Although I am not an expert in the field, I would say: the excess of metal ions (or nodes) is terminating the surface of the MOF crystals.

Any crystal stops growing when the surface is terminated (shielded against further growth). With such an excess you can control the crystal growth and stay in the nano regime.

To make this clearer: in the bulk three-dimensional NaCl crystal, the Na:Cl ratio is 1:1. If you look on the surface, there is either a coverage with Na+ (making the crystal positively charged) or Cl- (making up a negative potential). Since we are in the bulk, this coverage has no impact on the stoichiometry, which is 1:1 in both cases (but there is coverage). However, if you go to the nanoworld the surface gets more important and this has an impact on the stoichiometry.

Most of the MOF preparations involving hydrothermal or solvothermal methods produce serendipitous products. Very few chemists plan their reactions to get desired products and the rest leave it to the serendipity. In your case, you have to decide the course of the reaction by looking into the nature of metal precursor you have chosen and the organic ligands. If your metal reagent is not 100% pure or likely to form something else (say a stable complex or getting reduced) under the reaction conditions, then you have to use metal reagent in excess. So it requires critical analysis and evaluation of the reaction from both kinetic and thermodynamic point of view.

In conventional coordination chemistry, if the ligands are more reactive or known to undergo degradation/decomposition, for example, phosphines, then one should use slight excess of ligand. If the reaction is 1:1, then 1:1.1 or 1.2 should be considered. In your case, please look into these things.

Best

In general:

MOFs are "bulk" materials, usually contain more than one ion-pairs and not "isolated" ions/ion pairs unlike in a real solution. That means, on the nanometer scale the particles considerably fewer ions on the surface than the total number. And so, the ligand exchange is restricted to the surface. Additionally, not all the surface ligands can be exchanged.

An example (because I am not an expert of Cu3HHTP2, but I assume that it is general): rutile builds up tetragonal unit cells, where the unit cell dimensions are a = b = 0.458 nm, and c = 0.295 nm. The Ti2+ ions are surrounded by 6 oxygen atoms, forming an octahedron, and the unit cell contains more than one TiO2. Not to say anything about the intercalated other ions.

That means, > 1 nm particle sizes contain fewer replaceable, in the case of rutile the unit cell contains (as average, as Axel wrote) 5 TiO2 while on the surface (of the unit cell) only 8 (of 10) Oxygen is active. In theory on a ~1x1x1 nm particle, while the number of constituting TiO2 is 30, the number of surface O is only 20 (of 60). And usually, the nanoparticles are larger than 1 nm.

Of course, in (theoretical) two-dimensional cases the numbers are different.

Additionally, multidentate ligands further reduce potential substitution sites. So, the mentioned 2:1

Finally, Balakrishna's comment is an inappropriately undervalued comment.

Regarding your example.

There are many missing data, and I assume they are also hidden for you. For the solid synthesis, you mentioned 2:1 ratio. In solid syntheses, there is a size limit, which is generally higher than 1 nm. After that a) I have doubts and the 2D structure might be true only; b) using a 2:1 molar ratio seems to be too low.

Due to solubility issues in solution and lack of yield, isolation method, etc., the 10:1 can be reasonable.

Please, also note that thin films are not necessarily 2D structures, and the reactions are rarely so ideal as they are described in publications - as you are getting more experienced you will see the difference between a publication and real life. Many tricks remain hidden in publications, independently from the journal rank.

So, firstly try to reproduce the published method (=do strictly what is written even if you are not in agreement with the authors!), then, based on your experiments and comparison with the written matters, try to find explanations and understand the authors' synthetic method. This way usually works well. Thinking a lot of unknown things, then to conclude finally to repeat the published method is many times a waste of time.

Good luck,

Laszlo