JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 533 Molecular Non-thermal Excitation Spectrometry (MONES): a Procedure for the Determination of Non-metals Using Diatomic Molecules in the Non-thermal (FANES) Atomiser Part lm Determination of Fluoride and Chloride Ions by Magnesium Fluoride and Magnesium Chloride MONES* Plenary Lecture Klaus Dittrich and H. Fuchs Karl-Marx-University, Section of Chemistry, Analytical Division, DDR 70 10 Leipzig, Talstr. 35, GDR A method for the determination of the halides F- and CI- by non-thermal excitation of MgF and MgCl molecules in a glow discharge established in a hot graphite tube (2000 "C, 2 kPa) has been developed. The MgX molecules are formed by simultaneous evaporation of Mg- and F-containing species (Mg being the additive and F the sample).The detection limits are 0.5 ng of fluoride and 0.24 ng of chloride, respectively. The method is called ETE-MONES (molecular non-thermal excitation spectrometry with electrothermal evapora- tion) by analogy with the FANES technique (furnace atomisation non-thermal excitation spectrometry). Keywords: Furnace atomisation non-thermal excitation spectrometry; molecular non-thermal excitation spectrometry; non - th erm al o ve r-excita tion; molecule form a tion; n on -metals de term in ation The early analytical applications of the emission spectra of diatomic molecules utilised flames. 1-3 Later, following the trends of microanalysis, molecular emission cavity analysis (MECAj4 was developed, which is a chemiluminescence method using cavities and flames.The MECA technique has been used for the determination of halides through the measurement of InC1, InBr and In1 molecular emission.'-8 The detection limits reported for these methods were 5-50 ng for bromine and 1.3-SO ng for chlorine. Electrically heated graphite tubes have also been used for the measurement of molecular emission, and Gutsche and co-workers9%10 successfully determined halides by measuring the emission of InBr and InCl. Molecular absorption measurements have also been used for the determination of halides using diatomic molecules. Fuwa and co-workers determined chloride by molecular absorption of InCI'1 and AIC1.12 Dittrich and co-workers have also determined fluoride and chloride by MX molecular absorption (where M = Al, Ga, In or Mg) in an electrothermal atomiserl3-17 and reported detection limits in the ng region. The advantages of molecular absorption spectrometry with electrothermal evaporation (ETE-MAS) for the determination of non-metals have been reviewed,lS--'Y and the most attractive features are the simplicity of the methods and the low detection limits achieved for various elements.The development of furnace atomisation non-thermal excitation spectrometry (FANES) by Falk et al. 20321 provided a different approach to the determination of trace metals in micro-samples by atomic emission. However, in real sample analysis, depressions of the atomic emission signals of the Group 111 elements in the presence of halides was observed.z2 This was probably due to the relatively low vapour tempera- ture within the FANES atomiser tube at the time of volatilisation of the analyte and matrix species.As spectro- scopic studies of InF,23 GaF24 and MgCP molecules have been performed using conventional hollow-cathode dis- charges, it seemed reasonable to investigate the formation of these molecules in the FANES discharge. The initial results from this study have been reported recently26 for the * Presented at the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January, 1987. determination of F-, C1- and Br- ions by InX emission in the FANES source. These investigations gave rise to a new pseudonym, MONES, molecular non-thermal excitation spectrometry. The detection limits obtained for F-, C1- and Br- by MONES were superior to the values obtained earlier by ETE-MAS and other emission techniques.In this paper, the possibility of determining halides using MgX MONES measurements will be described. Experimental Apparatus A diagrammatic representation of a two-channel FANES/ MONES spectrometer is shown in Fig. 1. The two-channel 13A 15 Fig. 1. Diagram of the FANES/MONES spectrometer: 1 , power supply for heating and discharge with microcomputer control; 2, FANESMONES source (see Fig. 2); 3 A and B , two identical monochromators, SPM 2 (Zeiss Jena, GDR), with gratings of 1300 grooves mm-1 for 750-360 nm, resolution 2 nm mm-I and 2600 grooves mm-' for 360-180 nm, resolution 1 nm mm-I; 4, teletyper for programming of the microcomputer (1); 5 , Oscilloreg chart recorder (Siemens, Karlsruhe, FRG); 6, recorder, working in A, B or A-B mode; 7, two identical power supplies for PM tubes A and B, Type 4213 (Statron, Furstenwalde, GDR); 8, vacuum pump; 9, thermostat to ensure constant cooling; 10, gas; 11, quartz lenses, f = 80 mm; 12, HCL/EDL for wavelength adjustment only; 13A and €3, PM tubes, M12 FQC51 (Werk fur Fernsehelektronik, Berlin, GDR); 14, power supply for HCL; and 15, beam splitter534 v) C 3 4- .- 2 4 0 - P c .- I] 30- i .- v) C 20 .- JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 - 4 6 11 6 Fig. 2. Cross-section of the FANEYMONES excitation source: A, fixed parts and B, moveable parts; 1, closure with hole for sample support; 2, graphite tube and calhode for glow discharge; 3, sample support; 4, quartz window; 5 , anode for glow discharge; 6, power for tube heatitig; 7, graphite contact rings; 8, outlet to vacuum pump; 9, argor.inlet for inner Ar flow; 10, water cooling system; and 11, electrical insulator -4 .g 70 3 L- F .c 50 -P i 4- .- 30 al C c 0 c .- .; 10 .- I ( a ) 359.4 nm '1 5 t 268.94 nrn E L I I 359.0 360.0 268.0 269.0 w Wavelengthhm Fig. 3. Light emission of MgF molecules in glow discharges: T,,, . 2400 "C (MgF MONES); Mg2+ concentration, 2 pg per 10 p1; f-' concentration, A 0.1 pg per 10 pi and B 0.5 pg per 10 pi; ( u ) A band system and ( b ) B band system; broken line, non-specific background spectrometer can be used for simultaneom determination of two ions or for simultaneous background correction using the two lines (band) method. For routine simultaneous multi- elemental analysis, a multi-channel spectrometer is required.A cross-section of the FANEYMONES excitation source is shown in Fig. 2. Procedure The sample :lolution containing F- or C1- ions is mixed with a solution containing Mg2+ ions as the additive. The final concentration of Mg2+ should be higher than that of the X- ions in order to shift the equilibrium in the plasma. M g + X e M g X . . . . . . (1) A micro-volume (10-50 PI) of the solution is deposited in the graphite tube and the temperature - time - discharge pro- gramme is started: 1, drying phase; 2, first ashing phase at normal pressure; 3, evacuation phase (up to 5&100 Pa); 4, second ashing phase at low pressure; 5 , argon filling phase (1000-2000 Pa); 6, generation and stabilisation of the glow discharge (3-5 s); 7, evaporation phase (formation and excitation of the diatomic molecules); 8, light emission of the diatomic molecules and its measurement.The time for the cycle is 1-2 min, similar to that for ETA-AAS, as are the duration and the temperatures of the phases. The Zclration of the emission signal is relatively short, in the range 0.1-0.5 s. The intensity of emission is propor- tional to the concentration of diatomic molecules and hence to the concentration of X- ions in the original solution. 0 ' ' I -' 371.5 375.0 380.0 Wave1ength:nrn Fig. 4. 1900 "C (MgCI MONES); Mg2+ concentration, 3 pg per 10 pi; Cr-' concentration, 0.1 pg per 10 pl; A band system; broken line, non-specific background Light emission of MgCl molecules in glow discharges: T,,, u1 E 8 1 2 3 4 5 6 Mg2+ concentrationipg per 10 pl Fig.5. Influence of the amount of Mg2+ on the MgF MONES. Conditions: 359.4 nm; F-, 50 ng per 10 1.11; broken line, non-specific background Table 1. Optical properties of MgX m o l e ~ u l e s ~ ~ ~ 2 ~ Band Wavelength/ Excitation Molecule system Transition nm energyleV MgF . . . . A A2n+X2Z+ 368.6-346.8 3.5 B B2Z+ -+ X'Z' 274.2-263.0 4.6 C C'Z+ + X2Z+ 24Ch225 MgCl . . . . A A Z n j X 2 384-369 3.4 B B'X+X2C+ 273-266 4.6 Results and Discussion Optical Properties of MgX Molecules Table 1 shows the band systems for the MgX molecules. Using these known values, the spectra of MgX molecules were measured point by point with wavelength intervals (Ah) of 0.1 nm. The results are shown in Figs. 3 and 4.In Fig. 3, the band spectra of MgF for the A and B systems are shown. The band heads lie at 359.4 and 268.94 nm, respectively. These correspond to the 0 , O transitions. No results were found for the C system. For MgCl molecules, intense emission was found only for the A system (Fig. 4). The band head (0,O transition) lies at 377.9 nm. As the sample residue is evaporated at low pressure relatively low evaporation temper- atures are possible. These low temperatures are useful for quantitative molecule formation. No significant influence of the discharge on the molecule dissociation was observed, but no quantitative calculations were performed to establish the efficiency of molecule formation.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL.2 535 Optimisation of the Experimental Conditions Influence of Mgz+ concentration The influence of the Mg2+ concentration on the MgF MONES signal is shown in Fig. 5 . The mass of F- used in this experiment was 0.05 pg. The MgF MONES signal increased with the addition of up to 2 pg of Mg*+ to the sample, which corresponds to an ionic ratio of Mg2+ to F- of 30: 1 in the solution. It was impossible to calculate or estimate the ratio of the species in the plasma at the time of signal measurement, because of fractionated evaporation of the Mg- and F-contain- ing species. A further increase in the Mg*+ concentration did not lead to an additive enhancement of the molecular signal in contrast to MgF MAS.16 In ETE-MAS, 20-50 pg of Mg2f per sample volume gave the best analytical results.Similar results were obtained for solutions containing CI - (100 ng C1- per 10-pl sample volume). The optimum mass of Mg*+ for 100 ng of CI- was 3 pg. This corresponds to an ionic ratio in the solution for Mg2+ to C1- of 45 : 1. An increase in the mass of Mg2+ to 50 pg per 10 pl caused a large enhancement in the background emission signal for both halides. Influence of' thermal, guseous and electrical conditions The optimum conditions for the analytical determination of F- and C1- -concentrations are summarised in Table 2. As can be seen, none of the parameters have to be critically controlled. It is obviously necessary to avoid losses of Mg2+ or X- in the ashing phases, but such losses were found only at temperatures above 1200 and 1000 OC, respectively. Analytical Results for Pure Solutions Analytical results for the determination of trace amounts of F- and CI- are shown in Table 3 along with a comparison with our previous results for MgX MAS.16 The sample species were NaF and NaCI, and Mg(N03)* was the additive.The solutions were adjusted to pH 6. The detection limits were defined as the mass of analyte giving a signal equivalent to 0.25 mV in MONES and 0.01 A in AAS. From past experience, detection limits derived on these bases are similar to values obtained from 30 calculations where o is the standard deviation of repeated signals at concentrations close to the detection limits. Analytical Results for Fluoride and Chloride Determinations in the Presence of Other Halides Fig. 6 shows the dependencies of the MgF ( a ) and of the MgCl ( b ) MONES signals on the concentration of other halides. For MgF MONES there are two stages of interference.In the first interval of interferent concentration, for between 1O-I1 and 10-9 mol X-, it is clear that different halides have different interference effects, and these are in the order CI->Br->I-. This order corresponds to the dissociation energies of the molecules (Edi, MgF 4.7; MgC13.4; MgBr 3.0; and MgI 2.9 eV). The reason for this interference can probably be explained by MgXl +X,SMgX,+X, . . . . (2) In the second interval of interferent concentration (when the concentration of NaX is higher than 10-9 M) there are no differences between the interference of the different halides. In this instance the interference is probably caused by the equilibrium where M2 = Na.For MgCl MONES, the interference of fluoride added as NaF salt in the concentration range from 10-1' to 10-7 M is mainly caused by fluoride ions [as shown by equation (2)J. Fluoride forms the most stable MgX molecule. Because there is only a small difference between the NaBr and NaI interference, it can be concluded that the interference is mainly caused by Na+ ions [as shown by equation ( 3 ) ] . The results shown in Fig. 6 indicate that in the presence of other halides, interference free determination of fluoride and chloride by MgF- and MgCl- MONES is not possible. In spite of this fact it is possible to determine fluoride and chloride in the presence of other halides, when the interference concen- trations are known approximately.Table 4 gives the relative MgX, +M,=M,X, +Mg . . . . (3) Table 2. Optimum conditions for determination of fluoride and chloride ions by MgF and MgCl by ETE-MONES. Pyrolytically coated graphite tubes were used for all measurements Parameter MgF* MgCl Discharge current1mA . . 35 (25150) 30 (25135) Ar pressure1kPa . . . . 2 (1.312.8) 2 (1.312.8) Ashing temperature (10-100 Pa)/"C . . . . 1100 (90011200) 950 (90011000) Heatingrate/"Cs-1 . . . . > 2000 > 2000 Wavelength1nm . . . . 359.04 377.6 Evaporation temperature (2 kPa)/"C . . . . . . 2400 (230012450) 2100 (200012200) 268.44 * Values in parentheses are the conditions under which the MONES signal decreases to 90% of the maximum value. ~~ Table 3. Analytical results f o r MgX MONES compared with those from MgX Detection Ion Wavelength/ Excitation limit determined Molecule nm cncrgy1eV of X 1pg MONES- F- .. MgF 359.4 3.5 460 F - . . MgF 268.9 4.6 2400 CI . . MgCl 377.9 3.4 240 F - . . MgF 268.9 4.6 7500 CI- . . MgCl 377.9 3.4 9700 MAS- F - . . MgF 359.4 3.5 2300 Table 4. Analytical results for MgF and MgCl MONES in the prewnce of other halides Re la t ive detection limit, ofX in P.PJy Ion determined Matrix NaX F - . . . . . . . . NaCl 5000 F . . . . . . . . NaBr 1000 F-- . . . . . . . . NaI 600 C1 . . . . . . . . NaF 1400 CI . . . . . . . . NaBr 70 CI . . . . . . . . Nal 40 Q) U 0 - I I I L I 1 I I 10-10 10-9 10-8 10-7 10-10 10-9 1 0 - 8 10 7 Halide concentrationimol per 20 ul Interference of halide matrices on the intensity of ( a ) MgF Fig.6. and ( h ) MgCl ETE-MONES signals536 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, SEPTEMBER 1987, VOL. 2 detection limits for such determinations. The calculations of these detection limits are based on the following: concentra- tion of the interferent matrix at which the MgX MONES signal is depressed to 50%; and twice the measured detection limit (compare with Table 3). Conclusion Molecules of MgF and MgCl are formed at low pressure (2000 Pa) in hot graphite tubes (2000-3000°C) if both the metallic and non-metallic components are present. These molecules can be excited in the glow discharge of a FANES instrument and the intensity of the emitted light is proportional to the concentration of F- and C1-, if the metal concentration is high enough (an excess ratio of about 30 : 1).Comparing the results obtained with those previously found for F- and C1- determinations by InX MONES2” it can be seen that both types of molecules give similar possibilities for analytical applications. This means that the most appropriate metal component can be selected depending on the composi- tion of the sample. The MONES results with In or Mg are more than one order of magnitude more sensitive as the C1 FANES results obtained with our equipment (DL 8 ng C1-).26 So far we have not been able to detect an analytical signal for fluoride by FANES. As regards the interference found, it can be concluded that determinations of minor halides contents (10-1-10-2% F- and lO-*-lO-3% Cl-) in the presence of other halides are possible, but determinations of trace amounts of halides (<10-2O%) in the presence of other halides are not possible.Hence chemical separation procedures have to be combined with the MONES measurements, but these separation or enrichment methods ( e . g . , extraction or ion chromatography) do not need to be highly efficient, because the minor contents can be determined. A further major advantage of MONES measurements is the spectral selectivity between the halides and other elements. 1. 2. 3. 4. References Miller, W. A . , Philos. Mag., 1845, 27, 81. Salet, G., C.R. Acad. Sci., 1869, 68, 404. Salet, G., Bull. SOC. Chim. Fr., 1869, 11, 302. Belcher, R., Bogdanski, S. L., and Townshend, A., Anal. Chim. Actu, 1973, 67, 10. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.21. 22. 23. 24. 25. 26. 27. 28. 29. Belcher, R., Bogdanski, S. L., Kassir, L. M., Stiles, D. A . , and Townshend, A., Anal. Lett., 1974, 7, 751. Abdel-Kader, M. H. K., Peach, M. E., Ragab, M. H. T., and Stiles, D. A., Anal. Lett., 1979, 12, 1399. Osibanjo, O., and Aiyai, S. O., Anal. Chim. Acta, 1980, 120, 371. Abdel-Kader, M. H. K., Peach, M. E . , and Stiles, D. A . , J. Assoc. Off. Anal. Chem., 1979, 62, 114. Gutsche, B., Rudiger, K., and Herrmann, R . , Fresenius 2. Anal. Chem., 1977, 285, 103. Gutsche, B . , and Rudiger, K., Chromatogruphia, 1978, 11, 367. Yoshimura, E., Tanaka, Y., Tsunoda, K . , Toda, S., and Fuwa, K., Bunseki Kagaku, 1977, 26, 643. Tsunoda, K . , Fujiwara, K., and Fuwa, K., Anal. Chem., 1978, 50, 861. Dittrich, K., Anal. Chim. Acta, 1978, 97, 59 and 69. Dittrich, K., Anal. Chim. Acta, 1979, 111, 123. Dittrich, K., and Meister, P., Anal. Chim. Acta, 1980, 121,205. Dittrich, K., and Vorberg, B., Anal. Chim. Acta, 1982, 140, 237. Dittrich, K., Vorberg, B., Funk, J . , and Beyer, V., Spectro- chim. Acta, Part B, 1984, 39, 349. Dittrich, K., Prog. Anal. At. Specirosc., 1980. 3, 209. Dittrich, K., Crit. Rev. Anal. Chern., 1986, 26, 223. Falk, H., Hoffmann, E., and Ludke, Ch., Fresenius Z. Anal. Chem., 1981,307, 362. Falk, H . , Hoffmann, E., and Ludke, Ch., Spectrochim. A m , Part B, 1981, 36, 767. Dittrich, K., Hanisch, B . , and Stark, H. J., “Proceedings of the Second Hungaro-Italian Symposium on Spectrochemistry,” 1985, p. 361. Barrow, R. F., Glaser, D. V., and Zeemann, P. B., Proc. Phys. SOC. London, Sect. A, 1955, 67, 962. Barrow, R . F., Jaquest, J . A. T., and Thomson, E. W., Proc. Phys. SOC. London, Sect. A , 1954, 66, 528. Joffe, R. B., and Korovin, J. I., Zh. Prikl. Spektrosk., 1978.29, 197. Dittrich, K., Hanisch, B., and Stark, H. J., Fresenius Z. Anal. Chem., 1986,324, 497. Rosen, B., Editor, “International Tables of Selected Con- stants, Volume 7. Spectroscopic Data Relative to Diatomic Molecules,” Pergamon Press, Oxford, 1970, p. 250. Vorberg, B., Dissertation, Karl-Marx-University, 1981. Welz, B., “Fortschritte in der analytischen Atomspektrosko- pie,” Volume 11, Verlag Chemie, Weinheim, 1986, p. 97. Paper J7l7 Received January 13th, 1987 Accepted May I4th, I987