Development of the neutron salt-meter RANS- μ for non-destructive inspection of concrete structure at on-site use

Abstract

We have started the development of a transportable neutron salt-meter, we call it RANS-μ, combining a ²⁵²Cf neutron source and a prompt gamma neutron activation analysis. Trials of chloride detection measurement with RANS-μ were performed outdoor using removed bridges damaged by chloride attack at an outdoor yard in Public Works Research Institute (PWRI) and a test bridge in Fukushima Robot Test Field. For the measurement at PWRI, the results obtained by RANS-μ were compared with those of the automatic potentiometric titration by drilled powder, and then consistent results were obtained.

Keywords

non-destructive chloride measurement on-site prompt gamma-ray concrete structure

1. Introduction

In Japan, many concrete structures such as bridges are exposed to chloride attack by chloride ions (Cl⁻) containing sea breeze near coastal areas and anti-icing agents spread in cold and mountainous areas. As the chloride attack progresses, serious accidents like a bridge collapse may occur. It is commonly known that a steel bar corrosion starts when the chloride ion concentration around the steel bar exceeds the marginal chloride ion concentration of 1.2–2.5 kg/m³, which depends on the water cement ratio [4]. Thus, the chloride ion distribution in concrete structure from the surface to the steel bar, we call it as depth profile, is the critical index to diagnose the progress of the deterioration for the maintenance of concrete structures.

Several conventional methods measuring the depth profile [3,5,6] exist. But these methods damage structures because they must take cores or drilled powders from the structures. For that reason, there are several disadvantages such as taking long time, high cost for pre-measurement processing, limited coverage area and numbers, and no traceability for aging monitoring.

To overcome such disadvantages, a nondestructive method without core sampling and drilling plays an important role. To aim the practical use of nondestructive diagnostic method, we started the development for depth profile measurement inside concrete using Prompt Gamma-ray Neutron Activation Analysis (PGNAA) with the RIKEN Accelerator-driven compact Neutron Source, RANS [7], which has been actively operated to develop nondestructive testing technologies [2,8,14] and an in-vehicle pulse neutron source for onsite use. We have demonstrated the sensitivity of the chloride detection using concrete specimens [10] and our proposed methods deducing the depth profile [9,11,12] so far.

From urgent demand for salt measurement, in parallel with the accelerator-based neutron system for nondestructive method, we have started to develop the nondestructive technique by using a ²⁵²Cf neutron source (Cf source) which is a radioisotope (RI) [13]. Although a Cf source has a lower neutron intensity than an accelerator-driven neutron source, we focused on the advantages of the RI system in terms of small size, light weight, and handling flexibility to realize a practical system as early as possible.

In this paper we describe the feasibility study of the proposed system for salt measurement using Cf source, which we call salt-meter RANS-μ, through experiments at indoor and outdoor.

2. Required performance of salt-meter RANS-μ

The inspected points of the depth profile based on the bridge inspection manual for chloride attack in Japan are shown in Fig. 1. It is noted that “the underside of the bridge” is measured in priority to “the upper-side” such as roadway. For accessing the underside of bridges, a nondestructive measurement device is required to be light and small. The device is also required to display the result at the on-site location.

To satisfy these conditions, it is considered that a Cf source as a certified apparatus with indication of an activity of less than 3.75 MBq is suitable, which does not require to regulate a radiation-controlled area. Because the container size of a Cf source is very small and the radiation intensity is low, the volume and the weight for the radiation shield are possibly small and light. Since a neutron intensity around 10⁵ neutrons/s/4π is low, the efficient chloride detection system is a key point and required for realization of practical use. From these points, the design criteria for the system were suggested below.

Fig. 1.

The inspected points of a bridge for chloride ion concentration.

In this study for salt-meter RANS-μ, we aim to inspect areas such as “the exterior girder” of a bridge using a bucket type and/or corridor type bridge inspection vehicle and to evaluate whether measurement of the marginal chloride concentration of 1.2 kg/m³ could be feasible around a steel bar. From the views of radiation safety, easy operation and easy on-site setup without a radiation handling license and setting a radiation-controlled area, the requests to the salt-meter are as follows: (1) use of Cf source as certified apparatus with indication less than 3.75 MBq, (2) overall size within $100 \times 80 {cm}^{2}$ , total weight below 100 kg, and the lift-up function up-to around 180 cm based on the limit of size and weight of the bucket, (3) a detection sensitivity and accuracy of $1.0 \pm 0.2 {kg/m}^{3}$ at 7-cm-depth from concrete surface within 1 hour of measurement, (4) radiation dose lower than 1 μSv/h, which is within the allowed limit of radiation around the handler of a salt-meter and a bucket. Based on these criteria, the development of a salt-meter RANS-μ has begun.

3. Feasibility study for on-site use

In this section, we describe several trials for sensitivity of chloride detection with a simple geometry, for the chloride detection measurement on-site, and mounting test of RANS-μ system with the actual Cf source and relevant devices.

3.1. Current sensitivity for chloride detection with ²⁵²Cf neutron source

Fig. 2.

Experimental set-up with the Cf source and concrete plates adjusting chloride ion concentration. (a) Horizontal cut view of set-up (A). (b) Picture of set-up (B). (c) Definition of each layer.

The measurement for chloride detection sensitivity was performed with the Cf source. The experimental set-up is shown in Fig. 2. Figure 2(a) and (b) are a horizontal cut view and pictures of the experimental set-up, respectively. The measurements were performed by using a Cf source of 2.0 MBq with a Ge detector of relative efficiency 50% and a Cf source shield which consists of polyethylene, graphite, B₄C rubber, lead, and LiF tile. Two concrete plates with controlled Cl content were fabricated by mixing ordinary Portland cement with water in a weight ratio of 2:1 and coarse aggregate with crushed stone of 5–13 mm size, and by adding NaCl to be chloride ion concentrations of 1.8 kg/m³ and 3 kg/m³. The concrete plates of 3 cm thickness with a chloride ion concentration of 1.8 kg/m³ were set independently at 0 cm, 3 cm and 6 cm depth from the surface and defined as 1st-layer, 2nd-layer, and 3rd-layer as shown in Fig. 2(c). The other concrete plates did not contain salt. After a 2-hour measurement, the spectrum shown in Fig. 3 was obtained. The arrows with energies indicate ³⁵Cl-derived γ-rays via the ³⁵Cl(n,γ) reaction. The enlarged spectra around 6111 keV as an example of the ³⁵Cl-derived γ-rays are also shown in Fig. 3.

Fig. 3.

Gamma-ray spectrum obtained for a 2-hour measurement with concrete plates of 1.8 kg/m³ at 3rd-layer. The ³⁵Cl-derived γ-rays are indicated with “³⁵Cl” and its energy. S.E and D.E mean single escape peaks and double escape peaks, respectively. Gamma-rays derived from other nuclides are also displayed.

From the result, when using a Cf source of 3.75 MBq, one can expect that the chloride detection sensitivity is 1 kg/m³ at $7.5 \pm 1.5 cm$ -depth for a measuring time of 1.1 hour because the neutron intensity would be 1.8 times higher. The measuring time of 1 hour would be achieved by implementing an anti-Compton shield system with BGO scintillators and adopting a lead block with a taper expecte to increase the signal with the ³⁵Cl-derived γ-ray.

3.2. Chloride measurement at the outdoor yard

This section shows that the trials of on-site chloride detection measurement by salt-meter with the Cf source. The measurements were performed by using removal bridges damaged by chloride attack at the outdoor yard of the Public Works Research Institute (PWRI). Gamma-ray from the Nou and Araiso bridges were measured as shown in Fig. 4. The γ-ray spectra were obtained for each removal bridge for 30-min and 15-min measurement times. Then, the ³⁵Cl-derived γ-rays were observed as shown in Fig. 4.

Fig. 4.

Gamma-ray spectra obtained by measuring the removed bridge at outdoor. (a) The picture of Nou bridge and gamma-ray spectrum measuring for 30 minutes. (b) The picture of Araiso bridge and gamma-ray spectrum measuring for 15 minutes.

To deduce the chloride ion concentration of the measured area with the salt-meter, the γ-ray spectra using the concrete plates with chloride of 3 kg/m³ were utilized. They were obtained from the 1st-layer to the 3rd one similarly to those of 1.8 kg/m³ shown in Section 3.1. Assuming that chloride ions were distributed uniformly, the count rates of ³⁵Cl-derived γ-ray were simply added from the 1st to the 3rd layer. The count rate of γ-rays observed during the outdoor measurement was then compared with that of the indoor measurement. The relation is expressed by the formula: $\begin{matrix} (1) & D = (R_{out} / R_{in}) \times 3, \end{matrix}$ where D is the chloride ion concentration of the measured bridge, $R_{out}$ is the count rate of ³⁵Cl-derived γ-ray from the outdoor experiment shown in Fig. 4, and $R_{in}$ is the count rate of ³⁵Cl-derived γ-ray experiment obtained from the 1st to the 3rd layer indoor measurement shown in Fig. 2(c).

Using this formula and the count rates of ³⁵Cl-derived 1951 keV in this time, the chloride ion concentration were estimated to be $3.1 \pm 1.1 {kg/m}^{3}$ for the Nou bridge and $5.7 \pm 1.5 {kg/m}^{3}$ for the Araiso bridge as summarized in the Table 1.

To compare with the conventional automatic potentiometric titration method, the drilled powder near the RANS-μ measurement area at 3 positions and 3 depths were collected as shown in Fig. 5. For this test, the average of 9 points was compared assuming that chloride ion concentration was uniform. We obtained 3.27 kg/m³ for the Nou bridge and 5.72 kg/m³ for the Araiso bridge as shown in Table 1.s

Table 1

Comparison with the deduced chloride ion concentration at each count rate in unit of count per second (cps) of ³⁵Cl-derived 1951 keV γ-ray at each measurement and the average obtained from drilled powder

	Count rate of 1951 keV (cps)	Chloride ion concentration (kg/m³)

		Deduced From indoor measurement	Average obtained from drilled powder
Indoor measurement	0.035 (3)	–	–
Nou bridge	0.036 (13)	3.1 (1.1)	3.27
Araiso bridge	0.067 (17)	5.7 (1.5)	5.72

The results of the RANS-μ measurement are consistent with drilled analyses. This is the world’s first success for non-destructive chloride detection on-site measurement using PGNAA with a Cf source for real bridges damaged by chloride attack, although it is a removed bridge sample.

Fig. 5.

Conventional method of chloride detection. (a) and (b) show the positions where drilled powders were collecting respectively at the Nou and Araiso bridges. (c) The process of potentiometric titration method.

As next step for chloride measurements in near future, a method to increase the accuracy and deduce the depth profile by utilizing several ³⁵Cl-derived γ-rays such as 517, 788, 1165, (1951,) and 6111 keV will be developed and be verified by measuring more bridge samples. In addition, to achieve the objective accuracy of 20% (30% for the presented results in Table 1), more efficient shield structure and geometry around the Cf source and Ge detector to increase the ³⁵Cl-derived γ-ray detection yield and anti-Compton shield system to improve S/N ratio in γ-ray spectrum will be installed.

3.3. Chloride measurement on the bucket of the bridge inspection vehicle

This section shows RANS-μ at on-site measurement using the bucket type bridge inspection vehicle at the test bridge in Fukushima Robot Test Field [1]. In this measurement, RANS-μ with the concept verification model was used. RANS-μ was set on the bucket and moved to exterior girder as shown in Figs 6(a) and (b). Because the test bridge is not damaged by chloride attack, the same concrete plates with chloride ion concentration of 3 kg/m³ used at indoor experiment as shown in Section 3.1 were attached on exterior girder as shown in Fig. 6(c). Then, chlorine was detected as shown in Fig. 6(d). This is the world’s first success for non-destructive chloride measurement with an inspection vehicle at on-site bridge by PGNAA with the Cf source.

Fig. 6.

Pictures and γ-ray spectrum of the trial at Fukushima RTF. (a) RANS-μ was assembled in a bucket. (b) The bucket was moved to exterior girder. (c) The chloride measurement using the concrete plates with 3 kg/m³. (d) Obtained γ-ray spectrum.

Here, the count rate of 1951 keV was converted into chloride ion concentration. Despite using concrete plates of 3 kg/m³, it became a somewhat higher value of 4.3 kg/m³. It is considered that there are several geometrical differences between the measurements at Fukushima and at indoor. The differences may come from the gap and tilt between RANS-μ and concrete, and the lead block with a taper, as shown in Fig. 7. The effects of these differences have been estimated by simulation and will be confirmed experimentally. Then, these effects will be taken into account to perform more efficient chloride detection.

Fig. 7.

Differences between the measurement conditions (a) at Fukushima RTF and (b) indoor.

4. Summary and future

The salt-meter RANS-μ combining PGNAA and the Cf source has been developed for Cl depth profile measurement. ³⁵Cl-derived γ-rays were observed by RANS-μ with a Cf source of 2.0 MBq under the condition that the concrete plates of 1.8 kg/m³ were set at 6 cm from the surface. On-site chloride measurements with the Cf source and PGNAA were performed at the outdoor yard of PWRI by using removed bridges damaged by chloride attack, and the chloride detection succeeded. The chloride ion concentrations deduced from the count rate of ³⁵Cl-derived γ-ray are consistent with the conventional method using drilled powder and potentiometric titration. A chloride measurement utilizing the concrete plates with chloride ion concentration of 3.0 kg/m³ at the exterior girder of the test bridge in Fukushima Robot Test Field was also performed with RANS-μ on a bucket with success.

In the near future, to realize on-site chloride measurement of real concrete bridges with RANS-μ, we will install a γ-ray measurement system with optimized shield and anti-Compton shield system, and establish the method for quantitative evaluation of chloride depth profile.

Footnotes

Acknowledgements

A part of this research was carried out as “Research and development of salt-meter with neutron source for on-site non-destructive inspection of bridge structure”, the commissioned research of “Tohoku Regional Development Bureau” under technology research and development system of “The Committee on Advanced Road Technology” established by MLIT, Japan.

References

https://www.fipo.or.jp/robot/facility/infrastrucrure.

Ikeda

et al., Nondestructive measurement method to detect water/void inside slabs using compact neutron source by backscattered neutrons, Journal of Advanced Concrete Technology15 (2017), 603–609. doi:10.3151/jact.15.603.

JIS A 1154, Methods of test for chloride ion content in hardened concrete, (in Japanese).

JSCE, Standard Specifications for Concrete Structures-2012 “Design”, 2012 (in Japanese).

JSCE-G 574, Area analysis method of elements distribution in concrete by using EPMA, Concrete Journal62(1) (2006), 246–259, (in Japanese).

Kanada

et al., On-site elemental analysis of concrete by portal energy dispersive X-ray fluorescence analyzer, Concrete Journal44(6) (2006), 16–23, (in Japanese). doi:10.3151/coj1975.44.6_16.

Otake

et al., Research and development of a non-destructive inspection technique with a compact neutron source, Journal of Disaster Research12(3) (2017), 585–592. doi:10.20965/jdr.2017.p0585.

Taketani

et al., Visualization of water in corroded region of painted steels at a compact neutron source, ISIJ International57(1) (2017), 155–161. doi:10.2355/isijinternational.ISIJINT-2016-448.

Wakabayashi

et al., Proceedings of the concrete structure scenarios, JSMS18 (2018), 635–640, (in Japanese).

10.

Wakabayashi

et al., Feasibility study of nondestructive diagnostic method for chlorine in concrete by compact neutron source and PGA, Journal of Advanced Concrete Technology17 (2019), 571–578. doi:10.3151/jact.17.571.

11.

Wakabayashi

et al., EPJ Web of Conferences231 (2020), 05007. doi:10.1051/epjconf/202023105007.

12.

Wakabayashi

et al., Advances in construction materials, in: Proceedings of ConMat20, 2020, pp. 1882–1892.

13.

Wakabayashi

et al., Journal of Neutron Research23(2–3) (2021), 207–213. doi:10.3233/JNR-210015.

14.

Yoshimura

et al., Neutron transmission imaging of water penetration of fly ash concrete exposed in marine and inland environments, Asian Institute of Low Carbon Design (2019), 233–236.

Development of the neutron salt-meter RANS- μ for non-destructive inspection of concrete structure at on-site use

Abstract

Keywords

1. Introduction

2. Required performance of salt-meter RANS-μ

3.1. Current sensitivity for chloride detection with 252Cf neutron source

Footnotes

Acknowledgements

References

3.1. Current sensitivity for chloride detection with ²⁵²Cf neutron source