Design of high-power ISOL targets for radioactive ion beam generation

ARTICLE IN PRESS

Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Design of high-power ISOL targets for radioactive ion beam generation Y. Zhang1, G.D. Alton* Physics Division, Oak Ridge National Laboratory,2 Oak Ridge, TN, USA

Abstract In this report, we provide lists of refractory oxides, carbides and refractory metals suitable for use as targets for producing short-lived, proton-rich isotopes of elements (He through Pu) and neutron-rich isotopes of elements (As through Dy) for potential use at high-energy, ISOL-based radioactive ion beam facilities. Complex structure, highly permeable C matrices are described for coating with optimum thicknesses of any type of refractory target material (metal, carbide or oxide). Prescriptions are given for the design and fabrication of custom-engineered targets with diffusion lengths compatible with the release of isotopes of interest within their lifetimes. Computationally derived thermal analysis information is presented for selected low-density, fibrous, highly permeable targets, subjected to direct irradiation with 1 GeV, 100–400 kW proton beams. From these studies, internal thermal radiation is reconfirmed as an important heat transfer (cooling) mechanism within low-density, fibrous and composite targets. By utilization of the radiation cooling effect and beam manipulation techniques, in combination with placement of additional heat shielding on the exit end of targets, beam power depositional densities can be controlled and temperatures homogenized to acceptable levels within fast diffusion release, fast effusive-flow ISOL targets subjected to irradiation with 400 kW proton beams, as required at next-generation radioactive ion beam facilities. r 2004 Elsevier B.V. All rights reserved. PACS: 29.25.
t; 44.05.+e; 44.10.+i; 44.30.+v; 44.40.+a; 66.10.Cb; 66.30.
h Keywords: Target list; Fiber target; Foam target; Heavy ion target; Heating of target; Protective layer

1. Introduction During the past decade, important research has been successfully completed with radioactive ion *Corresponding author. Physics Division, Oak Ridge National Laboratory, Bldg. 6000, P.O. Box 2008, Oak Ridge, TN 37831-6368, USA. Tel.: +1-865-576-2648; fax: +1-865-5741268. E-mail addresses: [email protected] (Y. Zhang), [email protected] (G.D. Alton). 1 Also for correspondence. 2 Managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725.

beams (RIBs) at first-generation Isotope Separator On-Line (ISOL) facilities such as the Holifield Radioactive Ion Beam Facility (HRIBF) [1,2]. The availability of such beams has provided first-ever opportunities for addressing fundamentally important questions in nuclear physics and nuclearastrophysics that were previously inaccessible to experimental study using stable-beam/stable-target combinations. Despite species and intensity limitations imposed by available primary beam energy and power at the HRIBF, the nuclear structure properties [3] and decay modes of exotic nuclei have been measured [4], and cross-sections of nuclear

0168-9002/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2003.11.419

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

reactions of astrophysics importance in the synthesis of the elements [5–9] have been measured. Technical issues related to the design of robust, high-power, short diffusion length ISOL targets present the greatest impediment to success at ISOLbased facilities. The successes at the HRIBF are, in large part, attributable to the development of new technology, fast diffusion release, highly permeable targets [10–13] and high ionization efficiency ion sources [13–17]. These successes have invigorated interest in constructing next-generation facilities, such as the Rare Isotope Accelerator (RIA) facility, now under proposal for construction in the USA [18,19], that can provide the primary beam energy and power capabilities necessary for supplying a much broader spectrum of short-lived isotopes for research than presently available. However, the advent of such facilities will present further challenges to develop robust targets with short diffusion length and high permeability attributes that can withstand irradiation with 400 kW lightion beams for extended periods of time. Although important experiments have been successfully completed with radioactive beams of 100 particles/s or less at ISOL-based research facilities, in order to maintain viable, broad-based research programs in nuclear structure physics and nuclear astrophysics research at next-generation facilities, much higher radioactive ion beam (RIB) intensities will be required than presently available, ranging up to intensities now used in stable-beam experiments. However, experimentally useful RIB intensities of very short-lived species, produced by the ISOL method, are often difficult to generate, since they must be diffused from the interior of the target material, effusively transported to the ionization chamber of an ion source, ionized, extracted, mass analyzed, and accelerated to research energies in a time-span commensurate with their lifetimes. Unquestionably, target issues loom among the most challenging of problems associated with such research facilities. Robust production targets must be developed with design features commensurate with the release of a broad spectrum of isotopes of a large number of elements that can withstand irradiation with light ion beams at power levels up to 400 kW for extended periods of time for next-generation facilities such as the

73

RIA. Targets with these attributes pose serious challenges to personnel responsible for their design. One approach to avoiding the excessive heat generation problem imposed by direct radiation is by using the two-step target concept [20,21] whereby the production beam interacts with a primary target that emits high fluxes of neutrons. The neutrons then react with a close-coupled secondary target made of a fissionable material such as UC2 or ThC2. The secondary target can then be independently heated to the desired, controllable temperature. However, the two-step concept can only be used to produce neutron-rich species. For the production of proton-rich species, targets must be designed that can withstand direct irradiation with high-energy, high-power (400 kW), light-ion beams. The primary factors that govern the ultimate intensity of RIBs include: the energy, type, and intensity of the primary beam; the cross-section for production; the lifetime of the species of interest; the dimensions, densities and distributions of target materials; the maximum temperature at which targets can be operated; the dimensions, geometry and materials of construction of the vapor transport system; and the ionization efficiencies of ion sources chosen for RIB generation. Maximum production rates are set by reaction cross-sections for producing the species of interest and the practical limits of the primary beam intensity, in terms of maximum permissible power-density-ontarget that can be used without compromising the efficiency of the ion source or the physical integrity of the target. The principal means whereby short half-life radioactive species are lost between initial formation and utilization are associated with the times required to diffuse species of interest from the target material and to transport them to the ionization chamber of the source in relation to their lifetimes and therefore, delay times associated with these processes must be reduced to as low as practical values. These requirements suggest selection of highly refractory target materials that can be formed into short diffusion length, integral, highly permeable (low density) open structures. Targets that meet the fast diffusion release and effusive-flow criteria are necessarily fragile because of their low-density, low-thermal conductivity, low-heat capacity properties and therefore, can be easily

ARTICLE IN PRESS 74

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

destroyed by deposition of excessively high amounts of production beam power. Therefore, the irradiation process must be carefully managed to avoid exceeding the limiting temperature of the target material while homogeneously distributing the temperature over the volume of the target. This article describes a small part of an ongoing effort to design high-power ISOL targets for use at next-generation RIB facilities such as at the RIA. The RIA will utilize traditional ISOL as well as projectile fragmentation technologies for generation of short-lived accelerated RIBs for use in nuclear structure and astrophysics research. The objective of the present studies is to provide information appropriate for selection of refractory target materials and for the design and fabrication of robust, high-power ISOL targets with fast diffusion release and fast effusive-flow properties as required for generating useful radioactive ion (RIB) beam intensities of short-lived species at such facilities. Through these efforts, we seek to place target design technology on a firm scientific basis and avoid on-line trial-and-error processes that have characterized ISOL target development in the past. ISOL targets designed with careful attention given to the minimization of the diffusion release and effusive-flow times in combination with good thermal management of the power density distribution imparted to targets by the production beam can effectively enhance the intensities to useful values for a much wider range of shorter-lived species than otherwise available with conventional ISOL target design wisdom.

Sage6 were used to compute vapor pressures and thermal equilibrium compositions that provide limiting temperature information for candidate target materials. (The limiting temperature is defined as the temperature, above which, the ionization efficiency of the particular ion source is deleteriously affected and thus, serves as a fundamentally important parameter for material selection.) The Monte Carlo simulation code, GEANT4,7 was used to compute heat deposited in targets, in walls of target material reservoirs and in heat shielding surrounding the target by primary beams and by energetic secondary beams and fission fragments emitted during primary beam-target interactions. The power density information is then used as input to the finite element thermal analysis code, ANSYS,8 for extracting temperature distributions within a given target assembly.

2. Computer codes

6 FactSage is a chemical reaction and chemical equilibrium composition computer code developed by GTT Technologies, Herzogenrath, Germany and Ecole Polytechnique, Montreal, Canada H3C 3A7. 7 GEANT4 is a Monte Carlo code for calculating scattering effects, energy loss and production of secondary particles through nuclear reactions by energetic ion beams. The code tracks all primary and secondary particles until they are lost from the target material reservoir and accounts for energy deposition of all species, including fission reaction products. http://wwwinfo.cern.ch/asd/geant4. 8 ANSYS is a finite element computer code designed to simulate thermal transport, mechanical stress and electromagnetic problems; the code is marketed by ANSYS, Inc., Houston, PA.

Several thermo-chemistry codes including HSC,3 ThermoCalc,4 ChemSage,5 and FACT3

HSC is a chemical reaction and chemical equilibrium composition computer code and is a product of Outokumpu Research, Oy, Pori, Finland. 4 ThermoCalc is a chemical equilibrium composition and phase-diagram code developed by the Royal Institute of Technology, Stockholm, Sweden. 5 ChemSage is a chemical reaction and chemical equilibrium composition computer code and is a product of GTT Technologies, Herzogenrath, Germany.

3. Selection of target materials 3.1. Selection criteria Target-material selection begins by considering the physical, chemical, and thermal properties of the target material in relation to those of the product species. One of the principal problems lies in the availability of target materials that are sufficiently refractory so that they can be raised to the temperatures necessary for fast release of the product species without excessive vaporization/ sublimation of the target material itself. The choice

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

of material is further restricted by the requirement that the radioactive species be easily diffused from the target material and readily volatilized for subsequent transport to the ionization region of the source. Thus, in the ideal case, the radioactive species should possess physical and chemical properties almost opposite to those of the target material itself. For release of the species of interest, the species should not form refractory compounds within the target material, should rapidly diffuse to the surface, and upon reaching the surface, the species should be readily desorbed. These idealized differences in chemical and physical properties of the target/species combination are not often realizable, particularly for close-lying elements where their physical and chemical properties are often similar. However, the choice of high-limiting temperature materials alone does not guarantee fast, nor efficient, release of the species of interest from the target. It is imperative that the target materials have diffusion path-lengths (dimensions) commensurate with the fast release of the product species within its lifetime and that the diffusion coefficient be known for the particular target/ species combination. Since the diffusion coefficient, D, in a solid medium depends exponentially on the operational temperature of the target, it is desirable to heat the target to temperatures as high as practical. The upper limit to which any target can be operated depends on the vapor pressure of the target material in relation to the sensitivity of the ionization process to particle density within the ionization volume of the source, in the absence of other unwanted chemical reactions that could otherwise reduce the upper operational temperature. Thus, the limiting temperature of the target material is defined as the temperature at which the vapor pressure begins to deleteriously affect the ionization efficiency of the particular ion source used in the RIB generation process. Therefore, it is essential to know the upper temperature limit to which the target material can be raised; this value serves as one of the most important criteria for selecting a particular target material. For example, the temperature at which a given compound begins to dissociate compliments vapor pressure versus temperature data for these materials [12] where the limiting operating pressure of the

75

EBPIS (B2.67  10
2 Pa [13–15]) is used to define the vapor pressure limit to which a particular compound can be heated and hence, defines the limiting temperature of the particular material. (This criterion has been experimentally verified for several successful targets at the HRIBF.) 3.2. Other target-material selection considerations Obviously, target materials with the highest percentage of the production nuclei are desirable in order to maximize the production rate of the species of interest. Still other factors complicate the choice of target material, such as the presence of high atomic number nuclei that do not contribute to the production of the species of interest, but slow the production projectile down at a faster rate and, thereby, generate more heat in the target than necessary because of their larger
dE/dx (stopping power). For example, in choosing a particular target material for an application in which the production element is common in a group of refractory compounds such as O for the production of F isotopes with Al2O3, Y2O3, ZrO2 or HfO2, one must take into consideration increases in
dE/dx with atomic number, Z, and the higher consequent target temperatures for a given production-beam intensity due to increases in beam-deposited heat. Heavy nuclei may also produce unwanted, longlived radioactive by-product species that complicate the radioactive material-handling problem. 3.3. Target materials A partial listing of candidate target materials for the production of proton-rich and neutron-rich radioactive species through spallation or fission reactions, produced with 1 GeV protons, is given in Table 1. A broad range of proton-rich and neutron-rich nuclei can be produced in targets such as UC2 and ThC2 at high-energy facilities. Although the product species listed for a given target material can be produced through the respective nuclear reactions, many of these targets have not yet been tested on-line and not all of the species can be efficiently diffused from the particular target material because of unfavorable physical and chemical properties of particular

ARTICLE IN PRESS 76

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Table 1 Candidate spallation and fission target materials for highenergy (1 GeV) proton beam ISOL facilities Candidate target material

Limiting temp. ( K)

Product species

BeO C Al2O3 SiC TiC VC MnC2 Y2O3 Zr ZrC ZrO2 Nb NbC CeC2 CeS HfO2 Ta TaC W WC Re ThC2 UC2

2503 2240 1998 1973 2323 2203 2973 2223 2303 2648 2323 2353 2493 2663 2173 2473 2923 2613 3136 2753 2873 2923 2373

He - F He - N He - Si He - P He - V He - Cr He - Fe He - F; Al - Zr P - Nb P - Nb; He - N He - F; Si - Nb P - Mo He - N; P - Mo He - N; Cu - Nd He - Cl; Cu - Nd He - Ta He - W He - W He - Re He - Re He - Os He - Pa He - Np

species/target-material combinations. The list of target materials is not comprehensive nor is it intended to suggest that a particular production reaction/target material combination is the best combination for production and release of a given species. In cases where the species of interest cannot be released from the target matrix due to strong chemical reactions, alternative target materials must be considered. Since the species of interest must be expeditiously released from the volume of the target material within its lifetime through the sensitive temperature-dependent diffusion mechanism, all other parameters being equal, target materials with the highest limiting temperatures are the most desirable. Definitive evaluation of the viability of a given target material for production applications can only be made on-line at facilities such as the UNISOR.9

3.4. Fibrous compound materials In a few cases, fibrous materials with small diameters are commercially available at the purities required for ISOL target applications.10 These materials include Al2O3, SiO2, Sc2O3, TiO2, Y2O3, ZrO2, HfO2, Ta2O5 and rare-earth oxides. All of these candidate target materials are made of thin fibers (a few mm in diameter) and are highly permeable as required for fast effusive-flow transport following diffusion release. 3.4.1. Refractory metals Several metals, deposited or self-supported in thin layers or low-density metal foams, can be considered for direct use as spallation targets, including: Zr(2307 K); Nb(2618 K); Mo(2428 K); Tc(2443 K); Ru(2318 K); Rh(2013 K); Hf(2313 K); Ta(2918 K); W(3108 K); Re(2903 K); Os(2813 K); Ir(2443 K); Pt(2068 K). (The temperatures in parentheses represent the limiting temperature of the particular metal.) 3.5. Thermo-chemistry issues and effects The temperatures at which gaseous products begin to appear from the dissociation of target materials under high vacuum conditions and in isolation are found to closely agree with the vapor pressure criteria of the specific material in isolation; therefore, the limiting temperature of a particular target material can also be determined by computing thermal equilibrium composition versus temperature of the particular material under the high vacuum conditions that must exist during operation. Since materials that make up a target must be in contact with each other, thermal dissociation and chemical reactions between dissimilar compounds or elemental materials can alter the limiting temperature of the target through formation of volatile compounds. In cases where the target material and the matrix chemically react at elevated temperatures to form volatile compounds, the matrix must be pre-coated with a protective material such as W, Ir or Re to prevent

9

The UNISOR is an ISOL facility at the HRIBF that is primarily utilized for evaluating the release properties of candidate production targets.

10

NY.

Fibrous oxides are available from Zircar Products, Florida,

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

the undesired reaction process. Re is particularly appealing for this application because of its low adsorption-enthalpy properties for electronegative elements/molecules. However, according to thermo-chemical calculations as well as practical experience, W is a less-expensive solution to the problem for many target materials. 3.5.1. Oxides A number of oxides materials can be procured in fibrous forms with dimensions commensurate with fast diffusion release of relatively short halflife isotopes (see foot note 10). For oxides that are not available in fiber form, the compound materials can also be deposited in thin layers onto highly permeable substrates using a recently developed paint technique [22]. In cases where interactions between the target material and the substrate material, or the material of which the reservoir is constructed, reduce the limiting temperature to unacceptably low values, protective coatings must also be used that prevent contact between the reactants. Ta is the material customarily used to fabricate target material reservoirs and vapor transport systems for the ISOL production of short-lived radioactive species. For example, Ta reduces the limiting temperature of BeO or Al2O3 in contact with the surfaces of the material, in which case W is as effective and an inexpensive alternate to the choice of a noble refractory material such as Re. (BeO or Al2O3 are also customarily utilized as insulators for electrical isolation in high-temperature ion sources used to generate RIBs at such facilities and therefore, should not be used in direct contact with Ta.) Therefore, careful account must be taken of the high-temperature chemistry that can take place between compound and elemental materials that make up a particular target. Of course, oxides cannot be deposited onto carbon matrices as they will be converted to CO and CO2 and consequently destroy the substrate and leave the metal or metal carbide as residuals. As illustrations of how limiting temperatures are determined by use of the thermal equilibrium composition technique, the following examples are given: Fig. 1 displays thermal equilibrium composition versus temperature for the refractory oxide

77

BeO: (a) in isolation; (b) in contact with Ta; (c) in contact with W, and (d) in contact with C. As noted, the limiting temperature for BeO in isolation is B2140 K or in the presence of W is B2200 K. However, if operated in contact with Ta, its limiting temperature decreases to B1923 K and the material ‘‘burns’’ when in contact with C where the limiting temperature is seen to be B1423 K. For the case of Al2O3, the chemical reduction process is also rather severe. For example, the limiting temperature for Al2O3 in isolation is B2173 K while when in contact with Ta it is lowered to B1900 K and in the presence of W is again B2173 K. However, when in contact with C, the limiting temperature drops to B1323 K. Thermal equilibrium composition results for these respective reactions are displayed in Fig. 2. Other more electropositive metal-oxides are less susceptible to reduction processes. Table 2 provides a list of metal-oxides and limiting temperatures in isolation and in contact with Ta, W and C. For example, the limiting temperature of Y2O3 in isolation is B2233 K and in contact with Ta is only lowered to B2198 K, whereas its limiting temperature in the presence of W is B2233 K. For ZrO2 and HfO2, the effect is also less pronounced. For example, ZrO2 has a limiting temperature of 2330 K in isolation; 2173 K in the presence of Ta; and B2373 K in the presence of W. While the corresponding limiting temperatures of HfO2 are, respectively: B2473 K in isolation; B2323 K in the presence of Ta; and B2523 K in the presence of W. Of course, the use of oxides in the presence of C should always be avoided, since the oxides will ‘‘burn’’ as noted in Table 2. 3.5.2. Carbides The carbides of several metals are quite refractory and good choices for deposition on lowdensity, highly permeable matrices such as 2  reticulated-vitreous-carbon foam (RVCF). Specifically, spallation targets for high-energy RIB facilities targets such as RIA can be quickly and inexpensively fabricated from refractory carbides using the paint coating technique [22]. A list of candidate carbide targets is provided in Table 3. However, when deposited on RVCF, in the absence of volatile compound formation

ARTICLE IN PRESS 78

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 1. Thermal equilibrium composition of BeO as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. Note, the limiting temperature of the compound is B2140 K in isolation but lowered to B1923 K when in contact with Ta. When in contact with W, the limiting temperature is B2200 K. However, when in contact with C, the limiting temperatures decreases to B1423 K.

between the target material and the matrix, the upper limit to which the composite target can be heated is dictated by the limiting temperature of the most volatile component. Since C has a limiting temperature of B2353 K, it is the limiting component in composite carbide targets that have components with limiting temperatures in excess of this value. In some cases, exchange reactions can occur that affect the limiting temperature of the composite target. This effect is displayed in Fig. 3 for SiC: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. While neither Ta nor W affects the limiting temperature of SiC, as noted, C lowers the limiting

temperature of the compound in isolation from 1923 to 1773 K. The effect can be eliminated by pre-coating the RVCF matrix with 1–2 mm of Ta or W. UC2 is frequently used for the fission production of neutron-rich isotopes of Cu through Dy. Fig. 4 displays thermal equilibrium composition data for this compound: (a) in isolation; (b) in the presence of Ta; (c) in the presence of W and; (d) in the presence of C. As noted, the limiting temperature of UC2 in isolation, in the presence of Ta and in the presence of W are identically 2398 K; however, when in contact with C, the limiting temperature of the composite is dictated by the most volatile component, C, since it has a lower

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

79

Fig. 2. Thermal equilibrium composition of Al2O3 as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. Note, the limiting temperature of the compound is B2173 K in isolation but lowered to B1900 K when in contact with Ta. When in contact with W, the limiting temperature is again B2173 K but when in contact with C, the limiting temperature drops to B1323 K.

limiting temperature, i.e., 2353 K. ThC2 can also be used to produce both proton-rich and neutronrich radio-nuclei and since it has a higher limiting temperature, it is a viable alternate option for producing such nuclei. Thermal equilibrium composition data versus temperature are displayed in Fig. 5 for this compound: (a) in isolation; (b) in contact with Ta; (c) in contact with W and; (d) in contact with C. As noted, the limiting temperature of ThC2 is B2575 K in isolation or in contact with either Ta or W. However, when deposited onto C, the limiting temperature of the most volatile component dictates the limiting temperature of the system, i.e., C (B2353 K). The limiting temperatures of a variety of refractory carbides

in isolation, on Ta; on W; and on C are listed in Table 3. 3.5.3. Sulfides Only a few sulfides are sufficiently refractory to merit consideration for use as ISOL target materials. In particular, CeS is a preferred candidate target material for production of 33 Cl(t1/2=2.51 s), 34Cl(t1/2=32 min), 29P(t1/2= 4.14 s) and 30P(t1/2=2.5 min) through the respec32 34 tive reactions S(d,n)33Cl, S(p,n)34Cl, 32 29 32 30 S(p,a) P, S(d,a) P for potential use in important astrophysics experiments. CeS has a limiting temperature in isolation; in contact with Ta, and in contact with W of 2173 K. However,

ARTICLE IN PRESS 80

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Table 2 Limiting temperatures of various refractory oxides Metal-oxide

Limiting temperature (K) (in isolation)

Limiting temperature (K) (on Ta)

Limiting temperature (K) (on W)

Limiting temperature (K) (on C)

BeO Al2O3 CaO Y2O3 ZrO2 HfO2 Ce2O3 Pr2O3 Nd2O3 Sm2O3 Eu2O3 Gd2O3 Tb2O3 Dy2O3 Ho2O3 Er2O3 Tm2O3 Yb2O3 Lu2O3

2140 2173 1983 2233 2330 2473 2133 2073 2098 2083 2063 2453 2273 2373 2393 2423 2273 2133 2373

1923 1900 1703 2198 2173 2323 1983 1923 1948 1838 1948 2198 2138 2153 2138 2183 2013 1833 2213

2200 2173 1953 2233 2373 2523 2103 2073 2098 2098 2033 2473 2273 2378 2363 2423 2398 2133 2363

1423 1323 1183 1643 1653 1723 1598 1533 1483 1223 1148 1588 1598 1408 1433 1513 1438 — 1638

when in contact with C, it has a limiting temperature of B2073 K, as noted in Fig. 6. Thus, this compound cannot be deposited directly onto C without lowering the limiting temperature by 100 K (Fig. 6d). To protect against this occurrence, the carbon matrix can be coated with either Ta or W. ThS is also very refractory and has a limiting temperature of 2273 K in isolation, in contact with Ta, or in contact with W, as shown in Fig. 7. However, this compound cannot be deposited directly on C without consequences where the limiting temperature is 2173 K (see, Fig. 7d). Since Th is a fissionable material, this aspect of the element must be considered before choosing ThS for production of isotopes from S. However, since neither CeS nor ThS can be deposited directly onto C matrices such as RVCF without consequences, a thin layer of Ta or W can be deposited onto the RVCF before deposition of the respective compound to eliminate this problem.

4. Target design and fabrication In order to effectively solve target-related issues that loom as obstacles toward success at next-

generation RIB facilities, it is necessary to understand the underlying processes that ultimately limit the number of short-lived species that can be delivered with intensities adequate for viable research with these beams. Targets with the short diffusion lengths, high-permeability properties and controllable temperatures required for meeting the RIB intensity needs for a wide-range of species for nuclear and astrophysics research at ISOL-based facilities must be designed that can withstand irradiation with very high power primary beams. The new target technologies developed at the HRIBF include short diffusion length, low-density fibrous compound materials [10], low-density metal and ceramic foams [12], as well as composites targets formed by coating the surfaces of highly permeable matrices with the material of interest [11,12,22]. The target concepts advocated in this article, include the use of low-density target matrices, such as carbon-bonded-carbon-fiber (CBCF)11 or RVCF,12 to serve as the plating 11 Carbon-bonded carbon is fabricated at the Metals and Ceramics Division of the Oak Ridge National Laboratory. 12 Reticulated vitreous carbon is marketed by ERG, Inc., Oakland, CA.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

81

Table 3 Limiting temperatures of various refractory carbides Target material

Limiting temperature (K) (in isolation)

Limiting temperature (K) (on C)

Limiting temperature (K) (on Ta)

Limiting temperature (K) (on W)

C SiC TiC VC MnC2 YC2 ZrC MoC NbC LaC2 HfC TaC WC

2353 1923 2323 2203 2998 2273 2673 2448 2753 2373 2733 2893 2913

2353 1773 2273 2073 2356 2353 2353 2356 2373 2353 2353 2573 2573

2536 1923 2323 2183 2873 2353 2673 2448 2740 2373 2713 2793 2723

2543 1923 2333 2178 2973 2353 2673 2448 2773 2373 2733 2903 2913

matrices for the target material itself as well as a thermal conduit for redistribution of primary beam deposited heat to the surrounding medium. Since short-lived particles must swiftly diffuse from the target material, the diffusion lengths must be short (thin target materials) and the target temperature must be as high as practical. The diffusion process generally takes longer to effect than any other process and therefore is the principal mechanism that limits the intensities of short-lived species. Therefore, it is important to understand the process in order to be able to minimize delay times attributable to this mechanism. A brief description of the diffusion process is given below. (For more details on the subject, see, e.g., Refs. [23,24].) 4.1. Diffusion theory Nuclear reaction product species, embedded in a chemically dissimilar target material, are released from the material through a binary diffusion mechanism. Impurity atoms or vacancies will move through the solid until equilibrium is reached. The rate at which the diffusion process takes place depends exponentially on the temperature for solid diffusion couples. The net flux, J, of either atoms or vacancies is related to the concentration gradient, rn, by Fick’s first equa-

tion given by J ¼
Drn

ð1Þ

where D is the diffusion coefficient. Knowledge of D is required to estimate the release rates of particular species. In general, D must be measured by techniques such as described in Refs. [23,24] or by ion implantation methods, developed for this purpose [25–27]. Fortunately, large sets of data for certain diffusion couples are available from resources such as cited in Ref. [28] and within the data-bases associated with computer codes such as DICTRA.13 The ion implantation technique, employed at the HRIBF, is particularly useful for measuring diffusion coefficients for candidate species/target-material combinations [26,27]. The time-dependent form of Eq. (1) is known as Fick’s second equation. The form of this equation, which allows for creation of particles S(t) as well as loss of particles E(t), can be expressed as follows: @n=@t ¼ Dr2 n þ S
E

ð2Þ

where D is assumed to be independent of concentration. Small dimension (micron-scale) 13 DICTRA is a general software package for simulation of DIffusion Controlled TRAnsformations in multi-component systems code developed by the Royal Institute of Technology, Stockholm, Sweden.

ARTICLE IN PRESS 82

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 3. Thermal equilibrium composition of SiC as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. Note, the limiting temperature of the compound is B1923 K in isolation but lowered to B1773 K when in contact with C. However, when in contact with Ta or W, the limiting temperature is again B1923 K.

planar and cylindrical geometry targets are routinely used at the HRIBF. Only small dimension spherical geometry target materials that do not sinter are useful at such facilities, since sintering changes particle sizes rather quickly and consequently increases the diffusion time of short-lived species. A few materials such as SiC and BN do not sinter. Solutions to the respective timedependent form of Eq. (2), appropriate for the particular target-material geometry, can be found either by separation of variables, the use of Laplace or Fourier transformation techniques, or by standard numerical computational techniques or in combination. We assume that the crosssection is independent of production beam energy resulting in a uniform distribution of species in the path of the beam. For the more general case, the distribution function S must be chosen to

represent the actual distribution of the radioactive species within the target material. For a uniform distribution of particles such as assumed in the production of radioactive species with a primary ion beam of intensity, I, particles per second, the production rate density S is given by S ¼ nT ILs=V :

ð3Þ

In Eq. (3), nT is the number of target nuclei per unit volume; L is the length of the target material; s is the cross-section for production of the species of interest; and V is the volume of the beam-target overlap region. For production of radioactive species with halflife t1/2, E is given by E ¼ nl ¼ 0:693n=t1=2

ð4Þ

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

83

Fig. 4. Thermal equilibrium composition of UC2 as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. As noted, the limiting temperature of UC2 is B2398 K in isolation or in contact with either Ta or W. However, when deposited onto C, the limiting temperature of the most volatile component dictates the limiting temperature of the system, i.e., C (B2353 K).

where n is the number density of radio-nuclei in the target material. Solid-state diffusion. For solid-state diffusion, the diffusion coefficient D is dependent on the activation energy HA required to move the atoms or vacancies from site to site, and on the temperature T of the solid, according to D ¼ D0 expð
HA =kTÞ

ð5Þ

where k is Boltzmann’s constant and D0 is a constant. D is related to the vibration frequency and lattice parameters of the particular atom/ crystal couple. Typical values for this process range between 10–16 m2/s and 10
12 m2/s for target temperatures in the range of 1873–2073 K (see, e.g., Ref. [29]). HA can be extricated from experimental data by measuring the target

temperature dependence of D on T with methods such as described in Refs. [23,24] or by the use of ion implantation [25–27]. Examples of the solution to the diffusion equation for the release of neutron-rich 132Sn and 140,141,142,143,144Xe from thin layers of UC2 target material are shown, respectively, in Figs. 8 and 9. Liquid-state diffusion. In contrast to the case of diffusion within a solid target, the diffusion coefficient for particles in liquid targets is weakly dependent on the temperature according to pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi DEad 8k=pM T 3=2 ð6Þ where a is the coefficient of thermal expansion of the liquid target material; d is the diameter of the solute atom of mass, M. D has typical values of a few times 10
8–10
9 m2/s [29]. Because of the fact

ARTICLE IN PRESS 84

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 5. Thermal equilibrium composition of ThC2 as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. As noted, the limiting temperature of ThC2 is B2575 K in isolation or in contact with either Ta or W. However, when deposited onto C, the limiting temperature of the most volatile component dictates the limiting temperature of the system, i.e., C (B2353 K).

that the process is only weakly dependent on the temperature T, there is no particular gain achieved by heating the material significantly above its melting point other than through reducing surface desorption times as will be discussed later. As noted, diffusion coefficients for liquids are several orders of magnitude larger than those for their solid-state counterparts and therefore are especially attractive for short-lived species; unfortunately, few elemental metals have the vapor pressure characteristics commensurate with these applications. To take advantage of the fast diffusion properties of liquid-state targets, eutectic alloys can be employed as a means of reducing the melting points of refractory metals.

4.1.1. Target dimensional design criteria The form of the diffusion equation (Eq. (2)) can be solved for the appropriate target geometry (planar, cylindrical, or spherical) to derive simple, analytic expressions that relate the dimensions of the target material with the diffusion time, t, that will release B70% of the species of interest within its lifetime, provided that the diffusion coefficient is known at the operating temperature of the target material. Expressions for optimizing the target thickness of planar, cylindrical and spherical geometry targets are given below: A plate of thickness, x: xðcmÞ ¼ pfDtg1=2 :

ð7Þ

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

85

Fig. 6. Thermal equilibrium composition of CeS as a function of temperature: (b) in isolation; (c) in contact with Ta; (a) in contact with W; and (d) in contact with C. Note, the limiting temperature of the compound is B2173 K in isolation; in contact with Ta; and in contact with W. However, when in contact with C, the limiting temperature is lowered to B2073 K.

A rod of diameter, dr: dr ðcmÞD4:8fDtg1=2 :

ð8Þ

A sphere of diameter, ds: ds ðcmÞ ¼ 2pfDtg1=2 :

ð9Þ

Thus, by engineering target-material dimensions according to these basic relations, one can optimize the release rate of a particular radioactive species. These simple dimensional criteria permit custom engineering of the thickness of deposits of the production material that will ensure the diffusion release of B70% of the isotope of interest within its lifetime. Selected examples of target thicknesses required to release B70% species within their lifetimes are listed in Table 4. These criteria have been incorporated and experi-

mentally demonstrated on-line in important nuclear physics [3,4] and nuclear astrophysics [5–9] research for a number of highly successful, fast release fibrous [10] and highly permeable, composite production targets [11,12] at the HRIBF [1,2] and therefore, will be used as fundamental criteria for design of all targets considered in this report. 4.2. Target matrices Since refractory fibrous materials suitable for production applications are only available for a few materials, low-density, low-atomic number highly permeable matrices for coating with the appropriate production target material are desirable for deposition of the material of interest with thicknesses determined by the prescriptions

ARTICLE IN PRESS 86

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 7. Thermal equilibrium composition of ThS as a function of temperature: (a) in isolation; (b) in contact with Ta; (c) in contact with W; and (d) in contact with C. Note, the limiting temperature of the compound is B2273 K in isolation and when in contact with either Ta or W. However, it is lowered to B2173 K when in contact with C.

dictated by use of Eqs. (7), (8) or (9). Furthermore, it is desirable that the matrix have good thermal conductivity attributes so that the beam-deposited heat can be removed or redistributed at a controlled rate with appropriate thermal management techniques so that the target matrix can be operated at the maximum allowable primary beam intensity as dictated by the temperature limitation of the composite target. (CBCF) (see footnote 11) and RVCF (see footnote 12) offer machinable, highly permeable, matrices for deposition of generic target materials onto their surfaces. CBCF (rB10%r0) is made of cylindrical fibers (B6 mm diameter) sintered together at the points of intersection. Such high-permeability structure matrices require penetrating coating techniques.

RVCF has the physical integrity, fiber size and open structure for use as a matrix for deposition of thin layers of production target materials and presently is the preferred matrix material for fabricating composite production targets at the HRIBF. RVCF can be procured in the following forms: RVCF (uncompressed); 2  RVCF (compressed in one direction by a factor of 2); 4  RVCF (compressed by a factor of 2 in each of two directions); 6  RVCF (compressed by a factor of 2 in all three directions). (Compressed forms are desirable for applications where larger surface-to-volume ratios are required.) Uncompressed RVCF (rB2%r0) with 3.94 pores/mm is a continuous ligament tetrakai-decahedral open-structure with six square and eight hexagonal faces connected with equilateral triangular

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Release Rate (Particles/s)

3.0E+09 2.5E+09 2.0E+09

1 GeV 200 MeV

1.5E+09

60 MeV 1.0E+09 5.0E+08 1.0E+06 0

20

40

60

80

100

Time (s)

Fig. 8. Diffusion release of 132Sn from 10 mm thick, UC2/ 2  RVCF. D: 8  10
12 m2/s; T: 2273 K; proton beam energy: 60 MeV; 200 MeV and 1 GeV; beam intensity for each energy: 1 mA; average respective production cross-section: 8 mb (60 MeV); 4 mb (200 MeV); 2 mb (1 GeV); respective target length: 0.026 m (60 MeV); 0.305 m (200 MeV); 1 m (1 GeV).

Release Rate (Particles/s)

1.E+09 140

Xe

141

Xe

1.E+08 1.E+07 142 Xe

1.E+06 144

Xe

143

Xe

1.E+05 1.E+04 1.E+03 0

4

8

12

16

20

Time (s)

Fig. 9. Diffusion release of 140,141,142,143,144Xe from 10 mm thick, UC2/2  RVCF. D: 5.6  10
13 m2/s; T: 2373 K; proton energy: 1 GeV; beam intensity: 1 mA; target length: 1 m.

87

cross-section ligaments, base length: B74.5 mm (equivalent radius; 27.7 mm); the average leg length for the uncompressed material is 252.9 mm; and the equivalent radii of a unit cell is rD354 mm. The surface-to-volume ratio is an important aspect of this material since the surfaces of the ligaments are coated with the material of interest. For example, the respective surface-to-volume ratios are: 1  RVCF: B32.2 cm2/cm3; 2  RVCF: B64.4 cm2/cm3; 4  RVCF: B128.8 cm2/cm3; 6  RVCF: B193.2 cm2/cm3. These materials can be machined to the geometry desired for the particular application prior to depositing the specified thickness of target material onto the surface. Heretofore, 2  RVCF has been used as the preferred matrix for target fabrication at the HRIBF. This material is compressed by a factor of 2 in the z-direction and has a surface-to-volume ratio of 64.4 cm2/cm3 and therefore, is quite adequate for use with the relatively low production beam energies (p100 MeV) available at the facility where target lengths between 25 and 50 mm are typically required. However, at next-generation RIB facilities, such as at the RIA, where proton beam energies will be B1 GeV, targets made from this material may be 65–80 cm in length. Target lengths for RIA energy proton beams can be significantly shortened by choice of highly permeable matrix material with a much higher surfaceto-volume ratio such as 6  RVCF. By using 6  RVCF, RIA targets can be reduced by more than a factor of two in length (29 versus 65 cm for a one interaction length UC2 target) and thus, are more practically sized. However, thermal analyses of targets that utilize 6  RVCF,

Table 4 Permissible target thicknesses Species 140

Xe Xe 142 Xe 143 Xe 144 Xe 132 Sn 132 Sn 141

t1/2 (s) 13.6 1.73 1.22 0.3 1.15 40 40

Target material UC2 UC2 UC2 UC2 UC2 UC2 Thliq

Temp. ( K) 2273 2273 2273 2273 2273 2273 2023

D (m2/s)
13

5.6  10 5.6  10
13 5.6  10
13 5.6  10
13 5.6  10
13 8  10
12 1  10
8

x (mm)

dr (mm)

ds (mm)

8.66 3.09 2.6 1.29 2.52 56 1986.9

13.25 4.72 3.97 1.97 3.85 85.87 3036

17.32 6.18 5.2 2.58 5.04 112 3973.8

Selected radioactive species diffused from: plates of thickness x; rods of diameter dr, spheres of diameter ds.

ARTICLE IN PRESS 88

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

subjected to irradiation with 1 GeV, 100–400 kW proton beams, have not been performed to date and must be made in order to estimate the affect of primary beam deposited heat on the temperature distribution within such targets. Scanning Electron Micrographs (SEMs) of 2  RVCF and 6  RCVF are displayed, respectively, in Figs. 10a and b. 4.3. Target fabrication The challenge of the target design specialist is to fabricate robust short diffusion release, lowdensity compound fiber, metal foam, ceramic foam or low-density composite targets that can withstand irradiation with 1 GeV proton beams at power levels up to 400 kW as required for generating useful RIB intensities of a broad range of short-lived species for nuclear physics and nuclear astrophysics research at next-generation RIB facilities such as the Rare Isotope Accelerator (RIA). In order to achieve fast diffusion release and effusive transport to the ion source, the thicknesses of the target material must meet the design criteria specified in Eqs. (7)–(9) and form highly permeable integral structures. In this section, methods for fabrication of fibrous compound and composite targets are described.

4.3.1. Fibrous targets In a few cases, fibrous materials with small diameters are commercially available, (see footnote 10) including Al2O3, SiO2, Sc2O3, TiO2, Y2O3, ZrO2, HfO2, Ta2O5 and rare-earth oxides. All of these candidate target materials have thin fiber diameters and are highly permeable as required for high release efficiency. Fig. 11a–c, respectively, display scanning electron micrographs (SEMs) of small diameter Al2O3, ZrO2 and HfO2 fibrous materials that have been successfully used to produce and efficiently release 17,18 F for use in the HRIBF nuclear physics and astrophysics research programs [5–9]. These materials can also be used as spallation targets at next-generation high-energy, high-power RIB facilities as noted in Table 1. 4.3.2. Methods for coating matrices Although a number of methods have been developed for depositing films of production ISOL target materials onto surfaces with thickness commensurate with fast diffusion release of short-lived radioactive species for ISOL target applications, the methods are either inappropriate for coating interior surfaces of highly permeable complex structure matrices (non-infiltrating, nonpenetrating) or require complex chemical processes that are only available for depositing

Fig. 10. Highly permeable RVCF target matrices for depositing candidate production target materials. (a) 2  RVCF; pores/mm: 3.94; density: rD 4.44% r0; surface area/volume ratio: 64 cm2/cm3. (b) 6  RVCF; pores/mm: 3.94; density: r D 17.8% r0; surface area/ volume ratio: 193.2 cm2/cm3.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 11. SEMs of (a) Al2O3; (b) ZrO2; and (c) HfO2 fibrous materials used at the HRIBF to produce intense nuclear physics and nuclear astrophysics programs [4–9].

specific elemental or compound materials. Techniques are presently available that can be used to uniformly deposit specified thicknesses of the material in question onto the support matrix of choice. At this point in time, several target coating schemes are being used or are under consideration for use for this purpose, they include: chemical vapor deposition (CVD); chemical vapor infiltration (CVI); chemical reaction deposition (CRD); physical vapor deposition (PVD); electrolytic deposition (ED), electrophoresis deposition (EPD) and sol–gel coating (SGC) and a newly conceived paint coating (PC) method [22]. The high-permeability requirement aspect of fast release/transport targets necessitates the coating of internal surfaces of complex structures such as CBCF and RVCF and therefore, the method used for depositing the material must be penetrating. Since PVD is nonpenetrating, it is limited to line-of-sight coating of surfaces. The CVD/CVI method can be used to uniformly coat thin, complex structures with refractory metals and a few compounds for which precursors have been developed. However, the method is slow and expensive. The sol–gel method is also penetrating but slow and also relies on the development of chemical precursors to implement the coating process. Analogously, the CRD wet chemistry method is penetrating and can be used to coat complex structures such as RVCF with compounds such as UC2 [11]. However, the method relies on sequential complex chemical

89

17,18

F beams for the

reactions and therefore is slow and expensive and is also relegated to compound specific applications where the chemical deposition reactions and processing procedures have been established. Despite these handicaps, the CRD method, however, is a viable way to uniformly coat UC2 onto RVCF matrices to form fast effusive transport, fast diffusion release, thermally and mechanically stable production targets for neutron-rich isotopes of elements Cu through Dy. To date, these targets have produced shortlived species with intensities ranging between B1  104/s and B1  109/s from >130 isotopes of 28 elements for use at the HRIBF (see, e.g., Refs. [2,30]). A new, inexpensive, fast, reproducible, penetrating and close to universal alternative choice for deposition of custom-engineered thicknesses of target materials for spallation and fission production reactions was recently conceived. The method is based on suspension of finely divided (f: B1 mm) target material in a binder to form a paint that can be vacuum infiltrated to uniformly coat the surfaces of highly permeable, complex structure, fibrous matrices such as RVCF with many compound materials (e.g., oxides, carbides, and sulfides, etc.) and thereby form mechanically and thermally stable production targets. The coating operation is followed by thermal treatment under high vacuum conditions to drive off binding material and to sinter the material. For more details on the method, see Ref. [22] of these proceedings.

ARTICLE IN PRESS 90

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 12. (a) SEM of a carbon-bonded-carbon fiber (CBCF) matrix coated by chemical vapor infiltration with SiC; (b) SEM of RVCF matrix coated with Ta. (c) SEM of (RVCF) matrix CVD coated with W. As noted, these composite targets are highly permeable and designed in thickness to efficiently release a variety of short-lived, proton-rich radioactive species.

4.3.3. Composite targets The CVD/CVI process can be used to fabricate thin, uniform deposits of a variety of refractory metals as well as a limited number of compounds for potential use for the spallation production of proton-rich nuclei with high-energy proton beams, as suggested in Table 1. For example, RVCF matrices can be CVD/CVI (chemical-vapor-deposition/chemical-vapor-infiltration) coated with refractory metals, such as SiC(1923 K); Zr(2307 K); Nb(2618 K); Mo(2428 K); Tc(2443 K); Ru(2318 K); Rh(2013 K); Hf(2313 K); Ta(2918 K); W(3108 K); Re(2903 K); Os(2813 K); Ir(2443 K); Pt(2068 K), to form composite targets for use at such facilities. As examples of refractory composite-target fabrication using the CVD/CVI technique for potential use as spallation production targets for nextgeneration RIB facilities, SEMs of SiC, Ta and W, CVI deposited onto RVCF are displayed in Figs. 12a–c. As noted in Table 3, several carbides have the refractory character required for high-power target applications. SEMs of UC2 coated onto 2  RVCF by the CRD method and NbC and HfC coated onto 2  RVCF with the paint techniques are shown, respectively, in Figs. 13a–c. The CRD process is routinely used for preparing fission targets for production of neutron-rich isotopes for use at the HRIBF [11]. The new paint technique [22] offers a faster, more universal and less-expensive alternative than either the CVD/CVI or CRD processes for the infiltration coating of RVCF with

refractory metal-carbides, metal-oxides and metalsulfides of chemically active elements. As discussed previously, CeS is a preferred target material for production of 33Cl(t1/2=2.51 s), 34Cl(t1/2=32 min), 29 P(t1/2=4.14 s) and 30P(t1/2=2.5 min) through the respective reactions 32S(d,n)33Cl, 34S(p,n)34Cl, 32 S(p,a)29P, 32S(d,a)30P for potential use in important astrophysics experiments. SEMs of ZrO2/ W/2  RVCF, HfO2/W/2  RVCF and CeS/W/ 2  RVCF (B5 mm thick), prepared by the paint technique, are displayed, respectively, in Figs. 14a– c. (The W coating is used to prevent reactions between ZrO2 or HfO2 or CeS and their 2  RVCF matrices, as discussed previously.) (The release properties of any candidate target must be evaluated on-line before determining its suitability for the intended application.)

5. Thermal analyses of candidate target designs Since RIB intensities, produced through spallation and fission production reactions at highenergy ISOL facilities, increase with beam power, techniques must be developed for assessing the maximum beam power handling capabilities of prototype targets, prior to their fabrication. Thermal modeling techniques offer cost-effective means for evaluating the heat and temperature distributions imparted to targets during irradiation, and for assessing the merits of beam manipulation techniques for more uniformly

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

91

Fig. 13. SEM of carbide compounds coated onto RVCF matrices: (a) B10 mm of UC2 coated with the CRD method; (b) B6 mm of NbC coated with the paint technique; (c) B6 mm HfC coated with the paint technique.

Fig. 14. SEM of oxide and sulfide compounds deposited onto W/2  RVCF matrices with the paint technique: (a) ZrO2 coated with the paint technique; (b) HfO2 coated with the paint technique; (c) B5 mm CeS deposited onto 2  RVCF. The 2  RVCF matrices were pre-coated with B2 mm of W to prevent lowering of the respective limiting temperatures of the oxide and sulfide deposits.

distributing primary beam deposited heat over the volumes of targets at the highest beam power available without exceeding the limiting temperature of the specific target material. The previously described highly permeable targets are necessarily fragile and yet must withstand irradiation with 1 GeV, 400 kW proton beams at next-generation ISOL RIB facilities. Although such low-density fibrous materials generally have poor thermal conductivities, the beam power is deposited over a larger absorbing volume. Also, fibrous materials have high surfacearea/volume ratios and consequently, open structures that permit internal radiation into open channels that allow photon transport over the entire volume of the target and consequently a redistribution of beam-deposited heat over a substantially larger target volume than otherwise would be the case. Results derived from computa-

tional simulations, described in this article, reconfirm previous predictions and computational validation studies [31–33] that internal radiation is an effective means for redistributing primary beam power over extended regions of the target volume. Consequently, this mechanism permits heat transfer within low-density targets with extremely poor intrinsic thermal conductivities. Several ploys have been investigated for reducing the beam power density and homogenizing the temperature distribution over the entire volume of targets with the objective of arriving at designs that accommodate irradiation with primary beam power levels up to 400 kW, as required at the RIA, for example. The radiation transfer effect, in combination with the employment of primary beam manipulation techniques (focus and scanning) and placement of additional

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

radiation shielding on the exit end of targets, are found to be effective ways for reducing beam power deposition density to manageable levels and homogenizing the temperature distributions within practically sized targets to levels commensurate with their irradiation with 1 GeV, 400 kW proton beams, as required for use at the RIA [18,19]. Another, perhaps, less-practical technique that can be used for this purpose is to linearly increase the target material density along the length of the target, as described in Refs. [20,21]. However, since the target material density adversely affects the speed at which the diffusion and effusive-flow processes can take place, the method may not be practically useful. The Monte Carlo simulation code, GEANT4, (see footnote 7) was used to compute the heat deposited in targets by primary and secondary beams and in the walls of target material reservoirs and heat shielding surrounding the target by primary beams and by energetic secondary and fission fragments emitted during primary beam– target interactions. The power density information is then used as input information to the finite element thermal analysis code, ANSYS, (see footnote 8) for extracting temperature distributions within a given target assembly. All simulation studies were performed with ‘‘interaction length’’ targets. 5.1. Internal thermal radiation effects Low-density, small-diameter refractory materials have proved to be efficient and fast diffusion release/effusion-flow production target materials [2,10–13,30]. However, it is difficult to dissipate intense heat from the interior regions of lowdensity targets if only normal thermal conduction is considered, even for materials with the highest intrinsic bulk thermal conductivities. Since lowdensity, small-dimensioned fibrous targets required for fast diffusion release and fast effusiveflow transport have open channels over the entire target volume that provide partially optically transparent paths for photons to travel, thermal radiation can take place within the interior regions of such materials [31–33]. This heat-transfer mechanism is an effective means for redistributing

primary beam-deposited heat to other regions of the target, thereby permitting the use of higher power primary beams. 5.1.1. Thermal conductivities of fibrous materials Thermal conductivity data for low-density materials are not generally available and therefore, must usually be measured. However, a few sets of data are available from manufacturers of specific materials, such as fibrous oxides (see footnote 10) or from measurements [34,35]. The forms of a few sets of thermal conductivity data are displayed in Fig. 15. We note that the measured thermal conductivities increase with increasing temperature. This behavior is in complete contrast to the thermal conductivities of bulk materials that decrease with increasing temperature for most materials. These data can be represented by fitting to experimental data with the following expression: sDA þ BT 4 ¼ sF þ sR

ð10Þ

where s is measured the thermal conductivity; where sF=A is the thermal conductivity of the fiber and sR=BT4 is the contribution to the thermal conductivity by radiation between fibers within the fibrous material; and T is the absolute temperature of the fibrous target material. As noted, the thermal conductivity is composed of two terms. The first term is associated with the thermal conductivity of the fibrous material while the second term is due to thermal radiation.

0.9 Therm. Cond. (W/m.K)

92

Fit

0.6 Meas.

ZrO2 RVCF CBCF Ni-16um Al2O 3 ZrO2 RVCF CBCF Ni-16um Al2O 3

0.3

0.0 400

800

1200 T (K)

1600

2000

Fig. 15. Measured thermal conductivities of fibrous Al2O3, ZrO2, CBCF, 1  RVCF, 2  RVCF and Ni materials.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Therefore, it is obvious that thermal radiation effects take place within such open structures. This is an important mechanism for heat redistribution, especially at the elevated temperatures required for fast diffusion release from ISOL targets where thermal radiation dominates the heat transport process. Simulation studies of low-density materials that do not include the proper thermal conductivity behavior with temperature are obviously incorrect. In the absence of thermal conductivity data for fibrous materials, we take as the first term in Eq. (10), the thermal conductivity of the material, sF, computed from the bulk thermal conductivity, sBTC, of the fibrous material scaled by the ratio of the density of the fibrous material, rfiber, divided by the density of the bulk material, rbulk or A=sF(T)=C1[rfiber/rbulk]sBTC(T), where C1 is a constant. We take as the second term, B= C2esSBAS/rDT, where e is the emissivity of the material; sSB is Stephan–Boltzmann’s constant; C2 is a constant; DT=T
T0, As is the area of the radiating surface; and Dr is the average distance from the radiator to the absorbing surface. In the absence of thermal conductivity data, we formulate an analytical complement of Eq. (10) given by sD C1 ½rfiber =rbulk sBTC ðT Þ þ sRad ðT Þ ¼ C1 ½rfiber =rbulk sBTC ðT Þ þ C2 esSB As fT 4
T04 g=DrDT

ð11Þ

where T0 is the temperature of the surrounding surfaces to which the fibers radiate. 5.1.2. Thermal radiation modeling studies To simulate the contributions to heat transport by thermal radiation within the interior regions of low-density fibrous target materials, two simple 2D models were used to illustrate the effect. More accurate 3D simulations could not be effected with the computers available at the time of these studies because of the prohibitively large memory and excessive CPU time required to solve a particular problem. The maximum target temperature that a particular target material can tolerate, as computed with the 2D radiation models, is determined by the beam power, its distribution, the intrinsic thermal conductivity of the material and thermal

93

radiation between the elements of the target and the surrounding material and the walls of the reservoir. Schematic representations of the target models used to illustrate the importance of internal thermal radiation effects are displayed in Fig. 16. Forms of Eq. (11) were used to input thermal conductivity data for the respective target materials. The maximum target temperature that a particular target material can tolerate, as computed with the 2D radiation models, is determined by the beam power, its distribution, the intrinsic thermal conductivity of the material and thermal radiation between the elements of the target and the surrounding material and the walls of the reservoir. Table 5 lists deposited beam power, dP, and maximum acceptable beam current, IMAX, for a few single column, low-density, fibrous targets (dimensions: 0.04 m diameter; 0.80 cm (interaction length)) irradiated with 1 GeV proton beams, with and without internal thermal radiation effects present. As noted, according to these simulations, the maximum intensity required to raise the local temperatures to peak values X2273 K exceed 100 mA for targets with good thermal conductivity attributes and 70 mA for materials with poor intrinsic thermal conductivities such as UC2, ZrO2 and HfO2. These results further demonstrate that internal thermal radiation is the dominant means for dissipating heat within low-density fibrous target materials, at high temperatures. Thus, the use of highly permeable, low-density fibrous or composite RVCF matrix targets materials with ‘‘open’’ structures enhances internal radiation cooling. As an illustration of the effect

Fig. 16. Two-dimensional models used to illustrate the importance of internal radiation in redistributing primary beam deposited power within fibrous targets.

ARTICLE IN PRESS 94

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Table 5 Illustration of internal thermal radiation effects using 2D models beam energy: 1 GeV; target diameter: 0.04 m; target length: 0.80 m Target material

Ta/2  RVCF NbC/2  RVCF BeO/W/2  RVCF UC2/2  RVCF ZrO2 (fibers) HfO2 (fibers)

Without internal radiation

With internal radiation

IMAX (mA)

dP (kW)

IMAX (mA)

dP (kW)

86 64 66 31 9 7

14.2 11.7 9.9 6.2 1.3 1.2

245 120 100 70 82 80

40.5 21.6 14.5 14.0 11.8 13.4

of internal thermal radiation, a low-density, high thermal conductivity Ta/2  RVCF composite target (dimensions: 0.04 m diameter; 0.80 cm (interaction length)), without internal thermal radiation, can only withstand irradiation with 1 GeV, 86 mA proton beams. However, when internal radiation effects are included, targets made of this material can withstand irradiation with 245 mA proton beams of the same energy. Analogously, the power limits of targets, of the same size, made of poor thermal conductivity materials, such as ZrO2 and HfO2, without internal thermal radiation, are, respectively, 9 and 7 mA and 82 and 80 mA when thermal radiation effects are included. The magnitude of the internal heat deposition and consequently the temperature distribution is also affected by energetic nuclear reaction products such as fission fragments released from actinide targets e.g., UC2/RVCF (average energy: 85 MeV/ particle). Even though UC2/RVCF has a higher thermal conductivity than either ZrO2 or HfO2, this composite target can only accept irradiation with primary beam power levels up to 14 kW due to its low thermal conductivity and more importantly, the deposition of energy from fission fragments, as noted in Table 5. 5.2. CW irradiation of single-column targets with convergent incidence primary beams During target irradiation, a large fraction of primary beam particles may be lost during transit through long targets due to scattering processes with atoms for parallel, divergent or convergent

beams with less than optimum convergence angles. The use of converging incidence beams with optimal convergence angles permits full transmission of the primary beam through low-density, interaction length long targets with minimum scattering losses. The thickness, density and atomic number of the target material affect the number of nuclear reactions and angular distribution of scattered particles, and therefore, the size and convergence angle of the primary beam must be altered from target-to-target to minimize scattering losses and to ensure better power density distribution over the volumes of production targets. The beam power depositional density is lower at the exit (rear) ends of targets, due to nuclear reactions, beam scattering collisions and consequently, the temperature distribution will strongly vary along a given target. Even without dispersive scattering effects, changes in temperature along the target will necessarily occur due to loss of primary ion beam through nuclear reactions, since ISOL targets are nominally an interaction length long (the length of the target that will absorb 63% of the primary beam during passage). These processes cause serious temperature inhomogeneities that, in turn, can cause serious diffusion release and effusive-flow transport problems for short-lived species since these processes depend exponentially on target temperature. Therefore, techniques must be developed to ensure homogeneous temperature distributions over the entire volume of production targets. Several ploys have been investigated for reducing the beam power depositional density and

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

homogenizing the temperature distributions over the volumes of production targets. The optimal use of convergent beams, additional layers of heat shielding on the exit ends of the targets, in combination with the beam manipulation techniques (to be discussed later), can be quite effective in achieving more homogeneous temperature distributions within such targets. The proper placement of additional layers of heat shielding on the exit end of the target offers a simple and effective alternative method for homogenizing temperature distributions within such targets. Due to differences in scattering from target-totarget, the number of heat-shielding layers, and their positions will also vary from target-to-target. Scattering losses can be essentially eliminated in 0.04 m diameter, 0.8 m long targets by utilizing converging 1 GeV proton beams with angles chosen so that the total incident beam is transmitted through the target. An isometric drawing of a single-column target, irradiated with either parallel or convergent primary beams used in these simulation studies, is displayed in Fig. 17. All simulations were made for 0.04 m diameter, 0.8 m long targets, irradiated with 1 GeV, CW proton beams. The lengths of targets, unless otherwise specified, were determined by calculating the amount of target material required to reduce the beam to 1/e of its initial value through nuclear reactions. These targets are said to be interaction length long. Of course, for a fixed target length of 0.8 m, the density of the target material required to absorb 63% of the beam will vary according to the nuclear reaction cross-section for 1 GeV protons interacting with the particular target material. 5.2.1. Effects of additional radiation shielding Figs. 18 and 19, respectively, show heat distributions deposited by parallel incident, 1 GeV proton beams, in 0.04 m diameter, 0.8 m long ZrO2 targets. The peak power depositional density depends on the size of the beam at entrance to the target, for parallel beams of different radii, incident on an NbC/2  RVCF target, as seen in Fig. 20. By focusing a large diameter beam at a convergence angle that optimizes transmission, beam losses due to scattering can be minimized

95

Fig. 17. Schematic isometric representation of a single-column ISOL production target designed for irradiation with CW proton beams.

Fig. 18. Deposited beam power by parallel-incidence, 1 GeV, 1 mA proton beam in a ZrO2 fiber target. Density: 1.0  103 kg/ m3; length: 0.80 m; incident beam spot diameter: 0.01 m.

through reasonably diameter size targets, while ameliorating primary beam power depositional density within targets. The optimum convergent angle depends on the density and effective atomic number of the target material as well as the diameter and length of the target and thus will vary from target to target. This technique permits full or close to full utilization of the primary beam for producing radioactive species), while reducing the primary beam depositional density and improving the heat distribution within a given target since the beam will be confined within the diameter of the target (see, e.g., Fig. 21). The computed temperature distributions along a 0.04 m diameter, 0.8 m long MnC2/2  RVCF

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

96

2.0 Ro = 0.3 cm Ro = 0.5 cm Ro = 1.0 cm Converging

Sigma (cm)

1.5

1.0

0.5

0.0 0

Fig. 19. Deposited beam power by convergent-incidence, 1 GeV, 1 mA proton beam in a ZrO2 fiber target. Density: 1.0  103 kg/m3; length: 0.8 m; beam spot diameter: 0.02 m.

20

40 Z (cm)

60

80

Fig. 21. Radii (sigma) of parallel or convergent incidence, 1 GeV, 1 mA proton beams, of varying beam diameter at entrance, along a 0.04 m diameter, 0.8 m long, 1.2  103 kg/m3 NbC/2  RVCF target.

6.0

Ro = 0.3 cm Ro = 0.5 cm Ro = 1.0 cm

dE/dV (W/cc)

4.5

3.0

1.5

0.0 0

20

40 Z (cm)

60

80

Fig. 20. Beam power depositional density along a 0.04 m diameter, 0.8 m long, 1.2  103 kg/m3 NbC/2  RVCF target irradiated with a parallel entrant, 1 GeV, 1 mA proton beam of varying beam diameter at entrance.

composite target irradiated with a 1 GeV, 100 kW converging proton beam without and with layers of heat shielding added to the rear end of the target are displayed, respectively, in Figs. 22 and 23. As noted, the addition of two layers of heat shielding to the exit end of the target serves to homogenize the temperature within the target. The temperature distributions for NbC/2  RVCF composite targets of the same dimensions, without and with additional heat shielding, irradiated with 1 GeV, 100 kW proton beams are shown, respectively, in Figs. 24 and 25. The results of analogous simulations for composite UC2/2  RVCF targets of the

same dimensions, without and with additional heat shielding, under irradiation with 1 GeV, 70 kW proton beams, are shown, respectively, in Figs. 26 and 27. Due to scattering effects caused by changes in target density and atomic number of the target material, the number of layers of heat shielding required for homogenizing target temperatures will vary from target to target. Gradients in the temperature distributions are more pronounced in heavier targets because of increased scattering. Because of the higher density and higher atomic number, the UC2/2  RVCF target necessitates at least four layers of additional heat shielding at the rear of the target to homogenize the temperature over the target volume while lower density, lower atomic number targets, such as MnC2/2  RVCF and NbC/2  RVCF, only require two layers of additional shielding. Very low atomic number targets, such as BeO/W/2  RVCF, require no additional shielding. 5.2.2. Time domain simulation studies: scanning scenarios As noted in Table 5, conventional single-column targets made of low-density fibrous materials are limited to beam power levels p100 kW, far lower than the 400 kW production beam power levels proposed at the RIA. To accommodate such high beam power levels, the primary beam must be distributed over larger target cross-sectional areas

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

97

Fig. 22. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.0  103 kg/m3 MnC2/2  RVCF composite target irradiated with a 1 GeV, 100 kW converging proton beam without layers of heat shielding added to the rear end of the target.

Fig. 23. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.0  103 kg/m3 MnC2/2  RVCF composite target irradiated with a 1 GeV, 100 kW converging proton beam with layers of heat shielding added to the rear end of the target.

Fig. 24. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.2  103 kg/m3 NbC/2  RVCF composite target irradiated with a 1 GeV, 100 kW converging proton beam without layers of heat shielding added to the rear end of the target.

Fig. 25. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.2  103 kg/m3 NbC/2  RVCF composite target irradiated with a 1 GeV, 100 kW converging proton beam without layers of heat shielding added to the rear end of the target.

perpendicular to the direction of the primary beam, suggesting the use of beam manipulation techniques. In this section, computational methods are utilized to assess the merits of a number of

beam manipulation methods devised for more uniformly distributing 400 kW primary beamdeposited power over the volumes of low-density, fibrous, single-stage targets.

ARTICLE IN PRESS 98

Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

Fig. 26. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.6  103 kg/m3 UC2/2  RVCF composite target irradiated with a 1 GeV, 70 kW converging proton beam without layers of heat shielding added to the rear end of the target.

Fig. 27. Temperature distribution along a 0.04 m diameter, 0.8 m long, 1.6  103 kg/m3 UC2/2  RVCF composite target irradiated with a 1 GeV, 70 kW converging proton beam with layers of heat shielding added to the rear end of the target.

Since the scan technique involves the timedependent displacement of the primary beam, scan systems and beam transport systems must be designed and implemented near the target region that are commensurate with the scan operation. Time-domain converging incidence beam manipulation (scanning) techniques can be used to distribute the beam power over larger beam direction surface areas in order that these fragile targets withstand irradiation with 400 kW proton beams. This technique will permit varying the scan pattern and the scan period for a given scenario as well as changing the beam-direction surface area to accommodate foreseeable power levels required at high-energy facilities such as the RIA (e.g., 400 kW). By use of this technique, practically sized targets can be designed that will accommodate this power level. 5.2.3. The linear scanning technique (rectangular cross-section targets) A linear scan scenario, schematically illustrated in Fig. 28, was conceived to achieve an average power density commensurate with irradiation of

Fig. 28. Schematic isometric representation of a linear scan scenario for irradiating ISOL production targets with very high-power proton beams. Scan height: 0.24 m; target width: 0.04 m; target length: variable.

fibrous or composite fibrous targets with beams at power levels up to 400 kW. In this scenario, a circular cross-section beam with a Gaussian radial

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

profile is scanned over a rectangular cross-section target (height: 0.24 m; width: 0.04 m; length: 0.8 m) along the 0.24 m direction and the temperature distribution within the target computed in the time domain with ANSYS.7 Convergent incidence primary beams are required to both reduce the primary beam depositional density and to avoid losses of the beam due to scattering through the walls of the reservoir. The ends and edges of the target are cooled by radiation to the surrounding external surfaces, and consequently, temperatures

dx/dt (cm/s)

4.5

3.5

2.5

1.5 -12.0

-8.0

-4.0

0.0 x (cm)

4.0

8.0

12.0

Fig. 29. Beam scanning speed used to sweep a 1 GeV, 400 kW proton beam over a 1.2  103 kg/m3 NbC/2  RVCF composite target. Because of varying atomic number, densities, thermal conductivities and limiting temperatures, scan speeds will vary from target-to-target.

99

at the walls of the target-material reservoir will be lower than the internal regions of the target. This effect can be moderated by programming the scan speed to more uniformly heat the entire target volume. The scan speed profile used in the simulations is displayed in Fig. 29. In this scenario, a faster scan speed is used in the central region of the target while a slower speed is used near the edges of the target. Figs. 30 and 31 show, respectively, temperature profiles along the axis of a 0.04  0.24  0.8 m3, 1.2  103 kg/m3 NbC/ RVCF composite target irradiated with 1 GeV, 400 kW proton beam without and with additional heat shielding placed at the exit end of the target; the scan speed is varied according to the scheme described in Fig. 29 to try to more homogeneously distribute beam heat over the target volume. As noted in Fig. 32, the additional heat shielding greatly improves the temperature distribution over the target volume. Fig. 33 displays temperature versus time for several positions, x, along the scan direction of the target with radiation shielding on the exit end of the target. The cut-plane through the target was made at z=0.3 m at a scan period of 13 s. Figs. 34 and 35 show, respectively, temperature profiles along the axis of a 0.04  0.24  0.8 m3, 1.6  103 kg/m3 UC2/ RVCF composite target irradiated with 1 GeV, 300 kW proton beams without and with additional

Fig. 30. Temperature distribution along the mid-plane of an NbC/2  RVCF composite target without additional heat shielding. Target density: 1.2  103 kg/m3; scan height: 0.24 m; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 400 kW; incident beam: converging; 0.02 m diameter.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

100

2200

2450

2050

2250 Temp. (K)

Temp. (K)

Fig. 31. Temperature distribution along the mid-plane of a NbC/2  RVCF composite target with additional heat shielding. Target density: 1.2  103 kg/m3; scan height: 0.24 m; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 400 kW; incident beam: converging, 0.02 m diameter.

1900 Without With

1750

2050 x = 0 cm x = 10 cm x = 12 cm

1850

1650 0

1600 0

20

40 Z (cm)

60

80

Fig. 32. Comparison of the temperature distribution along the mid-plane of a NbC/2  RVCF composite target with and without additional heat shielding. Target density: 1.2  103 kg/ m3; scan range: 70.12 m; scan period: 13 s; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 400 kW; incident beam: converging, 0.02 m diameter.

heat shielding placed at the exit end of the target; the scan speed is again varied according to the scheme described in Fig. 29 to try to more homogeneously distribute beam heat over the target volume. The temperature profile along

5

10

15 t (s)

20

25

30

Fig. 33. Temperature versus time at different scan positions, x, in a 1.2  103 kg/m3 NbC/2  RVCF composite target under irradiation with a 1 GeV, 400 kW proton beam with additional heat shielding at the exit end of the target. Calculations made at a cut-plane of z=0.30 m. Scan range: 70.12 m; scan period: 13 s; incident beam: converging, 0.02 m diameter.

the axis of the UC2/2  RVCF target, taken at the mid-plane of the target, is shown in Fig. 36. Fig. 37 displays temperature versus time for three scan positions, x, along the scan direction with radiation shielding on the exit end of the target.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

101

Fig. 34. Temperature distribution along the mid-plane of a UC2/2  RVCF composite target without additional heat shielding. Target density: 1.6  103 kg/m3; scan height: 0.24 m; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 300 kW; incident beam: converging, 0.02 m diameter.

Fig. 35. Temperature distribution along the mid-plane of a UC2/2  RVCF composite target with additional heat shielding. Target density: 1.6  103 kg/m3; scan height: 0.24 m; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 300 kW; incident beam: converging, 0.02 m diameter.

The temperatures were calculated at the z=0.3 m position. The scanning period for the UC2/ 2  RVCF target is decreased from 13 s, as used for the NbC/2  RVCF target, to 6.5 s. As noted in Figs. 32 and 36, there are strong temperature gradients from the mid-plane to each side of the respective targets in the 0.04 m direction. The gradient problem can be ameliorated by programming the scan rates in both the 0.04 and 0.24 m

directions. Under these conditions, the target temperature is expected to reach a constant or approximately constant value over the full volume of the target. This approach will be studied in the future. As noted, the linear scan technique can be utilized to accommodate irradiation with beams commensurate with RIA applications while achieving more uniform heat distributions within

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

102 2500

Temp. (K)

2300 2100 1900 Without With

1700 1500 0

20

40 Z (cm)

60

80

Fig. 36. Comparison of the temperature distribution along the mid-plane of a UC2/2  RVCF composite target with and without additional heat shielding. Target density: 1.6  103 kg/ m3; scan period: 6.5 s; scan height: 0.24 m; target width: 0.04 m; target length: 0.8 m; proton beam: 1 GeV, 300 kW; incident beam: converging, 0.02 m diameter.

2450

Temp. (K)

2250

2050 x = 0 cm x = 10 cm x = 12 cm

1850

1650 0

2

4

6

8

10

12

14

t (s)

Fig. 37. Temperature versus time at different scan positions, x, in a 1.6  103 kg/m3 UC2/2  RVCF composite target under irradiation with a 1 GeV, 300 kW proton beam with additional heat shielding at the exit end of the target. Calculations made at a cut-plane of z=0.3 m. Scan range: 70.1 m; period: 6.5 s; incident beam: converging, 0.02 m diameter.

targets. Targets made of very low thermal conductivity materials, such as HfO2 and ZrO2 fibers, can withstand irradiation with 300–400 kW proton beams. Other higher thermal conductivity targets such as Ta/2  RVCF and BeO/W/2  RVCF can also withstand irradiation with 1 GeV, 400 kW proton beams. However, UC2/2  RVCF targets

Fig. 38. Schematic isometric representation of a rotational beam scan scenario for irradiating ISOL production targets with very high-power, convergent incidence, proton beams. Annular cylindrical geometry target; circular scan; annular target width: 0.04 m; inner diameter: variable; scan period: variable.

can only withstand irradiation with B300 kW proton beams because of its poor thermal conductivity properties, higher atomic number and the power deposited by fission fragments. This problem is exacerbated by the fact that the UC2/ 2  RVCF target has a higher density than other targets of the same thickness. 5.2.4. The rotating beam scanning technique: annular target scenario An alternative scan procedure is to rotate the beam around an annular target. A schematic illustration of this scan scenario is illustrated in Fig. 38. Although the rotating beam scan method is a more complex mechanism for improving the distribution of primary beam deposited heat, it is compatible in design with the indirect two-stage neutron-fission target proposed for use at the RIA [20,21]. The temperature distribution in the direct proton beam irradiation target scenario can be improved by providing heat shielding on the exterior of the target while cooling the inner annular surface. Fig. 39 shows temperature profiles taken along a 0.8 m long, annular (inner radius, Rin: 0.06 m; outer radius, Rout: 0.1 m) 1.2  103 kg/m3 NbC/ 2  RVCF composite target irradiated with a 1 GeV, 400 kW proton beam without and with additional heat shielding placed at the exit end of the target. The inner surface of the target is cooled.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

103

Fig. 39. Temperature distribution within a 1.2  103 kg/m3 NbC/2  RVCF annular target irradiated with a convergent incidence, 1 GeV, 400 kW proton beam with cooled inner surface. Inner target diameter: 0.06 m; target width: 0.04 m; target length: 0.8 m; scan period: 12 s. (a) end view; (b) without additional heat shielding; (c) with additional heat shielding.

The rotational scan speed of the proton beam is 12 s. Fig. 40 displays temperature versus time for several positions, x, along the mid-plane of the target (with radiation shielding on the exit end of the target. The cut-plane through the target was made at z=0.3 m.

As discussed previously, UC2/2  RVCF targets are heated to higher temperatures with the same primary beam power because of its poor thermal conductivity properties, higher atomic number, and especially because of the additional power transferred to the target matrix by

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

104 2250

Inner Median Outer

Temp. (K)

2150

2050

1950

1850 0

5

10

15

20

25

t (s)

Fig. 40. Temperature versus time for three radial positions within a 1.2  103 kg/m3 NbC/2  RVCF target irradiated with a convergent incidence, 1 GeV, 400 kW proton beam with additional heat shielding. Inner target diameter: 0.06 m; target width: 0.04 m; target length: 0.8 m; scan period: 12 s.

energetic fission fragments. Temperature profiles, taken along a 0.8 m long, annular (inner radius, Rin: 0.06 m; outer radius, Rout: 0.1 m) 1.6  103 kg/m3 UC2/2  RVCF composite target irradiated with a 0.01 m diameter, 1 GeV, 400 kW proton beam without and with additional heat shielding placed at the exit end of the target are shown in Fig. 41. The inner surface of the target is again cooled. The rotational scan speed of the proton beam is the same as for the NbC/ 2  RVCF target (12 s). Temperature versus time are displayed in Fig. 42 for several positions, x, along the mid-plane of the target (with radiation shielding). (These data were also taken at z=0.3 m.)

6. Conclusions This report provides lists of refractory oxides, carbides and refractory metals suitable for use as targets for spallation production of short-lived proton-rich isotopes of elements He through Pu and fissionable compounds such as UC2 and ThC2 for producing short-lived neutron-rich isotopes of Cu through Dy for potential use at next-generation, ISOL-based, RIB facilities.

Design specifications are provided for custom engineering targets with thicknesses compatible with the diffusion release of isotopes within their lifetimes. In addition, complex structure, highly permeable RVCF matrices are described that are amenable for coating optimum thickness of any target material (metals, carbides or oxides) during fabrication by use of a newly developed universal vacuum penetration method [22], also briefly described. We have developed dimensional criteria for achieving the release efficiency required of ISOL targets that will diffusively release 70% of the isotope of interest within its lifetime and have used them in the design of highly successful, fibrous [2,10,30] and highly permeable, composite production targets [2,11,13,30] for use at the HRIBF [1,2]. Thus, these criteria permit custom engineering of the thickness of deposits of the production material that will ensure the release of the species of interest within its lifetime and form the basis of a target design philosophy that has been successfully incorporated and experimentally demonstrated in a number of fast release targets at the HRIBF and therefore will be incorporated in the design of prototype RIA targets. In addition, a selected number of low-density fibrous targets are thermally analyzed using computational methods where internal radiation is a primary mechanism for heat redistribution. Such targets have open channels over their entire target volumes that provide partial optically transparent paths for photons to travel so that thermal radiation can take place within the interior regions of such structures when raised to high temperatures. During transit, primary beam scattering and nuclear reactions cause strong gradients in the beam power deposition and consequently, strong variations in temperature distributions within these targets. By utilization of the radiation cooling effect, converging beams and beam-manipulation techniques in combination with placement of additional radiation shielding on the exit ends of targets, temperatures can be homogenized and controlled to manageable levels within practically sized targets while avoiding devastating primary beam scattering losses during irradiation with 1 GeV,

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

105

Fig. 41. Temperature distribution within a 1.6  103 kg/m3 UC2/2  RVCF annular target irradiated with a convergent incidence, 1 GeV, 400 kW proton beam with cooled inner surface. Inner target diameter: 0.06 m; target width: 0.04 m; target length: 0.8 m; scan period: 12 s. (a) end view; (b) without additional heat shielding; (c) with additional heat shielding.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107

106 2400

Inner Median Outer

Temp. (K)

2300 2200 2100 2000 1900 0

5

10

15

20

25

t (s)

Fig. 42. Temperature versus time for three radial positions within a 1.6  103 kg/m3 UC2/2  RVCF target irradiated with a convergent incidence, 1 GeV, 400 kW proton beam with additional heat shielding. Inner target diameter: 0.06 m; target width: 0.04 m; target length: 0.8 m; scan period: 12 s.

400 kW proton beams, available at next-generation RIB facilities.

Acknowledgements The authors are indebted to present and past members of the HRIBF staff, who through their diligent efforts have contributed to the content of this paper through validation of the concepts embodied in targets described in this report by online testing and use of targets in executing important nuclear and nuclear astrophysics research.

References [1] G.D. Alton, J.R. Beene, J. Phys. G, Nucl. and Part. Phys. 24 (1998) 1347. [2] G.D. Alton, J.R. Beene, R.L. Auble, D. Stracener, Indian J. Phys. 76S (2002) 9. [3] D.C. Radford, et al., Phys. Rev. Lett. 88 (2002) 222501. [4] J. Gomez del Campo, et al., Phys. Rev. Lett. 86 (2001) 43. [5] D.W. Bardayan, et al., Phys. Rev. Lett. 83 (1999) 45. [6] D.W. Bardayan, et al., Phys. Rev. C 62 (2000) 042802. [7] D.W. Bardayan, et al., Phys. Rev. C 62 (2000) 055804. [8] D.W. Bardayan, et al., Phys. Rev. C 63 (2001) 065802.

[9] D.W. Bardayan, et al., Phys. Rev. Lett. 89 (2002) 262501. [10] G.D. Alton, in: J.L. Duggan, I.L. Morgan (Eds.), Application of Accelerators in Research and Industry, AIP Press, NY, CP392, 1997, p. 429. [11] G.D. Alton, H. Moeller, Phys. Div. Annual Report, ORNL Report-6057, hr021.pdf. [12] G.D. Alton, J.R. Beene, Y. Liu, Nucl. Instr. and Meth. A 438 (1999) 190. [13] G.D. Alton, Nucl. Instr. and Meth. A 382 (1996) 207. [14] G.D. Alton, D.L. Haynes, G.D. Mills, D.K. Olsen, Nucl. Instr. and Meth. A 328 (1993) 325. [15] G.D. Alton, Proceedings of the 2001 Particle Accelerator Conference, IEEE: 01CH37268C (2001) 59–63. [16] G.D. Alton, Y. Liu, C. Williams, S.N. Murray, Nucl. Instr. and Meth. B 170 (2000) 515. [17] H. Bilheux, Y. Liu, S.N. Murray, G.D. Alton, in: J.L. Duggan, I.L. Morgan (Eds.), Application of Accelerators in Research and Industry, AIP Conference Proceedings, CP576, 2001, p. 244. [18] G. Savard, Proceedings of the 2001 Particle Accelerator Conference, IEEE, New York, 01CH37268C (2001) 561. [19] http:/www.nscl.msu.edu/future/ria [20] W.L. Talbert, Two-Step Target for Fission Product Radioactive Beam Production, TechSource, Inc., Santa Fe, USA, March 2002, unpublished. [21] D. Ridikas, W. Mittig, A.C. Villari, Proceedings of the Fifth Shielding Aspects of Accelerators, Targets and Irradiation Facilities, Paris, France, 2000, p. 153. [22] G.D. Alton, J-C. Bilheux, A.D. McMillan, Nucl. Instr. and Meth. A, (2004) these proceedings. [23] P. Shewmon, Diffusion in Solids, 2nd Edition, A publication of the Minerals, Metals and Materials Society, Warrendale, PA, 1989. [24] R. Kirkaldy, D.J. Young, Diffusion in the Condensed State, The Institute of Metals, London, 1987. [25] R. Kirchner, Nucl. Instr. and Meth. B 70 (1992) 186. [26] G.D. Alton, J. Dellwo, H.K. Carter, I.Y. Lee, C.M. Jones, J. Kormicki, G. diBartolo, J.C. Batchelder, J. Breitenbach, J.A. Chediak, K. Jentoff-Nilsen, S. Ichikawa, in: G. Vourvopoulos, T. Paradellis (Eds.), Proceedings of the International Conference on Neutrons and their Applications, SPIE, Vol. 2339, Bellingham, WA, 1995, pp. 335–349. [27] G.D. Alton, J. Dellwo, Nucl. Instr. and Meth. A 382 (1996) 225. . [28] H. Landolt, R. Bornstein, W. Martienssen (Eds.), Numerical data and functional relationships in science and technology—new series, Group-3, Condensed Matter, Vol. 26; Diffusion in Solid Metals and Alloys, SpringerVerlag, Berlin, 1998. [29] A.K. Roy, R.P. Chhabra, Diffusivities in of metallic solutes in molten metals, in: D.R. Lide (Ed-in-Chief), The Handbook of Chemistry and Physics, 76th Edition, CRC Press, Boca Raton, 1995–1996, p. 258 (Chapter 6). [30] D. Stracener, et al., Nucl. Instr. and Meth. A, (2004) these proceedings.

ARTICLE IN PRESS Y. Zhang, G.D. Alton / Nuclear Instruments and Methods in Physics Research A 521 (2004) 72–107 [31] Y. Liu, G.D. Alton, Nucl. Instr. and Meth. A 438 (1999) 210. [32] Y. Liu, G.D. Alton, J.W, Middleton, AIP Conf. Proc. 475 (1998) 265–268. [33] Y. Zhang, Y. Liu, D. Bhowmick, G.D. Alton, Proceedings of the PAC 2001, Chicago, USA, June 2001, pp. 1607–1609. [34] G.D. Alton, J.W. Middleton, R. Dinwiddie, Thermal conductivity measurements of electroplated Ni/RVCF

107

composite targets, Phys. Div. Annual. Report ORNL6057, hr023.pdf. [35] R. Zee, G. Romanoski, H.S. Gale, H. Wang, Processing and Thermal properties of Filament Wound Carbon–Carbon Composites for Impact Shell Application, American Institute of Physics Conference Proceedings, Melville, NY, Vol. 552, 2001, pp. 727–732.