- Email: [email protected]
Development of FEL and SASE in the far-infrared region at ISIR, Osaka University
Available online at www.sciencedirect.com
Infrared Physics & Technology 51 (2008) 371–374 www.elsevier.com/locate/infrared
Development of FEL and SASE in the far-infrared region at ISIR, Osaka University Goro Isoyama *, Ryukou Kato, Shigeru Kashiwagi, Tetsuya Igo, Yutaka Morio Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan Available online 4 March 2008
Abstract Free electron laser (FEL) and self-amplified spontaneous emission (SASE) are being developed in the far-infrared region using the L-band electron linac at the Institute of Scientific and Industrial Research (ISIR), Osaka University. The L-band linac was recently remodeled extensively not only for higher operational stability and reproducibility but also for high power operation of FEL. After commissioning of the linac, we first began SASE experiment with a newly-developed strong-focusing wiggler. Recently we began FEL experiment and obtained lasing with the high peak power at 70 lm again after a long break. Ó 2007 Elsevier B.V. All rights reserved. Keywords: Linac; Stability; FEL; SASE; Infrared
1. Introduction We have been developing a far-infrared FEL and conducting basic research on SASE in the infrared region in parallel using the L-band electron linac, the RF frequency of which is 1.3 GHz, at the Institute of Scientific and Industrial Research (ISIR), Osaka University. The FEL based on the L-band linac is suitable for operation in the longer wavelength region because the bunch length is approximately 20 ps and it is much longer than that of the S-band linac, typically a few ps, used in the majority of infrared FELs. Since the first lasing of the FEL at 32–40 lm [1,2], we conducted research and development to expand the wavelength region towards the longer wavelength side and finally obtained lasing at 150 lm [3]. The maximum duration of the macropulse accelerated with the L-band linac was approximately 2 ls and due to the short pulse duration, the number of FEL amplification times was limited to approximately 50 only, so that we could not obtain power saturation of the FEL. Since the output power was low and unstable, the FEL could not be used for applica*
Corresponding author. Tel.: +81 6 6879 8485; fax: +81 6 6879 8489. E-mail address: [email protected] (G. Isoyama).
1350-4495/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.infrared.2007.12.030
tions, so that we suspended the development of FEL for a while. The L-band linac can accelerate a high-intensity single-bunch beam of 91 nC at maximum by use of the three-stage sub-harmonic buncher system. We have been conducting SASE experiments in the far-infrared region with the high-intensity single-bunch beam and the 1.96 m long wiggler for the FEL [4–6]. Recently, we extensively remodeled the L-band linac for higher stability and reproducibility of operation [7]. Taking advantage of this opportunity, we added a new operation mode for FEL in which the macropulse duration can be extended to 6 ls, three times longer than the previous value, so that the FEL can be operated at a high power level reaching saturation. We have re-commissioned the linac and the FEL. We will report the recent progress of the FEL and SASE experiments conducted with the L-band linac. 2. Renewed linac The linac is comprised of a thermionic electron gun, a three-stage sub-harmonic buncher system with two 108 MHz RF cavities and a 216 MHz cavity, a pre-buncher, a buncher, and a 3 m long accelerating tube. The last three components are excited by a single klystron of a
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
1.3 GHz frequency with the maximum peak power of 30 MW, the pulse duration of 4 ls, and the repetition frequency of 60 Hz in the normal mode. Another mode of operation is the long pulse mode for FEL, in which the RF pulse duration is extended to 8 ls with the output power of 25 MW and a repetition frequency of 30 Hz. The maximum acceleration energy of the linac is 40 MeV in the normal mode with no beam loading, while it is slightly lower in the multi-bunch mode due to the lower klystron power in the long pulse mode. Stability and reproducibility of linac operation are crucial for advanced studies with the linac and all the possible measures have been taken in the remodeling to reduce fast fluctuations and long-term drifts. Most influential power supply for the stability is the klystron modulator and it is designed and fabricated to reduce pulse-to-pulse fluctuations to less than 0.1%. A measured value is 2r = 0.06%. Voltage ripples on the flat top of the square pulse applied to the klystron should be less than 0.1% in order to make the amplitude and the phase of the RF power from the klystron constant over the macropulse duration. The undulation of the flat top in the normal mode is measured to be 0.12% (peak to peak), which is slightly larger than the design goal but close to it. 3. SASE experiment SASE has characteristics different from FEL in terms of the time structure and the peak power and hence SASE can be used as another light source in the far-infrared region, though it has not been realized. SASE is suitable for generating light in the longer wavelength region because the electron beam acts as an optical guide in the wiggler and there is no diffraction loss of light inevitable to the open optical resonator. We began basic study of SASE in the far-infrared region with a conventional wiggler used for the FEL. The period length of the wiggler is 6 cm and the number of periods is 32. We first measured characteristics of the single bunch electron beam used for SASE and then estimated the SASE performance based on the one-dimensional (1D) theory of SASE using the measured characteristics of the electron beam. It is concluded that the 1D theory is applicable for the electron energy lower than 20 MeV. The estimated parameters of SASE are as follows; the FEL parameter q = 0.02, the power gain length LG = 0.14 m, and the power gain Pout/Pin 106, so that the high intensity signal of SASE was expected. We conducted SASE experiments with the existing FEL system and an electron beam of the energy 11–14 MeV, the energy spread of 2–4%, charge per bunch <20 nC, the peak current >1 kA, and the normalized emittance of 150–200 p lm rad. We observed strong light signal when the single bunch beam went through the wiggler and measured the intensity as a function of the wiggler gap [4]. The measured functional dependence is well reproduced by the 1D theory of SASE and it is concluded that the measured strong light is SASE. We
then measured wavelength spectra of SASE and observed the second and the third harmonic peaks, which are due to non-linear harmonic generation of SASE, as well as the fundamental peak [5,6]. We have proposed and developed the edge-focusing wiggler, which is a Halbach type wiggler composed of permanent magnet blocks with edge angles, to focus the electron beam for FEL and SASE [8,9]. First, we made a 5-period model of the edge-focusing wiggler to demonstrate its performance and then we have developed a strong focusing (SF) wiggler based on the edge-focusing scheme. The main parameters of the SF wiggler are the same as those of the conventional wiggler used for our FEL and SASE experiments. The maximum peak magnetic field is 0.39 T, which is slightly higher than before, at the minimum gap of 3 cm. The wiggler consists of four FODO cells for strong focusing and the cell length is 0.48 m. The focusing and the defocussing units are single wiggler periods with edge angles ±5° and the drift space is four normal wiggler periods. The maximum field gradients are ±3.2 T/m at the gap of 3 cm. The SF wiggler has been installed on the FEL beam line in order to enhance not only the FEL parameter for SASE but also the gain of FEL. We have begun SASE experiments in the far-infrared region with the SF wiggler. Fig. 1 shows the wavelength spectrum of SASE measured first time with the SF wiggler. The energy and the energy spread of the single bunch beam is 10.3 MeV and 2–3%, respectively, charge per bunch is 30 nC, and the wiggler gap is 32 mm, for which the peak magnetic field is 0.354 T and the Kvalue is 1.4. It is a raw spectrum and the wavelength dependence of efficiency in the measurement system is not corrected. The intensity of the fundamental peak at 218 lm is as strong as or stronger than that with the conventional wiggler, and the second harmonic and the third harmonic peaks at 110 and 74 lm, respectively, due to non-linear harmonic generation of SASE are clearly seen. Intensity ratios of the second and the third harmonic peaks to the fundamental peak roughly estimated with the efficiency of the measurement system are 0.03 and 0.1, respectively.
30 Detector output [mV]
372
25
3rd harmonic (74μm)
20
2nd harmonic (~110μm)
fundamental (218μm)
15 10 5 0 50
100
150 Wavelength [μm]
200
Fig. 1. Wavelength spectrum of SASE with the strong focusing wiggler.
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
4. Development of FEL The infrared FEL system is the same as before except for the wiggler. The optical resonator is a concentric resonator consisting of two spherical mirrors with different curvature radii. It is 5.531 m long and four optical pulses bounce back and forth in the cavity. The two mirrors are installed asymmetrically to the wiggler due to physical limitations by the magnets and other components in the FEL beam line and the curvature radii are chosen so that the waist of the stored light is located at the center of the wiggler. The FEL light is taken out through a hole at the center of the upstream mirror and it is led with mirrors to the measurement room through the evacuated optical transport line, where the low vacuum in the optical transport line is separated with a synthetic diamond window from the high vacuum in the FEL beam line. The FEL light is analyzed with a grating monochromator and detected with a Ge:Ga photoconductive detector cooled with liquid helium. In FEL experiments, the linac is operated in the multibunch mode, in which the second 108 MHz and the 216 MHz cavities of the sub-harmonic buncher system are switched on and the 1.3 GHz RF power of 8 ls duration is provided to the accelerating structures so that the multi-bunch beam with time intervals of 9.25 ns between bunches is accelerated. In order to make the peak of the energy spectrum of the multi-bunch beam sharp and to make intervals between bunches precisely equal, the amplitude and the phase of the RF power must be constant over the macropulse period. It is said that the amplitude and the phase of the output power from a klystron do not vary in a pulse, provided that the flat top of the high-voltage square pulse applied to the klystron is flat and constant. The measured amplitude and the phase of the RF power, however, vary by 14.8% and by 13.5° in 7.2–7.4 ls, respectively, though undulation of the flat top is less than 0.12%. The variations are much larger than our expectation of <1% in amplitude and <1° in phase. Since reproducibility and stability of the high-voltage pulse from the klystron modulator are high and accordingly those of the amplitude and the phase variations of the RF power from the klystron are also high, we have adopted the feed-forward control to compensate for these variations. We, for the moment, use an analogue phase shifter with PIN diodes for phase control and an I–Q modulator for amplitude control. A part of the RF power is taken at a directional coupler on a waveguide from the klystron to the linac. The amplitude of the RF power is measured with a diode detector and the phase is measured with a phase detector, both of which have sufficiently fast response times, and the output signals are analyzed with a digital oscilloscope. A personal computer calculates waveforms to be applied to the phase shifter and the variable attenuator so that both amplitude and phase variations become constant and flat over the RF pulse. Among the various components in the RF line, the phase
373
shifter and a 300 W transistor amplifier for driving the klystron have relatively long response times. In order to compensate for the slow response times, we use overdrive technique to the control devices with a function generator. The response time and the delay time are measured for the actual system and they are s = 50 ns and Dt = 540 ns for the amplitude control and s = 780 ns and Dt = 640 ns for the phase control. The amplitude and the phase variations are reduced to 0.5% and 0.3°, respectively, after compensation made iteratively several times. The long pulse electron beam is accelerated with the corrected RF power and energy spectra are measured with a momentum analyzer by changing the macropulse duration. The filling time of the accelerating tube is approximately 2 ls and energy spectra for the electron beams with pulse durations shorter than it have broader peaks due to the transient effect, while the sharp peaks grow up on the lower energy side in the energy spectra of electron beams longer than the filling time or 2 ls. Thus the multi-bunch mode for FEL is successfully commissioned. We recently conducted FEL experiment again after long suspension with the renewed linac and the strong focusing wiggler. Fig. 2 shows time spectra of beam intensities measured with beam current monitors at the exit of the linac, and the entrance and the exit of the wiggler, respectively in the first upper row to the third row, as well as the light intensity measured with a Ge:Ga detector in the lowest row. The electron energy is 17.4 MeV and the beam current injected from the electron gun is 220 mA. The magnet gap of the wiggler is set at 32 mm. The electron beam accelerated with the linac is 6.5 ls long as can be seen in the first row in Fig. 2, while the electron beam measured at the entrance and the exit of the wiggler is approximately 5 ls long and the front part is lacking compared with the beam intensity measured at the exit of the linac, because the front part of the electron beam has higher energy due to the transient effect in the accelerating tube of the linac and it is scraped off in the FEL beam line. The bottom row shows evolution of the light intensity of FEL and it seems that the FEL reaches the power saturation. The peak power in the macropulse is roughly estimated to be 200 W. We have measured the wavelength spectrum of FEL with a grating monochromator and the peak is seen at 70 lm.
Fig. 2. Time structures of the multi-bunch beam (upper three rows) and the FEL light (bottom).
374
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
Thus, after a long break, we have succeeded in re-commissioning of the FEL with the renewed linac and have observed lasing at 70 lm with the high peak power. References [1] S. Okuda et al., Nucl. Instr. Meth. A 358 (1995) 244. [2] S. Okuda et al., Nucl. Instr. Meth. A 375 (1996) 329.
[3] [4] [5] [6] [7]
R. Kato et al., Nucl. Instr. Meth. A 445 (2000) 169. R. Kato et al., Nucl. Instr. Meth. A 445 (2000) 164. R. Kato et al., Nucl. Instr. Meth. A 475 (2001) 334. R. Kato et al., Nucl. Instr. Meth. A 483 (2002) 46. G. Isoyama et al., in: Proceedings of APAC 2004, Gyeongju, Korea, March 22–26, 2004, p. 237. [8] G. Isoyama et al., Nucl. Instr. Meth. A 507 (2003) 234. [9] S. Kashiwagi et al., Nucl. Instr. Meth. A 528 (2004) 203.
Infrared Physics & Technology 51 (2008) 371–374 www.elsevier.com/locate/infrared
Development of FEL and SASE in the far-infrared region at ISIR, Osaka University Goro Isoyama *, Ryukou Kato, Shigeru Kashiwagi, Tetsuya Igo, Yutaka Morio Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan Available online 4 March 2008
Abstract Free electron laser (FEL) and self-amplified spontaneous emission (SASE) are being developed in the far-infrared region using the L-band electron linac at the Institute of Scientific and Industrial Research (ISIR), Osaka University. The L-band linac was recently remodeled extensively not only for higher operational stability and reproducibility but also for high power operation of FEL. After commissioning of the linac, we first began SASE experiment with a newly-developed strong-focusing wiggler. Recently we began FEL experiment and obtained lasing with the high peak power at 70 lm again after a long break. Ó 2007 Elsevier B.V. All rights reserved. Keywords: Linac; Stability; FEL; SASE; Infrared
1. Introduction We have been developing a far-infrared FEL and conducting basic research on SASE in the infrared region in parallel using the L-band electron linac, the RF frequency of which is 1.3 GHz, at the Institute of Scientific and Industrial Research (ISIR), Osaka University. The FEL based on the L-band linac is suitable for operation in the longer wavelength region because the bunch length is approximately 20 ps and it is much longer than that of the S-band linac, typically a few ps, used in the majority of infrared FELs. Since the first lasing of the FEL at 32–40 lm [1,2], we conducted research and development to expand the wavelength region towards the longer wavelength side and finally obtained lasing at 150 lm [3]. The maximum duration of the macropulse accelerated with the L-band linac was approximately 2 ls and due to the short pulse duration, the number of FEL amplification times was limited to approximately 50 only, so that we could not obtain power saturation of the FEL. Since the output power was low and unstable, the FEL could not be used for applica*
Corresponding author. Tel.: +81 6 6879 8485; fax: +81 6 6879 8489. E-mail address: [email protected] (G. Isoyama).
1350-4495/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.infrared.2007.12.030
tions, so that we suspended the development of FEL for a while. The L-band linac can accelerate a high-intensity single-bunch beam of 91 nC at maximum by use of the three-stage sub-harmonic buncher system. We have been conducting SASE experiments in the far-infrared region with the high-intensity single-bunch beam and the 1.96 m long wiggler for the FEL [4–6]. Recently, we extensively remodeled the L-band linac for higher stability and reproducibility of operation [7]. Taking advantage of this opportunity, we added a new operation mode for FEL in which the macropulse duration can be extended to 6 ls, three times longer than the previous value, so that the FEL can be operated at a high power level reaching saturation. We have re-commissioned the linac and the FEL. We will report the recent progress of the FEL and SASE experiments conducted with the L-band linac. 2. Renewed linac The linac is comprised of a thermionic electron gun, a three-stage sub-harmonic buncher system with two 108 MHz RF cavities and a 216 MHz cavity, a pre-buncher, a buncher, and a 3 m long accelerating tube. The last three components are excited by a single klystron of a
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
1.3 GHz frequency with the maximum peak power of 30 MW, the pulse duration of 4 ls, and the repetition frequency of 60 Hz in the normal mode. Another mode of operation is the long pulse mode for FEL, in which the RF pulse duration is extended to 8 ls with the output power of 25 MW and a repetition frequency of 30 Hz. The maximum acceleration energy of the linac is 40 MeV in the normal mode with no beam loading, while it is slightly lower in the multi-bunch mode due to the lower klystron power in the long pulse mode. Stability and reproducibility of linac operation are crucial for advanced studies with the linac and all the possible measures have been taken in the remodeling to reduce fast fluctuations and long-term drifts. Most influential power supply for the stability is the klystron modulator and it is designed and fabricated to reduce pulse-to-pulse fluctuations to less than 0.1%. A measured value is 2r = 0.06%. Voltage ripples on the flat top of the square pulse applied to the klystron should be less than 0.1% in order to make the amplitude and the phase of the RF power from the klystron constant over the macropulse duration. The undulation of the flat top in the normal mode is measured to be 0.12% (peak to peak), which is slightly larger than the design goal but close to it. 3. SASE experiment SASE has characteristics different from FEL in terms of the time structure and the peak power and hence SASE can be used as another light source in the far-infrared region, though it has not been realized. SASE is suitable for generating light in the longer wavelength region because the electron beam acts as an optical guide in the wiggler and there is no diffraction loss of light inevitable to the open optical resonator. We began basic study of SASE in the far-infrared region with a conventional wiggler used for the FEL. The period length of the wiggler is 6 cm and the number of periods is 32. We first measured characteristics of the single bunch electron beam used for SASE and then estimated the SASE performance based on the one-dimensional (1D) theory of SASE using the measured characteristics of the electron beam. It is concluded that the 1D theory is applicable for the electron energy lower than 20 MeV. The estimated parameters of SASE are as follows; the FEL parameter q = 0.02, the power gain length LG = 0.14 m, and the power gain Pout/Pin 106, so that the high intensity signal of SASE was expected. We conducted SASE experiments with the existing FEL system and an electron beam of the energy 11–14 MeV, the energy spread of 2–4%, charge per bunch <20 nC, the peak current >1 kA, and the normalized emittance of 150–200 p lm rad. We observed strong light signal when the single bunch beam went through the wiggler and measured the intensity as a function of the wiggler gap [4]. The measured functional dependence is well reproduced by the 1D theory of SASE and it is concluded that the measured strong light is SASE. We
then measured wavelength spectra of SASE and observed the second and the third harmonic peaks, which are due to non-linear harmonic generation of SASE, as well as the fundamental peak [5,6]. We have proposed and developed the edge-focusing wiggler, which is a Halbach type wiggler composed of permanent magnet blocks with edge angles, to focus the electron beam for FEL and SASE [8,9]. First, we made a 5-period model of the edge-focusing wiggler to demonstrate its performance and then we have developed a strong focusing (SF) wiggler based on the edge-focusing scheme. The main parameters of the SF wiggler are the same as those of the conventional wiggler used for our FEL and SASE experiments. The maximum peak magnetic field is 0.39 T, which is slightly higher than before, at the minimum gap of 3 cm. The wiggler consists of four FODO cells for strong focusing and the cell length is 0.48 m. The focusing and the defocussing units are single wiggler periods with edge angles ±5° and the drift space is four normal wiggler periods. The maximum field gradients are ±3.2 T/m at the gap of 3 cm. The SF wiggler has been installed on the FEL beam line in order to enhance not only the FEL parameter for SASE but also the gain of FEL. We have begun SASE experiments in the far-infrared region with the SF wiggler. Fig. 1 shows the wavelength spectrum of SASE measured first time with the SF wiggler. The energy and the energy spread of the single bunch beam is 10.3 MeV and 2–3%, respectively, charge per bunch is 30 nC, and the wiggler gap is 32 mm, for which the peak magnetic field is 0.354 T and the Kvalue is 1.4. It is a raw spectrum and the wavelength dependence of efficiency in the measurement system is not corrected. The intensity of the fundamental peak at 218 lm is as strong as or stronger than that with the conventional wiggler, and the second harmonic and the third harmonic peaks at 110 and 74 lm, respectively, due to non-linear harmonic generation of SASE are clearly seen. Intensity ratios of the second and the third harmonic peaks to the fundamental peak roughly estimated with the efficiency of the measurement system are 0.03 and 0.1, respectively.
30 Detector output [mV]
372
25
3rd harmonic (74μm)
20
2nd harmonic (~110μm)
fundamental (218μm)
15 10 5 0 50
100
150 Wavelength [μm]
200
Fig. 1. Wavelength spectrum of SASE with the strong focusing wiggler.
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
4. Development of FEL The infrared FEL system is the same as before except for the wiggler. The optical resonator is a concentric resonator consisting of two spherical mirrors with different curvature radii. It is 5.531 m long and four optical pulses bounce back and forth in the cavity. The two mirrors are installed asymmetrically to the wiggler due to physical limitations by the magnets and other components in the FEL beam line and the curvature radii are chosen so that the waist of the stored light is located at the center of the wiggler. The FEL light is taken out through a hole at the center of the upstream mirror and it is led with mirrors to the measurement room through the evacuated optical transport line, where the low vacuum in the optical transport line is separated with a synthetic diamond window from the high vacuum in the FEL beam line. The FEL light is analyzed with a grating monochromator and detected with a Ge:Ga photoconductive detector cooled with liquid helium. In FEL experiments, the linac is operated in the multibunch mode, in which the second 108 MHz and the 216 MHz cavities of the sub-harmonic buncher system are switched on and the 1.3 GHz RF power of 8 ls duration is provided to the accelerating structures so that the multi-bunch beam with time intervals of 9.25 ns between bunches is accelerated. In order to make the peak of the energy spectrum of the multi-bunch beam sharp and to make intervals between bunches precisely equal, the amplitude and the phase of the RF power must be constant over the macropulse period. It is said that the amplitude and the phase of the output power from a klystron do not vary in a pulse, provided that the flat top of the high-voltage square pulse applied to the klystron is flat and constant. The measured amplitude and the phase of the RF power, however, vary by 14.8% and by 13.5° in 7.2–7.4 ls, respectively, though undulation of the flat top is less than 0.12%. The variations are much larger than our expectation of <1% in amplitude and <1° in phase. Since reproducibility and stability of the high-voltage pulse from the klystron modulator are high and accordingly those of the amplitude and the phase variations of the RF power from the klystron are also high, we have adopted the feed-forward control to compensate for these variations. We, for the moment, use an analogue phase shifter with PIN diodes for phase control and an I–Q modulator for amplitude control. A part of the RF power is taken at a directional coupler on a waveguide from the klystron to the linac. The amplitude of the RF power is measured with a diode detector and the phase is measured with a phase detector, both of which have sufficiently fast response times, and the output signals are analyzed with a digital oscilloscope. A personal computer calculates waveforms to be applied to the phase shifter and the variable attenuator so that both amplitude and phase variations become constant and flat over the RF pulse. Among the various components in the RF line, the phase
373
shifter and a 300 W transistor amplifier for driving the klystron have relatively long response times. In order to compensate for the slow response times, we use overdrive technique to the control devices with a function generator. The response time and the delay time are measured for the actual system and they are s = 50 ns and Dt = 540 ns for the amplitude control and s = 780 ns and Dt = 640 ns for the phase control. The amplitude and the phase variations are reduced to 0.5% and 0.3°, respectively, after compensation made iteratively several times. The long pulse electron beam is accelerated with the corrected RF power and energy spectra are measured with a momentum analyzer by changing the macropulse duration. The filling time of the accelerating tube is approximately 2 ls and energy spectra for the electron beams with pulse durations shorter than it have broader peaks due to the transient effect, while the sharp peaks grow up on the lower energy side in the energy spectra of electron beams longer than the filling time or 2 ls. Thus the multi-bunch mode for FEL is successfully commissioned. We recently conducted FEL experiment again after long suspension with the renewed linac and the strong focusing wiggler. Fig. 2 shows time spectra of beam intensities measured with beam current monitors at the exit of the linac, and the entrance and the exit of the wiggler, respectively in the first upper row to the third row, as well as the light intensity measured with a Ge:Ga detector in the lowest row. The electron energy is 17.4 MeV and the beam current injected from the electron gun is 220 mA. The magnet gap of the wiggler is set at 32 mm. The electron beam accelerated with the linac is 6.5 ls long as can be seen in the first row in Fig. 2, while the electron beam measured at the entrance and the exit of the wiggler is approximately 5 ls long and the front part is lacking compared with the beam intensity measured at the exit of the linac, because the front part of the electron beam has higher energy due to the transient effect in the accelerating tube of the linac and it is scraped off in the FEL beam line. The bottom row shows evolution of the light intensity of FEL and it seems that the FEL reaches the power saturation. The peak power in the macropulse is roughly estimated to be 200 W. We have measured the wavelength spectrum of FEL with a grating monochromator and the peak is seen at 70 lm.
Fig. 2. Time structures of the multi-bunch beam (upper three rows) and the FEL light (bottom).
374
G. Isoyama et al. / Infrared Physics & Technology 51 (2008) 371–374
Thus, after a long break, we have succeeded in re-commissioning of the FEL with the renewed linac and have observed lasing at 70 lm with the high peak power. References [1] S. Okuda et al., Nucl. Instr. Meth. A 358 (1995) 244. [2] S. Okuda et al., Nucl. Instr. Meth. A 375 (1996) 329.
[3] [4] [5] [6] [7]
R. Kato et al., Nucl. Instr. Meth. A 445 (2000) 169. R. Kato et al., Nucl. Instr. Meth. A 445 (2000) 164. R. Kato et al., Nucl. Instr. Meth. A 475 (2001) 334. R. Kato et al., Nucl. Instr. Meth. A 483 (2002) 46. G. Isoyama et al., in: Proceedings of APAC 2004, Gyeongju, Korea, March 22–26, 2004, p. 237. [8] G. Isoyama et al., Nucl. Instr. Meth. A 507 (2003) 234. [9] S. Kashiwagi et al., Nucl. Instr. Meth. A 528 (2004) 203.