The airgap is two thickness’s of 1.5mm flat rubber sheet cut to size.
When clamped up, the rubber kind of "grips" to the ferrite.
I originally clamped up the core halves with some long 10mm steel bolts, but fringing around the air gap caused massive eddy current heating of the bolts. Hot smells and the bolts were too hot to touch. soldering iron type temperatures !!!
So I found some half inch fibreglass rod, drilled and tapped the ends, and used short steel bolts which seem o/k. Although I had some nylon bolts on standby, just in case.
One interesting thing I discovered from all of this, was to keep the airgap well away from the windings. That is easy with a long legged U core, but with gapped EE cores, the fringing around the airgap is right in the middle of both windings.
That produced some ringing, and horrible cross regulation between +ve and -ve dc outputs.
As I can only regulate the +ve dc output, my problem was that the inverter alternately loads the +ve side, and then the -ve side, as it is a half bridge topology inverter.
But keeping the airgap fringing far away from the interleaved windings, produces super clean waveforms, no ringing, and absolutely terrific cross regulation.
I can load the secondary (unregulated) side to 2Kw and there is only a volt or two difference between the regulated and unregulated dc outputs, probably about 1% cross regulation, which is amazingly good. But the fringing around the airgap is huge and needs some caution.
Another thing to think about is using a large low ESR capacitor direct across the solar panel input. The panel is a weak current source and cannot supply high current peaks. I used some (white) Evox Rifa low ESR 2,200uF electrolytics to supply my 40A peak current.
These are also hiding behind the inductors, and supporting the flat and very inefficient heat sinks and the PCBs on the capacitor terminals.
These electrolytics are a very important feature of any switching power supply, buck, boost, or whatever you decide to use when running off solar panels.
.