Durable emission of positronium negative ions from Na- and K-coated W(100) surfaces

The emission of positronium negative ions from Na- and K-coated W(100) surfaces has been studied. The emission efficiencies (the fractions of incident slow positrons yielding the ions) for both samples were found to be as high as 1.5%. Although the efficiencies decreased with time after coating, the effects were more durable than that for Cs coating. In particular, the efficiency after Na coating was still higher than 0.5% after three days. The successful development of a durable Ps− source opens the door to a new era of experimental investigations on the Ps− ion.

. First ionization energy and maximum reduction of by alkali-metal coating on W(100).
First ionization Maximum reduction of by alkali-Element energy (eV) [6] metal coating on W(100) (eV) [ observed [1][2][3][4]. The energy required for Ps − emission from a metal surface, φ Ps − , is given by [1] φ where µ + and µ − are the positron and electron chemical potentials, respectively, and is the effect of the surface dipole barrier. E B is the total binding energy of Ps − , an energy required to break up Ps − into its constituent particles. It is equal to the sum of the positronium binding energy (6.80 eV) and the electron binding energy to positronium (0.33 eV). The value of φ Ps − is negative for W(100) and tungsten polycrystalline surfaces, and thus the Ps − ions are emitted spontaneously.
Although the emission efficiency, the fraction of slow positrons yielding Ps − , was less than 0.01% [5], a dramatic enhancement to 1.25% was achieved by coating the W(100) sample with a monolayer of Cs. The enhancement was attributed to a reduction in the surface dipole barrier, which resulted in an increase in the number of conduction electrons contributing to the Ps − formation. However, the efficiency decreased with time and was almost constant at about 0.1% after 6 × 10 4 s. This may be because the Cs layer on the sample surface reacted with residual molecules in the sample chamber. In order to perform new experiments with Ps − , a higherefficiency and more stable Ps − source is required.
The first ionization energies and the maximum reductions of by alkali metal coating on W(100) are listed in table 1. For Cs, the ionization energy is the lowest and reaction with residual molecules in the chamber is the greatest. On the other hand, K and Na are chemically less reactive than Cs and thus the coating may be effective for longer, although the efficiency is not expected to be as high as that of the Cs coating [8]. In this work, we have studied the effects of Na and K coating on W(100) samples on Ps − emission. Figure 1 shows, schematically, the sample chamber used in the present experiment. The apparatus was the same as that used in previous work, where the Ps − emission efficiencies for Cs-coated surfaces were measured [1][2][3][4], with the exception of the addition of Na and K deposition systems and a quartz crystal deposition monitor system. The base pressure was 2-6 × 10 −8 Pa, which was comparable with that in previous works.

Experimental procedure
Two identical samples, A and B, were used. They were W(100) foils of 2 µm thickness, manufactured at Aarhus University, and were supported on a backing of polycrystalline tungsten foils of 25 µm thickness. First, they were annealed in situ at 1800 K for 30 min by the passage of electric currents. Layers of Na and K were produced using alkali-metal dispensers purchased from SAES Getter SpA. Before the deposition, the samples were heated in situ at 1800 K for 5 min in order to clean the surfaces. The thickness of the layer was estimated using the deposition monitor system. The uncertainty in the thickness of the layer was estimated to be about 0.6 Å. The sample was biased at a voltage V s so that the positrons impinged on the sample with an energy of −eV s + 0.1 keV, where e is the charge of the positron. The value of V s was set at −1 kV for sample A and −3 kV for sample B. A grounded mesh with 89% transmission was placed at a distance d in front of the sample. The distance d, which could be changed using a linear transfer, was set at 2 mm for sample A and 3 mm for sample B, respectively. The Ps − ions emitted from the samples were accelerated by the electric field between the sample and the mesh, and hence the γ -rays arising from the Ps − self-annihilation were blue-shifted. The γ -rays were monitored using the Ge detector in coincidence with signals from the NaI scintillation detector placed behind the sample.

Results and discussion
Figure 2(a) shows the annihilation photon energy spectra for sample A. The thicknesses of the K and Na layers were 3.3 × 10 14 atoms cm −2 (2.4 Å) and 6.2 × 10 14 atoms cm −2 (2.5 Å), respectively, where the electron work functions become the lowest. Figure 2(b) shows the spectra for sample B with a Na coating of 3.4 × 10 14 atoms cm −2 (1.3 Å). The spectra for the samples without coating were accumulated for 1.0 × 10 5 s after annealing. Those with coating were taken for 1.0 × 10 5 s after deposition. The energies of the Ps − self-annihilation γ -rays are indicated by the vertical arrows. Small peaks due to the Ps − self-annihilation were observed for the samples without coating. After coating, a dramatic increase in the peak intensity was observed for both alkali metals. The data on time dependence of Ps − emission efficiency for samples A and B are shown in figure 3, together with those for Cs-coated W(100) in the previous work [1]. The Ps − emission efficiency for sample A increased after K coating and reached a maximum value of 1.5%, which is higher than that for the Cs-coated sample. Then the value decreased and was almost constant at about 0.1% after 1.2 × 10 5 s. The change after K coating was more gradual than that after Cs coating.
The efficiency for sample A increased slightly after Na coating and reached the maximum value 1.5%, which is almost the same as that for K coating, after 4 × 10 4 s. For sample B, the efficiency increased slowly after Na coating and reached the maximum value 1.4% after 6 × 10 4 s. The efficiencies for sample B with Na coating are slightly lower on average than those for sample A with Na coating. This might be due to the difference in the thickness of the Na layers and the positron incident energies [9]. Also, the vacuum conditions, i.e. residual molecules in the chamber which could not be detected using our experimental system, might be different. However, the efficiencies for both samples were higher than 1.0% even after 1.2 × 10 5 s, when the efficiency for the K-coated sample dropped to 0.1%. This means that  W(100) with Na coating has a higher durability than W(100) with Cs or K coating. Although the change in is smaller for Na coating, the efficiencies for Na-coated W(100) are as high as those for K-coated surfaces.
In summary, the change in efficiency for the both samples coated with alkali metals of higher ionization energies was more gradual as expected. Surprisingly, the maximum values for K and Na coating were higher than those for Cs coating. We speculate that the maximum value for the Cs-coated surface might be even higher but the decrease was too rapid to observe under the experimental conditions. Although the reason for the unexpectedly high efficiencies for Ps − emission from alkali-metal-coated surfaces has been discussed in previous papers [1,2], the details are not known. Further investigations are necessary, both experimentally and theoretically.
The durability of Ps − emission from a Na-coated surface enables us to realize new experiments for the Ps − ion. Recently, we succeeded in the first observation of the photodetachment of Ps − produced using the present method [10]. An energy-tunable positronium beam has also been achieved by the Ps − photodetachment technique [11].

Conclusions
We found that the Ps − emission from Na-coated W(100) is as effective as and the decrease is more gradual than those for Cs-and K-coated surfaces. The present technique has been applied for the first observation of Ps − photodetachment [10]. It can be used for other Ps − experimental investigations, e.g. the observation of Feshbach resonances of Ps − photodetachment [12][13][14] and the measurement of the Ps − decay rate with higher precision than in previous measurements [15,16].