re: Re: re:Re: Interferometer designs for DRO or servo control?

Jon wrote:

>Subject: Re: re:Re: Interferometer designs for DRO or servo control?



>Elliot Burke wrote:

>>2. the laser wavelength is not completely stable unless you have a (duh)
>>stabilized laser. There are two sorts of common lasers used for
>>interferometry, the HeNe and the diode. The HeNe has the remarkable
>>property that its wavelength is known to 1 ppm, so if care is taken (see

1.

>>above) distance measurements of this precision can be made. The HeNe can
>>mode hope, which can cause the fringe count to jump even then there is no
>>motion, but there are ways to prevent this I won't get into here.
>
>>The second common laser is the diode type. A cheap 650 nm laser pointer
>>will make a nifty source for your interferometer, but its wavelength is

not

>>exactly known (can be calibrated) and is unstable with temperature and

drive

>>current.
>
>The other problem with the common diode lasers is that the coherence
>length is
>just a few mm. Just as with HeNe, it is many times larger than the
>cavity length,
>but the cavity in the diode laser is just um long!

>Jon

Jon,
Short coherence length of laser diodes was true for the first visible laser
diodes I measured 5 years ago or so, the 670 nm ones.

When they switched to 650 nm a few years ago (for DVD?) the design was
changed, and now the coherence length is quite long. There are articles in
the optics testing literature of people making measurements with several
meters of optical path difference with these newer (but still cheap!)
diodes.

I have used them up to about 300 mm path difference with no reduction in
fringe contrast.

The coherence length has a complex relationship with cavity length. A short
(<6") HeNe has a very long coherence length, because there is only one mode
that fits within the gain bandwidth of the HeNe line (about 500 MHz wide,
IIRC). A longer cavity can support more modes within the 500 MHz. For an
example, at 632 nm the frequency is about 4.74 x 10^5 GHz. A 6" long cavity
has resonance at n x 1 GHz, where n is an integer, 237342 for the case of
the HeNe. The next mode, 237343, has a frequency 1 GHz higher, which is
outside of the gain bandwidth of the HeNe.

Long HeNe tubes can support more than one longitudinal modes. This does not
necessarily mean that they have short coherence length, however. The
coherence function is periodic, with peaks as optical path difference equals
the tube length.

Conventional diode lasers have cavities a few 100 µm long (<1 µm x <10 µm
transverse dimension). There are some complex cavity designs (DFB) that
narrow lines, but they aren't used in the cheap lasers. A cavity 100 µm
long of index of refraction about 3 has resonance at c/(n l) = 1000 GHz.
This is larger than the gain bandwidth for the amplifying materials used
now. So intrinsically these lasers operate in a single frequency mode. Of
course if the gain of the laser is reduced by reducing current or damaging
the material this may no longer be true.

It is quite easy to verify the long coherence length of laser diodes, and
amusing too to make sensitive instruments. It is important with diode
lasers not to reflect the laser light back into the laser junction, as this
will upset the cavity and make the output unstable. Use a circular
polarizer or bettter optical isolator to prevent this, or just don't use a
retroreflector for a reference mirror.

BTW, I use an optical flat to measure the flatness of machined surfaces. If
you get a good (shiny) finish, there is enough specular reflectance to see
fringes between an optical flat and the machined surface. They will
probably be of low contrast.

Elliot