betaFeedback

WorksPhysics and Astronomy

Nonlinear Anisotropy in Phase-Tuned Wide-Gap Halides

Abstract

Silver iodide (AgI) thin films offer a compelling platform for studying nonlinear optical phenomena due to their intrinsic noncentrosymmetric lattice and direct band gap. Here, we investigate the nonlinear optical properties of AgI thin films grown by physical vapor deposition that selectively produce zincblende (\zbAgI) and wurtzite (\wzAgI) phases. Using a combination of polarization-resolved second harmonic generation (SHG) and two-photon photoluminescence (2PPL) spectroscopy, we identify clear phase- and morphology-dependent anisotropic nonlinear responses. Triangular \zbAgI (111)(111) flakes exhibit sixfold SHG symmetry and isotropic 2PPL emission, while rod-shaped \wzAgI (101)(101) samples display twofold-symmetric patterns in both SHG and 2PPL, which are explained by theories that integrate explicit polarization analysis using second- and third-order nonlinear susceptibilities. We estimate χeff(2)χ^{(2)}_\mathrm{eff} of 0.45 pm/V (\zbAgI) and 0.16 pm/V (\wzAgI), respectively, after correcting for multiple internal reflections and generation of SHG within the AgI film and optical interference effect in the dielectric layers. These results identify AgI as a useful single-composition halide platform for studying phase-dependent nonlinearity and establish a realistic methodology for evaluating nonlinear susceptibilities in layered materials or thin films supported by a substrate.

Introduction

Nonlinear optical effects in crystalline materials provide powerful means of probing symmetry, band structure, and light–matter interactions. In particular, second harmonic generation (SHG) and two-photon photoluminescence (2PPL) have become indispensable tools for studying noncentrosymmetric materials and their symmetry-dependent responses [1]. The past decade has witnessed remarkable advances in two-dimensional (2D) materials such as transition-metal dichalcogenides (TMDs) and hexagonal boron nitride, where interlayer stacking and rotational alignment can reversibly break or restore inversion symmetry, thereby modulating the nonlinear optical response [2, 3, 4]. Such stacking-dependent SHG effects demonstrate that structural symmetry control at the atomic scale provides a robust handle for engineering nonlinearities in low-dimensional systems.

Beyond van der Waals semiconductors, halide compounds are emerging as a promising new class of nonlinear materials [5, 6, 7]. Their large polarizability, tunable electronic structure, and propensity to form noncentrosymmetric lattices give rise to strong χ(2)\chi^{(2)} and χ(3)\chi^{(3)} responses even in bulk phases. Lead halide perovskites and copper or silver halides have exhibited tunable second- and third-order nonlinearities due to structural phase transitions, excitonic resonances, and strain [8, 9, 10, 11]. However, in many halide systems, structural changes are accompanied by variations in chemistry, defect populations, or disorder, which complicates attributing the nonlinear optical response solely to the crystallographic phase. A material in which phase can be varied while composition remains fixed is therefore especially useful for separating structure-driven changes in nonlinear anisotropy from compositional effects. Silver iodide (AgI) provides such a platform. AgI exhibits polymorphism between zincblende (γ\gamma-AgI) and wurtzite (β\beta-AgI) under ambient conditions [12, 13, 14], while preserving the same Ag-I stoichiometry and tetrahedral coordination framework. Because both polymorphs are noncentrosymmetric and direct-gap semiconductors [15], the comparison is not limited to a symmetry-allowed versus symmetry-forbidden SHG case. Instead, it permits a more selective examination of how a change in long-range crystal symmetry modifies the angular response of both SHG and 2PPL within a common chemical system.

In this work, phase-dependent nonlinear optical response in AgI thin films synthesized by controlled physical vapor deposition is investigated using polarization-resolved SHG and 2PPL spectroscopy. Distinct angular responses are observed for triangular crystals assigned to γ\gamma-AgI(111) and rod-like crystals assigned to β\beta-AgI(101). We thus establish that in a single-composition halide, both symmetry and distinct morphological effects on χ(2)\chi^{(2)} (SHG) and χ(3)\chi^{(3)} (2PPL) can be analyzed simultaneously. The values of second-order nonlinear susceptibilities χeff(2)\chi_{\mathrm{eff}}^{(2)} were estimated to be 0.45 pm/V (γ\gamma-AgI) and 0.16 pm/V (β\beta-AgI), respectively, by using quartz as a reference and correcting for an interference effect coming from the AgI film and the dielectric substrate underneath. The calibration approach for χ(2)\chi^{(2)} adopted here should be useful to evaluate other emerging highly-nonlinear layered materials or thin films supported by a dielectric substrate.

Experimental

Silver iodide (AgI) microcrystals were synthesized in a custom-built quartz-tube physical vapor deposition system based on a horizontal furnace (MTI) with an 8-inch-long heating zone mounted on a sliding rail. Approximately 0.1 g of high-purity AgI powder was loaded on a ceramic plate, and three 1 cm ×\times 1 cm Si/SiO2_2 substrates with a 90 nm oxide layer were placed downstream on a separate ceramic plate. Before growth, the tube was evacuated for 5 min, purged with a 96% Ar and 4% H2_2 gas mixture up to 0.02 MPa, evacuated again for 5 min, and then placed under carrier-gas flow. We adopted reverse flow (substrate to powder) during the temperature ramp to prevent growth, and switched to forward flow (powder to substrate) at the target growth temperatures, which were 450 °C for triangular platelets, and 470 °C for rod-shaped micro-crystals, respectively. Growth was terminated by returning to reverse flow, rapidly sliding the furnace away from the substrate region, and then cooling to room temperature under carrier gas.

Image not extracted

Figure 1. Electron microscopy and EDS characterization of two AgI samples, a triangular platelet and a rod-like platelet, prepared as FIB lift-out cross-sectional lamellae. Triangular platelet: (a) STEM-ADF image of the cross section. (b) EDS elemental map showing Ag (magenta) and I (green). (c) High-resolution STEM image showing the [110][110] zone axis, consistent with γ\gamma-AgI. (d) SEM plan-view images of the triangular and rod-like platelets used for TEM sample preparation. The dashed yellow lines indicate the cutting locations for cross sectioning. Rod-like platelet: (e) STEM-ADF image of the cross section. (f) EDS elemental map. (g) High-resolution STEM image showing the [212ˉ][21\bar{2}] zone axis, consistent with β\beta-AgI.

Cross-sectional lamellae from representative triangular and rod-like crystals were prepared by focused-ion-beam lift-out using an FEI Scios DualBeam FIB/SEM, followed by final polishing at 2 keV to reduce surface damage. Scanning transmission electron microscopy was carried out on a probe-corrected Titan Themis 80–300 operated at 300 kV with a probe semi-convergence angle of 23.8 mrad and a beam current of 80 pA. Annular dark-field and high-angle annular dark-field imaging, together with Bruker SuperX energy-dispersive X-ray spectroscopy mapping, were used to evaluate morphology and elemental distribution. Phase assignment was based on lattice-resolved STEM, fast Fourier transforms, and zone-axis indexing, with comparison to expected interplanar spacings for candidate AgI polymorphs and, where applicable, simulated patterns. The triangular platelets were found to be consistent with γ\gamma-AgI with a nominal (111) exposed surface, whereas the rod-like crystals were found to be consistent with β\beta-AgI with a nominal (101) exposed surface. For the rod-like crystals, the TEM indexing was corroborated by XRD, which showed

a peak at 2θ≈25.3∘2\theta\approx25.3^\circ consistent with β\beta-AgI(101). Representative triangular platelets exhibited an approximately 500 nm cross-sectional thickness.

Polarization-resolved second-harmonic generation (SHG) and two-photon photoluminescence (2PPL) measurements were performed in a custom reflection-geometry setup described in Ref.[16] using a mode-locked Ti:sapphire laser (Tsunami, Spectra-Physics). An ultrafast pulsed laser with a pulse width of 80 fs and a near-infrared wavelength of 780 – 830 nm was used as an incident beam. A prism-based compressor was used to maintain comparable pulse conditions over the tuning range. A dichroic mirror directed the excitation beam through a 40× objective (NA 0.6), which gave a spot diameter of approximately 1.6 μ\mum at λ≈800\lambda\approx800 nm. The excitation power was set to 1.5 mW for polarization-dependent measurements. White-light illumination, a beam splitter, a tube lens, and a CCD camera were used for sample identification and alignment. The emitted nonlinear signal was collected through the same objective, routed through spectral filters to suppress the fundamental beam, and recorded with a spectrometer (Acton Research) equipped with a thermoelectrically cooled CCD (Horiba). The spectrometer with CCD was calibrated with a mercury pen lamp.

For angular measurements, the incident polarization was fixed by a linear polarizer in the excitation path, the sample was rotated with respect to a fixed pump polarization, and a second linear polarizer in the detection path selected the emitted component parallel to the incident polarization. The intensity of SHG and 2PPL was analyzed by integrating each emission over a corresponding spectral window. Rod-like crystals occasionally exhibited time-dependent signal degradation during rotation scans. In those cases, the analysis was restricted to the stable portion of the scan, and the angular traces were normalized to their own maxima before fitting.

Results and discussion

Phase-selective SHG and excitonic 2PPL

Image not extracted

Figure 2. Wavelength-resolved nonlinear emission from an individual AgI triangular flake. Photoluminescence spectra were collected as the excitation wavelength was tuned from 784 nm to 798 nm. Each trace exhibits two distinct features: (i) a band at 388–398 nm attributed to second-harmonic generation (SHG), which shifts according to λ/2\lambda/2 of the fundamental excitation; and (ii) another band at 412–420 nm assigned to excitonic emission, whose peak position remains nearly constant while its intensity varies with the pump wavelength. All intensities are plotted in arbitrary units (A.U.).

Figure 2 presents wavelength-resolved nonlinear emission spectra from an individual AgI triangular flake. The spectra reveal two distinct peaks: (i) a shorter-wavelength band at 388–398 nm corresponding to SHG, confirmed by its λout=λin/2\lambda_{\mathrm{out}} = \lambda_{\mathrm{in}}/2 dependence, and (ii) another peak at approximately 416 nm, assigned to the excitonic transition. The exciton emission varied from 412 nm to 440 nm, depending on the specific flake measured, likely originating from different crystal phases as detailed below. Since the fundamental laser energy is well below the estimated bandgap of around 3 eV in AgI [8], the observed exciton peak is attributed to 2PPL.

Reference-based estimate of χ(2)\chi^{(2)}

A quantitative estimate of the second-order nonlinear susceptibility was obtained by comparing the SHG intensity of AgI crystals with that of a commercial α\alpha-quartz reference, measured sequentially under the same nominal excitation, integration time, spectrometer settings, and polarization configuration. The incident fundamental power was 1.0 mW for AgI and 2.0 mW for quartz, which was accounted for using the quadratic power dependence of SHG. Minor refocusing and sample-angle adjustments were used to maximize the collected signal from each specimen (Figure 3).

Image not extracted

Figure 3. SHG spectra measured from representative AgI crystals with triangular and nanorod morphologies, together with a commercial quartz reference [Quartz (0001), 5 mm × 5 mm × 0.5 mm]. Insets show AFM topography images of the corresponding triangular platelet and nanorod flakes.

Because the AgI crystals possess thicknesses comparable to the optical wavelengths inside the material (e.g., 220 nm for the nanorods), an infinitesimally thin sheet approximation is insufficient. Following Lavrov et al., we employ a depth-integrated formalism [17] to account for the phase-mismatch and multiple internal reflections of the fundamental and second-harmonic fields within the bulk AgI/SiO2_2/Si stack. The absolute value of χAgI(2)\chi_{\mathrm{AgI}}^{(2)} is extracted by comparing the integrated AgI response to a bulk α\alpha-quartz reference, utilizing a model that incorporates the focused Gaussian geometry of the beam to appropriately handle the finite coherent buildup within the focal volume [18].

By taking the ratio of the generated SHG efficiencies, as described in detail in the Appendix E, the comparative expression for the effective bulk susceptibility becomes:

∣χAgI(2)∣=∣χq(2)∣[16λ2f2(N.A.)4nAgI(2ω)2dL2GMRnq2(nq+1)6IAgI(2ω)Iq(2ω)(PqPAgI)2]1/2,(1)\left|\chi_{\mathrm{AgI}}^{(2)}\right| = \left|\chi_{\mathrm{q}}^{(2)}\right| \left[ \frac{16\lambda^2 f^2(\mathrm{N.A.})^4 n_{\mathrm{AgI}}(2\omega)^2}{d_L^2 G_{\mathrm{MR}} n_{\mathrm{q}}^2(n_{\mathrm{q}}+1)^6} \frac{I_{\mathrm{AgI}}(2\omega)}{I_{\mathrm{q}}(2\omega)} \left(\frac{P_{\mathrm{q}}}{P_{\mathrm{AgI}}}\right)^2 \right]^{1/2}, \tag*{(1)}

where f=0.056f = 0.056 is the numerical integration constant for the focused Gaussian beam [18], N.A. is the objective numerical aperture, λ\lambda is the fundamental wavelength in vacuum, and dLd_L is the thickness of the AgI layer. The term GMRG_{\mathrm{MR}} is a dimensionless structure factor that integrates the depth-dependent local pump field enhancement and second-harmonic out-coupling influenced by multiple internal reflections within the target stack. IAgI(2ω)I_{\mathrm{AgI}}(2\omega) and Iq(2ω)I_{\mathrm{q}}(2\omega) denote the maximum SHG peak counts, and PAgIP_{\mathrm{AgI}} and PqP_{\mathrm{q}} are the respective incident pump powers. The detailed derivation of the integration and structure factor is deferred to Appendix E.

Using this approach, the extracted bulk susceptibilities were estimated as:

χAgI(zb)(2)≈0.45 pm/V,(2)\chi_{\mathrm{AgI(zb)}}^{(2)} \approx0.45\ \mathrm{pm/V}, \tag*{(2)}

for the 19 nm zincblende flake, and

χAgI(wz)(2)≈0.16 pm/V,(3)\chi_{\mathrm{AgI(wz)}}^{(2)} \approx0.16\ \mathrm{pm/V}, \tag*{(3)}

for the 220 nm wurtzite rod. These bulk values fall well within the typical range for wide-gap nonlinear crystals such as KDP (χ(2)∼0.42 pm/V\chi^{(2)} \sim0.42\ \mathrm{pm/V}) and quartz (χ(2)∼0.6 pm/V\chi^{(2)} \sim0.6\ \mathrm{pm/V}) [1]. We note that the same refractive-index values were used for zincblende and wurtzite AgI because phase-resolved optical constants for zincblende AgI are unavailable. Thus, the quoted values should be regarded as approximate comparative estimates.

Analysis of SHG and 2PPL Angular Dependence

To study phase-dependent nonlinear anisotropy in AgI, we have performed polarization-resolved SHG and 2PPL measurements for samples with different morphologies: triangular platelets and rod-shaped AgI. Distinct angular anisotropy was observed in SHG from those crystals, exhibiting a sixfold “flower” pattern for the triangular AgI (Figure 4(b)) compared to a twofold “butterfly” pattern for the AgI rod (Figure 4(e)). The 2PPL also showed distinct behavior depending on the morphologies: We observe an isotropic circular response for the triangular island (Figure 4(c)) and a twofold uniaxial pattern in the rod (Figure 4(f)). Note that although triangular AgI consistently displayed uniform angular-SHG patterns, the patterns derived from rod-shaped AgI were more diverse, sometimes also indicating the impact of beam damage.

Image not extracted

Figure 4. Polarization-resolved nonlinear-optical response of two AgI micro-crystals. (a) Optical micrograph of a triangular flake; (b) SHG polar plot from flake (a), exhibiting a six-lobed pattern that follows a cos⁡2[3(θ−ϕ0)]\cos^2[3(\theta-\phi_0)] dependence; (c) 2PPL response of the triangle. (d) Optical micrograph of a nanorod; (e) SHG polar plot recorded from the nanorod showing twofold symmetry aligning with simulation values for d1>d2d_1>d_2; (f) corresponding 2PPL revealing pronounced anisotropy, contrasting the isotropic response from (a) – maxima are aligned with the long axis of the rod.

To analyze the nonlinear anisotropy observed in zincblende and wurtzite phases of AgI with distinct morphologies, we employ second- and third-order nonlinear optical formalisms that account for both crystal symmetry and the experimental geometry.

Second-harmonic generation (SHG). The three orthogonal components of the second-order nonlinear polarization Pi,2ωP_{i,2\omega} are expressed as

Pi,2ω=2ε0∑j,kdijk(2)(2ω;ω,ω)EjEk,(4)P_{i,2\omega}=2\varepsilon_{0}\sum_{j,k}d_{ijk}^{(2)}(2\omega;\omega,\omega)E_jE_k, \tag*{(4)}

where ε0\varepsilon_0 is the permittivity of free space, dijk(2)=12χijk(2)d_{ijk}^{(2)}=\frac{1}{2}\chi_{ijk}^{(2)} is the second-order susceptibility tensor, and EjEkE_jE_k are components of the fundamental electric field.

For the zincblende phase of AgI (point group 4‾3m\overline{4}3m), the nonvanishing elements of dijkd_{ijk} are [1] dxyz=dyzx=dzxy≡d14d_{xyz}=d_{yzx}=d_{zxy}\equiv d_{14}, assuming dijk=dikjd_{ijk}=d_{ikj} symmetry:

dzb=[000d14000000d14000000d14].(5)\mathbf{d}_{\mathrm{zb}}= \begin{bmatrix} 0&0&0&d_{14}&0&0\\ 0&0&0&0&d_{14}&0\\ 0&0&0&0&0&d_{14} \end{bmatrix}. \tag*{(5)}

For the wurtzite phase (point group 6mm6mm), the nonvanishing elements are dzxx=dzyy=dyyz=dxxz≡d15d_{zxx}=d_{zyy}=d_{yyz}=d_{xxz}\equiv d_{15} and dzzz≡d33d_{zzz}\equiv d_{33} [1]:

dwz=[0000d150000d1500d15d15d33000].(6)\mathbf{d}_{\mathrm{wz}}= \begin{bmatrix} 0&0&0&0&d_{15}&0\\ 0&0&0&d_{15}&0&0\\ d_{15}&d_{15}&d_{33}&0&0&0 \end{bmatrix}. \tag*{(6)}

The xx, yy, and zz labels correspond to laboratory axes aligned with the crystallographic axes as defined in Figure 5.

Image not extracted

Figure 5. Relation between Cartesian and crystallographic coordinate systems, and definition of beam orientation angles. (a) Alignment of Cartesian and crystallographic axes in a cubic unit cell. (b) Alignment of Cartesian and crystallographic axes in a hexagonal unit cell. (c) Definition of the angles θ\theta, ϕ\phi, and δ\delta describing the orientation of the incident laser beam (wave vector k\mathbf{k}) and its optical electric field E\mathbf{E}.

To analyze polarization-resolved SHG when E is rotated within a crystallographic plane, we introduce an effective scalar susceptibility

deff(θ,ϕ,δ)=(e3x,e3y,e3z)⋅[e1x2e1y2e1z22e1ye1z2e1xe1z2e1xe1y],(7)d_{\mathrm{eff}}(\theta,\phi,\delta)=(e_{3x},e_{3y},e_{3z})\cdot\begin{bmatrix} e_{1x}^{2}\\ e_{1y}^{2}\\ e_{1z}^{2}\\ 2e_{1y}e_{1z}\\ 2e_{1x}e_{1z}\\ 2e_{1x}e_{1y} \end{bmatrix}, \tag*{(7)}

where e3e_3 denotes the SH polarization direction after the analyzer, and e1e_1 is the polarization of the fundamental beam. In our configuration e3=e1=ee_3=e_1=e, described by

ex=sin⁡ϕsin⁡δ−cos⁡θcos⁡ϕcos⁡δ,ey=−cos⁡ϕsin⁡δ−cos⁡θsin⁡ϕcos⁡δ,ez=sin⁡θcos⁡δ.(8)\begin{aligned} e_x&=\sin\phi\sin\delta-\cos\theta\cos\phi\cos\delta,\\ e_y&=-\cos\phi\sin\delta-\cos\theta\sin\phi\cos\delta,\\ e_z&=\sin\theta\cos\delta. \tag*{(8)} \end{aligned}

Rotation of polarization within the zincblende (111) plane corresponds to rotating δ\delta from 0° to 360° at fixed θ=54.74∘\theta=54.74^\circ, ϕ=45∘\phi=45^\circ, while for wurtzite (101) θ=62∘\theta=62^\circ and ϕ=30∘\phi=30^\circ. The scalar quantity deff(θ,ϕ,δ)d_{\mathrm{eff}}(\theta,\phi,\delta) can thus be computed and scaled to fit the SHG angular pattern.

Two-photon absorption (2PA) and two-photon photoluminescence (2PPL). The 2PA process [19, 20] originates from the imaginary part of the third-order nonlinear susceptibility, χ(3)\chi^{(3)}, which governs the third-order polarization

Pi(3)=ε0∑jklχijkl(3)EjEkEl.(9)P_i^{(3)}=\varepsilon_0\sum_{jkl}\chi_{ijkl}^{(3)}E_jE_kE_l. \tag*{(9)}

The time-averaged absorbed power [21] is

⟨E⋅P˙(3)⟩∝Im⁡[χeff(3)(θ,ϕ,δ)]∣E∣4,(10)\langle\mathbf{E}\cdot\dot{\mathbf{P}}^{(3)}\rangle\propto\operatorname{Im}\left[\chi_{\mathrm{eff}}^{(3)}(\theta,\phi,\delta)\right]|\mathbf{E}|^4, \tag*{(10)}

where we define the effective scalar susceptibility

χeff(3)(θ,ϕ,δ)≡∑ijklχijkl(3)ei(θ,ϕ,δ)ej(θ,ϕ,δ)ek(θ,ϕ,δ)el(θ,ϕ,δ).(11)\chi_{\mathrm{eff}}^{(3)}(\theta,\phi,\delta)\equiv\sum_{ijkl}\chi_{ijkl}^{(3)}e_i(\theta,\phi,\delta)e_j(\theta,\phi,\delta)e_k(\theta,\phi,\delta)e_l(\theta,\phi,\delta). \tag*{(11)}

This contraction reduces the fourth-rank tensor χijkl(3)\chi_{ijkl}^{(3)} to a complex scalar quantity that depends explicitly on the optical geometry (θ,ϕ,δ)(\theta,\phi,\delta). This E4E^4 scaling is consistent with both quantum-mechanical calculations of 2PA matrix elements [22] and macroscopic susceptibility analyses [23, 24]. The measured 2PPL intensity is assumed to be proportional to the absorbed two-photon energy and unpolarized due to inelastic scattering before emission, giving

I2PPL(θ,ϕ,δ)∝I2PA(θ,ϕ,δ).(12)I_{2\mathrm{PPL}}(\theta,\phi,\delta) \propto I_{2\mathrm{PA}}(\theta,\phi,\delta). \tag*{(12)}

Under degenerate excitation with linear polarization, the relevant quantity in Eq. (11) is the quartic field product EiEjEkElE_iE_jE_kE_l. Because the two absorbed photons are indistinguishable, the indices associated with the input fields (jj, kk, ll) enter symmetrically. We therefore group the quadratic field products into a six-component vector,

v(δ)=[Ex2 Ey2 Ez2 2EyEz 2EzEx 2ExEy]T,(13)v(\delta)=\left[E_x^2\ E_y^2\ E_z^2\ 2E_yE_z\ 2E_zE_x\ 2E_xE_y\right]^{\mathsf{T}}, \tag*{(13)}

so that the 2PA signal may be written compactly as a quadratic form

I2PA(θ,ϕ,δ)∝v(δ)TB v(δ).(14)I_{2\mathrm{PA}}(\theta,\phi,\delta)\propto v(\delta)^{\mathsf{T}}B\,v(\delta). \tag*{(14)}

Here, BB is a real, symmetric 6×66\times6 matrix whose entries are linear combinations of the independent components of Im⁡[χijkl(3)]\operatorname{Im}[\chi_{ijkl}^{(3)}] [24, 25]. The matrix BB encapsulates all crystal-symmetry constraints and fully determines the angular dependence of 2PA through the direction cosines (Ex,Ey,Ez)=(ex,ey,ez)(E_x,E_y,E_z)=(e_x,e_y,e_z) in Eq. (8). This is the same structure that underlies Voigt contraction for other fourth-rank optical tensors (e.g., Kerr and elastic tensors [26, 27]). Although not frequently used in the context of 2PA, it makes crystal-symmetry constraints (cubic vs. hexagonal) transparent. Angular anisotropy of 2PA under more general conditions is discussed, for example, in Refs. [23, 24, 28].

Expanding Eq. (14) gives

I2PA∝B11Ex4+B22Ey4+B33Ez4+2(B12Ex2Ey2+B13Ex2Ez2+B23Ey2Ez2)+B44(2EyEz)2+B55(2EzEx)2+B66(2ExEy)2.(15)I_{2\mathrm{PA}} \propto B_{11}E_x^4+B_{22}E_y^4+B_{33}E_z^4 +2\left(B_{12}E_x^2E_y^2+B_{13}E_x^2E_z^2+B_{23}E_y^2E_z^2\right) +B_{44}(2E_yE_z)^2+B_{55}(2E_zE_x)^2+B_{66}(2E_xE_y)^2. \tag*{(15)}

Following the tensor analysis of Inoue and Toyozawa [29] for two-photon absorption in semiconductors, the fourth-rank susceptibility entering the 2PPL process can be decomposed into a small number of scalar invariants fixed by crystal symmetry. For zincblende, the two-photon absorption tensor can be written as a sum of only two linearly independent invariants: one isotropic term and one cubic-anisotropy term.

β1(zb)=B11+2B12,β2(zb)=B11−(B11+2B12)(16)\beta_1^{(\mathrm{zb})}=B_{11}+2B_{12},\qquad \beta_2^{(\mathrm{zb})}=B_{11}-(B_{11}+2B_{12}) \tag*{(16)}

Here, β1(zb)\beta_1^{(\mathrm{zb})} sets the isotropic background, while β2(zb)\beta_2^{(\mathrm{zb})} controls the cubic anisotropy. In contrast, the lower hexagonal symmetry of wurtzite allows four independent invariants in the same decomposition. The corresponding 2PPL intensity then depends on four independent scalar coefficients, each a distinct linear combination of the underlying tensor elements.

β1(wz)=B33,β2(wz)=2B13+4B44,\beta_1^{(\mathrm{wz})}=B_{33},\qquad\beta_2^{(\mathrm{wz})}=2B_{13}+4B_{44},
β3(wz)=B11,β4(wz)=2B12+4B66−6B11(17)\beta_3^{(\mathrm{wz})}=B_{11},\qquad\beta_4^{(\mathrm{wz})}=2B_{12}+4B_{66}-6B_{11} \tag*{(17)}

Here, β3(wz)\beta_3^{(\mathrm{wz})} governs the in-plane isotropic part and β4(wz)\beta_4^{(\mathrm{wz})} the in-plane uniaxial (twofold) anisotropy, whereas β1(wz)\beta_1^{(\mathrm{wz})} and β2(wz)\beta_2^{(\mathrm{wz})} weight the out-of-plane and mixed (EzE_z) contributions.

Zincblende (4‾3m\overline{4}3m) form of BB. For cubic (zincblende) symmetry, the tensor simplifies to

Bzb=(B11B12B12000B12B11B12000B12B12B11000000B44000000B44000000B44),(18)B_{\mathrm{zb}}= \begin{pmatrix} B_{11} & B_{12} & B_{12} & 0 & 0 & 0\\ B_{12} & B_{11} & B_{12} & 0 & 0 & 0\\ B_{12} & B_{12} & B_{11} & 0 & 0 & 0\\ 0 & 0 & 0 & B_{44} & 0 & 0\\ 0 & 0 & 0 & 0 & B_{44} & 0\\ 0 & 0 & 0 & 0 & 0 & B_{44} \end{pmatrix}, \tag*{(18)}

a form commonly used for cubic semiconductors exhibiting zincblende-type symmetry [24, 28].

Wurtzite (6mm6mm) form of BB. For crystals with 6mm6mm symmetry, shear planes containing the c-axis are equivalent, giving

Bwz=(B11B12B13000B12B11B13000B13B13B33000000B44000000B44000000B66).(19)B_{\mathrm{wz}}= \begin{pmatrix} B_{11} & B_{12} & B_{13} & 0 & 0 & 0\\ B_{12} & B_{11} & B_{13} & 0 & 0 & 0\\ B_{13} & B_{13} & B_{33} & 0 & 0 & 0\\ 0 & 0 & 0 & B_{44} & 0 & 0\\ 0 & 0 & 0 & 0 & B_{44} & 0\\ 0 & 0 & 0 & 0 & 0 & B_{66} \end{pmatrix}. \tag*{(19)}

The results of theoretical fits to SHG and 2PPL using the models described above are shown as solid lines in Figure 4.

For zincblende (111), the SHG flower pattern is reproduced well, and the nearly circular 2PPL response is reproduced with ∣β2(zb)∣≪∣β1(zb)∣|\beta_2^{(\mathrm{zb})}| \ll|\beta_1^{(\mathrm{zb})}|, consistent with earlier results on GaAs(111) [23]. Fits to the rod data, on the other hand, require ∣β4(wz)∣≫∣β3(wz)∣|\beta_4^{(\mathrm{wz})}| \gg|\beta_3^{(\mathrm{wz})}| with a finite β2(wz)\beta_2^{(\mathrm{wz})}, consistent with the observed twofold pattern and its axis.

For the rod-like crystals, the agreement between model and experiment is qualitative rather than exact. The calculated pattern reproduces the dominant twofold character of the SHG response, but residual deviations remain in lobe amplitude and angular position (see Figure 4 and Appendix Fig. 7). Several factors may contribute to this mismatch. The analytical treatment of angular response assumes a uniform bulk crystal. It does not explicitly account for finite rod geometry, edge effects, or morphology-induced redistribution of the optical near field. For example, the electric field components polarized perpendicular to the long axis of the rod could be diminished due to geometrical effects [30, 31, 32, 33]. In addition, the rod measurements were more susceptible to degradation during acquisition, and small changes in alignment or sample condition can distort the angular trace. The theoretical fits for Wurtzite should therefore be interpreted as capturing the leading symmetry-derived anisotropy and the exact ratio of different tensor elements in χ(2)\chi^{(2)} or χ(3)\chi^{(3)} (through β\beta terms) can be quantitatively different from a uniform bulk case.

Although the optical trends correlate well with the structural assignments, the nonlinear anisotropy should not be interpreted as a sole phase fingerprint. The phase assignment is supported primarily by lattice-resolved STEM, zone-axis analysis, and, for the rod-like crystals, corroborated by XRD. Although the present data suggests that the crystallographic phase is a dominant factor in determining SHG and 2PPL anisotropy, morphology-related effects, particularly in the rod-like crystals, may contribute to the exact amplitude of lobes forming the measured anisotropy.

Summary

In conclusion, we have investigated the nonlinear optical properties of silver iodide (AgI) thin films using polarization-resolved second harmonic generation and two-photon photoluminescence spectroscopy. By tuning the physical vapor deposition (PVD) conditions, we selectively grew zincblende (γ\gamma-AgI) and wurtzite (β\beta-AgI) phases. Our measurements reveal that triangular γ\gamma-AgI(111) flakes exhibit sixfold SHG and isotropic 2PPL angular patterns, while rod-like β\beta-AgI(101) crystals display a butterfly-shaped SHG and twofold 2PPL angular patterns. These contrasting behaviors are consistent with theoretical analysis and underscore the role of crystal phase in dictating nonlinear optical behavior. The ability to engineer the phase and resultant morphology of AgI provides a promising platform for tailoring optical anisotropy and nonlinear light–matter interactions at the nanoscale.

Acknowledgments

We thank N. Nagaosa for helpful discussions. The work is supported by the Office of the Secretary of Defense for Research and Engineering under Award No. FA9550-23-1-0500 and the MonArk NSF Quantum Foundry from the National Science Foundation under Award No. DMR-1906383. L. L. S. was supported by the University of Arkansas Honors College Research Grant and the Chan and Chen Endowed Research Scholarship.

Transmission Electron Microscopy Analysis

Sample Preparation and Acquisition Conditions

Cross-sectional lamellae from a representative triangular platelet and a representative rod-like crystal were prepared using an FEI Scios DualBeam FIB/SEM (Thermo Fisher Scientific). Final polishing at 2 keV was used to minimize surface damage. STEM analysis was performed at 300 kV on a probe-corrected Titan Themis 80–300 (Thermo Fisher Scientific) with a probe semi-convergence angle of 23.8 mrad and a beam current of 80 pA to reduce beam damage. Elemental maps were acquired using a Bruker SuperX EDS system.

Lattice Indexing and Phase Assignment

Figure 6(d) shows a crystallographic model overlaid on the HRSTEM cross section acquired along the [2 1 2][2\ 1\ 2] zone axis, with lattice planes indexed to (101). The XRD trace in Figure 6(e) corroborates this assignment, which shows a peak at 2θ=25.3∘2\theta= 25.3^\circ consistent with β\beta-AgI(101) and the TEM analysis.

Image not extracted

Figure 6. Structural characterization of AgI nanorods (β\beta-AgI). (a)–(c) SEM images of representative AgI rods grown at 470∘C470^{\circ}\mathrm{C}, showing the elongated rod-like morphology. (d) HRSTEM image of a rod cross section acquired along the [212ˉ][21\bar{2}] zone axis, with lattice planes indexed to the (101)(101) family. (e) X-ray diffraction (XRD) pattern of the rod-like sample. The peak at 2θ=25.3∘2\theta= 25.3^{\circ} is consistent with the β\beta-AgI(101) reflection and agrees with the TEM-based indexing, indicating a preferred (101) orientation. The stronger peak near 2θ=28.5∘2\theta= 28.5^{\circ} is assigned to the Si (111) substrate.

Additional Polarization-Resolved SHG and 2PPL Data

Additional SHG and 2PPL data for AgI nanorods are shown in Figure 7. A predominantly twofold response was observed in thinner rods. The rod-shaped crystals frequently exhibited signs of degradation before 180∘180^\circ, which limited the available angular coverage in some scans [Fig. 7(a)]. The observed twofold pattern is consistent with simulations assuming a β\beta-AgI(101) out-of-plane surface.

Image not extracted

Figure 7. Polarization-resolved SHG and 2PPL from AgI rods. For two representative rod-shaped crystals, polar plots of normalized SHG (blue dots) and 2PPL (pink dots) intensity are shown as a function of in-plane pump-polarization angle δ\delta. SHG exhibits a butterfly-like twofold pattern, while 2PPL shows a uniaxial anisotropy with maximum intensity when the incident polarization is aligned with the long axis of the rod. All traces are normalized to their respective maxima.

Simulation Parameters for SHG

Zincblende

dzb=(000d14000000d14000000d14),d14=1.(20)\mathbf{d}_{\mathrm{zb}} = \begin{pmatrix} 0 & 0 & 0 & d_{14} & 0 & 0 \\ 0 & 0 & 0 & 0 & d_{14} & 0 \\ 0 & 0 & 0 & 0 & 0 & d_{14} \end{pmatrix}, \qquad d_{14}=1. \tag*{(20)}

Wurtzite

dwz=(0000d150000d1500d15d15d33000),(21)\mathbf{d}_{\mathrm{wz}} = \begin{pmatrix} 0 & 0 & 0 & 0 & d_{15} & 0 \\ 0 & 0 & 0 & d_{15} & 0 & 0 \\ d_{15} & d_{15} & d_{33} & 0 & 0 & 0 \end{pmatrix}, \tag*{(21)}

d15=100,d33=1.d_{15}=100,\quad d_{33}=1.

Simulation Parameters for 2PPL

Zincblende

Bzb=(B11B12B12000B12B11B12000B12B12B11000000B44000000B44000000B44).(22)\mathbf{B}_{\mathrm{zb}} = \begin{pmatrix} B_{11} & B_{12} & B_{12} & 0 & 0 & 0 \\ B_{12} & B_{11} & B_{12} & 0 & 0 & 0 \\ B_{12} & B_{12} & B_{11} & 0 & 0 & 0 \\ 0 & 0 & 0 & B_{44} & 0 & 0 \\ 0 & 0 & 0 & 0 & B_{44} & 0 \\ 0 & 0 & 0 & 0 & 0 & B_{44} \end{pmatrix}. \tag*{(22)}

The parameter map used in the simulations was

B11=1,B12=0,B44=0.5,(23)B_{11}=1,\qquad B_{12}=0,\qquad B_{44}=0.5, \tag*{(23)}

with

β1(zb)=1,β2(zb)=0.(24)\beta_{1}^{(\mathrm{zb})}=1,\qquad\beta_{2}^{(\mathrm{zb})}=0. \tag*{(24)}

Wurtzite

Bwz=(B11B12B13000B12B11B13000B13B13B33000000B44000000B44000000B66).(25)\mathbf{B}_{\mathrm{wz}} = \begin{pmatrix} B_{11} & B_{12} & B_{13} & 0 & 0 & 0 \\ B_{12} & B_{11} & B_{13} & 0 & 0 & 0 \\ B_{13} & B_{13} & B_{33} & 0 & 0 & 0 \\ 0 & 0 & 0 & B_{44} & 0 & 0 \\ 0 & 0 & 0 & 0 & B_{44} & 0 \\ 0 & 0 & 0 & 0 & 0 & B_{66} \end{pmatrix}. \tag*{(25)}

The parameter map used in the simulations was

B11=0.5,B12=0.2,B13=−0.2,B33=0.005,B44=0.12,B66=0.1,(26)\begin{aligned} B_{11} &= 0.5,\qquad B_{12}=0.2,\qquad B_{13}=-0.2,\\ B_{33} &= 0.005,\qquad B_{44}=0.12,\qquad B_{66}=0.1, \tag*{(26)} \end{aligned}

with

β1(wz)=0.005,β2(wz)=0.08,(27)\beta_{1}^{(\mathrm{wz})}=0.005,\qquad\beta_{2}^{(\mathrm{wz})}=0.08, \tag*{(27)}
β3(wz)=0.5,β4(wz)=−2.2.\beta_{3}^{(\mathrm{wz})}=0.5,\qquad\beta_{4}^{(\mathrm{wz})}=-2.2.

Details of the Comparative Estimate of χ(2)\chi^{(2)}

To accurately evaluate the second-order nonlinear susceptibility χAgI(2)\chi_{\mathrm{AgI}}^{(2)}, we modeled the SHG intensity using a depth-integrated formalism that accounts for multiple internal reflections within the finite AgI layer [17], followed by a direct comparison with a focused-beam calibration for bulk α\alpha-quartz [18].

Depth-Integrated Model for the AgI Layer

We treat the sample as a multilayer optical stack (Figure 8) consisting of air (0), the AgI layer (LL) of thickness dLd_L, a SiO2_2 oxide layer (oxox) of thickness doxd_{ox}, and the Si substrate (SS). The incident pump field (at frequency ω\omega) and the generated SHG field (at frequency 2ω2\omega) experience multiple internal reflections. Following an approach developed by Lavrov et al. [17], the reflected SHG field amplitude emitted back into the air (x<0x<0) is obtained by integrating the nonlinear polarization across the AgI thickness (0≤x≤dL0 \le x \le d_L):

Image not extracted

Figure 8. Schematic of the multilayer optical geometry used to estimate the effective second-order susceptibility of AgI on SiO2_2/Si. The incident pump field undergoes multiple reflections within the AgI and SiO2_2 layers, modifying the local fundamental field experienced by the AgI crystal. The generated second-harmonic field is emitted from the nonlinear AgI layer and is further affected by propagation and reflection in the optical stack.

E2ω(ref)=∫0dLiωnAgI(2ω)cχAgI(2)Einc2[u+(x)+u−(x)]2F2ω(x) dx(28)E_{2\omega}^{(\mathrm{ref})}=\int_{0}^{d_L}\frac{\mathrm{i}\omega}{n_{\mathrm{AgI}}(2\omega)c}\chi_{\mathrm{AgI}}^{(2)}E_{\mathrm{inc}}^2\left[u^+(x)+u^-(x)\right]^2F_{2\omega}(x)\,\mathrm{d}x \tag*{(28)}

where EincE_{\mathrm{inc}} is the incident pump electric field amplitude, nAgI(2ω)n_{\mathrm{AgI}}(2\omega) is the refractive index of AgI at the SHG frequency, and cc is the speed of light in vacuum. The dimensionless factors u+(x)u^+(x) and u−(x)u^-(x) describe the local enhancement of the forward and backward propagating fundamental fields at depth xx:

u+(x)=t0L(1)eikL(1)x1−rL0(1)rL,ox,S(1)e2ikL(1)dL(29)u^+(x)=\frac{t_{0L}^{(1)}e^{\mathrm{i}k_L^{(1)}x}}{1-r_{L0}^{(1)}r_{L,\mathrm{ox},S}^{(1)}e^{2\mathrm{i}k_L^{(1)}d_L}} \tag*{(29)}
u−(x)=t0L(1)rL,ox,S(1)eikL(1)(2dL−x)1−rL0(1)rL,ox,S(1)e2ikL(1)dL(30)u^-(x)=\frac{t_{0L}^{(1)}r_{L,\mathrm{ox},S}^{(1)}e^{\mathrm{i}k_L^{(1)}(2d_L-x)}}{1-r_{L0}^{(1)}r_{L,\mathrm{ox},S}^{(1)}e^{2\mathrm{i}k_L^{(1)}d_L}} \tag*{(30)}

Here, the superscript (m=1,2m=1,2) denotes the fundamental (ω\omega) and second-harmonic (2ω2\omega) frequencies, respectively. The wavevectors are kj(m)=2πmλnj(mω)k_j^{(m)}=\frac{2\pi m}{\lambda}n_j(m\omega), where λ\lambda is the fundamental vacuum wavelength. The coefficients tij(m)t_{ij}^{(m)} and rij(m)r_{ij}^{(m)} are the standard Fresnel transmission and reflection coefficients at normal incidence between media ii and jj. The effective reflection coefficient from the bottom interface structure (AgI/SiO2_2/Si) is given by:

rL,ox,S(m)=rL,ox(m)+rox,S(m)e2ikox(m)dox1+rL,ox(m)rox,S(m)e2ikox(m)dox(31)r_{L,\mathrm{ox},S}^{(m)}=\frac{r_{L,\mathrm{ox}}^{(m)}+r_{\mathrm{ox},S}^{(m)}e^{2\mathrm{i}k_{\mathrm{ox}}^{(m)}d_{\mathrm{ox}}}}{1+r_{L,\mathrm{ox}}^{(m)}r_{\mathrm{ox},S}^{(m)}e^{2\mathrm{i}k_{\mathrm{ox}}^{(m)}d_{\mathrm{ox}}}} \tag*{(31)}

Similarly, the out-coupling factor F2ω(x)F_{2\omega}(x), which dictates how the SHG field generated at depth xx propagates back to the air, is given by:

F2ω(x)=tL0(2)[eikL(2)x+rL,ox,S(2)eikL(2)(2dL−x)]1−rL0(2)rL,ox,S(2)e2ikL(2)dL(32)F_{2\omega}(x)=\frac{t_{L0}^{(2)}\left[e^{\mathrm{i}k_L^{(2)}x}+r_{L,\mathrm{ox},S}^{(2)}e^{\mathrm{i}k_L^{(2)}(2d_L-x)}\right]}{1-r_{L0}^{(2)}r_{L,\mathrm{ox},S}^{(2)}e^{2\mathrm{i}k_L^{(2)}d_L}} \tag*{(32)}

By defining the SHG intensity in SI units as I(2ω)=12ϵ0c∣E2ω(ref)∣2I(2\omega)=\frac{1}{2}\epsilon_0c\lvert E_{2\omega}^{(\mathrm{ref})}\rvert^2 and isolating the fundamental intensity Iinc(ω)=12ϵ0c∣Einc∣2I_{\mathrm{inc}}(\omega)=\frac{1}{2}\epsilon_0c\lvert E_{\mathrm{inc}}\rvert^2, the conversion efficiency for the AgI stack becomes:

IAgI(2ω)Iinc(ω)2=2ω2c3ϵ0nAgI(2ω)2dL2GMR∣χAgI(2)∣2(33)\frac{I_{\mathrm{AgI}}(2\omega)}{I_{\mathrm{inc}}(\omega)^2}=\frac{2\omega^2}{c^3\epsilon_0n_{\mathrm{AgI}}(2\omega)^2}d_L^2G_{\mathrm{MR}}\lvert\chi_{\mathrm{AgI}}^{(2)}\rvert^2 \tag*{(33)}

where we have factored out the thickness to define the dimensionless optical structure factor GMRG_{\mathrm{MR}}, which captures the total interference effect from the film and the dielectric layers beneath:

GMR=∣1dL∫0dL[u+(x)+u−(x)]2F2ω(x) dx∣2(34)G_{\mathrm{MR}}=\left\lvert\frac{1}{d_L}\int_{0}^{d_L}\left[u^+(x)+u^-(x)\right]^2F_{2\omega}(x)\,\mathrm{d}x\right\rvert^2 \tag*{(34)}

Quartz Calibration and Final Formula

For the bulk α\alpha-quartz reference, the SHG intensity under a focused Gaussian beam geometry is described in SI units by[18]:

Iq(2ω)Iinc(ω)2=128π2f2(N.A.)4cϵ0nq2(nq+1)6∣χq(2)∣2(35)\frac{I_q(2\omega)}{I_{\mathrm{inc}}(\omega)^2}=\frac{128\pi^2f^2(\mathrm{N.A.})^4}{c\epsilon_0n_q^2(n_q+1)^6}\lvert\chi_q^{(2)}\rvert^2 \tag*{(35)}

where f=0.056f=0.056 is the numerical integration constant for a focused Gaussian beam, N.A. is the objective numerical aperture, and nqn_q represents the quartz refractive indices. By taking the ratio of the AgI efficiency to the quartz efficiency, the fundamental constants (ϵ0\epsilon_0, cc, π\pi) cancel. Accounting for the differences in experimental incident pump powers (PAgIP_{\mathrm{AgI}} and PqP_q), the final analytical expression for the absolute susceptibility of AgI is:

∣χAgI(2)∣=∣χq(2)∣[16λ2f2(N.A.)4nAgI(2ω)2dL2GMRnq2(nq+1)6IAgI(2ω)Iq(2ω)(PqPAgI)2]1/2(36)\lvert\chi_{\mathrm{AgI}}^{(2)}\rvert=\lvert\chi_q^{(2)}\rvert\left[\frac{16\lambda^2f^2(\mathrm{N.A.})^4n_{\mathrm{AgI}}(2\omega)^2}{d_L^2G_{\mathrm{MR}}n_q^2(n_q+1)^6}\frac{I_{\mathrm{AgI}}(2\omega)}{I_q(2\omega)}\left(\frac{P_q}{P_{\mathrm{AgI}}}\right)^2\right]^{1/2} \tag*{(36)}

Numerical Evaluation

To compute χAgI(2)\chi_{\mathrm{AgI}}^{(2)}, we evaluated GMRG_{\mathrm{MR}} using the refractive indices, wavelengths, and thicknesses listed in Table 1. The numerical integration of the internal fields yields GMR(zb)=4.626G_{\mathrm{MR}}(\mathrm{zb})=4.626 for the dL=19d_L=19 nm zincblende flake, and GMR(wz)=0.049G_{\mathrm{MR}}(\mathrm{wz})=0.049 for the dL=220d_L=220 nm wurtzite nanorod. The vast reduction in GMR(wz)G_{\mathrm{MR}}(\mathrm{wz}) reflects the severe internal phase-mismatch and Fabry-Perot interference suppressing the out-coupled SHG at that specific thickness.

SymbolQuantityValueUnit
λ\lambdaFundamental wavelength800nm
χq(2)\chi_{\mathrm{q}}^{(2)}Quartz second-order susceptibility0.6 [16]pm/V
nq(ω)n_{\mathrm{q}}(\omega)Quartz refractive index at ω\omega1.4533 [34]–
nq(2ω)n_{\mathrm{q}}(2\omega)Quartz refractive index at 2ω2\omega1.4701 [34]–
nAgI(ω)n_{\mathrm{AgI}}(\omega)AgI refractive index at ω\omega2.045 [35]–
nAgI(2ω)n_{\mathrm{AgI}}(2\omega)AgI refractive index at 2ω2\omega2.385 [35]–
ffNumerical integration constant0.056–
N.A.Objective numerical aperture0.6–
PAgIP_{\mathrm{AgI}}AgI incident pump power1.0mW
PqP_{\mathrm{q}}Quartz incident pump power2.0mW
GMR(zb)G_{\mathrm{MR}}(\mathrm{zb})Zincblende structure factor4.626–
GMR(wz)G_{\mathrm{MR}}(\mathrm{wz})Wurtzite structure factor0.049–
IAgI(zb)(2ω)I_{\mathrm{AgI}(\mathrm{zb})}(2\omega)Zincblende SHG peak maximum2665counts
IAgI(wz)(2ω)I_{\mathrm{AgI}(\mathrm{wz})}(2\omega)Wurtzite SHG peak maximum480counts
Iq(2ω)I_{\mathrm{q}}(2\omega)Quartz SHG peak maximum584counts
dL(zb)d_{\mathrm{L}}(\mathrm{zb})Zincblende AgI thickness19nm
dL(wz)d_{\mathrm{L}}(\mathrm{wz})Wurtzite AgI thickness220nm

Table 1. TABLE I. Constants and parameters used in the χ(2)\chi^{(2)} estimation.

Note. The value of nAgI(ω)n_{\mathrm{AgI}}(\omega) at the fundamental wavelength was obtained by extrapolation of the form

n(λ)=1.93133808+0.05310596λ2+0.01243262λ4,n(\lambda) = 1.93133808 + \frac{0.05310596}{\lambda^{2}} + \frac{0.01243262}{\lambda^{4}},

where λ\lambda is expressed in μm\mu\mathrm{m}.

Using the experimental raw counts (IAgI(zb)=2665I_{\mathrm{AgI}(\mathrm{zb})} = 2665, IAgI(wz)=480I_{\mathrm{AgI}(\mathrm{wz})} = 480, Iq=584I_{\mathrm{q}} = 584) and applying the pump power scaling (Pq/PAgI)2=4(P_{\mathrm{q}}/P_{\mathrm{AgI}})^2 = 4, Eq. (36) directly yields:

χAgI(zb)(2)≈0.45 pm/V,(37)\chi_{\mathrm{AgI}(\mathrm{zb})}^{(2)} \approx0.45\ \mathrm{pm/V}, \tag*{(37)}
χAgI(wz)(2)≈0.16 pm/V.(38)\chi_{\mathrm{AgI}(\mathrm{wz})}^{(2)} \approx0.16\ \mathrm{pm/V}. \tag*{(38)}

These values reflect the intrinsic bulk-like nonlinear response, separated from the geometric optical artifacts caused by the optical interference effects.

References

References

  1. [1]R. W. Boyd, Nonlinear Optics, 4th ed. (Academic Press, 2020).
  2. [2]K. F. Mak, C. Lee, J. Hone, J. Shan, and T. F. Heinz, Atomically thin MoS₂: a new direct-gap semiconductor, Physical Review Letters 105, 136805 (2010).DOI
  3. [3]Y. Li, Y. Rao, K. F. Mak, Y. You, S. Wang, C. R. Dean, and T. F. Heinz, Probing symmetry properties of few-layer MoS₂ and h-BN by optical second-harmonic generation, Nano Letters 13, 3329 (2013).DOI
  4. [4]K. L. Seyler, J. R. Schaibley, P. Gong, P. Rivera, A. M. Jones, S. Wu, J. Yan, D. G. Mandrus, W. Yao, and X. Xu, Electrical control of second-harmonic generation in a WSe₂ monolayer transistor, Nature nanotechnology 10, 407 (2015).
  5. [5]F. Saouma, C. Stoumpos, J. Wong, M. Kanatzidis, and J. Jang, Selective enhancement of optical nonlinearity in two-dimensional organic-inorganic lead iodide perovskites, Nature Communications 8, 742 (2017).DOI
  6. [6]G. Yumoto, F. Harata, T. Nakamura, A. Wakamiya, and Y. Kanemitsu, Electrically switchable chiral nonlinear optics in an achiral ferroelectric 2D van der waals halide perovskite, Science Advances 10, eadq5521 (2024).DOI
  7. [7]J. Xu, X. L. Ji, J. Xiong, C. Yuan, S. Semin, T. Rasing, and X.-H. Bu, Halide perovskites for nonlinear optics, Advanced Materials 32, 1806736 (2020).DOI
  8. [8]A. Goldmann, Band structure and optical properties of tetrahedrally coordinated Cu-and Ag-halides, Phys. Stat. Sol. (B) 81, 9 (1977).DOI
  9. [9]R. Dinges, D. Fröhlich, and C. Uihlein, Two-photon absorption of anisotropic excitons in β-AgI, Phys. Stat. Sol. (B) 76, 613 (1976).
  10. [10]A. Alexander, M. Gupta, D. Kottilil, B. B. Rath, J. J. Vittal, and W. Ji, Anisotropic two-photon absorption and second harmonic generation in single crystals of silver (i) coordination complexes, ACS Appl. Mater. & Interfaces 13, 31891 (2021).DOI
  11. [11]M. Nakamura, Y.-H. Chan, T. Yasunami, Y.-S. Huang, G.-Y. Guo, Y. Hu, N. Ogawa, Y. Chiew, X. Yu, T. Morimoto, et al., Strongly enhanced shift current at exciton resonances in a non-centrosymmetric wide-gap semiconductor, Nature Communications 15, 9672 (2024).
  12. [12]G. Burley, Polymorphism of silver iodide, American Mineralogist: Journal of Earth and Planetary Materials 48, 1266 (1963).
  13. [13]J. Patnaik and C. Sundanara, Studies on gamma silver iodide, Journal of Physics and Chemistry of Solids 59, 1059 (1998).DOI
  14. [14]F. Sabath, C. Aleff, A. Latus, R. Bechstein, and A. Kühnle, Atomic-resolution imaging of the ice nucleating silver iodide surface: Does this polar surface reconstruct at the atomic scale?, Advanced Materials Interfaces 9, 20221065 (2022).DOI
  15. [15]A. Laref, W. Sekkal, A. Zaoui, M. Certier, and H. Aourag, Tight-binding calculations of electronic properties of AgI, Journal of applied physics 86, 4435 (1999).DOI
  16. [16]S. Puri, S. Patel, J. L. Caballos, L. E. Rosas-Hernandez, K. Reynolds, H. O. Churchill, S. Barraza-Lopez, B. S. Mendoza, and H. Nakamura, Substrate interference and strain in the second-harmonic generation from MoSe₂ monolayers, Nano Letters 24, 13061 (2024).DOI
  17. [17]S. D. Lavrov, A. V. Kudryavtsev, A. P. Shestakova, L. Kulyuk, and E. D. Mishina, Second harmonic generation in nanoscale films of transition metal chalcogenides: Taking into account multibeam interference, Optics and Spectroscopy 120, 808 (2016).DOI
  18. [18]L. M. Malard, T. V. Alencar, A. P. M. Barboza, K. F. Mak, and A. M. de Paula, Observation of intense second harmonic generation from MoS₂ atomic crystals, Physical Review B 87, 201401 (2013).DOI
  19. [19]M. Sheik-Bahae, A. A. Said, T.-H. Wei, D. J. Hagan, and E. W. Van Stryland, Sensitive measurement of optical nonlinearities using a single beam, IEEE Journal of Quantum Electronics 26, 760 (1990).
  20. [20]R. DeSalvo, M. Sheik-Bahae, A. A. Said, D. J. Hagan, and E. W. Van Stryland, Z-scan measurements of the anisotropy of nonlinear refraction and absorption in crystals, Optics Letters 18, 194 (1993).pubmed.ncbi.nlm.nih.gov/19802081
  21. [21]L. D. Landau and E. M. Lifshitz, Electrodynamics of Continuous Media, Vol. 8 (Pergamon Press, 1984).
  22. [22]M. A. C. Nascimento, The polarization dependence of two-photon absorption rates for randomly oriented molecules, Chemical Physics 74, 51 (1983).DOI
  23. [23]M. D. Dvorak, W. A. Schroeder, D. R. Andersen, A. L. Smirl, and B. S. Wherrett, Measurement of the anisotropy of two-photon absorption coefficients in zincblende semiconductors, IEEE Journal of Quantum Electronics 30, 256 (1994).DOI
  24. [24]D. C. Hutchings, Theory of the polarization dependence of two-photon absorption, Journal of Modern Optics 41, 2309 (1994).DOI
  25. [25]M. M. Karkhanechi, Polarization dependence of two-photon absorption in an oriented molecular crystal, Applied Optics 36, 7799 (1997).
  26. [26]J. F. Nye, Physical Properties of Crystals: Their Representation by Tensors and Matrices (Oxford University Press, 1985).
  27. [27]A. Yariv and P. Yeh, Optical Waves in Crystals: Propagation and Control of Laser Radiation (Wiley, 1984).
  28. [28]L. Krauss-Kodytek, W.-R. Hannes, T. Meier, C. Ruppert, and M. Betz, Nondegenerate two-photon absorption in ZnSe: experiment and theory, Physical Review B 104, 085201 (2021).
  29. [29]M. Inoue and Y. Toyozawa, Two-photon absorption and energy band structure, Journal of the Physical Society of Japan 20, 363 (1965).
  30. [30]G. Bautista, J. Makitalo, Y. Chen, V. Dhaka, M. Grasso, L. Karvonen, H. Jiang, M. J. Huttunen, T. Huhtio, H. Lipsanen, et al., Second-harmonic generation imaging of semiconductor nanowires with focused vector beams, Nano Letters 15, 1564 (2015).DOI
  31. [31]P. R. Wiecha, A. Arbouet, C. Girard, T. Baron, and V. Paillard, Origin of second-harmonic generation from individual silicon nanowires, Physical Review B 93, 125421 (2016).DOI
  32. [32]L. Neeman, R. Ben-Zvi, K. Rechav, R. Popovitz-Biro, D. Oron, and E. Joselevich, Crystallographic mapping of guided nanowires by second harmonic generation polarimetry, Nano Letters 17, 842 (2017).DOI
  33. [33]M. B. Johnston and H. J. Joyce, Polarization anisotropy in nanowires: Fundamental concepts and progress towards terahertz-band polarization devices, Progress in Quantum Electronics 85, 100417 (2022).DOI
  34. [34]I. H. Malitson, Interspecimen comparison of the refractive index of fused silica, J. Opt. Soc. America 55, 1205 (1965).DOI
  35. [35]G. Ghosh and E. D. Palik, Handbook of Optical Constants of Solids, Five-volume set: Handbook of Thermo-optic Coefficients of Optical Materials with Applications (Elsevier Science, 1997).

Paper details

vlnc:W_e43dxv9e
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents