Wikipedia:Reference desk/Archives/Science/2022 December 17

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

December 17

small electrical generators

I wonder what the obstacles are (i.e. why can't we get them) to making small electrical generators or fuel cells, e.g. for portable outdoor use. It's easy to get a 1 kilowatt generator that weighs 30 pounds, so why not a 100 watt generator that weighs 3 pounds? Things like this are on the internets now and then, but are always just around the corner, never "order now". Thanks. 2601:648:8200:990:0:0:0:497F (talk) 00:43, 17 December 2022 (UTC)

Gasoline and diesel gensets don't scale linearly. For example, I own a Chinese made 2 kW generator and a Chinese made 3.3 kW generator. They look identical, and occupy the same volume, but actually there are 2 key changes - the 3.3 kW unit has a larger cylinder bore, such that it is 215 cc capacity whereas the 2 kW unit is 185 cc. The 3.3 kW unit has a larger frame size generator. The difference in total weight is small - about 10%, due to the difference in generator frame size.

Honda Japan used to make a 300 Watt gasoline powered generator - it weighed about half of what my 2 kW unit weighs.

Basically, to make a gasoline engine half the power of another one, you basically need to halve the cylinder displacement volume. To do that, each length, width, and height dimension needs to be reduced by only the cube root of 1/2, i.e., reduce 20%. Stresses remain about the same (for instance, to run efficiently on standard gasoline, the compression ratio must stay about the same), so mass only comes down 20%. Similarly, to make an engine of 1/10 the power, it will end up reduced by the cube root of 0.1 i.e., dimensionally reduced by 53%.

As far as fuel cells go, they are like nuclear fusion power stations - i.e., they have been reported as just around the corner for almost as long as I have lived, which is 75 years. The trouble with fuel cells is that unless they are constructed with ultra pure materials, and fuel them with ultra pure fuel, they soon clog up with detritus. Dionne Court (talk) 02:47, 17 December 2022 (UTC)

And while combustion-powered generators get lower power-to-mass ratio (and lower efficiency) as you scale them down, this doesn't apply to batteries, which are therefore more attractive at small size. In particular as they don't produce noise and nasty fumes. It gets ever easier to recharge batteries as mains power sockets appear everywhere to charge our phones, e-bikes and electric cars, so I don't expect the minimum size of petrol-powered generators to drop. PiusImpavidus (talk) 11:34, 17 December 2022 (UTC)

Fusion is a few decades from now for 70 years cause no one ever throws them enough bones of money to make one big enough. Cube-square law. Sagittarian Milky Way (talk) 15:02, 17 December 2022 (UTC)

Although they've done something in the past few days that's never been done before [1]. 146.199.206.38 (talk) 17:10, 17 December 2022 (UTC)

Yeah - they have now and then been announcing they've done something new ever since the Soviets invented the tokamak in the 1950's, followed by British confirmation of improved Soviet tokamak results in the 1960's. Don't hold your breath for the world being saved from being cooked by emissions from fossil fuel burning. Unless WW3 sends us back to another Dark Age. Dionne Court (talk) 00:52, 18 December 2022 (UTC)

The Sn–Sb system

Are Sn–Sb alloys generally metallic conductors or semiconductors? (With no other metals added. I know for example that type metal was historically Pb–Sn–Sb, but that's not what I'm asking about.) Double sharp (talk) 13:51, 17 December 2022 (UTC)

Sn (tin) with 5% Sb (antimony) is used as a lead-free solder, so would need to have metallic conduction. Tin has a valency (number of outer shell electrons) of 4 and antimony has a valency of 5. For a binary alloy to be a semiconductor, the "median" valency would need to be 4, as in combining a valence 5 element with a valence 3 element, as with gallium (valence 3) and arsenic (valence 5).

There is a simple formula for calculating the resistivity of binary metallic alloys from the resistivity of each element and the ratio present, but I can't think of it right now. Maybe another poster will know. Dionne Court (talk) 15:11, 17 December 2022 (UTC)

Thanks! I wasn't sure because germanium arsenides (one row further up the periodic table) are semiconductors (10.1002/pssb.201552598), despite the wrong "median" valency. True, tin at ambient conditions is a metal (whereas germanium isn't), but grey tin becomes the stable state not too far below room temperature, so I thought it was better to ask. :) Double sharp (talk) 03:56, 18 December 2022 (UTC)

I'm not going to pay $12 to access 10.1002/pssb.201552598, nor travel to my university to access it for free, but I think you'll find that germanium arsenide is only a semiconductor at low fractions of arsenic, so that germanium's valency of 4 predominates. As the concentration of arsenic approaches 100%, it cannot be a semiconductor. Dionne Court (talk) 14:49, 18 December 2022 (UTC)

That paper calculates that GeAs and GeAs₂ are semiconductors and says that agrees with experiment, so it seems possible at quite high As fractions. But it seems to be a close thing, since they calculate that with a few Ge vacancies these can metallise (OTOH, with even more they can become semiconductors again). Anyway, pure As has both metallic (grey) and semiconducting (black) allotropes. Double sharp (talk) 16:19, 18 December 2022 (UTC)

Well, you can always ask Alexandra. Tigraan^{Click here for my talk page ("private" contact)} 18:39, 21 December 2022 (UTC)

Missing centers of the United States

We have mean center of United States population, median center of United States population, mean center of United States land and mean center of contiguous United States land, in all cases counting the land or people of the 50 or 48 states plus DC as a whole (including islands). Where are the missing centers? (the mean and median center of the contiguous population and the median center of the land and land minus Alaska and Hawaii, same definitions as the other 4). Are the 51 median land centers of the states+DC online anywhere? Or the 51 mean population centers and 51 median population centers? We have the 51 mean land centers. Sagittarian Milky Way (talk) 16:54, 17 December 2022 (UTC)

Note that there are different, non-equivalent definitions of "median centre".^[2] The cited source identifies the one used by the United States Census Bureau as following the "British tradition". It has the unappetizing property of being sensitive to isometric transformations and thereby ambiguous.  --Lambiam 17:27, 17 December 2022 (UTC)

Apparently mean is minimal sum of distances squared and geometric (other) median is minimal sum of distances. While the British tradition is sensitive to rotation and translation (i.e. the British median of an equilateral triangle of 3 people moves as it rotates, the middlemost people of an even number of ppl probably has an ambiguous gap in between) it also has the awesome property of "half on each side" in both latitude and longitude and in practice land is infinitely divisible and for large lands of millions population almost is: any gap for the whole country would be feet at most and smaller than any realistic error for a Census where the sum of double counts and missing people is c. millions. I've found some things like a map of the state population medians and a pause-able animation of cumulative county populations by latitude and longitude which allows figuring out the 2018 contiguous population median to county-level accuracy. Sagittarian Milky Way (talk) 22:37, 17 December 2022 (UTC)

So after trying many times to get it to successfully fall on exact right county out of 3,143 it appears the centers of median population are almost too close to tell signifying lots of persons near these meridians and parallels (i.e. Chicago metro area) and the median contiguous land is circa 2 counties south of the mean contiguous land and ~1-2 counties east. So maybe the north and west halves have better leverage or they're using a map projection where parallels sag in the middle. Sagittarian Milky Way (talk) 07:35, 18 December 2022 (UTC)

Patagopelta, Kelumapusaura, Notoceratops, Labocania, Alamosaurus...

I am really wanting to know what is the deal with this faunal interchange between dinosaurs of North America and South America. CuddleKing1993 (talk) 19:06, 17 December 2022 (UTC)

What you're interested in is biogeography and according to this paper there was probably a link between North and South America during the Campanian (Late?), via a line of volcanic islands, allowing "a dispersal route for hadrosaurids and other vertebrates". Mikenorton (talk) 22:56, 17 December 2022 (UTC)