定义：一种利用非线性材料实现激光频率为输入光频率三倍的效应

三倍频是一种非线性频率变换，得到的光的频率是输入激光的频率的三倍。原则上，这可以通过χ(3)非线性直接实现三次谐波产生，但是由于介质的χ(3)很小以及相位匹配的限制，使得利用这种方法实现三倍频很困难。
因此，三倍频通常利用级联非线性过程产生，首先对输入光束进行倍频，然后利用倍频光和基频光的和频实现三倍频过程，这两个过程均是基于非线性晶体材料的χ(2)非线性的。 

图 1 一个典型三倍频装置：波长为1064nm的红外输入利用一个倍频器实现了波长为532nm的绿米，然后利用一个和频器实现了355nm的光。 
三倍频的主要应用是紫外光的产生。最常见的是利用由 Nd:YAG或Nd:YVO4 激光器产生的1064nm的激光实现三倍频，从而产生波长为355nm的紫外光。一种常用的方法是使用两个LBO（三硼酸锂）晶体，或一个LBO和一个BBO晶体，一个用于相位匹配的二次谐波产生，另一个实现和频过程。
利用从一个调Q或锁模激光可以使得这个过程效率很高，当然利用腔内倍频和和频过程也可以实现连续光的高效三倍频过程。 
此外，利用掺钕激光器输出的波长为1.3um的激光的三倍频也可以实现蓝光的产生。 

电源转换效率 
理论上来说，三倍频过程的功率转换效率可以接近100%。当倍频器具有的2/3的转换效率时，那么倍频光的功率为剩余泵浦光的两倍，从而两束光具有相同的光子数。但是在实际过程中，倍频器的效率通常稍低（通常约40至50%），而且和频混频器的效率远低于100%。
对于后者来说原因有很多，如光强度较低、由光学损伤导致的设计上的限制、空间走离、脉冲脉宽的不匹配、时间走离等。转换效果最好的情况是输入光峰值功率较高而脉冲宽度较宽，光束质量高，而且带宽较窄的情况。在这种情况下，从红外光到紫外光的总转换效率可达30~40%的量级。 

非线性晶体退化 
为了实现高效的三次谐波产生，非线性晶体通常需要承受相当高的光强。但对于和频过程，这往往会产生一个问题：即使对于光强低于损伤阈值的光束，但强烈紫外光还是会导致非线性晶体和出射表面的增透膜的逐渐退化。在这种情况下，非线性晶体通常会变成消耗品，也就是说系统的工作寿命相当有限，从而需要进行晶体的更换。由于这个效应通常只受到光束直径的影响，因此可以通过移动非线性晶体来减少这个效应的影响。 
最终的晶体使用寿命取决于各种因素，包括非线性晶体材料的类型，材料质量，峰值和平均光强的大小，增透膜的类型，脉冲参数（例如脉冲宽度），以及环境空气的污染程度。例如，从机械零件油在紫外光的作用下会发生化学变化，从而会导致其在晶体材料上的沉积。 
需要注意的是，如果晶体上的模场面积等比例的增加，那么紫外光功率的增强并不会导致晶体工作的寿命的减小。 
对于给定的输出功率水平，可以通过在增强腔中防止三次谐波晶体实现晶体寿命的大幅提高。通过利用谐振的频率转换使得单程的转化效率降低，而作为非转换的泵浦光被“回收”。。然而，谐振增强的三倍频并不总是可以发生的（例如，由于激光源的相干性不足），并且还引入了系统额外的复杂性。

Definition: the phenomenon that an input laser beam generates a beam with three times the optical frequency

Alternative term: third-harmonic generation

More general terms: nonlinear frequency conversion

German: Frequenzverdreifachung

Frequency tripling is a process of nonlinear frequency conversion where the resulting optical frequency is three times that of the input laser beam. In principle, that can be achieved with a χ⁽³⁾ nonlinearity for direct third-harmonic generation [12–14], but this is difficult due to the small χ⁽³⁾ nonlinearity of optical media and also because of phase-matching constraints (except for tripling in gases). Therefore, frequency tripling is usually realized as a cascaded process, beginning with frequency doubling of the input beam and subsequent sum frequency generation of both waves, with both processes being based on nonlinear crystal materials with a χ⁽²⁾ nonlinearity.

Figure 1: A typical configuration for frequency tripling: an infrared input beam at 1064 nm generates a green 532-nm wave, and these two mix in a second crystal to obtain 355-nm light.

Applications

The main application of frequency tripling is the generation of ultraviolet light. Most common is the generation of 355-nm light by frequency tripling of a laser beam with 1064 nm, as obtained from a Nd:YAG or Nd:YVO₄ laser. A common approach is to use two LBO (lithium triborate) crystals, or an LBO and an BBO crystal, the first being phase-matched for second-harmonic generation and the second for sum frequency generation. It is easy to make this process efficient when using pulses from a Q-switched or mode-locked laser, but also possible in continuous-wave operation e.g. with intracavity frequency doubling and resonant sum frequency generation.

It is also possible to generate blue light by frequency tripling the output of a 1.3-μm neodymium laser.

Power Conversion Efficiency

Theoretically, the total power conversion efficiency of the frequency tripling process could be close to 100 % in a single pass through the crystals. For that, the frequency doubler should have a conversion efficiency of 2/3, so that the second-harmonic wave has twice the power of the remaining fundamental wave, and both have equal photon numbers. In practice, the efficiency of the frequency doubler is normally somewhat lower (often around 40 to 50 %), and in particular the sum frequency mixer is far from 100 % efficient. The latter problem can result from many effects, such as too low optical intensities, design limitations enforced by optical damage, effects of spatial walk-off, mismatch of pulse duration and/or temporal walk-off, etc. Tentatively, the conversion works best for high peak powers in not too short (e.g. picosecond) pulses, and when the beam quality is high and the optical bandwidth not too high. Overall conversion efficiencies from infrared to ultraviolet can then be of the order of 30 to 40%.

With numerical modeling, the whole frequency tripling process can be simulated with fairly high accuracy and reliability of the results.

Degradation of Nonlinear Crystals

For efficient single-pass third-harmonic generation, the nonlinear crystals need to be operated with fairly high optical intensities. For the sum-frequency mixer crystal, this is often a problem: the intense ultraviolet light can lead to gradual degradation of the crystal material and an anti-reflection coating on the exit surface, even for operation well below the threshold for instant laser-induced damage. Under such conditions, nonlinear crystals can become consumables, i.e., system parts with quite limited lifetime, which need to be replaced periodically. As normally only the volume of the beam is affected, one may often somewhat shift the crystal in order to use another region which has not yet been degraded. Such movements may also be done with some automatic crystal shifter; this method is sometimes called indexing of the crystal. Under good conditions, one may reach several hundred hours of operation on one spot of an LBO crystal and use dozens of spots for obtaining an overall lifetime of many thousand hours with a single crystal.

The achieved crystal lifetime can depend on various factors, including the crystal material type, the material quality, the peak and average intensity levels, details of anti-reflection coatings, other pulse parameters such as the pulse duration, and also contaminants in the ambient air. For example, hydrocarbons from some oils at mechanical parts may be chemically altered by the ultraviolet light and may lead to the deposition of absorbing material on the crystal and on other optics, e.g. a collimation lens.

Note that higher UV output powers do not necessarily decrease the crystal lifetime, if the mode areas in the crystal are increased in proportion to the power level. The operational intensities are then not higher than in lower-power devices. One may be tempted, however, to use higher powers for applying higher intensities with the goal of further increasing the power conversion efficiency. This can then reduce the crystal lifetime.

Substantial improvements in crystal lifetime for a given output power level are possible by using an enhancement cavity around the third-harmonic crystal, i.e., by doing resonant frequency conversion. This is because the single-pass conversion efficiency can then be lower, as the non-converted light is “recycled”. One may then work without an automatic crystal shifter. However, resonant enhancement is not always possible (e.g., for laser sources with insufficient coherence properties) and introduces additional complications.

Bibliography