Available online at www.sciencedirect.com
NIM B Beam Interactions with Materials & Atoms
Nuclear Instruments and Methods in Physics Research B 266 (2008) 3397–3405 www.elsevier.com/locate/nimb
Monte Carlo simulations of a portable prompt gamma system for nondestructive determination of chloride in reinforced concrete Ali Bellou Mohamed a, Mohamad Al-Sheikhly a,*, Richard Livingston a, Habeeb Saleh b a
Graduate Program in Nuclear Engineering, A. James Clark School of Engineering, University of Maryland, College Park, MD 20742, USA b Massey Cancer Center, Virginia Commonwealth University, Richmond, VA 23298-0037, USA Received 28 December 2007; received in revised form 17 April 2008 Available online 17 May 2008
Abstract In this work, the MCNP code was used to perform Monte Carlo simulations of the operation of a portable prompt gamma neutron activation (PGNA) system for chloride detection in reinforced concrete. The system consists of a moderated 252Cf neutron source, a high purity germanium (HPGe) gamma ray detector and a portable multichannel analyzer. The system maximum weight is 23 kg with a largest dimension of 31 cm. The simulations utilized a hybrid approach, which consisted of using MCNP simulations to model neutron transport and ray tracing for gamma ray transport, which considerably reduces computation time in comparison to a fully coupled neutron/ photon Monte Carlo simulations. The simulations have shown that the current moderator design effectively thermalizes the neutron energy spectrum. At low to moderate chloride concentrations, the hybrid simulation model of the PGNA chloride detector shows very good agreement with experimental data. The MCNP computations predicted that for a standard error of 10% in counting statistics, the detection of a 2000 ppm chloride concentration (the corrosion threshold) in reinforced concrete can be achieved in a seven minute counting period. This represents a significant improvement over the current standard destructive method of measuring chlorides in concrete. Over the range of water to cement (w/c) ratios normally found in concrete mixes (0.38–0.55), the chloride signal strength shows very little variation especially at the lower chloride concentrations. Thus for all practical purposes the chloride signal remains insensitive to the w/c ratio. Similarly, the chloride signal strength does not vary significantly if limestone coarse or fine aggregate is used in place of quartz. Ó 2008 Published by Elsevier B.V. PACS: 24.10.Lx; 25.20.Lj; 25.40.Lw Keywords: Monte Carlo; MCNP; PGAA; PGNA; Chloride detection; Reinforced concrete; Aggregates; Corrosion
1. Introduction The objective of this work is to use Monte Carlo simulations to evaluate the effect of the variations in water/cement ratio and aggregate type on the performance of a portable prompt gamma neutron activation (PGNA) system for the nondestructive determination of chloride concentration in reinforced Portland cement concrete (PCC) structures [1]. The need for such an instrument arises because the deterioration of the highway infrastructure due to corrosion of the reinforcing steel which is promoted by chloride ions [2]. *
Corresponding author. Tel.: +1 301 405 5214; fax: +1 301 314 2029. E-mail address:
[email protected] (M. Al-Sheikhly).
0168-583X/$ - see front matter Ó 2008 Published by Elsevier B.V. doi:10.1016/j.nimb.2008.04.021
There are several sources of the chloride ions including road deicing salts using sodium or potassium chlorides, concrete set accelerators and seawater either in the form of concrete mix water or as airborne droplets from ocean spray. The current standard method for chloride measurement in concrete is a destructive procedure based on wet chemistry [3,4]. A rotary hammer is used to remove a core sample, which is pulverized and the soluble chloride content is extracted with nitric acid solution. The solution is then analyzed for chloride ion concentration using potentiometric titration with a silver nitrate solution. Alternative methods analyze the solution with a chloride ion selective electrode or by atomic absorption spectroscopy [5]. More recently, a field method using an ion selective electrode has been proposed [6]. The analysis
3398
A.B. Mohamed et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 3397–3405
can take about 10–15 min. Although conventional methods are accurate when used to determine the chloride ions contents, they have severe limitations: they are destructive so that it is not possible to repeat a measurement at the same point later in time; moreover they cannot be used to sample exhaustively the entire structure under examination. For these reasons, an alternative nondestructive method to determine the chloride content in PCC is needed. Prompt gamma ray neutron activation (PGNA) analysis utilizes prompt gamma rays [7,8] emitted by nuclei in the target immediately after the (n, c) capture reaction with thermal neutrons. The reaction leaves the target nucleus in an excited state several MeV above the ground state. The compound nucleus then de-excites within 1014 s by emission of several gamma rays. The reaction 35Cl (n, c) 36 Cl is especially useful because of its large neutron capture cross-section, 33.2 barns [7,8] and also because it yields multiple gamma rays covering a wide range of energies[8]. In particular the 6110 keV gamma with a yield of 20% plays an important role in detecting chlorine in concrete. Among all elements in the prompt gamma ray database [9], only Ni and Mg produce an interfering gamma ray within a 5 keV window at 6105 keV. However, the yields of this gamma-ray photon from Ni and Mg as a result of neutron capture are 100 and 10,000 times lower than the yield from chlorine capture, respectively. Also, the high energy of the 6110 keV chlorine capture photon results in a mean free path in concrete of tens of centimeters, which makes it possible to probe large volumes of material [10]. Since PGNA is an elemental analysis technique, strictly speaking it detects chlorine. However, the concrete literature always uses the term ‘‘chlorides”. To be consistent with that literature, ‘‘chloride” will be used here in place of chlorine. In practice, this is not a problem since in concrete chlorine occurs only in the form of the chloride ion. A schematic diagram of the PGNA system is presented in Fig. 1. It consists of a neutron source in a polyethylene
moderator, approximately 0.95 g/cm3, a high purity Ge gamma-ray detector and a portable multichannel analyzer attached to a laptop computer. The neutron source is the radioisotope 252Cf. Previous work on the PGNA chloride detector used a NaI detector and because of its low resolution it was necessary to use a very intense neutron source, 400 lg of 252Cf, which in turn required massive shielding so that the entire system had to be truck-mounted [11]. Subsequently, the introduction of HPGe detectors has made it possible to reduce the source to 10 lg of 252Cf, leading to system weight of 23 kg, which is relatively portable [1]. Concrete is a complex material composed of a mixture of water, Portland cement and stone aggregates and the proportions can vary with the specific engineering application [12]. In particular, the water/cement (w/c) ratio is a key parameter. Concrete strength decreases with increasing w/c ratio, but low w/c ratios reduce the flowability of the fresh concrete and may not provide enough water to hydrate completely the Portland cement content. Consequently, concrete mixes are normally made with 0.38 < w/c < 0.55. Since the hydrogen content of the concrete can thus vary, the possibility exists that the degree of neutron thermalization in the concrete can also vary significantly enough to affect the chloride signal. It should be noted that the w/c ratio is essentially spatially uniform for a given batch of concrete. It is the batch-to-batch variation of the w/c that is of concern here. Another significant parameter is the type of aggregate used. Concrete is actually about 80% rock in the form of coarse and fine aggregates. Consequently, variations in their chemical composition can affect both the neutron and the gamma ray transport. Rock used as an aggregate comes in two major types: carbonate and siliceous. Carbonate rocks such as limestone or marble are primarily calcium carbonate, although dolomitic limestones can contain significant amounts of magnesium. Siliceous aggregates are based on silicate, but their compositions can vary widely depending
Fig. 1. Schematic diagram of the portable PGNA chloride detection system illustrating coupled neutron and gamma-ray transport. The system consists of a neutron source in a polyethylene moderator (approximately 0.95 g/cm3), a high purity Ge (HPGe) gamma-ray detector and a portable multichannel analyzer attached to a laptop computer.
A.B. Mohamed et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 3397–3405
on the geological processes that formed them. The compositions of the aggregates range from pure SiO2 for quartz to Si, Fe, Al, K, Na, Ca, etc. for granites and volcanic rocks. The approach used in this study is to investigate the influence of these two parameters on the detected chloride signal through Monte Carlo numerical simulations. The PGNA process involves coupled neutron and gamma-ray transport. However, to reduce the computational burden, only the neutron thermalization in the moderator and the flux distribution in the concrete were simulated by MCNP. The gamma-ray transport through the concrete to the detector has been modeled simply by ray tracing methods. A similar hybrid approach has been used to model the PGNA spectrum acquired by a NaI detector in a coal analyzer system [13]. 2. Operating principles of PGNA 2.1. Neutron production and moderation The neutron source is typically 252Cf in a stainless steel or zirconium capsule. An AmBe (Americium–Beryllium) neutron source has also been used for chloride detection in concrete [10], but 252Cf is preferred because of its lower gamma-ray background. The radioisotope source generates fast neutrons with typical energies in the 1–5 MeV range (Fig. 2), while PGNA requires thermal neutrons. Therefore, it is necessary to use a moderator to slow down the neutrons. Moderator design for field applications is not yet an exact science. Ideally, the moderator should maximize the delivery of
3399
thermal neutrons to the target while minimizing losses due to capture; while at the same time the radiation exposure to the operators should not exceed the allowable dose rate limit of 2.5 mrem/h set by the International Committee on Radiation Protection [14]. The moderator design currently used consists of a solid cylinder of polyethylene 15 cm in radius and 15 cm high with a hemispherical end cap 15 cm in radius (Fig. 3). For personnel shielding, the moderator is enclosed in a 1 cm thick layer of light lead, which is composed of 70% lead, 20% polyethylene and 10% boron enriched to 95%. 2.2. Gamma ray production by radiative capture in the concrete Gamma ray production interaction of neutrons and nuclei at a point ~ r in the concrete is given by: cð~ rÞ ¼ nð~ rÞry/ð~ rÞ; ð1Þ where r is the effective thermal capture cross section for the isotope of interest, nð~ rÞ is the number density, /ð~ rÞ is the total thermal flux at that point and y is the yield of the specific photon. The effective thermal capture cross section is defined as the average cross section over the neutron spectrum and can be expressed as [9]: R1 nðvÞrc ðvÞv dv ¼ 0 R1 r ; nðvÞ dv 0 where v is the neutron speed, n(v) dv is the number density of neutrons with speed between v and v + dv and rc(v) is the neutron speed-dependent capture cross section of the nuclide under consideration. If the neutron absorber obeys thep1/v ffiffiffiffi law, or in other words, the cross section varies as 1= E – which is the case for all concrete constituents – then vr(v) = constant. This leads to the fact that in the thermal energy range (