来宝网移动站

The Discharge Detector
About the same time that Ryce and Bryce were developing the thermionic ionization detector, Harley and Pretorious and (independently) Pitkethly and his co-workers were developing the discharge detector. By applying the appropriate potential, a discharge can be maintained between two electrodes situated in a gas providing the pressure is maintained between 0.1–10 mm of mercury. After the discharge has been initiated, the electrode potential can be reduced and the discharge will still continue. Under stable discharge, the electrode potential remains constant and independent of the gas pressure and the electrode current.The electrode potential, however, depends strongly on the composition of the gas. It follows, that the system could function as a GC detector. Pitkethly modified a small domestic neon lamp for this purpose and a diagram of his sensor is shown in figure 48.The lamp was operated at about 3 mm of mercury pressure with a current of 1.5. Under these conditions the potential across the electrodes was 220 V. Pitkethly reported that a concentration of 10-6 g/l gave an electrode voltage change of 0.3 V.

The noise level was reported to be about 10 mV thus at a signal–to–noise level of 2 the minimum detectable concentration would be about 3 x 10-11g/ml. This sensitivity is comparable to that of the FID and the argon ionization detector. The detector was claimed to be moderately linear with a linear dynamic range of three orders of magnitude but values for the response index were not reported. It was not apparent whether the associated electronics contained non linear signal modifying circuitry or not. Unfortunately, there were several disadvantages to this detector. One disadvantage was the erosion of the electrodes due to "spluttering" In addition, the electrodes were contaminated by sample decomposition and it was essential that it was used with a well–controlled vacuum system.

The Spark Discharge Detector
Lovelock noted that the voltage at which a spark will occur between two electrodes situated in a gas will depend on the composition of the gas between the electrode tips and suggested that this could form the basis for a GC detector. The system suggested by Lovelock is shown in figure 49.

The sensor consists of a glass tube in which two electrodes are sealed. The electrodes are connected in the circuit depicted in figure 49. The voltage across the electrodes is adjusted to a value that is just less than that required to produce a spark. When a solvent vapor enters the sensor, the sparking voltage is reduced and a spark discharge occurs. This discharges the capacitor until its voltage falls below that which will maintain the spark discharge. The capacitor is then charged up through the charging resistor until the breakdown voltage is again reached and another spark is initiated. Thus the spark frequency will be proportional to (or at least be a monotonic function of) the vapor concentration. The total counts in a peak will be proportional to the peak area and, if a digital–to–analog converter is also employed, the output will be proportional to the concentration in the detector and thus, plotted against time, will provide the normal chromatogram. This detector does not appear to have been developed further but is an interesting example of a sensor that, in effect, produces a digital output.The Radio Frequency Discharge Detector
When an RF discharge occurs across two electrodes between which the field is diverging (i.e. within a coaxial electrode orientation) a DC potential appears across the electrodes, the magnitude of which depends on the composition of the gas through which the discharge is passing. Karman and Bowman developed a detector based on this principle. A diagram of their detector is shown in figure 50.

The sensor consisted of a metal cylinder that acted as one electrode with a coaxial wire passing down the center that acted as the other. A 40 MHz radio frequency was applied across the electrodes and the DC potential that developed across them was fed via simple electronic circuit to a potentiometric recorder. The resistance capacity decoupling shown in their circuit appears hardly sufficient to achieve the removal of the AC signal in a satisfactory manner and consequently, the circuit shown in figure 50 may be only schematic. The column was connected directly to the sensor and the eluent passed through the annular channel between the central electrode and the sensor wall.

The response of the radio frequency discharge detector was reported as 106mV for a concentration change of 10-3 g/ml of methyl laureate. The noise level was reported to be 0.05 mV, which would give the minimum detectable concentration for a signal–to–noise ratio of 2 as about 6×10-10 g/ml. This detector had the advantage of operating at atmospheric pressure and so no vacuum system was required. The effect of temperature on the detector performance was not reported, nor was its linearity over a significant concentration range. This detector appears not to have been made commercially

The Ultrasound Whistle Detector The velocity of the propagation of sound through a gas depends on its density and, thus, the presence of a solute vapor in a gas changes the velocity of sound through it. This velocity change can be utilized as a basis for vapor detection in GC. The frequency of a whistle, consisting of an orifice which directs a stream of gas against a jet edge proximate to a resonant cavity, is related to the velocity of sound in the gas passing through it. A diagram of such a whistle is shown in figure 12. Nyborg et al. (38) showed that the frequency (fn) of the whistle could be described by the following equation. Testerman and McLeod designed and built a detector based on the whistle principle. In their sensor design, typical values taken for the dimensions in the diagram, and variables in the equation, were (t), 0.064 mm, (d), 0.74 mm, (h), 1.676 mm and (L) 3.81 mm. Under the flow conditions normally used for GC separations, frequencies ranging from 30-50 kHz (supersonic frequencies) were observed. The sensor contained two sound generators, one operating with pure carrier gas and the other with the eluent from the column. The two frequencies were allowed to beat together, the beat frequency being directly related to the frequency difference between the two whistles and consequently the density difference between the contents of the two sensors. An example of the use of the whistle detector to monitor the separation of a mixture of hydrocarbons is shown in figure 60. The sample size was 7.5 ml of gas mixture and the carrier gas flow rate was 180 ml /min. This chromatogram illustrates the effective use of the detector and the operating conditions shows its limitations. The sensitivity appears somewhat less than that of the katharometer but the very high flow rates necessary to activate the whistle restrict the use of this type of detector very severely. In the original report the linearity was stated to cover 2 orders of magnitude of concentration but with modern electronics it is likely that this linear range could be extended by at least another order of magnitude. Figure 60 The Separation of a Mixture of Hydrocarbons Monitored by the Supersonic Whistle Detector

Despite the relative casual arrangement of the adsorbent, it would appear that the adsorption was quite efficient and, with 10 mg charges on the column, an accuracy of 1% could be easily achieved. Later Bevan et al [45,46], reduced the size of the absorber and employed charcoal as the adsorbing material. Although this improved the performance of the detector and reduced the necessary sample size, the detecting system was never made commercially. Even after modification, its sensitivity was relatively poor and despite it being an absolute detecting system, it placed too many restrictions on the operation of the chromatograph and the samples that could be chromatographed to be generally useful.The Surface Potential Detector
The surface potential detector was developed by Griffiths and Phillips [47,48] in the early 1950s and consisted of a cell containing two parallel metal plates between which flowed the column eluent. One plate was mechanically attached to an oscillator that vibrated the plate at about 10 kHz. If the plates are identical, the surface charge on each plate is the same and so no potential is induced into the second plate by the vibrating plate. If however the surfaces are dissimilar, then the surface charge on each plate will differ and the vibrating plate will induce a potential on the other plate. A diagram of the detector is shown in figure 62.

Both plates were constructed of the same metal but one plate was coated with a monolayer of a suitable substance that would absorb any vapors present in the column eluent. The absorbing layer caused the charge on the two plates to be dissimilar and thus a potential appeared acrossthetwoplates which was balanced out by the bias potentiometer. When a solute vapor passes through the detector, some is distributed into the absorbent layer, changing the surface charge and thus inducing a change in potential between the electrodes. This produces an AC signal voltage that can then amplified, rectified and the output passed to a recorder (or to a data acquisition system). The signals provided by the detector sensor could be as great as several hundred millivolts.

The sensitivity of the detector was claimed to be similar to that of the katharometer (i.e. about 10-6 g/ml). Its response was partly determined by the distribution coefficient of the solute vapor between the carrier gas and the absorbing layer (and thus the chemical characteristic of the coating) as well as the chemical nature of the solute itself. As a consequence, the response varied considerably between different solutes. Within a given homologous series, however, the response increased with the molecular weight of the solute, but this was probably merely a reflection of the increase in the value of the distribution coefficient with molecular weight. Although an interesting alternative method of detection, this detector has been little used in GC and is not commercially available