定义：可以提高激光光束的光束质量的装置。

由于各种原因，激光光束的光束质量通常不是理想的，即其强度或者相位截面会发生畸变。根据情况的不同，可以采用不同类型的模清洁器来使光束截面“清洁“。其实模清洁器的叫法并不准确，因为清除的并不是某一个模式（它是一个数学概念，而不是物理量），而是清洁激光光束的相位或者强度截面。另一个常用的叫法为（至少非共振的模清洁器）是空间滤波器。 
存在共振和非共振的模清洁器，它们的工作原理和应用范围是不同的。共振模清洁器通常称为模清洁腔，有单独的词条进行描述。下面主要描述非共振的器件。 

图1：非共振模清洁器，包含两个透镜和一个针孔。 
所有的模清洁器存在的一个基本问题就是理论上无法实现采用一个光学装置将全部多模光束的功率集中到一个单模光束上（具有完美的光束质量）。但是，如果几乎所有的能量本来都集中在单模上，可以通过衰减其它模式来提高光束质量，而且这样不会产生很大的功率损耗。许多模清洁器都采用这一原理。另一种不同的原理是亮度转化器，是一个激光器装置，而不是简单光学元件的组合，可以改变光的波长。 

非共振模清洁器 
图1给出了很简单的装置示意图，可以认为是采用了傅里叶光学。左侧入射的光束先通过透镜聚焦到一个很小的点。在焦平面上，原始光束的不同空间傅里叶成分在空间上分离开：每一个空间傅里叶成分对应某一特定方向传播的平面波，例如传播方向稍微向上的成分的焦点也在较高的位置。因此，在焦平面的针孔可以有效的减小轴外的成分。第二个透镜使光束再准直，因此这时得到的光束具有很”干净的”强度和相位截面。由于每一空间成分的焦点都具有有限大小，清洁过程不是理想情况。简单的使聚焦程度增强不能解决问题，因为这样同时会使空间傅里叶成分在焦平面上的空间分离变小。但是，一定的聚焦程度是必要的，一个经验法则就是束腰尺寸要比原始光束尺寸小几倍。另外，真空的直径大小也必须合适；如果孔太小，不仅会引起附加损耗，还会使光束质量下降。 
当接近于高斯光束的原始光束（光束质量很高）产生了一些空间畸变，例如，由于反射镜上存在一些灰尘时，模清洁器可以很有效的工作。这一光束可以很有效的被清洁，而不会损失很多光功率。（注意这时存在高斯模式与高阶模式之间的叠加，会严重影响光束质量，但是携带的光功率不高。）但是，这一方法对于从激光器出射的光束质量差的空间多模光束则不能有效工作。不仅消除掉高阶模式会造成极大的功率损耗，而且将相近模式在空间上完全分离开也不理想。 
在高功率激光器，尤其是调Q激光器中存在的一个实际问题是，在针孔附近的光强非常大，可能会引起损伤。 与模清洁腔相比，非共振模清洁器排列相对容易，可以在多色光辐射情况下工作，但是光束质量提高的程度不能达到多模光束的提高程度。 

采用单模光纤进行模式清洁 
另一清洁激光光束的方法是将光传输进单模光纤。如果光纤足够长能够去除包层模式（几分米就足够了，与包层材料有关），那么出射的光纤模式接近于高斯截面。采用这种方法可以除掉任何光束畸变，但是代价就是存在功率损耗。（这与上面采用针孔方法中的损耗类似，尽管两种装置的工作原理是完全不同的。）存在的一个问题就是偏振态会发生改变，除非采用采用短光纤或者保偏光纤。

Definition: devices which can improve the beam quality of laser beams

Alternative term: spatial filters

For various reasons, laser beams often have an imperfect beam quality, i.e. distortions of their intensity and/or phase profile. Depending on the circumstances, various types of “mode cleaner” devices may then be used for “cleaning up” the beam profile. The term “mode cleaner” is actually imprecise, because what is cleaned is not a mode (which is a mathematical concept, not a physical entity), but rather the phase and/or intensity profile of a laser beam. Another frequently used term (at least for non-resonant mode cleaners) is spatial filter.

There are resonant and nonresonant types of mode cleaners, which are very different concerning their principle of operation and the range of application. Resonant mode cleaners, mostly called mode cleaner cavities, are discussed in a separate article. The following description concentrates on nonresonant devices.

A fundamental limitation for all mode cleaners is that it is physically not possible to have an optical device which concentrates all the optical power of a multimode beam into a single-mode beam (having perfect beam quality). However, if most of the beam power is in a single mode already, the beam quality can be improved by attenuating all other modes, and this will not lead to a substantial power loss. This principle is used by various mode cleaners. A different operation principle is that of a brightness converter, which however is a laser device rather than a combination of simple optical elements, and also changes the wavelength of the light.

Nonresonant Mode Cleaners

Figure 1: Nonresonant mode cleaner, consisting of two lenses and a pinhole.

Figure 1 shows the simple setup, which can be understood relatively easily on the basis of Fourier optics. The (somewhat “spoiled”) input beam coming from the left-hand side is focused to a small spot with a lens. In the focal plane, different spatial Fourier components of the original beam are to some extent spatially separated: each spatial Fourier component corresponds to a plane wave with a certain propagation direction, and a component which e.g. propagates slightly upwards will have its focus at a higher position. Therefore, the pinhole in the focal plane can effectively attenuate spatial components related to off-axis components. The second lens recollimates the beam, and the resulting beam has a “cleaned” intensity and phase profile.

Of course, the cleanup cannot be perfect, since each spatial component is focused to a spot with finite size. Note that simply focusing more strongly will not necessarily help, since this also reduces the spatial separation of spatial Fourier components in the focal plane. However, a certain degree of focusing is required; a rule of thumb is that the waist size must be a few times smaller than the original beam size. Further, the diameter of the pinhole must of course be appropriate; too small a hole, for example, leads not only to excessive power loss but also to a deterioration of the beam quality.

The beam cleanup can work very well when an originally close to Gaussian beam (with high beam quality) has acquired some spatially “fast” distortions e.g. by reflection on some mirrors with deposited dust particles. Such a beam is very effectively cleaned up without losing a significant amount of the optical power. (Note that one then has a coherent superposition of the Gaussian mode with fairly high-order modes, which can strongly affect the beam quality without carrying significant optical powers.) On the other hand, the method does not work very well on a really spatially multimode beam from a laser with poor beam quality. It is not only that removing the higher-order modes is then really associated with a significant power loss, but also that the discrimination of modes with similar orders is not perfect.

A practical problem for application with high-power lasers and particularly with Q-switched lasers is that the optical intensities at the pinhole can become fairly large, possibly leading to optical damage.

In comparison with mode cleaner cavities, nonresonant mode cleaners are relatively easy to align and can work with polychromatic radiation (limited by chromatic aberrations of the used lenses), but cannot reach the same amount of beam quality improvement for clearly multimode beams.

Mode Cleaning with a Single-mode Fiber

Another way to clean up a laser beam is to send the light through a single-mode fiber. If the fiber is sufficiently long to strip off light in the cladding modes (a few centimeters may be sufficient, depending on the cladding), the output will have the fundamental mode profile of the fiber, which is often close to Gaussian. Any beam distortions can be removed in this way, at the price of power losses at the fiber's input end. (This is similar to the losses at the above described pinhole apparatus, even though the physical working principles of the devices are different.) A problem may be that the polarization state can be scrambled, unless the fiber is short or a polarization-maintaining fiber is used.