Hey, I just happen to know some basic buoyancy math (an occupational necessity).
A cubic foot of volume displaces 63.2 pounds (63 in filtered/purified water like pool water; 65 in average seawater). There are 1728 cubic inches in a cubic foot. My calculator suggests that is 0.036#/cubic inch (or a bit more than 1/2 ounce.)
Now, boting obtuse maths aside, on to practical application.
Let us presume the 1/32 helo weighs an even pound. To support 1 pound in fresh water requires 1/63 cubic feet. Or, 1/126 c.f. in two floats each. 1/126 * 1728 = 13.75 cubic inches.
Let's call a prototype pontoon 4" in diameter, call that 12/8 to scale. Radius, thus, is 0.75. So, πR² gives us π0.75² or 1.76715" Dividing into our target is 13.75 ÷ 1.76715 = 7.78089 inches long.
Which is just short of 21' scale feet long--not terrible.
Except. we want the pontoons to not be submerged fully, we probably want them at least 50% out of the water. Sadly, that means doubling the dimensions, and 8' pontoons 43' long are not going to be any sort of scale.
This leaves us a problem. The easy answer would be to put the helo on a diet. A 1/2 lb. helo would sit on 50% submerged pontoons at our 4' x 21' scale pontoon.
Which is predicated upon the pontoons not adding much weight, and no additional nose weight being needed to prevent tail-sitting.
I'd be inclined to draw-moulding the pontoons from thin plastic sheet. For a flying model, I'd take the couple ounces of weight penalty to fill the pontoons with Low Expansion foam, too If I were going foam only, I'd recommend the "pink" stuff, the high-density closed-cell sheets. Closed cell means not taking on water, the high density means inherent strength (it's designed to have concrete poured over it).
But, that's just my 2¢; others' differ.