ON SNEEZING FLIES

The energy gain of nuclear power, as reported in our survey last month, is not particularly good. The energy made available at the terminals of the consumer is only 3.6 times larger than the energy invested in the production chain from uranium ore mining to electric power transmission; which is better than for oil burning power systems but worse than for gas, and some 4 times worse than for coal.

However, unlike the case of fossil fuels, the energy gain of nuclear power has the potential for substantial improvement. There is little hope that the techniques of coal, oil and gas production can be made substantially more efficient. For uranium. the losses are small in extraction and milling and practically zero tor transportation; the major loss occurs in processing. i.e., in separating the fissile uranium atoms from the non-fissile ones. ( "Fissile" means almost the same as "fissionable". )

In nature, uranium occurs as a mixture of isotopes. The vast majority (99.3%) is U 238, and though it can be bred into plutonium or fissioned by fast neutrons, it is useless for present reactors. The fissile isotope is U 235, of which natural uranium contains only a trace- 0.7%. To make the uranium usable in a power reactor, the fraction of the fissile U 235 has to be increased. This is called ''enrichment'' because it enriches the usable fraction from 0.7% to about 3%. (3%, by the way, is way below the enrichment level needed tor nuclear bombs, so that Nader's puerile horror stories about nuclear power plants being potential A-bombs end right here.)

Now isotopes of the same element. by definition, have different atomic weights, but exactly the same chemical properties. Which means that chemical reactions cannot be used to accomplish the separation, as is done in most other processing of raw materials. The separation must be performed by physical properties, i.e., utilizing only the different weight (mass) of the two isotopes. That would be difficult enough if the two masses had a ratio of 2: 1, but in fact the ratio is only 238: 235 = 1.0128. Enrichment, therefore, is one of the hairiest problems around.

All enrichment processes work with uranium in the form of a gas (uranium hexafluoride), and so far only two are practically feasible: diffusion and the centrifugal process.

In the diffusion process, the gas diffuses through a porous barrier. The molecules containing the lighter U 235 atoms make it through the barrier faster, resulting in very slightly enriched gas on the other side of the barrier. But hold on: The thing does not work like a sieve that will separate boulders from sand in every throw of the shovel. For one thing, the efficiency depends not on the ratio of the molecule masses, but on its square root, and that is a number even closer to one: 1.0043. Besides, the process is far less discriminatory than the action of a sieve, so that enrichment in a single stage is very, very slight. Thousands of stages are required to achieve substantial enrichment, and in each stage the gas has to be compressed and then cooled to remove the heat of compression. The sorry net result is this: The input energy for the diffusion process amounts to 5.2% of the equivalent energy contained in the U 235 atoms entering the plant; and of these, only 62.2% are recovered, the rest remains on the wrong side of the barrier and is lost.

That is the overriding reason why, at present, nuclear power has such a comparatively low energy gain in the overall production chain.

The centrifugal process works on exactly the same principle as a centrifuge separating cream from milk. The trouble is that cream is a lot heavier (denser) than milk, whereas the weight ratio of the two uranium isotopes differs from one only by a fly's sneeze. However, since the ratio, not its square root, is involved, the sneeze is some 3 times larger than the sneeze associated with diffusion, and that -Gesundheit! - is nothing to sneeze at.

The reason why the US is, at present, using only the diffusion process is largely historical. Both methods were investigated in the Manhattan Project in WW II, but the centrifugal process was abandoned because of the engineering problems of operating high-speed rotors (you need enormous speeds to separate flies' sneezes), the low capacity of individual centrifuges, and the large power input to overcome friction.

But in the three decades since then, the engineering problems have diminished, and the centrifugal process now definitely promises to be the more efficient of the two. The main trouble, as usual, is investment cost, the most conspicuous difference from a milk centrifuge.

But US enrichment capacity is slowly running out, and the AEC has therefore invited proposals for a centrifugal demonstration facility to be submitted by April 1975. Notifications of such propoosals have now been received from three oroganizations: the Exxon nuclear Co., the

Garrett Corp. and a joint venture by Electro-Nucleonics with Atlantic Richfield. The Tennessee Valley Authority is now considering such a proposal.

Though the centrifugal process promises to be more efficient than diffusion, they both work with flies' sneezes. Is there any hope of a more potent principle being exploited? Yes; laser separation. which took a small, but important step forward last summer. But, darn it, this is the fifth issue from which we had to omit the story for lack of space. Perhaps next month.