Laser induced damage threshold (LIDT) of β-barium borate (BBO) and cesium lithium borate (CLBO) – Overview

Highlights

• BBO and CLBO crystals are preferred for deep ultraviolet generation.

• Overview given of LIDT measurements of BBO and CLBO.

• LIDT is the crystallinity level and purity dependent.

• Pulse duration is a factor substantially affecting LIDT.

Abstract

Published data of laser induced damage thresholds (LIDT) of nonlinear BBO and CLBO crystals, which are used for the generation of deep-ultraviolet radiation, are collected and plotted against laser pulse duration. The LIDT measurement techniques are briefly described and their relevance relative to the intended application of the crystal is discussed. The intrinsic and extrinsic factors affecting the evolution of the LIDT are reported. Our own experience with the LIDT of BBO and CLBO crystals in the fifth harmonic generation (FiHG) process is presented. Lowering of the LIDT for crystals with antireflection coatings is treated.

Introduction

The most frequently used nonlinear optical crystals for generating deep ultraviolet radiation (DUV), i.e., under the wavelength of 300 nm, are borate crystals known under the following acronyms BBO (β-BaB₂O₄, beta barium borate) and CLBO (CsLiB₆O₁₀, cesium lithium borate). Both crystals have the relevant features well summarized in Ref. [1]: crystallographic noncentrosymmetry (otherwise second harmonic generation (SHG) would not be possible), transparency below 200 nm (down to 188 nm for BBO and to 180 nm for CLBO, see nonlinear optical (NLO) crystals producers), large (SHG) effective nonlinear coefficient (∼1.8 pm/V for BBO and ∼0.9 pm/V for CLBO) [2]), moderate birefringence Δn around 1 µm (Δn ∼ 0.05 [2]), chemical stability and large laser damage threshold. For high-power DUV, generated as harmonic frequencies of laser beams around 1 µm, the BBO and CLBO seem to be the most appropriate. However, both crystals have some advantages and disadvantages. BBO crystal has been known since 1985 [3], [4], [5], further see in Ref. [6], and CLBO since 1993, and they have extensively been studied since then, see e.g. Refs. [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]. A great advantage of BBO is its great effective nonlinear coefficient, twice that of CLBO, as given above. However, the CLBO has a much lower walk-off angle at the fourth (FHG) and fifth (FiHG) harmonics generation processes (about 30 mrad, compared to about 90 mrad of BBO at FHG [2]) which is relevant for longer crystals. Similarly further CLBO parameters are more favorable from the point of view of the deep UV generation: angular tolerance for SHG at λ = 515 nm is 1.2 mrad⋅cm for CLBO and 0.36 mrad⋅cm for BBO; that for SFG (sum frequency generation) at 1030/257 nm is 3.3/0.8 mrad⋅cm for CLBO and 0.65/0.15 mrad⋅cm for BBO [2]. As we shall see later, also the damage threshold for CLBO is by almost 20% higher than that of BBO [19]. A big disadvantage of CLBO is its hygroscopicity, the CLBO cracks immediately in pure water [20]. The cracking mechanism was explained in Ref. [21]. Using CLBO as a long-term harmonic converter means keeping it at a temperature close to 150 °C. However, the higher temperature is beneficial for lowering the nonlinear absorption in both materials, as given below. Therefore, the individual choice of a suitable crystal depends on preferred criteria (price, simple manipulation etc.).

Besides the laser induced damage threshold, which will be treated in detail below, one important thing must be mentioned: the nonlinear absorption of DUV radiation. It occurs mainly as two-photon absorption (TPA) [22], e.g. either as 2H-photon + 4H-photon, or 4H + 4H, or 4H + 5H, and also 5H + 5H and 5H + 1H. If the intensity of the impinging beam is kept below 10 MW/cm², the effect of the nonlinear absorption is negligible. In BBO crystals, no TPA was observed for 2H + 2H process at the intensity of 40 GW/cm² [23]. If the fundamental wavelength is 1030 nm, then the energy equivalents of the absorption of two photons mentioned previously are 7.2 eV (2H + 4H), 9.6 eV (4H + 4H), 10.8 eV (4H + 5H), 12.0 eV (5H + 5H) and 7.2 eV (5H + 1H). The cut-off wavelength of BBO is 189 nm (6.6 eV) and that of CLBO 180 nm (6.9 eV). It follows that the energy absorbed in the TPA processes exceeds the bandgap energy and is not irradiated, but turned into heat, in the case of CLBO crystals, see e.g. Ref. [24]. The heating of the crystal manifests itself in longitudinal and lateral temperature gradients [25]. Immediate consequences of the crystal heating are phase mismatching and decreased DUV output power and conversion efficiency. It holds for both crystals that a certain way to counteract the detrimental effect of the TPA is to operate the crystals at an elevated temperature, preferably at around 150 °C [10], [26], [27]. The TPA effect was detected very soon after the BBO and CLBO discovery and the measurement of the nonlinear coefficient β appeared in many publications, see e.g. Refs. [24], [26], [27], [28], [29], [30], [31], [32], [33], [34], mainly for nanosecond pulses. It was proved, see e.g. Ref. [31], that higher-quality crystals have lower nonlinear absorption. A question arises which of the crystals has a lower nonlinear coefficient β, and as such should be preferred for high power scaling. The answer is not simple as the nonlinear coefficient β depends on several factors, such as pulse duration, repetition rate, crystal temperature and beam polarization. In Ref. [32] e.g., the β-coefficient was measured for BBO, in both polarizations (o) and (e), and CLBO under the intensity range of 0.2–80 GW/cm² at 248 nm and pulse duration of 0.65 ps, at room temperature. The β-values were about 0.5 cm/GW (o) and 0.3 cm/GW (e) for BBO, and 0.5 cm/GW for CLBO, i.e. they are very similar. On the contrary, according to Ref. [35], the CLBO crystal should be preferred at lower pump intensities of ∼100 MW/cm² as its nonlinear coefficient is much lower than that of BBO, regardless of the temperature (0.5 cm/GW, 15 cm/GW, respectively). As given in Ref. [26], the higher temperature diminishes the nonlinear absorption, therefore the β-values will be lower at a temperature of about 150 °C, which is the recommended operation temperature for CLBO. It was found in Ref. [27] that the TPA at the elevated temperature was 3.5times lower than at room temperature for BBO crystal. Evidently, the preference of one crystal over the other is not clear-cut. As we shall see below, the nonlinear absorption is closely related to the damage threshold [19]. The lower the nonlinear coefficient is, the higher is the damage threshold. If a higher damage threshold is decisive for the crystal usage, then CLBO would probably be a better candidate than BBO.

In our paper we present an overview of the knowledge gained so far about the laser induced damage thresholds of BBO and CLBO – the most frequent DUV nonlinear crystals. We review the different LIDT measurements of both crystals, the effect of pulse duration, repetition rate of laser pulses and laser beam wavelength. The source literature which we used in our overview is listed in Table 1. We summarize briefly the techniques of the LIDT measurement and outline the physical processes leading to the crystal damage, as they have been identified up to now. Finally, we shall treat the problem of LIDT in the process of sum frequency generation, i.e. when three different wavelengths are present in the crystal simultaneously.