1. Introduction

Photo induced modifications of refractive index and absorption have been intensively studied in chalcogenide glassy (ChG) material systems [1]. The ability to change the optical properties of ChG under illumination with band gap light [2–5] makes them very suitable for photo fabrication/patterning as well as holographic recording [6–9]. However ChG materials also possess relatively weak thermal and mechanical resistance. In the same time, it was shown [10] that the addition of germanium Ge into the well-known arsenic sulfide As₂S₃ ChG system not only improves the mechanical and thermal properties of the glass, but also increases its photosensitivity. Furthermore, it was shown that huge (≈0.1) photo induced anisotropy may be achieved in GeAsS glassy films under polarized laser exposition [11]. Those two observations make them excellent candidates for high stability and high efficiency photo anisotropic components.

Gratings (or holograms) can be distinguished in two types: scalar and vector. The scalar holograms are recorded by two waves with the same polarization and thus they can be created by using any excitation mechanism. The vector gratings are obtained with two waves of orthogonal (e.g. circular) polarizations. Thus the spatial modulation of the total light polarization is used have and it requires the presence of photo anisotropic excitation mechanism. Such polarization gratings can be used for the measurement of Stokes parameters and can serve in real-time spectral photopolarimetry for polarization-detecting systems.

In this context, particularly interesting are reports in the literature about vector grating recording in ChG materials. Among them, vector gratings (with a spatial period of 2 µm and a maximal diffraction efficiency of 0.19%) were recorded in 1 µm thick As₂S₃ films by using Ar⁺ laser (operating at 514.5 nm) beams with perpendicular linear polarizations [12]. Later, Mitkova et al. [13] have investigated in details the properties of vector gratings in amorphous ChG Se₇₀Ag₁₅I₁₅ films (0.5 µm – 1.0 µm thick) by using two orthogonally polarized Ar⁺ laser beams operating at 488 nm. The achieved maximal diffraction efficiency was 1%. These vector holographic gratings were stable for six months at room temperature at 22°C. Two orthogonal circularly polarized recording beams were also used (instead of two orthogonal linearly polarized beams) to record vector gratings [13]. Different vector grating recording configurations were further applied to study the bulk and relief modulation mechanisms in As₂S₃ ChG thin films [8, 14].

In this paper, we report on the recording and study of diffraction gratings in Ge₂₅As₃₀S₄₅ ChG thin films. The behavior of recorded vector gratings is compared with that of scalar gratings, recorded in the films of same chemical composition. In addition, the stability (over time and storage temperature) of gratings inscribed in those films was evaluated and compared to those recorded in As₂S₃ thin films, fabricated by the same e-beam evaporation technique. High transient diffraction efficiencies and thermal resistance are demonstrated for Ge₂₅As₃₀S₄₅ films.

2. Experimental methods

Ge₂₅As₃₀S₄₅ films of ≈7 μm thickness were deposited onto BK7 glass substrates (held at room temperature) by electron beam evaporation method (with an electron beam voltage of 4 kV in a vacuum of 10⁻⁶ Pa) from the bulk glass of same composition Ge₂₅As₃₀S₄₅. The conservation of glass stoichiometry after the film deposition was confirmed by elemental dispersive analysis (EDAX). The obtained films were annealed at 350°C during 2 hours in an electrical oven in ambient atmosphere to remove the residual stress induced during the deposition. The glass transition temperature of the thin films T_g = 357°C was determined by thermal analysis (heating rate of 10°C/min). The optical band gap was calculated following the Tauc method (from a 3 μm thick film) to be E₀ = 2.42 eV while a coefficient absorption of ≈0.78x10⁴ cm⁻¹ was measured at λ = 514.5 nm and a penetration depth estimated to be in the micrometer scale. More details about the method employed for the fabrication of the Ge₂₅As₃₀S₄₅ bulk glass and thin films can be found in our previous work [15].

The experimental set-up, used to record vector (or polarization) gratings, is shown in Fig. 1. Grating recording was achieved with two polarized (for example, with left-hand (LCP) and right-hand (RCP) circular polarizations) continuous wave (CW) Ar⁺ laser (operating at 514.5 nm) beams of equal intensities of ≈4 W/cm² (for each beam).

 

Fig. 1 Experimental set-up for grating recording and study: Pump – Ar⁺ laser (514.5 nm); λ/2 – half-wave plate; WP – Wollaston prism; M – mirror; F – density filter, probe beam – He-Ne laser (632.8 nm); λ/4 – quarter-wave plate; S – sample; D1,2,3 – photodetectors.

Download Full Size | PDF

To record the scalar gratings, we first split (by means of a simple beam splitter; not shown here) the initial linearly polarized (s-type) beam of the Ar⁺ laser (“pump”) into two beams with the same polarization and intensity values (adjusted by means of a density filter F). The direct use of those beams allowed the recording of scalar gratings via spatial intensity modulation. The angle between these beams was ≈5° while the spot diameter size (at the ChG film sample S) was ≈2 mm. The inscription of vector gratings was then done by using two different recording configurations: first, by using two orthogonal linear (s- and p- polarized) beams, by means of a Wollaston prism (WP, Fig. 1). Then, another type of gratings was recorded by using two circular orthogonally polarized (LCP and RCP) beams, which were obtained with the help of two quarter wave plates (λ/4, Fig. 1). A commercial polarimeter (Thorlabs PA430) was used to measure the circularity of those beams and keep the corresponding deviations below 10%. The intensities of these components were equalized using the filter F. A vertically polarized (s-type) He–Ne laser beam, operating at 632.8 nm, was used as probe to study the recorded gratings; diffracted orders being simultaneously recorded by the photodetectors D1, D2 and D3, Fig. 1. All experiments were carried out at room temperature (≈22°C).

3. Results

As mentioned in the introduction, the addition of Ge into the binary As₂S₃ glass results in larger anisotropic photosensitivity along with the overall improvement of its mechanical and thermal properties thanks to an increased glass network connectivity. For instance, GeAsS glasses exhibit higher glass transition temperature (> 300°C) than the As₂S₃ (≈180°C). Therefore, a higher thermal resistance could also be expected for the holographic gratings recorded in those films (like the composition under study here, Ge₂₅As₃₀S₄₅) compared to other ChG films, such as the As₂S₃. To verify this assumption right at the beginning, thin films of both compositions were prepared by using the same e-beam evaporation method. The same vector gratings were recorded (by using RCP + LCP pumps) in Ge₂₅As₃₀S₄₅ and As₂S₃ ChG films, both of 7 µm thickness. The intensity of each pump beam was set to 4 W/cm² in both cases. Those films were then heated at different “storage” temperatures ranging from 50 to 350°C to study dynamically the diffraction efficiency change as a function of temperature. The samples were kept for 30 minutes at each storage temperature. Figure 2 shows the remnant diffraction efficiencies obtained for Ge₂₅As₃₀S₄₅ (black dashed) and As₂S₃ (red solid) films as function of storage temperature.

 

Fig. 2 Remnant diffraction efficiencies (see, e.g., point B, Fig. 4) of vector gratings recorded in Ge₂₅As₃₀S₄₅ (black dashed) and As₂S₃ (red solid) thin films as a function of temperature. Lines are guides to the eye. Vertical dashed lines show the glass transition temperatures of As₂S₃ (180°C) and Ge₂₅As₃₀S₄₅ (350°C).

Download Full Size | PDF

Since the refractive index of these films is high (≥ 2.4) [16], the diffraction efficiency was calculated by the following equation (to take into account the reflection losses):

As we can see, in Fig. 2, the grating efficiency of the As₂S₃ film is drastically decreasing to 0 (i.e. the grating is completely erased) at the temperature of 200°C. Whereas in the case of Ge₂₅As₃₀S₄₅, the grating still exhibits 0.52% of remnant efficiency at 300°C, which corresponds to 40% of its maximum efficiency, prior to vanishing at only 350°C. Let us recall that the glass transition temperatures of those compositions are 180°C and 350°C, respectively for As₂S₃ and Ge₂₅As₃₀S₄₅. It is important to note that the obtained remnant diffraction efficiency (measured after the pump beams were removed) for Ge₂₅As₃₀S₄₅ was almost one order of magnitude higher than that for As₂S₃ composition (i.e. 1.3% and 0.19%, respectively). Another interesting observation is the difference of slopes of erasure (noticeably higher for the As₂S₃ composition).

Gratings recorded in the Ge₂₅As₃₀S₄₅ films were further studied in more details to compare their behavior and involved mechanisms with those taking place in As₂S₃ films. Figure 3 presents the dependences of maximal transient diffraction efficiencies (see, e.g., the point A, Fig. 4) of scalar and vector gratings upon the pump intensity (recorded on the same set-up and conditions) in the Ge₂₅As₃₀S₄₅ film. The scalar grating was recorded with two linearly s-polarized pump beams, whereas the vector grating was recorded with two circular orthogonally polarized beams.

 

Fig. 3 Maximal transient diffraction efficiency (%) of vector (solid line) and scalar (dashed line) gratings recorded in the same Ge₂₅As₃₀S₄₅ film as a function of pump intensity. The vector grating was recorded by (RCP + LCP) polarized beams while the scalar grating was recorded with two linearly s-polarised beams. Broken lines are guide to the eye.

Download Full Size | PDF

 

Fig. 4 Dynamics of diffraction efficiency (%) evolution for scalar and vector gratings recorded on the same Ge₂₅As₃₀S₄₅ film. The scalar gratings (solid curve) were recorded by two linearly s-polarised beams, while the vector gratings were recorded by (s + p) polarized beams (dashed dot curve) and by circularly (RCP + LCP) polarized beams (dotted curve).

Download Full Size | PDF

As we can see, in Fig. 3, for scalar gratings, the efficiency was maximal at the pump intensity of ≈4 W/cm², for each beam. Significantly lower dependences were observed for vector gratings without noticeable improvement for higher pump intensities. Hence, the results presented hereafter were obtained at this pump intensity.

Figure 4 shows the kinetics of the diffraction efficiency of vector and scalar gratings as function of exposition time. The scalar gratings were recorded with two linearly s-polarized beams while the vector gratings were recorded either with two orthogonal (s + p) or with (RCP + LCP) polarizations. For vector gratings, we can observe, Fig. 4, a maximum transient diffraction efficiency of ≈12 ± 1% at a pump intensity Ieach = 4 W/cm², while a diffraction efficiency of ≈35 ± 3% is reached for the scalar gratings of the same ChG sample.

Optical microscopy images of obtained gratings are presented in Fig. 5. The grating’s period, observed for the case of scalar (s + s) recording (Fig. 5(a)) or (RCP + LCP) (Fig. 5(b)) cases, is ≈3.1 µm, while for the vector gratings obtained with two orthogonal linearly polarized pumps (s + p), the period is ≈1.3 µm, Fig. 5(c). Except for the polarization state of exposition beam, all the experimental conditions, i.e. pump intensity, angle between the recording beams, etc., were kept identical for this study. It is important to note that grating’s calculated period is 3 µm, considering the angle between the two pump beams and their wavelength.

 

Fig. 5 Optical microscope images (in unpolarized light) of recorded gratings on the same Ge₂₅As₃₀S₄₅ thin film: (a) scalar gratings written by (s + s) polarization beams; (b) vector gratings written by (RCP + LCP) polarization beams; (c) vector gratings written by (s + p) polarization beams.

Download Full Size | PDF

DekTak profilometry and atomic force microscopy (AFM) measurements did not reveal any relief modulation at the surface of recorded gratings.

Finally, as it is also well-known, the vector gratings, recorded by two orthogonal-circularly polarized beams, can be used for the measurement of the polarization ellipticity (or the Stokes parameter S3) of light with unknown polarization by means of the detection of + 1 and −1 diffracted orders [17]. Figure 6 shows the dependences of the diffraction efficiencies (recorded with two, left and right, circular polarized beams of Ar⁺ laser) of diffraction orders ( + 1 and −1) upon the angle of rotation of a quarter-wave plate, placed on the optical path of the incident probe beam (He-Ne laser). The recording beams’ total intensity was 8 W/cm² and time of exposure was ≈30 min. As we can see in the Fig. 6, the intensities of the diffracted orders indeed change correspondingly with the variation of the polarization state of the incident probe beam.

 

Fig. 6 Diffraction (in arbitrary units) of + 1 and −1 diffracted orders as a function of the rotation angle of the quarter-wave plate (the ellipticity of the incident probe beam polarization). Ge₂₅As₃₀S₄₅ film thickness is 7 µm. Broken lines are guides to the eye.

Download Full Size | PDF

4. Discussion

As expected, the addition of Ge to the glassy matrix of As-S has improved its performance compared to the original glass. As we can see, from the Fig. 2, the gratings efficiency of the As₂S₃ film starts to decrease at 100°C and vanishes completely at about 200°C, which is close to its glass transition temperature. While at those temperatures, the relative efficiency for Ge₂₅As₃₀S₄₅ gratings is still ≈100%. The progressive heating of the grating decreases its efficiency. However, the slope of this decrease is smaller and even at 300°C the Ge₂₅As₃₀S₄₅ grating still exhibits ≈40% of relative efficiency, which is still one order of magnitude higher than the relative efficiency for the As₂S₃ film. Very likely the reason of the “narrow” transition zone is related to the fact that the As₂S₃ matrix is closer to the stoichiometric equilibrium and its photosensitive units have more or less similar activation energies compared to those in the Ge₂₅As₃₀S₄₅ matrix. The complete ‘erasure’ of the grating occurs at 350°C for the Ge₂₅As₃₀S₄₅ film, which is close to its glass transition temperature. Such stability increases the number of potential applications of this glass composition as polarization selective optical element.

As presented in the previous section, the diffraction efficiency is higher in scalar gratings compared to those of vector gratings. There are many reports in the literature about the involved photoinduced anisotropic mechanisms in ChG films with regard to the recording of polarization gratings. Ozols et al. [18] assumed that the photo induced anisotropy originates from the generation and reorientation of defects (D centers) by exposing the film by both band gap and sub band gap lights [19, 20]. This hypothesis was supported by the EXAFS measurements performed in As-Se films. An increase of the next-nearest-neighbour distance around the Se atoms was found with its magnitude depending upon the direction of light polarization [21]. The main reason why the vector grating recording is less efficient than the scalar grating recording in ChG films is the much smaller concentration of D centers. It was shown that the vector recording involves indirect electronic transitions, whereas scalar recording involves direct ones [22]. The same authors have also shown an easy erasure of vector holographic gratings by one of the recording beams due to the fact that a spatially non uniform (periodic) D centers orientation distribution can be easily destroyed by light (the He – Ne laser photon energy is about twice the activation energy of D centers).

Furthermore, to understand the dynamic changes of the first order diffraction efficiency, observed for the scalar and vector gratings (Fig. 4), we should define which type of grating (or diffraction regimes) is created here. This depends upon several factors, i.e. geometrical ${\text{Q}}_0=2\text{π}\frac{{\text{dλ}}_0}{{\text{nΛ}}^2}$ and physical $\text{δ}{\varphi }_0=\frac{2\text{π}}{{\text{λ}}_0}\text{nd}\frac{\text{δε}}{4\text{ε}}$ [23], where d is the thickness of the sample, n is the average refractive index, $\text{Λ}$ is the gratings period, and $\epsilon$ is the dielectric constant of the material given by: $\frac{\text{ε}}{{\text{ε}}_0}={\text{n}}^2$.

For the so-called ‘thin grating’ diffraction regime, we should have ${\text{Q}}_0<1$ and $\text{δ}{\varphi }_0\le \text{π}$. In our case, ${\text{λ}}_0=0.514 \text{µm}, \text{d}=7 \text{µm}, \text{n}=2.4, \Delta \text{n}=0.03, \text{Λ}=1.3÷3.1$. Thus, ${\text{Q}}_0\equiv 1.04÷5.57$ (depending upon the period $\text{Λ}$) and $\text{δ}{\varphi }_0=0.0026\text{π}$. Therefore, the gratings studied here cannot be considered neither thin nor thick. This explains the recorded value of diffraction efficiency, which was ≈35%, while for the thin phase grating, the maximum theoretical value should be $\le$33.9% [24].

The oscillations we observe in the vector gratings (Fig. 4) can be related to the fact that there are different mechanisms contributing (they are not ‘pure’ vector gratings due to the presence of residual intensity modulation). The non-purity of the recorded vector gratings is also visible in the Fig. 6, where we can see some deformation of the curve and shifted maxima. Similar behavior was already reported in the study of vector gratings in As₂S₃ ChG films [8]. Such behavior was explained by the creation of superimposed undesirable scalar gratings during the vector grating recording. This hypothesis was confirmed by studying the kinetics of recording and optical erasure of vector gratings. The erasure was not possible to carry out there in the case of recording in (RCP + LCP) polarization configuration, which confirms the non-purity of recorded vector gratings.

Another interesting aspect of the present study is the period difference among the recorded gratings and particularly, the period doubling for the vector grating recorded by (s + p) beams compared to the case of recording with circular (RCP + LCP) polarized beams. The grating recording here might be described by the phenomenological model of anisotropy proposed by Fritzsche who suggested the presence (native or photo induced) of polarizable anisotropic units with positive anisotropy (cigar-like) [25]. Asatryan et al. have postulated in their study the existence (in As₂S₃ ChG films) of photo sensitive units, which have negative anisotropy of polarizability (disk-like) [14]. Based on this assumption, they have examined the grating recording for all configurations and calculated the period of corresponding relief modulation. Their analysis demonstrated that in the case of the cigar-like micro volumes, the period of relief modulation, recorded with (s + p) polarized beams, was two times smaller than the (RCP + LCP) recorded case. This calculation was done in the framework of photo induced contraction and dilatation that resulted in a relief modulation in the case of As₂S₃ films. We think that, even though no relief modulation was observed in our Ge₂₅As₃₀S₄₅ composition, the above mentioned defects (and periodic stress) should be also present in our Ge₂₅As₃₀S₄₅ composition, as the same phenomenon of period doubling occurs, depending on the method used for recording the vector gratings (see Fig. 5). Further investigation is required to verify this assumption.

5. Conclusion

In summary, the recording of vector and scalar holographic gratings in Ge₂₅As₃₀S₄₅ ChG thin films was experimentally studied and discussed with regard to anterior works conducted on As₂S₃ thin films. First, a significant improvement of the thermal resistance of the vector gratings, recorded in GeAsS films, was observed compared to those recorded in AsS films. Then, experimental results have shown different behaviors for the vector and scalar holographic recordings. The diffraction efficiency for scalar gratings recorded in GeAsS films at 514.5 nm, are higher than those of the vector gratings, as already reported in As₂S₃ thin films [18]. This may be due to the lower concentration of charged D centers compared to the concentration of sites of photo induced structural changes (in our case, homopolar As-As and Ge-Ge bonds), and/or to the indirect electronic transitions in the vector recording case, which are less efficient than direct transitions in the scalar recording case. In our experiments, maximal transient diffraction efficiencies of 12% and 35% were measured for vector and scalar grating excitations, respectively. The remnant efficiencies were 1.3% and 2.5% in the vector and scalar gratings, respectively. In addition, the obtained results confirm that the mechanisms previously proposed to explain the gratings’ period doubling in As₂S₃ thin films, as well as the non-uniform polarization selectivity of recorded gratings, also could be applied to the GeAsS compositionuthors have the option to upload a thumbnail image that will appear next to the published article on the Forthcoming, Current Issue, and Abstract pages. Please note that if authors do not choose a file, OSA Production Staff will choose an image from the submission. For precise representation of an article, we recommend that authors choose and upload the thumbnail image.