There are various techniques which can be used to produce a pattern with pentagonal symmetry.

One is to use tiles with special rules (or jigsaw puzzle-like keyed sides) for joining. This produces a Penrose tiling.

It is also possible to make a fractal tesellation, with different sizes of pentagons, down to the infinitely small, which consists of nothing but pentagons. This is illustrated by the faces of Dodecafoam, a three-dimensional fractal tesselation of dodecahedrons.

A less impressive technique, but still capable of drawing inspiration from the other two, is called substitution tiling.

Here is part of my first attempt at such a tiling:

In trying to make a tiling which is mostly composed of pentagons, the first step, with this technique, is to try to build larger and larger pentagons which are mostly composed of the small pentagon one starts with. This can be done fairly well, as this diagram shows:

On the smallest scale, one has one pentagon.

Next comes five pentagons, surrounding a sixth one in the centre. There are small wedge-like gaps on the edges.

If one takes that larger pentagon, and forms a yet larger one by substituting the larger pentagon into itself, replacing its component pentagons by its original shape, one gets the next larger pentagon. Some of the wedge-like gaps have now paired to form an additional tile, a small diamond, that we will need in addition to a pentagon.

Repeating this again, one gets the next larger pentagon. In addition to wedges, this has two other component shapes: a star, and an incomplete star which I'll call a "hat". Those two shapes could be built up from five and three wedges respectively, but I thought having them as additional tile types would look nicer, even if it made the tiling more complex.

Only one more basic tile type remains to be seen.

So far, we have repeated the process of expanding the pentagon only a few times. In addition to replacing the pentagon by a bigger version of itself, in the last iteration, we replaced the small diamonds by a bigger version of themselves, composed of a hat, a pentagon, and a star.

As we repeat the process further, we will need even larger versions of the diamond, and this illustration shows some of them:

Note that the fifth basic tile shape makes its appearance in this diagram. It is a larger pentagon, with each side exactly twice as long as the sides of the original pentagon.

To be able to continue this process indefinitely, we need larger versions of the double-size pentagon, the star, and the hat, as well as the diamond and the pentagon.

This illustration shows these: note that for these shapes,
unlike the diamond and the pentagon, there are no equivalents
after only one enlargement step. But there is an equivalent
for each one after *two* enlargement steps, and
because of the way in which the internal components of the
larger equivalents of the diamond are rearranged, one simply
switches to using a different version of the diamond as the
base when enlarging one step, instead of attempting to replace
all the diamond's internal components by enlarged versions of
themselves in one step instead of two.

Now, here is the full tesselation. It shows to what distance one must go to encounter an enlarged double-size pentagon.

Here is the same tesselation, with one difference. Although the diamond piece cannot be avoided, the "hat", with its lack of symmetry, was a piece I disliked. So, after producing the tesselation, I changed and adjusted some pieces so that every "hat" could be removed, by using decagons, which do have fivefold symmetry in the tesselation.

Finally, an incomplete diagram on a larger scale, with only the once-enlarged diamonds filled in.

This illustrates the fractal nature of this tesselation, although it is fractal only upwards, not downwards as shown here, unlike Dodecafoam, which is a true fractal. Also, it illustrates that the larger pentagons only create diamonds; thus, we do have a choice which order of diamond to use, and so the lack of once-enlarged hats, stars, and double pentagons is indeed not a problem, as claimed.

Before Kepler devised pentagonal tilings that had true pentagonal symmetry, designs of a simpler nature using pentagons were devised by Goethe. These could be put in an overall pattern having pentagonal symmetry, like the following:

but as you can see, the parts that are repeated are not pentagons: rather, a rectangular cell is being repeated, although in five directions in this case.