Designing high-performance infrared optoelectronic materials: indium-site substitution in LiInSe₂ with Al, Ga, Sn, and Sb

Abstract

In this study, we employed first-principles calculations based on density functional theory to systematically evaluate the impact of substituting Al, Ga, Sn and Sb atoms into LiInSe₂ crystals with R3m and I4₁/amd space groups. By adjusting the doping ratios of these elements, we analyzed their effects on the optoelectronic properties of LiInSe₂. The results reveal that Al doping reduces the formation energy in the I4₁/amd structure, indicating easier synthesis under conventional conditions. Moreover, Al incorporation increases the bandgap, thereby raising the excitation energy required for electronic transitions. In contrast, Ga, Sn and Sb dopants tend to increase the formation energy while narrowing the bandgap. Further analysis identified four effective doping pathways in the R3m structure, all exhibiting potential for enhanced optoelectronic performance. Notably, Sb doping-despite its higher formation energy and reduced bandgap compared to the intrinsic structure-significantly enhances the optical absorption response in 1.65–3.00 eV. The structural modifications induced by Al, Ga, and Sb doping contribute to improved crystal stability and broadened spectral response, underscoring the strong potential of these materials for infrared detector applications.

---

1 Introduction

Infrared detectors, as core optoelectronic devices, are widely used in fields such as military applications, medical diagnostics, and environmental monitoring. Current mainstream infrared detectors can be categorized into two types: cooled detectors (e.g., InSb, HgCdTe(MCT), and Type-II superlattice detectors(T2SL)) and uncooled detectors (e.g., vanadium oxide and amorphous silicon). Each type presents its own limitations in performance. Although InSb-based detectors are mature in applications, their narrow bandgap, fixed wavelength response, and low operating temperature hinder their suitability for engineering applications.¹ MCT detectors suffer from issues such as poor material uniformity, unstable interfaces, and complex epitaxial growth and post-processing.^2–5 T2SL offer advantages such as tunable bandgaps, low Auger recombination rates, large effective electron masses, and good material uniformity; however, their actual device performance has not yet met theoretical expectations.^6–11 Although uncooled infrared detectors are cost-effective, compact, and lightweight,^12,13 and thus attractive for low-performance applications, they are still inferior to cooled detectors in terms of sensitivity, noise level, and detection range.^14,15 To overcome the bottlenecks in traditional infrared materials-such as poor bandgap tunability, limited device stability,¹⁶ and complex fabrication processes-it is urgent to explore novel infrared detection materials with outstanding optoelectronic performance. Given the potential applications of chalcopyrite-type compounds in various infrared domains,¹⁷ we hypothesize that LiInSe₂ crystals could serve as promising alternative materials for infrared detection.^18–21 Therefore, this work adopts a theoretical design approach, utilizing elemental substitution and doping to investigate and optimize the LiInSe₂ system for infrared detection applications. Owing to its unique electronic, thermal, and structural characteristics,^22,23 LiInSe₂ exhibits excellent optical absorption properties. It offers high transmittance in commonly used infrared bands,²⁴ a broad transparency window, significant birefringence, a high laser damage threshold, and a low two-photon absorption coefficient,²⁵ making it a highly regarded candidate in the field of infrared detection.^26,27 Nevertheless, reports on LiInSe₂ remain scarce in the literature,²⁸ likely due to its electronic structure instability, limited optical performance, and intrinsic defects that cause optical losses and lead to laser-induced damage,^29–31 which impede its further application in infrared optoelectronic devices.³² Moreover, the difficulty in obtaining indium³³ restricts its large-scale production. Still, alternative strategies such as forming heterojunctions, tuning the elemental composition, and introducing foreign dopants³⁴ remain viable avenues for exploration. Since there are few existing studies on doping in LiInSe₂,³⁵ we selected several metal atoms for substitute doping in the LiInSe₂ crystal structure. Database analysis suggests that the introduction of Ga and Al could enhance structural stability and photoelectric conversion efficiency;^36,37 Sb doping may broaden the infrared transparency range and enhance nonlinear optical effects;³⁸ Sn doping is expected to regulate carrier concentration³⁹ and potentially improve optoelectronic properties.

1.1 Why Sn behaves differently

While Sn is often invoked to tune carrier concentration, our calculations reveal that In → Sn substitution in LiInSe₂ tends to collapse the band gap and yield semi-metallic states. This behavior originates from (i) the valence mismatch between Sn^2+/4+ and In³⁺ that promotes compensating defects or off-stoichiometry, and (ii) strong hybridization between Sn-5s/5p and Se-4p states that shifts the band edges and introduces band crossings. As a result, even when the formation energy can be small or negative in R3m for certain loadings, the resulting electronic structure is not suitable for MWIR photodetectors that require a finite gap. This motivates a focused discussion on Sn in the Results section. A key factor in the practical use of LiInSe₂ lies in its good carrier mobility, which supports efficient photoelectric conversion. Based on this, we conducted further research to reveal how various metal dopants influence its electronic structure and optical properties. Research on the optoelectronic properties of LiInSe₂ generally focuses on the I4₁/amd and R3m space groups, as they are more stable and easier to control under standard conditions. In our design, Al, Ga, Sb, and Sn atoms were introduced to partially replace atoms in LiInSe₂, resulting in four new compound types: LiAlInSe₂, LiGaInSe₂, LiSbInSe₂ and LiSnInSe₂. In this study, we used DFT-based first-principles calculations to investigate the intrinsic structures of LiInSe₂ under both space groups, as well as the doped structures with Al, Ga, Sb, and Sn. After structural optimization, we calculated properties such as formation energy, bandgap, density of states, and optical absorption. The results show that in the I4₁/amd space group, Al doping reduces the formation energy and thus improves structural stability, while Ga, Sb and Sn doping increase the formation energy. In the R3m space group, the formation energy of the Al-doped structure is −0.9938 eV, the lowest among all configurations, indicating the most stable structure. Sb doping follows, and Ga doping yields a positive but relatively small formation energy. In terms of bandgap, the intrinsic I4₁/amd structure has a relatively small bandgap of 0.3957 eV. Al doping increases the bandgap, whereas Ga, Sb and Sn doping decrease it. Sb and Sn doping even lead to gap closure, indicating that while Al doping enhances structural stability, it also raises the excitation energy required for electronic transitions. For the R3m space group, the intrinsic bandgap is 0.8283 eV. Doping with Al, Ga, and Sb (in two configurations) reduces the bandgap, which may be beneficial for improving optoelectronic performance. Regarding optical absorption, Al doping enhances the absorption of the I4₁/amd structure in the 0.1–0.8 eV range. In the R3m structure, Sb doping significantly boosts absorption in the 1.65–3 eV range, thereby expanding the potential application range of LiInSe₂.

2 Computational details

All calculations were based on the DFT method and employed the Projector Augmented-Wave (PAW) method. The electron exchange–correlation interaction was described using the Generalized Gradient Approximation (GGA) of the Perdew–Burke–Ernzerhof (PBE) functional. The plane-wave kinetic energy cutoff was set to 1.3 times the maximum ENMAX value among all pseudopotentials used in the system, and was validated through convergence tests to ensure the reliability of the computational results. The convergence criterion for structural optimization was set such that the force on each atom was less than 1 × 10⁻⁶ eV Å⁻¹. For k-point sampling, a Gamma-centered 4 × 4 × 4 uniform k-point mesh was used for structural optimization, total energy calculations, and density of states (DOS) to ensure efficient sampling of the Brillouin zone. The band structure calculations adopted the high-symmetry path k-point scheme, sampling along high-symmetry points in the Brillouin zone to obtain a clear band structure.⁴⁰ The carrier effective mass was computed using the finite-difference method, with its numerical parameters validated through convergence testing. The selected k-point step size (Δk = 0.015 Å⁻¹) ensures robust numerical stability of the results.

The I4₁/amd space group of LiInSe₂ corresponds to a layered tetragonal system, while the R3m space group belongs to the rhombohedral system. Their structure is shown in Fig. 1. As shown in Tables 1 and 2, after optimization, the lattice constants were determined to be a = b = 5.6589 Å, c = 11.3044 Å, α = β = γ = 90°, and a = 3.9588 Å, b = 3.4284 Å, c = 19.5983 Å, α = β = 90°, γ = 120°, respectively. To visualize the coordination of LiInSe₂ atoms, a 1 × 1 × 2 supercell was constructed, comprising 16 atoms (Li₄In₄Se₈) in the I4₁/amd space group and 12 atoms (Li₃In₃Se₆) in the R3m space group, both adopting a chalcopyrite structure.⁴¹

Fig. 1 (a) Structural models of LiInSe₂ with space groups I4₁/amd and R3m, along with the corresponding doped structures incorporating Al/Ga/Sn/Sb elements. The models labeled as origin represent the pristine structures, while those labeled as doped correspond to the four doped configurations. (b) Screening process of the doped structures used in this work.

Table 1 Lattice constants (a, b, c), electron effective mass (m_e), and hole effective mass (m_h) of pristine and doped LiInSe₂ with space group I4₁/amd. The “—” indicates structures that did not meet the screening criteria and thus were not calculated further

I41/amd	a (Å)	b (Å)	c (Å)	me	mh
LiIn1/2Al1/2Se2	5.5261	5.4640	11.3246	0.334	0.903
LiIn1/2Ga1/2Se2	5.5583	5.5807	11.1850	0.675	0.889
LiIn1/2Sn1/2Se2	5.6310	5.6204	11.3881	—	—
LiIn1/2Sb1/2Se2	5.6439	5.6368	11.5232	—	—
LiInSe2	5.6589	5.6589	11.3044	0.376	0.478

---

Table 2 Lattice constants (a, b, c), electron effective mass (m_e), and hole effective mass (m_h) of pristine and doped LiInSe₂ with space group R3m. The “—” indicates structures that did not meet the screening criteria and thus were not calculated further

R3m	a (Å)	b (Å)	c (Å)	me	mh
Liln1/3Al2/3Se2	3.7786	3.2723	19.3063	0.265	0.202
Liln2/3Ga1/3Se2	3.8967	3.3746	19.5002	0.929	0.829
Liln2/3Sn1/3Se2	3.9615	3.4308	19.7025	—	—
LiIn2/3Sb1/3Se2	3.9588	3.4284	19.6052	0.651	0.225
LiInSe2	3.9588	3.4284	19.5983	0.470	0.786

---

In this study, the in atomic content in the LiInSe₂ crystal structure was modified by substituting in atoms with Al, Ga, Sn, and Sb atoms, resulting in the formula LiIn_1−xM_xSe₂, where M represents different dopant atoms.⁴² For the I4₁/amd and R3m space groups, the doping concentrations were set to 50% (I4₁/amd), 33% (R3m), and 66% (R3m), respectively. Meanwhile, six different doping configurations were considered for the I4₁/amd structure (labeled as ab, ac, etc.), and three different doping configurations were considered for the R3m structure (labeled as 1, 2, and 3).

To screen high-l stability. Then, in order to identify semiconductor structures with appropriate bandgap values suitable for MWIR property studies, 23 structures with bandgaps larger than 0.25 eV and smaller than 1.05 eV were further selected. To identify materials with strong optical absorption, 15 structures exhibiting absorption coefficients greater than 10⁴ cm⁻¹ were then screened out. Finally, based on the calculated carrier effective masses, 6 structures were chosen for presentation and detailed analysis in the following study.

3 Results and discussion

The relative stability of all possible doped defect configurations in LiInSe₂ was systematically evaluated through formation energy calculations. For Al-, Ga-, Sn-, and Sb-doped defect models, the formation energy E_form was calculated as follows:⁴³

E_form = E_doped − E_pristine + E_M − E_x (1)

In this equation, E_doped and E_pristine represent the total energies of the doped and pristine systems, respectively. E_M denotes the reference state energy of the doping atom, while E_x corresponds to the reference state energy of the substituted atom.

The sign and magnitude of the formation energy reflect the energetic cost of the doping process and the stability of the doped structure. A negative formation energy indicates a thermodynamically favorable doping process, while a smaller formation energy suggests higher stability of the doped configuration. Fig. 2 illustrates the distribution of formation energies for the four dopants in both I4₁/amd and R3m space groups.

Fig. 2 Formation energies of all structures are represented by scatter plots. Blue squares denote the two pristine structures, yellow triangles indicate configurations meeting the screening criteria, and red triangles represent those failing the criteria. Structures are grouped by space group (I4₁/amd and R3m).

Subsequently, we calculated the bandgap energies of these structures for further screening, and the selected results based on our established criteria are summarized in Table 4.

Our calculations revealed that the electronic and optical properties of different doping sites in the I4₁/amd space group exhibited remarkable similarity. Consequently, we identified six representative structures for in-depth analysis. These include two from the I4₁/amd space group (LiIn_1/2Al_1/2Se₂-bc* and LiIn_1/2Ga_1/2Se₂-ab*) and four from the R3m space group (LiIn_2/3Ga_1/3Se₂-3*, LiSb_1/3In_2/3Se₂-2*, LiSb_1/3In_2/3Se₂-1*, and LiAl_1/3In_2/3Se₂-1*).

3.1 Band structure and band gap analysis

We systematically investigated the electronic structure and bonding characteristics of LiInXSe₂ (X = Al, Ga, Sn, Sb) using the GGA-PBE functional,⁴⁴ which revealed fundamental properties of these materials. Fig. 3 presents the calculated band structures obtained using the PBE functional, comprising six screened doped configurations and two pristine systems.

Fig. 3 Band structure diagrams of the calculated structures for: (a) pristine LiInSe₂ in the I4₁/amd phase, (b) pristine LiInSe₂ in the R3m phase, (c) I4₁/amd with 50% Al substitution at In sites, (d) R3m with 33% Sb doping at Wyckoff position 1, (e) I4₁/amd with 50% Ga substitution at In sites, (f) R3m with 33% Sb doping at Wyckoff position 2, (g) R3m with 33% Ga substitution at In sites, and (h) R3m with 33% Al substitution at In sites. The Fermi level is marked by red dashed lines in all panels.

The band structure calculations reveal that all LiInXSe₂ configurations exhibit distinct semiconductor characteristics: the I4₁/amd phase displays an indirect Γ–Z bandgap of 0.46 eV, while the R3m phase shows a direct Γ–Γ bandgap of 0.83 eV, demonstrating significant phase-dependent electronic structure variations.

Notably, elemental substitution induces significant bandgap modulation effects: Al doping increases the I4₁/amd bandgap to 0.67 eV, while Ga doping reduces the bandgaps to 0.36 eV (I4₁/amd) and 0.58 eV (R3m). This strong contribution of Ga 4p states near the CBM aligns with previous DFT predictions showing bandgap narrowing in Ga-doped LiInSe₂.^45,46 Sb doping decreases the R3m bandgap to 0.52 eV and 0.38 eV.

3.2 Why Sn doping drives semi-metallicity

Our band structure calculations reveal that In → Sn substitution in LiInSe₂ leads to near-gap closure and semi-metallic electronic states. This distinct behavior arises from the aliovalent nature of Sn, where Sn²⁺/Sn⁴⁺ valence mismatch with In³⁺ promotes compensating defects and shifts the Fermi level into conduction states. In addition, the strong hybridization between Sn 5s/5p and Se 4p orbitals introduces band crossings and localized states near the Fermi level. As a result, even when some Sn-doped R3m configurations show relatively low formation energies, as shown in Table 3, their electronic structures are unsuitable for mid-wave infrared detection applications that require a finite bandgap.

Table 3 Formation energies (in eV) of doped and pristine LiInSe₂ configurations in I4₁/amd and R3m space groups

I41/amd	E (eV)	R3m	E (eV)
LiInSe2	−12.3269	LiInSe2	−9.7102
Liln1/2Al1/2Se2-ab	−1.2050	Liln2/3Ga1/3Se2-1	0.4684
Liln1/2Al1/2Se2-ac	−1.1740	Liln2/3Ga1/3Se2-3	0.4763
Liln1/2Al1/2Se2-ad	−1.2050	LiIn2/3Sb1/3Se2-1	−0.2886
Liln1/2Al1/2Se2-bc	−1.2050	LiIn2/3Sb1/3Se2-3	−0.0013
Liln1/2Al1/2Se2-bd	−1.1761	LiIn2/3Al1/3Se2-1	−2.9831
Liln1/2Al1/2Se2-cd	−1.2050	LiIn2/3Al1/3Se2-2	−2.9831
Liln1/2Sn1/2Se2-ab	0.7044	LiIn2/3Al1/3Se2-3	−2.9831
Liln1/2Sn1/2Se2-ac	0.7014	LiIn1/3Al2/3Se2-1	−6.1739
Liln1/2Sn1/2Se2-ad	0.7044	LiIn1/3Al2/3Se2-2	−6.1789
Liln1/2Sn1/2Se2-bc	0.7044	LiIn1/3Al2/3Se2-3	−6.1789
Liln1/2Sn1/2Se2-bd	0.7014	LiIn2/3Sn1/3Se2-1	−2.8425
Liln1/2Sn1/2Se2-cd	0.7044	LiIn2/3Sn1/3Se2-2	−1.9963
Liln1/2Ga1/2Se2-ab	0.5930	LiIn2/3Sn1/3Se2-3	−1.9963
Liln1/2Ga1/2Se2-ac	0.6275	LiIn1/3Sn2/3Se2-1	−4.0279
Liln1/2Ga1/2Se2-ad	0.6102	LiIn1/3Sn2/3Se2-2	−3.1819
Liln1/2Ga1/2Se2-bc	0.6104	LiIn1/3Sn2/3Se2-3	−4.0279
Liln1/2Ga1/2Se2-bd	0.6275
Liln1/2Ga1/2Se2-cd	0.6102
Liln1/2Sb1/2Se2-ab	0.6235
Liln1/2Sb1/2Se2-ac	−1.5303
Liln1/2Sb1/2Se2-ad	−1.5769
Liln1/2Sb1/2Se2-bc	−1.5769
Liln1/2Sb1/2Se2-bd	−1.5303
Liln1/2Sb1/2Se2-cd	−1.5769

---

Table 4 Calculated bandgap energies (in eV) for pristine and doped LiInSe₂ configurations in I4₁/amd and R3m space groups. The doping concentrations are indicated by subscripts (e.g., LiIn_1−xM_xSe₂ with x = 1/2, 1/3, or 2/3). Configurations marked with an asterisk (*) were selected for in-depth analysis

I41/amd	Bandgap (eV)	R3m	Bandgap (eV)
LiInSe2	0.4638	LiInSe2	0.8283
Liln1/2Al1/2Se2-ab	0.5606	Liln2/3Ga1/3Se2-1	0.5763
Liln1/2Al1/2Se2-ac	0.5784	Liln2/3Ga1/3Se2-3*	0.5805
Liln1/2Al1/2Se2-ad	0.5572	LiIn2/3Sb1/3Se2-1*	0.5209
Liln1/2Al1/2Se2-bc*	0.6767	LiIn2/3Sb1/3Se2-2*	0.3764
Liln1/2Al1/2Se2-bd	0.5831	LiIn2/3Al1/3Se2-1*	1.0843
Liln1/2Al1/2Se2-cd	0.5591
Liln1/2Ga1/2Se2-ab*	0.3601
Liln1/2Ga1/2Se2-ac	0.6275
Liln1/2Ga1/2Se2-ad	0.6102
Liln1/2Ga1/2Se2-bc	0.6105
Liln1/2Ga1/2Se2-bd	0.6275
Liln1/2Ga1/2Se2-cd	0.6102

---

We attribute these effects primarily to varying degrees of hybridization between the p-orbitals of dopant elements (Al, Ga, Sb) and Se 4p states. The results demonstrate that Al doping effectively increases the bandgap, while Ga and Sb doping reduces it, with Sb doping at site 2 showing the most significant bandgap reduction effect. The band gap is smaller than that of Li doping substitution, which is conducive to improving the absorption capacity of mid-wave infrared.⁴⁷

3.3 PDOS contribution analysis

Fig. 4 presents the calculated density of states diagrams, revealing that the electronic states between −2 and 3 eV exhibit distinct orbital distribution characteristics. The top valence band region (−1 eV to Fermi level) is dominated by strongly hybridized Se 4p and in 5p orbital (about 70% contribution), with minimal Li-s orbital participation (about 5%), demonstrating the pronounced covalent character of In–Se bonds.⁴⁸ In the lower energy region (−2 to −1 eV), the DOS is primarily composed of Se 4p states with minor contributions from in 5s and Li 2p orbitals, consistent with reported behavior in isostructural LiGaSe₂. The bottom conduction band states mainly consist of Li 2 s and in 5s orbitals, while doping introduces significant Ga 4p/Al 3p contributions (about 35%), confirming their critical role in bandgap modulation. Notably, Ga and Sb orbitals exhibit greater influence near band edges compared to Al, with Sb-doped systems showing localized Sb 5p-derived impurity states near 1 eV that significantly modify the electronic structure. Our calculated redshift in the absorption edge is in line with prior experimental observations on stoichiometry variations in LiInXSe₂, which revealed similar spectral modifications due to intrinsic defects.⁴⁹ While dopant states contribute minimally to the valence band, they participate substantially in the conduction band, with about 35% enhancement in Ga/Al p-orbital contributions particularly highlighting their bandgap engineering capability. These results quantitatively establish the orbital-resolved electronic structure modifications induced by various doping schemes. Compared with the Pna2₁ space group, the high contribution of Se at the −2–0 eV energy level is consistent.⁵⁰

Fig. 4 (a) Presents pristine I4₁/amd structure, (b) pristine R3m structure, (c) I4₁/amd with 50% Al doping, (d) R3m with 33% Sb doping at site 1, (e) I4₁/amd with 50% Ga doping, (f) R3m with 33% Sb doping at site 2, (g) R3m with 33% Ga doping, and (h) R3m with 33% Al doping.

Our analysis reveals significant differences in doping-induced projected density of states(PDOS) variations between the I4₁/amd and R3m space groups of LiInXSe₂, uncovering crucial structure–property relationships. In the I4₁/amd configuration, Al doping causes notable bandgap widening, primarily due to reduced contributions from Al 3p orbitals near the conduction band minimum (CBM) compared to the native atoms. This leads to decreased electronic state density around the Fermi level, enhancing the material's wide-bandgap semiconductor characteristics-this finding fully consistent with band structure calculations showing CBM upshift, suggesting excellent potential for applications requiring larger bandgaps. In contrast, Ga doping exhibits the opposite effect, with its 4p orbitals making substantial contributions near the CBM that significantly increase the Fermi-level state density, thereby facilitating electron excitation to the conduction band and reducing the bandgap. This phenomenon is particularly prominent in the R3m space group, where the enhanced conductivity makes Ga-doped systems especially suitable for applications demanding high electrical conductivity.

Notably, Sb doping induces distinctive electronic structure modifications in both space groups, generating additional impurity states near band edges while simultaneously reducing the bandgap. Our analysis reveals enhanced contributions from Sb 5p orbitals at the valence band maximum compared to native atoms, suggesting improved hole transport properties and potential p-type conductivity. Particularly in the R3m structure, Sb doping causes remarkable alterations in the Fermi-level state density and introduces mid-gap states, which may significantly influence carrier recombination dynamics.

Comparative analysis between space groups demonstrates that the R3m configuration exhibits greater sensitivity to doping, manifested through more pronounced PDOS variations near the Fermi level, while the I4₁/amd structure maintains relatively stable electronic distribution with stronger electron localization characteristics. This feature most prominent in Al-doped systems. These systematic PDOS variations corroborate the band structure calculations, providing crucial doping-specific electronic structure modulation strategies for optimizing the optoelectronic performance and charge transport properties of LiInSe₂.

3.4 Analysis of light absorption capacity and absorption peak

In this study, the absorption peaks of the pristine LiInSe₂–I4₁/amd structure appear at 1.5 eV and 2.8 eV in the x, y direction, and at 1.2 eV and 2.5 eV in the z direction, while the pristine R3m configuration shows absorption peaks at 1.6 eV and 2.5 eV in the x, y direction, and at 1.6 eV in the z direction.

Upon doping, significant changes are observed. As shown in Fig. 5, in the Al-doped I4₁/amd structure, the original 1.5 eV peak in the x, y direction red shifts to 0.9 eV, and the 2.8 eV peak nearly disappears. In the z direction, the 1.2 eV peak also red shifts to 0.9 eV, while the 2.5 eV peak blue shifts to 3.0 eV. For Ga doping in I4₁/amd, the 1.5 eV peak shifts to 2.0 eV and the 2.8 eV peak moves beyond 3.0 eV in the x, y direction, whereas the 1.2 eV peak in the z direction almost vanishes, and the 2.5 eV peak slightly blue shifts to 2.6 eV.

Fig. 5 Optical absorption spectra of all selected structures: (a) 50% Al-doped I4₁/amd phase; (b) 33% Al-doped R3m phase; (c) 50% Ga-doped I4₁/amd phase; (d) 33% Ga-doped R3m phase; (e) 33% Sb-doped R3m phase at site 1; (f) 33% Sb-doped R3m phase at site 2. The insets show magnified views of the 0.1–0.8 eV range.

In the doped R3m structure, Al doping results in the disappearance of most peaks around 1.6 eV and 2.8 eV in both y and z directions. For Ga doping in R3m, the peaks in the x, y direction slightly red shift to 1.5 eV and 2.3 eV, and a similar red shift is observed for the 1.6 eV peak in the z direction. In the Sb-1 doped structure, the 1.6 eV peak in the x, y direction nearly disappears, while the 2.5 eV peak remains largely unchanged; similarly, the 1.6 eV absorption peak in the z direction almost vanishes. In contrast, for the Sb-2 doped structure, the 1.6 eV peak in the x, y direction blue shifts to 1.8 eV, and the 2.5 eV peak slightly shifts to around 2.7 eV, with no significant absorption response observed at 1.6 eV in the z direction.⁵¹

In this study, we analyzed the optical absorption properties of different configurations of LiInSe₂ and its doped structures. In the x, y direction, Al-doped structures did not exhibit a significant enhancement in absorption, while Ga-doped and Sb-doped structures showed noticeable improvements. In the z direction, Al-doped structures presented relatively minor changes, whereas Ga-doped and Sb-doped structures exhibited substantial enhancement in optical absorption.

For the I4₁/amd phase, Al doping caused a red shift of the absorption peaks, enhancing absorption in the low-energy region. This behavior can be attributed to the introduction of new hybrid states near the CBM by Al³⁺ ions, resulting in bandgap narrowing and increased low-energy absorption. In contrast, Ga doping led to a blue shift of the absorption peaks. For the R3m phase, Al doping induced an overall blue shift of the absorption peaks, while Ga doping had minimal effect. Sb doping significantly enhanced absorption in the 1.65–3 eV range.

This behavior is consistent with recent findings by Zhang et al., who demonstrated that suppressing deep-level defects in LiInSe₂ significantly enhances its MIR laser emission efficiency. This may be due to Al³⁺ raising the CBM and lowering the VBM, thereby increasing the optical transition onset. The weak hybridization between Ga 4s and Se 4p orbitals introduced no significant new electronic states. However, compared with the Ga-doped LiGaTe system reported in previous literature,⁵² our Ga-doped structure still exhibits reasonably good absorption performance. In contrast, the Sb 5p orbitals contributed more substantially to both the CBM and VBM, introducing new states and enabling more transitions, thus enhancing light absorption.

In summary, the doped structures generally exhibited improved optical absorption properties. For the I4₁/amd phase, doping significantly enhanced the low-energy absorption performance. For the R3m phase structures with reduced bandgaps, the optical absorption was also notably improved.

We compared the application characteristics of LiInSe₂, InSb, and HgCdTe in the field of infrared detectors. InSb is a narrow-bandgap semiconductor with relatively fixed intrinsic absorption characteristics in the MWIR spectrum (3–5 µm), renowned for its high quantum efficiency and electron mobility.⁵³ Although InSb has made progress in band engineering, such as nBn structures that can raise its operating temperature,⁵⁴ its band properties are fundamentally less tunable than those of doped LiInSe₂. HgCdTe, on the other hand, boasts a unique advantage with a continuously tunable bandgap by altering its composition (x), allowing it to address the infrared detection needs of multiple atmospheric windows in the 1–14 µm range.⁵⁴ However, HgCdTe's drawbacks include material growth instability and the difficulty of eliminating the non-ideal valence band offset, which severely limits its practical performanc.⁵⁵ In summary, LiInSe₂'s unique advantage lies in its optical absorption properties, which can be precisely tuned through doping. For example, doping with Al can cause the absorption peak to red shift to 0.9 eV, while doping with Sb can enhance absorption in the 1.65–3 eV range, providing tunable optical characteristics that could be explored in nonlinear optics and laser-related studies.

4 Conclusion

In this study, the effects of substituting In atoms with four different elements (Al, Ga, Sn, and Sb) on the crystal structure, electronic properties, and optical responses of LiInSe₂ were systematically investigated in two space groups (I4₁/amd and R3m) using first-principles calculations based on density functional theory (DFT). Interestingly, such space-group-dependent behavior was also observed in the LiIn_xAg_1−xSe₂ system, where a gradual transition from I4̄2d to Pna2₁ led to notable variations in optical responses.⁵⁶ Among 48 substitutional structures, six representative structures within specific property ranges were selected to study how atomic doping can modulate the optoelectronic performance of the material. The results show that among the four dopants, Al exhibits relatively low formation energy and increases the band gap while significantly enhancing the optical absorption above 2 eV, making it suitable for optoelectronic applications that require a wide band gap. Compared with AgInSe₂, LiInSe₂ offers superior optical nonlinearity and a broader infrared (IR) transparency window, as confirmed by Reshak and Brik's comparative study.⁵⁷ Sn doping results in a negative band gap, altering the material's semiconducting nature. Ga and Sb doping significantly reduce the band gap, improve carrier transport properties, and exhibit better absorption performance in the mid-wave infrared (MWIR) region. Particularly in the R3m space group, all doped structures except for Sn demonstrate enhanced optoelectronic performance, with Sb doping showing especially remarkable improvements. Through a comprehensive analysis of band structures, density of states, and absorption spectra, this study confirms the effectiveness of doping strategies in optimizing the performance of novel MWIR materials. In conclusion, this work not only provides theoretical support for designing MWIR photodetector materials and expands the potential application scope of LiInSe₂ in optoelectronics but also lays a foundation for future experimental fabrication and device optimization. Further exploration of co-doping schemes and structural optimization strategies may lead to even greater breakthroughs in material performance.

Author contributions

Chen J.: software validation formal analysis investigation data curation writing – original draft writing – review & editing visualization Luo S.: methodology software validation formal analysis investigation data curation writing – original draft writing – review & editing funding acquisition Yang Y.: software validation formal analysis investigation data curation writing – original draft writing – review & editing Zhang Y.: resources writing – review & editing visualization Han S.: resources writing – review & editing visualization Xiong C.: writing – review & editing Cheng Y.: writing – review & editing Cheng C.: writing – review & editing Guan X.: conceptualization, methodology, resources, writing – review & editing, visualization, supervision, project administration Lu P.: conceptualization supervision.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available on reasonable request from the corresponding author (e-mail: E-mail: guanxn@bupt.edu.cn). The data are not publicly available due to privacy and/or ethical restrictions.

Acknowledgements

This work was supported financially by the National Key Research and Development Program Projects of China (No. IRDT-24-02) and Infrared Detection National Key Laboratory Open Project (No. 2022YFB3204203). We are thankful for the helpful discussion with Prof. Pengfei Guan Beijing PARATERA Tech Corp., Ltd and the computational support from the Beijing Computational Science Research Center (CSRC).

References

1. O. Nesher, S. Elkind, A. Adin, I. Nevo, A. Yaakov, S. Raichshtain, A. B. Marhasev, A. Magner, M. Katz, T. Markovitz, et al., Infrared Technology and Applications XXIX, 2003, pp. 120–129.
2. H. Wen and E. Bellotti, J. Appl. Phys., 2016, 119, year.
3. R. Müller, M. Haertelt, J. Niemasz, K. Schwarz, V. Daumer, Y. V. Flores, R. Ostendorf and R. Rehm, Micromachines, 2020, 11, 1124.
4. D. O. Alshahrani, M. Kesaria, E. A. Anyebe, V. Srivastava and D. L. Huffaker, Adv. Photonics Res., 2022, 3, 2100094.
5. A. Dabrowska, M. David, S. Freitag, A. M. Andrews, G. Strasser, B. Hinkov, A. Schwaighofer and B. Lendl, Sens. Actuators, B, 2022, 350, 130873.
6. D. Smith and C. Mailhiot, J. Appl. Phys., 1987, 62, 2545–2548.
7. C. Cervera, I. Ribet-Mohamed, R. Taalat, J. Perez, P. Christol and J. Rodriguez, J. Electron. Mater., 2012, 41, 2714–2718.
8. Y. Chen, L. Lu, B. Zhu, R. Gong, J. Ou, X. Lou, L. Zhu and M. Dong, Opt. Express, 2025, 33, 22132–22144.
9. J. Yu, Y. Fu, L. Lu, W. Chen, J. Ou and L. Zhu, Micromachines, 2025, 16, 575.
10. Z. Guo, W. Wang and P. Zhu, Opt. Eng., 2024, 63, 107104.
11. M. Kopytko and A. Rogalski, IEEE Trans. Electron Devices, 2022, 69, 2992–3002.
12. J. Tissot, C. Trouilleau, B. Fieque, A. Crastes and O. Legras, Opto-Electron. Rev., 2006, 14, 25–32.
13. A. Das, M. L. Mah, J. Hunt and J. J. Talghader, Optica, 2023, 10, 1018–1028.
14. R. S. Balcerak, Infrared Technology and Applications XXV, 1999, vol. 3698, pp. 110–118.
15. M. Shen, X. Zhang, F. Cen, Q. Zhang, H. Zhang, G. Zhang, S. Jiang, Y. Chen, Y. Hu and K. Yao, Adv. Funct. Mater., 2025, 2423055.
16. X. Lyu, Highlights in Science, Engineering and Technology, 2022, vol. 27.
17. S. Das, P. Ranjan, K. Gaurav, P. K. Surolia and T. Chakraborty, Phys. B, 2022, 646, 414305.
18. T.-H. Ma, C.-H. Yang, Y. Xie, L. Sun, W.-Q. Lv, R. Wang, C.-Q. Zhu and M. Wang, Comput. Mater. Sci., 2009, 47, 99–105.
19. K. Wu, B. Zhang, Z. Yang and S. Pan, J. Am. Chem. Soc., 2017, 139, 14885–14888.
20. V. S. Khorkin, M. G. Milkov, S. N. Mantsevich, M. I. Kupreychik, S. I. Lobanov and V. N. Vedenyapin, Opt. Mater., 2024, 157, 116202.
21. L. Isaenko, L. Dong, K. Korzhneva, A. Yelisseyev, S. Lobanov, S. Gromilov, M. S. Molokeev, A. Kurus and Z. Lin, Inorg. Chem., 2023, 62, 15936–15942.
22. L. Isaenko, A. Yelisseyev, S. Lobanov, P. Krinitsin, V. Petrov and J.-J. Zondy, J. Non-Cryst. Solids, 2006, 352, 2439–2443.
23. V. V. Atuchin, L. I. Isaenko, S. I. Lobanov, A. A. Goloshumova, M. S. Molokeev, Z. Zhang, X. Zhang, X. Jiang and Z. Lin, Molecules, 2022, 27, 5078.
24. N. Jia, S. Wang and X. Tao, Acta Phys. Sin., 2018, 67, 244203.
25. L. Sun, C. Yang, T. Ma and C. Zhu, Prog. Chem., 2014, 26, 293.
26. I. Vasilyeva, A. Pochtar and L. Isaenko, J. Solid State Chem., 2014, 220, 91–96.
27. S. Wang, C. Ma, K. Xu, J. Zhang, F. Yang, X. Tao and S. Wang, Laser Photon. Rev., 2024, 18, 2400191.
28. G. Zhao, S. Zhu, B. Zhao, B. Chen, Z. He and S. Wan, J. Cryst. Growth, 2009, 311, 368–372.
29. I. Vasilyeva, R. Nikolaev, V. Malakhov and L. Isaenko, J. Therm. Anal. Calorim., 2007, 90, 601–605.
30. S. Wang, C. Ma, K. Xu, J. Zhang, F. Yang, X. Tao and S. Wang, Laser Photon. Rev., 2024, 18, 2400191.
31. Z. Zheng, H. Yu, M. Zhu, Z. Zhang, Z. Gao, M. Xu, R. Zhang and Y. Xu, Sci. Rep., 2024, 14, 24779.
32. S. Ali, B. U. Khan and T. Alshahrani, Mater. Chem. Phys., 2024, 305, 127847.
33. T. Werner, G. M. Mudd and S. M. Jowitt, Appl. Earth Sci., 2015, 124, 213–226.
34. C. Liu, Y. Yuan, X. Zhang, J. Su, X. Song, H. Ling, Y. Liao, H. Zhang, Y. Zheng and J. Li, Nanomaterials, 2020, 10, 1887.
35. Z. Zheng, H. Yu, M. Zhu, Z. Zhang, Z. Gao, M. Xu, R. Zhang and Y. Xu, Sci. Rep., 2024, 14, 24779.
36. S. Adachi, J. Appl. Phys., 1987, 61, 4869–4876.
37. W. Yan, S. Zhou, Q. Xie, B. Guo, C. Zhang and Z. Zhang, Acta Opt. Sin., 2012, 32, 0516003.
38. J. Han, H. Zhou and Q. Wang, Mater. Lett., 2006, 60, 252.
39. I. Suzuki, J. Ceram. Soc. Jpn., 2023, 131, 777–788.
40. X. Wang, J. Zhao, Z. Xu, F. Djurabekova, M. Rommel, Y. Song and F. Fang, Nanotechnol. Precis. Eng., 2020, 3, 211–217.
41. A. Reshak and M. Brik, J. Alloys Compd., 2016, 675, 355–363.
42. L.-H. Li, J.-Q. Li and L.-M. Wu, J. Solid State Chem., 2008, 181, 2462–2468.
43. S. Wang, Z. Zhang, C. Chen, H. Zhang, S. Shi, J. Xie, Y. Cai, S. Gao, W. Chen and C. Dong, et al., Chem. Eng. J., 2025, 504, 158638.
44. Y. Wang, Y. Zhao, J. Ni and Z. Dai, Adv. Funct. Mater., 2024, 34, 2410983.
45. S. Wang, C. Ma, K. Xu, J. Zhang, F. Yang, X. Tao and S. Wang, Laser Photon. Rev., 2024, 18, 2400191.
46. B. W. Wiggins, PhD Thesis, University of Central Florida, 2016.
47. Y. Zhang, H. Li, X. Wang and J. Liu, J. Alloys Compd., 2025, 950, 170918.
48. M. S. Khan, T. Alshahrani, B. U. Haq, S. Azam, G. Khan, H. Alrobei, Z. Abbas, M. Předota, M. A. Khan and M. Benaadad, J. Solid State Chem., 2021, 293, 121778.
49. S. Wang, C. Ma, K. Xu, J. Zhang, F. Yang, X. Tao and S. Wang, Laser Photon. Rev., 2024, 18, 2400191.
50. S. Wang, C. Ma, K. Xu, J. Zhang, F. Yang, X. Tao and S. Wang, Laser Photon. Rev., 2024, 18, 2400191.
51. L. Isaenko, L. Dong, A. Yelisseyev, S. Lobanov, K. Korzhneva, S. Gromilov, A. Sukhih, A. Pugachev, V. Vedenyapin and A. Kurus, et al., J. Alloys Compd., 2023, 969, 172382.
52. W. Li, D. Chu, Q. Wang, K. Yin, H. Zhang, S. Qu and P. Yao, J. Mater. Chem. C, 2023, 11, 11234–11242.
53. V. Nagarajan and R. Chandiramouli, Alex. Eng. J., 2014, 53, 437–444.
54. Q. Shi, S.-K. Zhang, J.-L. Wang and J.-H. Chu, J. Infrared Millim. Waves, 2022, 41 DOI:10.11972/j.issn.1001-9014.2022.01.010.
55. S. Kurtz, J. Bajaj, D. Edwall and S. Irvine, Semicond. Sci. Technol., 1993, 8, 941.
56. L. Isaenko, A. Yelisseyev, S. Lobanov, V. Vedenyapin, P. Krinitsyn and V. Petrov, Crystals, 2016, 6, 85.
57. A. Reshak and M. Brik, J. Alloys Compd., 2016, 675, 355–363.

---

Footnote

† These authors contributed equally to this work and share first authorship.

---