The latter - but that's still a nice jump in the solar cell world.

from ~350-700nm the absorption is enhanced by ~20%. For longer wavelengths, it's much higher (>100% enhancement). They said "the overall EQE is enhanced by 30.8%." on average

In the paper they show the enhancement of quantum efficiency by wavelength. It's pretty uniform (+10-20%) throughout the 350nm (blue) to 700nm (red) range - and is really enhanced at longer wavelengths (>100%, but there isn't as much energy to be harvested here, so not as exciting as it sounds)

Well, I work with semiconductors, and we absolutely use D for degree of electron motion.

Now this is almost off the front page. Hopefully someone reads it and learns something, haha.

I'm going to make this shorter than last time since it's now late (lost by accident) - but here goes:

Silicon has 4 of its 8 outer band electron states filled, ([Ne]+3s2+3p2) which hybridize into 4 sp3 orbinals in a Si crystal. If you start with a single Si atom, it has discrete energy levels away from its nucleus (aka orbitals). If you bring another Si atom (which has exactly the same energy states) closer and closer, eventually these energy states interact and split (i.e. when the covalent bond is formed). The split produces one higher in energy, and one lower in energy. The lower one ends up with both of the electrons (one spin up, one down), and the upper is empty. The lower energy state is shared (covalent bond) between both atoms. Adding more silicon atoms in a similar fashion, similar splits are seen in the new energy levels, but the splitting amount gets smaller each time. When you add thousands (or, in a real crystal, 10^23 !) of atoms, you essentially have tons of very very closely spaced (in energy) levels that are filled in the lower energy position, and tons that are empty in the higher energy position. The lower "band" of states is called the valence band in semiconductor terminology, and the upper band called the conduction band. The "band" is just tons and tons of very similar energy states. At 0K in a crystal, the valence band is full, and the conduction band is empty.

Depending on the material and the way these energy bands look, there can exist a forbidden energy gap between the two bands (valence and conduction) where no energy states exist. If you input energy into the system (heat, light, etc), you can excite the electrons into the higher energy conduction band. When this happens, the electron is only very weakly bound to the atom that it came from, because the underlying valence electrons screen the positively charged nucleus. With only a small amount of additional thermal energy (plenty at room temperature) - the electron is free to float around freely in the conduction band between other atoms. It is thus not bound to one atom. This free electron will fall to the lowest available conduction band energy state - and this energy state is shared amongst the atoms. It does not belong to one atom. The state exists because there are so many atoms. Because it's shared, it is considered "non localized" - it is everywhere at once (to a limit) -- it is behaving as a wave. (you can read more by searching for wave-particle duality).

In an insulator (glass, rubber, etc) the gap between valence and conduction band is very large -- too large for there to be any electrons in the CB at room temperature. There are thus no "free" electrons to contribute to current. In a metal, there is zero gap, and any tiny bit of thermal energy can excite electrons into the conduction band and you end up with a sea of free electrons - giving excellent electrical current abilities. Semiconductors like Silicon are somewhat in the middle -- there is a gap, but it's not too large. Some electrons will be excited at room temperature. The number of excited electrons can be modified vastly (many orders of magnitude) by doping (adding atoms with more or less available electrons like P or B) or by turning an electric field on or off (how a transistor works).

Holy s***. I just spent an hour typing out a response to you, and I changed my slashdot editing options and it lost my response after saving!. I'll retype it later after my wife uses the computer.

It's 2 dimensional to the electrons in the transistor - not 2 dimensional in material physical dimensions. Electrons in a semiconductor occupy a large area shared between many atoms (non localized).

Mentioned above - but electrons are not localized in a semiconductor. Many atoms share the electron - so if few enough atoms (in z direction) share the electron, the electron is 'confined' to 2 dimensions.

In a semiconductor, electrons are not localized. They exist as a wave -- usually mathematically as a wave packet to compromise between a wave and a particle (it is both) -- this wave can be very easily several nanometers. Additionally, electrons diffuse around a semiconductor (they are not bound to one atom) - and this diffusion length is much much larger than a few nm. When a material is just a couple of nanometers, the electrons cannot (statistically) move vertically, and the material is considered 2D (for electron transport - not for physical dimensions of the materials). Headline is 100% correct.

Electrons are stored in "trap" layers of Silicon between the gate of a transistor and the channel of the transistor. On either side of this trap layer is dielectric that blocks the electrons from leaving.

From my experience in a calibration lab for two different major electronics companies in the past few years, I can wholeheartedly say that Agilent products are generally the best of said brands. (Needing recalibration less often, better interfaces (IMO), less glitches in software, better build / support, etc.). That said, they are often the more expensive brand. At an academic research lab, this factor may take the most consideration depending on your funding sources and reliability. FWIW, we viewed most of the Tektronix equipment as junk and would opt to use the Agilent equipment when available (but, "junk" is a relative term).