DAQ system

As shown in Fig. 1(a) (green shaded area), the collection and processing of the signals from the PMTs and MagneTOF detector are realized through a series of NIM modules, such as a fast timing amplifier (ORTEC FTA 820A) and constant-fraction discriminator (CAEN N605). The bias voltages for the PMTs and MagneTOF detector are provided by high-voltage (HV) power suppliers (CAEN N1470). The signals processed by these NIM modules are further transformed into TTL signals before being sent to a time-to-digital converter (TDC: ChronoLogic TimeTagger4-2G) with a time resolution of 500 ps.

The TDC has four independent input channels in addition to a trigger channel. The corresponding timing sequence for these channels is shown in Fig. 2. After receiving a trigger signal (T₀; trigger frequency: e.g., 100 Hz), a user-defined time window (i.e., Δ T < 10 ms) is opened for each input channel of TDC. Event signals from detectors arriving within Δ T are recorded and stored in the buffer zone of TDC. For each recorded event signal, information on its arrival time (time tagging) relative to T₀, input channel, and trigger number since the start of TDC are recorded and saved. The events that occurred outside Δ T are excluded.

Fig. 2Full-screen View

(Color online) Time sequence of the TDC data. Event signals arriving within the time window (Δ T) following each trigger signal (T₀: black TTL signal) are collected and then delivered to a buffer zone as a group (yellow), while event signals arriving out of the Δ T are discarded (red). The time stamp for each recorded event is also tagged relative to T₀.

For a typical experiment in this work, where the ion beam is produced in bunched mode by an offline laser ablation ion source [18], the timing sequence for each involved device is shown in Fig. 3. A 532-nm Nd: YAG pulsed laser (Litron TRLi 250-100) with a repetition rate of 100Hz is used to drive the ablation ion source. This pulsed laser (named ablation laser) is externally triggered by a TTL signal from a Quantum Composers (QC9528) digital-delay pulse generator ( QC). The TDC is then triggered by a TTL signal that was t’ μs later than the ablation laser pulse. At the beginning of a CLS experiment, t’ must be adjusted such that the ion beam pulses (approximately 10 μs duration [18] in the present work) appear within the TDC time window (Δ T), corresponding to the estimated time for the ion beam pulse passing in front of the PMTs. Thus, the sensitivity (or signal-to-noise ratio) of the CLS experiment can be significantly improved in comparison to that performed with a continuous ion beam [3, 5, 12, 18]. This is because of the limited background events recorded within the narrow TDC time window Δ T (e.g., approximately 100 μs [18]), out of the average period of 10 ms between two pulses. The background events can be further suppressed by gating on the narrow TOF gate (∼ 10 μs) of the ion bunch, as detailed in Sect. 3.1. However, when an ion beam is delivered in continuous mode, such as that in Ref. [19], the TDC is triggered by a high-frequency TTL signal (e.g., 1 kHz). To avoid any event loss, the time window for TDC is set as Δ T=1 ms. Therefore, the measurement of the HFS spectrum with a continuous beam involves recording all signals coming from the detectors, including all background signals that may result from events such as the random scattering laser light, de-excitation after non-resonant collision excitation, and PMT dark current.

Fig. 3Full-screen View

(Color online) Time sequence for the CLS experiment performed with a bunched ion beam produced by a laser ablation ion source [18]. A TTL signal (dark blue) with a repetition rate of 100 Hz generated from a Quantum Composers (QC) digital-delay pulse generator is used to trigger the pulsed ablation laser. The actual pulse (red) of the 532-nm laser used for ion beam production arrives 490 μs later, which is measured by a photodiode detector. The TOF of the ion beam (light blue) is measured in the photon detection region. To fully cover the ion beam pulse, the time window for TDC (green) is defined as 100 μs and delayed by 8 μs relative to the ablation laser pulse.

Voltage scanning system

As mentioned above, in a CLS experiment, the HFS spectra of the isotope of interest can be measured by counting the LIF photons as a function of the probing laser frequency or scanning voltage. In the former case, the frequency of the laser can be tuned using a laser frequency control software and by a high-precision wavelength meter [17]. The program for the control and DAQ system must synchronize the probing laser frequency and the corresponding LIF photons to achieve HFS measurements. The scanning speed and step size of the probing laser frequency are often limited by the laser system and must be handled well. In fact, rapid tuning of the probing laser frequency with a large step size and a wide frequency range may cause loss of the lock of the laser cavity, which will strongly affect the laser stability and the measurement of the HFS spectrum. Nevertheless, these limitations can be overcome by using the following Doppler tuning approach.

In this approach, the velocity of the ion beam is varied by changing the voltage applied to the voltage scanning electrode mounted upstream of the charge exchange cell, as shown in Fig. 1. The applied voltage is preliminarily provided by a USB device (USB-3106, Measurement Computing), which is remotely controlled by the control and DAQ system. The output voltage of the USB-3106 ranges from -10 V to 10 V with a typical slew rate of 1.2 v/μs and typical duration of 5 μs, allowing fast voltage scanning. The output signal of the USB device is then linearly amplified via a DC HV amplifier (Trek 623 B) with a gain of 1000, maximum voltage scanning range of ± 2 kV, and slew rate greater than 300 v/μs. The amplified voltage is then applied to the voltage-scanning electrode to change the velocity of the ions. The program also records the applied voltage in real time using a voltage divider (Ohm-labs, KV-10A) and digital multimeter (Keysight 34470A) [19], as shown in Fig. 1.

Synchronization of voltage scanning with TDC events

An important application of the control and DAQ system is the measurement of the HFS spectrum using the voltage scanning technique, which involves recording LIF photons as a function of the scanning voltage. Because the latter is repeatedly varied in small steps, its precise synchronization with photon recording is crucial. Naturally, one would expect that a straightforward method would involve changing the voltage at an equal step size δ V after equal intervals of time δ T, as illustrated by Method 1” in Fig. 4(a). However, the time required for the program to execute each read” function, which reads the data from the buffer zone of the TDC, depends on the actual computational load and is not necessarily synchronized to the external trigger or each voltage step. Thus, this method ( Method 1”) can lead to a mismatch between the read data and each voltage step, as indicated by black diamonds in Fig. 4(a). This results in a measured spectrum that deviates from the real statistics.

Fig. 4Full-screen View

(Color online) Two methods for the control and DAQ system for recording and plotting the HFS spectrum using voltage scanning: (a) changing the voltage at equal step size (δ V) after equal time intervals (δ T), (b) changing the voltage at equal step size (δ V) after executing equal times of the read” function (i.e. n×read” with n a fixed number), associated with the effective read of the events from TDC. The intervals between the black lines, red lines, and blue brackets represent the period of a trigger (e.g. Δ t = 1 ms for 1 kHz trigger frequency), voltage step (δ V), and data in the buffer zone of the TDC for executing the read” function, respectively.

The Method 2” shown in Fig. 4(b) was adopted to overcome this problem. In this method, the control and DAQ systems subjectively switch the voltage after executing a certain number of read” function (n×read” with n a fixed number). Each n× read” corresponds to a clearly fixed scan voltage, but not necessarily the same time period, as shown in Fig. 4(b). Thus, for each voltage step (δ V) or n× read” step, the counts must be normalized to the number of covered triggers (namely, counts/trigger or count rate) before being used to plot the HFS spectrum (further details are provided in Sect. 3.1).

HFS measurement using voltage scanning

The control and DAQ systems were first applied to the measurement of the optical spectrum of the stable ⁴⁰Ca⁺produced in the bunched mode by a laser ablation ion source. As shown at the top of Fig. 1(a), the bunched ⁴⁰Ca ions with varying velocity (scanning voltage) can be resonantly excited in the photon detection region because of the interaction with a continuous-wave probing laser at a fixed frequency. The scanning voltage in this measurement was set to a range of 25 V, with a step size and step duration of 0.25 V and 0.02 s, respectively. The LIF photons were detected by the PMTs and processed by the NIM electronics before being sent to the TDC (Fig. 1(b)). As depicted in Fig. 3, only signals arriving within the TDC time window (100 μs in this measurement) were recorded. Thus, the background signals that primarily arise from the scattered laser light and PMT dark counts, were significantly reduced. The TOF spectrum of the bunched beam was obtained by plotting the LIF photon counts as a function of their arrival time relative to the start of the TDC time window, as shown in Fig. 6(b)). The optical spectrum of ⁴⁰Ca was obtained by plotting the LIF photon counts arriving within the TOF gate ( indicated by the vertical red lines in Fig. 6(b)) as a function of the scanning voltage. Figure 6(a) presents the measured optical spectrum of ⁴⁰Ca⁺ after Doppler correction and voltage-frequency conversion according to Eq. 1.

Fig. 6Full-screen View

(Color online) (a) Optical spectrum of the ⁴⁰Ca⁺ isotope measured for the $4s^2{\text{S}}_{1/2}$ → 4p²P_3/2 (D2) ionic transition. The ⁴⁰Ca⁺ beam was produced in the bunched mode by an offline laser ablation ion source [18]. The spectrum in (a) is obtained by gating on (b) TOF spectrum of bunched ⁴⁰Ca⁺ beam. For this measurement, the Method 1” discussed in Fig. 4 and in Sect. 2.3 was used. The optical spectrum in (a) is plotted using both counts and count rate. See text for more details.

It should be noted that the spectra in Fig. 6(a) were measured using Method 1, as discussed in Sect. 2.3 and Fig. 4(a), indicating that the voltage was switched by an equal step size δV after an equal time interval δ T. This may produce a mismatch between the triggers and voltage at the edge of every voltage step [see Sect. 2.3 and Fig. 4(a)]. However, this does not cause significant problems in the measured optical spectrum, as shown in Fig. 6(a) owing to the relatively low count rate within the narrow TOF gate (∼ 10 μs) shown in Fig. 6(b), and hence a low computational load for the bunched ion mode.

However, “Method 1” causes severe problems in the HFS spectrum of isotopes measured using continuous beams, for example, ³⁸K isotopes from BRIF [19], because hundreds or thousands of times more background events will be recorded by the TDC. To ensure the correct synchronization of photon events with the corresponding voltage step, the Method 2” in Fig. 4(b) was adopted, which implies that the voltage change was initiated by the control and DAQ system after executing the read” function n times. An example of the recorded total number of triggers for each voltage step δ V is presented in Fig 7(a), that shows the variation in the reading time for executing the same number of read” functions. This produces a large fluctuation in the total counts for each voltage step (converted into the relative frequency in Fig 7(b)) owing to the reading time variation. In this case, only the normalized counts per trigger (or count rate) for each voltage step were meaningful. These were used to plot the HFS spectrum of the ³⁸K isotope (Fig 7(c)). For each scan, we calculated the counts per trigger (or count rate), for example, Mi. The corresponding error in the counts per trigger (or count rate) Mi should be handled based on $\Delta M_i=\frac{\Delta N_i}{K_i}=\frac{\sqrt{N_i}}{K_i}=\frac{\sqrt{M_i{K}_i}}{K_i}=\sqrt{\frac{M_i}{K_i}},$ (2) where Ni and Ki are the count and number of triggers (or measurement time), respectively, for i^th scan.

Fig. 7Full-screen View

(Color online) HFS spectrum of the radioactive ³⁸K isotope measured for the $4s^2{\text{S}}_{1/2}\to 4p^2{\text{P}}_{1/2}$ (D₁) atomic transition. The ³⁸K beam was provided in the continuous mode by the BRIF RIB facility [19]. The Method 2” in Fig. 4 and described in Sect. 2.3 was adopted for this HFS spectrum measurement. (a) Number of triggers for each voltage step. (b) HFS spectrum of ³⁸K plotted using counts and (c) the count rate. See text for more details.

Notably, although no significant difference in the resonance peak of ⁴⁰Ca⁺ is observed in Fig. 6(a) owing to the use of a bunched beam, only the Method 2” is now routinely used for CLS experiments on ion beams in both bunched and continuous modes [18, 19, 23].

HV amplifier calibration

As mentioned above, the amplification coefficient of the HV amplifier needs to be calibrated regularly to ensure that the correct voltage is applied during voltage scanning. The online calibration of the amplification coefficient was performed using the function of the HV Calibration” as shown in Fig. 5. During calibration, the voltage applied to the voltage scanning electrode (namely, the measured voltage) was recorded as a function of the input voltage (namely, the scanning voltage) for the HV amplifier, which was provided by the USB device and controlled by the program of the control and DAQ system. Figure 8 shows a typical example of the measured voltage versus scanning voltage ranging from -1.9 V to 1.9 V. The scanning range, voltage step size, and switch time for each step can be set in the GUI ( HV calibration” in Fig. 5). By fitting the plot linearly, the amplification coefficient of the HV amplifier was measured, and the corresponding uncertainty was assessed. The actual linearity is excellent (see inset of Fig. 8).

Fig. 8Full-screen View

(Color online) Typical example for the calibration of the HV amplifier. The scanning voltage between -1.9 V and 1.9 V (x-axis) is provided by the USB device, and the measured voltage (y-axis) is the one applied to the voltage scanning electrode. The data is fitted linearly. See text for more details.

Monitor of the relative beam intensity

In general, the HFS spectrum is measured over a long scanning time to obtain sufficient statistics. Thus, monitoring the stability of different parameters (e.g., probing laser frequency, applied voltage, and beam intensity) in real-time is of particular importance. This is accomplished using the function of Real-time Plotting” in the GUI of Fig. 5. An example is presented in Fig. 9 for the relative beam intensity, corresponding to the ion counts measured by the MagneTOF detector mounted at the end of the CLS beamline (Fig. 1(a)) [18, 19]. During the HFS spectrum measurement, the signals from this detector were recorded by one of the four channels of the TDC and plotted as a function of time. This counting is very sensitive to variations in beam intensity, as shown in Fig. 9. The overlap of the probing laser beam with the ion beam can also be optimized. For example, LIF photon counts measured at a fixed scanning voltage corresponding to the position of the HFS resonance peak can be optimized with respect to the ion beam or probing laser beam adjustment.

Fig. 9Full-screen View

(Color online) Relative beam intensity monitored in real-time. The MagneTOF detector mounted below the third Faraday cup (FC3) (Fig. 1) measures the scattered particles arising from the beam and thus indirectly monitors the ion beam intensity. See text for more details.