The obvious advantage of using laser ultrasound is that it can be totally remote over relatively large distances. Below is a brief description of what laser generated / detected ultrasound is all about. Please read the references for further information as it is a very broad and area which would be impossible to thoroughly cover here.
Applications in which we have used lasers are covered in different research areas - please see weld inspection, plasma generation, on-line ultrasound.
Pulsed lasers can be used to generate ultrasound in a range of samples. It is generally described in terms of 2 extreme regimes, the ablative regime and the thermoelastic regime [1-3]. In reality it is a combination of both mechanisms where one will appear to dominate. In general a laser is a very strong ultrasonic source and can generate ultrasonic signals orders of magnitude higher than contact piezoelectric transducers.
Fast laser pulses generate high frequency broadband (wide frequency range in one pulse) ultrasonic signals. The exact frequency content depends on the laser and material properties. Pulsed lasers with rise times in the range of hundreds of nanoseconds up to picoseconds have been used to generate ultrasound.
In the ablative regime the laser energy is focused onto the surface of the sample which at sufficiently high energy densities will form a plasma. The plasma may consist of the vaporised sample and/or air breakdown depending on the sample and the experimental conditions. There are situations where a strong ablative-like source can be generated in a totally non-destructive way such as by using a pulsed CO2 laser to focused onto an aluminium sample .
In the thermoelastic regime the laser energy is directed onto the surface of the sample at a sufficiently low energy density to avoid ablation. The laser energy is absorbed within the skin depth of the sample and rapidly heats that small volume of the sample. While there may be no superficial damage to the sample there may be subsurface damage due to the absorption of laser energy. This is a major drawback when trying to generate on samples like carbon fibre composites where there is high absorption of the laser energy and poor thermal conductivity.
As described for the two extreme cases above laser generated ultrasound is not generally truly non-destructive, but for most industrial type samples any damage is usually on a microscopic scale and is negligible. The resulting ultrasonic waveforms and wavefronts from laser sources are extremely complicated. All ultrasonic modes are simultaneously generated so that a waveform will contain compression (longitudinal), shear and surface waves.
What lasers do we use for generation?
Nd:YAG - 200mJ, 10ns pulse, 1064nm
Nd:YAG - 400mJ, 10ns pulse, 532nm
TEA CO2 - 5J, 100ns pulse 10.6mm
Laser detection of ultrasound can be performed in a variety of ways, and these techniques are constantly being improved and developed. There is no 'best' method to use in general - it requires a knowledge of the problem and an understanding of what the various types of laser detector can do. Commonly used laser detectors fall into two categories, interferometric detection (Fabry Perot, Michelson, time delay, vibrometers and others) and amplitude variation detection such as knife edge detectors.
Knife edge type techniques
The idea behind a knife edge type detector is that when an ultrasonic waves arrives at a point on the surface of a sample illuminated by a beam of light, the reflected (or transmitted) light beam will move . The beam will move at the rate at which the ultrasonic wave disturbs the illumination point.
It is very difficult to measure exact wave amplitudes or true waveform shapes from the beam deflection approach but it can be configured to give easy or automated alignment in many cases. It also tends to be fairly inexpensive compared to other optical detection techniques.
Knife edge detectors are less sensitive, considerably less expensive and easier to align than interferometers.
There is a vast range of different approaches when in comes to interferometry, each with their own advantages and disadvantages. A good review to check out is the recent paper by Dewhurst et al , which gives a good idea of what a broad area it is.
To make a few generalisations about interferometers they can have high sensitivity, high bandwidth and work on rough surfaces on moving components. Unfortunately you don't get all the good features in one type of interferometer so again its down to looking at what the experimental conditions are, what do you need to measure, and how much money have you got.
By means of example the Michelson interferometer that we use in our labs has a bandwidth of 80MHz, and can detect displacements down to less than an Angstrom, and is pretty insensitive to low frequency background vibration . Unfortunately its near impossible to set up an automatic alignment for on-line applications and we need the surface of the sample to have high reflectivity.
Fabry-Perot interferometers can be used for on-line measurements and will work on surfaces with poor reflectivity. Unfortunately the bandwidth is limited and the sensitivity is lower than that of the Michelson interferometer.
References for further reading
 White R.M., "Generation of elastic waves by transient surface heating", J. Appl. Phys., vol. 34 , pp3559
 Birnbaum G. and White G.S. "Laser Techniques in NDE" in Research Techniques in Nondestructive Testing vol 7, Ed. R S Sharpe, New York: Academic, 1984, pp 259-265
 Hutchins D.A., "Ultrasonic generation by pulsed lasers" in: Physical Acoustics, (ed. W.P. Mason and R.N. Thurston) Academic Press N.Y., vol. 18., 1988, pp21-123
 C Edwards, G S Taylor and S B Palmer, The CO2 laser - a new ultrasonic source. J. Nondestr. Test. Eval., 1990, 5, pp 135-143
 Monchalin J-P, "Optical detection of ultrasound", IEEE Trans. UFFC-33, 1986, pp485-499
 Dewhurst R.J., Shan Q.,"Optical remote measurement of ultrasound", Meas. Sci. Tech., 1999, Vol.10, No.11, pp.R139-R168
 White R.G. and Emmony D.C., “Active feedback stabilisation of a Michelson interferometer using a flexural element”, J. Phys. E, vol.18, 1985, pp658-663