9.3.1 Composition of the gases and their flow-rates
The composition of the combustion gases determines the flame temperature and thus also the character of the spectrum and the intensity of the emitted light. Molecular emission, excited at low temperatures, is utilized in the flame photometric detector.
For this reason, a flame with excess of hydrogen is usually employed in the FPD.
It is desirable that the [H2] : [O,] molar ratio should be greater than 1.4. Free hydrogen is partially consumed in the combustion of eluted substances, but a greater part cools the flame and hence optimizes the excitation. In the initial FPD construc- tions, large gaseous mixture flow-rates, of the order of litres per minute, were usually employed [18]. In general, large gas flow-rates are not suitable, as the flame tempera-
156
ture increases with increasing amounts of combustion gases, resulting in a decrease in the molecular emission. An increased amount of combustion gas, with a constant elution volume of the test substance, leads to a decrease in the concentration of the eluted substance in the effective volume of the detector and thus also to a decrease in the measured signal. Thus newer FPD designs use lower gas flow-rates, e.g., 80 nil/min of H, and 20 ml/min of air.
The decrease in the air flow-rate is limited by the fact that a flame that contains excess of hydrogen is unstable and is easily extinguished. This frequently happens during elution of solvent from the column when the amount of oxygen drops below the nesessary level during solvent combustion. By decreasing the hydrogen and oxygen flow-rates, the continuum intensity is also decreased, the FPD background current decreases and a better signal-to-noise ratio is obtained. When oxidizing flames are used [61, 711, worse results are obtained than with reducing flames; on the other hand, oxidizing flames permit direct connection of the FID with the FPD.
Nitrogen is generally used as the carrier gas. The FPD signal can be increased by using helium, owing to more efficient heat removal. However, Bowman [ l l ] did not observe any change in the FPD signal when using helium as the carrier gas, possibly due to too high hydrogen and air flow-rates.
The carrier gas flow-rate is determined by the requirements placed on the separating part of the apparatus. In general the FPD signal increases with increasing carrier gas flow-rates, owing to a decrease in the flame temperature; this has been verified experimentally [ 331.
Carrier gases have not been especially purified for use in the flame photometric detector, with the exception of so-called sensitized measurements [3, 201. This design led to substantial narrowing of the linear dynamic range of the detector [68].
Impurities in the carrier gas result in the quenching of excited states and thus may cause a decrease in the FPD signal. Water molecules are ten times more efficient for quenching than the permanent gases (N,, 0,, CO,, etc.). Saturated hydrocarbons and hydrocarbons with double bonds exhibit effects similar to that of water. Purification of the carrier gas is unimportant as much more water is formed in the flame space during combustion than is introduced in the carrier gas.
9.3.2 Detector temperature
The detector temperature exerts a decisive influence on the background current and noise in the flame photometric detector. As the photomultiplier requires a low temperature, it is necessary to keep only the part of the detector from the column outlet to the end of the burner at an elevated temperature and to maintain the other parts of the detector at as low a temperature as possible, using suitable insulation.
Under these conditions, the most advantageous signal-to-noise ratio is obtained.
With increasing temperature, the signal-to-noise ratio decreases and the FPD response thus also decreases [54, 551. On an incresase in the detector temperature from
100 to 160°C, the response to hydrogen sulphide and sulphur dioxide decreases by half [53]. A further increase in the detector temperature, e.g., to 250 "C [25], requires cooling of the detector body with water.
9.3.3 Construction
The flame photometric detector measures the intensity of emission spectra originating in the flame space. By placing an interference filter in the radiation path, the wave- length corresponding to the monitored hetero-atom is selected. The light passed through the filter is amplified by the photometer and the resulting current is recorded.
Either diffusion or pre-mixed flames are employed in the flame photometric detector. Diffusion flames have lower temperatures and therefore are more suitable for monitoring the molecular emission of phosphorus, sulphur, the cyanide group or copper or indium halides. Under these conditions, the measurement is performed with a low background current and a low noise level. When a pre-mixed hydrogen - air flame is used, the noise increases by as much as forty times [ 2 2 ] .
Awe et al. [2] stated that, during the determination of heavy metal halides, they obtained identical results using flames with excess of hydrogen or excess of oxygen for the determination of Fe (373.5 nm), Pb (405.8 nm) and Sn (485 nm). The hydrogen flow-rate was 830 ml/min.
Flames with excess of hydrogen are frequently extinguished during solvent elution.
Thus burners [48] and control circuits have been modified for re-ignition of the flame For the determination of halogens, a container with indium or copper must be placed in the flame reaction zone. These containers are either mounted directly on the FPD burner in a similar way to the TlDA construction, or two flames, placed one above the other, are used, the lower flame saturating the burning mixture with the metal and emission being measured only after combustion in the upper flame [28]. In addition to metal salts or amorphous metals in containers, gauges or wires made of these materials are also placed in the flame. The larger the surface area, the better the performance. The metallic parts of the detector must be situated so that their radiation does not contribute to the FPD background signal.
A substantial portion of the FP D consists of the optical part before the interference filter. This part determines the selectivity of the device and increased attention should be paid to it. A necessary condition for proper functioning of the interference filter is that the Iight beam pass through the filter plane at right-angles. If this condition is not met, a broad wavelength interval is passed by the filter and costly filters with i.,,,, ,~ = 5 nm perform equally poorly as those with Arnaxl,* = 30 nm. Then par- ticularly the determination of phosphorus suffers from considerable interference from the C-C group.
The poorest measuring selectivity is obtained using mirrors. On the other hand, the highest signals are obtained owing to the low loss of light. The above difficulties
"61.
158
can be only partially removed by using a lens, as the flame is not a point light source.
Efforts to achieve maximal selectivity have led to the use of fibre optics [61], using which the sperical image of the flame is converted into a radiating point (see Fig. 9.1).
However, this design is less sensitive owing to loss of light through absorption. Success- ful design of the optical part is a necessary condition for the construction of a selective flame photometric detector.
Some workers have employed devices with shielded flames [48]. A glass cylinder is placed around the flame, which absorbs wavelengths shorter than 370 nm, thus suppressing the detector background current. However, it should be pointed out that these wavelengths should not be passed by the interference filters if the detector is correctly designed. If the background current decreases due to shielding, then the interference filter passes a broad range of wavelengths, i.e., the device is not selective.
Shielding decreases the emission intensity of the flame itself and of the hetero-atom monitored and the FPD signal is lower [53].
The choice of interference filter quality depends on the experimenter’s demands.
If the flame photometric detector is to be used for the detection of hetero-atoms, then it must be as selective as possible and consequently an interference filter with a small Amnx,,2 value must be employed. O n the other hand, if the FPD is to be used only as a single purpose instrument, e.g., for the determination of sulphur-containing substances in paper mills [27], an interference filter with a larger value can be used, as no interfering substances are present in the samples (the determination is not affected by substances that do not contain sulphur). By using a less selective device, an increase in the sensitivity is obtained.
FIG. 9.8. Scheme of the electric circuit of the flame photometric detector.
The photomultiplier is the measuring part of the FPD. Photomultipliers are selected according to their spectral characteristics, which determine the dependence of the gain on the wavelength of the incident radiation. The overall gain of the photomultiplier (up to lo’) is determined by the total voltage applied to the voltage divider. The gain is linear in wide voltage ranges and the value t o be employed is selected according to the most advantageous signal-to-noise ratio, usually from 700 to 1200 V. The photo- meter circuit is shown in Fig. 9.8. It must be emphasized that the photomultiplier is
a very sensitive measuring device with a limited life-time. This life-time is determined by the magnitude of the current drawn and therefore it is necessary to prevent strong irradiation of the phot3niultiplier when it is switched on.
The flame photometric detector can easily be connected with another detector.
Dual FP Ds, one part measuring at 394 r.m (sulphur) and the other at 526 nm (phospho- rus), have been described [lo, 311. Versino and Rossi [73] constructed a flame photo- metric detector that permits simultaneous measurement at three wavelengths corres- ponding to phosphorus (526 nm), sulphur (394nm)and chlorine (360 nm). The outputs are separated and yield two (three) recordings corresponding to phosphorus, sulphur and chlorine. This detector design is advantageous for pesticide analysis.
The flame photometric detector ici also often connected with a universal detector, e.g., a J C D [36, 471 or an FID [3, 61, 711. Connection of the FPD with the FID is especially easy, as ionization of organic substances also occurs in flames.
9.4 USE OF THE FLAME PHOTOMETRIC DETECTOR
The importance of the flame photometric detector in contemporary gas chromato- graphic instrumentation is steadily growing, as shown by a number of reviews [39, 50, 45, 60, 64, 751. The FPD is used increasingly frequently in air pollution analysis and in the analysis of pesticides. In general, the FP D can be imployed for the analysis of substance5 that contain phosphorus, sulphur, a halogen or the cyanide group. In addition to these substances, the FP D is also used for the determination of heavy metals. The applications of the flame photometric detector are surveyed in Table 9.3.
When the properties of the FP D are compared with those of other types of detector, it is found that the minimum detectable amounts are equal to [35, 591 or higher than those obtained with, for example, the TIDA or the ECD [ 7 3 ] . In all instances, a wider linear dynamic range and a better reproducibility is obtained with the FP D. The high selectivity of the flame photometric detector enables minimization or complete omission of purification procedures after the extraction of pesticides, thus consider- ably speeding up the analysis. The analytical results have excellent accuracy and pre- cision (see Fig. 9.9). It can be seen from the example given in Fig. 9.9 that only halo- gen-containing pesticides are monitored in the unseparated mixture using the FP D, while the ECD monitors the sum of pesticides and interfering substances. It can be assumed that further development in the construction of flame photometric detectors will be directed towards improvement of their selectivity and that they will become important in qualitative determination of pesticides using tabulated response ratios, The FPD is undoubtedly the most suitable detector for the determination of sulphur-containing compounds in air. No other detector permits the determination of sulphur dioxide, hydrogen sulphide, carbon disulphide, mercaptans, sulphides and RPlR9 R P I R C b RSlRCI.
160
TABLE 9.3
SURVEY OF FPD APPLICATIONS A N D CHARACTERISTICS
Substance determined
Minimum detectable amount
Concentration
range Ref.
SO29 H,S CH3SH SO2
cs2 Mercaptans sulphides SO,, H2S.
CH3SH, CS, sulphides s o , , H2S
so2
Mercaptans, sulphides thiophene
Total sulphur CS2
Pesticides
in air in air in petrol distillation fractions
in air in air in exhaust gases from petrol-burning engines in citrus oils in coal
10 PPb 10 PPb 0.05 ppm 0.1 pprn
0.001 ppm 10 PPm 0.003 pg 0.1 ppm
10- 1000 ppb 0-1 ppm 0.1 - 100 ppm
0.001 - 10 ppm 10- 1000 ppm
1 - 6.4::
5 x lo-" mole/sec ng level 1-lo4 ng
3 ppb
SO2 in presence of 20 ppb 0.1 - 500 ppm
0 3 , H2S, NO,,
co2
Parat hion, methylthion Malathion,
ethion, fenthion, methyltrithion, phorate and a number of others
PH3 in foodstuffs,
P in water
P
air, water
Pesticides
8 x lO-"g/sec
25 ng 25- 100 ng
5 Pg 0.2 ppb 40 Pg 0.05 ppb
53 18 20 58
67 30, 72 59 29
26 44 5 , 7, 8, 9, 14,17, 3 1, 55,66, 62
51
48 10
4 1 31 74 5, 6. 7, 8, 9, 10, 14, 17, 62, 66
Table 9.3 (continued)
Substance determined
Disyston Malathion Tetraethyl Methyl DEPP DePPT
Organo-halogens Chlorobenzene Organo-halogens,
chlorobenzene, m-diiodoben- zene, etc.
lindane, aldrin, D D T
DDT, dieldrin, in butter etc.
Lindane in foodstuffs Aldrin
Insecticides Organobromides Organoiodides CH,CN Organometallics Fe, Pb, Sn B
Si
Cr in urine
Ti, As, Zr Sn
pyrophosphate parathion
CCI, CCI,
Pesticides in milk, corn
c6 F6
Minimum detectable amount
Concentration
Ref.
range
1.9 x 10-8g/sec 2 x IO-'g/sec 7.7 x 10-'Og/sec 1.1 x Io-"g/sec 2.3 x 10-'3g/sec 1.6 x 10-'3g/sec 5 Ncg
4. I 0 - 9mole/sec 10-6g
16 ppm
0.07 pg
0.23 pg 0.022 pg 1 . 1 x 10-"g/sec 0.0007 pg 0.01 pg 0.047 pg
10-6g
0.44 ng lo-" mole
lo5
106
71 73 50 50 16 44 61 20 1 1 , 52
0.01-IOpg 11, 12, 52
33 40 32, 34 36,S2 0.01 - 1.4 pg 37
38 61 s x 103 73
41 2
2-20 ng 65
0-90 ng
1 o4 43
77
162
sulphoxides in concentrations of the order of 10 ppb after their separation. During analysis of such low concentrations in the atmosphere, the adsorption of gases on metallic parts of the instrument has been observed and these parts were thus replaced by PTFE [72]. Goretti and Possanzini [29] arrived at an interesting conclusion during
Vegetable sample extmct
a) FPD halogen-mode
interfen? 4 A
compoun s pesticides +. I
- - _ _ _ _ _ J ' .--_-- J C 4; - _ - _ _
FIG. 9.9. Analysis of chlorinated pesticides in a vegetable extract, performed with (a) the FPD and (b) the ECD [73].
determinations of sulphur-containing substances in exhaust gases from petrol engines;
they found that the amount of sulphur-containing substances in the analyzed mixture is negligible in spite of the fact that these substances could be smelled.
The flame photometric detector is also used for the determination of heavy metals.
The method has been applied in clinical analysis for the determination of chromium in amounts of up to 90 ng in urine [57].
9.5 LITERATURE
1 . Ackman R. G., Addison R. F.: Chem. Abstr. 75, 29 570v (1972) 2. Awe W. A., Hill H. H., Jr.: Anal. Chem. 45, 729 (1973) 3. Baldwin J. W.: Chem. Abstr. 69, 113 121h (1969)
4. Berek B., Westlake W. E., Gunther F. A.: J. Agric. Food Chem. 18, 143 (1970) 5. Beroza M., Bowman M. C.: J. Agric. Food Chem. 14, 625 (1966)
6. Beroza M., Bowman M. C.: Environ. Sci. Technol. 2, 450 (1968)
7. Bowman M. C., Beroza M.: J. Assoc. Off. Anal. Chem. 49, 1046, 1154 (1966) 8. Bowman M. C., Beroza M.: J. Agric. Food Chem. 15,465, 671 (1967)
9. Bowman M. C., Beroza M.: J. Assoc. Off. Anal. Chem. 50, 926, 940, 1228 (1967) 10. Bowman M. C., Beroza M.: Anal. Chem. 40, 1448 (1968)
11. Bowman M. C., Beroza M.: J. Chrornatogr. Sci. 7, 484 (1969)
12. Bowman M. C., Beroza M., Nickless G.: J. Chromatogr. Sci. 9, 44 (1971) 13. Braman R. S., Dynako A.: Pat. Ger. Offen. 1,900.309 (Sept. 4, 1969) 14. Brody S. S., Chaney J. E.: J . Gas Chromatogr. 4, 42 (1966)
15. Calcote H. F.: Ionization in High-Temperature Gases, Shuler K. E. (ed.), Academic Press, 16. Chovin P., Lebbe J., Moureu H. J.: J. Chromarogr. 6, 363 (1961)
17. Corley C., Beroza M.: J . Agric. Food Chem. 16, 361 (1968) 18. Crider W. L.: Anal. Chem. 37, 1770 (1965)
19. Crider W. L.: Anal. Chem. 41, 534 (1969)
20. Crider W. L., Slater R. W., Jr.: Anal. Chem. 41, 531 (1969)
21. Dagnall R. M., Thompson K. C., West T. S . : Analyst 92, 506 (1967) 22. Dagnall R. M., Thompson K. C., West T. S . : Analyst 93, 72 (1968) 23. Dagnall R. M., Pratt S. J., West T. S . , Deans D. R.: Talanta 16, 797 (1969) 24. Dagnall R. M., Smith D. J., West T. S . : Anal. Lett. 3, 475 (1970)
25. Dale W. E., Hughes C. C.: J. Gas Chromatogr. 6 , 603 (1968)
26. Darlage L. J., Block S. S., Weidner J. P.: J. Chromatogr. Sci. 11, 272 (1973)
27. Devonald B. H., Serenius R. S., Mchtyre A. D.: P u b . Pap. Mag. Can. 73, T 68 (1972) 28. Gilbert P. T.: Anal. Chem. 38, 1920 (1966)
29. Goretti G., Possanzini M.: J . Chromatogr. 77, 317 (1973) 30. Greer D. G., Bydalek T. J.: Environ. Sci. Techno(. 7, 153 (1973) 31. Grice H. W., Yates M. L., David D. J.: J. Chromatogr. Sci 8, 90 (1970)
32. Gunther F. A., Lopez-Roman A., Asai R. I., Westlake W. E.: Bull. Environ. Contam. Toxicol.
33. Gutsche B., Herrmann R. Z., Rudiger K.: Z . Anal. Chem. 241, 54 (1968) 34. Gutsche B., Herrmann R. Z.: Z . Anal. Chem. 242, 13 (1968)
35. Gutsche B., Herrmann R.: Z . Anal. Chem. 245, 274 (1969) 36. Gutsche B., Herrmann R.: Z . Anal. Chem. 249, 168 (1970) 37. Gutsche B., Herrmann R.: Z . Anal. Chem. 253, 257 (1971)
38. Gutsche B., Herrmann, R., Rudiger K.: Z . Anal. Chem. 258, 273 (1972) 39. Hartmann C. H.: Anal. Chem. 43, 113 A (1971)
40. Herrmann R., Gutsche B.: Analyst 94, 1033 (1969) 41. Hill H. H., Aue W. A,: J . Chromatogr. 74, 31 1 (1972)
42. Huyten F. H., Rijnders G. W. A.: 2. Anal. Chem. 205, 244 (1964) 43. Juvet R. S., Jr., Dubrin R. P.: Anal. Chem. 38, 565 (1966) 44. Kamada H., Kokubun N.: Bunseki Kagaku 17, 575 (1968) 45. McNair H. M., Chandler C . D.: J. Chromatogr. Sci. 11, 454 (1973) 46. Mizany A. I.: J. Chromatogr. Sci 8, 151 (1970)
47. Morrow R. W., Dean J. A., Shults W. D., Guerin M. R.: J . Chrornatogr. Sci. 7, 572 (1969) 48. Moye H. A.: Anal. Chem. 41, 1717 (1969)
49. Mulliken R. S.: Phys. Rev. 26, 1 (1925)
53. Natusch D. F. S . , Thorpe T. M.: Anal. Chem. 45, 1185 A (1973)
51. Okabe H., Splitstone P. L., Ball J. J.: J . Air Pollut. Contr. Ass. 23, 514 (1973) 52. Overfield C. V., Winefordner J. D.: J. Chrornatogr. Sci. 8, 233 (1970) 53. Pescar R. E., Hartmann C. H.: J. Chrornatogr. Sci. 11, 492 (1973)
54. Revel’skii I. A., Johnson V., Ilmoja K., Karavaeva V. G . , Loog E., Sirota T. S . : Ref. Zh.
55. Revel’skii I. A., Jonson V., Ilmoja K., Bezov V. M., Karavaeva V. G., Loog E., Sovakova 56. Ritschl R.: Z. Physik. 42, 172 (1927)
New York, London 1963, p. 107
4, 202 (1969)
Khirn. 1972, 12 N 422
T. M.: Ref. Zh. Khim. 1972, 12 N 448
164
57. Ross R., Shafik T.: J. Chromatogr. Sci. 11, 46 (1973)
58. Rupprecht W. E., Phillips T. R.: Anal. Chim. Acta 47, 439 (1969) 59. Scaringelli F. P., Rehme K. H.: Anal. Chem. 41, 707 (1969) 60. Selucky M. L.: Chromatographia 4, 425 (1971)
6 1. SevCik J.: Chromatographia 4, 195 (1 97 1 )
62. Shafik M. T.: Bull. Environ. Contam. Toxicol. 3, 309 (1968)
63. Sidebottom H. W., Badcock C. C., Jackson G. E., Calvert J. G., Reinhardt G. W., Damon 64. Soerensen 0.: Haus Tech., Essen, Vortagsveroeff. 283, 34 (1973)
65. Sowinsky E. J., Suffet I. H.: J. Chrornatogr. Sci. 9, 632 (1971) 66. Stevens R. K.: J. Assoc. Of. Anal. Chem. 50, 1236 (1967)
67. Stevens R. K., O'Keeffe A. E., Ortman G. C.: Environ. Sci. Technol. 3, 652 (1969) 68. Stubbs R. C.: Chem. Abstr. 76. 94 3212 (1972)
69. Sugiyama T., Suzuki Y., Takeuchi T.: J. Chromatogr. 77, 309 (1973) 70. Sugiyama T., Suzuki Y., Takeuchi T.: J . Chromatogr. 80, 61 (1973) 71. Svojanovskf V., Nebola R.: Chern. Listy 67, 295 (1973)
72. Tourres D. A.: Chromatographiu 5 , 441 (1972) 73. Versino B., Rossi G.: Chromatographia 4, 331 (1971) 74. West C. D.: Anal. Chem. 42, 811 (1970)
75. Winefordner J. D., Glenn T. H.: Advan. Chromatogr. 5, 263 (1968) 76. Winnett G.: J. Chromatogr. Sci. 8, 554 (1970)
77. Zado F. M., Juvet R. S . , Jr.: Anal. Chem. 38, 569 (1966) E. K . : Environ. Sci. Technol. 6, 72 (1973)
Most of the measuring devices employed in gas chromatography are based on the monitoring of various physical properties of eluted substances, e.g., thermal or electric conductivity, ionization potentials and luminescence. The signal of the coulometric detector alone corresponds to changes due to a chemical reaction of the eluted substance. For monitoring the course of the chemical reaction, a n electroana- lytical method known as coulometric titration is employed. Using this technique, the reagent is continuously regenerated and the electrolytic current is then a measure of the chemical reaction kinetics.