It can't be pulsed and single frequency. This is a logical impossibility. For the laser to have a single frequency it must be continuous. A pulsed laser must have some bandwidth. Single longitudinal mode lasers are readily available and can be "Q" switched. Alternatively continuous laser beams can be chopped outside the cavity.

It is not clear to me what, "a method to lock the cavity for locking" means. For a narrow bandwidth there are not any modes to lock to each other. There is just one longitudinal mode. I think a few more details of the requirements are needed.

It can't be pulsed and single frequency. This is a logical impossibility. For the laser to have a single frequency it must be continuous. A pulsed laser must have some bandwidth. Single longitudinal mode lasers are readily available and can be "Q" switched. Alternatively continuous laser beams can be chopped outside the cavity.

It is not clear to me what, "a method to lock the cavity for locking" means. For a narrow bandwidth there are not any modes to lock to each other. There is just one longitudinal mode. I think a few more details of the requirements are needed.

Acousto-Optic Modulators can run in the MHz regime. You need to design the laser to run that fast, including the seed source. Lasers with rep rates in GHz range have been reported:

Opt Lett. 1990 Oct 1;15(19):1070-2.

Compact mode-locked solid-state lasers at 0.5- and 1-GHz repetition rates.

Walker SJ, Avramopoulos H, Sizer Ii T.

Abstract

A compact laser system that combines a GaP mode locker and a diode-pumped Nd:YAG laser to produce 1064-nm pulses at 0.5- and 1-GHz repetition rates is described. Extension of this research using a Ti:sapphire-pumped Nd:YLF laser is also presented. Stable optical 1047-nm pulses with durations of 6.2 psec at a l-GHz repetition rate have been produced.

http://www.ncbi.nlm.nih.gov/pubmed/19771000 

I agree what  Anthony Dymoke-Bradshaw have described. I think you can buy mode locked CW Nd.YAG  laser. However it will be costly to procure such laser. thanks

One approach is to use a external cavity diode laser which is "single frequency" to start with, and then gain switch it to get transform limited pulses.  This approach is typically used to seed narrow linewitdh pulsed fiber lasers.  

If you would like to reach out to me directly we can discuss this in more detail.  I have done used/built these lasers many times and can recommend some sources. 

-Robert  

P.S. You need to define what you mean by single frequency.  1MHz linewidth is more or less a standard for "single frequency" external cavity diodes but with the right drive electronics I have seen less than 100kHz myself and others I know have gotten closer to 10kHz.   

Pulsed lasers do not run single frequency but they can be made to run single longitudinal mode (SLM). In a lamp pumped laser there is a chirp on the pulse on the order of tens of MHz due in some part to the rod changing length during the lamp pulse. DPSS should be better as the pumping is typically CW so the gain can equilibrate. You can approximate the linewidth assuming a Gaussian pulse as LW = 0.44/pulsewidth. So a 10 ns pulse has a theoretical transform limited bandwidth of 44 MHz. There are two other values, one - what you can actually measure, and two - what the manufacturer will specify. In practice people usually measure line widths < 2x transform limited when seeding pulsed lasers. I've only ever seen one group claim pulses within 10% of the transform limit.

Injection seeding of low repetition rate lamp pumped laser cavities using a single frequency CW seed laser (you'll see the term NPRO used often in the literature) is standard practice. You may be able to seed higher repetition rate diode pumped solid state lasers using a similar technique - depends on your pulse energy requirements. Eolite has some rod type fiber lasers that are capable of a mJ or so at these kinds of rep rates. http://www.eolite.com/Products/RodFiberLasers.aspx. 

Locking of a seeded electro-optically q-switched cavity is usually done in lamp pumped lasers by modulating the cavity length with a piezo mounted mirror and then looking at the build-up time reduction of the seeded pulse. The build-up time minimizes when the cavity length is an exact multiple of the seed laser wavelength. You need to mode match the seed beam to the cavity so as much as possible of the fundamental mode is physically overlapped with the seed beam - otherwise there will be mode competition and mode pulling. You will need to tune the seed frequency coarsely to be on the center of the slave gain spectral profile which needs to be done when everything has thermally equilibrated. You will also need to watch out for spatial hole burning if you are using an unpolarized gain medium - Nd:YAG for example. Two quarter wave plates on either side of the rod will eliminate hole burning if you are not using Brewster faces.

If your application is holography or some other phase sensitive technique then you can approximate the coherence length of the laser as 1/(2*bandwidth). A typical lamp pumped QS Nd:YAG might have a bandwidth of say 1 cm-1, so the coherence length is around 5 mm. When a pulsed lamp pumped laser is injection seeded, although no-one will specify this, an injection-seeded pulse is thought to be coherent over its full physical length. So if you have a 4 ns FWHM pulse from an injection seeded laser, the pulse is roughly 4 feet long FWHM, and if you were doing holography you could reasonably expect to get coherence over 4 feet. This has been demonstrated - for example Laser Reflections http://www.laserreflections.com/

Hope this helps!

What kind pulse width and peak power do you need? You could simply modulate the CW laser with an AOM or a EOM, and a pulse generator.

Pulsed single longitudinal mode with 50 -75 MHz line widths at 10 khz can be obtained by pulsed amplification of cw seed laser (another simpler option).  pump laser source for pulsed amplifier can be nsec pulse width and kHz rep rate. Fourier limit and chirping play a role in dictating the spectral width of the  amplifier output. longer pump laser pulse width can reduce the spectral width.

I agree with John Schaibley. You need to specify your requirements in terms of power and duration. I have done this for limited run lengths using a controlled intra cavity  pockels cell and partially opening it. This has gone up to Mhz frequencies for only 10-100 pulses of 5mJ output. Duration can be achieved by increasing flash lamp times and taking lower energies. This usually works for around 3 widely spaced modes. No perfect but can be done for reasonable cost.

For locking and seeding you can choose a Pound-Drever-Hall method, or its variation as in "2-µm Doppler lidar transmitter with high frequency stability and low chirp" by V. Wulfmeyer, M. Randall, A. Brewer, and R. M. Hardesty ; Optics Letters, Vol. 25, Issue 17, p. 1228 (2000) http://dx.doi.org/10.1364/OL.25.001228

Don't be discourage by the article's system pulse repetition rate (~200 Hz). We operated a similar system at a much higher rate.

May be it can be tried by using a single frequency CW laser along with a fast switching Pockel cell shutter.

I have achieved this using a infra cavity eltaon in a ruby pulse cavity by partially Q switching the Pockels cell. The system produced a train of pulses down at 1microsec separation over 500ms with around 10mJ/pulse single mode operation. The firing system cost around £20k.

Alternatively a continuous rate of pairs of pulses can be produced using twin or multiple headed diode pumped Yag lasers. But this is not a cheap solution.