- Email: [email protected]
Thin single crystal diamond detectors for alpha particle detection
Diamond & Related Materials 49 (2014) 96–102
Contents lists available at ScienceDirect
Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond
Thin single crystal diamond detectors for alpha particle detection F. Schirru c,⁎,1, D. Chokheli a,b,1,2, M. Kiš c,1 a b c
JINR, Russia HEPI, Tbilisi State University, Georgia GSI, Darmstadt, Germany
a r t i c l e
i n f o
Article history: Received 20 December 2013 Received in revised form 11 April 2014 Accepted 1 August 2014 Available online 10 August 2014 Keywords: Thin samples Diamond detectors Alpha Spectroscopy CVD diamond
a b s t r a c t Charged particle detectors based on thin single crystal diamond films (3.0 × 3.0 × 0.09 mm3) grown by chemical vapour deposition (CVD) technique are developed at Ruđer Bošković Institute and low field mobility, transit time and saturation velocity, measured by using a 210Po alpha source in vacuum at room temperature, are presented. The comparison of the charge transport properties obtained by time of flight (ToF) technique between detectors based on the same diamond samples, but with different metallization of the electrodes (Au and Al, 100 nm thick), and commercially purchased diamond detectors is performed. The fabricated diamond sensors showed a spectroscopic resolution of up to 1%, mobility up to 1808 ± 16 and 1914 ± 71 cm2/V s for electrons and holes respectively and carrier transit time below 2 ns for an applied electric field of 17.8 × 103 V/cm. Results also show variations of the charge transport properties according to the electrode metallization used and the different modality at which the detectors are irradiated. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The capability of single-crystal CVD (scCVD) diamond detectors to detect different types of radiation from the visible and ultraviolet wavelengths [1–3] up to the X- and γ-rays [4,5] and charged particles as well [6–8] has been widely demonstrated over the past few decades. The interest in using diamond as a radiation detector stems from its unique properties such as the high band gap (5.47 eV) [9] which leads to a low dark current and a high breakdown field (N 1 × 103 V/μm), the resilience to radiation damage due to the strong bonds of the crystal structure [10,11] and the room temperature thermal conductivity that is the highest of any other material [12]. A better understanding of the detection performance of the diamond detector can be achieved by performing a study of its main electronic properties, namely the low field mobility and the transit time of the charge carriers. The former describes the rate at which the carrier drift velocity increases with the applied electric field and mainly varies according to the experimental conditions adopted and the sample quality. In a typical experiment alpha-particles are used to induce currents. Values of low field mobility at room temperature of ~2300 cm2/V s and ~1700 cm2/V s, for holes and electrons respectively [13], have been reported. On the other hand, in experiments with laser-induced currents, mobility for holes was found to be in the range of 2000–2250 cm2/V s and for electrons in the range of 2200–2750 cm2/V s [14]. Other studies ⁎ Corresponding author. E-mail address: [email protected] (F. Schirru). 1 Formerly affiliated with Ruđer Bošković Institute, Zagreb, Croatia. 2 On leave.
http://dx.doi.org/10.1016/j.diamond.2014.08.001 0925-9635/© 2014 Elsevier B.V. All rights reserved.
have shown hole and electron mobility up to 3800 cm2/V s and 4500 cm2/V s respectively [15]. The electronic properties of a diamond sample can be studied by the time of flight (ToF) technique (also known as TCT or transient current technique) in which the duration of the current pulse induced on a read-out electrode by the drift of free charge carriers under the influence of an externally applied field is measured. For the ToF technique we assume that the effective charge carrier lifetime is much longer than the drift time tdr, so the current pulse width equals the time the charge carriers need to traverse the detector. In addition, it has to be recalled that the ToF measurements may depend on the charge density injected especially in the case of high charge densities. Considering a 5.4 MeV alpha particle, the charge carriers are generated in the first 16 μm below the diamond surface and, depending on the direction of the electric field applied, only one type of charge carrier will cross the whole detector thickness and mainly contribute to the signal. In first approximation, by knowing the drift time tdr and the thickness d of the diamond sample, the drift velocity of the carrier vdr can be easily calculated as v dr = d/tdr . On the other hand, by calculating the drift velocity as a function of the electric field applied, it is then possible to extract the low field mobility μ 0 and saturation velocity (maximum drift velocity) of the carrier vsat as [13]: vdr ¼
μ0E μ E 1þ 0 vsat
ð1Þ
Electronic properties of thick (usually 300–500 μm) scCVD diamonds have been studied using the ToF technique and discussed by many authors [16–19]. On this topic, however, little can be found
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
97
in literature about thin samples with thickness below 100 μm. Moreover, there are few papers which present a comparative study between the electronic properties of diamond devices and different types of electrodes [20–22]. For these reasons we report on the electronic properties (low field mobility, transit time and saturation velocity) and spectroscopic resolution of charge particle detectors based on 90 μm thin (scCVD) diamond samples metallized with gold (Au) and aluminium (Al) electrodes. The results, derived from measurements taken at room temperature by the ToF technique, are also compared with those obtained from commercially available scCVD diamond detectors.
2. Material and methods
Fig. 1. Schematic of the detector PCB on which the diamond samples Dmd#1 and Dmd#2 were mounted. The diamond detector signals were fed to the preamplifier by a standard socket SMA connector.
The material considered for this work consisted of: 3. Experimental setup • 2 × scCVD (electronic grade) diamond samples 3.0 × 3.0 × 0.09 mm3 purchased from Element Six and labelled as Dmd#1 and Dmd#2.3 The corresponding detectors were fabricated in house with Au or Al/Au electrodes in a sandwich configuration with squared shape and area of 4 mm2. • 2 × scCVD diamond detectors 4.6 × 4.6 × 0.5 mm3 purchased from CIVIDEC with both surfaces fully covered by 300 nm Al electrodes. These detectors, labelled B10044 and B10045 were used for comparison purposes only. Prior to deposition of Au electrodes, Dmd#1 and Dmd#2 were at first thoroughly cleaned in acetone, hysopropanol and de-ionized water. Deposition of 100 nm Au electrodes was performed by using a shadow mask and a SC7620 ‘Mini’ Sputter operating at a working pressure of 10−3 Torr. Once the measurements on the sample Dmd#1 were completed, Au contacts were removed from the diamond sample using a solution of aqua regia. New Al/Au electrodes with a thickness of 100/30 nm were then deposited onto Dmd#1 by thermal evaporation at the working pressure of 10−5 Torr. First a 100 nm Al layer was deposited followed by 30 nm of Au layer which served to prevent oxidisation of the Al itself. After deposition of electrodes, samples were installed and fixed by vinyl-acetate non-conductive glue onto an in-house built printed circuit board (PCB) based on standard Fr-4 whose schematic is shown in Fig. 1. In order to minimize the influence of the properties of the PCB on the detector performances, dimensions of its copper pads, used to transfer the signals from the detectors to the preamplifier via SMA connector, were minimized. The noise generated by the PCB itself was in fact reduced by decreasing its capacitance and increasing as much as possible the resistance of the housing. The sample electrodes were then connected to a gold wire of 25 μm in diameter (mounted on a K&S 4123 wedge bonder) by silver loaded conductive epoxy. The second end of the wire was next connected to the PCB copper pads by fast drying silver paint. After successful fabrication of the diamond detectors, a set of measurements was then performed with the aim to: • investigate the spectroscopic and charge transport properties of the diamond detectors Dmd#1 and Dmd#2 fabricated with Au electrodes; • compare the properties of spectroscopic resolution of Dmd#1 and Dmd#2 with those obtained from the diamond detectors B10044 and B10045; • compare the charge transport properties of the diamond sensor Dmd#1 fabricated first with Au and then with Al/Au electrodes.
3
Samples provided by Dr. Elèni Berdermann (GSI, Darmstadt, Germany).
Spectroscopic and charge transport properties of the diamond detectors were investigated by irradiating the samples with a 5.4 MeV alpha-particles from 210Po source whose activity was about 0.02 μCi when the experiments started. The devices were mounted, along the alpha source, inside a newly developed vacuum chamber (details can be found in a forthcoming paper [23]). Measurements were recorded at the pressure of ~ 10−2 Torr. Irradiations of the devices were performed in two different modes: by placing the biased or the grounded electrode in front of the alpha source. Henceforth these irradiation modes will be called “forward” or “backward”, respectively. The diamond sensors were positioned carefully in front of the alpha source by means of a translation stage purchased from OWIS®. The induced charge carriers i.e. the generated signal was measured either with a charge sensitive preamplifier (CANBERRA 2004) for spectroscopic characterization or with a current sensitive (DBA IV) preamplifier for a ToF characterization by using respectively RG-58 and RG-316 50 Ω impedance cables whose lengths (40 cm) are as short as possible to minimize the loss of signal. For spectroscopic measurements (Fig. 2 (left)), the signal is further fed into the amplifier (CANBERRA AFT Research Amplifier 2025) and then displayed and acquired by an oscilloscope (Tektronix DPO4054) with a sampling time of 400 ps. For ToF measurements (Fig. 2 (right)), the DBA IV output is directly connected to the oscilloscope for signal visualization and acquisition. In both configurations, a NIM crate HV based supply C.A.E.N. N1470, connected to PC via USB/RS488 link, is used to supply the proper bias voltage to the diamond detectors. Measurements on Dmd#1 and Dmd#2 detectors were performed in both bias polarities and in the voltage range of 8–160 V which corresponds to a maximum applied electric field of 17.8 kV/cm. Due to the geometry of the electrodes, the devices B10044 and B10045 were biased up to 300 V only which corresponds to an electric field of 6 kV/cm. 4. Data processing Data processing of the detector signals was performed according to the type of preamplifier used during the experiments. For each voltage applied, a set of N = 1000 events was recorded from the oscilloscope and then stored in a file. The alpha particle rate measured by the detectors was below 10 s −1 as theoretically expected due to the geometry of the experimental setup. 4.1. Data processing with the charge sensitive preamplifier The signal generated in the detector is directly integrated online by the charge sensitive preamplifier (CANBERRA 2004) and processed by the AFT Research Amplifier (CANBERRA 2025). The latter gives in output a semi-Gaussian pulse shape and a waveform of the resulting signal is stored in the PC via the digital oscilloscope Textronix
98
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Fig. 2. Schematic of the experimental setup using: (left) a charge sensitive amplifier (Camberra 2004 in chain with Camberra 2025) and (right) a current sensitive (DBA IV) preamplifier. In this case the DBA IV is connected to a HAMEG HMP 2030 programmable power supply.
DPO4054. The charge sensitive preamplifier allows better noise condition for the calculation of the spectroscopic resolution of the diamond devices. On the other hand, due to the high RC constant of the preamplifier, the information about the rise time of the signal and, as a consequence, the transit time of the carriers, is completely lost. Reconstruction of the integrated signal is performed offline fitting the oscilloscope waveform by a combination of Gaussian and polynomial functions (Eq. (2)) since it represents a good approximation of the recorded pulse. 2
y ¼ y0 þ A exp
−ðx−μ Þ 2σ 2
! ð2Þ
The integral of each fitted pulse Pi (that is proportional to the total charge Qi collected in the single event) is then plotted to get the total charge distribution (Fig. 3) whose Gaussian fit returns the information about the diamond detector resolution for that particular bias voltage applied. The detector resolution was calculated according to the ratio: R¼
FWHM μ′
ð3Þ
where FWHM = 2.35σ′, σ′ and μ′ are respectively the standard deviation and mean of the charge distribution found by the Gaussian fit. 4.2. Data processing with the current sensitive preamplifier In this case, integration of the detector signals, estimation of the pulse rise time and transport properties of the charge carriers are
Fig. 3. Charge distribution of the diamond detector Dmd#2 referred to an applied electric field of 5.5 kV/cm. For this case a spectroscopic resolution of 1.9% was achieved.
performed offline. Calculation of the diamond detector resolution is carried out as explained in the previous subsection by using the Eqs. (2) and (3). As for the transit time of the charge carriers, this is calculated by taking the width of each pulse after a preliminary pulse baseline adjustment. Then, the generated transit time distribution at that particular voltage applied is fitted with a Gaussian function to estimate the mean pulse width. 5. Results The following subsections describe the measurements performed on the diamond detectors. At first their leakage current, charge collection efficiency and spectroscopic resolution under alpha particle irradiation were investigated. Next, the electronic properties of the charge carriers were estimated by the ToF technique and compared with those obtained for devices with Au and then with Al/Au electrodes. 5.1. Leakage current The leakage current of the diamond detectors was the first parameter investigated. Measurements were performed both in darkness and under illumination conditions by using a metallic enclosure. Conditions of illumination were achieved by a white LED mounted inside the box and connected to a voltage supply of 5 V via 100 Ω current limiting resistor. The luminous intensity of the LED corresponds to 80–110 cd in 20 mA direct current applied with color temperature of 10,000 K. Fig. 4 shows the leakage current measured in darkness and under conditions of illumination of the diamond detectors Dmd#1 and Dmd#2 with Al (blue and cyan data points) and Au (orange and red data points) electrodes in the bias range of ± 160 V. Both detectors present similar behavior of the leakage current and its magnitude does not change significantly regardless the fact that the measurements are taken under dark conditions or not. Such results suggest the possibility to use the detectors in the presence of light. It can be also observed that the leakage current recorded from the metallic box itself with the diamond detector disconnected (green data points on the graph) is very low and basically does not influence the detector performance. Considering the maximum applied electric field to the devices (17.8 kV/cm), the diamond detector with Au electrodes showed the highest value of leakage current. Taking a value of 1.04 × 10− 10 A, it yields a bulk resistance of ~ 1.5 × 1012 Ω. Next, the bulk resistivity of the diamond sample can be estimated as ρ = R(S/l) where R is the resistance of the diamond (calculated for example at + 160 V), S is the electrode area and l is the detector thickness. Assuming S = 4 mm2 and l = 90 μm, it yields a bulk resistivity for the sample of ~(6.7 ± 1.4) × 1012 Ω cm. For the case of the commercial devices B10044 and B10045, their measured leakage current did not exceed 10−11 A for a maximum applied electric field of ± 6 kV/cm.
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Fig. 4. Leakage current measurements of the diamond detectors with Au (orange and red data points) and Al (blue and cyan data points) electrodes. The green points represent the leakage current measured from the metallic enclosure.
5.2. Charge collection efficiency and spectroscopic resolution in vacuum The charge collection efficiency (CCE) of the diamond detectors was the next parameter investigated. Measurements were taken at different bias voltage irradiating the detectors in forward mode. In this case therefore, at the ground electrode, holes are collected applying a positive bias while electrons are collected applying a negative bias. Data were acquired with the experimental setup reported in Fig. 2 (left). Fig. 5 (left) shows the CCE for both types of charge carriers as a function of the applied electric field up to 17.8 × 103 V/cm. Data are related to the diamond sensor Dmd#2 with Au electrodes. The CCE on diamond was estimated by first irradiating a Hamamatsu S-5821 PIN silicon diode which has 100% of CCE and then taking into account the different values of energy required to create an electron–hole pair in the two materials that are usually quoted as 3.6 eV for silicon and 13 eV for diamond [24]. As shown in Fig. 5 (left), the CCE of the device Dmd#2 increases with the applied electric field as more charge carriers are collected at the electrodes. For a bias applied N 50 V the device reaches a 100% of CCE and remains constant in the range of 50–160 V. Also the spectroscopic resolution of the detector (shown in Fig. 5 (right) and calculated according to Eq. (3)) presents analogous behavior and remains constant to a value of 1.2% for applied biases N 50 V.
99
In a subsequent test, keeping the bias at 160 V but irradiating the device at an incidence angle of ±45°, the resolution worsened to 2.0%. This fact could be mainly related by the thicker electrode which the alpha particle has to cross due to the geometrical configuration but also by the different spread of charge across the sample due to the shorter penetration length and the different angle of the applied electric field. This measurement however shows that even for wide incident angles, the detector resolution is still acceptable. The data reported on Fig. 5 (left) and 5 (right) clearly show a correlation between values of CCE and spectroscopic resolution: a decrease of the CCE corresponds also to a decrease of the spectroscopic resolution of the device. The measurements performed on the device Dmd#1 led to the same results and conclusions. For a meaningful comparison between the different diamond detectors, Table 1 reports the CCE and the spectroscopic resolution of the devices calculated for the same magnitude of applied electric field, that is − 6.0 × 103 V/cm. Although for such an electric field also the commercial sensors show 100% of CCE [25], the best spectroscopic resolution is achieved by the detectors made of thinner samples. This could be justified by taking into account that with thicker material it increases the probability of carrier trapping which also causes a bigger variation in the charge collection at the electrodes. Moreover, the commercial devices have thicker electrodes which could also contribute to the worse resolution observed. Anyway, these results may suggest the preference in using thinner samples for alpha spectroscopy purposes, provided that the thickness is large enough to detect the alpha particle. 5.3. Spectroscopic resolution in air Fig. 6 reports the spectroscopic resolution in air of the device Dmd#1 with Al electrodes as a function of the distance from the alpha source. Due to the particular construction of the diamond detector holder, the closest distance at which the devices could be irradiated was 5 mm. For this measurement, the alpha source housing and the electrode geometry of the detectors served as natural collimators of 6 and 3 mm in diameter respectively. In this way only alpha particles which performed almost a straight path toward the detector were collected. The device showed a resolution of 4% up to 1 cm away from the alpha source and then decreased. This measurement was obtained under illumination applying a positive voltage of 160 V on the sensor and irradiating it in backward mode (therefore the signal is mainly due to the electrons). Opposite to devices like those based on silicon, diamond detectors offer the possibility to perform measurements in air and under condition of illumination while keeping a reasonable good value of spectroscopic resolution.
Fig. 5. (Left) Charge collection efficiency CCE and (right) resolution of the diamond sensor Dmd#2 as a function of the electric field applied. The inset on the left graph represents the estimated CCE for holes up to an applied electric field of 2.5 kV/cm. Measurements were taken placing the ground electrode in front of the alpha source.
100
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Table 1 CCE and spectroscopic properties of the investigated diamond detectors. The parameters were obtained for an applied electric field of −6.0 × 103 kV/cm. Device
Electr.
Thickn. (nm)
CCE (%)
Res. (%)
Dmd#1
Au Al/Au Au Al Al
100 100/30 100 300 300
100 100 100 100 100
1.2 1.8 1.2 1.9 3.8
Dmd#2 B10044 B10045
In vacuum and at the same experimental conditions, the resolution achieved by the detector was 1.3%. However, even in the presence of air, a good spectroscopic resolution can be achieved if the detector is located close to the alpha source in order to minimize the beam straggling. Similar results were also obtained for the device Dmd#2 with Au electrodes. 5.4. Charge transport properties The next part of the experiments was focused on the ToF low field mobility measurement of the charge carriers. Irradiation of the devices was performed at room temperature (300 K) using the experimental setup reported in Fig. 2 (right). These experiments aimed also at comparing, according to the mode of irradiation of the devices, the values of mobility obtained for electrons and holes with the deduced spectroscopic resolution. However, it should be recalled that, in this case, the spectroscopic resolution is calculated offline by integration of the current pulse and therefore the values are expected to be somehow worse than those reported in Table 1. Fig. 7 shows the drift velocity of the charge carriers against the applied electric field for the Dmd#2 device with Au electrodes. Measurements were taken irradiating the device in forward mode and therefore the contribution to the signal was mainly given by electrons or holes according to the applied bias negative or positive. Increasing the applied electric field, the recorded current pulses become narrower as shown in (Fig. 8) and, as a consequence, the drift velocity becomes higher. The drift velocity vdr reported on the graph was calculated by the knowledge of the thickness of the diamond sample divided by the drift time estimated from the current pulse width as described in subsection 4.2. For an electric field of + 6.0 × 103 V/cm, the drift time of electrons and holes calculated for the device Dmd#2 was ~ 1.5 ns while ~ 7 ns was estimated for the commercial devices made up of thicker diamond samples. The variation in the drift time between different measurements was found to be in the order of 5% which has to be attributed more to
Fig. 6. Resolution of the device Dmd#1 with Al electrodes as a function of the distance from the alpha source. Measurements were taken in air applying a positive bias voltage. Irradiations were performed placing the ground electrode in front of the alpha source.
Fig. 7. Drift velocity against the electric field of the detector Dmd#2 irradiated placing the biased electrode in front of the alpha source. The measured values (data points) were fitted using Eq. (1).
the random errors than to the amount of absorbed dose due to the low irradiation rate. According to Eq. (1), the fit performed on the data points led to a value of mobility for electrons of 1628 ± 12 cm2/V s and a saturation velocity of (1.199 ± 0.007) × 107 cm/s. As for the holes the values found were 1539 ± 81 cm 2 /V s and (1.521 ± 0.078) × 107 cm/s. The device Dmd#2 was then irradiated in backward mode at the same experimental conditions. In this configuration for a positive or negative bias applied the signal is mainly given by the electrons or holes respectively. Values of 1658 ± 19 and 1914 ± 71 cm2/V s for low field mobility and values of (1.255 ± 0.012) × 107 cm/s and (1.411 ± 0.042) × 107 cm/s for the saturation velocity were found for electrons and holes. These results show that the mobility of holes is also influenced by the mode of irradiation. Considering the data reported on Table 2, the observed increase in the holes mobility leads also to an improvement of the spectroscopic resolution of the detector. It has to be noted that the fit related to the holes in Fig. 7 is worse than that obtained for electrons. This can be linked to the fact that if the holes have shorter drift times then, the oscilloscope used to extract and process the waveforms might not be fast enough (especially in the high electric field region) to sample correctly the real pulse which leads to an underestimation of the drift velocity of the charge carriers. In these measurements, according to the polarity of the bias applied and the position of the detector with respect to the alpha source, the same type of charge carrier drifts inside the diamond in opposite directions. Since the electrodes of the device are fabricated in similar
Fig. 8. Time evolution of the recorded pulses from the device Dmd#1 (carrier: holes) at different voltage applied. Each pulse represents an average over 100 events.
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102 Table 2 Electronic properties and resolution of the developed devices Dmd#1 and Dmd#2 calculated according to the different mode of irradiation and electrode metallization. The spectroscopic resolution is calculated for an applied electric field of ±17.8 × 103 V/cm. Mobility and saturation velocity are calculated by fitting the drift velocity against the applied electric field. According to the mode of irradiation (fw = forward, bw = backward), main contribution to the generated signal stems from electrons (e) or holes (h). Device
Electr.
Irrad. mode
Carrier
Sat. vel. (107 cm/s)
Mobility (cm2/V s)
Res. (%)
Dmd#1
Au
fw fw fw fw fw fw bw bw
e h e h e h e h
1.162 1.305 1.090 2.113 1.199 1.521 1.255 1.411
1463 1424 1808 1036 1628 1539 1658 1914
4.1 3.6 3.3 3.2 4.1 3.9 4.0 3.5
Al Dmd#2
Au Au
± ± ± ± ± ± ± ±
0.005 0.019 0.004 0.087 0.007 0.078 0.012 0.042
± ± ± ± ± ± ± ±
15 44 17 61 12 82 19 71
way and since the experimental conditions are the same, a change in the mobility of the charge carriers could be linked only to the different way at which they cross the detector bulk. If this is the case, a complete separate study is required to understand in detail the properties of the diamond bulk. 5.5. Role of the electrode metallization on the charge transport properties The device Dmd#1 was next used to study the electronic properties of the charge carriers as a function of the different type of metallization used to create the electrodes. In this case the irradiations were performed in forward mode only. Fig. 9 reports the drift velocity against the electric field applied for the case of the diamond device Dmd#1 fabricated first with Au and then Al electrodes. As reported in Table 2, the mobility for electrons and holes, estimated by fitting the curves reported in Fig. 9, changes considerably according to the type of electrode metallization. With Au electrodes, charge carriers show similar low field mobility that is, 1463 ± 15 cm2 /V s for electrons and 1425 ± 44 cm2/V s for holes. With Al electrodes, however, the calculated values are 1036 ± 61 cm2/V s for holes and 1808 ± 17 cm2/V s for electrons. According to Fig. 9, the total transit time of the charge carriers could be influenced, in the low electric field region, by the Schottky barrier which Al forms on diamond [26]. For higher applied electric field, the highest drift velocity observed is related to the device with Al electrodes which also shows the best spectroscopic resolution. It should be taken into account that the obtained values of mobility are calculated by
101
fitting the data points with Eq. (1) which does not include the effect of the Schottky barrier and therefore they can be considered only approximately. 6. Conclusions This paper presented the spectroscopic and electronic properties of charge particle detectors based on 90 μm thick samples and provided of Au or Al electrodes. The electronic properties of the devices were measured in vacuum and at room temperature with the ToF technique using a 210Po alpha source. The results were compared with those obtained from two commercially purchased diamond detectors 500 μm thick. The first part of the experiments was focused on the evaluation of the spectroscopic resolution and CCE of the developed detectors fabricated with Au electrodes. It was found that, compared to the commercial detectors, the devices based on thinner diamond samples showed better spectroscopic performance with a resolution achieved for alpha particles up to 1.2% for a CCE of 100%. On the same set of experiments it was also observed a correlation between the CCE and the spectroscopic resolution of the devices. In air and under condition of illumination, the detectors showed a good resolution of 4% up to 1 cm of distance from the alpha source which could make these devices a portable tool for alpha spectroscopy measurements. The second part of the experiments was focused on the evaluation of the electronic properties of the devices and in particular the low field mobility and saturation velocity of the charge carriers. According to the mode at which the devices were irradiated, variations in the mobility were observed. In particular, increase in the mobility led to an improvement of the spectroscopic resolution of the devices. The device Dmd#1 with Au electrodes showed similar values of mobility of the charge carriers. The same sample, provided of Al electrodes, showed a change in the total transit time of the charge carriers which could be linked to the fact that Al forms a Schottky barrier on diamond. Prime novelty statement This paper shows the electronic properties of thin single crystal diamond samples obtained by the time of flight technique. Compared to what (at our knowledge) is present in literature, this paper gives a new contribution by showing the electronic properties of the sensors as a function of the modality of irradiation (front or back electrode) and electrode metallization used to fabricate the devices. Acknowledgment This work was performed under the EU-FP7-REGPOT-2010-5 grant, agreement number 256783. The authors would like to acknowledge Dr. Elèni Berdermann for supplying the diamond samples and to express the deepest appreciation to Dr. Tome Antičić, Dr. Milko Jakšić and Dr. Stjepko Fazinić (RBI) for providing useful suggestions and the necessary facilities to perform all the presented measurements. The authors also want sincerely thank Andrija Gajski, Damir Španja, Željko Periša and Natko Skukan (RBI) for the great technical support and valuable help in the experiment maintenance. References
Fig. 9. Comparison of the drift velocity against the electric field for the Dmd#1 fabricated with Au and later Al electrodes. Irradiations were performed placing the device with the biased electrode in front of the alpha source.
[1] L.A. Vermeulen, A.J. Harris, J. Appl. Phys. 49 (1978) 913. [2] R. Vaitkus, T. Inushima, S. Yamazaki, Appl. Phys. Lett. 62 (1993) 2384. [3] M. Marchywka, J.F. Hocnendez, M.W. Geis, D.G. Sacker, D. Moses, T. Goldberg, Appl. Opt. 30 (1991) 5011. [4] J.W. Coleman, J.R. Lierzer, R.D. Petrasso, Rev. Sci. Instrum. 66 (1995) 792. [5] R.B. Spielman, Rev. Sci. Instrum. 66 (1995) 867. [6] S. Han, S.G. Prussin, L.S. Pan, D.R. Kania, S.M. Lane, G. Dennis, R.S. Wagner, K. Harris, Diamond Relat. Mater. 2 (1993) 835. [7] M. Franklin, et al., Nucl. Inst. Methods Phys. Res. A 315 (1992) 39. [8] F. Borchelt, et al., Nucl. Inst. Methods Phys. Res. A 354 (1995) 318. [9] C.D. Clark, P.J. Dean, P.V. Harris, Proc. R. Soc. Lond. Ser. A 277 (1964) 312.
102
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
[10] C. De Angelis, M. Bucciolini, M. Casati, M. Bruzzi, S. Lagomarsino, S. Sciortino, I. Lovik, S. Onori, Radiat. Prot. Dosim. 120 (1–4) (2006) 38. [11] M. Bucciolini, E. Borchi, M. Bruzzi, M. Casati, P. Cirrone, G. Cuttone, C. De Angelis, I. Lovik, S. Onori, L. Raffaele, S. Sciortino, Nucl. Inst. Methods Phys. Res. A 552 (1–2) (2005) 189. [12] C.J.H. Wort, R.S. Balmer, Mater. Today 11 (1–2) (2007) 22. [13] H. Pernegger, et al., J. Appl. Phys. 97 (2005) 073704. [14] M. Nesladek, et al., Diamond Relat. Mater. 17 (2008) 1235. [15] J. Isberg, J. Hammersberg, E. Johansson, T. Wikström, D.J. Twitchen, A.J. Whitehead, S. E. Coe, G.A. Scarsbrook, Science 297 (2002) 1670. [16] S. Gkoumas, A. Lohstroh, P.J. Sellin, Diamond Relat. Mater. 18 (2009) 1338. [17] N. Tranchant, D. Tromson, C. Descamps, A. Isambert, H. Hamrita, P. Bergonzo, M. Nesladek, Diamond Relat. Mater. 17 (2008) 1297. [18] H. Jansen, D. Dobos, T. Eisel, H. Pernegger, V. Eremin, N. Wermes, J. Appl. Phys. 113 (2013) 173706.
[19] M. Pomorski, E. Berdermann, M. Ciobanu1, A. Martemyianov, P. Moritz, M. Rebisz, B. Marczewska, Phys. Status Solidi A 202 (11) (2005) 2199. [20] A. Galbiati, S. Lynn, K. Oliver, F. Schirru, T. Nowak, B. Marczewska, J.A. Dueñas, R. Berjillos, I. Martel, L. Lavergne, IEEE Trans. Nucl. Sci. 56 (4) (2009) 1863. [21] F. Schirru, B.S. Nara Singh, L. Scruton, M.A. Bentley, S.P. Fox, A. Lohstroh, P.J. Sellin, A. Banu, M. McCleskey, B.T. Roeder, E. Simmons, A.A. Alharbi, L. Trache, M. Freer, D. Parker, JINST 7 (2012) P05005. [22] M.A.E. Abdel-Rahman, A. Lohstroh, P.J. Sellin, Phys. Status Solidi A 208 (9) (2011) 2079. [23] D. Chokheli, Development of a multipurpose vacuum chamber for experiments with proton and light ion beams (2014) (in preparation). [24] D.R. Kania, M.I. Landstrass, M.A. Plano, Diamond Relat. Mater. 2 (1993) 1012. [25] C. Weiss, A CVD Diamond Detector for (n, a) Cross Section Measurements, PhD Thesis, University of Technology, Wien, 2014. [26] D. Doneddu, O. J. Guy, D. Twitchen, A. Tajani, M. Schwitters, P. Igic, 2006, 179–182, http://dx.doi.org/10.1109/ICMEL.2006.1650924.
Contents lists available at ScienceDirect
Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond
Thin single crystal diamond detectors for alpha particle detection F. Schirru c,⁎,1, D. Chokheli a,b,1,2, M. Kiš c,1 a b c
JINR, Russia HEPI, Tbilisi State University, Georgia GSI, Darmstadt, Germany
a r t i c l e
i n f o
Article history: Received 20 December 2013 Received in revised form 11 April 2014 Accepted 1 August 2014 Available online 10 August 2014 Keywords: Thin samples Diamond detectors Alpha Spectroscopy CVD diamond
a b s t r a c t Charged particle detectors based on thin single crystal diamond films (3.0 × 3.0 × 0.09 mm3) grown by chemical vapour deposition (CVD) technique are developed at Ruđer Bošković Institute and low field mobility, transit time and saturation velocity, measured by using a 210Po alpha source in vacuum at room temperature, are presented. The comparison of the charge transport properties obtained by time of flight (ToF) technique between detectors based on the same diamond samples, but with different metallization of the electrodes (Au and Al, 100 nm thick), and commercially purchased diamond detectors is performed. The fabricated diamond sensors showed a spectroscopic resolution of up to 1%, mobility up to 1808 ± 16 and 1914 ± 71 cm2/V s for electrons and holes respectively and carrier transit time below 2 ns for an applied electric field of 17.8 × 103 V/cm. Results also show variations of the charge transport properties according to the electrode metallization used and the different modality at which the detectors are irradiated. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The capability of single-crystal CVD (scCVD) diamond detectors to detect different types of radiation from the visible and ultraviolet wavelengths [1–3] up to the X- and γ-rays [4,5] and charged particles as well [6–8] has been widely demonstrated over the past few decades. The interest in using diamond as a radiation detector stems from its unique properties such as the high band gap (5.47 eV) [9] which leads to a low dark current and a high breakdown field (N 1 × 103 V/μm), the resilience to radiation damage due to the strong bonds of the crystal structure [10,11] and the room temperature thermal conductivity that is the highest of any other material [12]. A better understanding of the detection performance of the diamond detector can be achieved by performing a study of its main electronic properties, namely the low field mobility and the transit time of the charge carriers. The former describes the rate at which the carrier drift velocity increases with the applied electric field and mainly varies according to the experimental conditions adopted and the sample quality. In a typical experiment alpha-particles are used to induce currents. Values of low field mobility at room temperature of ~2300 cm2/V s and ~1700 cm2/V s, for holes and electrons respectively [13], have been reported. On the other hand, in experiments with laser-induced currents, mobility for holes was found to be in the range of 2000–2250 cm2/V s and for electrons in the range of 2200–2750 cm2/V s [14]. Other studies ⁎ Corresponding author. E-mail address: [email protected] (F. Schirru). 1 Formerly affiliated with Ruđer Bošković Institute, Zagreb, Croatia. 2 On leave.
http://dx.doi.org/10.1016/j.diamond.2014.08.001 0925-9635/© 2014 Elsevier B.V. All rights reserved.
have shown hole and electron mobility up to 3800 cm2/V s and 4500 cm2/V s respectively [15]. The electronic properties of a diamond sample can be studied by the time of flight (ToF) technique (also known as TCT or transient current technique) in which the duration of the current pulse induced on a read-out electrode by the drift of free charge carriers under the influence of an externally applied field is measured. For the ToF technique we assume that the effective charge carrier lifetime is much longer than the drift time tdr, so the current pulse width equals the time the charge carriers need to traverse the detector. In addition, it has to be recalled that the ToF measurements may depend on the charge density injected especially in the case of high charge densities. Considering a 5.4 MeV alpha particle, the charge carriers are generated in the first 16 μm below the diamond surface and, depending on the direction of the electric field applied, only one type of charge carrier will cross the whole detector thickness and mainly contribute to the signal. In first approximation, by knowing the drift time tdr and the thickness d of the diamond sample, the drift velocity of the carrier vdr can be easily calculated as v dr = d/tdr . On the other hand, by calculating the drift velocity as a function of the electric field applied, it is then possible to extract the low field mobility μ 0 and saturation velocity (maximum drift velocity) of the carrier vsat as [13]: vdr ¼
μ0E μ E 1þ 0 vsat
ð1Þ
Electronic properties of thick (usually 300–500 μm) scCVD diamonds have been studied using the ToF technique and discussed by many authors [16–19]. On this topic, however, little can be found
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
97
in literature about thin samples with thickness below 100 μm. Moreover, there are few papers which present a comparative study between the electronic properties of diamond devices and different types of electrodes [20–22]. For these reasons we report on the electronic properties (low field mobility, transit time and saturation velocity) and spectroscopic resolution of charge particle detectors based on 90 μm thin (scCVD) diamond samples metallized with gold (Au) and aluminium (Al) electrodes. The results, derived from measurements taken at room temperature by the ToF technique, are also compared with those obtained from commercially available scCVD diamond detectors.
2. Material and methods
Fig. 1. Schematic of the detector PCB on which the diamond samples Dmd#1 and Dmd#2 were mounted. The diamond detector signals were fed to the preamplifier by a standard socket SMA connector.
The material considered for this work consisted of: 3. Experimental setup • 2 × scCVD (electronic grade) diamond samples 3.0 × 3.0 × 0.09 mm3 purchased from Element Six and labelled as Dmd#1 and Dmd#2.3 The corresponding detectors were fabricated in house with Au or Al/Au electrodes in a sandwich configuration with squared shape and area of 4 mm2. • 2 × scCVD diamond detectors 4.6 × 4.6 × 0.5 mm3 purchased from CIVIDEC with both surfaces fully covered by 300 nm Al electrodes. These detectors, labelled B10044 and B10045 were used for comparison purposes only. Prior to deposition of Au electrodes, Dmd#1 and Dmd#2 were at first thoroughly cleaned in acetone, hysopropanol and de-ionized water. Deposition of 100 nm Au electrodes was performed by using a shadow mask and a SC7620 ‘Mini’ Sputter operating at a working pressure of 10−3 Torr. Once the measurements on the sample Dmd#1 were completed, Au contacts were removed from the diamond sample using a solution of aqua regia. New Al/Au electrodes with a thickness of 100/30 nm were then deposited onto Dmd#1 by thermal evaporation at the working pressure of 10−5 Torr. First a 100 nm Al layer was deposited followed by 30 nm of Au layer which served to prevent oxidisation of the Al itself. After deposition of electrodes, samples were installed and fixed by vinyl-acetate non-conductive glue onto an in-house built printed circuit board (PCB) based on standard Fr-4 whose schematic is shown in Fig. 1. In order to minimize the influence of the properties of the PCB on the detector performances, dimensions of its copper pads, used to transfer the signals from the detectors to the preamplifier via SMA connector, were minimized. The noise generated by the PCB itself was in fact reduced by decreasing its capacitance and increasing as much as possible the resistance of the housing. The sample electrodes were then connected to a gold wire of 25 μm in diameter (mounted on a K&S 4123 wedge bonder) by silver loaded conductive epoxy. The second end of the wire was next connected to the PCB copper pads by fast drying silver paint. After successful fabrication of the diamond detectors, a set of measurements was then performed with the aim to: • investigate the spectroscopic and charge transport properties of the diamond detectors Dmd#1 and Dmd#2 fabricated with Au electrodes; • compare the properties of spectroscopic resolution of Dmd#1 and Dmd#2 with those obtained from the diamond detectors B10044 and B10045; • compare the charge transport properties of the diamond sensor Dmd#1 fabricated first with Au and then with Al/Au electrodes.
3
Samples provided by Dr. Elèni Berdermann (GSI, Darmstadt, Germany).
Spectroscopic and charge transport properties of the diamond detectors were investigated by irradiating the samples with a 5.4 MeV alpha-particles from 210Po source whose activity was about 0.02 μCi when the experiments started. The devices were mounted, along the alpha source, inside a newly developed vacuum chamber (details can be found in a forthcoming paper [23]). Measurements were recorded at the pressure of ~ 10−2 Torr. Irradiations of the devices were performed in two different modes: by placing the biased or the grounded electrode in front of the alpha source. Henceforth these irradiation modes will be called “forward” or “backward”, respectively. The diamond sensors were positioned carefully in front of the alpha source by means of a translation stage purchased from OWIS®. The induced charge carriers i.e. the generated signal was measured either with a charge sensitive preamplifier (CANBERRA 2004) for spectroscopic characterization or with a current sensitive (DBA IV) preamplifier for a ToF characterization by using respectively RG-58 and RG-316 50 Ω impedance cables whose lengths (40 cm) are as short as possible to minimize the loss of signal. For spectroscopic measurements (Fig. 2 (left)), the signal is further fed into the amplifier (CANBERRA AFT Research Amplifier 2025) and then displayed and acquired by an oscilloscope (Tektronix DPO4054) with a sampling time of 400 ps. For ToF measurements (Fig. 2 (right)), the DBA IV output is directly connected to the oscilloscope for signal visualization and acquisition. In both configurations, a NIM crate HV based supply C.A.E.N. N1470, connected to PC via USB/RS488 link, is used to supply the proper bias voltage to the diamond detectors. Measurements on Dmd#1 and Dmd#2 detectors were performed in both bias polarities and in the voltage range of 8–160 V which corresponds to a maximum applied electric field of 17.8 kV/cm. Due to the geometry of the electrodes, the devices B10044 and B10045 were biased up to 300 V only which corresponds to an electric field of 6 kV/cm. 4. Data processing Data processing of the detector signals was performed according to the type of preamplifier used during the experiments. For each voltage applied, a set of N = 1000 events was recorded from the oscilloscope and then stored in a file. The alpha particle rate measured by the detectors was below 10 s −1 as theoretically expected due to the geometry of the experimental setup. 4.1. Data processing with the charge sensitive preamplifier The signal generated in the detector is directly integrated online by the charge sensitive preamplifier (CANBERRA 2004) and processed by the AFT Research Amplifier (CANBERRA 2025). The latter gives in output a semi-Gaussian pulse shape and a waveform of the resulting signal is stored in the PC via the digital oscilloscope Textronix
98
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Fig. 2. Schematic of the experimental setup using: (left) a charge sensitive amplifier (Camberra 2004 in chain with Camberra 2025) and (right) a current sensitive (DBA IV) preamplifier. In this case the DBA IV is connected to a HAMEG HMP 2030 programmable power supply.
DPO4054. The charge sensitive preamplifier allows better noise condition for the calculation of the spectroscopic resolution of the diamond devices. On the other hand, due to the high RC constant of the preamplifier, the information about the rise time of the signal and, as a consequence, the transit time of the carriers, is completely lost. Reconstruction of the integrated signal is performed offline fitting the oscilloscope waveform by a combination of Gaussian and polynomial functions (Eq. (2)) since it represents a good approximation of the recorded pulse. 2
y ¼ y0 þ A exp
−ðx−μ Þ 2σ 2
! ð2Þ
The integral of each fitted pulse Pi (that is proportional to the total charge Qi collected in the single event) is then plotted to get the total charge distribution (Fig. 3) whose Gaussian fit returns the information about the diamond detector resolution for that particular bias voltage applied. The detector resolution was calculated according to the ratio: R¼
FWHM μ′
ð3Þ
where FWHM = 2.35σ′, σ′ and μ′ are respectively the standard deviation and mean of the charge distribution found by the Gaussian fit. 4.2. Data processing with the current sensitive preamplifier In this case, integration of the detector signals, estimation of the pulse rise time and transport properties of the charge carriers are
Fig. 3. Charge distribution of the diamond detector Dmd#2 referred to an applied electric field of 5.5 kV/cm. For this case a spectroscopic resolution of 1.9% was achieved.
performed offline. Calculation of the diamond detector resolution is carried out as explained in the previous subsection by using the Eqs. (2) and (3). As for the transit time of the charge carriers, this is calculated by taking the width of each pulse after a preliminary pulse baseline adjustment. Then, the generated transit time distribution at that particular voltage applied is fitted with a Gaussian function to estimate the mean pulse width. 5. Results The following subsections describe the measurements performed on the diamond detectors. At first their leakage current, charge collection efficiency and spectroscopic resolution under alpha particle irradiation were investigated. Next, the electronic properties of the charge carriers were estimated by the ToF technique and compared with those obtained for devices with Au and then with Al/Au electrodes. 5.1. Leakage current The leakage current of the diamond detectors was the first parameter investigated. Measurements were performed both in darkness and under illumination conditions by using a metallic enclosure. Conditions of illumination were achieved by a white LED mounted inside the box and connected to a voltage supply of 5 V via 100 Ω current limiting resistor. The luminous intensity of the LED corresponds to 80–110 cd in 20 mA direct current applied with color temperature of 10,000 K. Fig. 4 shows the leakage current measured in darkness and under conditions of illumination of the diamond detectors Dmd#1 and Dmd#2 with Al (blue and cyan data points) and Au (orange and red data points) electrodes in the bias range of ± 160 V. Both detectors present similar behavior of the leakage current and its magnitude does not change significantly regardless the fact that the measurements are taken under dark conditions or not. Such results suggest the possibility to use the detectors in the presence of light. It can be also observed that the leakage current recorded from the metallic box itself with the diamond detector disconnected (green data points on the graph) is very low and basically does not influence the detector performance. Considering the maximum applied electric field to the devices (17.8 kV/cm), the diamond detector with Au electrodes showed the highest value of leakage current. Taking a value of 1.04 × 10− 10 A, it yields a bulk resistance of ~ 1.5 × 1012 Ω. Next, the bulk resistivity of the diamond sample can be estimated as ρ = R(S/l) where R is the resistance of the diamond (calculated for example at + 160 V), S is the electrode area and l is the detector thickness. Assuming S = 4 mm2 and l = 90 μm, it yields a bulk resistivity for the sample of ~(6.7 ± 1.4) × 1012 Ω cm. For the case of the commercial devices B10044 and B10045, their measured leakage current did not exceed 10−11 A for a maximum applied electric field of ± 6 kV/cm.
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Fig. 4. Leakage current measurements of the diamond detectors with Au (orange and red data points) and Al (blue and cyan data points) electrodes. The green points represent the leakage current measured from the metallic enclosure.
5.2. Charge collection efficiency and spectroscopic resolution in vacuum The charge collection efficiency (CCE) of the diamond detectors was the next parameter investigated. Measurements were taken at different bias voltage irradiating the detectors in forward mode. In this case therefore, at the ground electrode, holes are collected applying a positive bias while electrons are collected applying a negative bias. Data were acquired with the experimental setup reported in Fig. 2 (left). Fig. 5 (left) shows the CCE for both types of charge carriers as a function of the applied electric field up to 17.8 × 103 V/cm. Data are related to the diamond sensor Dmd#2 with Au electrodes. The CCE on diamond was estimated by first irradiating a Hamamatsu S-5821 PIN silicon diode which has 100% of CCE and then taking into account the different values of energy required to create an electron–hole pair in the two materials that are usually quoted as 3.6 eV for silicon and 13 eV for diamond [24]. As shown in Fig. 5 (left), the CCE of the device Dmd#2 increases with the applied electric field as more charge carriers are collected at the electrodes. For a bias applied N 50 V the device reaches a 100% of CCE and remains constant in the range of 50–160 V. Also the spectroscopic resolution of the detector (shown in Fig. 5 (right) and calculated according to Eq. (3)) presents analogous behavior and remains constant to a value of 1.2% for applied biases N 50 V.
99
In a subsequent test, keeping the bias at 160 V but irradiating the device at an incidence angle of ±45°, the resolution worsened to 2.0%. This fact could be mainly related by the thicker electrode which the alpha particle has to cross due to the geometrical configuration but also by the different spread of charge across the sample due to the shorter penetration length and the different angle of the applied electric field. This measurement however shows that even for wide incident angles, the detector resolution is still acceptable. The data reported on Fig. 5 (left) and 5 (right) clearly show a correlation between values of CCE and spectroscopic resolution: a decrease of the CCE corresponds also to a decrease of the spectroscopic resolution of the device. The measurements performed on the device Dmd#1 led to the same results and conclusions. For a meaningful comparison between the different diamond detectors, Table 1 reports the CCE and the spectroscopic resolution of the devices calculated for the same magnitude of applied electric field, that is − 6.0 × 103 V/cm. Although for such an electric field also the commercial sensors show 100% of CCE [25], the best spectroscopic resolution is achieved by the detectors made of thinner samples. This could be justified by taking into account that with thicker material it increases the probability of carrier trapping which also causes a bigger variation in the charge collection at the electrodes. Moreover, the commercial devices have thicker electrodes which could also contribute to the worse resolution observed. Anyway, these results may suggest the preference in using thinner samples for alpha spectroscopy purposes, provided that the thickness is large enough to detect the alpha particle. 5.3. Spectroscopic resolution in air Fig. 6 reports the spectroscopic resolution in air of the device Dmd#1 with Al electrodes as a function of the distance from the alpha source. Due to the particular construction of the diamond detector holder, the closest distance at which the devices could be irradiated was 5 mm. For this measurement, the alpha source housing and the electrode geometry of the detectors served as natural collimators of 6 and 3 mm in diameter respectively. In this way only alpha particles which performed almost a straight path toward the detector were collected. The device showed a resolution of 4% up to 1 cm away from the alpha source and then decreased. This measurement was obtained under illumination applying a positive voltage of 160 V on the sensor and irradiating it in backward mode (therefore the signal is mainly due to the electrons). Opposite to devices like those based on silicon, diamond detectors offer the possibility to perform measurements in air and under condition of illumination while keeping a reasonable good value of spectroscopic resolution.
Fig. 5. (Left) Charge collection efficiency CCE and (right) resolution of the diamond sensor Dmd#2 as a function of the electric field applied. The inset on the left graph represents the estimated CCE for holes up to an applied electric field of 2.5 kV/cm. Measurements were taken placing the ground electrode in front of the alpha source.
100
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
Table 1 CCE and spectroscopic properties of the investigated diamond detectors. The parameters were obtained for an applied electric field of −6.0 × 103 kV/cm. Device
Electr.
Thickn. (nm)
CCE (%)
Res. (%)
Dmd#1
Au Al/Au Au Al Al
100 100/30 100 300 300
100 100 100 100 100
1.2 1.8 1.2 1.9 3.8
Dmd#2 B10044 B10045
In vacuum and at the same experimental conditions, the resolution achieved by the detector was 1.3%. However, even in the presence of air, a good spectroscopic resolution can be achieved if the detector is located close to the alpha source in order to minimize the beam straggling. Similar results were also obtained for the device Dmd#2 with Au electrodes. 5.4. Charge transport properties The next part of the experiments was focused on the ToF low field mobility measurement of the charge carriers. Irradiation of the devices was performed at room temperature (300 K) using the experimental setup reported in Fig. 2 (right). These experiments aimed also at comparing, according to the mode of irradiation of the devices, the values of mobility obtained for electrons and holes with the deduced spectroscopic resolution. However, it should be recalled that, in this case, the spectroscopic resolution is calculated offline by integration of the current pulse and therefore the values are expected to be somehow worse than those reported in Table 1. Fig. 7 shows the drift velocity of the charge carriers against the applied electric field for the Dmd#2 device with Au electrodes. Measurements were taken irradiating the device in forward mode and therefore the contribution to the signal was mainly given by electrons or holes according to the applied bias negative or positive. Increasing the applied electric field, the recorded current pulses become narrower as shown in (Fig. 8) and, as a consequence, the drift velocity becomes higher. The drift velocity vdr reported on the graph was calculated by the knowledge of the thickness of the diamond sample divided by the drift time estimated from the current pulse width as described in subsection 4.2. For an electric field of + 6.0 × 103 V/cm, the drift time of electrons and holes calculated for the device Dmd#2 was ~ 1.5 ns while ~ 7 ns was estimated for the commercial devices made up of thicker diamond samples. The variation in the drift time between different measurements was found to be in the order of 5% which has to be attributed more to
Fig. 6. Resolution of the device Dmd#1 with Al electrodes as a function of the distance from the alpha source. Measurements were taken in air applying a positive bias voltage. Irradiations were performed placing the ground electrode in front of the alpha source.
Fig. 7. Drift velocity against the electric field of the detector Dmd#2 irradiated placing the biased electrode in front of the alpha source. The measured values (data points) were fitted using Eq. (1).
the random errors than to the amount of absorbed dose due to the low irradiation rate. According to Eq. (1), the fit performed on the data points led to a value of mobility for electrons of 1628 ± 12 cm2/V s and a saturation velocity of (1.199 ± 0.007) × 107 cm/s. As for the holes the values found were 1539 ± 81 cm 2 /V s and (1.521 ± 0.078) × 107 cm/s. The device Dmd#2 was then irradiated in backward mode at the same experimental conditions. In this configuration for a positive or negative bias applied the signal is mainly given by the electrons or holes respectively. Values of 1658 ± 19 and 1914 ± 71 cm2/V s for low field mobility and values of (1.255 ± 0.012) × 107 cm/s and (1.411 ± 0.042) × 107 cm/s for the saturation velocity were found for electrons and holes. These results show that the mobility of holes is also influenced by the mode of irradiation. Considering the data reported on Table 2, the observed increase in the holes mobility leads also to an improvement of the spectroscopic resolution of the detector. It has to be noted that the fit related to the holes in Fig. 7 is worse than that obtained for electrons. This can be linked to the fact that if the holes have shorter drift times then, the oscilloscope used to extract and process the waveforms might not be fast enough (especially in the high electric field region) to sample correctly the real pulse which leads to an underestimation of the drift velocity of the charge carriers. In these measurements, according to the polarity of the bias applied and the position of the detector with respect to the alpha source, the same type of charge carrier drifts inside the diamond in opposite directions. Since the electrodes of the device are fabricated in similar
Fig. 8. Time evolution of the recorded pulses from the device Dmd#1 (carrier: holes) at different voltage applied. Each pulse represents an average over 100 events.
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102 Table 2 Electronic properties and resolution of the developed devices Dmd#1 and Dmd#2 calculated according to the different mode of irradiation and electrode metallization. The spectroscopic resolution is calculated for an applied electric field of ±17.8 × 103 V/cm. Mobility and saturation velocity are calculated by fitting the drift velocity against the applied electric field. According to the mode of irradiation (fw = forward, bw = backward), main contribution to the generated signal stems from electrons (e) or holes (h). Device
Electr.
Irrad. mode
Carrier
Sat. vel. (107 cm/s)
Mobility (cm2/V s)
Res. (%)
Dmd#1
Au
fw fw fw fw fw fw bw bw
e h e h e h e h
1.162 1.305 1.090 2.113 1.199 1.521 1.255 1.411
1463 1424 1808 1036 1628 1539 1658 1914
4.1 3.6 3.3 3.2 4.1 3.9 4.0 3.5
Al Dmd#2
Au Au
± ± ± ± ± ± ± ±
0.005 0.019 0.004 0.087 0.007 0.078 0.012 0.042
± ± ± ± ± ± ± ±
15 44 17 61 12 82 19 71
way and since the experimental conditions are the same, a change in the mobility of the charge carriers could be linked only to the different way at which they cross the detector bulk. If this is the case, a complete separate study is required to understand in detail the properties of the diamond bulk. 5.5. Role of the electrode metallization on the charge transport properties The device Dmd#1 was next used to study the electronic properties of the charge carriers as a function of the different type of metallization used to create the electrodes. In this case the irradiations were performed in forward mode only. Fig. 9 reports the drift velocity against the electric field applied for the case of the diamond device Dmd#1 fabricated first with Au and then Al electrodes. As reported in Table 2, the mobility for electrons and holes, estimated by fitting the curves reported in Fig. 9, changes considerably according to the type of electrode metallization. With Au electrodes, charge carriers show similar low field mobility that is, 1463 ± 15 cm2 /V s for electrons and 1425 ± 44 cm2/V s for holes. With Al electrodes, however, the calculated values are 1036 ± 61 cm2/V s for holes and 1808 ± 17 cm2/V s for electrons. According to Fig. 9, the total transit time of the charge carriers could be influenced, in the low electric field region, by the Schottky barrier which Al forms on diamond [26]. For higher applied electric field, the highest drift velocity observed is related to the device with Al electrodes which also shows the best spectroscopic resolution. It should be taken into account that the obtained values of mobility are calculated by
101
fitting the data points with Eq. (1) which does not include the effect of the Schottky barrier and therefore they can be considered only approximately. 6. Conclusions This paper presented the spectroscopic and electronic properties of charge particle detectors based on 90 μm thick samples and provided of Au or Al electrodes. The electronic properties of the devices were measured in vacuum and at room temperature with the ToF technique using a 210Po alpha source. The results were compared with those obtained from two commercially purchased diamond detectors 500 μm thick. The first part of the experiments was focused on the evaluation of the spectroscopic resolution and CCE of the developed detectors fabricated with Au electrodes. It was found that, compared to the commercial detectors, the devices based on thinner diamond samples showed better spectroscopic performance with a resolution achieved for alpha particles up to 1.2% for a CCE of 100%. On the same set of experiments it was also observed a correlation between the CCE and the spectroscopic resolution of the devices. In air and under condition of illumination, the detectors showed a good resolution of 4% up to 1 cm of distance from the alpha source which could make these devices a portable tool for alpha spectroscopy measurements. The second part of the experiments was focused on the evaluation of the electronic properties of the devices and in particular the low field mobility and saturation velocity of the charge carriers. According to the mode at which the devices were irradiated, variations in the mobility were observed. In particular, increase in the mobility led to an improvement of the spectroscopic resolution of the devices. The device Dmd#1 with Au electrodes showed similar values of mobility of the charge carriers. The same sample, provided of Al electrodes, showed a change in the total transit time of the charge carriers which could be linked to the fact that Al forms a Schottky barrier on diamond. Prime novelty statement This paper shows the electronic properties of thin single crystal diamond samples obtained by the time of flight technique. Compared to what (at our knowledge) is present in literature, this paper gives a new contribution by showing the electronic properties of the sensors as a function of the modality of irradiation (front or back electrode) and electrode metallization used to fabricate the devices. Acknowledgment This work was performed under the EU-FP7-REGPOT-2010-5 grant, agreement number 256783. The authors would like to acknowledge Dr. Elèni Berdermann for supplying the diamond samples and to express the deepest appreciation to Dr. Tome Antičić, Dr. Milko Jakšić and Dr. Stjepko Fazinić (RBI) for providing useful suggestions and the necessary facilities to perform all the presented measurements. The authors also want sincerely thank Andrija Gajski, Damir Španja, Željko Periša and Natko Skukan (RBI) for the great technical support and valuable help in the experiment maintenance. References
Fig. 9. Comparison of the drift velocity against the electric field for the Dmd#1 fabricated with Au and later Al electrodes. Irradiations were performed placing the device with the biased electrode in front of the alpha source.
[1] L.A. Vermeulen, A.J. Harris, J. Appl. Phys. 49 (1978) 913. [2] R. Vaitkus, T. Inushima, S. Yamazaki, Appl. Phys. Lett. 62 (1993) 2384. [3] M. Marchywka, J.F. Hocnendez, M.W. Geis, D.G. Sacker, D. Moses, T. Goldberg, Appl. Opt. 30 (1991) 5011. [4] J.W. Coleman, J.R. Lierzer, R.D. Petrasso, Rev. Sci. Instrum. 66 (1995) 792. [5] R.B. Spielman, Rev. Sci. Instrum. 66 (1995) 867. [6] S. Han, S.G. Prussin, L.S. Pan, D.R. Kania, S.M. Lane, G. Dennis, R.S. Wagner, K. Harris, Diamond Relat. Mater. 2 (1993) 835. [7] M. Franklin, et al., Nucl. Inst. Methods Phys. Res. A 315 (1992) 39. [8] F. Borchelt, et al., Nucl. Inst. Methods Phys. Res. A 354 (1995) 318. [9] C.D. Clark, P.J. Dean, P.V. Harris, Proc. R. Soc. Lond. Ser. A 277 (1964) 312.
102
F. Schirru et al. / Diamond & Related Materials 49 (2014) 96–102
[10] C. De Angelis, M. Bucciolini, M. Casati, M. Bruzzi, S. Lagomarsino, S. Sciortino, I. Lovik, S. Onori, Radiat. Prot. Dosim. 120 (1–4) (2006) 38. [11] M. Bucciolini, E. Borchi, M. Bruzzi, M. Casati, P. Cirrone, G. Cuttone, C. De Angelis, I. Lovik, S. Onori, L. Raffaele, S. Sciortino, Nucl. Inst. Methods Phys. Res. A 552 (1–2) (2005) 189. [12] C.J.H. Wort, R.S. Balmer, Mater. Today 11 (1–2) (2007) 22. [13] H. Pernegger, et al., J. Appl. Phys. 97 (2005) 073704. [14] M. Nesladek, et al., Diamond Relat. Mater. 17 (2008) 1235. [15] J. Isberg, J. Hammersberg, E. Johansson, T. Wikström, D.J. Twitchen, A.J. Whitehead, S. E. Coe, G.A. Scarsbrook, Science 297 (2002) 1670. [16] S. Gkoumas, A. Lohstroh, P.J. Sellin, Diamond Relat. Mater. 18 (2009) 1338. [17] N. Tranchant, D. Tromson, C. Descamps, A. Isambert, H. Hamrita, P. Bergonzo, M. Nesladek, Diamond Relat. Mater. 17 (2008) 1297. [18] H. Jansen, D. Dobos, T. Eisel, H. Pernegger, V. Eremin, N. Wermes, J. Appl. Phys. 113 (2013) 173706.
[19] M. Pomorski, E. Berdermann, M. Ciobanu1, A. Martemyianov, P. Moritz, M. Rebisz, B. Marczewska, Phys. Status Solidi A 202 (11) (2005) 2199. [20] A. Galbiati, S. Lynn, K. Oliver, F. Schirru, T. Nowak, B. Marczewska, J.A. Dueñas, R. Berjillos, I. Martel, L. Lavergne, IEEE Trans. Nucl. Sci. 56 (4) (2009) 1863. [21] F. Schirru, B.S. Nara Singh, L. Scruton, M.A. Bentley, S.P. Fox, A. Lohstroh, P.J. Sellin, A. Banu, M. McCleskey, B.T. Roeder, E. Simmons, A.A. Alharbi, L. Trache, M. Freer, D. Parker, JINST 7 (2012) P05005. [22] M.A.E. Abdel-Rahman, A. Lohstroh, P.J. Sellin, Phys. Status Solidi A 208 (9) (2011) 2079. [23] D. Chokheli, Development of a multipurpose vacuum chamber for experiments with proton and light ion beams (2014) (in preparation). [24] D.R. Kania, M.I. Landstrass, M.A. Plano, Diamond Relat. Mater. 2 (1993) 1012. [25] C. Weiss, A CVD Diamond Detector for (n, a) Cross Section Measurements, PhD Thesis, University of Technology, Wien, 2014. [26] D. Doneddu, O. J. Guy, D. Twitchen, A. Tajani, M. Schwitters, P. Igic, 2006, 179–182, http://dx.doi.org/10.1109/ICMEL.2006.1650924.