What is the purpose of using 8.33 kHz instead of 25 kHz frequency spacing?
The purpose is to accomondate more dedicated frequencies within the airband VHF range (117.975 to 137 MHz).
Increasing number of stations made this necessary, if the smaller division was not implemented, it was estimated that only 70% of future requirements for frequencies in Europe can be met.
833radio.com has further information and an nice table describing the channel spacing:
For simplicity, the actual 8.33 channels are displayed as rounded to .005 values. For example to use the frequency 118.0083MHz, you dial 118.010
If you are referring to 25 kHz spacing versus 8.33 kHz spacing, it gives more channel options in radio transceivers.
Ex1. There are 41 channels in the frequency space of 118.0 and 119.0 MHz, inclusively, with 25 kHz spacing.
Ex2. There are 121 channels in the frequency space of 118.0 and 119.0 MHz, inclusively, with 8.33 kHz spacing.
Since current airband transceivers are analog instead of digital, we needed some way for accommodating the increased amount of use on the airbands. We use analog receivers to this day because of their nature of transmitting and receiving as much as possible regardless of the situation or environment. The transmission may be faint, staticky or garbled, but it can still be received.
On the other hand digital transceivers like cell phones can use much fewer frequencies to accommodate many times the amount of information and users by compressing the information. A digital transmission is usually much clearer and more distinct with less distortion even over great distances with lower power. But, digital transmissions are an all or nothing proposition. Any loss or interruption in the flow of the data may render that portion of the information unreceivable. One word in a transmission may be entirely lost, changing the entire meaning of the transmission without the receiver knowing.
Ex. “Descending one five thousand” is a different meaning with different consequences than “Descending five thousand”.
25 spacing is used in the United States. Europe and other countries use 8.33 spacing. All NEW avionics and airband transceivers that I know of feature user selectable spacing.
Channel spacing for voice communication on the airband was originally 200 kHz until 1947, providing 70 channels from 118 to 132 MHz. Some radios of that time provided receive-only coverage below 118 MHz for a total of 90 channels. From 1947–1958 the spacing became 100 kHz; from 1954 split once again to 50 kHz and the upper limit extended to 135.95 MHz (360 channels), and then to 25 kHz in 1972 to provide 720 usable channels. On 1 January 1990 the frequencies between 136.000 and 136.975 MHz were added, resulting in 760 channels.
Increasing air traffic congestion has led to further subdivision into narrow-band 8.33 kHz channels in the ICAO European region; all aircraft flying are required to have communication equipment for this channel spacing. Outside of Europe, 8.33 kHz channels are permitted in many countries but not widely used as of 2012.
As already stated in other answers, the narrower the spacing, the more channels which can be fit within a given spectrum. Channels have to have some amount of spacing because signal modulated onto a carrier will create sideband; the exact nature of the sideband depends on the signal to be modulated and the type of modulation. In general, some sideband is inevitable because it actually represents the input signal. Rejecting sideband in a receiver generally excludes information (lower quality of demodulated signal). In order to allow for multiple channels to co-exist, transmitters must confine their sideband emissions so as not to interfere with transmissions on adjacent channels and receivers have to be capable of recovering the transmitted signal with acceptable fidelity; this means that receivers have to accept some amount sideband while rejecting frequencies too far away from that assigned to the channel.
The progression from wide spacings to narrower spacings reflects technological improvements in radio design; transmitters improve to confine their emissions to narrower sidebands while receivers improve to recover acceptable signals from the narrower transmissions. Congestion of the RF spectrum is what drives to need to increase the number of channels, either by allocating new portions of the spectrum or by fitting more channels into the existing spectrum.