**5. Low-temperature LPE growth and characterization of dilute nitride GaAsN and InGaAsN thick layers**

Dilute nitride III-V-N alloys with nitrogen content in the range of few percent, such as GaAsN and InGaAsN, are of considerable interest for application in multijunction solar cells.

The incorporation of nitrogen into group V sublattice causes profound effect on the band gap and properties of the dilute nitride material strongly differ from those of the conventional III-V alloys. While in conventional alloys a smaller lattice constant increases the band gap, the mixing of GaAs with few molar percent of GaN leads to giant reduction of its band gap due to the smaller covalent radius and large electronegativity of N atoms. The large changes in the electronic structure in dilute III–V nitrides could be explained by the band anticrossing model (BAC). The interaction between the localized levels introduced by a highly electronegative impurity, such as N in GaNxAs1−x, and the delocalized states of the host semiconductor causes a restructuring of the conduction band into E+ and E− subbands, which in this case effectively lowers the conduction band edge of the alloy.

temperature at the interface between the liquid phase and substrate can not be measured and common it is determined by measurements of the source component solubility

Fig. 4.3. Slide boat for growth from finite melt: 1, boat body with container for melts;

Slide boats with different design modification are used for growth of variety structures in different multicomponent system. A boat made of two different materials, sapphire (for body) and graphite (for slider), is suggested by Reynolds and Tamargo (Reynolds and Tamargo, 1984). This design reduces temperature variations around the perimeter of the substrate which contribute to unwanted 'edge' growth effects. Slide boats with narrowed melt contact for epitaxy of extremely thin epilayers have been used to grow active layer in single-quantum well lasers by (Alferov et al, 1985 ) and later by (Kuphal, 1991). Also a modified slide boat can be used for multilayer periodic structures growth (Arsent'ev et al, 1988). The use of two growth chambers with narrow slits makes it possible to produce such structures by means of repeated reciprocating movements of the slider with the substrate situated underneath these slits. Another variant of an LPE boat (Mishurnyi et la, 1997), which is a combination of the 'sliding' and 'piston' designs has been used successfully to grow InGaAsSb, AlGaAsSb and various multilayer structures on the basis of these materials.

2

3

1

**5. Low-temperature LPE growth and characterization of dilute nitride GaAsN** 

Dilute nitride III-V-N alloys with nitrogen content in the range of few percent, such as GaAsN and InGaAsN, are of considerable interest for application in multijunction solar

The incorporation of nitrogen into group V sublattice causes profound effect on the band gap and properties of the dilute nitride material strongly differ from those of the conventional III-V alloys. While in conventional alloys a smaller lattice constant increases the band gap, the mixing of GaAs with few molar percent of GaN leads to giant reduction of its band gap due to the smaller covalent radius and large electronegativity of N atoms. The large changes in the electronic structure in dilute III–V nitrides could be explained by the band anticrossing model (BAC). The interaction between the localized levels introduced by a highly electronegative impurity, such as N in GaNxAs1−x, and the delocalized states of the host semiconductor causes a restructuring of the conduction band into E+ and E− subbands,

which in this case effectively lowers the conduction band edge of the alloy.

2, slider with container for finite melts; 3, slider for the substrates.

(Mishurnyi et al, 1999) .

**and InGaAsN thick layers** 

cells.

Figure 5.1. shows the relationship between the lattice constant and band-gap energy in some III-V semiconductor alloys. In the case of InGaNAs adding In to GaAs increases the lattice constant, while adding N to GaAs decreases the lattice constant. In the same time the incorporation of In and N in GaAs leads to reduction of the band gap energy in the new alloy. Consequently, by adjusting the contents of In and N in quaternary InGaNAs alloys can be grown lattice-matched to GaAs layers because In and N have opposing strain effects on the lattice and make it possible to engineer a strain-free band gap layers suitable for different applications.

Fig. 5.1. Relationship between lattice constant and bad gap energy for some III-V semiconductor alloys

Recently a development of the spectral splitting concentrator photovoltaic system based on a Fresnel lens and diachronic filters has a great promise to reach super high conversion efficiencies (Khvostiokov et al. 2010). Module efficiency nearly 50% is expected for the system with three single-junction solar cells connected in series with band gap of 1.88-1.42-1.0 eV. The development of three optimized AlGaAs, GaAs and InGaAsN based cells is the best combination for application in such system if PV quality of the quaternary InGaAsN could be reached by LPE growth.

In this paper low-temperature LPE is proposed as a new growth method for dilute nitride materials. Because of its simplicity and low cost many experiments on GaInAsN and GaAsN growth under different condition and with different doping impurities could be made using LPE. The systematic study of their structural, optical and electrical properties by various methods make it possible to find optimized growth conditions for InGaAsN quaternary compounds lattice matched to GaAs substrate.

### **5.1 Growth and characterization of GaAsN layers**

GaAsN compounds were grown by the horizontal graphite slide boat technique for LPE on (100) semi-insulating or n-type GaAs substrates. A flux of Pd-membrane purified hydrogen at atmospheric pressure was used for experiments. No special baking of the system was done before epitaxy. Starting materials for the solutions consisted of 99.9999 % pure Ga, polycrystalline GaAs and GaN. The charged boat was heated at 750oC for 1 h in a purified H2 gas flow in order to dissolve the source materials and decrease the contaminants in the

Dilute Nitride GaAsN and InGaAsN Layers Grown by Low-Temperature Liquid-Phase Epitaxy 83

The effect of N-N interstitial is very small and can be neglected. Also the formation of an isolated N interstitial is unlikely due to a high formation energy (Li et al. 2001) and their concentrations in GaAsN is very small. While the substitutional NAs atoms compress the lattice constant , the N-As complexes expand the lattice constant of GaAsN in the growth direction, as shown in Figure 5.3. So, XRD results may underestimated the N composition

Fig. 5.3. Incorporation of N-atom in As-sublattice: a) as substitutional atom NAs; b) as As-N

a b

XRD rocking curves are recorded in the symmetrical (004) reflection. Fig. 5.4 shows the experimental XRD rocking curves of two GaAsN samples, 1.2 μm thick, with N composition of 0.3% and 0.62% and may consider that they are fuly relaxed. The N content determines the line shape of the main peak of the spectra: it manifests itself as a broad shoulder evolving into a weak separate peak shifted away to the right from the (004) GaAs substrate reflection. Our data show that N compositions measured by the two methods, SIMS and XRD, agree well and XRD measurements by using Vegard's law could be used to determine the lattice constant of GaAsN layers containing low N concentrations. These results are in a good agreement with the calculations from the theoretical model and the experimental results for small N concentration in the GaAsN reported in the literature. The deviation from Vegard's law has been observed for nitrogen concentration levels above 2.9 mol % GaN in

33,1 33,2 33,3 33,4 33,5 33,6

O m ega, degree

Fig. 5.4. XRD rocking curves for GaAsN samples with differenrt N content.

0.62% N

0.3% N

due to the N-As and N-N split interstitials.

the layer ( Spruytte at all., 2001; Li et al. 2001).

100

1000

10000

Intensity

split interstitial

melt. Epitaxial GaAsN layers 0.8-1.5 thick were grown from different initial temperatures varied in the range 560-650 ºC at a cooling rate of 0.6 ºC/min.

### **5.1.1 Structural characterization**

XRD and SIMS techniques are used to determine N concentration in grown samples.

While SIMS measures the total nitrogen content in the layer, XRD determines the change in the lattice constant due to the substitution of nitrogen atoms on As-sublattice sites.

The N composition from XRD results could be estimated assuming Vegard's law. In many cases the Vegard's law is a good approximation for the lattice parameter dependence on the composition. The deviation from Vegard's law dependences on many parameters, for instance, the difference in the atom bond length, different atom electronegativity and elastic constants of the components in the alloy. For the ideal case N incorporates predominantly as substitutional NAs atoms in As- sublattice substituting As atoms. However, it is known that there are some other N configurations: N-As split interstitial; N-N split interstitial; and isolated N interstitial. Figure 5.2 presents the main configurations of N in GaAsN as substitutional atom NAs and as As-N and N-N split interstitials, respectively.

The influence of these N-related complexes on the lattice constant can be calculated on the base of the theoretical model of Chen (Chen et al. 1996) for analyzing the correlation between lattice parameters and point defects in semiconductors. According this model the lattice strain caused by the substitutional NAs is given by the following relation:

( ) 2( ) *N As Ga As a xr r a rr* (5.1.1)

where: *rN, rGa,, rAs* the covalent radii; (1 ) / (1 ) , and ν is the Poissn ratio

The lattice strained caused by split interstitial is:

$$\frac{\Delta a}{a} = \mu \frac{\mathbf{x} (d\_b - r\_{\odot a} - r\_{\text{Ar}})}{\mathbf{2} (r\_{\odot a} + r\_{\text{Ar}})} \tag{5.1.2}$$

Where db is the distance of the N-As complex from its nearest neighbours:

$$d\_s = \frac{\sqrt[4]{3}}{3}r\_u + \sqrt{(r\_u + r\_{Gu})^2 + \frac{2}{3}r\_u^{\frac{2}{3}}} \tag{5.1.3}$$

where *rsi* =( *rN* + *rAs*)/2 is an effective bond radius.

Fig. 5.2. The main configurations of nitrogen atoms in GaAsN N-atom, As-atom Ga-atom

melt. Epitaxial GaAsN layers 0.8-1.5 thick were grown from different initial temperatures

While SIMS measures the total nitrogen content in the layer, XRD determines the change in

The N composition from XRD results could be estimated assuming Vegard's law. In many cases the Vegard's law is a good approximation for the lattice parameter dependence on the composition. The deviation from Vegard's law dependences on many parameters, for instance, the difference in the atom bond length, different atom electronegativity and elastic constants of the components in the alloy. For the ideal case N incorporates predominantly as substitutional NAs atoms in As- sublattice substituting As atoms. However, it is known that there are some other N configurations: N-As split interstitial; N-N split interstitial; and isolated N interstitial. Figure 5.2 presents the main configurations of N in GaAsN as

The influence of these N-related complexes on the lattice constant can be calculated on the base of the theoretical model of Chen (Chen et al. 1996) for analyzing the correlation between lattice parameters and point defects in semiconductors. According this model the

> *a xr r a rr*

*a xd r r a rr* 

Where db is the distance of the N-As complex from its nearest neighbours:

( ) 2( ) *N As Ga As*

( ) 2( ) *b Ga As Ga As*

As

N

(5.1.1)

(5.1.2)

N

N

<sup>3</sup> 3 2 2 2 ( ) 3 3 *<sup>b</sup> si si Ga si d r rr r* (5.1.3)

XRD and SIMS techniques are used to determine N concentration in grown samples.

the lattice constant due to the substitution of nitrogen atoms on As-sublattice sites.

substitutional atom NAs and as As-N and N-N split interstitials, respectively.

lattice strain caused by the substitutional NAs is given by the following relation:

, and ν is the Poissn ratio

The lattice strained caused by split interstitial is:

where *rsi* =( *rN* + *rAs*)/2 is an effective bond radius.

N-atom, As-atom Ga-atom

Fig. 5.2. The main configurations of nitrogen atoms in GaAsN

varied in the range 560-650 ºC at a cooling rate of 0.6 ºC/min.

**5.1.1 Structural characterization** 

where: *rN, rGa,, rAs* the covalent radii;

 

 (1 ) / (1 ) 

NAs

The effect of N-N interstitial is very small and can be neglected. Also the formation of an isolated N interstitial is unlikely due to a high formation energy (Li et al. 2001) and their concentrations in GaAsN is very small. While the substitutional NAs atoms compress the lattice constant , the N-As complexes expand the lattice constant of GaAsN in the growth direction, as shown in Figure 5.3. So, XRD results may underestimated the N composition due to the N-As and N-N split interstitials.

Fig. 5.3. Incorporation of N-atom in As-sublattice: a) as substitutional atom NAs; b) as As-N split interstitial

XRD rocking curves are recorded in the symmetrical (004) reflection. Fig. 5.4 shows the experimental XRD rocking curves of two GaAsN samples, 1.2 μm thick, with N composition of 0.3% and 0.62% and may consider that they are fuly relaxed. The N content determines the line shape of the main peak of the spectra: it manifests itself as a broad shoulder evolving into a weak separate peak shifted away to the right from the (004) GaAs substrate reflection. Our data show that N compositions measured by the two methods, SIMS and XRD, agree well and XRD measurements by using Vegard's law could be used to determine the lattice constant of GaAsN layers containing low N concentrations. These results are in a good agreement with the calculations from the theoretical model and the experimental results for small N concentration in the GaAsN reported in the literature. The deviation from Vegard's law has been observed for nitrogen concentration levels above 2.9 mol % GaN in the layer ( Spruytte at all., 2001; Li et al. 2001).

Fig. 5.4. XRD rocking curves for GaAsN samples with differenrt N content.

Dilute Nitride GaAsN and InGaAsN Layers Grown by Low-Temperature Liquid-Phase Epitaxy 85

FTIR absorption spectra of an as grown GaAs1-xNx layer on a n-GaAs substrate is plotted in Fig. 5.6. A peak at 472.6 cm-1, attributed to a local vibrational mode of nitrogen at arsenic site

Electrical parameters of undoped GaAs and GaAsN layers with different nitrogen content grown on seminsulating (001) GaAs substrates are measured in the temperature range

Figure 5.7. shows the temperature dependence of the Hall-concentration *nH* on reciprocal temperature for two layers GaAsN with nitrogen concentration of 0.2% and 0.5%, respectively in comparison with undoped GaAs. It is seen that all samples are of n-type and for layers containing nitrogen electron concentration increases about one order of magnitude. This could be explained by the assumption that nitrogen behaves mainly as an isoelectronic donor, which arises from the local heterojunction scheme GaAs-GaN according to Belliache (Bellaiche et al., 1997). The results shown in figure indicate that the free carrier concentration increases strongly with the N concentration. The increase in *nH* has also been observed in GaNxAs1−x doped with S (Yu et al., 2000a) and in Ga1−3x In3xNxAs1−x alloys doped with Se (Skierbiszewski at al., 2000). This large increase of the free electron concentration can be quantitatively explained by a combination of the band anticrossing model (Shan et al, 1999) and the amphoteric defect model (Walukiewicz, 1989). The later suggests that the maximum free carrier concentration in a semiconductor is determined by the Fermi energy with respect to the Fermi-level stabilization energy EFS which is a constant for III-V semiconductors. Since the position of the valence band in GaAsN is independent of N concentration, the giant downward shift of the conduction band edge toward EFS and the enhancement of the density of states effective mass in GaAsN lead to much larger concentration of uncompensated, electrically active donors for the same location of the Fermi energy relative to EFS. In order to explain the large enhancement of the doping limits in dilute nitride alloys both the effects of band gap reduction and the increase in the effective mass have to be taken into account (Yu et al., 2000 b; Skierbiszewski at al., 2000).

Fig. 5.7. Free carrier concentration as a function of inverse temperature for as grown GaAs,

2 4 6 8 10 12 14

1000/T, K-1

 0% N 0.2% N 0.5% N

and two GaAsN layers with different N content

10<sup>17</sup>

10<sup>18</sup>

Hall concentrations, cm-3

in GaAs is clearly seen.

**5.1.2 Electrical characterization** 

80 – 300 K using van der Pauw geometry.

Unlike XRD used for assessing the incorporation of nitrogen in GaAs1-xNx alloys grown by LPE the nitrogen bonding configurations and local atomic structures have been studied using x-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) spectroscopy. The XPS spectra have been measured over a range of binding energies from 1 to 550 eV. The X-ray photoelectron spectra of N 1s photoelectron and Ga LMM Auger lines recorded from the as grown GaAs1-xNx samples prepared in different temperature ranges are shown in Fig 5.5. It is clearly seen the Ga Auger peak around 391 eV and the N 1*s* level photoemission peak of the samples. The variation of the intensity of the N 1*s* peak with respect to the Ga LMM peaks reflects is due to the different nitrogen content of the samples. Sample grown from higher initial epitaxy temperature of 650 ºC contains 0.2% N and exhibits a N 1s peak with lower intensity and lower binding energy in comparison with the N 1s peak intensity of the sample grown in the lower temperature range (600-570 ºC) with 0.5% N content. It has been established that lower epitaxy temperatures favours nitrogen incorporation in the layers. The N 1s spectra of the samples indicate that nitrogen atoms exist in a single-bonded configuration, the Ga-N bond, and interstitial nitrogen complexes is not observed, in contrast to data of high nitrogen content GaAsN samples where the additional nitrogen complex associated peak is recorded (Spruytte at all., 2001).

Fig. 5.5. XPS spectra of two GaAsN samples with different N content.

Fig. 5.6. FTIR spectrum of as grown GaAsN sample.

FTIR absorption spectra of an as grown GaAs1-xNx layer on a n-GaAs substrate is plotted in Fig. 5.6. A peak at 472.6 cm-1, attributed to a local vibrational mode of nitrogen at arsenic site in GaAs is clearly seen.

### **5.1.2 Electrical characterization**

84 Solar Cells – New Aspects and Solutions

Unlike XRD used for assessing the incorporation of nitrogen in GaAs1-xNx alloys grown by LPE the nitrogen bonding configurations and local atomic structures have been studied using x-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) spectroscopy. The XPS spectra have been measured over a range of binding energies from 1 to 550 eV. The X-ray photoelectron spectra of N 1s photoelectron and Ga LMM Auger lines recorded from the as grown GaAs1-xNx samples prepared in different temperature ranges are shown in Fig 5.5. It is clearly seen the Ga Auger peak around 391 eV and the N 1*s* level photoemission peak of the samples. The variation of the intensity of the N 1*s* peak with respect to the Ga LMM peaks reflects is due to the different nitrogen content of the samples. Sample grown from higher initial epitaxy temperature of 650 ºC contains 0.2% N and exhibits a N 1s peak with lower intensity and lower binding energy in comparison with the N 1s peak intensity of the sample grown in the lower temperature range (600-570 ºC) with 0.5% N content. It has been established that lower epitaxy temperatures favours nitrogen incorporation in the layers. The N 1s spectra of the samples indicate that nitrogen atoms exist in a single-bonded configuration, the Ga-N bond, and interstitial nitrogen complexes is not observed, in contrast to data of high nitrogen content GaAsN samples where the

additional nitrogen complex associated peak is recorded (Spruytte at all., 2001).

393.1

Fig. 5.5. XPS spectra of two GaAsN samples with different N content.

XPS Intensity, arb.units

Absorbance, arb. units

Fig. 5.6. FTIR spectrum of as grown GaAsN sample.

388 390 392 394 396 398 400 402 404 406

420 430 440 450 460 470 480 490 500 510 Wavenumber, cm-1

bonding energy, eV

397.6

397.3

0.2% N

0.5% N

Electrical parameters of undoped GaAs and GaAsN layers with different nitrogen content grown on seminsulating (001) GaAs substrates are measured in the temperature range 80 – 300 K using van der Pauw geometry.

Figure 5.7. shows the temperature dependence of the Hall-concentration *nH* on reciprocal temperature for two layers GaAsN with nitrogen concentration of 0.2% and 0.5%, respectively in comparison with undoped GaAs. It is seen that all samples are of n-type and for layers containing nitrogen electron concentration increases about one order of magnitude. This could be explained by the assumption that nitrogen behaves mainly as an isoelectronic donor, which arises from the local heterojunction scheme GaAs-GaN according to Belliache (Bellaiche et al., 1997). The results shown in figure indicate that the free carrier concentration increases strongly with the N concentration. The increase in *nH* has also been observed in GaNxAs1−x doped with S (Yu et al., 2000a) and in Ga1−3x In3xNxAs1−x alloys doped with Se (Skierbiszewski at al., 2000). This large increase of the free electron concentration can be quantitatively explained by a combination of the band anticrossing model (Shan et al, 1999) and the amphoteric defect model (Walukiewicz, 1989). The later suggests that the maximum free carrier concentration in a semiconductor is determined by the Fermi energy with respect to the Fermi-level stabilization energy EFS which is a constant for III-V semiconductors. Since the position of the valence band in GaAsN is independent of N concentration, the giant downward shift of the conduction band edge toward EFS and the enhancement of the density of states effective mass in GaAsN lead to much larger concentration of uncompensated, electrically active donors for the same location of the Fermi energy relative to EFS. In order to explain the large enhancement of the doping limits in dilute nitride alloys both the effects of band gap reduction and the increase in the effective mass have to be taken into account (Yu et al., 2000 b; Skierbiszewski at al., 2000).

Fig. 5.7. Free carrier concentration as a function of inverse temperature for as grown GaAs, and two GaAsN layers with different N content

Dilute Nitride GaAsN and InGaAsN Layers Grown by Low-Temperature Liquid-Phase Epitaxy 87

32,4 32,6 32,8 33,0 33,2 33,4 33,6

Omega, degree

The local structure of the InGaAsN is defined by Raman spectrum. In the Raman spectrum, presented in fig. 5.10. does not observed N-induced local mode LO2 , assigned to the vibration of isolated nitrogen atom bonded to four Ga neighbors (NAsGa4). Instead of this two LVM peaks at 454 and 490 cm-1 originated from In-N bonds in a local Ga3In1-N or Ga2In2-N configurations are appeared. Similar LVM peaks have been reported in the literature for as grown InGaAsN layers by MBE (Mintairov et al., 2001, Hashimoto et al., 2003) and for some MBE and MOCVD samples after annealing (Pavelescu et al.,2005; Kurtz et al. 2001). The experimentally observed local modes could be explained by theoretical analyses of the microscopic lattice structures related to the incorporation of N in InGaAsN alloys. The Monte Carlo simulation (Kim & Zunger, 2001) reveal that in InGaAsN quaternary alloys the "small atom–large atom" bond configuration i.e. "large cation-small anion" In-N + "small cation-large anion" Ga-As is preffered for better lattice-matched of the alloy to GaAs substrate, because introduces less strain. On the other hand, the cohesive energies of GaN is larger than that of InN, so the highly strained Ga-N + In-As configuration is preferred in terms of bond energy. In LPE growth under near to equilibrium conditions In-N bonds are more favorable since they reduce the sum of local strain plus chemical bond energies. The introduction of In changes N environment by formation short-range-ordered nitrogen centered N-InnGa4-n (0 ≤ *n*≤ 4) clusters in InGaAsN alloy. In Ga-rich InGaAsN quaternary the most probably realized are the nearest –neighbor pair defects NAs-InGa in which one of the Ga atom in the neighborhood of N is replaced by a large size heavier InGa (NAsInGaGa3) and also a formation of a second nearest-neighbor complex NAsInGa(2)Ga2 where two of four Ga atoms is replaced by two large-site and heavier InGa in the vicinity of NAs. The calculations using Green's function technique (Talwar, 2007) relieve the splitting of a triple degenerate NAs near to 471 cm-1 into a non-degenerate LVM ~ 462 cm-1 and a double degenerate LVM at 490cm-1 for the nearest –neighbor complex and three bands near to 481, 457, and 429 for second nearest-neighbor complex . The surface roughness of the samples has been examined by atomic force microscopy (AFM). A three-dimensional AFM image of an as grown 1.3 μm-thick InGaAsN layer is presented in Fig. 5.11. The measured root-mean-

100 Intensity, a. u.

Fig. 5.9. XRD rocking curves of InGaAsN sample

square (RMS) roughness on 1-micron area is 0.42 nm.

Figure 5.8. presents the temperature dependencies of the Hall-mobility for the same samples. The mobility of the dilute GaAsN samples is considerably lower due to space charge scattering contributions induced by N-related defects added to well-known scattering mechanisms such as phonon and ionized impurity scattering. The mobility maximums of both curves are almost at the same temperature with a relatively small difference of about 20K, which is an indication for scattering specificity. It is seen a well expressed low-temperature mobility decrease which could be explained by the temperature dependence of the GaAs conduction band edge energy, which is closer to the N defect levels at lower temperatures, increasing the scattering cross-section.

Fig. 5.8. Temperature dependence of Hall electron mobility for GaAs (full squares), and two GaAsN with: 0.2%N (full circles); 0.5%N (full triangles)

The mobility values of the dilute GaAsN samples is lower than those of the undoped GaAs layer but considerable higher than mobility values obtained in n-type GaAsN films with similar free electron concentration grown by MOCVD and MBE.

### **5.2 Growth and characterization of InGaAsN layers**

Dilute InGaAsN layers have been prepared using the same technique as for GaAsN growth. A series of nearly-lattice matched InGaAsN epilayers 1.3-1.5 μm thick have been grown from In-rich solution containing 1.5 at.% polycrystalline GaN as a nitrogen source in the temperature range 615 – 580 oC at a cooling rate 0.6 oC/min.

### **5.2.1 Structural characterization**

Typical XRD rocking curves for grown layers are plotted in the Fig. 5.9.

Two prominent peaks associated with the GaAs substrate and the quaternary InGaAsN layer are observed. The lattice mismatch ∆*a*/*a*o determined from the XRD spectrum is ~ 0.1%. The *In*-concentration of the layers measured separately by X-ray microanalyses is 6.4%. Using Vegard's law the *N*- content in InGaAsN layers is determined to be 2.8%

Fig. 5.9. XRD rocking curves of InGaAsN sample

Figure 5.8. presents the temperature dependencies of the Hall-mobility for the same samples. The mobility of the dilute GaAsN samples is considerably lower due to space charge scattering contributions induced by N-related defects added to well-known scattering mechanisms such as phonon and ionized impurity scattering. The mobility maximums of both curves are almost at the same temperature with a relatively small difference of about 20K, which is an indication for scattering specificity. It is seen a well expressed low-temperature mobility decrease which could be explained by the temperature dependence of the GaAs conduction band edge energy, which is closer to the N defect levels

50 100 150 200 250 300

Fig. 5.8. Temperature dependence of Hall electron mobility for GaAs (full squares), and two

The mobility values of the dilute GaAsN samples is lower than those of the undoped GaAs layer but considerable higher than mobility values obtained in n-type GaAsN films with

Dilute InGaAsN layers have been prepared using the same technique as for GaAsN growth. A series of nearly-lattice matched InGaAsN epilayers 1.3-1.5 μm thick have been grown from In-rich solution containing 1.5 at.% polycrystalline GaN as a nitrogen source in the

Two prominent peaks associated with the GaAs substrate and the quaternary InGaAsN layer are observed. The lattice mismatch ∆*a*/*a*o determined from the XRD spectrum is ~ 0.1%. The *In*-concentration of the layers measured separately by X-ray microanalyses is

6.4%. Using Vegard's law the *N*- content in InGaAsN layers is determined to be 2.8%

T, K

0% N 0.2% N 0.5% N

at lower temperatures, increasing the scattering cross-section.

GaAsN with: 0.2%N (full circles); 0.5%N (full triangles)

**5.2 Growth and characterization of InGaAsN layers** 

**5.2.1 Structural characterization** 

temperature range 615 – 580 oC at a cooling rate 0.6 oC/min.

Typical XRD rocking curves for grown layers are plotted in the Fig. 5.9.

similar free electron concentration grown by MOCVD and MBE.

Hall mobility, cm2/ V.s

The local structure of the InGaAsN is defined by Raman spectrum. In the Raman spectrum, presented in fig. 5.10. does not observed N-induced local mode LO2 , assigned to the vibration of isolated nitrogen atom bonded to four Ga neighbors (NAsGa4). Instead of this two LVM peaks at 454 and 490 cm-1 originated from In-N bonds in a local Ga3In1-N or Ga2In2-N configurations are appeared. Similar LVM peaks have been reported in the literature for as grown InGaAsN layers by MBE (Mintairov et al., 2001, Hashimoto et al., 2003) and for some MBE and MOCVD samples after annealing (Pavelescu et al.,2005; Kurtz et al. 2001). The experimentally observed local modes could be explained by theoretical analyses of the microscopic lattice structures related to the incorporation of N in InGaAsN alloys. The Monte Carlo simulation (Kim & Zunger, 2001) reveal that in InGaAsN quaternary alloys the "small atom–large atom" bond configuration i.e. "large cation-small anion" In-N + "small cation-large anion" Ga-As is preffered for better lattice-matched of the alloy to GaAs substrate, because introduces less strain. On the other hand, the cohesive energies of GaN is larger than that of InN, so the highly strained Ga-N + In-As configuration is preferred in terms of bond energy. In LPE growth under near to equilibrium conditions In-N bonds are more favorable since they reduce the sum of local strain plus chemical bond energies. The introduction of In changes N environment by formation short-range-ordered nitrogen centered N-InnGa4-n (0 ≤ *n*≤ 4) clusters in InGaAsN alloy. In Ga-rich InGaAsN quaternary the most probably realized are the nearest –neighbor pair defects NAs-InGa in which one of the Ga atom in the neighborhood of N is replaced by a large size heavier InGa (NAsInGaGa3) and also a formation of a second nearest-neighbor complex NAsInGa(2)Ga2 where two of four Ga atoms is replaced by two large-site and heavier InGa in the vicinity of NAs. The calculations using Green's function technique (Talwar, 2007) relieve the splitting of a triple degenerate NAs near to 471 cm-1 into a non-degenerate LVM ~ 462 cm-1 and a double degenerate LVM at 490cm-1 for the nearest –neighbor complex and three bands near to 481, 457, and 429 for second nearest-neighbor complex . The surface roughness of the samples has been examined by atomic force microscopy (AFM). A three-dimensional AFM image of an as grown 1.3 μm-thick InGaAsN layer is presented in Fig. 5.11. The measured root-meansquare (RMS) roughness on 1-micron area is 0.42 nm.

Dilute Nitride GaAsN and InGaAsN Layers Grown by Low-Temperature Liquid-Phase Epitaxy 89

3 6 9 12 15

Fig. 5.12. Temperature dependence of free carrier concentrations for undoped InGaAs and

The temperature dependence of Hall mobility for undoped InGaAs and InGaAsN layers grown from In-rich solution is similar to those for the GaAsN layers grown from Ga-rich

50 100 150 200 250 300

Fig. 5.13. Hall mobility as a function of temperature for undoped InGaAs sample

T, K

50 100 150 200 250 300

Fig. 5.14. Hall mobility as a function of temperature for lattice matched InGaAsN sample

T, K

1000/T, K-1

 InGaAs InGaAsN

InGaAs

InGaAsN

2x10<sup>17</sup>

solution as it is shown in Figures 5. 13 and 5. 14.

Hall mobility, cm2/ V.s

2100

2150

2200

Hall mobility, cm2/ V.s

2250

2300

Hall concentrations,cm-3

lattice matched InGaAsN samples

5x10<sup>18</sup>

Fig. 5.10. Raman spectrum of as grown InGaAsN layer.

Fig. 5.11. AFM image of the surface of as grown InGaAsN.

### **5.2.2 Electrical characterization**

In the Fig. 5.12 are plotted the temperature dependence of Hall concentrations *nH* for lattice matched InGaAsN in comparison with a metamorphic InGaAs layer. For undoped InGaAs, *nH* decreases linearity in the explored temperature range, 80 to 300K, typical for slightly degenerate III-V semiconductors. However, for N-containing films, two distinct temperature regimes with different temperature dependence of *nH* are observed. The Hall electron concentration decreases as the temperature decreases down to about 200 K, indicating the presence of thermally activated deep donor levels within the dilute nitride bandgap. The saturation of nH at low temperature (T < 200K) is attributed to fully ionized shallow donors. This behavior could be explained by the presence of two donor levels in the InGaAsN bandgap, one being a shallow N isoelectronic donor and the second a thermally activated deeper donor, presumably N-related deep-level defects typically associated with different N-N pair and N-cluster states ( Zhang & Wei, 2001).

**Raman Intensity, a.u.** 

Fig. 5.10. Raman spectrum of as grown InGaAsN layer.

Fig. 5.11. AFM image of the surface of as grown InGaAsN.

N-N pair and N-cluster states ( Zhang & Wei, 2001).

**5.2.2 Electrical characterization** 

**LO()+LA(X)**

400 440 480 520

**LVM**

**Raman Shift, cm-1**

In the Fig. 5.12 are plotted the temperature dependence of Hall concentrations *nH* for lattice matched InGaAsN in comparison with a metamorphic InGaAs layer. For undoped InGaAs, *nH* decreases linearity in the explored temperature range, 80 to 300K, typical for slightly degenerate III-V semiconductors. However, for N-containing films, two distinct temperature regimes with different temperature dependence of *nH* are observed. The Hall electron concentration decreases as the temperature decreases down to about 200 K, indicating the presence of thermally activated deep donor levels within the dilute nitride bandgap. The saturation of nH at low temperature (T < 200K) is attributed to fully ionized shallow donors. This behavior could be explained by the presence of two donor levels in the InGaAsN bandgap, one being a shallow N isoelectronic donor and the second a thermally activated deeper donor, presumably N-related deep-level defects typically associated with different

Fig. 5.12. Temperature dependence of free carrier concentrations for undoped InGaAs and lattice matched InGaAsN samples

The temperature dependence of Hall mobility for undoped InGaAs and InGaAsN layers grown from In-rich solution is similar to those for the GaAsN layers grown from Ga-rich solution as it is shown in Figures 5. 13 and 5. 14.

Fig. 5.13. Hall mobility as a function of temperature for undoped InGaAs sample

Fig. 5.14. Hall mobility as a function of temperature for lattice matched InGaAsN sample

Dilute Nitride GaAsN and InGaAsN Layers Grown by Low-Temperature Liquid-Phase Epitaxy 91

Algora C., Oritz E., Rey-Stolle I, Diaz V., Pena R., Andreev V. M., Khvostikov V. P.,

Andreev V. M., Kazantsev A. B., Khvostikov V. P., Paleeva E. V., Rumyantsev V. D. and

Andreev V. M., Khvostikov V. P., Larionov V. R., Rumyantsev V. D., Paleeva E. V., Shvarts

Arsent'ev I.N., Bert N.A., Vasil'ev A.V., Garbuzov D.Z., Zhuravkevich E.V., Konnikov S.G.,

Ayers J.E., Schowalter L.J., and S.K. Ghandhi (1992). Post-growth thermal annealing of GaAs

Bauer, E . (1958). Phaenomenologische Theorie der Kristallabscheidung an Oberflaechen,

Bellaiche L., S. H. Wei, Zunger A., (1997) Band gaps of GaPN and GaAsN alloys, *Appl. Phys.* 

Chen N. F., Wang Y., He H, and Lin L. (1996). Effects of point defects on lattice parameters

Dhar S., Halder N., Mondal A., Bansal B., Arora B. M. (2005). Detailed studies on the origin

epitaxy, *Semiconductor Science and Technology*, Vol. 20, No. 12. pp. 1168-1172. Fahy S., O'Reilly E. P. (2004), Theory of electron mobility in dilute nitride semiconductors, *Physica E*: *Low-dimensional systems and nanostructures*, Vol. 21, No 1-2, pp. 881-885. Frank F.C. and Van der Merwe J.H. (1949). One-dimensional dislocations. II. Misfitting

Grabow M.H. and Gilmer,G.H. (1988). Thin film growth modes, wetting and cluster

Guter W., Schöne J., Philipps S., Steiner M., Siefer G., Wekkeli A., Welser E, Oliva E, Bett A.

Johnston S. W., Kurtz S. R., Friedman D. J., Ptak A. J., Ahrenkiel R. K., and Crandall R. S.

Hashimoto A., Kitano T., Nguyen A.K., Masuda A., Yamamoto A., Tanaka S., Takahashi M.,

*Lett.*, Vol.16, No 5, pp 339–340.

*Tekh. Poluprovodn*. Vol. 33, N0 9, pp 1070-1072.

*Zeitschrift für Kristallographie*, Vol. 110, pp 372-394.

nucleation, *Surf. Science*, Vol. 194, No 3, pp 333-346

*Phys. Lett.*, Vol 14, No 4, pp 264–266.

*Lett*., Vol. 70, No 26, pp.3558-3560.

August 15, 1949), pp 216 -225.

0018-9383.

p.130-135.

No1-2, pp. 329-335.

8521.

023504.

072109.

lasers with uncoated mirrors prepared by liquid-phase epitaxy, *Sov. Tech. Phys.* 

Rumyantsev V. D.(2001). A GaAs solar cell with an efficiency of 26.2% at 1000 suns and 25.0% at 2000 suns, *IEEE Trans. on Elec. Dev.*, Vol. 48, No 5, pp 840-844. ISSN:

Sorokina S. V. (1996). Quantum-well AlGaAs heterostructures grown by lowtemperature liquid-phase epitaxy, *Materials Chemistry and Physics*, Vol. 45, No 2, p

M. Z.(1999). High-efficiency AlGaAs/GaAs concentrator (2500 suns) solar cells *Fiz.* 

Kosogov A.O., Kochergin A.V., Faleev N.N. and Flaks L.I.(1988). Periodic multilayer In-Ga-As-P structures fabricated by liquid-phase epitaxy, *Sov. Tech.* 

on Si(001) grown by organometallic vapor phase epitaxy, *J. Cryst. Growth*, Vol. 125,

of semiconductors, *Physical Review B*, Vol. 54, No 12, (September 1996), pp 8516-

of nitrogen-related electron traps in dilute GaAsN layers grown by liquid phase

monolayers and oriented overgrowth, *Proc. R. Soc. London A*, Vol. 198, No 1053, (

and Dimroth F.(2009). Current-matched triple-junction solar cell reaching 41.1% conversion efficiency under concentrated sunlight, *Appl. Phys. Lett*. Vol. 94, No 2,

(2005). Observed trapping of minority-carrier electrons in *p*-type GaAsN during deep-level transient spectroscopy measurement, Appl. Phys. Lett. Vol 86, No 7,

Moto A., Tanabe T., Takagishi S., Raman characterization of lattice-matched

The mobility of the metamorphic InGaAs structure is low, down to 500 cm2/V.s, since it possibly contains threading dislocations of high density and the latter causes relatively poor material quality. High values over 2000 cm2/ V.s for Hall mobility exhibits the lattice matched to GaAs substrate InGaAsN sample. These values are about the theoretical limit predicted by Fahy and O'Reilly (Fahy and O'Reilly, 2004) and among the highest reported for lattice matched thick InGaAsN layers.
