betaFeedback

WorksPhysics and Astronomy

Anisotropic Third Harmonic Generation in Two-Dimensional Tin Sulfide

Abstract

The in-plane anisotropic properties of two-dimensional (2D) group IV monochalcogenides provide an additional degree of freedom which can be used in future optoelectronic devices. Here, it is shown that the third harmonic generation (THG) signal produced by ultrathin tin (II) sulfide (SnS) is in-plane anisotropic with respect to the incident linear polarization of the laser field. We fit the experimental polarization-resolved THG (P-THG) measurements with a nonlinear optics model, which accounts for the orthorhombic crystal structure of 2D SnS. We calculate the relative magnitudes of the {chi}^(3) tensor components by recording and simultaneously fitting both orthogonal components of the P-THG intensity. Furthermore, we introduce a THG anisotropy ratio, whose calculated values compare the total THG intensity when the excitation linear polarization is along the armchair crystallographic direction with the case when it is along the zigzag direction. Our results provide quantitative information on the anisotropic nature of the THG process in SnS, paving the way to a better understanding of anisotropic nonlinear light-matter interactions, and the development of polarization-sensitive nonlinear optical devices.

Introduction

The nonlinear optical properties of 2D layered materials have been recently attracting considerable interest for both fundamental studies and technological applications, based on light generation of additional frequencies and their modulation, that can also be used for material characterization.[1, 2, 3, 4, 5, 6] Second harmonic generation (SHG) and THG by 2D group-VI transition metal dichalcogenides (TMD) crystals, such as MoS2MoS_2, WS2WS_2, MoSe2MoSe_2 and WSe2WSe_2, have already been at the center of this interest, providing useful information on their structural properties.[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12] For example, THG microscopy has enabled rapid visualization of grain boundaries in monolayer MoS2MoS_2.[11]

Recently, another family of layered 2D materials has been gaining growing attention, namely the group-IV monochalcogenides, also known as group IV-VI metal monochalcogenides, and denoted by MX with M=SnM = Sn or GeGe and X=SX = S or SeSe.[13, 14, 15] MXs are layered, orthorhombic, semiconducting 2D materials, known as phosphorene analogues,[13, 14, 15, 16] given that they share similar puckered or wavy lattice structures with phosphorene, a 2D format of black phosphorus.[17] [18] Importantly, MXs feature in-plane anisotropic physical properties,[13, 14, 15] [19, 20] originating from their in-plane structural anisotropy, with puckered structure along the armchair (AC) crystallographic direction.[15] Properties with reported in-plane anisotropic response include carrier mobility,[19] optical absorption, reflection, extinction, and refraction,[20] and Raman spectral behavior.[13] The in-plane anisotropic response is exhibited along the distinguished in-plane AC and zigzag (ZZ) crystallographic directions, offering an additional degree of freedom in manipulating their properties.[13, 14, 15] [19, 20] For example, polarization-sensitive photodetectors have been presented, based on the intrinsic linear dichroism of GeSeGeSe,[20] and black phosphorus.[21] Furthermore, monolayer MXs are predicted to be multiferroic with coupled ferroelectricity and ferroelasticity, and large spontaneous polarization.[22,23] Indeed, in-plane ferroelectricity has been demonstrated for monolayer SnSSnS at room temperature.[24] Moreover, access to the valley-related degree of freedom has been reported.[25,26]

Recently, studies on the nonlinear optical properties of MXs have also been generating interesting research.[24,27-33] In particular, monolayer MXs have been theoretically predicted to produce giant optical SHG.[27,28] Moreover, polarized SHG spectroscopy on monolayer SnSSnS,[24] efficient and anisotropic SHG in few-layer SnSSnS,[29] SHG imaging of ultrathin SnSSnS,[30] and wavelength-dependent SHG from few-layer ferroelectric SnSSnS,[31] have been reported. Finally, the nonlinear optical absorption properties of SnSSnS,[32] and SnSeSnSe,[33] have been studied, revealing their saturable absorption properties.

However, the THG properties of 2D SnSSnS remain unexplored. THG is a process in which three incident photons with frequency ω\omega generate coherent radiation with frequency 3ω3\omega.[34] Notably, THG, in contrast to SHG, does not require non-centrosymmetry, and thus can also be observed in centrosymmetric crystals. Importantly, polarization-dependent anisotropic THG has been reported for various 2D materials, including germanium selenide (GeSe), [35] which belongs to the family of MXs, black phosphorus (BP),[36-38] silicon phosphide (SiP), [39] germanium arsenide (GeAs) [40], arsenic trisulfide (As2_2S3_3),[41] and rhenium disulfide (ReS2_2).[42]

In this work, we investigate the P-THG process in ultrathin SnS, produced via liquid phase exfoliation (LPE),[30,43-47] and characterized with various techniques to contain monolayer and bilayer crystals (see Experimental Section). Our methodology is based on nonlinear optical imaging, which has been recently demonstrated as a powerful tool to explore the properties of 2D materials.[30,48-53] With respect to the rotating orientation of the excitation linear polarization, the THG signal is found to be in-plane anisotropic. By using a polarizing beam splitter in front of two orthogonally placed detectors, we simultaneously record the intensity of the parallel and perpendicular polarization components of the THG signal. We then simultaneously fit these two sets of experimental P-THG measurements to the theoretical model, obtaining a single set of parameter values, allowing us to calculate the relative magnitudes of the χ(3)\chi^{(3)} tensor components. As we demonstrate, this approach provides increased precision and decreased ambiguity in the above calculation,[54] given the fact that the theoretical model contains five free parameters, and thus, several combinations of values could fit into the model. Indeed, the extraction of the third-order coefficients has been reported in literature as a challenging endeavour, in both 2D materials,[37] and bio-tissues.[55] We also introduce and calculate a THG anisotropy ratio, which compares the total THG intensity when the excitation linear polarization is along the AC direction, to the THG intensity when the polarization is along the ZZ direction. All the above analysis is performed for different 2D SnS crystals belonging in the same field of view. By using laser raster-scanning and the acquisition of spatially resolved THG intensities, forming images, we obtain the means of direct comparison regarding the anisotropic nonlinear optical response between different SnS crystals. Besides this, in case of 2D MXs crystals of larger size, this technique can additionally offer large-area information on the presence of grain boundaries and defects.[11] [48, 49] The demonstrated technique is all-optical, minimally invasive and rapid, and can become a useful tool towards fundamental studies and optoelectronic applications of 2D materials with in-plane anisotropy.

Results and Discussion

Theoretical Formulation of P-THG in 2D SnS

The crystal structure of 2D SnS is schematically illustrated in Figure 1a.[56] In order to describe the interaction of the laser excitation field with an orthorhombic MX crystal, and the generation of the third harmonic field, we employ the Jones formalism.[30,48-53,57] We particularly consider two coordinate systems: the laboratory frame (XX, YY, ZZ), and that defined by the crystal plane (xx, yy, zz), where z∥Zz \parallel Z (Figure 1b). The laser beam propagates along ZZ-axis, normally incident to the crystal, and is linearly polarized along the sample plane, oriented at an angle φ\varphi with respect to XX-axis. In the experiment, the angle φ\varphi is controlled via a rotating halfwaveplate. The xx-axis is considered to be parallel to the AC crystallographic direction and oriented at angle θ\theta with respect to XX-axis (Figure 1b).

Schematic illustration of the crystal structure of orthorhombic 2D SnS and the experimental coordinate systems

Figure 1. a) Schematic illustration of the crystal structure of orthorhombic 2D SnS.[56] b) Schematic illustration of the two coordinate systems adopted in our experimental configuration: the laboratory one (XX, YY, ZZ) and the crystal one (xx, yy, zz), where zz and ZZ coincide. The angles φ\varphi and θ\theta describe the orientation of the laser field, and the AC crystallographic direction, relative to XX laboratory axis, respectively. In the detection path, a polarizing beam splitter (PBS) analyzes the generated THG field into its two orthogonal components, P∥3ωP_{\parallel}^{3\omega} and P⊥3ωP_{\perp}^{3\omega}.

In our system, the excitation field past the half-wave retardation plate can be expressed in laboratory coordinates by the Jones vector (E0cos⁡φE0sin⁡φ)\begin{pmatrix} E_0 \cos\varphi\\ E_0 \sin\varphi\end{pmatrix}, where E0E_0 is the amplitude of the electric field. The expression of this vector in crystal coordinates can be derived by multiplying the excitation field with the rotation matrix (cos⁡θsin⁡θ−sin⁡θcos⁡θ)\begin{pmatrix} \cos\theta& \sin\theta\\ -\sin\theta& \cos\theta\end{pmatrix}, giving Eω=(E0cos⁡(φ−θ)E0sin⁡(φ−θ))E^\omega= \begin{pmatrix} E_0 \cos(\varphi-\theta) \\ E_0 \sin(\varphi-\theta) \end{pmatrix}.

The 2D MXs belong to the orthorhombic point group C2vC_{2v} (mm2),[13] [24, 27] and their contracted third-order nonlinear susceptibility tensor, χ(3)\chi^{(3)}, features the following non-zero elements: χ11,χ16,χ18,χ22,χ24,χ29,χ33,χ35,χ37\chi_{11}, \chi_{16}, \chi_{18}, \chi_{22}, \chi_{24}, \chi_{29}, \chi_{33}, \chi_{35}, \chi_{37}.[34,35,58] In this notation, the first subscript refers to: 1: xx, 2: yy, 3: zz, and the second to: 1: xxxxxx, 2: yyyyyy, 3: zzzzzz, 4: yzzyzz, 5: yyzyyz, 6: xzzxzz, 7: xxzxxz, 8: xyyxyy, 9: xxyxxy, 0: xyzxyz. As a result, the THG equation is expressed as:[34,35,58]

(Px3ωPy3ωPz3ω)=ε0(χ110000χ160χ18000χ220χ240000χ29000χ330χ350χ37000)(ExωExωExωEyωEyωEyωEzωEzωEzω3EyωEzωEzω3EyωEyωEzω3EzωEzωExω3ExωExωEzω3EyωEyωExω3ExωExωEyω6ExωEyωEzω)(1)\begin{pmatrix} P_x^{3\omega}\\ P_y^{3\omega}\\ P_z^{3\omega} \end{pmatrix} = \varepsilon_0 \begin{pmatrix} \chi_{11} & 0 & 0 & 0 & 0 & \chi_{16} & 0 & \chi_{18} & 0 & 0\\ 0 & \chi_{22} & 0 & \chi_{24} & 0 & 0 & 0 & 0 & \chi_{29} & 0\\ 0 & 0 & \chi_{33} & 0 & \chi_{35} & 0 & \chi_{37} & 0 & 0 & 0 \end{pmatrix} \begin{pmatrix} E_x^\omega E_x^\omega E_x^\omega\\ E_y^\omega E_y^\omega E_y^\omega\\ E_z^\omega E_z^\omega E_z^\omega\\ 3E_y^\omega E_z^\omega E_z^\omega\\ 3E_y^\omega E_y^\omega E_z^\omega\\ 3E_z^\omega E_z^\omega E_x^\omega\\ 3E_x^\omega E_x^\omega E_z^\omega\\ 3E_y^\omega E_y^\omega E_x^\omega\\ 3E_x^\omega E_x^\omega E_y^\omega\\ 6E_x^\omega E_y^\omega E_z^\omega \tag*{(1)} \end{pmatrix}

Considering that the pump laser beam is polarized along the sample plane, i.e., Ezω=0E_z^\omega= 0, Equation 1 is reduced to:

(Px3ωPy3ω)=ε0(χ11(Exω)3+3χ18Exω(Eyω)2χ22(Eyω)3+3χ29Eyω(Exω)2)(2)\begin{pmatrix} P_x^{3\omega}\\ P_y^{3\omega} \end{pmatrix} = \varepsilon_0 \begin{pmatrix} \chi_{11}(E_x^\omega)^3 + 3\chi_{18}E_x^\omega(E_y^\omega)^2\\ \chi_{22}(E_y^\omega)^3 + 3\chi_{29}E_y^\omega(E_x^\omega)^2 \tag*{(2)} \end{pmatrix}

and therefore, only terms including four independent χ(3)\chi^{(3)} elements survive. χ11\chi_{11} and χ22\chi_{22} are known as the on-axis χ(3)\chi^{(3)} elements,[41] where the row number equals the column number, and they represent contributions to the THG signal when all three incident fields are parallel to xx-axis (AC direction) or yy-axis (ZZ direction), respectively. On the other hand, χ18\chi_{18} and χ29\chi_{29} are known as the off-axis elements,[41] where the row number is not equal to the column number, and they represent contributions to the THG signal when the incident fields are not all parallel to each other.

We then transform Equation 2 back to laboratory coordinates. By using a polarizing beam splitter in front of the detectors, we collect the two orthogonal components of the THG field, P∥3ω=PX3ωP_{\parallel}^{3\omega}=P_X^{3\omega} and P⊥3ω=PY3ωP_{\perp}^{3\omega}=P_Y^{3\omega}. The intensities of these two orthogonal THG components, I∥3ω=∣P∥3ω∣2I_{\parallel}^{3\omega}=\lvert P_{\parallel}^{3\omega}\rvert^2 and I⊥3ω=∣P⊥3ω∣2I_{\perp}^{3\omega}=\lvert P_{\perp}^{3\omega}\rvert^2, are calculated to be:

I∥3ω=a[(cos⁡2(φ−θ)+3bsin⁡2(φ−θ))cos⁡θcos⁡(φ−θ)−(csin⁡2(φ−θ)+3dcos⁡2(φ−θ))sin⁡θsin⁡(φ−θ)]2(3)I_{\parallel}^{3\omega} = a\left[ \left(\cos^2(\varphi-\theta)+3b\sin^2(\varphi-\theta)\right)\cos\theta\cos(\varphi-\theta) - \left(c\sin^2(\varphi-\theta)+3d\cos^2(\varphi-\theta)\right)\sin\theta\sin(\varphi-\theta) \right]^2 \tag*{(3)}
I⊥3ω=a[(cos⁡2(φ−θ)+3bsin⁡2(φ−θ))sin⁡θcos⁡(φ−θ)+(csin⁡2(φ−θ)+3dcos⁡2(φ−θ))cos⁡θsin⁡(φ−θ)]2(4)I_{\perp}^{3\omega} = a\left[ \left(\cos^2(\varphi-\theta)+3b\sin^2(\varphi-\theta)\right)\sin\theta\cos(\varphi-\theta) + \left(c\sin^2(\varphi-\theta)+3d\cos^2(\varphi-\theta)\right)\cos\theta\sin(\varphi-\theta) \right]^2 \tag*{(4)}

where:

b=χ18χ11,c=χ22χ11,d=χ29χ11(5)b=\frac{\chi_{18}}{\chi_{11}},\quad c=\frac{\chi_{22}}{\chi_{11}},\quad d=\frac{\chi_{29}}{\chi_{11}} \tag*{(5)}

are the relative magnitudes of the χ(3)\chi^{(3)} tensor components, and α\alpha is a multiplication factor that depends on the square of the χ(3)\chi^{(3)} tensor element χ11\chi_{11}, and the amplitude of the electric field. Details on the above calculation are presented in section S1 in Supporting Information.

The total P-THG intensity can then be obtained through:

I3ω=I∥3ω+I⊥3ω(6)I^{3\omega} = I_{\parallel}^{3\omega} + I_{\perp}^{3\omega} \tag*{(6)}

or equivalently:

I3ω=a{cos⁡2(φ−θ)[cos⁡2(φ−θ)+3bsin⁡2(φ−θ)]2+sin⁡2(φ−θ)[csin⁡2(φ−θ)+3dcos⁡2(φ−θ)]2}(7)I^{3\omega} = a\left\{\cos^{2}(\varphi-\theta)\left[\cos^{2}(\varphi-\theta)+3b\sin^{2}(\varphi-\theta)\right]^{2}+\sin^{2}(\varphi-\theta)\left[c\sin^{2}(\varphi-\theta)+3d\cos^{2}(\varphi-\theta)\right]^{2}\right\} \tag*{(7)}

We then present numerical simulations of the theoretical P-THG intensities I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, produced by a 2D MX, described by Equation (3) and Equation (4), respectively. In particular, in Figure S1-Figure S4, Supporting Information, we plot these P-THG modulations in polar diagrams, as function of the orientation of the linearly polarized excitation angle φ\varphi, for different values of the parameters of the model, namely the relative magnitudes of the χ(3)\chi^{(3)} tensor components bb, cc and dd (Figure S1-Figure S3, Supporting Information), and the AC direction θ\theta (Figure S4, Supporting Information). The two P-THG intensity components are predicted to exhibit two-lobe patterns. It is noted that for specific values of these parameters, a change from a two-lobe to a four-lobe pattern could, in principle, be observed (Figure S6, Supporting Information). Furthermore, we find that for different values of these parameters, the shapes of the P-THG modulations change. Considering that I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega} are found to be sensitive to changes of the values of the relative magnitudes of the χ(3)\chi^{(3)} tensor components bb, cc and dd, we conclude that we can calculate the values of these quantities, upon measuring the two P-THG intensity components and subsequently fitting to the above nonlinear optics model.

Image not extracted

Figure S1. Numerical simulations on the effect of different values of the χ(3)\chi^{(3)} tensor ratio bb (defined by Equation 5) on the P-THG intensity components produced by a MX, described by Equation 3 and 4. We plot I∥3ωI_{\parallel}^{3\omega} (left) and I⊥3ωI_{\perp}^{3\omega} (right), in polar diagrams, as function of the orientation of the linearly polarized excitation angle φ\varphi.

Polar diagrams showing the parallel and perpendicular P-THG intensity components for different values of the tensor ratio $d$

Figure S3. Numerical simulations on the effect of different values of the χ(3)\chi^{(3)} tensor ratio dd (defined by Equation (5)) on the P-THG intensity components produced by a MX, described by Equation (3) and Equation (4). We plot I∥3ωI_{\parallel}^{3\omega} (left) and I⊥3ωI_{\perp}^{3\omega} (right), in polar diagrams, as function of the orientation of the linearly polarized excitation angle φ\varphi.

Polar diagrams showing the parallel and perpendicular P-THG intensity components for different AC angles

Figure S4. Numerical simulations on the effect of different values of the AC angle θ\theta on the P-THG intensity components produced by a MX, described by Equation (3) and Equation (4). We plot a) I∥3ωI_{\parallel}^{3\omega} and b) I⊥3ωI_{\perp}^{3\omega}, in polar diagrams, as function of the orientation of the linearly polarized excitation angle φ\varphi.

Numerical simulation of the smallest detectable change in the AC angle of a 2D SnS crystal

Figure S6. Numerical simulation which investigates the smallest detectable change in the AC angle θ\theta of a 2D SnS crystal, based on the P-THG measurements. In order to perform this evaluation, we have generated “ideal” P-THG values, for both I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, using fixed values of the parameters of the model. In these values, we have applied 1000 realizations of Poisson noise, and performed the same simultaneous fitting procedure adopted to fit the real experimental data. However, in this case, only the parameter θ\theta is fitted. In this way, we have extracted the angle θ\theta, for each realization of Poisson noise. Here we plot the standard deviation value, σθ\sigma_{\theta}, that resulted from the 1000 realizations of Poisson noise, as function of the values of the angles θ\theta that were used to generate the simulated data. Results for three different quality of fittings R2R^2 are reported.

Additionally, we have estimated the smallest detectable change in the AC crystallographic direction of a 2D SnS crystal, based on the P-THG measurements. For this purpose, we have performed numerical simulations using “ideal” P-THG data, subject only to Poisson noise, which is commonly used to describe the effect of photon noise in our detectors, namely photomultiplier tubes.[48] See Figure S5 and Figure S6, and the relevant discussion in section S3 in Supporting Information

Numerical simulation of P-THG measurements and fitting curves

Figure S5. Numerical simulation which investigates the precision with which our theoretical model and fitting process fit the data and calculate the AC angle θ\theta. In order to perform this evaluation, we have generated “ideal” P-THG values, for both I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega} (denoted by black dots in the polar plots), using fixed values of the parameters of the model (θ=50∘\theta=50^\circ in this example). In these values, we have applied 1000 realizations of Poisson noise, and performed the same simultaneous fitting procedure adopted to fit the real experimental data, for each realization; however, in this case, only the parameter θ\theta is fitted. In this way, we have extracted the angle θ\theta, reported on the right in red color, calculated as the average value that resulted from the set of the 1000 realizations of Poisson noise; as well the respective fitting curve (denoted by red color in the polar plots). The average quality of fitting R2R^2 is also reported.

The dependency of the two P-THG intensity components to the parameters of the model, namely θ\theta, bb, cc and dd, constitutes a direct link between the P-THG properties and the in-plane anisotropy of orthorhombic MXs. To account for such anisotropy, we introduce a THG anisotropy ratio (AR), which compares the total THG intensities for incident field polarization along the AC and ZZ crystallographic directions. For φ=θ\varphi= \theta, i.e., when the linear polarization of the incident field coincides with the AC direction, we obtain IAC3ω∼χ112I_{AC}^{3\omega} \sim\chi_{11}^{2}. On the other hand, for φ=θ+90∘\varphi= \theta+ 90^{\circ}, i.e., when the linear polarization of the incident field coincides with the ZZ direction, we obtain IZZ3ω∼χ222I_{ZZ}^{3\omega} \sim\chi_{22}^{2}. Therefore, we have:

AR=IZZ3ωIAC3ω=(χ22χ11)2=c2(8)AR = \frac{I_{ZZ}^{3\omega}}{I_{AC}^{3\omega}} = \left(\frac{\chi_{22}}{\chi_{11}}\right)^2 = c^2 \tag*{(8)}

an expression that correlates the AR with the χ(3)\chi^{(3)} tensor ratio cc, which is defined in Equation (5).

We also note that the produced THG field is also linearly polarized at an angle θTHG\theta_{THG} given by:

θTHG=tan⁡−1{χ22tan⁡2(φ−θ)+3χ29χ11+3χ18tan⁡2(φ−θ)tan⁡(φ−θ)}(9)\theta_{THG} = \tan^{-1}\left\{\frac{\chi_{22}\tan^2(\varphi-\theta)+3\chi_{29}}{\chi_{11}+3\chi_{18}\tan^2(\varphi-\theta)}\tan(\varphi-\theta)\right\} \tag*{(9)}

Consequently, by using Equation (9), we could, in principle, calculate the direction θTHG\theta_{THG} of the linear polarization of the THG field.

Experimental P-THG Imaging

In Figure 2, we present a schematic illustration of the experimental setup, used for P-THG imaging of ultrathin SnS (see Experimental Section for a detailed description of the setup, as well as the sample preparation and characterization). In particular, a fs laser beam is focused onto the sample under study, while a pair of galvanometric mirrors is used for laser raster-scanning over the sample area. We record spatially resolved THG images while rotating the angle of the linear polarization of the fundamental beam. In the detection path, we have used a polarizing beam splitter cube, coupled with two orthogonally placed detectors. Based on this experimental configuration, we simultaneously record the intensity of the two orthogonal THG field components, in a single-shot experiment.

Figure 2. Schematic illustration of the nonlinear optical setup, which is based on a fs laser beam coupled to a microscope. The fundamental, 1028 nm, pulses are converted to 1542 nm, by means of an optical parametric oscillator. 2D SnS crystals, excited by the laser beam, generate third harmonic radiation centered at 514 nm. The THG signal is recorded while changing the angle of the linear polarization of the excitation beam, via a rotating half-wave plate, performing P-THG imaging. A pair of galvanometric mirrors enables raster-scanning of stationary SnS crystals, obtaining THG images of the sample area. The THG signal passes through a polarizing beam splitter, allowing the simultaneous measurement of the intensity of the two orthogonal components of the THG field. Inset: Log-scale plot of the THG intensity, produced by an ultrathin SnS crystal, as function of the incident pump power. Black points with the error bars represent the experimental data, and the red line the linear fitting.

In the inset in Figure 2, we present the power-law dependence of the THG intensity produced by an ultrathin SnS crystal, as function of the excitation laser power. The slope 3, in the double log-scale plot, confirms the THG process.[35-42]

In Figure 3, we present a series of images of the intensity of the parallel THG component, I∥3ωI_{\parallel}^{3\omega}, produced by ultrathin SnS crystals in the same field of view, recorded while rotating the pump linear polarization angle φ\varphi, indicated by the orange arrow. As can be observed, the THG signals modulate upon φ\varphi variation. Video S1 (I∥3ωI_{\parallel}^{3\omega}) and S2 (I⊥3ωI_{\perp}^{3\omega}), Supporting Information, show all ninety THG images recorded, with φ∈[0∘,360∘)\varphi\in[0^\circ, 360^\circ) with step of 4∘4^\circ. We mark four regions of interest (ROIs) containing ultrathin SnS crystals. Additionally, we have studied two more similar crystals in a different field of view, namely ROIs 5 and 6 shown in Figure S7, Supporting Information, for a total sample of six ultrathin SnS crystals under study.

Experimental P-THG images of ultrathin SnS crystals

Figure 3. Experimental P-THG images of ultrathin SnS crystals belonging in the same field of view. Particularly, we present the intensity of the parallel component of the THG field, I∥3ωI_{\parallel}^{3\omega}, for different values of the orientation φ\varphi of the linearly polarized excitation field, illustrated by the orange arrows. The value of φ\varphi is varied from 0° to 180° with step of 20°. We mark four ROIs corresponding to ultrathin SnS crystals. Brighter color indicates higher THG intensity.

In Figure 4a,Figure 4b, we present, for the field of view shown in Figure 3, the sum of all ninety collected THG intensity images of I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, respectively. In Figure 4c,Figure 4d, we plot, in polar diagrams, the modulation of I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, respectively, for the SnS crystals in ROIs 1 and 3, as function of the angle φ\varphi of the linear polarization of the excitation beam. The two P-THG intensity components are found to exhibit two-lobe patterns, complying with the theoretical prediction. Furthermore, it is evident that these two SnS crystals produce different P-THG modulations (the one is rotated with respect to the other), in accordance with the theoretical prediction. This difference is the signature of nonlinear optical anisotropy between the two crystals.

Sum of parallel and perpendicular P-THG intensities and experimental polar plots

Figure 4. Sum of a) I∥3ωI_{\parallel}^{3\omega} and b) I⊥3ωI_{\perp}^{3\omega} P-THG intensities, for all orientations φ\varphi of the excitation linear polarization, corresponding to the field of view shown in Figure 3. We mark the four ROIs we have focused in our analysis, corresponding to ultrathin SnS crystals present in the same field of view. Brighter color indicates higher THG intensity. Experimental polar plots of c) I∥3ωI_{\parallel}^{3\omega} and d) I⊥3ωI_{\perp}^{3\omega} P-THG intensities as function of the angle φ\varphi, comparing the P-THG modulations in ROI 1 (in black) and ROI 3 (in magenta).

Experimental Fitting Analysis

We then fit the P-THG experimental data with our theoretical model, to extract the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc, and dd. It is important to note that fitting only one P-THG intensity component (either I∥3ωI_{\parallel}^{3\omega} or I⊥3ωI_{\perp}^{3\omega}) is not sufficient to obtain reliable fitting results. This is clearly demonstrated in Figure S8, Supporting Information, where, for ROI 3, we have fitted only one intensity component, namely I⊥3ωI_{\perp}^{3\omega} with Equation (4), and then tested the fitting values to the other component, namely I∥3ωI_{\parallel}^{3\omega} through Equation (3). It is observed that although there is a set of parameter values that fit very well I⊥3ωI_{\perp}^{3\omega} (quality of fitting 99.3%), the same set does not fit I∥3ωI_{\parallel}^{3\omega} (quality of fitting 32.2%), producing a four-lobe pattern instead of a two-lobe one.

Figure S8. Example showing experimentally that fitting only one P-THG intensity component (either I∥3ωI_{\parallel}^{3\omega} or I⊥3ωI_{\perp}^{3\omega}) is not sufficient to obtain reliable fitting results. On the left, we show THG images from the same field of view, where the THG intensity is integrated for all orientations φ\varphi of the excitation linear polarization. In the polar plots (center), black points represent the experimental P-THG data, for the SnS crystal shown with the red square. Here we have fitted I⊥3ωI_{\perp}^{3\omega} with Equation (4) (red curve), and then we have tested the fitting parameter values (reported on the right) on I∥3ωI_{\parallel}^{3\omega}, through Equation (3) (green curve). Although this set of fitting parameter values fits very well the one component (quality of fitting 99.3%), it does not fit the other (quality of fitting 32.2%), producing a four-lobe pattern instead of a two-lobe one.

In this work, we have resolved this issue, by simultaneously fitting both I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, obtaining one set of parameter values, with quality of fitting, R2R^2, equal to the average (R2=(R∥2+R⊥2)/2R^2 = (R_{\parallel}^2 + R_{\perp}^2)/2). It is noted that all of the experimental fitting results presented here exhibit average quality of fitting larger than 85%. We also note that the whole procedure of loading and simultaneously fitting both experimental datasets, for an individual crystal, requires less than one minute, rendering our methodology a rapid characterization tool (see also Experimental Section for the data analysis). Following the above approach, in Figure 5 and S9, Supporting Information, we present the simultaneous fitting to both P-THG intensity components, for all six ultrathin SnS crystals under study. The agreement between the experimental data and the theoretical fitting is remarkable.

Figure 5. Experimental P-THG intensity modulations, I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, as function of the linearly polarized excitation angle φ\varphi, presented in polar plots, for the ultrathin SnS crystals 1-4, corresponding to the ROIs marked in Figure 4. Black points represent the experimental data, while the blue and red lines represent the simultaneous fitting with Equation (3) (I∥3ωI_{\parallel}^{3\omega}) and 4 (I⊥3ωI_{\perp}^{3\omega}), respectively. Using this fitting process, we are able to calculate the relative magnitudes of the χ(3)\chi^{(3)} tensor components, for each case, summarized in Table 1.

In Table 1, we summarize the fitting values for the relative magnitudes of the χ(3)\chi^{(3)} tensor components, for each SnS crystal under study. We also report the average values based on this sample of crystals, along with the standard deviation (std) values, as the errors of the average values. In order to obtain a graphical perspective of these average results, in Figure S10, Supporting Information, we plot them as Gaussian distributions, with the μ\mu and σ\sigma Gaussian parameters equal to the average and std values, respectively.

Crystalb=χ18/χ11b = \chi_{18}/\chi_{11}c=χ22/χ11c = \chi_{22}/\chi_{11}d=χ29/χ11d = \chi_{29}/\chi_{11}
10.27±0.020.27 \pm0.020.99±0.020.99 \pm0.020.46±0.020.46 \pm0.02
20.40±0.020.40 \pm0.020.93±0.030.93 \pm0.030.49±0.030.49 \pm0.03
30.46±0.020.46 \pm0.021.00±0.031.00 \pm0.030.40±0.020.40 \pm0.02
40.21±0.020.21 \pm0.020.65±0.050.65 \pm0.050.16±0.010.16 \pm0.01
50.16±0.020.16 \pm0.020.74±0.030.74 \pm0.030.20±0.010.20 \pm0.01
60.09±0.010.09 \pm0.010.82±0.020.82 \pm0.020.40±0.020.40 \pm0.02
Average0.26±0.130.26 \pm0.130.86±0.130.86 \pm0.130.35±0.130.35 \pm0.13

Table 1.

Gaussian distributions of the relative magnitudes of the tensor components

Figure S10.

Table 1. Summary of the values of the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc and dd, which we have experimentally calculated by simultaneously fitting both P-THG intensity components, I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, with Equation (3) and 4, respectively. We present results for the ultrathin SnS crystals 1-6, corresponding to the ROIs marked in Figure 4 and S7, Supporting Information. The average values, along with the std values, as the errors of the average values, are also reported.

The results of Table 1 indicate the anisotropic nature of the χ(3)\chi^{(3)} tensor and the THG process in ultrathin SnS crystals. We conclude that the on-axis nonlinear susceptibility tensor elements, χ11\chi_{11} and χ22\chi_{22}, are found to be considerably larger with respect to the off-axis elements, χ18\chi_{18} and χ29\chi_{29}. The latter, expressed through bb and dd, are found to feature relatively similar magnitudes. These experimental results could be useful in future theoretical studies to evaluate or constrain theoretical models for THG in MXs.

It is noted that the values of the relative magnitudes of the χ(3)\chi^{(3)} tensor components exhibit variations among different crystals. Such variations have also been observed in other 2D materials, and have been attributed to deformation or defects in the crystal lattice, introduced during the exfoliation process.[35,40,41]

For reference, in Table S1, Supporting Information, we present a summary of literature reports, to our knowledge, with experimental results on the relative magnitudes of the χ(3)\chi^{(3)} tensor components, for other 2D materials. We observe that our conclusions are in qualitative agreement with most literature findings.

CrystalCrystal structureb=χ18/χ11b = \chi_{18}/\chi_{11}c=χ22/χ11c = \chi_{22}/\chi_{11}d=χ29/χ11d = \chi_{29}/\chi_{11}
SnS (This work)Orthorhombic0.26±0.130.26 \pm0.130.86±0.130.86 \pm0.130.35±0.130.35 \pm0.13
GeSe (ref. [2]) (*)Orthorhombic0.26±0.010.26 \pm0.010.54±0.060.54 \pm0.060.4±0.10.4 \pm0.1
Black phosphorus (ref. [3])Orthorhombic0.460.460.50.50.350.35
Black phosphorus (ref. [4])OrthorhombicAlmost zeroComparable with ddComparable with cc
SiP (ref. [5])Orthorhombic-0.72±0.080.72 \pm0.08-
GeAs (ref. [6]) (*)Monoclinic0.21±0.040.21 \pm0.040.48±0.040.48 \pm0.040.19±0.010.19 \pm0.01
As2_2S3_3 (ref. [7])Monoclinic0.290.290.600.600.290.29

Table S1. Summary of literature reports, to our knowledge, with experimental results on the relative magnitudes of the χ(3)\chi^{(3)} tensor components, for other 2D materials [2, 3, 4, 5, 6, 7]. (*) For GeSe and GeAs, the values we present here are the average values we have calculated based on a sample of two (ref. [2]) and seven (ref. [6]) crystals, respectively, which are reported in the corresponding studies (along with the std values as the errors of the average values).

In Figure 6, we present both P-THG intensity components, together with the total P-THG intensity, I3ω=I∥3ω+I⊥3ωI^{3\omega} = I_{\parallel}^{3\omega} + I_{\perp}^{3\omega}, for the ultrathin SnS crystal which corresponds to the red dashed square shown in Figure 4. In Figure S11, Supporting Information, we present the corresponding polar diagram for ROI 4 shown in Figure 4. The I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega} experimental data are simultaneously fitted with (3) and 4, respectively, while the I3ωI^{3\omega} data are fitted with (6). The total P-THG intensity is found to be clearly anisotropic. Specifically, different angles φ\varphi of the excitation linear polarization produce different THG intensities. Therefore, the above result confirms the in-plane anisotropic nature of the THG process in ultrathin SnS crystals. Furthermore, it offers the possibility of controlling the emitted THG intensity by tuning the angle φ\varphi.

Image not extracted

Figure 6. Experimental P-THG intensity modulations as function of the linearly polarized excitation angle φ\varphi, presented in a polar plot, for the ultrathin SnS crystal which corresponds to the red dashed square shown in Figure 4. Blue and red points represent I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, respectively, while the black points represent the total P-THG intensity, I3ω=I∥3ω+I⊥3ωI^{3\omega} = I_{\parallel}^{3\omega} + I_{\perp}^{3\omega}. The solid curves represent the respective theoretical fittings, with Equation (3), 4 and 6.

Polar plot of P-THG intensity components and total P-THG intensity

Figure S11.

THG Anisotropy Ratio

Furthermore, through Equation (8), we have introduced a THG AR, which compares the total THG intensities for incident field polarization along the main crystallographic axes AC and ZZ. This AR has been correlated with the χ(3)\chi^{(3)} tensor ratio cc, through Equation (8), and thus, can be directly calculated using the fitting results of cc, reported in Table 1. The results are summarized in Table 2, for each ultrathin SnS crystal under study. The AR is found to be, on average, not equal to unity, suggesting that the total THG intensity is not equal when the excitation is along the AC direction compared to the case when it is along the ZZ direction. This AR allows to compare and classify different orthorhombic 2D materials, based on their degree of nonlinear optical anisotropy.

CrystalAnisotropy ratio (AR)
10.98±0.040.98 \pm0.04
20.86±0.060.86 \pm0.06
31.00±0.061.00 \pm0.06
40.42±0.070.42 \pm0.07
50.55±0.040.55 \pm0.04
60.67±0.030.67 \pm0.03
Average0.75±0.220.75 \pm0.22

Table 2. Summary of the values we have experimentally calculated for the AR, through Equation (8). We present results for the ultrathin SnS crystals 1-6, corresponding to the ROIs marked in Figure 4 and S5, Supporting Information. The average value, along with the std value, as the error of the average value, are also reported.

Finally, we address the effect of the thickness of a SnS crystal on its THG intensity. This can be realized by using a constructive interference model, where each layer contributes constructively in the detected THG signal.[30,50,59] In this context, in the case of a N-layer SnS crystal with zero twist-angle, the produced THG intensity is found to be analogous to N2N^2, where N is the number of layers (see Equations (15)-(19), Supporting Information). This analysis allows to estimate the thickness of the SnS crystals (see Table S2 and the relevant discussion in § 17 in Supporting Information).

CrystalMax. THG intensity (×105\times10^{5}) (arb. units)Number of layers NN
116.42.9
21.91
317.93.1
419.83.2
59.42.2
66.51.8

Table S2. Summary of the maximum THG intensity detected from the ultrathin SnS crystals 1-6, corresponding to the ROIs marked in Figure 4 and S7. As monolayer is characterized the crystal with the minimum intensity, i.e., ROI 2, while the number of layers NN for the other crystals is estimated using Equation (19).

Conclusion

In conclusion, we have presented an all-optical, minimally invasive and rapid methodology, based on P-THG nonlinear optical imaging, to characterize the anisotropic properties of ultrathin SnS crystals, produced via LPE. We observe a variation of THG emission intensities for different polarization angles of the excitation field, demonstrating the in-plane anisotropic nature of the THG process. By using a polarizing beam splitter and two orthogonally placed detectors, we simultaneously record the intensity of the two orthogonal components of the THG field, while rotating the direction of the linear polarization of the excitation beam, enabling P-THG imaging. We then simultaneously fit the intensities of the two orthogonal components with a nonlinear optics model, which accounts for the orthorhombic crystal structure of 2D SnS. This approach enables the calculation of the relative magnitudes of the χ(3)\chi^{(3)} tensor components with high precision. Our results indicate that the on-axis nonlinear susceptibility tensor elements, χ11\chi_{11} and χ22\chi_{22}, are considerably larger than the off-axis elements, χ18\chi_{18} and χ29\chi_{29}, while the latter feature relatively similar magnitudes. We have also introduced and calculated a THG AR, which compares the total THG intensity upon excitation parallel to the AC direction compared to the case where it is parallel to the ZZ. Our results provide quantitative information on the effect of the in-plane anisotropy of ultrathin SnS on its nonlinear optical properties. We envisage that this work can introduce a useful method for studying the THG response of in-plane anisotropic 2D materials, towards fundamental studies, as well as the realization of polarization-sensitive nonlinear optical devices.

Experimental Section

Nonlinear Optical Imaging Setup and Analysis: The nonlinear optical imaging setup is based on a fs laser beam coupled to a microscope (Figure 2). The 1028 nm fs laser beam (FLINT FL1 Yb Oscillator, 1028 nm, ≈76 MHz, ≈36 fs, Light Conversion) passes through an optical parametric oscillator (APE Levante IR) which converts it to 1542 nm. The angle of the linear polarization of the excitation field is rotated, performing P-THG imaging, using an achromatic half-wave plate (AHWP10M-1600, Thorlabs), placed in a motorized rotation stage. The beam is guided into an inverted microscope (Axio Observer Z1, Carl Zeiss) by a pair of silver-coated galvanometric (galvo) mirrors (6215H, Cambridge Technology), allowing to raster-scan stationary samples. A pair of achromatic doublet lenses suitably expands the beam diameter to fill the back aperture of the objective lens (Plan-Apochromat 20×/0.820\times/0.8 NA, Carl Zeiss).

At the motorized turret box of the microscope, at 45∘45^\circ just below the objective lens, a silver mirror is used, which is insensitive to the polarization state of the laser beam. The objective lens tightly focuses the beam onto the sample, which, following light-matter interaction, produces THG radiation, which is collected in the forward detection geometry by a condenser lens (achromatic-aplanatic, 1.4 NA, Carl Zeiss). A polarizing beam splitter cube (CCM1-PBS251, Thorlabs) analyzes the THG signal into two orthogonal components, which are filtered by suitable short-pass (FF01-680/SP, Semrock) and narrow bandpass (FF01-514/3, Semrock) filters, placed just in front of the detectors, to cut off residual laser light and any other unwanted signal. The two orthogonally placed detectors, collecting the THG radiation, are based on photomultiplier tube modules (H9305-04, Hamamatsu). The diffraction-limited spatial resolution is ≈1.176 μm (0.61λexc/NA0.61\lambda_{\mathrm{exc}}/\mathrm{NA}, with λexc=1542\lambda_{\mathrm{exc}} = 1542 nm, NA = 0.8).

The galvanometric mirrors and the photomultiplier tubes are connected to a connector block (BNC-2110, National Instruments Austin), which is interfaced to a PC through a DAQ (PCI 6259, National Instruments). The coordination of the detector recordings with the galvanometric mirrors for the image formation, as well as the movement of the motors, are carried out using LabView (National Instruments).

Each image presented here consists of 500×500500 \times500 pixels. For the data analysis, the open-source Python programming language,[60] and the open-source ImageJ image analysis software,[61] are used. We note that in all experimental polar plots presented in this work: i) each ROI of pixels is treated as one pixel with intensity equal to the mean intensity in the ROI, and ii) we have selected a sample region without SnS crystals, whose intensity has been considered as noise and has been subtracted from our measurements. Regarding the adopted fitting procedure, the relative magnitudes of the χ(3)\chi^{(3)} tensor components are bounded in [0,1][0,1]. Furthermore, experimentally, the throughput of the parallel and perpendicular P-THG intensities is different, and therefore, different scaling factors, α\alpha, were used during the fitting procedure.

Sample Preparation and Characterization: The LPE method is employed to isolate an ultrathin layer of SnS sheets;[30,43-47] see Supporting Information. The isolated ultrathin SnS sheets are characterized with UV-Vis spectroscopy, atomic force microscopy (AFM), Raman spectroscopy, and photoluminescence (PL) spectroscopy.

Figure S12, Supporting Information, represents the UV-Vis extinction spectra of isolated SnS dispersion. The recorded extinction spectra of SnS exhibit a broad absorption profile in UV-Vis-NIR range, with a shoulder at 420 nm. The displayed extinction profile is consistent with the previously reported results.[43-46] This observation suggests the successful isolation of ultrathin layer of SnS sheets.

UV-Vis extinction spectra and AFM topography image of isolated SnS sheets

Figure S12. a) UV-Vis extinction spectra of isolated SnS sheets (in acetone). b) AFM topography image of SnS sheets. The scale bar illustrates 330 nm.

Moreover, AFM images are acquired to investigate the dimensionality of the isolated SnS sheets (Figure S12, Supporting Information). The height profiles in the AFM images vary from 0.8 nm to 1.4 nm. The thickness of monolayer SnS is reported to be below 1 nm,[46] while the AFM measurements of liquid phase exfoliated SnS crystals have been reported to overestimate their thickness because of the solvent overlayer.[43] Therefore, our AFM measurements immensely suggest the presence of monolayer and bilayer crystals.

Furthermore, Raman spectroscopy is employed to quantify the layer number of the SnS crystals. Raman spectra of isolated SnS sheets reveal the expected optically active phonon modes;[45,46,61] Figure S13, Supporting Information. Compared to the bulk crystal, the Raman peaks of the phonon modes of the ultrathin SnS sheets feature spectral blueshifts (Figure S13, Supporting Information), which are consistent with our previous results of monolayer SnS sheets;[46] (see Supporting Information).

Raman and $\mu$-PL spectra of SnS

Figure S13. a) Raman spectra of bulk and monolayer SnS. b) μ\mu-PL spectra of isolated SnS monolayers. The spectra are collected from the cluster/ensemble of the SnS monolayers in reflection mode. The laser excitation source is 543 nm (2.28 eV), with laser spot size ≈1 μm\approx1\,\mu\mathrm{m} and incident laser power 640 μ\muW.

Finally, PL spectra of isolated thin layer SnS sheets are recorded, depicting a single broad emission peak, centered at ≈1.95\approx1.95 eV (Figure S13, Supporting Information). The peak center of the PL emission closely matches the energy gap obtained from monolayer SnS.[46] All the above characterization techniques are further discussed in the Supporting Information.

Supporting Information

Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgments

E.S., G.K., L.M., S.P., and G.M.M. acknowledge financial support by the EU-funded DYNASTY project, ID: 101079179, under the Horizon Europe framework programme. E.S. and S.P. acknowledge funding by the Hellenic Foundation for Research and Innovation (H.F.R.I.) in the context of the project "Investigation of the protective roles of caloric restriction mimetics in myelin disruption via a novel, non-invasive advanced imaging approach_ID14772_ΜΑΥΑ". G.M.M. thanks Artem Larin (ITMO University) for a useful discussion on analyzing data related with nonlinear optical processes; Dionysios Xydias (IESL-FORTH) and Maria Kefalogianni (IESL-FORTH) for their contribution while sharing the nonlinear optical setup; and Matina Vlahou (IESL-FORTH) for designing the illustration of the nonlinear optical setup.

Author Contributions

S.P., L.M., E.S., and G.K. guided the research. G.M.M. and S.P. conducted the nonlinear optical experiments. L.M., G.M.M., and S.P. derived the theoretical equations. G.M.M., S.P., L.M., and A.L. conducted the data analysis. A.S.S. prepared and characterized the samples. A.L. provided technical support. G.M.M., L.M., S.P., E.S., and A.S.S. prepared the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supporting Information

Theoretical formulation of P-THG in 2D SnS

The THG equation in two-dimensional (2D) SnS is expressed in crystal coordinates as in Equation (2), main manuscript.

We then transform Equation (2), main manuscript, back to laboratory coordinates by multiplying with the rotation matrix (cos⁡θ−sin⁡θsin⁡θcos⁡θ)\begin{pmatrix} \cos\theta& -\sin\theta\\ \sin\theta& \cos\theta\end{pmatrix}, obtaining:

(PX3ωPY3ω)∼((χ11cos⁡2(φ−θ)+3χ18sin⁡2(φ−θ))cos⁡θcos⁡(φ−θ)−(χ22sin⁡2(φ−θ)+3χ29cos⁡2(φ−θ))sin⁡θsin⁡(φ−θ)(χ11cos⁡2(φ−θ)+3χ18sin⁡2(φ−θ))sin⁡θcos⁡(φ−θ)+(χ22sin⁡2(φ−θ)+3χ29cos⁡2(φ−θ))cos⁡θsin⁡(φ−θ))(10)\begin{pmatrix} P_{X}^{3\omega} \\ P_{Y}^{3\omega} \end{pmatrix} \sim \begin{pmatrix} \left(\chi_{11}\cos^{2}(\varphi-\theta)+3\chi_{18}\sin^{2}(\varphi-\theta)\right)\cos\theta\cos(\varphi-\theta)-\left(\chi_{22}\sin^{2}(\varphi-\theta)+3\chi_{29}\cos^{2}(\varphi-\theta)\right)\sin\theta\sin(\varphi-\theta) \\ \left(\chi_{11}\cos^{2}(\varphi-\theta)+3\chi_{18}\sin^{2}(\varphi-\theta)\right)\sin\theta\cos(\varphi-\theta)+\left(\chi_{22}\sin^{2}(\varphi-\theta)+3\chi_{29}\cos^{2}(\varphi-\theta)\right)\cos\theta\sin(\varphi-\theta) \tag*{(10)} \end{pmatrix}

By using a polarizing beam splitter in front of the detectors, we collect the two orthogonal components of the THG field. These components, P∥3ω=PX3ωP_{\parallel}^{3\omega}=P_{X}^{3\omega} and P⊥3ω=PY3ωP_{\perp}^{3\omega}=P_{Y}^{3\omega}, can be described by multiplying with the Jones matrix (cos⁡2ζsin⁡ζcos⁡ζsin⁡ζcos⁡ζsin⁡2ζ)\begin{pmatrix} \cos^{2}\zeta& \sin\zeta\cos\zeta\\ \sin\zeta\cos\zeta& \sin^{2}\zeta\end{pmatrix} of a linear polarizer. Here, ζ\zeta is the angle of the transmission axis of the polarizer with respect to XX-axis, and we set ζ=0∘\zeta=0^{\circ} for P∥3ωP_{\parallel}^{3\omega}, and ζ=90∘\zeta=90^{\circ} for P⊥3ωP_{\perp}^{3\omega}, i.e., axis of transmission parallel to XX and YY-axis, respectively. Then, the intensities of these two orthogonal THG components, I∥3ω=∣P∥3ω∣2I_{\parallel}^{3\omega}=|P_{\parallel}^{3\omega}|^{2} and I⊥3ω=∣P⊥3ω∣2I_{\perp}^{3\omega}=|P_{\perp}^{3\omega}|^{2}, are calculated to be:

I∥3ω∼[(χ11cos⁡2(φ−θ)+3χ18sin⁡2(φ−θ))cos⁡θcos⁡(φ−θ)−(χ22sin⁡2(φ−θ)+3χ29cos⁡2(φ−θ))sin⁡θsin⁡(φ−θ)]2(11)I_{\parallel}^{3\omega}\sim\left[\left(\chi_{11}\cos^{2}(\varphi-\theta)+3\chi_{18}\sin^{2}(\varphi-\theta)\right)\cos\theta\cos(\varphi-\theta)-\left(\chi_{22}\sin^{2}(\varphi-\theta)+3\chi_{29}\cos^{2}(\varphi-\theta)\right)\sin\theta\sin(\varphi-\theta)\right]^{2} \tag*{(11)}
I⊥3ω∼[(χ11cos⁡2(φ−θ)+3χ18sin⁡2(φ−θ))sin⁡θcos⁡(φ−θ)+(χ22sin⁡2(φ−θ)+3χ29cos⁡2(φ−θ))cos⁡θsin⁡(φ−θ)]2(12)I_{\perp}^{3\omega}\sim\left[\left(\chi_{11}\cos^{2}(\varphi-\theta)+3\chi_{18}\sin^{2}(\varphi-\theta)\right)\sin\theta\cos(\varphi-\theta)+\left(\chi_{22}\sin^{2}(\varphi-\theta)+3\chi_{29}\cos^{2}(\varphi-\theta)\right)\cos\theta\sin(\varphi-\theta)\right]^{2} \tag*{(12)}

These relationships are expressed in terms of the absolute values of the χ(3)\chi^{(3)} tensor elements. Instead, they can be expressed in terms of dimensionless ratios of the χ(3)\chi^{(3)} tensor elements, obtaining Equation (3) and (4), main manuscript.

We also note that the produced THG field is also linearly polarized at an angle θTHG\theta_{THG} given by:

θTHG=tan⁡−1{χ22tan⁡2(φ−θ)+3χ29χ11+3χ18tan⁡2(φ−θ)tan⁡(φ−θ)}(13)\theta_{THG}=\tan^{-1}\left\{\frac{\chi_{22}\tan^{2}(\varphi-\theta)+3\chi_{29}}{\chi_{11}+3\chi_{18}\tan^{2}(\varphi-\theta)}\tan(\varphi-\theta)\right\} \tag*{(13)}

which for the special case of φ=θ+45∘\varphi=\theta+45^{\circ} yields:

θTHG=tan⁡−1{χ22+3χ29χ11+3χ18}(14)\theta_{\mathrm{THG}}=\tan^{-1}\left\{\frac{\chi_{22}+3\chi_{29}}{\chi_{11}+3\chi_{18}}\right\} \tag*{(14)}

Numerical simulations on the effect of different values of the parameters on the P-THG intensity components in MXs

Varying the ratios of the χ(3)\chi^{(3)} tensor components

In the following figures, we present numerical simulations on the effect of different values of the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb (Figure S1), cc (Figure S2) and dd (Figure S3), defined by (5), on the polarization-resolved third harmonic generation (P-THG) intensity components produced by a group IV monochalcogenide (MX), described by (3) and 4. In each case, we vary only one ratio of the χ(3)\chi^{(3)} tensor components as reported in each figure, while keeping the rest ratios fixed to the average values we have eventually, experimentally calculated from our analysis, reported in Table 1. In all cases, the armchair (AC) direction θ\theta is kept fixed to 50°.

Polar diagrams showing the parallel and perpendicular P-THG intensity components for different values of the tensor ratio $c$

Figure S2. Numerical simulations on the effect of different values of the χ(3)\chi^{(3)} tensor ratio cc (defined by Equation (5)) on the P-THG intensity components produced by a MX, described by Equation (3) and Equation (4). We plot I∥3ωI_{\parallel}^{3\omega} (left) and I⊥3ωI_{\perp}^{3\omega} (right), in polar diagrams, as function of the orientation of the linearly polarized excitation angle φ\varphi.

Image not extracted

Figure 7 (PDF p. 23).

Varying the armchair angle

In Figure S4, we present numerical simulations on the effect of different values of the AC angle θ\theta on the P-THG intensity components produced by a MX, described by Equation (3) and Equation (4). The relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc and dd are kept fixed to the average values we have eventually, experimentally calculated from our analysis, reported in Table 1.

Estimation of the smallest detectable change in the armchair angle

Here, we aim to estimate the smallest detectable change in the AC crystallographic direction of a 2D SnS crystal, based on the P-THG measurements. For this purpose, we have performed numerical simulations using “ideal” P-THG data, subject only to Poisson noise, which is commonly used to describe the effect of photon noise in our detectors, namely photomultiplier tubes.[1] In particular, we have generated ninety P-THG values, for both I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega} through (3) and 4, respectively, for φ∈[0∘,360∘)\varphi\in[0^\circ, 360^\circ) with step of 4∘4^\circ, using fixed values of the parameters of the model, namely: the AC angle, the scaling factors α\alpha (a∥=2a_{\parallel}=2 and a⊥=1a_{\perp}=1), and the relative magnitudes of the χ(3)\chi^{(3)} tensor components b, c and d (equal to the average values we have eventually, experimentally calculated from our analysis, reported in Table 1). In these “ideal” P-THG values, we have applied 1000 realizations of Poisson noise, and then performed the same simultaneous fitting procedure adopted to fit the real experimental data, for each realization. However, in this case, only the parameter θ\theta is fitted, and not α\alpha, b, c and d. In this way, we have extracted the angle θ\theta, for each realization of Poisson noise. We have repeated this analysis for various angles θ\theta. An example is presented in Figure S5, for θ=50∘\theta=50^\circ.

In the context we have described, as the smallest detectable change in the AC angle, we have utilized the standard deviation value, σθ\sigma_\theta, that resulted from the 1000 realizations of Poisson noise. For “ideal” data (quality of fitting R2=100%R^2=100\%), σθ\sigma_\theta is found to be zero (blue points in Figure S6). Deviation from the “ideal” data, manifested through the Poisson noise, provides error in the calculation of the angle θ\theta. Interestingly, σθ\sigma_\theta is found to depend on the angle θ\theta. It is estimated to be in the range 1.5∘<σθ≤2.3∘1.5^\circ< \sigma_\theta\le2.3^\circ for quality of fitting R2>92%R^2>92\%, and in the range 2.1∘<σθ≤3.4∘2.1^\circ< \sigma_\theta\le3.4^\circ for quality of fitting R2>85%R^2>85\% (green and red points, respectively, in Figure S6).

Experimental P-THG imaging

In Figure S7, we present the sum of all ninety collected P-THG intensity images of I∥3ωI_{\parallel}^{3\omega}, for all orientations φ\varphi of the excitation linear polarization, corresponding to a field of view containing ultrathin tin (II) sulfide (SnS) crystals. The two crystals we have focused in our analysis are marked.

Sum of collected P-THG intensity images showing ultrathin tin (II) sulfide crystals

Figure S7.

Image not extracted

Figure S7 (PDF p. 27). Sum of all ninety collected P-THG intensity images of I∥3ωI_{\parallel}^{3\omega}, for all orientations φ\varphi of the excitation linear polarization, corresponding to a field of view containing ultrathin SnS crystals. The two crystals we have focused in our analysis are marked. Brighter color indicates higher THG intensity.

Experimental fitting analysis

In Figure S8, we experimentally demonstrate that fitting only one P-THG intensity component is not sufficient to obtain reliable fitting results. An example is presented, for the crystal shown with the red square in the THG images (left), which corresponds to crystal 3 shown in Figure 4. The white square in the same images shows a region without SnS crystals, whose intensity has been considered as noise and has been subtracted from our measurements, following the methodology adopted in the main manuscript.

In Figure S9, we present the simultaneous fitting to both P-THG intensity components, for the crystals 5 and 6, corresponding to the regions of interest (ROIs) shown in Figure S7.

Experimental P-THG intensity modulations for ROIs 5 and 6

Figure S9. Experimental P-THG intensity modulations, I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, as function of the linearly polarized excitation angle φ\varphi, presented in polar plots, for the ultrathin SnS crystals 5 and 6, corresponding to the ROIs marked in Figure S7. Black points represent the experimental data, while the blue and red lines represent the simultaneous fitting with Equation 3 (I∥3ωI_{\parallel}^{3\omega}) and Equation 4 (I⊥3ωI_{\perp}^{3\omega}), respectively. With this fitting, we are able to calculate the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc and dd, for each case, summarized in Table 1.

Relative magnitudes of the χ(3)\chi^{(3)} tensor components

In Table 1, we have reported the average values of the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc and dd, calculated from the fitting results on the sample of the six ultrathin SnS crystals under study. In Figure S10, we plot these average values as Gaussian distributions, with the μ\mu and σ\sigma Gaussian parameters equal to the average and standard deviation (std) values, respectively.

Figure S10. Graphical representation of the average values we have experimentally calculated for the relative magnitudes of the χ(3)\chi^{(3)} tensor components, bb, cc and dd, which are here denoted by RR. These average values, reported in Table 1, are plotted here as Gaussian distributions, with the μ\mu and σ\sigma Gaussian parameters equal to the average and std values, respectively.

In Table S1, we summarize literature findings, to our knowledge, which report experimental results on the relative magnitudes of the χ(3)\chi^{(3)} tensor components, for other 2D materials.

Total P-THG intensity

In Figure S11, we present both P-THG intensity components, together with the total P-THG intensity, I3ω=I∥3ω+I⊥3ωI^{3\omega} = I_{\parallel}^{3\omega} + I_{\perp}^{3\omega}, for the ultrathin SnS crystal which corresponds to ROI 4 shown in Figure 4. The I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega} experimental data are simultaneously fitted with Equation (3) and Equation (4), respectively, while the I3ωI^{3\omega} data are fitted with Equation (6).

Figure S11. Experimental P-THG intensity modulations as function of the linearly polarized excitation angle φ\varphi, presented in a polar plot, for the ultrathin SnS crystal which corresponds to ROI 4 shown in Figure 4. Blue and red points represent I∥3ωI_{\parallel}^{3\omega} and I⊥3ωI_{\perp}^{3\omega}, respectively, while the black points represent the total P-THG intensity, I3ω=I∥3ω+I⊥3ωI^{3\omega}=I_{\parallel}^{3\omega}+I_{\perp}^{3\omega}. The solid curves represent the respective theoretical fittings, with Equation (3), Equation (4) and Equation (6).

THG from N-layer SnS crystals

In order to address the thickness-dependent THG intensity generated from a layered group IV MX crystal, with N number of layers, we can use a constructive interference model.[8, 9, 10] The third harmonic field arising will have the form of vector superposition:

E3ω=E13ω+E23ω+…+EN3ω(15)\mathbf{E}^{3\omega}=\mathbf{E}_{1}^{3\omega}+\mathbf{E}_{2}^{3\omega}+\ldots+\mathbf{E}_{N}^{3\omega} \tag*{(15)}

where the indices denote the third harmonic contribution from the corresponding layers. The total THG intensity produced by the N-layer structure, will then be:

I3ω=∣E1∣2+∣E2∣2+…+∣EN∣2+2E1⋅E2+…+2EN−1⋅EN(16)I^{3\omega}=\lvert\mathbf{E}_{1}\rvert^{2}+\lvert\mathbf{E}_{2}\rvert^{2}+\ldots+\lvert\mathbf{E}_{N}\rvert^{2}+2\mathbf{E}_{1}\cdot\mathbf{E}_{2}+\ldots+2\mathbf{E}_{N-1}\cdot\mathbf{E}_{N} \tag*{(16)}
I3ω=I1+I2+…+IN+2I1I2cos⁡δ1,2+…+2IN−1INcos⁡δN−1,N(17)I^{3\omega}=I_{1}+I_{2}+\ldots+I_{N}+2\sqrt{I_{1}I_{2}}\cos\delta_{1,2}+\ldots+2\sqrt{I_{N-1}I_{N}}\cos\delta_{N-1,N} \tag*{(17)}

where δi,j\delta_{i,j}, i,j=1,2,…,Ni,j=1,2,\ldots,N denote the relative angle between layers ii and jj, i.e., the twist-angles, and the frequency index 3ω3\omega is suppressed for simplicity. If we assume for simplicity that the THG intensity from the individual layers is equal (I1=I2=…=IN=IML3ωI_{1}=I_{2}=\ldots=I_{N}=I_{ML}^{3\omega}), and that the layers are aligned (i.e., all twist-angles are zero), we find that:

I3ω=NIML3ω+IML3ωN(N−1)(18)I^{3\omega}=NI_{ML}^{3\omega}+I_{ML}^{3\omega}N(N-1) \tag*{(18)}
I3ω=N2IML3ω(19)I^{3\omega}=N^{2}I_{ML}^{3\omega} \tag*{(19)}

Therefore, we find that the THG intensity from 2D crystals with zero twist-angle is analogous to N2N^{2}, where N is the number of layers. It is valid for 2D crystals, where each layer contributes constructively in the detected THG signal. It has been experimentally verified for WSe2_2 crystals.[11]

The above analysis allows to interpret the reason behind the considerably higher THG intensity exhibited by some SnS crystals compared to others, as illustrated in Figure 4a, Figure 4b and Figure S7. We note that the integrated THG intensity from each SnS crystal shown in these figures (i.e., sum of all ninety THG intensities acquired for all the different excitation polarization angles φ∈[0∘,360∘)\varphi\in[0^\circ,360^\circ) with step of 4∘4^\circ) is not polarization-dependent anymore, and thus differences in the THG intensities among the different crystals in the same image are attributed to differences in their number of layers. In order to quantify these differences, Table S2 summarizes the maximum THG intensities detected in each ROI. We characterize as monolayer the crystal with the minimum intensity, i.e., ROI 2. Then, the number of layers N for the other crystals can be estimated using Equation (19) (N=I3ω/IML3ωN=\sqrt{I^{3\omega}/I_{ML}^{3\omega}}), with the results summarized in Table S2. We find that, according to the above constructive interference model, the N2N^{2} times THG signal dependency, which appeared in our experimental data, could occur from bilayer and trilayer SnS crystals.

Interestingly, in these ROIs, we observe a variation in the THG intensity and consequently in the calculated number of layers (i.e., we find not exactly N=2N=2 and N=3N=3 in Table S2, but rather N=2±0.2N=2\pm0.2 and N=3±0.2N=3\pm0.2). This variation may be attributed to solvent overlayer, which could change the z-positions of the SnS crystals and place them slightly out of focus.[10] This solvent residual is known to affect the measurement of the thickness of liquid phase exfoliated SnS flakes.[10, 12]

Material preparation and characterization

Preparation of SnS sheets

In order to isolate an ultrathin layer of SnS sheets, the liquid phase exfoliation method is employed.[10, 12, 13, 14, 15, 16] In particular, tin (II) sulfide granular (>99.99%>99.99\%, Lot No-40105, Sigma Aldrich) trace metal basis (3mg/mL) is dissolved in acetone (≥99.5%\ge99.5\%, Honeywell), and the glass bottle is sealed under molecular nitrogen (N2\mathrm{N}_{2}) atmosphere. The SnS solution is ultrasonicated (Elma S 30 H, Elma Schmidbauer GmbH), using a bath sonicator with power 80 W and frequency 37 kHz, for 20 h. The ultrasonicated solution is centrifuged at 8000 rpm for 15 min. The suspended solution containing SnS is collected and used for further experimental investigations.

Material characterization

The isolated ultrathin SnS sheets are characterized with UV-Vis spectroscopy (Figure S12), atomic force microscopy (AFM) (Figure S12), Raman spectroscopy (Figure S13), and photoluminescence (PL) spectroscopy (Figure S13). The results presented in Figure S12 and S13 are discussed in the main manuscript.

Raman spectroscopy

The layer number of SnS thin sheets is quantified using Raman spectroscopy. Figure S13 represents the Raman spectra of bulk SnS and the thin layer of SnS sheets. The spectra depict several vibrational modes. In particular, the SnS crystal belongs to the orthorhombic point group C2v\mathrm{C}_{2v} (mm2), with 24 phonon modes at the center of the Brillouin zone, Γ\Gamma, expressed as:[14, 17]

Γ=4Ag+2B1g+4B2g+2B3g+2Au+4B1u+2B2u+4B3u(20)\Gamma= 4\mathrm{A}_{g} + 2\mathrm{B}_{1g} + 4\mathrm{B}_{2g} + 2\mathrm{B}_{3g} + 2\mathrm{A}_{u} + 4\mathrm{B}_{1u} + 2\mathrm{B}_{2u} + 4\mathrm{B}_{3u} \tag*{(20)}

where Ag\mathrm{A}_{g}, B1g\mathrm{B}_{1g}, B2g\mathrm{B}_{2g} and B3g\mathrm{B}_{3g} are optically active Raman modes. The room temperature Raman spectra of our isolated SnS sheets feature the expected optically active phonon modes (Figure S13). The phonon modes peak at ≈100\approx100, ≈164\approx164 and ≈194\approx194 cm−1^{-1}, and are associated with Ag(1)\mathrm{A}_{g}(1), B3g\mathrm{B}_{3g} and Ag(2)\mathrm{A}_{g}(2) modes, respectively. The isolated thin layer of SnS sheets reveal the prominent Raman peaks at ≈167\approx167 and ≈194\approx194 cm−1^{-1}, corresponding to B3g\mathrm{B}_{3g} and Ag(2)\mathrm{A}_{g}(2) modes, respectively. It is observed that the peak positions feature blueshifts compared to their bulk counterpart. The estimated shifts are ≈5\approx5 and ≈2\approx2 cm−1^{-1} for B3g\mathrm{B}_{3g} and Ag(2)\mathrm{A}_{g}(2) modes, respectively. The observation of blueshift in Raman modes is due to the reduction of the dielectric screening effect upon thinning the bulk crystal layer. These spectral shifts in phonon modes are consistent with our previous results of monolayer SnS sheet,[15] affirming the presence of isolated monolayer SnS sheets.

Experimental setups for material characterization

Extinction spectroscopy

Extinction spectra of the isolated SnS dispersions are acquired with a Perkin Elmer, Lambda 950 UV/Vis/NIR spectrometer.

AFM

Surface topography images of isolated SnS sheets are captured with tapping mode AFM (Multimode AFM from Digital Instruments, Bruker), with controller Nanoscope IIIa. The sample is prepared by drop casting isolated solution on a cleaned SiO2_2 substrate. All the samples are dried and placed on the microscope stage to scan.

Raman spectroscopy

A Nicolet Almega XR μ\mu-Raman analysis system (Thermo Scientific Instruments) is used to collect the Raman spectra. All the spectra are collected with a backscattering geometry, with excitation laser source of 473 nm. The laser beam is focused with a 50×50\times long working distance objective lens with 1800 g(mm)−1\mathrm{g}(\mathrm{mm})^{-1} grating. The equipment is calibrated with SiO2\mathrm{SiO}_2 before collecting the sample spectra.

PL spectroscopy

A home-built backscattering μ\mu-PL setup is used to collect the PL spectra. A laser source of 543 nm (2.28 eV) is used to excite the sample, and a 50×50\times Mitutoyo objective lens focuses the beam to ≈1 μm\approx1\,\mu\mathrm{m} spot size. The sample is prepared with the drop casting method on SiO2\mathrm{SiO}_2 substrate and dried under vacuum. All measurements are performed under atmospheric pressure.

Videos

Description of video 1

We present a video of the experimental P-THG imaging of the ultrathin SnS crystals which belong to the field of view shown in Figure 3. We record images of I∥3ωI_{\parallel}^{3\omega} while rotating the orientation φ\varphi of the linear polarization of the laser, with φ∈[0∘,360∘)\varphi\in[0^\circ, 360^\circ) with step of 4∘4^\circ.

Description of video 2

We present a video of the experimental P-THG imaging of the ultrathin SnS crystals which belong to the field of view shown in Figure 3. We record images of I⊥3ωI_{\perp}^{3\omega} while rotating the orientation φ\varphi of the linear polarization of the laser, with φ∈[0∘,360∘)\varphi\in[0^\circ, 360^\circ) with step of 4∘4^\circ.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  1. [1]S. Psilodimitrakopoulos, L. Mouchliadis, I. Paradisanos, A. Lemonis, G. Kioseoglou, E. Stratakis, Light: Sci. Appl. 2018, 7, 18005.
  2. [2]A. Dasgupta, J. Gao, X. Yang, Laser Photonics Rev. 2020, 14, 1900416.
  3. [3]M. J. L. F. Rodrigues, C. J. S. de Matos, Y. W. Ho, H. Peixoto, R. E. P. de Oliveira, H.-Y. Wu, A. H. C. Neto, J. Viana-Gomes, Adv. Mater. 2016, 28, 10693-10700.
  4. [4]A. Autere, C. R. Ryder, A. Säynätjoki, L. Karvonen, B. Amirsolaimani, R. A. Norwood, N. Peyghambarian, K. Kieu, H. Lipsanen, M. C. Hersam, Z. Sun, J. Phys. Chem. Lett. 2017, 8, 1343.
  5. [5]H. Sar, J. Gao, X. Yang, Sci Rep 2021, 11, 6372.
  6. [6]H. Sar, J. Gao, X. Yang, Sci Rep 2020, 10, 14282.
  7. [7]R. P. N. Tripathi, X. Yang, J. Gao, Opt. Express 2022, 30, 22661-22670.
  8. [8]W.-T. Hsu, Z.-A. Zhao, L.-J. Li, C.-H. Chen, M.-H. Chiu, P.-S. Chang, Y.-C. Chou, W.-H. Chang, ACS Nano 2014, 8, 2951-2958.
  9. [9]S. Psilodimitrakopoulos, L. Mouchliadis, I. Paradisanos, G. Kourmoulakis, A. Lemonis, G. Kioseoglou, E. Stratakis, Sci. Rep. 2019, 9, 14285.
  10. [10]G. M. Maragkakis, S. Psilodimitrakopoulos, L. Mouchliadis, A. S. Sarkar, A. Lemonis, G. Kioseoglou, E. Stratakis, Adv. Optical Mater. 2022, 10, 2270038.
  11. [11]H. G. Rosa, Y. W. Ho, I. Verzhbitsky, M. J. F. L. Rodrigues, T. Taniguchi, K. Watanabe, G. Eda, V. M. Pereira, J. C. V. Gomes, Sci Rep 2018, 8, 10035.
  12. [12]J. R. Brent, D. J. Lewis, T. Lorenz, E. A. Lewis, N. Savjani, S. J. Haigh, G. Seifert, B. Derby, P. O’Brien, J. Am. Chem. Soc. 2015, 137, 12689.
  13. [13]A. S. Sarkar, E. Stratakis, J. Colloid Interface Sci. 2021, 594, 334.
  14. [14]A. S. Sarkar, A. Mushtaq, D. Kushavah, S. K. Pal, npj 2D Mater. Appl. 2020, 4, 1.
  15. [15]A. S. Sarkar, I. Konidakis, E. Gagaoudakis, G. M. Maragkakis, S. Psilodimitrakopoulos, D. Katerinopoulou, L. Sygellou, G. Deligeorgis, V. Binas, I. M. Oikonomou, P. Komninou, G. Kiriakidis, G. Kioseoglou, E. Stratakis, Adv. Sci. 2022, 10, 2201842.
  16. [16]A. S. Sarkar, A. Kumari, Anchala, N. Nakka, R. Ray, E. Stratakis, S. K. Pal, Appl. Phys. Lett. 2021, 119, 241902.
  17. [17]S. Zhang, N. Zhang, Y. Zhao, T. Cheng, X. Li, R. Feng, H. Xu, Z. Liu, J. Zhang, L. Tong, Chem. Soc. Rev. 2018, 47, 3217.

Paper details

vlnc:W_ptbsstu5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 8

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