Temporal profile monitor based on electro-optic spatial decoding for low-energy bunches

The measurement of electron bunch temporal profile is one of the key diagnostics in accelerators, especially for ultrashort bunches. The electro-optic (EO) technique enables the precise longitudinal characterization of bunch electric field in a single-shot and nondestructive way, which can simultaneously obtain and analyze the time jitter between the electron bunch and the synchronized laser. An EO monitor based on spatial decoding for temporal profile measurement and timing jitter recoding has recently been demonstrated and analyzed in depth for low-energy bunches at the Tsinghua Thomson scattering X-ray source. A detailed description of the experimental setup and measurement results are presented in this paper. An EO signal as short as 82 fs (rms) is observed with $100\mu \mathrm{m}$ gallium phosphide for a 40 MeV electron bunch, and the corresponding length is 106 fs (rms) with $300\mu \mathrm{m}$ zinc telluride. Owing to the field-opening angle, we propose a method to eliminate the influence of energy factor for bunches with low energy, resulting in a bunch length of $\sim 60\mathrm{fs}$ (rms). The monitor is also successfully applied to measure time jitter with approximately 10 fs accuracy. The experiment environment is proved to be the main source of the slow drift, which is removed using feedback control. Consequently, the rms time jitter decreases from 430 fs to 320 fs.

Figure 1

Schematic description of spatial decoding EO detection (top view).

Figure 2

Schematic layout of the TTX beam line and laser system.

Figure 3

Experimental setup in vacuum interaction chamber and laser transmission path for the spatial decoding EO detection (top view).

Figure 4

Measurements of the single-shot spatial decoding EO detection at crossed polarization with $300\mu \mathrm{m}$ ZnTe. (a) Typical raw CCD image of the probe laser profile. The left end of the laser is stopped by the crystal mounting structure, and the right is cut by the crystal edge. (b) Signal image with laser background subtracted. (c) The corresponding projected distribution of images (a) and (b). The projection of the laser profile without electron beam is plotted in dotted black curve. The bunch head is on the left. The EO trace shows a narrow spike followed by a long decreasing tail.

Figure 5

Time calibration for the EO measurements. Each data point represents the peak position in the laser profile of a single-shot EO signal for different laser delays. The red curve represents the slow drift of time jitter, and the linear fit in black line provides a calibration factor of $13.65\mathrm{fs}/\mathrm{pixel}$.

Figure 6

EO signal amplitude of the same bunch at varying distance between laser pulse and electron beam. The amplitude of EO signal decreases rapidly with the increase of distance. The calculated data in blue square is in good agreement with the measured point. A fit according to the experimental data provides the correction basis. Continuous change of distance is realized by adjusting laser delay and taking advantage of the relative time jitter between laser pulse and electron beam.

Figure 7

Correction of spatial decoding EO measurement by distance between probe laser and electron bunch. The bunch head is on the left side. Signal-to-noise ratio is relatively low for small signal amplitude with long bunch. EO signals before and after correction are shown in blue and black lines, respectively. The TDS measurement in dotted red line is taken as the reference of the real bunch distribution.

Figure 8

Measurements of two different bunch longitudinal profiles by transverse deflecting structure (TDS) and spatial decoding EO detection with $300\mu \mathrm{m}$ ZnTe and $100\mu \mathrm{m}$ GaP. Three measured results under the same accelerator conditions for long bunches (a) and short bunches with tiny inner structure (b). The bunch head is on the left. At the bottom of the drawing (a), the second peak located at 2.2 ps behind the main one is caused by the internal reflection of THz pulse in the GaP crystal.

Figure 9

Single-shot spatial decoding EO measurement of overcompressed twin bunch with $100\mu \mathrm{m}$ GaP. Bunch head is on the left. A Gaussian fit to the front bunch is shown in red line at an rms length of 82 fs. For GaP, broadening of Coulomb field is the main reason of resolution limitation, and an rms bunch duration of 64 fs is obtained by a deconvolution shown in green dashed line.

Figure 10

Measured bunch length (red point) for the same bunches at varying distance with $300\mu \mathrm{m}$ ZnTe (a) and $100\mu \mathrm{m}$ GaP (b) and the calculated curve of Coulomb field width (blue line) with distance for different bunch lengths. Inset: deviation between measured results and calculated curve of different bunch lengths. According to the minimum deviation, the obtained bunch length is 61 fs and 68 fs for ZnTe and GaP, respectively. The large length jitter found at a far distance is mainly caused by the small signal amplitude. Note that measured data with distance under 2 mm is not considered in the deviation calculation for ZnTe.

Figure 11

Single-shot EO measurements of five consecutive electron bunches. (a) EO signal on the laser profile (background subtracted). Each bright section represents the signal, and its width and location indicate the bunch duration and relative arrival time, respectively. (b) Uncorrected longitudinal bunch profile delivered from the projection of the corresponding signal. The arrival time is defined as the time of the peak position determined by a Gaussian fit to the main peak.

Figure 12

Online bunch profile monitor. (a) Corrected longitudinal bunch profile of EO signal in Fig. 11 according to Eq. (7). (b) Directly measured bunch length of 500 consecutive shots. (c) Arrival time corresponding to the distance required in the following length calibration. An early arrival time means a far distance in the experimental setup. (d) Real bunch length obtained by the method in Sec. 4e.

Figure 13

(a) Beam arrival time measured in the EO experiments over an hour. The slow drift is shown in solid red line. A time jitter of 430 fs rms is found from the data. The resulting fast-term jitter is 270 fs rms without slow drift. (b) Frequency spectrum of the data point by Fourier transform in the low-frequency region. The frequencies of the five peaks are $f_1=0.000495\mathrm{Hz}$, $f_2=0.000908\mathrm{Hz}$, $f_3=0.00129\mathrm{Hz}$, $f_4=0.00252\mathrm{Hz}$, $f_5=0.00186\mathrm{Hz}$, and $f_6=0.00309\mathrm{Hz}$.

Figure 14

Temperature and humidity in the laser room and their Fourier transform. (a) In the long-time recording, temperature and humidity show a period of $\sim 13.2$ and $\sim 40.2\min$, respectively. (b) The frequencies of the four peaks are $f_1=0.000415\mathrm{Hz}$, $f_2=0.000844\mathrm{Hz}$, $f_3=0.00126\mathrm{Hz}$, and $f_4=0.00251\mathrm{Hz}$.

Figure 15

(a) Beam arrival time measured with feedback control in the EO experiments over an hour. The slow drift is shown in solid red line. The rms time jitter is decreased to 320 fs. The corresponding fast-term jitter without slow drift is, however, increased to 300 fs rms. (b) Frequency spectrum of arrival time by Fourier transform in the low frequency region. Vertical axis is set in the same scale with Fig. 13 for comparison. The peak frequency is $f_3=0.00129\mathrm{Hz}$.

Figure 16

Adjustment of the rf gun phase (blue line) and slow drift (red line) during the feedback process. The initial phase is 30°, and the variation range is approximately 4°.