Preparation of iron selenide films by selenization technique

Journal of Crystal Growth 241 (2002) 313–319

Preparation of iron selenide films by selenization technique

N. Hamdadoua, J.C. Bern"edeb,*, A. Khelila

aUniversit!e d’Oran, Laboratoire de Physique des Mat!eriaux et Composants pour l’Electronique, BP 1524 El Mnaouer Oran, AlgeriabLPSE-FSTN Universit!e de Nantes, Facult!e des Sciences et des Techniques, 2, rue de la Houssin"ere, PB 92208,

44322 Nantes Cedex 3, France

Received 10 February 2002; accepted 6 March 2002

Communicated by R. Kern

Abstract

Iron selenide films have been grown by selenization of evaporated iron thin films. The substrates temperature during

selenization was 573 or 723K. It is shown that whatever the substrate temperature is, for a partial pressure of Se of

3.75� 10�2 Pa, there is selenide compound formation. It is shown that the films are mainly composed of tetragonal

FeSe while some crystallites of orthorhombic FeSe2 are also present. XPS shows that before etching there is some

superficial selenium excess while some oxidation of Fe and Se is put in evidence at the surface of the films. However,

after etching the superficial oxidation disappears. After annealing for 2 h at 773K under vacuum the films become

crystallized in the hexagonal structure of FeSe, while no FeSe2 is detected. The Se/Fe ratio tends towards 1. When

annealed in the same conditions, but in selenium atmosphere, the films obtained are FeSe2 films. r 2002 Elsevier

Science B.V. All rights reserved.

PACS: 68.55.�a; 81.15.Ef

Keywords: A2. Selenization; A3. Physical vapor deposition processes; A3. Thin films; B2. Semiconducting iron selenide

1. Introduction

Due to the unusual structures and electronicproperties of transition metal chalcogenides, atten-tion has been paid to these materials with variousapplications. In the case of iron chalcogenidescompounds, much attention has been done topyrite FeS2 because of its potential as absorber insolar cells [1,2]. However, few is known for iron

selenide compounds, FeSe crystallize in the tetra-gonal structure (JCPDS 85-0735) and in thehexagonal structure (JCPDS 75-0608), while FeSe2crystallize in the orthorhombic marcasite-typestructure (JCPDS 21-0432) and in the cubicstructure (JCPDS 48-1881). The energy gap ofFeS2 (D1 eV) [3] made them quite interesting fortandem solar cells, while the use of FeSe in newmagnetic materials is under study [4].If low temperature chemical process has been

described to obtain transition metal selenides inbulk configuration [5], works on thin films are notnumerous. Molecular beam epitaxy has been usedto grow FeSe on GaAs single crystals [4]. In the

*Corresponding author. Tel.: +33-2-51-12-5530; fax: +33-2-

51-12-55-28.

E-mail address: [email protected]

(J.C. Bern"ede).

0022-0248/02/$ - see front matter r 2002 Elsevier Science B.V. All rights reserved.

PII: S 0 0 2 2 - 0 2 4 8 ( 0 2 ) 0 1 2 5 0 - 2

present work, a more simple process is used togrow polycrystalline iron selenide (FeSe, FeSe2) onglass substrates.

2. Experiments

2.1. Thin film growth

The films were synthesized by self-selenizationof Fe films deposited onto glass substrates. Thesubstrates used were soda lime chemically cleaned.The substrate temperature was controlled by achromel-alumel thermocouple attached with silverpaste to the sample surface. The substrate washeated using infrared lamps. The deposition rateand the deposited thickness were measured withthe help of a quartz monitor. The substrates wereat room temperature during iron deposition. Afilm of iron of thickness 50–100 nm was depositedby Joule effect using a tungsten crucible at a rateof 0.25 nm s�1 in a vacuum of 10�4 Pa. Then inorder to prevent any contamination effect the filmswere covered by a thin (100 nm) amorphousselenium film. Then the samples were heated at573 and 723K and selenized for half an hour in theselenization chamber. During selenization a broadhome made selenium crucible was used. In order toobtain stable selenization conditions the crucibletemperature was checked with a thermocouple insuch a way that the deposition rate was stable atabout 2.5 nm s�1.

2.2. Sample characterization

The structure of the films was examined using ananalytical X-ray system-type DIFFRACT ATV3.1 Siemens which uses a graphics programEVA. The wavelength l was 0.15406 nm.The full-width at half-maximum (FWHM) of thediffraction peaks was given directly by theDIFFRACT AT programs.1

The quantitative X-ray photoelectron spectro-scopy (XPS)2 studies were based on the determina-

tion of Fe2p3/2 and Se3p peak area with 0.125 and0.44, respectively, as sensitivity factors, which weregiven by the manufacturer, Leybold. The Se3p hasbeen used because there is some coverage effectbetween the Se3d and Fe3p peaks. The depthprofile was realized by recording successive XPSspectra obtained after ion etching for short pe-riods (B1min). Using an ion gun, sputteringwas accomplished at pressures of less than5� 10�4 Pa with a 10mA emission current and5 kV beam energy. The Ar+ ion beam was restoredover the entire sample surface. At the surface ofthe films there is a carbon–carbon bond corre-sponding to surface contamination. In the appa-ratus used, this C–C bond has a well-definedposition at 284.4 eV and the carbon peak was usedas a reference to estimate the electrical chargeeffect, as the film samples were on insulating glasssheet.Electronic microprobe analysis (EMPA) was

performed using a JEOL 5800 LV scanningelectron microscope equipped with a PGT X-raymicroanalysis system, in which X-rays weredetected by a germanium crystal.3

The surface topography and the cross-section ofthe films were observed with a field effect scanningelectron microscope JEOL F 6400.4

3. Experimental results and discussion

The selenization was performed by supplying Seat substrate temperature 573 and 723K. Thesetemperatures have been chosen because Takamuraet al. [4] have shown that below 653K theselenization attempt of Fe thin films failed, whileabove this temperature FeSe forms.In the present work it can be seen (Fig. 1) that

the films are composed of small crystallites of ironselenides. Two phases are crystallized, mainlytetragonal FeSe and some orthorhombic FeSe2.It should be noted that if the films appear poorly

1Measurements have been carried out at the LCS-IMN.2Measurements have been carried out at Nantes (University

CNRS).

3Measurements have been carried out at the micro-char-

acterization center of the FSTN.4Measurements have been carried out at the micro-char-

acterization centre of the FSTN.

N. Hamdadou et al. / Journal of Crystal Growth 241 (2002) 313–319314

crystallized since the peak intensity is small andthe back ground clearly visible, the FWHM of thediffraction suggests that the crystallites size D isnot so small (Table 1), which means that crystal-

lites should be embedded in a microcrystalline oramorphous matrix. Moreover, it appears that theFeSe crystallites are oriented along the (h k o)direction.The bulk atomic ratio estimated from EMPA

analysis is slightly higher than unity (1.1) and doesnot vary with the substrate temperature duringselenization. If there is only a small seleniumexcess in the bulk there is three Se for one Fe at thesurface as shown by XPS. Moreover, there is someoxygen contamination (after one minute of etching10 at. of oxygen) even if the relative oxygen atomicconcentration decreases steeply when the etchingtime increases (Fig. 2c).In fact it can be seen (Fig. 2) that there is a

superficial oxidation effect. The binding energy ofthe Fe2p3/2 peak corresponds to iron oxide [6](Fig. 2a), while the shape of the Se3d peak shouldbe attributed to partial oxidation (Fig. 2b). More-over, it can be seen that the sodium peak Na1 s isvisible (Fig. 2d), which means that sodium issuedfrom the soda lime substrate diffuses through thefilms. Similar behavior is well known to improvethe film properties in the case of chalcopyriteselenides such as CuInSe2 [7]. After etching there isa decrease of the binding energy of the Fe2pdoublet which can be attributed to a decrease ofthe oxidation state of this element. The seleniumpeak Se3d has been decomposed before and afteretching. The Se3d peak is asymmetrical since it is a3d3/2�5/2 doublet. Even if we have to be carefulwith the decomposition of the selenium peak, sincethe spin-orbit splitting of the Se3d doublet is only0.83 eV [8], while the relative intensities of the twocomponents (0.67 and 1) were fixed by thetheoretical ratio of their degeneracies 2J þ 1; the

10 20 30 40 50 60

FeS

e(20

1)FeS

e 2(1

21)

FeS

e(00

1)

FeS

e(01

1)

Inte

nsit

y (a

.u.)

10 20 30 40 50 600

50

100

150

200

250

FeS

e 2 (1

21)FeS

e (1

10)

FeS

e (0

01)

2θ (deg.)

2θ (deg.)

Inte

nsit

y (a

.u.)

(a)

(b)

Fig. 1. X-ray diffractogramms of iron films after selenization

half an hour at: (a) 573K; (b) 723K.

Table 1

X-ray diffraction and EMPA studies

Substrate temperature Ts (K) Post annealing Diffraction peaks visible D (nm) Fhko F00c At. Se/Fe

573 — FeSe+1 FeSe2 Tetragonal 35 80 10 1.11

Under vacuum FeSe Hexagonal 30 5 1.09

Under Se pressure FeSe2 Orthorhombic 40 65 2.17

450 — FeSe+1 FeSe2 Tetragonal 12 80 1.12

Under vacuum FeSe Hexagonal 35 — — 1.08

Under Se pressure FeSe2 Orthorhombic 45 90 2.28

N. Hamdadou et al. / Journal of Crystal Growth 241 (2002) 313–319 315

signal of the Se3d is high enough (at least 8500 cpsat is maximum) to propose a decomposition. It canbe seen (Fig. 3a) that to achieve a good agreementbetween experimental and theoretical curve, threeselenium contributions are needed.The central one (Se3d3/2=55 eV) corresponds to

neutral Se, the one situated at 53.7 eV can beattributed to FeSe while the one at 58.5 eVcorresponds to selenium oxide. The neutral Secorresponds to the selenium excess detected at thesurface of the films. After 1min of etching(Fig. 3b) the peak situated at 53.6 eV correspondsto FeSe while the other one confirms that some Seexcess is still present, while the oxide contributionhas disappeared. This is in good agreement withthe dramatic decrease of the intensity of theoxygen peak after etching. The Na1s peak hasnearly disappeared which confirms the accumula-

tion at the film surface. Similar qualitative resultshave been obtained whatever the value of Ts is.After this approach it can be said that even at

quite low temperature (T s ¼ 573K), iron selenidecan be grown by selenization. However, the filmsare not highly crystallized and moreover they areconstituted of, at least, two phases.In order to improve the film quality they have

been post-annealed. Two post-annealing processeshave been used. In the first process, the films arepost-annealed in vacuum at 773K for 2 h. In thesecond process, simultaneously with the samplesome Se powder is introduced into the Pyrex tubewhich is then sealed under vacuum. Here also wecontinue annealing for 2 h at 773K.When annealed at 773K under vacuum it

appears that the crystalline properties of the filmshave been modified (Fig. 4). The peak intensity is

Fig. 2. XPS spectrum of Fe2p (a), Se3d (b), O1s (c), Na1s (d), before (—) and after 1min etching (- - - ).


higher, moreover the films structure has beenmodified. Before annealing the FeSe films arecrystallized in the tetragonal structure and after inthe hexagonal structure. Moreover, the texturationof the crystallites has disappeared. The seleniumexcess has slightly decreased while the grain sizedoes not change. A visualization of the surface ofthese films allows to see that the films are formedof well-faceted crystallites, while some cracks arevisible (Fig. 5).When annealed in the presence of selenium, the

films are crystallized in the orthorhombic FeSe2structure (Fig. 6) and as shown by EMPA theatomic ratio Se/Fe is slightly higher than two and

Fig. 3. Decomposition of the S3d peak (a) before etching, (b)

after etching, experimental curve (- - -), theoretical curve (—)

for different contributions.

10 20 30 40 50 600

50

100

150

200

250

300

350

400

(101

)

(102

)

(110

)

26 (deg.)

Inte

nsit

y (a

.u.)

Fig. 4. X-ray diffractogramm of FeSe film after annealing

under vacuum 2h at 723K.

Fig. 5. SEM visualization of a FeSe film after annealing under

vacuum 2h at 723K.

10 20 30 40 50 600

200

400

600

800

1000

(110

)

(111

) (1

20)

(211

) (2

20)

26 (deg.)

Inte

nsit

y (a

.u.)

Fig. 6. X-ray diffractogramm of FeSe2 after annealing under Se

atmosphere 2 h at 723K.


the grain size is broader (Table 1). Here also thefilms appear clearly polycrystalline, however, nocracks are present in the films (Fig. 7).Therefore, the post-annealing treatment modi-

fies strongly the crystalline properties of the filmsand their composition, depending on the annealingconditions.Because the vapor pressure of Se is as high as

0.1 Pa at 523K, the Se cannot stay on a substrateheated at 573K or more if it does not reactschemically with the substrate. Presently, seleniumreacts with the iron films at 573K while more than653K are needed in the case of Takemura et al. [4].During such selenization process there is acompetition between the selenium escape speedfrom the substrate, which also increases with thesubstrate temperature and the iron seleniumreaction rate which also increases with thetemperature. In fact, the probability of contactbetween atoms of iron and selenium, i.e. thereaction probability, increases with the seleniumpressure in the selenization chamber. Therefore,the higher reactivity of our films by comparisonwith those of Takemura is related to our higher,selenium working pressure (D3.75� 10�2 Pa in thepresent work and D3.75� 10�4 Pa for Takemuraet al.). Such high selenium partial pressure allowsnot only FeSe formation but also as shown inTable 1 and Fig. 1 the formation of some FeSe2microcrystallites, which can justify the atomicratio Se/Fe higher than one. Probably, theselenium first reacts with iron to give FeSe and

then FeSe2, which can justify that at the filmssurface the Se concentration is far higher than inthe bulk. At the end of the selenization processthere is a gradient concentration of Se in the film.During the post-annealing at 773K for 2 h invacuum, an equilibrium state is reached.There is a redistribution of the selenium all over

the bulk of the film with disappearance of the bulkSe gradient concentration. This atom migrationmovement induces a new crystallization process.The films become crystallized in the hexagonalstructure, the selenium excess decreases slightlyand only one phase is present in the films, the FeSehexagonal phase.When annealed at 773K for 2 h in selenium

pressure, there is a permanent selenium feeding ofthe film up to the formation of stoichiometricFeSe2 films as shown by X-ray diffraction. The Seoverpressure allows not only the growth oforthorhombic FeSe2 but also it induces someselenium excess. When post-annealed under va-cuum the change of crystalline structure inducesatoms migration in the films which induces cracksapparition. When annealed under selenium pres-sure, the atom migration takes place with perma-nent atoms feeding of the films which avoid suchcracks formation.

4. Conclusion

We have shown that FeSeX films (X ¼ 1 and/or2) can be grown by selenization of an iron pre-evaporated film. The efficiency of the selenizationprocess depends not only on the substrate tem-perature but also on the selenium partial pressureduring the process. It is shown that the phase FeSeor FeSe2 can be chosen by using adequate post-annealing conditions. These post-annealing condi-tions allow also to improve the crystallizationproperties of the films.

Acknowledgements

This work has been financially supported by theFrance, Alg!erie, contrat CMEP.00MDU510.

Fig. 7. SEM visualization of FeSe2 after annealing under Se

atmosphere 2 h at 723K.


References

[1] A. Ennaoui, S. Fiechter, Ch. Pettenkofer, N. Alonso-Vante,

K. Buker, M. Bronold, Ch. Hopfner, H. Tributsch, Solar

Energy Mater. Solar Cells 29 (1993) 289.

[2] N. Hamdadou, A. Khelil, J.C. Bern"ede, Mater. Chem.

Phys., to be published.

[3] T. Harada, Phys. Soc. Japan 67 (1998) 1352.

[4] Y. Takemura, H. Suto, N. Honda, K. Kakuno, K. Saito,

J. Appl. Phys. 81 (1997) 5177.

[5] Z.H. Han, Y.P. Li, J. Lu, S.H. Yu, H.Q. Zhao, Y.T. Qian,

Mater. Res. Bull. 35 (2000) 1825.

[6] E. Baba Ali, J.C. Bern"ede, N. Barreau, Mater. Chem. Phys.

63 (2000) 208.

[7] C. Heske, R. Fink, D. Jacob, E. Umbach, in: H.S. Stephens

(Ed.), Proceedings of the 13th European Photovoltaic Solar

Energy Conference, Nice France, Bedford UK, 1995,

p. 2003.

[8] C. Cardinaud, G. Turban, B. Cros, M. Ribes, Thin Solid

Films 205 (1991) 165.


Documents

Preparation of iron selenide films by selenization technique