Is it just talking about grain sizes? EDIT: just read the article, yeah they're talking about grain boundaries. Basically when you have that level of size con

Is it just talking about grain sizes? EDIT: just read the article, yeah they're talking about grain boundaries. Basically when you have that level of size control you start to engage the well-known materials science strengthening mechanisms for impeding dislocation slip. EDIT2: excerpt from the Hertzberg fracture book (ch2-sec2): --- > In 1934 Taylor, Orowan, and Polanyi postulated independently the existence of a lattice defect that would allow the cube in Fig. 2.1 to slip at much lower stress levels. > By introducing an extra half plane of atoms into the lattice (Fig. 2.3), they were able to show that atom bond breakage on the slip plane could be restricted to the immediate vicinity of the bottom edge of the half pane (called the dislocation line). > As the dislocation line moves through the crystal, bond breakage across the slip plane occurs consecutively rather than simultaneously as was necessary in the perfect lattice (Fig. 2.1). > The consecutive nature of bond breackage is shown in Fig 2.4 where the extra half plane is shown at different locations during its movement through the crystal. > The end result of the movement of this half plane is the same as shown in Fig 2.1---the upper half of the cube has been translated relative to the bottom half by an amount equal to the distance between equilibrium atomic positions vec(b). > The major difference is the fact that it takes much less energy to break one bond at a time than all the bonds at once. > This concept is analogous to moving a large floor rug across the room. > If you have ever tried to grab the edge of the rug and pull it to a new position, you know that it is nearly impossible to move a rug in this manner. > In this case, the "theoretical shear stress" necessary to move the rug is strongly dependent on the frictional forces between the rug and the floor. > If you persisted in your task you probably discovered that the rug could be moved quite easily in several stages by first creating a series of buckles at the edge of the rug and then propagating them, one at a time, across the rug shuffling your feet behind each buckle. > In this way you were able to move the rug by increments equal to the size of the buckle. Since the only part of the rug to move at any given time was the buckled segment, there was no need to overcome the frictional forces acting on the whole rug. > Since the lattice dislocation is a similar work-saving "device," one may reconcile the large errors between theoretical and experimental yield strengths (Table 2.1) by assuming the presence of dislocations in the crystals that were examined. --- EDIT 3: And if the rug is creased to conform to a ramp or a wall, in order to continue moving the buckle would need to change direction, causing extra resistance. That's basically how grain boundary strengthening works. It forces the dislocations to change direction or even split off into different directions in order to continue, costing extra energy. EDIT 4: Metallurgist have been trying to figure out for decades what is the best way of making grain sizes as small as possible -- through forging, alloying, heat treatment etc. But fundamentally it mainly comes down to the solidification -- if your cooling rate is extremely fast, then you have lots of excess energy for initial nucleation centers, and grains grow out from the sites into each other. The more centers you have, the smaller the grains are. And what determines the cooling rate far more strongly than anything? The bulk volume of the material being solidified. What better way to minimize the solidification volume than to just cool a tiny bit of material at a time rather than the whole volume at once? That's the main idea of the nanoscale 3D printing.