Laser-ultrasonics uses lasers to generate and detect ultrasonic waves. It is a non-contact technique used to measure materials thickness, detect flaws and materials characterization. The basic components of a laser-ultrasonic system are a generation laser, a detection laser and a detector.
Ultrasound generation by laser 
The generation lasers are short pulse (from tens of nanoseconds to femtoseconds) and high peak power lasers. Common lasers used for ultrasound generation are solid state Q-Switched Nd:YAG and gas lasers (CO2 or Excimers) . The physical principle is of thermal expansion (also called thermoelastic regime) or ablation. In the thermoelastic regime the ultrasound is generated by the sudden thermal expansion due to the heating of a tiny surface of the material by the laser pulse. If the laser power is sufficient to heat the surface above the material boiling point, some material is evaporated (typically some nanometres) and ultrasound is generated by the recoil effect of the expanding material evaporated. In the ablation regime, a plasma is often formed above the material surface and its expansion can make a substantial contribution to the ultrasonic generation. consequently the emissivity patterns and modal content are different for the two different mechanisms.
The frequency content of the generated ultrasound is partially determined by the frequency content of the laser pulses with shorter pulses giving higher frequencies. For very high frequency generation (up to 100sGHz) fs lasers are used often in a pump-probe configuration with the detection system (see picosecond ultrasonics).
Ultrasound detection by laser 
Ultrasound may be detected optically by a variety of techniques. Most techniques use continuous or long pulse (typically of tens of microseconds) lasers but some use short pulses to down convert very high frequencies to DC in a classic pump-probe configuration with the generation. Some techniques (notably conventional Fabry–Pérot detectors) require high frequency stability and this usually implies long coherence length. Common detection techniques include: interferometry (homodyne or heterodyne  or Fabry–Pérot) and optical beam deflection (GCLAD) or knife edge detection.
With GCLAD, (Gas-coupled laser acoustic detection), a laser beam is passed through a region where one wants to measure or record the acoustic changes. The ultrasound waves create changes in the air's index of refraction. When the laser encounters these changes, the beam slightly deflects and displaces to a new course. This change is detected and converted to an electric signal by a custom-built photodetector. This enables high sensitivity detection of ultrasound on rough surfaces for frequencies up to 10 MHz.
In practice the choice of technique is often determined by the physical optics and the sample (surface) condition. Many techniques fail to work well on rough surfaces (e.g. simple interferometers) and there are many different schemes to overcome this problem. For instance, photorefractive crystals and four wave mixing are used in an interferometer to compensate for the effects of the surface roughness. These techniques are usually expensive in terms of monetary cost and in terms of light budget (thus requiring more laser power to achieve the same signal to noise under ideal conditions).
At low to moderate frequencies (say < 1 GHz) the mechanism for detection is the movement of the surface of the sample, at high frequencies (say >1 GHz) other mechanisms may come into play (for instance modulation of the sample refractive index with stress).
It is interesting to note that under ideal circumstances most detection techniques can be considered theoretically as interferometers and, as such, their ultimate sensitivities are all roughly equal. The reason for this is that in all these techniques the interferometry is used to linearize detection transfer function and when linearized maximum sensitivity is achieved. Under these conditions photon shot noise dominates the sensitivity and this is fundamental to all the optical detection techniques. However, the ultimate limit is determined by the phonon shot noise and since the phonon frequency is many orders of magnitude lower than the photon frequency the ultimate sensitivity of ultrasonic detection can be much higher. The usual method for increasing the sensitivity of optical detection is to use more optical power. However, because the shot noise limited SNR is proportional to the square root of the total detection power this has limited effect and it is easy to reach damaging power levels before getting an adequate SNR. Consequently optical detection frequent has lower SNR than non optical, contacting techniques. Optical generation (at least in the firmly thermodynamic regime) is proportional to the optical power used and it is generally more efficient to improve the generation rather than the detection (again the limit is the damage threshold).
Techniques like CHOTs can overcome the limit of optical detection sensitivity by passively amplifying the amplitude of vibration before optical detection and can result in an increase in sensitivity by several orders of magnitude.
Industrial applications 
Well established applications of laser-ultrasonics are composite inspections for the aerospace industry and on-line hot tube thickness measurements for the metallurgical industry.
- C.B. Scruby and L.E. Drain, Laser Ultrasonics, (Adam Hilger: Bristol), 1990.
- J.W. Wagner and J.B. Spicer, ‘Theoretical Noise-Limited Sensitivity of Classical Interferometry,’ Journal of the Optical Society of America B, Vol. 4, no. 8, p. 1316, 1987.
- J.-P. Monchalin and R. He’on, ‘Laser Generation and Optical Detection with a Confocal Fabry-Perot Interferometer,’ Materials Evaluation, Vol. 44, 1986, p. 1232
- J.N. Caron, Y. Yang, J.B. Mehl, and K.V. Steiner, ``Gas coupled laser acoustic detection for ultrasound inspection of composite materials", Materials Evaluation, Vol. 58, No. 5, 2001, p. 667.
- J.P. Monchalin, “Laser-ultrasonics: from the laboratory to industry,” Review of Progress in Quantitative Nondestructive Evaluation, 23A, eds. D. O. Thompson and D. E. Chimenti, AIP Conference Proceedings, vol. 700, American Institute of Physics, Melville, NY, pp. 3–31 (2004).