin films is the possiblereason for the overestimation of overall Eg and a fraction of ZnSewithin CZTSe can make the measured Eg of CZTSe increase from1.0 to 1.5 eV or more [38]. Similarly, the measured Eg of CZTS thinfilms will be also larger than the true value if ZnS occurs in CZTSthin films. In fact, ZnS is hardly distinguishable from CZTS by XRDpattern since their diffraction peaks are too similar to distinguish.On the other hand, though Raman scattering is used to detect

cient for CZTS thin films with various Tsub (a) 300 1C, (b) 350 1C, (c) 400 1C and

flection spectra. Insets: optical reflection spectra of CZTS thin films with different

ted mean values (blue lines). (For interpretation of the references to colour in this

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–29132912

impurity phase in CZTS thin films [26], our Raman scatteringresults (Fig. 3) cannot rule out the presence of ZnS phase in CZTSthin films because the broad Raman peak of 329 cm�1 is likely tooverlap the Raman peak of ZnS (350 cm�1). Hence, the determi-nation of Eg may imply that a small amount of ZnS occurs in CZTSthin films when Tsubr400 1C and the presence of ZnS in CZTS ispossibly responsible for the larger Eg values of CZTS thin filmswith lower growth temperature ( i.e. Tsubr400 1C). The drop in Eg

may suggest that the content of ZnS phase in CZTS thin filmsgradually decreases with the increase in growth temperature(Tsub) and only the smallest amount of ZnS occurs in CZTS thinfilms for Tsub¼450 1C since its Eg is just in the vicinity of 1.5 eV.For CZTS thin films with Tsub¼350 and 400 1C, the existence ofsmall Cu2�xS phase should not contribute to the larger Eg, becausethe band gap of Cu2�xS is only 1.3 eV [39]. Finally, it can beconfirmed from the optical measurement that the sample withTsub¼ 450 1C is most dominated by CZTS kesterite structure andits band gap (EgE1.5 eV) makes it very suitable as the photo-voltaic absorber of thin-film solar cells.

3.3.2. Absorption coefficient of CZTS thin films

SE is a very powerful approach to obtain optical constantspectra (including refractive index n and extinction coefficient k)and the optical absorption coefficient (a) can be calculated usingthe well-known relation:

a¼ 4pk

lð4Þ

where l is the wavelength of photon, k is the extinctioncoefficient.

Fig. 7 shows the optical absorption coefficient (a) of CZTS thinfilm with Tsub¼450 1C and the determination of Eg by means of SEmethod. It should be noted that optical constant spectra are notdirectly measured but extracted by fitting the experimental datausing suitable structural and physical models for the SE method.In our case, we adopt air/CZTS thin film/Mo-coated substrate tri-layers structural model and Lorenz Oscillator model to fit theexperimental data. As can be seen in Fig. 7, the band gap Eg ofCZTS thin film with Tsub¼450 1C is estimated to be 1.43 eV andthe absorption coefficient exceeds quickly 104 cm�1 when photonenergy is higher than Eg. Obviously, Eg determined by the REmethod is slightly smaller than that of reflection spectra mea-surement. But it should be pointed out that the discrepancy iswithin 0.1 eV and can be originated from the influence of differentmeasurement methods. Therefore, the SE method further con-firms that CZTS thin film grown at Tsub¼450 1C is one of the most

Fig. 7. Optical absorption coefficient of CZTS thin film with Tsub¼450 1C and the

determination of Eg (inset) from spectroscopic ellipsometry.

promising absorber materials for thin-film solar cells due to itshigh absorption coefficient (4104 cm�1) and optimal band gap(EgE1.5 eV).

4. Conclusion

Polycrystalline CZTS thin films have been successfully grownon the heated Mo-coated glass substrates directly from the singleCZTS compound target by Pulsed Laser Deposition. All of CZTSthin films deposited at different Tsub are Cu-rich and S-deficient interms of EDX results. XRD and Raman scattering results revealthat CZTS thin films have strong [1 1 2] texture regardless of Tsub

whereas a little fraction of Cu2�xS occurs in CZTS samples withTsub¼350 and 400 1C. The lattice parameters calculated from XRDpatterns are relatively smaller and the main Raman peaks of CZTSshift from 338to 329 cm�1 position, suggesting that the com-pressive internal strain exists in all of CZTS samples. AFM showsthe homogeneous and dense surface morphology and XRD ana-lysis by the Scherrer formula gives the grain sizes of CZTS thinfilms. Analysis of reflection spectra for CZTS thin films, depositedat different Tsub, reveals that Eg gradually decreases from 1.98 to1.53 eV with increasing Tsub and the larger Eg for CZTS sampleswith lower Tsub(r400 1C) is presumably attributed to the pre-sence of the fraction of ZnS in these CZTS compounds. The opticalcoefficient of CZTS thin film with Tsub¼450 1C is evaluated to behigher than 104 cm�1 by means of the SE method. Hence, opticalmeasurements together with XRD and Raman spectroscopydemonstrate that CZTS thin film grown at Tsub¼450 1C by thePLD technique is most dominated by CZTS kesterite structure andit has ideal band gap (EgE1.5 eV) and high optical absorptioncoefficient.

Acknowledgment

The authors would like to thank the guidance of ProfessorZhigao Hu group in optical measurements, the beneficial discus-sion of Associate Professor Shiyou Chen on experimental data andthe help of Professor Xiaodong Tang group in AFM measurement.This project was financed by specialized Research Fund for theDoctoral Program of Higher Education of China (Grant no.20100076120009), the National Science Foundation for Post-doctoral Scientists of China (Grant no.20080440083), the MajorState Basic Research Development Program of China (Grant no.2007CB924901), the Science and Technology Commission ofShanghai Municipality Project (Grant nos. 11ZR1411400,09ZR1409200 and 10DJ1400200), the National Natural ScienceFoundation of China (Grant nos. 60990312 and 61076060) andPCSIRT in University.

References

[1] A. Weber, H. Krquth, S. Perlt, B. Schubert, I. Kotschau, S. Schorr, H.W. Schock,Multi-stage evaporation of Cu2ZnSnS4 thin films, Thin Solid Films 517 (2009)2524–2526.

[2] K. Wang, O. Gunawan, T. Todorov, B. Shin, S.J. Chey, N.A. Bojarczuk, D. Mitzi,S. Guha, Thermally evaporated Cu2ZnSnS4 solar cells, Appl. Phys. Lett. 97(2010) 143508-1-3.

[3] H. Katagiri, K. Saitoh, T. Washio, H. Shinohara, T. Kurumadani, S. Miyajima,Development of thin film solar cell based on Cu2ZnSnS4 thin films, Sol. EnergyMater. Sol. Cells 65 (2001) 141–148.

[4] J. Seol, S. Lee, J. Lee, H. Nam, H. Nam, K. Kim, Electrical and optical propertiesof Cu2ZnSnS4 thin films prepared by rf magnetron sputtering process, Sol.Energy Mater. Sol. Cells 75 (2003) 155–162.

[5] K. Jimbo, R. Kimura, T. Kamimura, S. Yamada, W. Maw, H. Araki, K. Oishi,H. Katagiri, Cu2ZnSnS4-type thin films solar cells using abundant materials,Thin Solid Films 515 (2007) 5997–5999.

L. Sun et al. / Solar Energy Materials & Solar Cells 95 (2011) 2907–2913 2913

[6] F. Liu, Y. Li, K. Zhang, B. Wang, C. Yan, Y. Lai, Z. Zhang, J. Li, Y. Liu, In situgrowth of Cu2ZnSnS4 thin films by reactive magnetron co-sputtering, Sol.Energy Mater. Sol. Cells 94 (2010) 2431–2434.

[7] C.P. Chan, H. Lam, C. Surya, Preparation of Cu2ZnSnS4 films by electrodeposi-tion using ionic liquids, Sol. Energy Mater. Sol. Cells 94 (2010) 207–211.

[8] J.J. Scragg, P.J. Dale, L.M. Peter, Synthesis and characterization of Cu2ZnSnS4

absorber layers by an electrodeposition-annealing route, Thin Solid films 517(2009) 2481–2484.

[9] H. Yoo, J. Kim, Comparative study of Cu2ZnSnS4 film growth, Sol. EnergyMater. Sol. Cells 95 (2011) 239–244.

[10] S.C. Riha, B.A. Parkinson, A.L. Prieto, Solution-based synthesis and character-ization of Cu2ZnSnS4 Nanocrystals, J. Am. Chem. Soc. 131 (2009)12054–12055.

[11] K. Moriya, K. Tanaka, H. Uchiki, Cu2ZnSnS4 thin films annealed in H2Satmosphere for solar cell absorber prepared by ULSED laser deposition, Jpn.J. Appl. Phys. 47 (2008) 602–604.

[12] S.M. Pawar, A.V. Moholkar, I.K. Kim, S.W. Shin, J.H. Moon, J.I. Rhee, J.H. Kim,Effect of laser incident energy on the structural, morphological and opticalproperties of Cu2ZnSnS4 (CZTS) thin films, Curr.ent Appl. Phys. 10 (2010)565–569.

[13] K. Sekiguchi, K. Tanaka, K. Moriya, H. Uchiki, Epitaxial growth of Cu2ZnSnS4

thin films by pulsed laser deposition, Phys. Status Solidi C 3 (2006)2618–2621.

[14] T.K. Todorov, K.B. Reuter, D.B. Mitzi, High-efficieny solar cell with earth–abundant liquid-processed absorber, Adv. Mater. 22 (2010) 1–4.

[15] I. Repins, M.A. Contreras, B. Egaas, C. DeHart, J. Scharf, C.L. Perkins, B. To,R. Noufi, 19.9%-efficient ZnO/CdS/CuInGaSe2 solar cell with 81.2% fill factor,Prog. Photovolt: Res. Appl 16 (2008) 235–239.

[16] D.B. Chrisey, G.K. Hubler, Pulsed Laser Deposition of Thin films, John Wileyand Sons, Inc., New York, 1994 pp.14.

[17] P.K. Singh, J. Narayan, Pulsed-laser evaporation technique for deposition ofthin films: physics and theoretical model, Phys. Rev. B 41 (1990) 8843–8861.

[18] R.A. Wibowo, E.S. Lee, B. Munir, K.H. Kim, Pulsed laser deposition ofquaternary Cu2ZnSnSe4 thin films, Phys. Status Solidi A 204 (2007)3373–3379.

[19] S. Chen, X.G. Gong, A. Walsh, S.H. Wei, Defect physics of the kesterite thin-film solar cell absorber Cu2ZnSnS4, Appl. Phys. Lett. 96 (2010) 021902-1-3.

[20] S. Chen, J.H. Yang, X.G. Gong, A. Walsh, S.H. Wei, Intrinsic point defects andcomplexes in the quaternary kesterite semiconductor Cu2ZnSnS4, Phys. Rev.B 81 (2010) 245204.

[21] S. Britting, A. Borkowski, C. Dutsch, R. Simon, K.U. Vanosten, Development ofnovel target materials for Cu(In,Ga)Se-based solar cells, Plasma Process.Polym. 6 (2009) S25–28.

[22] H. Yoo, J. Kim, Growth of Cu2ZnSnS4 thin films using sulfurization of stackedmetallic films, Thin Solid Films 518 (2010) 6567–6572.

[23] A. Weber, R. Mainz, H.W. Schock, On the Sn loss from thin films of thematerial system Cu–Zn–Sn–S in high vacuum, J. Appl. Phys. 107 (2010)013516-1-6.

[24] A. Redinger, D.M. Berg, P.J. Dale, S. Siebentritt, The consequence of Kesteriteequilibria for efficient solar cells, J. Am. Chem. Soc. 133 (2011) 3320–3323.

[25] M. Altosaar, J. Raudoja, K. Timmo, M. Danilson, M. Grossberg, J. Krustok,E. Mellikov, Cu2Zn1-xCdxSn(Se1-ySy)4 solid solutions as absorber materials forsolar cells, Phys. Status Solidi. A 205 (2007) 167–170.

[26] P.A. Fernandes, P.M.P. Salome, A.F. da Cunha, Growth and Raman scatteringcharacterization of Cu2ZnSnS4 thin films, Thin Solid Films 517 (2009)2519–2523.

[27] J. Muller, J. Nowoczin, H. Schmitt, Composition, structure and optical proper-ties of sputtered thin films of CuInSe2, Thin Solid Films 496 (2006) 364–370.

[28] M.A. Contreras, B. Egaas, D. King, A. Swartzlander, T. Dullwever, Texturemanipulation of CuInSe2 thin films, Thin Solid Films 361–362 (2000) 167–171.

[29] M.A. Contreras, M.J. Romero, R. Noufi, Characterization of Cu(In,Ga)Se2

materials used in record performance solar cells, Thin Solid Films 511–512(2006) 51–54.

[30] M.A. Contreras, B. Egaas, K. Ramanathan, J. Hiltner, A. Swartzlander,F. Hasoon, R. Noufi, Progress toward 20% efficienct in Cu(In,Ga)Se2 poly-crystalline thin-film solar cells, Prog. Photovolt.: Res. Appl 7 (1999) 311–316.

[31] N. Ott, G. Hanna, U. Rau, J.H. Werner, H.P. Strunk, Texture of Cu(In,Ga)Se2 thinfilms and nanoscale cathodoluminescence, J. Phys.: Condens. Matter 16(2004) S85–S89.

[32] C.S. Batett, Structure of Metals, Crystallographic Methods, Principles andData, McGraw-Hill, New York, 1956, pp 156.

[33] N.M. Shah, C.J. Panchal, V.A. Kheraj, J.R. Ray, M.S. Desai, Growth, structuraland optical properties of copper indium diselenide thin films deposited bythermal evaporation method, Sol. Energy 83 (2009) 753–760.

[34] M. Bar, S. Nishiwaki, L. Weinhardt, S. Pookpanratana, O. Fuchs, M. Blum,W. Yang, J.D. Denlinger, W.N. Shafarman, C. Heshe, Depth-resolved band gapin Cu(In,Ga)(S,Se)2 thin films, Appl. Phys. Lett. 93 (2008) 244103-1-3.

[35] V. Kumar, S.K. Sharma, T.P. Sharma, V. Singh, Band gap determination in thickfilms from reflectance measurements, Opt. Mater. 12 (1999) 115–119.

[36] S. Chen, X.G. Gong, A. Walsh, S.H. Wei, Crystal and electronic band structureof Cu2ZnSnX4 (X¼S and Se) photovoltaic absorbers: first-principles insights,Appl. Phys. Lett. 94 (2009) 041903-1-3.

[37] S. Chen, A. Walsh, J.H. Yang, X.G. Gong, L. Sun, P.X. Yang, J.H. Chu, S.H. Wei,Compositional dependence of structural and electronic properties of theCu2ZnSn(S,Se)4 alloys for thin film solar cells, Phys. Rev. B 83 (2011) 125201.

[38] S. Ahn, S. Jung, J. Gwak, A. Cho, K. Shin, K. Yoon, D. Park, H. Cheong, J.H. Yun,Determination of band gap energy (Eg) of Cu2ZnSnSe4 thin films: on thediscrepancies of reported band gap values, Apl. Phys. Lett. 97 (2010) 021905-1-3.

[39] B.J. Mulder, Optical properties of an unusual form of thin chalcosite (Cu2S)crystals, Phys. Status Solidi A 15 (1973) 409–413.