Home Search Collections Journals About Contact us My IOPscience

The neutron cross section standards, evaluations and applications

This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2011 Metrologia 48 S328 (http://iopscience.iop.org/0026-1394/48/6/S09)

View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 129.6.99.26 The article was downloaded on 28/11/2012 at 19:44

Please note that terms and conditions apply. IOP PUBLISHING METROLOGIA Metrologia 48 (2011) S328–S345 doi:10.1088/0026-1394/48/6/S09 The neutron cross section standards, evaluations and applications

A D Carlson

National Institute of Standards and Technology, STOP 8463, Gaithersburg, MD 20899, USA E-mail: [email protected]

Received 21 April 2011, in final form 19 June 2011 Published 28 October 2011 Online at stacks.iop.org/Met/48/S328

Abstract The measurement of a neutron cross section can be simplified if it is measured relative to a neutron cross section standard. Then it is not necessary to measure the neutron fluence. The standards in effect provide determinations of the neutron fluence. These standards are not defined standards but instead working standards. As such it is important to make measurements to improve the quality of the standards. Also better techniques for evaluating the standards must be promoted. The history of the evaluation process for the standards is discussed from the first rather primitive one to the latest very sophisticated evaluation that was an international cooperative effort. An overview of the main detectors used for applications of the standards is given. (Some figures in this article are in colour only in the electronic version)

1. Introduction section or fluence measurements that can be made without using a standard. These measurements are either independent In the measurement of a cross section, for thin samples, the or only very weakly dependent on any cross sections. These event counting rate for the cross section to be measured is include those using transmission (to obtain the total cross given by section), certain types of spherical shell transmission (to obtain R = εφNσ (1) the inelastic cross section), associated particle and associated where ε is the efficiency for detecting the event, φ is the neutron activity measurements. Also certain detectors such as the fluence, N is the number of atoms in the sample and σ is the ‘Black Detector’ [POE72] allow accurate determinations of cross section being measured. the neutron fluence. The measurements of neutron cross sections are Neutron cross section standards were used very early in significantly easier if a neutron cross section standard can be the cross section measurement process. Unfortunately many used. Then if a detector that implements the standard reaction of the ‘standards’ were not appropriate. And in some cases, is placed in the same neutron beam where the detector for which improved candidates were suggested [CAR70, CIE77] but not a cross section to be measured is located, for the conditions of adopted, largely as a result of the need to significantly improve equation (1), one obtains their databases which would have required significant effort. An idealized standard should have the following R ε N σ = x s s σ (2) characteristics: x R ε N s s x x • It should be possible to use the nuclide in elemental where the subscript x refers to the detector for which a cross form (not in a compound), be chemically inert and not section is to be measured and subscript s refers to the standard radioactive. detector. Using the standard it is not necessary to make a direct • It should be easy to fabricate into various shapes. measurement of the neutron fluence. The standards in effect • It should be readily available; not expensive. remove the need for measuring the fluence. However, then it is • It should have few (or no) other channels open that could necessary to measure the standards very accurately since any cause interference with the reaction of interest. measurement relative to a standard is limited in accuracy to • For activation standards, there should be suitable half-lives that of the standard. There are, however, a few types of cross and decay schemes for the product.

0026-1394/11/060328+18$33.00 © 2011 BIPM & IOP Publishing Ltd Printed in the UK & the USA S328 The neutron cross section standards, evaluations and applications • Monotopes are preferred. For the various evaluations in the library, the data had been • For capture standards, the total gamma-ray energy should measured relative to a standard cross section, but different be large with high multiplicity yet a hard capture gamma- evaluators were not necessarily using the same values for the ray spectrum. standard cross sections. Thus the term standard had a different • In the standards energy region, the cross section should be meaning for that work compared with what it means now. It large with a minimal amount of structure. was rapidly recognized that proper and consistent standards • For capture standards, preferably one resonance should must be used in the evaluation process. The evaluations for have appropriate resonance parameters so the saturated ENDF/B-I were basically used to check the processing codes resonance technique can be used for neutron fluence which had been written. determination. Normally the conversion of a cross section ratio measurement 2.2. ENDF/B-II to the cross section of interest is not done with a single measurement of that standard. Instead the available data on the Greater attention was given to the standards used for cross standards are evaluated to improve the accuracy and provide section evaluations in this version. A Normalization and uncertainty information that is generally much better than can Standards Subcommittee was established to work on the be obtained from a single standards measurement. These standards. The ENDF/B-II library did not predict several standards become the basis for the evaluation of cross sections integral benchmark measurements as well as the ENDF/B-I for the neutron cross section libraries. library, even though much of the ENDF/B-II data were believed to be appreciably ‘better’ than the ENDF/B-I data. Neither 2. Historical perspective of these libraries was believed adequate for reactor design applications. There were many early standards but as the need for better cross sections, hence better standards arose, the standards with less desirable features were removed. Except for the most recent 2.3. ENDF/B-III evaluation of the standards, the responsibility for evaluating the standards was the US Cross Section Evaluation Working The scope of responsibilities of the Normalization and Group (CSEWG) that produces the ENDF library. Within that Standards Subcommittee had become very broad. All Group the actual evaluation of the standards was usually done activities within CSEWG Subcommittees involving standards by a Standards Subcommittee. The standards were always and dosimetry cross sections required interaction with this made available internationally for all versions of the ENDF/B Subcommittee. Any interaction outside the CSEWG, e.g. with standards. The evaluations were accepted internationally to the International Atomic Energy Agency (IAEA), on standards ensure that the same standards are used worldwide in all efforts was also handled by this Subcommittee. It appeared as major evaluation projects. The ENDF/B standards were made though anything that was known relatively well had to have available outside of America even for ENDF/B-V while other some sort of approval from the Normalization and Standards ENDF/B-V evaluations were not. Seldom is a single standards Subcommittee. measurement used in a critical calculation. Instead all the The ENDF/B-III efforts led to laboratories/individuals standards data are evaluated to provide the best results. The taking on the responsibility for specific evaluations for which standards have an important role in the definition of a data they had expertise and interest. library. Almost all the cross sections in a given data library The ENDF/B-III evaluation process for the 235U(n, f) cross depend on the standard cross sections. There is therefore section had a problem since the database showed a trend very large leverage for improvements to the standards since in which the cross section appeared to be decreasing with an improvement in a standard causes an improvement in every time. This is illustrated in [POE70]. Actually the trend was cross section relative to that standard. This applies to cross not unreasonable. The earlier measurements were subject to section measurements that have been made as well as those large backgrounds which were difficult to remove completely. that will be made. These backgrounds add signal (counts) to the apparent fission When ENDF/B was in its infancy, the number of standards, response, thus making the cross section appear too high. their energy ranges of applicability and their accuracy were However, there were many discussions with experimenters not well established. The need for better standards has led who had the larger cross sections and were convinced that they to significant improvements and very sophisticated evaluation had the ‘correct’ values for the cross sections. All the data procedures such as those that were used for the ENDF/B-VI were used in the evaluation process. and ENDF/B-VII standards evaluations. For ENDF/B-III, there were a number of reactions that were very seriously considered for standards but 2.1. ENDF/B-I not accepted (e.g. 233U(n, f), and a number of capture The initial evaluation efforts for the standards were for the reactions). The ENDF/B-III standard cross sections were for ENDF/B-I library. The standards could not be used properly H(n, n), 3He(n, p), 6Li(n, t), 10B(n, α), 12C(n, n), Au(n, γ)and for this library. The focus was on getting as many evaluations 235U(n, f). For the first time an ENDF/B report [DRA72] as possible that were of reasonable quality into the library. providing summaries describing the standards was published.

Metrologia, 48 (2011) S328–S345 S329 A D Carlson 2.4. ENDF/B-IV or those measured relative to the H(n, n) standard which were converted to cross sections using the new hydrogen evaluation. There was a significant movement towards more objective Then the 10B+n standard cross sections were evaluated. The evaluation techniques for the standards with this version of only 10B+n data that were used were absolute measurements ENDF/B. However, at that time these techniques were largely and those relative to the H(n, n) and 6Li(n, t) cross sections focused on the light-element standards with the use of R-matrix which were converted using the new hydrogen and lithium analyses. For the heavy-element standards, older evaluation evaluations. This process was continued for each of the methods were used. For example, a ‘Task Force’ was used for standards. This method for using ratio measurements does not the evaluation of the 235U(n, f) cross section. The evaluation use all the information available. It does not include absolute involved a very large piece of graph paper with all the and ratio data on the basis for which they were measured. For measurements and their uncertainties plotted on it. The Task example, a ratio of the 10B(n, α) to the 6Li(n, t) cross sections Force made suggestions as to how they felt the curve should go, would be used in the 10B(n, α) cross section evaluation but not based on their understandings of the various experiments. It in the 6Li(n, t) evaluation. A proper evaluation would ensure gave them freedom to favour (or discriminate against) certain that the ratio measurement would have an impact on both cross data sets based on the quality of the work generally done by sections. Also most evaluations were just smooth curves drawn those experimental groups. Such evaluations are difficult to through the experimental data sets. document and it is not clear how to determine meaningful The difficulties with the hierarchical evaluation procedure uncertainties and covariance information. It was clear that and the success already realized using comprehensive objective a more modern objective procedure needed to be developed. data combination techniques in the ENDF/B-V standards The 10B(n, α γ) cross section was added as a new 1 evaluation for the 235U(n, f) cross section led to the seeking standard. This is an important standard since it can be out of a more global approach for ENDF/B-VI standards than implemented by detecting the gamma ray, which does not had been used earlier. Least-squares methods should be used change in energy with the energy of the incident neutron. to combine the input data consistent with the experimental Also there were changes in the energy ranges for the H(n, n), uncertainties. The method should be able to handle the full 6Li(n, t), 12C(n, n), Au(n, γ) and 235U(n, f) cross sections information content of the database. Thus data should be compared with the ENDF/B-III evaluations. Documentation evaluated simultaneously to ensure proper use of the available for these standards is given by Magurno [MAG75]. information. Ratio measurements of standard cross sections should have an impact on each of the cross sections in the 2.5. ENDF/B-V ratio. Correlations among the experimental data should be The movement towards more objective evaluations led to a taken into account in the simultaneous evaluation. It was simultaneous evaluation of the 235U(n, f) cross section by also important to retain fits to theory in the evaluation of Poenitz [POE81]. It was composed of an evaluation of the the light-element standards. Such fits can provide coupling shape of the cross section and a separate evaluation of the to reaction theory and give a smooth meaningful analytical normalization for the shape of the cross section. The members expression for the energy dependence of the cross sections. of the Normalization and Standards Subcommittee (that Data in addition to angle-integrated neutron cross sections became the Standards Subcommittee) selected the experiments such as differential cross sections, polarizations and charged- which were used for the determination of the normalization particle measurements involving the same compound nucleus factor for the shape evaluation. This evaluation was a first step can then have a significant impact on the standard cross towards an evaluation process that would provide consistent sections. This could be implemented with R-matrix analyses. sets of cross sections for all the standards. In R-matrix analyses, different reactions leading to the same No standards were added but there were changes in the compound nucleus are linked by unitarity to the standard cross energy ranges for the 12C(n, n) and Au(n, γ) cross sections section. This condition imposes constraints on the standard compared with the ENDF/B-IV evaluations. The evaluations cross section which are particularly strong near resonances. are documented in [CAR82]. For this evaluation it was decided to combine the output of a simultaneous evaluation using generalized least-squares with separate R-matrix analyses. An important condition was 2.6. ENDF/B-VI that there cannot be any correlations between the database For the ENDF/B-VI evaluation of the standards, considerable used for the simultaneous evaluation and the database used effort was devoted to improved evaluation procedures. In for the R-matrix evaluations. This procedure took advantage previous evaluations for ENDF/B, a hierarchical approach of the strengths of the two different analysis modes that can was followed. The lighter element cross section standards make use of separate classes of experimental information to were generally considered to be better known. The H(n, n) impact on the evaluation of the standard cross sections. This cross section was considered the best known standard and led to a consistent evaluation in which correlations and ratio was evaluated first and independently of the other standards. measurements were properly taken into account. To satisfy This standard is considered so well known that measurements the correlation condition, the boron and lithium experimental relative to it are often called absolute measurements. The data were separated into two uncorrelated groups, one for use in 6Li(n, t) cross section evaluation was performed next. The the R-matrix analyses and the other for use in the simultaneous only 6Li(n, t) data that were used were absolute measurements analysis.

S330 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications Table 1. Data types used in the simultaneous evaluation database. Table 2. R-Matrix evaluation data types. 7Li Database 11B Database Data type Example 6Li Total 10B Total 235 6 10 (1) Absolute cross section σnf ( U) Li(n, n) Integral data B(n, n) Integral data 6 6 10 (2) Cross section shape c ∗ σnα( Li), Li(n, n) Differential data B(n, n) Differential data c is unknown 6Li(n, n) Polarization data 10B(n, n) Polarization data 238 235 6 10 (3) Absolute cross section ratio σnf ( U)/σnf ( U) Li(n, t) Integral data B(n, α0) Integral data 235 6 6 10 (4) Ratio shape c ∗ σnf ( U)/σnα( Li), Li(n, t) Differential data B(n, α0) Differential data 6 10 c is unknown Li(n, t) Polarization data B(n, α1) Integral data 10 = 10 10 4 10 (5) Sum of cross sections σnα( B) σnα0( B) + σnα1( B) He(t, n) Differential data B(n, α1) Differential data 235 252 4 7 (6) Spectrum averaged σnf ( U) averaged over Cf He(t, t) Differential data Li(α0, α0) Differential data 4 7 cross section spontaneous fission spectrum He(t, t) Polarization data Li(α, α1) Differential data 235 10 7 (7) Absolute ratio of cross σnf ( U)/σnα( B), where Li(α, n) Differential data 10 10 10 section/sum of σnα( B) = σnα0( B) + σnα1( B) cross sections (8) Shape of type (5) data (9) Shape of type (7) data with the associated variance–covariance data. In addition to the 235U(n, f) data, this evaluation included accurate cross sections which had been measured relative to the neutron cross All the standards except the H(n, n), 3He(n, p) and C(n, n) section standards. Thus they would have an impact on the cross sections were evaluated using a simultaneous evaluation determination of the standards. 6 10 and R-matrix analyses. For the H(n, n) standard, the cross Evaluations of the Li+n and B+n cross sections were 6 section was considered so well known that data on the other produced from R-matrix analyses [HAL87]. The Li+n 10 nuclides would have very little impact on it. This cross section and B+n analyses were done using a large database that was thus treated as absolute in the evaluation. For the 3He(n, p) included the types of measurements shown in table 2.For and C(n, n) cross sections, very few ratio measurements to the ENDF/B-VI standards evaluation process, a separate code other standards existed so little would be gained by putting [PEE93] was used to combine the simultaneous evaluation them into the simultaneous evaluation and R-matrix analyses and R-matrix analyses and produce the final cross sections evaluation process. Separate R-matrix evaluations were and covariances. Figure 1 shows schematically the standards performed for the H(n, n) [DOD87], 3He(n, p) [HAL87] and evaluation procedure. Due to the nature of the R-matrix C(n, n) [FU90] cross sections. program, all experiments which are correlated and all ratio The input data for the simultaneous evaluation were measurements (except those to the hydrogen standard) were put composed of two independent subsets. The first into the first data subset, which was used in the simultaneous of these subsets was a large database of pointwise evaluation. In the R-matrix analyses, the experimental data measurements, assembled by Poenitz. The reactions in that were weighted based on the quoted relative uncertainties and database included 6Li(n, t), 6Li(n, n), 6Li(n, tot), 10B(n, α), it was assumed that no correlations other than the overall 10 10 10 10 B(n, α1γ), B(n, α0), B(n, n), B(n, tot), Au(n, γ), normalization were present among the data from a particular 235U(n, f), 238U(n, f), 238U(n, γ) and 239Pu(n, f). Significant experiment. effort was used to understand those data. Often corrections It was found that very unusual results can be obtained with that were not made by the measurers were applied. Also discrepant correlated data. For example, combining two highly uncertainties and correlations between data within a data set correlated discrepant data points can produce a result which is as well as correlations between different data sets were used not between the two input values. In an attempt to remove in defining the database. This database included many types problems associated with discrepancies for the simultaneous of measurements that are shown in table 1. Total cross section evaluation, data greater than three standard deviations from measurements for 6Li and 10B were contained in the database the output results were downweighted. This had the effect of since the scattering and reaction data are interrelated in these reducing χ 2/(degree of freedom) to essentially 1. measurements. Cross section data for the 238U(n, f), 238U(n, γ) No standards were added but there were changes in the 239 6 10 10 and Pu(n, f) reactions were included since they improved energy ranges for the H(n, n), Li(n, t), B(n, α), B(n, α1γ), the quality of the standards. This is a result of accurate and 235U(n, f) cross sections compared with the ENDF/B-V absolute measurements of these cross sections and many ratio evaluations. More information on the evaluation process is measurements of them to the standards. Measurements of the contained in [CAR93]. 235U and 239Pu fission cross sections in the 252Cf spontaneous A concern about this evaluation was the rather small fission neutron spectrum were also included in the database. uncertainties which resulted from the evaluation process. This These data had been obtained with high accuracy and were will be discussed further in section 2.7. only weakly dependent on the uncertainties in the 252Cf spontaneous neutron fission spectrum. They had an effect on 2.7. The international evaluation of the neutron standards the normalization of the evaluated cross sections. (ENDF/B-VII) The second subset which was used as input to the simultaneous evaluation was an evaluation of the thermal This evaluation, in contrast to previous evaluations of the data for 233U, 235U, 239Pu and 241Pu by Axton [AXT86] standards, was done internationally so that full use of

Metrologia, 48 (2011) S328–S345 S331 A D Carlson

THERMAL DATA FOR for direct use in the GMAP code were the ratio measurements. 6 10 10 233U, 235U, 239Pu, 241 Pu For the Li(n, t), B(n, α) and B(n, α1γ)R-matrix work, the cross sections obtained from the RAC and EDA analyses were THERMAL CONSTANTS not identical. The cross sections from the RAC and EDA EVALUATION analyses were averaged (unweighted) and used as the R-matrix input to GMAP. The covariance matrix used with these central values was that from the RAC code since its results appeared 610Li+n, B+n, Au(n,γ) more physically reasonable. At each energy point, half the 235U(n,f), 238U(n,f) difference between the RAC and EDA results was treated as 238U(n,γ ), 239Pu(n,f) a model uncertainty that was added to the RAC covariance of uncertainties. This then takes into account the differences SIMULTANEOUS obtained between the RAC and EDA analyses. EVALUATION The data from this international standards evaluation were adopted for use in the EDNF/B-VII evaluation. However, not all the covariance data from this evaluation are being COMBINING 610 RESULTS used. Only the cross energy correlations between data Li+n, B+n, CHARGED PROGRAM PARTICLE DATA points for a given standard reaction (lower triangle of the square covariance matrix) are being used in the ENDF/B-VII library. Due to the evaluation procedure used for the standards R-MATRIX ANALYSES evaluation, cross material and cross reaction correlations were also obtained. Clearly the preferred option should be to use the most complete covariance information that is available. Cross- Figure 1. Evaluation procedure for the ENDF/B-VI standards. material and cross reaction correlations can be very large in some cases. For example, most fission cross sections have worldwide capabilities would be available for the evaluation. been measured relative to the 235U(n, f) cross section. If those The CSEWG formed a Task Force to investigate how correlations are not taken into account in practical calculations to perform a new evaluation of the standards. The of nuclear systems involving those fission cross sections, the Working Party on International Evaluation Cooperation of covariance weighting will lead to incorrect results and incorrect the Nuclear Energy Agency Nuclear Science Committee uncertainties. Using complete covariance information will be formed a new Subgroup to promote international cooperation more complicated to implement, but it is necessary if improved on the nuclear data standards. The IAEA formed a results are needed. Coordinated Research Program (CRP) focused on improving Again as for the ENDF/B-VI standards evaluation, some the standard cross sections. The largest contribution to concern was expressed over the small uncertainties that were the evaluation process was made by the IAEA CRP. The obtained. A study [CAR09] was made of the process used CRP included membership from Austria, Belgium, China, for obtaining these evaluations with the conclusion that using Germany, Japan, the Republic of Korea, Russia and the USA. the assumptions used in the evaluations, the uncertainties These groups worked cooperatively to update the previous appear to be reasonable. An important result of this work work by including standards measurements made since the is that it is essential to consider the covariances, not just the ENDF/B-VI evaluation was completed and to improve the variances, in applications of cross sections to practical systems. evaluation process. The use of models in the standards evaluations leads to a redistribution of the uncertainties between variances and off- This evaluation was initiated with an extensive effort diagonal covariances of the uncertainty matrix with a reduction involving a number of code comparisons, both R-matrix and of the variances. As a result, the uncertainties (the square root model-independent, to ensure that the results obtained were of the variances) are reduced but the uncertainty of integral not code-dependent. quantities dependent on the evaluated data, if covariances are It was decided that a combination procedure similar to used in their calculation in a wide energy region, should be that used for the ENDF/B-VI standards evaluation would be conserved in general. used to obtain the standards. The simultaneous evaluation As for the ENDF/B-VI standards, in addition to the cross code, GMA, used in the ENDF/B-VI evaluation was used after sections for the standards, improved values were obtained for a modification [SMI04] to minimize the PPP effect. This the thermal constants, and the 10B(n, n), 6Li(n, n),238U(n, γ) code is GMAP. A separate combination code was not used. and 239Pu(n, f) cross sections. 3 All the standards except the H(n, n), He(n, p) and C(n, n) Plots of the ENDF/B-VII standards are shown in figures 2 cross sections were evaluated using the GMAP code. This and 3. Extensive comparisons with the ENDF/B-VI standards code provides the combining procedure using input from two and experimental measurements are given in [CAR09]. The separate R-matrix analyses, RAC and EDA, and a thermal standards and their energy ranges for ENDF/B-VII are shown constants evaluation in addition to the direct data sets normally in table 3. The thermal cross sections for the thermal standards used. The R-matrix input and thermal constants data were are given in table 4. treated like the additions of other data sets to the GMAP The 238U(n, f) cross section was accepted as a new standard code. For this evaluation, the only lithium and boron data for ENDF/B-VII. The boron and fission standards now have an

S332 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications

100 Table 3. The neutron cross section standards. Reaction Standards energy range H(n, n) 1 keV to 20 MeV 3 3He(n, p) 0.0253 eV to 50 keV 10 He(n,p) 6Li(n, t) 0.0253 eV to 1 MeV 10 αγ 10B(n, α) 0.0253 eV to 1 MeV B(n, 1 ) 10 10 B(n,α ) B(n, α1γ) 0.0253 eV to 1 MeV C(n, n) 0.0253 eV to 1.8 MeV 6 γ) 1 Li(n,t) Au(n, 0.0253 eV, 0.2 MeV to 2.5 MeV 235

Cross Section / barns U(n, f) 0.0253 eV, 0.15 to 200 MeV 238U(n, f) 2 MeV to 200 MeV

0.1 0.001 0.01 0.1 1 data are still being analysed so these conclusions may not be Neutron Energy / MeV final. In the high-energy region the uncertainties in all the data sets are rather large. Improved data are still necessary. Figure 2. The low neutron energy cross section standards. Data below 1 keV are not shown. A continuing program to improve the standards should be maintained. As a follow-on to the international standards 100 evaluation effort, an IAEA Data Development Project, ‘Maintenance of the Neutron Cross Section Standards’ was H(n,n) recently initiated to provide a mechanism for allowing new 10 experimental data and improvements in evaluation procedure 235 C(n,n) U(n,f) to be used in new evaluations of the neutron standards. This project will help to ensure that the process is in place for 1 238U(n,f) producing improved results for the next evaluations of the neutron cross section standards. γ

Cross Section / barns Au(n, ) 0.1 2.8. Higher energy standards

The NEANDC/INDC [CON92] has recommended the VL40 0.01 0.01 0.1 1 10 100 solution for the hydrogen scattering cross section as a standard Neutron Energy / MeV from 20 MeV to 350 MeV. VL40 is an energy-dependent partial-wave representation [ARN92] of combined pp and np Figure 3. The high-energy neutron cross section standards. Data elastic scattering data below 400 MeV. However, the most below 10 keV are not shown. recent measurements of the hydrogen angular distribution in the 100 MeV energy region are not consistent near 180◦ in extended energy range compared with that of the ENDF/B-VI evaluations. the CMS. Larger cross sections were measured at Uppsala The database for the standards still needs improvement. University [RAH98] (96 MeV and 162 MeV) and at PSI For example above about 20 MeV there are inconsistencies in [HUE80, FRA00] (from about 280 MeV to 580 MeV), both the more recent measurements of the 238U/235U fission cross using pseudo-monoenergetic sources, compared with the section ratio. There is a difference of nearly 10% at 200 MeV calculations. The work at Indiana University [SAR06]at between the Lisowski et al results [LIS91] and those obtained 194 MeV agrees with the calculations. The Uppsala data have a steeper angular shape at back angles and are larger by Shcherbakov et al [SHC02] for this ratio. The recently ◦ published results of Nolte et al [NOL07] tend to support the by as much as 10% at 180 in the CMS, compared with the Lisowski et al data; however, the uncertainties on the Nolte calculations. In figure 5, the Uppsala data at 162 MeV and the et al measurements are rather large. In figure 4 these three Indiana University data at 194 MeV are shown, each compared data sets are shown. Measurements of this ratio have been with calculations using the PWA93 partial-wave analysis. The made by two experimental groups at the N TOF facility. The PWA93 results [STO93] agree excellently with the VL40 data from both groups, Calviani et al [CAL07] and the Audouin solution. The Indiana work was done using neutrons tagged by et al [AUD07], are still preliminary. The Calviani et al data are detection of the associated protons from the D(p, n)2p reaction generally in good agreement with both the Lisowski et al results and these data have smaller uncertainties than the Uppsala data. and those obtained by Shcherbakov et al. However, at the This discrepancy has led to large increases in the uncertainty higher energies, the Calviani et al data fall between the other associated with the use of this standard cross section at back two data sets except at the highest energies where they agree angles. The smaller uncertainties of the Indiana data suggest somewhat better with the the Lisowski et al measurements. that the VL40 data are more reliable. The Audouin et al results are in better agreement with the The PSI group indicates they have done all they can with Lisowski et al data. This is particularly evident at the highest their experiment and its analyses. Nothing further can be neutron energies. Both the Calviani et al and the Audouin et al expected from that group to resolve the discrepancy.

Metrologia, 48 (2011) S328–S345 S333 A D Carlson Table 4. Thermal cross section standards. 3 6 10 10 235 Standard He(n, p) Li(n, t) B(n, α) B(n, α1γ) Au(n, γ) U(n, f) Cross section/barns 5316.00 938.47 3842.56 3600.86 98.66 584.33

1.0 example, it is convenient to have a standard of the same type as the cross section to be measured. Then, by replacing 0.9 the sample to be measured with the standard sample, the 0.8 cross section measurement can be made without additional 0.7 detectors or electronics and only a simple beam monitor is 0.6 required to relate the neutron beam intensity for the two runs. The situation is somewhat similar for fission measurements 0.5 where a fission standard can be used in ratio measurements 0.4 Nolte et al. using convenient detectors such as double fission chambers. 0.3 Cross Section Ratio Cross Lisowski et al. Then each of the detectors is essentially identical and requires 0.2 Shcherbakov et al. similar electronics. There is also a need for many standards 0.1 because none of our standards satisfies all the requirements for a standard for all neutron energies of interest, e.g. the structure 0.0 0 20 40 60 80 100 120 140 160 180 200 220 in the cross section or the difficulties in implementing the cross Neutron Energy / MeV section.

Figure 4. High-energy measurements of the 238U/235U fission cross section ratio by Nolte et al, Lisowski et al and Shcherbakov et al. 3. Measurement techniques for determining and applying the standards

For measurements of cross sections relative to a standard, both the detector for which a cross section is to be determined and the detector utilizing the standard cross section should be in the same neutron beam simultaneously. The two detectors can be close together or a great distance apart as long as they both see the same neutron beam. Measurements are often made at different times for the reaction rate and standard detector, but it is important for those measurements that a well characterized monitor be used for intercomparison of the two runs. Also for such work, the beam conditions must be very carefully controlled. In the following sections applications will be given in a limited way. Further information on the applications is given in the references in each section. Since the standards can be used to determine the neutron fluence, the paper by [NOL11] contains important information on some of the applications of the standards that complements this work.

3.1. H(n, n)H Figure 5. Measurements of the np scattering angular distribution by the Uppsala and Indiana groups compared with PWA93 calculations. The hydrogen scattering cross section is believed to be the most accurately known of the standards. The total cross section has Although there is an indication that the discrepancy may been measured to an accuracy of a few tenths of a per cent. be resolved in the 100 MeV to 200 MeV energy region, as a The total cross section is composed of the total scattering result of the Indiana data, the PSI data which cover a very large cross section and the capture cross section. But the capture and higher energy range still stand as measured. Further work cross section is extremely small in the energy region where the should be done to understand this problem. scattering cross section is a standard. Thus the total scattering There are no official standards at energies above 350 MeV. cross section can be known very well. Several techniques have been used to utilize this cross section as a standard.

2.9. Why are so many standards needed? 3.1.1. Gas proportional counters. In the low-energy region, There are a number of reasons for having so many standards. from about 1 keV to 1 MeV, gas proportional counters are In capture and scattering cross section measurements, for used. The pulse height distribution from an ideal hydrogen

S334 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications gas proportional counter in a beam of monoenergetic neutrons scatter off protons in the scintillator deviates significantly from is proportional to the centre-of-mass angular distribution for rectangular. Thus it is necessary to obtain response functions neutrons scattering on hydrogen. This relationship applies in order to unfold the measured pulse height distribution. This [BAR54] even at relativistic energies. However, the use of usually leads to unacceptable levels of uncertainty. these counters is limited to moderate neutron energies due to A special scintillator-based detector, the dual-thin wall and end effects, i.e. to situations in which the ranges of the scintillator, used by [DIA83, DIA92] has moderate efficiency, reaction products are small in comparison with the dimensions fast timing and small corrections. With this detector the escape of the counter. of proton recoils from a scintillator detector is eliminated At the lower energies the pulse height distribution for an experimentally using a second scintillator placed behind the ideal counter is nearly rectangular since the centre-of-mass first one. If there is a coincidence between the two scintillators, angular distribution for neutrons scattering on hydrogen is the outputs are added (the sum coincidence mode), otherwise nearly isotropic. Extrapolating the pulse height distribution only the output of the first one is used. Thus the total of a proportional counter to zero energy and summing the energy of a proton recoil is measured and edge effects are events in the distribution yields a result which is directly essentially eliminated. In the sum coincidence mode, the pulse proportional to the product of the neutron fluence and the height distribution has a peaked response for a given neutron hydrogen total elastic cross section, with little information on energy. This is because large angle, lower energy proton the hydrogen angular distribution being required. The timing recoils generally have larger pathlengths in the first scintillator jitter of proportional counters limits the use of these detectors and, because of their shorter ranges, have difficulty reaching in time-of-flight experiments in the MeV energy region unless the second scintillator. The sum coincidence mode detector very long flight paths are used. Both hydrogen and methane response therefore consists of the more forward scattered, (CH4) filled counters are used. Relative to hydrogen counters, higher energy proton recoil events. These recoil events have the methane counter has higher electron mobility (less timing energies more strongly correlated with the incident neutron jitter) but lower gas multiplication (lower gain). Methane energy and are less likely to fall below the pulse height counters suffer due to carbon recoils which require large bias. The response of the detector system can be accurately electronic biases and therefore sizable corrections for events calculated from the cross sections and information on the which fall below the bias. Thus for the higher neutron energies, properties of the scintillator. in principle, information on the angular distribution is needed. 3.1.4. Time projection chambers. A new method for using 3.1.2. Proton recoil telescopes. Proton recoil telescopes the hydrogen cross section as a standard using a time projection [JOH60] represent the other extreme in that this detector has detector (TPC) is being developed [HEF09]. Such detectors a relatively well-defined peak so that the error associated have been used in particle physics applications for more with the number of counts below the bias channel is reduced. than 25 years and are now being adapted for neutron data The response of the detector is directly proportional to the measurements. The design for neutron applications differs product of the fluence and the cross section for neutron–proton from that used in particle physics applications so additional scattering into the appropriate solid angle of the detector. work has been done. This type of detector can provide much better timing than The detector is similar to an ionization chamber. However, is possible with the hydrogen gas proportional counter but it can measure charged-particle trajectories in the active suffers from low efficiency. The detector in its simplest form volume in three dimensions. From the specific ionization for a is a radiator of a hydrogenous material such as polyethylene, track it can distinguish between different particle types. It will appropriate proton collimators to define the solid angle and a be used to measure and increase the accuracy of the hydrogen proton detector. This form of the telescope detector is only scattering angular distribution. A working gas of hydrogen appropriate for the lowest energy neutrons. At higher energies could be used in the TPC. Then protons can be detected for background from various sources, including the detector itself, virtually the full angular range, taking into account the energy of the proton recoil. Also if necessary, solid hydrogenous requires that additional detectors be placed in front of the targets could be used with an appropriate heavy gas for final detector and coincidences between the detectors be ionization. The TPC with its sophisticated readout planes and used. Hydrogen scattering angular distribution measurements electronic systems could then be used to determine the energy have been made using these telescopes with many telescopes and direction of the recoil protons. Thus edge effects and looking at the same hydrogenous scatterer [BOU02]. This energy losses of the protons in the target could be determined type of measurement provides important information since and their uncertainty could be minimized. Also background hydrogen scattering is not isotropic, particularly at higher could be reduced since particle identification can be used. Tests neutron energies. are now underway to understand the performance of the TPC and ensure that the results obtained have minimal systematic 3.1.3. Organic scintillator detectors. It is possible errors. to implement the hydrogen scattering cross section with hydrogen-based scintillators. The problem with these 3.2. 3He(n, p)T detectors, compared with the proportional counters, is that the light output is not linear with proton recoil energy. Thus This standard was widely used in the past but now is only the pulse height distribution resulting from neutrons that rarely used. The very high cost of 3He gas further reduces the

Metrologia, 48 (2011) S328–S345 S335 A D Carlson motivation for using this standard. The 3He(n, p)T Q value 3.3.1. Scintillators. The scintillator technique has been is 764 keV which allows the use of the standard in a number employed with 6LiI and 6Li glass scintillators. Then the of different detectors. The main detectors used are 3He gas response is a result of both the triton and alpha particles. The proportional counters, ionization chambers, gas scintillators 6LiI detectors have been used in the past [GAB59] but they are and liquid scintillators. The 3He gas proportional counter is the now seldom used due to the problems with neutron activation most popular detector. For each of these detectors the response of and resonance structure in iodine. The 6Li glass scintillator is due to both reaction products. When a neutron is absorbed has become the most popular means of using the 6Li(n, t)4He by a 3He nucleus, both the reaction products, the proton and reaction. A thin piece of this scintillator properly coupled to triton, will produce ionization (or light) as they lose energy in an appropriate photomultiplier tube can provide fast timing, 3 the He gas. The total ionization (or light) produced is closely adequate efficiency, good pulse height resolution, acceptable proportional to the total energy of the reaction products, i.e. the corrections for multiple scattering, small losses of reaction Q value plus the incident neutron energy, that produces a pulse products near the surface and high transmission of neutrons in height peak. The sum of the counts in the peak is proportional the glass. Experience has shown that 0.1 mm to 1 mm is the to the 3He(n, p) cross section. optimum thickness for many experiments. Using the same size scintillator has allowed experimenters to compare the various 3.2.1. Proportional counters and ionization chambers. correction factors, such as multiple scattering, which they have Similar to the hydrogen gas proportional counter, the use of independently deduced, thus leading to greater confidence in these counters is limited to moderate neutron energies due to these corrections. A number of cross section measurements wall and end effects but in this case there are two reaction have been made where no light pipe or optical coupling was products being detected. Typically gas additives are added to used so that the amount of scattering material in the beam the gas composition to minimize wall effects and optimize could be minimized. Special problems are encountered in the spectral resolution over a wide energy range. These 6 mixtures can be tailored for optimization of operating voltage, linac experiments with Li glass detectors due to the intense pulse rise time, pulse jitter time, gas gain, spectral resolution gamma flash from the neutron-producing target. The glasses and gamma-ray sensitivity. These counters [COS70] and are very sensitive to these gamma rays. The intense light ionization chambers [RUD74] are difficult to use for time- output of the scintillator can cause saturation and after-pulsing of-flight measurements at the higher neutron energies due to of the photomultiplier tube. One of the important reasons for the timing jitter and low neutron detection efficiency of these selecting moderately thin glasses is to minimize this sensitivity detectors. These counters have good pulse height resolution. but yet have an acceptable neutron detection efficiency. The The highest resolution has been obtained with gridded effects of the gamma flash are generally handled in one of ionization chambers [RUD74] . This type of counter has a three ways. The first method is to reduce the intensity of the resolution of 16 keV FWHM for thermal neutrons and 35 keV gamma flash by moving the detector further from the target or FWHM for 1 MeV neutrons. High resolution such as this putting thick filters in the beam. The second method involves provides good response functions for unfolding of neutron the use of electronic off-gating techniques to desensitize the spectra. photomultiplier for a period of time which includes the gamma flash burst [LAM75]. The third method utilizes common mode 3.2.2. Gas and liquid scintillators. Gas scintillators using rejection with another detector system (possibly using a 7Li 3He [BEH88, AAM66] have been designed and fabricated. glass detector) that detects the gamma flash but not the neutron These detectors have good timing and neutron efficiency, but so a minimization of the gamma flash can be obtained with at the present state of the art the pulse height resolution is electronic subtraction. only moderate. More work needs to be done to understand The high efficiency of 6Li glasses for gamma rays can the factors limiting their resolution. It is hoped that with some introduce a gamma-ray background problem. Low-energy effort the performance of these counters could be significantly gamma rays, however, can be biased out with pulse height improved. discrimination. Older attempts at pulse shape discrimination High efficiency along with good timing and pulse height to significantly reduce high-energy gamma-ray events without resolution have been obtained with liquid 3He scintillators affecting the neutron counting rates have been unsuccessful [VAN76]. In principle, such a detector would be an excellent [NEI70]; however, work [SCH91] suggests it is possible with way to implement the 3He standard. In practice, however, special stripping methods. The use of 6Li glasses in absolute the rather complicated and expensive equipment required may cross section measurements requires that the 6Li density be severely limit its use. known. Investigations of various 6Li glasses [LAM77]have revealed non-uniform 6Li densities and depletion near the 3.3. 6Li(n, t)4He edges. Also, there is concern about the stability of the 6Li The 6Li(n, t)4He reaction with a Q value of 4.78 MeV provides density of a given piece of glass depending on how it is handled a large amount of energy which can be used in neutron and with what materials it comes into contact. These properties detection. The reaction has been used with scintillators, make it necessary to treat the glasses carefully and characterize ionization chambers, solid-state detectors and track etch films. the glasses well for certain experiments. The glasses may be A review of detectors using the 6Li(n, t)4He cross section as a used in relative cross section measurements with considerably standard has been made by Weston [WES77]. less concern.

S336 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications 3.3.2. Ionization chambers. Ionization chambers employing to obtain information on the 6Li(n, t)4He angular distribution electrodes coated with 6Li compounds have been used in cross [HAR77]. For the geometry involving the solid-state detector section measurements [FRI74, BAR66]. These detectors offer out of the beam, the small solid angle implies a low count good timing and low sensitivity to gamma rays; however, they rate and only one of the reaction products would normally suffer from low neutron efficiency and a rather complicated be detected. Also the angular distribution must be known. response. The low neutron efficiency is a result of the small These complications have restricted its use in cross section 6Li thicknesses which must be used due to the short range of the measurements with high neutron flux at relatively low neutron reaction products. The complicated response of the detector energies. This method was used by [LEM71]. Measurements is due to a number of factors including the thickness of the of the 6Li(n, t)4He angular distribution with telescopes using 6Li layer which relates directly to the loss of events resulting solid-state detectors have recently been made by [DEV08]. from reaction products which do not reach the ionization chamber gas. Another factor in this response is the range 3.3.4. Track etch films. Track etch films for use with of the reaction products in the gas. In chambers of realistic the 6Li(n, t)4He reaction have a rather limited use. In this spacing, which provide fast time response, the tritons emitted measurement limited solid angle geometry was employed 6 perpendicular to the Li layer do not stop in the gas. Thus the to avoid the problem with shallow tracks from alpha energy deposited by the tritons is dependent on the angle of particles emitted nearly parallel to the 6Li deposit. It was 6 emission relative to the Li layer. The pulse height distribution possible to identify with 100% efficiency the alpha particles is then a function of the triton angular distribution and its incident upon the films. This technique requires accurate neutron energy dependence. The angular distribution is not determination of solid angles and in principle is dependent isotropic even for neutron energies in the keV region. If a on the angular distribution of the alpha particles. In the bias on the pulse height distribution can be selected which measurement of Engdahl et al [ENG81], the dependence on allows all or most of the tritons to be counted then the this angular distribution was minimized by measuring the response and its energy dependence are not very important angular distribution with the equipment used for the cross factors. However, if such a bias cannot be used, complicated section measurement. No timing information is available with corrections to the data must be made based on factors such this technique. as the angular distributions. The most recent work on this reaction using ionization chambers has used Frisch gridded 3.4. 10B(n, α)7Li and the 10B(n, α γ)7Li standards types [KNI83, ZHA06]. These detectors allow determination 1 of the energy and direction of the recoil products. Thus edge The two standards are correlated since the 10B(n, α)7Li cross effects and losses in the target can be determined and their section is the sum of the cross sections for the reactions 10 7 10 7 uncertainty could be minimized. Also background can be B(n, α0) Li and B(n, α1γ) Li. These reactions have Q reduced since particle identification can be used. It is important values of 2.792 MeV and 2.314 MeV, respectively. The to make measurements to ensure that each detector using this 10B(n, α)7Li reaction is implemented using detectors which 7 technique does not have bias. The performance of a simple will detect the α0 or α1 particles and also possibly the Li ionization chamber is well understood. But the Frisch gridded nuclei. The main detectors that have been used in cross detector has both mechanical and electronic aspects that could section measurements are proportional counters, ionization 10 7 cause an inefficiency. chambers and gamma-ray detectors (for the B(n, α1γ) Li cross section). Boron scintillators [THO62] have been 3.3.3. Solid-state detectors. Solid-state detectors combined investigated; however, the boron compounds poison the with thin 6Li layers have been used in two basically different scintillator so that low pulse heights are produced. These geometries. In one case both the 6Li and the detector are detectors are not being used in cross section measurements. in the neutron beam, while in the other case only the 6Li The use of solid-state detectors with 10B films is very similar 6 layer is in the beam. The geometry involving the detector experimentally to their use with Li films, as discussed in in the beam allows the 6Li layer to be placed very close to section 3.3.3. The lower Q value, however, is a limiting factor. the detector. Using two detectors placed very close together with a lithium compound placed on one of the diodes, it is 3.4.1. Proportional counters. Proportional counters have possible to detect both reaction products simultaneously and been popular due to the simplicity of their use [SOW70]. 10 do summing. The excellent energy resolution of solid-state Typically, the counter gas used is BF3. When a neutron detectors has permitted the in-beam detectors to be used as is absorbed by a 10B nucleus, both the reaction products— neutron spectrometers [JON77]. The in-beam detectors are the 7Li nucleus and an alpha particle—will produce ionization 10 similar to the ionization chamber with respect to timing and as they lose energy in the BF3 gas. The total ionization efficiency. They do have lower counting rates due to the produced is closely proportional to the total energy of the small areas of solid-state detectors which are often used. The reaction products, i.e. the Q value plus the incident neutron amount of material in the neutron beam is a problem with energy. With low gas pressure the pulse height distribution will 10 7 10 7 these detectors. This results in corrections for scattering and show peaks for the B(n, α0) Li and B(n, α1γ) Li reactions. reactions in the components of the solid-state detector. Also, These peaks will be separated by the Q value difference of solid-state detectors are damaged after exposure to a high dose 478 keV. The sum of the counts in the peaks is proportional of neutrons. These detectors have been used in experiments to the 10B(n, α)7Li cross section. With higher pressure it is

Metrologia, 48 (2011) S328–S345 S337 A D Carlson difficult to separate the peaks due to electron attachment to The problem is more severe for NaI detectors since some of the 10 the BF3 gas but it is not necessary to resolve the groups. gamma-ray lines (e.g. 438 keV from Na and 417 keV from I) As with the previous discussion of proportional counters, they are relatively close to the 478 keV line from 10B+n. The rather are limited to moderate neutron energies due to wall and end poor resolution does not permit the peaks to be completely effects. The timing jitter of these counters is a severe limitation resolved. It is important to make scattered neutron background to the use of these devices for time-of-f1ight experiments at measurements with this detector. This can be done by replacing high neutron energies. the 10B sample with a scattering sample such as carbon having a thickness to give equivalent scattering. The use of Ge(Li) 3.4.2. Ionization chambers. Ionization chambers have been detectors significantly reduces the scattered neutron problems 10 10 designed which contain BF3 gas as well as solid B deposits. since the gamma rays from inelastic scattering in germanium 10 A typical ungridded chamber contains a 20% BF3 +80% are relatively far from the 478 keV line and can be well resolved argon gas mixture. The argon decreases the range of the from it with the good pulse height resolution which is possible. reaction products and improves the timing of the detector. For ungridded chambers the peak response is lost due to 3.5. C(n, n)C the dependence of the pulse size on track orientation of the alpha and 7Li ions. The use of gridded chambers removes The natural carbon differential elastic scattering cross section this problem. Sophisticated gridded chambers have been is used as a standard for cross section measurements below designed that also provide angular distribution data on the 1.8 MeV.Above this energy the presence of resonance structure reaction products so corrections can more accurately be made limits the usefulness of the cross section as a standard. This for losses in the 10B deposits [GOP00˝ ]. Ionization chambers structure does, however, provide a convenient method for containing 10B deposits on backing material have much the checking the consistency of neutron energy scales at different same response as those containing solid 6Li deposits, as laboratories. The carbon standard has frequently been used was discussed in section 3.3.2. The low Q value for the for the measurement of neutron scattering cross sections 10B(n, α)7Li reaction results in large corrections for particle where direct substitution of samples and maintaining the same loss when foils thick enough to give useful counting rates neutron detector makes the determination relatively easy. High are used. For many of the older measurements not using quality carbon samples are readily available at very low cost. sophisticated gridded chambers, corrections for lost events are The carbon standard has been preferred over the hydrogen a strong function of the angular distribution of the reaction standard by some measurers since the neutron energy loss (in products. Measurements [SEA76, HAM09] of these angular the laboratory system) in scattering from carbon is much less distributions exist. Friesenhahn et al [FRI74] have produced than the loss in scattering from hydrogen. Also it is not possible ◦ thin self-supporting boron films for which there is a high to use hydrogen scattering at laboratory angles beyond 90 . probability that both the reaction products will escape from For neutron energies below 2 MeV the only non-elastic the film so that summing of the alpha and 7Li energies can reaction on carbon is capture, which is negligible above thermal occur. This improves the response function and reduces the energies. Thus for practical purposes the total cross section is corrections for lost events. identical to the elastic cross section. Measurements of the total cross section are important as a means of normalizing 3.4.3. Gamma-ray detectors. Cross section measurements relative angular distribution data in evaluations of the standard. 10 7 Determinations of the carbon angular distribution relative to relative to the B(n, α1γ) Li cross section have been made only with counters which detect the 478 keV gamma ray from hydrogen scattering such as those of [HOL75] have been made this reaction. It is not practical to implement this cross section using polyethylene samples and measuring scattered neutron by detecting the reaction products produced when 7Li is left in energies in order to determine whether the scattering occurred its first excited state. This arises due to the frequent difficulty from carbon or hydrogen. in separating these events from those leaving 7Li in its ground state. However, a number of measurements [DAV61, MAC68], 3.6. Au(n, γ) 10 7 [SEA76]ofthe B(n, α1γ) Li cross section have relied on the detection of the reaction products and the separation of the Gold has excellent properties as a capture standard. Sample ground state and first excited state groups. Both NaI and Ge(Li) preparation is particularly easy since the material is stable, 10 7 monoisotopic and easy to fabricate. The standard has been detectors have been used to implement the B(n, α1γ) Li cross section. The NaI detectors provide higher efficiency applied with two different methods, activation and prompt but a poorer pulse height distribution and somewhat worse gamma-ray detection. timing than Ge(Li) detectors; however, the timing depends on the operating conditions. Neutron activation of the iodine in 3.6.1. Activation. The activation method for gold is well NaI detectors can cause problems particularly when adequate understood. The 198Au nucleus which is formed has a well shielding is not possible. Neutrons that scatter from the 10B known decay scheme. A 2.3 day isomeric state at 812 keV sample into the gamma-ray detector can produce background in 198Au which has been seen in 197Au(d, p)198Au has not due to inelastic scattering in the detector materials. This been observed in 197Au(n, γ)197Au. Thus the cross sections background is most important at high neutron energies where measured by activation and prompt gamma-ray detection are many gamma rays from inelastic scattering may be present. the same. The 198Au half-life of 2.7 days is somewhat long

S338 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications so that long run times are required to produce acceptable on the pulse size only. The use of this detector then requires a activity but this is generally not a large problem. Special care determination of the weighting function and applying it to each must be employed in non-thermal activation measurements to pulse either online or offline by storing the pulse height data. reduce the detection of low-energy background neutrons. The The efficiency of these detectors is considerably larger than that rather large thermal capture cross section is a disadvantage of the Moxon–Rae detector, typically 15% to 20%. Events for in this regard. The use of coverings such as cadmium is which more than one gamma ray per capture are detected can important for reducing this background but their effects on be corrected [MAC67] or handled with a somewhat different the measurements must be well understood. The experimental representation [CZI69]. In principle total energy detectors setup for an activation measurement is relatively simple. A require a zero energy discrimination level for the detector. foil of gold is placed near a monoenergetic source (or one Practical considerations related to low-energy background which has a well known energy distribution) and irradiated events and gamma rays from inelastic scattering require a until a satisfactory activity has been produced. In a capture non-zero discrimination level. Corrections must be made for cross section measurement relative to the gold standard the this bias. Also, where the detector does not subtend a large foil containing the material to be measured is often placed solid angle, changes in the gamma-ray angular distribution near the gold so the irradiations can be done simultaneously. which may occur at high neutron energies must be taken into The effects of neutron scattering from the target materials, consideration. The small size of these detectors and the use support structure, etc must be fully evaluated. The activity is of non-hydrogenous liquid scintillators, such as deuterated then measured with a suitable gamma-ray detector. Typically, benzene C6D6, allow lower scattered neutron backgrounds NaI and Ge(Li) detectors are used. These counters are often than are possible with the large liquid scintillators. They also absolutely calibrated. When absolutely calibrated counters are easier to shield, have lower ambient backgrounds and have are not available, shape measurements or data relative to the somewhat better timing than large liquid scintillators. Initial thermal cross section are frequently obtained. uses of these detectors involved Monte Carlo simulations of the detection process used in calculating the response functions 3.6.2. Prompt gamma-ray detection with large detectors. that are used in calculating the weighting functions. It was The prompt gamma rays emitted following neutron capture shown [SOW88, BRU80] that this procedure was not correct. also provide a convenient means of measuring the capture Hard spectra led to the largest errors. Recent uses of total cross section. The different detector types which are energy detectors use the corrected weighting function. used can be characterized by their response to the various Prompt gamma-ray determinations of the gold cross gamma-ray cascades which follow neutron capture. Large section have also been made by the analysis of measurements liquid scintillators [FRI71, POE75] which almost completely of the energy distribution of the capture gamma rays [JOL79]. surround the sample have been used. Essentially all the capture This is not a very practical way to implement the standard. gamma rays are absorbed in the detector and its efficiency 235 238 is virtually independent of cascade mode. The detectors are 3.7. U(n, f) and U(n, f) large and somewhat expensive. Although the detectors can be The techniques used for measurements and utilization of the made relatively insensitive to neutron scattering by loading the fission standards apply to both the 235U(n, f) and 238U(n, f) scintillator with a neutron absorbing material such as boron, standard cross sections. The 238U(n, f) cross sections standard the uncertainty in this correction is often the largest limitation has the advantage that it has a very low cross section for in the use of these detectors. More recent detectors of this epithermal and thermal neutrons so that backgrounds from type use a large number of BaF2 detectors [HEI01, GUE09]or those neutrons can be minimized. The cross section does, similar scintillators, thus the detector volume is smaller, with however, have a very rapid reduction below about 2 MeV so high efficiency, and most of the gamma rays are detected. it is only considered a standard above 2 MeV. The 235U(n, f) cross section at thermal is important since it is a rather 3.6.3. Total energy-type detectors. Total energy detectors large, accurately known standard. Detectors which have have a response which is essentially independent of gamma-ray been employed to implement these fission cross sections are cascade mode. The earliest type of this detector is the Moxon– basically of passive or active type. For the active technique, Rae detector [MOX63]. It is designed so that its efficiency is either the very energetic fission fragments or fission neutrons closely proportional to the energy of the incident gamma ray. are counted using gaseous ionization chambers, scintillation Then the overall efficiency for detecting a capture event is chambers, solid-state detectors, spark chambers, avalanche only dependent on the total gamma-ray energy produced, i.e. counters and thin film breakdown counters. In principle, the neutron energy plus the binding energy of a neutron in the detection could also be done using the gamma rays produced, compound nucleus being studied. This detector suffers from but generally the detection of gamma rays is not very practical very low efficiency (approximately a few per cent). Another due to the possibility of detecting gamma rays from capture. total energy-type detector which was originally suggested Methods have been devised to separate the fission and capture by Maier–Leibnitz has become a very popular detector for gamma rays based on multiplicity and total gamma-ray energy capture cross section measurements [MAC67, CZI69]. For emitted but the methods are difficult to use and are generally this detector an average response function is generated which not as accurate as the other techniques. is proportional to energy. This is done by weighting each The characterization of the sample (N) in equation (1)is pulse from the detector with a function which is dependent extremely important for fission deposits. One should reduce

Metrologia, 48 (2011) S328–S345 S339 A D Carlson the contaminants in the deposit to a minimum and know the fragment in the deposit, η is the average velocity of the fission amounts of the contaminants accurately so their effect can fragment relative to the recoil velocity and a2 is the coefficient be calculated. In an unfavourable case, a trace amount of a in the angular distribution of the fission fragments in the CMS substance with a large cross section can produce a contribution in the expression ω = 1+a2P2(cos θ). comparable to the cross section which is to be measured. The non-relativistic expression for η is For thin deposits the presence of contaminants is typically   1/2 determined from mass spectrometry and the number of nuclei TAf η = /(An +1) (5) is determined by alpha particle counting. Tf Various techniques have been used to determine N. where T is the neutron kinetic energy, T is the average kinetic For sufficiently heavy targets, weight measurements may be f energy of the fission fragment, A is the average atomic mass satisfactory. It can also be done using alpha particles from f of the fragment and A is the atomic mass of the target nucleus. the nucleus of interest or a spike (contaminant) which has a n Evaluated values of R from Rustichelli [RUS73]have favourable rate. It is important to have very accurate half- frequently been used. This work is somewhat dated. More life data for these determinations. Usually the alpha particle work needs to be done on determining R in various compounds disintegration rate is too high or the configuration of the for the fissioning of a wide range of nuclides. detector is not appropriate to allow these determinations to be To reduce the losses one should use as thin a deposit made with the detector being used to measure the fission cross as is practical. Also note that if equivalent deposits are put section. Low geometry measurements using well understood on both sides of the backing material, the calculation of the systems with appropriate samples can provide assays for a losses for that system is simplified since it does not contain number of fission samples with an uncertainty of less than the momentum term. However, for measurements with high- one per cent. energy neutrons, one should have both the deposits facing Isotope dilution mass spectrometry [JAF77] provides the away from the neutron source. Then the momentum term highest accuracy in mass determination. This is a destructive reduces the net loss correction. The correction above does technique. By determining relative masses, by alpha or fission not take into account any losses associated with the electronic rate measurements, of reference deposits compared with those discriminator setting. Typically such a setting must be imposed destroyed, one can characterize the reference foils accurately. to reduce or eliminate the counts due to noise, alpha particles, With well-defined standard deposits such as these, the mass of etc. However, this does cause some legitimate fission events a foil to be used in an experiment can be deduced by direct to be lost. The correction for those events is typically made by comparison with the reference. Comparisons of deposits used assuming that the contribution is constant in the pulse height at various US laboratories [POE80] have led to improvements distribution for low pulse heights and equal to the value in the in the characterization of those deposits. valley of the pulse height distribution. This correction is based Except for measurements with thick targets such as those on a number of measurements made with various thicknesses used when prompt neutron detection is used, the deposit should of deposits which are well known and extrapolating to zero be very uniform in thickness so that the energy losses of the thickness. The uncertainty in this procedure may be large so fragments as they emerge from the deposits can be minimized. the objective is to reduce the correction to a minimum. For If the deposit surface is rough with many ‘hills and valleys’ it thin deposits this correction can be less than 1%. is very difficult to estimate the energy loss and the number of In equation (1), the efficiency ε includes the detector fragments which will be lost. It is best if these thin deposits solid angle factor. The smaller the solid angle, the larger the are made by evaporation onto very smooth backings which are correction may be for the anisotropy of the angular distribution rotated so as to produce deposits which are quite uniform. of the fission fragments. Until recently there were only a Except for the case where the solid angle of the detector few measurements of these angular distributions and most of is limited, the fractional losses of fission events for which those data, e.g. [SIM68], are rather old. The appearance of both fragments are totally absorbed in the fission deposit and well designed Frisch gridded ionization chambers has allowed backing can be calculated using the treatment of [CAR74]. new data to be obtained for a number of nuclides. For This treatment includes the effect of the recoil motion caused example, investigations by Hambsch and his group [VIV97] by the incident neutron as well as the angular distribution of at IRMM have yielded fission fragment angular distribution the fission fragments. It is assumed that the fission deposit has measurements. Also direct measurements using telescopes to a uniform thickness, t, with a flat surface. Then measure the angular distribution have been made [PRO04].
 t These data can be used directly in equations (3) and (4) for F = + η (1 − a /2) (3) calculating fragment losses. R 2 2 There is also a correction as a result of neutrons which
 t Rη2 scatter from the backing into the fission deposit and cause B = − η + (1 − a2/2) (4) fission events. This multiple scattering effect causes an 2R 2t increase in the fission rate compared with what would be where F is the fraction of the events which are lost when the observed if no backing material were there. In principle there deposit is facing towards the neutron source, B is the fraction is also a similar correction for the fission deposits. However, of the events which are lost when the deposit is facing away for the thin deposits used in fission chambers, such scattering from the neutron source, R is the average range of a fission effects are normally very small.

S340 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications 3.7.1. Passive detectors. Passive detectors cannot provide of dE/dx for 5 MeV alpha particles compared with fission timing information. They include emulsions, and special fragments. The range of the alpha particles is approximately films can be used to observe and count the fission fragment twice the average range of the fission fragments. Also due tracks. Track etch polyester film techniques have been to the large positive charge of a fission fragment, the energy applied to fission cross section measurements largely by the loss at the beginning of its track is at its maximum. This Michigan group [DAV78, MAH83]. Similar work with mica is the opposite of what occurs for the alpha particles where detectors has been done by [KUK74]. For these measurements the energy loss peaks near the end of the track. Thus by a limited solid angle geometry was used so as to avoid reducing the product of gas pressure and the spacing between problems associated with the fission fragments emitted nearly the plates in the ionization chamber, a more favourable ratio parallel to the 235U deposit. Geometrical factors are necessary of signal from the fission fragments compared with the pile- to accurately determine the solid angle and the results up of alpha particles can be obtained. It is not necessary to are dependent on the fission fragment angular distribution. stop the fission fragments in the counting gas. In fact with a Activation measurements are also a passive technique for spacing of half the average range of the fission fragments, a which the measured activity of an appropriate fission product satisfactory fission pulse is formed and the alpha particles only after a period of exposure is related directly to the number of produce a lightly ionizing track. Typical ion chambers have fissions that occurred. The use of activation requires detailed plate spacings of 0.5 cm to 1 cm with gas pressures of 1 atm to information about level schemes and transition probabilities. 2 atm of a mixture of 90% argon and 10% CO2. The smaller spacings give faster timing. Also, the use of CO2 reduces the 3.7.2. Active detectors, gas ionization chambers. Gaseous electron collection time significantly compared with the use ionization chambers (fission chamber) are the most commonly of pure argon. These detectors are frequently used in a back- used fission fragment detectors. There have been a large to-back or stacked design so that one or more measurements number of different designs for fission chambers. Gaseous of fission cross sections relative to the fission standard can be compounds of 235U have not been successful as counting made simultaneously. The amount of material in the beam even gases. Detectors are designed to detect the fragments from for the stacked designs can be made small so that corrections deposits on backing plates in the chamber. The most common for neutron transmission and scattering are minimal. of these chambers detect fragments going into a 2π solid angle. The conditions are quite different compared with 3.7.3. Active detectors, scintillation chambers. The 235U the use of ionization chambers discussed in the previous scintillation counters, particularly gas scintillation counters, sections since the energies of the fragments are much greater have been used in cross section measurements. Gaseous so corrections for losses in the deposits can be much smaller. scintillators and fission chambers are both limited in efficiency There are a number of experiments where the solid angle for due to the thin deposits which must be used. The scintillators the fragments is 4π. The counting gas, plate spacing and do offer better timing than is possible with fission chambers voltage used in these detectors should be carefully considered. but have poorer pulse height resolution and larger corrections A publication by Wender et al [WEN93] describes the design for lost fragments. Generally, heavy noble gases such as argon, of a fission chamber for fission cross section measurements. together with nitrogen, are used in these detectors since larger The pulse height distribution observed from a fission chamber pulse heights are then produced. In one experiment [KAR78], depends primarily on the deposit thickness and the geometrical very thin targets were used so that coincidences between the conditions under which the fission fragments are collected. fission fragments could be made, thus removing alpha particle For deposits which are very thin compared with the fission induced events. In those measurements, several detectors were fragment ranges, the familiar double humped fission fragment used simultaneously so relative measurements could be made. energy distribution is observed if the plate spacing and the The need for a number of photomultiplier tubes increases pressure of the gas are great enough to stop the fragments. the complexity of doing work with these detectors. These Under these circumstances, the energies of the light and heavy detectors are not used in linac experiments. This may be a fragments are about 100 MeV and 70 MeV, respectively. As result of the increased sensitivity of the heavy gases to the the deposit thickness increases, the average energy of these intense gamma flash. fragments is reduced and the energy distribution worsens. Solid scintillators containing 235U have been made Thicknesses of the order of 100 µgcm−2 are used. These [CAT77]. Such detectors need further development but offer thicknesses lead to very low efficiencies, typically a small ultimate advantages in timing, efficiency and a reduction in the fraction of a per cent. Alpha particles, with energies of about correction for lost fission events. 5 MeV, are present. They can cause a background that may need to be reduced. The fact that the energy is so much 3.7.4. Active detectors, solid-state detectors. Solid-state greater for the fission fragments compared with the alpha detectors have been used in conjunction with thin 235U deposits particles means that the alpha particles can often be eliminated in two different modes. In one mode, the deposit is located at or electronically using an energy discrimination level higher than near the surface of the detector [BAR76]. The performance of the energy of the alpha particles. For those cases where the the detector is similar to that of a fission chamber though it has alpha activity is very large, so that pile-up becomes a problem, somewhat better timing and pulse height resolution. Since this a number of methods can be used to reduce their effect. One type of detector is used directly in a neutron beam, corrections method takes advantage of the different ranges and values for scattering and reactions in the solid-state detector material

Metrologia, 48 (2011) S328–S345 S341 A D Carlson must be made. Also, neutron damage to the detector itself 3.7.7. Active detectors, TPC detectors. A TPC detector for occurs. This damage is in addition to that which occurs from fission work, similar to the one discussed in section 3.1.4, fission fragments and the ever present alpha particles. Limited has been constructed and is now going through diagnostic active areas of solid-state detectors restrict the count rates tests. These tests are checking the hardware and software possible with these detectors. The other mode of operation necessary for the operation of the detector. The objective is for solid-state detectors uses limited solid angle counting of measurements of fission cross section to an uncertainty of less fission fragments from 235U deposits [DER73]. In this mode than 1%. It should also be possible to obtain fission fragment the detector is out of the beam so that radiation damage and masses, energies and their angular distributions. Used in the corrections for neutron reactions and scattering can be conjunction with neutron and/or gamma-ray detectors it should minimized. This mode suffers from the need to know fission be possible to do important correlation studies. fragment angular distributions and solid angles accurately. The count rates are low due to small detector areas (small solid 4. Summary and future activities angle) and the need for thin 235U deposits. Over the years improved measurements and evaluations have 3.7.5. Active detectors, avalanche detectors, spark chambers been made for the neutron cross section standards. The and thin film breakdown counters. These detectors are useful present results from an extensive international cooperative fission fragment detectors since they are insensitive to high effort produced a set of self-consistent cross sections. The rates due to alpha particles, gamma rays and electrons which main detectors that have been used to implement those standard are emitted by fission deposits. Avalanche detectors [KAZ79] cross sections were discussed. are somewhat awkward to use, but they are very fast and As a follow-on to that evaluation effort, an IAEA have nearly 100% efficiency. Spark chambers are difficult Data Development Project, ‘Maintenance of the Neutron to construct and keep operational. They also do not have Cross Section Standards’, was recently initiated to provide high efficiency. Thin film breakdown counters [EIS95] are a mechanism for allowing new experimental data and inherently very simple and compact detectors that do not improvements in evaluation procedure to be used in new involve gas so the problems with maintaining pressure are evaluations of the neutron standards. This project will help removed. The amount of material in the neutron beam is to ensure that we are prepared for the next evaluations of the small so beam distortion and neutron scattering are small. The neutron cross section standards. Within this project, there detectors are fast and use relatively low voltages. However, is an effort to add additional cross sections relative to which the efficiencies are strongly voltage dependent and deviate cross sections can be measured. For example, measurers who significantly from 100%. detect individual gamma rays, e.g. gamma rays from inelastic scattering, would prefer to have a standard or reference cross section having individual gammas. A ‘reference’ cross section 3.7.6. Active detectors, prompt fission neutron detection. is being investigated for this application. Prompt fission neutron detection provides an additional measurement technique for fission cross section determination [GAY74]. Fission neutron detection is generally done with References large high efficiency neutron detectors that view the fission sample and are outside the neutron beam. If these detectors [AAM66] Aamodt R L et al 1966 High pressure 3He gas are placed near the sample, high count rates can be obtained. scintillation neutron detector Rev. Sci. Instrum. Relatively thick samples can be used when detecting the 37 1338–40 neutrons which accompany fission due to the low level of [ARN92] Arndt R and Workman R L 1992 The H(n, n)H cross section from 20 to 350 MeV Nuclear Energy Agency attenuation the neutrons experience. However, with thicker Report NEANDC-311 ‘U’ (INDC(SEC)-101) samples, the thin sample approximation given in equation (1) [AUD07] Audouin L et al 2007 Neutron-induced fission cross is not accurate. More appropriate expressions should be used sections measurements at n TOF Proc. Int. Conf. on which take into account multiple scattering and self-shielding Nuclear Data for Science and Technology (Nice, France) pp 421–4 effects. The response of the detectors is proportional to the 233 ¯ [AXT86] Axton E J 1986 Evaluation of thermal constants of U, product of the average number of neutrons per fission, ν, and 235U, 239Pu and 241Pu and the fission neutron yield of the fission cross section. The energy dependence of ν¯ must 252Cf Report GE/PH/01/86 Commission of the be taken into account. For high incident neutron energies, the European Communities, Joint Research Centre, backgrounds increase due to the scattered neutrons producing CBNM, Geel, Belgium pulses higher than the discriminator bias level. Gamma-ray [BAR54] Barschall H H and Powell J L 1954 Energy distribution of nuclear reaction products Phys Rev. 96 713 discrimination must be used to ensure that the response is [BAR66] Barry J F 1966 Cross section for the reaction 6Li(n, α)T due only to neutrons. It is more difficult to make such a Proc. Conf. on Neutron Cross Section Technology Book measurement absolute since solid angles and neutron detector 2 AEC-CONF-660303 (Washington, DC) pp 763–5 efficiencies must be accurately determined. The method has [BAR76] Barton D M et al 1976 Measurement of the uranium-235 fission cross section over the neutron energy range 1 to been used mostly at low neutron energies where the energy 6 MeV Nucl. Sci. Eng. 60 369–82 dependences referred to above are extremely small. This [BEH88] Behrens J W et al 1988 Development of a 3He/Xe gas method has not been used recently. scintillation counter to measure the 3He(n, p)T cross

S342 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications section in the intermediate energy range Proc. Int. Conf. [DRA72] Drake M K 1972 ENDF/B-III cross section measurement on Nuclear Data for Science and Technology (Mito, standards Brookhaven National Laboratory Report Japan) pp 371–4 BNL-17188 (ENDF-179) [BOU02] Boukharouba N et al 2002 Measurement of the n–p [EIS95] Eismont V P et al 1995 Thin film breakdown counters and elastic scattering angular distribution at En = 10 MeV their applications Radiat. Meas. 25 151–6 Phys. Rev. C 65 014004 [ENG81] Engdahl J C et al 1981 Absolute measurement of the [BRU80] Brusegan A et al 1980 Neutron capture cross section 6Li(n, α)3H cross section at 23 keV Nucl. Sci. Eng. 78 measurements of 56Fe Proc. Int. Conf. on Nuclear Cross 44–52 Sections for Technology (Knoxville, TN) NBS SP-594 [FRA00] Franz J et al 2000 The PSI np data and their effect on the pp 163–7 πNN coupling constant Phys. Scr. T87 14–21 [CAL07] Calviani M et al 2007 Measurement of neutron induced [FRI71] Friesenhahn S J et al 1971 A facility for the measurement fission of 235U, 233U and 245Cm with the FIC detector at of neutron-capture cross sections from 0.01 eV to the CERN n TOF facility Proc. Int. Conf. on Nuclear 1 MeV Nucl. Instrum. Methods 93 477–92 Data for Science and Technology (Nice, France) [FRI74] Friesenhahn S J et al 1974 The (n, α) cross sections of 6Li pp 335–8 and 10B between 1 and 1500 keV Report [CAR70] Carlson A D 1970 Neutron-capture-cross-section AEC-INTEL-RT 7011-001 standards Proc. NEANDC Symposium on Neutron [FU90] Fu C Y 1990 Evaluated cross sections for neutron Standards and Flux Normalization (Argonne, IL) scattering from natural carbon below 2 MeV, including CONF-701002 pp 285–312 R matrix fits to 13C resonances Nucl. Sci. Eng. 106 [CAR74] Carlson G W 1974 The effect of fragment anisotropy on 494–500 fission chamber efficiency Nucl. Instrum. Methods [GAB59] Gabbard F et al 1959 Study of neutron reactions 119 97–110 Li6(n, α)H3,F19(n, γ)F20, and I127(n, γ)I128 Phys. Rev. [CAR82] Carlson A D and Bhat M R 1982 ENDF/B-V Cross 114 201–8 section measurement standards Brookhaven National [GAY74] Gayther D B et al 1974 Measurement of the 235U cross Laboratory Report BNL-NCS-51619 section in the energy range 1 keV to 1 MeV Proc. Panel [CAR93] Carlson A D et al 1993 The ENDF/B-VI neutron cross on Neutron Standard Reference Data (Vienna: IAEA) section measurement standards National Institute of IAEA-PL-246-2/26 201-211 Standards and Technology Report NISTIR-5177 [GOP00]˝ Gopfert˝ A et al 2000 A twin ionization chamber setup as (ENDF-351) detector for light charged particles with energies around [CAR09] Carlson A D et al 2009 International evaluation of 1 MeV applied to the 10B(n, α)7Li reaction Nucl. neutron cross sections Nucl. Data Sheets 110 3215–324 Instrum. Methods Phys. Res. A 441 438–51 [CAT77] Catalano E and Czirr J B 1977 Development of a new [GUE09] Guerrero C et al 2009 The n TOF total absorption scintillator containing uranium Nucl. Instrum. Methods calorimeter response to γ -cascades following neutron 143 61–2 capture in minor actinides Proc. Capture Gamma-Ray [CIE77] Cierjacks S 1977 237Np and 238U as possible standards for Spectroscopy and Related Topics (Kolkata) (Melville, the MeV region Proc. Int. Specialists Symp. on Neutron NY: AIP) pp 372–5 Standards and Applications (Gaithersburg, MD) [HAL87] Hale G M 1987 private communication, Los Alamos (National Bureau of Standards) NBS SP vol 493 National Laboratory, Los Alamos, NM 10 7 pp 278–89 [HAM09] Hambsch F -J and Ruskov I 2009 The B(n, α0) Li and 10 7 [CON92] Conde´ H 1992 Nuclear data standards for nuclear B(n, α1) Li alph-particle angular distributions for measurements Nuclear Energy Agency Report En spallation source Nucl. [DER73] Deruytter A J et al 1973 The accurate fission cross Instrum. Methods Phys. Res. A 459 229–46 section of 235U from 0.002 to 0.15 eV and its reference [HOL75] Holt R J et al 1975 Observation and analysis of elastic value at 2200 m s−1 J. Nucl. Energy 27 645–76 neutron scattering from 12C Proc. Conf. on Nuclear [DEV08] Devlin M et al 2008 Differential cross section Cross Sections and Technology (Washington, DC) measurements for the 6Li(n, t)alpha reaction in the few SP-425 vol I pp 246–9 MeV region Los Alamos National Laboratory Report [HUE80] Huerster W et al 1980 Neutron–proton charge exchange LA-UR-08-06048 scattering between 190 MeV and 590 MeV Phys. Lett. [DIA83] Dias M D et al 1983 An absolute neutron flux detector for B 90 367–70 the 1–20 MeV energy region Proc. Int. Conf. on [JAF77] Jaffey A H 1977 On quantitative sample preparation of Nuclear Data for Science and Technology (Antwerp, some heavy elements Proc. Int. Specialists Symp. on Belgium) (Dordrecht: Reidel) pp 875–7 Neutron Standards and Applications (Gaithersburg, [DIA92] Dias M D 1992 Improvements in the response calculation MD) (National Bureau of Standards) NBS SP vol 493 of the dual thin scintillator neutron detector Nucl. pp 206–11 Instrum. Methods A 321 317–21 [JOH60] Johnson C H 1960 Recoil telescope detectors [DOD87] Dodder D C and Hale G M 1987 private communication Fast Neutron Physics Part I ed J B Marion to CSEWG, Los Alamos National Laboratory, Los andJLFowler (New York: Interscience) Alamos, NM pp 247–95

Metrologia, 48 (2011) S328–S345 S343 A D Carlson [JOL79] Joly S et al 1979 Neutron capture cross sections of Standards and Technology Report NISTIR 5177 rhodium, thulium, iridium and gold between 0.5 and (ENDF 351) 3.0 MeV Nucl. Sci. Eng. 70 53–9 [POE70] Poenitz W P 1970 Interpretation and comparison of [JON77] Joneja O P et al 1977 Surface barrier spectrometers for standard cross sections Proc. NEANDC Symp. on calibration of fast neutrons in MeV range Proc. Int. Neutron Standards and Flux Normalization Specialists Symp. on Neutron Standards and (Argonne, IL) CONF-701002 pp 331–40 Applications (National Bureau of Standards) NBS SP [POE72] Poenitz W P 1972 The Black Neutron Detector ANL vol 493 pp 61–6 Report ANL-7915 [KAZ79] Kazerouni M A and Kappeler¨ F 1979 A fast spherical [POE75] Poenitz W P 1975 Measurements of the neutron capture avalanche fission detector with intrinsic cross sections of gold-197 and uranium-238 between 20 α-discrimination Nucl. Instrum. Methods and 3500 keV Nucl. Sci. Eng. 57 300–8 164 439–45 [POE80] Poenitz W P et al 1980 The evaluation of 235U(n, f) above [KAR78] Kari K and Cierjacks S 1978 Measurement of the fast 100 keV for ENDF/B-V and the implications of a fission cross sections of 239Pu and 240Pu Proc. Int. Conf. unified 235U mass scale Proc. Int. Conf. on Nuclear on Neutron Physics and Nuclear Data for Reactors and Cross Sections for Technology (Knoxville, TN) NBS Other Applied Purposes (Harwell, UK) (Paris: OECD SP-594 pp 483–7 Nuclear Energy Agency) pp 905–9 [POE81] Poenitz W P 1981 Evaluation methods for neutron cross [KNI83] Knitter H-H et al 1983 Angular distribution measurements section standards Proc. Conf. on Nuclear Data for the reaction 6Li(n, t)4He Nucl. Sci. Eng. Evaluation Methods and Procedures (Upton, NY), 83 229–41 Brookhaven National Laboratory Report [KUK74] Kuks I M et al 1974 Measurement of fission cross BNL-NCS-51363 vol 1 pp 249–90 sections of 235U by neutrons with energy 2.5 MeV and [PRO04] Prokofiev A et al 2004 A new facility for high-energy help of associated particle method Proc. 1973 Kiev neutron-induced fission studies Proc. Int. Conf. on Conf. (Kiev) vol 4 pp 18–20 Nuclear Data for Science and Technology (Santa Fe, [LAM75] Lamaze G P et al 1975 After pulse suppression for 8850 NM) pp 800–3 and 8854 photomultipliers Proc. Conf. on Nuclear [RAH98] Rahm J et al 1998 np Scattering measurements at Cross Sections and Technology (Washington, DC) 162 MeV and the πNN coupling constant Phys. Rev. C SP-425 vol I pp 106–7 57 1077–96 [LAM77] Lamaze G P 1977 Special problems with 6Li [RUD74] Rudstam G et al 1974 Delayed neutron emission from glasses Proc. Int. Specialists Symp. on Neutron separated fission products Nucl. Instrum. Methods Standards and Applications (Gaithersburg, MD) 120 333–44 (National Bureau of Standards) NBS SP vol 493 [RUS73] Rustichelli F 1973 Evaluation of the ranges of fission pp 37–42 products Proc. Symp. on Applications of Nuclear Data [LEM71] Lemley J R et al 1971 High resolution fission cross in Science and Technology (Paris) IAEA Report section of uranium-235 from 20 eV to 100 keV Nucl. IAEA-SM-170/16 pp 559–73 Sci. Eng. 43 281–5 [SAR06] Sarsour M et al 2006 Measurement of the absolute [LIS91] Lisowski P W et al 1991 Fission cross sections in the differential cross section for np elastic scattering at intermediate energy region Proc. of a Specialists’ 194 MeV Phys. Rev. C 74 044003 Meeting on Neutron Cross Section Standards for the [SCH91] Schneider D M and Hubner B G 1991 Neutron/gamma Energy Region above 20 MeV (Uppsala, Sweden) discrimination in a lithium-6 glass scintillator in an Report NEANDC-305 ‘U’ pp 177–86 MWD tool Nuclear Science Symp. and Medical [MAC67] Macklin R L and Gibbons J G 1967 Capture cross Imaging Conf. (Santa Fe, NM) pp 1113–7 section studies for 30–220 keV neutrons using a new [SEA76] Sealock R M and Overley J C 1976 10B(n, α)7Li, 7Li∗ technique Phys. Rev. 159 1007–12 differential cross section measurements between 0.2 [MAC68] Macklin R L and Gibbons J G 1968 Study of and 1.25 MeV Phys. Rev. C 13 2149–58 10 7 7 ∗ B(n, α) Li, Li for 30

S344 Metrologia, 48 (2011) S328–S345 The neutron cross section standards, evaluations and applications [THO62] Thomas G E 1962 A boron-loaded liquid scintillation spallation source Nucl. Instrum. Methods Phys. neutron detector using a single photomultiplier Nucl. Res. A 336 226–31 Instrum. Methods 17 137–9 [WES77] Weston L W 1977 Instruments for use of 6Li as a [VAN76] Van Staa R et al 1976 A liquid 3He detector for fast standard Proc. Int. Specialists Symp. on Neutron neutrons Nucl. Instrum. Methods 136 241–8 Standards and Applications (Gaithersburg, MD) [VIV97] ViveSF´ et al 1997 Investigation of the 238U(n, f) process (National Bureau of Standards) NBS SP vol 493 below and above the threshold in the fission cross pp 43–6 section Proc. Int. Conf. on Nuclear Data for Science [ZHA06] Zhang G et al 2006 Measurement of differential and and Technology (Trieste, Italy) pp 479–81 angle-integrated cross sections of the 6Li(n, t)4He [WEN93] Wender S A et al 1993 A fission ionization reaction in the MeV neutron energy range Nucl. detector for neutron flux measurements at a Instrum. Methods Phys. Res. A 566 615–21

Metrologia, 48 (2011) S328–S345 S345