**5. Determination of LS parameters of polycrystalline chalcogenide films by optical spectroscopy (low-temperature photoluminescence)**

Low-temperature photoluminescence (PL) is one of the most reliable tools applied for investigation of longitudinal, native, impurity and point defect ensembles in semiconductors. High resolution of the method makes it possible to examine not only bulk materials (bulk chalcogenide semiconductor are now good studied [75-87, 90-93, 96, 99-100, 102-104]), but also thin films, in particular, chalcogenide semiconductor thin layers. In this part we present data obtained by studying low-temperature PL spectra of ZnTe, CdTe and ZnS films. These results allowed monitoring and adding new results to those given by the IS method.

#### **5.1. СdTe films**

Fig. 5 (a, b) illustrates the typical spectra of these films. As shown, the spectra for both types of the films are significantly similar. A modest energetical displacement of lines in spectra from epitaxial films comparing to those from the polycrystalline layers films deposited on glass may be caused by presence of sufficient macrodeformations in the layers CdTe/BaF2. PL spectra from CdTe layers have lines originated from optical transfers with participation of free and bound excitons, transfers valence band – acceptor (е-А), donor-acceptor transfers (DAP), the radiation caused by presence of dislocations or DP (donor pairs, DP) (Y - stripes); the spectra also have a set of lines corresponding to optical transitions where phonons take place (LO - phonon replica) [87-99].

**Figure 5.** PL spectra registered at *T*=4,5 К for polycrystalline films CdTe/glass (а), prepared at *Te* = 893 К and various *Ts*, К: 473 К (1); 523 К (2); 623 К (3); 823 К (4) and for the epiraxial layer CdTe/BaF2 (b): *Te* = 893 К, *Ts* = 798 К

Activation energies relative to the valence band (while the most samples were of p-type conductivity) were calculated using expression (26) (in analogy with description above).

520 Advanced Aspects of Spectroscopy

bound with single- *Zni*

method.

**5.1. СdTe films** 

*Te* = 893 К, *Ts* = 798 К

place (LO - phonon replica) [87-99].

All the LS found here were not identified because of absence of corresponding reference data. Only the levels with activation energy *Е1* = 0.15 eV and *Е2* = (0.22÷0.25) eV may be

**5. Determination of LS parameters of polycrystalline chalcogenide films** 

Low-temperature photoluminescence (PL) is one of the most reliable tools applied for investigation of longitudinal, native, impurity and point defect ensembles in semiconductors. High resolution of the method makes it possible to examine not only bulk materials (bulk chalcogenide semiconductor are now good studied [75-87, 90-93, 96, 99-100, 102-104]), but also thin films, in particular, chalcogenide semiconductor thin layers. In this part we present data obtained by studying low-temperature PL spectra of ZnTe, CdTe and ZnS films. These results allowed monitoring and adding new results to those given by the IS

Fig. 5 (a, b) illustrates the typical spectra of these films. As shown, the spectra for both types of the films are significantly similar. A modest energetical displacement of lines in spectra from epitaxial films comparing to those from the polycrystalline layers films deposited on glass may be caused by presence of sufficient macrodeformations in the layers CdTe/BaF2. PL spectra from CdTe layers have lines originated from optical transfers with participation of free and bound excitons, transfers valence band – acceptor (е-А), donor-acceptor transfers (DAP), the radiation caused by presence of dislocations or DP (donor pairs, DP) (Y - stripes); the spectra also have a set of lines corresponding to optical transitions where phonons take

**Figure 5.** PL spectra registered at *T*=4,5 К for polycrystalline films CdTe/glass (а), prepared at *Te* = 893 К and various *Ts*, К: 473 К (1); 523 К (2); 623 К (3); 823 К (4) and for the epiraxial layer CdTe/BaF2 (b):

interstitial Zn atom.

and double charged 2 *Zni*

**by optical spectroscopy (low-temperature photoluminescence)** 




The lines due to exciton recombination in CdTe single crystals are well-known. Authors [13] show energy level diagram for the exciton localized on neutral donors or acceptors and possible transfers between these levels. Commonly the elements of 3rd Group (Ga, In, Al) and 7th Group (Cl, Br, I) are shallow donors in CdTe and ZnTe, and acceptors are the elements of 1st Group and 5th Group (Li, Na, Cu, Ag, Au, N, P, As). These elements are typical excessive impurities in compounds II-VI. Authors [13] also give the ionization energies of principal dopant impurities in CdTe: for the donors (13.67÷14.48) meV, for

acceptors 56 (N) - 263 (Au) meV. We have used these values for further interpretation of the experimental results.

Unlike ZnTe condensates the peak bound with a free exciton recombination at energies *Ei* = 1.596 eV [12] had not been observed for CdTe films. However, the spectra showed a line caused by the recombination of exciton localized on neutral acceptor A0X - *Ei* = (1.583÷1.588) eV (1.589 eV [90, 92, 95, 99, 100]). This line indirectly demonstrated that the investigated films were of *p*-type conductivity and correspondingly low concentration of dopant impurities. Maybe there is a reason for absence of the peak bound with the exciton localized on the neutral donor D0X – 1.593 eV [13, 90, 100] in the registered spectra. The excessive impurity (Li, Na) commonly is an acceptor in II-VI compounds which produces shallow LS near the valence band.

In some PL spectra of CdTe films we have observed the peak due to the phonon repetition of the line from the bound exciton (A0X)-LO at *Ei* = 1.567 eV. The similar peak with *Ei*=1.568 eV and *Ei*=1.570 eV was also observed in [90, 100]. It should be noted that the excitation energy of the longitudinal phonon in CdTe is LO(Г) – 21.2 meV [13, 90, 92]. This value is almost coinciding with that observed experimentally (21 meV) showing our correct interpretation of the experimental data.

The most intensive peak of 1.545 eV was observed in PL spectra from polycrystalline films. The similar peak with energies *Ei*=1.55 eV and *Ei*=1.545 eV was registered by authors [92, 94, 97, 99, 103]. The common interpretation says that this peak is caused by the electron transition between the conduction band and acceptor (е-A) (a single-charged vacancy *VCd* [94] or other shallow acceptor [92, 99]). Nevertheless, authors [13, 95, 103] point out this radiation as a consequence of p resenting donor-acceptor pairs (DAP) where the acceptor is a native defect ( *VCd* ) [13, 103] or another uncontrolled shallow impurity [95]. Authors [97] have found the activation energy of corresponding donors and acceptors: 8 meV and 47meV.

Results of investigated polycrystalline CdTe films in hetero structures CdTe/ZnS under air and vacuum annealing have given [97] another interpretation. It is supposed that the luminescence with 1.55 eV is due to oxygen presence in the material. In which form it exists in the material (substitutional impurity or oxide phase) is not established. However, authors [102] have studied the LS in CdTe single crystals by thermo electronic spectroscopy and demonstrated the energy level 0.06 eV bound with a complex ( *V O Cd Te* )- .

Analyzing our results allows us to conclude that the peak *Ei*=1.545 eV is rather due to the electron transitions between the conduction band and acceptor (a single-charged vacancy or DAP). Really if this peak was caused by oxygen we could it observed in PL spectra from both polycrystalline and epitaxial films but there are no such a peak in PL spectra from the films CdTe/BaF2. Besides that no structural method had revealed the oxygen in these compounds The films under investigation have shown no registered donor impurities of considerable concentration, so the interpretation of this peak as a consequence of the DAP presence is lesser probable than that a consequence of e-A transition.

In some cases the PL spectra from the polycrystalline films showed an asymmetric peak 1.545 eV indicating that in reality it may be a superposition of two nearest lines. Mathematical analysis showed that the most probable position of the additional peak is *Ei* = 1.538 eV. The similar peak was observed in spectra from the epitaxial films CdTe/BaF2. The line with the same energy was revealed by authors [93] in PL spectra from deformed CdTe single crystals and is supposed to be caused by defects generated in the material due to slip of principal Cd(g)-dislocations. Authors [13, 92] explain the peak *Ei*= 1.538 eV as one of unknown nature. Similar interpretation is also in [89] where the line *Ei*= 1.539 eV is caused by DAP (here the acceptor is sodium, *NaCd* ). The next peak *Ei* = 1.525 eV is likely is the phonon repetition of the previous one (е-A)-LO [96].

522 Advanced Aspects of Spectroscopy

experimental results.

near the valence band.

a native defect ( *VCd*

47meV.

interpretation of the experimental data.

acceptors 56 (N) - 263 (Au) meV. We have used these values for further interpretation of the

Unlike ZnTe condensates the peak bound with a free exciton recombination at energies *Ei* = 1.596 eV [12] had not been observed for CdTe films. However, the spectra showed a line caused by the recombination of exciton localized on neutral acceptor A0X - *Ei* = (1.583÷1.588) eV (1.589 eV [90, 92, 95, 99, 100]). This line indirectly demonstrated that the investigated films were of *p*-type conductivity and correspondingly low concentration of dopant impurities. Maybe there is a reason for absence of the peak bound with the exciton localized on the neutral donor D0X – 1.593 eV [13, 90, 100] in the registered spectra. The excessive impurity (Li, Na) commonly is an acceptor in II-VI compounds which produces shallow LS

In some PL spectra of CdTe films we have observed the peak due to the phonon repetition of the line from the bound exciton (A0X)-LO at *Ei* = 1.567 eV. The similar peak with *Ei*=1.568 eV and *Ei*=1.570 eV was also observed in [90, 100]. It should be noted that the excitation energy of the longitudinal phonon in CdTe is LO(Г) – 21.2 meV [13, 90, 92]. This value is almost coinciding with that observed experimentally (21 meV) showing our correct

The most intensive peak of 1.545 eV was observed in PL spectra from polycrystalline films. The similar peak with energies *Ei*=1.55 eV and *Ei*=1.545 eV was registered by authors [92, 94, 97, 99, 103]. The common interpretation says that this peak is caused by the electron transition between the conduction band and acceptor (е-A) (a single-charged vacancy *VCd*

[94] or other shallow acceptor [92, 99]). Nevertheless, authors [13, 95, 103] point out this radiation as a consequence of p resenting donor-acceptor pairs (DAP) where the acceptor is

have found the activation energy of corresponding donors and acceptors: 8 meV and

Results of investigated polycrystalline CdTe films in hetero structures CdTe/ZnS under air and vacuum annealing have given [97] another interpretation. It is supposed that the luminescence with 1.55 eV is due to oxygen presence in the material. In which form it exists in the material (substitutional impurity or oxide phase) is not established. However, authors [102] have studied the LS in CdTe single crystals by thermo electronic spectroscopy and

Analyzing our results allows us to conclude that the peak *Ei*=1.545 eV is rather due to the electron transitions between the conduction band and acceptor (a single-charged vacancy or DAP). Really if this peak was caused by oxygen we could it observed in PL spectra from both polycrystalline and epitaxial films but there are no such a peak in PL spectra from the films CdTe/BaF2. Besides that no structural method had revealed the oxygen in these compounds The films under investigation have shown no registered donor impurities of considerable concentration, so the interpretation of this peak as a consequence of the DAP

demonstrated the energy level 0.06 eV bound with a complex ( *V O Cd Te* )-

presence is lesser probable than that a consequence of e-A transition.

) [13, 103] or another uncontrolled shallow impurity [95]. Authors [97]

.

The PL line *Ei* = 1.497 eV was observed in [99] on monocrystalline CdTe samples under doping by ion implantation. As this line has appeared in the samples doped with oxygen only the authors suggest it is caused by the presence of this impurity. Other authors suppose this line is due to electron transitions between the conduction band and the level of the substitutional impurity acceptor *AgCd* (*EV+*0.107 eV) [13] or by native defect 2 *VCd* (*Ev+*0.111 eV) [13].

The wide radiation stripe in polycrystalline films at the energy 1,45 eV is separated in single peaks based on results from the PL of epitaxial films. They are shown in Fig. 5.

The peak 1.476 eV in [96-98] is due to longitudinal defects (dislocations and DP, a so-called Y-stripe). Authors [90, 98] assume the Y-stripe at (1.46-1.48) eV is caused by longitudinal defects (dislocations). Authors [99] make it more precicely: this peak is caused by the recombination of exciton localized on slipped Cd-dislocations. Authors [93] have investigated The photoluminescence of deformed CdTe single crystals and showed that the peak *Еi*=1.476 eV is not caused by the Cd-dislocations but is due to the electron states of 600 Те(g)-dislocations (-dislocations). So that, the number of authors have the same opinion that this line in PL spectra is caused by the longitudinal defects. We also agree with this interpretation.

The lines 1.453 eV, 1.433 eV and 1.413 eV which are good resolved in spectra from the epitaxial films CdTe/BaF2 are very similar to 1LO, 2LO, 3LO repetitions of the peak *Ei* = (1.473÷1.476) eV. However, the energy difference of these lines (*Е*=0.0200 eV) does not coincide with the energy of longitudinal optical phonons in CdTe 0.0212 eV making it difficult to interpret the corresponding peaks unambiguously. At the same time, the analogous set of lines with the LO structure and the energy difference 0.0200 eV in the range *E* = (1.39÷1.45) eV was observed by authors [101]. They have studied polycrystalline CdTe films deposited by vacuum evaporation at *Ts* = (723÷823) К on glass and aluminum substrates.

Authors [100] have examined undoped and doped with donor impurities (Al, In) CdTe single crystals and also have observed the PL stripe in the energy range *E* = (1.380÷1.455) eV containing four lines with LO structure. The authors interpreted them as electron transition between DAP and their phonon repetitions. Authors [81, 91] suppose the wide peak 1.46 eV is due to the excitons localized in longitudinal defects, probably dislocations (*Y*-stripe). The lines 1.455 eV, 1.435 eV and 1.415 eV were observed in [94] from the polycrystalline CdTe films prepared by the gas-transport method.

As we see the most authors have an unique opinion: the set of lines in the range *Е*=(1.413÷1.476) eV is due to longitudinal defects (rather dislocations), and their intensity [93] can be a measurement unit of these defects in the material.

For polycrystalline films (Fig. 5, а) the LO structure of the stripe caused by the longitudinal defects at energies ~1.45 eV has practically not been observed, maybe because of superposition with additional lines of another origin.

The defect complexes in the material (A-centers) are also resolved by the PL in the same energy range (it can be considered as a partial case of DAP). According to [13, 97] А- centers

 <sup>2</sup> *V D Cd* where Cl is a donor produce the line and its LO-phonon repetitions with energies 1.454, 1.433, 1.412, 1.391, 1.370, 1.349 and 1.328 eV. However, as is seen in Fig. 5, this stripe is displaced relatively to that observed experimentally, so the experimental PL spectra of CdTe films can be completely explained by these complexes only. The narrower stripe with peaks 1.458, 1.437, 1.417 and 1.401 eV produces the А-center where indium is a donor. This stripe has the better coincidence with experimental one but is also displaced. Besides that, it is difficult to explain why the А-complex is observed in the polycrystalline films and is not observed in the epitaxial layers while the charge mixture for both types of the films is the same. Thus we suppose the interpretation of the wide stripe in the energy range *Е* = (1.413÷1.4760) eV due to longitudinal defects is more reliable.

Under change of condensation conditions of polycrystalline samples we have observed the change of intensity for a stripe due to prolonged defects (~1.45 eV). As shows Fig. 5, as the substrate temperature increases from 473 К to 623 К the intensity of this stripe is decreasing and then it increases as the Ts increases. These results have a good correlation with data of investigation of CdTe film substructure [49], this fact points out an enhance of the structural quality (lowering vacancy concentration) of the bulk crystallites in condensates under elevating substrate temperature up to *Ts*=623 К, but this quality becomes lower as the substrate temperature increases over 623 K.

As the substrate temperature elevated (*Ts*>723 К) the optical properties of CdTe films were strongly degraded forming a number of additional peaks in the PL spectra which finally become a bell-like curve without possibility to identify the separate lines. Morphological studies demonstrated further increase of the crystallite sizes in this temperature range. However, the volume of these crystallites becomes a high-defective one.

Table 6 summarizes results of PL spectra interpretation for CdTe films showing their high optical quality.

## **5.2. ZnTe films**

Fig. 6 illustrates typical PL spectra of ZnTe films registered at 4.5 K. A number of lines is observed, their energies are indicated in the Fig. 6 and are listed in Table 7. Analysis and interpretation of the PL peaks are carried out according to reference data [75-89].

The low-temperature PL spectra of ZnTe films show a set of peaks originated from: i) optical transitions under participation of free (X) and bound on neutral donor (D0X) and acceptor (A0X) excitons; ii) transitions valence band – acceptor impurity (е-А), iii) radiation due to presence of longitudinal defects (dislocations, *Y*-stripe); iv) optical transitions where phonons of different type are participating (LO (0.0253 eV), TO, LA (0.0145 eV), TA (0.007 eV) -repetition).

524 Advanced Aspects of Spectroscopy

<sup>2</sup> *V D Cd*

optical quality.

**5.2. ZnTe films** 

(*Y*-stripe). The lines 1.455 eV, 1.435 eV and 1.415 eV were observed in [94] from the

As we see the most authors have an unique opinion: the set of lines in the range

For polycrystalline films (Fig. 5, а) the LO structure of the stripe caused by the longitudinal defects at energies ~1.45 eV has practically not been observed, maybe because of

The defect complexes in the material (A-centers) are also resolved by the PL in the same energy range (it can be considered as a partial case of DAP). According to [13, 97] А- centers

Under change of condensation conditions of polycrystalline samples we have observed the change of intensity for a stripe due to prolonged defects (~1.45 eV). As shows Fig. 5, as the substrate temperature increases from 473 К to 623 К the intensity of this stripe is decreasing and then it increases as the Ts increases. These results have a good correlation with data of investigation of CdTe film substructure [49], this fact points out an enhance of the structural quality (lowering vacancy concentration) of the bulk crystallites in condensates under elevating substrate temperature up to *Ts*=623 К, but this quality becomes lower as the

As the substrate temperature elevated (*Ts*>723 К) the optical properties of CdTe films were strongly degraded forming a number of additional peaks in the PL spectra which finally become a bell-like curve without possibility to identify the separate lines. Morphological studies demonstrated further increase of the crystallite sizes in this temperature range.

Table 6 summarizes results of PL spectra interpretation for CdTe films showing their high

Fig. 6 illustrates typical PL spectra of ZnTe films registered at 4.5 K. A number of lines is observed, their energies are indicated in the Fig. 6 and are listed in Table 7. Analysis and

interpretation of the PL peaks are carried out according to reference data [75-89].

 where Cl is a donor produce the line and its LO-phonon repetitions with energies 1.454, 1.433, 1.412, 1.391, 1.370, 1.349 and 1.328 eV. However, as is seen in Fig. 5, this stripe is displaced relatively to that observed experimentally, so the experimental PL spectra of CdTe films can be completely explained by these complexes only. The narrower stripe with peaks 1.458, 1.437, 1.417 and 1.401 eV produces the А-center where indium is a donor. This stripe has the better coincidence with experimental one but is also displaced. Besides that, it is difficult to explain why the А-complex is observed in the polycrystalline films and is not observed in the epitaxial layers while the charge mixture for both types of the films is the same. Thus we suppose the interpretation of the wide stripe in the energy

*Е*=(1.413÷1.476) eV is due to longitudinal defects (rather dislocations), and their intensity

polycrystalline CdTe films prepared by the gas-transport method.

[93] can be a measurement unit of these defects in the material.

range *Е* = (1.413÷1.4760) eV due to longitudinal defects is more reliable.

However, the volume of these crystallites becomes a high-defective one.

superposition with additional lines of another origin.

substrate temperature increases over 623 K.

We calculated activation energies of corresponding processes using the expression (26). The gap of ZnTe crystal at 4.5 K was supposed to be *Eg* = 2.394 eV. As the examined material was of p-type conductivity the activation energies were counted down relative to the valence band. Table 7 summarizes these data.

**Figure 6.** Photoluminescence spectra registered at *T*=4.5 К for ZnTe films prepared at *Te* = 973 К and various *Ts*, К: 573 К (1); 673 К (2); 773 К (3)

Optical transitions with energy (2.381÷2.383) eV were observed in [68, 75-82, 84-86] where authors had studied monocrystalline or bulk polycrystalline ZnTe of high structural and optical quality. These transitions are commonly relating to a free exciton (X). Earlier [82] the PL line of *Ei* = (2.374÷2.375) eV was suggested to be caused by the exciton bound on neutral acceptor (zinc vacancy *VZn*). Further [76-78] it was shown that other acceptor centers take part in forming such an excitonic complex, in particular, acceptor centers due to uncontrollable impurities (Li, Cu) in ZnTe are of interest. However, in the most recent works [68, 81] this line is ascribed to the exciton localized on shallow neutral donor (atoms of uncontrollable impurities from 3rd and 7th Groups of the Periodical System (In, Ga, Al, Cl, Br. I)). These impurities form in the gap of the material more narrower levels than the acceptor ones. The line with *Ei*=2.371 eV which is energetically closed to that considered above is due to radiation of bound excitons [76-78, 81]; nevertheless the impurity (acceptor) in this complex has (obviously) somewhat larger energy level causing other energy of the stripe. These acceptors are native defects and uncontrollable excessive impurities (Li, Na, Ag, Cu).


According to [81] LiZn is the most probable candidate while it forms in the gap of the material energy level 60.6 meV.

**Table 7.** Principal lines in PL spectra of ZnTe films and their interpretation

It should be noted that the presence of excitonic lines in PL spectra from high-temperature ZnTe condensates points out their high optical and crystal quality. These lines are of sufficient intensity in the spectra from the films deposited at the substrate temperature *Ts*=573 К and the larger intensity for condensates prepared at *Ts*=673 К. Excitonic lines in the spectra from low-temperature condensates and layers manufactured at *Ts>*773 К were not registered. Thus, the results of PL studies indicate that the films deposited at the substrate temperatures *Ts=* (623÷673) К are the most optically perfect layers. These data are coinciding with the results of investigations of substructural characteristics of ZnTe films reported earlier [54]. According to these data the dependence of the CSD sizes on the substrate temperature is a curve with the maximum at *Ts* = (600÷650) К. The minimal dislocation concentration is also observed in the films under these temperatures.

The line *Е<sup>i</sup>* = 2.34 eV belonging [66] to *VZn* is not observed in spectra of the radiation recombination in ZnTe films. This fact is also confirming high stoichiometry of the films under study.

The set of nearest lines in the energy range *Е* = (2.30÷2.33) eV and *Е* = (2.17÷2.25) eV authors [76-79] ascribe to the electron transitions from the conductance band to the shallow acceptor levels formed by Li or Cu atoms and their phonon repetitions (LO – 25.5 meV). There are stripes 2*S*Li, 3*Sb*Li, (e-A) Li, 2P Li, 4*Sb*Li, 4*Sb*Li-LO, 2*S*Cu, 3*Sb*Cu, 4*Sb*Cu, 2*Sb* Cu - LO, 2*Sb* Cu - 2LO and others. Experimental and theoretical values of the activation energy for ground and excited states for the main excessive impurities in ZnTe (lithium and copper) are reported in [76]. They are in the energy range *E* = (0.0009÷0.0606) eV for Li and *E* = (0.001÷0.148) eV for Cu. However, in [65] the line *Е<sup>i</sup>* = 2.332 eV is supposed to be due to other excessive impurity *NaZn*, and in [82] this line is due to the native defect *VZn*. Another optical transition *Е<sup>i</sup>* = 2.27 eV authors [77] ascribe to the Ag impurity 2S Ag.

What about the peaks in the energy range *Е* = (2.10÷2.21) eV. These transitions were for the first time observed in [75-79] and authors had called them *Yi*-lines. They are ascribed to the distortions of the crystalline lattice of the material near incoherent twin boundaries, dislocations and other longitudinal defects where the dangling bonds are formed in the semiconductor material. So that, the lines *Еi* = 2.159 eV and *Еi=*2.194 eV can be interpreted as *Y2* (2.155 eV) and *Y1* (2.195 eV) [75]. They are due to longitudinal defects and the change of their intensity may point out the change of these defects concentration in the material. Somewhat other energy position of the line due to oxygen (2.06 eV) is reported in [66]. Thus, analysis of the reference data has forced us to conclude that PL lines in the energy interval *Е* = (1.835÷2.055) eV are rather caused by oxygen, its complexes and phonon repetitions. If it is true, the analysis of PL spectra from ZnTe films indicates the increase of the oxygen content in the samples under increasing the condensation temperature. Actually, if there is no oxygen in the samples prepared at 573 К, its concentration in high-temperature films (*Ts*=773 К) is sufficiently larger. Oxygen concentration in the material strongly depends on the vacuum conditions under the film preparation and the charge mixture quality.

#### **5.3. ZnS films**

526 Advanced Aspects of Spectroscopy

2.375 2.375

2.331 2.334; 2.332

2.208 0.186

2.159 0.235

Radiation line *Еi*, eV

under study.

material energy level 60.6 meV.

Reference data, *Еi*, eV

According to [81] LiZn is the most probable candidate while it forms in the gap of the

2.383 2.381÷2.383 0.011 Exciton *X, n=*1 [68, 76, 82]

2.371 2.374; 2.375 0.023 Exciton *A0X, A-Li,Cu* [68]

2.301 2.307 0.093 (е-А)-LO *А - LiZn* [76-77] 2.270 2.270 0.124 е-А *А - AgZn* [76] 2.233 2.230 0.161 е-А *А - CuZn* [75, 79]

2.194 2.195; 2.19 0.200 е-А *Y1* [75]

2.151 2.155 0.243 е-А *Y2* [75, 76, 78, 84]

It should be noted that the presence of excitonic lines in PL spectra from high-temperature ZnTe condensates points out their high optical and crystal quality. These lines are of sufficient intensity in the spectra from the films deposited at the substrate temperature *Ts*=573 К and the larger intensity for condensates prepared at *Ts*=673 К. Excitonic lines in the spectra from low-temperature condensates and layers manufactured at *Ts>*773 К were not registered. Thus, the results of PL studies indicate that the films deposited at the substrate temperatures *Ts=* (623÷673) К are the most optically perfect layers. These data are coinciding with the results of investigations of substructural characteristics of ZnTe films reported earlier [54]. According to these data the dependence of the CSD sizes on the substrate temperature is a curve with the maximum at *Ts* = (600÷650) К. The minimal dislocation

The line *Е<sup>i</sup>* = 2.34 eV belonging [66] to *VZn* is not observed in spectra of the radiation recombination in ZnTe films. This fact is also confirming high stoichiometry of the films

The set of nearest lines in the energy range *Е* = (2.30÷2.33) eV and *Е* = (2.17÷2.25) eV authors [76-79] ascribe to the electron transitions from the conductance band to the shallow acceptor levels formed by Li or Cu atoms and their phonon repetitions (LO – 25.5 meV). There are stripes 2*S*Li, 3*Sb*Li, (e-A) Li, 2P Li, 4*Sb*Li, 4*Sb*Li-LO, 2*S*Cu, 3*Sb*Cu, 4*Sb*Cu, 2*Sb* Cu - LO, 2*Sb* Cu - 2LO and others. Experimental and theoretical values of the activation energy for ground and excited states for the main excessive impurities in ZnTe (lithium and copper)

*Е*, eV Recombination type Interpretation

*А0X* [76-77] *А- VZn* [82] *D0X, D- In* [68, 86]

е-А *А - LiZn, NaZn* [75, 76]

Activation energy,

2.379 0.019 Exciton

0.060 (0.061 – Li [75, 78]) (0.063 – Na [75])

**Table 7.** Principal lines in PL spectra of ZnTe films and their interpretation

concentration is also observed in the films under these temperatures.

Low-temperature photoluminescence is the most reliable tool for examining wide gap materials providing minimization of overlapping peaks due to various recombination processes. The typical PL spectra from ZnS films at 4.7 K are shown in Fig. 7. The detailed analysis of the PL spectra (identification of complex broadened lines) was carried out by ORIGIN program. Maximums of the peaks revealed by this analysis (Fig.7) are noted by vertical lines.

It should be noted that the PL spectra registered at various temperatures of experiment have no sufficient distinctions except those with somewhat larger line intensities in spectra obtained at 77 K. Analysis of the spectra shows that for ZnS films deposited at *Т<sup>s</sup>* = (393- 613) K the peaks with *λ<sup>i</sup>* = 396 nm (*Ei* = 3.13 eV) and *λ<sup>i</sup>* = 478 nm (*Ei*=2.59 eV) are dominating. Further working-out of the spectra demonstrated that the peak *λi*= 396 nm is asymmetric (Fig. 7) what is may be explained by the superposition of two closely placed lines. The spectra also have low intensity peaks with *λ<sup>i</sup>* =603 nm (*Ei*=2.06 eV) and *λ<sup>i</sup>* =640 nm (*Ei*=1.94 eV).

PL spectra from the films prepared at higher *Тs* is sufficiently changed. There is a number of overlapping peaks where the most intensive ones are in the wavelength range *λi* = (560÷620) nm.

Under interpretation of PL spectra from ZnS films we have calculated the activation energies of processes causing the corresponding lines. We also have suggested the PL radiation took place under transfers of electrons from the conduction band (or shallow donors) to the deep LS in the gap of the material. Then the optic depth of the energy level of the defect (Δ*E*) relative to the valence band causing the spectral peak may be found from (26) supposing the optical gap of the material at 4.5 K is *Eg* = 3.68 eV.

Taking into account that the chalcogenide films were not doped in-advance one can suggest that the lines in spectra are due to transfers of carriers between conduction band and LS caused by the native point defects, their complexes and uncontrolled impurities. We made an attemption to identify these LS according to reference data [104-108] (Table 8). As is shown there is a good correlation of our results and those obtained by other authors for ZnS single crystals.

**Figure 7.** Typical PL spectra for ZnS films (a) and the example of the peak differentiation (b)

The investigations have shown that the Schottky defect *VZn* is a dominant defect type in ZnS films prepared at low substrate temperatures *Ts* = (393-613) K. As *Ts* increases the number of single-charged Zn vacancies in the condensates decreases, and concentration of doublecharged Zn vacancies increases. In the films deposited at higher substrate temperatures *Ts*=(653-893) K single-charged S vacancies *VS* and double-charged S vacancies 2 *VS* and interstitial Zn atoms *Zni* are dominating.

Such features of the PD ensemble in the samples are obviously caused by processes of condensation and re-evaporation of Zn and S atoms from the substrate. Actually, at low *Ts* the defect formation in the films is determined by higher S pressure comparing to Zn pressure in the mixture vapor providing Zn vacancy formation in ZnS condensates. As *Ts* increase the PD ensemble in the material is determined by the more rapid re-evaporation of the same S atoms from the substrate resulting in production of Zn-beneficiated films. Sulfur vacancies and interstitial Zn atoms are being dominant defects in such condensates.


single crystals.

radiation took place under transfers of electrons from the conduction band (or shallow donors) to the deep LS in the gap of the material. Then the optic depth of the energy level of the defect (Δ*E*) relative to the valence band causing the spectral peak may be found from

Taking into account that the chalcogenide films were not doped in-advance one can suggest that the lines in spectra are due to transfers of carriers between conduction band and LS caused by the native point defects, their complexes and uncontrolled impurities. We made an attemption to identify these LS according to reference data [104-108] (Table 8). As is shown there is a good correlation of our results and those obtained by other authors for ZnS

**Figure 7.** Typical PL spectra for ZnS films (a) and the example of the peak differentiation (b)

*Ts*=(653-893) K single-charged S vacancies *VS*

interstitial Zn atoms *Zni* are dominating.

condensates.

The investigations have shown that the Schottky defect *VZn* is a dominant defect type in ZnS films prepared at low substrate temperatures *Ts* = (393-613) K. As *Ts* increases the number of single-charged Zn vacancies in the condensates decreases, and concentration of doublecharged Zn vacancies increases. In the films deposited at higher substrate temperatures

Such features of the PD ensemble in the samples are obviously caused by processes of condensation and re-evaporation of Zn and S atoms from the substrate. Actually, at low *Ts* the defect formation in the films is determined by higher S pressure comparing to Zn pressure in the mixture vapor providing Zn vacancy formation in ZnS condensates. As *Ts* increase the PD ensemble in the material is determined by the more rapid re-evaporation of the same S atoms from the substrate resulting in production of Zn-beneficiated films. Sulfur vacancies and interstitial Zn atoms are being dominant defects in such

and double-charged S vacancies 2 *VS*

and

(26) supposing the optical gap of the material at 4.5 K is *Eg* = 3.68 eV.

**Table 8.** Results of PL spectra working-out and their comparison with reference data (solid values are for peaks of maximum intensity)

The PL spectra of ZnS films have also revealed low intensity lines from the activator impurity (*Cu*) and, possible, *<sup>i</sup> S* [106] or a complex defect 2 (, ) *O Vs Zn* [107]. Results of studying low-temperature PL allowed constructing energy position model of native point defects in zinc sulfide films prepared by quasi-closed space technique (Fig. 8).

**Figure 8.** The model of the level positions for native point defects in band gap of ZnS films

To explain the experimental results we have used quasi-chemical formalism for modeling the point defects ensemble in the examined chalcogenide films in dependence on physical technical conditions of layer condensation. This method concerns all defects, electrons and holes as components of thermodynamic equilibrium in the bulk crystal (complete equilibrium of the point defects). Then the modeling procedure reduces to solving set of equations which describe penetration of point defects into solid from the gas state along with the equation of electroneutrality and intrinsic conductivity equation [109-110]. The most complete spectrum of the native defects was taken into account under modeling the point defects ensemble. Calculations were carried out for the complete defects' equilibrium as well as for their quenching. Under modeling we have used energies of native defects formation obtained «*ab initio*» in [56-60]. Reference data of ionization energies of acceptor and donor centers of point defects in CdTe, ZnS, and ZnTe were used along with results of our experiments. Main data of modeling are presented in [111-116].
