The Schottky is fine where it is. It's the free-wheel diode for the inductor current.
But you could add a zener across the drain-source of the MOSFET to clamp any inductive spikes.

Post #12 is where I figured out it wasn't a zener.
A Shottky symbol so much resembles a zener symbol...especially when hand written.

 

So with an inductor 47uH, would I be able to lower my frequency without going to crazy sized capacitors? I'm scavenging inductors from old PSU's since they are pretty expensive to buy new. I think at 31Khz the circuit called for 9uH or there abouts.

You could lower the frequency but I don't think that will significantly help your spiking problem.
That could be due to the inductance of the circuit leads.
Do you have good short ground leads with the decoupling cap negative going directly to the MOSFET drain and short leads between the MOSFET drain and the anode of the free-wheel diode?
Connecting the diode anode directly to the MOSFET drain is the best since that absorbs the inductive spike from the wiring inductance in the drain circuit.

Minimizing spikes involves looking at the complete path for the high current, from the input filter through the inductor and capacitor and load, through the transistor and back to the input filter. That current path should be as short as possible.
Also look at the current path through the free-wheel diode when the transistor is off.

 

I have the leads relatively short. The components are not direct connect as they are too large, the entire circuit is containted in a 2"x2" square. The physical layout is identical to the schematic I posted, Decoupling cap is connected directly to the load negative, then there is a 1" lead to the diode / FET source (they are within 1/8" of each other). I could potentially move the cap quite a bit closer on the ground side, which would increase it's lead length to the inductor.

What if I added a snubber between the Capacitor and free wheeling diode?

 

Is it a trick of the light, or does the red wire at the bottom left corner of the pic have solder splashes on it which might be shorting to a pad?

 

Is it a trick of the light, or does the red wire at the bottom left corner of the pic have solder splashes on it which might be shorting to a pad?

Trick of the light, you are seeing some 3145 RTV that I used to hold the wires in place so they didn't get pinched accidentally.

The circuit worked great right up until I actually hard switched on it and got the transient voltage spikes.

 

Overvoltage and ringing comes from the parasitic inductance which is related to the wire loop: "input source plus - transistor drain - transistor source - diode cathode - diode anode - source minus" in a buck converter. The circuit shown in post #3 is what I would expect as buck converter circuit.

If you reduce switching speed by putting a gate resistor (the larger, the slower), the overvoltage should decrease. On the other hand, the switching losses will increase and the converter efficiency will go down. Switching frequency (PWM) should have no impact on the overvoltage and ringing.

 

Is there a practical difference between switching on the positive leg and the negative leg? I chose the way I did because I had a N-Mosfet, should I be looking at a P-Mosfet?

If I add a snubber to the exsisting circuit to clamp the spikes, would I add it across the diode?

 

I have a test circuit working-ish. It's still using 12v Vcc on the driver and the IRFZ44N's, I just gave the driver it's own source and threw a 18v zener across it for some protection, while I just test and learn until my new parts arrive.

This circuit has a 2200uF cap in parallel with the load instead of a 1000uF, and the inductor is a bit larger (again, available parts). I have removed the input filter circuit and only left a 240uF cap to help with switching load.

@DickCappels I've added a snubber across the schottky, .33uF (I think a MEF cap? The yellow square guys in a PSU) with 10 ohms (not truly calculated, more of a rough estimate and available parts). The 40-50v ringing is gone, yay! Thank you Mr. Cappels! Still some light ringing and other unexpected results, but I think that they are workable. Someone more experienced than I am could probably explain why and if what I am seeing is expected and or due to my 'roughing it' when adding the snubber. I don't think that it will have any adverse side effects on the controller or peltier, more just curious about it.
This from PSU ground to the load: (ac coupling with PWM at 20 - 4.5ish volts output)

This is across the load using the drain of the FET as ground for the scope:

The duty cycle does not select the voltage in a linear curve as I expected. Is this due to me limiting the available current by regulating the voltage I'll allow? This was a test done with 4 6ohm 50 watt resisters in parallel to attempt to simulate a peltier. Not perfect, but it at least gives me a starting point with a high current load.
0-255 is 0-100% in the micro controller I am using. The chart was generated from 34 test points.