VoxSolaris: The Voice of the Sun
Solar voltaics

General Principle of the solar voltaic cell

Direct production of electricity from sunlight depends on the photoelectric effect, discovered by Heinrich Rudolf Hertz in 1887 and later explained by Einstein. When a metal surface is exposed to photons (electromagnetic radiation) above a certain threshold frequency, the photons are absorbed and electrons are emitted from the surface. The energy or velocity of the emitted electrons (the voltage), depends on the frequency of the radiation. The number of emitted electrons (the current) depends on the intensity of the sunlight.

No electrons are emitted with radiation below the threshold frequency because the electrons are imparted insufficient energy by the photons to overcome an electrostatic barrier that exists at the surface. The barrier is caused in the case of a metal, by the termination of the crystalline structure. The threshold which is termed the 'work function' is expressed as a voltage and is specific to the surface of the material. Different materials have different work functions and as energy spent overcoming the work function arises in the material as useless heat, materials with low work functions are favored.

The photoelectric effect is not a great deal of use on its own as the emitted electrons either fall back into the material or form a 'space charge' above the surface. To extract a useful electric current we need something to act as a sink for the electrons. This is called the anode while the surface emiting the electrons is the cathode. This arrangement is termed a diode and for it to work, the properties of the anode and cathode must differ. Otherwise both would chuck electrons at each other and the net effect would be all the energy wasted as heat. All solar voltaic cells then, are essentially diodes.

Solar Economics

The most established solar voltaic technology uses crystalline silicon. Silicon is a semi-conductor which can be 'doped' with impurities to make it either more or less inclined to emit or receive electrons. It is therefore a very good material for making diodes out of. Comercially available silicon crystal solar panels convert between 10 and 12 percent of the sunlight falling on them into electricity. This sounds poor but as sunlight reaching the Earth's surface averages 1000 watts per square meter during the day, we can get 100 watts per square meter. Not bad for a device that has no moving parts, makes no noise and largely sits on the roof where it wastes no space. Silicon crystal is quite 'radiation hard' and so deterioration from exposure to sunlight is slow. Solar panels based on this material are widely accepted as having useful lifespans in excess of 20 years.

Silicon is very abundant but alas, does not occur in elemental form and is very difficult to extract from compounds to a high purity, making it expensive. The techniques for this have been honed considerably and we now see solar panel bulk prices at $1 per peak watt or less. A solar panel's peak watt rating is the output under idea conditions which are rarely if ever achieved. We have daylight for an average off 12 hours per day, some 4,380 hours per year. In even in a good location one could expect to get 1,800 hours worth of peak wattage for the year. Poorer locations have typical values closer to 1,000 hours. At this level each peak watt of panel generates 1kWhr per year. And since panels have lifespans of the order of 20 years one can translate the capital cost into a cost per kWhr as the annual payments on a 20 year loan divided by the number of kWhr's produced by the installation in a year. Assuming these payments to be $5.25 per month per $1,000 (with a good credit rating), that is $63 per 1,000kWhr so 6.3 cents per kWhr. In better climates where one might get 1,500 peak hours per year, that would drop to 4.2 cents. Few grids charge as little as 6.3 cents let alone 4.2 cents.

But solar panels produce electricity only when the sun feels like shining while our usage is often when the sun is nowhere in sight. Storing power for later use in batteries increases the cost of actual usable power in two ways; the battery has a capital and sometimes a maintenance cost - and the battery does not give back all the power that is put in. The latter is easier to calculate. If the battery gives back 80 percent of what you put in then that means the power you get from the battery is costing you 25 percent more. If you were looking at 1,000 peak hours and 6.3 cents you are now looking at 8 cents. But the capital costs are the worse part. Lead acid might cost as little as $150 per kWhr to buy but you will be doing well to get 500 full charge cycles out of it. That's a massive 30 cents a kWhr. If you go for the 10kWhr Tesla Powerwall at $350 per kWhr you'll get a lot more than 1000 full charge cycles but estimates of the storage costs are still around the 20 cents mark. These figures are lousy and make a mockery of the costs attributed to the panels. The only saving grace is the power you can make use of without storing. Selling power back to the grid is the primary method for this although not the only one. If you need air conditioning for example, there will be some correlation between the sun deciding to shine and you deciding you are a bit too warm. And at that point you are paying just the 6.3 cents.

Solar in the Future

All the above said, the days when solar panels on roofs owed their existence to government subsidies or die-hard environmental enthusiasts not deterred by the price, are essentially over. Solar pretty much stands on its own. And one of the best things that has come of the fall in price of solar are the millions of homes in parts of developing countries nowhere near a grid who are now with power and with Internet into the bargain. With the price as it now is, there will become slowly more and more of it. Good but alas not good enough. We need a lot more solar and for that to happen, solar has to do more then match the grid, it has to beat the grid hands down. So what are the odds of solar becoming cheaper still? And more importantly, what is the price/performance outlook on the battery front?

Silicon crystal based solar panels are not quite at the point where they cheap enough to be used as ordinary non-solar panels in partition walls or similar. Another 50 percent reduction and they would more or less be there. We are doubtful that they will ever get much cheaper than 30 cents per peak watt meaning the raw cost of the electricity they generate would be about 2 cents per kWhr. Lithium ion will fall to $150 cutting storage costs to 10 cents per kWhr so home solar systems will become very competitive with current grid prices.

Other battery types, particularly the Sodium Sulfur Flow Battery will in our view come to dominate the grid storage offering storage costs of the order of 2 cents per kWhr. Small scale community grids will slowly profligate, particularly in areas with no or poor grid coverage. The routine outages seen in many parts of the world will become a thing of the past. Electricity companies will make a lot less money from selling grid electricity to homes but before you rush to eject their shares from your pension portfolio you might want to consider what else they will be doing.

Electricity companies sell most of their electricity to large corporate buyers and that market will increase massively. The thing to consider is where. As the the Icelandic model has shown, you can export electricity by hosting energy hungry industries in situ. And in some countries you get 1,800 peak hours per year, in rare cases 2,000. In those countries you could see wholesale prices at 3 cents per kWhr, unheard of today even in Iceland. It won't just be aluminum smelters jumping in. It will be a vast range of chemical and raw material companies as well as car manufacturers etc. Big companies don't just try to lower their tax bill. They try to lower any costs, particularly the big items like energy bills. The UK government might want to contemplate what solar will be doing in 10 years time before they commit to Hinkley Point and the agreement to buy all the electricity it produces for the first 35 years of operation at 9.25 pence (about 14 cents) per kWhr.