Everyone heard about lithium tantalate pyroelectric mediated deuterium ion acceleration fusion? It's desktop scale stuff for neutron product rates in the hundreds per second range (~500). Small, but useful for tinkerers.
It's a type of piezoelectric crystal that exhibits spontaneous polarization of voltage in the direction of one crystal face as the result of some odd cancelings of positive and negative ions in the crystal shape/lattice (3m, trigonal pyrimidal)-this is called the pyroelectric effect. With simple thermal stresses to the crystal (but you can't go above a certain Curie temperature which varies for each material, about 600C for this case), like heating it up to 100C and letting it cool over 15 minutes, the random polarization of multiple small domains in the monocrystal tend to align which each other following a power law distribution rate that varies with a couple of the 3m crystal tensor asymmetries (it also shows typical Barkhausen noise and jumps just like magnetic domains do in slowly changing B-fields, which is nifty). Anyway, this complex polarization process/equilibrium can be pushed toward very high levels of homogeneity in domain polarization by inducing energy into the crystal lattice with thermal expansion stresses. The crystal lattice concentrates this energy considerably (gist), and can show induce voltages gradients in the gigavolts/meter range from the positive to negative faces of the z-axis of the crystal.
In practice this can lead to voltages at the surface great enough that if you stick a thin electrically conductive needle (200nm is the optimal diameter for tungesten) to one of the (flat) z-axis faces it'll generate a local voltage gradient of 100keV or so and accelerate positive ions very fast (or shoot off high energy electron beams from the -z face, but that's another story). There's ample energy, in fact, due to some non-linear properties of the mean free path of light ions at low vaccum, that you can do dueterium-deutrium fusion with plastics that have had their hydrogen replaced with it's heavier 2 neutron isotope...but only if you attach the aformentioned dueterated plastic onto the -z face of another crystal and run them in bipolar configuration (one is heating while another cooling).
[http://www.nature.com/nature/journal/v434/n7037/abs/nature03575.html]
Unfortunately lithium tantalate crystals of the 'congruent'type not stochiometerically equal to litium tantalate but have something like 1% less lithium...it's one of those tweaks, the crystal domain polarization is more homogenous this way that strictly making the crystal as chemistry suggests) and of optical quality (good homogeneity) are really expensive, like $500 for a tiny unpolished z-axis cut wafer 2cm in diameter if the company will even talk to you.
The stereotypical oxide crystal growth method is the Czochralski; for boule growth use quartz crucibles and a seed crystal to pull the desired shape out of the molten oxide (while spinning the seed crystal at a very consistant rate around 30 turns a minute). This uses induction based heating, which is awesome and easy to do with cheap IGBTs I already have (Insulated gate bipolar transistors). I know that ferroelectric single-domain surface layers are formed in z-cut LiTaO3 following heat treatment above the Curie temperature and they can become crystalline at a temperature as low as 450° C (which is autoachieved by the feed-forward balance of a good H-bridge induction heater (resistance goes up with temp, etc)). Double quartz crucibles though are not so awesome. I don't think I can do this.
Looking further I discovered all sorts of high vaccum ion-this-that deposition methods. These were also not cool.
So, I gave up on growing boules in the basement and focused on something which I think is considerably cooler, formation of thin layers of monocrystaline lithium tantalate. There are a couple ways to do this, the most common being: disolve everything (ex: tantalum(V) ethoxide and lithium acetate) in ethanol, pour it onto a very steadily spinning disk of polished sapphire with the c-face oriented up, or however you want the thin film monocrystal to be oriented. Up is good. Polished c-cut optically homogenous sapphire is not cool. Some fellows say that similarly c-cut silicon which has the same type of 3m crystal point structure (approximately) will also work. I haven't confirmed this yet but I'm kind of counting on it.
http://dx.doi.org/10.1016/j.jallcom.2005.01.063
In this paper, a new sol–gel method was developed to prepare the LiTaO3
powder and thin film. In this method, hydrogen peroxide aqueous solution
(H2O2) reacts with tantalum and lithium ethoxides ethanol solution. The
property of transformation into pure LiTaO3 was demonstrated through
annealing the precursor powder dried from the solution. Nano-sized lithium
tantalate particles with dimension of 40–50 nm were obtained at a sintering
temperature of 700 °C. Such aqueous solution is suitable for fabrication of
lithium tantalate thin film. LiTaO3 thin films on Si and Pt/Si were
fabricated through heat treatment of the spin coated wet films. The obtained
thin film is characterized through measuring its ferroelectric and dielectric properties.
Also, from elsewhere (before I started copying doi numbers):
"Oxidation experiments were conducted separately with ambient air and O2 at
atmospheric pressures at 600 deg. C. Results indicated an increase in green
absorption from 0.01 cm-1 with congruent material, and a drop down to 0.0008
for oxygen annealed. IR absorption at 1064 nm, the UV edge absorption band,
and photorefraction stayed relatively constant for all three samples."
Green absorbtion is okay with me because I do not plan to shine polarized lasers through the material and bend them with index of refraction tricks caused by matlab generated pyrimidal macroscale photomasks and subsequent domain wall polarizations canceling out (in the math too) to create an effective total index of refraction of 1 from each positive negative domain wall border switch, and efffectively make a lot of little prisms that bend the extraordinary part of the polarized light, but I digress. The point is a little oxygen contamination is probably okay.
http://www.springerlink.com/content/u12642u96004l1lu/ - Powder Production
Preparation of oriented lithium tantalate thin films using molecularly
modified tantalum(V) ethoxide and lithium acetate
http://www.springerlink.com/content/v472170217857161/
http://dx.doi.org/10.1016/S0040-6090(99)01102-5
We have successfully prepared highly c-axis oriented sol–gel derived
LiNb1−xTaxO3 films on Si(111) substrates. The orientations of LiNb1−xTaxO3
films for different values of the tantalum composition, x, are investigated.
Highly c-axis oriented films were obtained in the composition range of
0<x<0.33 and the degree of the orientation was found to be relatively low
when 0.5<x<1. The c-axis orientation of the films is believed to result from
a weak effect caused by the 3-fold symmetry match between the film and the
Si(111) substrate. C-axis oriented LiNb1−xTaxO3 films are very useful in
integrated optical devices, pyroelectric devices, and
metal-ferroelectric-semiconductor non-volatile memory applications.