5.3.1 Gas flow-rate
The effect of the hydrogen flow-rate is frequently discussed. It is generally concluded that there is a certain optimum flow-rate for a given system, corresponding to the maximum detector response. This can be explained on the basis of the detection mechanism: with increasing hydrogen flow-rate, the temperature of the space where
96
eluted substances are cracked increases, leading to an increase in the response [30];
the subsequent decrease in the response with a further increase in the flow-rate is probably caused by the fact that the flame becomes too long and thus heats the burner less effectively. Increasing the hydrogen flow rate also causes cooling of the burner owing to the high thermal conductivity of hydrogen. This concept has been verified by measurement of the dependence of the FID response on the burner temperature;
it has been demonstrated that the ionization current increases with increasing burner temperatures [72]. In order to achieve the maximum degree of cracking, the use of an H, - N2 mixture is recommended [64].
Under FID conditions, argon, helium or helium with 0.03% of ammonia, in addi- tion to nitrogen, are used as carrier gases [5l]. When the magnitudes of the signal are compared for various carrier gases under identical experimental condititions, it is found that the signal for helium is always smaller than that for nitrogen [22, 38, 391.
These observations are in agreement with the theory of the ionization mechanism described above.
FIG. 5.2. The dependence of the FID response on the carrier gas flow-rate;
A - the flow-rate through the detector is varied; B - the flow-rate through the detector is maintained constant by adding pure carrier gas before the detector inlet.
FIG. 5.3. Scheme of the FID.
Although it is frequently maintained that the FID response is independent of the flow-rate of the carrier gas, some workers have demonstrated experimentally that the contrary is true. The flow-rate of the carrier gas is adjusted, considering the optimum separation of the analyzed mixture. The FID response, however, increases with increasing flow-rate, reaches a maximum and then decreases [33,64]. This dependence has been observed with various designs of FID and can be explained by the fact that
the shape of the flame changes with increasing flow-rate of the carrier gas; the flame becomes drawn out and its surface area increases. As diffuse flames are employed in the FID system, their temperature increases, leading t o an increase in the FID response.
If additional carrier gas is fed into the system after the column with a flow-rate adjusted so that the flow-rate through the detector is constant, a constant FID response is obtained (Fig. 5.2).
An increase in the FID response by adjustment of the flow-rate of carrier gas or hydrogen generally causes an increase in the noise level because the burner becomes hot and metal ions are emitted. It can be assumed that the cracking reaction time would be lengthened and consequently a greater FID response obtained if an external thermostat were connected to the part of the detector through which hydrogen and the eluted substances are led.
The flow-rate of air has only a very small effect on the FID response. As diffuse flames are employed, it is necessary to use excess of oxygen (several times the stoichio- metric amount), up to 600 cm3/min. The measurement is subject to a higher noise level when low flow-rates of air are employed. In addition to air, oxygen or an oxygen- nitrogen mixture is sometimes used; the flow-rate is then smaller as the stoichiometry of the burned mixture remains the same.
5.3.2 Geometry of the FID
The FID contains a small diffuse flame into which the eluted substance is fed. The ions formed migrate towards the electrodes, to which a small voltage is applied. An FID system is shown in Fig. 5.3.
The burner is either a metallic jet or a quartz capillary with an internal diameter of about 0.5 mm [74]. Quartz is advantageous because of the suppresion of the emission of secondary ions from the burner, i.e., it leads to a decrease in the detector noise. On the other hand, the response is usually lower than with metallic burners [90] as there is no substantial increase in the temperature of the pre-reaction space due to the poor heat conduction and therefore the cracking process is suppressed.
A system employing a burner with two separate jets has also been described [91].
The carrier gas and the eluted substance were introduced using one jet and a mixture of hydrogen and oxygen with the other. The detector dead volume was thus reduced and danger of explosion reduced but, due to the above suppresion of the cracking process, the FID signal was small. It is generally preferable that the eluted substance be introduced with the hydrogen flow and be heated to as high a temperature as possible before it enters the flame.
The FID noise increases during prolonged measurements, usually owing to the deposition of high-boiling compounds on the surface of the collecting electrode, insulating layers on the electrode, etc. The detector must then be cleaned mechanically in order to obtain the initial conditions. In addition to mechanical cleaning of elec-
98
trodes, fluorinated compounds such as alcohols and ethers have been injected into the system [4S]. Presumably the hydrogen fluoride formed reacts with the deposited compounds to produce volatile fluorides, which are the evaporated.
The electric field is used for collection of the charged particles. In most FID systems, a d.c. field is employed, the burner being one electrode with other electrode placed above it. The burner is usually the cathode and is insulated from the body of the detector. When a quartz capillary is employed, a metallic ring placed a t its orifice, or at or below the tip of the burner, serves as the cathode in order to suppress secondary ion emission and thus the noise as much as possible [97]. An a.c. electric field has also been tested [32] but has not enjoyed wider use.
The shape of the electrodes and their geometry and polarity are the most frequently discussed detector parameters. Several shapes of collecting electrode are depicted schematically in Fig. 5.4. It has been found that detectors with cylindrical collecting electrodes, i.e., systems b, d and e in Fig. 5.4, exhibit the lowest noise [83], while systems a and c have a high noise level due to heating of the collecting electrode [22].
The latter systems have a further disadvantage in the deposition of solid particles of high-boiling substances [34] on the electrode surface; the noise is consequently ncreased and the signal decreased. When silylated samples are treated, SiO, or 2 Si02.P,0, is deposited on the surface of the electrode, as has been shown by using
C d e
a b
FIG. 5.4. Some FID collecting electrode shapes and positions with respect to the burner;
for a description, see text.
Debye-Scherrer X-ray diffraction spectroscopy and emission spectroscopy [62]. The shape of the electrode determines the electric field gradient. The parallel plate arrangement is not suitable in the FID system, as the lines of force of the electric field are poorly developed and defined, because of the large voltage gradient in the flame space. This arrangement also requires that the positive and negative ions travel long distances, making the FID signal very variable [78]. Similarly, a very inhomogeneous electric field is obtained when symetrically placed semi-cylindrical electrodes are employed (see Fig. 5.4e) [SO]. The highest field intensity lies in the gap between the
edges of the two senti-cylindrical electrodes, while it is very low in the flame because of its location in the centre of the space enclosed by the electrodes. In Fig. 5.5 is depicted the electric field between the burner and the electrode formed by a coarse grid (A) when a third ring electrode with a potential equal to that of the anode (B, C) is placed in the system. It can be seen that the lines of force of the electric field are changed in the presence of the third electrode and hence the collection of charged particles was significantly affected. This phenomenon simulates changes in the flame shape due to an increase in the gas flow-rate. For this reason, contemporary FID designs contain an anode with an adjustable height [46, 691.
A
v i r A
FIG. 5.5. The electric field character in the FID system; A - collecting electrode - burner; B - collecting electrode, auxiliary electrode,burner; C- collectingelectrode, auxiliary electrode, burner.
FIG. 5.6. The volt-ampere curve for the FID with a cylindrical collector (1, 15 mm;
2, 25 mm) and with parallel plates (3, 8 mm; 4, 12 mm).
The magnitude of the polarization voltage is selected from the saturation regions of the volt-ampere curve. The voltage usually does not exceed 300 V and depends on the construction of the detector [18]. When parallel electrodes are employed, a higher voltage must be applied in order to attain the saturatiori current region (Fig. 5.6).
Although measurement employing a voltage between the saturated and multiplication regions of the Geiger-Muller curve has also been proposed [23], the absolute increase
100
in the ionization current is unfavourably affected by a considerable increase in the noise level and therefore it is probably unsuitable for the analysis of unknown samples.
The collecting electrode has frequently been a subject for discussion. In general, it is preferable that the FID burner be the cathode and the collecting electrode the anode.
As noted above, the highest charge concentration lies in the lower part of the flame and cations with low mobilities must thus travel only a short distance, while mobile ions travel further. It is obvious that the magnitude of the FID signal is independent of the polarity of the electrode in the parallel-plate arrangement [74], as the path lengths of all ions are equal in a symmetrical system. From a practical point of view, the electronic circuity is simpler when the polarization voltage is brought to the cathode (burner) and the anode is grounded. The detector noise is then substantially decreased, as capacitative effects in the cable between the anode and the amplifier are minimal, the device is not too sensitive to changes in the resistance of the girder insulation and higher cathode polarization voltages can be employed [ 5 9 ] .
FIG. 5.7. Schemes of the voltage and current amplifiers.
The ionization current is measured with a voltage or current amplifier (Fig. 5.7) [59]. In both instances, very low ionization currents can be measured (down t o 10-13A). However, voltage amplifiers require a high input resistance, R , , of the order of 1014 R. The resistance between the anode and cathode and the connecting cables, R,, which is depicted by dashed lines in the scheme, must have approximately the same value. The maximum amplification of the voltage drop on the measuring resistor, R,, is thus utilized and the output voltage is then given by the equation
The gain of a current amplifier is determined by the magnitude of the feedback resistor, R,. For the magnitude of the output voltage, it then holds that
When the two expressions for the output voltage are compared, it is evident that the functioning of the current amplifier is not significantly dependent on the input resis- tance, insulating properties of the connecting cables, insulation of the girder, etc. For these reasons, current amplifiers are more advantageous and are employed more frequently.
5.4 F ID APPLICATIONS
The FID is probably the most widely used measuring device in gas chromatrography.
Its application is also frequent in liquid chromatography [27 -29,45, 63, 93, 94, 1001, where the sample is usually transported on a moving wire. The evalutation of thin- layer chromatographic separations is based on a similar principle [76]: the plate is placed in a heated mantle into which a carrier gas is fed. With changing temperature gradient, the separated substances are gradually evaporated and introduced into the FID.
The simplicity of the design and the reliability of the FID have resulted in its use for teaching demonstration purposes; the elution of substances can also be observed visually on the basis of changes in the shape or colour of the flame [19,79]. A number of analyzers are also based on the application of the FID [43, 49, 70, 861. These analyzers have been designed not only for analyses in the earth’s atmosphere but also for measurements in extraterrestrial space [66, 1041.
The FID is employed in almost all chemical and biochemical fields. Its universality and high sensitivity for organic substances permitted trace determinations from its early stages [65]. The FID is most sensitive for hydrocarbons. As discussed above, the response of the FID strongly depends on the structure of the eluted substances and on the presence of hetero-atoms in their molecules. These dependences have been utilized, for example for following the degree of polymerization of styrene [77].
The presence of oxygen or sulphur in the molecules of the eluted substance always decreases the response of the FID. Nevertheless, this detector can’ still be used for the analysis of such substances, e g., organic acids [I031 present in tobacco [lo21 and esters of fatty [47] and dicarboxylic acids [9]. Amines [75] and pyridine derivatives [51] have also been analyzed using the FID. The detector has made possible the analyses of various bichemically important substances, such as steroids [85, 1051.
These substances, which usually have large molecules, must be pre-treated prior to the analysis, for example by silylation, preventing the use of some detectors, such as the ECD. The FID is thus also employed in pesicide analysis [6, 8, 961.
102
An FID has not been specially constructed for the analysis of inorganic substances.
It follows from the detection mechanism that the eluted substance must form a methyl radical in order to be detectable by the FID. However, the use of the FID for the analysis of various inorganic substances has been repeatedly reported in the literature.
For example, the FID is sometimes used for the determination of metals [4, 7, 8, 901.
The measurement is then performed at high flame temperatures, using, for example, pre-mixed flames or multiple burners. Direct ionization of the metals by the TIDA mechanism is employed. The FID also responds to water [67, 981, rare gases [98], oxygen [40, 88, 981, nitrogen [40], nitrogen oxides [88] and carbon dioxide [98].
In general, these responses are not the result of the FID mechanism described above, but are caused by a decrease in the flame temperature and an increase in the electron capture probability, as reflected in the negative signals which are usually obtained.
5.5 LITERATURE
1. Ackman R. G.: J. Gas Chromatogr. 6, 497 (1968) 2. Andeatch J. A,, Feinland R.: Anal. Chem. 32, 1021 (1960)
3. Anderson A., Jerele S., Shimanskaya M. V.: Latv. PSR Zinat. Akad. Vestis, Kim. Ser. 1971, 4. Araki S., Suzuki S., Hobo T., Yamada M.: Bunseki Kagaku 19, 493 (1970)
5. Askew W. C., Maduskar K. D.: J. Chromatogr. Sci. 9, 702 (1971)
6. Aue W. A.: Aduances in Chemistry, Series No. 104, Pesticides Identification at the Residue 7. Aue W. A., Hill H. H.: J. Chromatogr. 74, 319 (1972)
8. Aue W. A., Hill H. H.: A n a l . Chem. 45, 729 (1973) 9. Binder H., Lindner W.: J. Chromatogr. 77, 175 (1973)
480
Level, 4, 39 (1971)
10. Blades A. T.: J . Chromatogr. Sci. 8, 414 (1970) I I . Blades A. T.: J. Chromatogr. Sci. 10, 693 (1972) 12. Blades A. T.: J. Chromatogr. Sci. 11, 251 (1973) 13. Blades A. T.: J. Chromatogr. Sci. 11, 267 (1973)
14. Blu G., Lazarre F., Guiochon G.: Anal. Chem. 45, 1375 (1973) IS. BoEek P., Novak J., Janik J.: J. Chromatogr. 43, 431 (1969) 16. B o k k P., Novak J., Janak J.: J. Chromatogr. Sci. 8, 226 (1970) 17. BoEek P., Novak J., Janik J.: J. Chromatogr. 48, 412 (1970)
18. Bolton H. C., McWilliam I. G.: Proc. R. SOC. London, Ser. A 321, 361 (1971) 19. Brabson G. D.: J. Chem. Educ. 49, 71 (1972)
20. Bruderreck H., Schneider W., HalBsz I.: Anal. Chem. 36, 461 (1964)
21. Calcote H. F.: Ionization in High-Temperature Gases, Shuler K. E., Ed., Academic Press, New York, London 1963, p. 107
22. Chizhov L. V., Shorygin A. P., Petukhova E. A,: Gazoc. Khromatogr. 1969, 66 23. Clardy E. K.: US Pat. 3,542,516
24. Clementi S., Savelli G., Vergoni M.: Chromatographia 5, 413 (1972) 25. Douglas D. M., Schaeffer B. A.: J. Chrornatogr. Sci. 7, 433 (1969) 26. Dressler M.: J. Chromatogr. 42, 408 (1969)
27. Dubsky H., Pajurek J., KrejEi M.: Chem. Listy 67, 93 (1973) 28. Dubsky H.: Chem. Listy 67, 533 (1973)
39. Dutta J., Ghosh A., Hoque M., Ghosh A,: Chem. Ahstr. 69, 32 820h (1968) 30. Eggertsen F. T., Stross F. H.: Thermochim. Acta 1, 451 (1970)
31. Fenimore C. P.: The International Encyclopedia of Physical Chemistry and Chemical Physics, Vol. 5., Chemistry in Premixed Flames (Trotman-Dickenson A. F., Ed.), Pergamon Press, Oxford 1964
32. Fertig G. H.: Ger. Offen. 1,803,616 (Jul. 12, 1969)
33. Folmer 0. F., Jr., Haase D. J.: Anal. Chirn. Acta 48, 63 (1969) 34. Forster E. P., Weiss A. H.: J. Chromatogr. Sci. 9, 266 (1971)
35. Franklin J. L., Munson M. S. B., Field F. H.: Ionization in High-Temperature Gases 36. Gaspar G.: Meres. Automat. 18, 17 (1970)
37. Gray P., Herod A. A,, Jones A,: Chern. Rec. 71, 247 (1971)
38. Green L.: Hewlett-Packard Application Note GC-2-73, Operating Conditions for Optimum 39. Green L. E., Mikkelsen L.: Hewlett-Packard Notes, Factors Affecting Linearity of the Flame 40. Gupta M. C., Mathur M., Chandra K., Bhattacharya S . N.: J. Chromatogr. Sci. 11, 373 41. Hainova O., BoEek P., Novak J., Janak .I.: J. Gas Clzromatogr. 5, 401 (1967)
42. Harley J., Nel W., Pretorius V.: Nature 181, 177 (1958) 43. Hartmann K.: Wasser, Luft Betr. 15, 21 I (1971) 44. Hill D. W.: Chem. Abstr. 73, 115 954p (1971) 45. Huber J. F. K.: J. Chromatogr. Sci. 7 , 172 (1969)
46. Ibragimov I. A., Farzane N. G., Feldleifer M. B.: Chem. Abstr. 7 0 , 64 093k (1969) 47. Iverson J. L.: J. Assoc. Off Anal. Chem. 53, 1214 (1970)
48. Kaczaj J.: US Pat. 3,531,256 (Sept. 29, (1970) 49. Kadlec K.: Vod. Hospod. B21, 17 (1971)
50. Kaiser R., Stoll W., Fischer K.: Chromatographia 2, 20 (1969)
51. Kan I. I., Sembaev D. K., Suvorov B. V.: Z h . Anal. Khim. 25, 374 (1970) 5’. Karasek F. W., Kane D. M.: J. Chromarogr. Sci. 10, 673 (1972)
53. Karasek F. W., Kane D. M.: Anal. Chem. 45, 576 (1973) 54. Karmen A,, Kelly E. L.: Anal. Chem. 43, 1912 (1971)
55. Katachevtsev G. J., Tal’rose V. L.: Chem. Absrr. 78, 149 425k (1973) 56. Kalmanovskii V. I.: Anal. Abstr. 20, 4491 (1971)
57. King I. R.: Ionization in High-Temperature Gases (Shuler K. E., Ed.) Academic Press, 5 s . King W. A,, Dupre G. D.: Anal. Cheni. 41, 1936 (1969)
59. Knapp 0.: Chromatographia 2, 67 (1969) 69, Knapp 0.: Chromatographia 2, 1 I I (1 969)
61. Kozeiko T. A,, Mashkevich A. I.: Zacod. Lab. 39, 25 (1973) 62. Lakeland B. R., McDermontt I. T.: J. Chromatogr. 38, 392 (1962) 63. Lapidus B. M., Karmen A,: J. Chromatogr. Sci. 10, 103 (1972) 64. Levy R. L., Walker J. Q., Wolf C. J.: A t ~ a l . Chem. 41, 1919 (1969) 65. Lovelock J. E.: Anal. Chem. 33, 162 (1961)
66. Lucero D. P., Smith P. H., Johnson R. D.: I S A Trans. 10, 58 (1971) 67. Lucero D. P.: J. Chromatogr. Sci. 10, 463 (1972)
68. Lucero D. P., Smith P. H.; J. Chromatogr. Sci. 10, 544 (1972) 69. McCoy R. W., Cram S . P.: J. Chromatogr. Sci. 7, 17 (1969) 70. McNair H. M., Chandler C. D.: J. Chroniatogr. Sci. 11, 454 (1973) 71. McWilliam I. G., Dewar R. A,: Narirre 181, 760 (1958)
(Shuler K. E., Ed.), Academic Press, New York, London 1963, p. 67
Performance of the Model 571 1 A Flame Ionization Detector Ionization Detector
( I 973)
New York, London 1963, p. 197
104
72. McWilliam I. G.: J. Chromatogr. 51, 391 (1970)
73. Mead A. S., Speakman F. P.: Chromatographia 5 , 21 1 (1969) 74. Micheletti S. F., Bryan G. T.: Anal. Chim. Acta 48, 51 (1969) 75. Moffat A. C., Horning E. C.: Anal. Letters 3, 205 (1970)
76. Mukherjee K. H., Spaans H., Haahti E.: J. Chrornatogr. 61, 317 (1971)
77. Nikolaev A. F., Belogorodskaya K. V., Andreev A. I., Rumyantsev J. G.: Vysokomol.
78. NovAk J., BoEek P., Reprt L., Janhk J.: J . Chrornutogr. 51, 385 (1970) 79. Nowak A. V., Malmstadt H. V.: J. Chem. Educ. 45, 519 (1968) 80. Nunnikhoven R.: Fresenius’ Z. Anal. Chem. 236, 79 (1968) 81. Oster H., Oppermann F.: Chromatogruphia 2, 251 (1969)
82. Perkins G., Jr., Laramy R. E., Lively L. D.: Anal. Chem. 35, 360 (1963) 83. Prescott B. O., Wise H. L., Chesnut D. A.: US Pat. 3,451,780 (Jun. 24, 1969) 84. Rossiter V.: J. Chrornatogr. Sci. 8, 164 (1970)
85. Rutten G.A.F.M., Luyten J.A.: J . Chrornatogr. 74, 177 (1972)
86. Sawicki E., Corey R. C., Dooley A. E., Monkman J. L., Ripperton L. A,, Sigsby J. E., White L. D.: Health Lab. Sci. 8, 248 (1971)
87. Schaeffer B. A.: Anal. Chem. 43, 448 (1970) 88. Schaeffer B. A,: J. Chromatogr. Sci. 10, 110 (1972)
89. Sokolov D. N., Golubeva L. K.: Zacod. Lab. 35, 143 (1969) 90. Sokolov D. N., Vauin N. A.: Zucod. Lab. 36, 513 (1970) 91. Spencer S. F., Mikkelsen L.: US Pat. 3,399,974 (Sept. 3, 1968)
92. Sternberg J. C., Gallaway W. S., Jones D. T. L.: Gas Chromatography (Brumer N., Callen 93. Stevens R. H.: J. Gas Chromatogr. 6, 375 (1968)
94. Stolyhwo A., Privett 0. S., Erdahl W. L.: J. Chrornatogr. Sci. 11, 263 (1973)
95. Sugden T. M.: Ionization in High-Temperature Gases (Shuler K. E., ed.), Academic Press, 96. Svojanovskq J., Nebola R.: Chem. Listy 67, 295 (1973)
97. Taft E. M.: US Pat. 3,399,339 (Aug. 27, 1968) 98. Thombs D. A,: Chromatographia 6, 111 (1973)
99. Umstead M. E., Woods F. J., Johnson J. E.: J . Chrornatogr. Sci. 8, 375 (1970) Soedin., Ser. A15, 436 (1973)
J. E.. Weiss M. D., eds.), Academic Press 1962
New York, London 1963, p. 145
100. Van Dijk J. H.: Brit. Pat. 1,292,754 (Oct. 11, 1972)
101. Van Tiggelen A.: Ionization in High-Temperature Gases, (Shuler K. E., ed.), Academic Press, New York, London 1963, p. 165
102. Van der Wal S. Sj.: Column 3, 5 (1970)
103. Watanabe S., Nakasato S., Kuwayama H., Sasamoto Y., Shiraishi S., Seino H., Nagai T., Negishi M., Hayano S.: Yukuguku 22, 95 (1973)
104. Wilhite W. F., Burnell M. R.: Proceedings of the Ninth National Instrument Society Of
America, Analysis Instrumentation Symposium, April 29 - May I , Houston, Texas 1963 105. Yannone M. E., Mueller J. R., Osborn R. H.: Chromatogruphia 3, 13 (1970)
Detectors denoted in the literature as thermionic detectors ( J I D ) , alkali flame ionization detectors (AFID), chemi-ionization detectors (CID), nitrogen detectors (NFID) and phosphorus detectors belong to the group of ioi ization detectors in which thermal energy is used as a source of ionization energy. However, the detection mechanism does not involve measurement of the ionization current of the eluted substance as with other ionization detectors (ECD, FID, PID, HeD); instead, changes in the ionization of an alkali metal present in the detector effective space are monitored.
A detector using an alkali metal salt was first described in 1964 [38]. Since then, this detector has been one of the most frequently employed devices, especially in the analysis of phosphorus- and nitrogen-containing compounds. Its high sensitivity for these hetero-atoms led to its being called the phosphorus or nitrogen detector.
The measurement of other atoms, such as sulphur, the halogens, arsenic, antimony, tin and lead, is not very sensitive and the varying magnitude and character of the detector response for these substances have often been studied.
A common feature of all TIDA constructions is the measurement of the ionization current as the sum of all of the ionization processes that take place in the detector, alkali metal ionization predominating. Thermal energy for the ionization process is liberated by combustion or by electrical heating. The detection mechanism has been explained from several points of view. It has been assumed that the combustion products of phosphorus-containing substances react with the alkali metal or with its salt [38], the detector signal corresponding to the ionization of the evaporating salt.
It has also been suggested that the ionization results from collisions of the alkali metal atoms with intermediates (e.g., radicals) formed in the flame [50]. Another mechanism involves evaporation of the alkali metal salt in the flame space as a result of interaction with photons [38, 471. Brezhnikov et al. [ 9 ] , who have contributed significantly to TIDA research, did not propose an unambiguous detection mechanism and pointed out that clarification of the mechanism is a condition for further TIDA development [8]. The mechanism has been partially elucidated by experiments with a flameless detector, called the chemi-ionization detector [59]. The results demonstra- ted unambiguously that ionization occurs in the gaseous phase and that neither hydrogen nor combustion products need be present. Ionization depends on the amount of alkali metal present in the gaseous phase, which in turn depends on the temperature of the surroundings. TIDA anomalies have been explained on the basis of evaluation of hetero-atom-selective interactions [61].