Vacuum ultraviolet photon detector with continuously adjustable resolution for inverse photoemission spectroscopy LIU Shu-Hu(刘树虎)1 HONG Cai-Hao(洪才浩)1 ZHAO Yi-Dong(赵屹东)1GENG Dong-Ping(耿东平)1,2 ZHENG Lei(郑雷)1ZHAO Xiao-Liang(赵晓亮)1 LI Hua-Peng(李华鹏)1 1
Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201800, China
Abstract We present a vacuum ultraviolet (VUV) band-pass photon detector for inverse photoemission spectroscopy. A SrF2 window is used due to its high-energy cutoff of the optical transmission at E= 9.7eV, and acetone is selected as filling gas with the photoionization threshold also at E= 9.7eV. The structure of the detector described in detail is based on a Geiger-Müller type counter with a MgF2 window and argon as amplification gas. Its energy resolution can be tuned continuously from 46meV with a normal temperature situation to 105meV at 215K. Meanwhile, the signal intensity of the detector is adjusted accordingly to find an optimal operation program for our inverse photoemission system under constructing. The ratio of acetone vapor and argon is varied carefully. Background signals and the response of time are analyzed. The detector is calibrated by deuterium lamp in combination with a grating monochromator. Key words: Photon Detector, Inverse Photoemission, Resolution, SrF2, Acetone, Double Windows, Cooling Equipment. PACS: 07.60.Dq; 07.07.Df; 68.35.bd
1. Introduction Inverse photoemission spectroscopy (IPES) is the most direct technique to probe empty electronic states in solids and at clean and adsorbate covered surfaces. It is a complementary method with photoemission spectroscopy (PES) to map the energy band structure across fermi level. In an IPES experiment, low energy electronic beam is utilized to inject into the unoccupied part of conduction band of solids, subsequently, Such excited states will de-excited forward to lower lying unoccupied final states with radiating vacuum ultraviolet (VUV) range photons. Once injecting electron energy is scanned with fixed special energy photons detected, the map of spectra of energy band above fermi level will be showed. Unfortunately, IPE progress has a smaller interactive cross section than PE, it is presented by low count rate in an isochromat VUV detector. Moreover, PES has an energy resolution about 10meV, which is far better than IPES. Up to now, it is important to realize high performance VUV detector that will be reported in this paper. Ultraviolet photon detectors which are used to record the amount of the photons emitted from samples by electrons delaying from initial states to lower energy level states above the fermi level Ef have experienced a long period of development [1-4]. Those detectors may be divided into two kinds. One can be termed narrow bandpass detectors that commonly utilize the cutoff of fluoride  crystals being at ~10 eV to filter high energy photons and electron generator with thin film or some gas as barriers of low energy photons entry [6,7]. These gas filling bandpass detectors employ Geiger-Müller counter geometry structure and they were originally used to capture Lyman-alpha radiation with iodine vapor. The electric field strength ε in this structure at a radius r is written by 1) E-mail: [email protected]
2) E-mail: [email protected]
ε(𝑟) = 𝑟 ln(𝑏⁄𝑎). (1) Where V is the bias applied on anode, (a) and (b) are the cathode inner and anode diameter . In this electric field, avalanche would be formed when some electrons run to special radii r. However, Iodine vapor filling VUV detectors suffer a poor resolution and the count rate strongly depends on the temperature 8%/℃ [9,10]. The other one makes photons to be detected by photomultiplier or Microchannel plate (MCP) combination with a grating spectrometer [11, 12]. Though the later have a higher resolving power, The Geiger- Müller type bandpass detectors are more popular because they cost much less than grating spectrometer. In fact, Bandpass ultraviolet detectors also can be constructed without filling gas. N. Sanada et al had took advantage of CuBe electron multiplier with the first dynode as the photocathode to detect photons. But, comparison had been done for the two kinds bandpass detectors, the result was that gas detector had a greater sensitivity than solid-state-detector by a factor of 20 . All in all, Gas/window Geiger-Müller type bandpass detectors are more popular at present. The CaF2 is generally selected as the window of bandpass detector with an energy cutoff at 10.08eV, which makes a resolution about 300meV for IPES combined with some filling gas at a mean energy 10 eV. This resolution can be varied to 130meV at present has be reported by heating the CaF2 making a movement of the cutoff straightforward to lower energy . However, heating CaF2 in an ultrahigh vacuum system is not a good idea for surface experiments. Of course, MgF2 crystals which have a cutoff at 10.97eV also have been tried as the window of detector with different filling gas as CaF2 detectors. Though the detector’s mean energy has been improved by 0.8eV, the resolution of the MgF2 detector is approximately 370meV. This result would be not better than CaF2 detector. We have to mention a double CaF2 window detector which had achieved a resolution about 115meV by using krypton absorption lines . It was a very exciting progress for greatly improving the resolution of inverse photoemission spectrometer, but another two methods make us expect more. One is to dope strontium element into CaF2 monocrystalline which accomplishes a resolution about 80meV by mixing suitable ratio of the three elements . The other one is to cool SrF2 window to shift the cutoff toward high energy. By this method, 165meV resolution of inverse photoemission spectrometer had been obtained. However, Gases used for absorbing Ultraviolet photons have played an important role in the detector for inverse photoemission experiments. Researchers will generally fill two kinds gas to a bandpass ultraviolet detector for an inverse photoemission spectrometer. Generally, acetone vapor and argon would be selected as function gas for our independent design photon detector. The work gas acetone which usually has a photoionization potential at approximate 9.7eV makes a function filter to impede low energy photons to product signal, which makes only high energy photons detected. Moreover, Ar+ de-excitation and recombination with electrons at the surface of the cathode wall may cause radiation ultraviolet photons which products signal as same as photons from samples used for inverse photoemission experiments, while, acetone is a kind of polyatomic molecules gas that is able to stop false count pulses from photon or ion feedback process. However, argon affects as multiplier gas to form significant amount of negative charges to the amplifier input. Furthermore, the more the amount of charge, the better signal to noise ratio will be accessed. We have constructed a double window Geiger-Müller type vacuum ultraviolet photon detector with SrF2 crystal as low energy cutoff and acetone as filling gas. As has been known, the cutoff energy of SrF2 lies around 9.7eV as same nearly as acetone ionization potential, which is took advantage of acquiring continuously adjustable resolution by varying the temperature of SrF2 window. We have had the ability to control the temperature change of SrF2 from room condition to 210K ±0.2K at any point by a simple method with the SrF2 crystal placed in an ultrahigh vacuum system. We concentrated on investigating the characteristics of SrF2 /acetone VUV detector, such as resolution, and had made efforts to the measurement of the absolute spectral response of the detector. Note on count voltage range limit of detector electronics should be stressed for electronic field potential distorted by space charge effect. Related phenomena have been observed and “dead time” will be discussed in the following.
2. The structure of detector
Fig. 1. (Color online) Cross sectional design drawing of VUV photon detector: (a) vacuum Safe High Voltage (SHV) feedthrough, (b) hole punched for acetone vapor in, (c) apertures of cooling gas in and out, (d) double sided CF flange, (e) tungsten wire, (f) 316L stainless-steel cathode cylinder, (g) PTFE spacer, (h) MgF2 window, (i) insulator PEEK, (j) copper ring bracket, (K) SrF2 window. Color was used to distinguish components.
The VUV detector geometric structure is similar to end-window Geiger-Müller type counter. It consists of electropolished stainless steel cylindrical tube with inner diameter of 27mm welded to a double sided DN40 conflat flange. It was punched two 3mm diameter holes to let cooling gas flow and one 1/4 in diameter to fill function gas for detector. Tungsten wire anode of 1mm in diameter was placed at the center of cathode cylinder by polytetrafluorethylene (PTFE) spacer. The spacer was apart from frontier of W wire about 70mm for keeping anode standing at center and preventing it from vibrating. The double sided CF flange was sealed by a special vacuum safe high voltage feedthrough with silicone ring anti-corrosion from acetone vapor. MgF2 crystal of 2mm thickness and 26.5mm diameter with 90% transmittance at 9.7eV was glued to cathode tube inner by Torr Seal sealant epoxy. The distance between tungsten cutting edge and MgF2 window is 5-10mm. The lateral of MgF2 window is Polyetheretherketone (PEEK) insulatorbracket tightened to cathode tube external to fix oxygen-free copper which is used to transfer heat of cooling gas and SrF2 window. 316L stainless steel tube of 3mm outside diameter and 2mm inner diameter was welded to copper bracket by silver solder through double sided CF flange.
3. The apparatus of temperature control
Fig. 2. Schematic of refrigeration system for temperature controlling, showing the (a) compressed Nitrogen cylinder, (b) regulator, (c) screwed bonnet needle valve and pipe, (d) solenoid valve, (e) Dewar container, (f) temperature controller.
This temperature control system was constructed for cooling SrF2 window with sufficient accuracy and less cost. Commercial available cylinder with 13Mpa compressed Nitrogen would be used as gas source to cool the crystal via 6mm diameter polyethylene tube connected by steel pipes with tube fittings. However, 77K liquid nitrogen Dewar was assembled between gas source and cooled element with normal Nitrogen in and cryogenic gas out. Note must be mentioned is that copper pipe would be set up for heat exchange between Nitrogen and liquid nitrogen. The key to make the window temperature stable at any point is to control airflow opening and closing. It is carried out by PI temperature controller and screwed bonnet needle valve.
4. Experiment A series of test and measurement system had been assembled for charactering the performance of the detector (fig. 3.)
Fig. 3. (Color online) Measurement system scene photo.
The importance is that the system is entirely independent of inverse photoemission spectrometer and relatively easy to be implemented in a general laboratory. This is because Maniraj. M et al had made great efforts to do many remarkable experiments with synchrotron light source on the measurement of acetone detector, but a lot of extra works had to be done for reducing the intensity of light taking into account the sensitivity of gas detector. We employed a 150W large power commercial deuterium lamp (L1835, Hamamatsu, Japan) as photon source which had much broader spectrum than argon microwave discharge source from 110nm to 400nm. This deuterium could provide a stable output for a long time with extremely small fluctuation ratio. However, the 150W deuterium is high flux for the detector to work. Increasing distance between deuterium lamp and the detector to make most of photons absorbed by the air, especially oxygen, had been tried though it was difficult to settle down the position which the detector should be placed on. The consumption and supplement of oxygen by the air flow in the optical path significantly impact on the penetration distance of UV photons. Fortunately, pulses could be observed with uncertain count frequency that shows the design of our VUV detector working well. However, the energy resolution of detector cannot be obtained by the method. So we arranged a monochromator (VM-502, Princeton Instruments, USA) using sinebar drived scanning system with single point resolution 0.1nm. The optical path was covered between them by nitrogen or extracted to vacuum 10E-2Pa with scroll pump. Nonetheless, the slit must be opened nearly to “zero” cautiously for preventing photons breaking all acetone molecular down before doing the experiment. Acetone/Ar2 VUV-detector can be operated in a very broad range of gas filling ratio. Argon pressure had been changed starting from 5KPa. It is an extremely low atmosphere condition for acquiring sufficiently reliable signals with accommodative signal to noise ratio. As the pressure promoted to 15KPa, no pulse can
be observed even the cosmic background signal in spite of anode bias applied to 1500V volts that is withstand limit of our electronics. It is worth noting that acetone employed during the measurement was 300Pa. As the same as Ar2, acetone gas was carefully added to the detector by slowly rotating Swagelok Integral-Bonnet Needle Valve with pressure range from 100Pa to 1500Pa, subsequently, Ar2 would be filled into detector so that the ultimate pressure was in the region of 10KPa. Both acetone and argon gas pressure was measured by ceramic capacitor film absolute pressure transmitter. The gauge takes advantage of resisting corrosion from acetone vapor. Its working range was set covering 0.2Pa-26KPa for obtaining more accurate measurement and it had to be calibrated when exposed to the atmospheric pressure. Calibration can be performed in a vacuum system with pressure lower than 0.13Pa or 1X10-3torr. We must point out that all of above experiment were operated under the condition of 250K SrF2 crystal. However, the resolution had been revealed by changing the temperature of SrF2 from 300K to 215K by the gas-cooled device with 10KPa argon and 300Pa acetone gas. The voltage was applied to tungsten anode around 1000V and anode pulse signals were shaping-amplified by homemade amplifier with gain range from 1/2 to 20. The source of background signal had been analyzed by using scintillator detector system (fig. 4). Two scintillator detectors were placed on both sides of VUV detector upper and down. Anode voltages were changed from 400V to 1500V. They were manipulated to monitor coupling signal with RIGOL four-channel digital oscilloscope once high-energy cosmic particles through them. These three detectors were placed in parallel with the ground. As the same time, anode positive bias of VUV detector would be varied to try to access pulse signal.
Fig. 4 (Color online) Schematic diagram for scintillator system
The other important experiment is to normalize the spectroscopy of VUV detector photon response because the deuterium light source possesses different photon flux as energy changed. A silicon photodiode (AUX-100G, IRD, USA) was selected as detector to measure the deuterium monochromatic spectrum as slits width larger than the previous optical design.
5. Result and discussion The most important characteristic of VUV photon detector is the counting capacity for surffering low light signal radiation condition. It is resolved by mixing ratio of argon and acetone vapor. The results are shown in Fig. 5 and Fig. 6 fitted by gaussison function. Ar gas pressure contributes smaller impact than acetone vapor on count. 10Kpa argon obtains 30% count ratio increment than 5Kpa gas filled in the peak position. The count is certain confident of being consistent taking account of experimental uncertainty with 5Kpa and 15Kpa argon compared. While, the detector exhibits more sensitivities with acetone vapor. From fig. 6, information revealed that acetone of 300Pa pressure creats three times more count rate than 1500Pa condition. It is a excellent charify that acetone molecules play a quench gas role as well as detect gas. Because acetone as polyatomic organic molecule posesses a large of number non-radiative excited states, such as rotational and vibrational states, to capture VUV range photons with dissociating itself. However, Considering voltage
applied to anode and detector lifetime, 300-600Pa acetone vapor mixed with 10Kpa argon is found to be the best combination.
Fig. 5. (Color online) Response curves of detecter from the different Argon pressure with 300Pa acetone and SrF 2 cooled to 250K. The pressure of 5KPa Argon is represented by the thinest green line, for 10KPa condition by the most thickest red line, and 15KPa by middle size black line.
Fig. 6. (Color online) Significant change illustrated by purple dash dot arrow. Colors and sizes of the solid line is used to identify different acetone vapor pressure filled.
In terms of increasing count rate, cooling SrF2 window is a effective method. Fig.7 shows the corresponding variation of the sensitivity and resolution of the detector as the temperature of crystal was reduced. Comparing the two solid curves, Resolution exhibits more slower growth than sensitivity does. Moreover, This comparsion sharply indicates that less resolution deterioration can rapidly upgrade the sensitivity of detector, It is considerably significant for inverse photoemission experiment of different samples. Simply put, Various samples would produce different photon signal intensity so that resulution and sensitivity have to be adjusted to adapt it.
Fig. 7. (Color online) According variations of resolution and sensitivity with the change of SrF2 crystal window temperature showed. The blue square point solid line represents resolution status, for sensitivity by red round point solid line. Resolution is comfirmed by FWHM and sensitivity is accessed through count rate in the partly peak position of response curve schematized in the Fig. 9.
However, there is no doubt that resolution estimated by Full Width at Half Max (FWHM) is a significant characteristic of detector performance. The cutoff of SrF2 crystal has been measured in vacuum ultraviolet photon beam station of Beijing Synchrotron Radiation Facility (BSRF), China (Fig. 8). Inco indicates that transmittance of SrF2 crystal is only 5% at 9.69eV under normal temperature condition. As we all know, acetone photoionization threshold also is nearby 9.69eV energy level, so the probability of creating signal of detector would be extremely small. The detector photon response “window” can be opened by cooling SrF2 leading cutoff to shift forward higher energy level though resolution exhibits deteriorative trend (Fig. 9). Obviously, the resolution can be continuously adjusted from 46meV at room temperature to 105meV once SrF2 temperature cooled to 215K. Besides the position of high energy cutoff of SrF2 crystal, the other factor contributing to FWHM is acetone photoionization efficiency. It is impacted by gas-filled pressure and lowerlying Rydberg states. The latter would be excited by photons with energy higher than photoionization threshold together [17, 18]. It is illustrated by the rising edges in the Fig. 9. For this reason, more researches should be executed to solute it for obtaining better resolution, as the same time, higher count rate, in the further.
Fig. 8. (Color online) Transmittance of SrF2 crystal window obtained at 300K condition in vacuum ultraviolet station of BSRF, China. Steep side is observed nearby 9.7eV photon energy. The flux measurement of front and behind positions of SrF2 window is not at the same time, but is normalized by storage ring electron beam current.
Fig. 9. (Color online) The response curves of SrF2 window with different temperature from 295K to 215K realized by homemade cooling apparatus. Photon entering window was opened by reducing SrF2 temperature and the cutoff of acetone vapor was marked by purple dash dot arrow. The resolution was measured by FWHM.
The resolution impacted by deuterium lamp has also been calibrated. Because Photon flux is not same at different energy when light exits from slits. The result show that morphology of response curve is unchanged but the resolution is found to be smaller by 10meV (Fig. 10).
Fig. 10. (Color online) The response of detector normalized by deuterium lamp spectroscopy.
The other calibration about cosmic radiation background revealed an expected result. No signal was observed in the VUV detector when scintillator detector coupling signal occurred, while the channel of VUV detector obtained a large of pulse signals as baseline of oscilloscope substantial jitter though coupling signals being rarely numbers. It signifies that the detector has low noise. This ensures the reliability of detector as operated under low count rate and high resolution condition. As the same as others lectures had pointed out , we found that the detector did not work in Geiger region and the amplitude of pulse signals showed considerable gap among them. Fig. 11 was acquired by recording signal pulses in a relative long time with oscilloscope and the small illustrate is the morphology of a single pulse. Due to that, every pulse is seemed to be a black vertical line in the figure. It is a significant information to discuss about dead time and recovery time. In a general Geiger-Müller type counter, the lower threshold selects pulse to be counted. Only low count limit changed, the recovery time would be extended or shortened accordingly. However, the varying of amplitudes of VUV detector estimated is random without the condition of high pulse followed by gradually increasing low pulse as general G-M counter dose. The space
charge effect and photons entering in different transverse position of the detector are considered contributing the phenomena. In a word, recovery time is not a meaningful and exact statement to reflect the response time of detector here, we fully believe, and the pulse width defined as dead time is measured around 20-50μs confirmed as no counting time with the detector geometry described prior. Interestingly, when we perform oscilloscope to take picture recording signal amplitudes in a long time with changing energy from 9.9eV to 9.6eV one by one, we observed that amplitudes would be reduced in higher count rate region but increased in lower count rate region in Fig. 11. That is an important condition minded to set up counter voltage threshold values.
Fig. 11. (Color online) Relationship between photon energy and pulses amplitude and quantities of detector. The region of sparse pulses shows the distribution of amplitude, yet. The time scale of oscilloscope was configured to scan all pulses in a specified time.
6. Conclusion The performance of Geiger-Müller type VUV photon counter has been investigated and we have realized much simpler program to change the resolution from around 50meV to 100meV with low cost. 600Pa acetone vapor mixed by 10KPa argon with around 1000V bias applied is determined as the most appropriate combination for high sensitivity of detector. The relation between sensitivity and resolution have been elaborated, yet. The background signal analysis, count ratio normalization and response of time have been executed to ensure the actual count capability of the detector. Meanwhile, it is important that researches of VUV photon detector do not need to be done in a UHV spectrometer system, of course, and synchrotron light may be replaced by deuterium lamp.
Reference 1. 2. 3. 4. 5. 6. 7. 8. 10. 9.
D Funnemann and H Merz. J. Phys. E: Sci. Instrum., 1986, 19 554. M. Budke et al. Review of Scientific Instruments, 2007, 78(8): 083903. M. Maniraj et al. AIP Conference Proceedings, 2011, 1349, 497-498. R. T. Brackmann et al. Review of Scientific Instruments, 1958, 29(2): 125. N. Sanada et al. Review of Scientific Instruments, 1993, 64(12): 3480. P. M. G. Allen, P. J. Dobson, and R. G. Edgell, Solid State Commun.,1985, 55, 701. J. A. Lipton-Duffin et al. Review of Scientific Instruments, 2002, 73(9): 3149. Glenn F. Knoll. Radiation Detection and Measurement. Third Edition. Ann Arbor, Michigan: John Wiley ＆Sons,Inc., 1999, 162-163. V. Dose et al. Journal of Physics B: Atomic and Molecular Physics, 1969, 2(12): 1357. G. Denninger, V. Dose, and H. Scheidt. Applied physics, 1979, 18(4), 375-380.
11. H. Yoshida. Rev Sci Instrum 84(10): 103901. 12. T. Fauster et al. Review of Scientific Instruments, 1985, 56(6): 1212. 13. Hill, I. G. and A. B. McLean. Review of Scientific Instruments, 1998, 69(1): 261. 14. M. Budke et al. Review of Scientific Instruments, 2007, 78(11): 113909. 15. R. Stiepel et al. Review of Scientific Instruments, 2005, 76(6): 063109. 16. M. Maniraj et al. Rev Sci Instrum., 2011, 82(9): 093901. 17. J. H. Kim et al. (2012). Hyperfine Interactions, 2013, 216(1), 85-88. 18. Lixia Wei, et al. J. Phys. Chem. A, 2005, 109, 4231-4241. 19. S. Banik et al. Review of Scientific Instruments, 2005, 76(6): 066102.