A new large area MCP-PMT for high energy detection

20-inch Large area photomultiplier tube based on microchannel plate (MCP-PMT) is newly developed in China. It is widely used in high energy detection experiments such as Jiangmen Underground Neutrino Observatory (JUNO), China JinPing underground Laboratory (CJPL) and Large High Altitude Air Shower Observatory (LHAASO). To overcome the poor time performance of the existing MCP-PMT, a new design of large area MCP-PMT is proposed in this paper. Three-dimensional models are developed in CST Studio Suite to validate its feasibility. Effects of the size and bias voltage of the focusing electrodes and MCP configuration on the collection efficiency (CE) and time performance are studied in detail using the finite integral technique and Monte Carlo method. Based on the simulation results, the optimized operating and geometry parameters are chosen. Results show that the mean ratio of photoelectrons landing on the MCP active area is 97.5%. The acceptance fraction of the impinging photoelectrons is close to 100% due to the emission of multiple secondary electrons when hitting the MCP top surface. The mean transit time spread (TTS) of the photoelectrons from the photocathode is 1.48 ns.

for high CE).A pair of MCPs are placed at the top opening of the conical barrel.Implementing millions of MCP channels in the three-dimensional model is impossible.Hence, a simplified model based on two MCPs without channels is adopted in our simulation.The glass envelope handle cooperates with the Electrode I, Electrode II and MCPs to generate the accelerating and focusing electric field which benefits CE and time performances.
Simulations are conducted to validate the feasibility.The new MCP-PMT model is simulated in CST Studio Suite 10 .The electric field, electron trajectories, energies and velocities are calculated based on the Finite Integral Technique and Monte Carlo method.The feasibility and effectiveness of this simulation approach has already been validated [11][12][13] .
Photoelectron trajectories from the photocathode to MCP are well simulated.2000 photoelectrons uniformly distributed from 28.5° (corresponding to the photocathode edge) to 90°on photocathode are sampled by Monte Carlo method.The initial energy of electrons exiting the photocathode obeys β(1,4) distribution with mean value 0.15 eV in the range 0.0-0.6 eV.The emitted azimuth is uniform distributed over the range of 0-2π.Initial elevation follows Lambert cosine distribution from 0° to 90°.Owing to the short distance and high potential difference between the first MCP-in and anode, the electron transit time through the MCP is around several hundred picoseconds and TTS is just tens of picoseconds which thus are negligible.In our simulation, the transit time distribution between the photocathode and the first MCP is evaluated.
CE is defined as a ratio of the number of photoelectrons from the photocathode collected by MCP channels to the total number.Large area MCP-PMT employs coated MCP which is coated with high secondary electron yield (SEY) material on the flat inter-channel area and electrode penetrating in channels area, to obtain high CE 11,14 .A great deal of secondaries can be excited in the high SEY area and finally be collected by the MCP channels, which makes the acceptance fraction of the impinging photoelectrons close to 100% 11 .In our simulation, it is impossible to obtain the exact CE value attributing to the simplified MCP model approximated to a flat surface without channels.Thus, only the ratio of photoelectrons landing on the simplified MCP input flat surface (active area, CE a ) is evaluated.
The dependence of CE a and time performance on the bias voltage and size of the two focusing electrodes and MCP is systematically investigated.Photocathode voltage is 0 V.Only one parameter is varied at a time, while the others are kept constant, with values listed in Table 1.

Electrode I
The dependence of CE a and time performance on the applied voltage (U I ), diameter (D I ) and height (H I ) of the electrode I is investigated.U I , H I and D I are varied from 0 to 800 V, 20 mm to 170 mm and 240 mm to 400 mm, respectively.For each computing, only one parameter is varied at a time, while the others are kept constant, with values listed in Table 1.Sample results with values of mean ± SD are graphically represented in Figs. 2, 4 and 5.
In Fig. 2, decreasing CE a and TTS are observed as results of increasing U I .Figure 3 shows the electric field distributions in the PMT cavity for U I = 0 V (a) and 800 V (b).For U I = 0 V the electric field shows better focusing characteristic.It is helpful to obtain a high CE a .For U I = 800 V, the electric field between the photocathode and the top opening of electrode I is more uniform, which benefits the consistency of transit time which means short TTS.The CE a at U I = 0 V and 50 V is 100%.The shortest mean TTS is 0.93 ns at U I = 800 V.
It is obvious in Fig. 4 that CE a gradually increases to a maximum of 100% at D I = 360 mm and then reduces to some extent with increasing D I .Besides, a decreased TTS by increasing D I is observed.The shortest mean TTS is 1.79 ns at D I = 400 mm.For smaller D I , the electric field shows better focusing property.Larger D I is benefit for the uniformity of electric field, and finally TTS.As can be seen in Fig. 6 that with the increasing U II , CE a increases to 100% at U II = 3300 V, and the mean TTS decreases until its minimum 1.73 ns at U II = 1900 V and then increases.With the increasement of U II , the electric field is divergent first, then approximately uniform (benefit for short TTS), finally focused (conducive to high CE a ).
It is shown in Fig. 7  Increasing H II has a slight negative effect on CE a (exhibited in Fig. 8) but no significant effect on TTS.The mean TTS changes from 1.74 to 1.85 ns over the range of 30 mm ≤ H II ≤ 210 mm.At H II (fixed value) = 150 mm, mean TTS is 1.79 ns.The highest mean CE a is 99.2% at H II = 30 mm.

MCP
Dependence of CE a and TTS on the input face of the top MCP applied voltage (U MCP ) is well simulated.The total voltage applied on the two MCPs is 1000 V.Only U MCP is varied, while the others are kept constant, with values listed in Table 1.As is shown in Fig. 9 that with the increase of U MCP , CE a and TTS increase.Electric field for higher U MCP shows better focusing property but poorer uniformity.The highest mean CE a is 99.6% at U MCP = 3500 V.The shortest mean TTS is 1.54 ns at U MCP = 500 V.

Optimized model
Considering both high CE a and good time performance requirements, U I = 100 V, D I = 400 mm, H I = 130 mm, U II = 2000 V, D II-b = 380 mm, H II = 30 mm and U MCP = 2500 V are employed in the MCP-PMT model.Other parameter values are listed in Table 1.Simulation results show that the new design has a high CE and outstanding time performance.Simulation results show that the mean CE a of the photoelectrons from the whole photocathode is 97.5%, which mean that CE of the coated MCP-PMT is expected to be 100%.

Conclusion
This work presents a new large area MCP-PMT design with good time performance.A novel focusing system with a cylindrical barrel electrode and a conical barrel electrode are designed.Three-dimensional models are developed in CST Studio Suite to validate the feasibility and effectiveness.Dependences of CE and time performance on the size and bias voltage of the focusing electrodes and MCP configuration are systematically investigated.Based on the simulation results, a set of operating and geometry parameters are chosen for the new MCP-PMT design considering both high CE and good time performance.Results show that the mean ratio of photoelectrons landing on the channelless MCP input flat surface is 97.5%.The mean TTS value of photoelectrons from the whole photocathode achieves 1.48 ns.It will be a good candidate for the detection experiments with high CE and high time resolution requirements.

Figure 1 .
Figure 1.Schematic diagrams of the existing (left side) and new (right side) large area MCP-PMTs.

Figure 5
Figure5exhibits the simulated CE a and TTS, which result from the variation of H I over the range of 0-170 mm.CE a and TTS basically decrease with the increasing H I .From the visual electric distribution in the PMT cavity, it is observed that the electric field for lower H I shows better focusing property.Higher H I is benefit for the electric field uniformity.CE a is 100% for 0 mm ≤ H I ≤ 110 mm.The shortest mean TTS is 1.44 ns at H I = 170 mm.

Figure 3 .
Figure 3. Electric field distributions in the PMT cavity for U I = 0 V (a) and 800 V (b).

Figure 4 .
Figure 4. CE a and TTS versus D I over the range of 240 mm ≤ D I ≤ 400 mm.

Figure 5 .
Figure 5. CE a and TTS versus H I over the range of 0 mm ≤ H I ≤ 170 mm.

Figure 6 .
Figure 6.CE a and TTS versus U II over the range of 0 V ≤ U II ≤ 3500 V.
that increasing D II-b has positive effects on CE a and TTS.Electric field for smaller D II-b shows better focusing property.Larger D II-b is benefit for the field uniformity.D II-b = 380 mm is the optimum value, for which the mean CE = 96.8% and mean TTS = 1.7 ns.

Figure 7 .
Figure 7. CE a and TTS versus D II-b over the range of 120 mm ≤ D II-b ≤ 380 mm.

Figure 8 .
Figure 8. CE a versus H II over the range of 30 mm ≤ H II ≤ 210 mm.

Figure 9 .
Figure 9. CE a versus U MCP over the range of 500 V ≤ U MCP ≤ 3500 V.

Figure 10 .
Figure 10.Electron trajectories from the photocathode to the MCP input face.

Figure 12 .
Figure 12.Incident photoelectron distribution on the MCP input face.90° Latitude corresponding to the top point.Considering the symmetry of the MCP-PMT, photoelectrons are emitted from half of the photocathode.(a).Incident photoelectron position distribution on the MCP input face.(b).Incident radius distribution on the MCP input face.