Seeding the Discharge

In an attempt to obtain discharge pulse lengths shorter than the gas pulse to improve the cooling of the molecular beam (Section 3.3.4), Lewandowski and co-workers introduced a tungsten filament close to the pulsed valve in their experimental chamber [208]. By applying a high current to the filament, the group was able to achieve a reliable and reproducible discharge for HV pulses as short as 1 μs. They also showed that the presence of the hot filament reduced the minimum voltage required to obtain a stable discharge at short HV pulses by a factor of ~ 4, which in turn produced a significantly colder molecular packet. This method of stabilizing the discharge has been

adopted by other groups using short discharge pulse lengths in their experiments [198, 209-210].

Conventional wisdom would therefore suggest that a source of electrons, such as a filament, would need to be added to our experimental chamber for the discharge to work at short HV pulses. During the first attempts at running scattering experiments with short HV pulses, however, the discharge was found to still strike reliably even at HV pulses lower than 60 μs without the need for a hot filament. Through later optimization experiments involving the gas baffle (discussed in section 3.6.2), it was eventually discovered that the short discharge pulses were being stabilized by the wide range pressure gauges (Edwards WRG-S-DN40CF) mounted on the main chamber. These gauges are Penning-type gauges, also known as cold-cathode ion gauges. The cold-cathode gauges generate electrons between the cathode and the anode of the gauge. The electrons then ionize gas molecules between the two electrodes, which are then accelerated to the grounded cathodes. The pressure is then measured based on the current required to neutralize the ionized gas molecules during each time unit. As both the cathode and anode of the gauges are directly exposed to the vacuum chamber, it is possible for the generated electrons to escape the gauge and enter the vacuum chamber, where they can then access the discharge device, and thus stabilize the discharge strike. This explanation was confirmed by checking the discharge stability after disconnecting the gauge, as shown in Figure 3.36 (a) when comparing with Figure 3.36 (c) and (d). Further testing showed that having two pressure gauges on opposite hemispheres of the main chamber (as shown in Figure 3.37) was optimal for obtaining a stable discharge. This indicates that the discharge device can potentially be stabilized by a pressure gauge on the opposite hemisphere of the main chamber, provided there are no obstructions in between, such as a large gas baffle (described in section 3.6.2). It also suggests that only one pressure gauge, placed close to the discharge device, might be sufficient to stabilize the discharge, though it may require further adjustment of the other conditions.

Figure 3.36: Representative persistence plots of the discharge stability using Ne as the carrier gas under optimal in situ conditions, with pressure gauges on or off, or with the

bias voltage present or absent, as described in the captions under each plot. ‘Close gauge’ position is shown in Figure 3.37 below

Figure 3.37: Photo of main chamber in the lab, with pressure gauges position highlighted with respect to the valve and discharge device port, set at 45° angle of

The presence of one or more wide range gauges ensured that the discharge would strike reliably at short HV pulses. However, on their own, the gauges were not enough to guarantee a reproducible discharge over the entire HV pulse under all experimental conditions tested, and especially when using Ne as the carrier gas (as shown, for example, in Figure 3.6). In order to obtain a reproducible discharge that strikes at the start of the HV pulse, two methods were employed to further stabilise the discharge

in situ.

Figure 3.38: Representative persistence plots of the discharge conditions of a He beam, showing stabilizing effect of using the filament (right column) compared to no filament (left column). The discharge faceplates used are the no channel faceplate (top row) and

the 6 mm channel faceplate (bottom row)

In the first instance, a hot filament was inserted in the main chamber close to the discharge device, as per the approach pioneered by Lewandowski et al. [208]. The current applied to the filament was set at between 0.4 and 1.4 A, adjusting it throughout the course of the day and also between days to obtain a stable discharge. This method was used primarily for the single-point detection set-up reported in this thesis. The stabilizing effect of the hot filament was tested for discharge devices with different faceplate designs (Section 3.4.1). A comparison of the persistence plots (Figure 3.38)

shows that the hot filament is very effective at stabilizing the discharge when a faceplate with no channel is attached to the discharge device (Figure 3.38 (a) and (b)). When a faceplate with a channel is attached to the discharge device, however, no improvement in the discharge conditions is observed. (Figure 3.38 (c) and (d)). This was thought to be due to the constrained line of sight to the electrodes that results from using the faceplates with the channel.

There were also concerns that the light produced from the glowing filament when a high current is applied would also be picked up by the optical detectors used to capture the LIF signals. This was not a problem for the more conventional PMT experiments reported in Chapter 4, as the gas baffle used was large enough to block most of the light going into the detection region of the chamber, and no additional baseline signal was observed. However, it became more of a concern when switching to the image acquisition set up, as the gas baffle used in these experiments was smaller. In this case, it was found that the discharge could also be stabilized by biasing the front electrode with a positive voltage of up to 230V, provided by a second power supply that the fast switch switches to when the high negative pulse is not active, to attract the electrons produced by the gauges to the discharge device. This effect is best shown in Figure 3.36 (b) comparing with Figure 3.36 (d), for an unoptimized molecular beam with neon as the carrier gas.