1. Introduction

Analyzing fluorescence on the ultrashort time scale reveals novel insights into fundamental processes such as relaxation mechanisms in gases, liquids, or solids, for instance biological tissue. A temporal resolution below 100 fs is favorable, which lies far beyond the electronic device limits and is achievable by sophisticated optical techniques only. A prevalent technique is optical Kerr gating (OKG), where the polarization of the signal light is transiently rotated by an optically induced birefringence [1]. It has been applied for time resolved detection of ultrafast fluorescence [2–4], and for structure analysis by time-of-flight measurements [5, 6] as well as for the suppression of fluorescence background in resonance Raman spectroscopy [7]. Since the induced birefringence depends directly on the refractive index modification, materials with a large nonlinear refractive index n₂ are considered to be best suited for efficient OKG. Moreover, for sub-100 fs gating times, the origin of n₂ has to be dominantly of electronic nature and therefore solids should be used as gating material. Various glasses, consisting of heavy atom glass matrices such as tellurites [3, 5, 8, 9] or chalcogenides [8, 10] have been investigated as OKG candidates. Unfortunately, these glasses possess small bandgaps, limiting broadband switching towards the UV region. They furthermore may give rise to significant competing processes such as multiphoton absorption. Large bandgap materials however, consisting of light atoms such as fluorides or fused silica suffer from very inefficient OKG due to small nonlinear refractive indices [11, 12]. Recently, a novel type of SiO₂-Al₂O₃-La₂O₃ (SAL) glasses was found to show a relatively large n₂ value, accompanied by a rather wide bandgap and low dispersion [13], due to a small content of heavy La₂O₃ in a light SiO₂-Al₂O₃ glass matrix.

In the present work SAL glass is evaluated for OKG and compared to commonly used large n₂ gating materials, in particular heavy flint glass (N-SF56), tellurite glass (Te-glass) and poly-crystalline zinc sulfide (ZnS) [14]. Additionally SAL glass gating properties are compared to those of fused silica (F300). The materials are tested with respect to their gating efficiency at peak intensities of the femtosecond (fs) laser gate pulse ranging from 50 to 700 GW/cm². Such intensities are achieved by present day high power oscillators [15], providing pulse repetition rates in the 0.1 to 1 MHz range.

2. Experimental

2.1. Materials selection

Crucible melted SAL glasses with La₂O₃ contents between 10 and 24 mol% were used. The composition and manufacturing process of the SAL glass samples (SAL1-4, thickness d = 1.0 mm, n₂ [10⁻⁷ cm²/GW] values [13]: SAL1: 5.8 ± 0.4, SAL2: 6.7 ± 0.4, SAL3: 8.1 ± 0.2, and SAL4: 9.3 ± 0.1) are given in [13, 16, 17]. The Te-glass consisting of 75 mol% TeO₂, 20 mol%, ZnO, and 5 mol% K₂O (d = 0.6 mm, n₂ = (50 ± 5)×10⁻⁷ cm²/GW determined after [13]) was produced by the crucible melting technology according to [18]. The high purity raw materials (oxides, carbonates) were thoroughly mixed, homogenized and subsequently melted in a covered gold crucible in ambient atmosphere between 800 and 900 °C. After several hours of melting, the liquid glass batch (20 g) was cast into preheated moulds and slowly cooled to room temperature. The chemical glass composition was determined by quantitative electron probe microanalysis (EPMA) using energy dispersive X-ray analysis (EDX, JEOL, JXA 8800L). ZnS (d = 1.17 mm, n₂ = (130 ± 20)×10⁻⁷ cm²/GW determined after [13]) was purchased from Vitron Spezialwerkstoffe GmbH, N-SF56 (d = 0.6 mm, n₂ = 26×10⁻⁷ cm²/GW [19], p. 463) from Schott AG and fused silica (F300) from Heraeus Quarzglas GmbH & Co. KG (d = 0.6 mm, n₂ = (3.0 ± 0.1)×10⁻⁷ cm²/GW [13]). Linear and nonlinear refractive indices were determined by transmission and reflection measurements using a photo spectrometer (PerkinElmer, Lambda 900) or by the Z-Scan method [13].

2.2. Kerr gating experiment

The linearly polarized pulses of the fs laser system described in [13] (central wavelength λ_L = 800 nm, pulse duration τ_L = 70 fs, repetition rate f_R = 1 kHz) were partially converted by an optical parametric amplifier (Coherent Opera solo) to signal pulses (λ_s = 530 nm, τ_s = 70 fs, pulse energy E_s = 1nJ/pulse) and their polarization rotated about 45°. The other part of the laser output served as gate pulses of adjustable energy 0.2 to 2 μJ and delay time −300 fs ≤ Δt ≤ 600 fs with the signal preceding the gate pulse for Δt > 0. Both beams were focused into the OKG sample nearly collinear at an angle of 2° between the beams using two lenses of focal lengths f_g = 250 mm (800 nm) and f_s = 100 mm (530 nm) with coincident focal planes at the center of the irradiated sample to yield focal radii w_i(1/e²) and Rayleigh lengths L_R,i of w_s = (25 ± 1) μm and w_g = (45 ± 1) μm as well as L_R,s = (3.7 ± 0.3) mm and L_R,g = (8.0 ± 0.4) mm for the signal and the gate pulses, respectively. Gate beam intensities 0.05 ≤ I_P ≤ 0.7 TW/cm² were achieved in the focal range.

Behind the OKG medium the signal pulse was coupled into a spectrometer (Ocean optics USB4000) and its energy transmission evaluated by spectral integration (450 to 600 nm) to yield the polarization integrated signal (PIS) as a function of Δt and I_P. Additionally, the polarization selective signal (PSS) was determined at 90° polarization angle relative to the gate beam by placing an analyzer in front of the spectrometer. The optical losses of the signal beam in the analyzer were measured and taken into account. All PIS and PSS values were normalized to the PIS transmission obtained with gate pulses of low intensity at large delay Δt > 0.

3. Results

Figure 1(a) shows typical PIS and PSS delay time traces obtained at I_P = 0.62 TW/cm² for the SAL4 sample. PIS stays constant at ≈ 1. The Gaussian fitted PSS curve displays the definition of the OKG efficiency η as the PSS transmission maximum. It is shifted to Δt ≈ 70 fs and its width amounts to 135 fs. In contrast the PSS maximum of fused silica (100 GW/cm², magnified × 10) lies at Δt = 0 and its width amounts to 90 fs. The PIS and PSS delay time traces for the Te-glass as a high n₂ material recorded at various I_P values are complex: The PIS values in Fig. 1(b) decrease steadily for increasing intensity I_P whereas the PSS signal increases in the low intensity region (I_P up to ≈ 300 GW/cm²), but decreases again for large I_P values.

 

Fig. 1 (a) OKG signal pulse transmission (PIS and PSS) as a function of the pulse delay Δt for SAL4 glass. The Gaussian fit of the PSS visualizes the OKG efficiency η. The trace PSS(Δt) of fused silica (magnified ×10) is shown as a reference. (b) PIS(Δt) traces (dashed lines) and PSS(Δt) traces (solid lines) of the Te-glass for selected I_P values in GW/cm² (color coded).

Download Full Size | PDF

The η values of the SAL glasses and of fused silica increase steadily with increasing gate pulse intensity I_P. The former show η values between 50 % and 70 % for high intensities in Fig. 2(a), whereas η of the latter stays below 30 % within the investigated I_P range. For small intensity values, η of the high index materials in Fig. 2(b) is larger than that of the low index glasses. After surpassing a characteristic intensity I_max, however, it decreases again.

 

Fig. 2 OKG efficiency η as a function of I_P for (a) SAL glasses and F300 and (b) high n₂ materials (Te-glass, N-SF56 and ZnS). The dashed lines are guides for the eye. In parentheses: n₂ [10⁻⁷ cm²/GW]. The solid vertical lines indicate I_max.

Download Full Size | PDF

To achieve additional insight into the OKG behavior the PIS is displayed as a function of I_P for two representative delay times in Fig. 3(a) (Δt >500 fs: signal pulse preceding gate pulse (unfilled symbols) and Δt = 0: simultaneous pulses (filled symbols)). As a common feature the PIS (Δt = 0) are smaller than PIS (Δt > 500 fs). For the SAL glass and F300 no significant decrease of the PIS was detected within the investigated I_P range at any delay time. Supplementary, important material parameters, such as sample thicknesses d, the group velocity mismatch of gate and signal pulses Δn_g (in parentheses) and the self-focusing distance of the gate beam z_sf ( $z_{\mathit{sf}}=\left(2n_0{w}_g^2{\left(P/P_{\mathit{Cr}}\right)}^{-1/2}\right)/{\lambda }_g$ with w_g: beam radius within L_R, $P_{\mathit{Cr}}=\left(0.146{\lambda }_g^2\right)/\left(n_0{n}_2\right)$ [20], P: peak pulse power, n₀: refractive index at 800 nm) of the gating pulse as a function of I_P are displayed in Fig. 3(b).

 

Fig. 3 a) Normalized PIS transmission as a function of gate pulse intensity I_P at Δt > 500 fs (unfilled symbols) and Δt = 0 fs (filled symbols) for Te-glass (circles), N-SF56 glass (squares) and ZnS (triangles), PA: permanent attenuation, TA: transient attenuation. b) Self-focusing distances z_sf according to [20] (solid lines), group velocity mismatch Δn_g between gate and signal pulses in parentheses and sample thicknesses (dashed lines).

Download Full Size | PDF

4. Discussion

The most impressive result of the present work is the high OKG efficiency of the SAL glasses shown in Fig. 2(a) up to η ≈ 0.7 which increases steadily with rising intensity I_P and/or La₂O₃ content. Looking at the almost identical ratios of the nonlinear refractive index n₂(F300):n₂(SAL4) ≈ n₂(SAL4):n₂(N-SF56) ≈ 3 it was a priori an open question, if the SAL glass properties are closer to those of fused silica or those of N-SF56: The relation between η and n₂ in Fig. 2(a) is a monotonous function for the SAL glasses and fused silica. For high n₂ valued materials in Fig. 2(b), the monotonous η :n₂ relation seems only valid in the low intensity region up to a characteristic intensity I_max. At higher I_P values, however, their OKG behavior turns out to be complex as it is evident e.g. from the η decrease for I_P > I_max. This complex behavior can be rationalized as the superposition of Kerr rotation, self-focusing of the gate laser beam, combined multiphoton absorption induced by the high gate pulse intensity, and photodarkening of the sample. This helps to understand why SAL glass has advantages in OKG at moderate pump beam intensities. The analysis simplifies, if the investigated samples are categorized into two groups, the SAL glasses and F300 with n₂ < 10⁻⁶ cm²/GW and N-SF56, Te-glass and ZnS with n₂ > 10⁻⁶ cm²/GW.

Looking at the OKG signals of main interest (=PSS) in Fig. 1(a) the symmetric broadening and the shift of the PSS towards positive delay times of SAL4 with respect to fused silica (Δn_g = 0.02) is attributed to a slightly stronger temporal walk-off in the SAL glass (Δn_g = 0.05). This is in agreement with the findings by Yan et al. [21] and Ziolek et al. [22].

For high n₂ materials PSS broadening in Fig. 1(b) is stronger (e.g. ≈ 0.4 ps in case of Te-glass), due to the larger group velocity mismatch between the signal and gate pulses (Δn_g = 0.16). Although its PSS trace is symmetric for small I_P values, increasing I_P leads to an asymmetric shape in Fig. 1(b). Evidently, the rising I_P causes additional nonlinear optical effects, which are observed as PIS reductions. These include permanent effects (PA, delay time regions Δt < −200 fs and Δt > 500 fs), which are attributed to photodarkening, and asymmetric transient signal beam attenuation (TA, −150 fs ≤ Δt≤ +500 fs). I_P dependent changes in the PSS in Fig. 1(b) parallel those in PIS and finally limit the gating efficiency η.

Focusing on the differences between the low and high n₂ material groups their absorption band edges are found in the hν > 5.5 eV and hν < 3.7 eV regions in Fig. 4, i.e. above and below the limits of the gate pulse intensity induced combined two-photon (hν_g(800 nm) + hν_s(530 nm) = 1.55 eV + 2.34 eV = 3.89 eV) and three-photon absorption (2hν_g(800nm) + hν_s(530 nm) = 2.10 eV + 2.34 eV = 5.44 eV): the high n₂ materials are amenable to efficient multiphoton absorption. As depicted in Fig. 3(a) this leads to strong effects in transient signal beam absorption (TA) and photodarkening (PA) of N-SF56 and Te-glass, which reduces the OKG signal. Additionally, the increasing gate pulse intensity reduces the self-focusing distance z_sf, thus leading to smaller spot sizes and consequently stronger nonlinear absorption in the rear part of the sample. Figure 3(b) shows that z_sf is close to the sample thickness for high n₂ materials within the investigated I_P range. This may enhance absorption tremendously. The combination with their temporal walk-off leads to the asymmetric shape of the transient absorption, since the gate and signal pulses overlap deeper in the sample for larger delay times [22]. Due to the negligible nonlinear counter effects in SAL glass and fused silica, self-focusing does not play a critical role and only may enhance, but not reduce the gating efficiency [21]. Thus the OKG similarity between fused silica and SAL is a result of the large band gaps, preventing significant nonlinear absorption even at relatively large I_P values.

 

Fig. 4 Absorption spectra of the investigated materials. The black lines at 3.9 eV and 5.44 eV indicate the energies for combined two- (hν_g+hν_s) and three-photon (2hν_g+hν_s) absorption. The absorption spectrum of fused silica in this energy range is negligible [23].

Download Full Size | PDF

5. Conclusion

In the present work, newly developed SiO₂-Al₂O₃-La₂O₃ (SAL) glasses were evaluated by their OKG efficiency η. SAL4 glass of efficiency η ≤ 70 % is superior to fused silica (F300) in the full gate beam intensity I_P range. Even compared to high n₂ value materials SAL4 glass is superior above material specific I_P values of 150 GW/cm² (ZnS), 350 GW/cm² (Te-glass) and 450 GW/cm² (N-SF56 glass). This is due to its relatively wide bandgap preventing unwanted photodarkening and/or combined multiphoton absorption.