2. Experimental set-up
All measurements in this work have been made in a pulsed low-pressure discharge lamp that belongs to the experimental set-up depicted in Fig. 1. All information about the excitation unit that generates the discharge and its synchronization with the detection systems is provided by Gigosos et al. (1994). We will only present here the most significant general information and the additional specific details concerning this experiment.
The plasma has been generated by discharging a capacitor bank of 20 µF, charged up to 8.5 kV, on the electrodes of the lamp filled with a continuous laminar flow of a He+SiH4 mixture. The global pressure has been Pa at a helium rate of 11.5 ml min-1 and a silane rate of 0.075 ml min-1. The lamp consists essentially of a cylindrical Pyrex tube (150 mm long and 18 mm of inner diameter) which ends on two annular electrodes closed by two corresponding optical windows. Silicon evaporation from tube walls and from windows has not been detected.
The basic lamp constitution, described in detail by del Val et al. (1998), has been customized for this work to account for the particular features of the plasmogen gas. Contrary to working gases of other elements, silane plasma produces a great amount of waste (not only electrode sputtering) on very few discharges, even though its concentration in the gas mixture is low and the vacuum conditions in the system are well cared for. To minimize the effect of waste on the lamp windows we have designed a protecting chamber system put before each window. Each 30 mm long chamber, which is electrically isolated from the close electrode, has a complex internal structure that limits the amount of waste that reaches the window and projects a very narrow shielding gas curtain on its inner surface. The gas feeding of the lamp has been symmetrically carried out through both chamber mouthpieces and the gas exit through the electrode ones. With respect to the transmittance of the windows the protecting chamber system extends the experiment duration from tens to hundreds of discharges. Longitudinal pressure gradients have not been detected in this 210 mm long lamp and careful measurements reveal that it preserves a high axial homogeneity and a very good cylindrical symmetry of electron density and temperature. An extensive description of this system is given by González (1999).
The gas in the lamp has been continuously preionized between consecutive plasma pulses in order to assure the good repetitiveness of discharges. To minimize the appearance of waste due to silane not the whole gas has been preionized but only a local region in the vicinity of the lamp cathode. A new electrode has been located in the protecting chamber next to the cathode to get the local preionization (González 1999). A variable voltage source feeds this arrangement to make possible a fine-tuning of the current level that ionizes the gas without unnecessary waste production. The local preionization voltage used in this work has been 500 V. Temperature and pressure of the plasma precursor gas stay well known with this kind of preionization, and it is therefore possible to make an accurate determination of the initial particle density in the lamp.
Interferometric and spectroscopic end-on measurements have been simultaneously performed all over the plasma life on two parallel plasma columns of 3-mm diameter defined by pinholes depicted in Fig. 1. Both columns are placed 2 mm off the lamp axis and in symmetrical positions referred to it. The lamp is placed in one of the arms of a Twyman-Green interferometer simultaneously illuminated by a He-Ne (632.8 nm) and an Ar+ laser (488.0 nm). In one simply discharge is therefore possible to obtain the whole plasma refractivity changes due to only free electrons and, from them, the electron density temporal evolution. The spectra have been obtained by using a 1.5 m focal length Jobin-Yvon monochromator with a 1200 lines mm-1 holographic reflection grating, equipped with an optical multichannel analyser (OMA) and a 512 channel detector (EG&G 1455R-512-HQ). Mirror M3 in Fig. 1 is employed to detect self-absorption and to reconstruct the unabsorbed spectral profiles (González 1999) if necessary.
Careful relative and absolute intensity calibrations of the spectroscopic system in the first and second diffraction orders have been performed; exhaustive details are presented by González (1999). The first one provides a transmittance function with an uncertainty around 4% and it has been employed in temperature calculations by means of the Boltzmann-plot methods. The second one supplies a calibration function with an uncertainty about 10% which has been used in temperature determination via absolute emission intensity measurement method. A wavelength calibration of the spectrometer provides the dispersion for any wavelength and diffraction order with an uncertainty lower than 1%. The inverse linear dispersion at nm results 12.73 pm channel-1 in the first order of diffraction and 4.91 pm channel-1 in the second order.
© European Southern Observatory (ESO) 2000
Online publication: December 5, 2000