Semiconductors are also Semiresistors

By definition, a semiconductor is any material with a mediocre electrical conductivity. In other words, more conductive than glass but say less conductive than copper. So, everything that's not so great at insulating or conducting could be considered a semiconductor. This vagueness is not helpful to understanding Semiconductors. The broad dictionary definition includes half the periodic table.  The most common today are: silicon, germanium, and compounds of gallium, usually in the form of slices of artificially grown crystals. Diodes, LEDs, transistors and integrated circuits are all made from semiconductor material.

This is really important to understand because you are reading this on a device (mobile or otherwise) that contains a number of semiconductors it relies upon to function. But they are not quite so new or complicated. The Nobel prize was bestowed upon Marconi and Karl Braun in 1909 for the discovery of semiconductor rectification. [source] But yes, that was a semiconductor.

In 1874, Braun had discovered that current only passed freely in one direction through lead sulfide (galena). Sadly he didn't find much use for that until radio was invented two decades later. But all this led directly to the invention of the transistor. Greenleaf Pickard got his patent in 1906 on a silicon-based rectifier.
 In 1907, Physicist George W. Pierce demonstrated the rectification properties of carborundum diodes after experimenting with metals and semiconductors. This rectifier required the application of current to function but was more mechanically stable. In 19276 Julius Lilienfeld proposed a metal–semiconductor triode rectifier using copper-sulfide semiconductors. These were actually early semiconductor field effect transistors.there is some debate as to whether any of  his devices worked.. but that's not actually required for a US patent.

Bell labs and other groups were looking for a solid-state equivalent to the vacuum tube. When they found one, it was probably an accident, examining the diode point-contact. The first transistors were of that type. But strangely they didn't invent the point-contact transistor first. William Shockley produced a field-effect theory in 1939 and Bell Labs announced the invention of the junction field-effect transistor in 1951 and began licensing it. What happened to that point contact transistor, that simple semiconductor?

Pickar's point-contact rectifier was awfully close back in 1906. In other more patentable words, it had a lot of prior art. But lacking the vocabulary, crystal oscillation wasn't referred to as a semiconductor. But by 1920 related works with crystals had already been made by Oleg V. Lossev, William H. Eccles, Frank W. Jordan,Eugen Nesper, Julius Lilienfeld, Robert G. Adams, William T. Ditcham and many others. Hugo Gernsback predicted that crystals would someday replace valves in circuits. Lilienfeld actually patented it in 1923. But the semiconductor was "invented" 3 decades later. More here

Semiconductors have low resistance in one direction and high resistance in the other, the inverse of their conductive properties. the same is true of their insulation properties. Braun's rectifier was a diode, the simplest possible semiconductor device. The transistor is just two diodes back to back, and the integrated circuit is made of hundreds of these etched onto a silicon chip.

3 Penny Radio

In science the Penny takes more abuse than any other denomination of American currency. A century ago Herbert Willis famously bombarded pennies with radiation just for giggles... and testing alpha rays. They were the currency of tchotchkes: penny candy, penny post cards, penny arcades and so on. We crush them into oblate shapes in hand cranked machines at truck stops to imprint them with hearts and images of tourist attractions.

The ultimate indignity was in radio, we soldered them. In his autobiography, Steve Wozniak mentions the archetypal use of the penny as the "crystal" in the crystal radio. In that application it's more accurate to say that the cuprous oxide on it's surface is the detector but it's still a penny. It's a use that dates back to at least WWII in early foxhole radios that used American Pennies, German Pfennigs, and old copper French centimes. The 2003 book Sneaky Uses for Everyday Things by Cy Tymony revisits that old design.

But getting back to the foxhole radios, the most common crystal in use by consumers around WWI was probably galena. Metallic detectors were known to exist. Soldiers innovated one that used razor blades which we call the foxhole radio. Iron(III) oxide, aka rust, is a semi conductor. This was discovered by Greenleaf Pickard in 1906. He also patented the  silicon crystal detector that year. It's likely that few, if any men in the trenches were aware of his discovery, more likely that they discovered it independently through trial and error. They already knew that galena could be used... someone probably had the sense to experiment with other metals. It wasn't until 1927 that Lars Olai Grondahl and Paul H. Geiger invented the copper oxide rectifier. More here. they both worked at the Union Switch & Signal Co.  While investigating the corrosion of copper switches they discovered that current flowed more easily one way than the other and made the crucial deduction.

That radio discovery has a modern descendant in the 3-Penny Radio. In this application the Penny is not the detector. The Pennies don't do anything electrically significant in this design. But that's OK. Projects like this are for kids, and half of the adventure is really just soldering practice. It's based on older designs, lacking even an oscillator. But it concedes to kid friendliness with an integrated circuit to handle AGC (automatic gain control) so that when junior hooks up the earphone the results will be more listenable. AM is so much noisier now than it was a century ago AGC is necessary to make the radio experience at all compelling. 

Meteor Scatter

Meteor scatter communications are also known as Meteor burst communication. It's a types of signal bounce possible only under specific circumstances. When a meteor enters the atmosphere of Earth an area of the E-layer is ionized.  This area is roughly cylindrical and like the tail of the meteor, it is long. When it first forms it is able to reflect and scatter radio signals.This ionization is brief, lasting from a few seconds to about a minute. The effect is observable from about 25 MHz up to about 10 GHz. Meteor skip allows a radio signal to bounce upwards of 900 miles.

 The earliest recorded observation of meteor scatter was by Hantaro Nagaoka of Nagasaki, Japan. He was then the chief physicist at the Institute of Physical and Chemical Research in Tokyo.  In 1929 he published a paper on his findings titled "The Possibility of radio Transmissions Being Disturbed by Meteoric Showers."  It sounds dry, but the 1929 Volume 5 issue of the Imperial Japanese Academy of Tokyo became a classic. The NRAO described his breakthrough

Nagaoka was impressed with the coincidence of height (~100 km) between the ionospheric layer and where visual meteors occur. Not only might meteors directly cause quick ionization, but he argued that long-lived fine dust resulting from disintegration of the original meteor particles would act as nuclei for collecting ions or electrons from the surroundings, thus causing a region of less ionization, or a "disturbance"

In 1931, Greenleaf Whittier Pickard made his own observations on the phenomena. He noticed that bursts of long distance propagation occurred during meteor showers. He published his paper "A Note on the Relation of Meteor Showers and Radio Reception" in July 1931.

 At about the same time Albert. Melvin Skellett a researcher at Bell Labs was theorizing a relationship between the kinetic energy of a meteor and the ionization of the ionosphere. Skellett spent 2 years testing his ideas with his colleagues. In November of 1932, during the Leonids meteor shower they tested the meteoric effect. More here. He assumed three crucial and correct details:

1. Meteors as a class could contribute, along with cosmic rays, the bulk of the energy needed to maintain charged particles in the ionosphere.
2. A large meteor could produce as many as 1,000,000 electrons and a critical mass of ions. (enough to effect radio) 
3. The variation in the strength of shortwave radio signals indicates turbulence in the ionosphere caused by the constant mixing action of the meteors.

The first signifigant attempt to use meteor scatter for anything useful was by the Communications Defense Centre in Canadia (CRC)  In the 1950s they launched project, "JANET."  It was code named JANET for Janus, the Roman god of the doorways, beginnings, endings and time. Janus is often depicted with two heads, facing opposite directions. I take that as a metaphor for transmission and reflection. In JANET they sent bursts of data when conditions were ideal trying to catch those short partial minute windows. to detect the window they broadcast a 90 MHz carrier wave and monitored it for sudden changes in strength.  They bounced signals from their radar research station in  Saskatchewan all the way to Toronto, a distance over 600 miles. The system was operational from 1955 until about 1960. The average character error rate was only 1.5 percent!  the Friends of CRC sumarized their data and bandwidth allocation:

"It may seem that the Janet technique was somewhat wasteful of spectrum space because of the relatively high instantaneous signaling rates which are required to achieve modest average information rates. This defect was offset by the fact that Janet systems, using the same frequency assignments, can be located much closer to each other than can other systems operating over comparable distances."

It is generally the case that frequencies between 50 to 80 MHz are ideal for meteor scatter communication. The FM band itself  (88.1 – 107.9 MHz) also works quite well. During meteor showers, laypeople can still perform basic experiments and get recognizable results. All you need is to know when the metoer showers are. Here's what fun to look for in 2011

• Quadrantids - January 3
• Lyrids - April 21
• Eta Aquarids -May 5
• Perseids - August 13
• Draconids - October 8
• Orionids - October 21
• Leonids - November 17
• Geminids - December 13

Perikon detector

Strange as it may seem, the crystal in crystal radios wore out. The rectifier was a contact metal–semiconductor point-contact junction, aka a Schottky diode. I mentioned this once before so I wont get too much into the diode. But suffice it to say that not every part of the crystal would rectify the circuit. Rectification in this case means "act like a diode." More here.
Because the whisker was made of wire or a needle, and the crystal was softer, the surface was worn by use. OK hypothetically not every "crystal" was softer, but they were almost inherently of different harnesses. The very soft Graphite was used as a whisker sometimes. and at the other end, carborundum was considered durable. So either the whisker or the crystal was experiencing wear and usually it was the crystal. So hobbyist chose a crystal and whisker based both on it's sensitivity as a detector but also it's longevity. The hunt was on from the start for a "permanent detector."

Lets look at the hardness of some known whiskers and crystals:
WHISKERS
Graphite: 1-2 Mohs
Bornite: 3 Mohs
Iron: 4 Mohs
Steel: 7-8 Mohs
Bronze: 3 Mohs

CRYSTALS
Galena: 2.5 Mohs
Zincyte: 4 - 4.5 Mohs
Molybdenite: 1 - 1.5 Mohs
Iron Pyrite: 6-7 Mohs
Carborundum: 9 Mohs
Silicium: 7-8 Mohs
Chalcopyrite: 2.5 Mohs

As narrow as that variation is, remember that the Mohs scale is only a 10 point same topping out with diamond. not every whisker and crystal worked well together and even when they did, their mismatched harnesses could be problematic. Some combinations were proprietary and many of those were discovered by Mr. Greenleaf Whittier Pickard. He found both the Zincite /Bornite rectifier and the Zincite / rectifier. He called it the Perikon Detector.

He also tested versions with iron pyrite and silicon crystals. Silicon was novel enough he patented it's use in 1906. His Zinc Oxide detector was patented in 1909. He developed one with Molybdenite the same year. More here.