After mentioning the phosphate charge curve in #25 it occurred to me that this really is a good way of gaining insight as to why a phosphate buffer may or may not work at controlling mash pH so I've attached the curve here. This is really more for the OP and others who have asked 'why not' than it is to illustrate the point WRT #24. The red curve shows the charge on 1 mmol of phospho. Phosphoric acid, monbasic phosphate ion, dibasic phosphate ion and tribasic phosphate ion are all components of 'phospho' or, put another way, the number of millimoles of phospho in a mash is the sum of the number of millimoles of phosphorous in those species. The relative occurance of each of those species is a function of the pH. If you add some phosphoric acid to water and then start dropping in a strong base (such as NaOH) the OH- ions from that will absorb protons from the acid and water will form:
H3(PO4) + OH- ---> H2(PO4)- + H2O
If you had 1 mmol of phosphoric acid originally and added 1/2 mmol of NaOH its 1/2 mEq of OH would take the protons from half the phosphoric acid and form the monobasic phosphate ion which has a single charge. The phosphoric acid molecule, H3(PO4), has no charge whereas the monobasic ion has the single charge and thus, the effect of adding 1/2 mEq of strong base to 1 mmol of phosphoric acid is to decrease the charge on the 1 mmol of phospho from 0 to -1/2. Looking at the graph we see that charge of -1/2 corresponds to pH 2.12. This is the first pK of phosphoric acid. So having arrived at pH 2.12 we have a system which contains a half millimole of phosphoric acid and a half millimole of monobasic phosphate ion. Lets add another 0.2 mmol strong base (total 0.7). More phosphoric converts to the monobasic ion. Reading from the curve, -0.7 total charge means the pH shifts to about 2.5. Now add another 0.2 mEq of the base (total 0.9). That must shift the total charge to -0.9. There is no where else for the negative charge on the hydroxyl ions to go. For -0.9 mEq phospho charge the pH is about 3. Now another 0.2 for total -1.1. The corresponding pH is 6.3!
For the first 0.2 mol increment (from pH 2.12) of base the pH shifted about 0.38. For the second it shifted about 0.5. For the third it shifted 3.3. Clearly the system is much more sensitive to base additions (it works the same way if we were to add strong acid) at pH 3 than it is at pH 2. By sensitive we mean that a small acid or base addition brings about a large change in pH. Where the curve is steep it means that a small pH shift requires a large acid or base addition. Where it is shallow it means that a small acid or base addition produces a large pH shift. As the phosphate system is flat over the range of mash pH it is clear that it is a poor buffer choice.
Someone asked why adding 5.2 to water at high pH would lower the pH. The blue curve shows the charge on 1 mmol of carbo (carbonic plus bicarbonate plus carbonate). Suppose we had water with 1 mmol carbo (about 50 ppm as CaCO3 alkalinity) at pH 8. The charge on the carbo system in 1 L of this water would be -1. Now suppose we add 1 mmol of 5.2 which is a mix of mono and dibasic ions set for pH 5.9. In distilled water its charge would be -1.05. Let us see what would happen if the alkalinity of the water dominated and the pH went to 8. The carbo would neither lose nor gain charge but the phospho would lose (negative charge increase) about 0.81 mEq. Conversely, if the 'buffer' were strong enough to pull the pH of the combined system to 5.9 it would neither gain nor lose charge but the carbo system would have to gain about 0.75 mEq. So clearly the pH will fall somewhere between 5.9 and 8. Examination of the chart shows that at pH 6.8 the carbo system will have gained charge of about 0.24 while the phospho system will have lost about 0.24. This, then, is where the pH will settle. Obviously, the final pH will depend on the relative amounts of phospho and carbo. At pH 6.4 it seems the carbo charge gain would be about 5 times the phospho loss so if we augment phospho by a factor of 5 we would expect the pH of this mix to be around 6.4.
It is exactly this line of reasoning that allows us to predict mash pH from knowledge of water, malt and salt/acid/base/buffer additions titration curves. It should also help illustrate why phosphate salts are not good candidates for mash pH adjustment (be they in the form of 'pH adjustors' or raw salts) but that phosphoric acid is.