All we've been saying, really, and continue to say, is that the input variables are numerous and you have not isolated them for your results. You are presenting results and then drawing conclusions. It's little different to speculation when you have no way of falsifying the results you're presenting.
You also dismissed the results that confirmed what I had been saying about temperature differentials - your results exactly confirmed the variations that I predicted. You say the differences are too small to matter, but you haven't tested anything to verify that.
It's a good subject and a good test, but your results are not comprehensive - there's room for more testing, not least of which is to isolate the various factors that could be contributing to try to determine what it is that's actually causing the issue, rather than assuming the answer.
Quite the opposite, really.
Long-established knowledge gives us the following facts as a starting point:
1. That paper is a hygroscopic material, and that a sheet of paper changes its dimensions significantly as the relative humidity of the surrounding air changes.
2. That the coefficient of thermal expansion of paper, taken as a whole sheet, is extremely low - less than glass, metal, concrete, plastics of all kinds and even graphite. Part of this is due to the porous nature of paper - individual fibres may expand (although the coefficient of thermal expansion even of individual fibres is very low in the longitudinal direction, and still fairly low across the fibre) but, when they expand, they simply fill the gaps between the fibres, rather than causing the sheet of paper to expand as a whole.
3. That the major source of stress at an interface between two layers made of different materials is the difference in dimensional change between the two materials when exposed to the same environmental conditions. This is regardless of what the two materials are - paper and coating, copper and zinc, concrete and steel.
4. That the response of the individual layers to stress at the interface depends on the Young's modulus and yield points (elastic limit and fracture point) of the material, as well as the strength of the bond between the two materials.
5. That exposure to ultraviolet light and atmospheric pollutants can alter the physical characteristics of a material, particularly a material based on organic molecules.
From your arguments, you obviously disagree with point (2), but the results of my second experiment demonstrate that point 2 isn't even relevant when considering the results of the first experiment.
From these principles, we can infer the following hypotheses:
1. That the Young's modulus and yield points of a material (in this case, the inkjet coating) can change as the material is altered on exposure to ultraviolet light and atmospheric pollutants. Therefore, a stress that doesn't affect a new coating (e.g. reducing the RH to 10%) could possibly cause failure in a coating whose stress-strain characteristics have been altered by UV light and pollutants.
2. That failure of a coating material can occur either within the coating material itself (cracking, buckling and disintegration) due to the stress in the material exceeding the yield points of the material, or at the junction between the coating material and the substrate (delamination and 'flaking'), due to failure of the bond between the two materials.
3. That the vast majority of dimensional change in paper is due to hygroscopic effects rather than thermal expansion.
Obviously, you disagree with the third hypothesis. But the second test I performed renders that argument irrelevant (explanation further down).
The first test establishes the following:1. That different papers fail at different rates, depending on their general structure - coating on polythene on paper, or coating on baryta on paper, or coating directly on paper.
2. That the mechanism of failure seems to differ between papers - delamination of the coating on RC papers, delamination and cracking of the baryta layer in the baryta papers, and no obvious failure seen in the coating-on-paper papers (although other studies have shown that microporous layers develop microscopic cracks under heavy ink loads).
You contended that the test is irrelevant, because putting prints up against a window exposes them to much greater
peak temperatures as compared to a framed, matted print hanging on a wall away from the window and exposed to sunlight for 1-2 hours at a time. The second test refutes that.
The second test establishes the following:1. That the temperature of an unframed piece of white paper remains near the ambient room temperature (1 or 2 degrees warmer) even when in direct sunlight.
2. That the front and back of a piece of paper remain within around half a degree of each other when one side is exposed to strong solar irradiation.
3. That the air temperature in the small gap between the paper and the window is similar (less than 2 degrees warmer) to the air temperature at the back of the room, out of direct sunlight.
You contend that these small temperature differences matter, however small they may be and, therefore, invalidates the results. However, the other part of the test completely invalidates your argument. The tests involving black paper and framed paper establish:
4. That dark paper (e.g. prints) becomes a lot hotter than white paper when exposed to the same level of solar irradiation, greatly exceeding the surrounding air temperature. Therefore, a typical print would become much hotter than my unprinted paper samples when exposed to the same level of sunlight.
5. That prints enclosed in a frame behind glass get a lot hotter than unframed prints or paper, even when the prints/frames are well back from the window. Therefore, a framed print gets a lot hotter than an unframed print when exposed to sunlight. That's not to say that there aren't other protective features of framing that make it worthwhile, but that protection from the thermal effects of direct sunlight isn't one of them.
The fourth and fifth findings of the second test render all the arguments about 'extreme temperatures' completely invalid, since they show that a piece of printed paper well back from the window (but still receiving an hour or so of direct sunlight) and enclosed within a frame gets much hotter than a white piece of paper up against the window - much more than the 1 or 2 degrees above room air temperature experienced by the unframed white paper.
Essentially, what the second set of tests show is that the 'extreme temperature' argument against the results of the first test are invalid, since they show that a framed print will be exposed to much higher temperatures than a white, unframed piece of paper ever will.
