Column-like Films of Silicon for Battery Applications

About a year-ago I relayed the story of Chan's work on using silicon nanowires as a potential anode material for Lithium-ion batteries. Silicon can store some ten-times more charge than Carbon, the current industry standard, but this comes at the expense of a huge volume change. The difference in volume between charged and uncharged is 300 % for Si. Just to give you an idea, the alloy of Lithium and Silicon that's formed is Li₁₄Si₄ from metallic Si. The article from Nature Nano suggested that forming the silicon into high aspect-ratio wires would allow the silicon freedom to expand along the long-axis of the wire and hence be less likely to physically break and no longer have a physical, conductive pathway to the anode.

In general, the lifetime of a battery is determined by how great of a volume change it undergoes when cycling from the charged to uncharged state and vice versa. Volume changes imply stress and the gradual introduction of defects that can trap electrons and reduce electrical conductivity. For LiFePO₄ cathodes, the volume change is around 4-7 %, but this is a crystalline material. Silicon nanowires are amorphous (i.e. poorly ordered) and the introduction of defects on cycling is not necessarily an issue.

The previous work was truly proof of principal, but unlikely for variety of reasons to be a direct path to commercialization. There's some new work out by a local group that expands on the work of Chen. Fleischeur et al. tried their specialty, glancing angle vapour deposition, to form a thin-film of Silicon composed of many, regular pillars. (Disclosure: our research group collaborates with the group that did this research. I personally do not, however.) In glancing angle deposition, the substrate (onto which the film is deposited) is at nearly right angles to the incoming vapour stream. In thin film deposition, one tends to see small clusters form first due to surface tension. As the clusters grow, they amalgamate together and form a (porous) solid thin film. When the substrate is at high angles of incidence, the first clusters to form shadow any smaller trees and grab more than their fair share of the incoming mass stream. Hence glancing-angle deposition typically forms column-like thin films.

The glancing-angle fabrication method has a number of potential advantages over Chan's technique:

1. Chan's thin film process relied on a gold catalyst ($$$), whereas the GLAD process only requires a thin layer of chromium for adhesion on Si substrate and none at all on a stainless steel substrate.
   
2. Glancing-angle deposition can easily control the spacing of pillars by patterning the substrate.
3. The glancing-angle films were "robust" when I asked the author about it. He said hitting the batteries with a hammer had no effect on performance, so presumably the pillars were not breaking.
   
4. Glancing-angle deposition requires a microscopically smooth surface for proper column formation.

Let's look at some results. The question is, how durable are these anodes compared to graphite? The charge curve is really all we are interested in.

Figure 1: Chan et al. charge capacity after 10 total cycles.
Figure 2: Fleischauer et al. charge capacity after cycling up to cycle 70. I don't recall the reason for the discontinuities but I vaguely recall it had something to do with the test electronics.

Both authors show a very large drop in charge capacity after the first recharge. This means there is some sort of irreversible change to the material occuring from film fabrication to charged and uncharged Si. Then there is a progressive loss in capacity. Evidently Chan and company are less confident in their material as they are only showing results up to ten cycles. Average capacity fade for Fleischeur's battery was found to be 0.3 % per cycle. If we extrapolate, that would imply it would take approximately 750 cycles for the charge capacity of the silicon anode to drop below that of a conventional graphite one. Obviously, that's not good enough for commercial applications.

Overall, I think that this is an important step in terms of fabrication and longevity. We are still looking at a minimum of a decade before any such silicon Li-ion batteries hit the shelves; this is progress on that path.