Bunch-by-bunch linac beam energy profile diagnostics using storage ring phase rotation and synchrotron radiation

This paper reports a new method of measuring the bunch-by-bunch energy profile of a linear accelerator (linac) electron beam, using a dual-sweep streak camera in the storage ring. This method records the evolution of the time profile in the storage ring following injection. A high-resolution linac beam energy profile is reconstructed from the time profile measured at one-quarter of a synchrotron oscillation period following injection. Conversion between the time profile and energy profile is based on results from a longitudinal phase space tracking simulation. By separating the streak camera images of the different bunches, a bunch-by-bunch energy profile can also be measured.

Figure 1

Layout of the SPring-8 linac, booster synchrotron, and NewSUBARU storage ring.

Figure 2

Layout of the BL6 mirrors and optical equipment before the streak camera.

Figure 3

Streak camera image (accumulation of 50 shots). (a) shows the longitudinal transient of the injected beam, and (b) shows the intensity profile $T({\tau }_1)$ at $T_{\mathrm{MEAS}}$. The pulse width is 0.25 ns and the rf voltage is 130 kV. The bunch charge is approximately 0.03 nC. The broken lines in (a) indicate the slice width (about $1.9\mu \mathrm{s}$) for (b).

Figure 4

Result of a particle tracking simulation for the NewSUBARU rf bucket ($V_{\mathrm{rf}}=130\mathrm{kV}$). A uniform rectangular distribution, with a 20 ps bunch length and 1.0% energy spread, are used for the linac beam. The center energy is $+0.3\%$. The thick lines represent the beam occupancy at injection ($t=0$) and measurement ($t=T_{\mathrm{MEAS}}=53.3\mu \mathrm{s}$).

Figure 5

The energy profile $D({\delta }_0)$ of the linac beam, as reconstructed from the time profile $T({\tau }_1)$ of Fig. 3b using Eqs. (4, 5).

Figure 6

Streak camera images with a 1 ns linac pulse width. (a) uses a normal operation voltage ($V_{\mathrm{rf}}=130\mathrm{kV}$) and (b) uses a high rf voltage ($V_{\mathrm{rf}}=400\mathrm{kV}$). The pulse contains three bunches, #1, #2, and #3, which are consecutively injected into the storage ring.

Figure 7

Streak camera image of bunches #1 and #2, for simulation results with (a) a normal operation voltage and (b) a high rf voltage. The specified beam parameters are 0.5% for the full energy spread and 20 ps for the bunch length. The image is a trace of 200 particles per bunch.

Figure 8

Energy profiles of the front bunch (#1) and rear bunch (#2) for various rf fields, using an ECS.

Figure 9

Nonlinearity correction functions for the rf bucket. The solid line is the scaling function $f_0^{-1}$, which is the initial energy displacement ${\delta }_0$ for the calculated time displacement ${\tau }_1$. The broken line is the density correction function, which is the differential of $f_0$ normalized by its value at ${\delta }_0=0$.

Figure 10

Energy profiles reconstructed from (a) multishot camera images (accumulations of 50 shots) and (b) a single-shot camera image. The slice width of the oscillation image along the slow-sweep axis is $2.0\mu \mathrm{s}$ (5 revolutions) in (a) and $8.1\mu \mathrm{s}$ (20 revolutions) in (b).

Figure 11

Calculated contributions of the nonlinear error effects in ${\delta }_0$. The solid line shows $\Delta {\delta }_{\Delta T_{1/4}}$, the contribution due to variation of $T_{1/4}$, as given by Eq. (b3). The broken line shows the contribution due to the initial bunch length $\Delta {\delta }_{\Delta {\tau }_0}$ as given by Eq. (b4).