Status of the Experiment

Figure 5: Layout of the Polarized Cold Neutron Facility at PSI.

The cold neutron flux density of unpolarized neutrons, measured at the border of the SINQ shielding is around 10⁹ (cm^{2 .} s ^. mA)^-1; the "thermal equivalent flux" would be 3 ^. 10⁹ (cm^{2 .} s ^. mA)^-1. The total number of unpolarized neutrons with the characteristic cold spectrum exceeds 10¹¹ (s ^. mA)^-1. The commissioning data measured at the location of the experiment show a flux of polarized cold neutrons of 2 ^. 10⁸ (cm^{2 .} s ^. mA)^-1. The characteristic cold neutron spectrum and the wavelength dependence of the beam polarization was then determined (Fig. 6) using the time-of-flight technique, a polarization analyzer and a combination of high efficiency spin flippers [18]. The total intensity of the polarized neutron beam exceeds 10¹⁰ (s ^. mA)^-1. At present the SINQ source operates routinely at 1.2 mA proton beam. The average polarization in the central part of the beam exceeds 97% (Fig. 6).

Figure 6: (a) Neutron wavelength spectrum measured using the TOF technique and a thin ³He transmission detector. (b) Wavelength dependence of the beam polarization.

Detailed investigations of the beam properties, i.e. the flux and polarization distributions and divergence, are in progress. Fig. 7 presents the first results from the measurements of the horizontal beam profiles and the computer representation of the beam intensity for the polarized, compressed beam at the experimental station.

Figure 7: Beam profiles (right) and a representation of the beam intensity obtained by interpolation (left). Note different scales for "z" and "x" directions, which indicate the distance from the beam exit and the horizontal position across the beam, respectively.

Prototype Tracking Detector

The requirements imposed by the conditions described in previous sections can be summarized as follows: the detector should have low mass and should be constructed of low Z materials. This leads to the concept of a gas detector with all electrodes of thin wire grids since foil cathode planes would lead to unacceptably large multiple scattering. Accordingly, the gas mixture should be based on helium. Our prototype multiwire proportional chamber (MWPC) is 20 x 20 cm² large with an active area of 13 x 13 cm². The chamber has a modular construction which allows for detailed investigation of the cell geometry and easy extension of the detector by adding more electrodes.

The results of the laboratory investigations of the prototype MWPC are described in detail in Ref.  [19]. The experience made by testing this detector in the real environment influenced by neutron beam led to a slight modification of the cell geometry. In the following, the main features of the newest version of the prototype MWPC are briefly summarized.

Each anode plane consists of 28 wires with 5 mm spacing. Cathode planes have wires spacing of 2.5 mm. These particular values have been chosen as an optimum for the moderate spatial resolution required by the experiment, which is limited by multiple small angle scattering of the electrons in the gas.

The calculated RMS angle , describing the multiple scattering, varies between 3^o and 15^o for electron energies between 250 and 750 keV, 10 cm path length and 5 ÷ 15% admixture of Methylal vapor into helium gas. The spacing and thickness of the wires affect also the overall (optical) transparency of the chamber. Also, capture of stray neutrons on the wires should be avoided to prevent delayed -activities. We have found that 25 m Ni-Cr alloy (80-20%) wires, for anode and cathode planes, are optimal for our purposes, providing an optical transparency above 85% for a detector with 6 anode and 12 cathode planes. The electrode frames and distancers are made of standard glass fiber, screwed together and sealed up by flat rubber gaskets. Two kinds of cathode planes were built. The "active" cathodes allow for a readout of charge from individual wires, while the "inactive" ones are used only for the creation of the electrostatic field in the chamber.

Electronics and Data Readout System

The electronic system of the detector consists of several stages. These include pulse amplification and shaping, discrimination, analog multiplexing and digitization. All the electronics up to the digitization stage was designed and customproduced. This has allowed for easy optimization and significant reduction of costs. The developed electronics proved to be simple, versatile and robust. It behaved stable not only in the laboratory but also during lengthy tests with the neutron beam. Below, only the most important stages are described.

Fast trigger

The experimental conditions characterized by the presence of background radiation, consisting of scattered neutrons, secondary -rays and electrons, require inclusion of information from the wire chambers in the event selection process.

We note that in the neutron decay experiment the electrons scattered from the analyzer foil will cross a set of two multi-plane wire chambers placed on the opposite sides of the neutron beam. Thus, the plane multiplicity for the most interesting events, which exhibit a scattering vertex signature, will exceed by the factor of two the plane multiplicity for the most frequent background events, where the charged particle crosses only one chamber ¹.

The fast electronic modules for on-line event selection have been constructed using ECL logic chips. Fig. 8 shows the logic diagram. The gate signal is generated by the scintillator hodoscope detectors (Fig. 3). Simultaneously, the analog sum of the two photomultiplier signals from the same scintillator slab, is directed into the ADC to measure the electron energy. The plane hit signals, from the anode planes, yield an analog sum in the plane multiplicity unit with the amplitude proportional to the plane multiplicity. The desired minimal multiplicity is selected by comparing this signal with an external threshold voltage. The gate generated by the plastic scintillator detectors provides timing reference for the TDC common stop signal.

Using the AND and OR logic of the two chambers responses, the fast trigger circuit recognizes within about 30 ns two kinds of events: the candidates for "V-tracks" (scattering in the analyzing foil took place) and single track events in one of the two chambers.

Extensive laboratory test and test run with the neutron beam have shown that the fast trigger system performs very well and can be extended and implemented in the ultimate detector for the experiment.

Figure 8: Logic diagram of the fast trigger system. The Master Gate signal is generated by the Plane Multiplicity strobed with the Scintillator Gate.

Figure 9: Multiplexer based on the mono-flop delays and its timing graphs. The leading edge timing information of each signal is conserved.

Multiplexer

The experiment should operate with about 2500 sense wires. At the expected counting rates of few hundred events per second, allocation of separate digitizing channels to all sense wires would be impractical and expensive. Therefore, we have developed a multiplexing scheme which exploits the multi-hit capability of the LeCroy 3377 TDC module. Several comparator output signals are delayed by multiple of 230 ns and the whole group of wires is plugged into a single TDC channel. As a result, a separate time window of 200 ns length is allocated to each sense wire. For a given group of wires these windows cover the TDC conversion range and do not overlap.

The implementation of the multiplexer is shown in Fig. 9. The delays are generated by two mono-flops connected in series. The first mono-flop, triggered by the leading edge of the wire comparator output, generates a pulse with a length equal to the desired delay time. The second pulse is triggered by the trailing edge of the former pulse and generates a standard signal, which enters the OR gate. Thereby, wire signals reach the start input of the TDC in a well defined sequence to be processed in a multi-hit mode of the TDC. It is valuable to note that complete timing information within 200 ns after the trigger is preserved. No additional dead-time in the readout system is introduced by the presence of multiple hits at the TDC input. Further advantages of such mono-flop solution are the flexibility in setting of the delay time and a low cost. The multiplexer exhibits a very good stability of the delays and a negligible time jitter (less than 1 ns). No increase of noise and no lost hits were found in the laboratory tests.

In this set-up, the LeCroy 3377 multi-hit TDC module plays a very important role in the chamber readout and digitizing system. It serves as a component of the multiplexer, as a latch for the hit pattern and as a digitizer for the pulse amplitude [19].

Performance of the Prototype Detector

The gas mixtures used in our studies are based on helium. Due to the low primary ionization probability in helium and the small volume of the sensitive cell in our chamber, the efficiency is primarily determined by the organic component. In most of the laboratory experiments we used Methylal CH₂(OCH₃)₂ which provides a high primary ionization. The quantity of Methylal was varied between 5% and 40% and effects were investigated separately for different wire thicknesses and cell geometry. In addition, small quantities of isobutane, which extends the efficiency plateau, were added. The optimal gas mixture besides securing the stable operation of the detector, should also minimize the energy loss and the effects of multiple scattering. For electrons in the energy range 250÷780 keV, multiple scattering angles up to 15^ofor 10 cm long tracks are acceptable in the proposed experiment. This can be achieved by minimizing the amount of organic components of the gas mixture.

For our preferred cell geometry (25 m thick Ni/Cr wires, 5 mm anode wire spacing, 2.5 mm cathode wire spacing, 4 mm distance between anodes and cathodes), we achieved very good results for helium mixed with 5% of Methylal and 5% of isobutane. The corresponding efficiency curve is shown in Fig. 10. This gas mixture guarantees a stable operation of the chamber in a high counting rate environment which was proved in the test measurement with the neutron beam.

Figure 10: Relative efficiency function (absolute efficiency is about 2÷3% lower) of a single anode plane measured for two gas mixtures: 5% of Methylal, 5% of isobutane, 90% of helium and 10% of Methylal, 5% of isobutane, 85% of helium. The cell geometry is explained in the text.

A very important performance parameter of the chamber is the spatial resolution of the reconstructed scattering vertex. Apart from the obvious geometrical relations, as distances between the wires and planes, the resolution of the vertex position is affected by the size of the hit clusters. We have developed a method of separation of the overlapping hit clusters using the timing information carried by the wire pulses. The details are described in Ref. [19].

Fig. 11 presents the results of the vertex reconstruction accuracy for two positions of the scattering foil. This test was performed with the ⁹⁰Sr -source. Events were selected according to the requirements of the R-coefficient measurement, e.g. appropriate range of scattering angles, electron energy, cluster and plane multiplicities. Due to the lever arm effect, for the incoming and outgoing tracks, the width of the distribution of the reconstructed vertices increases with the distance of the scattering foil from the chamber. To minimize the multiple scattering of the electrons in the gas a small distance is recommended. Placing the foil at about 20 mm from the last electrode gives a very good separation of events scattered in the foil from those created by the presence of other scatterers (chamber wires, scintillator hodoscope etc.).

Figure 11: Reconstructed vertices of the electron scattering events according to the requirements of the planned neutron decay experiment: a) anode projection, b) cathode projection.

Identification of Electrons from Neutron Decay

The necessity of tracking the low energy electrons from neutron decay imposes the requirement that the detector window must be very thin. This, in turn, forces the neutron beam, passing in front of the tracking detector, to be transported in gas. Transporting the beam in gas causes neutron scattering and capture reactions. Both these processes are source of severe electromagnetic background since stray neutrons will ultimately be captured too. The optimal gas with respect to the above problem is pure helium, where the capture cross section vanishes and the scattering cross section is sufficiently low. Nevertheless, some neutrons will be scattered off the beam (about 0.3% per 1 m of beam length) and the optimization of the experimental environment (choice of materials and geometry) is required to prevent them from inducing background.

The best customary available material known to capture slow neutrons without producing electromagnetic radiation is LiF based polymer with lithium enriched to 90% of ⁶Li [20]. Such a material produces only 3 photons per 10⁴ captured slow neutrons. In our test experiments we had to our disposal a very limited amount of this effective shielding material. It was therefore only used in critical places. Generally, the shielding against stray neutrons was made with the material containing natural boron. However, the main component of the radiation following the capture of slow neutrons in such a material are 480 keV -rays produced at a rate of about 1 photon per 1 captured neutron.

We have performed a series of test measurements aiming at the determination of the real background conditions and identification of electrons originating from neutron decay. A sketch of the experimental setup is shown in Fig.  12. The neutron beam was collimated and transported in a box made of organic plates containing natural boron ("borated epoxy plates") and filled with pure helium. The thickness of the walls was chosen as to prevent stray neutrons to penetrate into the air. Finally, the neutron beam was stopped in a 2.4 mm thick ⁶LiF polymer layer.

The box was equipped with 12 m thick mylar windows which allow the low energy electrons to be detected outside the box. The electrons were detected by our prototype MWPC backed with three plastic scintillator detectors, each 15 x 5 x 1 cm³ large. The trigger logic selected the events characterized by at least one signal in the plastic detector and the plane multiplicities n_y=3, n_z=3 in both projections respectively. We varied the vertical position of the detector with respect to the beam. In the following, we discuss the spectra of events projected onto the vertical symmetry plane of the beam along the reconstructed direction.

Figure 12: A sketch of the experimental setup.

Figure 13: Distribution of the intersection points of the reconstructed electron tracks and the vertical symmetry plane of the beam and selected slice projection spectra (A,B,C,D). The z-axis is parallel to the beam direction, while the y-axis is vertical. The geometrical situation is shown in the inset on the left. The arrows indicate the main directions of observation associated with the given slice.

Fig.13 corresponds to the measurement where the detector was placed 10 cm below the horizontal symmetry plane of the beam. Only the lower part of the beam profile can be observed in the slice projection A. The slices B and C represent the distributions of the background radiation. The rate increase towards the sides is due to the fact that the detector "sees" the boron epoxy plate instead of the ⁶LiF layer which is placed only in the middle. This interpretation is confirmed by the inspection of the energy spectra in Fig. 14. In slice A, corresponding to the situation where the neutron beam was in the foreground, one clearly observes the characteristic -spectrum of neutron decay superimposed on the two-component background. The low energy component can be attributed to boron. The magnitude of this component varies depending on whether the direction of observation has mainly in the background borated epoxy plate (slice C) or mainly ⁶LiF (slice B).

Fig. 15 contains the result of the symmetrical arrangement: the axis of the detector coincides with the horizontal symmetry plane of the beam. Also here the electrons from neutron decay are well seen (slice A). Projection D reflects the vertical profile of the beam superimposed on the wider background distribution, while the energy spectrum is seen in the projection B. We note that the correction for the angular acceptance does not change these conclusions significantly. As expected, the characteristic -decay distribution disappears from the energy spectra (Fig. 16) for directions not pointing to the neutron beam.

Figure 14: Energy distributions for detected electrons with cuts on tracks originating from various regions. While slice A represents the energy distribution for the direction of observation of the neutron beam, slices B and C show mainly background. (More information in the text).

Figure 15: Same as Fig. 14 but for the detector placed at the same height as the neutron beam.

Figure 16: Same as Fig. 14 but for the detector lowered by 19 cm with respect to the neutron beam.

Figure 17: Result of the background subtraction procedure.Upper part compares the appropriately normalized spectra from Figs.  14A and 14B. The result of subtraction of these two spectra is shown in the lower part of the figure. The signal-to-background ratio is found to be about 2:1. The solid line represents the theoretical energy distribution of -decay electrons corrected for energy losses in gases and windows.

In order to get an impression of the signal-to-background ratio in the present configuration, we arbitrarily assume that the energy distribution of the background is the same as measured directly in the vicinity of the beam (Fig. 14, slice B)¹. The result of the background subtraction procedure can be seen in Fig. 17. The solid line represents the theoretical energy distribution of the -decay electrons with included corrections for the energy losses only. The average signal-to-background ratio achieved in this study is  2:1. The recorded counting rate of the electrons from neutron decay (  20 s^-1mA^-1) is somewhat lower than the expected one (  40 s^-1mA^-1). Nevertheless, we estimate this result as satisfactory since not all efficiency factors are included in the analysis.

The obtained result is in no way the ultimate one. The background can be further reduced by replacing the borated epoxy plates in the neutron transport system (He-box) with ⁶LiF polymer and by proper optimization of its geometry. Such a modification will bring a significant improvement especially in the low energy range. This expectation was confirmed in a measurement where the whole beam was dumped into ⁶LiF layer at the exit of the collimator (70 cm away from the detector). The counting rate of the detector was four times lower as compared to the measurement where the neutron beam traversed the He-box. Additionally, the low energy component of the background was absent (Fig. 18).

Concluding this section we state that:

• The electrons originating from neutron decay can be clearly identified (signal-to-background ratio of 2:1) in the realistic conditions.
• There exists an obvious potential for further reduction of background by optimizing materials and geometry of the neutron transport system (He-box).

Figure 18: The energy spectrum of electrons collected with the ⁶LiF beam dump placed at the exit of the beam collimator. Note that the low energy component is absent.