Li2CdGeSe4 and Li2CdSnSe4: biaxial nonlinear optical materials with strong infrared second-order responses and laser-induced damage thresholds influenced by photoluminescence

Jian-Han Zhangab, Daniel J. Clarkc, Ashley Weilanda, Stanislav S. Stoykoa, Yong Soo Kimd, Joon I. Jangc and Jennifer A. Aitken*a
aDepartment of Chemistry and Biochemistry, Duquesne University, Pittsburgh, PA 15282, USA. E-mail:
bSanming University, School of Resources and Chemical Engineering, Sanming, Fujian, China 365004
cDepartment of Physics, Applied Physics and Astronomy, Binghamton University, Binghamton, NY 13902, USA
dDepartment of Physics and Energy Harvest-Storage Research Center, University of Ulsan, 44610 South Korea

Received 4th January 2017 , Accepted 9th June 2017

First published on 13th June 2017

Two new biaxial, diamond-like semiconductors, Li2CdGeSe4 and Li2CdSnSe4, were prepared via high-temperature, solid-state synthesis. Single crystal X-ray diffraction and X-ray powder diffraction coupled with Rietveld refinement were used to refine the crystal structures and assess the phase purity, respectively. Both compounds adopt the lithium cobalt(II) silicate structure type. Strong second-order nonlinear optical (NLO) susceptibility, phase matchability, relatively high thermal stability, and excellent transparency deem both materials potential infrared (IR) NLO candidates. Li2CdGeSe4 and Li2CdSnSe4 display optical bandgaps of approximately 2.5 and 2.2 eV, respectively. Li2CdSnSe4 exhibits a strong, red-light emission under 1064 nm excitation, allowing the compound to release energy that accumulates by two-photon absorption under Nd:YAG laser radiation. Therefore, Li2CdSnSe4 shows a high laser-induced damage threshold (LIDT) of 0.7 GW cm−2. This special phenomenon is remarkable and may open a new avenue in searching for promising IR NLO materials with large LIDTs.


Infrared (IR) nonlinear optical (NLO) crystals are of great importance in laser technology and are increasingly desired in military, medical, commercial and industrial applications, such as remote sensing of chemical and biological weapons,1 laser communication,2 diagnosis and monitoring of diseases by breath analysis,3 detection of air and water pollutants,4 and industrial process control.5 However, it is particularly challenging to find suitable materials working in the IR region that demonstrate a collective set of desirable attributes over a wide range of wavelengths. In addition to highly efficient second-order nonlinearity, practical NLO materials must possess other critical characteristics, such as high laser-induced damage thresholds (LIDTs) and noncritical phase matching, among others.6 To date, only a few crystals are commercially available for IR NLO applications, such as AgGaS2,7 AgGaSe2[thin space (1/6-em)]8 and ZnGeP2.9 It is interesting to point out that these ternary materials feature a noncentrosymmetric (NCS) diamond-like structure, which is highly favourable to produce a large second-order NLO response. Yet these materials possess some significant drawbacks, such as challenging crystal growth10 and two-photon absorption (2PA), which leads to relatively low LIDTs, limiting their use in high-powered laser applications.11

In the past decades, several strategies have been devised for the pursuit of new candidate NLO materials. One of the widely used approaches is to focus on the synthesis of materials containing building blocks that tend to exhibit acentricity due to phenomena such as second-order Jahn–Teller (SOJT) effects, stereochemical expression of lone pairs,12–16 and conjugated π-orbital systems.17–27 By combining SOJT-distorted transition-metal cations and stereochemically active lone pairs into the same crystalline material, some oxides with high second harmonic generation (SHG) responses have been recently reported, for example, BaTeMo2O9,12 BiO(IO3),13 Li2Ti(IO3)6,14 K(VO)2O2(IO3)3[thin space (1/6-em)]15 and BaNbO(IO3)5.16 In conjugated π-orbital systems, borates play a key role and prolific achievements have been made thus far including the newly discovered compounds Cs3Zn6B9O21,17 Cs2B4SiO9,18 Pb2B5O9I,19 K3B6O10Cl,20 Na2CsBe6B5O15,21 Cd4BiO(BO3)3,22 Pb2(BO3)(NO3),23 Na2Be4B4O11,24 Li4Sr(BO3)2,25 Rb3Al3B3O10F,26 and Ba3(ZnB5O10)PO4.27 However, generally speaking, these oxides are not transparent in the IR region and, therefore, not well suited for IR NLO applications, but rather more appropriate for the ultraviolet and visible regions. Halides and chalcogenides with high transparency in the IR regime are considered ideal materials, and newly reported compounds such as Cs2HgI2Cl2,28 Rb3Ta2AsS11,29 γ-NaAsSe2,30 La4InSbS9,31 RbMn4In5Se12,32 Ba3CsGa5Se10Cl2,33 Ba23Ga8Sb2S38,34 Ba2Ga8MS16 (M = Si, Ge),35 Na2In2MS6 (M = Si, Ge),36 [A3X][Ga3PS8] (A = K, Rb; X = Cl, Br),37 BaGeOSe2,38 (Hg6P3)(In2Cl9), and (Hg8As4)(Bi3Cl13)39 have emerged as strong contenders for IR NLO applications.

Although these avenues have proven productive for the discovery of new NLO materials, it must be mentioned that the success of these tactics is contingent upon the packing of the acentric building blocks in an NCS fashion, which does not always take place. In order to ensure an NCS crystal structure, we focus on the inherently NCS diamond-structured materials and work to optimize the resulting physicochemical properties by implementing changes in the elements residing in specific crystallographic positions. Our plan is to target quaternary diamond-like semiconductors (DLSs) with the general formula I2–II–IV–VI4. Compared to the ternary I–III–VI2 or II–IV–V2 materials, the quaternary DLSs allow the introduction of a more diverse set of ions with differing radii and electronegativity values providing even greater opportunity for the deliberate modification of properties, e.g. bandgap, leading to enhanced nonlinearity. Using this concept, we have previously reported several promising diamond-like compounds, such as Cu2CdSnS4, α/β-Cu2ZnSiS4,40 Li2CdGeS4,41 Li2CdSnS4,42 Li2CoSnS4, Li2MnGeS4,43 Li2ZnGeSe4 and Li2ZnSnSe4.44 Among which, Li2CdGeS4 exhibits a phase-matched SHG efficiency that is the highest among IR NLO materials with bandgaps larger than 3.0 eV.41 Li2MnGeS4 exhibits a huge LIDT, >16 GW cm−2 (λ = 1064 nm; τ = 30 ps) without undergoing any multi-photon absorption (MPA).43 Both Li2ZnGeSe4 and Li2ZnSnSe4 are highly promising IR NLO candidates with strong χ(2) values (∼20 pm V−1) and moderate LIDTs simultaneously.44 Here, we report the synthesis, crystal structures and physicochemical characterization, including the NLO properties, of two new, biaxial quaternary DLSs, Li2CdGeSe4 and Li2CdSnSe4, where the LIDT of the latter is influenced by two-photon-induced photoluminescence (PL). We find that Li2CdSnSe4 has a greater LIDT than commercially available AgGaSe2 and AgGaS2 crystals.



All of the elements obtained from commercial sources were of analytical purity and used without further purification: (1) lithium metal rod (99.9%, Sigma-Aldrich); (2) cadmium powder (99.9%, Strem); (3) germanium pieces (99.999%, Strem) were ground by a diamonite mortar and pestle into a powder prior to use; (4) tin powder (99.99%, Cerac); and (5) selenium powder (99.99%, Strem).


Traditional high-temperature, solid-state reactions were carried out to prepare Li2CdGeSe4 and Li2CdSnSe4. The starting materials were weighed in stoichiometric portions, Li (0.0158 g), Cd (0.1287 g), Ge (0.0831 g) and Se (0.3614 g) for Li2CdGeSe4 and Li (0.0138 g), Cd (0.1124 g), Sn (0.1187 g) and Se (0.3156 g) for Li2CdSnSe4, in an inert atmosphere (argon) glove box. The reactants were inserted in graphite tubes inside 12 mm o.d. fused-silica tubes. The tubes were flame-sealed under a pressure of approximately 10−3 mbar and placed into a programmable furnace. The reactions were heated to 190 °C in 2 h, held at 190 °C for 10 h, and then heated up to 675 °C in 5 h, held at 675 °C for 100 h and slowly ramped down to 350 °C in 100 h. The furnace was then shut off and the samples were allowed to cool to room temperature radiatively. The reaction vessels were opened under ambient conditions. Using an optical microscope, Li2CdGeSe4 was observed as pink-red polyhedral-shaped polycrystals and inspection of Li2CdSnSe4 revealed orange, irregularly shaped crystals. A digital camera was used to obtain optical images, taken directly under the optical microscope, see Fig. 1.
image file: c7qi00004a-f1.tif
Fig. 1 Digital images of Li2CdGeSe4 and Li2CdSnSe4 taken under an optical microscope. Crystal sizes range from tenths to a couple of mm on an edge.

Single crystal X-ray diffraction

A Bruker SMART Apex2 CCD single crystal X-ray diffractometer using graphite monochromatized Mo-Kα radiation (λ = 0.71073 Å) and running with a tube power of 50 kV and 30 mA was used to collect single crystal X-ray diffraction data for 40 seconds per frame at room temperature. The program SAINT45 was used to integrate the data and SADABS46 was utilized to carry out the multiscan absorption correction. Two space groups were indicated as possible, Pnma (no. 62) and Pna21 (no. 33), based on the careful inspection of systematic absences; however, DLSs are inherently NCS due to the alignment of all tetrahedra along one crystallographic direction; therefore, Pna21 was selected. The structures of Li2CdGeSe4 and Li2CdSnSe4 were solved by direct methods and refined by the full-matrix least-squares fitting on F2 by SHELX-97.47 The Flack parameters for Li2CdGeSe4 and Li2CdSnSe4 refined to 0.030(8) and 0.04(2), respectively. Both structures were tested for additional symmetry elements using PLATON, but none were found.48 Except for Li(1) and Li(2) in Li2CdSnSe4 that were refined isotropically, the atomic displacement parameters of all atoms were anisotropically refined. Attempts to collect data on a higher quality crystal of Li2CdSnSe4 were met with great difficulty. While the R factors and residual electron density for the refinement of Li2CdSnSe4 are not as good as those obtained for the germanium analogue, the Rietveld refinement results strongly support the reported structure, as discussed later. Crystallographic data and structure refinement details for both compounds are gathered in Table 1. Atomic coordinates and equivalent isotropic displacement parameters are listed in Table S1. Selected bond distances and angles are provided in Table S2.
Table 1 Crystallographic data and structure refinement details for Li2CdGeSe4 and Li2CdSnSe4
Empirical formula Li2CdGeSe4 Li2CdSnSe4
a R1 = R1 = ∑||Fo| − |Fc||/∑|Fo|; wR2 = {∑w[(Fo)2 − (Fc)2]2/∑w[(Fo)2]2}1/2.
Formula weight 514.71 560.81
Space group, Z Pna21 (no. 33), 4 Pna21 (no. 33), 4
T (K) 293 K 293 K
a (Å) 14.1441(2) 14.4508(3)
b (Å) 8.3046(1) 8.3986(2)
c (Å) 6.6805(1) 6.8032(1)
V3) 784.70(2) 825.68(3)
ρcal (g cm−3) 4.357 4.511
μ (Mo Kα) (mm−1) 24.980 23.125
F(000) 888 960
Reflections collected/unique 10[thin space (1/6-em)]838/1737 10[thin space (1/6-em)]208/1817
Completeness to θ = 27.14° 100.0% 98.7%
GOF on F2 0.948 1.163
Flack parameter 0.030(8) 0.04(2)
R1, wR2 [I > 2σ(I)]a 0.0143, 0.0296 0.0384, 0.0977
R1, wR2 (all data)a 0.0283, 0.0296 0.0437, 0.1023

X-ray powder diffraction (XRPD) and Rietveld refinement

A PANalytical X'Pert Pro MPD X-ray powder diffractometer using Cu-Kα radiation (λ = 1.541871 Å) and operating with a tube power of 45 kV and 40 mA was used to collect XRPD data over an angular range of 5° to 145° 2θ with a step size of 0.0083556° and scan rate of 0.010644° s−1. The incident beam optics consisted of a 0.02 rad soller slit, a 1/4° divergent slit and a 1/2° anti-scatter slit; whereas, the diffracted beam optics incorporated a 0.02 rad soller slit, a 1/4° anti-scatter slit and a nickel filter. The samples were ground under inert atmosphere using an agate mortar and pestle for twenty minutes, and then loaded into a backfilled sample holder for data collection. The X'Pert HighScore Plus software package49 and the International Centre for Diffraction Data (ICDD) powder diffraction file (PDF) database were used for phase identification.50

Rietveld refinements were performed using the GSAS software package with the EXPGUI interface.51 The crystal structures of Li2CdGeSe4 and Li2CdSnSe4 resulting from the single crystal X-ray data sets were used as the starting structure models. The backgrounds were fit with a shifted Chebyschev polynomial,52 and the peak shapes were modelled using Lorentzian terms, the Lorentzian isotropic crystallite size broadening (LX) and Lorentzian isotropic strain broadening (LY), within the type-3 profile function. Lattice parameters (Table S3), atomic coordinates and isotropic displacement parameters were also refined.

Scanning electron microscopy and energy dispersive spectroscopy

Energy dispersive spectroscopy (EDS) was completed using a Hitachi S-3400N scanning electron microscope (SEM) furnished with a Bruker Quantax model 400 energy dispersive spectrometer using an XFlash® 5010 EDS detector with a 129 eV resolution. Data were collected at a working distance of 10 mm and an accelerating voltage of 15 kV for 5 min live time.

Differential thermal analysis

Samples were prepared for differential thermal analysis (DTA) and data were collected as described elsewhere.53 For each sample, a charge of ∼50 mg was sealed in a carbon-coated, fused-silica ampoule and flame-sealed under a vacuum of ∼10−3 mbar. The sample and the reference material, Al2O3, of comparable mass were heated at a rate of 10 °C min−1 from 25 °C to 1000 °C under a constant flow of nitrogen gas, held at the high temperature for 1 min and cooled at a rate of 10 °C min−1 to 100 °C. A second cycle was attempted to determine the reversibility of the thermal events; however, severe attack of the fused-silica ampoule occurred; therefore, only one cycle of data is reported here. We did not observe any attack of the ampoule with only one heating/cooling cycle.

UV-vis-NIR and FT-IR spectroscopy

The optical diffuse reflectance spectra of polycrystalline powder samples were measured at room temperature using a Varian Cary 5000 UV-vis-NIR spectrometer equipped with a Harrick Praying Mantis diffuse reflectance accessory. Powdered samples were loaded into a sample cup. Barium sulphate was used as a 100% reflectance standard in the range of 0.19–2.5 μm. The measurements were carried out using a scan rate of 600 nm min−1. The reflectance spectra were transformed to absorption spectra by implementation of the Kubelka–Munk equation.54

A Nicolet 380 FT-IR spectrophotometer equipped with an attenuated total reflectance accessory was used to collect the IR spectra, comprised of 64 scans and collected over the range of 400 cm−1 to 4000 cm−1. The OMNIC software was used for data collection and analysis. The sample penetration depth is near the lower limit of the particle sizes for the studied samples, ∼2 μm, as a result of the instrument configuration. The instrument employs a diamond crystal in optical contact with the sample. Therefore, the influence of thickness on the intensity of the obtained spectra is insignificant.55

Second-order nonlinear optical (NLO) property measurements

In order to examine the phase matchable (PM) behaviour, polycrystalline Li2CdGeSe4 and Li2CdSnSe4 were prepared for SHG measurements using a series of sieves, resulting in samples of discrete particle size ranges of ∼2–20 μm, 20–45 μm, 45–63 μm, 63–75 μm, 75–90 μm, 90–106 μm, 106–125 μm and 125–150 μm. Each batch of particles was enclosed in quartz capillary tubes. All capillary tubes were flame sealed to prevent exposure of the sample to air and moisture during NLO measurements. Each tube was loaded into a homemade sample holder that was mounted on a Z-scan translation stage. SHG data were collected as a function of particle size using fundamental wavelengths from λ = 2100–3100 nm in steps of 200 nm in order to determine the onset wavelength of phase matchability for the title compounds.

Since both samples were deemed PM, broadband SHG experiments were conducted at room temperature for the largest particle size range. The incident wavelength was tuned from 1100–3700 nm in steps of 200 nm with an input pulse energy of 15 μJ. The corresponding SHG range is λSHG = 550–1850 nm. The relative SHG signals, spectrally resolved in a broad wavelength range, were precisely calibrated with the known and measured efficiencies of all optical components. The SHG efficiencies of the samples were compared to that of a ground commercially-obtained, optical-quality (OQ) single crystal and a “homemade” microcrystalline (MC) specimen of the benchmark NLO material, AgGaSe2 (χ(2) = 66 pm V−1) that was ground to a powder and sieved in the same manner as described above for the samples. The χ(2) values were estimated using SHG responses in the lower energy region, λ = 3300–3700 nm, where both of the compounds and the reference are PM, not undergoing absorption and the SHG responses become essentially static. Additional details regarding the instrument setup can be found elsewhere.41

Using an incident laser with λ = 1064 nm and a pulse width, τ, of 30 ps the LIDTs were assessed for both compounds. The SHG responses were measured as the input laser intensity was varied by over two orders of magnitude. It was confirmed that the thermal load on the samples by the laser pulses tuned below the bandgap was negligible due to its slow repetition rate of 50 Hz.

Photoluminescence (PL) measurements

Two-photon-induced PL experiments were conducted at room temperature. A Nd:YAG laser was used as the excitation source, λ = 1064 nm, with a repetition rate of 50 Hz and a pulse width of 30 ps. This laser also synchronously pumps an EKSPLA PG403 optical parametric oscillator to generate continuously tunable coherent radiation for the broadband SHG measurements.

The excitation beam was focused onto the sample using a positive lens (f = 7.5 cm) and the corresponding spot size was directly measured to estimate the input fluence if required. The PL signal was collected onto a fiber-optic bundle in a reflection geometry with a combination of collection lenses. Time-integrated PL signals of the sample were spectrally resolved through the use of a fiber-optic cable coupled to a selective-grating Horiba HR320 spectrometer equipped with a charge-coupled device camera.

Electronic structure calculations

Electronic band structure and density of states (DOS) of Li2CdGeSe4 and Li2CdSnSe4 were calculated using the total-energy code CASTEP, which uses density functional theory and the plane wave pseudopotential method to solve the one-electron Kohn–Sham equations.56 In order to treat the exchange and correlation effects, the Perdew–Burke–Ernzerhof generalized gradient approximation (PBE-GGA) was used.57 The interactions between the ionic cores and the electrons were described by the norm-conserving pseudopotential.58 The following orbital electrons were treated as valence electrons: Li-2s1, Cd-4d104p65s2, Ge-4s24p2, Sn-5s25p2 and Se-4s24p4. The number of plane waves included in the basis set was determined by a cutoff energy of 600 eV, and the numerical integration of the Brillouin zone was performed using a 3 × 3 × 4 Monkhorst–Pack k-point sampling for both compounds. The distance cut-off for Mülliken bond population analysis was set as 3.5 Å. The other calculating parameters and convergence criteria were the default values of the CASTEP code. All calculations were performed using the structures of Li2CdGeSe4 and Li2CdSnSe4 that were obtained from the room-temperature, single-crystal X-ray structures as reported herein, without volume or geometry optimization.

Results and discussion

Crystal structures

The new, quaternary DLSs, Li2CdGeSe4 and Li2CdSnSe4, adopt the lithium cobalt(II) silicate, Li2CoSiO4, structure-type crystallizing in the NCS space group Pna21. This structure can be considered as a superstructure of lonsdaleite, the rare hexagonal diamond, see Fig. 2.59 While the wurtz-stannite (Pmn21) structure is the most usual structure-type for quaternary DLSs derived from hexagonal diamond, the wurtz-kesterite (Pn) structure is ranked less common and the lithium cobalt(II) silicate (Pna21) structure type is considered relatively rare. The only chalcogenides possessing the lithium cobalt(II) silicate structure-type are Ag2CdGeS4 and Li2MnGeS4, which have been recently reported by our group.43,53 All of these quaternary DLS structure-types derived from lonsdaleite can be described as a hexagonal, closest-packed array of chalcogenide anions with the cations occupying half of the tetrahedral holes. The patterns in which those cations fill the holes dictate the specific structure type. In each case, the placement of the IV–VI4 tetrahedra is identical and the difference lies only in the ordering of the I and II cations within the [IV–VI4]3− molecular framework. In the lithium cobalt(II) silicate structure, the pattern of the I and II cations yields a doubling along the a axis in comparison to the wurtz-stannite and wurtz-kesterite structures, Fig. 2. Since the two title compounds are isostructural, only the structure of Li2CdGeSe4 will be described in detail herein.
image file: c7qi00004a-f2.tif
Fig. 2 The crystal structure of the rare hexagonal diamond, lonsdaleite (left) and the three superstructures, wurtz-stannite (Pmn21), wurtz-kesterite (Pn), and lithium cobalt(II) silicate (Pna21) that are derived thereof and have been observed for quaternary DLSs of the general formula I2–II–IV–VI4, where the Roman numerals indicate the number of valence electrons each element contributes to the compound. The typical elements found in these formulae are as follows: I = Li, Cu or Ag; II = Mn, Fe, Co, Ni, Zn, Cd or Hg; IV = Si, Ge or Sn; and VI = S, Se or Te.

Eight crystallographically independent atoms make up the asymmetric unit of Li2CdGeSe4, one Cd, one Ge, two Li and four Se atoms, all of which are tetrahedrally coordinated and reside on general crystallographic positions, Fig. 3. All of the Se anions are coordinated by one Cd2+, one Ge4+ and two Li+ cations, which is in harmony with Pauling's electrostatic valence principle,60 resulting in local charge neutrality. All of the M–Se bond distances and angles are similar to those catalogued for other DLSs.44,61 The bond valence sum (BVS) calculations produced values of 1.00 and 1.16 for Li(1) and Li(2) respectively, 2.05 for Cd(1) and 3.99 for Ge(1), which are in agreement with the anticipated valence states, Table S4.[thin space (1/6-em)]62 The refined structure model obtained from Rietveld refinement using XRPD data is in accord with that found using single-crystal X-ray diffraction, see Table S3.

image file: c7qi00004a-f3.tif
Fig. 3 A representative Oak Ridge thermal ellipsoid plot (ORTEP) of the eight crystallographically independent atoms in Li2CdGeSe4. The thermal ellipsoids are drawn at 50% probability.

A view of the tetrahedral packing illustrates that each cation in Li2CdGeSe4 coordinates to four tetrahedral selenium anions, and all of the tetrahedra align along the c axis. Due to this alignment, the structure is NCS, see Fig. 4 upper right, and produces a huge macroscopic dipole moment that is favourable for a sizeable SHG response. It is also interesting to compare the cation tetrahedral packing pattern of the recently reported zinc analogue, Li2ZnGeSe4 (Pn), with the wurtz-kesterite structure type44 to that exhibited by Li2CdGeSe4 (Pna21), Fig. 4. The 3D tetrahedral packing arrangements in Li2ZnGeSe4 and Li2CdGeSe4 are composed of a 2D stacking in an ABAB mode along the a and c axes, respectively. In Li2ZnGeSe4 (Pn), the A and B layers are composed of GeSe4 (red), Li(2)Se4 (dark green), Li(1)Se4 (light green), and ZnSe4 (blue) tetrahedral which are off-set from one another by two tetrahedral units. Whereas in Li2CdGeSe4 (Pna21), the composition of the A and B layers are still the same, but some of the Li(2)Se4 (dark green) and CdSe4 (blue) tetrahedra switch their positions leading to a different ordering of these cations along the b and a axes respectively in Li2ZnGeSe4 and Li2CdGeSe4, see Fig. 4 bottom. Therefore, we propose that the radius of the II cation has an effect on the structure-type adopted. In agreement with this idea, we have recently demonstrated that access to three hexagonally derived quaternary diamond-like structures can be controlled by II element selection in the Li2–II–Ge–S4 systems.63 These observations could provide some hints for further explorations of other DLSs when a specific structure-type is desired.

image file: c7qi00004a-f4.tif
Fig. 4 Comparison of the tetrahedral packing patterns for wurtz-kesterite type (Pn) Li2ZnGeSe4[thin space (1/6-em)]44 and the lithium cobalt(II) silicate type (Pna21) Li2CdGeSe4.

X-ray powder diffraction (XRPD) and Rietveld refinement

XRPD data for polycrystalline samples of Li2CdGeSe4 and Li2CdSnSe4 in combination with Rietveld refinements allowed the structures, as determined by single crystal X-ray diffraction, to be confirmed and the phase purity to be assessed. The results are shown in Fig. 5. All diffraction peaks were indexed to the lithium cobalt(II) silicate (Pna21) structure and no additional structural models were added to the refinements. The results of Rietveld refinements yielded excellent discrepancy values, Rp = 0.0581, wRp = 0.0803, and χ2 = 1.556 for Li2CdGeSe4 and Rp = 0.0560, wRp = 0.0768, and χ2 = 1.400 for Li2CdSnSe4, which establishes that both materials have been prepared with high phase purity. The unit cell parameters attained from Rietveld refinements, which are listed in Table S3, concur with those obtained from single crystal X-ray diffraction.
image file: c7qi00004a-f5.tif
Fig. 5 Rietveld refinement results for Li2CdGeSe4 and Li2CdSnSe4 using conventional XRPD data.

Thermal stability

Differential thermal analysis (DTA) diagrams of both Li2CdGeSe4 and Li2CdSnSe4 were found to have one endothermic peak during heating and two exothermic peaks during cooling. As shown in Fig. 6, upon heating the endothermic peak for Li2CdGeSe4 appears at 686 °C, and upon cooling the exothermic peaks occur at 785 and 643 °C, respectively. The 686 °C endothermic peak was identified as the melting of Li2CdGeSe4 with partial decomposition. The first exothermic peak at 785 °C may come from the recrystallization of a binary and/or ternary by-product resulting from the decomposition, and the second exothermic peak at 643 °C is likely the recrystallization of Li2CdGeSe4. For Li2CdSnSe4, all three events are similar, but occur at relatively lower temperatures, see Fig. S1. The XRPD data obtained for both materials before and after DTA were compared to assess the reversibility of the thermal events, see Fig. S2. The diffraction patterns of the DTA residues indicate that the major phase is still Li2CdGeSe4 or Li2CdSnSe4; however, there exist additional peaks, some of which can be indexed to CdSe, indicating a partial decomposition, see Fig. S2.
image file: c7qi00004a-f6.tif
Fig. 6 DTA diagram for the heating (red)/cooling (blue) cycle of Li2CdGeSe4.

UV-vis-NIR and FT-IR spectroscopy

Diffuse reflectance UV-vis-NIR spectroscopy was used in conjunction with attenuated total reflectance (ATR) FT-IR spectroscopy to assess the windows of optical clarity for Li2CdGeSe4 and Li2CdSnSe4. As shown in Fig. 7, both compounds exhibit measurable absorption in the visible region due to their bandgaps. In the region from ∼0.7 to 25 μm, both materials show wide optical clarity windows with transparency >60%. In fact, their transparency seems to extend beyond the detection limit of FT-IR (25 μm), indicating great potential as THz generators.
image file: c7qi00004a-f7.tif
Fig. 7 Transmittance spectra for Li2CdGeSe4 (top) and Li2CdSnSe4 (bottom) in UV-vis-NIR and mid-IR regions.

The UV-vis-NIR optical diffuse reflectance spectra converted to absorption for both materials are shown in Fig. 8. Extrapolation of the absorption edges to the respective baselines provides estimated optical bandgaps of ∼2.5 eV for the pink-red coloured Li2CdGeSe4 and 2.2 eV for orange crystals of Li2CdSnSe4. The pink-red colour of the former seems to arise from significant mid-gap transitions (Urbach tailing) as evident in Fig. 8. It should be noted that we had great trouble in obtaining optical absorption edge data consistent with the colour of these samples. Crystals ground and left under ambient conditions appear to darken in colour and give rise to narrower optical absorption edges. It seems that the crystals have some reaction to air/moisture resulting in a surface degradation that causes colour changes in these samples that in turn seriously distorts the near-gap absorption behaviour, yielding extrinsically modified “bandgaps”. Shown in Fig. 8 are the widest optical band edges that we could observe for these compounds when fresh samples were measured immediately after being obtained and with minimal grinding. The same samples left under ambient conditions for one week provide significantly narrower optical bandgaps of approximately 1.8 eV for both samples, see Fig. S3. Photographs of the as-prepared crystals can be seen in Fig. 1.

image file: c7qi00004a-f8.tif
Fig. 8 Optical diffuse reflectance UV-vis-NIR spectra converted to absorption for Li2CdGeSe4 (left) and Li2CdSnSe4 (right).

Interestingly, X-ray powder diffraction patterns of the ground samples left under ambient conditions for one week or more still show the characteristic XRPD patterns for pure-phase samples. This further supports the conclusion that the degradation only takes place on the surface. It is not surprising that the surface degradation has a profound effect on the diffuse reflectance measurement, since this is a more surface sensitive technique; yet, it does not cause any noticeable effect on the X-ray powder diffraction data, since X-ray diffraction better investigates the bulk sample. It should also be noted that crystals of both compounds were prepared for NLO measurements using fresh samples with no grinding. After sieving, these samples were stored and analysed in sealed, evacuated quartz tubes so they retained their original colour and had limited exposure to ambient conditions. We also emphasize that this surface degradation does not alter the SHG coefficients that are determined at mid-IR, which is far away from the bandgaps of the compounds; however, the LIDT may be affected as detailed below.

Electronic structure calculations

The calculated electronic band structures and corresponding density of states (DOS) for Li2CdGeSe4 and Li2CdSnSe4 are shown in Fig. 9. The state energies of the conduction band minimum (CBMin) and valence band maximum (VBMax) at each k point are listed in Table S5. For both compounds the highest point in the valence band and the lowest point in the conduction band lie at the G point of the irreducible Brillouin zone (IBZ); therefore, both compounds are calculated to be direct-gap semiconductors. Although quite similar, the calculated bandgap of Li2CdGeSe4, 1.98 eV, is a bit lower than that of Li2CdSnSe4, 2.11 eV. Compared to the experimentally determined optical bandgaps, the calculated bandgaps are underestimated. The underestimation is usual for this calculation method; accurate bandgap values are not expected from DFT calculations using the PBE-GGA functional due to the inability to accurately determine the exchange–correlation potential energy term.64 These slight variances between experiment and theory, however, should not inhibit our qualitative analysis of the DOS.
image file: c7qi00004a-f9.tif
Fig. 9 Calculated electronic band structure and the corresponding total and partial densities of states (DOS) for Li2CdGeSe4 and Li2CdSnSe4. The dotted line denotes the Fermi level. High-symmetry points in the irreducible Brillouin zone (IBZ) are displayed in an arbitrary path.

To better understand the origin of the bandgaps, the total and partial DOS were calculated as plotted in Fig. 9. The DOS plots for Li2CdGeSe4 and Li2CdSnSe4 are strikingly similar. Near the Fermi level, the VBMax shows states with contributions from Li p and Cd p orbitals that hybridize with the Se p orbitals. The states at the CBMin chiefly arise from the Ge/Sn s and Se p orbitals; however, there are also quite significant contributions from the Se s, Li s and p, as well as the Cd s and p states. Therefore, the Se p states dominate the whole Fermi level region and overlap fully with Li p, Cd p, and Ge/Sn s orbitals.

The Mülliken bond population analyses allow for a more quantitative bond assessment. The calculated bond orders of Cd–Se, Ge–Se and Sn–Se are 1.05–1.22, 1.13–1.41 and 0.99–1.26, respectively, significantly larger than those of the Li–Se bonds (0.04–0.22), (Table S4), which signify mainly ionic bonding interactions between the Li cations and Se anions.

The bandgaps observed for the Li2CdGeSe4 and Li2CdSnSe4 are significantly narrower, by more than 0.6 eV, than their sulphide analogues, Li2CdGeS4 (3.15 eV) and Li2CdSnS4 (2.87 eV), which is to be expected. The crystal radius of Se2− is significantly larger than that of S2− resulting in larger volume unit cells, longer metal–chalcogen bonding distances, and higher energy p-orbitals that result in considerably reduced bandgap energies. However, the bandgap values of the title compounds are comparable to the zinc analogues, Li2ZnGeSe4 (2.5 eV) and Li2ZnSnSe4 (2.0 eV).44 Therefore, while the II cation influences the structure type adopted by these quaternary DLSs, it does not play a meaningful role in defining the bandgap energy for these materials. The relatively narrow bandgaps of the title compounds compared to the sulphide analogues leads to the expectation of large second-order NLO susceptibilities.65

Nonlinear optical (NLO) properties

The naturally polar crystal structures of DLSs, in this case crystallizing in the Pna21 space group, prompted us to investigate the macroscopic NLO properties of Li2CdGeSe4 and Li2CdSnSe4. To be specific, the phase matchability, second-order NLO susceptibility and LIDT were determined.
Phase matchability. Particle size dependence of the SHG responses was assessed for Li2CdGeSe4 and Li2CdSnSe4 to determine the wavelengths at which these DLSs are PM. A positive trend in the SHG response as a function of increasing particle size indicates that a material is type-I PM at the applied wavelength, while a decrease in the SHG intensity with increasing particle size, or an initial increase followed by a decrease, is a characteristic of non-PM (type-I) materials.66 As shown in Fig. S4 and S5, both Li2CdGeSe4 and Li2CdSnSe4 are type-I PM with onset wavelengths around 2700 and 2900 nm, respectively, and the phase matchability of both materials likely extends to the far-IR or even terahertz (THz) region, as evidenced by their relatively static behaviours.
Second harmonic generation (SHG). SHG was measured as a function of broadband wavelength for both DLSs, from which the second-order NLO susceptibility, χ(2), was assessed as detailed below. The SHG responses were probed using the largest particle size (125–150 μm) because the compounds were deemed PM. As shown in Fig. 10, both compounds have greater SHG responses than the microcrystalline silver gallium selenide, AgGaSe2 (MC), reference in the short-wavelength region; but, the reference shows stronger SHG when the conversion wavelength exceeds 1450 nm (where the incident wavelength is 2900 nm). All compounds show larger SHG responses for longer wavelengths than shorter wavelengths. From Fig. 10, one can see that the SHG responses in the long-wavelength region are four orders of magnitude larger than those in the short-wavelength regime. This arises essentially from the bandgaps of the materials. As shown in Fig. 10, the bandgaps of the title compounds fall within the range of λSHG, resulting in substantial absorption of SHG radiation as well as 2PA of the fundamental beam. Using AgGaSe2 (χ(2)R = 66 pm V−1) as a reference, based on the Kurtz powder method,67
image file: c7qi00004a-t1.tif(1)

image file: c7qi00004a-f10.tif
Fig. 10 Broadband SHG responses of Li2CdGeSe4 and Li2CdSnSe4 compared to the AgGaSe2 (MC) reference with conversion wavelength at λ = 550–1850 nm. χ(2) is estimated in the long-wavelength regime, where the SHG responses plateau.

In eqn (1), χ(2) values for Li2CdGeSe4 and Li2CdSnSe4 were estimated in the long-wavelength regime where the SHG responses plateau, and both samples and the reference are PM with minimal 2PA. Using AgGaSe2 (MC) yielded χ(2) values of 56.3 and 55.7 pm V−1, for Li2CdGeSe4 and Li2CdSnSe4, respectively. It should be noted, however, that AgGaSe2 (MC) exhibits a lower SHG response in comparison to AgGaSe2 (OQ), an optical quality AgGaSe2 reference.41a,43 Accordingly, we have assigned the upper limit of the χ(2) in comparison to AgGaSe2 (MC), as explained above, and a lower bound on the χ(2) values for Li2CdGeSe4 and Li2CdSnSe4 as 25.6 and 25.3 pm V−1, respectively, by comparison to AgGaSe2 (OQ). While these values are lower than that of AgGaSe2, and AgGaS2,65 they exceed those of several commercially available IR NLO crystals, such as the lithium-containing DLSs LiInS2 and LiInSe2.68 Furthermore Li2CdGeSe4 and Li2CdSnSe4 display slightly improved SHG responses compared to those observed for Li2ZnGeSe4 and Li2ZnSnSe4.44

Laser-induced damage threshold (LIDT). While lasers have been greatly improved over the years in their ability to generate high-power, ultra-short pulses, the other optical components of a device often inhibit the high-performance expected of the system because of their inability to handle high power, due to laser-induced damage.69 Therefore, optimizing the LIDT is key in the pursuit of new IR NLO crystals for high-power laser applications.70 Yet critical evaluation of LIDT is often problematic, because the published studies are difficult to compare for the reason that the measurements are usually performed under different conditions (i.e. pulse width (τ), wavelength (λ), spot size, etc.).71 Therefore, here we compare only data that have been collected under identical conditions.

Generally, LIDT is dependent on MPA when the incident laser pulse duration is <50 ps and scales with bandgap energy. Our LIDT was measured at λ = 1064 nm (ħω = 1.17 eV) and τ = 30 ps. The SHG response was monitored as the input laser intensity was varied over two orders of magnitude. Note that the fundamental Nd:YAG line can induce 2PA in both samples and the reference. In the absence of 2PA and damage, the SHG intensity is expected to follow a squared power law. However, in the presence of 2PA, the power law is modified to

image file: c7qi00004a-t2.tif(2)
where I is the input laser intensity, β is the 2PA coefficient and d is the particle size. The I-dependent SHG responses and LIDT fits are shown in Fig. 11, where the solid traces indicate a squared power dependence in the absence of 2PA and the dotted traces are the fits using the modified squared law expressed in eqn (2). The 2PA coefficient, β, is found to be 12 and 5 cm GW−1, respectively for Li2CdGeSe4 and Li2CdSnSe4. A larger β value for Li2CdGeSe4 clearly indicates that 2PA arises via mid-gap states located around 2.0–2.3 eV; note that 2PA is not allowed if Urbach tailing is absent (2ħω < 2.5 eV). The measured 2PA efficiencies directly affect the LIDT values of Li2CdGeSe4 and Li2CdSnSe4 that were estimated to be 0.3 and 0.7 GW cm−2, respectively. These values are higher than that observed for AgGaSe2 (MC and OQ), for which LIDT = ∼0.2 GW cm−2.43 Remarkably, the LIDT for Li2CdSnSe4 even exceeds that of AgGaS2 (∼0.3–0.4 GW cm−2),43 although AgGaS2 has a notably wider bandgap of 2.62 eV.72 These realizations along with the visual observation of PL during the SHG measurements prompted us to further evaluate the PL response of Li2CdSnSe4 to gain insight on its unexpectedly high LIDT.

image file: c7qi00004a-f11.tif
Fig. 11 Intensity-dependent SHG of Li2CdGeSe4 and Li2CdSnSe4 used to determine the 2PA coefficients and the LIDTs.
Two-photon-induced photoluminescence (PL). The PL response of Li2CdSnSe4, first measured under band-to-band 2PA transition,73 exhibits a strong, red-light emission. Thus, time-integrated PL spectra were measured under various input laser intensities at λ = 1064 nm, as shown in Fig. 12. It should be noted that the intensity (I) of the PL emission at around 650 nm is higher than that of the SHG response (532 nm). The PL spectra were empirically fit with a double Gaussian of the form:
image file: c7qi00004a-t3.tif(3)
where λi and Γi are the parameters determining the location and width of each peak (i = A or B). Each constituent Gaussian (red and green traces) is plotted in Fig. 13 at I = 10.01 GW cm−2, along with the overall fit (blue trace) from eqn (3). The spectrally integrated counts of each PL peak are plotted (dots) as a function of I in Fig. 14. The two overlaid solid lines represent I2 dependence. Note that the data points begin to deviate from the lines for I > 1 GW cm−2, which is reasonably consistent with the LIDT for Li2CdSnSe4 obtained from SHG experiments, as shown in Fig. 11. For the lower I regime, I-dependence confirms that the PL is induced by the simultaneous absorption of two photons.

image file: c7qi00004a-f12.tif
Fig. 12 Two-photon-induced PL in Li2CdSnSe4 as a function of incident laser intensity at 1064 nm.

image file: c7qi00004a-f13.tif
Fig. 13 The PL spectrum of Li2CdSnSe4 at I = 10.01 GW cm−2 overlaid with a double Gaussian fit.

image file: c7qi00004a-f14.tif
Fig. 14 The intensity (I)-dependence of the A and B peaks, from the PL spectra in Fig. 12 superimposed by the square-law fits (lines). Laser-induced damage is evident for I > 1 GW cm−2, roughly consistent with the LIDT value obtained by the SHG method.

Relationship between bandgap, LIDT and PL

The process of laser damage in semiconductors and insulators is initialized when the application of an incident laser of sufficient energy promotes electrons from the valence band to the conduction band, creating free electrons. Depending upon the bandgap of the material and the wavelength of the incident radiation, this phenomenon can take place via one-photon absorption (1PA), 2PA, or three-photon absorption (3PA) etc. Once in the conduction band, the free electrons can absorb incident photons, thus obtaining additional energy. When enough energy has been absorbed, these free electrons will begin to collide with one another and transfer energy to bound electrons in the lattice, thus producing additional free electrons. This process is termed an electron avalanche. Although some energy that builds up is lost as heat, which is dissipated by the lattice, sooner or later the material will no longer be able to handle this energy, resulting in irreversible damage that changes, usually dramatically, the optical properties of the crystal. The first signs of laser-induced damage can usually be visualized via a microscope focused on the surface of the sample as a plasma formation and/or pitting,71 ultimately resulting in photodarkening, burning or cracking of the crystal that can be visually noticeable by the unaided eye.

In the short-pulse-width regime, the photoionization is the dominant phenomenon, because the energy is absorbed faster by the electrons than it is transferred to the lattice. In other words, an electron avalanche is more likely to occur with longer pulse duration. Both the photoionization and collisional excitation mechanisms of laser-induced damage have efficiencies that directly correlate to the bandgap of the material being investigated, and it has been shown for a variety of materials that LIDT is directly proportional to bandgap energy. For example, Gallais and coworkers74 have shown, using both experimental data and simulations, that the LIDT at both 800 nm and 400 nm (τ = 100 fs) is proportional to bandgap energy using a variety of thin film materials with bandgaps ranging from 1 eV for silicon to 10 eV for calcium fluoride, CaF2. The LIDT values increase in a “regular and continuous way”. In our case, damage occurs via 2PA as explained earlier. Even with a larger bandgap of 2.5 eV, a lower LIDT (or a higher β value) for Li2CdGeSe4 is understandable because of efficient mid-gap transitions, when compared with Li2CdSnSe4. This indicates that “optical gap” should be considered for 2PA, which is not necessarily the same as the fundamental gap, when assessing the LIDT. This also implies that one can achieve a very high LIDT of the Ge-containing compound when it is prepared as mid-gap-state free (2PA free) via post-growth processing (annealing etc.).

However, we note that the high LIDT of Li2CdSnSe4 cannot be explained solely by this relation, especially when compared with AgGaS2. As mentioned above, the PL emission of the Sn-containing compound seems to have a dramatic effect on the LIDT. Although the bandgap of Li2CdSnSe4 (2.2 eV) is slightly lower than that of AgGaS2 (2.62 eV), the LIDT is almost twice that of the ternary chalcogenide, which indeed does not exhibit PL emission under the same experimental conditions. We therefore deem that the strong, red-light emission serves to release energy from the compound, which is absorbed by a 2PA process under laser radiation. This energy dissipation allows the Sn-containing compound to maintain a consistent SHG response under increasing laser intensity, delaying an electron avalanche process that eventually leads to laser-induced damage in the sample.


In summary, the two new, biaxial quaternary DLSs, namely, Li2CdGeSe4 and Li2CdSnSe4, synthesized by high-temperature, solid-state reactions adopt the lithium cobalt(II) silicate structure type. Li2CdGeSe4 and Li2CdSnSe4 have strong SHG responses with the upper bound of χ(2) values being 56.3 and 55.7 pm V−1 and lower limits of 25.6 and 25.3 pm V−1, respectively, these results outshine those of commercially available Li-containing DLSs, such as LiInS2 (6.8,68a 11.16,68b 23.8,68c and 1568d pm V−1) and LiInSe2 (1768c and 2268e pm V−1), which suffer from difficulties in crystal growth as reflected in the different χ(2) values that have been reported for differently coloured samples containing various defects. Additionally, both DLSs show phase matchability, possess decent thermal stability, and exhibit wide windows of optical transparency making them attractive materials for IR NLO applications.

Moreover, the LIDTs of 0.3 and 0.7 GW cm−2 (λ = 1064 nm; τ = 30 ps), for Li2CdGeSe4 and Li2CdSnSe4 respectively, are higher than the benchmark AgGaSe2.43 The LIDT for Li2CdSnSe4, which is even higher than that of AgGaS2, is related to its strong, red-light emission under excitation at 1064 nm, which is induced by 2PA. This phenomenon helps to dissipate energy buildup from irradiation delaying an electron avalanche process. As a result, despite a slightly smaller optical bandgap, Li2CdSnSe4 presents an LIDT, which is about double that of AgGaS2.

Considering the research up to this time on quaternary chalcogenides, it can be concluded that the choice of the chalcogenide critically affects the bandgap energy and consequently the resulting NLO properties. For example, direct comparison of the Li2–II–IV–Se4 compounds presented here, as well as those previously published,44 with their sulphur-containing counterparts reveals clear trends in terms of optical nonlinearity, PM onset wavelength, and LIDT, which can be correlated to Eg. While the narrower-gap Li2-II-IV-Se4 compounds exhibit stronger optical nonlinearity (i.e. higher χ(2)), the wider-gap Li2CdGeS4 and Li2MnGeS4 have higher LIDTs, as well as broader PM ranges. However, as demonstrated here, improved LIDTs can be expected for selenides that exhibit multiphoton-induced PL. In the case of Li2CdSnSe4, the two-photon-induced PL serves to raise the LIDT; applying this idea to other compounds may open up a new avenue toward engineering LIDTs in DLS crystals.


This work was supported by the National Science Foundation of United States under Grants DMR-1201729 and DMR-1611198. Y. S. K. acknowledges support from the Basic Science Research Program (2015R1D1A3A03019609), Priority Research Center Program (2009-0093818), and Basic Research Lab Program (2014R1A4A1071686) through the National Research Foundation of Korea (NRF), funded by the Korean Government. The authors acknowledge Gooch and Housego (Ohio) for providing the optical-quality AgGaS2 and AgGaSe2 single crystal reference materials. A special thanks to Gary Catella and Dr. Carl Brunetta for useful discussions regarding practical NLO materials.

Notes and references

  1. R. J. Clewes, C. R. Howle, D. J. M. Stothard, M. H. Dunn, G. Robertson, W. Miller, G. Malcolm, G. Maker, R. Cox, B. Williams and M. Russell, Proc. SPIE, 2012, 8456, 85460X CrossRef.
  2. W. Lu, L. Liu, J. Sun and W. Pan, Opik, 2008, 119, 388–394 Search PubMed.
  3. (a) C. Wang and P. Sahay, Sensors, 2009, 9, 8230–8262 CrossRef CAS PubMed; (b) S. Persijn, F. Harren and A. van der Veen, Appl. Phys. B, 2010, 100, 383–390 CrossRef CAS.
  4. (a) V. Vaicikauskas, M. Kaucikas, V. Swedas and Z. Kuprionis, Rev. Sci. Instrum., 2007, 78, 023106 CrossRef CAS PubMed; (b) Y. M. Andreev, P. P. Geiko, G. M. Krekov and O. A. Romanovskii, Proc. SPIE, 1991, 1811, 367–370 CrossRef.
  5. D. J. Bamford, D. J. Cook, S. J. Sharpe and A. D. Van Pelt, Appl. Opt., 2007, 46, 3958–3968 CrossRef CAS PubMed.
  6. P. G. Schunemann, Proc. SPIE, 2007, 6455, 64550R CrossRef.
  7. (a) A. Jayaraman, V. Narayanamurti, H. M. Kasper, M. A. Chin and R. G. Maines, Phys. Rev. B: Solid State, 1976, 14, 3516–3522 CrossRef CAS; (b) A. Harasaki and K. Kato, Jpn. J. Appl. Phys., 1997, 36, 700–703 CrossRef CAS.
  8. (a) G. C. Catella, L. R. Shiozawa, J. R. Hietanen, R. C. Eckardt, R. K. Route, R. S. Feigelson, D. G. Cooper and C. L. Marquardt, Appl. Opt., 1993, 32, 3948–3951 CrossRef CAS PubMed; (b) M. C. Ohmer and R. Pandey, MRS Bull., 1998, 23, 16–20 CrossRef CAS.
  9. (a) V. G. Dmitriev, G. G. Gurzadyan and D. N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Springer, New York, 1999 Search PubMed; (b) N. C. Giles, L. H. Bai, M. M. Chirila, N. Y. Garces, K. T. Stevens, P. G. Schunemann, S. D. Setzler and T. M. Pollak, J. Appl. Phys., 2003, 93, 8975–8981 CrossRef CAS.
  10. (a) P. Korczak and C. B. Staff, J. Cryst. Growth, 1974, 24–25, 386–389 CrossRef CAS; (b) G. W. Iseler, J. Cryst. Growth, 1977, 41, 146–150 CrossRef CAS; (c) K. T. Zawilski, P. G. Schunemann, S. D. Setzler and T. M. Pollak, J. Cryst. Growth, 2008, 310, 1891–1896 CrossRef CAS.
  11. (a) A. Harasaki and K. Kato, Jpn. J. Appl. Phys., 1997, 36, 700–703 CrossRef CAS; (b) I. B. Zotova and Y. J. Ding, Opt. Commun., 2001, 198, 453–458 CrossRef CAS.
  12. (a) H. S. Ra, K. M. Ok and P. S. Halasyamani, J. Am. Chem. Soc., 2003, 125, 7764–7765 CrossRef CAS PubMed; (b) Z. Gao, X. Tian, J. Zhang, Q. Wu, Q. Lu and X. Tao, Cryst. Growth Des., 2015, 15, 759–763 CrossRef CAS.
  13. S. D. Nguyen, J. Yeon, S. H. Kim and P. S. Halasyamani, J. Am. Chem. Soc., 2011, 133, 12422–12425 CrossRef CAS PubMed.
  14. H. Y. Chang, S. H. Kim, P. S. Halasyamani and K. M. Ok, J. Am. Chem. Soc., 2009, 131, 2426–2427 CrossRef CAS PubMed.
  15. C. F. Sun, C. L. Hu, X. Xu, B. P. Yang and J. G. Mao, J. Am. Chem. Soc., 2011, 133, 5561–5572 CrossRef CAS PubMed.
  16. C. F. Sun, C. L. Hu, X. Xu, J. B. Ling, T. Hu, F. Kong, X. F. Long and J. G. Mao, J. Am. Chem. Soc., 2009, 131, 9486–9487 CrossRef CAS PubMed.
  17. H. Yu, H. Wu, S. Pan, Z. Yang, X. Hou, X. Su, Q. Jing, K. R. Poeppelmeier and J. M. Rondinelli, J. Am. Chem. Soc., 2014, 136, 1264–1267 CrossRef CAS PubMed.
  18. H. Wu, H. Yu, S. Pan, Z. Huang, Z. Yang, X. Su and K. R. Poeppelmeier, Angew. Chem., Int. Ed., 2013, 52, 3406–3410 CrossRef CAS PubMed.
  19. Y. Z. Huang, L. M. Wu, X. T. Wu, L. H. Li, L. Chen and Y. F. Zhang, J. Am. Chem. Soc., 2010, 132, 12788–12789 CrossRef CAS PubMed.
  20. H. Wu, S. Pan, K. R. Poeppelmeier, H. Li, D. Jia, Z. Chen, X. Fan, Y. Yang, J. M. Rondinelli and H. Luo, J. Am. Chem. Soc., 2011, 133, 16317–16317 CrossRef CAS.
  21. S. Wang and N. Ye, J. Am. Chem. Soc., 2011, 133, 11458–11461 CrossRef CAS PubMed.
  22. W. L. Zhang, W. D. Cheng, H. Zhang, L. Geng, C. S. Lin and Z. Z. He, J. Am. Chem. Soc., 2010, 132, 1508–1509 CrossRef CAS PubMed.
  23. J. L. Song, C. L. Hu, X. Xu, F. Kong and J. G. Mao, Angew. Chem., Int. Ed., 2013, 54, 3679–3682 CrossRef PubMed.
  24. H. Huang, L. Liu, S. Jin, W. Yao, Y. Zhang and C. Chen, J. Am. Chem. Soc., 2013, 135, 18319–18322 CrossRef CAS PubMed.
  25. S. G. Zhao, P. F. Gong, L. Bai, X. Xu, S. Q. Zhang, Z. H. Sun, Z. S. Lin, M. C. Hong, C. T. Chen and J. H. Luo, Nat. Commun., 2014, 5, 4019–4019 CAS.
  26. S. Zhao, P. Gong, S. Luo, S. Liu, L. Li, M. A. Asghar, T. Khan, M. Hong, Z. Lin and J. Luo, J. Am. Chem. Soc., 2015, 137, 2207–2210 CrossRef CAS PubMed.
  27. H. Yu, W. Zhang, J. Young, J. M. Rondinelli and P. S. Halasyamani, Adv. Mater., 2015, 27, 7380–7385 CrossRef CAS PubMed.
  28. G. Zhang, Y. J. Li, K. Jiang, H. Y. Zeng, T. Liu, X. G. Chen, J. G. Qin, Z. S. Lin, P. Z. Fu, Y. C. Wu and C. T. Chen, J. Am. Chem. Soc., 2012, 134, 14818–14822 CrossRef CAS PubMed.
  29. T. K. Bera, J. I. Jang, J. B. Ketterson and M. G. Kanatzidis, J. Am. Chem. Soc., 2009, 131, 75–77 CrossRef CAS PubMed.
  30. T. K. Bera, J. I. Jang, J. H. Song, C. D. Malliakas, A. J. Freeman, J. B. Ketterson and M. G. Kanatzidis, J. Am. Chem. Soc., 2010, 132, 3484–3495 CrossRef CAS PubMed.
  31. H. J. Zhao, Y. F. Zhang and L. Chen, J. Am. Chem. Soc., 2012, 134, 1993–1995 CrossRef CAS PubMed.
  32. H. Lin, L. Chen, L. J. Zhou and L. M. Wu, J. Am. Chem. Soc., 2013, 135, 12914–12921 CrossRef CAS PubMed.
  33. P. Yu, L. J. Zhou and L. Chen, J. Am. Chem. Soc., 2012, 134, 2227–2235 CrossRef CAS PubMed.
  34. M. C. Chen, L. M. Wu, H. Lin, L. J. Zhou and L. Chen, J. Am. Chem. Soc., 2012, 134, 6058–6060 CrossRef CAS PubMed.
  35. B. W. Liu, H. Y. Zeng, M. J. Zhang, Y. H. Fan, G. C. Guo, J. S. Huang and Z. C. Dong, Inorg. Chem., 2015, 54, 976–981 CrossRef CAS PubMed.
  36. S. F. Li, B. W. Liu, M. J. Zhang, Y. H. Fan, H. Y. Zeng and G. C. Guo, Inorg. Chem., 2016, 55, 1480–1485 CrossRef CAS PubMed.
  37. B. W. Liu, H. Y. Zheng, X. M. Jiang, G. Wang, S. F. Li, L. Xu and G. C. Guo, Chem. Sci., 2016, 7, 6273–6277 RSC.
  38. B. W. Liu, X. M. Jiang, G. Wang, H. Y. Zeng, M. J. Zhang, S. F. Li, W. H. Guo and G. C. Guo, Chem. Mater., 2015, 27, 8189–8192 CrossRef CAS.
  39. X. M. Jiang, M. J. Zhang, H. Y. Zeng, G. C. Guo and J. S. Huang, J. Am. Chem. Soc., 2011, 133, 3410–3418 CrossRef CAS PubMed.
  40. K. A. Rosmus, J. A. Brant, S. D. Wisneski, D. J. Clark, Y. S. Kim, J. I. Jang, C. D. Brunetta, J. H. Zhang, M. N. Srnec and J. A. Aitken, Inorg. Chem., 2014, 53, 7809–4811 CrossRef CAS PubMed.
  41. (a) J. A. Brant, D. J. Clark, Y. S. Kim, J. I. Jang, J. H. Zhang and J. A. Aitken, Chem. Mater., 2014, 26, 3045–3048 CrossRef CAS; (b) J. I. Jang, D. J. Clark, J. A. Brant, J. A. Aitken and Y. S. Kim, Opt. Lett., 2014, 39, 4579–4582 CrossRef CAS PubMed.
  42. J. W. Lekse, M. A. Moreau, K. L. McNerny, J. Yeon, P. S. Halasyamani and J. A. Aitken, Inorg. Chem., 2009, 48, 7516–7518 CrossRef CAS PubMed.
  43. J. A. Brant, D. J. Clark, Y. S. Kim, J. I. Jang, A. Weiland and J. A. Aitken, Inorg. Chem., 2015, 54, 2809–2819 CrossRef CAS PubMed.
  44. (a) J. H. Zhang, D. J. Clark, J. A. Brant, C. W. Sinagra III, Y. S. Kim, J. I. Jang and J. A. Aitken, Dalton Trans., 2015, 44, 11212–11222 RSC; (b) and erratum in preparation.
  45. Bruker, SMART and SAINT, Bruker AXS Inc., Madison, Wisconsin, USA, 1998 Search PubMed.
  46. G. M. Sheldrick, SADABS, University of Göttingen, Germany, 2002 Search PubMed.
  47. G. M. Sheldrick,SHELXTL, Crystallographic Software Package, Version 5.1, Bruker-AXS, Madison, WI, 1998 Search PubMed.
  48. A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13 CrossRef CAS.
  49. X'PertHighScorePlus, PANalytical B.V., Almelo, The Netherlands, 2006 Search PubMed.
  50. J. Faber and T. Fawcett, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 325–332 CrossRef.
  51. (a) A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory Report LAUR, 1994, pp. 86–748 Search PubMed; (b) B. H. Toby, J. Appl. Crystallogr., 2001, 34, 210–213 CrossRef CAS.
  52. Handbook of Mathematical Functions, ed. M. Abramowitz and I. A. Stegun, Dover Publications, Dover, NY, 1965, ch. 22 Search PubMed.
  53. C. D. Brunetta, W. C. Minsterman III, C. H. Lake and J. A. Aitken, J. Solid State Chem., 2012, 187, 177–185 CrossRef CAS.
  54. P. Kubelka and F. Munk, Z. Tech. Phys., 1931, 12, 593–601 Search PubMed.
  55. R. J. McGowan, Anal. Chem., 1963, 35, 1664–1665 CrossRef CAS.
  56. (a) M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717–2744 CrossRef CAS; (b) V. Milman, B. Winkler, J. A. White, C. J. Pickard, M. C. Payne, E. V. Akhmatskaya and R. H. Nobes, Int. J. Quantum Chem., 2000, 77, 895–910 CrossRef CAS.
  57. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868 CrossRef CAS PubMed.
  58. J. S. Lin, A. Qteish, M. C. Payne and V. Heine, Phys. Rev. B: Condens. Matter, 1993, 47, 4174–4180 CrossRef.
  59. F. P. Bundy and J. S. Kasper, J. Chem. Phys., 1967, 46, 3437–3446 CrossRef CAS.
  60. L. Pauling, J. Am. Chem. Soc., 1929, 51, 1010–1026 CrossRef CAS.
  61. O. V. Parasyuk, L. D. Gulay, L. V. Piskach and I. D. Olekseyuk, J. Alloys Compd., 2002, 335, 176–180 CrossRef CAS.
  62. (a) I. D. Brown and D. Altermatt, Acta Crystallogr., Sect. B: Struct. Sci., 1985, 41, 244–247 CrossRef; (b) N. E. Brese and M. O'Keeffe, Acta Crystallogr., Sect. B: Struct. Sci., 1991, 47, 192–197 CrossRef.
  63. J. A. Brant, K. P. Devlin, C. Bischoff, D. Watson, S. W. Martin, M. D. Gross and J. A. Aitken, Solid State Ionics, 2015, 278, 268–274 CrossRef CAS.
  64. (a) R. W. Godby, M. Schluther and L. J. Sham, Phys. Rev. B: Condens. Matter, 1987, 36, 6497–6500 CrossRef CAS; (b) J. P. Perdew, Int. J. Quantum Chem., 1986, 19, 497–523 Search PubMed; (c) D. Koller, F. Tran and P. Blaha, Phys. Rev. B: Condens. Matter, 2011, 83, 195134 CrossRef.
  65. (a) A. G. Jackson, M. C. Ohmer and S. R. LeClair, Infrared Phys. Technol., 1997, 38, 233–244 CrossRef CAS; (b) J. A. Aitken, J. A. Brant, D. J. Clark, Y. S. Kim and J. I. Jang, Nonlinear Optics: Fundamentals, Applications and Technological Advances, NOVA Science Publishers, New York, 2014 Search PubMed.
  66. K. M. Ok, E. O. Chi and P. S. Halasyamani, Chem. Soc. Rev., 2006, 35, 710–717 RSC.
  67. S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798–3813 CrossRef CAS.
  68. (a) G. D. Boyd, H. M. Kasper and J. H. MacFee, J. Appl. Phys., 1973, 44, 2809–2812 CrossRef CAS; (b) G. M. H. Knippels, A. F. G. van der Meer, A. M. MacLeod, A. Yelisseyev, L. Isaenko, S. Lobanov, I. Thénot and J. J. Zondy, Opt. Lett., 2001, 26, 617–619 CrossRef CAS PubMed; (c) W. Chen, J. Cousin, M. W. Sigrist, X. Gao, J. J. Zondy, L. Isaenko, A. Yelisseyev and S. Lobanov, Laser and Electro-Optics Society, 19th Annual Meeting of the IEEE, Oct., Montreal, Canada, 2006, pp. 88–89 Search PubMed; (d) A. Yelisseyev, L. Isaenko, S. Lovanov and J. J. Zondy, Advanced Solid State Lasers, in OSA Trends in Optics and Photonics Series, ed. H. Injeyan and C. Marshall, Optical Society of America, Washington, D.C., 2000, vol. 34, pp. 561–596 Search PubMed; (e) M. Ebrajim-Zadeh and I. T. Gorokina, NATO science for peace and security series B: physics and biophysics. Mid-infrared coherent sources and applications, Springer, Dordrecht, The Netherlands, 2008, pp. 347–375 Search PubMed.
  69. M. Sozet, J. Néauport, E. Lavastre, N. Roquin, L. Gallais and L. Lamaignère, Opt. Lett., 2015, 40, 2091–2094 CrossRef PubMed.
  70. F. K. Hopkins, Opt. Photonics News, 1998, 9, 32–38 CrossRef CAS.
  71. H. Kildal and G. W. Iseler, Appl. Opt., 1976, 15, 3062–3065 CrossRef CAS PubMed.
  72. G. C. Catella and D. Burlage, MRS Bull., 1998, 23, 28–36 CrossRef CAS.
  73. D. J. Clark, L. Yuan, C. O. Otieno, G. Zhou and J. I. Jang, Solid State Commun., 2014, 181, 9–14 CrossRef CAS.
  74. L. Gallais, D.-B. Douti, M. Commandré, G. Batavičiūtė, E. Pupka, M. Ščiuka, L. Smalakys, V. Sirutkaitis and A. Melninkaitis, J. Appl. Phys., 2015, 117, 223103 CrossRef.


Electronic supplementary information (ESI) available: Crystallographic information files (CIF) for the title compounds Li2CdGeSe4 and Li2CdSnSe4; tables of crystallographic data; lattice parameters obtained from Rietveld refinements; results of bond valence calculations; the state energies of the conduction band min. and valence band max. at each k-point; DTA diagram for Li2CdSnSe4; XRPD patterns before and after DTA; plots of absorption edges before and after exposure to ambient conditions; plots of SHG particle-size dependence. CCDC 1524708 and 1524808. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7qi00004a
Current address: Department of Physics, Sogang University, Seoul, South Korea.

This journal is © the Partner Organisations 2017