A concept for detecting the presence of special nuclear material (235U or 239Pu) concealed in intermodal cargo containers is described. It is based on interrogation with a pulsed beam of 7MeV neutrons that produce fission events and their β-delayed neutron emission or β-delayed high-energy γ radiation between beam pulses provide the detection signature. Fission product β-delayed γ-rays above 3MeV are nearly 10 times more abundant than β-delayed neutrons and are distinct from natural radioactivity and from nearly all of the induced activity in a normal cargo. Detector backgrounds and potential interferences with the fission signature radiation have been identified and quantified.

There is an urgent need to improve the reliability of screening cargo containers for illicit nuclear material that may be hidden there for terrorist purposes. A screening system is described for the detection of fissionable material hidden in maritime cargo containers. The system makes use of a low-intensity neutron beam for producing fission and the detection of the abundant high-energy /spl gamma/ rays emitted in the /spl beta/-decay of short-lived fission products and /spl beta/-delayed neutrons. The abundance of the delayed /spl gamma/ rays is almost an order of magnitude larger than that of the delayed neutrons normally used to detect fission, and they are emitted on about the same time scale as the delayed neutrons, i.e., /spl sim/1 min. The energy and temporal distributions of the delayed /spl gamma/ rays provide a unique signature of fission. Because of their high energy, these delayed /spl gamma/ rays penetrate low-Z cargoes much more readily than the delayed neutrons. Coupled with their higher abundance, the signal from the delayed /spl gamma/ rays escaping from the container is predicted to be as much as six decades more intense than the delayed neutron signal, depending upon the type and thickness of the intervening cargo. The /spl gamma/ rays are detected in a large array of scintillators located along the sides of the container as it is moved through them. Measurements have confirmed the signal strength in somewhat idealized experiments and have also identified one interference when 14.5-MeV neutrons from the D, T reaction are used for the interrogation. The interference can be removed easily by the appropriate choice of the neutron source.

Ion microtomography (IMT) is based on measuring the residual energy of ions as they pass through the sample. However, for some samples, the maximum range of the ions can prevent acquiring a complete projection data set for the density reconstruction, i.e., only a limited set of angles will allow the beam to penetrate the sample. Limited angle data sets from flat planar objects are a commonly occurring class for limited data reconstructions in IMT. Such a target can pose an additional problem in that the ion beam may not ''see'' its edges - the target is not finite in extent to the beam. Two reconstruction methods are examined to treat limited data for IMT. The first uses mathematical models of the sample to produce simulated ray-sums that are incorporated into the reconstruction of experimentally measured data. The second method produces a partial reconstruction of the region of interest. A technique is also devised, using a marker, for determining the projected center of rotation - a necessary parameter for the reconstruction algorithms - which cannot be determined from the limited data by usual methods. Examples are presented for two specific planar samples: one made of an artificial heart valve material and the other a metal matrix composite.

Nuclear muscopy uses focused MeV ion mubeams to non-destructively characterize materials and components with mun scale spatial resolution. Although a number of accelerator-based mubeam methods are available for materials analysis, this paper centers on the techniques of Ion mutomography (IMT) and Particle-Induced X-ray Emission (PIXE). IMT provides quantitative three-dimensional density information with mun-scale spatial resolution and 1% density variation sensitivity. Recently, IMT has become more versatile because greater emphasis has been placed on understanding the effects of reconstruction artifacts, beam spatial broadening, and limited projection data sets. PIXE provides quantitative elemental information with detection sensitivities to 1 μg/g or below in some instances. By scanning the beam, two-dimensional maps of elemental concentration can also be recorded. However, since X-rays are produced along the entire path of the ion beam as it penetrates the sample, these measurements only give depth-averaged information in general. PIXE tomography (PIXET) is the natural extension from conventional PIXE analysis to the full three-dimensional measurement and forms the bridge linking the complementary techniques of PIXE and IMT. This paper presents recent developments and applications of these ion beam techniques in a diverse range of fields including characterizing metal-matrix composites, biological specimens and inertial confinement fusion targets.

The ion microprobes now found in many accelerator laboratories were developed to perform quantitative elemental microanalysis with high sensitivity. The rapid evolution on this instrument of new computer based techniques has led to the development of high resolution quantitative 3D ion microtomography. This technique offers unique opportunities to examine internal structure of microscopic specimens.

We have developed a system of quantitative radiography in order to produce quantitative images displaying homogeneity of parts. The materials that we characterize are synthetic composites and may contain important subtle density variations not discernable by examining a raw film x-radiograph. In order to quantitatively interpret film radiographs, it is necessary to digitize, interpret, and display the images. Our integrated system of quantitative radiography displays accurate, high-resolution pseudo-color images in units of density. We characterize approximately 10,000 parts per year in hundreds of different configurations and compositions with this system.

The density of a cosmic dust particle is an important parameter in calculating the orbital evolution timescale and temperature profile on atmospheric deceleration. Sutton and Flynn inferred the densities of 12 stony micrometeorites from the stratospheric dust collection using synchrotron x-ray fluorescence. This method used irons mass determinations to infer the particle mass and optical microscope measurements to determine volumes. These results coupled with density measurements by others suggest that stony micrometeorite densities fall into two distinct density groups, with mean values of 0.6 and 1.9 g/cm{sup 3}. Such a bimodal distribution has dramatic implications for natural segregation of dust by gravitational resonances and interpretation of degree of heating in terms of proportions of asteroidal and cometary populations. The synchrotron-inferred densities have estimated uncertainties of {plus minus} 25% due largely to the difficulty in determining the volumes of these irregular objects. We are presently exploring the value of Scanning Transmission Ion Microscope (STIM) for directly mapping the density distribution in individual stratospheric particles and other micrometer-sized objects. In this approach, the energy loss experienced by protons as they traverse the specimen is used to produce 3-dimensional microtomographic images of the internal structure of objects. One of the key advantages ofmore » this technique is the ability to focus this charged particle beam to below 100 nm and thereby attain extremely high resolution tomographic images.« less

Data acquisition and analysis rates and data set size seriously affect the practicality of conducting ion microtomography studies that map out material densities of extended objects in three dimensions. For the promise of today's fine spatial resolution capabilities to be met, huge data sets must be acquired, processed, and analyzed efficiently. In this paper, we present the results of an experimental parametric study exploring the limits of minimizing the amount of data required for tomographic reconstruction. We consider data acquisition system resolution and the effect of limited accuracy of stopping powers. We show graphically the specific results of variation of parameters affecting minimum data set size.

Proton Energy Loss (PEL), has proven to be a valuable method of material characterization. Unlike x-ray methods which are sensitive almost exclusively to the high Z elements of a material, PEL provides information on all elements. With the current system we are capable of gathering a 2-D mapping of areal density or a tomographic slice through a material. Multiple slices can be combined to form a full three-dimensional rendered image of the object. We have used the PEL system to characterize materials and to study machining and assembly techniques. 4 refs., 6 figs.

LLNL programs utilize a low density, cellular carbon material. The manufacturing process for this material is described elsewhere. The material is called replica carbon'' because the carbon structure replicates that of a precursor material. For some applications, it is essential that replica carbon have a highly uniform density throughout a bar or finished component. We have for several years utilized quantitative (low-energy) radiography to image finished bars. This is a valuable tool for characterizing the material. It is a good material for revealing local flaws and pockets of high or low density. In order to better understand this possible density gradient and to determine where in the production process it arises, Lewis performed an x-ray gaging study. Lewis presented areal density plots of four carbon bars. The areal densities were measured through the thinnest dimension of the bar. These data show that the bars are low in areal density at the center, higher near the edges, and highest in the corners. These data are known to be free from scatter effects, but leave the question of the three-dimensionality of the density profile. This became an especially important issues when a hypothesis was presented that the bars might have a uniformmore » inner core surrounded by a higher density rind of variable thickness. In order to evaluate the density variations in three dimensions, we performed x-ray computed tomography of one of the bars from the Lewis study, and on a second bar chosen to be representative of the best possible quality as evaluated by other methods including radiography. 7 refs., 4 figs.« less

ABSTRACTTwo NDT techniques were used to characterize low-density, microcellular, carbon foams fabricated from a salt replica process. The two techniques are x-ray computed tomography (CT) and ion microtomography (IMT); data are presented on carbon foams that contain high-density regions. The data show that densities which differ by <10% are easily observable for these low density (<100 mg/cm3) materials. The data reveal that the carbon foams produced by this replica process have small density variations; the density being ∼30% greater at the outer edges than when compared to the interior of the foam. In addition, the density gradient is found to be rather sharp, that is the density drops-off rapidly from the outer edges to a uniform one in the interior of the foam. This edge build-up in carbon density was explained in terms of polymer concentrating on the foam exterior during drying which immediately followed a polymer infusion processing step. Supporting analytical data from other techniques show the foam material to be >99.9 % carbon