Properties of sputtered ZnS and ZnS:A (A = Er, Yb) films grown at low substratetemperaturesCarlos A. OrtizAstrid L. Giraldo-BetancurMartín A. Hernández-LandaverdeMarius Ramírez-CardonaArturoMendoza-Galván and Sergio Jiménez-Sandoval

Citation: Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 35, 031505 (2017); doi:10.1116/1.4978946View online: http://dx.doi.org/10.1116/1.4978946View Table of Contents: http://avs.scitation.org/toc/jva/35/3Published by the American Vacuum Society

Properties of sputtered ZnS and ZnS:A (A 5 Er, Yb) films grown at lowsubstrate temperatures

Carlos A. OrtizCentro de Investigaci�on y de Estudios Avanzados del IPN, Unidad Quer�etaro, Libramiento NorponienteNo. 2000, Fracc. Real de Juriquilla, 76230 Quer�etaro, Mexico and Centro de F�ısica Aplicada y Tecnolog�ıaAvanzada, Universidad Nacional Aut�onoma de Mexico, Boulevard Juriquilla 3001, Juriquilla,76230 Quer�etaro, Mexico

Astrid L. Giraldo-BetancurCONACyT-Centro de Investigaci�on y de Estudios Avanzados del IPN, Unidad Quer�etaro, LibramientoNorponiente No. 2000, Fracc. Real de Juriquilla, 76230 Quer�etaro, Mexico

Mart�ın A. Hern�andez-LandaverdeCentro de Investigaci�on y de Estudios Avanzados del IPN, Unidad Quer�etaro, Libramiento NorponienteNo. 2000, Fracc. Real de Juriquilla, 76230 Quer�etaro, Mexico

Marius Ram�ırez-CardonaCentro de Investigaciones en Ciencias de la Tierra y Materiales, Universidad Aut�onoma del Estado deHidalgo, Pachuca, 42039 Hidalgo, Mexico

Arturo Mendoza-Galv�an and Sergio Jim�enez-Sandovala)

Centro de Investigaci�on y de Estudios Avanzados del IPN, Unidad Quer�etaro, Libramiento NorponienteNo. 2000, Fracc. Real de Juriquilla, 76230 Quer�etaro, Mexico

(Received 1 October 2016; accepted 7 March 2017; published 21 March 2017)

Radio-frequency magnetron sputtering was used to deposit ZnS thin films doped with rare earth

elements (Er and Yb) on glass substrates at two different temperatures (room temperature and

180 �C). The crystal structure, surface morphology, and optical properties of the deposited films

were studied as a function of the deposition variables as a first step in the possible application of

lanthanide-doped ZnS as buffer layer and/or window layer in thin film photovoltaic devices. From

the x-ray diffraction patterns, it was found that the samples have preferred orientation depending

on the substrate temperature and on the doping level. Moreover, optical constants were determined

from spectroscopic ellipsometry measurements and found that samples of undoped ZnS, in particu-

lar, exhibit optical constants that are in good agreement with those reported in the literature for the

two ZnS crystalline phases. The band gap ðEgÞ was calculated using the extinction coefficient ðkÞ,which was reduced according to the degree of doping. Also, based on ultraviolet-visible spectros-

copy measurements, it was determined that the films have a transmittance of 80% for wavelengths

above 350 nm, and a shift of the absorption edge to longer wavelengths when the degree of doping

was increased. Finally, photoluminescence measurements were performed at room and low temper-

ature using below bandgap excitation to probe defect states within the gap, as well as the lumines-

cent characteristics of the Er3þ and Yb3þ ions in the ZnS host. VC 2017 American Vacuum Society.

[http://dx.doi.org/10.1116/1.4978946]

I. INTRODUCTION

The use of cadmium sulfide (CdS) layers is highly wide-

spread in the fabrication of reproducible and efficient hetero-

junction solar cells based on Cu(In,Ga)Se2 (CIGS).1

However, important efforts are devoted to replace this layer

with other materials2 because light of wavelength lower than

�517 nm cannot be transmitted to the absorbing layer. This

results in a quantum efficiency drop for the short wavelength

region due to the optical absorption loss in the CdS layer. In

addition, the possibility of having Cd-free buffer layers is

highly desirable from an environmental safety standpoint. In

comparison with CdS, among several alternative buffer

layers, zinc sulfide (ZnS) has become a successful candidate

at producing high-efficiency CIGS solar cells.3

The advantages of ZnS, in comparison to CdS, include its

band gap (3.56 eV)1,2 larger than that of CdS (2.4 eV),4

which enables the transmission of higher-energy photons to

the absorbing layer, a better lattice-matching to CIGS thin

film absorbers,5 its natural abundance, low cost, and, very

importantly, it is cadmium-free.

ZnS has been doped with different elements, including

lanthanides, which essentially exist in their most stable oxi-

dation state as trivalent ions exhibiting unique luminescent

properties such as photonic upconversion: the ability to con-

vert near infrared long-wavelength excitation radiation into

shorter visible wavelengths.6 Recently, ZnS has been consid-

ered an attractive host material for the implementation of

that process. Upconversion is one of the approaches pro-

posed to overcome the Shockley–Queisser efficiency limit in

the performance of photovoltaic cells.7,8 Several growth

techniques, such as chemical bath deposition, metal-organic

a)Author to whom correspondence should be addressed; electronic mail:

sergio.jimenez@cinvestav.mx

031505-1 J. Vac. Sci. Technol. A 35(3), May/Jun 2017 0734-2101/2017/35(3)/031505/8/$30.00 VC 2017 American Vacuum Society 031505-1

chemical vapor deposition, and atomic layer epitaxy have

been applied to grow high quality ZnS films for electrolumi-

nescent displays and solar cells.

The studies of Er and/or Yb doped ZnS are rather scarce

and in most cases, have involved the growth of ZnS nanopar-

ticles. Tong et al. reported the growth of Er-doped ZnS

nanostructures synthesized via a chemical solution route.9

Near infrared emission was studied in free standing ZnS

quantum dots in an aqueous solution doped with Er and

Yb.10 Similar studies were performed in ZnS nanoparticles

synthesized by homogeneous precipitation where several

lanthanides were incorporated, including Er and Yb.11

Previous studies in high-resistivity zincblende ZnS crystals

were aimed to analyze the recombination mechanisms under

Yb activation.12 The properties of lanthanide-doped ZnS

films have been poorly explored to date and remain as an

open field for the determination of their properties, espe-

cially for applications where the use of thin films is crucial.

In particular, radio-frequency (RF) sputtering, a cost-

effective deposition technique, yields proper control over

stoichiometry, uniformity, and it is currently a deposition

method widely used in industry.

For the present work, RF magnetron sputtering was used

to deposit ZnS thin films doped with rare earth elements (Er

and Yb) on glass substrates at two different temperatures,

RT and 180 �C. These temperatures were chosen because

low temperature processing is attractive for device applica-

tions in order to reduce diffusion processes, which in turn

may have a detrimental effect on devices performance.

Three doped ZnS targets were prepared for each lanthanide

with concentrations of 0.9, 1.9, and 2.9 at. % and one more

with a higher concentration of Er (3.88 at. %), in addition to

a ZnS target used to deposit undoped reference films. The

structural properties were characterized in detail by x-ray

diffraction (XRD), while the optical properties were studied

by UV-Vis, ellipsometry, and photoluminescence (PL) spec-

troscopies. The surface morphology was analyzed by field

emission scanning electron microscopy (FESEM).

II. EXPERIMENTAL DETAILS

ZnS films were deposited by RF magnetron sputtering on

Corning 2947 glass substrates. Prior to deposition, the sub-

strates were subjected to a chemical pretreatment, consisting

of (1) cleaning with detergent and rinsing with distilled

water, (2) immersion in a chromic acid solution (H2SO4,

K2Cr2O7, and H2O) for 24 h, (3) immersion in a H2O:NO3

(3:1) boiling solution for 3 h; and finally, (4) rinsing with

distilled water. To avoid contamination before deposition,

the substrates were stored in clean ethanol immediately after

the treatment described above. The target was prepared from

Sigma-Aldrich powdered ZnS. Each 2 in.-diameter sputter-

ing target was made by cold pressing 15 g of ZnS and Er or

Yb at three doping levels (0.5, 1.0, or 1.5 g). Additionally, a

target with a higher concentration of Er (2 g) was prepared.

For samples grown at room temperature, the depositions

were performed at a working pressure of 9 mTorr and an

effective forward power of 55 W for 90 min. The growth

time was extended to 120 min for samples grown at 180 �C.

XRD patterns were obtained in a Rigaku Ultima IV

diffractometer, using CuKa1;2 radiation (k1¼ 1.54059 A, k2

¼ 1.54442 A), parallel beam geometry in asymmetric config-

uration, fixed incident angle of 2�, and a scanning velocity of

1.5�/min. In order to carry out the phase identification pro-

cess and cell parameters calculations, Jade 9.0VR

was used.

FESEM images were acquired in a Helios Nano Lab 650 to

study the films surface morphology. The optical characteri-

zation was performed using spectroscopic ellipsometry.

These measurements were taken using a Jobin-Yvon UviselDH 10 phase modulated ellipsometer at an incident angle

of 70� in the spectral range of 1.5–5.0 eV. To avoid light

reflection from the backside of the substrate, it was mechani-

cally polished. The optical spectra were analyzed using an

air-roughness-film–substrate model (Fig. 1) in the FILM

WIZARDTM software (SCI, Inc.). The surface roughness was

represented by the Bruggeman effective medium approxima-

tion. For the films’ dielectric function, a generalized Lorentz

expression was used.13 Finally, PL measurements were car-

ried out using a Dilor microspectrometer model Labram IIequipped with an Arþ laser (488.0 nm), a thermoelectrically

cooled CCD detector, a diffraction grating (1800 g/mm), and

a confocal microscope. The sub-band gap laser energy was

used to probe electronic states within the band gap, as well

as the luminescent characteristics of the Er3þ and Yb3þ ions

in the ZnS host.

III. RESULTS AND DISCUSSION

A. Crystal structure analysis—X-ray diffraction

1. Diffraction patterns

Glancing incidence x-ray diffraction was used to identify

the different crystalline phases. Probable phases to be

formed in the films are ZnS polymorphs and, at a minor scale

(because of the low lanthanide concentrations) formations of

Er2S3, ErS2, ErS, Yb2S3, or stable alloys in the ZnEr or

ZnYb solid solution series. In our films, grown at two differ-

ent substrate temperatures and constant pressure, the coexis-

tence of the two well-known stable polymorphs of ZnS were

detected, i.e., sphalerite,14 cubic with space group F�43m,

and wurtzite,15 hexagonal described by the space group

FIG. 1. Layer model used to fit the optical transmission and ellipsometric

data (uniaxial model).

031505-2 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-2

J. Vac. Sci. Technol. A, Vol. 35, No. 3, May/Jun 2017

P63mc. No evidence was detected about the formation of the

above mentioned sulfides or of Zn-lanthanide alloys. The

hexagonal ZnS polymorph has been indexed in this work

using the unit-cell parameters of the 2H polytype. The dif-

fraction patterns presented in Fig. 2 indicate the existence of

preferential growth.

Figures 2(a) and 2(b) show the diffraction patterns of the

ZnS samples doped with erbium deposited at RT and 180 �Cusing a logarithmic scale in the vertical axis to enhance low

intensity peaks. The observed peaks were indexed using

the PDF cards ICDD 01-071-4763—sphalerite and ICDD

01-070-2552—wurtzite 2H, which are plotted at the top and

at the bottom of each figure, respectively, as reference.

According to these patterns, a mixture of cubic (C) and

hexagonal (H) phases is exhibited in all cases. Due to the

structural similarities between the sphalerite and wurtzite

phases, some diffraction peaks cannot be resolved because

the reflections occur at nearly identical angles. For instance,

the main diffraction peak (111)C of sphalerite is difficult to

distinguish from the wurtzite (002)H since their positions in

2h differ by only 0.12�. From Figs. 2(a) and 2(b), it may be

stated that the films are a mixture of cubic and hexagonal

phases that grew preferentially parallel to the [111]C/[002]H

direction. It is likely that the sample of undoped ZnS grown

at RT [Fig. 2(a)] is predominantly cubic as the main peak is

closer to the (111)C direction and no other reflections coin-

cide with any purely hexagonal peak. Similarly, it is plausi-

ble that the undoped film grown at 180 �C [Fig. 2(b)] is

predominantly hexagonal, as the most intense peak is closer

to the (002)H reflection and the (103)H peak appears as the

second most intense peak besides the (102)H reflection.

These two peaks do not have any close cubic counterpart.

The effect of dopants on the crystalline structure is better

defined than that of the substrate. From the diffraction

patterns in Fig. 2(a), it is possible to conclude that the incor-

poration of Er favored the growth of the wurtzite phase.

Indeed, the (103)H diffraction peak, which is distinctive of

the hexagonal phase, is accentuated as the dopant level

increases. For the samples grown at 180 �C [Fig. 2(b)], the

predominance of the hexagonal phase is evident. As men-

tioned above, the overlap of the cubic and hexagonal contri-

butions in the main peak, and at least in two other peaks, is a

limiting factor in discarding the existence of either cubic

or hexagonal phases. For the samples doped with Yb, the

diffraction patterns exhibit a similar behavior to that of the

films doped with Er. This is not surprising since the chemical

properties of lanthanides are alike, thus producing similar

effects when incorporated in the ZnS host. Moreover, the

difference in ionic radii of the trivalent cations of Er and Yb

is only �2% (0.890 and 0.868 A, respectively).16

2. Temperature evolution of crystal structure

Since one of the potential applications of ZnS is as win-

dow or buffer layer in photovoltaic structures, it is desirable

that no significant structural changes occur that alter detri-

mentally the material properties, in particular, with tempera-

ture. To investigate the structural stability of the films

diffraction patterns, it was measured at various temperatures

(25, 70, 110, 150, 190, 250, 300, and 350 �C), both in air and

vacuum. In situ x-ray diffraction measurements were carried

FIG. 2. (Color online) Diffraction patterns of ZnS samples doped with different quantities of erbium in the target at room temperature (a) and 180 �C (b). The

red squares at the top represent the crystallographic directions and relative intensities reported for sphalerite (cubic—ICDD 01-071-4763) and black hexagons

at the bottom for wurtzite 2H (hex.—ICDD 01-070-2552).

031505-3 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-3

JVST A - Vacuum, Surfaces, and Films

out using a temperature chamber (Rigaku HT-1500). The

heating rate was set to 10 �C/min with a stabilization time of

5 min, and a scanning velocity of 1.2�/min. According with

those parameters, each diffraction pattern was obtained in

approximately 35 min.

The results of these experiments were consistent with the

films being stable up to 350 �C regardless of the ambient.

That is, the diffraction patterns were basically unmodified by

the sample holder temperature inside the previously vac-

uumed chamber or in ambient air. This is desirable for prac-

tical applications since a prospective device will not be

affected by these variables.

B. Morphological analysis—Field emission scanningelectron microscopy

In order to study the surface morphology of the samples,

dependent on doping level and substrate temperature, micro-

graphs using field emission scanning electronic microscope

were obtained. Due to the small size of the aggregates

(25–30 nm), it was necessary to use high resolution equip-

ment. In order to have a reasonable estimate of the average

aggregate diameter, the specialized software (SPIPTM) was

used. Figures 3(a) and 3(b) show the micrographs of

undoped ZnS grown at RT, taken with different magnifica-

tions (32 500� and 75 000�, respectively). The average

aggregate diameter (calculated using SPIP) for this sample

was 29:768:4 nm. According to Figs. 3(a) and 3(b), the sur-

face presented a homogeneous morphology. The small light

spots in the images correspond to local charging effects that

occurred during the measurement process. Figure 3(c) shows

a micrograph of the sample with the highest doping level of

Er (3.88 at. %) grown at RT. The surface roughness is lower

due to its smaller aggregates size (24:469:9 nm) compared

with the undoped film. In contrast, in Fig. 3(d), the micro-

graph that corresponds to the same doping level but grown at

180 �C presents an evident modification in its surface mor-

phology with respect to undoped ZnS. The texture presented

FIG. 3. FESEM micrographs of (a) undoped ZnS grown at room temperature (32 500�). (b) Undoped ZnS fabricated at RT (75 000�). (c) ZnS doped with

erbium (3.88 at. %) grown at room temperature (32 500�). (d) ZnS doped with erbium (3.88 at. %) at 180 �C (32 500�). All images were acquired using a

voltage of 5 kV and a current of 0.34 nA.

031505-4 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-4

J. Vac. Sci. Technol. A, Vol. 35, No. 3, May/Jun 2017

may be described as stacked flakes. For the samples doped

with Yb, it was possible to obtain only two micrographs due

to difficulties encountered in focusing the surface with rea-

sonable clarity. From those micrographs, it is possible to

state that the surface morphology of Yb-doped films pre-

sented a similar behavior to the case of Er-doped samples.

In brief, as a result of the doping level for the samples

grown at RT and 180 �C, a slight reduction in the aggregates’

average diameter occurred. Moreover, the substrate tempera-

ture effect in the surface morphology consisted in a change

on the aggregates’ geometry [Fig. 3(d)] besides the expected

increase in aggregates’ average diameter for the higher sub-

strate temperature, in agreement with the well-known

Thornton’s diagram.17 In this case, however, the crystallites

and their formed aggregates present mostly random orienta-

tions instead of the typical columnar forms typically

observed in sputtered films.18

C. Optical analysis

1. UV-vis spectroscopy

The thickness of the films and the homogeneity in chemi-

cal composition could be assessed using the spectra of trans-

mittance and reflectance. To do so, four zones of equal area

(�2.34 cm2) from the center to one end (1, 2, 3, and 4,

respectively) were defined to carry out the measurements in

each of them. In Fig. 4(a), spectra for the undoped ZnS sam-

ple grown at RT are presented. According to these spectra,

the amplitude of the oscillations does not vary significantly

between zones, which evidences sample homogeneity in

chemical composition. In contrast, the period of the oscilla-

tions presents variations among zones. This behavior is due

to differences in film thickness, that is, a smooth variation

exists from the center (thickest) toward the edges, as it

normally occurs in sputtered films. The same tendency in

frequency and amplitude of the oscillations is replicated

throughout the whole set of samples which means that the

films are homogeneous in chemical composition and inho-

mogeneous in thickness. In order to calculate the thickness

of the zones, specialized software (FILM WIZARD) was used

using the layer model (Fig. 1) described above.

In Figs. 4(b)–4(d), the thicknesses calculated for each

zone of undoped and the most doped (Er—3.88 at. % and

Yb—2.81 at. %) ZnS samples fabricated at RT and 180 �Care presented. The number that is close to each point, corre-

sponds to the root mean square error (RMSE) associated with

the minimization process that yielded the best fit between the

experimental and calculated data. In general, the thickness of

the samples fabricated at 180 �C was lower than that of films

grown at RT, due mainly to atom re-evaporation occurring

FIG. 4. (Color online) Transmittance and reflectance spectra per zone of undoped ZnS film fabricated at room temperature. (a) Thicknesses obtained from fit-

ting the optical data for different center-to-edge spots on the films surface. (b) ZnS undoped, (c) ZnS doped with erbium (3.88 at. %) and ZnS doped with

yterbium (2.81 at. %) (d). Black squares correspond to samples fabricated at room temperature, and red circles correspond to films grown at 180 �C. The num-

ber close to each datum point is the RMSE after the error minimization process.

031505-5 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-5

JVST A - Vacuum, Surfaces, and Films

during growth as a consequence of the higher temperature. It

is interesting to note that the lowest RMSE corresponds to

the thinner regions, that is, the fitting process was more accu-

rate in these cases. This may be attributed to the fact that the

number of crystalline defects increases with film thickness,

while the model used to fit the spectra does not take into

account the effects of crystalline imperfections on the optical

properties. This is also the reason why the thicknesses

obtained from spectroscopic ellipsometry in Sec. III C 2 are

quoted for the thinner regions four in Table I.

2. Spectroscopic ellipsometry

In order to obtain optical constants ðn; kÞ of the films as

well as the optical band gap ðEgÞ, ellipsometry measure-

ments were carried out in the same zones measured previ-

ously by UV-vis spectroscopy. After the error minimization

process was performed in all the experimental data, zone 4

was selected to determine the thickness, surface roughness,

and optical constants since, similarly to the UV-Vis case,

this zone yielded the lowest RMSE in the spectra fitting pro-

cedure carried out with the FILM WIZARD software.

In Table I, the thickness, roughness, and RMSE are sum-

marized for all samples. According to these values, in gen-

eral, roughness increases with the substrate temperature, in

agreement with the FESEM findings, and the results pub-

lished by Hwang et al.19 When the substrate temperature is

higher, the grain size is enhanced yielding rougher surfaces.

So as to verify the optical constants obtained for the

undoped ZnS sample grown at room temperature, a compari-

son with the literature was done and a good correspondence

was found. Based on the values obtained for k, the absorp-

tion coefficient (aÞ for the set of samples was calculated

from the relation a ¼ 4pk=k. Since ZnS is a direct band gap

semiconductor, the equation for the absorption coefficient

near the optical absorption edge is [Eq. (1)]20

ah� ¼ Aðh� � EgÞ1=2: (1)

Thus, plotting ðah�Þ2 as a function of h�, fitting the lin-

ear portion of the resulting curve and finding the intercept

with the x-axis, the band gap could be determined.21 In

Figs. 5(a) and 5(b), the band gap calculated for the samples

of ZnS undoped fabricated at room temperature and 180 �Care shown. The obtained values, 3.55 eV (RT) and 3.71 eV

(180 �C), are in agreement with those reported for cubic

(sphalerite),22,23 and hexagonal (wurtzite 2 H)23,24 zinc

sulfide. These results are consistent with those found

through X-ray diffraction in the sense that the dominant

crystalline structure of the undoped ZnS film grown at

room temperature is sphalerite and for the film grown at

180 �C wurtzite.

In the case of samples doped with Er, the band gap values

are also shown in Figs. 5(a) and 5(b) for RT and 180 �Cfilms, respectively. In general, a reduction of ðEgÞ is pre-

sented as a function of the lanthanide concentration. This, in

turn, indicates that Er atoms were incorporated in the ZnS

host during the growth process, instead of forming metallic

segregations which would not have any effect on Eg. The

same behavior is presented in the samples doped with Yb.

TABLE I. Thickness, roughness, and RMSE for the set of samples. The

RMSE was obtained after error minimization in the experimental data fitting

procedure.

Sample Thickness (nm) Roughness (nm) RMSE

ZnS/M1a 172.4 5.94 1.2503

ZnS/M2 149.8 2.62 1.1876

ZnS:Er (0.97 at. %)/M3a 178.0 4.62 0.5442

ZnS:Er (0.97 at. %)/M4 130.2 2.19 0.5800

ZnS:Er (1.94 at. %)/M5a 189.1 2.06 0.6818

ZnS:Er (1.94 at. %)/M6 196.5 3.98 0.6908

ZnS:Er (2.91 at. %)/M7a 307.8 2.06 0.6818

ZnS:Er (2.91 at. %)/M8 416.5 5.25 1.1314

ZnS:Er (3.88 at. %)/M9a 186.1 2.10 1.6196

ZnS:Er (3.88 at. %)/M10 278.4 3.15 0.8670

ZnS:Yb (0.93 at. %)/M11a 129.0 2.66 0.6947

ZnS:Yb (0.93 at. %)/M12 121.6 3.10 0.5131

ZnS:Yb (1.87 at. %)/M13a 202.9 2.98 2.2511

ZnS:Yb (1.87 at. %)/M14 172.4 5.94 0.6910

ZnS:Yb (2.81 at. %)/M15a 183.6 3.20 0.9857

ZnS:Yb (2.81 at. %)/M16 172.4 5.94 0.6447

aSamples labeled in bold italics were grown at room temperature.

FIG. 5. (Color online) Energy band gap ðEgÞ estimated for ZnS doped sam-

ples with different erbium concentrations deposited at room temperature (a)

and 180 �C (b).

031505-6 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-6

J. Vac. Sci. Technol. A, Vol. 35, No. 3, May/Jun 2017

3. Photoluminescence

In order to investigate possible electronic states within

the bandgap, as well as the luminescent characteristics of the

Er3þ and Yb3þ ions in the ZnS matrix, PL measurements

were carried out at low temperature (� 100 K) and at room

temperature. The excitation source was an Arþ laser emitting

at 488.0 nm (2.54 eV). As reference, the PL signal from pure

lanthanide elements (99.9%) was previously measured. It is

well known that the emission properties of the lanthanide

ions are not modified substantially when they are incorpo-

rated in a host. Indeed, trivalent cations Er3þ or Yb3þ are

formed when the two 6s and one of the 4f electrons are used

to form bonds with the surrounding atoms. The PL emission

from the ions is a consequence of transitions between 4forbitals, that is, f-f transitions. Since the 4f orbital (partially

filled) is effectively shielded from the surrounding environ-

ment by the filled 5s25p6 subshells, the disturbance to the 4felectronic levels participating in the PL process is effectively

uncoupled from the host atoms. Such shielding is responsible

for the specific lanthanide luminescent properties, particu-

larly, narrow emission bands with long excited-state

lifetimes.

The room temperature PL spectrum of metallic Er exhib-

its four narrow well defined electronic transitions. The elec-

tronic configuration of Er is ½Xe�4f 126s2, that is, all the

sublevels are filled except the 4f subshell, which is occupied

by 12 electrons. The number of possibilities, or microstates,

to localize those electrons is described by the following

equation:25

G!

N! G� Nð Þ! ; (2)

where N is the number of electrons in a particular sublevel

and G the maximum number of electrons that can occupy the

sublevel. In the case of the partially filled sublevel of Er

(N¼ 12 and G¼ 14), there are 91 microstates to allocate the

electrons. Therefore, it is expected that there are several

components or lines for the four identified transitions in the

measured spectra of pure Er. For pure Yb according to its

electronic configuration (½Xe�4f 146s2Þ, no luminescence sig-

nal was obtained since all the sublevels are filled which

yields only one associated microstate. Thus, electronic tran-

sition is not possible, as experimentally observed.

In Fig. 6(a), the low temperature PL spectra of undoped

ZnS sample is presented. The PL signal of ZnS is a wide

band centered at �600 nm (2.07 eV), whose origin may be

associated with transitions between levels involving zinc

interstitials (Zni) and sulfur vacancies (VS) of the type

Zn2þi ! V2þ

S .26 The steplike feature between 600 and

625 nm is of instrumental origin. In the case of the Er doped

samples, no transition associated with Er3þ was observed;

only a broad band similar to that of pure ZnS presented in

Fig. 6(a) was detected. The fact that the Er3þ emissions coin-

cide with the broad PL band of ZnS shows that the emission

of the host is in this case stronger than that of the Er3þ ions.

Even the 4� I9=2 ! 4

� I15=2 emission is absent in the region

between 800 and 850 nm where the ZnS signal is low,

although such Er3þ emission is the weakest. These results

indicate that the energy transfer from the host atoms to the

Er3þ ions is nonexisting or it occurs with such low efficiency

that the Er3þ emissions do not surpass the host PL intensity.

More specifically, the photoexcited electrons from the host

atoms do not make transitions to the lanthanide excited states

(or very weakly), which would enhance the Er3þ emissions

upon returning to the ground state through photon emission.

In Fig. 6(b), the room temperature PL spectra of ZnS:Yb

samples grown at 180 �C are presented for various doping

levels. The only possible transition 2F5/2-2F7/2 reported for

the cation Yb3þ (Ref. 6) centered at 982 nm is present con-

firming that Yb atoms are incorporated in the ZnS matrix as

trivalent cations; otherwise, no transition would be observed.

As mentioned above, in metallic Yb, all the sublevels are

filled so that electronic transition is not possible. It is conve-

nient to mention that the weakness of the emission in Fig.

6(b) is affected by the detector quantum efficiency, which

for 980 nm is quite low, near to its lowest limit.

FIG. 6. (Color online) Low temperature (100 K) photoluminescence spectra

of ZnS film grown at room temperature (a). The discontinuities in the spec-

tra were due to instrumental artifacts. Room temperature photoluminescence

spectra of undoped ZnS and ZnS:Yb films grown at 180 �C for various dop-

ing concentrations (b).

031505-7 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-7

JVST A - Vacuum, Surfaces, and Films

IV. CONCLUSIONS

The use of low substrate temperatures (RT and 180 �C) to

deposit ZnS and lanthanide (Er and Yb) doped ZnS films by

sputtering on glass substrates yielded preferentially oriented

films along the lowest surface energy directions: (111)-cubic

and (002)-hexagonal. Whole pattern fittings of the undoped

and Er doped samples carried out using FullprofVR

indicated

the existence of four different phases in different propor-

tions: (1) cubic, (2) hexagonal, (3) cubic doped, and (4) hex-

agonal doped. The hexagonal wurtzite phase was promoted

by increasing the substrate temperature and by the dopants

incorporation. The average grain size for the films grown at

room temperature changed slightly upon doping, from

29:768:4 to 24:469:9 nm for the film with 3.88% at of Er.

The growth habits were modified in the case of the doped

films (Er or Yb) grown at 180 �C, presenting a flaky appear-

ance, as opposed to the spherelike grains observed in the

doped and undoped RT films. The films’ thickness and

roughness was obtained from a spectra fitting procedure of

the spectroscopic ellipsometry data. As expected, the rough-

ness values were higher for the films grown at 180 �C than

those of the RT films. Upon doping, the band gap was

slightly reduced, the effect being larger for the films grown

at 180 �C. The room temperature strong photoluminescence

band centered at �600 nm (2.07 eV) was ascribed to transi-

tions between levels involving zinc interstitials and sulfur

vacancies Zn2þi ! V2þ

S .

ACKNOWLEDGMENTS

The authors would like to thank F. Rodr�ıguez-Melgarejo,

E. Urbina-Alvarez, C. Z�u~niga-Romero, J. M�arquez-Mar�ın,

and O. Castillo-Vargas, for technical assistance and to E. G.

Santill�an-Rivero for bibliographical assistance. C. A. Ortiz

thanks CONACyT-Mexico for his M. Sc. Scholarship.

Partial financial support of Conacyt through Grant No.

257166 is acknowledged.

1K. Orgassa, U. Rau, Q. Nguyen, H. W. Schock, and J. H. Werner, Prog.

Photovoltaics: Res. Appl. 10, 457 (2002).2M. M. Islam, S. Ishizuka, A. Yamada, K. Sakurai, S. Niki, T. Sakurai, and

K. Akimoto, Sol. Energy Mater. Sol. Cells 93, 970 (2009).3T. Nakada and M. Mizutani, Jpn. J. Appl. Phys., Part 2 41, L165 (2002).4C. Kittel, Introduction to Solid State Physics (Wiley, New York, 2004).5L. X. Shao, K. H. Chang, and H. L. Hwang, Appl. Surf. Sci. 212, 305

(2003).6F. Wang et al., Nature 463, 1061 (2010).7W. Shockley and H. J. Queisser, J. Appl. Phys. 32, 510 (1961).8F. Wang and X. Liu, Chem. Soc. Rev. 38, 976 (2009).9Y. Tong et al., Mater. Chem. Phys. 151, 357 (2015).

10X. Zhang, X. Dong, L. Li, Z. Wang, Y. Liu, D. Dong, Y. Zhang, and G.

Kai, Proc. SPIE 6782, 67822J (2007).11X. Wei, W. Wang, and K. Chen, Dalton Trans. 42, 1752 (2013).12H. Przybylinska, K. Swiytek, A. Stupor, A. Suchocki, and M. Godlewski,

Phys. Rev. B 40, 1748 (1989).13E. Zawaideh, U.S. patent 5,999,267 (7 December 1999).14B. J. Skinner, Am. Mineral. 46, 1399 (1961).15C. Frondel and C. Palache, Am. Mineral. 35, 29 (1950).16R. D. Shannon, Acta Cryst. 32A, 751 (1976).17J. A. Thornton, J. Vac. Sci. Technol. 11, 666 (1974).18C. V. Thompson, J. Appl. Phys. 58, 763 (1985).19D. H. Hwang, J. H. Ahn, K. N. Hui, K. S. Hui, and Y. G. Son, Nanoscale

Res. Lett. 7, 26 (2012).20S. M. Sze and K. K. Ng, Physics of Semiconductor Devices (Wiley,

Hoboken, 2007).21H. K. Sadekar, N. G. Deshpande, Y. G. Gudage, A. Ghosh, S. D. Chavhan,

S. R. Gosavi, and R. Sharma, J. Alloys Compd. 453, 519 (2008).22R. A. Soref and H. W. Moos, J. Appl. Phys. 35, 2152 (1964).23D. Vogel, P. Kruger, and J. Pollmann, Phys. Rev. B 54, 5495 (1996).24A. Ebina, E. Fukunaga, and T. Takahashi, Phys. Rev. B 12, 687 (1975).25D. A. McQuarrie and J. D. Simon, Physical Chemistry: A Molecular

Approach (University Science Books, Sausalito, CA, 1997).26D. Kurvatov, V. Kosyak, A. Opanasyuk, and V. Melnik, Physica B 404,

5002 (2009).

031505-8 Ortiz et al.: Properties of sputtered ZnS and ZnS:A 031505-8

J. Vac. Sci. Technol. A, Vol. 35, No. 3, May/Jun 2017