4H-SiC

0 downloads 0 Views 5MB Size Report
Oct 13, 2017 - Schottky Diode Radiation Detectors with a Sensitive ..... Characteristics of the pulse height spectra of the 4H-SiC detector in response to the ...
sensors Article

The Fabrication and Characterization of Ni/4H-SiC Schottky Diode Radiation Detectors with a Sensitive Area of up to 4 cm2 Lin-Yue Liu 1,2 , Ling Wang 3 , Peng Jin 2 , Jin-Liang Liu 2 , Xian-Peng Zhang 2 , Liang Chen 2 , Jiang-Fu Zhang 2 , Xiao-Ping Ouyang 1,2,4, *, Ao Liu 3 , Run-Hua Huang 3 and Song Bai 3, * 1 2

3

4

*

School of Nuclear Science and Technology, Xi’an Jiaotong University, No. 28, Xianning West Road, Xi’an 710049, China; [email protected] State Key Laboratory of Intense Pulsed Radiation Simulation and Effect, Northwest Institute of Nuclear Technology, Xi’an 710024, China; [email protected] (P.J.); [email protected] (J.-L.L.); [email protected] (X.-P.Z.); [email protected] (L.C.); [email protected] (J.-F.Z.) State Key Laboratory of Wide-Bandgap Semiconductor Power Electronic Devices, Nanjing Electronic Devices Institute, No. 524 East Zhongshan Road, Nanjing 210016, China; [email protected] (L.W.); [email protected] (A.L.); [email protected] (R.-H.H.) Shaanxi Engineering Research Center for Pulse-Neutron Source and its Application, Xijing University, Xi’an 710123, China Correspondence: [email protected] (X.-P.O.); [email protected] (S.B.)

Received: 16 July 2017; Accepted: 25 September 2017; Published: 13 October 2017

Abstract: Silicon carbide (SiC) detectors of an Ni/4H-SiC Schottky diode structure and with sensitive areas of 1–4 cm2 were fabricated using high-quality lightly doped epitaxial 4H-SiC material, and were tested in the detection of alpha particles and pulsed X-rays/UV-light. A linear energy response to alpha particles ranging from 5.157 to 5.805 MeV was obtained. The detectors were proved to have a low dark current, a good energy resolution, and a high neutron/gamma discrimination for pulsed radiation, showing the advantages in charged particle detection and neutron detection in high-temperature and high-radiation environments. Keywords: 4H-SiC; radiation detection; large sensitive area; Schottky diode

1. Introduction Since the first silicon carbide (SiC) detector was developed nearly sixty years ago [1–3], the potentials of SiC detectors have been recognized for their better endurance to elevated temperatures and radiation-induced damage than conventional silicon or germanium detectors. Many other semiconductors have been used to fabricate detectors at the same time: CdTe, CdZnTe, GaAs, and AlInP are focused on photon detection [4–6]; diamond is suitable for neutron, photon, and charged particle detection and has ultra-high radiation resistance but with tiny dimension, uneven quality, and high cost [7,8]. By now, SiC detectors have been demonstrated to have a high resolution in the detection of charged particles [9–14], photons [15–18], and neutrons [19–22]. Particularly, because of their outstanding operations in applications in intense radiation fields and harsh environments, such as alpha particle monitoring and neutron detection in actinide waste-tank environments [23] and neutron and gamma-ray monitoring of spent nuclear fuel assemblies [24,25], and because the technology has matured in terms of material growth and device fabrication, they have been considered preferable substitutions for conventional silicon radiation detectors. However, compared with commercial silicon detectors whose sensitive areas are usually in the range of 0.78–7 cm2 , even up to 70 cm2 in some applications, the largest sensitive area of an SiC Sensors 2017, 17, 2334; doi:10.3390/s17102334

www.mdpi.com/journal/sensors

required. Most high-quality SiC detectors are made with epitaxial SiC material. The low dark current is necessary for SiC detectors, both in spectroscopic and in current mode detection. The fabrication of large-area SiC detectors is a difficult task due to the defects in epitaxial material and micro-pipes in the SiC substrate, which will cause excessive leakage current and a reduction in breakdown voltage, Sensors 2017, 17, 2334 2 of 9 thus resulting in the degradation of the response properties of SiC detectors. We fabricated a passel of SiC Schottky diode chips with a size of 1 cm × 1 cm using lightly doped 4H-SiC epitaxial material 20 μm thick, and assembled of alarge-area SiC[26] detectors, with four chips in a 2[27]. ×2 spectrometry detector is onlytwo 0.36groups cm2 for single chip or 0.81 each cm2 for a splicing device array on a PCB plate a ceramic The properties of thegood detectors were experimentally Small detectors haveand been studiedcase. sufficiently and show performance in charged studied, particle and the following results were achieved: a dark current of 15–60 nA at 600 V, an optimum energy monitoring, etc., but usually they have low efficiency in radiation beams with large diameters or resolution of 3.22% for alpha particles, a rise time of 9.4 ns, and a neutron/gamma discrimination of large radiation emission angles, and thus need more time to accumulate sufficient counts to ensure 126. the results meet the statistical requirements. For the detection in a large radiation field, larger that detectors are required. 2. Experimental SectionSiC detectors are made with epitaxial SiC material. The low dark current is Most high-quality necessary for SiC detectors, both in spectroscopic and in current mode detection. The fabrication of 2.1. The Fabrication of 4H-SiC Detectors large-area SiC detectors is a difficult task due to the defects in epitaxial material and micro-pipes in the SiC substrate, which willlightly cause excessive leakage 4H-SiC current and a reduction in breakdown voltage,vapor thus The high-quality doped epitaxial material was grown via chemical resulting in(CVD) the degradation of the4H-SiC response of substrate SiC detectors. We(Φfabricated passel SiCa deposition on commercial N+properties conducting wafers 10.2 cm ×a350 μm,of and −3, supplied Schottky diode chips with a size of 1 cm lightly by doped 4H-SiC Semiconductor epitaxial material µm target nitrogen doping concentration of×10119cm cmusing TankeBlue Co.20Ltd., Beijing, China). The epitaxial layers were 20 μm thick and with target doping thick, and assembled two groups of large-area SiC detectors, each withnitrogen four chips in a 2 concentrations × 2 array on a of 1–5plate × 1014 cma−3.ceramic The topcase. Schottky wasof formed by the deposition of 100 nm nickel on epitaxial PCB and The barrier properties the detectors were experimentally studied, and the layers via results thermal vacuum evaporation, and was protected multi-layers of monox/silicon nitride following were achieved: a dark current of 15–60 nA atby 600 V, an optimum energy resolution of (50 nm/50 nm)particles, that covered electrode. The bottom ohmic contact was acquired by 3.22% for alpha a rise the timenickel of 9.4 ns, and a neutron/gamma discrimination of 126. evaporation of Ni/Au and then annealing at 900 °C in nitrogen. The front contact was protected by 2. Experimental Section multi-floating rings from high voltage damage. Figure 1a shows a schematic diagram of a 4H-SiC Schottky diode detector. 2.1. The Fabrication of 4H-SiC Detectors Normally, the yield of an SiC detector will be limited by the concentration of the defects in the The high-quality lightly material chemical vapor deposition detector [26]. Detectors of adoped largerepitaxial diameter4H-SiC are more likelywas to grown containvia more defects in their active (CVD) on commercial 4H-SiC N+ conducting substrate wafers (Φ 10.2 cm × 350 µm, and a target area, which will degrade their response properties, such as excessive leakage current. Initially, we 3 , supplied by TankeBlue Semiconductor Co. nitrogen doping concentration of 1019 of cma−diode Ltd., Beijing, attempted to make an area scale-up with a sensitive area up to 25 mm2, and following 2 China). The epitaxial layers were 20 µm thick and with target nitrogen doping concentrations of encouraging results, fabricated a passel of diode chips with 100 mm in sensitive area equivalents. 14 − 3 1–5 × 10 cm two . The topofSchottky barrier was four formed byconnected the deposition of 100 on We assembled groups detectors, each with chips in parallel in anm 2 × nickel 2 array— epitaxial layers vacuum evaporation, and was by multi-layers of monox/silicon one group on a via PCBthermal plate, the other on a ceramic case. The protected back electrode was connected by a welding nitride nm/50 nm)electrode that covered nickel electrode.with TheAu bottom contact was process(50 and the front was the linked by bonding wires.ohmic Figure 1b,c are theacquired picturesby of ◦ evaporation of Ni/Au and to then annealing at 900 C in nitrogen. The front contact wasdetector protected the diode chips connected a PCB plate and a ceramic shell, respectively. Each SiC hasbya multi-floating rings from 1a shows a schematic diagram a 4H-SiC sensitive volume of 20 mmhigh × 20 voltage mm × 20damage. μm and aFigure dead layer of Ni/SiO 2/Si3N4 (100 nm/50ofnm/50 nm) Schottky diode detector. without considering the dead region in the SiC near the Schottky contact.

Figure 1. (a) Schematic diagram of a 4H-SiC Schottky diode detector; photograph of a 4H-SiC detector in a 2 × 2 array mounted on a multi-layer PCB plate (b) and in a ceramic case (c) with a total sensitive area of 4 cm2 .

Normally, the yield of an SiC detector will be limited by the concentration of the defects in the detector [26]. Detectors of a larger diameter are more likely to contain more defects in their active area, which will degrade their response properties, such as excessive leakage current. Initially, we attempted to make an area scale-up of a diode with a sensitive area up to 25 mm2 , and following encouraging results, fabricated a passel of diode chips with 100 mm2 in sensitive area equivalents. We assembled two groups of detectors, each with four chips connected in parallel in a 2 × 2 array—one group on a

Sensors 2017, 17, 2334

3 of 9

Sensors 2017, 17, PCB plate, the2334 other on a ceramic case.

9 The back electrode was connected by a welding process and3 of the front electrode was linked by bonding with Au wires. Figure 1b,c are the pictures of the diode chips Figure 1. (a) Schematic diagram of a 4H-SiC Schottky diode detector; photograph of a 4H-SiC detector connected to a PCB plate and a ceramic shell, respectively. Each SiC detector has a sensitive volume in a 2 × 2 array mounted on a multi-layer PCB plate (b) and in a ceramic case (c) with a total sensitive of 20 mm × 20 2mm × 20 µm and a dead layer of Ni/SiO2 /Si3 N4 (100 nm/50 nm/50 nm) without area of 4 cm . considering the dead region in the SiC near the Schottky contact.

2.2. 2.2. Measurements Measurements Both Both the the forward forward I-V I-V and and C-V C-Vcurves curvesof ofthe thedetector detectorwere weremeasured measuredusing usingAgilent AgilentB1500A B1500A Power Power Device Analyzer/ Curve Tracer. The dark current was measured by Keithley 6517A Ampere Meter in Device Analyzer/ Curve Tracer. The dark current was measured by Keithley 6517A Ampere Meter ainshielded copper box in in darkness. AA PS350 a shielded copper box darkness. PS350high highvoltage voltagesupply supply(Stanford (Stanfordresearch researchsystem system Inc., Inc., Sunnyvale, CA, USA) was used to provide the reverse bias. Sunnyvale, CA, USA) was used to provide the reverse bias. The SiC detectors to charged particles was was studied experimentally with the alpha The response responseofofthe the SiC detectors to charged particles studied experimentally with the sources in a vacuum chamber in Nuclear Institute of Northwest Technology (NINT) (NINT) in Xi’an,inChina. alpha sources in a vacuum chamber in Nuclear Institute of Northwest Technology Xi’an, 243 One alpha mixed (E243 α = 5.275 MeV, branch ratio of 87.5%) and 244Cm (Eα =244 China. Onesource alphawas source waswith mixedAm with Am (Eα = 5.275 MeV, branch ratio of 87.5%) and 5.805 Cm 3 Bq, the other was 239Pu (Eα = 5.157 MeV, MeV, a branch ratio of 76.4%) with a radioactivity of 1.8 × 10 3 239 (Eα = 5.805 MeV, a branch ratio of 76.4%) with a radioactivity of 1.8 × 10 Bq, the other was Pu branch ratio MeV, of 73.3%) withratio a radioactivity of 1.2a ×radioactivity 105 Bq. Both of alpha sources were prepared via the (Eα = 5.157 branch of 73.3%) with 1.2 × 105 Bq. Both alpha sources electro-deposition of oxidized isotopes on stainless-steel plates—one with a diameter of 10 mm anda were prepared via the electro-deposition of oxidized isotopes on stainless-steel plates—one with the other of As shown in Figure 2, the sources were concentrically with the diameter of 30 10 mm. mm and the other of 30 mm. Asalpha shown in Figure 2,positioned the alpha sources were positioned detector’s sensitive layer, 80 mm away from the detector. The signals from the detector were concentrically with the detector’s sensitive layer, 80 mm away from the detector. The signals from the amplified by an Ortec-142B Pre-Amplifier and an Ortec-672 Amplifier with a shaping time of 1 μs detector were amplified by an Ortec-142B Pre-Amplifier and an Ortec-672 Amplifier with a shaping and 50,aand an Ortecbymultichannel analyzer (MCA) and Gammatimeaofgain 1 µsofand gainwere of 50,then andanalyzed were thenby analyzed an Ortec multichannel analyzer (MCA) and Vision software. The reverse bias voltages of 0–500 V were applied to the detector by the PS350 Gamma-Vision software. The reverse bias voltages of 0–500 V were applied to the detector bybias the supply through the Ortec 142B preamplifier. PS350 bias supply through the Ortec 142B preamplifier.

Figure Figure 2. 2. Experimental Experimentalsetup setupfor foralpha alpha particle particle detection detection with with the the SiC SiC detector. detector.

The The response response time time of of aa semiconductor semiconductor detector detector is is one one of of the the key key parameters parameters in in pulsed pulsed radiation radiation detection. It can be determined in the detection of prompt pulsed radiation from a source fast detection. It can be determined in the detection of prompt pulsed radiation from a source fast enough enough to as aa delta delta(δ) (δ)source. source.InInthe theexperiment experimentdescribed described here, a pulsed sub-nanosecond Xto be be assumed assumed as here, a pulsed sub-nanosecond X-ray ray source a pulsed UV laser device provided by NINT The pulsed sub-nanosecond source andand a pulsed UV laser device provided by NINT were were used.used. The pulsed sub-nanosecond X-ray X-ray source emits a pulsed X-ray beam on average lower than 100 keV, with a rise time around 600a source emits a pulsed X-ray beam on average lower than 100 keV, with a rise time around 600 ps and ps and a repetition frequency of 1 UV Hz. laser The UV laser device (EKSPLA 355 nmUV-light pulsed repetition frequency of 1 Hz. The device (EKSPLA PL2251C)PL2251C) emits 355emits nm pulsed UV-light with a pulse-width of 30 ps and a maximum energy of 20 mJ in each shot. The response with a pulse-width of 30 ps and a maximum energy of 20 mJ in each shot. The response waveforms waveforms were by4104 a Tektronix 4104 Oscilloscope 1 GHz; 4sample-rate: GS/s) were recorded byrecorded a Tektronix Oscilloscope (bandwidth: 1(bandwidth: GHz; sample-rate: GS/s) and a4Lecroy and a Lecroy 6100A(bandwidth: Oscilloscope1(bandwidth: 1 GHz;10sample-rate: 10 GS/s) through cables. well-shielded 6100A Oscilloscope GHz; sample-rate: GS/s) through well-shielded cables. 3. Results and Discussion 3. Results and Discussion 3.1. Electric Parameters 3.1. Electric Parameters The result of the forward I-V test is shown in Figure 3a. The curve exhibits a rectification character. According to the I-V characteristics and the Bethe equation, we find the ideality factor is The result offorward the forward I-V test is shown in Figure 3a. The curve exhibits a rectification 1.422 ± 0.005, which indicates the current is not just dominated by thermionic current—the diffusion character. According to the forward I-V characteristics and the Bethe equation, we find the ideality current recombination current are contributing factor is and 1.422 ± 0.005, which indicates the current istoo. not[27] just dominated by thermionic current—the diffusion current and recombination current are contributing too. [27] Figure 3b shows the C-V curve acquired at 1 MHz. Figure 3c is the curve of 1/C2 vs. V. The effective doping concentration (Neff) of the 4H-SiC epitaxial layer was calculated to be (2.721 ± 0.004)

Ni/Au electrode, and improving the surface roughness of the SiC material near the front Ni electrode, and we then measured the dark current of the three detectors in the ceramic shell (second batch). We found that the dark current decreased to 15.2 nA, 38.8 nA, and 58.6 nA with an uncertainty within 1%, respectively, at a reverse bias of 600 V. The dark current of the second group of detectors was much Sensors lower 2017, 17,than 2334 those of conventional silicon PIN detectors of the same dimensions (higher than 4 of19 μA) [28].

Figure parametersofofthe the4H-SiC 4H-SiC diode: Forward of one chip; (b) C-V ofchip one ; Figure 3. 3. Electric Electric parameters diode: (a)(a) Forward I-VI-V of one chip; (b) C-V curvecurve of one 2-V plot of one chip; (d) Reverse I-V (Dark current) of an SiC detector from the PCB group chip ; (c) 1/C 2 (c) 1/C -V plot of one chip; (d) Reverse I-V (Dark current) of an SiC detector from the PCB group (half (half in black) anddetectors three detectors from the ceramic shell(open group (open right-triangle in blockblock circle circle in black) and three from the ceramic shell group right-triangle in red, open red, open star in black, and open up-triangle in green) from the four pixel structure at a reverse bias star in black, and open up-triangle in green) from the four pixel structure at a reverse bias of 600 V. of 600 V.

Figure 3b shows the C-V curve acquired at 1 MHz. Figure 3c is the curve of 1/C2 vs. V. The effective 3.2. Alpha-Particle Detection–Steady State Measurement doping concentration (Neff ) of the 4H-SiC epitaxial layer was calculated to be (2.721 ± 0.004) × 1014 cm−3 In the detection of charged and contact ion beams, the such as Schottky protons and the built-in Vbi potential of particles the Schottky was once found tocharged be 1.229particles, ± 0.007 eV. The and alpha (α) particles, incident oneV. theFigure SiC material, ionization occur, causing the incident barrier height was aboutare 1.513 ± 0.009 3d shows the darkwill current of an SiC detector from charged particles to lose part or all of their energy, resulting in the formation of electrons and holes the PCB plate group (first batch). The dark current is 0.48 µA at a reverse bias of 600 V, which is higher (called charged carriers).We The charged drift inoptimization the bias fieldtoofthe theother detector and collected than what we expected. then made carriers some technical group of are the detectors, by the electrodes. SRIM concentration code [29], we of calculated the energy of the incident emitted including reducingUsing the doping the SiC epitaxial layer, selecting SiCparticles wafers with low 239 243 244 from Pu,adjusting and AmCm sources after they of passed through the Si 3N4/SiO2/Ni layer defectthe density, the annealing temperature the bottom Ni/Au electrode, andentrance improving the (dead and found thatmaterial all of their energyand waswelost in measured the activethe volume of the surfacelayer) roughness of the SiC nearresidual the frontkinetic Ni electrode, then dark current detector. of the three detectors in the ceramic shell (second batch). We found that the dark current decreased to Figure the response spectra of the detector the respectively, alpha particles theofsource 15.2 nA, 38.84 shows nA, and 58.6 nA with an uncertainty withinto1%, at aemitted reverseby bias 600 V. 243Am-244Cm at the reverse bias voltages of 0, 100 V, 200 V, 300 V, 400 V, and 500 V. It is worth of The dark current of the second group of detectors was much lower than those of conventional silicon PIN noting that detector attained(higher similarthan alpha response detectors of the same dimensions 1 µA) [28]. spectra and worked stably at reverse bias voltages no less than 100 V, but measurable numbers lost amounts of incident events induced by 3.2. Alpha-Particle Measurement alpha particles at aDetection–Steady reverse bias of State 0. Figure 5a gives the peak centroid as a function of reverse bias voltage. Thedetection peak centroid in theparticles spectrumand at 0ion V (without reverse is about 5% lower those In the of charged beams, once the bias) charged particles, such than as protons in the other spectra at the reverse bias voltages of 100–500 V. Fitting the peaks obtained above by the and alpha (α) particles, are incident on the SiC material, ionization will occur, causing the incident charged particles to lose part or all of their energy, resulting in the formation of electrons and holes (called charged carriers). The charged carriers drift in the bias field of the detector and are collected by the electrodes. Using SRIM code [29], we calculated the energy of the incident particles emitted from the 239 Pu, and 243 Am-244 Cm sources after they passed through the Si3 N4 /SiO2 /Ni entrance layer (dead layer) and found that all of their residual kinetic energy was lost in the active volume of the detector. Figure 4 shows the response spectra of the detector to the alpha particles emitted by the source of 243 Am-244 Cm at the reverse bias voltages of 0, 100 V, 200 V, 300 V, 400 V, and 500 V. It is worth noting that the detector attained similar alpha response spectra and worked stably at reverse bias voltages no less than 100 V, but measurable numbers lost amounts of incident events induced by alpha particles at a reverse bias of 0. Figure 5a gives the peak centroid as a function of reverse bias voltage. The peak

Sensors 2017, 17, 2334

5 of 9

centroid in the spectrum at 0 V (without reverse bias) is about 5% lower than those in the other spectra 5 of 9 5 of 9 voltages of 100–500 V. Fitting the peaks obtained above by the Gaussian function, we got thefunction, full width By dividing the By FWHM by the centroid, Gaussian we at gothalf the maximums full width at(FWHMs). half maximums (FWHMs). dividing the peak FWHM by the Gaussian function, we got the as full width at half maximums (FWHMs). By dividing the FWHM by the we got the energy resolution a function of reverse bias (Figure 5b). The best energy resolution is peak centroid, we got the energy resolution as a function of reverse bias (Figure 5b). The best energy peak centroid, weV.got therise energy resolution as a function of reverse bias (Figure 5b). The best energy at 200 V and 300 The of energy resolution at reverse bias voltages above 400 V can be due to resolution is at 200 V and 300 V. The rise of energy resolution at reverse bias voltages above 400 V resolution is atthe 200 V and 300 V.white The rise of energy resolution at reverse bias voltages above 400 V the increase detector’s whichwhite could increase the detector’s electronic noise and can be due in to the SiC increase in the SiC noise, detector’s noise, which could increase the detector’s can be due to the increase in the SiC detector’s white noise, which could increase the detector’s broaden the alpha peaks. electronic noise and broaden the alpha peaks. electronic noise and broaden the alpha peaks. Sensors 2017, 17, 2334 Sensors 17, 2334 at the 2017, reverse bias

243Am and 244Cm at reverse Figure 4. Response Response of the the 4H-SiC detector detector to alpha alpha particlesemitted emitted by243 244 Cm at reverse Figure Figure 4. 4. Response of of the 4H-SiC 4H-SiC detector to to alphaparticles particles emittedby by 243Am Am and and 244Cm at reverse bias voltages of 0, 100 V, 200 V, 300 V, 400 V, and 500 V. bias voltages of 0, 100 V, 200 V, 300 V, 400 V, and 500 V. bias voltages of 0, 100 V, 200 V, 300 V, 400 V, and 500 V.

Figure 5. Characteristics of the pulse height spectra of the 4H-SiC detector in response to the alpha Figure 5. 5. Characteristics Characteristics of the pulse pulse height spectra spectra of the the244 4H-SiC detector detector in in response response to to the alpha alpha Figure 4H-SiC Am (blackheight open triangle)ofand Cm (red open circle) source (a)the channel particles emitted from theof243the 243 244 (black open and 244Cm (red open source (a) channel particles emitted emitted from from the the 243Am particles Am openoftriangle) triangle) (red open circle) circle) source number of alpha peak centroid as(black a function appliedand reverseCm bias voltages ranging from(a) 0 tochannel 500 V; number of alpha peak centroid as a function of applied reverse bias voltages ranging from number of resolution alpha peakascentroid as aoffunction applied reverse bias voltages ranging from 00 to to 500 500 V; V; (b) energy a function reverseof bias voltage. (b) energy resolution as a function of reverse bias voltage. (b) energy resolution as a function of reverse bias voltage.

The response spectra to 239Pu, 243Am, and 244Cm alpha particles at a reverse bias of 200 V is shown The response spectra to 239239 Pu, 243Am, and 244Cm particles at a reverse bias of 200 V is shown 243counts 244alpha in Figure 6a, which spectra is expressed of the alpha particles as a function of channel The response to by Pu,the Am, and Cm alpha particles at a reverse bias of number. 200 V is in Figure 6a, which is expressed by the counts of the alpha particles as a function of channel number. Three alpha-particle can be clearly observed. Thealpha energy of alpha as of a function shownsharp in Figure 6a, which peaks is expressed by the counts of the particles asparticles a function channel Three sharp alpha-particle peaks can be clearly observed. The energy of alpha particles as a function of observed peaksharp centroid’s channel number is shown in Figure 6b. The energy andofchannel number number. Three alpha-particle peaks can be clearly observed. The energy alpha particles of observed peak centroid’s channel number is shown in Figure 6b. The energy and2 channel number of centroidofofobserved the three peak peakscentroid’s are linearly correlated withisa shown correlation factor 6b. (R ),The very close to 1. as the a function channel number in Figure energy and of the centroid of the three peaks are linearly correlated with a correlation factor (R2), very close to 21. The average deviation is 1.33 keV over the range of 5.157 MeV and 5.805 MeV. channel number of the centroid of the three peaks are linearly correlated with a correlation factor (R ), The average deviation is 1.33 keV over the range of 5.157 MeV and 5.805 MeV. was made with the peaks Figure 6a, and theand FWHMs of the three very Gaussian close to 1.fitting The average deviation is 1.33 keVacquired over thein range of 5.157 MeV 5.805 MeV. Gaussian fitting was made with the peaks acquired in Figure 6a, and the FWHMs of the three peaks were attained: 183.5 keV for 239Pu, 190.2 keV for 243Am, and 187.7 keV for 244Cm. Many factors peaks were attained: 183.5 keV for 239Pu, 190.2 keV for 243Am, and 187.7 keV for 244Cm. Many factors may contribute to the results: the statistical broadening (about 5.9 keV for 239Pu, 6.0 keV for 243Am, may contribute to the results: the statistical broadening (about 5.9 keV for 239Pu, 6.0 keV for 243Am, and 6.3 keV for 244Cm) [14,30], the energy straggling of the dead layer (about 11 keV) [31], the and 6.3 keV for 244Cm) [14,30], the energy straggling of the dead layer (about 11 keV) [31], the electronic noise (about 10 keV), etc. Excluding the influence of statistical broadening, the dead layer’s electronic noise (about 10 keV), etc. Excluding the influence of statistical broadening, the dead layer’s straggling, and the electronic noise, we attained the inherent FWHMs of 182.8 keV for 239Pu, 189.5 straggling, and the electronic noise, we attained the inherent FWHMs of 182.8 keV for 239Pu, 189.5

Sensors 2017, 17, 2334

6 of 9

keV for Am, and Sensors 2017, 17, 2334

187.0 keV for 244Cm, as well as an optimum energy resolution of about 3.22%6 of at9a reverse bias voltage of 200 V. 243

Sensors 2017, 17, 2334

6 of 9

keV for 243Am, and 187.0 keV for 244Cm, as well as an optimum energy resolution of about 3.22% at a reverse bias voltage of 200 V.

239 Pu, 243 Am, Figure6.6.Response Response spectra of4H-SiC the 4H-SiC detector to the alphafrom particles from ofthe sources of Figure spectra of the detector to the alpha particles the sources 239Pu, 243Am, and 244Cm at a reverse bias of 200 V: (a) alpha counts as a function of channel number; (b) 244 and Cm at a reverse bias of 200 V: (a) alpha counts as a function of channel number; (b) linear fitting linear of alpha particles’ energy vs. channel number of peak centroid. of alphafitting particles’ energy vs. channel number of peak centroid.

3.3. Response Time—Pulsed Radiation Detection Gaussian fitting was made with the peaks acquired in Figure 6a, and the FWHMs of the three 244 Cm. Many peaksThe wereresponse attained:waveforms 183.5 keV for keV for and X-rays 187.7 keV factors of 239 thePu, SiC190.2 detector to 243 theAm, pulsed andfor UV-light are shown in 239 243 Am, may contribute to theheight results:ofthe broadening (aboutsources 5.9 keVwere for normalized, Pu, 6.0 keVthe forresponse Figure 7. If the pulse thestatistical detector for the two pulsed 244 Cm) and 6.3 keV for [14,30], the energy of the dead layer (aboutwaveforms 11 keV) [31], waveforms would be little different. Thestraggling rise time for X-ray and UV-light is the 9.4 electronic ns and 8.0 Figure 6. Response spectra of the 4H-SiC detector to the alpha particles from the sources of noise (aboutthe 10 keV), etc.for Excluding the UV-light influence of statistical broadening, straggling, ns, while FWHM X-ray and waveforms are both 84 the ns.dead The layer’s difference can be 239Pu,243Am, and 244Cm at a reverse bias of 200 V: (a) alpha counts as a function 239 243 Am, ofPu, channel number; attributed to the fact that excitation of charged carriers occurred in the whole sensitive volume and the electronic noise, wethe attained the inherent FWHMs of 182.8 keV for 189.5 keV for(b) 244 linear keV fitting offor alpha particles’ vs. channel number of peakofcentroid. for X-rays, while UV-light, it energy only inenergy the thinresolution layer sensitive near the incident and 187.0 for Cm, as well as anoccurred optimum of aboutvolume 3.22% at a reverse bias surface.of 200 V. voltage 3.3. Response Time—Pulsed Radiation Detection According to Dikinson’s theory [32], the rise time and the FWHM of an SiC detector can be 3.3. Response Time—Pulsed Radiation Detection improved significantly by increasing thedetector detector’s sensitive This effectively The response waveforms of the SiC to the pulsedthickness. X-rays and UV-light are achieves shown ina faster time response. Figure 7. If the pulse height of the detector for the two pulsed sources were normalized, the response The response waveforms of the SiC detector to the pulsed X-rays and UV-light are shown in waveforms would beheight little different. The rise and UV-light waveforms is 9.4 and 8.0 Figure 7. If the pulse of the detector fortime the for twoX-ray pulsed sources were normalized, thensresponse ns, while the FWHM fordifferent. X-ray and both 84 ns. The is difference be waveforms would be little TheUV-light rise time waveforms for X-ray andare UV-light waveforms 9.4 ns andcan 8.0 ns, attributed to the fact that the excitation charged carriers occurred in the whole sensitive volume while the FWHM for X-ray and UV-lightofwaveforms are both 84 ns. The difference can be attributed for X-rays, while forexcitation UV-light,ofit charged only occurred in occurred the thin layer ofwhole sensitive volume near the to the fact that the carriers in the sensitive volume forincident X-rays, surface. while for UV-light, it only occurred in the thin layer of sensitive volume near the incident surface. According to Dikinson’s theory [32], [32], the the rise rise time time and and the the FWHM FWHM of of an SiC detector can be improved significantly by increasing the detector’s detector’s sensitive thickness. This This effectively effectively achieves a faster time response.

Figure 7. Response waveforms of a 4H-SiC detector with a dimension of 20 mm × 20 mm × 20 μm at a reverse bias of 400 V: response waveform to fast pulsed X-ray (blue open circle) and to ultra UVlight (red open triangle).

3.4. Neutron/Gamma Discrimination One of the most important applications of SiC detectors is neutron detection. SiC detectors have a relatively high radiation resistance. It was reported that the dose threshold for the onset of damage in an SiC film detector could be three orders of magnitude higher than that in a silicon PIN detector. 7. waveforms of 4H-SiC detector with withaadimension dimensionof of20 20mm mm××discrimination), 20 mm mm × × 20 [33] Figure Besides, SiC detectors haveof a aahigh neutron/gamma discrimination (n/γ Figure 7. Response Response waveforms 4H-SiC detector 20 20 μm µm at atwhich a reverse bias of 400 V: response waveform to fast pulsed X-ray (blue open circle) and to ultra UVmakes SiC detectors tools for neutrontodetection complex a reverse bias of 400good V: response waveform fast pulsedinX-ray (blue fields. open circle) and to ultra UV-light light (red open triangle). (red open triangle).

3.4. Neutron/Gamma Discrimination One of the most important applications of SiC detectors is neutron detection. SiC detectors have a relatively high radiation resistance. It was reported that the dose threshold for the onset of damage in an SiC film detector could be three orders of magnitude higher than that in a silicon PIN detector. [33] Besides, SiC detectors have a high neutron/gamma discrimination (n/γ discrimination), which

Sensors 2017, 17, 2334

7 of 9

3.4. Neutron/Gamma Discrimination One of the most important applications of SiC detectors is neutron detection. SiC detectors have a relatively high radiation resistance. It was reported that the dose threshold for the onset of damage in an SiC film detector could be three orders of magnitude higher than that in a silicon PIN detector [33]. Besides, SiC17,detectors have a high neutron/gamma discrimination (n/γ discrimination), which makes Sensors 2017, 2334 7 of 9 SiC detectors good tools for neutron detection in complex fields. We Westudied studiedthe then/γ n/γ discrimination of the detectors with a thickness thickness of 20 µm μm and for the neutrons neutrons of of 14 14 MeV and γ-rays of 1.25 MeV using MCNP-4C Code [34], and the results are shown in Figure 8. The discrimination for the neutrons of 14 MeV and the γ-rays of 1.25 MeV is 126, over nine times The n/γ n/γ discrimination times higher higher than than that of of aa silicon silicon detector detector (300 (300 µm μm in in thickness) thickness) and and seven seven times times higher higher than that of of aa diamond diamond detector detector (300 µm μm in thickness) according to the results acquired in our former research [35], respectively. respectively. In neutron neutron detection, detection, γ-rays γ-rays always always exist exist in in the the background. background. The SiC detector detector with aa thin discrimination, thin sensitive sensitive volume volume can can attain attain aa low low response response to to background backgroundradiation radiationand andhigh highn/γ n/γ discrimination, and As aa result, the thin detector shows great advantages and then then attain attain aa high high signal/noise signal/noise ratio. As advantages in in neutron neutron detection detection in in complex complex radiation radiation fields. fields.

Figure 8. 8. n/γ n/γ discrimination neutrons of of 14 14 MeV MeV Figure discrimination of of an an SiC SiC detector detector with with aa thickness thickness of of 20 20 μm, µm, for for the the neutrons and the γ-rays of 1.25 MeV. and the γ-rays of 1.25 MeV.

4. Conclusions Conclusions 4. 2 successfully developed using highLarge-areaSiC SiCdetectors detectors with a sensitive Large-area with a sensitive areaarea of 4 of cm42 cm werewere successfully developed using high-quality quality epitaxial SiC materials andinused the detection of alpha particles and pulsed X-rays/UVepitaxial SiC materials and used the in detection of alpha particles and pulsed X-rays/UV-light. light. The experiment and simulation indicate that the detectors have a thin sensitive volume, low The experiment and simulation indicate that the detectors have a thin sensitive volume, a lowadark dark current, a good energy resolution, and a high n/γ discrimination, though their dimensions are current, a good energy resolution, and a high n/γ discrimination, though their dimensions are similar similar with conventional Si detectors. These large-area SiCoffer detectors offer an important for with conventional Si detectors. These large-area SiC detectors an important option for theoption detection the detection in large radiation fields, the application of SiC detectors will thus no longer be affected in large radiation fields, the application of SiC detectors will thus no longer be affected by the limitation bydimensions. the limitation of the dimensions. With theresistance excellentand radiation resistance and outstanding highof With excellent radiation outstanding high-temperature endurance, temperature endurance, SiC in detectors be more useful in radiation detection in harsh SiC detectors will be more useful radiationwill detection in harsh environments and intense radiation fields. environments and intense radiation fields.

Acknowledgments: The National Natural Science Foundation of China (Grant Nos. 11605140, 11435010) financially supportedThe this National work. TheNatural authorsScience want toFoundation thank Xu-Hui Li, and Yong-Ning He in Acknowledgments: of Wang, China Fang-Pei (Grant Nos. 11605140, 11435010) Xi’an Jiaotong University their The helpauthors in I-V and C-V Yang Li, Chun-Lei Su, and Yi-Hua financially supported thisfor work. want tomeasurements, thank Xu-Hui and Wang, Fang-Pei Li, and Yong-Ning HeDai in in NINT for the supply of the pulsed X-ray/UV-light source and the alpha source. Xi’an Jiaotong University for their help in I-V and C-V measurements, and Yang Li, Chun-Lei Su, and Yi-Hua Author Contributions: Lin-Yue and Xiao-Ping Ouyang designed experiment, Dai in NINT for the supply of theLiu pulsed X-ray/UV-light source and thethe alpha source. finished alpha-particle detection, and wrote the main manuscript text. Ling Wang, Ao Liu, Run-Hua Huang, and Song Bai designed and Author Contributions: Lin-Yue LiuJin-Liang and Xiao-Ping Ouyang Zhang, designed theChen, experiment, finished alpha-particle fabricated the SiC detector. Peng Jin, Liu, Xian-Peng Liang and Jian-Fu Zhang carried out some of theand measurements. All authors reviewed the manuscript. detection, wrote the main manuscript text. Ling Wang, Ao Liu, Run-Hua Huang, and Song Bai designed and fabricated the SiC Jin, Xian-Peng Zhang, Liang Chen, and Jian-Fu Zhang Conflicts of Interest: Thedetector. authors Peng declare noJin-Liang conflict ofLiu, interest. carried out some of the measurements. All authors reviewed the manuscript.

Conflicts of Interest: The authors declare no conflict of interest.

References 1.

Babcock, R.; Ruby, S.; Schupp, F.; Sun, K. Miniature Neutron Detectors; Westinghouse Electric Corporation Materials Engineering Report No. 5711-6600-A; Westinghouse Electric Corporation: Pittsburg, PA, USA, 1957.

Sensors 2017, 17, 2334

8 of 9

References 1.

2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13.

14.

15.

16.

17. 18. 19. 20. 21.

22.

Babcock, R.; Ruby, S.; Schupp, F.; Sun, K. Miniature Neutron Detectors; Westinghouse Electric Corporation Materials Engineering Report No. 5711-6600-A; Westinghouse Electric Corporation: Pittsburg, PA, USA, 1957. Babcock, R.; Chang, H. Silicon Carbide Neutron Detectors for High-Temperature Operation. In Reactor Dosimetry; International Atomic Energy Agency: Vienna, Austria, 1963; Volume 1, p. 613. Franceschini, F.; Ruddy, F.H. Silicon Carbide Neutron Detectors, Properties and Applications of Silicon Carbide; Gerhardt, R., Ed.; InTech: Rijeka, Croatia, 2011. Takahashi, T.; Watanabe, S. Recent progress in CdTe and CdZnTe detectors. IEEE Trans. Nuclear Sci. 2001, 48, 950–959. [CrossRef] Butera, S.; Gohil, T.; Lioliou, G.; Krysa, A.B.; Barnett, A.M. Temperature study of Al0.52 In0.48 P detector photon counting X-ray spectrometer. J. Appl. Phys. 2016, 120, 024502. [CrossRef] Kozhevnikov, D.; Chelkov, G.; Demichev, M.; Gridin, A.; Smolyanskiy, P.; Zhemchugov, A. Performance and applications of GaAs:Cr-based Medipix detector in X-ray CT. J. Instrum. 2017, 12, C01005. [CrossRef] Rebai, M.; Cazzaniga, C.; Croci, G.; Tardocchib, M.; Cippob, E.P.; Calvanic, P.; Girolamic, M.; Trucchic, D.M.; Grossob, G.; Gorinia, G. Pixelated Single-crystal Diamond Detector for fast neutron measurements. J. Instrum. 2015, 10, C03016. [CrossRef] Liu, L.Y.; Ouyang, X.P.; Zhang, J.F.; Jin, P.; Su, C.L. Properties comparison between nanosecond X-ray detectors of polycrystalline and single-crystal diamond. Diam. Relat. Mater. 2016, 73, 248–252. [CrossRef] Szalkai, D.; Ferone, R.; Issa, F.; Klix, A. Fast Neutron Detection with 4H-SiC Based Diode Detector up to 500 ◦ C Ambient Temperature. IEEE Trans. Nuclear Sci. 2016, 63, 1491–1498. [CrossRef] Abubakar, Y.M.; Lohstroh, A.; Sellin, P.J. Stability of Silicon Carbide Particle Detector Performance at Elevated Temperatures. IEEE Trans. Nuclear Sci. 2015, 62, 2360–2366. [CrossRef] Moscatelli, F. Silicon carbide for UV, alpha, beta and X-ray detectors results and perspectives. Nucl. Inst. Methods Phys. Res. A 2007, 583, 157–161. [CrossRef] Bruzzi, M.; Nava, F.; Pini, S.; Russo, S. High quality SiC applications in radiation dosimetry. Appl. Surf. Sci. 2001, 184, 425–430. [CrossRef] Ruddy, F.H.; Seidel, J.G.; Sellin, P. High-resolution alpha spectrometry with a thin-window silicon carbide semiconductor detector. In Proceedings of the IEEE Nuclear Science Symposium Conference Record, Orlando, FL, USA, 24 October–1 November 2009; pp. 2201–2206. Liu, L.Y.; Liu, J.L.; Chen, L.; Zhang, Z.B.; Jin, P.; Ruan, J.L.; Chen, G.; Liu, A.; Bai, S.; Ouyang, X.P. Properties of 4H silicon carbide detectors in the radiation detection of 86 MeV oxygen particles. Diam. Relat. Mater. 2017, 73, 177–181. [CrossRef] Dubecký, F.; Gombia, E.; Ferrari, C.; Zat’ko, B.; Vanko, G.; Baldini, M.; Kováˇc, J.; Baˇcek, D.; Kováˇc, P.; Hrkút, P.; et al. Characterization of epitaxial 4H-SiC for photon detectors. J. Instrum. 2012, 7, P09005. [CrossRef] Nava, F.; Vittone, E.; Vanni, P.; Verzellesi, G.; Fuochi, P.G.; Lanzieri, C.; Glaser, M. Radiation tolerance of epitaxial silicon carbide detectors for electrons, protons and gamma-rays. Nucl. Inst. Methods Phys. Res. A 2003, 505, 645–655. [CrossRef] Lees, J.E.; Barnett, A.M.; Bassford, D.J.; Stevens, R.C.; Horsfall, A.B. SiC X-ray detectors for harsh environments. J. Instrum. 2011, 6, C01032. [CrossRef] Bertuccio, G.; Caccia, S.; Nava, F.; Preti, F. Ultra low noise epitaxial 4H-SiC X-ray detectors. Mater. Sci. Forum 2009, 615–617, 845–848. [CrossRef] Ha, J.H.; Kang, S.M.; Kim, H.S.; Park, S.H.; Lee, N.H.; Song, T.Y.; Lee, J.H.; Park, H.; Kim, J. 4H-SiC PIN-type semiconductor detector for fast neutron detection. Prog. Nucl. Sci. Technol. 2011, 237–239. [CrossRef] Seshadri, S.; Dulloo, A.R.; Ruddy, F.H.; Seidel, J.G.; Rowland, L.B. Demonstration of an SiC neutron detector for high-radiation environments. IEEE Trans. Electron Devices 1999, 46, 567–571. [CrossRef] Giudicea, A.L.; Fasolo, F.; Durisi, E.; Manfredotti, C.; Vittone, E.; Fizzotti, F.; Zanini, A.; Rosi, G. Performances of 4H-SiC Schottky diodes as neutron detectors. Nucl. Inst. Methods Phys. Res. A 2007, 583, 177–180. [CrossRef] Wu, J.; Lei, J.; Jiang, Y.; Chen, Y.; Rong, R.; Zou, D.; Fan, X.; Chen, G.; Li, L.; Bai, S. Feasibility study of a SiC sandwich neutron spectrometer. Nucl. Inst. Methods Phys. Res. A 2013, 708, 72–77. [CrossRef]

Sensors 2017, 17, 2334

23. 24.

25. 26. 27.

28. 29. 30. 31. 32.

33. 34. 35.

9 of 9

Ruddy, F.H.; Seidel, J.G.; Chen, H.; Dulloo, A.R.; Ryu, S.H. High-resolution alpha-particle spectrometry using silicon carbide semiconductor detectors. IEEE Trans. Nucl. Sci. 2006, 53, 1713–1718. [CrossRef] Dulloo, A.R.; Ruddy, F.H.; Seidel, J.G.; Flinchbaugh, T.; Davison, C.; Daubenspeck, T. Neutron and Gamma Ray Dosimetry in Spent-Fuel Radiation Environments Using Silicon Carbide Semiconductor Radiation Detectors. In Reactor Dosimetry: Radiation Metrology and Assessment; ASTM STP 1398; American Society for Testing and Materials: West Conshohoken, PA, USA, 2001; pp. 683–690. Natsume, T.; Doi, H.; Ruddy, F.; Seidel, J.G.; Dulloo, A.R. Spent Fuel Monitoring with Silicon Carbide Semiconductor Neutron/Gamma Detectors. J. ASTM Int. 2005, 3, 1–8. Ruddy, F.H.; Dulloo, A.R.; Seidel, J.G.; Palmour, J.W.; Singh, R. The charged particle response of silicon carbide semiconductor radiation detectors. Nucl. Inst. Methods Phys. Res. A 2003, 505, 159–162. [CrossRef] Wu, J.; Li, M.; Jiang, Y.; Li, J.; Zhang, Y.; Gao, H.; Liu, X.; Du, J.; Zou, D.; Fan, X.; et al. Performance of a 4H-SiC Schottky diode as a compact sized detector for neutron pulse form measurements. Nucl. Inst. Methods Phys. Res. A 2015, 771, 17–20. [CrossRef] Zhang, X.; Ouyang, X.; Chen, Y.; Zhang, Z.; Tian, G.; Chen, L.; Liu, J. A Si-PIN-stack detector for 14 MeV pulsed neutrons measurement. Nucl. Inst. Methods Phys. Res. A 2012, 693, 1–5. [CrossRef] Wu, J.; Jiang, Y.; Lei, J.; Fan, X.; Chen, Y.; Li, M.; Zou, D.; Liu, B. Effect of neutron irradiation on charge collection efficiency in 4H-SiC Schottky diode. Nucl. Inst. Methods Phys. Res. A 2014, 735, 218–222. [CrossRef] Yamaya, T.; Asano, R.; Endo, H.; Umeda, K. Measurement of the Fano factor for protons on silicon. Nucl. Inst. Methods 1979, 159, 181–187. [CrossRef] Ziegler, J.F. SRIM-2003. Nucl. Inst. Methods Phys. Res. B 2004, 219–220, 1027–1036. [CrossRef] Dickinson, W.C.; Lauzon, A.F.; Neifert, R.D.; Lent, E.M. Response Function and Sensitivity of Double-Diffused Silicon Detectors in High γ-Dose Rate Fields; Report of Lawrence Livermore National Laboratory, UCRL-14405; Lawrence Livermore National Laboratory: Livermore, CA, USA, 1965. Kuckuck, R.W. Semiconductor Detectors for Use in the Current Mode; Report of Lawrence Livermore National Laboratory, UCRL-51011; Lawrence Livermore National Laboratory: Livermore, CA, USA, 1971. MCNP4C Mont Carlo N-Particle Transport Code System: Report of Los Alamos National Laboratory; Los Alamos National Laboratory: Los Alamos, NM, USA, 2000. Liu, L.Y.; Ouyang, X.P.; Zhang, Z.B.; Zhang, J.F.; Zhang, X.P.; Zhong, Y.H.; Wang, W. Polycrystalline chemical-vapor-deposited diamond for fast and ultra-fast neutron detection. Sci. China Technol. Sci. 2012, 55, 2640–2645. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).