- Email: [email protected]
Technology development of high purity germanium crystals for radiation detectors
Journal Pre-proofs Technology Development of High Purity Germanium Crystals for Radiation Detectors N. Abrosimov, M. Czupalla, N. Dropka, J. Fischer, A. Gybin, K. Irmscher, J. Janicskó-Csáthy, U. Juda, S. Kayser, W. Miller, M. Pietsch, F.M. Kießling PII: DOI: Reference:
S0022-0248(19)30611-6 https://doi.org/10.1016/j.jcrysgro.2019.125396 CRYS 125396
To appear in:
Journal of Crystal Growth
Received Date: Revised Date: Accepted Date:
24 January 2019 24 October 2019 29 November 2019
Please cite this article as: N. Abrosimov, M. Czupalla, N. Dropka, J. Fischer, A. Gybin, K. Irmscher, J. JanicskóCsáthy, U. Juda, S. Kayser, W. Miller, M. Pietsch, F.M. Kießling, Technology Development of High Purity Germanium Crystals for Radiation Detectors, Journal of Crystal Growth (2019), doi: https://doi.org/10.1016/ j.jcrysgro.2019.125396
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier B.V.
Technology Development of High Purity Germanium Crystals for Radiation Detectors N. Abrosimov, M. Czupalla, N. Dropka, J. Fischer, A. Gybin, K. Irmscher, J. Janicsk´o-Cs´athy, U. Juda, S. Kayser, W. Miller, M. Pietsch, F.M. Kießling Leibniz-Institut f¨ur Kristallz¨uchtung (IKZ), Max-Born-Str. 2, 12489 Berlin, Germany
Abstract High Purity Germanium (HPGe) crystals are indispensable for radiation detection and fundamental research. Among other applications they will be used in the Large Enriched Germanium Experiment for Neutrinoless double beta Decay (LEGEND) experiment for the search of neutrinoless double beta decay of 76 Ge. Ongoing research activities are done in collaboration with members of the existing GERDA experiment. In this context the Leibniz-Institut f¨ur Kristallz¨uchtung (IKZ) is engaged in the development of HPGe Czochralski crystal growth including purification and characterization of the material. The Ge starting material is purified in a horizontal multi-zone furnace. Two inch Ge single crystals were grown by the Czochralski method and analyzed with respect to electrically active impurities by means of Hall-e ect measurements and Photo-thermal Ionization Spectroscopy. The dislocation density is assessed by combining defect etching and optically microscopy imaging. Lateral Photovoltage Scanning is applied to visualize doping striations, which bear information on the shape of the solid-liquid phase boundary during growth. PACS: Keywords: A1:Impurities, A1:Purification, A2:Czochralski method, B2:Semiconducting germanium
1
1. Introduction
29 30
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Radiation detectors made of High Purity Germanium (HPGe) 31 are in the front line of fundamental research since their first 32 development. Mainly used as gamma-ray detectors they also 33 play an important role in the field of the search for neutrinoless 34 double-beta decay, nuclear physics and are used for dark matter 35 searches as well. A recent overview the reader can find in [1]. 36 One of the open questions in the field of particle physics is 37 the nature of neutrinos, if neutrinos are their own antiparticles. 38 Neutrinoless double-beta decay (0 2!) could be an evidence for 39 lepton number violation and it would be a phenomena beyond 40 the Standard Model of particle physics that could potentially 41 contribute to our understanding of the origins of the universe. 42 For the search of the 0 2! decay, among other detector types, 43 a large number of HPGe detectors are used. The germanium 44 isotope 76 Ge is one of the rare nuclei that undergoes double 45 beta decay and very conveniently one can use it to fabricate 46 an ionization detector with the best resolution among all detec47 tors used in particle physics. In this way, the searched decay 48 takes place in the detector itself ensuring very high detection e!ciency. The current running experiments GERDA [2] and the Majo- 49 rana Demonstrator [3] deployed each about 40 kg HPGe detectors fabricated from material enriched in the 76 Ge isotope. The 50 recently formed LEGEND collaboration [4] is aiming to deploy 51 enriched HPGe detectors first 200 kg and later with 1 ton total 52 mass. Therefore we expect over the next decade an increasing 53 54 demand for HPGe crystals. 55
Corresponding author Email address: [email protected] (F.M. Kießling) Preprint submitted to Elsevier
56 57
For these future HPGe detectors rather large, at least 3” Ge single crystals, are required with concentrations of the electrically active impurities of " 1010 cm!3 and dislocation densities well below 104 cm!2 [5]. In order to meet the extraordinary purity conditions, special growth equipment has to be built with inductive heating, which makes the control of the thermal field much more di!cult than in resistive heating systems (see e.g. [6]). The detectors for GERDA and the Majorana Demonstrator were produced entirely by commercial producers. To address the special requirements of future low background experiments and the potential shortage of production capacity for the tonscale experiment, several research institutions started developing HPGe crystal growth. Among others the University of South Dakota demonstrated the ability to grow HPGe crystals and recently they also succeeded to produce detectors [7, 8, 9]. Several years ago a similar development started at IKZ originally with the goal to produce crystals for the GERDA experiment. In this paper, we report on the recent developments and results of the project. 2. HPGe crystal growth technology The main ingredient of a HPGe detector is a germanium crystal with a net concentration of electrically active impurities in the range of 109 1010 cm!3 . Only at such low net doping concentrations, it becomes possible to fully deplete the detectors diode structure over centimeter-sized regions by applying reverse bias voltages of up to a practical limit of 5 kV. Hence, the trade-o between achievable crystal purity and reverse bias limitations influences the optimum in crystal and detector sizes. December 3, 2019
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
The next important requirement is a low density of structural defects, in particular dislocations, which may otherwise cause charge trapping in radiation detectors. Dislocation-free crystals on the other hand, are not suited for detectors, because in such crystals the divacancy-hydrogen complex may unimpededly be formed. This defect with an acceptor level at Ev 0.07 eV can be observed in the DLTS spectrum and attains a concentration of 2 1011 cm 3 [10, 11]. In weakly dislocated crystals, however, vacancies are gettered by the dislocations and the formation of the divacancy-hydrogen complex is e ectively suppressed. Experience shows that good quality detectors can be produced only from crystals with a dislocation density between 103 to 104 cm 2 [12]. It is important to note that 0 2 experiments in the future will require detectors with a mass of about 2 kg or more with good energy resolution and optimized for pulse shape discrimination. Such detectors are not produced routinely by the commercial manufacturers but they are occasionally made on request. All these requirements imply that crystals with diameters 75 mm of exceptional quality should be grown. The large diameter113 poses a challenge for constructing the growth equipment due114 115 to the limitations with respect of the high purity. From the literature [11, 13] is clear that HPGe crystal growth needs a dedicated Czochralski furnace with a gas system that116 allows for the use of pure hydrogen. In addition the furnace117 should provide the right thermal conditions for the growth of118 119 crystals with a small number of dislocations. Because the segregation coe cient of Al is known to be close120 to one during crystal growth (see for example [11, 13]) and the121 removal of other impurities is limited by the segregation dur-122 ing normal freezing, the purity requirement has to be achieved123 124 before the crystal growth process. 125
90
2.1. Zone refining
126 127
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
Zone refining or zone melting is the method of choice for pu-128 rifying germanium beyond the standard electronic grade purity.129 Pioneering work on this field was done by W.G. Pfann [14]. 130 Our self-built zone refiner consists of four fused silica tubes131 mounted in a supporting frame made of aluminium profiles132 which can be tilted by an angle between 0 and 10 degree. The133 germanium ingot is heated by a radio frequency coil to produce134 the molten zone which is moved along the bar. Typical zone ve-135 locities are between 1 and 5 mm min [14, 15]. The zone refiner136 is built and certified for use with pure H2 gas. 137 The power of the RF generator is adjusted as a function of the138 position of the induction coil. The purpose of varying the power139 during the process is to eliminate the problem of non-constant140 zone length: at the beginning and at the end of the ingot less141 power is required than in the middle for the same zone length142 due to the changing heat conduction conditions along the ingot.143 To prepare high purity material for crystal pulling we start144 with 6N electronic grade germanium which first we purify in145 graphite boats and afterwards in fused silica boats following146 the procedure described in [15]. Graphite boats are used for the147 ease of operation whenever is possible but for the final purifi-148 cation step silica boats are required with soot coating made by149 2
Figure 1: Simulated global temperature distribution during crystal pulling.
burning high purity (6N) methane gas. From the zone refined ingot we cut out the cleaner part and we load it in the crucible for crystal pulling. 2.2. Crystal growing The utilized crystal puller is a standard Czochralski furnace made by the company N¨urmont - Germany. It was equipped with a custom built gas system for the use with pure hydrogen. The hydrogen gas is purified with a palladium cell before entering the furnace. The heat is generated by inductive heating of a graphite susceptor that holds a 4 inch fused silica crucible. All internal parts of the furnace were designed by IKZ. The construction was preceded by a detailed simulation of the transport phenomena taking place in the furnace. A realistic 3D CAD model was used to define the geometry for the CFD simulation done with the commercial ANSYS software. A simulated temperature field in the furnace obtained by 3D calculation can be seen in Fig.1. Initially, the furnace was equipped for the growth of 2 inch crystals. The development is ongoing to upgrade the furnace for 3 inch and larger crystals. For the process development we employed axis-symmetric calculations performed with the Elmer software [16]. From a number of crystals we prepared vertical slices for Lateral Photovoltage Scanning (LPS). With LPS the shape of the solid-liquid interface during the growth can be reconstructed [17]. Because the method is based on the measurement of local di erences in resistivity a certain amount of charge carriers are required therefore the purity of Ge should be less than required for detector applications. The LPS picture was compared to the simulation. After tuning the material parameters of the simulated components the solid-liquid interface could be reproduced well as shown in Fig.2. With the optimum determined by the numerical simulation the crystal and crucible rotation speed was chosen to be 15 rpm and 2 rpm respectively. The pulling speed is kept constant during the process at 0.5 mm min. During growth the diameter of the crystal is controlled only by adjusting the power of the
Figure 3: Micrographs of etch pits from two crystals. One with a dislocation density of about 1000 cm 2 and another one with about 50000 cm 2 .
EPD [cm-2]
<100>
{100} <110> R
R/2
Middle
8000
Slice A <100> 7000
Slice A <110> 6000
Slice B <100>
5000
Slice B <110>
4000
Figure 2: LPS image of a crystal with the simulated solid-liquid interface overlay.
3000
Slice C <100>
2000
Slice C <110> 0
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
RF generator. To measure the diameter a computer program continuously analyses the digitized video image of the growing crystal. The diameter displayed on a screen serves as orientation for the operator, and later will be used to automatize the 179 whole process. 180 Limited by the size of the 4 inch crucible the crystals grown are between 0.7 to 1.5 kg with diameters from 30 to 60 mm. All181 182 crystals were grown in the 100 direction. Despite that the crystals are grown always with the Dash183 necking technique [18] the target for the number of dislocations184 is di cult to achieve due to the high heat conductivity of the H2 185 gas. To optimize the crucible-to-heater position we varied the186 position by 10 mm each time (by trial and error) and measured187 the EPD. The optimum in the furnace set-up was reached when188 189 the dislocation density went below the target of 104 cm 2 . 190 191
165
3. Crystal characterization
192 193
166 167 168 169 170 171 172 173 174 175 176 177 178
After the crystal is removed from the Czochralski furnace194 samples are cut for analysis. Usually three slices are cut from195 each crystal which are then used to produce the samples for196 the Hall e ect measurement, Photothermal Ionization Spec-197 troscopy (PTIS) and Etch Pit Density (EPD) measurement. In198 addition, from some crystals longitudinal slices were cut to per-199 form LPS measurements mentioned in Section 2. 200 For EPD analysis crystal slices undergo chemo mechanical201 polishing after which they are etched to reveal the defects. The etch pits are then visible under the microscope and counted (Fig.3). Figure 4 shows an example crystal with the location of the analysis slices and the radial orientation dependent measurement points of the etch pits. 3
R/2
R
Figure 4: Radial and axial distribution of the dislocation density measured at the points indicated in the drawing.
Significant di erences of EPD in crystallographic directions could not be found but in more than 2 3 of all investigated crystals the EPD decreases towards the tail. The later can be seen for the middle and R 2 position. The EPD values near the crystal surface (R) do not follow this trend. For van der Pauw-Hall-e ect measurement we had 8 mm x 8 mm x 1 mm samples cut. The measurement is performed first at room temperature and then with the sample submersed in liquid nitrogen. The results are shown in Table 1. The charge carrier concentration, which is equal to the net concentration of donor and acceptor impurities in Ge at 77K, is still above the requirements. To reduce it further (by an order of magnitude) we are implementing the techniques described in [15]. As the Hall e ect measurement yields only the unintentional net doping concentration we need PTIS spectra to identify the chemical nature of the donor and acceptor impurities present in the crystal. PTIS can also provide a hint about the compensation level. On the other hand it is di cult to extract quantitative information from the PTIS spectra. Therefore we use the results of the Hall e ect measurements to reconstruct the impurity profile in the crystal along the growth direction with the knowledge of the impurities shown by PTIS. Assuming a well-mixed melt during the entire growth process the concentration for a particular impurity i is given by the Scheil equation [19]: Ci (x)
Ci0 ki (1
x)ki
1
Sample
Resistivity [ cm] 145 1608 154
A B C
Mobility ! [cm2 "V s] 4#4 104 4#7 104 3#3 104
Carrier conc. [cm 3 ] 1#5 1012 1#3 1011 12 !1#5 10
226 227 228 229 230 231
-3
Al
Concentration [cm ]
Intensity [a.u.]
Table 1: Hall e ect measurement results of three samples taken from one crys232 tal. The negative sign stands for n-type conductivity. The measurements were done at liquid nitrogen temperature (77K). A
Al
0.06 0.04
B
Ga
B
Ga
Al P
0.02
P
P P
0
Intensity [a.u.]
50
60
70
80
Ga
0.4 0.3
Al
90
As a crosscheck we compare the longitudinal variation of the resistivity computed from the net concentration given by the model with the resistivity measurements on the crystal done at liquid nitrogen temperature. Mainly due to the rapidly decreasing boron acceptor and increasing phosphorous concentration the crystal turns n-type towards the tail. As one can see in Figure 6 the model describes rather well the resistivity profile. B Al |P - Ga| |N - ND|
1013
P
N
A
1012
100 110 Wavenumber [cm-1]
1011
B
Ga Al
0.2 0.1
P
Al
B
P
1010
P P
0
Intensity [a.u.]
−0.1 50
60
70
80
90
100 110 Wavenumber [cm-1]
109
C
P
0.4
0
0.2
60
70
80
90
100 110 Wavenumber [cm-1]
Figure 5: PTIS spectra of three samples taken from the same crystal and from the same slice as the Hall measurement samples in Table 1. The strong spectrum lines belonging to an impurity are indicated above the peak. The spectra were recorded at 7 K. The segregation of impurities during crystal growth can be clearly seen.
205 206 207 208 209 210 211
213 214 215 216 217 218 219 220 221 222 223 224 225
Ga #
Resistivity calculated
104
Resistivity measured
103
102
were x is the fraction of the melt solidified, Ci (x) is the concentration at x, Ci0 is the initial concentration of the impurity in the melt and ki is the segregation coe!cient. The measured net concentration is the sum of all donor and acceptor impurities at the point where the sample was taken. Knowing that we have three major components we can build a simple model: an acceptor impurity with k $ 1 (boron, k " 17 [20]), an acceptor impurity with k " 1 (aluminium, see [15, 21],) and a donor impurity with k % 1 (phosphorous k " 0#08 233 [20]). Thus we have C(x) " CB # CAl ! CP
212
0.8 1 Solidified fraction
Ga
Resistivity [ Ω cm]
0 50
204
0.6
P P
Ga
203
0.4
P
0.1
202
0.2
0.3
234 235
Since the equilibrium segregation coe!cients of phospho-236 rous and gallium are very similar (k " 0#087 [20]) alone from237 the Hall measurement their relative concentration cannot be de-238 duced. Therefore the term CP Ga includes the gallium acceptor,239 resulting in a compensated phosphorous concentration. 240 With Hall e$ect measurements at three di$erent positions241 along the growth direction we have a system of equations with242 three unknown Ci0 . Solving the system we obtain the concentra-243 tion of B, Al and the compensated P concentration in the melt. 244 Fig. 5 shows the PTIS spectra taken with samples from crys-245 tal slices almost at the same positions as the three Hall-e$ect246 measurements in Table 1. The spectra confirm the abundance247 of Al, B, and P in the crystal. In addition, the presence of gal-248 lium can be seen. 249 4
10
b)
0.2
0.4
0.6
0.8
1 Solidified fraction
Figure 6: a) Calculated impurity distribution of the crystal. b) Comparison of the resistivity estimated with the model and the measured resistivity in the opening direction of the p-n junction.
4. Summary Growing HPGe crystals for the purpose of radiation detectors is a challenging task especially what concerns the purity of the crystals. As the sensitivity goals of the germanium based 0&2' experiments are becoming more ambitious the requirements on the size and quality of the HPGe crystals are also becoming more stringent. Production of crystals from germanium enriched in the 76 Ge isotope is an additional challenge that could be better addressed with a dedicated production chain. For this reason IKZ set up the complete production chain needed to grow HPGe crystals starting with commercially available 6N material and later from isotopically modified germanium. This includes a zone refiner and a Czochralski furnace with H2 atmosphere. In this paper we presented our preliminary results with 2” crystals as part of the R&D e$ort to reach high purity with 3” crystals grown in a Czochralski furnace still under construction.
267
We measured the net charge carrier concentration by means309 of Hall e ect measurement and we determined the chemical310 identity of the electrically active impurities with PTIS on se-311 312 lected samples. In addition dislocation density was estimated313 with the EPD method. LPS measurements on longitudinal cuts314 315 were used to verify the applied numerical simulations. 316 After the reduction of the shallow level impurities by another317 order of magnitude we are planning to investigate deep levels318 and charge trapping. Performing Deep Level Transient Spec-319 troscopy (DLTS) will be the final step towards producing de-320 321 tector grade germanium. 322 The detector production for the LEGEND-1000 experiment323 will require hundreds of kilograms of isotopically enriched ger-324 manium turned into crystals. Besides the orders placed at com-325 326 mercial producers smaller batches could be regularly processed327 by IKZ as well. Alternatively the technology developed at IKZ328 could be later transferred to a commercial crystal grower for large scale production.
268
Acknowledgment
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
271
The project is supported financially by the German Federal Ministry for Education and Research (BMBF) under the grant number 05A17BC1.
272
References
269 270
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
[1] N. Fourches, M. Zieli´nska, G. Charles, High Purity Germanium: From Gamma-Ray Detection to Dark Matter Subterranean Detectorsdoi:10.5772 intechopen.82864. [2] M. Agostini, et al., Background-free search for neutrinoless double- decay of 76 Ge with GERDA, Nature 544 (2017) 47. [3] C. Aalseth, et al., Search for neutrinoless double- decay in 76 Ge with the Majorana Demonstrator, Phys. Rev. Lett. 120 (2018) 132502. [4] N. Abgrall, et al., The large enriched germanium experiment for neutrinoless double beta decay (LEGEND), AIP Conf. Proc. 1894 (1) (2017) 020027. [5] M. V. Sande, L. V. Goethem, L. D. Laet, H. Guislain, Dislocations in highpurity germanium and its relation to !-ray detector performance, Appl. Phys. A 40 (1986) 257–261. [6] V. A. Moskovskih, P. V. Kasimkin, V. N. Shlegel, Y. V. Vasiliev, V. A. Gridchin, O. I. Podkopaev, The low thermal gradient Cz technique as a way of growing of dislocation-free germanium crystals, J. Crystal Growth 401 (2014) 767–771. [7] X.-H. Meng, G.-J. Wang, M.-D. Wagner, H. Mei, W.-Z. Wei, J. Liu, G. Yang, D.-M. Mei, Fabrication and characterization of high-purity germanium detectors with amorphous germanium contacts, J. Instrum. 14 (02) (2019) P02019. [8] G. Wang, Y. Sun, G. Yang, W. Xiang, Y. Guan, D. Mei, C. Keller, Y.-D. Chan, Development of large size high-purity germanium crystal growth, J. Crystal Growth 352 (1) (2012) 27 – 30. [9] G. Wang, M. Amman, H. Mei, D. Mei, K. Irmscher, Y. Guan, G. Yang, Crystal growth and detector performance of large size high-purity ge crystals, Mat. Sci. in Semiconductor Processing 39 (2015) 54 – 60. [10] E. Haller, G. Hubbard, W. Hansen, A. Seeger, Divacancy-hydrogen complexes in dislocation-free high-purity germanium., Radiation e!ects in semiconductors, Institute of Physics Conference Series 31 (1976) 309 – 318. [11] E. E. Haller, W. L. Hansen, F. S. Goulding, Physics of ultra-pure germanium, Adv. Phys. 30 (1) (1981) 93–138. [12] P. A. Glasow, E. E. Haller, The e!ect of dislocation on the energy resolution of high-purity germanium detectors, IEEE Trans. Nucl. Sci. 23 (1) (1976) 92–96.
5
[13] R. N. Hall, T. J. Soltys, High purity germanium for detector fabrication, IEEE Trans. Nucl. Sci. 18 (1) (1971) 160–165. [14] W. G. Pfann, Zone Melting, Wiley & Sons, 1966. [15] G. S. Hubbard, E. E. Haller, W. L. Hansen, Zone refining high-purity germanium, IEEE Trans. Nucl. Sci. 25 (1) (1978) 362–370. [16] CSC - IT CENTER FOR SCIENCE LTD., CSC - Elmer (2019). URL https://www.csc.fi/web/elmer [17] N. V. Abrosimov, A. L¨udge, H. Riemann, W. Schr¨oder, Lateral photovoltage scanning (LPS) method for the visualization of the solid-liquid interface of Si1 x Ge x single crystals, J. Crystal Growth 237-239 (2002) 356–360. [18] W. C. Dash, Growth of silicon crystals free from dislocations, J. Appl. Phys. 30 (1959) 459–474. [19] E. Scheil, Bemerkungen zur Schichtkristallbildung, Zeitschrift Metallkunde 34 (1942) 70–72. [20] F. A. Trumbore, Solid solubilities of impurity elements in germanium and silicon, Bell System Technical Journal 39 (1) (1960) 205–233. [21] E. E. Haller, W. L. Hansen, G. S. Hubbard, F. S. Goulding, Origin and control of the dominant impurities in high-purity germanium, IEEE Trans. Nucl. Sci. 23 (1) (1976) 81–87.
I confirm that: a. there's no financial/personal interest or belief that could affect the objectivity b.
no potential conflicts exist.
Highlights !"complete production chain needed to grow Ge crystals with extreme high purity !"Ge has been purified in a horizontal multi-zone furnace !"2“ Ge single crystals were grown in a sophisticated designed Cz-puller !"crystals are analyzed by means of Hall-effect measurements and Photo-thermal Ionization Spectroscopy !"dislocation distribution have been evaluated from etch pits !
S0022-0248(19)30611-6 https://doi.org/10.1016/j.jcrysgro.2019.125396 CRYS 125396
To appear in:
Journal of Crystal Growth
Received Date: Revised Date: Accepted Date:
24 January 2019 24 October 2019 29 November 2019
Please cite this article as: N. Abrosimov, M. Czupalla, N. Dropka, J. Fischer, A. Gybin, K. Irmscher, J. JanicskóCsáthy, U. Juda, S. Kayser, W. Miller, M. Pietsch, F.M. Kießling, Technology Development of High Purity Germanium Crystals for Radiation Detectors, Journal of Crystal Growth (2019), doi: https://doi.org/10.1016/ j.jcrysgro.2019.125396
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier B.V.
Technology Development of High Purity Germanium Crystals for Radiation Detectors N. Abrosimov, M. Czupalla, N. Dropka, J. Fischer, A. Gybin, K. Irmscher, J. Janicsk´o-Cs´athy, U. Juda, S. Kayser, W. Miller, M. Pietsch, F.M. Kießling Leibniz-Institut f¨ur Kristallz¨uchtung (IKZ), Max-Born-Str. 2, 12489 Berlin, Germany
Abstract High Purity Germanium (HPGe) crystals are indispensable for radiation detection and fundamental research. Among other applications they will be used in the Large Enriched Germanium Experiment for Neutrinoless double beta Decay (LEGEND) experiment for the search of neutrinoless double beta decay of 76 Ge. Ongoing research activities are done in collaboration with members of the existing GERDA experiment. In this context the Leibniz-Institut f¨ur Kristallz¨uchtung (IKZ) is engaged in the development of HPGe Czochralski crystal growth including purification and characterization of the material. The Ge starting material is purified in a horizontal multi-zone furnace. Two inch Ge single crystals were grown by the Czochralski method and analyzed with respect to electrically active impurities by means of Hall-e ect measurements and Photo-thermal Ionization Spectroscopy. The dislocation density is assessed by combining defect etching and optically microscopy imaging. Lateral Photovoltage Scanning is applied to visualize doping striations, which bear information on the shape of the solid-liquid phase boundary during growth. PACS: Keywords: A1:Impurities, A1:Purification, A2:Czochralski method, B2:Semiconducting germanium
1
1. Introduction
29 30
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Radiation detectors made of High Purity Germanium (HPGe) 31 are in the front line of fundamental research since their first 32 development. Mainly used as gamma-ray detectors they also 33 play an important role in the field of the search for neutrinoless 34 double-beta decay, nuclear physics and are used for dark matter 35 searches as well. A recent overview the reader can find in [1]. 36 One of the open questions in the field of particle physics is 37 the nature of neutrinos, if neutrinos are their own antiparticles. 38 Neutrinoless double-beta decay (0 2!) could be an evidence for 39 lepton number violation and it would be a phenomena beyond 40 the Standard Model of particle physics that could potentially 41 contribute to our understanding of the origins of the universe. 42 For the search of the 0 2! decay, among other detector types, 43 a large number of HPGe detectors are used. The germanium 44 isotope 76 Ge is one of the rare nuclei that undergoes double 45 beta decay and very conveniently one can use it to fabricate 46 an ionization detector with the best resolution among all detec47 tors used in particle physics. In this way, the searched decay 48 takes place in the detector itself ensuring very high detection e!ciency. The current running experiments GERDA [2] and the Majo- 49 rana Demonstrator [3] deployed each about 40 kg HPGe detectors fabricated from material enriched in the 76 Ge isotope. The 50 recently formed LEGEND collaboration [4] is aiming to deploy 51 enriched HPGe detectors first 200 kg and later with 1 ton total 52 mass. Therefore we expect over the next decade an increasing 53 54 demand for HPGe crystals. 55
Corresponding author Email address: [email protected] (F.M. Kießling) Preprint submitted to Elsevier
56 57
For these future HPGe detectors rather large, at least 3” Ge single crystals, are required with concentrations of the electrically active impurities of " 1010 cm!3 and dislocation densities well below 104 cm!2 [5]. In order to meet the extraordinary purity conditions, special growth equipment has to be built with inductive heating, which makes the control of the thermal field much more di!cult than in resistive heating systems (see e.g. [6]). The detectors for GERDA and the Majorana Demonstrator were produced entirely by commercial producers. To address the special requirements of future low background experiments and the potential shortage of production capacity for the tonscale experiment, several research institutions started developing HPGe crystal growth. Among others the University of South Dakota demonstrated the ability to grow HPGe crystals and recently they also succeeded to produce detectors [7, 8, 9]. Several years ago a similar development started at IKZ originally with the goal to produce crystals for the GERDA experiment. In this paper, we report on the recent developments and results of the project. 2. HPGe crystal growth technology The main ingredient of a HPGe detector is a germanium crystal with a net concentration of electrically active impurities in the range of 109 1010 cm!3 . Only at such low net doping concentrations, it becomes possible to fully deplete the detectors diode structure over centimeter-sized regions by applying reverse bias voltages of up to a practical limit of 5 kV. Hence, the trade-o between achievable crystal purity and reverse bias limitations influences the optimum in crystal and detector sizes. December 3, 2019
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
The next important requirement is a low density of structural defects, in particular dislocations, which may otherwise cause charge trapping in radiation detectors. Dislocation-free crystals on the other hand, are not suited for detectors, because in such crystals the divacancy-hydrogen complex may unimpededly be formed. This defect with an acceptor level at Ev 0.07 eV can be observed in the DLTS spectrum and attains a concentration of 2 1011 cm 3 [10, 11]. In weakly dislocated crystals, however, vacancies are gettered by the dislocations and the formation of the divacancy-hydrogen complex is e ectively suppressed. Experience shows that good quality detectors can be produced only from crystals with a dislocation density between 103 to 104 cm 2 [12]. It is important to note that 0 2 experiments in the future will require detectors with a mass of about 2 kg or more with good energy resolution and optimized for pulse shape discrimination. Such detectors are not produced routinely by the commercial manufacturers but they are occasionally made on request. All these requirements imply that crystals with diameters 75 mm of exceptional quality should be grown. The large diameter113 poses a challenge for constructing the growth equipment due114 115 to the limitations with respect of the high purity. From the literature [11, 13] is clear that HPGe crystal growth needs a dedicated Czochralski furnace with a gas system that116 allows for the use of pure hydrogen. In addition the furnace117 should provide the right thermal conditions for the growth of118 119 crystals with a small number of dislocations. Because the segregation coe cient of Al is known to be close120 to one during crystal growth (see for example [11, 13]) and the121 removal of other impurities is limited by the segregation dur-122 ing normal freezing, the purity requirement has to be achieved123 124 before the crystal growth process. 125
90
2.1. Zone refining
126 127
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
Zone refining or zone melting is the method of choice for pu-128 rifying germanium beyond the standard electronic grade purity.129 Pioneering work on this field was done by W.G. Pfann [14]. 130 Our self-built zone refiner consists of four fused silica tubes131 mounted in a supporting frame made of aluminium profiles132 which can be tilted by an angle between 0 and 10 degree. The133 germanium ingot is heated by a radio frequency coil to produce134 the molten zone which is moved along the bar. Typical zone ve-135 locities are between 1 and 5 mm min [14, 15]. The zone refiner136 is built and certified for use with pure H2 gas. 137 The power of the RF generator is adjusted as a function of the138 position of the induction coil. The purpose of varying the power139 during the process is to eliminate the problem of non-constant140 zone length: at the beginning and at the end of the ingot less141 power is required than in the middle for the same zone length142 due to the changing heat conduction conditions along the ingot.143 To prepare high purity material for crystal pulling we start144 with 6N electronic grade germanium which first we purify in145 graphite boats and afterwards in fused silica boats following146 the procedure described in [15]. Graphite boats are used for the147 ease of operation whenever is possible but for the final purifi-148 cation step silica boats are required with soot coating made by149 2
Figure 1: Simulated global temperature distribution during crystal pulling.
burning high purity (6N) methane gas. From the zone refined ingot we cut out the cleaner part and we load it in the crucible for crystal pulling. 2.2. Crystal growing The utilized crystal puller is a standard Czochralski furnace made by the company N¨urmont - Germany. It was equipped with a custom built gas system for the use with pure hydrogen. The hydrogen gas is purified with a palladium cell before entering the furnace. The heat is generated by inductive heating of a graphite susceptor that holds a 4 inch fused silica crucible. All internal parts of the furnace were designed by IKZ. The construction was preceded by a detailed simulation of the transport phenomena taking place in the furnace. A realistic 3D CAD model was used to define the geometry for the CFD simulation done with the commercial ANSYS software. A simulated temperature field in the furnace obtained by 3D calculation can be seen in Fig.1. Initially, the furnace was equipped for the growth of 2 inch crystals. The development is ongoing to upgrade the furnace for 3 inch and larger crystals. For the process development we employed axis-symmetric calculations performed with the Elmer software [16]. From a number of crystals we prepared vertical slices for Lateral Photovoltage Scanning (LPS). With LPS the shape of the solid-liquid interface during the growth can be reconstructed [17]. Because the method is based on the measurement of local di erences in resistivity a certain amount of charge carriers are required therefore the purity of Ge should be less than required for detector applications. The LPS picture was compared to the simulation. After tuning the material parameters of the simulated components the solid-liquid interface could be reproduced well as shown in Fig.2. With the optimum determined by the numerical simulation the crystal and crucible rotation speed was chosen to be 15 rpm and 2 rpm respectively. The pulling speed is kept constant during the process at 0.5 mm min. During growth the diameter of the crystal is controlled only by adjusting the power of the
Figure 3: Micrographs of etch pits from two crystals. One with a dislocation density of about 1000 cm 2 and another one with about 50000 cm 2 .
EPD [cm-2]
<100>
{100} <110> R
R/2
Middle
8000
Slice A <100> 7000
Slice A <110> 6000
Slice B <100>
5000
Slice B <110>
4000
Figure 2: LPS image of a crystal with the simulated solid-liquid interface overlay.
3000
Slice C <100>
2000
Slice C <110> 0
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
RF generator. To measure the diameter a computer program continuously analyses the digitized video image of the growing crystal. The diameter displayed on a screen serves as orientation for the operator, and later will be used to automatize the 179 whole process. 180 Limited by the size of the 4 inch crucible the crystals grown are between 0.7 to 1.5 kg with diameters from 30 to 60 mm. All181 182 crystals were grown in the 100 direction. Despite that the crystals are grown always with the Dash183 necking technique [18] the target for the number of dislocations184 is di cult to achieve due to the high heat conductivity of the H2 185 gas. To optimize the crucible-to-heater position we varied the186 position by 10 mm each time (by trial and error) and measured187 the EPD. The optimum in the furnace set-up was reached when188 189 the dislocation density went below the target of 104 cm 2 . 190 191
165
3. Crystal characterization
192 193
166 167 168 169 170 171 172 173 174 175 176 177 178
After the crystal is removed from the Czochralski furnace194 samples are cut for analysis. Usually three slices are cut from195 each crystal which are then used to produce the samples for196 the Hall e ect measurement, Photothermal Ionization Spec-197 troscopy (PTIS) and Etch Pit Density (EPD) measurement. In198 addition, from some crystals longitudinal slices were cut to per-199 form LPS measurements mentioned in Section 2. 200 For EPD analysis crystal slices undergo chemo mechanical201 polishing after which they are etched to reveal the defects. The etch pits are then visible under the microscope and counted (Fig.3). Figure 4 shows an example crystal with the location of the analysis slices and the radial orientation dependent measurement points of the etch pits. 3
R/2
R
Figure 4: Radial and axial distribution of the dislocation density measured at the points indicated in the drawing.
Significant di erences of EPD in crystallographic directions could not be found but in more than 2 3 of all investigated crystals the EPD decreases towards the tail. The later can be seen for the middle and R 2 position. The EPD values near the crystal surface (R) do not follow this trend. For van der Pauw-Hall-e ect measurement we had 8 mm x 8 mm x 1 mm samples cut. The measurement is performed first at room temperature and then with the sample submersed in liquid nitrogen. The results are shown in Table 1. The charge carrier concentration, which is equal to the net concentration of donor and acceptor impurities in Ge at 77K, is still above the requirements. To reduce it further (by an order of magnitude) we are implementing the techniques described in [15]. As the Hall e ect measurement yields only the unintentional net doping concentration we need PTIS spectra to identify the chemical nature of the donor and acceptor impurities present in the crystal. PTIS can also provide a hint about the compensation level. On the other hand it is di cult to extract quantitative information from the PTIS spectra. Therefore we use the results of the Hall e ect measurements to reconstruct the impurity profile in the crystal along the growth direction with the knowledge of the impurities shown by PTIS. Assuming a well-mixed melt during the entire growth process the concentration for a particular impurity i is given by the Scheil equation [19]: Ci (x)
Ci0 ki (1
x)ki
1
Sample
Resistivity [ cm] 145 1608 154
A B C
Mobility ! [cm2 "V s] 4#4 104 4#7 104 3#3 104
Carrier conc. [cm 3 ] 1#5 1012 1#3 1011 12 !1#5 10
226 227 228 229 230 231
-3
Al
Concentration [cm ]
Intensity [a.u.]
Table 1: Hall e ect measurement results of three samples taken from one crys232 tal. The negative sign stands for n-type conductivity. The measurements were done at liquid nitrogen temperature (77K). A
Al
0.06 0.04
B
Ga
B
Ga
Al P
0.02
P
P P
0
Intensity [a.u.]
50
60
70
80
Ga
0.4 0.3
Al
90
As a crosscheck we compare the longitudinal variation of the resistivity computed from the net concentration given by the model with the resistivity measurements on the crystal done at liquid nitrogen temperature. Mainly due to the rapidly decreasing boron acceptor and increasing phosphorous concentration the crystal turns n-type towards the tail. As one can see in Figure 6 the model describes rather well the resistivity profile. B Al |P - Ga| |N - ND|
1013
P
N
A
1012
100 110 Wavenumber [cm-1]
1011
B
Ga Al
0.2 0.1
P
Al
B
P
1010
P P
0
Intensity [a.u.]
−0.1 50
60
70
80
90
100 110 Wavenumber [cm-1]
109
C
P
0.4
0
0.2
60
70
80
90
100 110 Wavenumber [cm-1]
Figure 5: PTIS spectra of three samples taken from the same crystal and from the same slice as the Hall measurement samples in Table 1. The strong spectrum lines belonging to an impurity are indicated above the peak. The spectra were recorded at 7 K. The segregation of impurities during crystal growth can be clearly seen.
205 206 207 208 209 210 211
213 214 215 216 217 218 219 220 221 222 223 224 225
Ga #
Resistivity calculated
104
Resistivity measured
103
102
were x is the fraction of the melt solidified, Ci (x) is the concentration at x, Ci0 is the initial concentration of the impurity in the melt and ki is the segregation coe!cient. The measured net concentration is the sum of all donor and acceptor impurities at the point where the sample was taken. Knowing that we have three major components we can build a simple model: an acceptor impurity with k $ 1 (boron, k " 17 [20]), an acceptor impurity with k " 1 (aluminium, see [15, 21],) and a donor impurity with k % 1 (phosphorous k " 0#08 233 [20]). Thus we have C(x) " CB # CAl ! CP
212
0.8 1 Solidified fraction
Ga
Resistivity [ Ω cm]
0 50
204
0.6
P P
Ga
203
0.4
P
0.1
202
0.2
0.3
234 235
Since the equilibrium segregation coe!cients of phospho-236 rous and gallium are very similar (k " 0#087 [20]) alone from237 the Hall measurement their relative concentration cannot be de-238 duced. Therefore the term CP Ga includes the gallium acceptor,239 resulting in a compensated phosphorous concentration. 240 With Hall e$ect measurements at three di$erent positions241 along the growth direction we have a system of equations with242 three unknown Ci0 . Solving the system we obtain the concentra-243 tion of B, Al and the compensated P concentration in the melt. 244 Fig. 5 shows the PTIS spectra taken with samples from crys-245 tal slices almost at the same positions as the three Hall-e$ect246 measurements in Table 1. The spectra confirm the abundance247 of Al, B, and P in the crystal. In addition, the presence of gal-248 lium can be seen. 249 4
10
b)
0.2
0.4
0.6
0.8
1 Solidified fraction
Figure 6: a) Calculated impurity distribution of the crystal. b) Comparison of the resistivity estimated with the model and the measured resistivity in the opening direction of the p-n junction.
4. Summary Growing HPGe crystals for the purpose of radiation detectors is a challenging task especially what concerns the purity of the crystals. As the sensitivity goals of the germanium based 0&2' experiments are becoming more ambitious the requirements on the size and quality of the HPGe crystals are also becoming more stringent. Production of crystals from germanium enriched in the 76 Ge isotope is an additional challenge that could be better addressed with a dedicated production chain. For this reason IKZ set up the complete production chain needed to grow HPGe crystals starting with commercially available 6N material and later from isotopically modified germanium. This includes a zone refiner and a Czochralski furnace with H2 atmosphere. In this paper we presented our preliminary results with 2” crystals as part of the R&D e$ort to reach high purity with 3” crystals grown in a Czochralski furnace still under construction.
267
We measured the net charge carrier concentration by means309 of Hall e ect measurement and we determined the chemical310 identity of the electrically active impurities with PTIS on se-311 312 lected samples. In addition dislocation density was estimated313 with the EPD method. LPS measurements on longitudinal cuts314 315 were used to verify the applied numerical simulations. 316 After the reduction of the shallow level impurities by another317 order of magnitude we are planning to investigate deep levels318 and charge trapping. Performing Deep Level Transient Spec-319 troscopy (DLTS) will be the final step towards producing de-320 321 tector grade germanium. 322 The detector production for the LEGEND-1000 experiment323 will require hundreds of kilograms of isotopically enriched ger-324 manium turned into crystals. Besides the orders placed at com-325 326 mercial producers smaller batches could be regularly processed327 by IKZ as well. Alternatively the technology developed at IKZ328 could be later transferred to a commercial crystal grower for large scale production.
268
Acknowledgment
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
271
The project is supported financially by the German Federal Ministry for Education and Research (BMBF) under the grant number 05A17BC1.
272
References
269 270
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
[1] N. Fourches, M. Zieli´nska, G. Charles, High Purity Germanium: From Gamma-Ray Detection to Dark Matter Subterranean Detectorsdoi:10.5772 intechopen.82864. [2] M. Agostini, et al., Background-free search for neutrinoless double- decay of 76 Ge with GERDA, Nature 544 (2017) 47. [3] C. Aalseth, et al., Search for neutrinoless double- decay in 76 Ge with the Majorana Demonstrator, Phys. Rev. Lett. 120 (2018) 132502. [4] N. Abgrall, et al., The large enriched germanium experiment for neutrinoless double beta decay (LEGEND), AIP Conf. Proc. 1894 (1) (2017) 020027. [5] M. V. Sande, L. V. Goethem, L. D. Laet, H. Guislain, Dislocations in highpurity germanium and its relation to !-ray detector performance, Appl. Phys. A 40 (1986) 257–261. [6] V. A. Moskovskih, P. V. Kasimkin, V. N. Shlegel, Y. V. Vasiliev, V. A. Gridchin, O. I. Podkopaev, The low thermal gradient Cz technique as a way of growing of dislocation-free germanium crystals, J. Crystal Growth 401 (2014) 767–771. [7] X.-H. Meng, G.-J. Wang, M.-D. Wagner, H. Mei, W.-Z. Wei, J. Liu, G. Yang, D.-M. Mei, Fabrication and characterization of high-purity germanium detectors with amorphous germanium contacts, J. Instrum. 14 (02) (2019) P02019. [8] G. Wang, Y. Sun, G. Yang, W. Xiang, Y. Guan, D. Mei, C. Keller, Y.-D. Chan, Development of large size high-purity germanium crystal growth, J. Crystal Growth 352 (1) (2012) 27 – 30. [9] G. Wang, M. Amman, H. Mei, D. Mei, K. Irmscher, Y. Guan, G. Yang, Crystal growth and detector performance of large size high-purity ge crystals, Mat. Sci. in Semiconductor Processing 39 (2015) 54 – 60. [10] E. Haller, G. Hubbard, W. Hansen, A. Seeger, Divacancy-hydrogen complexes in dislocation-free high-purity germanium., Radiation e!ects in semiconductors, Institute of Physics Conference Series 31 (1976) 309 – 318. [11] E. E. Haller, W. L. Hansen, F. S. Goulding, Physics of ultra-pure germanium, Adv. Phys. 30 (1) (1981) 93–138. [12] P. A. Glasow, E. E. Haller, The e!ect of dislocation on the energy resolution of high-purity germanium detectors, IEEE Trans. Nucl. Sci. 23 (1) (1976) 92–96.
5
[13] R. N. Hall, T. J. Soltys, High purity germanium for detector fabrication, IEEE Trans. Nucl. Sci. 18 (1) (1971) 160–165. [14] W. G. Pfann, Zone Melting, Wiley & Sons, 1966. [15] G. S. Hubbard, E. E. Haller, W. L. Hansen, Zone refining high-purity germanium, IEEE Trans. Nucl. Sci. 25 (1) (1978) 362–370. [16] CSC - IT CENTER FOR SCIENCE LTD., CSC - Elmer (2019). URL https://www.csc.fi/web/elmer [17] N. V. Abrosimov, A. L¨udge, H. Riemann, W. Schr¨oder, Lateral photovoltage scanning (LPS) method for the visualization of the solid-liquid interface of Si1 x Ge x single crystals, J. Crystal Growth 237-239 (2002) 356–360. [18] W. C. Dash, Growth of silicon crystals free from dislocations, J. Appl. Phys. 30 (1959) 459–474. [19] E. Scheil, Bemerkungen zur Schichtkristallbildung, Zeitschrift Metallkunde 34 (1942) 70–72. [20] F. A. Trumbore, Solid solubilities of impurity elements in germanium and silicon, Bell System Technical Journal 39 (1) (1960) 205–233. [21] E. E. Haller, W. L. Hansen, G. S. Hubbard, F. S. Goulding, Origin and control of the dominant impurities in high-purity germanium, IEEE Trans. Nucl. Sci. 23 (1) (1976) 81–87.
I confirm that: a. there's no financial/personal interest or belief that could affect the objectivity b.
no potential conflicts exist.
Highlights !"complete production chain needed to grow Ge crystals with extreme high purity !"Ge has been purified in a horizontal multi-zone furnace !"2“ Ge single crystals were grown in a sophisticated designed Cz-puller !"crystals are analyzed by means of Hall-effect measurements and Photo-thermal Ionization Spectroscopy !"dislocation distribution have been evaluated from etch pits !