Planning ahead
Today is National Donut Day. Tomorrow is National Trails Day. So be sure to plan your hike before you eat yourself into a sugar coma.
What is thin film silicon, anyway?
In a conversation with a friend of mine, it became clear that terms like “thin film” are often used in ways that may seem contradictory or downright wrong, especially to people from outside the solar and integrated circuit industries. So this post is an attempt to clarify the terminology. As such, I ask the indulgence of those readers who are experts: this is not meant to be a definitive technical review.
To make electronics grade silicon, you start by reducing quartz sand to get (impure) silicon metal, then react it with HCl to make tricholorosilane gas (TCS). Because it’s a gas, you can purify TCS by distillation and all the other usual gas purification methods. Then you decompose the gas at high temperature to get high purity polysilicon. This is a “thin film” process, in that it produces conformal coatings, but it’s optimized to make bulk polysilicon rods, not thin coatings. This kind of bulk deposition is not at all what people have in mind when they talk about thin film polysilicon. Instead, the polysilicon rods are broken into chunks and used as feedstock for ingot furnaces. There are two basic approaches relevant to integrated circuit and solar cell manufacturing. Both produce “bulk” silicon wafers:
- Make Czochralski ingots and slice them, producing single crystal Si wafers for use in either integrated circuits or top-quality solar cells.
- Melt silicon in a crucible, cool, and slice into wafers. This process, with less control over cooling conditions than Czolchralski growth, produces multi-crystalline Si. The exact crystallinity depends on the cooling conditions. These wafers are used for most bulk silicon solar cells.
Quartz sand is also the feedstock for silane gas production, except that you pull the finished product out before the polysilicon re-deposition step. Silane gas prices tend to move in parallel with wafer prices because both materials come from the same place. Silane gas, in turn, is used to deposit thin layers of silicon onto a target substrate, usually glass, with one of three outcomes:
- High temperature CVD onto temperature-tolerant glass gives thin film polysilicon. Crystallinity depends on growth conditions. This is the approach usually used for displays and other performance-sensitive thin film transistor (TFT) applications.
- Or you can deposit CVD silicon at lower temperatures, then anneal with a laser to get the crystallinity you want. This process also produces thin film or laser-annealed poly-Si, suitable for display and transistor applications. It allows you to use less expensive glass, but the annealing process adds some of that cost back.
- Chemical vapor deposition of silicon at still lower temperatures gives thin film amorphous silicon. This is the least expensive of the three, allowing use of relatively inexpensive glass, but also produces the lowest silicon quality. This is the primary method used to make thin film silicon solar cells.
Among these three methods, there’s a cost vs. performance tradeoff. More expensive methods deliver better crystal quality, and in particular better device mobility: the transistors switch faster and carriers encounter less resistance. Paying the cost premium makes sense for displays.
For solar cells, the tradeoff is a little different. Even thin film polysilicon cells aren’t as efficient as bulk silicon cells, so the added glass cost doesn’t bring enough benefit.
“Thin film silicon,” could thus mean either polysilicon on glass for displays, or amorphous silicon on glass (or sometimes metal) for solar cells. Either usage is correct, so it’s important to be sure which one is really being discussed.
Cats and science. How could I resist?
‘A Cool Cat’, photo by John Zimmerman from Giant Molecules by Herman F. Mark (Life Science Library 1966/1968) (my book), Time-Life’s guide to the exciting new world of industrial plastics & polymers. ‘Perched atop a 4,000° F. flame, this kitten is protected from the heat by a slab of silicone called RTV 615, a transparent rubber that resists intense heat by an unusual process called ablation—a slow, layer-by-layer decomposition.’
(via freshphotons)
Tight capacity creates opening for Intel
If you depend on a foundry, and that foundry’s capacity at the technology you need is limited, then you’re sort of stuck. Some analysts think Intel is in a great position to take advantage of a capacity crunch at Samsung and TSMC. Partly because their own chips are unaffected while competitors might face shortages, and partly because potential foundry customers may seek them out.
Fiscal austerity? What’s in it for Greece?
The slow motion trainwreck that is the European financial crisis continues, with an election in Greece that largely repudiated the terms of that country’s bailout by the rest of the Eurozone. The election has inspired stern reminders that Greece has no choice but to proceed with further austerity measures.
Really? Greece is currently in the throes of a five-year recession, with unemployment in excess of 20% and no improvement in sight. How long can any government, democratic or not, continue to inflict that much misery on its citizens? Wars and revolutions have begun over much less.
Sovereign nations always have a choice. History shows that countries who default on their debts can do just fine in the long term. While certainly a Greek default would be bad for the rest of Europe, it’s not so clear than it would be any worse for Greece than the status quo. If I were a Greek voter, I’d be tempted to take my chances with drachmas over euros, too.
“ The location of actin nucleators are micropatterned onto a circle using deep UV lithography on a glass coverslip (see Reyman et al 2010). Actin polymerization is then induced by applying actin monomers, profilin, and the Arp2/3 complex. A dense and branched meshwork of filaments assemble on the circle (bright yellow) while non-branched filaments grow out of the circle and form parallel bundles. 7% of actin monomers are labelled with Alexa568, which allows the filaments to be imaged with classical epifluorescence microcopy ”
“Wrinkles in a thin film radiating from a drop of water: The thickness and elasticity of the film can be determined from the number and length of the wrinkles.”
The Economics of 3D
Francoise von Trapp offers a very astute analysis of the obstacles to acceptance of 3D integration. It’s not the technology at this point, it’s the cost, particularly on the assembly and test end of things.
Freelance business practices
While I understand how tough it can be to deal with the deadbeat clients represented in the World’s Longest Invoice, I really have to wonder about some of the business practices on display.
I mean really. $49,000? There is no chance that I’ll ever have an unpaid invoice that large. For a project that large, there would be an upfront deposit and three or four intermediate payments. No payment, no further work on the project.
In more than ten years, I’ve had two clients balk at that policy. The first one is the reason why I don’t make exceptions: it took three months, two registered letters, and the threat of a lawsuit to actually get paid. The second time, I declined the project and the potential client was out of business within three months.
Clients should pay their invoices, but freelancers should figure out how to limit their exposure, too.
#getpaidnotplayed
Client: I can’t pay you.
Me: Why?
Client: Because my client hasn’t paid me.
Me: I can see how that could be frustrating.
What are you owed? Bill deadbeats everywhere and join the World’s Largest Invoice to help freelancers #getpaidnotplayed.
