1.
https://www.reddit.com/r/microgrowery/c ... for_a_4x4/ PFD seems to have totally disappeared, but he already did basically that, adding 440 and 460 to Samsung LEDs with their own pump at 450nm, plus 660nm diodes, and had the measurements for it, too. Result: a wider, smoother blue bump, instead of the narrow spike. Those things made tastier more aromatic lemon and common thyme, oregano, and dark purple opal basil, than the sun itself, IME. The 660 comes in right where most whites start to drop off, filling it in. As long as there are enough emitters spread out enough, you basically get blanket coverage. At 120 degrees (typical bare white LED), you want a distance to canopy of no less than 2/3 the widest horizontal distance between LEDs, FI. 445nm might be a good add on to 460nm or longer blue pump, but most are 450 or 440. If you want to try going down that road, try to keep the added blues 10-15nm off, so that they don't just make for a bigger main spike - that can result in a lot more waste heat inside the leaves, and could also signal to the plant that there's too much light, retarding leaf growth.
2. Generally yes, though there are exceptions. With some series, the CCTs are different enough. FI, if you look at the new Bridgelux EB gen 3 data sheets, you'll see they are clearly using a very different phosphor recipe, and I would guess a different pump, from 4000k to 5000k, though 3000k to 4000k is just the typical change in blue/green/red balance. With 80 CRI Samsungs, you'd need to include most CCTs from 2700k to 6500k to really get something different.
3A. Those add to the output in spectral ranges where whites are usually weak, but where plants still show a good bit of response. Going back to the linked page, note the 660nm spike. Without the added LEDs, it would follow the gentle drop off that begins around red-orange, instead. Blue and red are also pretty efficient as single color LEDs, while cyan and green don't tend to be. In addition, sub-480nm blues, right down through UV-B, can cause plants to perform expensive actions like making secondary pigments, much more efficiently than just blasting them with more overall light. Since sunlight comes with strong blue during intense mid-day sun, and that usually comes with UV, more nearer-UV blue can get some nice stress responses going.
3B. Down to around 400nm or so, there is still a pretty strong photosynthetic response, so it can also add to PPF efficiently. If you look at most data sheet graphs, from about 10nm shorter than the pump, and then around 640nm on up, output is often negligible, despite there being receptors, pigments, and chlorophyll peaks below ~450nm and above ~640nm. Even high CRI only go so far, in this regard, often still having little in the 660-700 range, where the McCree curve shows a strong response, there is a second red chlorophyll peak, and where there can be some synergistic effects with the shorter wavelength blues, and far red (it seems 5-10% far red is plenty for the Emerson Effect, and even 80 CRI emitters often get into that range, now).
3C. Having said all that, while there isn't anything super conclusive quite yet, plants seem to adapt well to artificial lighting, with wonky spectrums, when it comes to photosynthesis. So, while different added wavelengths can be good for use with triggering desired morphology, improved spectrums of light sources don't improve growth/weight by a significant amount, once a, "good enough," point is reached. If you have enough blue, red, and far red for the Emerson effect, PPF is more or less PPF. Plankton have more research on that, since knowledge of their environmental light changes predates white LEDs (effects of water depth). While not directly applicable to plants, they adapt to varying amounts of blue and green, including shifting peak sensitivity in chlorophylls a and b (or at least the related photosynthetic action) for longer blue and cyan wavelengths, as shorter wavelengths got filtered out. Plants are unlikely to be so dynamic, but similar adaption when grown from seedlings under artificial light would seem like a reasonable hypothesis.