1. RTGs are desperately inefficient. Figures vary, but I seem to recall coming across 1% efficiency as fairly typical. The advantage of RTGs is extreme reliability and long life - e.g. a Strontium-90 source (readily available in huge quantities as a by-product of nuclear power or weapons production) has a half life of around 30 years, fairly respectable heat production, and limited gamma/X-ray emission. As a result, such devices have been widely used (particularly in Russia & the old Soviet Union, where cost and accessibility were issues, and safety less so) in military and special industrial applications (e.g. powering sensors at remote military bases, powering scientific equipment on marine buoys, etc.
For space use, higher performance isotopes tend to be used - Plutonium 238 is the typical isotope, as it has very high spontaneous heat production and as it's an alpha emitter requires less shielding than beta sources like Sr90. Pu238 is difficult to obtain and rare - I think it normally has to be made by irradiating other artificial elements in a reactor under controlled circumstances - it's not obtainable in useful quantities as a by product of nuclear power or weapons.
Aside from the fact that the thermoelectric effect is pretty inefficient - there are fundamental limits to efficiency of conversion of thermal to non-thermal energy. The Carnot limit is the application of the 2nd law of thermodynamics, which relates the maximum efficiency at which energy can be harnessed to the temperatures of the 'hot' and 'cold' side of a device. The larger the temperature gradient, the better the efficiency. For example, a device which has a heat input at 100C, and a heatsink temperature of 40C, would have a theoretical perfect efficiency of 19%. In practice, this level cannot be reached. This is why power plant designers go to such enormous efforts to get the temperatures up as high as possible (and is why nuke plant efficiency is in the high 20s, low 30s per cent, whereas nat gas is high 50s - you can't let the reactor get as hot as a gas turbine).
Of course, if your heat source is plentiful enough, then it may be worth building a device to make use of that gradient. There's a power plant technology called OTEC (ocean thermal energy conversion) - this makes use of the fact that in the tropics, deep ocean water is much colder than the surface. Even though it's only 20 degrees or so, and the Carnot efficiency is only a few per cent, the amount of heat that you can extract through your plant is large enough to make this viable.
2. Long half life elements were present when the earth was formed (e.g. uranium and throium). Uranium-238 has a half life of 4.5 billiion years, U235, about 1.5 billion, and thorium about 10 billion. Resources of uranium were significantly richer in the young earth. In fact, the isotopic properties were richer - with more of the fissile U235. In fact, about 1.5 billion years ago, natural uranium then would be what we currently regard as 'slightly enriched uranium'. While current natural uranium won't support a nuclear reaction without careful engineering, enrichment makes it much easier. In fact, in Gabon, there was a rich uranium ore deposit that did undergo a spontaneou nuclear reaction (at about that time, when it got flooded with water).
Shorter half life isotopes may be part of the decay chain of other elemetns. E.g. Uranium decays via Radium, etc. Radium has a relatively short half life about 1500 years, so you get an equilibrium forming according to the ratio of half lives - so uranium will contain about 0.3 ppm of radium, which can be concentrated. Same story with Radon, which is an intermediate of uranium and thorium decay.
Short half life elements can be made in a variety of ways - fission of uranium, and subsequent purification; decay of a parent isotope; irradiation of an element with heavy particles (e.g. neutrons, protons or deuterium nuclei) in a particle accelerator. The technique used depends upon the existance of a suitable reaction, and the degree of purity needed.
E.g. Technetium 99m is a widely used medical isotope used for many types of body/brain scan. It has a half life of about 6 hours, so is usually made on site. This is done in an elution generator: This is a vial filled with resin beads containing the isotope Molybdenum-99. The Mo99 decays into Tc99m, which is not electrochemcially attracted to the resin, so will leach out if the vial is filled with water. The Mo99 itself is made by purification of fission products of uranium that has been irradiated in a reactor.
Another medical example is Fluorine 18 which is used for PET scans. This is made by irradiating heavy water (containing Oxygen 18) with protons in a particle accelerator (cyclotron). F18 has a half life of under 2 hours, which makes PET scanning logistically difficult and very expensive (as the F18 must be shipped from the accelerator facility to hospitals by the quickest method possible).
3. I'm not sure what that would be detecting. Most pocket radiation detectors detect high energy beta and gamma rays. These rays are emitted from the source material - the material itself doesn't have to escape. Similarly, gamma rays can never be completely blocked, some will always get through - but you can cut them down to such low levels that it doesn't really matter. Beta (and alpha) particles have a defined range - beta can go up to several yards in air, alpha may only manage a few inches. And there is a further problem, if you want to detect these rays - you need to stop them *in* your detector. If your detector is small (key chain sized) it's not going to do a very good job, and will have terrible sensitivity.
One of the things that a lot of people forget is that radiation sources obey the inverse square law - so if you're 10 times as far away, you receive 1% of the radiation. This means that radiation levels fall off very, very fast - if you want a good chance of detecting a radiation source, you need to be close. If you were able to detect radioactive sources on a truck, from a house near a highway - you can bet that it will be doing a good job of cooking the driver.
Indeed, it's not clear what your friend would have been detecting on a truck. Most shipments of industrial radioactive materials are heavily shielded (there are very strict laws on the amount of shielding) and it would be unlikely that such a source would be reliably detectable even standing directly outside the truck. Even high-level nuke waste from power stations is shipped so heavily shielded that you're unlikely to detect it from more than about 30 or 40 feet.
Indeed, the fact that the detector kept alarming for 15 minutes, suggested it detected something in the nearby environment (or it was simply a false positive) rather than anything related to the truck, as the truck would move rapidly out of range.
Facebook
Twitter