Development of Array-Type Secondary Electron Emission Monitor Toward Single-Shot Measurement of Extraction Efficiency of FEL Oscillators

Abstract

To enable the single-shot measurement of extraction efficiency, a key parameter of an FEL oscillator, we developed an array-type secondary electron emission monitor capable of measuring the temporal evolution of the electron beam energy distribution in a macro-pulse at KU-FEL. The monitor consists of 24 ribbon-shaped electrodes and 2 shielding electrodes, and it is positioned after the energy analyzer magnet and just before a beam dump. The beam energy evolutions in a macro-pulse with and without FEL lasing were measured in a single shot with approximately 100 ns temporal resolution. From the results obtained, the extraction efficiency of FEL oscillators can be evaluated.

1. Introduction

In a Free Electron Laser (FEL), a relativistic electron beam passes through an undulator with a periodic magnetic field, where light is amplified via the interaction between the electron beam and the electromagnetic field of the radiation. The FEL wavelength is determined by the electron energy, the undulator period length, and the magnetic field strength, which together enable broad wavelength tunability. FELs can produce highly intense, coherent light across a wide range of wavelengths, ranging from microwaves to X-rays [1].

FELs are employed in a wide range of application-oriented research across diverse fields. In renewable energy research, FELs enable the study of photovoltaic materials and processing, thereby supporting the development of more efficient solar energy technologies [2]. In manufacturing, FELs are valuable due to their unique properties of wavelength tunability, coherence, and high intensity, making them effective tools for precision manufacturing and materials processing [3]. In the medical field, FELs show promise for applications such as surgery and cancer therapy, particularly for the precise targeting and destruction of cancerous cells [4]. In materials science, FELs are used to investigate the structure and dynamics of materials at the atomic and molecular scales, advancing the development of new materials with enhanced properties [5]. Through these diverse contributions, FELs play a vital role in promoting sustainable development.

An infrared FEL named Kyoto University Free Electron Laser (KU-FEL) [6] has been developed to promote energy-related research in the Institute of Advanced Energy, Kyoto University. The schematic diagram of KU-FEL is shown in Figure 1. In KU-FEL, the optical cavity is used to store the undulator radiation generated from the electron beam. The stored light is then amplified by decelerating the subsequent electron beams through the FEL interaction between the electron beams and the stored light. This design is known as an oscillator-type FEL. Currently, KU-FEL has achieved the highest extraction efficiency worldwide, reaching 9.4% [7] in photocathode operation. This parameter represents the ratio of the kinetic energy of the electron bunch converted into the energy of the laser field.

1.1. Extraction Efficiency of Oscillator-Type Free Electron Laser

An FEL converts the kinetic energy of electron beams to the energy of the electromagnetic wave (i.e., a laser field) through FEL interaction in an undulator. Extraction efficiency η is defined as the proportion of kinetic energy extracted from the electron beam and transferred to the laser field during a single pass through the undulator. This is an important parameter that determines the performance of FEL oscillators, as a higher extraction efficiency results in stronger FEL output power. Furthermore, FEL lasing with a micro-pulse width of only a few optical cycles has been demonstrated under high-extraction-efficiency conditions [8]. Under such conditions, the micro-pulse duration is known to be inversely proportional to the extraction efficiency [9]. Here, it should be noted that the overall FEL efficiency from electron beam to FEL output power depends not only on the extraction efficiency but also on other parameters. Detailed theoretical analysis has been made and reported by A. Curcio [10].

To determine the extraction efficiency, an accurate evaluation of the energy transferred from the electron beam to the laser field is essential. A detailed method for measuring this is reported in reference [11]. In FEL interaction, energy is transferred exclusively between the electron beam and the laser since there is no laser medium that would otherwise cause energy loss via heating. According to the law of conservation of energy, the kinetic energy lost by the electron beam must equal the energy gained by the laser field. Figure 2 shows a schematic of the electron beam energy distribution with and without FEL lasing. In the figure, E₀ and E₁ denote the average kinetic energy of the electron beam without and with FEL lasing, respectively. As a result of FEL amplification, the energy spread of the electron beam increases, while its average energy decreases. The extracted energy, ΔE (=E₀ − E₁), is the difference between these two conditions. Therefore, the extraction efficiency can be calculated as:

$$\eta ={\displaystyle \frac{\Delta E}{E_0}},$$

(1)

with assuming the same total charge for the condition with and without FEL lasing. Consequently, the extraction efficiency can be evaluated by measuring the electron beam energy distribution with and without FEL lasing. In several facilities, this method has been used [11,12].

In an FEL oscillator driven by a normal-conducting linac, the FEL lasing process persists throughout the electron beam macro-pulse. At KU-FEL, the electron beam macro-pulse duration is approximately 7 μs. Within this macro-pulse, FEL starts from shot noise and eventually reaches saturation. The electron beam energy distribution undergoes significant changes within the macro-pulse due to FEL lasing, specifically the energy transfer from the electron beam to the laser light. To determine the instantaneous extraction efficiency of an oscillator-type FEL driven by a normal-conducting accelerator, it is necessary to measure the temporal evolution of the electron beam energy distribution, both with and without FEL lasing, with a temporal resolution of approximately 100 ns. This resolution corresponds to three roundtrips in a 5 m optical cavity. For this measurement, a Faraday cup and a slit placed after a dipole magnet have been employed [11]. The energy distributions of the electron beams have been measured by scanning the excitation current of the dipole magnet (thereby controlling the electron energy that can pass through the slit) while simultaneously measuring the electron beam current profiles with the Faraday cup. In previous research [7], a maximum electron energy decrease of 16% was observed when the extraction efficiency was 9.4%. It is expected that higher extraction efficiency will lead to an even greater reduction in electron energy. Therefore, for the future development of an oscillator FEL aiming for an extraction efficiency beyond the current record, the measurement energy range needs to accommodate a decrease larger than 16%.

The conventional method for measuring extraction efficiency, which relies on scanning the excitation current of a dipole magnet, can measure the energy decrease larger than 16% but presents several challenges. Firstly, it makes it difficult to determine the extraction efficiency under conditions where FEL lasing is unstable. Secondly, it requires extensive averaging to mitigate the influence of the shot-by-shot fluctuation in FEL intensity. Thirdly, the process of scanning and averaging results in a prolonged measurement time, making machine tuning based on extraction efficiency impractical. The higher the extraction efficiency becomes, the longer the measurement time is required. Furthermore, due to the necessity of averaging, the shot-by-shot fluctuation of the extraction efficiency cannot be measured. These issues can be solved by replacing the combination of a Faraday cup and a slit with a 1D profile monitor that offers high temporal resolution. As a promising candidate for such a monitor, we propose the use of a 1D profile monitor based on Secondary Electron Emission (SEE).

1.2. Secondary Electron Emission (SEE) Monitor

When high-energy charged particles strike a metal surface, they scatter and transfer energy to the electrons within the metal, causing these electrons to be ejected from the surface as secondary electrons. This phenomenon is known as Secondary Electron Emission (SEE). SEE has been utilized in measuring the beam profiles of charged particles [13,14,15,16,17]. SEE-based profile monitors are generally categorized into two types. The first is the wire scanner, which uses a single or a few wires with high-precision linear translation [14]. This setup allows for the measurement of very small beam sizes, down to the diameter of the wire itself. However, the wire should be scanned for reconstructing the beam profile then multiple shots are required for one measurement. The second type is the multi-wire or multi-ribbon profile monitor, which utilizes an array of many wires or ribbon electrodes [13,15,16,17]. The multi-ribbon configuration specifically enables real-time or single-shot measurement of the beam profile. The multi-ribbon-type 1D profile monitor, utilizing SEE (hereafter referred to as the SEE monitor), is schematically illustrated in Figure 3. As depicted, the SEE monitor for 1D profile measurement comprises a 1D array of electrodes and two thin-film shield electrodes. A positive bias voltage is applied to these shield electrodes. When the thickness of the shield and ribbon electrodes is thin enough and the kinetic energy of the incident charged particle beam is sufficiently high, the charged particle beam can pass through these three-layered electrodes without strong perturbation. Secondary electrons emitted from the 1D-array electrodes are collected by the shield electrodes. Similarly, secondary electrons generated by shield electrodes themselves are collected by the shield electrodes. An electric current flows through the external resistor, its magnitude is proportional to the amount of emitted secondary electrons in each 1D-array electrode. Since the yield of secondary electrons is directly proportional to the amount of incident charged particle beam on each 1D-array electrode, a 1D profile of the charged particle beam can be reconstructed from the voltage signals appearing across the 1D-array electrodes in single-shot.

To date, numerous multi-wire SEE monitors have been developed for measuring the 1D profile of charged particle beams [18,19]. However, to the best of our knowledge, this type of monitor has not yet been employed for measuring the extraction efficiency of oscillator-type FELs. As previously discussed, measuring the extraction efficiency of an oscillator-type FEL can be accomplished by monitoring the temporal evolution of the electron beam energy distribution within a macro-pulse. As illustrated in Figure 1, there is a bending magnet after the undulator to separate the electron beam and the FEL beam. After passing through this bending magnet, there is a strong correlation between the electron beam’s energy and its horizontal beam position. By placing a 1D-array-type SEE monitor before the beam dump, the energy distribution of the electron beam after FEL interaction can be readily measured by utilizing this strong correlation between the energy and horizontal position. The remaining concerns are the energy resolution, the temporal resolution, and signal sensitivity (or signal-to-noise ratio) of the monitor. In most cases, the array-type SEE monitors have been used for applications that do not require high temporal resolution. Slow amplifiers are often used to increase signal intensity. However, since the secondary electron emission is a prompt reaction to the strike of high-energy charged particles, the temporal resolution of the SEE monitor can, in principle, be high.

In this research, we successfully demonstrated the single-shot measurement of the extraction efficiency of an oscillator-type FEL. This was achieved using a novel 1D-array-type SEE monitor, which consists of 24 ribbon-shaped electrodes and two shielding electrodes. We experimentally evaluated the temporal resolution, signal sensitivity, and energy resolution of this newly developed device. Results of some demonstration experiments of single-shot extraction efficiency measurement are also reported.

2. Design of SEE Monitor for Extraction Efficiency Measurement of KU-FEL

A 1D-array-type SEE monitor was designed based on the conventional method employed for the extraction efficiency measurement in previous research [7,11]. In the previous study [7], a slit width of 3 mm was used before the Faraday cup, yielding an energy resolution of approximately 1%. To achieve a comparable energy resolution in our design, the width of each electrode was chosen to be 3 mm, with a spacing of 0.8 mm between adjacent electrodes. With this arrangement, each electrode is expected to cover the energy range of approximately 1.3%. Consequently, a monitor incorporating 24 electrodes can cover an approximate energy range of 30%. The overall measurable energy range is also influenced by the dimensions of the vacuum chamber, the electron extraction window, and the good field region of the dipole magnet. For this research, a new vacuum chamber with a substantial width (90 mm, same with the width of the iron yoke of the dipole magnet) was specifically designed and fabricated. Furthermore, an electron extraction window made of a Titanium alloy with a thickness of 300 μm was also designed and fabricated with considering the balance between the scattering of the electron beam and mechanical robustness. Due to budgetary constraints, the existing dipole magnet cannot be replaced. The iron yoke of this dipole magnet has an approximate width of 90 mm, with its good field region spanning approximately 40 mm.

Figure 4 represents the schematic of the designed 1D-array-type SEE monitor. The ribbon electrodes are installed and connected to coaxial connectors via insulated copper wires. These ribbon electrodes are fabricated from 50-μm-thick, 3 mm-wide titanium strips. With 0.8 mm spacing between electrodes, they are capable of fully covering the 90 mm width. The two shielding electrodes are made of 11-μm-thick aluminum foil and are positioned to cover the ribbon electrodes. Due to space limitations, the device incorporates 12 coaxial connectors on both the top and bottom sides, with ribbon electrodes connected from both directions. Photographs of the shielding electrode and the ribbon electrodes are shown in Figure 5.

3. Experimental Setup and Condition

The fabricated and constructed 1D-array-type SEE monitor was installed before the beam dump, as shown in Figure 1. Photographs of the electron beam extraction window and the installed monitor are provided in Figure 6. In contrast to conventional setups where SEE monitors were typically operated under vacuum conditions [15,16,17], in this study the SEE monitor was placed in air. This configuration introduces an inherent challenge: scattering from the electron extraction window and the ambient air can increase the electron beam size at the SEE monitor. To mitigate this effect, the distance between the SEE monitor and the extraction window was minimized to approximately 3.5 cm. Additionally, the extraction window was fabricated from a titanium alloy with a thickness of 300 µm to further reduce scattering. The shielding electrode was positively biased with +300 V relative to the electric ground of the facility. To acquire the electric current signals from each ribbon electrode, a high-speed 16-channel Analog-to-Digital Converter (ADC) (Manufacturer: Spectrum Instrumentation GmbH, Model: DN2.592-16, City: Grosshansdorf, Country: Germany) was used. This ADC offers an analog bandwidth of 10 MHz and a sampling rate of 20 MS/s, and was operated without additional amplification. To minimize the time constant and capture fast variation in secondary electron emission from each ribbon, 50 Ω termination resistors were employed. The ADC channels were connected to the ribbon electrodes in sequential order: channel #1 was connected to the 5th electrode, channel #2 to the 6th, and so forth, with channel #16 connected to the 20th electrode. The 1st electrode was positioned at the outermost corner side, meaning that higher energy electrons would strike electrodes and channels with lower numbers.

The demonstration experiments were conducted under the thermionic cathode operation mode of KU-FEL. The electron-bunch repetition rate was 2856 MHz, and the macro-pulse duration of the electron beam was 7 μs. The electron beam energy was adjusted to be 35 MeV, and the undulator gap was set to 15.15 mm. Under these conditions, the peak wavelength of the FEL was 8.17 μm.

4. Results

4.1. Basic Performance Test

Initially, we performed a basic performance test of the 1D-array-type SEE monitor without FEL lasing. The excitation current of the dipole magnet was adjusted to be 14.6 A. The results are shown in Figure 7. The signal voltage from the SEE monitor was observed to be highest on channel #5, reaching approximately 1.55 V at 6.5 μs. This signal was significantly higher than environmental noise, which was measured to be around 4 mV. Consequently, the maximum S/N (signal-to-noise) ratio was calculated to be 388 (1.55 V/0.004 V). This confirms that the signal sensitivity of the array-type electrode SEE monitor is sufficiently high. Regarding the response speed of the SEE monitor, the system was able to capture fast transitions in the signal voltage, on the order of 100 ns, both at the beginning and end of the macro-pulse. Therefore, it was confirmed that the constructed system possesses a sufficiently high response speed to enable single-shot measurement of the extraction efficiency in KU-FEL.

The energy calibration of the SEE monitor was carried out by varying the excitation current of the dipole magnet. Figure 8a illustrates the dependence of the integrated voltage measured by each electrode over the macro-pulse on the excitation current of the dipole magnet. Based on the results, the electrode dependence on the excitation current giving maximum charge is calculated as shown in Figure 8b. According to the relationship between current and channel shown in Figure 8b, the current variation in each electrode was about 1.2%. The linear fitting of the measured data yielded the following relationship between the channel and the excitation current:

I = 13.62 + 0.18 × ch,

where I represents the excitation current of the dipole magnet (in Ampere) and ch denotes the channel number of the fast ADC.

For the single-shot measurement of the energy distribution, it is essential to establish a conversion equation that relates the fast-ADC channel number directly to the electron energy, assuming a fixed dipole excitation current. When the electron beam energy is fixed at E₀, the electron radius of curvature R′, within the dipole magnet operating at the excitation current of I′, can be described by the following relationship:

$$\mathit{R}\prime = {\mathit{R}}_0 \times {\displaystyle \frac{I_0}{I\prime }},$$

where R₀ is the radius of curvature of the reference orbit and I₀ is the excitation current of the dipole magnet to have the radius of curvature R₀ with the given electron energy E₀. Conversely, when the excitation current of the dipole magnet is fixed at I₀, the electron radius of curvature R′, within the dipole magnet with the electron energy E′, can be described as:

$$\mathit{R}\prime = {\mathit{R}}_0 \times {\displaystyle \frac{E\prime }{E_0}}.$$

When the radius of curvature (R′) is the same under both conditions, we can connect these two expressions to establish a connection between them as follows:

$$\mathit{R}\prime = {\mathit{R}}_0 \times {\displaystyle \frac{I_0}{I\prime }}={\mathit{R}}_0 \times {\displaystyle \frac{E\prime }{E_0}},$$

(2)

$$\mathit{E}\prime ={\mathit{E}}_0 \times {\displaystyle \frac{I_0}{I\prime }}.$$

(3)

Based on the derived relationship, the electron beam’s energy is inversely proportional to the excitation current (I) of the dipole magnet. Therefore, the equation relating E and I can be expressed as: E = $A/I$, where A is a constant. The constant A was calculated as 511.35 (MeV·A), using the reference energy E₀ = 35.0 MeV given by the current 14.61 A. Thus, the conversion from the channel to the energy is as follows:

$$\mathit{E} = A/I={\displaystyle \frac{511.35}{13.62+0.18(ch-{ch}_0)}}.$$

(4)

Here, ch₀ denotes the channel that corresponds to the reference orbit. By performing a Gaussian fitting of the result from channel 5 shown in Figure 8a, the half-width of the fitting curve was calculated to be 0.27 A. Consequently, the energy resolution was calculated as 1.9% (0.27 A/14.53 A). The energy resolution of the conventional slit–Faraday cup method was approximately 1%, where a 60 mm long aluminum slit with a thickness of 3 mm was placed at a relatively distant location. In contrast, the SEE monitor was installed 3.5 cm from the extraction window. Although its resolution is lower than that obtained in the conventional method, it enables compact, single-shot, non-scanning measurements of the temporal evolution of the energy distribution.

4.2. Results of Single-Shot Measurement

The 2D maps of the single-shot measurement with and without FEL lasing are shown in Figure 9. The signal intensity increased over time, and the average signal intensity was observed to be highest at channel 5 without FEL lasing (Figure 9a). This is due to the electron current increase due to the backbombardment effect in the thermionic RF gun [20]. In addition, the fast transition of the electron beam energy at the beginning of the macro-pulse can be clearly captured. On the other hand, an increase in energy spread and a reduction in the average energy, similar to previous research [11], can be observed when the FEL was lasing (Figure 9b). The time-varying feature of the electron energy distribution can be clearly captured by the SEE monitor.

The typical temporal evolutions of the average energy with and without FEL lasing are shown in Figure 10a. The average energy started to change with FEL lasing at 3.25 μs. The extraction efficiency was evaluated based on Equation (1) and is shown in Figure 10b. The highest extraction efficiency was 5%. The extraction efficiency measured at the FEL wavelength of 11.4 µm under the thermionic operation by using the conventional method was 5.5% [11]. The obtained highest extraction efficiency (5%) at the FEL wavelength of 8.17 µm by using SEE monitor in this research was comparative with the previous one. The extraction efficiency monotonically increased until 6.5 µs and slightly decreased after 6.5 µs. This could be caused by a slight mismatch of the electron bunch interval against the optical cavity length. Real-time observation of these kinds of defects in the FEL lasing process is quite important for optimizing operational conditions of oscillator FELs.

5. Conclusions

In this research, a 1D array-type SEE monitor with 24 ribbon electrodes and 2 shielding electrodes was developed for the single-shot measurement of the extraction efficiency of KU-FEL. As the basic performance test, the signal voltages (S/N ratio) of the SEE monitor and the response time of the developed system were investigated. As a result, it was confirmed that the developed system has a sufficiently high S/N ratio and temporal resolution for single-shot measurements of the temporal evolution of energy distribution. The single-shot measurements of the temporal evolution of electron beam energy distribution with and without FEL lasing were performed. As a result, the time-varying energy distribution could be clearly captured using the developed system. The highest extraction efficiency was evaluated as 5% when the electron beam energy was 35 MeV and the FEL peak wavelength was 8.17 μm. The developed system can be used for single-shot measurement of the extraction efficiency of KU-FEL and other oscillator-type FEL facilities. The system makes it easy to find some defects in the FEL lasing process, which is quite sensitive against the temporal evolution of the electron bunch interval and the average energy in a macro-pulse. The developed SEE monitor will enable finer tuning of the oscillator FELs by adjusting the bunch interval and average energy with controlling the phase and amplitude of RF pulses fed to accelerating structures, paving the way for achieving higher extraction efficiency.