- Email: [email protected]
Ion Beam Induced Charge characterization of epitaxial single crystal CVD diamond
Diamond & Related Materials 16 (2007) 940 – 943 www.elsevier.com/locate/diamond
Ion Beam Induced Charge characterization of epitaxial single crystal CVD diamond C. Manfredotti a,⁎, M. Jaksic b , S. Medunic b , A. Lo Giudice a , Y. Garino a , E. Colombo a , Marco Marinelli c , E. Milani c , G. Verona-Rinati c a
Experimental Physics Department and Center of Excellence, NIS, University of Torino, Italy, Via Giuria 1, 101125 Torino, Italy b Ruder Boskovic Institute, Zagreb, Croatia c INFN — Dip. Ingegneria Meccanica, Università di Roma “Tor Vergata”, Italy Available online 29 December 2006
Abstract IBIC (Ion Beam Induced Charge) technique has been used in order to characterize single crystal epitaxial CVD diamond film with respect to homogeneity and stability of the response (in terms of charge collection efficiency, cce) as a function both of counting rate and of the number of counts per unit surface area. The maximum shift of cce peak, under a 1.2 MeV proton microbeam, is 1.5% for counting rates from 43 to 4330 Hz, while the homogeneity, evaluated as the standard deviation with respect to the average value of cce over strip-like regions 60–100 μm wide and 800–1200 μm long, is 0.5%. Counting rates per unit surface area were between 30 and about 15,000 Hz/mm2. A total number of counts per unit area up to 9 106 counts/mm2 was reached without noticing any polarization effect due to trapped charge. Moreover, the functionality of a new kind of bulk electrode, realized by a boron doped buffer layer laterally contacted with Ag paste, has been checked by measuring cce at different proton ranges. © 2006 Elsevier B.V. All rights reserved. Keywords: Single crystal CVD diamond; IBIC; Homogeneity; Energy resolution
1. Introduction In the past, selected natural diamond nuclear detectors [1] were found to display very nice performances both in terms of speed and energy resolution, together with well known properties of resistance to radiation damage. This was not the case of artificial diamond, which, as deposited by CVD, was polycrystalline and, as obtained by HPHT, was strongly doped. However, because of the relatively weak radiation resistance of Si, CVD diamond started to be investigated for nuclear applications. In fact, from 1995, CVD diamond was tried to be used as a track detector in large collider experiments such as LHC [2] and this challenging application contributed a lot to improve its detector quality [3]. Nevertheless, due to its polycrystalline nature, the performances of CVD diamond in terms of homogeneity of response and consequently of energy resolution were very poor [4].
⁎ Corresponding author. Tel.: +39 0116707306; fax: +39 0116691104. E-mail address: [email protected] (C. Manfredotti). 0925-9635/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2006.11.062
The situation changed suddenly with the advent of homoepitaxial CVD diamond, which displayed superior properties in terms of carrier mobilities and trapping times [5]. However, even if it was documented that energy resolution was now very good [6], no report appeared concerning : 1) the homogeneity of the charge collection efficiency (cce) over relatively large areas of the detector; 2) the stability of the detector as a function of counting rate and of the total number of counts; 3) the presence of polarization effects and 4) the need of priming or pumping [2], largely used for the initial stabilization and improvement of diamond response. Recently [7], we reported about preliminary measurements on energy resolution and of homogeneity over relatively large areas by using alpha and proton microbeams, with the indication of no need of priming, of the relative absence of polarization effects and of a good stability up to counting rates of about 700 Hz. In the meantime, the material quality has been continuously improved and this work is a progress report mainly dedicated to the homogeneity of cce up to surface areas more than 1 mm2 and to the stability of the full energy peak up to counting rates as high as
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
Fig. 1. Lay out of the detector, a 17 μm thick CVD diamond film equipped with a top circular Al electrode, 150 nm thick and 2 mm in diameter. The film is deposited over a p+ type CVD diamond film, doped with boron, which is electrically connected by four lateral Ag paste contacts and constitutes the back electrode. Both films are epitaxial, but they were deposited in different chambers to avoid contamination. The contacts are annealed at high temperature in order to reach the necessary mechanical strength. The HPHT substrate is 300 μm thick with a surface area of 3 × 3 mm2.
4500 Hz and total number of counts of more than 3 105 over areas from 100 × 100 μm2 to 1.2 × 1.2 mm2. This work has been carried out with a microbeam of protons of energies between 1.2 and 1.6 MeV, also to test the functionality of the back electrode, realized with a new technology. 2. Experimental The diamond sample investigated in this work was epitaxially grown by MW CVD on a 300 μm low cost HPHT diamond substrate, which had a nominal N concentration of 1016 cm− 3 and which was electrically insulating. Substrate temperature was 800 °C, methane concentration 1% in hydrogen gas with a nominal purity of 99.9999%, total gas flow rate 200 sccm, growth velocity maintained at about 1 μm/h. The sample was carefully characterized by X-ray diffraction and Scanning Electron Microscopy, confirming the good single crystal quality of the grown samples. By CL measurements it was possible to conclude that at least in the first few micrometers below the surface the epitaxial layer was both boron and nitrogen free. The thickness of this epitaxial layer of was evaluated to be 17 μm by a profilometer. Circular Al electrodes, of diameter of 2 mm and of thickness 150 nm, were deposited on the top surface (see Fig. 1) of this layer, which was grown over a buffer layer of crystalline diamond strongly doped with boron, used as a back electrode. The electrical connection was created by four contacts of Ag paste, suitably annealed at high temperature, placed at the periphery of the sample. This kind of contacts avoids the problem of lapping away the substrate, a problem which becomes very difficult when the active layer is very thin, as in the present case. The behaviour of these contacts was checked to be quite good, apart from some time needed in order to obtain a new stable bias voltage across the detector. Some results concerning the characterization of these contacts will be presented in a future paper. At + 85 V bias voltage the current intensity was below 10 pA, i.e. the global resistivity was well above 1011 Ω cm. The present measurements were carried out with a proton microbeam with a spot diameter of the order of few μm at Ruder
941
Fig. 2. Multichannel spectra obtained from 1.2 MeV protons at counting frequencies from 42 to 4332 Hz (or counts per second, cps). In terms of counting rates per unit surface area the values are from 30 to 15400 cps/mm2, while in terms of total counts the value are from 9500 to 9.3 106 counts/mm2.
Boskovic Institute of Zagreb. The selected energies were 1.2, 1.4 and 1.6 MeV. The beam intensity was generally kept low in order to get a counting rate of no more than 50–100 Hz, but in some cases this rate was increased up to more than 4500 Hz. Selected scanning zone areas varied from 100 × 100 μm2 up to 1.2 × 1.2 mm2. Electronic shaping time was 0.25 μs in all the cases. 3. Results and discussion One of the main concerns about new nuclear detectors is of course the stability, i.e. the constancy of the response in time even at very long counting times and at relatively high counting rates. The stability may be affected either by transient effects, the meaning of which is that the response of the detector is in some way influenced by its previous history and the detector itself needs some kind of initial treatment in order to obtain the same starting conditions, or by polarization effects, which are due to the growth of an opposite internal electrical field in regions characterized by strong trapping (see for instance Refs.
Fig. 3. Position of the peak energy of the multichannel spectrum produced by 1.2 MeV protons as a function of counting frequency (in Hz).
942
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
Fig. 4. Average fluctuations with respect to the average value of the total collected charge (in arbitrary units) over pixel strips about 10 pixels wide and 120 pixels long, considered as ROI in the maps of cce. The strips are both in X and in Y direction or profile, as indicated over the X-axis and are taken at the center and towards both ends of the map. The group of data refer to maps which are either 800 × 800 μm2 wide (4520 Hz) or 1.2 × 1.2 mm2 wide (1528 and 4332 Hz). All the maps are carried out over squared 128 × 128 pixels regions and each pixel width may vary from about 1 μm to 10 μm according to the spot size of microbeam.
[8–10]). In the present case the stability is very high, as shown in Fig. 2, where 6 spectra obtained in different times with 1.2 MeV energy protons, with no initial priming, are reported for different counting frequencies from 43 to 4332 Hz. What is observed is a very small displacement of the order of 1.5% towards larger frequencies, but within the experimental errors. Measurements were carried out generally over more than 5 min, but this kind of analysis was done by taking only the first 1000 counts, in order to not compare data with very different total amounts of counts. In any case the stability was checked at the highest counting rates up to about 500,000 counts. The energy resolution was about 93 keV, in a good agreement with data coming from alpha particles measurements [11]. What has to be observed is that the peak is almost symmetric and that the tail at the left due to trapping is very low. A better way to present stability is given in Fig. 3, where the peak position together with FWTM (calculated in this case as the full width at one-tenth of the maximum) as a function of counting frequency is displayed and where the stability of both the peak and of its total width can be appreciated in detail. Together with stability, a problem can be presented also by the homogeneity of the response. A way to appreciate homogeneity could be represented by profiles of cce, but a profile along a row or a column could be worsened by pixels in which, for statistical reasons, the number of counts is lower. In order to improve statistics, the profiles are derived not from ROIs of only one line (either a row or a column), but from rectangular ROIs. As an example, Fig. 4 shows the mean values of fluctuations (expressed as standard deviations in this case) of the total collected charge (in arbitrary units) with respect to the average value, along ROIs about 60 μm wide and 800 or 1200 μm long. SDs are no more than 0.5% at counting frequencies of 1528 and 4520 Hz. Even the differences between
Fig. 5. An example of a cce map with indications of strips (rectangles) over which average fluctuations are calculated as indicated in Fig. 4. The map is 800 × 800 μm wide.
the averages in different regions (see the values reported at 4332 Hz) are very similar. These data confirm the very good homogeneity of our sample and the smallness of inhomogeneous broadening of energy resolution. In fact, values of the order of 1.2% are reached with alpha particles at energies above 5 MeV [7]. In order to clarify better the way the data were obtained, in Fig. 5 we report a map of the total number of counts, weighted each by the related charge signal, in a region 1.2 × 1.2 mm wide. Since each pixel should have the same probability of being hit, at a high number of counts the total accumulated signal should be proportional to cce. In order to obtain cce profiles and to present data as those reported in Fig. 3 with a sufficient statistics, the data are summed in columns for profiles of type X) or in rows (for profiles of type Y). The last point is concerning this new kind of electrode, realized by a doped buffer layer laterally connected. At 1.2 MeV, the proton range inside diamond should be of about 12 μm, but at 1.6 MeV it can reach 16 μm, which is almost the thickness of the
Fig. 6. Multichannel spectra taken over the same region at different proton energies. The almost linear behaviour of peak position as a function of proton energy is presented in the insert. The different spectra are not normalized with respect to their total number of counts.
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
active region. The corresponding multichannel spectra are presented in Fig. 6, together with the behaviour of peak position as a function of beam energy. Linearity is relatively good, and therefore the electrode does not affect charge collection by trapped charge or other effects. The electrode works properly — i.e. the accumulated charge is suddenly dispersed and the polarization is negligible. Moreover, the electrode seems to be at the forecast distance from top electrode. 4. Conclusions A single crystal epitaxial CVD diamond detector has been tested in terms of stability and homogeneity of response to a microbeam of 2 MeV protons over regions up to more than 1 mm2 and for counting frequencies larger than 4500 Hz, with more than encouraging results. Moreover, a new kind of electrode realized with a doped buffer layer has been successfully tested. The stability is always better than 1.5% and the homogeneity, as evaluated along strip-shaped ROIs over lengths of the order of the diameter of the sample, is better than 0.5% in any condition. References [1] C. Canali, E. Gatti, S.F. Koslov, P.F. Manfredi, C. Manfredotti, F. Nava, A. Quirini, Nuclear Instruments and Methods A 160 (1979) 73.
943
[2] For a complete overview, see for instance D. Meier, CVD diamond sensors for particle detection and tracking, Thesis, Duesseldorf 1999. [3] W. Adam et al., EPS 2003, RD42 contribution. [4] C. Manfredotti, F. Fizzotti, Lo Giudice, C. Paolini, E. Vittone, R. Lu, Nuclear Instruments & Methods in Physics Research. Section B, Beam Interactions with Materials and Atoms 187 (2002) 566. [5] J. Isberg, et al., Science 297 (2002) 1670. [6] C. Manfredotti, Diamond and Related Materials 14 (2005) 531. [7] A. Balducci, M. Jaksic, A. Lo Giudice, C. Manfredotti, Marco Marinelli, S. Medunic, G. Pucella, G. Verona-Rinati, Diamond and Related Materials 14 (2005) 1988. [8] C. Manfredotti, F. Fizzotti, E. Vittone, P. Polesello, F. Wang, Physica Status Solidi. A, Applied Research 154 (1996) 327. [9] C. Manfredotti, F. Fizzotti, E. Vittone, M. Boero, P. Polesello, S. Galassini, M. Jaksic, S. Fazinic, I. Bogdanovic, Nuclear Instruments & Methods in Physics Research. Section B, Beam Interactions with Materials and Atoms 100 (1995) 133. [10] C. Manfredotti, F. Fizzotti, K. Mirri, P. Polesello, E. Vittone, M. Jaksic, T. Tadic, I. Bodganovic, V. Valkovic, T. Pochet, Diamond and Related Materials 6 (1997) 320. [11] Marco Marinelli, private communication.
Ion Beam Induced Charge characterization of epitaxial single crystal CVD diamond C. Manfredotti a,⁎, M. Jaksic b , S. Medunic b , A. Lo Giudice a , Y. Garino a , E. Colombo a , Marco Marinelli c , E. Milani c , G. Verona-Rinati c a
Experimental Physics Department and Center of Excellence, NIS, University of Torino, Italy, Via Giuria 1, 101125 Torino, Italy b Ruder Boskovic Institute, Zagreb, Croatia c INFN — Dip. Ingegneria Meccanica, Università di Roma “Tor Vergata”, Italy Available online 29 December 2006
Abstract IBIC (Ion Beam Induced Charge) technique has been used in order to characterize single crystal epitaxial CVD diamond film with respect to homogeneity and stability of the response (in terms of charge collection efficiency, cce) as a function both of counting rate and of the number of counts per unit surface area. The maximum shift of cce peak, under a 1.2 MeV proton microbeam, is 1.5% for counting rates from 43 to 4330 Hz, while the homogeneity, evaluated as the standard deviation with respect to the average value of cce over strip-like regions 60–100 μm wide and 800–1200 μm long, is 0.5%. Counting rates per unit surface area were between 30 and about 15,000 Hz/mm2. A total number of counts per unit area up to 9 106 counts/mm2 was reached without noticing any polarization effect due to trapped charge. Moreover, the functionality of a new kind of bulk electrode, realized by a boron doped buffer layer laterally contacted with Ag paste, has been checked by measuring cce at different proton ranges. © 2006 Elsevier B.V. All rights reserved. Keywords: Single crystal CVD diamond; IBIC; Homogeneity; Energy resolution
1. Introduction In the past, selected natural diamond nuclear detectors [1] were found to display very nice performances both in terms of speed and energy resolution, together with well known properties of resistance to radiation damage. This was not the case of artificial diamond, which, as deposited by CVD, was polycrystalline and, as obtained by HPHT, was strongly doped. However, because of the relatively weak radiation resistance of Si, CVD diamond started to be investigated for nuclear applications. In fact, from 1995, CVD diamond was tried to be used as a track detector in large collider experiments such as LHC [2] and this challenging application contributed a lot to improve its detector quality [3]. Nevertheless, due to its polycrystalline nature, the performances of CVD diamond in terms of homogeneity of response and consequently of energy resolution were very poor [4].
⁎ Corresponding author. Tel.: +39 0116707306; fax: +39 0116691104. E-mail address: [email protected] (C. Manfredotti). 0925-9635/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2006.11.062
The situation changed suddenly with the advent of homoepitaxial CVD diamond, which displayed superior properties in terms of carrier mobilities and trapping times [5]. However, even if it was documented that energy resolution was now very good [6], no report appeared concerning : 1) the homogeneity of the charge collection efficiency (cce) over relatively large areas of the detector; 2) the stability of the detector as a function of counting rate and of the total number of counts; 3) the presence of polarization effects and 4) the need of priming or pumping [2], largely used for the initial stabilization and improvement of diamond response. Recently [7], we reported about preliminary measurements on energy resolution and of homogeneity over relatively large areas by using alpha and proton microbeams, with the indication of no need of priming, of the relative absence of polarization effects and of a good stability up to counting rates of about 700 Hz. In the meantime, the material quality has been continuously improved and this work is a progress report mainly dedicated to the homogeneity of cce up to surface areas more than 1 mm2 and to the stability of the full energy peak up to counting rates as high as
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
Fig. 1. Lay out of the detector, a 17 μm thick CVD diamond film equipped with a top circular Al electrode, 150 nm thick and 2 mm in diameter. The film is deposited over a p+ type CVD diamond film, doped with boron, which is electrically connected by four lateral Ag paste contacts and constitutes the back electrode. Both films are epitaxial, but they were deposited in different chambers to avoid contamination. The contacts are annealed at high temperature in order to reach the necessary mechanical strength. The HPHT substrate is 300 μm thick with a surface area of 3 × 3 mm2.
4500 Hz and total number of counts of more than 3 105 over areas from 100 × 100 μm2 to 1.2 × 1.2 mm2. This work has been carried out with a microbeam of protons of energies between 1.2 and 1.6 MeV, also to test the functionality of the back electrode, realized with a new technology. 2. Experimental The diamond sample investigated in this work was epitaxially grown by MW CVD on a 300 μm low cost HPHT diamond substrate, which had a nominal N concentration of 1016 cm− 3 and which was electrically insulating. Substrate temperature was 800 °C, methane concentration 1% in hydrogen gas with a nominal purity of 99.9999%, total gas flow rate 200 sccm, growth velocity maintained at about 1 μm/h. The sample was carefully characterized by X-ray diffraction and Scanning Electron Microscopy, confirming the good single crystal quality of the grown samples. By CL measurements it was possible to conclude that at least in the first few micrometers below the surface the epitaxial layer was both boron and nitrogen free. The thickness of this epitaxial layer of was evaluated to be 17 μm by a profilometer. Circular Al electrodes, of diameter of 2 mm and of thickness 150 nm, were deposited on the top surface (see Fig. 1) of this layer, which was grown over a buffer layer of crystalline diamond strongly doped with boron, used as a back electrode. The electrical connection was created by four contacts of Ag paste, suitably annealed at high temperature, placed at the periphery of the sample. This kind of contacts avoids the problem of lapping away the substrate, a problem which becomes very difficult when the active layer is very thin, as in the present case. The behaviour of these contacts was checked to be quite good, apart from some time needed in order to obtain a new stable bias voltage across the detector. Some results concerning the characterization of these contacts will be presented in a future paper. At + 85 V bias voltage the current intensity was below 10 pA, i.e. the global resistivity was well above 1011 Ω cm. The present measurements were carried out with a proton microbeam with a spot diameter of the order of few μm at Ruder
941
Fig. 2. Multichannel spectra obtained from 1.2 MeV protons at counting frequencies from 42 to 4332 Hz (or counts per second, cps). In terms of counting rates per unit surface area the values are from 30 to 15400 cps/mm2, while in terms of total counts the value are from 9500 to 9.3 106 counts/mm2.
Boskovic Institute of Zagreb. The selected energies were 1.2, 1.4 and 1.6 MeV. The beam intensity was generally kept low in order to get a counting rate of no more than 50–100 Hz, but in some cases this rate was increased up to more than 4500 Hz. Selected scanning zone areas varied from 100 × 100 μm2 up to 1.2 × 1.2 mm2. Electronic shaping time was 0.25 μs in all the cases. 3. Results and discussion One of the main concerns about new nuclear detectors is of course the stability, i.e. the constancy of the response in time even at very long counting times and at relatively high counting rates. The stability may be affected either by transient effects, the meaning of which is that the response of the detector is in some way influenced by its previous history and the detector itself needs some kind of initial treatment in order to obtain the same starting conditions, or by polarization effects, which are due to the growth of an opposite internal electrical field in regions characterized by strong trapping (see for instance Refs.
Fig. 3. Position of the peak energy of the multichannel spectrum produced by 1.2 MeV protons as a function of counting frequency (in Hz).
942
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
Fig. 4. Average fluctuations with respect to the average value of the total collected charge (in arbitrary units) over pixel strips about 10 pixels wide and 120 pixels long, considered as ROI in the maps of cce. The strips are both in X and in Y direction or profile, as indicated over the X-axis and are taken at the center and towards both ends of the map. The group of data refer to maps which are either 800 × 800 μm2 wide (4520 Hz) or 1.2 × 1.2 mm2 wide (1528 and 4332 Hz). All the maps are carried out over squared 128 × 128 pixels regions and each pixel width may vary from about 1 μm to 10 μm according to the spot size of microbeam.
[8–10]). In the present case the stability is very high, as shown in Fig. 2, where 6 spectra obtained in different times with 1.2 MeV energy protons, with no initial priming, are reported for different counting frequencies from 43 to 4332 Hz. What is observed is a very small displacement of the order of 1.5% towards larger frequencies, but within the experimental errors. Measurements were carried out generally over more than 5 min, but this kind of analysis was done by taking only the first 1000 counts, in order to not compare data with very different total amounts of counts. In any case the stability was checked at the highest counting rates up to about 500,000 counts. The energy resolution was about 93 keV, in a good agreement with data coming from alpha particles measurements [11]. What has to be observed is that the peak is almost symmetric and that the tail at the left due to trapping is very low. A better way to present stability is given in Fig. 3, where the peak position together with FWTM (calculated in this case as the full width at one-tenth of the maximum) as a function of counting frequency is displayed and where the stability of both the peak and of its total width can be appreciated in detail. Together with stability, a problem can be presented also by the homogeneity of the response. A way to appreciate homogeneity could be represented by profiles of cce, but a profile along a row or a column could be worsened by pixels in which, for statistical reasons, the number of counts is lower. In order to improve statistics, the profiles are derived not from ROIs of only one line (either a row or a column), but from rectangular ROIs. As an example, Fig. 4 shows the mean values of fluctuations (expressed as standard deviations in this case) of the total collected charge (in arbitrary units) with respect to the average value, along ROIs about 60 μm wide and 800 or 1200 μm long. SDs are no more than 0.5% at counting frequencies of 1528 and 4520 Hz. Even the differences between
Fig. 5. An example of a cce map with indications of strips (rectangles) over which average fluctuations are calculated as indicated in Fig. 4. The map is 800 × 800 μm wide.
the averages in different regions (see the values reported at 4332 Hz) are very similar. These data confirm the very good homogeneity of our sample and the smallness of inhomogeneous broadening of energy resolution. In fact, values of the order of 1.2% are reached with alpha particles at energies above 5 MeV [7]. In order to clarify better the way the data were obtained, in Fig. 5 we report a map of the total number of counts, weighted each by the related charge signal, in a region 1.2 × 1.2 mm wide. Since each pixel should have the same probability of being hit, at a high number of counts the total accumulated signal should be proportional to cce. In order to obtain cce profiles and to present data as those reported in Fig. 3 with a sufficient statistics, the data are summed in columns for profiles of type X) or in rows (for profiles of type Y). The last point is concerning this new kind of electrode, realized by a doped buffer layer laterally connected. At 1.2 MeV, the proton range inside diamond should be of about 12 μm, but at 1.6 MeV it can reach 16 μm, which is almost the thickness of the
Fig. 6. Multichannel spectra taken over the same region at different proton energies. The almost linear behaviour of peak position as a function of proton energy is presented in the insert. The different spectra are not normalized with respect to their total number of counts.
C. Manfredotti et al. / Diamond & Related Materials 16 (2007) 940–943
active region. The corresponding multichannel spectra are presented in Fig. 6, together with the behaviour of peak position as a function of beam energy. Linearity is relatively good, and therefore the electrode does not affect charge collection by trapped charge or other effects. The electrode works properly — i.e. the accumulated charge is suddenly dispersed and the polarization is negligible. Moreover, the electrode seems to be at the forecast distance from top electrode. 4. Conclusions A single crystal epitaxial CVD diamond detector has been tested in terms of stability and homogeneity of response to a microbeam of 2 MeV protons over regions up to more than 1 mm2 and for counting frequencies larger than 4500 Hz, with more than encouraging results. Moreover, a new kind of electrode realized with a doped buffer layer has been successfully tested. The stability is always better than 1.5% and the homogeneity, as evaluated along strip-shaped ROIs over lengths of the order of the diameter of the sample, is better than 0.5% in any condition. References [1] C. Canali, E. Gatti, S.F. Koslov, P.F. Manfredi, C. Manfredotti, F. Nava, A. Quirini, Nuclear Instruments and Methods A 160 (1979) 73.
943
[2] For a complete overview, see for instance D. Meier, CVD diamond sensors for particle detection and tracking, Thesis, Duesseldorf 1999. [3] W. Adam et al., EPS 2003, RD42 contribution. [4] C. Manfredotti, F. Fizzotti, Lo Giudice, C. Paolini, E. Vittone, R. Lu, Nuclear Instruments & Methods in Physics Research. Section B, Beam Interactions with Materials and Atoms 187 (2002) 566. [5] J. Isberg, et al., Science 297 (2002) 1670. [6] C. Manfredotti, Diamond and Related Materials 14 (2005) 531. [7] A. Balducci, M. Jaksic, A. Lo Giudice, C. Manfredotti, Marco Marinelli, S. Medunic, G. Pucella, G. Verona-Rinati, Diamond and Related Materials 14 (2005) 1988. [8] C. Manfredotti, F. Fizzotti, E. Vittone, P. Polesello, F. Wang, Physica Status Solidi. A, Applied Research 154 (1996) 327. [9] C. Manfredotti, F. Fizzotti, E. Vittone, M. Boero, P. Polesello, S. Galassini, M. Jaksic, S. Fazinic, I. Bogdanovic, Nuclear Instruments & Methods in Physics Research. Section B, Beam Interactions with Materials and Atoms 100 (1995) 133. [10] C. Manfredotti, F. Fizzotti, K. Mirri, P. Polesello, E. Vittone, M. Jaksic, T. Tadic, I. Bodganovic, V. Valkovic, T. Pochet, Diamond and Related Materials 6 (1997) 320. [11] Marco Marinelli, private communication.