Bunch Shape Monitor (BSM) Diagnostics

• BSM is a longitudinal beam diagnostic that converts the temporal structure of charged-particle bunches into phase-dependent signals using techniques like RF phase scans and secondary-electron emission.
• The method integrates data from BPMs, wire scanners, and Gaussian fitting to extract key parameters such as rms bunch length and longitudinal emittance with practical implications for accelerator tuning.
• Optical implementations, such as beamstrahlung-based monitors at colliders, offer a non-invasive alternative with sub-picosecond resolution for reconstructing the accurate longitudinal bunch profile.

Searching arXiv for recent and foundational work on Bunch Shape Monitors and related diagnostics. A Bunch Shape Monitor (BSM) is a longitudinal beam diagnostic that measures the time profile of a charged-particle bunch by converting temporal structure into a phase-dependent detectable signal. In the Fermilab Linac Transition section, the BSM measures the longitudinal time profile of a $400\ \mathrm{MeV}$, $\sim 25\ \mathrm{mA}$ $\mathrm{H}^-$ beam by using secondary electrons emitted from a negatively biased wire and phase-selective transmission through an RF deflector; the resulting phase scan yields the bunch profile in degrees at $201\ \mathrm{MHz}$ (Sharankova et al., 16 Sep 2025). In a broader diagnostic sense, the term has also been applied to non-invasive optical methods in which the time structure of beamstrahlung at an interaction point is used to infer bunch timing and length, as proposed for SuperKEKB with an ultrafast large-angle beamstrahlung monitor (Carlo et al., 2017). These implementations differ in hardware and observables, but both treat bunch shape as a longitudinal distribution reconstructed from a calibrated phase or delay scan.

1. Definition and diagnostic role

In the Fermilab implementation, the BSM is part of a diagnostic chain that reconstructs longitudinal emittance and longitudinal Courant–Snyder parameters by combining three measurements: BSM bunch length versus upstream cavity phase, Beam Position Monitor (BPM) phasing data, and transverse beam sizes from a wire scanner (WS) (Sharankova et al., 16 Sep 2025). The BSM therefore serves not merely as a profile monitor, but as the downstream observable in an RF-scan procedure analogous to a transverse quadrupole scan.

The essential role of the BSM in this setting is to provide a direct measurement of the longitudinal time profile in the middle of the linac. The measured rms bunch length downstream of the upstream Buncher cavity is compared with TraceWin simulations while varying the Buncher phase. Because the cavity introduces both longitudinal focusing and transverse RF defocusing, interpretation of the BSM signal requires correlation with independent BPM and WS diagnostics rather than isolated use of the BSM trace alone (Sharankova et al., 16 Sep 2025).

A broader interpretation of the BSM concept appears in collider diagnostics. For SuperKEKB, the ultrafast large-angle beamstrahlung monitor functions as a BSM by measuring the time profile of beamstrahlung emitted at the interaction point and mapping that profile to bunch timing and bunch length. This suggests that, in advanced accelerator diagnostics, “BSM” can denote either an interceptive secondary-electron instrument or a non-invasive optical monitor, provided that the device reconstructs the longitudinal bunch shape from a time-resolved observable (Carlo et al., 2017).

2. Secondary-electron BSM principle in the Fermilab Linac

The Fermilab BSM operates by driving a negatively biased tungsten wire at $-10\ \mathrm{kV}$ through the $\mathrm{H}^-$ beam. When $\mathrm{H}^-$ ions strike the wire, they liberate secondary electrons that retain the instantaneous time structure of the beam. These electrons are repelled from the wire toward a primary slit, enter an RF deflector cavity, and are transmitted through a secondary slit only when they are at zero phase with respect to the deflector RF. Electrons at other phases are deflected away from the acceptance. The transmitted electrons are then amplified in an electron multiplier tube, and the detected signal is proportional to electron density at the selected phase (Sharankova et al., 16 Sep 2025).

By scanning the RF phase of the deflector, the instrument maps the bunch longitudinal profile as a function of RF phase. In the reported setup, the RF deflector operates at the Drift Tube Linac (DTL) frequency, and longitudinal profiles are reported in degrees at $201\ \mathrm{MHz}$, with plots labeled “BSM phase (deg@201 MHz).” The practical observable is therefore the detected signal as a function of deflector phase, and the analysis extracts bunch length by fitting a Gaussian to the measured BSM curve at each upstream cavity setting (Sharankova et al., 16 Sep 2025).

No explicit deconvolution kernel or analytical response function is applied beyond the phase scan and Gaussian fitting, and the analysis assumes that BSM acceptance and electron optics remain stable over the scan. Time resolution, dynamic range, and geometric parameters such as wire diameter and slit dimensions are not specified. The phase-to-time conversion is, however, explicit: at $201.49\ \mathrm{MHz}$, $1^\circ$ corresponds to $\sim 25\ \mathrm{mA}$0, so an rms width $\sim 25\ \mathrm{mA}$1 in degrees converts as $\sim 25\ \mathrm{mA}$2 (Sharankova et al., 16 Sep 2025).

3. Fermilab Linac configuration and measurement workflow

The Fermilab Linac comprises a $\sim 25\ \mathrm{mA}$3 DTL with five Alvarez tanks to $\sim 25\ \mathrm{mA}$4 and an $\sim 25\ \mathrm{mA}$5 side-coupled linac to $\sim 25\ \mathrm{mA}$6. The Transition section contains two bunching cavities, termed Buncher and Vernier, four FODO quadrupoles, three wire scanners, a BSM, and BPMs. During the measurements, the $\sim 25\ \mathrm{mA}$7 beam current was approximately $\sim 25\ \mathrm{mA}$8–$\sim 25\ \mathrm{mA}$9. The BSM is located near WS D03 and downstream of the Buncher and Vernier. BPMs are placed upstream of each quadrupole, operate at $\mathrm{H}^-$0 as a second-harmonic system, and report horizontal and vertical positions together with relative RF phase (Sharankova et al., 16 Sep 2025).

The measurement sequence begins with BPM phasing scans. All downstream cavities are turned off to create a longitudinal drift, the Buncher phase is scanned over $\mathrm{H}^-$1 at $\mathrm{H}^-$2, and downstream BPM phase differences are recorded as a function of distance. Because BPM phases are not absolutely calibrated for cable and electronics delays, the analysis uses BPM differences and double subtraction. The ideal phase difference for a beam with speed $\mathrm{H}^-$3 over distance $\mathrm{H}^-$4 is

$\mathrm{H}^-$5

and small beam-velocity variations around $\mathrm{H}^-$6 give

$\mathrm{H}^-$7

The measured quantity is the double-subtracted phase

$\mathrm{H}^-$8

where the subscript $\mathrm{H}^-$9 denotes the value with the cavity off (Sharankova et al., 16 Sep 2025).

The energy modulation induced by the Transition-section cavity is approximated as

$201\ \mathrm{MHz}$0

and the measured BPM phase-difference curves versus scanned cavity phase are fitted to a cosine. The phase offset $201\ \mathrm{MHz}$1 corresponding to maximum acceleration is identified at the minimum of the $201\ \mathrm{MHz}$2 curve, after which the beam phase is taken as $201\ \mathrm{MHz}$3. The amplitude $201\ \mathrm{MHz}$4 is extracted from the linear increase of the cosine amplitude with BPM distance according to

$201\ \mathrm{MHz}$5

In the reported study, this procedure yielded $201\ \mathrm{MHz}$6 in $201\ \mathrm{MHz}$7 degrees and an inferred Buncher energy modulation amplitude of approximately $201\ \mathrm{MHz}$8 (Sharankova et al., 16 Sep 2025).

The BSM scans are then performed with the Vernier cavity turned off to preserve a drift after the Buncher. The Buncher phase is scanned, and at each phase and voltage the BSM deflector phase is scanned at $201\ \mathrm{MHz}$9 to obtain the longitudinal profile. Gaussian fits provide the rms bunch length. In parallel, WS D03 measures transverse profiles, which are also fit with Gaussians to obtain rms beam sizes in $-10\ \mathrm{kV}$0 and $-10\ \mathrm{kV}$1 versus Buncher phase (Sharankova et al., 16 Sep 2025).

4. Reconstruction of longitudinal phase space

The analysis uses TraceWin to connect the BSM bunch-length scan, BPM-derived Buncher settings, and WS transverse sizes to the incoming longitudinal phase space. Conceptually, the procedure is analogous to a quadrupole scan: the Buncher cavity supplies a tunable longitudinal focusing strength, and the downstream rms bunch length serves as the fit observable. The effective longitudinal focusing produced by the cavity is proportional to

$-10\ \mathrm{kV}$2

so portions of the $-10\ \mathrm{kV}$3 curve that are symmetric around $-10\ \mathrm{kV}$4 represent the same focusing strength. To make this explicit, the study replots the measured bunch length against $-10\ \mathrm{kV}$5 rather than cavity phase alone (Sharankova et al., 16 Sep 2025).

The longitudinal covariance matrix is written as

$-10\ \mathrm{kV}$6

with longitudinal Twiss parameters

$-10\ \mathrm{kV}$7

In practice, TraceWin varies $-10\ \mathrm{kV}$8 upstream of the Buncher to match the measured $-10\ \mathrm{kV}$9 versus $\mathrm{H}^-$0 curve at the BSM location. The paper does not provide an explicit RF-gap transport matrix; instead, the dependence of energy gain on $\mathrm{H}^-$1 and focusing on $\mathrm{H}^-$2 is treated as sufficient for the scan parameterization used in TraceWin (Sharankova et al., 16 Sep 2025).

The workflow is therefore a three-diagnostic inference chain. BPM scans determine cavity phase and amplitude relative to the beam, BSM scans provide rms bunch length as a function of the calibrated focusing strength, and WS scans quantify the transverse RF defocusing produced by the same cavity fields. This suggests that the BSM reconstruction is only as reliable as the consistency among longitudinal phasing, transverse optics, and the simulation model that links them (Sharankova et al., 16 Sep 2025).

5. Measured performance, reconstructed parameters, and systematic limits

The BSM-measured rms bunch length at the Fermilab location varied from about $\mathrm{H}^-$3 to $\mathrm{H}^-$4 at $\mathrm{H}^-$5 over the phase scan, corresponding to approximately $\mathrm{H}^-$6–$\mathrm{H}^-$7 rms. As designed, the minimum bunch length occurred near $\mathrm{H}^-$8 (Sharankova et al., 16 Sep 2025).

Representative profiles reported in the study are summarized below.

Beam phase and voltage	RMS width at 201 MHz	Time-equivalent rms
$\mathrm{H}^-$9, $\mathrm{H}^-$0	$\mathrm{H}^-$1	$\mathrm{H}^-$2
$\mathrm{H}^-$3, $\mathrm{H}^-$4	$\mathrm{H}^-$5	$\mathrm{H}^-$6
$\mathrm{H}^-$7, $\mathrm{H}^-$8	$\mathrm{H}^-$9	$201\ \mathrm{MHz}$0

Fitting the measured bunch-length scan with TraceWin, using the BPM-derived Buncher parameters, yielded longitudinal parameters that were reported as consistent across two data sets, dated 12-18-2024 and 02-12-2025. The best-fit values were an rms normalized longitudinal emittance $201\ \mathrm{MHz}$1, $201\ \mathrm{MHz}$2, and $201\ \mathrm{MHz}$3. One figure annotates $201\ \mathrm{MHz}$4 as $201\ \mathrm{MHz}$5, while the text summarizes $201\ \mathrm{MHz}$6; both are presented in the source description, and no uncertainties are quoted for the reconstructed longitudinal parameters (Sharankova et al., 16 Sep 2025).

The same cavity fields that focus longitudinal motion produce transverse defocusing. In a thin-lens description, the longitudinal and transverse focal strengths have opposite sign, and this was observed experimentally: the smallest transverse WS sizes occurred at cavity phases where the bunch length was longest, consistent with strong transverse defocusing when longitudinal focusing was strongest. TraceWin reproduced the trend but differed by more than two standard deviations in magnitude in the $201\ \mathrm{MHz}$7 plane, indicating remaining optics or calibration uncertainties. The transverse effect also appeared in the BSM integral signal, with fewer electrons collected when the beam was transversely defocused into the BSM acceptance (Sharankova et al., 16 Sep 2025).

Several systematic limitations constrain interpretation. BPM phases are only relatively calibrated, so absolute phase is not directly known and must be eliminated through double subtraction. The Buncher RF regulation limited the maximum gradient, preventing a full scan through the minimum and into overfocusing; compared with a standard quadrupole scan, this reduces the accuracy of the longitudinal reconstruction. The BSM integral signal depends on transverse focusing and electron-optics acceptance, which couples a nominally longitudinal measurement to transverse optics. Detailed BSM resolution, acceptance, and electron-optics calibration are not given, and the analysis relies on Gaussian fits and relative comparisons rather than an explicit instrument-response model (Sharankova et al., 16 Sep 2025).

6. Optical BSM concept at colliders: the beamstrahlung-based monitor

A distinct BSM architecture was proposed for $201\ \mathrm{MHz}$8 colliders in the form of the Ultrafast Large Angle Beamstrahlung Monitor. In this system, the measured quantity is not secondary-electron current but the time profile of beamstrahlung emitted during beam–beam collision at the interaction point. The observed beamstrahlung power $201\ \mathrm{MHz}$9 encodes the longitudinal overlap of the two bunches, modulated by crossing geometry, transverse beam sizes, observation angle, and polarization. In a pure-overlap picture, the intensity envelope can be represented as

$201.49\ \mathrm{MHz}$0

and for Gaussian longitudinal profiles the envelope becomes Gaussian with rms width $201.49\ \mathrm{MHz}$1 and mean shifted by the timing offset $201.49\ \mathrm{MHz}$2 (Carlo et al., 2017).

The SuperKEKB proposal exploits polarization-dependent asymmetry. The $201.49\ \mathrm{MHz}$3-polarized component of the beamstrahlung pulse exhibits skewness that is approximately linear in the normalized longitudinal separation $201.49\ \mathrm{MHz}$4, with slope $201.49\ \mathrm{MHz}$5, whereas the $201.49\ \mathrm{MHz}$6-polarized component remains essentially symmetric. Operationally, the measured $201.49\ \mathrm{MHz}$7-polarized pulse is fit for rms width and skewness; the skewness gives the relative timing offset and the width gives bunch length, after correction for instrument resolution and known target-bunch length (Carlo et al., 2017).

Temporal measurement is performed by sum-frequency generation in a nonlinear crystal. A synchronized femtosecond laser pulse samples the beamstrahlung pulse, and scanning the laser delay $201.49\ \mathrm{MHz}$8 yields

$201.49\ \mathrm{MHz}$9

so that the recorded upconverted-photon rate reproduces the beamstrahlung time profile convolved with the instrument response. For a $1^\circ$0 BBO crystal and an $1^\circ$1, $1^\circ$2 Ti:Sapphire laser, the estimated overall time resolution is $1^\circ$3, which is $1^\circ$4 of a $1^\circ$5 beamstrahlung pulse. The proposal further estimates approximately $1^\circ$6 upconverted photons per second, an acquisition time of $1^\circ$7–$1^\circ$8 minutes for a full scan, statistical timing precision of about $1^\circ$9, and overall timing and length accuracy of approximately $\sim 25\ \mathrm{mA}$00 (Carlo et al., 2017).

Within the source material, this optical monitor is explicitly compared with traditional BSMs. Conventional BSMs, including secondary-electron devices in hadron linacs, RF deflecting cavities, and streak cameras, are described as invasive or as requiring dedicated beamline hardware, as measuring away from the interaction point with associated transport-systematics, and as having time resolutions of order $\sim 25\ \mathrm{mA}$01 for streak cameras. The beamstrahlung-based optical BSM is instead non-invasive, operates during physics runs, measures directly at the interaction point, and achieves sub-picosecond resolution. A plausible implication is that the term “Bunch Shape Monitor” now spans a diagnostic class defined more by reconstruction target—longitudinal bunch structure—than by a single hardware principle (Carlo et al., 2017).