Using the Terahertz Spectrum for Wireless Communication

holy_calamity writes "A first step to allowing wireless data transfer over a currently unused part of the electromagnetic spectrum is reported in New Scientist. Terahertz radiation exists between radio and infrared. A new filter created at the University of Utah can filter out particular frequencies, a prerequisite for using it for data. The abstract of the paper in the journal Nature is freely available."

ridiculously expensive

[evwah]

I regularly work with equipment that produces signals up to 50 GHz and let me tell you... components get much higher in cost the higher in frequency they go. a 3 foot 40GHz cable can cost hundreds of dollars and a 100GHz connector can cost a thousand dollars or more on its own. I imagine that producing and transmitting signals in the terahertz range is not economically viable for most companies.

Re:It might just take a while

[Anonymous Coward]

...so communication would have to be rather short-haul as in LAN.

Perhaps TFA should have mentioned that.
 
Wait...

Re:Hmm,

[evwah]

micrwave frequencies are usually considered to be the upper end of the radio frequency spectrum... the former being about 1G-300GHz, and the later covering 3Hz-300GHz.

Re:Geek into English.

[rhythmx]

Basically it says that putting the holes in a fractal pattern give much better results than holes in more 'normal' pattern. The rest is Calculus explaining how they can generate patterns that are really good at transmitting a certain frequency.

Sounds really interesting. I wonder if any of this applies to antenna design at average RF.

Re:Hmm,

[NixieBunny]

The microwave spectrum really ends at about 30 GHz, with the frequencies from 30G-300GHz called millimeter wave, and those from 300 GHz up called submillimeter. Terahertz technology is quite in its infancy. There was a terahertz conference last week, so the office I work in was pretty well cleared out. (I work on spectrometers that use what we consider low frequencies, The other thing about terahertz waves is that they behave quasi-optically, being focused by teflon lenses and blocked by cardboard. So it's not a radio band that one would use for cellphones.

Re:ridiculously expensive

[John Miles]

I regularly work with equipment that produces signals up to 50 GHz and let me tell you... components get much higher in cost the higher in frequency they go.

Depends on how precise you want to be. Conducting and measuring signals in that region of the spectrum with low-loss gear can be tough. Generating and receiving them isn't, necessarily. Not many people realize that some of the very first wireless communications experiments were done in the 60 GHz range, two years before Marconi [nrao.edu].

Re:ridiculously expensive

[Anonymous Coward]

Terahertz signals can be generated using microwave vacuum tubes (BWOs and Gyrotrons) and transmitted through waveguides, so cables aren't really necessary. Reception might be more of an issue, although that's not something I know much about. There is or at least has been work on terahertz generation at the University of Utah as well.

Not strictly true

[femto]

... a prerequisite for using it for data

It's not strictly true that you need to have bandpass filters to transmit information. There are other ways to select individual users without frequency division multiplexing. For example:

• Do it in the time domain (ultra wideband) using narrow pulses. Each user transmits at a different time.
• Use a spreading sequence to spread the signal so it takes up the entire band, with no need for a narrow filter (CDMA). Each user has a different sequence.
• Use multiple antennas to do space encoding. Users are separated in space, not frequency.

The gotcha is that you need some way of sampling the band. One way is to to use a bandpass filter, mixer and slow sampler. Another is to directly sample (using RTDs???) or in the case of UWB just detect pulses. Bandpass filters are the conventional way of doing it, but not the only way.

Re:Not strictly true

[NixieBunny]

Bandpass filters are not typically used with the astronomical receivers I'm familiar with. They use a local oscillator operating a few gigahertz above or below the interesting signal and just mix it down to microwave. The usual receiver sees the imagefrequency as well as the desired frequency, but the latest generation uses a sideband-separating mixer with hybrid couplers at RF and IF ports to allow separate reception of upper and lower sidebands. The group I work in was the first to apply such receivers to actual astronomical use.

Re:It might just take a while

[DarkAxi0m]

or... it could be for Atacama Large Millimeter/submillimeter Array - National Radio Astronomy Observatory :P

http://www.alma.nrao.edu/ [nrao.edu]

Google can be your friend too.. .

Re:Optical...

[Mad Bad Rabbit]

Once you get close to the frequency of infrared light... Why not just make the jump, and go with light instead? They're both going to be line-of-sight anyhow, with anything that blocks light very likely also blocks THz rf.

Actually no; terahertz rays [wikipedia.org] can go through wood, sheetrock, masonry, etc. (but not metal or water).

Clueless....

[j_square]

Another example of how the tabloids (Nature & Science) publish things that have been known for ages... There seems to be a trend that you can get anything published there, since the peer review is done by totally clueless physicists who do not know anything about the state of the art.

The concept of making filters by cutting holes in a sheet of metal has been known for ages. Using periodic (or in this case quasiperiodic) metallic patterns is called Frequency Selective Surfaces (FSS). There are numerous books and tons of publications in IEEE transactions, etc. in this area.

I did etched FSS filters for 375 GHz around 1982, and the concept was already pubslished in books by then.

Old stuff. Too many scientists, too much money, too little brain.

Re:ridiculously expensive

[91degrees]

A TV remote uses Terahertz frequencies. The components you need depend on what you're using it for.

Re:Geek into English.

[mdsolar]

In a way this is a pretty standard result. One can reduce the ringing in a Fourrier transform by including non-periodic sampling. What is provacative is the implication that there is some flaw in the surface plasmon interpretaion. Namely, they point to straight interference as being important rather than the constrained response of the surface electons.
--
Get solar: http://mdsolar.blogspot.com/2007/01/slashdot-users -selling-solar.html [blogspot.com]

Re:It might just take a while

[FuzzyDaddy]

http://www.gigabeam.com/technology.cfm [gigabeam.com] has a nice plot of atmospheric absorption versus wavelength. For reference, 100 dB/Km = 3 dB/30 meters - or 50% signal strength loss per 30 meters, not counting the 1/r^2 factor.

Also, generating and modulating signals, with current technology, is done by firing very expensive lasers at very customized pieces of semiconductor materials. As for receivers, NixieBunny would know better then me what the current technology cost and noise figures would be.

All of which to say, this is an interesting article, but it's about 1% of the way towards communications in this band.

Don't get me wrong - this is a cool paper, looks like good work, and this might have some very interesting technological applications. But the perpetual question of "what is it good for?" that every reporter asks (it's got to be a law or something) about every scientific advance misses the point. We don't know what it's good for, but it expands our knowledge of the world, and that can only help us.

Using it for something is the job of the next genius. These guys did enough by getting it to work. Someone else will have to figure out what it's good for.

I prefer THz for scanning!

[Freqfreq]

... been doin' teraherz for years - it's just "in fashion" now.

Publication with some terahertz images of concealed weapons on people (towards the article end):

http://stl.uml.edu/PubLib/DickinsonDSS2006.pdf [uml.edu]

lots of other THz articles if you chop back the URL to PubLib/