The adsorbed material reduces the work function of the surface and, thus, electron emission is increased and the current collected at the anode rises. The NPD has a very high sensitivity, i.e., about an order of magnitude less than that of the electron capture detector (ca.10-12 g/ml for phosphorus and 10-11 g/ml for nitrogen).
The Electron Capture Detector
Lovelock's work on ionization detectors culminated in the invention of the electron capture detector. However, the electron capture detector operates on an entirely different principle from that of the argon detector. A low energy b-ray source is used in the sensor to produce electrons and ions. The first source to be used was tritium absorbed into a silver foil but, due to its relative instability at high temperatures, this was quickly replaced by the far more thermally stable 63Ni source. 
The Thermionic Ionization Detector
Electrons produced by a heated filament can be accelerated by an appropriate potential so that they attain sufficient energy to ionize any gas or vapor molecules in their path. In 1957, the early days of gas chromatography, Ryce and Bryce modified a standard vacuum ionization gauge to examine its possibilities as a GC detector. A diagram of the device is shown in figure 47.
The sensor consisted of a vacuum tube containing a filament, grid and anode, very similar in form to the thermionic triode valve. The tube was operated under reduced pressure and an adjustable leak was arranged to feed a portion of the column eluent into the gauge. The sensor was fitted with its own pumping system and vacuum gauge and the usual necessary cold traps. Helium was used as a carrier gas and the grid collector–electrode was set at +18 V with respect to the cathode and the plate at -20 V to collect any positive ions that are formed. As the ionization potential of helium is 24.5 volts, the electrons would not have sufficient energy to ionize the helium gas. However, most organic compounds have ionization voltages lying between 9.5 and 11.5 V and consequently would be ionized by the 18 V electrons and provide a plate current. The plate current was measured by an impedance converter in much the same way as the FID ionization current. The detection limit was reported to be 5 x 10-11 moles, but unfortunately the actual sensitivity in terms of g/ml is not known and is difficult to estimate.
