Device behavior of an In/p-Ag(Ga,In)Te2/n-Si/Ag heterojunction diode

Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing

Materials Science in Semiconductor Processing 34 (2015) 138–145

http://d1369-80

n Corrnical Un

E-m

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

Device behavior of an In/p-Ag(Ga,In)Te2/n-Si/Agheterojunction diode

E. Coşkun a,b,c,n, H.H. Güllü a,c, İ. Candan a,c, Ö. Bayraklı a,c,M. Parlak a,c, Ç. Erçelebi a,c

a Department of Physics, Middle East Technical University, 06800 Ankara, Turkeyb Department of Physics, Çanakkale Onsekiz Mart University, 17100 Çanakkale, Turkeyc Center for Solar Energy Research and Applications (GÜNAM), METU, Ankara 06800, Turkey

a r t i c l e i n f o

Keywords:Thin filmsHeterojunctionsDepositionElectrical transportThermal analysis

x.doi.org/10.1016/j.mssp.2015.02.04301/& 2015 Elsevier Ltd. All rights reserved.

esponding author at: Department of Physiciversity, 06800 Ankara, Turkey.Tel.: þ90 28ail address: [email protected] (E. Coşkun

a b s t r a c t

In this work, p-(Ag–Ga–In–Te) polycrystalline thin films were deposited on soda-limeglass and n-type Si substrates by e-beam evaporation of AgGa0.5In0.5Te2 crystallinepowder and the thermal evaporation of Ag powder, sequentially in the same chamber.The carrier concentration and mobility of the Ag–Ga–In–Te (AGIT) film were determinedas 5.82�1015 cm�3 and 13.81 cm2/(V s) as a result of Hall Effect measurement. The opticalanalysis indicated that the band gap values of the samples were around 1.58 eV. Thestructural analysis was carried out by means of X-ray diffraction. Current–Voltage (I–V)measurements depending on the sample temperature were performed to investigate thedevice characteristics and the dominant conduction mechanism in an In/p-AGIT/n-Si/Agstructure. The series and shunt resistances were calculated by the help of parasiticresistance analysis as 5:73 and 1:57� 104 Ω cm2, respectively at room temperature. Theideality factors and barrier heights were evaluated as a function of sample temperature. Inthe low bias region, the thermionic emission together with the generation–recombinationmechanism was investigated as the dominant transport mechanism; however, in the highbias region, space charge limited current was analyzed as the other effective mechanismin the carrier conduction. The built-in potential of the device was also determined by thehelp of capacitance–voltage measurements.

& 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Thin film photovoltaic device technology is low costalternative to the traditional silicon-based solar cells, whichis offering large scale material usage. Moreover, there hasbeen a great deal of interest in the study of the chalcopyritesemiconductors in this technological point of view. Espe-cially, ternary and quaternary chalcopyrite compounds basedon the combinations of I, III and VI groups of elements have

s, Middle East Tech-6 2180018x1933.).

taken considerable attention because of their tunable opticaland electrical characteristics [1]. High absorption coefficientand suitable band gap values make these compoundspromising materials for absorber layer in solar cell applica-tions. Therefore, these chalcopyrite absorbers are quitefavorable for the production heterojunction and tandemsystems [2]. The polycrystalline I–III–VI2 chalcopyrite com-pound thin films, especially CuInSe2 (CIS) [3] and theirquaternary counterpart Cu(In,Ga)Se2 (CIGS) [4] are takingconsiderable attention for the next generation photovoltaicdevices [5]. Thin-film solar cells with CIGS module applica-tions reach the highest photovoltaic conversion efficiency[6]. However, Cu based compounds causes shorting effect inphotovoltaic devices due to its larger diffusion coefficient

Table 1EDXA results for crystal powder and AGIT films.

Ag (at%) Ga (at%) In (at%) Te (at%)

Powder 29 13 12 46AGIT 13 11 14 62

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145 139

and there are many works to overcome these obstacles forthin film structures [7]. Therefore, other chalcopyrite materi-als having very similar electrical and optical propertiesbecame popular [8]. Among the several materials of thisgroup, AgInTe2 (AIT) and AgGaTe2 (AGT) have proved to bestable and efficient absorber materials for polycrystallinethin film heterojunction solar cells [9–13].

In this work, we focused on the quaternary system Ag–Ga–In–Te (AGIT) since it allows tailoring of the opticalband gap and other properties. This system is locatedbetween direct band gap the ternary semiconductingchalcopyrite compounds AIT and AGT. Very little work onAGIT thin films has been reported. The physical propertiesof the chalcopyrite semiconductor AGIT has confirmed itspotential for various thin film applications [14–16], but asfar as it is known there is no such a work published on thedevice characteristics of this structure. Therefore, tounderstand and get the information about the devicebehavior and properties of p-AGIT thin films, the p-AGITwas deposited on n-Si wafer and soda lime glass sub-strates. Device characterization of In/p-AGIT/n-Si/Ag struc-tures was investigated. This study can be a guide for futurestudies about the photo-voltaic device applications of AGITchalcopyrite material.

2. Experimental details

AGIT thin films were deposited on soda-lime glass andmono-crystalline n-type Si (111) wafers having the resis-tivity 1–3 (Ω cm) substrates to produce a p-AGIT/n-Siheterojunction structure. Thin film deposition was carriedout in a system including e-beam and thermal evaporationfacilities in the same chamber. The stoichiometric highpurity Ag, Ga, In and Te elements were sintered in anevacuated quartz ampoule in a vertical furnace at 1050 1Cfor 2 days. Then, the pre-sintered quartz ampoule wasplaced to a Crystalox MSD-4000 model three zone verticalBridgman–Stockbarger system and a special temperatureprofile was used to obtain the single crystal AgGa0.5In0.5Te2compound. The temperatures of three zones of furnacewere adjusted to the values of 1100, 800, and 600 1C,respectively and following to a 72 h of translation from topto bottom zone with a translation speed of 1.0 mm/h.

AgGa0.5In0.5Te2 crystal powder was crunched in the finegrains before used as the evaporation powder. The surface ofthe wafers was subjected to a cleaning procedure with HF:H2O¼1:10 solution in order to remove the native oxide layerand then rinsed in deionized water consecutively blown dryin N2. During the deposition process, successive layer bylayer deposition method were used to get a stoichiometricand homogenous film structure by using AgGa0.5In0.5Te2and high purity Ag powders as the evaporation sources ofe-beam and thermal evaporations, respectively. The sub-strate temperature during deposition was kept at aroundTS¼200 1C and the thickness of the samples controlled andmonitored in-situ by Inficon XTM/2 deposition monitor.Before starting the deposition process, the back surface ofthe n-Si wafer was coated with Ag by thermal evaporation,as a back ohmic contact, and annealed at 450 1C underthe nitrogen atmosphere to enhance the ohmicity of thecontacts. The fabrication of the p-AGIT/n-Si heterojunction

structure was completed after the deposition of the trans-parent indium front contact by the thermal evaporationusing the dot-patterned copper masks. Again, following toIn contact deposition, the samples were annealed at 100 1Cunder the nitrogen atmosphere to improve the contactbehavior. The thickness of the films was measured electro-mechanically following to the deposition by Vecoo Dektak6 M thickness profilometer and it was around 385 nm. Thecompositional analysis of the samples was carried out bymeans of a JSM-6400 Scanning Electron Microscope (SEM),equipped with NORAN System 6 X-ray Microanalysis Systemand Semafore Digitizer detector that operated at 25 kV. Theoptical transmittance of the AGIT films deposited on soda-lime glass substrate was measured by PerkinElmer Lambda45 model UV/vis spectrometer at room temperature todetermine the band gap of the AGIT layer. The devicecharacteristics of the fabricated In/p-AGIT/n-Si/Ag hetero-junction structures were investigated by carrying out thetemperature-dependent dark current–voltage (I–V), capaci-tance–voltage (C–V) measurements. These measurementswere performed with the computer-controlled measurementsetup and a Keithley 2401 sourcemeter, Hewlett Packard4192 A LF model impedance analyzer, the Model 22 CTICryogenics closed-cycle helium refrigeration system and thetemperatures of the sample and the substrate heater wasmonitored by LakeShore DRC-91C temperature controller.

3. Results and discussion

3.1. Structural, electrical, optical and photocurrent analysis

The atomic ratios of the deposited AGIT films are shownin Table 1. The films are Ag-deficient and Te-rich. This couldbe because of different vapor pressures and the segregationcoefficients of the constituent elements.

The surface roughness morphology of the AGIT films wasinvestigated by AFM analysis. As seen from Fig. 1 for typicalsample, the measured surface roughness is almost uniformwith the value of 16.4 nm. The obtained X-ray diffractionpatterns for the films show that they have polycrystallinecharacteristics corresponding to a tetragonal AGIT structurewith a strong preferred orientation along with (112) planedirection at 2θ �24.41 (see Fig. 2).

However, the film has binary Ga2Te5 crystal phase 2θ�27.51 [17]. All of the as-grown films at substrate temperatureof TS¼200 1C are in polycrystalline nature with Ga2Te5 phase.Although some elements according to EDXA are present inexcess as compared to their ideal stoichiometry, no metallicelements were detected. The secondary phases in the qua-ternary structure are originated from the partial reaction ofthe constituent excessive elements of In, Ga and Ag with Teduring the evaporation of the crystalline source. Furthermore,the detailed structural and compositional analyses of AGIT

Fig. 1. AFM image (1 mm�1 mm) of the AGIT film for a typical sample.

Fig. 2. X-ray diffraction pattern of the AGIT film.Fig. 3. The plot of T versus λ (black line) and R versus λ (red line) for atypical AGIT film. The inset shows (αhυ)2 versus (hυ) plot. (For inter-pretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145140

thin films deposited by thermal evaporation were carried andreported in [18].

The room temperature electrical resistivity of the filmwas measured by Van der Pauw method as 95.65 Ω cm,and the carrier conduction type was determined by hotprobe method as p-type. The carrier concentration andmobility of the film were determined as 5.82�1015 cm�3

and 13.81 cm2/(V s) by using Hall Effect measurement.The optical behaviors of Ag–Ga–In–Te films were

investigated by using the spectral transmission and reflec-tion measurements. The calculated absorption coefficientwas around 105 cm�1 at 703 nm for all samples. Ag–Ga–In–Te films naturally have direct interband transition and aTauc plot was used to calculate the band gap values of thefilms by using the expression [19]

αhνð Þ ¼ C hν�Eg� �1=2 ð1Þ

where C is a wavelength independent constant, h is thePlanck constant, α is absorption coefficient, ν is frequencyand Eg is the optical band gap. As understood from the inset

of Fig. 3, the valance band splits three bands due to the p-likevalance band states [1,20]. This threefold degeneracy is liftedby the combined influence of spin–orbit interactions andnoncubic crystal field [1]. The band gap values were obtainedas Eg1¼1.31, Eg2¼1.49 and Eg3¼1.59 eV. This Eg3 value can beaccepted as the main band to band transition because ofhaving longest linearity for the absorption edge.

This splitting can also be observable from the photo-current measurement result of p-AGIT/n-Si device (seeFig. 4). The peak separation process was made by PeakFitprogram and the calculated correlation coefficient is equalto R2¼0.996. By using the separated peaks, the band gapvalues were calculated by using [21]

Iph2p hν�Eg

� � ð2Þ

relation as Eg1¼1.37, Eg2¼1.41 and Eg3¼1.58 eV. EgSi¼1.19 eV belongs to n-Si wafer for which the band gap valuewas reported as 1.12 eV [22].

Fig. 6. Plots of rectification factor (RF) versus V at different sampletemperatures.

Table 2The device parameters of the p-AGIT/n-Si heterojunction diode calculatedfrom the temperature dependent I–V analysis.

T (K) RS (Ω cm2) RSh (Ω cm2) n I0 (A) qϕb (eV) R2

220 27:8 3:21� 106 1.43 1:40� 10�10 0.65 0.999

240 16:6 7:79� 105 1.28 5:76� 10�10 0.68 0.999

260 10:1 1:92� 105 1.15 2:55� 10�9 0.71 0.999

280 6:46 5:21� 104 1.08 1:13� 10�8 0.73 0.999

300 5:73 1:57� 104 1.04 4:57� 10�8 0.75 0.999

320 4:46 4:46� 103 1.03 1:82� 10�7 0.76 0.998

340 3:40 1:65� 103 1.01 6:44� 10�7 0.78 0.997

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145 141

The main band to band transition energies are almostthe same for both optical transmittance and photocurrentanalysis, but the splitting band energies have some devia-tions that could be related with possible curve fittingerrors for middle peak in the photocurrent analysis. Ascompared with the reported value Eg¼1.37 eV of an AGITfilm [14], the calculated main band gap value for this filmis higher than the reported value. The difference betweenthem could be the result of the different atomic concen-trations of the constituent elements of the studied films.

3.2. Device characterization

In order to determine the device parameters of an In/p-AGIT/n-Si/Ag sandwich structure, I–V characteristics of thesample in the temperature range of 220–340 K weremeasured and are plotted in Fig. 5.

It shows that the rectifying behavior of the depositedheterojunction structure indicates a typical p–n junctiondiode characteristic [23]. To check the ohmicity of Ag/n-Sicontacts, I–V measurements on these contacts were doneand plotted in the inset of Fig. 5. The rectification factors(RF), which is the ratio of forward and reverse currents(IF/IR), were calculated about 102 at 0.2 V for all sampletemperatures. The voltage dependence of rectificationfactor in the studied temperatures is shown in Fig. 6.

It indicates that, the RF value increases with increasingvoltage at constant temperature. On the other hand, whenthe voltage is constant, the RF value decreases with increas-ing sample temperature, because of the effect of interfacestates and non-homogeneous trap distribution in the bulk of

Fig. 4. Spectral photocurrent plot of p-AGIT/n-Si heterostructure. The redline shows the total calculated photocurrent. (For interpretation of thereferences to color in this figure legend, the reader is referred to the webversion of this article.)

Fig. 5. I–V plots of p-AGIT/n-Si heterojunction diode at different sampletemperatures. The inset shows I versus V plot for n-Si wafer with Agcontact.

heterostructure [24,25]. In Fig. 5, at high forward and reversebias regions (greater than 70.3 V) I–V dependence deviatesfrom linearity because of the series and shunt resistanceeffects [23,26]. The low series resistance (RS) results in highpower and high speed of devices [27], but the high shuntresistance (RSh) is the result of leakage current because of thecontacts and the surface inhomogeneities [28]. In this study,RS and RSh values were calculated from the parasitic resis-tance RP ¼ ∂V=∂I

� �in the voltage range of 0.3–0.7 V for

forward and reverse bias, respectively, and they are listedin Table 2. It shows that their values decrease with increasingtemperature. Increase in temperature can cause either bondbreaking or de-trapping mechanism so these events mayincrease the density of the free charge carriers and decreasethe RS and RSH values [29,30].

In the forward bias region in between 0–0.2 V (seeFig. 5), I–V behavior exhibits two different regions for thevoltage ranges of 0–0.09 V and 0.1–0.2 V. The semiloga-rithmic plot of forward bias current increases with increas-ing bias voltage as seen in Fig. 5. Therefore, the forward I–Vdata was analyzed by using the equation:

I¼ I0 expqVnkT

� ��1

� �ð3Þ

where I0 is the reverse saturation current, V is the biasvoltage, n is the ideality factor and k is the Boltzmanconstant [31]. The reverse saturation current is given bythe following relation:

I0 ¼ AAnT2exp �qφb

kT

� ð4Þ

In this equation, ϕb is the potential barrier, A is the devicearea (Affi15� 10�3 cm2), Richardson constant is given as

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145142

An ¼ 4πqmnk2=h3 ¼ 120 ðmn=mÞ ðA=ðcm2 K2ÞÞ [31]. The diodeideality factor values of the device at studied temperaturewere calculated from the slope of semilogarithmic plot ofcurrent with voltage (see Fig. 7) by using the relation

1n¼ kT

qd lnðIÞð ÞdV

ð5Þ

By the way, the saturation current and the barrier heightvalues were also determined by using Eqs. (3) and (4),respectively. The calculated device parameters are listed inTable 2. It shows that ϕb values increase with increasingtemperature because more carriers have enough energy tosurmount the barriers and contributed to the conductionwith increasing temperature that is the reason of increasingpotential barrier height with the temperature [32]. Theideality factor n can be used to obtain information aboutcurrent transport mechanism of the junction. Calculated nvalues are higher than unity as seen from Table 2 and Fig. 8.This implies that there can be other transport mechanismsdifferent from the pure thermionic emission mechanism inwhich n¼1 [22].

If interface recombination [33] was the dominant trans-port mechanism, the reverse saturation current was equal tothe equation I0 ¼ I00exp �Ea=ðnkTÞ

� �[34]. In this equation,

I00 shall be a reference current that includes all terms of thereverse saturation current, I0, with negligible temperaturedependence, Ea is activation energy and n should be unity. Eais approximately equal to (qVbiþ∂p), where Vbi is built-inpotential and ∂p is the energy difference between the Fermilevel and the top of the valance band [35]; however thecalculated activation energy from Fig. 9 is 0.46 eV, which issmaller than the built-in potential of AGIT thin films (seeTable 5). Therefore, interface recombination could not be thedominant conduction mechanism. To determine the possible

Fig. 7. Plots of ln(I) versus V in the range of 0.02–0.1 V at different sampletemperature.

Fig. 8. The variation of ideality factor (n) as a function of temperatures.

conduction mechanism based on recombination–generationin depletion region, the ln(I0/T5/2) versus 1/T was plottedin Fig. 10.

The activation energy obtained from ln(I0/T5/2) versus 1/Tgraph should be approximately half of the band gap [35], butthe calculated activation energy from the Fig. 10 is equal to0.40 (eV), which is smaller than the half of band gap.Therefore, these values imply that recombination–generationmechanism could not also be the dominant conductionmechanism in the structure. On the other hand, the linearityof the ln(I0/T2) versus 1/T plot was checked for the validity ofthermionic emission as the dominant transport mechanism[36]. As a result of analysis using Fig. 11, the calculated slopevalues around one indicated that the thermionic emissioncould be the dominant conduction mechanism.

Consequently, it was deduced that in the voltage rangeof 0.01–0.09 V, the thermionic emission is the dominanttransport mechanism and the recombination–generationin depletion region contributes to thermionic emission forthe studied heterojunction structure. There are similarreported results in the literature for p-ZnTe/n-CdTe andn-AgIn5Se8/p-Si heterojunction diodes [37,38]. In order todetermine the conduction mechanism in the voltage rangeof 0.1–0.2 V, log I versus log V graph for different tempera-tures is plotted in Fig. 12.

Fig. 9. Plot of ln(I0) versus 1000/T for the p-AGIT/n-Si structure.

Fig. 10. Plot of ln(I0/T5/2) versus 1000/T for the p-AGIT/n-Si structure.

Fig. 11. Plot of ln(I0/T2) versus 1000/T for the p-AGIT/n-Si structure.

Fig. 12. Plots of Log(I) versus Log(V) at different sample temperatures.

Table 3The SCLC analysis results of the p-AGIT/n-Si heterojunction diode.

T (K) m l TT (K) R2

220 5.07 4.07 894.72 0.996240 4.80 3.80 912.89 0.995260 4.48 3.48 904.06 0.994280 4.15 3.15 881.73 0.994300 3.78 2.78 833.04 0.995320 3.32 2.32 743.62 0.996340 2.80 1.80 612.19 0.996

Fig. 13. Plots of ln(I) versus 1000/T for the p-AGIT/n-Si devices atdifferent bias voltage.

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145 143

The linearity of the slopes obtained from this plotimplies that the transport mechanism in this voltageregion could be related with space charge limited current(SCLC) depending on the IpVm relation. The current–voltage–temperature dependence of SCLC processes insemiconductors is largely defined by the distribution oflocalized trap levels within the band gap region [39]. Byusing the Fig. 12, the power exponent (m) values of voltagewere calculated as higher than 2 (see Table 3). Thisindicates the existence of the space charge limited cur-rents with the exponential trap distribution [36,40]. Thecurrent–voltage dependence for SCLC analysis relatedwith exponential trap distribution can be expressed as[39,41,42]

I¼ qAμNC

d2lþ1

ε ε0qN0kTT

� �l

V lþ1 ð6Þ

where ε is the dielectric constant, ε0 is the permittivity offree space, μ is the mobility of charge carriers, NC is theeffective density of states at the electronic band edge, d isthe thickness of the sample, N0 is the trap density per unitenergy range at the conduction band, l is a parametergiven by l¼m�1¼TT/T, TT is a critical temperature thatinfluences the exponential trap distributions and this trapdistributions per unit energy range at an energy E belowconduction band is given by

NðEÞ ¼N0exp � EkTT

� �ð7Þ

The total trap concentration NT is expressed by sum-ming the entire distribution of Eq. (7) as [43, 44]

NT ¼N0 k TT ð8ÞIn order to determine both NT and N0, ln(I) versus 1/Tat different bias voltages is plotted in Fig. 13. It is clearfrom the figure that there is an exponential variation with

temperature depending on the applied voltage as expectedfor the exponential distribution of traps [39]. The slope ofeach plot was analyzed by using the equation [43]

d lnðIÞð Þd 1=T� �¼ TT ln

ε ε0V

qd2NT

!ð9Þ

The obtained NT and N0 values are tabulated in Table 4.They increase with increasing applied voltage that couldbe related with the effect of thermally generated carriersto the conduction. Eqs. (7) and (8) reveal that over 63% oftraps situate within kTT of the conduction band edge andover 99% within 5kTT. The mean kTT value was calculatedas 0.071 eV; according to the trap distribution and bandgap value, the entire distribution of traps situate withinthe uppermost 26% of the band gap. TT values showdecreasing with increasing temperature, which means thatthe entire distribution of traps lies in the narrow energyrange within the band gap.Fig. 13. The room temperaturefrequency dependent C–V measurements were carried on(see Fig. 14) to investigate electrical characteristics of thejunction region.

Because of very low resistivity value of Si wafer withrespect to AGIT film, it was assumed that the fabricatedjunction behave as one sided abrupt junction, so C–Vcharacteristic can be described as [32]

C�2 ¼ 2Vbi�V�kT

q

� qA2ε ε0NA

24

35 ð10Þ

where Vbi is built in potential and NA is acceptor density inp-AGIT. C�2 versus V graphs for all frequencies are plottedin the inset of Fig. 14, and by using the slopes andintercepts of the curves in the voltage interval where theyare in linear behavior, qVbi and NA values were calculatedand tabulated in Table 5. The linearity of C�2–V in thevoltage range of 0.4–0.7 V indicates that NA is uniform inthe depletion region [23]. The calculated NA values can beassumed in good agreement with the values determinedfrom the Hall Effect measurements.

4. Conclusions

In this study, characterizations of AGIT film and thedevice behavior of the In/p-AGIT/n-Si/Ag sandwich structurewas investigated. The carrier concentration and mobilityof the AGIT film were determined as 5.82�1015 cm�3 and

Table 4The NT and N0 values obtained by using the SCLC analysis for different voltage values at the high bias region in the studied temperature range.

T (K) NT (�1021 m�3) N0 (�1041 m�3) V¼0.10 V

V¼0.10 V V¼0.14 V V¼0.16 V V¼0.18 V V¼0.20 V

220 2.34 2.67 3.09 3.96 4.37 1.89240 2.33 2.65 3.08 3.94 4.35 1.85260 2.33 2.66 3.09 3.95 4.36 1.87280 2.34 2.68 3.10 3.97 4.39 1.93300 2.37 2.71 3.14 4.02 4.45 2.06320 2.41 2.77 3.22 4.12 4.56 2.35340 2.47 2.87 3.33 4.27 4.72 2.93

Fig. 14. Plots of C versus V for the p-AGIT/n-Si devices of differentfrequencies. The inset shows C�2 versus V for the p-AGIT/n-Si device ofdifferent frequencies.

Table 5C–V analysis results of the p-AGIT/n-Si device.

f (kHz) NA (�1015cm�3) qVbi (eV) R2

1� 100 2.74 0.55 0.997

5� 100 2.70 0.54 0.998

1� 101 2.68 0.53 0.999

5� 101 2.69 0.54 0.998

1� 102 2.71 0.55 0.998

5� 102 2.69 0.53 0.997

1� 103 2.71 0.54 0.998

5� 103 2.70 0.54 0.999

1� 104 2.68 0.53 0.998

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145144

13.81 cm2/(V s) The band gap of the p-AGIT layer wasdetermined as Eg¼1.58 eV which is desired energy valuefor an absorber layer in solar cell applications. The diodestructure showed quite good device behavior having rectifi-cation factor was about 6.0�102 at 0.2 V. The series andshunt resistances were found to be on the order of 100–1 and103–6 Ω cm2, respectively. The small RS values mean the highpower and high speed of diode, whereas the large RSh valueimplies the decrease of the lost of current. I–V characteristicsof the device structure showed two different voltage regionsimplying the existence of two different carrier transportmechanisms. The analysis indicated that between 0 and0.1 V, thermionic emission mechanism could be dominantconduction mechanism together with the recombination–generation in depletion region, on the other hand in thevoltage range of 0.1–0.2 V, the space charge limited currentmechanism seemed to be another carrier transport mec-hanism. In the SCLC model, the trap distribution has an

important role on carrier conduction. As result of the SCLCanalysis, it is possible to say that in this device structure, theexponential trap distribution is more favorable than a singlediscrete trapping level model. The qVbi values of the fabri-cated devices were determined as about 0.54 eV from C–Vcharacterization.

Acknowledgments

This work was financed by Middle East Technical Uni-versity (METU-BAP) under Grant no. BAP-01-05-2013-005.Also, one of the authors would like to thank to TUBITAK-BIDEB for the financial supports during this study.

References

[1] J.L. Shay, J.H. Werrnick, Ternary chalcopyrite semiconductors:growth, in: Electronic Properties and Applications, Pergamon,Oxford, 1975.

[2] M.C. Lux-Steiner, A. Ennaoui, C.H. Fischer, A. Jager-Waldau, J. Klaer,R. Klenk, R. Konenkamp, T. Matthes, R. Scheer, S. Siebentritt,A. Weidinger, Thin Solid Films 361 (2000) 533.

[3] A. Rockett, R.W. Birkmire, J. Appl. Phys. 70 (1991) R81.[4] U. Rau, H.W. Schock, Appl. Phys. A-Mater. Sci. Process. 69 (1999) 131.[5] M.A. Contreras, K. Ramanathan, J. Abushama, F. Hasoon, D.L. Young,

B. Egaas, R. Noufi, Prog. Photovolt.: Res. Appl. 13 (2005) 209.[6] P. Reinhard, A. Chirila, P. Blosch, F. Pianezzi, S. Nishiwaki,

S. Buecheler, A.N. Tiwari, IEEE J. Photovolt. 3 (2013) 572.[7] M. Igalson, P. Zabierowski, D. Przado, A. Urbaniak, M. Edoff,

W.N. Shafarman, Sol. Energy Mater. Sol. Cells 93 (2009) 1290.[8] J. Krustok, A. Jagomagi, M. Grossberg, J. Raudoja, M. Danilson, Sol.

Energy Mater. Sol. Cells 90, 1973 (2006).[9] A. El-Korashy, M.A. Abdel-Rahim, H. El-Zahed, Thin Solid Films 338

(1999) 207.[10] H.G. Chandra, O.M. Hussain, S. Uthanna, B.S. Naidu, Phys. Scripta 63

(2001) 422.[11] S. Arai, S. Ozaki, S. Adachi, Appl. Opt. 49 (2010) 829.[12] C. Julien, I. Ivanov, A. Khelfa, F. Alapini, M. Guittard, J. Mater. Sci. 31

(1996) 3315.[13] A. Yusufu, K. Kurosaki, A. Kosuga, T. Sugahara, Y. Ohishi, H. Muta,

S. Yamanaka, Appl. Phys. Lett. (2011)061902-1.[14] H. Karaagac, M. Parlak, J. Alloy. Compd. 503 (2010) 468.[15] E.A. El-Sayad, Status Solidi A 201 (2004) 103.[16] G.B. Sakr, E.A. El-Sayad, Egypt. J. Solids 32 (2009) 9.[17] Joint Committee on Powder Diffraction Standards (JCPDS) Card no.

65-2452.[18] E. Coskun, H.H. Gullu, M. Parlak, J. Low Temp. Phys. 178 (2015) 162.[19] J. Tauc, Mater. Res. Bull. 3 (1968) 37.[20] V.V. Kindyak, A.S. Kindyak, V.F. Gremenok, I.A. Victorov, Thin Solid

Films 293 (1997) 75.[21] X. Mathew, J. Phys. D: Appl. Phys. 33 (2000) 1565.[22] P. Aurang, O. Demircioglu, F. Es, R. Turan, H.E. Unalan, J. Am. Ceram.

Soc. 96 (4) (2013) 1254.[23] S.M. Sze, K.N. Kwok, Physics of Semiconductor Devices, third ed.

John Wiley & Sons Inc., United States of America, 2007.

E. Coşkun et al. / Materials Science in Semiconductor Processing 34 (2015) 138–145 145

[24] K. Yilmaz, M. Parlak, C. Ercelebi, Semicond. Sci. Technol. 22 (2007) 1268.[25] J. MartinezPastor, A. Segura, J.L. Valdez, A. Chevy, J. Appl. Phys. 62

(1987) 1477.[26] A.A.M. Farag, M. Cavas, F. Yakuphanoglu, Mater. Chem. Phys. 132

(2012) 550.[27] A. Bengi, S.J. Jang, C.I. Yeo, T. Mammadov, S. Ozcelik, Y.T. Lee, Solid

State Electron. 61 (2011) 29.[28] I.S. Yahia, M. Fadel, G.B. Sakr, F. Yakuphanoglu, S.S. Shenouda,

W.A. Farooq, J. Alloy. Compd. 509 (2011) 4414.[29] K.S. Karimov, M.M. Ahmed, S.A. Moiz, M.I. Fedorov, Sol. Energy

Mater. Sol. Cells 87 (2005) 61–75.[30] H. Uslu, A. Bengi, S.S. Cetin, U. Aydemir, S. Altindal, S.T. Aghaliyeva,

S. Ozcelik, J. Alloy. Compd. 507 (2010) 190.[31] D.K. Schroder, Semiconductor Material and Device Characterization,

second ed. John Wiley & Sons Inc., New York, 1998.[32] V. Janardhanam, H. Lee, K. Shim, H. Hong, S. Lee, K. Ahn, C. Choi,

J. Alloy. Compd. 504 (2010) 146.

[33] A. Rothwarf, IEEE Trans. Electron Devices ED-29 (10) (1982) 1513.[34] R. Scheer, J. Appl. Phys. (2009)104505-1.[35] J. Yoo, A.L. Fahrenbruch, R.H. Bube, J. Appl. Phys. 68 (1990) 4694.[36] M.A.M. Seyam, H.T. El-Shair, G.F. Salem, Eur. Phys. J.: Appl. Phys. 41

(2008) 221.[37] G.K. Rao, K.V. Bangera, G.K. Shivakumar, Solid State Electron. 56

(2011) 100.[38] M. Kaleli, M. Parlak, Ç. Erçelebi, Semicond. Sci. Technol. 26 (2011)

105013.[39] R.D. Gould, J. Appl. Phys. 53 (1982) 3353.[40] F. Yakuphanoglu, N. Tugluoglu, S. Karadeniz, Physica B 392 (2007) 188.[41] M.A. Lampert, Rep. Prog. Phys. 27 (1964) 329.[42] A.A.M. Farag, A. Ashery, E.M.A. Ahmed, M.A. Salem, J. Alloy. Compd.

495 (2010) 116.[43] A.O. Odour, R.D. Gould, Thin Solid Films 270 (1995) 387.[44] R.D. Gould, M.S. Rahman, J. Phys. D: Appl. Phys 14 (1981) 79.