1 Introduction

With the development and application of the radioactive ion beam (RIB), new physical phenomena have been revealed in exotic nuclei which are distant from the β-stability line, such as the halo or skin structures [1, 2] and the new magic numbers [3]. To overcome the low intensity and quality of the RIB, the detection efficiency has been improved experimentally by using a large solid-angle covered detector array. The silicon detectors, especially the single- or double-sided silicon strip detectors (SSSDs or DSSDs) with efficient energy and spatial resolutions for charged particles, have been widely applied in building silicon detector arrays with a large and continuous angular coverage. A number of silicon detector arrays, such as MUST [4, 5], GLORIA [6], TIARA [7], EXODET [8], EXPADES [9], and MITA [10], have been developed for experimental studies motivated from the RIB.

A preamplifier is a key electronic module for realizing impedance matching between the detector and spectroscopy amplifier; it plays an important role in reducing external interference and consequently improving the signal-to-noise ratio. The ORTEC charge-sensitive preamplifier modules 142A, 142B, and 142C [11] that feature low noise and fast rising time (T_rise) were designed mainly for the optimal matching of each individual charged-particle detector. However, processing the hundreds of signals produced by the silicon detector array is challenging for the traditional discrete preamplifier modules. Consequently, specially designed electronic systems are required to support these complex detector arrays. The integrated Mesytec MPR-16 module [12] provides 16-channel preamplifiers specially designed for SSSD or DSSD. However, owing to its large size, it is difficult to install it close to the detectors in a limited chamber space. To reduce the crosstalk and noise, a special preamplifier module, that is not commercialized to match other silicon detector arrays, has been designed for MUST [4, 5]. Preamplifiers based on the application-specific integrated circuit (ASIC), such as VATA from IDEAS [13], were used for the silicon detectors. The IDEAS integrated circuits, however, are available only as bare dice or packaged chips, making it difficult for VATA to match the existing amplifier modules and data acquisition systems.

For the individual readout of each strip of the large-scale silicon detector array, the application of the highly integrated and low-cost preamplifier modules is definitely required. Therefore, a compact 16-channel integrated charge-sensitive preamplifier named the smart preamplifier (SPA), based on the commercial integrated circuit operational amplifier (OPA), has been designed and manufactured. With the compact size feature, the SPA can be placed in a vacuum chamber with water cooling and directly connected to the silicon detectors to significantly reduce noise. The performance was tested by using the alpha sources and further confirmed by the in-beam experiments performed using both the HI-13 tandem accelerator of the China Institute of Atomic Energy (CIAE) and the Radioactive Ion Beam Line in Lanzhou (RIBLL1) of the Institute of Modern Physics (IMP), Chinese Academy of Sciences.

2 Circuit design for the preamplifier

The SPA circuit mainly consists of three core parts: the bias circuits composed of R₁, R₂, and C₁, the input part composed of input and test capacitors C₀ and C₂, and the amplification-stage circuit composed of an OPA and resistor–capacitor feedback network R_fC_f.

A simplified block diagram of the SPA02 is shown in Fig. 1 (a). The bias voltage for the detector is supplied via the bias resistor R₂. To reduce the noise coming from the high voltage cable, the first-order RC filter circuit consisting R₁ and C₁ was employed. The maximum high voltage can reach 500 V, which is enough for most silicon detectors. C₀ is a coupling capacitor which couples the detector signal to the input end of the OPA and isolates the direct current (DC) to guarantee a stable quiescent point for the post stage circuit. The function of C₂ is the same as that of C₀ but for the test signal. The PA labelled in Fig. 1 is a high-performance commercial integrated circuit OPA. OPA657 [14] manufactured by Texas Instruments was applied, which is characteristic of a low input voltage noise of 4.8 nV/$\sqrt{\text{Hz}}$, high gain bandwidth of 1.6 GHz, and DC precision of a trimmed junction field effect transistor (JFET)-input stage to give an exceptionally high input impedance. The R₀ series resistor at the output terminal provides a matching load for the subsequent equipment. The resistor-capacitor feedback network R_fC_f forms a charge integrating and bleeding circuit that determines the energy sensitivity and pulse decay time. The energy sensitivity can be easily changed on request. The performance of SPA02 for the low capacitance silicon detectors is efficient and will be discussed later.

Fig. 1.Full-screen View

Simplified block diagrams of SPA02 (a), SPA03 (b), and SPA02-16 (c).

To match the high capacitance (approximately 1000 pF) SSSD or DSSD, such as a SSSD with a thickness of 20 μm (W1-type from Micron Semiconductor Ltd. [15]), it is necessary to ensure that the dynamic input capacitance of the preamplifier, which depends on the corresponding transconductance, is considerably larger than the detector capacitance. To address this requirement, another preamplifier type, named the SPA03 was developed, as shown in Fig. 1 (b). For the amplifier to have a sufficiently high transconductance, the commercial JFET BF862 [16] from Philips Semiconductors with the advantages of a high transition frequency (640 MHz) and a high transconductance (g_m=40 mS) was used in the first amplification stage. A common source circuit of the field effect transistor (FET) employed in the first stage features a high input impedance, low noise and wide bandwidth. The gain value of the FET connected with the common source circuit is A₁ = - g_mR₄ =-40 mS × 680 Ω=-27.2. Because of the high gain in the first stage, the capacitance reflected to the input from the drain to gate capacitance will be significant [17] owing to the Miller effect. The high transconductance reduces the series noise contribution, which is the most significant factor to receive a good output signal for a high capacitance detector. The current feedback OPA AD8001 [18] developed by Analog Devices was used as the second amplification stage to meet the requirement of a fast response and low power consumption. The AD8001 features a low power consumption of 50 mW, a wide bandwidth of 800 MHz, and a slew rate of 1200 V/μs.

Eight SPA02s were installed to a ceramic circuit substrate with a size of 24 mm×80 mm to easily match the silicon detector array with hundreds of channels. This unit shown in Fig. 2 (a) was named the SPA02-08. Further, to match both the strip detector with 16 strips and the main amplifier module with 16 channels, the SPA02-16 (Fig. 1 (c)) was assembled using two pieces of SPA02-08 positioned back-to-back, as shown in Fig. 2 (c). The Rogers ceramic printed circuit board (PCB) was adopted to provide efficient heat conduction. Standard 2.54-mm pin connections were employed for the input and output signals as well as the common input of the ±6 V power supply, bias, and test signals. SPA03-08 (Fig. 2 (b)) and SPA03-16 can be assembled similarly. With this compact design, the silicon detectors can be connected to the SPAs using short cables in actual use for improving the signal quality.

Fig. 2.Full-screen View

(Color online) Photos of (a) SPA02-08, (b) SPA03-08, and (c) SPA02-16. "±" is the power supply of ±6 V, "H" represents the bias for the detector with a 100-MΩ protection resistor, "T" denotes the test pulse input, protection is represented by "P", and "G" and other areas without labels are the earthed regions.

3 Performance test for the preamplifier

We tested the SPAs in detail using a pulse generator and alpha source, to have a better understanding of the performance of the preamplifiers. Both the positive and negative input signals are acceptable for SPA02 and SPA03. The baseline drift of the output signal is typically less than ±20 mV and at most ±50 mV over the full output range of ±2 V with a 50 Ω terminal. Based on the test results, the SPAs were further applied in the in-beam experiments, and showed good quality and high stability.

3.1 Response linearity

An exponential attenuation signal with an adjustable pulse amplitude of 0–1300 mV and frequency of 1 kHz, supplied by the precise pulse generator AFG3022C manufactured by Tektronix [19], was used for the linearity test. The output amplitude (V_out) was obtained by the CAEN N1568A [20] spectroscopy amplifier and CAEN V785 [21] peak-sensing analog-to-digital conversion (ADC). Fig. 3 shows the result of the linearity test; good linearities are observed for both SPA02 and SPA03 with determination coefficients of R²=0.99953 and R²=0.99990, respectively.

Fig. 3.Full-screen View

(Color online) Linearity test between V_out and V_test for SPA02 and SPA03.

3.2 Influence of input capacitance

An exponential attenuation test signal with 1 kHz and 800 mV was fed to SPA02 and SPA03 with R_f = 100 MΩ and C_f=1 pF, respectively. The amplitude and rising time of the output signal were obtained using an oscilloscope DPO 4034B manufactured by Tektronix [22] with a 50 Ω terminal. Equivalent noise was extracted by fitting the output spectrum acquired by N1568A [20] and V785 [21]. To clearly demonstrate the performance of the SPA and compare with that of the preamplifiers manufactured by ORTEC and Mesytec, the equivalent noise was calibrated using the ²³⁹Pu and ²⁴¹Am alpha sources and keV was used as the physical unit when quoting the equivalent noise of the preamplifiers.

The dependences of V_out, T_rise, and equivalent noise (Noise) on the input capacitance (C_in) are shown in Fig. 4. It can be seen from Fig. 4 (a) that the V_out of SPA03 changes gradually compared with that of SPA02 with the increase of C_in. Fig. 4 (b) shows that the T_rise of SPA02 and SPA03 increases with the increase of C_in. Both SPA02 and SPA03 have a fixed rising time of generally less than 6 ns at C_in=0 pF within the full output range. The "Noise Slope", defined as the variation of the equivalent noise with C_in, is shown in Fig. 4 (c). The equivalent noises of both SPA02 and SPA03 are less than 1.5 keV without external C_in. The equivalent noise of SPA03 is lower than that of SPA02, especially at large C_in, as shown in Fig. 4 (c). The parasitic capacitance of a silicon strip with an area of 3 mm×50 mm for a SSSD with a thickness of 20 μm is approximately 1000 pF. Therefore, SPA03 is more suitable for high-capacitance detectors. The relevant parameters of the rising time, equivalent noise, and noise slope for SPA02 and SPA03 along with the 142A/B/C and MPR-16 are listed in Table 1.

Fig. 4.Full-screen View

(Color online) Variations of the output amplitude (V_out), rising time (T_rise), and equivalent noise (Noise) with the external input capacitance (C_in) for SPA02 and SPA03 with R_f=100 MΩ and C_f=1 pF for the input pulse hight V_in= 800 mV

TABLE 1.Full-screen View

Performance of SPA02, SPA03, 142A/B/C, and MPR-16.

Preamplifier type	Tarise (ns)	Equivalent Noisea(keV)	Noise Slope (eV/pF)
SPA02	6	1.5	397
SPA03	6	1.5	159
142A	5	1.6	180
142B	5	1.4	176
142C	5	2.2	124
MPR-16	12	4.0	400

Show more

a These values are obtained under C_in=0 pF.

3.3 Energy resolution and energy dynamic range

A triple mixed alpha source composed of ²³⁹Pu, ²⁴¹Am, and ²⁴⁴Cm was used to evaluate the energy resolution of the Si-PIN detector with an effective area of 5 mm×5 mm (C_det 9 pF) equipped with the SPA02. A typical energy spectrum is shown in Fig. 5. The types of alpha sources, energies, and energy resolutions are marked. An energy resolution of 0.42% was obtained for 5.803 MeV alpha particles from ²⁴⁴Cm by fitting the main peak. For a cooperative production of the DSSD manufactured by CIAE and the Beijing Kelixing Photoelectric Technology [23], with an active area of 48 mm × 48 mm and a thickness of 300 μm, the energy resolutions for a triple mixed alpha source have been achieved at 0.65–0.80% for the junction strips and 0.85–1.00% for the ohmic strips of the DSSD equipped with the SPA02.

Fig. 5.Full-screen View

(Color online) Energy spectrum of the alpha particles from the triple mixed source measured by a Si-PIN detector with SPA02.

A ²³⁹Pu alpha source was used to measure the energy resolution of a SSSD with a thickness of 300 μm equipped with SPA02-16 or SPA03-16 modules. The connection between the SSSD and SPA-16 is shown in Fig. 6 (a). The alpha source was placed at a distance of 10 cm from the SSSD with an effective area of 50 mm × 50 mm. The output pulse shape and noise of the SPA02-16 module with R_f=100 MΩ and C_f=1 pF for the 5.153 MeV alpha particle detected by the SSSD with a thickness of 300 μm and a bias of -50 V are shown in Fig. 6 (b). The pulse shape for the 5.153-MeV alpha particle features T_rise of 24 ns, output amplitude of 105 mV, and noise amplitude of 2 mV with a 50 Ω terminal. The output of SPA02 or SPA03 was fed directly to N1568A [20] and then to V785 [21].

Fig. 6.Full-screen View

(Color online) (a) Photos of SSSD of thickness 300 μm with an SPA02-16 module and cooling plate. (b) Pulse shape and noise of the 5.153 MeV alpha particles measured by the SSSD equipped with SPA02-16 (R_f=100 MΩ and C_f=1 pF).

To achieve better energy resolution and to maintain operational stability, the SPA modules were cooled to about 10∘C using a circulating cooling machine. Fig. 7 shows the measured α-energy spectra using SPA02 (a) and SPA03 (b), respectively. The obtained energy resolutions are 0.73% and 0.67% for SPA02 and SPA03, respectively. The small bumps at 5.499 MeV originate from contamination with ²³⁸Pu.

Fig. 7.Full-screen View

(Color online) α energy spectra measured by the SSSD with SPA02 (a) and SPA03 (b).

The energy dynamic range (EDR) was extracted using the output amplitude of the 5.486-MeV alpha particle (α-amp) from ²⁴¹Am in the Si-PIN with SPA02 or SPA03 by varying C_f. The C_f of SPA02, corresponding energy sensitivity, EDR, and the output amplitude are listed in Table 2. The energy sensitivity of SPA03 is almost identical to that of SPA02 and is not shown here. Compared with the fixed EDR for the preamplifiers made by both ORTEC and Mesytec, the EDR of the current SPA can be easily changed by modifying C_f to match the different requirements.

TABLE 2.Full-screen View

Relation of the output amplitude of the alpha particles (α-amp), energy dynamic range, and energy sensitivity with C_f for SPA02.

Cf (pF)	α-ampa (mV)	EDR (MeV)	Sensitivity (mV/MeV)
1	108	100	20.0
2	60	183	10.9
3	38	289	6.9
4	28	392	5.1
5	25	439	4.6

Show more

a α-amp was obtained using the oscilloscope DPO 4034B with a 50 Ω terminal.

3.4 Application to in-beam experiments

The integrated SPA modules were applied to in-beam experiments performed at the HI-13 tandem accelerator of CIAE [24-26]. To study the one-proton transfer reaction of ⁷Li+²⁰⁸Pb, a detector array containing 16 Si-PIN detectors, four multi-layer ionization-chamber (IC) telescopes [10], and two silicon detector telescopes was used. The multi-layer IC telescope consists of one IC, one 60-μm-thick DSSD (W1-type from Micron Semiconductor Ltd. [15]), and two quadrant silicon detectors (QSDs) with thicknesses of 300 μm and 1000 μm, respectively. The silicon detector telescope includes one 20-μm-thick SSSD, one 60-μm-thick DSSD, and one-1000-μm-thick QSD. Therefore, two SPA03-16 and seventeen SPA02-16 modules were applied for the two 20-μm-thick SSSDs and other detectors, respectively.

The photo of the assembled detector array with the SPA modules and cooling copper ring is shown in Fig. 8 (a). Fig. 8 (b) represents a typical ΔE-E_R spectrum obtained by the silicon detector telescope located at θ_lab=144°-171° at E_lab=28.55 MeV. The corresponding kinetic energy spectrum of ⁶He is shown in Fig. 8 (c). The excitation energies of ²⁰⁹Bi can be calculated using the kinetic energies of ⁶He, and are shown in Fig. 8 (c). The ⁴He, ⁶He, and ⁷Li bands, as well as the different low-lying excitation states of ²⁰⁹Bi can be clearly separated. These results further indicate the good performance of both SPA02 and SPA03.

Fig. 8.Full-screen View

(Color online) Photo of the detector array with (a) the SPA02-16 modules and cooling copper ring and (b) ΔE-E_R spectrum obtained by the silicon detector telescope for ⁷Li+²⁰⁸Pb at E_lab=28.55 MeV with θ_lab=144°-171°, and (c) the projected energy spectrum of the ⁶He band where the peaks are labeled corresponding to the excitation states of ²⁰⁹Bi in the unit of MeV.

The integrated SPA modules were also applied to the in-beam experiments [27-35] performed at RIBLL1 of IMP. To study the decay properties of the nuclei near the proton drip line, a detector array [30] was composed of three layers of DSSDs with different thicknesses, four 300-μm-thick QSDs, and one 1546-μm-thick QSD. Two SPA03-16 and six SPA02-16 modules were employed to match the 40-μm-thick DSSD and other silicon detectors, respectively.

The photo of the detector array equipped with the SPA02-16 and SPA03-16 modules for studying the β-decay of ²⁷S is shown in Fig. 9 (a). Fig. 9 (b) shows a typical two-dimensional particle identification spectrum of ΔE-ToF for the secondary beams, where ΔE is given by the QSD with the SPA02 and the ToF is determined by two plastic scintillation detectors located in the RIBLL1 beam-line. The cumulative proton energy spectrum from ²⁷S β-decay measured by the two DSSDs with thicknesses of 40 μm and 304 μm is shown in Fig. 9 (c). The good quality of the particle identification and the observed proton peaks demonstrate good performance of the SPA02-16 and SPA03-16 modules in the RIB experiment.

Figure 9:Full-screen View

(Color online) (a) Photo of the detector array with the modules of SPA02-16 and SPA03-16 equipped with the cooling PCB plate. (b) Typical two-dimensional particle identification spectrum of ΔE-ToF for the secondary beams. (c) Cumulative β-decayed proton energy spectrum for ²⁷S β-decay measured by the 40-μm- and 304-μm-thick DSSDs.

4 Summary

In this study, we designed and manufactured two types of integrated charge-sensitive preamplifier modules, namely, SPA02-16 and SPA03-16, to match the silicon detector array with a large solid angle coverage, which is widely used in current nuclear experiments. Generally, the SPA02 was produced to support silicon detectors with thicknesses larger than 60 μm, whereas the SPA03 type with external FET was designed to match thin silicon detectors with large parasitic capacitances. The EDR of the SPAs can be easily adjusted by modifying the feedback capacitor C_f for the different experimental requirements.

Both the SPA02 and SPA03 have a fast response and low noise, which is typically less than 6 ns for the rising time and 1.5 keV for the equivalent noise without the external capacitor. The good performance of both SPA02 and SPA03 was further confirmed using an alpha source. A typical energy resolution of 0.6–0.8% was achieved for the silicon strip detectors equipped with the SPAs using the ²³⁹Pu alpha source. A maximum EDR of 439 MeV was reached with the C_f of 5 pF. The SPA modules were applied in the in-beam experiments performed at CIAE and IMP; the excellent results further demonstrated the flexibility and stability of the integrated SPA modules.