Artai’yst, May, 1 9 6 8 , Vol. 93, +p. 281-285 28 1 Preparation of Microwave-excited Electrodeless Discharge Tubes for Titanium, Vanadium and Zirconium for Use as Spectral-line Sources BY R. M. DAGNALL, R. PRIBIL, JUN.* AND T. S. WEST (Chemistry Department, Imperial College, London, S . W.7) The preparation of microwave-excited electrodeless discharge tubes as spectral-line sources for the atomic-absorption and atomic-fluorescence spectroscopy of titanium, vanadium and zirconium is described. A long, cylindrical, resonant cavity has been found particularly useful for elements such as these. IN a previous communication,l a general method for the preparation of electrodeless discharge tubes for selenium and tellurium was described, and their use as spectral-line sources was evaluated for atomic-absorption and atomic-fluorescence spectroscopy.2 Since then, several other elements have been studied in detail, viz., antimonyJ3 bismuth,4 arsenics and tin.* Source tubes for many other elements have also been similarly prepared.However, there are several elements, e.g., titanium, vanadium and zirconium, and their compounds, which do not possess the vapour pressure requirements of about 1 mm of mercury at 200” to 300” C to fit into this general scheme. In such instances a different approach must be made to obtain a stable, intense and long-lived narrow-line spectral source. These elements also require special attention because of the large number of atomic lines in their spectra. A microwave-excitation cavity different from that normally used was also investigated in this study, and it was found to offer some advantages, both for these and some other elements.EXPERIMENTAL PREPARATION OF DISCHARGE TUBES- The tube vessels were prepared in essentially the same manner as that previously described.1 The principal differences, viz., materials used, their amount, tube length and operating conditions are given below for individual elements. SPECTRAL EVALUATION- A Hilger and Watts quartz spectrograph E742 was used for recording and assigning the radiation emitted from each tube. Records were obtained on Kodak B10 plates. A Unicam SPSOOA flame spectrometer, with an E.M.I. 9529B photomultiplier, was used for recording spectra and monitoring the stability at the principal resonance lines. DISCHARGE-TUBE EXCITATION- Excitation was achieved by using an Electro-Medical Supplies Limited “Microtron 200” high frequency generator of output 2450 25 MHz (up to 200 watts) connected to a resonant cavity.Cavity Model No. 214L-Obtained from Electro-Medical Supplies Limited. This cavity was used in all of our former studies and has been adequately described and illustrated elsewhere. This cavity is cylindrical in shape (about 5 cm in diameter and 13 cm long) , has an adjustable re-entrant gap, co-axial with the discharge tube, and a tuning probe offset, and parallel to the discharge tube (cf. Fig. 1). The cavity may be air-cooled and also possesses adequate viewing apertures, either at right angles to, or in line with, the discharge tube. The cavity can be used in any position and is readily interchangeable with the 214L cavity.Initiation was obtained with a “Tesla” vacuum tester. Cavity Model No. 21OL-Obtained from Electro-Medical Supplies Limited. * Present address : Jaroslav Heyrovskf Polarography Institute, C.S. A.V., Prague, Czechoslovakia. 0 SAC and the authors.282 DAGNALL, PRIBIL AND WEST : MICROWAVE-EXCITED ELECTRODELESS [A%W&St, VOl. 93 Type ‘C’ connector U I” Air inlet (Reproducsd by cwrtesy of Ekctro-Medical Supplies Limited, London.) Fig. 1. Gas-discharge cavity No. 210 L It was found that the discharge tubes could be operated in almost any position in this cavity and that the position of the tuning probe made little difference to the intensity of the discharges. Even when it was removed completely the tubes remained alight.This is in marked contrast to the critical nature of the tuning in the 214L cavity. In addition, the discharge tubes are well screened from draughts, which normally alter spectral characteristics. A further factor that adds to the usefulness of the 210L cavity, and which is absent in the 214L cavity, is the almost total “immersion” of the tube (up to about 7 cm in length) in the microwave field of excitation. Hence, a tube prepared from relatively involatile or very volatile materials is able to reach equilibrium conditions because, no matter how much movement of material takes place within the tube, it is always within the microwave field. In addition, the cylindrical cavity acts as an effective thermostat for the tube. The only disadvantage of the 210L cavity is that it is rather less efficient than the 214L cavity, and about 25 per cent.more power is required to initiate most tubes. In general, discharge tubes for most elements would only operate with the 210L cavity if they could be started at powers below about 30 to 40 watts with the 214L cavity. The intensities of the various discharges examined were similar in both cavities under optimum operating conditions. The best operating conditions for each element are given below. TITANIUM- Titanium discharge tubes were made by introducing about 5mg of titanium metal (as crushed sponge) and about 10 mg of iodine into a quartz tube about 7 to 8 cm in length. The tube was then filled with argon at atmospheric pressure from a cylinder, and the reaction between titanium and iodine was promoted by heating the tube gently with a bunsen burner, while keeping the top of the tube cool with moist asbestos string.This stage reduces the “running-in” period of the tube, and the very volatile titanium tetra-iodide (boiling-point 377.1” C) is deposited on the cool upper portions. The tube was then re-evacuated, flushed with argon and sealed under a pressure of argon of about 2 to 4 mm of mercury. These tubes still required a “running-in” period of between 1 and 2 hours when first prepared and only about 10 to 20 minutes afterwards. In most tubes a considerable amount of the less volatile black titanium &-iodide was observed after some time. We have not, as yet, established the lifetime of these tubes. The tubes thus prepared were examined in both types of cavity.The optimum excitation powers were about 35 watts when using the 214L cavity and about 50 watts when using the 210L cavity. Under these conditions, and in the absence of draughts, a stable, intense source showing no background emission from argon or iodine (except for the 206-2 nm line) was obtained. The stability when the 210L cavity was used showed less than +2 per cent. variation in response, while with the 214L cavity it was about +3 per cent. Also, the intensity of the discharge was slightly greater with the 210L cavity. The colour of the discharge was blue, and a spectrographic examination showed that it contained all of the expected atomic lines for titanium. The Unicam SPSOOA spectrometer is not capable of resolving all of the atomic lines, but the relative intensities of those linesMay, 19681 DISCHARGE TUBES FOR SPECTRAL-LINE SOURCES 283 which could be resolved were similar to those obtained in the arc, rather than the spark, spectrum.* Table I shows the relative intensities of the major lines observed when the Unicam SPSOOA flame spectrometer is used.TABLE I TITANIUM TUBE SPECTRUM Wavelength, nm 319.991- 334.19 336.461- 337.15 363.661- 364.277 366.361- 372-98 374.01 Relative intensity* (recorder reading) 16 16 15 25 50 70 73 60 66 66 50 396.63$ 90 396*29$ 396.43$ 85 398-25s 398.98 90 399.867 100 Tube operated at 50 watts in a 210L cavity; slit width, 0.007 mm; * Uncorrected for detector/monochromator response characteristics. t Spectral lines used for atomic-absorption studies.9 1 Unresolved lines with Unicam SPSOOA.gain, 2,O; and placed 25 cm from monochromator slit. VANADIUM- Attempts to prepare vanadium tubes in the way usually recommended failed and they could only be prepared by using commercially available vanadium trichloride. About 10 mg of vanadium trichloride were introduced into a quartz tube (about 6 to 7 cm in length) and, after flushing with argon, the tube was heated gently with a bunsen burner. When the first yellow fumes of free chlorine appeared, the heating was stopped, the tube was re-evacuated and then allowed to cool under vacuum. The tube was finally sealed under a pressure of argon of between 7 and 10mm of mercury. As usual, these tubes required to be “run-in” for about 1 hour when first prepared.Although most of the material remaining in the tube is vanadium trichloride, a small amount of vanadium dichloride (apple-green, boiling-point 1377” C) is also produced. The colour of the discharge was pink - violet, and the cylindrical 210L cavity is recom- mended for best operation at about 40 watts, without cooling. The success of the 210L cavity is once more thought to be because of its ability to act more efficiently as a thermostat for the tube than the 214L cavity, which yielded only an unstable discharge. Even with the 210L cavity, it is advisable to shield it from external draughts. Table I1 shows the principal lines observed with a Unicam SP9OOA spectrometer and their relative intensities. A spectrographic examination confirmed that the spectrum was primarily caused by atomic vanadium.Under these conditions a stable discharge was obtained. ZIRCONIUM- Zirconium tubes were prepared by using about 5mg of high purity zirconium metal and about 10 mg of iodine in a manner similar to that for titanium. In this instance, heating the metal and iodine under an atmosphere of argon to produce the moderately volatile zirconium tetra-iodide (vapour pressure 1 mm of mercury at 264’ C) was found to be essential because, otherwise, only a purple iodine discharge was obtained. The optimum pressure of argon as carrier gas corresponded to between 3 and 5 mm of mercury.284 DAGNALL, P ~ I B I L AND WEST : MICROWAVE-EXCITED ELECTRODELESS [Analyst, Vol. 93 TABLE I1 VANADIUM TUBE SPECTRUM Wavelength, Relative intensity * nm (recorder reading) 306.641- 16 100 26 20 26 40 66 45 3 18-34? $ 3 18*40t$ 318-64t 370.367 870.47t 370-60t 383*90t 386*64t$ 386*68t$ } 411.18 437.92 438-47t 439-00t 46 448.89 16 487.66 36 i Tube operated at 40 watts in 210L cavity; slit width, 0.012mm; * Uncorrected for detector/monochromator response characteristics.t Spectral lines used for atomic-absorption studies.* $ Unresolved lines with Unicam SP9OOA. gain, 2,O; and placed 25 cm from monochromator slit. The zirconium discharge was light blue in colour and was best obtained in the 214L cavity, at about 40 watts, without cooling. The tubes tended to overheat in the cylindrical cavity, with consequent instabiiity. The stability of these tubes at 40 watts was good, as is to be expected with an element or compound with a vapour pressure of 1 mm of mercury at about 200" to 300" C.In this instance the principal resonance lines at 318.34, 318-40 and 318.504 nm were not resolved by the Unicam SPSOOA spectrometer (Table 111), but no background emission was observed. TABLE 111 ZIBCONIUM TUBE SPECTRUM Wavelength, Relative intensity* nm (recorder readings) 208-64t 16 301.17t 36 302.967 20 339.20 30 360*98t 6 351*96? 3 364.77t 30 360-12t 100 362*39i 90 386.39t 40 389.03t 80 Tube operated a t 40 watts in 214L cavity; slit width, 0408mm; * Uncorrected for detector/monochromator response characteristics. t Spectral lines used for atomic-absorption studiese gain, 2,O; and placed 25 cm from monochromator slit. DISCUSSION No absolute intensity measurements or comparative measurements were made with hollow-cathode lamps, but we have no reason to suspect that these particular sources are any less intense than those previously described.For example, under operating conditions suitable for atomic absorption (i.e., operating powers of about 30 watts), we have found seleaium, antimony and arsenic electrodeless discharge tubes to be between 20 and 800 times more intense than conventional hollow-cathode lamps. Further, an increase in the operating powerMay, 19681 DISCHARGE TUBES FOR SPECTRAL-LINE SOURCES TABLE IV EFFECT OF POWER ON INTENSITY OF SOURCES Intensity (recorder reading) 285 I Power, * Titanium Vanadium? Zirconium watts (363-55nm) (318.4nm) (360.1 2nm) 20 15 0 0 30 25 18 2 40 32 28 5 50 65 36 32 60 > 100 68 50 70 > 100 83 82 Titanium tube with slit width, 0.005 mm and gain, 2,O; vanadium tube with slit width, 0.018 mm and gain, 2,O; zirconium tube with slit width, 0.004 mm and gain, 2,O.* Uncorrected for reflected power loss. t The 318.34 and 318-52 nm lines were unresolved from this line. also results in further increases in intensity. Table IV shows such a dependence for titanium, vanadium and zirconium at the lines normally used for atomic absorpti~n.~ Again, we have not established the lifetime of these sources, but many titanium, vanadium and zirconium sources have been prepared, and most of them have been run for up to about 50 hours without failing or markedly changing their operating conditions or intensity. Microwave-excited electrodeless discharge tubes are simple and inexpensive to prepare, and they do not require sophisticated or expensive operating equipment.The tubes, when correctly prepared, show no background-emission spectrum from the canier gas, but only the atomic spectrum of the metal concerned. The emission is normally as stable as that from a conventional hollow-cathode lamp and can be 2 to 3 orders of magnitude more intense. This is an important consideration in atomic-absorption spectroscopy because it overcomes many problems associated with the hollow-cathode lamp that has been used almost exclusively in this technique. In some instances, e.g., iron hollow-cathode lamps, a non-absorbing line is close to the resonance line, and the subsequent need for narrow slit widths results in a general loss in sensitivity and a prematurely curved calibration graph.This problem has been overcome by the use of a microwave-excited electrodeless discharge tube for iron.1° A similar situation also arises when an ionised metal or rare-gas line lies close to the resonance line, e.g., nickel and cobalt. In other instances, hollow-cathode lamps must sometimes be operated at low currents to avoid a loss in sensitivity caused by self-absorption effects. Yet again, some elements that can be detected in high temperature flames, such as nitrous oxide - acetylene, give only weak emissions in hollow-cathode lamps, e g . , silicon and boron. The use of resonance detectors also requires an intense source to eliminate noise at the amplifier output. While the recent introduction of high intensity hollow-cathode lamps is directed at solving these problems,ll their cost is high and little is known of their lifetime.Electrodeless discharge tubes prepared in this laboratory have been used following storage periods of more than 1 year, without any noticeable variation in output or the necessary operating conditions. These properties thus fulfil the essential requirements for spectral- line sources in both atomic-absorption and atomic-fluorescence spectroscopy. We are grateful to the Science Research Council for the award of a research grant in aid of this work, to I.M.I. Ltd. for the gft of pure titanium and zirconium metal, and to Electro-Medical Supplies Limited, London, for permission to reproduce Fig. 1. One of us (R.P.) also thanks the Czechoslovak Academy of Sciences, Prague, for study leave. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. REFERENCES Dagnall, R. M., Thompson, K. C., and West, T. S., Talanta, 1967, 14, 551. 3 8 , Ibid., 1967, 14, 1151. --- , Ibid., 1967, 14, 1467. --- , Atomic Absorption Newslefter, 1967, 6, 117. We$, T. S:, Endeavour, 1967, 26, No. 97, p. 44. Brode, W. R., “Chemical Spectroscopy,” Second Edition, J. Wiley and.Sons Inc., New York, 1943. Amos, M. D., and Willis, J . B., Spectrochim. Acta, 1966, 22, 1325. Marshall, G., and West, T. S., Talanta, 1967, 14, 823. Sullivan, J. V., and Walsh, A., Spectrochim. Acta, 1965, 21, 721. Received October 12th, 1967 3 , Ibid., 1967, 14, 667. , I W . , in the press. --- --- --- ,